EPRI
Electric Power
Research Institute
Keywords:
Nitrogen oxides
Combustion control
Denitrification
Flue gas treatment
Fossil fuel boilers
EPRI GS-7447
Volume 1
Project 2154
Proceedings
November 1991
                    Proceedings: 1991  Joint
                    Symposium on Stationary
                    Combustion  NOX Control
                    Volume 1

-------
                              REPORT      SUMMARY
                              Proceedings: 1991 Joint Symposium on Stationary
                              Combustion NOX Control
                              Volumes 1 and 2
                              Proceedings of this 1991 symposium, sixth in a biennial series on NOX
                              control, provide an overview of current NOX control activities. The 66
                              presentations in these two volumes contribute significantly to the
                              development of cost-effective and reliable control systems for fossil-
                              fuel-fired power plants.
INTEREST CATEGORY

Fossil plant air quality
  control

KEYWORDS

Nitrogen oxides
Combustion control
Denitrification
Flue gas treatment
Fossil fuel boilers
OBJECTIVE  To foster an international exchange of information on developments
in NOX control technologies for stationary combustion processes.


APPROACH  EPA and EPRI cosponsored the sixth joint NOX control symposium,
held March 25-28, 1991, in Washington, D.C. Approximately 500 representatives of
electric utilities, equipment vendors, R&D groups, and government agencies heard
66 speakers report on control of NOX emissions from stationary combustion
processes. Reports focused  on developments since the 1989 symposium that per-
tain to electric utility power plants and other stationary combustion sources. They
described progress in combustion technologies, selective catalytic reduction
(SCR), and selective  noncatalytic reduction (SNCR).


KEY POINTS
• R&D in the United States to reduce NOX emissions from conventional pulverized-
coal-fired boilers is oriented  mainly toward retrofit combustion modifications. Low
NOX burners (LNBs) with or without the addition of overfire air (OFA) continue to
be the preferred approach, both economically and technically, for  tangentially fired
and wall-fired units. Reburning remains the only widely discussed option for
cyclone boilers.
• Demonstrations of full-scale retrofit LNB and LNB/OFA systems have increased
considerably in the past two  years. The trend in these demonstrations is toward
increasing staging of air and fuel. With controls, emission levels (short-term mea-
surements) for tangentially fired boilers are commonly 0.30 to 0.50 Ib/MBtu, and
those for wall-fired boilers range from 0.45 to 0.60 Ib/MBtu. Continuously achiev-
able levels would be higher.
• Many presentations suggested that the maximum NOX reduction achievable with-
out significantly affecting boiler operations depends on fuel characteristics, specifi-
cally on reactivity, nitrogen content, and fineness. A number of speakers reported
increases in unburned carbon (UBC) in fly ash when using combustion modifica-
tion techniques to control NOX. The increase depends on the above  properties and
the amount of staging. Except for high-reactivity coals, UBC increases ranged
from 2 to 5%.
• SNCR technologies using NH3 or aqueous  urea are receiving increased  attention
in the United States and  Europe. Full-scale tests indicate that NOX emission reduc-
tions up to 50% are possible with NH3 slip below 5 to 10 ppm. Optimization of
EPRI GS-7447S Vols. 1 and 2
Electric Power Research Institute

-------
reagent mixing at 1700 to 1900°F and accurate temperature measure-
ments are critical in obtaining these results.
• Experience with SCR reported by one utility in Germany indicates no
significant catalyst activity decrease, attainment of design NOX reduction
levels (75 to 80%), and control over NH3 slip, usually to less than 1 ppm.
« Retrofit capital costs for SCR on a conventional coal-fired boiler in the
United States are estimated at approximately $100/kW. Operating costs
are estimated at 5 to 7 mills/kWh and are dominated by catalyst replace-
ment costs.


PROJECT
RP2154
Project Manager: Angelos Kokkinos
Generation and Storage Division

For further information on EPRI research programs, call
EPRI Technical Information Specialists (415) 855-2411.

-------
 DISCLAIMER OF WARRANTIES AND

 LIMITATION OF LIABILITIES

 THIS REPORT WAS PREPARED BY THE ORGANIZATION(S) NAMED BELOW AS AN
 ACCOUNT OF WORK SPONSORED OR COSPONSORED BY THE ELECTRIC POWER
 RESEARCH INSTITUTE, INC. (EPRI). NEITHER EPRI, ANY MEMBER OF EPRI, ANY
 COSPONSOR, THE ORGANIZATION(S) NAMED BELOW, NOR ANY PERSON ACTING
 ON BEHALF OF ANY OF THEM:
 (A) MAKES ANY WARRANTY OR REPRESENTATION WHATSOEVER, EXPRESS OR
 IMPLIED, (I) WITH RESPECT TO THE USE OF ANY INFORMATION, APPARATUS,
 METHOD, PROCESS, OR SIMILAR ITEM DISCLOSED IN THIS REPORT,  INCLUDING
 MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE, OR (II) THAT SUCH
 USE DOES NOT INFRINGE ON OR INTERFERE WITH PRIVATELY OWNED RIGHTS,
 INCLUDING ANY PARTY'S INTELLECTUAL PROPERTY, OR (III) THAT THIS REPORT
 IS SUITABLE TO ANY PARTICULAR USER'S CIRCUMSTANCE; OR
 (B) ASSUMES  RESPONSIBILITY FOR  ANY DAMAGES OR OTHER  LIABILITY
 WHATSOEVER (INCLUDING ANY CONSEQUENTIAL DAMAGES, EVEN IF EPRI OR ANY
 EPRI REPRESENTATIVE HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES)
 RESULTING FROM YOUR SELECTION OR USE OF THIS REPORT OR ANY INFORMATION,
APPARATUS, METHOD, PROCESS, OR SIMILAR ITEM DISCLOSED IN THIS REPORT

ORGANIZATION(S) THAT PREPARED THIS REPORT:
ELECTRIC POWER RESEARCH INSTITUTE
Printed on Recycled Paper
                                                          Proceedings:  1991 Joint
                                                          Symposium on Stationary
                                                          Combustion NOX Control
                                                          Volume 1
                                                          GS-7447, Volume 1
                                                          Proceedings, November 1991
                                                          March 25-28, 1991
                                                          Washington, D.C.
                                                         Symposium Cochairpersons
                                                         A. Kokkinos
                                                         ELECTRIC POWER RESEARCH INSTITUTE

                                                         R. Hall
                                                         U.S. ENVIRONMENTAL PROTECTION AGENCY
Prepared for
U.S. Environmental Protection Agency
Air and Energy Research Laboratory
Combustion Research Branch
Research Triangle Park,  North Carolina 27711

EPA Branch Chief
R. Hall

Electric Power Research Institute
3412 Hillview Avenue
Palo Alto, California 94304

EPRI Project Manager
A. Kokkinos

Air Quality Control  Program
Generation and Storage  Division

-------
Electric Power Research Institute and EPR1 are registered service marks of Electric Power Research Institute, Inc.

Copyright S1 1991 Electric Power Research Institute. Inc All rights reserved

                              ORDERING INFORMATION

Requests for copies of this report should  be directed to Research Reports Center
(RRC), Box  50490, Palo  Alto, CA 94303, (415) 965-4081. There is no charge for reports
requested by EPRI member  utilities and affiliates,  U.S. utility associations,  U.S. government
agencies (federal, state,  and  local), media, and foreign organizations with which EPRI  has
an information exchange agreement  On request,  RRC will  send  a catalog of EPRI reports.

-------
                                        ABSTRACT

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 NOX 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 NOX controls for oil- and gas-
fired boilers.  The symposium proceedings  are published in two 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 Iow-N0x 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 EPRI'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 two 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
•     Session 4A:    Combustion NOX Developments I
•     Session 4B:    Large Scale SCR Applications

-------
Papers from the following sessions are contained in Volume 2:

•     Session 5A:   Post Combustion Developments I
•     Session 5B:   Industrial/Combustion Turbines on NOX Control
•     Session 6A:   Post Combustion Developments II
•     Session 6B:   Combustion NOX Developments II
•     Session 7A:   New Developments 1
•     Session 7B:   New Developments II
•     Session 8:     Oil/Gas Combustion Applications

An appendix listing the symposium attendees is included in both volumes.
                                           VI

-------
                                      CONTENTS


Paper                                                                            Page

  SESSION 1:         BACKGROUND
                      Chair:  I. 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. Pagan  1-15

"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. Offen, 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-NOy Burner," J. Pedersen and
M. Berg                                                                          2-37

"Application of Gas Reburning-Sorbent Injection Technology for Control of
NOX and SO2  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 350MW 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 NOX Combustion Technologies for Coal
Fired Utility Boilers," T. Uemura, S.  Morita, T. Jimbo, K.  Hodozuka, and H. Kuroda          2-109
                                            VII

-------
Paper
   SESSION 3:         LARGE SCALE COAL COMBUSTION II
                      Chair:  D. Eskinazi, 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
   SESSION 4A:        COMBUSTION NOX DEVELOPMENTS I
                      Chair: W. Linak and D. Drehmel, EPA

 "An Advanced Low-NOx Combustion System for Gas and Oil Firing," R. Lisauskas
 and C. Penterson                                                                 4A-1

 "NOX 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. Beer, 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:  E. Cichanowicz, EPRI

 "Understanding the German and Japanese Coal-Fired  SCR Experience," P  Lowe,
 W. Ellison, and M. Perlsweig                                                        4B-1

 "Operating  Experience with Tail-End-and High-Dust DENOX-Technics at the Power Plant
 of Heilbronn," H.  Maier and P Dahl                                                  4B-17
                                            VIII

-------
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 NOX 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 NOX 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 NOX 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
                                            IX

-------
Paper                                                                            Page


"Reburning Applied to Cogeneration 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. Lisauskas, R. Dunnette, K. Nakazato,  P. Duggan, and D. Linz       5B-89


   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. Kalinowski                                             6A-43


   SESSION 68:        COMBUSTION  NOX  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

-------
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 NOX 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. DoleZal                                                         7B-43

"Computer Modeling of N2O Production by Combustion Systems," R. Lyon, J. Cole,
J. Kramlich, and Wm. Lanier                                                        7B-63
   SESSION 8:         OIL/GAS 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. Bisonett 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-N0x Burner System,"
S. Kerho, D. Giovanni, J. Yee, and D. Eskinazi                                          8-85

"Demonstration of Advanced Low-N0x 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
                                             XI

-------
     Session 1




   BACKGROUND











Chair:  I. Torrens, EPRI

-------
NOx EMISSIONS REDUCTION IN THE FORMER
     GERMAN DEMOCRATIC REPUBLIC

         B. Kassebohm
     Stadtwerke Dusseldorf AG
        LuisenstraGe 105
     4000 Dusseldorf 1, Germany

         S. Streng
        Lentjes AG
      Hansa-Allee 305
     4000 Dusseldorf, Germany

-------
                       NOx EMISSIONS REDUCTION IN THE FORMER
                           GERMAN DEMOCRATIC REPUBLIC
ABSTRACT
Looking at a map of the European continent, we can see three areas of high NOx
emission concentration: the industrial regions of western and eastern Germany, and
the industrial area between Poland and Czechoslovakia. Unlike the SO,-, emission,
which, due to the prevalent wind currents in Europe,  is concentrated and settles
on the southern part of Scandin3via, the NOx immission always comes from a nearby
source.

It is remarkable that these three equally-large environmental burdens are to be
found in such completely different political and economic systems. Using the
population figures and gross national product as a basis, we for example, discover
that three times as many people and a three times higher GNP cause the emission in
western Germany. The air pollution in East Europe, therefore, is mainly being
caused by inefficiency and energy wastage.

In order to effectively reduce the emission of pollutants in these countries,
therefore, it is not enough to make use of familiar primary and secondary techno-
logies, but especially efficiency must be increased and energy saved.
                                         1-3

-------
                      NOx EMISSIONS REDUCTION IN THE FORMER
                          GERMAN DEMOCRATIC REPUBLIC
INTRODUCTION
In the former German Democratic Republic, as in all the other communist-governed
countries of the Eastern bloc, the economic development after the Second World War
was completely under state control. In place of the forces of a free market with
the flexible reactions of private initiative, the economic goals were determined in
long-term state plans. In these plans, not only requirements and demand, but also
the prices for raw materials and finished products, were regulated. This awkward
system was lacking any private initiative, not least because it was no advantage
for the individual. As a result, everybody only did what they had been told to do.
Now, after 40 years, we can see the serious damage this has done to the economic
system of the former GDR. Adequate profits, which could have been used to finance
the renovation or improvement of production facilities, or even measures for
environmental protection, were not allowed. The raw materials were mainly limited
to those found in their own country, or from communist neighbours. The prices for
raw materials and products did not cover their costs. The constant lack of products
for everyday life, and the effort for each individual to obtain them, also took
their minds off serious deficiencies such as adequate environmental protection.

ENERGY CONSUMPTION AND POLLUTANT EMISSIONS
Fig.  1 shows the relation between energy consumption and GNP for various countries.
Here  we can see clearly that the former GDR is an energy waster compared to the
Federal Republic of Germany, due, as mentioned previously, to their antiquated
production equipment and methods, as well as low energy prices laid down by the
government. Fig. 2 shows this clearly using primary energy consumption. The
difference becomes particularly noticeable when we consider that the Federal
Republic of Germany has 60 million inhabitants, and the former GDR only 16 million.
The consumption of primary and final energy per capita is accordingly high; whereas
the old fashioned, antiquated production methods lead to lower electricity con-
sumption.
                                         1-4

-------
It is characteristic of the communist systems that they endeavour to be economically
self-sufficient, and this is also true with regard to energy. In the former GDR,
since 1978, this has increasingly led to 85 % of the energy needs being met by
native brown coal. With this fuel 70 % of the electricity was produced, and more
than 65 % of heating needs met. At the same time, the brown coal met the need for
gas and largely also the need for fuel through hydrogenation. Incinerating the brown
coal, which here has a high proportion of water, salts, and sulphur, led to a high
emission of pollutants, as no money was spent on holding them back.

Fig. 3 shows a comparison of specific figures for the emission of the pollutants
NOx, S02 and dust per capita of population, and for NOx in the former GDR the
sectors involved, such as power plants, industry, dwelling heating and transport.
This picture was, and is, not just typical for the former GDR, but rather for the
entire Eastern bloc and especially for the industrial conurbation in the triangle
between the GDR, Poland and the CSFR which continues into Hungary, Romania and
Bulgaria. The damage afflicted upon the vegetation, and buildings, in these areas,
mainly due to SCU is well known, due to the ease with which S02 spreads, is not
limited to its place of origin. Southern Scandinavia is, due to the air currents,
the European collecting tank for transported S02, and as the region is rocky with
only a thin earth covering, with fir monocultures and lakes, it is unable to neu-
tralise these large amounts. As a result, the lakes are dead, and the forests'
rate of growth is reduced. The map of Europe in Fig. 4 shows the extensive dis-
tribution and concentration of SOV, in Scandinavia. With S02 especially, it is fairly
easy to follow up on imports and exports, and the result seen in Fig. 5, corres-
ponding to the seasonal air currents, is a familiar one. This shows an annual
average, whereas extreme conditions, which exist in Europe in winter with prevalent
eastern winds, have at times already led to disastrous smog conditions through im-
ported S02 and dust.

REGARDING THE EMISSION OF THE POLLUTANT NOx
The NOx which is measured recorded in the air of our population centres is, contrary
to the unchanging SO^SO^, not an import, but rather is produced on the spot. Pro-
vided it leaves the lower levels, NO changes to N02, and then through UV-influence
it may change to N20, or be a cause for the reduced formation of ozon, and then
elude identification. Fig. 6 shows a map of the European continent with regard to
NOx emission, and as typical points of interest the three industrial urban areas:
Rhein-Ruhr, southern former GDR and southern Poland, CSFR and Hungary. There  is  no
doubt that NOx has formed here due to the dense population, industrial work with
                                         ^-5

-------
various fossil fuels, and motor vehicle traffic. Fig. 7 shows the sources of heaviest
emission. The region displayed near the Polish border emits approximately the same
amount of NOx as Norway. In the Federal Republic of Germany, we can assume that, in
spite of the measures taken to decrease the emission of NOx in stationary industrial
furnaces, motor vehicles, and household fireplaces, which are already resulting in
decreases of 80 %, this heavy emission is the result of anthropogenic activity
from its roughly 60 million inhabitants. The situation is different regarding the
concentration of emission in the southern part of the former GDR and the CSFR. A
maximum  of  16 million people live in this area, but their energy consumption,
specifically due to unefficient production methods and production plants, as well
as  a certain energy wastage through subsidised prices, is substantially  higher.

A PROGNOSIS FOR REDUCING THE EMISSION OF NOx
After  dissolving the GDR and integrating it into the Federal Republic of Germany,
the region  will as from July 1st, 1992 come under the environmental laws in the FRG.
 The desolate condition  of the eguipment alone though calls  for a time limit for
 conversion. These  limits for various pollutants from stationary coalfired sources
 are:
      •     sulphur dioxide  (S02) from Jan. 1st, 1994     <  200 mg/mj
      •     nitrous oxide  (NOx)   from July 1st, 1996     <  200 mg/m3
      •     dust                  from July 1st, 1996     <  50 mg/m3 for new
                                                         <  80 mg/m3 for
                                                            existing plants.

 The total emission of  NOx  in the GDR before unification was approximately
 700 000 t/a, whereby 400 000 t  were attributed to the stationary sources. The  re-
 maining 300 000  t  came  from a comparatively small amount of train, lorry, bus,  and
 car traffic. The  latter will now align itself quickest to the west European level,
 as the entire motor vehicle production in  the former GDR has come  to a  standstill,
 and a spontaneous  exchange  for  western vehicle types with catalysts and  minimum
 pollution has begun. The substantial emission of hydrocarbons up to now, which
 was a considerable burden  for urban areas, and which was due to the common  motor
 vehicle types with two-stroke engines, will also be  improved by this development.

 As shown in Fig.  8,  short-term  relief from the pollutants from stationary  fossil-
 fired power plants,  though, is  possible, as about 50 % of the oldest plants, some
 of which are up  to 50  years old, can be closed down. This is feasible  due  to the
 present economic  recession, but also through savings due to the higher,  market-
 conforming  prices  for  energy. This would leave the best power plants technically
                                          1-6

-------
and economically-speaking, which could continue to use brown coal. For brown coal
mining it is planned to reduce the amount extracted by about a third from 300
million t per year at present to about 185 million t per year, in order to protect
the environment, but on the other hand not to increase unemployment. The remaining
power plant capacity of around 12 000 MW will then have to be retrofitted with flue-
gas cleaning equipment by the dates mentioned above. This especially concerns
improving the electrostatic dust filter and the installation of desulphurisation
systems. There is a good chance that using brown coal with a low calorific value
and high water content, it will be possible to decrease the NOx emission to below
100 ppm NOx just through so called primary methods in the furnace.

As regards the development of the economy in the former GDR it is valid to expect
to recover, and to reach the niveau in the FRG, very quickly. The associated
increase in energy requirements will be met by construction new power plants in
good time. These plants will be built using the latest concepts with hard coal as
a fuel, possibly with integrated gasification or natural gas, but in any case a
combined gas/steam power plant. Fig. 9 shows what pollutant decreases could be
achieved in the former GDR using power plants with new technology compared to the
existing brown coal plants.

The stationary dwelling heating systems must also be retrofitted to reduce the
amount of pollution through NOx emission, whereby approximately 23 %  of dwelling
heating in the former GDR is already being provided by district heating from
central heating plants. The energy consumption here is also suspected of including
up to 50 % wastage, as neither the subsidized energy prices nor the antiquated
buildings are incentives to save energy. With fair market prices and improved
building insulation, capacity and fuel can be saved, and in addition, the pollution
level decreases. It is hoped that the unemployed capacity of the district heating
plants will then be used to switch new costumers from the individual heating systems
It is expected that the majority of these ecologically harmless fuel natural gas,
as soon as a pipework has been set up. A maximum of ten years has been set for
completing this measure.

All in all, the experts are optimistic enough to say that the pollution problem in
the former GDR will have been brought down to the level in the Federal Republic of
Germany in three or four years for traffic, in about five to six years for power
                                         1-7

-------
plants, combined heat and power plants, and heating plants, and at ten years at the
latest for dwelling heating systems. The technology is available for this; it just
depends on whether the economy manages to recover  in this set time to the niveau
in the Federal Republic of Germany.
REFERENCES
1.  Ministerium  fur  Umweltschutz, Naturschutz,  Energie  und Reaktorsicherheit
    der  DDR,  Berlin.  Umweltbericht der DDR,  Februar  1990

2.  Ministerium  fur  Umweltschutz, Naturschutz,  Energie  und Reaktorsicherheit
    der  DDR,  Berlin.  Fristenplan zur Ubernahme  der GroGfeuerungsanlagenverordnung
    zum  1.  Juli  1992

 3.  Volkskammer  der  DDR,  Berlin. Umweltrahmengesetz  der DDR, Juli 1990
                                         1-8

-------
Energie Consumption

   per Capita
   100 -
                                       Canada


                                            )


50 -
Arg
n _
m
CSFR
Poland -*~s
Greece
entina Q
(J ^\raq
1
Great

P
Spam
Australia
O FRC
Britain „ ^^
U Q
Q(J France
Japan

Norway
: O"

Sweden


            i              50


         ^   Brazil, Chile, Tunisia

          China, India, Peru, Egypt
                       100
                                                     GNP per Capita
Fig 1: Gross  Net  Production arid
        Energy Consumption per Capita  (USA =  100)

Source: BWK 363.7  (1990)
        FRG
GDR
                 Total

Fig 2: Energy Consumption

Source: Handelsblatt (1990)
    K\\
    K"

 FRG GDR

Primary—
 energy
                                                         kWh

                                                       6500
                                              FRG  GDR  FRG GDR
                                               Final-
                                               energy
                                Electrical-
                                  energy
                 Consumption per Capita
                                  1-9

-------
                                             41,9% Power Stations
                                             '/S/S///SSS/////S/S//SS/S//S/S/S.
                                                42,9% Traffic
     Nitrogen-    Sulfuroxide   Dust
       oxide
Fig  3:   Emission per Capita in  1988
Source: Iiistitut fur Umweltschutz Berlin, DDR (1990)
1,7% Dwelling Heating

 \ Nitrogenoxide
               2     4
          Fig  4: Measured  S04 Immission

                     in  Europe ( ^

          Source: NATO/CCMS Studie 1979
                                  1-10

-------
Great Britain
D aiimark
36 Ireland , — -^_ » 3
c-.-. _i \ rt
X~."-- <: T "A* ^P
Netherland'S \ /
13 „ — > r h
^~ V '-v / -C= Poland
) > 2fl
s7 \
Belgium / 7
25 LUX.—/ r_,
t~-- 1 s P GDR
; ^ %r 122
^ (
( '^J"\^
\ Vx .h^j CSFR
^~~7 • •-- 61
France ~~7 \
45 / /r
/ /
! \
( -f — •> i^5^^ 6
^ '~^^^yr"J'~' H> ' Austria

*"" ^ '""*
Switzerland (I Yugoslavia
3 Italy °
36
Great Britain ^ ,
T , 4 Danmtirk
12 Ireland .. . n
V -- ~\ /-.
X V-. ~^'--Jf
Netherland x j- ('
41^ ^-—i^ ^ Poland
"*^> \ ''' : ~*' 32
1 ^7
Belgium j )
20 Lux—- f
-*«, ) y ^- GDR
J 1 53
( ^
/ \
s) V*==^ CSFR

France ~7 J
57 /
/ ^'
1 \-^Au stria

^—-' \_^v r'-''"' \
/ | ^-' \
1 \
Switzerland 1 Yugoslavia
6 Italy 3
7
Fig 5: SO 4 Exchange in  1984 (Figures x lOOGt)
Source: Globus 6358
        =  500 -  700 mg per m

        = 1500 - 2000 mg per rn 2

        = 2000 - 2500 mg per rn ^
     Fig 6: Estimated Annual Deposition
            of NOx in  Europe
     Source: Acid Magazine 1990
                           1-11

-------
IV)
Actual Situation *
- Existing Power Plant Capacity
     (among these Nuclear
     Availability   ~ 86%    i.e.
     Peak load
                                                                      Danmark
                                          23.580  MW
                                           2.200  MW)
                                          19.800  MW
                                          17.900  MW
Hypothesis for the Future:
- Saving  Potential in Power  Plant  Capacity
          ~  50%       i.e. ~ 12.000 MW
Increase  in Efficiency and/or
Emission  Reduction by
    - Renewal (till 2020)
       Cost       ~ 30 Bill DM
                 or
     - legal Retrofitting by
     Flue  Gas Scrubbing   (till 01.01.94)
     Denitrogenisation    (till 01.07.96)
     Dust Precipitation    (till 30.06.96)
       Cost       ~ 10 Bill DM

Fig 8: Power  Generation of the GDK
       in 1988
* Source: IZE
NetherlafidAJ'    Q
          GDR   \
•  -  r    16,5 Mio   \
Hamtmrs v  Inhabitants  \
      .^.08.000 km2(
                                                                                        \ Poland
                                                                                      Berlin x
                                                            w_^  FRG
                                                       France  ;>  62 Mio
                                                               / Inhabitants
                                                               J- 249.000 km
                                                            .  . I 3 Mill t NO x/a    D
                                                            Switzerland
                                                                                  Austria
                                                       Fig 7: Points  of Max NOx-  Emission
                                                       Source: DIW (1985)

-------
  a
  o
 •rH
  03
  03
 
-------
"TOP-DOWN" BACT ANALYSIS AND
RECENT PERMIT DETERMINATIONS
          John R. Cochran
          Morgen E. Pagan
           Black & Veatch

-------
                         ABSTRACT

New  EPA  requirements  for  "top-down"  best available  control
technology (BACT) analyses have resulted in determinations that re-
quire more stringent control technologies. Accordingly, these per-
mit decisions include nitrogen oxide (NOX) emission limits significant-
ly lower than applicable New Source Performance Standards. However,
with careful consideration of acceptable site-specific impacts, obtain-
ing a reasonable BACT determination is still possible.

This paper presents a step-by-step approach for conducting a top-
down BACT analysis, and summarizes important considerations that
will lead to a more effective BACT analysis. In addition, recent per-
mit decisions regarding NOX emission rate and control technology
requirements for combined cycle combustion turbine and coal fueled
power plants  are summarized and examined to ascertain the basis
for decisions. Guidance from this paper will help applicants in prepar-
ing an accurate and comprehensive BACT analysis for their proposed
projects.
                             1-17

-------
                            "TOP-DOWN" BACT ANALYSIS AND
                             RECENT PERMIT DETERMINATIONS

                                           John  R. Cochran
                                           Morgen E. Fagan
                                           Black & Veatch
                INTRODUCTION

 On December 1, 1987, the EPA Assistant Administrator
 for Air and Radiation, J. Craig Potter, issued a memoran-
 dum  implementing a  number  of program initiatives
 aimed at improving the effectiveness of the Clean Air
 Act  new  source review program. Among  these  in-
 itiatives was the implementation of a "top-down" ap-
 proach to  determine  the  best  available  control
 technology (BACT) under the Prevention of Significant
 Deterioration (PSD) program of the Clean  Air Act.
 Primarily,  the top-down approach requires that the most
 stringent  feasible control  technology available,
 designed to achieve the lowest achie/able emission rate
 (LAER) be evaluated  first in a BACT analysis. This
 technology would  then represent BACT unless it could
 be reasonably demonstrated on the basis of site-specific
 energy, environmental, and economic impacts that this
 level of control is not warranted. The next most stringent
 level of control would then be evaluated. This process
 would continue  until a technology could  not  be
 eliminated on the  basis of energy, environmental, and
 economic  considerations,  in  which  case  this
 technology  is BACT for  the project.

 The  EPA has indicated that the intent of the new top-
 down BACT  procedure is not to establish a national
 BACT standard, but to avoid "bottom-up" evaluations
 that  do not consider LAER technologies and result in
 the  use of  control technologies designed for  com-
 pliance with  New Source  Performance Standards
 (NSPS). Accordingly, permit decisions since implemen-
 tation of the guideline have resulted in NOX emission
 limits significantly lower than applicable NSPS. Top-
 down BACT analysis has made  it increasingly difficult
 for  new  sources  to  avoid a requirement  for  post-
 combustion  NOX  control  systems.  However,  with
 careful consideration of site-specific impacts, it is still
possible to obtain a BAG" determination appropriate
for a proposed project.

   NEW SOURCE  PERFORMANCE STANDARDS

Baseline  air  emission  performance  requirements
(emission limits) for a number of new source types are
established by the  United States Government in the
Code of Federal Regulations, Chapter 40, Part 60 (40
CFR 60). The emission requirements dictated by the
NSPS establish the minimum level  of acceptable air
emission control. Table 1 provides  a listing of NSPS
for  coal  fueled  steam generators  and combustion
turbines.

          BACT  PROGRAM OBJECTIVES

The definition of a  BACT requirement is an emission
limitation based on the maximum reduction for a pollu-
tant  regulated by  the Clean  Air Act,  which, on a
case-by-case  basis  taking into account energy,  en-
vironmental,  and economic impacts, is determined
to be achievable  through  application of available
methods.

The  primary  objective of the BACT determination
process is to minimize consumption of PSD air quali-
ty increments, thereby enlarging the potential for future
economic growth without significantly degrading air
quality. To avoid setting national control  technology re-
quirements, BACT guidelines dictate evaluating feasi-
ble control technology alternatives  on a case-by-case
basis while considering site-specific impacts. The in-
tent is that this case-by-case approach will encourage
adoption  of  improvements   in  emission  control
technology more rapidly than would occur through
uniform  control technology  requirements  or  New
Source Performance Standards.
Mr. Cochran and Mr. Fagan may be contacted at (913) 339-2000.
                                                  1-18

-------
                                               Table 1
                        Nitrogen Oxide Emission NSPS for Selected Source Types
  Steam Generating Units Larger than
  250 MBtu/h (Subpart  Da source)
    Bituminous, Anthracite,  and Lignite
    Subbituminous Coal

  Steam Generating Units With Heat Inputs Between
  100 and 250 MBtu/h  (Subpart Db source)
    Spreader Stoker and Fluidized Bed Boilers
    Pulverized Coal
    Lignite

  Stationary Gas Turbines*
    Rated Load
    Peak Load

  "Corrected lo 15 percent  oxygen minus corrections for heat rate and fuel bound nitrogen.
                                                                                       Emission  Limit
                                   0.60 Ib/MBtu
                                   0.50 Ib/MBtu
                                   0.60 Ib/MBtu
                                   0.70 Ib/MBtu
                                   0.60 Ib/MBtu
                                       75  ppm
                                       150  ppm
As previously discussed, NSPS provide the baseline re-
quirement establishing the minimum acceptable level
of control for a BACT determination. The BACT analysis
was required to evaluate alternatives between the NSPS
base line and the most stringent control technology
which  provide the maximum emissions reduction.

In response to guidance documents issued by the EPA
in December 1978,  new  source permit applicants
prepared so-called "bottom-up" 8ACT analyses. These
analyses started with the NSPS base line and then com-
pared  the feasibility  of  more  stringent control
technologies  to the NSPS base line. Because of either
regulatory procedures or inconsistencies in the process,
most of these  bottom-up BACT analyses resulted  in per-
mit determinations at or near NSPS limits. Since the
EPA interprets that the intent of Congress for implemen-
ting the BACT process was to drive technology, emis-
sion limits near NSPS were an unacceptable result of
the process.

With the adoption of a requirement for top-down BACT
analysis in  December  1987,  the   EPA  recognized
that  the  original  BACT analysis guidance  did  not
adequately focus  the  BACT  process  and attain  the
objective of adequately addressing the most stringent
level of control (LAER technology). Draft EPA top-down
guidance documents indicate that the burden of proof
for a top-down analysis is to disprove LAER. However,
the  EPA  maintains that the  fundamental  purpose
of the  top-down approach is  to  arrive at consistent
determinations  that  adequately  consider  LAER
technology.

Theoretically, either the top-down or bottom-up BACT
procedure should result in the same permit determina-
tion. The same principles apply in both cases. However,
the real result of top-down guidance is a shifting of the
"burden of proof." In  bottom-up BACT analyses, the
presumption  lies in favor of an NSPS determination.
Therefore, a  more stringent control technology must
be proven to be warranted. In top-down BACT analyses,
the presumption is to install LAER based  technology.
Since EPA's interpretation of the BACT process is to drive
technology, this presumption has led to more stringent
determinations.
                                                 1-19

-------
    TOP-DOWN BACT ANALYSIS PROCEDURE

A BACT analysis must be perrormed for each  new,
modified,  or reconstructed  emissions  source. The
applicability criteria  requiring a BACT analysis vary
among states and EPA regional jurisdiction. In general,
BACT is required for pollutants whose potential emis-
sions exceed significant emission rates established by
the EPA.

The EPA has recommended that a BACT analysis follow
the general requirements of EPA's draft "Top-Down" Best
Available  Control  Technology Guidance Document,
March 15, 1990. The following discussion describes a
step-by-step approach to performing  a BAG" analysis
that meets  EPA requirements.  Figure  1  graphically
depicts this step-by-step  approach.

STEP 1—DETERMINE SOURCE AND EVALUATION
CRITERIA
One of the most important steps in a  BACT analysis is
to  accurately define source  technical and economic
characteristics.  Evaluation criteria typically  used in a
BACT  analysis  are listed  belcw:

     •   Technical Evaluation Criteria.
        —   Type of Combustor.
        —    Fuel Burn Rate.
        —    Fuel Analysis
        —    Emission  Rates  (Controlled  and
               Uncontrolled).
        —    Flue Gas  Flow Rates.
        —   Site-Specific Constraints.
     •   Economic Evaluation Criteria.
        —   Commercial Operation Date
        —   Economic Recovery Period.
        —   Capital Cost Contingency Factor.
        —   Escalation Rate (Capital and  O&M).
        —   Fixed Charge Rate.
        —   Present Worth Discount Rate.
        —    Indirects Cost Factor.
        —    Interest During Construction.
        —    Capacity Factor.
        —    Fuel Cost.
        —    Incremental Capacity  Charge.
        —    Energy Cost.
        —    Additive Cost.
        —    Waste Disposal Cost.
These  technical and  economic criteria should  be
accurately determined before  any  substantial efforts
are made on the BACT analysis. Subsequent evalu-
ation  of alternative control technologies is greatly
dependent on these evaluation criteria.

Technical criteria  are primarily used  to determine
potential  emissions,  air  quality control equipment
effectiveness, and equipment sizes. The two primary
criteria   that have  a  major  impact  on  pollutant
emission rates and equipment type and size are  the
fuel  quality and  the  maximum  anticipated  fuel
burn rate. For a coal fueled  application, a specific
fuel  source or  at  least a  potential   range of  fuel
properties  needs  to  be  determined  early in  the
analysis  process. For any type  of source, a maximum
fuel burn rate should  also be established early in  the
analysis  process. Since this fuel  burn rate directly
affects  the  amounts  of pollutants  emitted  and  the
subsequent  mass emission  limits,  it  is critical  that
this parameter  be established with some margin to
account  for uncertainties  inherent  in conceptual
design.  Since costs  are  closely dependent on  fuel
quality  and fuel  burn rates,  economic portions of
the BACT  analysis  will  have  to be recalculated
whenever these parameters change. This recalculation
could delay the submittal of a PSD permit application.

Economic evaluation criteria are also important to  the
BACT  analysis since  varying certain  criteria  can
significantly affect the  final conclusions. It is important
that the economic criteria be project-specific If project-
specific criteria are not available, typical values can be
used that are representative  of the  current economic
trends. Since economics is not  an exact science, some
variation in the evaluation criteria could  be considered
in the analysis to provide a range of cost impacts (sen-
sitivity  analysis).

Until recently, economic evaluation criteria have  not
received close scrutiny. However, the economic analysis
has become a focal point of BACT analyses. Therefore.
it is very important that an  applicant be capable of
defending economic evaluation  criteria. Like fuel quali-
ty and fuel burn rate, economic criteria should also be
carefully  selected to ensure the  accuracy of the evalua-
tion and  to  prevent delays associated  with changing
criteria.
                                                  1-20

-------
                                           Step 1
                                  Determine Services and
                                     Evaluation Criteria

                                   •  Fuel Data
                                   •  Economic Data
                                   •  Site-Specific Information
            Step 2
Review Recent Permit Decisions

   • BACT/LAER Clearinghouse
      Documents
   • Other Projects
            Step 3
     Identify Alternatives

     • Technically Feasible
     • Technology Transfer
     • Similar Applications
            Step 4
  Economic/Energy Evaluation

  • Rank Alternatives
  • Engineering Economics
  • "Top Down"
  • Incremental Cost
            Step 5
 Environmental Consideration

Other Pollutant Emissions
Waste Disposal
Contamination of Other Waste
  Products
Comparison of Air  Quality Impacts
Public Health and Safety
  Considerations
                                           Step 6
                               Recommend BACT Alternatives

                                 • Technical
                                 • Energy
                                 • Environmental
                                 • Economic
                                          Figure 1
                             Top-Down BACT Analysis Technique
                                           1-21

-------
STEP 2—REVIEW RECENT PERMIT  DECISIONS
The next step m the analysis is to review recent permit
decisions to determine the  LAER control alternative.
The 8ACT/LAER Clearinghouse documents provide a
good reference for this activity. These documents (1985
and 1990 editions) contain a comprehensive listing of
permit decisions and the  associated  control  effec-
tiveness. Clearinghouse  documents can be obtained
from the EPA. These documents are also helpful as an
indicator of what level of control  (emission  limit)
might represent BACT or, at least, provide a range of
control effectivness  that should be considered in the
analysis.

Typically,  EPA's position is that if a permit has been
previously issued requiring a specified emission limit
or technology, then this is sufficient justification to
assume that the control  technology or emission limit
is feasible  or achievable.  However,  it  should  be
remembered that some of  the more stringent  BACT
determinations were arrived  at for  various reasons.
Because of project  schedule  requirements and fiscal
health, a number of applicants may have conceded to
a regulatory agency proposed BACT  determination to
expedite the permitting process. Other applicants may
have accepted the use of a technology or emission rate
to get below significance levels or to meet increment
consumption or ambient air quality standards.  In ad-
 dition, a number of so
-------
Engineering economics are the generally  accepted
method of evaluation. Capital and annual operating (in-
cluding maintenance) costs are presented, as well as
total annual costs (levelized fixed charges on capital
plus levelized annual operating  costs) of the various
alternatives. Costs presented should be comprehensive,
reflecting fully integrated  systems. Operating  costs
should reflect expenditures for maintenance, additive,
energy, demand, waste disposal, and operating person-
nel. In  addition, if a control alternative  negates the
potential for sale of waste products, the cost analysis
should reflect this impact.  For control alternatives that
affect unit reliabilities, cost estimates for replacement
power should also be included. Total  annual costs are
used to determine incremental cost-effectiveness (in-
cremental  total levelized  annual cost divided by  in-
cremental  annual  emissions) of the various control
levels and  technologies being considered. Incremen-
tal  costs, not total removal costs,  are the true indicator
of cost-effectiveness of a particular control alternative
as  compared  to   the  next  less effective  control
alternative.

STEP 5—ENVIRONMENTAL EVALUATION OF
CONTROL ALTERNATIVES
Environmental  impacts of  the  various  alternatives
should  also be included. Environmental impacts that
should be considered for inclusion in the BACT analysis
include the following:

    •    Increased  emission of  other  pollutants
         resulting from use of a  control  alternative.
    •    Handling and storage of hazardous materials.
    •    Hazardous waste disposal of spent catalysts.
    •    Contamination of waste products that could
         be sold for reuse
    •    Comparison of proposed BACT air quality im-
         pacts with impacts resulting from  use of a
         more stringent control technology.

STEP fr-RECOMMEND tt^CT ALTERNATIVE
This step basically summarizes Steps 4 and 5. The most
effective emission  control technology capability not
previously  eliminated   for  technical,   energy,  en-
vironmental,  and  economic reasons  is  then  pro-
posed  as  BACT. Generally,  the 8ACT analysis  and
recommendation  are  documented  in  the  PSD
application.

 IMPORTANT BACT  ANALYSIS CONSIDERATIONS

Several  important  considerations  should  be  incor-
porated into planning an effective  BACT analysis.

The  economic analysis should be based  on a total
levelized annual cost, including capital and operating
costs. Levelized costs  reflect the effect of escalation and
present worth discounting of future annual  expen-
ditures, resulting in an  equivalent of constant dollars
over the evaluation period. Levelized costs more ac-
curately represent financial impacts over the life of the
project than do first year costs only. Therefore,  it is im-
portant to have good  representative economic evalua-
tion  criteria, since these criteria significantly affect the
results of the  analysis.  The  economic evaluation of
alternate technologies  capable of various  degrees of
effectiveness should also be compared on an incremen-
tal basis. Incremental costs accurately reflect the true
economic effectiveness of a  technology.

Various control technologies  require  additive  or
catalyst. Special consideration must be given to any
technology that requires an  additive or catalyst that
might have hazardous or deleterious environmental ef-
fects, for instance, ammonia generally can be used with
relatively little risk. However,  an accidental  spill could
have catastrophic consequences on the safety of per-
sonnel and surrounding communities. For instance, in
densely populated  areas, emissions of unreacted am-
monia  (ammonia  slip)  could be  a significant en-
vironmental disadvantage Accordingly, such considera-
tions should be included in  the environmental  and
economic portions of the BACT analysis.

During the top-down  BACT analysis the selection of a
particular technology or emission level  may result in
an increase in other pollutants. A good example of this
is carbon monoxide  (CO) and volatile organic com-
pounds (VOO which  are inversely related to combus
tion control of NOX emissions. Combustion controls
                                                  1-23

-------
that are effective in lowering NOX emissions, such as
staged combustion and water injection in combustion
turbines, result in increased emissions of CO and VOC.
The environmental importance of these other pollutants
must be  evaluated as compared to NOX  emissions
reductions.

The results from ambient air quality impact modeling
should also be included  in the BACT analysis to deter-
mine if the project will  emit pollutants at a rate that
exceeds PSD significance levels or ambient air quality
limits.  If  the proposed emissions from the facility are
below, ambient air quality modeling significance values,
there would  be no quantifiable benefit from using  a
more effective control technology. Demonstrating that
emissions from the facility will be below ambient air
quality significant impact criteria will provide a good
argument against  the imposition of a LAER technology
in a BACT situation.

 Depending on  the nature of the project  and  the type
of combustion   technology  being   considered,
sometimes  there  is  a  potential  net  environmental
benefit from implementation of a project. For instance
 a cogeneration plant providing steam supply  to an in-
 dustrial user may result in the retirement of process
steam  boilers. These process steam boilers probably
 have higher emission rates than the cogeneration plant
 and are likely to discharge pollutants at relatively low
elevations, resulting in reduced dispersion. Despite the
cogeneration plants use of significantly larger boilers,
 ambient air quality impacts may be reduced as a result
of relatively lower emission rates and increased disper-
 sion. This would be an extremely important site-specific
consideration that should be included in the analysis.

 It is not unusual for the BACT process to exceed a year
 to resolve a contested BACT. Therefore, it is recom-
 mended  that a conservative BACT schedule be assumed
 if a project plans to propose and defend BACT at some
 level less than  a  LAER technology.

         RECENT BACT DETERMINATIONS

To evaluate the effect of the top-down process, it is
beneficial to review recent BACT determinations. The
following discussion  are summaries of BACT analyses
and determinations for NOX emissions reduction at
several coal fueled and combustion turbine combined
cycle  projects.

COMBINED CYCLE COMBUSTION TURBINE
PROIECTS
Over the years, combustion turbine manufacturers have
improved their product by substantially lowering NOX
emissions. However, the requirement to use selective
catalytic reduction (SCR) systems on combined cycle
units  in some cases is mandated by state  and federal
regulatory agencies.

A recent combined cycle project in Florida obtained
a draft permit from the state that did not require an SCR
system so long as the capacity factor for the facility re-
mained below  60 percent. This determination  was
based on excessive control technology costs (as com-
pared to other similar applications) for  use at capacity
factors less than 60 percent. Subsequently, the Florida
governor  and cabinet approved the  draft permit.
However,  under pressure from  the  EPA,  the Florida
Department of  Environmental  Regulation issued the
final permit allowing the use of combustion controls
only if the capacity factor is limited to 25 percent or
less. Alternatively, at higher capacity factors the  per-
mit dictates the  installation and operation  of a SCR
system. This determination was made despite the fact
that the incremental costs of an SCR system  on the plant
with a 25-percent capacity factor limitation are much
higher  than  generally  accepted  incremental  cost
thresholds. Currently, the applicant is contesting this
determination.

This does not appear to be an isolated incident. There
is some indication from other projects that the high cost
of an  SCR  system on a combined cycle plant (as com-
pared to the cost of control alternatives  for other types
of plants) is not a significant factor in regulatory agen-
cy BACT determinations.

COAL FUELED HAWAIIAN COGENERATION PLANT
This project consists of two 90 MW bituminous coal
fueled  circulating  fluidized  bed  (CFB)  boilers
scheduled  or commercial  operation  in  1992.  The
project will sell electrical power to a Hawaiian utility
and process steam to  a local  refinery.  The Hawaiian
                                                   1-24

-------
Department of  Health  worked  closely  with  EPA
Region  IX  during  the  PSD  permitting  of  this
facility.

A  SCR system was  identified as the  most stringent
method of NOX  control, with selective non-catalytic
reduction  (SNCR)  and  combustion  controls  also
evaluated as available control technologies. The BACT
analysis  recommended  that SCR and  SNCR be
eliminated  because  of technical, economic,   en-
vironmental,  and energy considerations.  The NOX
BACT recommended by the applicant for the project
was CFB combustion controls to meet an emission  limit
of 0.36 Ib/MBtu.

EPA Region IX contested this proposed determination
on the basis of reasonable SNCR economics and as not
being representative of BACT, considering the number
of SNCR installations on CFB boilers in California. EPA
Region IX strongly suggested, and the project accepted,
the use of a SNCR system  designed to meet a NOX
emission Iimitationof0.il  Ib/MBtu.

COAL FUELED MICHIGAN  POWER PLANT
This 45 MW project consists of one CFB boiler burn-
ing bituminous coal. The original BACT analysis com-
pared SCR, SNCR, and combustion control options for
NOX  emission control. Based on economic, energy,
and environmental considerations, combustion controls
designed to limit  NOX emissions to 0.35 Ib/MBtu were
recommended as BACT. The Michigan  Department of
Natural Resources (DNR) accepted the proposed BACT
and issued  a draft permit for public comment.

During the public  comment period,  EPA Region V
issued an official protest rejecting the DNR's determina-
tion of no post-combustion controls. The EPA recom-
mended that a SNCR system designed for  maximum
NOX reduction efficiency be required as representative
of BACT. The EPA  referenced numerous California per-
mits requiring SNCR to limit NOX emissions to 0.039
Ib/MBtu.

In response, the applicant prepared a BACT analysis ac-
cepting the use of SNCR, but contesting a requirement
for maximum control efficiency. The revised BACT
countered that the California plants burned extremely
low-sulfur (less than 0.50 percent) bituminous coals not
available in the Midwest. According to information pro-
vided by the SNCR manufacturer, burning low-sulfur
coals limited the  technical  effectiveness of SNCR
systems to approximately 0.12 Ib/M8tu. However, the
revised BACT also indicated that with the chlorine con-
tents of Midwestern coals, use of SNCR to meet a 0.12
Ib/MBtu emission  limit would result in an ammonia
chloride plume (resulting from ammonia slip emis-
sions). To avoid the potential for an ammonia chloride
plume, SNCR effectiveness must be decreased to result
in a NOX emission of 0.16 Ib/MBtu. The revised BACT
analysis compared the relative economics  and  en-
vironmental  effects of these two alternate emission
limits and recommended a 0.16 Ib/MBtu emission limit.
As a result of this analysis, the DNR (with agreement
by the EPA) issued a final permit  at the 0.16 Ib/MBtu
emission limit.

COAL FUELED  POWER PLANT
This project will consist of a bituminous coal fueled
CFB boiler. Once again an SCR system was identified
as the most  stringent  method of NOX control with
SNCR and combustion controls also being evaluated
as available control technologies.  The BACT analysis
recommended that SCR  and SNCR  be  eliminated
because of technical,  economic, environmental, and
energy considerations. The applicant  recommended
that BACT for the  project was combustion controls.

The  state regulatory  agency and  the regional  EPA
disputed the validity of this BACT selection. In response,
the applicant  provided substantial financial data  sup-
porting that an SNCR  determination would result in
cancellation of the project. In addition, the applicant
demonstrated that if the project was implemented as
recommended, other aspects of the project would  lead
to ambient air quality improvements. This information
convinced both  the state agency and the EPA that the
project-proposed  BACT  determination  was  valid
when overall environmental  benefits and  relative
project  economics were  taken into  consideration.
Therefore, BACT for control of NOX emissions from this
project was determined  to be combustion controls.
This  determination  is  a  good  example  of  site-
specific considerations  controlling  a  BACT
determination.
                                                 1-25

-------
                 CONCLUSION

The primary objective of the EPA in issuing guidelines
requiring top-down BACT analysis was to gain  con-
sistency in the process, and to drive plants away  from
NSPS determinations  towards permit  requirements
more representative of  the current state of air quality
control technology. However, the  EPA also maintains
that it is still an objective of the BACT program not to
dictate national BACT standards, but to evaluate the site
specifics of  a given project.

At this time,  it appears that the EPA has been successful
at achieving the objective of permit determinations
more stringent than NSPS. This is especially evident
on recent permit decisions for combustion turbine com-
bined cycle facilities. In this situation, EPA and  state
regulatory agencies are mandating the use of selective
catalytic reduction systems capable of achieving NOX
emissions 80 percent below NSPS. An additional il-
lustration of this success has been the requirement for
SNCR systems  at a number of coal fueled power plants.
Should these trends continue, the BACT process will
essentially accomplish  emission requirements reflec-
tive of a  de facto NSPS.
For an  applicant  to effectively  dispute  regulatory
agency-proposed  BACT  requirements,  the  BACT
analysis must be pertormed in a careful, objective man-
ner. This will first require  an  adequate schedule  to
prepare and defend a non-LAER BACT proposal. In ad-
dition,  the applicant  must carefully research the
background and status of comparison permit deter-
minations. Economic and environmental considerations
must be fully developed to provide adequate arguments
disputing regulatory agency control technology man-
dates. Finally, the applicant must carefully establish and
represent project site specifics.

It is becoming a concern that environmental regulators
are intent on maximizing pollutant reductions from
new plants without regard for site-specific considera-
tions.  Therefore, it appears that the BACT process  is
becoming more closely related to the LAER process.
However, as illustrated by some of the examples, it  is
still possible  to  use  well  developed  site-specific
arguments to convince the regulatory agencies that the
proposed BACT is a reasonable control technology re-
quirement.
                                                 1-26

-------
RETROFIT COSTS AND PERFORMANCE OF NOX CONTROLS AT
          200 U.S. COAL-FIRED POWER PLANTS
                     T. E. Emmel
                      M. Maibodi
                  Radian Corporation
                3200 E. Chapel Hill Road
        Research Triangle Park, North Carolina  27709

-------
            RETROFIT COSTS AND PERFORMANCE OF NOX CONTROLS AT
                        200 U.S. COAL-FIRED POWER PLANTS
ABSTRACT
This paper presents the results of a study conducted under the National Acid Precipitation
Assessment Program by the U.S. Environmental Protection Agency's Air and Energy
Engineering  Research Laboratory. The objective of this research program was to significantly
improve engineering cost estimates currently being used to evaluate the economic effects of
applying sulfur dioxide (SO2) and nitrogen oxide (NOX) controls at 200 large  SO2-emitting
coal-fired utility plants.  To accomplish the objective, procedures were developed and used
that account for site-specific retrofit factors.  The site-specific information was obtained from
aerial photographs, publicly available data bases, and input from  utility companies.  Cost
results are presented for the following control technologies: low NOX combustion or natural
gas reburn, and selective catalytic reduction. Although the cost estimates provide useful  site-
specific cost information on retrofitting NOX controls, the costs are estimated for a specific
time period and do not reflect future changes in boiler and coal characteristics (e.g., capacity
factors and fuel prices) or significant developments in control technologies that would change
the cost and performance estimates.
                                        1-29

-------
            RETROFIT COSTS AND PERFORMANCE OF NOX CONTROLS AT
                            200 U.S. COAL-FIRED PLANTS
INTRODUCTION
The objective of the National Acid Precipitation Assessment Program (NAPAP) study of 200
U.S.  coal-fired power plants was to improve cost estimates being used to evaluate the
economic effect of retrofitting sulfur dioxide (SO2) and nitrogen oxide (NOX) controls at coal-
fired  utility plants.  Although study resources were primarily focused on the retrofit cost of
SO2  controls, cost estimates were developed for low NOX combustion controls [low NOX
burners (LNB), overfire air (OFA), or natural gas reburning (NGR)] and selective catalytic
reduction (SCR) for the boilers at each plant.

Figure 1  shows the phases in which the NAPAP study of 200 plants was conducted.  In
Phase I,  detailed site-specific procedures were developed with input from a technical advisory
committee.  In Phase II, these procedures were used to evaluate retrofit costs at 12 plants
using data collected from site visits (1). Based on the results of this effort,  simplified
procedures were developed to estimate site-specific costs without conducting site visits.  For
LNB and OFA, performance-estimating procedures for NOX  reduction were developed  with
input from a consultant (2).  SCR procedures were tested and revised based on the results of
a parallel program effort  in  which five coal-fired power plants in Germany were evaluated (3).
In Phase III, the simplified procedures were used to estimate NOX control cost and
performance for 188 plants. The results of this effort were sent to each utility company for
review and comment.  In Phase  IV, the review comments from the utility companies and the
NAPAP advisory committee were incorporated into the final 200-plant study report (4).

PERFORMANCE AND COST-ESTIMATING PROCEDURES
Figure 2  presents the cost-estimating methodology used to develop  inputs to the  Integrated
Air Pollution Control System (IAPCS) cost model (5).  For each plant, a boiler profile was
                                        1-30

-------
developed based primarily on public information from Energy Information Administration
Form 767, Powerplants Database (boiler design) (6), and aerial photographs obtained from
state and federal agencies.

Low NOX combustion (LNC) was evaluated for all dry-bottom boilers, with application of LNB
on wall-fired units and OFA on tangential-fired units. NGR was evaluated for wet-bottom
boilers and unconventional firing types because applying LNB was considered infeasible and
OFA would not reduce emission rates sufficiently.  Performance estimates were developed to
account for non-ideal situations that will occur when retrofitting LNB and OFA.  As discussed
below, the NOX reduction estimates are based  on the boiler volumetric heat release rate.  No
cost adjustments were made to  reflect site-specific  situations.  For NGR, a NOX reduction of
60% was assumed for all boilers.

SCR was evaluated for all boilers with two types of SCR systems  considered:  hot-side and
cold-side. Both configurations have wide commercial application  in Japan and Germany.
During the course of this study,  very limited data were available on the long-term
performance of hot-side  systems on coal-fired  applications,  and no commercial or recent pilot
scale data were available for hot-side systems  using U.S. coal. Therefore, cold-side SCR
systems were selected for most boilers.

Cold-side SCR systems are located downstream of particulate and SO2 control systems,
thereby reducing or eliminating the  catalyst poisoning effects of sulfur (SO3), chlorides,
arsenic, and alkali metals, which are found to a higher degree in U.S. coals than in coals
used overseas. The cold-side system configuration also minimizes unit downtime and
replacement power costs, and facilitates combining smaller  units into one system, thereby
allowing economy-of-scale benefits.  A disadvantage of cold-side  systems is that the catalytic
reactor is located after the air heater,  so that the flue gas must be reheated to 650-700°F
However, the capital and energy costs associated with flue  gas reheat are somewhat offset
by lower  catalyst costs.

Hot-side SCR systems require that the catalyst reactor be located in the flue gas path
between the economizer and the air heater to take advantage of the high flue gas
                                         1-31

-------
temperature (650-700°F), but because of the lack of data mentioned above, hot-side systems
were applied only to boilers with hot ESPs or where space constraints prohibited the use of a
cold-side system.

Low NOX Burners and Overfire Air
When applying combustion controls, many boiler and coal parameters affect uncontrolled
NOX emissions and achievable NOX reductions.  However, accurate data for most of these
parameters are not available. Additionally, well documented data on the performance of LNC
retrofits on U.S. boilers are limited. After a review of the detailed procedures used to
estimate NOX reduction performance at 12 plants where site visits were conducted,  boiler
volumetric heat release rate was chosen to estimate NOX reduction performance. For most
boilers, furnace volume information was found in Powerplants Database.  Based on data from
four LNB  retrofits, the following correlation was developed expressing NOX emission reduction
as a function of boiler furnace volume and unit power generation (2):
      NOXEFF =  68.8 * (V/MW)                                                     (A)
      where:     NOXEFF   =   NOX removal efficiency (percent)
                V         =   Furnace volume (1000 ft3)
                MW       =   Boiler rating (megawatts)
 Although this equation can yield NOX reduction values lower than 25% and greater than 55%,
 25 and 55% were used as lower and upper limits in this study.

 If the furnace volume was not known, the following equations relating furnace volume to
 boiler rating were used for boilers constructed before (Equation B) and after (Equation C) the
 1971 New Source Performance Standards (NSPS) (2):
      For boilers constructed before the 1971 NSPS, V = 0.596 *  MW                   (B)
      For boilers constructed after the 1971 NSPS, V = 0.844 * MW                     (C)

 Therefore, substituting either Equation B or C into Equation A for furnace volume gives
 roughly 40% NOX removal efficiency for wall-fired boilers constructed before the NSPS was
 promulgated and roughly 55% NOX removal efficiency for wall-fired boilers constructed after
 the NSPS promulgation.
                                        1-32

-------
Even less data were available on NOX reductions that can be achieved when retrofitting OFA
on tangential-fired boilers.  A review of furnace volume data for tangential-fired boilers
showed that the furnace volumes for pre-NSPS boilers are 40% smaller than the post-NSPS
boilers.  OFA is generally capable of achieving a  15 to 35% NOX reduction (7).  It was
assumed that OFA can reduce uncontrolled NOX  emissions by 35% for tangential boilers that
were in service after 1974 or that had furnace volumes similar to post-NSPS boilers.  For
boilers in service before 1974, a NOX emission reduction of 25% was assumed.  For boilers
firing coals with high slagging tendencies, NOX emission reductions were reduced by 5%
(i.e., 25 - 5  = 20).

Natural Gas  Reburninq
NGR is included in this analysis, although it is not as commercially developed as the other
NOX control technologies.  Including NGR in the study provides a  moderate  NOX control level
(relative to SCR) where LNBs are inapplicable (cyclone furnaces, slagging wall-fired units,
unusual firing types. The NOX reduction performance of NGR would be affected by some of
the same factors discussed previously for LNC.

The Gas Research Institute is hoping to achieve NOX reductions as high as 75% on high-
NOx-emitting boiler types.  However, because of the lack of commercial demonstration
performance data,  a single estimate of 60% NOX  reduction was  used in this  study.  To
achieve 60% NOX reduction, it was assumed that 15% of the boiler heat input would be
injected into the upper furnace as natural gas. Capital costs include the installation of natural
gas and OFA injection ports into the upper furnace, reburn gas supply piping, and controls.

Selective Catalytic Reduction
The major equipment items for an SCR system include the catalyst, ammonia system,
controls, air preheater modifications or flue gas reheater, ductwork, and fan. The catalyst
volume is based on the flue gas flow rate and an 80% NOX reduction. The SCR equipment
cost estimates were developed from EPRI (8) and EPA (9) studies.

The IAPCS cost algorithms are based on new unit installation.  In  order to adjust these costs
for specific retrofit situations, scope adders (additional equipment  costs) and retrofit factors
(difficulty multipliers) were used to adjust the costs. Scope adder costs considered were:

                                        1-33

-------
          duct and building demolition,
          new duct work,
          new roads and replacement and demolished facilities, and
          new air heater (hot-side) or flue gas reheater (cold-side).

The EPRI flue gas desulfurization (FGD) retrofit guidelines (10) were used to develop costs
for the first three items. New roads and replacement of facilities were handled as increases
in general facilities.  New air heater and flue gas reheater costs are based on  a vendor quote
for a 500-MW plant and scaled by  a 0.6 factor (9).

Access/congestion and underground obstruction factors were applied to the catalytic reactor
area.  The EPRI FGD retrofit guideline factors for the SO2 and flue gas handling area were
used. The scope adjustments and retrofit difficulty factor were input to the IAPCS model to
generate the site-specific retrofit cost estimates.

IAPCS COST MODEL RESULTS
The site-specific model inputs developed for each NOX control technology were input to the
IAPCS cost model, along with other boiler and coal characteristics.  The model generated
capital, operating and maintenance, and levelized annual costs of control and emission
reductions. Table 1 summarizes the economic bases used to develop the cost estimates.
Economic assumptions such as inflation rate, cost of money, cost of consumables, and
expected plant life are from the 1986 EPRI Technical Assessment Guide (1J) escalated to
1988  dollars.

For each control technology, cost  per ton of NOX removed (Figures 3 and 5) and annual cost
(Figures 4 and 6) are plotted versus the sum of controlled megawatts.  In each figure, the x-
axis (sum of megawatts) is the cumulative sum of the boiler size sorted in order from the
lowest to the highest cost to control. Also identified on each curve are the 25, 50, and 75
sum of megawatt percent points for the boilers included in the figures.  Each  point on the
curves represents a specific boiler cost result.  The first point represents the boiler that had
                                        1-34

-------
the lowest cost.  The last point represents the boiler that had the highest cost. The curves
turn up sharply because each curve was developed starting with the boiler having the lowest
control cost and  ended with the boiler have the highest control cost. The cost results .do not
represent the average or cumulative cost of control.

Costs developed in this report are based on economic assumptions that may not represent a
particular utility company's economic guidelines. The  cost results are static (not  dynamic)
and represent a single year (1985 base year or another year specified by the individual utility
company) with regard to capacity factor,  coal sulfur, and pollution control characteristics.

Low NO^ Combustion Cost Results
Figures 3 and 4 summarize the unit cost  and the annual cost, respectively, of retrofitting LNB
at 228 boilers, OFA at 214 boilers, and NGR at 81 boilers.  In general, boilers having low unit
costs and annual costs are large, and have high capacity factors and high NOX reduction
efficiencies. Boilers having high unit costs and annual costs are small and have  low capacity
factors and low NOX reduction efficiencies.

LNB were applied to wall-fired dry-bottom boilers. The boiler characteristics  of the low, mid,
and high unit cost are:

                                     Low $/ton         Mid $/ton        High $/ton
 NOX Unit Cost  $/ton                      50             150              1315
 Boiler size  MW                         1000             640               45
 NOX Reduction  %                         53              43               40
 Capacity Factor  %                        83              48               28
 NOX Removed - tons/yr                 15828           4586               183

Of the 228  boilers, 16% were estimated to have high (45 to 55%) NOX reduction  efficiencies;
61% were estimated to have moderate (35 to 45%) NOX reduction efficiencies; and 23% were
estimated to have low (25 to 35%) NOX reduction efficiencies.
                                        1-35

-------
OFA controls were applied to tangential-fired boilers.  The boiler characteristics of the low,
mid, and high unit cost are:

                                     Low $/ton        Mid $/ton         High $/ton
 NOX Unit Cost - $/ton                      27             100            1248
 Boiler size  MW                          865             350              29
 NOX Reduction  %                         35              25              25
 Capacity Factor   %                        79              56              18
 NOX Removed - tons/yr                 6895            1271              39

Of the 214 boilers, 4% were estimated to have high (26 to 35%) NOX reduction efficiencies;
73% were estimated to have moderate (25%)  NOX reduction efficiencies; and 23% were
estimated to have low (15 to 24%) NOX reduction efficiencies.

NGR controls were applied to wet-bottom boilers and boilers having unusual firing types
(e.g., roof-fired). The cost of NGR is much greater than LNB and  OFA because of the fuel
price differential between natural gas and coal.  The cost results presented here are based
on a fuel price  differential of $2 per million Btu and 15% natural gas substitutes.  Reducing
the fuel price to $1 per million Btu reduces the unit cost  by -50%. Not included in the unit
cost  is the benefit of the 15% reduction in SO2 due to the 15% fuel substitution.  If the SO2
reduction were included in the unit costs, the unit cost of NGR would be reduced by 15 to
45 percent.

Selective Catalytic Reduction Cost Results
In this study, cost estimates for SCR were developed for 624 boilers:  577 boilers with cold-
side systems and 47 boilers with hot-side systems.  Figures 5 and 6 summarize the cost
estimates for application of SCR. For cold-side systems, a significant energy penalty occurs
with flue gas reheating, (equivalent to 120°F reheat).  This cost was not included in this study
because the earlier version of the IAPCS model did not estimate this cost.  Reheat costs
estimated by the most  recent version of IAPCS (12) increase the annual cost of control by 20
to 30% for cold-side  systems.
                                        1-36

-------
Costs presented here are for a 3-year catalyst life.  However, clean gas applications of SCR
may have much longer catalyst life.  Annual and unit costs estimated for a 7-year catalyst life
are 10 to 20% less than those with 3-year catalyst life.

For SCR, the boiler characteristics of the low, mid,  and high unit cost are:

                                     Low $/ton        Mid $/ton         High $/ton
 NOX Unit Cost - $/ton                     710            1810            6091
 Boiler size   MW                          217             543              45
 NOX Reduction - %                         80             80              80
 Capacity Factor   %                        94             49              28
 NOX Removed - tons/yr                 10,546            8331              366

CONCLUSION
There is a high degree of uncertainty regarding the cost and performance of retrofitting NOX
controls on U.S. boilers because of the very limited commercial application experience.
Passage of the Clean Air Act Amendments of 1990 require that NOX emissions from wall- and
tangential-fired boilers meet NOX emission limits of  0.5 and 0.45 pounds per million Btu by
applying LNB. These emission limits do not apply  to cell-fired and wet-bottom boilers.  The
results of this study show that the application of LNB and OFA are likely to have a wide
range of effectiveness for retrofit applications.  As result, it can be expected that many units
may have difficulty achieving emission  limits of 0.5  and 0.45 pounds per million Btu through
the application of LNB and OFA.  The study results and study database can provide
guidance in evaluating the effectiveness of retrofitting LNC controls, but because of the
complexities of combustion modification controls the results of this study could vary widely
from the NOX reductions that can be achieved in practice.
                                        1-37

-------
REFERENCES

1     Emmel T  E., S. D. Piccot, and B. A. Laseke.  Ohio/Kentucky/TVA Coal-Fired Utility
     S02 and NOX Control Retrofit Study.  EPA-600/7-88/014 (NTIS PB88-244447/AS), U.S.
     Environmental Protection Agency, Research Triangle Park, North Carolina, 1988.

2.    Smith, L. L,  Energy Technology Consultants, Inc.  Evaluation of Radian/EPA NOX
     Reduction  Estimation Procedures.  Radian Corporation, Research Triangle Park,
     North Carolina, 27709.  February 1988.

3.    Emmel, T.E., et al.  Comparison of West German and U.S. Flue Gas Desulfurization and
     Selective Catalytic Reduction Costs.  EPA-600/7-90-009, U.S. Environmental Protection
     Agency, Research Triangle Park, North Carolina 27711. April 1990.

4.    Emmel,  T. E., and M. Maibodi.  Retrofit Costs for SO2 and NOX Control Options at 200
     Coal-Fired Plants.   EPA-600/7-90-021a, U.S. Environmental Protection Agency,
     Research Triangle Park, North Carolina 27711.  November 1990.

5.    Palmisano, P. J., and B. A. Laseke. User's Manual for the Integrated Air Pollution
     Control  System Design and Cost Estimating Mode, Version II, Volume I. EPA-600/8-
     86/031 a (NTIS PB87-127767), U.S. Environmental Protection Agency, Research
     Triangle Park, North Carolina, 1986.

6.    Elliot, T. C.,  ed. Powerplants Database, Details of the Equipment and Systems  in Utility
     and Industrial Powerplants, 1950-1984. McGraw-Hill, Inc., New York, New York, 1985.

7.    Thompson, R. E.,  and M.  W. McElroy.  Guidelines for Retrofit Low NOX Combustion
     Control. In Proceedings:  1985 Symposium on Stationary Combustion NOX Control,
     Volume 1, EPA-600/9-86-021a (NTIS PB86-225042), July 1986.

8.    Bauer, T. K., and P G. Spendle. Selective Catalytic Reduction for Coal-Fired Power
     Plants:  Feasibility and Economics.  EPRI CS-3603,  Electric Power Research Institute,
     Palo Alto,  California, 1984.

9.    Burke, J. M., and  K. L. Johnson.  Ammonium Sulfate and Bisulfate Formation in Air
     Preheaters.  EPA-600/7-82-025a (NTIS PB82-237025), U.S. Environmental Protection
     Agency, Research Triangle Park, North Carolina, 1982.

10.  Shattuck,  D. M., et al. Retrofit FGD Cost Estimating Guidelines.  EPRI Report CS-3696,
     Electric  Power Research Institute, Palo Alto, California, 1984.

11.  Electric  Power Research Institute.  Technical Assessment Guide (TAG), Volume 1.
     Electricity  Supply-1986.  EPRI Report P-4463-SR, Palo Alto, California,'1986.

12.  Maibodi. M,  et al.   Integrated Air Pollution Control System, Version 4.0, Volume 2:
     Technical  Documentation  Manual.  EPA-600/7-900-022b, U.S. Environmental Protection
     Agency, Research Triangle Park, North Carolina 27711.  December 1990.
                                       1-38

-------
                                      Table 1.

            ECONOMIC BASES USED TO DEVELOP THE COST ESTIMATES



                     Item                                     Value

 Operating labor                                  19.7  $/person-hour

 Natural gas to coal fuel price difference            $2/million Btu
 Electric Power                                   0.05 $/kWh

 Catalyst cost                                    20,290  $/ton

 1988 constant dollar levelization factors

     Operating and maintenance                   1.0
     Capital carrying charges3                     0.105
aBook life - 30 years; Tax life  20 years; Depreciation Method  Straight Line; and Discount
 Rate  6.1%.
                                       1-39

-------
                           Phase I
                  Develop Detailed Procedures
                           Phase II
                Select 12 Plants and Develop Cost
                    & Performance Estimates

                Revise Procedures Based on Utility
                 and Advisory Committee Input

                 Develop Simplified Procedures
Low NOx Burner and Over Fire
Air NOx Reduction Procedure
 Recommendations by ETEC
 Results of Site Visits and
Review of Cost of Selective
 Catalytic Reduction at 5
 German Coal-fired Plants
                           Phase III

              Develop Cost & Performance Estimates
                    For  Boilers at 200 U.S.
                     Coal-fired Power Plants
                           Phase IV

              Incorporate Utility and NAPAP Review
                    Comments into 200 Plant
              	Study Final Report
       FIGURE 1.   200 PLANT STUDY  TECHNICAL APPROACH.
                              1-40

-------
                         Site Specific Information Sources
    Aerial
 Photographs
     SCR
Retrofit Factors
  and Scope
 Adder Costs
Energy Information Administration
         - 767 Form
   Boiler/Coal Characteristics
     Estimated NOx Reduction
                                  Cost Model Inputs
Utility Comments and
 Other Data Sources
                         Integrated Air Pollution Control System
                                 Cost Model Outputs
Boiler/Coal Parameters
                                                                 1
              Capital Costs   O & M Costs   Annualized Costs   Emission Reduction
                  FIGURE 2.   SITE-SPECIFIC COST ESTIMATION METHODOLOGY.

-------
Q
LU

o
2
01
DC
o
o
o
                  NGR

                  LNB

              X  OFA

              1988 CONSTANT DOLLARS
           75% of Boilers

50% of Boilers
                 25% of Boilers
       400
       200
                                20,000
                      40,000
60,000
                                            SUM OF MW
           FIGURE 3.  SUMMARY OF COST  PER TON OF NOx REMOVED  RESULTS FOR LOW

                     NOx COMBUSTION.

-------
GO
           .c
           X
           s^
           I-
           o
           O
                    6
                    5 -
                    0
A  LNB
X  OFA
•  NGR
1988 CONSTANT DOLLARS
                                                         75% of Total MW
                                                              V	
                                        50% of Total MW
                          25% of Total MW
                                                                  40.000

                                                       SUM OF MW
                                                               60.000
                     FIGURE 4.  SUMMARY  OF  ANNUAL COST RESULTS FOR LOW  NOx COMBUSTION

-------
Q
LU

o
5
LU
DC
 X
O
z
c
o
O
O
7,000





6,000





5,000





4,000





3,000





2,000





1,000
             1988 CONSTANT DOLLARS
                25% of Boilers
                        \
                                                                     3 YEAR CATALYST LIFE
               \    i    i    i    i    i    i    i    r   t    i    i    \    i    i    i    i    i    i

                20,000  40,000  60,000  80,000  100,000 120,000 140,000  160,000  180,000  200,000


                                             SUM OF MW
             FIGURE 5. SUMMARY  OF ANNUAL COST  RESULTS FOR SELECTIVE CATALYTIC

                        REDUCTION.

-------
en
j»


^-»
I-

o
o
         Z
         z
                 24
                 20
                 16
                 12
                  8
                  0
                       1988 CONSTANT DOLLARS
                                                                                3 YEAR CATALYST LJFE
    i    i    r

0     20.000
                                                i    r
                                 60.000
                                                          100.000
i    i

 140.000
180.000
                                                     SUM OF MW
                      FIGURE  6.  SUMMARY OF COST PER TON OF  NOx  REMOVED RESULTS FOR

                                 SELECTIVE CATALYTIC REDUCTION.

-------
NITROGEN OXIDES EMISSION REDUCTION PROJECT

      Larry Johnson, Project Manager
    Case Overduin, Supervising Engineer
        Southern California Edison
         2244 Walnut Grove Avenue
        Rosemead, California 91770

-------
                    NITROGEN OXIDES EMISSION REDUCTION PROJECT
ABSTRACT

Utilities in the Southern California South Coast Air Basin are subject to regula-
tions requiring over a 77% reduction in NOx from their oil/gas fired units and over
50% reduction from stationary gas turbines.  This paper describes Edison's efforts
in developing a strategy to meet these new requirements and in parallel pursuing
new technologies which potentially will save Edison and our ratepayers significant
costs while still  meeting the requirements of the South Coast Air Quality Management
District.
                                        1-49

-------
                    NITROGEN OXIDES EMISSION REDUCTION PROGRAM
INTRODUCTION

On  August  4,  1989 the  South  Coast  Air Quality  Management District  in Southern
California adopted Rules 1134 and 1135 which require  significant  NOx reductions on
the  utilities  in  the   air  basin.   In  order  to comply  with the  new requirements
Southern California  Edison  has  assembled  the  Nitrogen  Oxides  Emission  Reduction
Project which has two main objectives.

         Comply with Rules 1134 and 1135

    •    Reduce the costs of complying with the Rules


The analysis and  planning  involved in an  effort  to  meet  these two  objectives is
discussed below as well as the results to date in both cost and performance.


BACKGROUND

Although Rules 1134 and 1135 were adopted in August  1989,  studies  and alternatives
for meeting  various  levels  and timetables  for  NOx  reduction  were under evaluation
over a  year  prior to final  adoption.  Also,  studies  and alternatives  continue to
be evaluated due  to a number of factors as  follows:

         Rule 1135 was revised on December  21, 1990

         Further  revisions are expected May this year

         Results of new technologies will be forthcoming

    •    Results of unit retrofits vary with technology


The basics of Rules 1135 and  1134  are  shown  in Figure 1.   Because  the  majority of
the impact  and  cost  of complying is associated with Rule 1135,  the balance of  this
paper will  deal  with this aspect.
                                        1-50

-------
ANALYSIS OF CONTROL OPTIONS

Control Technologies

Following an evaluation  of applicable NOx  reduction  control  systems,  three mature
technologies were  chosen  for  potential  application  on  29 SCE  steam generators.
These technologies are:

Combustion  Modifications.   Combustion  modifications  involve   the  replacement  or
upgrading of existing  burners with Low NOx Burners which employ various methods of
staged  combustion  to  mitigate  thermal  NOx  formation.   Low  NOx  burners  can  be
combined  with  flue gas  recirculation  to  the  windbox where such  a  system does not
exist already.

Urea  Injection.   This  technology  involves  the  injection  of  Urea  (Nh^CON^)
into  the  furnace  exit  and/or boiler convection  pass.   If the  temperature in these
boiler  locations  is  between 1650 and  1850°F,  the  Urea reacts  with the  NOx  in the
combustion  flue  gases  to  form nitrogen,  water,  and  carbon dioxide.   If the flue
gas   temperature   is   too   high   (>2000°F)   NHj   radicals   formed   from   the
di sassoci ation  of  the  urea  will  oxidize  to  form  additional  NOx.    Should  the
temperature  be  too  cold   (<1500°F)  the   NH^  radials  will   recombined  to  form
ammonia  which  will  "slip"  through unreacted  diminishing the  effectiveness  of the
Urea  system.   Because  of this temperature  sensitivity system  performance  is  very
dependent on boiler type and geometry.

Selective  Catalytic  Reduction.   Selective  Catalytic  Reduction  or  SCR  involves the
injection  of  NH3  in  the flue  gas  to convert  the NOx  to  innocuous   nitrogen  and
water.   The reaction  is  effective  at  gas temperatures between  600  to  750°F  in
the  presence  of a  catalyst.    This  temperature  typically  exists  at   the  boilers
economizer  outlet   and   just  upstream  of  the  air-preheater.   Catalyst  volume
requirements are  such  that  typically a  major  retrofit of  the boiler  backend  is
required to accommodate  the reactor containing the catalyst and associated ducting.


Removal Performance

The removal performance  is dependent on the boiler being  treated  and  in particular
with  urea  injection,  substantial  temperature and flow testing is  needed to predict
NOx  removal  efficiencies   as  a  function  of  load.   For  analysis  and  initial
selection  of  NOx  controls   required  to  meet  the  new NOx  limits the  following
average removal efficiencies have been assumed:

      Technology                                    Removal  Efficiency

      Combustion Modifications
      without Flue Gas Recirculation                       10%

      Combustion Modifications
      with Flue Gas Recirculation                          30%

      Urea Injection                                       35%

      Selective Catalytic Reduction                        90%
                                        1-51

-------
Technology Costs

Capital   costs  for  NOx removal  technologies  vary  in  particular for  SCR systems
which  are  highly  dependent  on  the  amount  of   boiler   retrofit  required  to
accommodate  the  SCR reactor  and ducting.  When  expressed  as  capital requirements
per unit capacity the following averages were estimated for SCE generating units.
    Urea injection
    Combustion Modifications w/FGR
    Combustion Modifications w/o FGR
    Selective Catalytic Reduction
                                                      $3/kW
                                                     $28/kW
                                                     $9.30/kW
                                                  $100 to $120/kW
Generating Units Subject to NOx Reduction

The generation  system subject  to  NOx reduction  regulation  consists  of 28  units
with  a  total  capacity  of  6626 MW.   All  units  are conventional oil  and gas  fired
steam generators and vary in size  from 480 MW supercritical  units built  in  the mid
sixties  to  33.5 MW  drum type  generators  constructed in  the  early  fifties.   The
units  are  located   throughout  the  South  Coast   Air  Basin  at  seven  generating
stations as indicated in Table 1.


NOx Control  Selection Methodology

With  three  basic control  technologies  and  seven  control  technology  combinations
available  for   possible   application  on  30  units,  a  myriad  of  control-unit
combinations  can  be  applied  to  meet  the  new  lower  NOx   emission  limit  of
Rule  1135.  Considerable  savings  can be achieved  by  applying  controls  selectively
rather than across the board.

To find the lowest  cost  control  solution the technique of linear  programming (LP)
was  used which  is   a  mathematical  technique  for  solving  complex  allocation and
planning problems.  LP selected which controls were to be applied to  what  units  to
attain  the  lowest theoretical  Rule compliance cost.   This mathematical  selection,
adjusted for  operational  and   construction  considerations  and  constraints,  formed
the basis for SCE's compliance plan.


SUMMARY OF COMPLIANCE PLAN AND COST

SCE  submitted  a  Compliance  Plan  to fulfill  the  requirements  of  the SCAQMD  to
identify the  type  and  location of  NOx controls planned  to  be  installed to meet or
exceed  the  emission  limitations and compliance schedule.  The  plan  identifies the
installation  of  urea  injection  of  20  units,  the  implementation   of  combustion
                                      retrofit  of  9  units  with  SCR  systems.   An
                                     Figure 2 and  was  designed  to meet  the interim
                                      .25 Ib/MWHR system emission  level required by
 modifications  of  9  units,  and  the
 implementation  schedule  is shown in
 emission  levels  as well the ultimate
 the  end of  1999  as  illustrated  in Figure 3.

 The  compliance cost has been estimated at $673 million  installed  cost.
 Demonstration Technologies

 The  second  major  objective  of the  project  is  to  reduce  the  overall  cost  of
 compliance.   The methodology  used  in arriving  at  a  compliance plan as  described
 above  does  produce  the most  cost   effective  scenario for meeting  the  rule  using
                                        1-52

-------
proven  technologies.   However,  additional  cost  savings may  be realized  by  using
more  advanced  NOx control  technologies  which are  ready  for  demonstration  level
testing.   Many   new   technologies   were  reviewed  and  four  were  selected  for
demonstration on the Edison system.

    •    Advanced Low NOx Burners

    •    High Energy Urea Injection

         Selective Catalytic Reduction Air Preheater

    •    Economizer (In-Duct) SCR


Figures 4-7  show  schematically  the  basics of  each  technology.   The Low NOx Burner
(LNB) was  installed  on 1-row of Alamitos  Generating  Station  Unit  5 (480  MW).   By
installing just  1-row (replacing onewhich  is out of service) the operability of the
unit  is not  affected and  yet the stability and  potential  NOx  reduction capability
of  the  new burner can  be tested.  The high energy urea injection demonstration was
installed on Huntington Beach Unit 2  (215 MW).  The primary  purpose was  to assess
the  ability  to inject  urea into a  narrow cavity  and achieve  a high  level  of NOx
reduction  (50-60%).    The 'basic difference  between  this   type  of urea  injection
and the  previously discussed  system  is   the  use  of  large  blowers/compressors  to
enhanced  the  injection  and  mixing.    The   selective   catalytic   reduction  air
preheater  utilizes  replacement  SCR  baskets in place of  the existing  plate baskets
used  in  Lungstrum type air preheaters.   At  least  two previous  installations  were
tried  in   Germany  on  coal  fired  units.    One  half  or  1-wheel  of  Mandalay  Unit  2
(215  MW)  is  being  modified with  this   system.    The  last  demonstration  is  the
economizer  or  in-duct  SCR  which  will  be  installed  on Redondo Unit   8  (480 MW).
This  system  utilizes advanced catalyst design and basically attempts to install  as
much  catalyst  as  possible  between  the   economizer  and  the air preheater  without
having an appreciable affect on unit performance.

The  installation  of  these  demonstrations  is  projected  to  cost over  $20  million.
However, the potential  savings  assuming   some of these  technologies  are successful
and can be retrofitted on units compatible with a given  technology  is  projected  to
be  between  $100-200  million.   The  reasons  for this  large potential   is  apparent
when  you compare  the  average  capital  costs  for  conventional   SCR  at  $100-$120/kW
versus  the  demonstration costs  shown  on  Figure  8.    By  combining  technologies  as
shown in Figure 9 a reduction sufficient  to  meet the  requirements  of  the  rule can
be met at theoretically much less cost.


Status of Project

Compliance with Rule  1135 is proceeding  per  the compliance  scenario   as  shown  in
Figure 3.  To  date,  this has been accomplished for the most part with  optimization
of  the  existing  units.    Installation  has   been  completed  on two   320  MW  urea
injection  systems  for  a  reduction  of 357,, over  the  load  range.   Eight additional
urea  systems are  in  various stages  of construction  with  installation by  July  of
this year.

Two  of  the  four   demonstrations,  LNB and High  Energy Urea   Injection  have been
completed and  initial  testing  concluded.  The LNB  installation should  provide  in
excess of  15%  reduction  if utilized  on  a complete boiler.   The  high energy urea
system as  tested  in  a  narrow cavity has  not  provided  significant  NOx reduction.
Reduction typically ranged  between  20-25%.  It does  not  appear at this time to be
a viable option  for  narrow cavity injection  where  downstream tubes cool  the flue
gases immediately after the urea injection occurs.


                                        1-53

-------
The installation  of the  SCR  air preheater  system  is  complete  with  startup  and
testing  to  begin   shortly.   The  economizer  SCR  project   construction   has  been
deferred to the fall of this year due to permitting problems.


Conclusions

Each  system,  boiler,  and  the  specific  NOx   reduction  requirements  have  to  be
analyzed  very  carefully  to match  technologies  with  individual  units  and  their
associated  costs  in order  to  achieve  a cost effective program  for  NOx reduction.
For  Edison's case  no  one  technology   is  the  solution  for  cost   effective  NOx
reduction.   Although  a   solution  can  be  found  utilizing   existing  technologies,
variations  in  performance of  these  technologies  as  well  as  potential  advancement
in  NOx  technologies make  it  imperative that  planning  remain flexible within  the
constraints  of time  and  schedule.   The  project  must  integrate  these  elements  as
well  as  operating constraints, permitting requirements and  budgeting constraints.
In essence the project  team needs to be well-rounded within   the  company  as well  as
outside  in  order  to successfully complete  any  major systemwide pollution  reduction
program.
                                       1-54

-------
                     SUMMARY - RULE 1135


Requires  .25  Ib. NOx/MWh by  December 31, 1999

Averaging  time  is calendar day

Systemwide  averaging

Incremental compliance schedule  starting in 1990  through 1999

Edison  to  install and operate  an SCR unit  on  a  480  MW  steam generator by
December  31,  1993


                     SUMMARY - RULE 1134

Requires  15 ppm  NOx by December  31,  1995 for combined cycle units  >60  MW

Requires  certified continuous  in-stack monitoring

Exempts peaking  units operating  less than 200  hours/year



    FIGURE 1. SUMMARY - RULES  1135,1134
                             1-55

-------
                         Table  1

              BOILERS SUBJECT TO RULE 1135
Station

Alami tos
Alami tos
Alami tos
Alami tos
Al ami tos
Alami tos
El Segundo
El Segundo
El Segundo
El Segundo
Etiwanda
Etiwanda
Etiwanda
Eti wanda
Hi ghgrove
Hi ghgrove
Hi ghgrove
Hi ghgrove
Huntington Beach
Huntington Beach
Huntington Beach
Huntington Beach
Redondo
Redondo
Redondo
Redondo
Redondo
San  Bernardino
San  Bernardino
Unit

  1
  2
  3
  4
  5
  6
  1
  2
  3
  4
  1
  2
  3
  4
  1
  2
  3
  4
  1
  2
  3
  4
 1-4
  5
  6
  7
  8
  1
  2
Maximum Rated
Capacity, MN

     175
     175
     320
     320
     480
     480
     175
     175
     335
     335
     132
     132
     320
     320
      32.5
      32.
      44.
      44.
     215
     215
     215
     225
     292
     175
     175
     480
     480
      63
      63
                                   Total
                     6626
                           1-56

-------
Compliance Schedule





Technology/
Activity Descriptions
Rule 1135
Urea Injection (5,600 MW)
Combustion Mods (1,900 MW)
SCR (3,200 MW)
Repower (1,000 MW)
1989
HIS



1990




1991


SSilS
;;Elflf
1992


SfHH
!x\X>vW:X\}»
^XXlvSXl;
1993



^
1994
ilSSSS



1995




1996

11


1997


•OV:;:;.^'.^
S;->"':>>S".;xv
1998



SSiS
1999



S;S:WS
1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999






FIGURE 2

-------
NOx Lb./MWh
  1.50
       NOx Emissions Compliance Schedule
                       (Rule 1135)
  1.10
  1.00


   .75


   .50


   .25
                   Proposed Rule 1135
Compliance Illustration
        I    I    I    I
                     I    I    I     I
    1989  90   91   92   93   94   95  96   97   98   99  2000
                       FIGURE 3

-------
                          Low NOx Burner
                         Typical Installation
en
CD
                                   Flue
                                   Gas
                                                       To stack
                                            Air
                                         Preheater
Forced Draft
   Fan
                               FIGURE 4

-------
                 High Energy Urea Injection
                        Demonstration Project
 Super Heater
 Urea
Injection
 Ports
Reheater
                            Super Heater
                             Economizer

                            To Air Heater
                             High Pressure
                             Urea Injection
                                Pump   Ok
                                          Urea Day Tank

                                     Air Compressor
                           FIGURE  5

-------
CD
      Burners
                             SCR/Air Preheater
                            Typical Installation
                                           Economizer
Air preheater

SCR modules
              To
              injection grid
                                            To stack   vaporizer
                                FD Fan
                                            CD
                                          Liquid ammonia
                                           storage tank
                           Air
                         blower
                                   FIGURE 6

-------
                     Economizer SCR
                    Typical Installation
                  Rue
                  Gas
 Boiler
Burners ^' •
        HI Catalyst
          Modules
                        Economizer
To ammonia
injection grid
                                     To stack
                                             Ammonia
                                             vaporizer
                          Air   Forced Draft
                        Preheater   Fan
                                  Liquid ammonia
                                    storage tank
             Air
            blower
                           FIGURE 7

-------
Proposed Demonstration Technologies
 NOx Controls         $/kW      % Reduction



 High energy urea        25          50-60



 Economizer SCR        34          50-80



 SCR air preheater       23           40



 Burners               10           15
               FIGURE 8

-------
Suggested Scenarios for NOx Reduction

NOx (ppm)

 120


 100


  75


  50
  25
Baseline
                New limit
              SCR
              Economizer
         I	I	I	I
        25   50   75  100%
          Load (percent)
                FIGURE 9

-------
THE GLOBAL ATMOSPHERIC BUDGET OF NITROUS OXIDE
                   Joel S. Levine
             Atmospheric Sciences Division
            NASA Langley Research Center
               Hampton, Virginia 23665

-------
             THE GLOBAL ATMOSPHERIC BUDGET OF NITROUS OXIDE
ABSTRACT

While only a trace constituent in the atmosphere at a concentration of about 0.31 parts
per million by volume, nitrous oxide  is very important. Nitrous oxide is  a greenhouse  gas
that  impacts global climate and  also  leads to the  chemical destruction of stratospheric
ozone, which  shields the Earth  from biologically  lethal  solar ultraviolet radiation  (200-
300  nm).  Nitrous oxide is increasing  in the atmosphere at  a rate of 0.2-0.3% per year.
Fundamental uncertainties exist in  our understanding of global sources of nitrous  oxide.
Recent measurements have downgraded the global  production  of nitrous oxide by  two
sources once believed important—fossil fuel combustion and biomass burning. Suggestions
for new sources of nitrous oxide include the "fertilization" of natural soils by nitrate formed
from atmospheric  nitric oxide which enhances  biogenic soil emissions of nitrous oxide  and
enhanced biogenic soil emissions of  nitrous oxide following surface burning.
                                       1-67

-------
            THE GLOBAL ATMOSPHERIC BUDGET OF NITROUS OXIDE
INTRODUCTION

Nitrous oxide (N2) with a concentration of only about 0.31  parts per million by volume is the
most abundant atmospheric nitrogen species after molecular nitrogen (N2). Nitrous oxide is
a very long-lived gas with an atmospheric lifetime of about 150 years (1). Nitrous oxide is an
important atmospheric constituent for two reasons—it is a greenhouse gas that traps Earth-
emitted infrared or heat energy (2) and it leads to the chemical destruction of stratospheric
ozone (1). A single nitrous oxide molecule has the greenhouse warming potential of about
250 carbon  dioxide molecules  (3) with strong absorption bands at 520-660 cirr1,  1200-
1350 cm-1,  and 2120-2270 cm-1 (4). Nitrous oxide is chemically inert in the troposphere
and is only destroyed once it diffuses into the stratosphere.  The atmospheric destruction of
nitrous oxide is due to photolysis and reaction with excited atomic oxygen (O(1D))  (1):

(1)  N2O + hi/ -> N2 + O(1D), A < 341  nm
(2)  N2O + O(:D) -* NO + NO
(3)  N2O + O^D) -> N2 + O2

The photolysis of  nitrous oxide  (reaction (1)) is responsible for about 90% of its destruction
with reactions (2)  and (3) each accounting for about 5% of its destruction  (1).  Reaction (2)
leads to  the production of nitric oxide  (NO) which  leads to the  chemical  destruction of
stratospheric ozone (O3) through the nitrogen oxide catalytic cycle (1):

(4)  NO + O3 - NO2 + O2
(5)  NO2 + O -+ NO + O2

The net reaction of reactions (4) and  (5) is: O3 +  O -> 2O2. The nitrogen oxide catalytic cycle
is responsible for about  70% of the  global chemical destruction of stratospheric ozone (5).
Stratospheric ozone absorbs solar ultraviolet radiation (200-300 nm) and shields the  Earth's
surface from this  biologically  lethal radiation.   Measurements indicate that atmospheric
concentrations of  nitrous oxide  are increasing at a rate of about 0.2 to 0.3%  per year (6).

GLOBAL SOURCES  OF NITROUS OXIDE

The major uncertainty in our understanding of nitrous oxide concerns its global sources.  The
magnitude of its global sources must balance  its rate of global destruction  plus its rate of
                                        1-68

-------
atmospheric accumulation. Photolysis (reaction (1)) and reaction with excited atomic oxygen
(reactions (2) and (3)) destroy about 10.5 ± 3.0 Teragram of N in the form of N2O per year
(1 Teragram of or 1 Tg = 106 metric tons = 1012 grams) (7). The atmospheric accumulation
of nitrous oxide requires an additional 3.5 ± 0.5 Tg N per year (7). Hence, the total global
production of nitrous oxide must be about 14 ± 3.5 Tg  N  per year.

Estimates of global  sources of nitrous oxide are summarized in Table 1.  Inspection  of
Table 1 indicates that soils, either natural or unfertilized, and fertilized agricultural fields are
important global sources of nitrous oxide.  Nitrous oxide  is a free intermediate  in microbial
denitrification in anaerobic environments.  Denitrification involves the reduction of soil nitrate
(NO3~) to nitrate (NO2~) and then to nitrous oxide. However, almost all of the nitrous oxide
produced by denitrification in anaerobic environments is consumed by microorganisms which
use nitrous oxide as an oxidant.  Significant quantities of nitrous oxide are also produced in a
variety of aerobic or partially aerobic soil environments via  nitrification. Nitrification involves
the oxidation of reduced soil nitrogen, such as a ammonium (NH4+), to nitrous oxide. In the
oceans it is unclear whether nitrous oxide is primarily produced from nitrification in the oxygen-
rich surface waters or from denitrification in the oxygen-deficient  deep waters.  In fertilized
agricultural fields the use of nitrate and ammonium fertilizers enhances the production  of
nitrous oxide via denitrification and nitrification, respectively.  The  conversion percentage  of
fertilizer  nitrogen to  nitrous oxide ranges from 0.01  to about 2% (15). The annual global
production of nitrogen fertilizer in 1990 was estimated at about 55 Tg N and is increasing
with time (15).  The leaching of  nitrogen fertilizers from agricultural fields into ground water
may result in additional  emissions of nitrous oxide (16).

The total global production of nitrous oxide of 11.2-16.1 Tg N per year (Table 1) is consistent
with the amount needed to balance the rate of global destruction of nitrous oxide and its rate
of accumulation in the atmosphere (10.5-17.5 Tg N per year).  However,  that was before
measurements of nitrous  oxide  from combustion sources collected and stored  in sampling
bottles prior to analysis using a gas  chromatograph/electron capture  detector were found
to be questionable due to  the  presence of an artifact which caused nitrous  oxide in the
sampling bottles to increase with time (17). In the sampling bottles, fossil  fuel combustion
products, nitric oxide, sulfur dioxide, and water vapor formed nitrous oxide at a level of several
hundred  parts per million only several days after collection  (17). Since all determinations of
combustion-produced nitrous oxide were based on the chemical analysis of samples collected
and stored in such  bottles, all  such  measurements are questionable.  Real  time, in situ
continuous analyzers for nitrous oxide measurements in fossil fuel burners recorded nitrous
oxide concentrations of only 5 parts per million or less (17,18,19). These significantly lower
nitrous oxide concentrations correspond to a fossil fuel combustion source of nitrous oxide of
                                         1-69

-------
only between 0.1-0.3 Tg N per year(2Q), rather than the earlier estimates of about 3.2 Tg N
per year (12).

The question arose whether the nitrous oxide  artifact discovered in  fossil fuel combustion
sampling bottles also affected biomass burn samples which are collected in identical sampling
bottles. Measurements did indeed confirm that  on occasion, concentrations of nitrous oxide
increased in sampling bottles several  days after collection (21). To assess the concentration
of nitrous oxide produced  in biomass burning without the artifact effect,  a real time, in situ
measurement technique was  developed which consisted of a gas chromatograph/electron
capture detector  flown in a  helicopter  directly over a 300-hectare fire  (22).   This new
measurement technique yielded a  mean emission  ratio of  nitrous oxide to carbon dioxide
production in biomass burning of 0.015% (22). To arrive at the global nitrous oxide production
per year due to biomass burning, this emission ratio is multiplied by the global carbon dioxide
production per year due to biomass burning. For a value of 2850 Tg carbon in the form of
carbon dioxide produced per year by  biomass burning, the corresponding annual production
of nitrous oxide is about 1  Tg  N (22).  For 1425  Tg carbon in the form of carbon dioxide, the
corresponding annual production of nitrous oxide is  about 0.5 Tg N.

Using the new artifact-free  estimates for nitrous oxide production due to fossil fuel combustion
and biomass burning, the global production of nitrous oxide  is reduced from 11-16 Tg N per
year (see Table 1) to 7.5-12.6 Tg  N, and the global budget of nitrous oxide is no longer in
balance.

The missing nitrous oxide needed to balance  the global destruction and the atmospheric
accumulation of nitrous oxide may be due either to an underestimate of the strength of
known sources or it may be that there are as yet unknown global sources of nitrous oxide.
Two new sources of nitrous oxide have recently been suggested.  Measurements of biogenic
emissions of both nitrous oxide and  nitric oxide from temperate soils following burning indicate
a significant enhancement in the emission of both of these  gases (23, 24). Prior  to surface
burning, the biogenic emissions of nitrous oxide were not detected, which indicates a nitrous
oxide emission of below  2 ng  N  rrr2 s~l, the minimum  detectable emission of the gas
chromatograph,  to  more  than 20  ng N rrr2 s"1 following burning  (24).  Measurements
indicated that the enhanced post-burn nitrous oxide  emission persisted for at least 6 months
after the burn (23). The post-burn enhancement of nitrous oxide is  believed to  be related
to the measured  post-burn enhancement of soil ammonium (NH4+) (24).  Measurements
indicate that soil  ammonium  increased  by  more than a factor  of 3,  while the soil nitrate
(NO3-) decreased after burning (24). Soil ammonium is the substrate utilized by nitrifying
microorganisms in the production of both nitrous oxide and nitric oxide (24,25,26,27).  Recent
                                        1-70

-------
measurements indicate that on a global basis,  biomass burning is much more widespread
than previously believed and  that it appears to be increasing with time (28,29,30).  More
information is needed before the impact of enhanced post-fire nitrous oxide emissions on the
global production of nitrous oxide may be  accurately assessed.

Another source not previously evaluated  is enhanced emission of nitrous  oxide resulting
from "fertilization"  of atmospheric  nitrate on natural  soils (31).  The atmospheric nitrate
that stimglates  microbial production  of nitrous oxide results from nitric  oxide chemically
transformed to nitrate in the atmosphere. Hence, it may be that nitric oxide  may enhance the
biogenic emission of nitrous oxide in natural soils (31). The impact of nitric oxide-produced
nitrate on the global production of nitrous  oxide has not yet been assessed. It is ironic that
nitric oxide emissions  produced in  part to reduce nitrous oxide emissions in various fossil
fuel combustion schemes may eventually  lead to enhanced emissions of nitrous oxide from
the soil.

The buildup of atmospheric greenhouse  gases,  carbon dioxide,  nitrous  oxide, methane,
and chlorofluorocarbon 11 and 12  will lead to global  warming (20). A global temperature
increase will have a positive feedback on  soil emissions of nitrous oxide. The production of
nitrous oxide in soil via denitrification and nitrification  increases with soil temperature (27).
Furthermore, global warming  may result in increased  drought conditions (20).  Both higher
temperatures and drought conditions are  conducive to an increased frequency of burning.
Increased burning will lead  to enhanced production of nitrous oxide both  as a  direct
combustion product of burning and as post-burn enhanced biogenic soil emissions of nitrous
oxide.  Hence, global warming may very well lead to  enhanced emissions of nitrous oxide
due to biogenic production in  soil and by biomass burning combustion.

REFERENCES

1. R. P Turco.   "The  Photochemistry  of the  Stratosphere."   The Photochemistry of
   Atmospheres (J. S. Levine, editor).  Orlando: Academic Press,  Inc., 1985, pp. 77-128.

2. W. R.  Kuhn.   "Photochemistry, Composition, and Climate."  The Photochemistry of
   Atmospheres (J. S. Levine, editor).  Orlando: Academic Press,  Inc., 1985, pp. 129-163.

3. C. S. Silver and R. S.  DeFries. One Earth  One Future:  Our Changing Global Environ-
   ment. Washington,  D.C.:  National Academy Press, 1990, pp. 64-67.

4. J. F.  B. Mitchell. "The Greenhouse Effect and Climate Change." Reviews of Geophysics,
   27, 1988, pp. 115-125.
                                       1-71

-------
 5. R. P. Wayne.  Chemistry of Atmospheres.  London:  Oxford University Press,  1985,
   pp. 113-173.

 6. R. Prinn, D. Cunnold, R. Rasmussen, P. Simmonds, F. Alyea, A. Crawford, and R. Rosen.
   "Atmospheric Trends and Emissions of Nitrous Oxide  Deduced  from Ten Years  of
   ALE/GAGE Data." Journal of Geophysical  Research, 1990.

 7. Word  Meteorological Organization.    Global Ozone  Research and Monitoring Project
   Report No. 16:  Atmospheric Ozone 1985,  1985, Volume I, pp. 77-84.

 8. P A. Matson and P. M. Vitousek. "Cross-System Comparisons of Soil Nitrogen Transfor-
   mations and Nitrous Oxide Flux in Tropical Forest Ecosystems." Global Biogeochemical
   Cycles, 1, 1987, pp. 163-170.

 9. F. Luizao, P. Matson, G. Livingston, R. Luizao, and P. Vitousek.  "Nitrous Oxide Flux
   Following Tropical Land Clearing." Global Biogeochemical Cycles, 3, 1989, pp. 281-285.

10. J. Schmidt,  W. Seiler, and R. Conrad.  "Emission of Nitrous Oxide from Temperature
   Forest Soils into the  Atmosphere." Journal of Atmospheric Chemistry,  6, 1988, pp. 95-
   115.

11. R. D. Bowden, P. A. Steudler, J. M. Melillo,  and J. D. Aber.  "Annual Nitrous Oxide Fluxes
   from Temperate Forest  Soils in Northeastern United States."  Journal of Geophysical
   Research. 1990.

12. W. M. Hao, S. C.  Wofsy, M. B. McElroy,  J.  M. Beer, and M. A. Togan.  "Sources of
   Atmospheric Nitrous Oxide  from  Combustion." Journal of Geophysical Research,  92,
   1987, pp. 3098-3104.

13. J. H. Butler, J. W. Elkins, T. M. Thompson,  and K. B. Egan. "Tropospheric and Dissolved
   N2O in the West Pacific and East Indian Oceans During the El Nino-Southern Oscillation
   Event of 1987." Journal of Geophysical Research, 1990.

14. P. J. Crutzen, A. C. Delany, J. Greenberg,  P. Haagenson,  L. Heidt,  R. Lueb, W. Pollock,
   W. Seiler, A. Wartburg,  and P  Zimmerman.  "Tropospheric Chemical  Composition
   Measurements  in Brazil  During the  Dry  Season." Journal of Atmospheric Chemistry, 2,
   1985, pp. 233-256.

15. R. Conrad, W. Seiler, and G. Bunse. "Factors Influencing  the Loss of Fertilizer Nitrogen
   into the  Atmosphere As N2O." Journal of  Geophysical Research, 88,  1983, pp. 6709-
   6718.
                                       1-72

-------
16.  D.  Ronen, M.  Mordeckai,  and E. Almon.  "Contaminated  Aquifers Are A  Forgotten
    Component of the Global N2O Budget." Nature. 355, 1988, pp. 57-59.

17.  L. J. Muzio and J. C. Kramlich.  "An Artifact in the Measurement of N2O from Combustion
    Sources." Geophysical Research Letters, 15, 1988, pp. 1369-1372.

18.  L. J. Muzio,  M. E. league, J. C. Kramlich,  J. A. Cole,  J. M.  McCarthy,  and R. K.
    Lyon.   "Errors in Grab Sample  Measurements of  N2O  from  Combustion Sources."
    Journal Air Pollution Control Association. 39, 1989, pp. 287-293.

19.  T. A. Montgomery,  G. S. Samuelson, and L.  J. Muzio. "Continuous Infrared  Analysis
    of N2O in  Combustion  Products."  Journal Air Pollution Control Association. 39, 1989,
    pp. 721-726.

20.  J. T. Houghton, G. J. Jenkins, and J. J. Ephraums. Climate Change:  The IPCC Scientific
    Assessment. Cambridge: Cambridge University Press, 1990.

21.  W. R. Gofer III, J. S. Levine, E. L. Winstead, and B. J. Stocks. "Gaseous Emissions from
    Canadian Boreal Forest Fires." Atmospheric Environment, 24A, 1990, pp. 1653-1659.

22.  W. R. Gofer III, J. S. Levine, E. L. Winstead, and B. J. Stocks. "New Estimates of Nitrous
    Oxide Emissions from Biomass Burning." Nature, 349, Feb. 21, 1991.

23.  I. C. Anderson, J.  S. Levine,  M. A. Poth, and P. J. Riggan.   "Enhanced  Biogenic
    Emissions  of Nitric  Oxide and Nitrous  Oxide Following  Surface Biomass  Burning.
    Journal of Geophysical Research,  93, 1988, pp. 3893-3898.

24.  J. S. Levine, W.  R.  Gofer  III, D. I. Sebacher, E. L. Winstead, S. Sebacher, and P. J.
    Boston.  "The  Effects of Fire  on  Biogenic Soil Emissions of Nitric Oxide and Nitrous
    Oxide." Global Biogeochemical Cycles, 2, 1988, pp. 445-449.

25.  J. S. Levine, T. R. Augustsson, I.  C.  Anderson, J. M.  Hoell, and D. A.  Brewer.  "Tropo-
    spheric Sources of NO^:  Lightning and Biology."  Atmospheric Environment, 18, 1984,
    pp. 1797-1804.

26.  I. C. Anderson and J. S. Levine. "Relative Rates of Nitric Oxide and Nitrous Oxide Produc-
    tion by  Nitrifiers, Denitrifiers, and  Nitrate Respirers." Applied and Environmental Micro-
    biology, 51, 1986, pp. 938-945.

27.  I. C. Anderson and  J. S. Levine.   "Simultaneous Field  Measurements of  Biogenic
    Emissions  of Nitric Oxide  and  Nitrous Oxide." Journal of Geophysical Research, 92,
    1987. pp. 965-976.
                                       1-73

-------
28.  J. S. Levine.  "Global Biomass  Burning: Atmospheric, Climatic, and Biospheric Impli-
    cations."  EOS, Transactions of  the American Geophysical Union, 71, 1990, pp.  1075-
    1077.

29.  J. S. Levine.  "Atmospheric Trace Gases:  Burning Trees and  Bridges."  Nature. 346,
    1990, pp. 511-512.

30.  J. S. Levine.  Global Biomass Burning.  Cambridge,  Massachusetts: MIT Press, Inc.,
    1991.

31.  J. W.  Ekins.  Presented at NASA/NOAA/EPA Workshop on Scientific Basis of Global
    Warming Potential Indices, Boulder, Colorado, November 14-16, 1990.
                                       1-74

-------
                       Table 1
      Estimates of Global Sources of Nitrous Oxide
                 Units: Tg N per year

Natural soil emissions:
  Tropical forests (8)                           3.7
  Tropical forests transformed                0.8-1.3
    to pastures (9)
  Temperate forests (10,11)                  0.01-1.5
Combustion of fossil fuels  (12)                  3.2
Ocean (13)                                  1.4-2.6
Biomass burning  (14)                          1.6
Fertilized agricultural fields  (15.)               0.01-1.1
Fertilizer leaching into groundwater (16)       0.5-1.1
Total                                      11.22-16.10
                         1-75

-------
            Session 2




 LARGE SCALE COAL COMBUSTION
Chair:  B. Martin, EPA and G. Off en, EPRI

-------
     DEVELOPMENT AND  EVOLUTION  OF THE ABB  COMBUSTION ENGINEERING
                   LOW  NOX CONCENTRIC FIRING SYSTEM
         John Grusha, Manager of Firing Systems Engineering
Michael  S.  McCartney. Director,  Fuel  Systems and Controls Engineering
              ABB  Combustion  Engineering Services,  Inc.

-------
                         Development and Evolution of the
            ABB Combustion Engineering Low NOx Concentric Firing System
                John Grusha,  Manager of Firing Systems Engineering
       Michael  S.  McCartney,  Director, Fuel  Systems and Controls Engineering
                     ABB Combustion Engineering Services, Inc.
In the 1989 EPA EPRI Symposium in San Francisco, Combustion Engineering and Ferco
reported on the CE/Mitsubishi Heavy Industries  (MHI) Pollution Minimum (PM) coal
retrofit at Kansas P&L Lawrence #5.  Those reports documented MCR NOx levels of
less than  .3 Ibs/mmBtu which has been recently  improved by the plant personnel  to
less than  .25 Ibs/mmBtu over the top 50% of the load range (Figure 1).  Clearly
the Lawrence demonstration met and in some aspects, exceeded the expectation of
the project sponsors.  Many of the program participants were also duly impressed
with the complexity of the P.M. retrofit.  In virtually all cases, the coal PM
burner requires a replacement of the original windbox enclosure which is a major
task.  In the case of Lawrence #5, windbox replacement was probably easier than
can be anticipated with most other units, due to the accessibility around the
unit.

During the same time period, ENEL, the Italian  national utility, installed a coal
PM burner on the 320 MW Fiume Santo.  This unit was under construction at the time
the PM was substituted for the standard windbox.  Faced also with the need to
retrofit existing tangentially fired, multi-fuel units, ENEL was receptive to
demonstrating an alternative to the PM technology that was more tailored for
retrofit.  With the obvious market in the U.S.  generated by the pending Clean Air
legislation, together with a committed host site in Italy, ABB/CE committed to an
accelerated R&D program to develop an advanced  low NOx multi-fuel firing system
specifically for retrofit to existing tangentially fired units.

The intent of this paper is to report on the development, demonstration and
subsequent evolution of this low NOx technology for tangential, multi-fuel fired
boilers from the original  concept called Clustered Concentric Tangential Firing
System (CCTFS)  to the incorporation of its clustering feature into the
commercially established LNCFS product line.
                                       2-3

-------
CCTFS Concept

Classic NOx theory as is generally accepted in the industry identifies three NOx

formation mechanisms called prompt, thermal and fuel bound NOx.  At the risk of

oversimplification of these very complex kinetics, one can reasonably state that

time, temperature and the availability of elemental nitrogen and 02 are the

primary variables for all three mechanisms.  Within the constraints of existing

units, time and temperature are not practical methods of control.  Time for

example, is fixed by the volume flow rate of the products of combustion and the

volume of the existing furnace.  Temperature is also a function of the degree of

air preheat, net fuel heat input and the amount of heat absorbing surface in the
furnace.  Since the  air preheat temperature, net heat fired and heat absorbing

surface are fixed, temperature also is eliminated as a practical method of NOx

control.  Thus the control of elemental nitrogen and oxygen is, by default, the
most  cost-effective  means of controlling NOx.  The practice of temporarily

withholding oxygen from the combustion process is generally referred to as

staging.  The use of staging is the common denominator of virtually all low NOx

burner designs for both wall and tangential systems and was the basis of the

original CCTFS concept.


CCTFS  stands for Concentric Clustered Tangential Firing System.  The concept
behind CCTFS is to stage the combustion process at three points throughout the
history of the fuel  in the furnace.

      1.   Early Staging:  A concept called clustering is used to produce
          early staging by the coal nozzles being grouped together without
          intermediate air  (Figure 2).  The clustered fuel nozzles are
          separated  by large distances and large intermediate air
          compartments.  The theory is that fuel nozzles when placed next to
          each other will entrain  less air as the fuel enters the furnace
          which will result in a more fuel rich environment during ignition
          and the early devolatilization process.  This temporary surplus of
          ignited fuel depletes the available oxygen and in turn, forces the
          kinetic path of the fuel bound nitrogen to N2.  It is essential
          with this  concept to have early coal ignition producing
          devolatilization and fuel bound nitrogen release within the time the
          fuel jets  can stay air-lean within the furnace. The fuel nozzles
          are designed with a flame attachment feature borrowed from existing
          LNCFS technology to encourage stable flame propagation from the
          furnace back to a point  close, near the fuel nozzle tip.

      2.   Intermediate Staging:  No early staging technique can effectively
          control fuel NOx from coal to any high degree, primarily because
          large quantities of nitrogen can evolve from the fuel well after
          complete devolatilization and well after the point in time that
          fuel and air nozzles can maintain local fuel/air ratios.  To
          achieve intermediate staging a "close coupled" overfire air
          compartment is incorporated into the top of the tangential


                                      2-4

-------
          windboxes.   This  portion  of  the  CCTFS  is  a  copy  of  the  original
          overfire  air first  used on tangential  designs  since the early
          1970s.  Because  intermediate and late  stage techniques  drive
          overall stoichiometry  below  1.0,  the  intermediate air nozzles  are
          designed  to  direct  combustion  air away from the  center  of  the
          furnace and  toward  the waterwalls.  This  technique  called
          concentric firing is well established  as  a  method to keep  oxygen on
          the  waterwalls  and  increase  lower furnace heat absorption.   In  and
          by  itself, it does  not reduce  NOx,  but it plays  a crucial  role  in
          that it offsets  the tendency of  high  quantities  of  overfire  air to
          slag the  lower  furnace and increase furnace outlet  temperature.  It
          too  was borrowed  from  existing LNCFS  technology.

      3.   Late Staging:  In theory, the  most  effective location for
          introducing  the  air required to  stage  and then complete combustion
          is  as close  to  the  furnace outlet as  possible.   This maximizes  the
          time the  fuel is  staged and  minimizes  the time in an excess  air
          condition.   Since the  fuel must  be  burned to completion at some
          point before the  furnace  outlet,  this  late  staging  device  must  be
          the  most  effective  mixing system in the furnace.  With  CCTFS,  this
          is  done with "separated overfire" (SOFA)  windboxes  located higher
          in  the furnace.   Each  is  equipped with multiple  air nozzles  with
          the  ability  to  tilt in both  the  vertical  plane (Pitch)  as  well  as
          the  horizontal  plane  (Yaw).   The air  to the separated overfire  air
          is  of sufficient  quantity to keep the  mid furnace around 1.0
          stoichiometry and was  boosted  in the  laboratory  development, to
          approximately 20  in w.g.


 Laboratory  Development of CCTFS

 The CCTFS was  developed in  ABB/CE's Kreisinger  Development Laboratory  (KDL)  in

 Windsor,  CT.   The test facility  is  a 50  mmBtu/hr (15  mwt)  facility sized  to

 simulate  large boiler  residence  times.  The temperatures are  replicated  by

 selective refractory lining (Figure 3).  One of the program's main objectives was

 to test  and evaluate the  various candidate designs  from the standpoint of a

 variety  of  different coals.  It  was felt that many  problems in the past  were

 caused by reaching  conclusions  on the  basis of  one  fuel,  only to  find  that  it was

 not repeatable on another fuel.   In addition, the host utility,  ENEL,  had a  fuel

 policy that required a very wide range of  potential fuels,  ranging from  South

 African  to  U.S. high volatile bituminous fuels.   The  fuels tested are  summarized

 below:
Source                HHV            FC/VM          %N           %S           %Ash


Ashland,  KY           13430          2.0            1.4          .8            9.3

Virginia               14150          2.0            1.7          .8            6.6

Utah                  11740          2.7            1.5          .6           16.1

W.  Virginia           13310          5.0            1.4          2.3          13.3
                                       2-5

-------
The program was built on prior work on concentric firing and close coupled OFA,
with the primary interests of the program centered on the development of a optimum
separated OFA system and clustering arrangement.  In the interest of time and the
ABB-CE "proprietary" interest, a complete description of the test matrix and
results are beyond the intent of this paper.  However, some of the KDL conclusions
and supporting data is presented to understand the evolution of CCTFS.  These
conclusions are as follows:
     1.   Splitting the overfire air distribution between the close coupled
          and the separated OFA positions produced the same or lower NOx than
          placing all the OFA in the separated positions.  This is shown in
          Figure 4.  The optimum OFA distribution was fuel specific.
     2.   The ability to vary the yaw (angular motion in horizontal plane)
          had a clear impact on combustion efficiency.  Figure 5 shows one
          example of the effect of yaw angle on carbon in ash values.
     3.   Moderate levels of OFA had the effect of slightly reducing furnace
          outlet temperatures; however, larger quantities had the reverse
          effect of raising them significantly.  The CFS nozzles reduced
          furnace outlet temperature resulting in a zero degree net change in
          furnace outlet temperatures when CFS and high quantities of OFA
          were used together  (Figure 6).
     4.   The clustering techniques reduced NOx approximately 15% under
          conditions with low quantities of OFA.  At 30% OFA, there was no
          significant difference in NOx or combustion efficiency between
          clustered and non-clustered configurations.
The CCTFS configuration was capable of achieving approximately 200 ppm NOx
corrected to 3% 0~ (60% reduction)  while operating at 3.1% 02 on  an Eastern high
volatile bituminous coal.   The test clearly showed the dominance  of overfire air
flow on NOx levels.   All  other features of the KDL demonstrated CCTFS were
successful in maintaining  furnace outlet temperatures,  surplus 0? on the
waterwalls and minimizing  unburned  carbon,  but the quantity of OFA determined NOx
levels.  The lesson learned was that NOx reduction from clustering was not
multiplicative or even additive to  the OFA contributions.   Larger quantities of
OFA appear to not only prevent the  formation of NOx,  but reduce NOx formed in the
burner area.

Fusina #2 Description
ENEL's Fusina #2 unit is a balanced draft,  radiant-reheat, tangentially fired,
boiler designed and manufactured by Franco Tosi Legano, Italy under license from
ABB-CE.   It is a multi-fueled unit  capable of full load operation on either coal,
oil or natural gas.  Four CE 623 RS exhauster-type pulverizers are used to supply
four levels of tilting coal nozzles which are located in the four corner
windboxes.  In addition, four elevations of oil and natural gas firing equipment

                                        2-6

-------
also exist.  A side elevation view of Fusina #2 is shown in Figure #7.  A brief
summary of unit design conditions are as follows:

     Megawatt rating - 160 MW
     Main steam flow - 1,119,800 Ibs/hr
     Throttle pressure   2,100 PSIG
     SH/RH temperatures - 1,005/1,005°F
     Contract year - 1967

The unit was designed to fire low sulphur bituminous type coal, moderate sulfur #6
oil and natural gas.

The original firing system windbox arrangement consisted of four elevations of 12"
coal nozzles equally spaced throughout the height.  Each windbox is 16" wide and
approximately 21' 7-3/4" high.  Located between the lower three coal elevations
and above the uppermost elevation is the oil and gas firing equipment.

In order to incorporate the Clustered Concentric Tangential  Firing System (CCTFS),
the original windbox arrangement had to be reconfigured.  These changes to the
original windbox are shown in Figure #8.  The most significant change  was in
clustering or close arrangement of both the upper two and lower two coal
elevations.  Additional design requirements of the reconfigured CCTFS  windboxes
included a close-coupled overfire air system, "flame attachment" coal  nozzle tips
and offset concentric air nozzle tips similar to the KDL CCTFS arrangement.   In
order to incorporate all of these changes, oil  and natural  gas firing  equipment
levels had to be relocated.

Equally important as the windbox modification,  was the addition of a separated
overfire air system (SOFA) for late staging and completion of the combustion
process.  This consisted of four smaller windboxes located approximately 15'
directly above the main windbox.  Each SOFA windbox included three tilting air
nozzle tips, with the capability of tilting vertically + 30° independent of the
main windbox.  Also incorporated into the separated overfire air system is the
unique feature of being able to horizontally adjust each nozzle tip either toward
or against the main fireball rotation see Figure #9.

The sizing criteria for both the close-coupled and separated overfire air system
at Fusina was based on the results of the CCTFS development program in KDL.  The
close-coupled overfire air was designed to deliver 10% of the combustion air,
                                        2-7

-------
while the separated overfire air was sized for 20% of the total combustion air.
Also as a result of the KDL program, high pressure boost fans capable of 25 in.
w.g. pressure were installed on the separated overfire air system in order to
optimize the discharge velocity of the separated overfire air independent of the
main windboxes.  Its intended benefit was to allow utilization of higher
percentages of staged combustion air while maximizing upper furnace fuel/air
mixing and minimizing CO and unburned carbon in the flyash.

In  addition to the overfire air, each coal elevation had flame attachment type
coal nozzle tips to help promote the early initiation of coal ignition under
oxygen deficient conditions, an established low NOx requirement.

To  direct combustion air down along the waterwall, offset air nozzle tips were
located within the CCTFS windbox, to minimize reducing atmospheres and control
F.O.T. under staged coal firing conditions.

During the evaluation of the CCTFS performance, four different coals were
evaluated.  These  ranged from two South African medium volatile coals to two U.S.
eastern bituminous high volatile coals.  A tabulation of their analyses are listed
in  Figure #10.   (Also listed for later comparison is the analysis of the western
bituminous coal  fired with  a LNCFS system at PSCC, Valmont.)

Fusina #2 Test Results - Coal
Approximately  150  tests were conducted with the CCTFS at Fusina #2.  Four
different coals  were evaluated.  Throughout the parametric testing, all the
features of the  CCTFS were  evaluated against NOx, CO, carbon loss and other boiler
performance.   As expected,  each coal exhibited distinct characteristics.

Figure #11 presents the effect  of firing zone stoichiometry on NOx for the
different types  of coal with the CCTFS.  For clarification, firing zone
stoichiometry  is that percentage of total combustion air introduced at or below
the uppermost  fuel elevation.   Combustion air introduced above this upper
elevation is considered overfire air.

The results show that with  approximately 90% firing zone stoichiometry, the CCTFS
produced nearly  50% NOx reduction.  The results further show that the South
African coals  (TCOA, AM coal),  produced overall higher NOx emissions than the  U.S.
eastern bituminous coals.
                                        2-8

-------
Besides the NOx reduction capability of the CCTFS, the effect of firing zone
stoichiometry on unburned carbon was investigated.  This is a common concern among
many boiler operators.  As the percentage of total air to the main firing zone was
decreased for NOx reduction, the unburned carbon increased for the same coal
fineness, see Figure #12.
The first series of tests conducted with the South African coals (TCOA, AM Coal),
using a coal fineness of 85% through 200 mesh (3.8% on 100 mesh), resulted in an
increase of unburned carbon from 9% to 12% when the firing zone stoichiometry was
reduced from approximately 124% down to 94%.  This same trend was demonstrated
with the U.S. coals which had a coal fineness on the first series of tests of 87%
through 200 mesh (3.6% on 100 mesh).  With these U.S.  coals,  the unburned carbon
again increased, but to a lesser degree.  This unit,  however,  had the pulverizer
capability to increase coal fineness.  Referring again to Figure #12, improving
the coal fineness from 87% to 93% through the 200 mesh (1% on 100 mesh) on the
U.S. coals reduced the unburned carbon under staged firing conditions to below the
unstaged baseline values.  This same trend was demonstrated with the South African
coal s as well .

Figure #13 shows the effect of increasing the separated overfire velocity for
three of the coals.  The TCOA coal was further evaluated under both normal and
higher coal fineness.  The results showed that increasing the pressure behind the
staged combustion air had little effect on improving the percent unburned carbon
in the flyash.   Further it showed no improvement in NOx emissions and CO emissions
never exceeded 40 pptn throughout the test program.

Fusina Test Results - Oil and Natural Gas
In addition to evaluating the low NOx capability of the CCTFS system when firing
coal, the effectiveness of its separated and close coupled overfire air system was
tested on reducing NOx emissions when firing either No. 6 oil  or natural gas.

As with coal firing on tangentially fired units, the use of an overfire air  system
is very effective in reducing NOx emissions.  Diverting a percentage of the  total
combustion air away from the primary combustion zone,  reduces both the thermal and
fuel NOx conversion by way of the staging process.  Figure #14 illustrates the
effectiveness of staged combustion when firing heavy oil.  With heavy oil, 50% of
the NOx formed can be a result of the fuel nitrogen in the oil.  It therefore, has
an important effect on final NOx emissions.  This graph presents the effect  of two
fuel nitrogen levels on NOx emissions as a function of firing zone stoichiometry.
Still higher nitrogen values would expectedly produce higher NOx levels.

                                        2-9

-------
In addition to fuel  nitrogen, operating 02 levels were evaluated. Figure #15
compares the effects of varying CL on NOx and CO, during staged and unstaged
firing conditions.  Varying the CL between 2% and 3% without OFA had a minimal
effect on NOx production.  At the same 02 levels, CO emissions remain relatively
unchanged.

Utilizing overfire air to stage the oil combustion process, again did not appear
to demonstrate a significant sensitivity of NOx production to operating 02 levels.
However, under deep staged conditions at low Op levels, CO emissions increased.
It should be noted such  increases are sensitive to such parameters as atomization,
oil quality, viscosity,  and the mixing efficiency of air and fuel during the early
stages  of the combustion process.  Throughout the low NOx oil firing tests
reported particulate levels never exceeded .1 Ibs/mmBtu prior to the precipitator.

Unlike  coal  and oil, natural gas has no fuel nitrogen.  All of its NOx production
is therefore thermal NOx which is solely dependent on 0? availability and
temperature.  The results from Fusina demonstrate that as the firing zone
stoichiometry is  reduced by the utilization of staged combustion, NOx emissions
decreased  significantly.  Figure #16 illustrates the results of  staged combustion
on natural  gas firing  at Fusina.  NOx reduction efficiencies of  70% were achieved.
CO increased, but never  exceeded 200 ppm (corr. 3% 0?).

Low NOx Concentric Firing System (LNCFS) Results at Public Service of Colorado
Co.,  Valmont #5
Similar goals of  NOx reduction were targeted for a coal fired unit at PSCC,
Valmont #5.  To meet this objective an LNCFS was proposed and installed.  Valmont
#5 being  somewhat similar in size to Fusina #2 made for a good comparison of the
two low NOx firing systems.

Unit  Description
PSCC  Valmont #5 is a tangentially fired boiler manufactured by ABB-CE, and capable
of full  load operation when  firing either pulverized coal or natural gas.  Side
elevation  views of Valmont #5  is shown on Figure #17.  The unit  design conditions
are as  follows:
      MW rating -  165
      Steam  flow   1,230,000 Ibs/hr
      Throttle pressure - 1,800 PSIG
      SH/RH  temperatures  - 1,005/1005°F
      Contract year   1961

                                       2-10

-------
Valmont #5 fires low sulfur, western bituminous coal, a typical analysis is listed
in Figure #10.

The original  windbox arrangement at Valmont #5 shows very little contrast from the
initial Fusina arrangement.  Similarly four elevations of 18" coal nozzles are
supplied by four CE 743 RS exhauster-type pulverizers.  Each windbox is 22" wide
and approximately 21' - 3-1/4" high.  Between the four coal  elevations exist three
levels of gas firing equipment which are capable of full load.  The LNCFS similar
in concept to CCTFS, combines the NOx reducing capabilities of furnace combustion
air staging with early fuel devolatilization and offset air nozzles to control CL
availability; thereby, reducing total NOx emissions.  The major components of the
LNCFS at Valmont #5 are:

     Separated overfire air system
     Offset concentric auxiliary nozzle tips
     Flame attachment coal nozzle tips

Figure #18 comparatively illustrates the modifications made to the original
Valmont windbox to incorporate the LNCFS.

The major differences between the CCTFS and the LNCFS are seen primarily in the
fact that the coal nozzles are not clustered together.  Similar to the CCTFS was
the utilization of flame attachment coal nozzles, but only a separated overfire
air system without boost was incorporated.  Both systems' separated overfire air
arrangements had horizontal and vertical adjustment features proven to provide
optimal fuel  air mixing during staged combustion operation to minimize CO
increases and 02 imbalance.

Valmont #5 Test Results
The effect of firing zone stoichiometry (FZS) on NOx with CCTFS and LNCFS are
compared in Figure #19.  With the units being similar in size, but firing notably
different coals, the overall percent reduction efficiencies of both systems are
shown to be comparable in performance capabilities.  For the same firing zone
stoichiometry due to staged combustion, percent NOx reductions of 50% with the
LNCFS closely approximated the CCTFS performance.

The unburned carbon results from Valmont #5 were superimposed on the CCTFS data
results and are shown in Figure #20.  The unburned carbon with the LNCFS did not
increase with firing zone stoichiometry changes even with poorer overall coal
                                       2-11

-------
fineness.   This is attributed primarily to the fact that western bituminous coals
are highly reactive and do not tend to exhibit unburned carbon easily even under
aggressively staged firing conditions.  This demonstrates the point that unburned
carbon levels under staged conditions are very dependent on coal type.  The less
reactive coals such as the South African coals and Eastern U.S. bituminous coals
used at Fusina were more inclined to an increase in unburned carbon under staged
low NOx firing conditions.  Further the sub-bituminous and lignitic coals which
are highly reactive should exhibit very little if any change in unburned carbon.

No opportunities to test the low NOx gas firing capabilities at Valmont were
avai1 able.

Summary and Conclusions
By comparing the CCTFS configuration results from both the lab and the field
demonstration  at Fusina with the LNCFS results from Valmont, it is clear that the
firing zone stoichiometry overshadowed all other variables, except for final CL,
in determining the outlet NOx levels.  This is not to say that OFA in and by
itself constitutes a low NOx firing system.  High quantities of OFA by itself will
increase  furnace outlet temperatures and depending on the specific fuel properties
may  increase lower furnace slagging and increase unburned carbon loss.  Moreover,
if an OFA  system is a poor mixing system, carbon monoxide, 0? unbalance and a
whole host of  other potential consequences are possible.

Thus the design approach for low NOx retrofit systems on a tangentially fired unit
is focused on  first achieving the best mixing from the OFA systems and secondly,
manipulating the method of fuel  and air introduction to counterbalance the
potentially adverse effects of staging.  In general, we find that a tangential
system can, if designed properly, accommodate large quantities of OFA without
realizing  these negative side effects, with one notable exception.  The increase
in unburned carbon reported in this paper at Fusina under high overfire air flow
modes is  the inevitable consequence of aggressive furnace staging on the less
reactive,  agglomerating bituminous-type coals.  As evidenced by the Valmont data,
this phenomena is less apparent with weakly agglomerating western bituminous or
sub-bituminous fuels.  Where no deterioration in unburned carbon is acceptable,
modification to the pulverizer system will be required to reduce the particle size
of fuel to the furnace.  Modification to pulverizers such as ABB/CE's dynamic
classifier is  designed to reduce the mass fraction of the largest sized fuel
particles  without having to reduce the size of all size fractions.  With devices
such as the dynamic classifier,  it is actually possible to reduce unburned carbon
in ash levels  to a point lower than those measured prior to the low NOx retrofit.

                                       2-12

-------
The retrofit system that will be offered by ABB-CE for coal fired units affected
by the Phase I of the Clean Air Act will be a blend of the Fusina CCTFS advanced
OFA as well as the fuel and offset concentric air nozzle configurations
demonstrated on both Fusina and Valmont.  Because clustering will be dropped as a
NOx control technique, the CCTFS name will also be dropped.  The LNCFS name will
be used for all coal fired retrofits.  This is logical since LNCFS has always
combined OFA, flame attachment coal nozzle tips and offset concentric air nozzle
configurations.

Clearly not all units will require the same percent reduction to meet the .45
Ibs/mmBtu required by the Clean Air Act for tangentially-fired units.  Where the
percent reduction does not require maximum quantities of OFA, a single level of
separated OFA will be utilized as  in Valmont or integrated in the main windbox as
close coupled OFA.  The higher levels of OFA within the LNCFS configuration are
shown in Figure #21 as Level 1, 2  and 3 LNCFS.  In Level 1 and 3, the top two
elevations continue to be clustered which sounds like a contradiction with what
was stated before.  The use of clustering in these situations is simply to make
room within the existing windbox enclosure for close coupled OFA.

All configurations of LNCFS can utilize the existing main windbox which greatly
simplifies and reduces the cost as compared to the PM which was discussed briefly
in the beginning of this paper.  There will  always be a few situations where the
original windboxes cannot be salvaged because they have deteriorated beyond repair
with age.  However, in most cases, the box can be reused.  The cost advantage can
be seen in Figure #22 which shows  the approximate D&E cost of three levels of
LNCFS versus the PM, which always  requires a windbox replacement.  The cost
assumes a 200 MW, 4 corner unit and does not include dynamic classifiers.

As a final note, it should be stated that ABB-CE believes that LNCFS can meet the
Clean Air requirements for virtually all of the Phase I affected units. In
addition, the overfire air technology successfully demonstrated at Fusina #2 is
available for those oil and gas fired units also requiring NOx reductions. On new
boiler construction where the constraints of existing windboxes are obviously not
an issue, ABB-CE plans to continue utilizing the PM technology.
                                       2-13

-------
E
CL
D.

o"
           KP&L Lawrence #5
           400 MW CCRR Unit
           Subbit  A Coal
                                     BASELINE
                  BOILER LOAD (MW)

  Figure 1. NOx vs. Boiler Load with 'PM' Firing
  System
  ORIGINAL
  WINDBOX
  CCTFS
ARRANGEMENT
Figure 3. Boiler Simulation Facility
PRE RETROFIT
                               - SEPARATED OFA WITH
                                ADJUSTABLE YAW
                                -COUPLED OFA
                                ^COAL NOZZLE CLUSTER
                                i-OFFSET AIR NOZZLES
                     POST RETROFIT
                                                         305 -
                                                         300 —
                                                                                                 <0^,


                                                                                                 ^
                                                        Figure 4. NOx vs. OFA Quantity & Elevation
               Figure 2. Clustering
                                                 2-14

-------
                                      FIREBALL
                                      ROTATION
                         ADJUSTABLE
                           YAW -	
                                     FURNACE
                                     PLAN VIEW
           0  YAW    -15 YAW    -15 YAW
     30% SEPARATED OFA, 3.1% O2KENTUCKY H! VOL. BIT.

Figure 5. Separated OFA YAW vs. % Carbon in Ash
    SEPARATED
      OFA
      20
       20
       30
   CFS ANGLE
DEG. FROM FUEL q_
                         *22
AVG. F.O.T.
                                          1276  C
                                          1255  C
                                          1235  C
                                          1275  C
   Figure 6. Effect of Separated OFA and CFS on
   Furnace Outlet Temperature
                                         Figure 7. Fusina #2 Side Elevation
                                                  2-15

-------
      ESS
                                EEEB
                                      Separated
                                      Dverfire
                                     Close
                                     CoupLecl
                                     Overfire

(As Received]
HHV
HOIST
VM
FC
ASH
(DAF)
C
H
N
S
°2
FC/VH
TCOA

1091
a. 6
21.7
55-0
14.5

83.29
4.42
2.03
0 63
9 63
2.53
AM COAL

11844
7.50
24.54
52 76
15.10

71.10
3 82
1 63
0 40
7.74
2-14
MC CALL

14170
1.3
27.8
64 5
6.4

87.1
5.2
1.5
0.9
5.3
2.32
ARCH MINERAL

12731
7.53
34.43
50.38
7.66

76.60
5.16
1 51
0 85
7 5
1.69
W BIT

10957
9 26
34.78
43 86
12.10

79.52
5.80
1.74
0.52
12.33
1.26
                                                             Figure 10. Coal Analysis, CCTFS
                                                        400 I-
                                                     E  300
                                                                     95         105         115

                                                                         Firing Zone Stoichiometry

                                                                      160 MW COAL/OIL/GAS UNIT
                                                     Figure  11. Effect of Firing Zone Stoichiometry on
                                                     NOx for Different Types of Coal with CCTFS
 PREVIOUS  ARRANGEMENT
                           MODIFIED ARRANGEMENT
Figure 8. Changes to Original Windbox, ENEL,
Fusina #2
 CE Separated Overfire Air Assembly
  Patent  Pendinq
Figure 9. CE Separated Overfire Air Assembly
(Coal Fineness thru 200 Mesh)
                                                             95   100  105   110   115   120  125
                                                                 Firing Zone Stoichiometry

                                                                     160 MW COAL/OIL/GAS UNIT
                                                                                               Coal

                                                                                            *  TCOA

                                                                                            °  AMCoal

                                                                                            '->  MoCall

                                                                                            a  Arch Mineral

                                                                                            NOTE.
                                                                                            BASELINE CARBON
                                                                                            VALUES AVERAGED
                                                                                            BETWEEN 7-9%
                                                    Figure 12.  Effect of Coal Fineness on Unburned
                                                    Carbon vs. Firing Zone Stoichiometry with CCTFS
                                                 2-16

-------
                                                           100-
                                  COAL
                                  	 TCOA               .3

                                  -=- TCOA (High Fineness)    o   50 "
                                                        2
                                  ^— MoCall

                                   —  Arch Mineral
     100 150 200 250 300 350 400 450 500
            VELOCITY (FT/SEC)
         160 MW COAL/OIL/GAS UNIT
0.7      0.8      0.9       1
             Firing Zone Stoichiometry

            160 MW COAL/OIL/GAS UNIT
Figure 13. Effect of SOFA Air Velocity on Unhurried    F'9ure 16^fect of Firin9 Zone Stoichiometry on
  '•'                                   •*                   K r\ij- nnf-1 f*ri iui4-l~t  I K DCTC?
Carbon in Flyash with CCTFS
                                                         NOxandCOwithLNBFS
            Fuel - Heavy Oil; 02-2.5%
  O   100 "
                                     Particulates < 0.1
       0.7     0.8
                     0.9      1       1.1
                     Firing Zone Stoichiometry

                    160 MW COAL/OIL/GAS UNIT
                                           12      13
 Figure 14. Effect of Fuel Nitrogen on NOx vs. Firing
 Zone Stoichiometry with LNBFS
             - 	  -  — —  	—  -	. 350
              Fuel - Heavy Oil (N-0.34%. S-0 81%)
  m   250
      150
  O   100
       50
                                            ,100  §
                                           - 50
                1.5      2      2.5

                      Percent O2

                160 MW COAL/OIL/GAS UNIT
                                                              Figure 17. Valmont #5 Side Elevation
  Figure 15.  Effect of O2 on NOx and CO Levels
                                                   2-17

-------
                                                         12 - —   -—  	   —   —
                                                           (Coal Fineness thru 200 Mesh)
                                         Separated
                                         Over-fire
      EuB
 ORIGINAL ARRANGEMENT
                                   FOB
Gas



Offset A,r



Coal



Offset A,r
                                        Offset ,

                                        Gns
                                        an
                                        Gas

                                        Offset ,
                              MODIFIED  ARRANGEMENT
        Figure 18. Windbox Modifications
                                  —=— WBit (Valmont)

                                  CCTFS
                                   — " Arch Mmeral(Fusina)

                                   --- McCall (Fusma)
     30  85 90 95 100 105 110 115 120 125
          Firing Zone Stoichiometry
Figure 19. Effect of Firing Zone Stoichiometry on
NOx for Different Types of Coal
                                                                                        —^- WBit (Valmont)

                                                                                        CCTFS
                                                                                           Arch Mmeral(Fusina)

                                                                                           McCall (Fusina)
                                                          80 85 90  95 100 105 110 115 120 125 130
                                                               Firing Zone Stoichiometry
                                                      Figure 20. Effect of Coal Fineness on Unburned
                                                      Carbon vs. Firing Zone Stoichiometry (200 Mesh)
Standard
Windbox


AIR
COAL
AIR
COAL
AIR
COAL
AIR
COAL
OIL
COAL
AIR
% Reductions'
LNCFS
Level 1


OFA
OFA
COAL
COAL
CFS
COAL
CFS
COAL
OIL
COAL
AIR
25-32
Level 2
OFA
OFA
AIR
COAL
CFS
COAL
CFS
COAL
CFS
COAL
OIL
COAL
AIR
33-40
Level 3
OFA
OFA
OFA
OFA
COAL
COAL
CFS
COAL
CFS
COAL
OIL
COAL
AIR
41 -50
                                                            Figure 21.  Low NOx Retrofit Options
S/KW INSTALLED
n o m o 01 O o i





J
H
                         LNCFS      LNCFS     LNCFS     PM
                         Level 1      Level 2     Level 3

               Figure 22. NOx Reduction Systems, Retrofit Cost
               Comparison
                                                  2-18

-------
     PERFORMANCE OF A  LARGE CELL-BURNER UTILITY
BOILER RETROFITTED WITH FOSTER WHEELER LOW-NOX BURNERS
                      T.  L.  Lu
                    R.  L.  Lungren
            Arizona Public Service Company
                   Phoenix, Arizona
                     A.  Kokkinos
           Electric  Power Research  Institute
                 Palo Alto, California

-------
ABSTRACT
A comprehensive boiler testing program was performed on Units 4 and 5 of the
Four Corners Steam Electric Station to compare the NOX emissions and thermal
performance of a unit retrofitted with low-NOx burners (Unit 4)  with a
"sister" unit still equipped with its original turbulent burners (Unit 5).
Built in the late 1960s, Units 4 and 5 are 800-MW Babcock & Wilcox
supercritical, once-through boilers designed for firing of a western
subbituminous coal.  In 1989, Unit 4 was retrofitted with low-NOx circular
burners designed by Foster Wheeler Energy Corporation; Unit 5 was left
unmodified while awaiting its scheduled retrofit in 1991.  Major objectives of
the comparative testing program were to establish the NOX emissions levels and
to assess any changes in the performance and operability of Unit 4 due to the
installation of low-NOx burners.

Testing included measurement of NOX, CO, and S02 emissions,  unburned  carbon,
gas temperature leaving the economizer, and heat absorption in various boiler
circuits at different levels of unit load and excess air, and with different
burner air register adjustments.  Test results indicate that the low-NOx
burners reduced NOX emissions from Unit 4 by 55% compared with the unmodified
Unit 5, without any detrimental effect on boiler performance, efficiency, or
operability.

This paper should be of interest to any utility evaluating potential NOX
reductions and boiler performance effects that could be anticipated by
retrofitting these low-NOx burners to pulverized-coal-fired utility boilers
with "cell" burners or conventional circular turbulent burners.
                                      2-21

-------
INTRODUCTION
The Arizona Public Service Company (APS)  operates five coal-fired units at the
Four Corners Steam Electric Station located near Farmington, New Mexico.  The
units operate under a state environmental  regulation which limits emissions of
nitrogen oxide (NOX)  to  0.70 Ib/MBtu  from  coal-fired utility boilers.   This
regulation was promulgated in 1972, several years after the two units that
produce the highest NOX  emissions--Units  4 and  5--went into  commercial
operation.

Babcock & Wilcox  (B&W) boilers manufactured during the late 1960s, like Units
4 and 5, were equipped with closely spaced, two- or three-nozzle "cell"
burners specially designed to maximize combustion intensity and produce
extremely high heat releases in a compact burner zone.  These combustion
features result in very high flame temperatures, heavy slagging in the
furnace, and NOX emissions around 1.20 Ib/MBtu  at full load.

Between 1972 and  1984, APS conducted several testing programs and NOX control
technology  studies on Units 4 and 5 in an attempt to achieve compliance with
the  state regulation.  None of these efforts were successful or even
promising.   In 1985, APS  identified the Foster Wheeler Energy Corporation
(FWEC)  Controlled-Flow/Split-Flame (CF/SF) low-NOx burner as a promising NOX
control technology for possible application to the Four Corners boilers.
Subsequent  pilot-scale burner testing programs and design engineering studies
supported a  retrofit of CF/SF low-NOx burners on Units 4 and 5.1   In  1987,  the
retrofit was approved for major overhauls of Units 4 and 5 scheduled for 1989
and  1991, respectively.

BOILER  PLANT DESCRIPTION
Units 4 and  5 are identical B&W opposed-fired, supercritical, once-through,
pressurized  boilers.  Each  is capable of a maximum continuous rated output of
5,445,000 Ib/h main sueam flow at  1000/1000'F.  The units fire a western,  low-
sulfur, high-ash, subbituminous coal with the characteristics shown in
Table 1.  Units 4 and 5 were originally designed by BSW with nine pulverizers
serving 18  three-nozzle cell burners.  The closely spaced cell burners

                                      2-22

-------
illustrated in Figure 1 were arranged in a nonuniform firing pattern in the
furnace.

Retrofit of the FWEC CF/SF low-NOx burners (Figure  2)  required  major design
modifications to the Unit 4 boiler including
   •    Conversion to eight pulverizers and 48 low-NOx burners, arranged in
        four rows of six burners on each firing wall
   •    New lower furnace waterwall panels designed for a conventional,
        widened burner spacing
   •    Replacement of most of the burner piping
   •    Installation of a new pulverizer/burner control system

These construction modifications were completed during a major two-month
overhaul of Unit 4 in the spring of 1989.

TEST PROGRAM DESCRIPTION
Following installation of the low-NOx burners,  APS  and the Electric Power
Research Institute (EPRI) entered into a cooperative agreement to test and
compare the modified boiler's performance and emissions with the performance
and emissions of unmodified Unit 5.

Specific objectives of the testing program were
   •    To assess any changes in Unit 4 boiler performance and operability
        with the new low-NOx burners
   •    To investigate the effects of inner and outer air register positions
        and burner inner nozzle adjustments on flame shape and stability, NOX
        emissions, and boiler absorption rates, particularly in the secondary
        superheater and pendant reheater sections
   •    To evaluate the effects of different unit loads and furnace excess
        oxygen (02)  levels on NOX  emissions

Tenerx Corporation was hired to collect emissions data on NOX,  excess 02,
carbon monoxide (CO), and sulfur dioxide (S02),  to  measure gas  temperatures
leaving the economizer, and to collect and analyze  coal and ash samples.  APS
                                     2-23

-------
engineers collected all  flow,  pressure,  and temperature data needed to
evaluate boiler absorption performance.   APS and Tenerx conducted a site
inspection of the boilers prior to the testing program to establish acceptable
sampling locations and testing procedures.  Coal fineness and fuel/air balance
testing were performed on the  fuel supply systems of both units prior to the
main testing program to ensure acceptable boiler test conditions.  All
permanent plant instrumentation used in  the testing was checked for
calibration and recalibrated where necessary.

Gaseous emissions and flue gas temperatures were measured in the flue gas
ducts between the economizer outlet and  the air preheater inlets.  Gaseous
emissions were collected from an 18-point grid in Unit 4 and an 8-point grid
in Unit 5.  Gas temperatures were measured from a 27-point thermocouple grid
in Unit 4 and a 24-point thermocouple grid in Unit 5.  A computer-based data
acquisition system was used to collect all thermocouple readings.

TESTING PROCEDURE
The comparative testing program followed the test matrix shown in Table 2 to
evaluate the units' emissions and thermal performance over a range of
operating conditions.  The test plan consisted of a series of 15 parallel
tests on Units 4 and 5, along with six tests conducted only on Unit 4.  Test
variables included unit load,  furnace excess 02 level,  burner tip position,
and inner/outer air register position.  The following tests were conducted:
    •     Four full-load parallel tests were run on both units at standard
         burner tip position of +3 inches at low  (1.8-2.4%), normal (2.7-
         2.9%), and high  (3.4-3.6%) excess 02 levels.
    •     Four full-load, parallel tests were run on both units with burner tip
         positions moved to zero and -3 inches at low (1.8-2.4%) and high
         (3.4-3.6%) excess 02 levels.
    •     Four full-load tests were run on Unit 4 only while varying inner and
         outer air  register positions.  These tests were performed at  normal
         excess 02 levels (2.7-3.0%).   Optimum inner and outer air register
         positions were identified based on NOX emission levels.  Two
         additional full-load tests were then run on Unit 4 with both  air
         registers at optimized positions at low  (2.0%) and normal (2.7%)
         excess 02 levels.
                                      2-24

-------
   •    Four 75% load, parallel tests were run on both units under standard
        burner operating conditions with all mills in service (AMIS) and with
        top mill out of service (MOOS) at normal excess 02 levels (3.2-3.7%);
        and with AMIS at high excess 02 levels (4.4-4.5%).
   •    Three 50% load, parallel tests were run on both units under standard
        burner operating conditions with two top MOOS at normal  excess 02
        levels (4.7-5.0%), and then with two top MOOS 02 (5.4-5.5%).

Each test lasted about 4-5 hours —1.5 hours of process stabilization,  and 3-4
hours of actual  testing.  Emissions were monitored and recorded  as single-
point samples and as composite samples.  Emissions testing equipment consisted
of a chemiluminescent NOX analyzer,  infrared analyzers for CO and C02,  a
zirconia cell analyzer for 02,  and a DuPont S02  analyzer.

Two fuel analyses were performed during each test.  Coal samples were
collected immediately downstream of the coal silos before the coal entered
each mill feeder.  These samples were riffled together to produce an "average"
coal sample, and higher heating value (HHV), proximate, and ultimate analyses
were performed.   Mineral analyses were also conducted on selected coal
samples.

Bottom ash samples were collected once per test from a selected  bottom ash
hopper.  Fly ash samples were collected from two selected baghouse hoppers and
one economizer hopper for each unit.  Samples were analyzed for  mineral
constituents, fusion temperature, and carbon carryover (loss on  ignition,
LOI).  Size, quantity, and elemental analyses were performed on  selected
bottom ash and baghouse fly ash samples.

At the end of Test No. 1 a severe leak in the first point high-pressure
feedwater heater of Unit 5 occurred, and the heater had to be valved out of
service.  Testing revealed that NOX emissions from Unit 4 were approximately
the same with this first point heater in or out of service.  Based on  this,
the Unit 4 first point heater was also valved out of set/ice for the remaining
parallel tests to allow a fair performance comparison between Units 4  and 5.
During the six tests on only Unit 4, both first point heaters were in  service.
                                     2-25

-------
EMISSION TEST RESULTS
Because the secondary  air  supply  to  the  burners  out  of service on Unit 5 could
not be shut off,  a staging effect was  present  that could  explain the lower NOX
emissions from Unit 5  under low-load conditions.   Figure  3  illustrates NOX
emissions versus  stoichiometric  air  ratio  to correct for  the staging effect.

The FWEC low-NOx  burners installed on  Unit 4 achieved  an  average 50% reduction
in NOX emissions  versus  those  from the unmodified Unit 5  when  operating at
full load.  Under 75%  load and normal  (3.4-3.7%)  excess 02  levels,  the
reduction in NOX  emissions was 47% with  AMIS,  and 40%  with  the top MOOS.  With
the two top MOOS  at 50% load conditions, the  staging effect on Unit 5 NOX
emission reduction was so  obvious that only  an average 17%  NOX reduction was
observed on Unit  4.

Figure 4 illustrates NOX emissions versus  unit load  at various operating
excess 02 levels.  A correlation  analysis  indicated  a  definite correlation
between NOX emissions  and  unit load  for  Unit 5,  while  the Unit 4 data did not
show significant  correlation.   Reducing  unit  load can only  reduce thermal NOX,
which is a small  portion of total NOX  emissions.  Because the  turbulent
burners on Unit 5 were operated at higher peak flame temperatures than the
low-NOx burners on Unit 4, Unit  5 produced more  thermal NOX than Unit 4,  and
was more sensitive to  unit load changes.

Figure 5 indicates that changing the burner  tip position  had little effect on
NOX emissions. When the burner  tip  position is  adjusted, the  primary air
velocity is changed because primary  airflow  is constant.   Adjustments are used
to optimize the primary air/secondary  air ratio to  minimize shear-induced
turbulence.  They may  also cause major changes in flame shape.  APS had
previously identified  the  optimum burner tip position as  +3 inches.

The effects of inner and outev air register  position on the performance of  the
low-NOx burners in Unit 4  are  illustrated  in  Figure  6.  Inner air registers
regulate the amount of swirl in the  secondary  air near the  burner tip and
control the point of flame ignition.  Outer  air registers  impart initial  swirl
                                     2-26

-------
to the secondary air and control the overall flame shape and size/strength of
the internal recirculation zone.  Minimum NOX emissions levels  occurred at an
inner air register position of 10' open and an outer air register position of
35-40* open.  With both inner and outer air registers in their optimum
positions, NOX emission levels were 0.44 Ib/MBtu at normal  excess 02  level  and
0.42 Ib/MBtu at low excess 02 level.

Figure 7 illustrates the NOX emissions versus burner zone liberation  rate.
The effect of staging on Unit 5 NOX emissions fs obvious.

S02 flue gas values ranged from 605 to 914 ppm for Unit 4,  and  546 ppm to  761
ppm for Unit 5.  CO emissions on Unit 5 ranged from 32 to 75 ppm, while Unit 4
CO emissions ranged from 32 to 75 ppm except on one test with low excess 02,
where average CO emissions were 185 ppm.

Analyses of coal and ash samples taken during the testing program revealed the
consistency of the coal fired in Units 4 and 5.  The coal is fairly reactive,
so there was little difference in the unburned carbon levels between  Units 4
and 5.

BOILER PERFORMANCE TEST RESULTS
Based on preliminary analyses of boiler performance data, it appears  that
there was no detrimental effect on boiler performance, efficiency, or
operability as a result of the installation of low-NOx burners  on Unit 4.
Table 3 presents the results of a typical, full-load, boiler performance test.
Further analysis is required to explain the substantial differences between
some of the comparative data.  APS plans to conduct boiler performance and
emissions testing on Unit 5 after the installation of low-NOx burners.  This
will provide additional data for a better comparison of the boiler performance
before and after the burner retrofit.

Furnace Exit Gas Temperature (FEGT)
The FEGTs at full-load operations calculated using the back-calculation method
ranged from 2541 to 2680'F for Unit 4, and 2647 to 2850'F for Unit 5.  The
                                     2-27

-------
difference in FEGT for Unit  4,  which  ranges  from 100  to 209'F lower than Unit
5 FEGT,  is due to increased  furnace  heat  absorption  as  a result of reduced
levels of slagging in the furnace.   The low-NOx  burners control  high-
temperature flame regions which promote slagging.

Heat Absorption Rates in Boiler Circuits
Changes  in the burner firing arrangement  and the retrofit of low-NOx burners
have created a new furnace heat absorption  pattern.   Figure 8 illustrates a
typical  comparison of heat absorption rates  in  the  boiler circuits for Units  4
and 5.

Only enthalpies for the boiler circuits are  shown  in  the figure because the
water/steam rates for Units  4 and 5  are almost  identical.  Unit 4 data shows
an increase of 31-66% in the upper furnace  heat  absorption rate compared with
Unit 5.   This is due to reduced levels of slagging  in the furnace.  A  decrease
of 33-43% in the primary superheater heat absorption  rate is also indicated in
the Unit 4 data.  The heat absorption rates  for  the  upper furnace and  primary
superheater are being investigated further  to find  out  why there are
substantial differences between the  units.   The  much  higher upper furnace heat
absorption rate increases the primary superheater  outlet steam temperature
such that the superheater spray flow requirement for  Unit 4 is increased by
148-180% when compared with  Unit 5.   Units  4 and 5  data show insignificant
changes  in the heat absorption rates  in other boiler  circuits such as  the
lower furnace, secondary superheater,  superheater  enclosure, reheater, and
economizer

Main Steam and Hot Reheat Temperatures
Data indicate a slight increase in the main  steam  temperature for Unit 4.
Main steam temperatures are  in the range  of  990-1004'F  for Unit 4 versus 983-
992'F for Unit 5.  However,  there is  a substantial  decrease in the hot reheat
temperature for Unit 4.  Hot reheat  temperatures are  in the range of 947-
977'F for Unit 4 and 975-101TF for  Unit  5.   This  is  due to the pendant
reheater inlet gas temperature for Unit 4,  which is  124-177'F lower than the
temperature for Unit 5.  The effect  of lower hot reheat temperature for Unit  4
on turbine cycle efficiency  is being  investigated.
                                     2-28

-------
Boiler Efficiencies
Boiler efficiency increased slightly, in the range of 0.72-1.52%, for Unit 4.
This is due to lower air preheater inlet gas temperatures as a result of lower
economizer inlet water temperatures.  During the only test that Units 4 and 5
had with both first point high-pressure feedwater heaters in service (Test
No. 1), the air preheater outlet gas temperatures were almost the same.
Therefore, it can be concluded that the boiler thermal efficiencies remained
the same after the retrofit.

CONCLUSION
The boiler emissions and thermal performance testing program comparing the
performance of Units 4 and 5 at the Four Corners Steam Electric Station
revealed that the retrofit of low-NOx burners to Unit 4 reduced NOX  emissions
by about 50% at normal, full-load operating conditions without any detrimental
effect on boiler performance, efficiency, or operability.

The average level of NOX emissions from Unit 4 was 0.53 Ib/MBtu,  well  under
the applicable state of New Mexico air-quality standard of 0.70 Ib/MBtu.

ACKNOWLEDGMENTS
Special thanks to Four Corners operations, maintenance, and engineering
personnel for their assistance in conducting the testing, to Charles Allen for
his technical assistance, and to Paul Thompson of Tenerx for conducting the
testing.

REFERENCE
1.    Vatsky, Joel, and Allen, Charles, "Predicting Boiler and Emissions
      Performance by Comparative Turbulent/Low-N0x Burner Testing on a  Large
      Testing Facility."  Proc. 1989 Joint Symposium on Stationary Combustion
      NOv Control.  2,  23 (1989).
                                     2-29

-------
              WINOBOX
                                    NORMAL FIRING POSITION
                - COAL AND
                 PRIMARY AIR
                        COAL NOZZLE
                           Figure 1
            FWEC Three - Nozzle Cell Burner
                     MOVABLE SLEEVE
                       DAMPER
                                  SPLIT FLAME COAL NOZZLE
                          Figure 2
FWEC Controlled - Flow/Split - Flame Low - NO Burner
                            2-30

-------
NOx, LB/MBTU
1.3
                  Figure 3
    EFFECT OF STOICHIOMETRY ON
             NOX EMISSIONS
1.2

1.1

1

0.9

0.8

0.7

0.6

0.5

0.4
 0.9
              1.1      1.2     1.3
              STOICHIOMETRIC AIR RATIO
                                  1.4
1.5
                 Figure 4
     EFFECT OF BOILER LOAD ON
            NOx EMISSIONS
NOX
1.4
1.2
1
0.8
0.6
0.4
0.2
n
, LB/MBTU n


__- 	 •
a
¥





— 	 —












+ __
-f '
A
M
O



__ > —






	 —






+d
— •






m 	
, 	
H3
k* 	
•
O
	




-e- UNIT*
-f UNIT 5
N UNIT 4-HIGH O2
D UNIT 5-HIGH 02
X UNIT *-LOW 02
• UNIT 5-LOW O2
A UNIT 5-STAOINQ

400   450   500   550   600  650
             LOAD, MW (N.tl
                         700   750
                                  800
                    2-31

-------
                   Figure 5
EFFECT OF BURNER NOZZLE POSITION ON
              NOX EMISSIONS
 FULL LOAD NO*, LB/MBTU
1.4
1.3
1.2
1.1
1
0.9
0.8
0.7
0.6
0.5
0.4
n •>


-e- LOWER EXCESS 02
-i- HIGHER EXCESS 02
n NORMAL EXCESS 02
























; C










}





















t










t

   -4    -3-2-10     1    2
               BURNER TIP POSITION, INCHES
                   Figure 6
  EFFECT OF AIR REGISTER POSITION ON
              NOX EMISSIONS
  FULL LOAD NO*, LB/MBTU
1.3-
1.1 -
•)
0.9-
0.8-
0.7 -
0 6
0.5 -
0.4 -
0 3 -





-e- INNER AIR REGISTER
-t- OUTER AIR REGISTER
* INNER REG w/LOWER O2
n OUTER REG w/LOWER O2


1 	
(



'^— -E


L ^^
1 	 ""







f-iccrrTTTTr'
fi









> 	 -o
	 -,_-^
, — -^ — -^-t









r^
























        10    20    30    40    50    80

               AIR REGISTER POSITION, % OPEN
70
     80
                     2-32

-------
                 Figure 7
 EFFECT OF BURNER ZONE LIBERATION
        RATE ON NOX EMISSIONS
  NOx, LB/MBTU
1.2-
1 -
0.8-
0 6
0.4-
0.2-
0 -


o UNIT 4
+ UNIT 5
« UNITS-STAGING


















>
a
o










+
n
o
o



+

o
ef>
e
S>
o









      50   100   150   200   250   300   350
         BURNER ZONE UBERATION RATE, KBTU/HR-FT2
400
ENTHALPY (BTU/LB)
1500
                  Figure 8
      COMPARISON OF SECTIONAL
           HEAT ABSORPTION
                   2-33

-------
                                  Table 1

                           COAL ANALYSES SUMMARY
       Test
Proximate Analysis

     % Moisture
     % Ash

     % Volatiles

     % Fixed Carbon

     Energy Content, Btu/lb

     % Sulfur

     MAP, Btu/lb
     % Air Dry Loss
 Condition
As Received
Dry Basis
As Received
Dry Basis
As Received
Dry Basis
As Received
Dry Basis
As Received
Dry Basis
  Unit 4
   14.67
   16.33
   19.14
   26.47
   31.02
   42.53
   49.84
 9561
11205
    0.85
    1.00
13857
    9.03
                                                               Unit  5
   14.34
   16.32
   19.05
   31.32
   36.88
   37.75
   44.07
 9602
11210
    0.72
    0.84
13848
    7.57
Ultimate Analysis

     % Moisture
     % Carbon

     % Hydrogen

     % Nitrogen

     % Sulfur

     % Ash

     % Oxygen
As Received
Dry Basis
As Received
Dry Basis
As Received
Dry Basis
As Received
Dry Basis
As Received
Dry Basis
As Received
Dry Basis
   14.67
   53.87
   63.13
    3.86
    4.52
      10
      29
    0.85
    1.00
   16.33
   19.14
    9.32
   10.92
   14.34
   54.55
   63.68
      96
      62
      15
      34
    0.72
    0.84
   16.32
   19.05
    8.96
   10.47
                                   2-34

-------
Table 2
TEST MATRIX
Test Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Load
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
75%
75%
75%
75%
50%
50%
50%
50%
50%
Units
4&5
4&5
4&5
4&5
4
4
4
4
4&5
4&5
4&5
4&5
4&5
4&5
4&5
4&5
4&5
4&5
4&5
4&5
4&5
Excess Air
Normal
Normal
Low
Normal
Normal
Normal
Normal
Normal
Low
High
Low
High
Normal
High
Normal
Normal
Normal
High
Normal
Normal
Normal
Test Condition
Std. Nozzle & Register Positions
Std. Nozzle & Register Positions
Std. Nozzle & Register Positions
Repeat Std.
Outer Register Open
Outer Register Closed
Inner Register Open
Inner Register Closed
Burner Nozzle In (0)
Burner Nozzle In (0)
Burner Nozzle In (-3)
Burner Nozzle In (-3)
Std., All Mills In Service
Std., All Mills In Service
Std., 1 Mill Out of Service
Std.
Std., 2 Mills Out of Service
Std., 2 Mills Out of Service
Repeat Std.
Optimal Nozzle Register Positions
Optimal Nozzle Register Positions
2-35

-------
                             Table 3
            TYPICAL  COMPARATIVE  BOILER  PERFORMANCE DATA

      Parameter                            Unit           Unit
Load, MW                                  760            752
Excess 02,  % Wet                             2.82           2.74
Feedwater Flow, 1000 Ib/h                5478           5493
Superheater Spray Flow, 1000 Ib/h         433            286
Main Steam Temp., "F                     1004            992
Hot Reheat Temp., 'F                      965            999
Furnace Exit Gas  Temp.,'F                2593           2775
NOX,  Ib/MBtu                                0.49           1.15
                               2-36

-------
DESIGN AND APPLICATION RESULTS OF A NEW EUROPEAN LOW-NOX BURNER

                             J. Pedersen
                       Burmeister & Wain Energy
                           23 Teknikerbyen
                       DK-2830 Virum, Denmark

                               M. Berg
                       ELKRAFT Power Company
                            5 Lautruphoej
                      DK-2750 Ballerup, Denmark

-------
                      DESIGN AND APPLICATION RESULTS
                   OF A NEW EUROPEAN LOW-NOX BURNER

                                  J.  Pedersen
                            Burmeister & Wain Energy
                                23 Teknikerbyen
                            DK-2830  Virum, Denmark

                                    M. Berg
                            ELKRAFT Power Company
                                  5 Lautruphoej
                            DK-2750 Ballerup, Denmark
ABSTRACT

To meet the new NOX regulations in Denmark, the ELKRAFT Power Company decided in
1987 to retrofit existing coal-fired  units with low-NOx burners. In cooperation with the boiler
company Burmeister & Wain Energy, a new burner was developed to meet the specifications
of the first boiler to be retrofitted, and was tested at full scale in 1988 in an experimental
facility in the USA. In 1989 the front  wall-fired Asnaes Unit 4 (285 MWe) was retrofitted with
24 coal/oil low-NOx burners of the new design, each with individual control of the combustion-
air flows. Daily operating experience, and specific testing with a range of world coals, has
demonstrated  greater than 50% reduction in NOX emissions, whilst maintaining carbon in  fly-
ash levels below 5%. Due to the stable and well-attached nature of the flames, the control
range on coal-firing has also been considerably extended. Evidence of flame impingement
on furnace walls, slag  deposition  or corrosion has not been observed.  Following this
successful program several additional boilers will be retrofitted with the new low-NOx burners
in the near future.
                                      2-39

-------
                     DESIGN AND APPLICATION RESULTS
                   OF A NEW EUROPEAN LOW-NOX BURNER
INTRODUCTION

In Denmark, more than 90% of the power is generated from coal, and having no domestic
coal resources, the coals are imported from all over the world.

To meet the NOx-regulations, the existing power stations in Denmark have to reduce the total
amount of nitrogen oxide emission per year gradually until the year 2005 in which a 50%
reduction, in relation to the 1980-level, will be demanded. New units will be further restricted
in their NOx-emission.

The ELKRAFT Power Company, covering the eastern part of Denmark and generating about
half of the  Danish power supply, decided in 1987 to retrofit a number of units with low-NOx
burners. The Asnaes Power Station Unit 4 was the first unit to be retrofitted in 1989.

In the light of the low-NOx burner technology existing at that time, it was decided to develop
a new low-NOx burner for Asnaes Unit 4. The development program was sponsored by the
ELKRAFT  Power Company  and  carried  out in cooperation with the boiler and burner
manufacturer Burmeister & Wain Energy and the Asnaes Power Station.
ASNAES UNIT 4

Asnaes Unit 4, with a capacity of 270 MWe net, was commissioned as an oil-fired unit in
1968 and was converted to coal/oil-firing in 1978.

The boiler is of the Benson type and is front wall-fired. It is a two-pass system with reheater
and dry bottom furnace. The boiler has no flue gas recirculation system, and is designed for
full load when firing heavy fuel oil or bituminous coal. The coals used are imported from
countries around the world. The design coal is high volatile Polish coal. Due to the fly ash
                                      2-40

-------
utilization in the cement industry, the unburnt carbon is limited to a maximum of 5%. Boiler
dimensions are shown on Figure 1, and performance data are listed in Table  1.

The 24 burners are located on the front wall in 4 levels of 6 burners each. Each burner level
is supplied with pulverized coal from its respective coal mill of the ball-ring type. During
normal operation, three coal mills, 18 burners, will maintain full load of the unit. Combustion
air is controlled and supplied separately to each individual burner.
LOW-NOX BURNER DESIGN CRITERIA

In the design of the new  low-NOx burners, a number of design criteria were taken into
account. For application to Asnaes Unit 4, the number of burners and their location were to
be maintained. However, since the existing burner openings were designed for the original
unit operating on oil-firing only, it was decided that larger openings would be installed on the
boiler front wall. The size of the new burner openings  was,  however, restricted  by the
presence of vertical support tubes. Also, the fact that future boiler retrofits would need  to be
accomplished without modifications to the firing wall was taken into consideration in sizing
the new burners.

The  restrictions caused by the vertical support tubes were also one reason for the new
burners being developed to  achieve optimum NOX control at full  boiler load with all 24
burners in service. This corresponds to a nominal fuel input of 28.8  MW per burner. It was
required, however, that full load should still be attainable on 18 burners only, corresponding
to a fuel  input of 38.3 MW per burner, but without necessarily maintaining optimum NOX
performance.

When firing coal, the low-NOx burner should have a load  range from 38.3 MW down to 16
MW fuel input and should have a reliable flame scanner signal over the full load range.

The  layout of the new burner was also influenced by the decision not to change the coal mill
system. Furthermore, the burners should also be able to obtain full boiler load on heavy fuel
oil.

The goal for the burner development program was to achieve a NOx-emission at Asnaes with
full boiler load and 3% excess O2 at the economizer outlet, of less than 440 ppm (3% O2,
dry) with less than 5% unburned carbon in the fly ash. The pre-retrofit full load baseline  NOX-
emission  for Asnaes Unit 4 was 740 ppm (3% O2, dry).

Only low-NOx burners were taken into consideration for NOX reduction on Asnaes 4. Because
of limited residence time in  upper furnace and the possibility of low mixing efficiency, the use
of overfire air was out of the question.
                                        2-41

-------
BURNER DEVELOPMENT PROGRAM

In order to retrofit the Asnaes boiler with an optimum burner, a full-scale test burner of the
internal  staging type was  optimized in a "Coal Burner Test  Facility"  in  USA  in 1988,
simulating the same conditions as in Asnaes 4.

The test burner was designed with a high degree of flexibility and provided for a wide range
of variables in the  configuration  of combustion-air registers and in the design of the  coal
injector. The test burner design was based upon common design features employed  in
existing Burmeister & Wain coal burners, such as the axial movable turbolator for swirl
control.

The optimized low-NOx burner is based upon staged combustion with the flame attached to
a flame holder, mounted at the exit of the coal pipe. Typically, NOx-emissions will be reduced
when a detached flame (with flame stand-off) is changed to an attached flame, anchored by
the flame holder. The flame holder establishes local recirculation zones and promotes mixing
between coal and secondary air.  Secondary air swirl ensures a well-attached flame. Sample
data from  the burner test, shown on Figure 2, shows more than 50% NOx-reduction, when
changing burner settings.

The  burner design  is such that the tertiary air turbolator setting controls the flame shape.
With a flame length well below the firing depth of the Asnaes 4 furnace, the performance of
the optimized test burner at nominal load was 260 ppm  NOX (3% O2, dry) with less than 5%
carbon in the fly ash. The burner was also tested at high load, corresponding to three-mill
operation  at Asnaes 4, and at low load corresponding  to minimum load on the coal  mills.
Burner testing was conducted primarily with  U.S. (Pennsylvania) coal and Polish coal with
limited tests on  heavy fuel oil (no. 6). The development  program was terminated  with
operation of the optimized test burner to determine such  parameters as secondary air/tertiary
air-flowsplit and turbolator settings. Good flame scanning signals were demonstrated over
the full load range.
BOILER RETROFITTING

At the end of 1989 Asnaes Unit 4 was brought back into operation with 24 of the low-NOx
burners installed in new boiler wall openings.

The new burner design  is based on the recommendations developed in  the burner test
program, and has been designated as the "BWE Type 4 AF" low-NOx burner. The burner
design is illustrated schematically in Figure 3, and shows the division of the combustion air
into secondary air and tertiary air streams, each with an axial movable turbulator for swirl
control.

In forward  position, the  axial turbolator  creates a  maximum swirl of the airflow.  While
retracted, a certain part of the airflow bypasses the turbolator vanes and thus weakens the
swirl  of the  total flow. Both  turbolators in the burner are provided with drives and are,
independently of each other,  adjusted automatically or from the control room.
                                       2-42

-------
A "LAND" flame monitor of the cross-correlation type is mounted on each burner for detection
of both coal flames and fuel oil flames.

The burners are provided with oil lances with steam atomization and high voltage ignition
lances.

The 24 combustion air ducts were modified during the burner retrofitting. The duct section
after each combustion air venturi was changed to include secondary and tertiary air ducts,
each with a control damper.

The secondary and tertiary air flow to each burner is measured by the common venturi. The
secondary  air  flow  is measured downstream and  separately by  an annubar  probe;
subsequently, the tertiary air flow is calculated. The secondary  and tertiary airflow, as well
as the so-called SA/TA-flowsplit, is controlled automatically by the two dampers. The coal
and air supply to a given burner level is illustated on Figure 4.

A small part of the secondary air is supplied to the core air pipe as cooling and sealing air
around the  oil lance. The core airflow to each burner was adjusted during the commissioning
by a manually operated damper and a pitot tube.

Due to the  high degree of automation on Asnaes Unit  4, and because of the knowledge of
optimum burner settings obtained in the test program, the new burners were commissioned
in a very short period of  time. In fact, only one half day was required to set and to verify
satisfactory operation of all  24 burners.

With coal-firing, the ideal secondary air tubolator position is forward for generation of high
swirl.  At high burner  loads, the tertiary airturbolator position is retracted in order to reduce
swirl,  whilst at low burner loads, the tertiary air turbolator  position is forward. For fuel oil-
firing, both  turbolator settings are  retracted for swirl reduction.

During boiler operation on coal, the six burners in one burner level are operated with identical
burner settings. All burner levels in service are operated similarly. During daily operation, the
burners are operated automatically over the full load range  for coal-firing, as well as for fuel
oil and combined coal/oil-firing.
FIELD TESTS

During the first three months of 1990, the burners were tested with a range of different coal
types,  i.e. Polish, U.S., W.  Canadian and Colombian coals representing a volatile content
from 20.2% to 32.9% (as received) and an ash content from 6.8% to 18.6%. Typical coal
analyses are presented in Table  2.
                                         2-43

-------
For the field test program at Asnaes 4, additional measurement and sampling equipment was
used. NO, SO2, O2, CO, CO2 and H2O in-situ instruments were installed at the two induced
draught fan outlets. Pulverized coal was sampled from each of the coal pipes with a rotary
multiprobe sampler. Coal samples were taken from the coal feeders during operation. Fly ash
was sampled from the two vertical flue gas ducts before the ESP, which  is located on the
boiler house ceiling.  Suction  pyrometers  were also installed  for measurement  of gas
temperature and O2% in upper furnace.

All gas emission data, the most  important boiler data, firing system data and burner data
were sampled and averaged with an on-line data logging processor.

During the burner test program, the NOx-emission dependence on excess O2 and unit load
was  recorded for 4 and 3 mill operation respectively. Also the dependence on turbolator
settings and SA/TA flowsplit was tested.
RESULTS

The main field test was carried out on Polish coal, since this coal type was the design coal
for Asnaes 4 and was used during the burner development program.

The following  results summarize the tests with Polish coal, at full boiler load and are shown
in Figure 5:

0    4-mill operation: NOX = 370 ppm  (3% O2, dry) and < 5% carbon in
     fly ash at 3.0% O2 (dry), at the economizer outlet.

•    3-mill operation: NOX = 410 ppm  (3% O2, dry) and < 5% carbon in
     fly ash at 3.5% O2 (dry), at the economizer outlet.

These actual NOx-emissions on Asnaes 4 are close to the NOx-emissions estimated from the
testburner data, considering differences in the burner zone heat release rate.

With 3  mill operation, a cooling air flow through the burners out of service amounts to an
airflow at full boiler load, corresponding to  0.5% excess O2, at economizer outlet.

The distribution of pulverized coal to the  burners, at full  boiler load was measured to a
deviation from the average of 10-15%.  For individual burners, however, the deviation could
be up to 25%.

The average fineness of the coal at full boiler load for Polish coal was found to be:

•    4-mill operation: 20% > 90 microns  (170 mesh) and
                    0.9% > 250 microns (60 mesh)

•    3-mill operation: 26% > 90 microns  (170 mesh) and
                    0.5% > 250 microns (60 mesh)
                                       2-44

-------
Load was found to have only a moderate influence on the NOx-emission as shown in Figure
6. These results represent data with all burners in service and constant excess air at the
economizer outlet.

The secondary/tertiary air split was also tested over a nominal control  range for full boiler
load and 3 and 4 coal mills in operation. As shown in  Figure 7, this parameter has only a
small impact, indicating a satisfactory selection of optimum design parameters.

The tertiary airturbolator position has also a negligible effect on the NOx-formation, as shown
in Figure 8, but has a marked impact on flame shape.

Burner testing with the other coal types (U.S., W. Canada, Colombia) shows similar trends
as for Polish coal, but with minor variations in the  absolute NOX levels,  as shown  in Figure
9.
OPERATING EXPERIENCE

Reliable ignition of the heavy fuel oil flames with the high voltage igniter has been proven
with the new low-NOx burner. Also,  when firing coals within the normal range of control,
stable ignition characteristics have been demonstrated with attached and well-defined flames.
With normal burner settings,  flame lengths  are  less  than 9 meters, and  are easily
accommodated into the furnace without impingement on the rear wall.

Acceptable performance on Polish, US, Canadian, and Colombian coals, in addition to a
number of mixed coals types, has been demonstrated since the initial  commissioning of the
burners.

The NOX reduction  level achieved on this unit is demonstrated by a comparison of daily
averages of NOX emissions during routine operation for periods before and after the burner
retrofit. This comparison is shown in Figure 10. For the period January  to March 1989, which
was prior to the retrofit, the NOX emissions averaged 740 ppm (3% O2, dry). With the new
low-NOx burners, and over the same period in 1990, NOx emissions averaged 345 ppm (3%
O2, dry). This corresponds to an average NOX reduction of greater than 50%.

Daily average NOX emissions before and after the retrofit are further compared in Figure 11.

One further important consequence of installing the new low-NOx burners has been the ability
to reduce the minimum unit load with coal-firing from 180 MWe net, to  130  MWe net. The
control range of the unit before and after the burner retrofit is compared in Figure 12, where
the extended range is a direct result of improved flame stability with the new burner design.
This represents a potential  for considerable savings in heavy fuel oil consumption during
start-up and low load operation. Minimum coal load on the unit is now obtained with two coal
mills in operation, and is limited only by the available primary air temperature for the coal
drying process.
                                        2-45

-------
Since the commissioning of the  new low-NOx burners, the boiler has been operated with
individual flame monitoring  with no problems and maintaining strong scanner signals.

For the types of coals used, no slag deposition has been found on the furnace walls or in the
burnerthroats. Soot blowing once every 24 hours has been sufficient. On one occasion when
firing U.S. coals, the boiler was operated for 60 h without  soot blowing, with no problems at
all. The consistency of the bottom  ash  is very porous compared with the pre-retrofitting
bottom ash.

The boiler operation has therefore been significantly improved  by retrofitting the low-NOx
burners. Boiler efficiency and capacity have, however, not  been reduced compared with
conditions before the burner retrofit.
CONCLUSION

The results of more than 1 year of operation with the new low-NOx burner at Asnaes Unit 4
can be summarized as follows:

•    NOx-emission reduced by 50%

•    Unburned carbon in ash  less than 5%

•    Stable flames at all boiler loads

•    No flame impingement on furnace walls

•    No slag deposition  in the furnace

•    Increased control range on coal-firing

•    Excellent flame  scanner signals

•    Good correlation between test-burner and field-burner results
                                       2-46

-------
FURTHER RETROFITTING PROGRAMS

Following the positive experience with Asnaes Unit 4 in 1989, further ELKRAFT units will be
retrofitted with the Burmeister & Wain, type 4AF Low-NOx burners:

1991:     Amager Unit 1, 140 MWe, 12 burners, front wall fired drum boiler


1991:     Asnaes Unit 2, 155 MWe, 12 burners, opposed fired drum boiler


1992:     Amager Unit 2, 140 MWe, 12 burners, front wall fired drum boiler


1992:     Asnaes Unit 5, 725 MWe, 48 burners, opposed fired Benson boiler.

These retrofit programs will be based on the burner designs and experience developed in
Asnaes Unit 4, and will be achieved without changes to the existing burner openings, or
major modifications to the existing firing systems.


ACKNOWLEDGEMENTS

The authors wish to thank the following organizations for generously providing equipment and
services.  The authors  also wish to thank the  personnel for all its efforts and services
rendered to make this project possible:

     Energy and Environmental Research Corporation

     Riley Research Center

     Land Combustion
                                       2-47

-------
                                           41.75
         Figure 1.   Asnaes 4 boiler dimensions.
                            TEST  BURNER
NOx
ppm
400
300
200
100
"5
<
-c
u

-i
I
-i
i!
$f •
- \ ™
•-]•-
:'^
— £
•--!
r










*
si
i-
r
-








.
t:
-
r^
-
7
-


I
-T*
-
_"
^1
_ f
-t'
4'
-I
^-^
~4
-t -
:-L
-r
t
- i


^



J1
L


\
:




_



^-


_
^
.
-
""
t



-

A-
—

—
-
-,-
—
-





--

^-,
U-
•r.
;
•v
.




::


-

-
-_
-
f











-__
-
r







/
-

r
n
- _
-
-
-






-
-











s
—

-
"_
-

-

-








^




"



fW

-
--

A
1
--
-_

-



^

-


v/







-


-

\-
w

V
'-
-






r

-
A
-


-"-








U
r*
-
-
--
-"
-




V



[_*
-
t

-:-
-

1
pt.
n

•^



L-
_
-

—
v





>«.

*•
^
•
-
H-



i

-


•v



w^
;-_
—








-


^-
-;
-

--
T




v»

L
A.


—







•s.


v


--
-
^
77




I/

/
T





:
5









ATTACHED FlArtE


DIRECTION OF CHART SPEED
ml n .
      Figure 2.    NOX reduction by flame attachment.
                          2-48

-------
                          TERTIARY
                            AIR
             IGNITER
                     FOR SUIRL CONTROL
           L-FLAME  HOLDER
   PRIMARY AIR
       +
PULVERIZED COAL
           Figure 3.   "BWE type 4 AF" attached flame low-NOx burner.
                            ASNAES  POWER  STATION  UNIT  NO. 4
                                 COAL AND AIR SUPPLY TO BURNERS
                                                      ii
TERTIARY
  AIR
    65X
  PRIMARY AIR
      +
PULVERIZED COAL
                                                  COAL MILL
              Figure 4.   Coal and combustion air supply to a burner level.
                                        2-49

-------
                       ASNAES POWER  STATION UNIT  NO. 4
                               FULL LOAD.  POLISH COAL
  NOX ppm
(37. 02 .DRY)

       500 -
       400 -
       300
        ?00
        100
             2.5
  NOX ppm
(3% Oj .DRY)

       500
              400
              300
              200 -
              100 •
                oH
                                                       CARBON IN ASH
                    •   4 MILLS
                    A   3 MILLS
                                                                    -7
                                                                    -6
                                                                    -5
                                                                    -4
                                                                    -3
                                                                    -2
                                                                    - t
                                                                     0
                              3.0
                                               3.5
                                                        4.0   X 0,  (DRY)
             Figure 5.    Relation between NOx-emission and excess O2%.
                           ASNAES  POWER  STATION  UNIT NO.  4
                               POLISH COAL. 4 MILLS IN SERVICE
                                                       CARBON  IN ASH
                                                             -10
                                                             -9
                                                             -6
                                                             -7
                                                             -6
                                                             -5
                                                             -4
                                                             -3
                                                             -2
                                                             - 1
                      160     160    200    220    240    250
                                                              MU
                                                               e .net
               Figure 6.   Relation between NO -emission and unit load.
                                         2-50

-------
                       ASNAES POWER STATION UNIT  NO. 4
                               FULL LOAD. POLISH COAL
 NOx ppm
C3X 02 .DRY)
       500 -
       400 •
       300 •
       200 •
       100 -
•   4 HILLS
A   3 MILLS
                          SECONDARY A!R / TERTIARY AIR   FLOUSPLIT
           Figure 7.    Relation between NO-emission and SA/TA-flowsplit.
                       ASNAES POWER  STATION UNIT  NO.  4
                               FULL LOAD.  POLISH COAL
NOX ppm
C3X 02 .DRY)
500 -
400 -
300 -
200 -
100 -
0 -

A-=- 	 * 	 L
•- 	 • 	 A

• 4 MILLS
A 3 MILLS
60 70 80 90 100 %
Figure 8. Relation between NOx-emission and TA-turbolator positi
                                         2-51

-------
                      ASNAES POUER STATION UNIT NO.  4
                           FULL  LOAD. 4 MILLS IN SERVICE
NOX  pp
      500 -
      400
      300  -
      200 -
      100 -
                          O    CANADIAN COAL
                          O    COLUMBIAN COAL
                          D    U.S. COAL
                          •    POLISH COAL
                 2.5
                                  3.0
                                                    3,5
                                                          4.0  % 02 [DRY)
                                     ECONOMIZER OUTLET
                   Figure 9.    Relation between NOx-emission and
                   excess O2% for different coal types.
                         ASNAES POUER STATION  UNIT NO.  4
                         COMPARISON OF DAILY AVERAGE  NOx EMISSIONS
mg/MJ  (LHV)
 500
                                    NOx EHISSION
           500
           400
           300
           200
           too
Ib/mmBtu  (HHV)
       I,
        _ fy.   *  i   y   • •
      •rt    •  ••_>*•        •
      •  *    ," .  V   •  «*L  ^
                                                      OLD BURNERS
                                                      JANUARY-MARCH 1989
                                                      NEU BURNERS
                                                      JANUARY-MARCH 1990
                                                                          --  0.5
                                                                          -"-  0
                               EACH DOT REPRESENTS DAILY AVERAGE
           Figure  10.   Comparison of NO -emission with new and old burners.
                                        2-52

-------
N0y  ppm
(3%  02 .DRY)
 1000
  900 -
  600
  700 -

  600

  500 -

  400

  300 -

  200

  100 -
ASNAES  POWER  STATION  UNIT  NO.  4

       COMPARISON OF NOX EMISSIONS


             N0» EMISSION
               PRE
                           POST
             RETROFITTING ASNAES 4
                     EC-DIRECTIVE     U.S. CLEAN AIR ACT
                     NEW COAL-FIRED   DRY-BOTTOn.  UALL-FIRED
                     BOILERS >50 MU   (0.5 Ib/mmBtu)
             Figure 11.  Pre- and post-retrofitting data.
               ASNAES POWER STATION UNIT NO.  4

              COMPARISON OF  CONTROL RANGES ON COAL  FIRING
               MWe.nel

               300

               200

               250

               240

               220
                 PLANT LOAD
              200 •

              180
               150 -

               140
               120 •
               100
                           PRE
                                        POST
                         RETROFITTING ASNAES 4

                   Figure 12.   Unit control range.
                                 2-53

-------
                                  Table 1

                            PERFORMANCE DATA
Generator output

Net output

High-pressure steam, outlet



Reheat steam, outlet
                                                285 MWe

                                                270 MWD
                                                       c

                                                235 kg/sec
                                                190 bar
                                                545°C

                                                40.5 bar
                                                545°C
Coal Type
Ultimate, as received
                                   Table 2
                              COAL ANALYSES
           Polish
U.S.
Canadian   Colombian
Proximate, as
Moisture
Ash
Volatiles
Fixed carbon
received
o/
/o

-------
APPLICATION OF GAS REBURNING-SORBENT INJECTION TECHNOLOGY
           FOR CONTROL OF NOX AND S02  EMISSIONS
                        W. Bartok
                       B.A. Folsom
                       T.M. Sommer
                       J.C.  Opatrny
                        E.  Mecchia
                        R.T. Keen
      Energy and Environmental Research Corporation
                         18  Mason
                Irvine, California  92718
                         T.J.  May
                       M.S.  Krueger
                  Illinois  Power Company
                  500 South 27th Street
                 Decatur.  Illinois   62521

-------
              APPLICATION OF GAS  REBURNING-SORBENT INJECTION TECHNOLOGY
                        FOR CONTROL OF NOX AND S02  EMISSIONS
 ABSTRACT

 A  Clean  Coal  Technology project  is  being  carried out by Energy  and  Environmental
 Research  Corporation  (EER) to demonstrate Gas  Reburning-Sorbent  Injection  (GR-SI)
 technology for NOX and SO? emission control from coal  fired  utility boilers.   Phase
 I,  Design and Permitting, was completed in 1989 for three coal  fired utility  boiler
 host  sites  in II1inois--tangential ,  wall, and cyclone  fired  units.   The overall
 objectives of the program  are to  reduce NOX emission by 60%  and SOz emission  by  50%
 while  maintaining  or improving  operability  and not causing adverse  environmental
 impacts.   In view  of the  Clean  Air Act  Amendments  of 1990,  the niche for this
 technology appears to be relatively  small, older,  low  capacity  factor  units  firing
 coals  of medium to high sulfur content.

 This  paper describes the  design  and  installation  of the  GR-SI  and  ancillary
 equipment for a  71  MWe (net)  tangentially fired boiler (Illinois Power  Hennepin Unit
 No. 1), which burns 3.0 wt% sulfur Illinois coal.   The  detailed  design  of the GR-SI
 system was based  on  process specifications obtained through mixing, heat transfer
 and kinetic  modeling.   Four  sets of  four tangential  natural  gas injectors with
 recirculated  flue gas  as  carrier  have been  installed above the  existing coal
 burners for gas reburning, followed by four reburn  air ports in the upper furnace.
 Hydrated lime sorbent will be injected with transport air through  six  injectors at
 the elevation of the boiler nose, four on  the front wall and two  on side walls  (at
 low load the reburn air  ports will  be used for  sorbent injection).  Flue gas duct
 humidification has been  installed to  upgrade  the  performance of the electrostatic
 precipitator  (ESP) with sorbent injection.   The spent sorbent/fly  ash  mixture will
 be sluiced to  an  existing ash  pond after neutralization by CO?.

This  paper  compares  predicted   results  with  data   collected  during  initial
operations.
                                       2-57

-------
             APPLICATION  OF  GAS  REBURNING-SORBENT INJECTION TECHNOLOGY
                       FOR  CONTROL OF NOX AND S02  EMISSIONS
INTRODUCTION

Gas Reburning-Sorbent  Injection (GR-SI) n-12) is  the  combination  of  two developmental
technologies  being tested  by  Energy  and Environmental Research  (EER)  at Illinois
Power Company's Hennepin Power Station  Unit  1.   This project  is  part  of the U.S.
Department of Energy's Clean Coal  Technology  program.  Co-funding the project with
DOE  are  the  Gas  Research   Institute  and  the  Illinois  Department  of  Energy and
Natural  Resources.

The  basis of  the  GR-SI   technique  (shown  schematically  in  Figure  1)   is  the
introduction  of a calcium-based sorbent  and  natural  gas  into  the boiler to reduce
both sulfur dioxide  ($02) and  nitrogen oxide  (NOX) emissions.  The specific goal of
the project is  to demonstrate that S02  reductions of 50%  and NOX reductions of 60%
are  attainable on  an  economically feasible  basis.   Among the  advantages  of this
technology is  that  many coal fired  boi 1 er--especial ly  small,  older  ones — can be
retrofitted to  use  it  at a  reasonable cost.

 Illinois  Power's  Hennepin   Unit  1  is a  tangentially fired  pul verized-coal  boiler
 utilizing  3  wt% sulfur  Illinois bituminous coal.  As  part of  the  same  project.  GR-
 SI  technology  will  be tested on a cyclone-fired boiler  at  City Water.  Light,  and
 Power's Lakeside Unit 7 in  Springfield.  Illinois.

 The  GR-SI  process has  two  distinct steps.  The  first step,  gas  reburning,  modifies
 boiler  combustion by  firing only 80-85  percent  of the coal  fuel  in the  lower
 furnace.   The  remaining fuel requirement is provided by injecting  natural  gas  above
 the  primary  combustion zone.   This  gas  creates a  slightly  fuel-rich  "reburning
 zone" above  the  main  coal   flame zone.    The  gas reacts with NOX to form molecular
 nitrogen.  Finally,  overfire air is injected  into the upper  furnace  to  complete the
 combustion process without  generating additional  NOX.

 The  second  step  occurs above  the  reburning  zone  when  a  calcium-based  sorbent
 (hydrated  lime in this project)  is injected into the upper furnace.   The SO? in the
 furnace gases  reacts with  the sorbent to form calcium sulfate,  which  passes  through
                                        2-58

-------
the  convective  pass  of the boiler with coal fly  ash  and  unreacted  sorbent,  to be
collected by the electrostatic  precipitator.


Preliminary  tests  indicate  that  the most  important  factors  controlling  the
effectiveness of the  GR-SI  process  are:


     1.   The ratio of  air to  fuel  in  the  reburning zone.   Overall NOX reduc-
          tions  are  highest  when that  ratio in  that  zone is 0.9.   The  test
          program will  verify  the optimum quantity of  reburning  gas  for NOX
          reduction.

     2.   Residence time distribution  of the. reactants in  the reburning zone.
          Mean gas-phase residence time in  the zone  is 0.3 to 0.5 seconds.  By
          operating the system under varying  load  conditions, the influence of
          residence time on reburning effectiveness  will be quantified.

     3.   Temperature  of  the  furnace.   Sulfur capture  (sulfation)  occurs  at
          temperatures  between 1600°and 2200°F.   Temperatures  in  excess of
          2300°F  significantly  reduce  the  sulfation  rates  because of
          "deadburning."   Tests  conducted at varying  load conditions  will
          quantify the influence of furnace temperature on sulfur removal.
GAS REBURNING DESIGN


Another critical  factor to the success of both processes  is  the rapid  and complete
mixing of the  injected  reactants  with the furnace gases.   To  ensure  this,  a  1/12
scale  isotherma"!  flow  model  of  the  Hennepin  boiler  was   built  to  test   the
effectiveness of  the  proposed  injector designs.  The model  was  based on furnace gas

velocity measurements  made during  the field  evaluation  tests.   Smoke and  soap
bubbles  were  used  in  the model  to  make  the  jets  visible,  and   quantitative
dispersion  measurements  were made  using methane as the tracer.   The physical  models

were verified  using  computer  models  of  the  heat  transfer  characteristics  of  the

boiler.
SORBENT INJECTION  DESIGN


Sorbent Injection is used  to  reduce  S02  emissions from the combustion  of  sulfur-

containing fuels.   At  Hennepin Unit  1,  the SI  system  consists of the  following

steps:


     1.   Hydrated lime is  injected into the upper furnace.

     2.   The flue  gases are  humidified  prior to  solids collection in  the
          electrostatic precipitator  (ESP) to  enhance the performance  of the
          ESP by increasing the conductivity of the  solids and decreasing the
          temperature of the gases.
                                       2-59

-------
     3.    An acid  is  injected  into  the ash sluice line to  control  the pH of
          the effluent water to a  value between  6  and 9.

The process design for sorbent injection involved the selection of the appropriate
temperatures, furnace  injection points and proper velocities  for  good dispersal.
mixing and contact of  sorbent particles and SO? gases.

Sorbent is  used  in excess  quantities  because sorbent  utilization  is incomplete.
The collected mixture of fly  ash and  sorbent is alkaline.   Uhen the fly  ash is
hydraulically conveyed  to  an  ash pond,  the resulting  pH  is above 9.0.   Acidity
control  can  be   accomplished   by  the  addition  of an acid  such  as  sulfuric,
hydrochloric, acetic  or  C02-   The  choice  of the acid  depends upon  existing and
anticipated  ground water  quality  vs.  regulatory  levels, corrosiveness  vs.  system
materials,  and  cost.    For  the Hennepin  installation,  pH  control  by  liquid CO?
injection  was selected.

The GR-SI  injector specifications  are shown  in Figure 2.  They include four sets of
four  tangential   natural  gas  injectors with  recirculated  flue  gas as  carrier
installed  above  the  existing  coal  burners  for gas reburning.  followed  by  four
reburn air ports  in the upper furnace.   Hydrated lime sorbent will  be  injected with
transport air through  six  injectors  at the elevation of the  boiler  nose,  four on
the front wall and two on side walls (at  low  load injectors  located  in the reburn
air ports will be used  for  sorbent  injection).    Flue gas  duct  humidification has
been installed to upgrade the performance of the  ESP  with sorbent  injection.
 GR-SI ENGINEERING DESIGN

 Based on  detailed  process  design  studies  completed  for the Hennepin boiler, GR-SI
 system performance requirements were identified.  Detailed engineering followed to
 identify new equipment and  modifications to  existing  equipment necessary to  achieve
 those performance  requirements.   The GR-SI  installation  is shown in the schematic
 di agram of Fi gure 3.

 The Gas  Reburning-Sorbent  Injection and auxiliary  system design included detailed
 engineering work  in the  following  areas:

          Sorbent injection system
          Gas reburning  system
          Flue gas humidification  system
          Ash handling system  modifications
          Power distribution system  modifications
          Control system modifications
          Sootblowing system modifications

                                       2-60

-------
 To  meet  performance  requirements  for  the  sorbent  injection  system,  subsystems  were
 designed to store, meter, convey, and inject sorbent  into  the  Hennepin  boiler upper
 furnace.   A  new  bolted,  skirted  sorbent storage  silo was  erected  to provide
 approximately 3 days supply at the nominal operating  condition  (Ca:S-2.0).  Sorbent
 will  be  delivered  by truck to  the site,  pre-pulverized, and pneumatically conveyed
 into  the storage  silo.   The  metering  and dilute-phase  conveying  subsystem will
 deliver  2,300-11,200 Ib/hr of  sorbent  to be distributed evenly  to the  injection
 nozzles.   Additional injection  air  is  introduced with the sorbent/transport  air
 stream at  the  injection  nozzles  located  on the  front and  side walls of the boiler
 upper furnace.

 The performance requirements  for the gas  reburning subsystems were  designed to
 meter, convey, and inject natural gas, recirculated flue gas, and overfire air into
 the Hennepin  boiler furnace.   The GR system is  designed to  supply  2244 SCFM or
 approximately 20%  of the boiler  fuel  requirements at full  load.  Recirculated flue
 gas  is mixed  with  the  natural  gas at the point  of injection into the boiler.  The
 flue  gas stream provides the jet  momentum  necessary  to insure  that  adequate mixing
 of  the  natural  gas and  the  combustion  products occurs in  the boiler  furnace.   A
 total of  3-5  percent of the flue gas is recirculated in the  Hennepin  Unit.   The
 overfire air  system  provides  heated  combustion  air to four  ports  located  on  three
 of  the  four  furnace walls.   This  air  stream  provides  the  oxygen  necessary  to
 complete burnout of the remaining combustibles in the furnace.

 To  maintain particulate emissions below  the allowable limit, the performance of  the
 existing ESP  must  be enhanced during sorbent injection.  This is  accomplished by
 using a  single  pass  multi-spray  humidification  system.   The system  is designed to
 deliver approximately 60 gpm of filtered  river water  to cool the flue gases to 70°F
 above  saturation.    The test  program will  identify the  optimum  requirement  of
 cooling water.  Compressed  air  is used to atomize  the water.  The existing flue  gas
 breeching  was  modified  to  provide  2.0  seconds   of  retention  time  in  the
 humidification spray chamber  prior to entering  the  ESP.  The  installation  of  the
 spray chamber required that the  ID fans  be relocated  as well as the replacement of
 the majority  of the ESP  inlet and outlet  ductwork.

 The flue  gases  contain  a  mixture of solid particulates  as they exit  the  boiler
 during GR-SI  operation.   The particulate is  a  mixture  of  normal coal  ash  and
 partially sulfated  sorbent.   Analysis  of  ash handling alternatives available to the
 Hennepin  site determined that  wet handling to  an  on-site pond was  the most cost
 effective approach.   Modifications to the  existing  sluice system  included  a  new
 hydroveyor with  a  built-in  ram  for  cleaning,  sluice  line  replacement and  a  new
microprocessor based  controller.

A power distribution system was  designed  to provide  power to  all  the GR-SI  system
equipment.   The power  is fed from the plant's  existing 2300 V switchgear.   A  new

                                       2-61

-------
1000 KVA transformer  reduces the  voltage to 480 V for distribution.  That supply is
distributed to the GR-SI  equipment  through  various overload and  fault  protection
equipment.   All motors  rated  from  1  HP  to 200 HP are supplied  480  V.  and smaller
motors are  supplied 110  V.  The maximum peak demand expected during GR-SI operation
is 750 kW.

The  control   of  the   GR-SI  systems  was  studied extensively  during  the  design
engineering effort.   Control  logic  was  determined to  provide  system  operation  in  a
safe  and  efficient  manner.   Necessary  control  system  hardware  and  software
modifications  were  identified  and  equipment   selected.    A   Westinghouse  WDPF
microprocessor system  was selected  in  order to allow for  future  expansion  and
interaction with  any  future controls upgrade.

After  reviewing  the  cleaning  capability  of  the existing  16  sootblowers,  it  was
determined  that several  areas  of  the boiler may be subject to heat transfer surface
fouling due to the increased  ash loading  in the  flue gas  during  sorbent injection
operation.   The areas of concern  are  all  in  the  downflow convective  section  of the
boiler.  Eight sootblower  locations were identified for  the installation  of the new
sootblowers,  which are supplied by a new sootblower  air  compressor.
PERFORMANCE TESTING

To determine performance of the process, a  continuous  emissions  monitoring  system
(CEMS) is  needed.   The  CEMS  is  capable of  monitoring from the economizer or  from
the stack  breeching.  At the  economizer, eight  4-in.  sampling  ports  are utilized.
Each contains two phase discrimination probes designed to  reduce  particles  in  the
gas stream.  Eight  of  the probes are inserted 1/3 of the way into  the  duct,  and  the
other  eight  2/3 of the way.  to  provide a  sampling grid  representative  of  the
economizer duct  flow.

The phase  discrimination  probe  design  is  shown schematically in  Figure 4.   The
design of the probe  is  such that heavy  dust  particles in the gas are unable  to make
the two  90° bends required  to enter the  annular  space  and  flow to the  monitoring
system.   A  ratio  of  approximately  10:1  by-pass  gas  to  sample system  gas  is
maintained.  The relatively clean  gas  from  each probe passes through  a  filter  and
rotameter system, and  is then  pumped to the test trailer via a heated  Teflon  sample
line.   The CEMS analyzers  used in  the test trailer are shown  in Figure 5.   All
components  outside the  sample duct   are  heated  to  at  least 250°F  to prevent
condensation  in  the  sampling system.
                                                                                is
Phase discrimination  probes  are  not  required at  the stack  breeching.   This
because  of the very low  dust  content and good mixing from the  ID fan just upstream.
Thus,  only a  single probe is  used for sampling.
                                       2-62

-------
 The test trailer is equipped with  a  computerized  data  acquisition system which is
 tied to all  of  the monitors.   Every  3  seconds the computer takes  a  reading from
 each instrument.  The computer  logs the  average of  the  readings  each  minute.  The
 computer displays the  raw signals  and  calculated emission data.  Two hour trends of
 average data are also available.   All  data  are  downloaded to tape  from the hard
 disk daily  for  later reduction  and  analysis.
 BASELINE MEASUREMENTS

 Baseline data measurements  began  on  October  1.  1990.  Over  the  period  October 1,
 1990 through January 23, 1991, economizer sampling was performed  on  49  days  for a
 total  of 406 hours, with stack breeching data collected on  19  days  for  a  total of
 166 hours.   Hennepin Unit  1  is  a  cycling  unit and  is  generally  under  dispatch
 control  .   Thus, the load varies widely from day to day and the unit is frequently
 off-line from 10  P.M.  to 6 A.M.  during  the week  and  more often than  not  is  off
 during weekends.   Thus,  it  provides  a  changing environment for NOX formation due to
 the lack of steady state during its  normal  mode  of  operation.  The distribution of
 economizer  sampling  hours versus load  is shown below:

              MWe(gross)   hr                   MUe(gross)     hr
               30-34        3                     55-59       89
               35-39       12                     60-64       71
               40-44       19                     65-69       80
               45-49       29                     70-75       53
               50-54       50

 Figure 6  shows  a plot of NOX concentration versus  load under baseline conditions.
 The  data points  represent average  values of NOX versus boiler load grouped  in 5  MW
 increments.   The solid line represents the  linear  regression trend line with  the
 dashed lines indicating ±1 standard deviation.
GAS REBURNING RESULTS

Only short  term GR  runs  have  been made to  date (February 1991)  to  initiate the
optimization  of the operation,  which will  be  followed by  optimization  of SI
operating conditions and long term tests  of the  combined  technology.   Six-hour  runs
were performed  on  a number  of occasions.   Parameters that were  varied include:
natural  gas, flue  gas  recirculation,  overfire  air (OFA)  flow rates,  and  reburning
zone stoichiometry  at different boiler loads.  Figure 7 shows a plot  of  NOX  and 0?
versus  time  during  a  typical  GR run.   It can  be  seen  that at  a reburning  zone
stoichiometry of 0.9,  the  NOX  emission level  drops  from  the baseline condition of
                                       2-63

-------
about 400 ppm to between 125 and 150 ppm (  all  values  corrected to 3% Oz dry basis).
Thus,  reductions  of NOX  on  the order  of  65% are  indicated by these  prel iminary
tests.   The  degree of  reduction  slightly  exceeds  the stated  target of  60% NOX
reduction by  GR.   For  comparison,  the predicted  reductions in NOX  calculated using
EER's  kinetic models are  shown  in  Figure 8.   Figure  7 clearly indicates the strong
correlation between decreases in NOX emission and  lowering  the reburn stoichiometric
ratio  or  the  overall  level  of excess air,  respectively.   The  results  also suggest
that increasing the rate of overfire air flow tends to decrease NOX emissions.  This
effect will be  investigated further.
 The  variation of NOX emissions with reburning zone stoichiometry is shown in Figure
 9  at a  unit  load of  69 MWe.  A linear relationship is exhibited, ranging from about
 400  ppm NOX  at  a reburning  zone  stoichiometric  ratio  of  1.20  to a  range of 125-225
 ppm  at  0.90  in  line with the NOX  reduction  levels  indicated  in  Figures  7  and 8.

 There is another  effect  worth  noting.   The OFA ports were  added to  the boiler for
 gas  reburning.   Under  baseline  (non-GR)  operating conditions,  about  6000-8000 SCFM
 cooling air  passes through them to prevent thermal damage.   In one  test cooling air
 was  shut off for  a  few  minutes  and it  was  found that the measured  NOX increased
 about 9-10%.  When cooling air  flow was restored,  the NOX level  decreased  to its
 former  value of 400 ppm.

 This observation  suggests that  evaluation of  gas reburning performance  must  take
 into account the effect  of  this  cooling  air  on  baseline  NOX levels (i.e.. the true
 baseline may be about  10% higher than  the level  measured with  cooling, air passing
 through the  OFA ports).
 FUTURE PLANS

 A statistically designed matrix of tests is  being completed  to evaluate the effects
 of  02.  load,  flue  gas  reci rcul ati on.  coal  burner and  gas  injector tilts,  and
 stoichiometric  ratio  on the  reburning  process.   This optimization  of  operating
 conditions will be  followed  by a similar set  of sorbent  injection tests  aimed  at
 optimizing the overall  GR-SI  process  for the Hennepin unit.

Once optimization of  conditions  is  achieved,  a  long term  (12-month)  test will  be
conducted  at Hennepin.  which  will  evaluate the  emission  and  thermal  performances  of
the GR-SI  control  technology,  in addition  to  potential  boiler  impacts such  as
slagging,  fouling, and  tube wastage.   Other discharges will also be monitored
                                       2-64

-------
A  similar GR-SI  test  program will  be conducted  on  the  cyclone  fired boiler  at
Lakeside, where construction is in progress.
SUMMARY

A  full  scale Gas Reburning-Sorbent  Injection  NOX-S02  emission control system  has
been installed on a  71  MWe(net)  tangential  boiler  fired with  3 wt%  sulfur  Illinois
bituminous coal.  Preliminary GR  test  results  indicate that  the predicted  level  of
60% reduction in NOX  should  be  attainable.   Startup  and preliminary  testing  of  the
SI  component of the  system will  be completed  in early  1991.   Current  economic
projections  indicate  that  the  combined technology may have broad applicability  to
older,  relatively  small boiler units  requiring  emission  reduction  as a result  of
the Clean  Air Act  Amendment of 1990.   The  projected capital  cost of this type  of
installation (about  $90/kW for GR-SI.  $30/kW  for GR)  is  lower  than  that   of
scrubbers, while  operating costs of 6-9  mills/kWhr  may  be kept  within  reasonable
bounds for units operating at low capacity factor.
ACKNOWLEDGEMENTS

This paper  is  based on work  funded  by the U.S.  Department  of Energy. Pittsburgh
Energy Technology Center,  through Cooperative  Agreement  No.  DE-FC-22-87PC79796; the
Gas  Research  Institute through  Contract  No.   5087-254-1494; and  the  State  of
Illinois.  Department  of  Energy  and  Natural  Resources through  Coal  and  Energy
Development Agreement EERC-2.
REFERENCES

1.   Reed, R.D., "Process for Disposal of  Nitrogen Oxide." John Zink  Company.  U.S.
     Patent 1.274,637. 1979.
2.   Sternling. C.V., et al.. 14th Symposium (International)  on  Combustion,  p.  897.
     The Combustion Institute,  1973.
3.   Takahashi, Y., et al.,  "Development of Mitsubishi  MACT  In-Furnace  NOX  Removal
     Process,"   Paper  presented at  U.S.-Japan NOX  Information  Exchange,  Tokyo,
     Japan, May 25-30, 1981.
4.   Ogikami,  N.,  et al., "Multistage Combustion Method  for  Inhibiting Formation  of
     Nitrogen  Oxides," U.S.  Patent 4.395.223,  1983.
5.   Greene,  S.B.,  et al.,  "Bench-Scale Process  Evaluation  of Reburning  and  Sorbent
     Injection for  In-Furnace NOX/SOX  Reduction." EPA-600/7-85-012, March, 1985.
6.   Greene,  S.B.,  et al.,  "Bench-Scale Process  Evaluation  of Reburning  and  Sorbent
     Injection for  In-Furnace NOX Reduction."  ASME Paper No. 84-JPGC-APC-9. 1984.

                                       2-65

-------
8.
9.
10.
11.
     Seeker, W.R.,  et al.,  "Controlling Pollutant  Emissions  from  Coal  and  Oil
     Combustors Through  the  Supplemental  Use of  Natural  Gas," Final  Report,  GRI
     5083-251-0905.  1985.

     England,  G.C.,  et  al.,  "Field  Evaluation  Humidi f i cati on for  Precipitator
     Performance  Enhancement,"  presented at  the  7th  Symp. on  the  Transfer  and
     Utilization  of  Particulate  Control Technology,  Nashville, TN,   March 22-25,
     1988.
     Bartok.  W.  and  B.A.  Folsom, "Control  of NOX  and  SOz  Emissions by Gas Reburning-
     Sorbent  Injection,"   American  Institute  of  Chemical  Engineering  Annual
     Meeting,  New  York,  November 1987.
                                     .

     Folsom, B.A..  et  al . ,  "Field  Evaluation  of Gas  Reburni ng-Sorbent  Injection
     Technology for NOX and SOX Emission Control for  Coal  Fired  Utility  Boilers,"
                            Conference  and  Exposition,  Washington,  D.C.,  February
                      X
     15th Energy Technology
     17-19,  1988.
     Bartok,  W., et al .  , "Gas  Reburning-Sorbent Injection  for  Controlling  SOX
     NOX  in  Utility  Boilers," Env. Progress SCI).  18.  1990.
                                                                               and
12.   Bartok, W.,  et  al.,  "Design  Modeling  of  a  Nitrogen  Oxide-Sulfur  Dioxide
     Emission Control  Process," Toxic and Hazardous  Substance  Control,  in  press.
                                      2-66

-------
           SORBENT
        (HIGH LOAD)
     OVERFIRE AIR + ,
SORBENT (LOW LOAD)

       GAS 20%+FGR •

          COAL 80%-
                                      COAL 80%'
                    SORBENT
                    OVERFIRE AIR
                    GAS20%+FGR
                      TANGENTIAL
                                                       CYCLONE
            Figure 1.  GR-SI Configurations  for Two Types of Boilers
         SORBENT*
   EL 553'   0"
   SIX 3" INJECTORS
     240 FT/S
     4 ON FRONT WALL
     2 ON SIDE WALLS
   TRANSPORT AIR IS 3%  OF
   TOTAL COMBUSTION AIR
         REBURN  GAS
   EL 520'    6"
   FOUR INJECTOR ASSEMBLIES
     TANGENTIAL/TILTING
     EACH HAS FOUR
     45/s"  x 1"  NOZZLES
     ON 11" CENTERS
     NG + FGR
     415 FT/S
   FGR IS 3% OF  TOTAL  FLUE
   GAS FLOW
             REBURN AIR
          EL  530'   9"
          FOUR  15" x 30" PORTS
           110  FT/S
           575" F
           ADJUSTABLE VANES
EXISTING
  COAL
BURNERS
                                     * LOW LOAD SI THROUGH  REBURN AIR INJECTORS
  Figure  2.   Summary of  Injector Specifications  for  Tangentially  Fired Boiler
                                     2-67

-------
ro
en
oo
      TRUCK
     UNLOADING
              TRANSPORT AIR
               BLOWER
                                                                                  ASH LINE
                                                                                  PH CONTROL
                   HENNEPIN  UNIT 1  TANGENTIALLY  EIRED  BOILER
                                   Figure 3.  GR-SI Process Schematic

-------
BY PASS GAS
(VIA WATER SEAL
VACUUM PUMP)
                       GAS SAMPLE TO
                          CEMS
                      (VIA THOMAS PUMP)
                              A

             Figure  4.   EER Phase Discrimination  Gas  Sample Probe
            From
            Plant "
           Sources
i   i—i        i
    Heafed "   Perma-
   Enclosure     Pure
                 Drier
                                                       INSTRUMENT AIR
                Figure 5.  Continuous  Emission Monitoring System
                                       2-69

-------
ro
->i
o
          n
          it
          a
          H
          •o
dP
n

OB


B
         O
         55
               500
               475
               450
               425
               400
               375
     350
              325
              300
                       30-34
                         35-39
40-44
45-49
                                                                     50-54
                                                                       55-59
                                                                                            60-64
                                                                                                       65-69
                                                                                                                   70-75
                                                          Boiler  Load,  MW  (Gross)
                                          AVERAGE VALVES
                                                                LINEAR REGRESSION TREND     ..........  STD .  DEV.
                                                       Figure  6.   Baseline NOX Emissions

-------
ro
                 500
                 450
              —  400
              a
              a
                 350
              O  300
              I
              x
             o
                 250
                 200
                 150
                 100
                    8:27
                                                                                                                                W
                                                                                                                                M
                                                                                                                                o
                                 9:27
                                             10:27
                                                         11:27
                                                                      12:27
                                                                                  13:27
                                                                                              14:27
                                                                                                           15:27
                                                                                                                       16:27
                       RS  = Reburn Stoichiometry
— NOX
— 02
                                                                                          All flowrateg expressed in SCFM
                       1/10/91  8   08:30  hrs.
                                              Figure  7.   Hennepin Unit  1  Gas Reburning Test

-------
1000
 200
      CYCLONE
    •QVERRRE AIRI

                 L_
                                  TANGENTIAL
                                  OVERFIRE AIR
            0.2
0.4
0.6
0.8
1.0
              RESIDENCE TIME (SECONDS)
                Figure 8.  Predicted NOX Control
                        2-72

-------
               550
                                                                                             Boiler Load = 69 MW (Gross)
               500
                                                                                                             O
               450
               400
                                                                                                     O
                                                                                                          O
               350
ro

-^
OJ
           I
           a.
            X

           o
               300
               250
               200
               150
               100
                  0.85
                              0.90
                                         0.95
                                                           O
                                                                 O
                                                    1.00
                                                               1.05
                                                                          1.10
                                                                                      1.15
                                                                                                 1.20
                                                                                                            1.25
                                                                                                                       1.30
                                                   Reburning  Zone  Stoichiometric  Ratio
                                     Figure 9.  Effect of  Reburning Zone Stoichiometry  on  NOX  Emissions

-------
RETROFITTING OF THE ITALIAN ELECTRICITY BOARD'S
             THERMAL POWER BOILERS

      R. Tarll, A. Benantl, G. De MIchele
        Ente  Nazionale  Energia Elettrica
                     Italy
                 A. Piantanida
                     F.T.C.
                     Italy
                   A.  Zennaro
                  ANSALDO ABB
                     Italy

-------
ABSTRACT

ENEL  is carrying  out  research to  improve  the  environmental  impact of  thermal
power  stations.  Particularly,  as regards NOx  emissions,  ENEL is about to  adopt
mainly  combustion  modification technique  and low-NOx burners  and to  install  SCR-
type abatement systems.
The paper  presents the first results of the demonstration program started  in  89
and presently in progress.
1.    FOREWORD

In 1990 over 877. of  the overall  electricity  demand  (235 TWh) in Italy was met by
ENEL power  stations  fired with  fossil fuels  (oil,  coal  and natural gas)  Coal-
fired stations  supplied 28,5 TWh, thus covering 12% of the demand.
To  meet  the  requirements  concerning  NOx  emissions,  ENEL, as   announced  in a
previous paper  /!/ intended to first attain  maximum reduction through combustion
modifications  on  all planned  and  operating plants, and  then  add high-dust SCR
systems  to  plants  fired  with  low-sulphur  oil.  A  comprehensive demonstrative
program and a number of preliminary results were presented.
This paper  illustrates  the more  important data and  conclusions of  the  tests
performed   so   far  by   ENEL,  jointly   with  the   national    steam-generator
manufacturers  (Ansaldo  and F.T.C.),  to  reduce  nitric oxide  emissions  through
modification of the  combustion system.  It should be remembered that the boilers
being modified  are  mainly of  two  types,  that is,  the wall-firing  type  made by
Ansaldo under  licence  from Babcock and Wilcox, USA,  and the  tangential-firing
type made by F. Tosi under licence from Combustion Engineering.
Table 1  shows   the configuration of the  burners  of the  operating  boilers  that
will be modified.  The  modifications will  involve an  overall installed capacity
of 23,815 MW,  including 29% from coal.
The units can be divided into nine different groups, depending on the combustion
system configuration (five for oil-fired units and four for coal-fired units).
In the meantime the  Italian Government  issued new and more restrictive emission
limits   that,  in the  case  of  NOx,  are  200 mg/Nm^ for  new and  existing  power
stations with  a capacity  greater  than  500  MWth. For  the  sake of  clarity,   the
results  of  the above-mentioned  demonstrative program are described separately
for the  two  different types of boiler (wall-firing and tangential-firing)
                                       2-77

-------
2.     ACTIVITIES UNDER WAY ON WALL-FIRING BOILERS
2.1   Oil and Gas Firing

In  the  above  mentioned  paper,  ENEL  announced  that  tests  were  being  run,
concerning the  BOOS  technique,  on oil-  and gas-firing unit;  certain  data were
presented for cell burners in gas firing units (ROSSANO #4) and for cell burners
in oil firing units (CASELLA #3) .   A short communication was given for the tests
that ENEL was running,  at that  time,  on SERMIDE #1  (gas  firing,  axial burners
unit); this test is now complete.
ANSALDO  installed 6  new  NOx  ports  on  CASELLA #3  (the  6  upper burners  were
eliminated).
A  test  is  now under  way  on SERMIDE #2  concerning  the BOOS  technique  on axial
burners In oil firing units.
Another test  is  under way in PIOMBINO //4 concerning  the  BOOS technique for oil
firing in coal designed units.  More details about the results of these tests are
given further on.
2.1.1 Low NOx Combustion Tests with Oil Units.   In April 1990, ANSALDO installed
6 new NOx ports on boiler #3 at the CASELLA power station, eliminating the 6 old
upper burners,  previously used as  NOx ports;   after  this modification  the NOx
dropped  from 950  mg/Nm3  of  the  18  burner  configuration.  (02  about  0,9%,  GR
regulating)  to  550 mg/Nm3  with 02    1.4% and gas recirculation  (GR)  dampers
opened at 20%.
The  new  NOx  ports,  in comparison  with  the old  burner air  register,  slightly
improved NOx  reduction:  but an 0.2%  reduction  in 02 was  possible  (at  the same
NOx value) as compared to the 12 burner configuration.
The unit has been operating in the 12 burner mode, waterwall gas analysis showed
a reducing  atmosphere;  anyway,  up to  now,  we  have no  evidence  of corrosion on
the  waterwalls.   The  same  thing  can be  said  for  two  units  of  the  same power
station  that,  for  many  months,  have been operating  with  the  BOOS  12  burner
configuration.
During April  1989,  ENEL  performed a test to verify  the application of the BOOS
technique to  ROSSANO  #4;  this is  an  oil   and  gas-firing  cell burner unit; the
test was performed firing oil.
NOx  emissions are  reduced  from   775  mg/Nm3  (18  burners)  to  540  mg/Nm3  (12
burners)  with 02 =  1,6%  and GR dampers opened 70% with CO =  84  mg/Nm3 (fig. 1,
2)   The  opening of  the  GR  dampers has  a  smaller effect on NOx  reduction as
compared to the reduction achieved  on the  same  unit during the gas-firing test.
                                      2-78

-------
 The  poor quality of the oil burnt  during the test required a  small  increase  of
 02 to  keep  CO  and particulate  at acceptable values.
 ENEL is  now running  a  test  on  the PIOMBINO #4 oil-firing,  coal-designed  unit,  to
 verify the  application of the BOOS technique  on this  type of plant.  This  unit
 has  30 burners,  divided into  15  cells  of 2 burners each.   The upper burner  of
 each cell is used as a NOx port,  while the oil  flowrate has been doubled  in the
 lower  one;  in  another low  NOx mode,  all the cells in  the upper row act  as NOx
 ports,   while  the   fuel  flow rate  of   the   lower  one   has   been   increased.
 Preliminary data show a  NOx  emission  reduction of about  20%  from an  original
 value  of about 430 mg/Nm3 with 02 = 2%.
 2.1.2 Low NOx  Combustion Tests for Boilers  Equipped  with  Axial Burners.   Since
 the end of November  1990,  ENEL has been  running  a  test on  SERMIDE #  2;  this unit
 is  an  axial  burner,  with  an  oil-   and  gas-fir ing  boiler.  This  is another
 application  of  the  BOOS  technique   (the   fuel   used  for  the  test  is   oil) .
 Preliminary  data show NOx  is reduced  from 950 mg/Nm3  (02"!%)  to about 520 mg/Nm3
 with  02     1,1%.  Waterwall  gas  analysis  is being performed on this  unit and
 further and  more  complete details will  be  ready at the end  of the test period
 (March 1991).
 As part  of  the demonstrative  program for low NOx burners,  ANSALDO installed a
 burner named TEA (designed jointly by  ENEL and ANSALDO) on the MONFALCONE unit #
 4  (oil-firing  with  axial burners before  modification).  Preliminary data show a
 reduction of about  40%  (fig.  3).   The test program has not  been completed yet
 and  the   installation of  NOx ports   for  the  application  of post combustion
 technique on low NOx burner combustion system is scheduled for next year.
 The second  part of  the  demonstration program involves  the  installation of the
 Babcock and  Wilcox  low  NOx burner, XCL type,  on  SERMIDE  #  1;  the  installation
 has been completed.  A test phase is under way and will last at least two months.
 The test  program  has not  been completed yet and  the  installation  of NOx  ports
 for the  application of  post  combustion technique on  low  NOx burner combustion
 system is scheduled for  next year.
 The second  part of  the  demonstration program involves  the  installation of the
 Babcock and  Wilcox  low  NOx burner, XCL type,  on  SERMIDE  #  1;  the  installation
has been completed.  A test phase is under way and will last at least two months.
2.1.3 Low NOx Combustion Tests  with  Gas-Fired Axial Burners. The test unit  used
to  verify the  application  of  the  BOOS  technique  to gas-fired,  axial burner
boilers, is SERMIDE # 3 (320 MW) .
Operation with 12 burners and addition of air  through  the upper burners  leads  to
a considerable reduction of  NOx emissions as compared  to the  18-burner operating
                                       2-79

-------
mode (fig. 4).   At 320 MW,  with 02 =  1.07  %  and GR dampers opened 65%,  the  NOx
emissions decreased from 685 mg/Nm3  to 170  mg/Nm3 (75% reduction), while  CO  was
40 mg/Nm3   Without GR (dampers  closed)  the NOx emissions were    300 mg/Nm3  (02
  0.8%)
2.2   Low NOx Combustion Tests with Coal

During  1990  ANSALDO  modified the  combustion  system  of  Unit  #  4  at  the VADO
LIGURE power station; this unit was originally equipped with  30 burners  arranged
in  15  two-register cells.   According  to the technology  developed  by B & W,the
lower burner of  each cell  is  still firing coal, while  the  other  one is used  to
introduce  the  over-fire  air.   The coal  pipes of  the lower burners  are  enlarged
to  accomodate  the increased  coal  flow, and the  upper air  registers have been
modified internally
The  first  test program was completed  in 1990,  firing American bituminous coal;
the  data were  compared to the data  we obtained in  1988,  firing  the same coal.
As  far  as emissions  are concerned, NOx at 330  MW after  modification are 876
mg/Nm3  (02   4%)  and carbon in flying  ashes is about  8%; (5 mills  in  service);
with  the  original burners in  the  same  operating  condition,  NOx  emissions were
1200 mg/Nm3  and  carbon in flying  ashes was  6%.  Operating  in the same way, but
with higher 02  ("  5%), NOx  can be  reduced from  1440 mg/Nm3 to 970 mg/Nm3, while
carbon  in  flying  ashes  increases  from  5,5  (original burner configuration)   to
7,7% (new burner configuration) (see fig. 5,  6)
3.    ACTIVITIES UNDER WAY ON TANGENTIAL FIRED BOILERS

For  this  type  of  boilers  ENEL  adopted  the  Combustion  Engineering  low  NOx
combustion system; this system was tested at the FUSINA power station.


3 .1   FUSINA Project

Unit # 2 of the  FUSINA power station has a 160 MW, multi-fuel tangential boiler,
usually firing coal.  During 1989 the unit was modified to reduce NOx emissions;
a  new  combustion system  /3/  was installed by  F.T.C. (the  Italian  licensee  of
COMBUSTION  ENG.)   During the  test  period  (February-July  1990),   low-volatile
South African and high volatile U.S.  bituminous coal, oil and gas were burned.
While firing  a S.A.  coal (TCOA),  a decrease in NOx  emission from 930 mg/Nm3  to
500 mg/Nm3 was achieved  (02 was about  4.1%  in both cases)   An  increase in mill
                                       2-80

-------
fineness  (from  85%  to  90%  on  200 mesh)  was  necessary  to  limit the  carbon loss  in
flying ashes  to about  7% in the low NOx configuration  (fig. 6)
The  NOx  reduction obtained by  firing an American coal  (MC  CALL)  was 47%  (from
740  mg/Nm3  to about 390 mg/Nm-') with 02    4%  and unburned coal in  flying  ashes
was  about 6%  with the  low  NOx system (fig.  7).  In both cases we  found  that NOx
emissions are not affected by coal fineness and that high NOx reduction can  be
achieved only by  a high over-fire air flow  rate  (~ 140 t/h).
We noted  that U.S.  coals produce  less NOx than S.A.  coals ("  200  mg/Nm3  less  in
the  high  NOx   without OFA    configuration and    100 mg/Nm3 in  the low NOx
high  OFA    configuration)    Increasing  the mill  fineness was  the  only way  to
reduce carbon loss  in  the  flying  ashes  to acceptable values   (6-8%),  for  all the
coals we tested (S.A.  and American)
During the  oil  firing  test,  while  maintaining  the temperature of  the  convective
parts  of the boiler,  the appropriate  value  was the  most  important   problem
(FUSINA # 2 has no GR  system);  so excess  air had to be high in order to  increase
these temperatures.   NOx  emissions  are reduced from    500  mg/Nm3  (without the
OFA  mode) to    220 mg/Nm3  (high  OFA mode) ;   C>2  was about  2,4% and  CO was low
in both cases (~  30 ppm)
As  far  as  emissions  in  gas-firing  are  concerned,  NOx  emissions  in the  base
configuration  (without OFA)  were  about 360 mg/Nm3 (with 02    1,5%  and  CO  =  40
ppm);  in the  best  operating low-NOx  condition,  NOx  emissions   were about   95
mg/Nm3  with  02    1,65 and CO = 60 ppm;  NOx emission reduction was 74% (fig. 8)
4.     REBURNING

The aim of ENEL's program  is  to evaluate  the application of reburning technology
both on standard oil and coal units.

Three phases have been planned:

a)     Bench-scale tests on a 50 kW furnace
b)     Experiments on a 15 MW Combustion Engineering Boiler Simulator
c)     Demonstration on the 35 MWe Santa Gilla # 2 unit.

The bench-scale  experiments  were  concluded and new,  interesting  results were
obtained.    The  modification of  Santa Gilla  //  2  was completed and  tests  are
scheduled for next April.
The  first  part  of  the  experiments  on  the  CE   simulator  was  completed  and
concerned  the application  of  gas  reburning  to   an  oil-designed  boiler;  the
results were promising (fig. 10).
                                       2-81

-------
5.     CONCLUSIONS

ENEL's efforts in the field of NOx reduction are producing their first results.
The  application  of  the  BOOS  technology  is  giving  good  results  in oil/gas-
designed boilers whatever  burner  is installed  (axial or  circular).  In terms  of
NOx  reduction,  the  application  of BOOS  in the  case of  coal-designed boilers
firing oil and coal is less effective.
Good  performances  are  obtained  by using  an oil-low-NOx  burner  and further
improvements are expected thanks to the installation of NOx ports.
Application of OFA on tangentially  fired boilers  gave a very good NOx  reduction
with gas, good performances were obtained with oil and coal.
Reburn tests  on a  15  MW oil-firing boiler  simulator confirmed the data obtained
using a bench-scale apparatus and indicated a strong decrease of NOx production.
ENEL's demonstrative  program  on  the application of low NOx combustion  technique
is  going  on  and  will be completed in two years,  at the  same  time,  on  the basis
of  the  results being obtained, the  first  applications are being made and will
involve, over  the next ten years,  all units of ENEL's  power stations.
ACKNOWLEDGMENTS

The  Authors  wish to  thank Dr.  G.  Bianchi  for his  contribution  to the  present
paper.
 REFERENCES

 /!/   B. Billi, E. Marches!, R. Tarli
      Retrofitting of existing thermoelectric plants
      GEN-UPGRADE 90 Symposium, March 1990, Washington DC, USA.

 /2/   A. Benanti, G. De Michele,  A. Piantanida, R. Tarli, A. Zennaro
      Retrofitting  of  the  Italian  Electricity  Board's  Thermal  Power  Plant
      Boilers. GEN-UPGRADE 90 Symposium, March 1990, Washington DC, USA.

 /3/   Towle D.P  et al.
      An update  on NOx Emission Control  Technologies  for Utility Coal,  Oil  and
      Gas Fired Tangential Boilers.  AFRC Meeting, March  '91, Hartford,  CT, USA.
                                       2-82

-------

OIL FIRED BOILER
CH1VASSO Unit 6 [•]
PIACENZA Units 3,4
LA CASELLA Units 1,2,3,4
OSTIGLIA Units l.S.3,4
TURBIGO LEV. Unit 1 f«J
Unit 2
Units 3,4
SERMIDE Units 1,2,3,4
TAVAZZANO Units 5,6
nONFALCONE Units 3,4
PORTO TOLLE Units 1,2,3,4
TORREV. SUD Units 2,3,4
TORREV. NORD Units 1,2,3,4
ROSSANO Units 1,2,3,4
TERMINI I. Units 4,5
PRIOLO G. Units J.2
TOML CAPACITY OTyj
CO^L^f_IRED_ BOILEFl
LA SPEZIA Units 1,2,4
Unit 3
VADO LIGURE Units 1,2,3
Unit 1
PIOMBINO Units 1,2,3,4
FUSINA Units 3,4
BRINDISI NORD Units 1,2
Units 3,4
S.FILIPPO Unit 5
Unit 6
SULCIS Units 1,2,3 (•)
TOT^L owMcirr mw
FRONT-REAR BURNERS
TW
REGISTER
CELLS
















O




33O
/ .260

640


32O

2.57O
THREE
REGISTER
CELLS


1 .260


320







/ .280
640

3.520


60O
99O




640
320


2.550
PARALLEL
FLOW







1 .280
64O
640


2.640



5.200












O
TANGENTIAL
FIRING
BURNERS

66O

1 .3SO


660



2.64O
960



640
6.880

/ .235




640





1 .875
            f»3  UNITS EQUIPPED WITH CIRCULAR BURNERS ON THE
               FRONT UALL (TOTAL CAPACITY (.220 MW
fob./.  BURNER  CONFIGURATION  OF ENEL's EXISTING BOILERS  (CAPACITY >200
                                        2-83

-------
           900

           eoo-

           700-

           600-
       \  500-
        Cn
        e
       ^  400-
           300-

           200-

           100-
D IB BURNERS
O IS BURNERS (DAI1PERS CLOSED]
+ 12 BURNERS (DAHPERS OPENED 25X1
                           \
                                 \
              0
     100   300    300
      LOAD.   (HU)
Fig.]  -  ROSSANO #4 - OIL FIRING
         NOx  C3% 021  AS A FUNCTION  OF THE OPENING
         OF THE DAMPERS OF  THE UPPER  BURNERS
           900
           900-

           700-

        |  600-

        \  500-
        cn
        E
        ^  400-

        O  300-
           100-
                a 18 BURNERS
                + 12 BURNERS (DMPERS OPENED 25X1
                     1	1	\	1	\	
                    !00    POO    JOO

                      LOAD,  CHU)
                       400
Fig.2 -  ROSSANO #4  - OIL FIRING
          H1NIMUH NOx   [3% 0?J  VALUES  IN
          ACCEPTABLE  OPERATING CONDITIONS
                        2-84

-------
        \
        CD
1200-



1000-



 600-



 600-
        2 400-
          200-
               & AXIAL BURNER

               o TEA  BURNER
                                NOx
                                       -250
                                       -200
                                       -150
                                      - 100
                                      -50
                                            o
                                            e
                                            en
                                            e
             0     0.5    1.0    1.5    2'.0
                       0S,
    Fig. 3  -  MONFALCONE POUER STATION
             COHPARISON BETWEEN NOx  AND CO
             C3% 02J  OF AXIAL AND  TEA BURNER
              Id BURNERS


              12 BURNERS
100
         150
           200          280


              LOAD,  (HU1
                                              320
    Fig. 4  -  SERtllDE  #3  (320  HU) GAS  FIRING
             NOx  (3%  021 EMISSIONS  VERSUS LOAD
                        2-85

-------
1000



 900-



 800-
\
Dl

vS  700-



0  600-
2;


   500-
                     CARBON LOSS
        CO
 100



•80 'j

   <.

-60 "I







-20
                                        o
   ^ to
. ;2 £JLU

    - to
   to -=c
• /O to
   O CJ5
   -j 2:

•6  ^ x
   O -j
   OQU.
                                           -4
                                              CO •
Fig. 5 - \/>ADO LIGURE #4  (330  MU- 5 MILLS]
         NOx  (3%  02; EMISSIONS VERSUS  0?
         NOx PORTS OPENING=50mm COAL ASHLAND

         GR=50% DAMPERS OPENING

' — ,
0
e
2:
\
O^
e
^
X.
Q
2:




/ uuu -
900-

800-


700-

600-


500-



	 NU X




	 --^
^^^ CARBON LOSS
~^^ 	
** >^^^ 	 ^ — — - '
•-

CO 	 •

- too
-
-80 'S
6
2;
-60 ^
V 	 ,
i- 40 v
O
u
-20
-


^~ 1 c ^
_ V
to
- 10 to
o
	 1
~s 2:
o
CD
-6 Q:
5
-4
                     I   I    I   I   I
    240 250 260 270 280 290 300 310 320 330

              LOAD,  (HU)
Fig.6 - VADO LIGURE #4- LOU  NOx BURNERS

         NOx  (3%  0?)  EMISSIONS VERSUS  LOAD

         NOx PORTS OPENING=50mm COAL ASHLAND
         GR=50% DAMPERS OPENING
                       2-86

-------
           500
         6  450
        S:
        \
         O)
         6
        O  400
           350
               FINESS:90X ON 200 HESH
               OFA i. 155 t/h
    CO
 16  LU
    3:
    CO
                                       - 10
                                       -8
    o
    OD
    CC
    -^
    O
                        Op,  (XJ

Fig.7 - FUSINA  #2  (170  HU - COAL  TCOAJ
         INFLUENCE OF  EXCESS AIR  ON NOx  (3%
         AND  CARBON IN FLYING ASHES-HIGH OFA
         HODES-HIGH FINESS
           450
         S  400-
        \
         en
         e
        O  350-
           300
               FINESS:B7X ON 200 HESH
               OFA i IdS t/h
- ie

- 16



- 12

- 10

- 8

 6
    *•<
                                           LU
                                           U.
                                           Q
                                           CQ
                        Op,  CXJ

Fig.6  -  FUSINA  Jt2  [170  M - COAL  McCALLl
         INFLUENCE OF  EXCESS AIR  ON NOx  (3% 0?J
         AND CARBON IN FLYING ASHES-HIGH  OFA
         HODES-LOU FINESS
                       2-87

-------
        300
        200-
      \
      o>
      e
      o too
        350
                       NO* NO OFA
                   NO* HIGH OFA CiHO I/hi
               I   \    I   I   I    1
          0.6   I  1.2  /.4 t.6  1.8  2  2.2
Fig. 9 -  FUSINA #2 - GAS  FIRING
         INFLUENCE OF EXCESS AIR ON NOx
             Otf  -HIGH AND LOU OFA MODES
    60-
    40-
    20-
             REBURNING
          1.0
                 T
                 0.9      0.6
                    sr. f-j
0.7
Fig. JO  -  DEPENDENCE OF  NOx  REDUCTION
          FROM REBURN  OR MAIN BURNER
          ZONE STOICHIOMETRY OBTAINED
          ON CE-15MU-BSF
                    2-88

-------
RETROFIT EXPERIENCE USING LNCFS ON 350 MW AND 165 MW
            COA1 FIRED TANGENTIAL BOILERS

             T.G. Hunt and R.R. Hawley
         Public Service Company of Colorado
                  Denver,  Colorado

             R.C. Booth and B.P.  Breen
             Energy Systems Associates
              Pittsburgh,  Pennsylvania

-------
                   Retrofit Experience Using LNCFS
          on 350MW and 165MW Coal Fired Tangential Boilers
                     T.  G.  Hunt and R.  R.  Hawley
                 Public Service Company of Colorado
                          Denver,  Colorado
                     R.  C.  Booth and B.  P.  Breen
                      Energy Systems Associates
                      Pittsburgh, Pennsylvania
Abstract

     Public Service Company of  Colorado has  installed ABB Combustion
Engineering's Low NOx Concentric Firing System  (LNCFS)  on both 165MW
and  350MW  coal  fired  tangential   boilers.  The  modifications  were
completed in  1990  as  part of  a voluntary program  to  reduce nitrogen
oxide  (NOx)  emissions.   The  LNCFS  included  new  burners,  control
modifications, and separated overfire air ports.
     Energy System Associates  (ESA)  completed an extensive test program
on each unit both before  and  after  the retrofits.  The test data from
the 165MW unit  has shown that  NOx  was reduced  by 52%  from  0.664  to
0.316 Ib/MMBtu at optimum full load conditions with minimal impact to
carbon monoxide  or unburned  carbon.  Testing was  completed  at  many
different conditions so that the LNCFS could be  optimized for low NOx
operation with minimal operator supervision. Baseline NOx testing on
the 350MW unit  has been completed  and preliminary post-installation
testing has shown a NOx reduction of 47%  from a  baseline of 0.533 to
approximately 0.28 Ib/MMBtu at  full  load. Acceptance  testing has not
been completed due to operational problems with the unit.

Introduction

     Denver Colorado is a beautiful  city  at  the  base  of the majestic
Rocky Mountains but as rapid growth occurred in  the 70's and 80's the
city has also become known for occasional visible pollution problems.
Local politicians foresaw the  importance of clean air not only to local
residents but also as a requirement  to maintaining a  healthy growing
economy.  With cooperation and  financing from Public Service Company of

                                 2-91

-------
Colorado (PSCC)  and other private and public entities, a comprehensive
study of Denver's brown cloud was accomplished in the  late 1980's. This
study concluded that  the major contributors to  the  brown  cloud were
automobiles and fireplaces. Coal fired  power  plants  were responsible
for only 1% of the  direct particulate that interfered with visibility.
The study did find that secondary particulate of ammonium sulfate and
nitrates contributed up to  43%  of the brown cloud, but the study could
not attribute these secondary pollutants to any  source.
     All PSCC's Denver metropolitan power plants  were installed before
NOx regulations were implemented. In September 1988 PSCC announced that
it  would take  a  voluntary pro-active  position to  this  study.  The
Company announced that it would retrofit major coal fired metropolitan
power plants with NOx controls  to allow  a minimum 20% NOx reduction by
the end of 1991 in addition to significant SO2 removal modifications.
PSCC has a mix of boiler types in the metro area including top, wall,
and tangentially fired units.  The goal of this program was to retrofit
combustion modifications that would allow the highest removal that was
economically feasible.  The equipment  was  to be  installed  as  soon as
possible but within the existing scheduled unit outages. The first two
units to be modified were  the  165MW  Valmont 5  and  350MW  Cherokee 4
tangentially fired units.  Modifications were completed in May 1990 on
Valmont  5  and in November  1990  on Cherokee  4.  The  remaining wall and
one top  fired unit will be modified by  December  of 1991.

Original Investigation

     Public  Service Company of  Colorado organized  in the late 1980's
a  three  step program to determine the  best method  for obtaining NOx
reductions. The first step was a complete NOx assessment of all major
metropolitan units  to determine current emissions  and secondarily to
find operational means to  reduce NOx emissions.  The second  step was to
perform  an in-depth  analysis  of the  data,  investigate  the cost and
availability of NOx control modifications,  and finally to recommend NOx
control  measures for the metropolitan  coal fired units. The  final step
was implementation of the  NOx control  plan through installation of the
recommended  modifications.
     PSCC  personnel completed  a comprehensive study of all the units
to find the most economical method to  implement NOx reductions. All
known  options  were  considered  for  the  tangential  units including
operational modifications,  overfire air, Low NOx Concentric Fire System
 (LNCFS),  Pollution  Minimum  (PM)  system,  gas  co-firing,  selective
catalytic  reduction  (SCR),  and  selective  non-catalytic reduction
 (SNCR).  The original NOx  assessment  testing  found  that  operational
modifications  could be effective but NOx reductions were minimal at
high  loads and were  dependent  upon close  operator  attention.  It was
determined   that   operational   modification  would   not  meet  PSCC's
requirements.  As  combustion  modifications  would be required before

                                 2-92

-------
strongly considering SCR or SNCR and neither technology has been proven
on large US coal fired boilers,  these technologies were not seriously
considered.
     The  effect of  each  applicable method  for  NOx  reduction  was
compared and a cost per ton of NOx removed was calculated. As PSCC was
not striving to meet any specific regulatory requirements, the analysis
was used to select  the technology that would provide the highest level
of economical NOx reduction.  LNCFS met PSCC's requirments for Cherokee
4 and Valmont 5 and was recommended for installation
     Before  proceeding with  the  recommended  modifications,  Energy
System Associates (ESA) reviewed the NOx reduction study and concurred
with the recommendation.

Low NOx Concentric Firing System Description

     ABB  Combustion  Engineering   (ABB-CE)   developed  the  Low  NOx
Concentric Firing System  (LNCFS) in the early  1980's  to increase NOx
removal to higher levels than achievable with overfire air alone. The
system is composed of three main features:

          1. Separated Overfire Air
          2. Offset Concentric Air Nozzle Tips
          3. New Coal Nozzle Tips

     The separated  overfire  air  allows  diversion of up to 30% of the
combustion air  above  the  main burner area.  This  allows combustion to
occur at lower stoichiometric ratios in the  main burner area and thus
reduces both the thermal  and fuel NOx.
     The auxiliary  air nozzles  used as  part  of LNCFS  are modified to
offset a portion of  the secondary air approximately 22 degrees  from the
furnace  diagonal.  This accomplishes  two  functions.  The  first  is by
directing  a  portion of the  secondary air away  from the  main flame,
excess air during the first stages  of combustion  is lower and thus NOx
generation  is  lower.  The second  is that the offset  auxiliary air
blankets  the walls with  a high O2 stream  and  thus  can  lessen the
affects of substoichiometric  combustion on tube corrosion. The velocity
of air across the tube wall also reduces wall  slagging.
     The coal nozzle tips are modified to  bring the flame front closer
to the nozzle. Some NOx is formed in a standard burner as the coal is
injected  into the  furnace before  combustion  is  initiated.  The coal
devolitizes  in this hot high O2  zone and NOx is  formed. By initiating
combustion  sooner,   less  oxygen  is  available  to  combine  with the
volatile nitrogen.
                                 2-93

-------
Unit Description

     Valmont 5 and Cherokee  4 are tangentially-fired boilers installed
in 1964 and 1968  respectively. Low  sulfur  western bituminous coal is
normally fired in both units but any combination of natural gas or coal
may be fired.  Valmont 5  uses a Bailey pneumatic control system and has
manual control of  the secondary  air dampers.  Manual  loading stations
to operate the overfire  air  dampers, overfire air tilt, and concentric
fire air were  added.  A  new Distributed Control System  is  planned in
1991 and automated controls  will be added.  Cherokee 4 uses a Bailey 721
electronic  analog control  system  and  has  automatic  control of  the
secondary  air  dampers.  A  Westinghouse  WDPF  system was  installed
recently for data acquisition. The new controls required to control all
dampers and tilts were added to the WPDF on Cherokee as an additional
drop on the system.  The table below lists major features of the units.

Electrical Generation
Steam Flow
Steam Pressure
Superheat /Reheat
Valmont 5
165MW
1,230,000 Ib/hr
1800 psig
1005°F/1005°F
Cherokee 4
350MW
2,587,000 Ib/hr
2400 psig
1005°F/1005°F
     Combustion Engineering Services,  Inc submitted a proposal for the
LNCFS  for  both units in  September  1989  and was  authorized  to begin
design in November 1989. The Valmont proposal included replacement of
several  sections  of  tube panels in an area  above the  burners due to
excessive tube  leaks  caused  by  hydrogen  embrittlement.  The LNCFS was
installed at Valmont  during  a planned six week outage  and was placed
in-service in May 1990. The work was accomplished using double shifts
due  to  the  short schedule. The  approximate  installed cost  of  the
Valmont  system including all PSCC  overheads was  2.5 million or $15/KW.
The  LNCFS was  installed at  Cherokee during a planned ten week outage
and was placed in-service in  November  1990.  One shift per day was used
at  Cherokee  due to the extended  schedule. The approximate installed
cost  of  the  Cherokee  system including  all  PSCC overheads  was  4.0
million  or $11.5/KW.  In addition to  the LNCFS retrofit, the economizer
was  replaced by another supplier  on the  Cherokee unit.
     Figures 1 and 2  compare  the windbox arrangement  of the Valmont and
Cherokee units before and  after the LNCFS  modifications were completed.
The  major  modification  in  the  windbox  arrangement,  other  than  the
addition of  separated  overfire  air, is  that  the original  single
auxiliary air and gas fuel compartments were  split into three separate
compartments to allow for two concentric fire air nozzles in the new
                                 2-94

-------
ORIC
I
n
C3
cr
a.
m
•d
EL
a
[E1
EEE
VALMONX #5
3 1 HAL LNCFS




TOP AUXILIARY AIR
COAL


COAL,
COAL
2 GAS •* AUXILIARY AIR
COAL
LOWER AUXILIARY AIR
m
III!!
TUTU

1
^ILJJ
-j— i
ii
rffff
ffl
TffT
ffl
1 	 1 .
m
m
H
OV^RKIRED AIR
TOP AUXILIARY AIR
COAL
CF AIR
CF AIR
COAL
CF AIR
CF AIR
COAL
CK AIR
2 GAS + AUXILIARY AIR
CF AIR
COAL
LOWER AUXILIARY AIR


DRI
L~~l

1 — I
"dJ-"
-JT1
nzzr
1
CHEROKEE #-4
SINAI, LNCFS

TO P AUXIL IARY AI R
COAL
COAL
A QAS - AUXILIARY AIR
COAL

COAL
4 GAS -i- AUXILIARY AIR
COAL


WT
im\
TTTTTI
Jltm
n
2
-LLLL
frff

nn.
HE
ffi
ffff
rm
LiiJ

TOP AUXILIARY AIR
COAL
CF AIR
CF AIR
COAL
CF AIR
3 OAS + AUXILIARY ATR
COAL
CF AIR
CF- AIR
COAL
CF AIR
3 GAS ^ A1JX1 LIAKY Al R
CF- ATR
COAL


Figure 1
Figure 2
installation. This limited the space available for the gas spuds on the
Cherokee unit so the original four gas spuds were reduced in number to
three per gas compartment. A new ignitor system was also purchased for
the Cherokee 4 unit  as the original  igniters  operations  had not been
meeting expectations.

Guarantees

     As part of  the contract  for  the design  and  installation  of the
LNCFS,  substantial  performance  guarantees  were  implemented.  The
guarantee for NOx removal was based on a sliding scale of percent NOx
removal dependent upon baseline values that were to be obtained before
the  shutdown.  In  addition to  NOx  removal,   boiler  efficiency  and
unburned carbon levels were guaranteed. Small allowances for decreased
boiler efficiency and increased unburned carbon levels were allowed for
variances in the testing  accuracy. All guarantees were applicable only
at the  full  load  conditions and  were based  on a  series  of  tests
completed shortly before and after the modifications.

Testing

     Energy Systems  Associates  (ESA)  of Pittsburgh,  Pennsylvania was
selected to perform  emission testing  as they  had performed other NOx
testing for PSCC  and  demonstrated a thorough knowledge of NOx formation
and  NOx  reduction  methods.  Although  both PSCC  and ABB  Combustion
                                 2-95

-------
Engineering  Services  have  very  capable  emission  testing  groups
qualified for this type of work,  an independent contractor was required
to  ensure impartiality.  In  addition  to  conducting the  acceptance
testing at full  load,  ESA performed sufficient testing  to  define an
operating procedure  across the  load range  that  would  minimize  NOx
emissions. ESA provided all equipment required for the measurements of
nitrogen oxides,  oxygen, carbon monoxide, and carbon dioxide. They also
provided the necessary equipment to collect EPA method 17 particulate
samples from  the outlet duct. These samples were used  to  determine
unburned  carbon  loss.  Plant  personnel  conducted boiler  efficiency
testing by the standard short form ASME method.

Valmont 5 Results

     In general the design, installation,  and testing on Valmont 5 was
completed on  schedule and  without significant problems. During  the
installation of the system it  was  discovered  that four of the existing
coal nozzles were damaged  beyond  repair.  The delivery  of new nozzles
was  expedited  and  were installed  as part of the outage. Soon after
startup it was discovered  that the new  coal  nozzle  tips  were binding
in  one corner of  the boiler. The unit  was brought off-line  and  a
portion of the nozzle  tip  side material  was  removed  to eliminate the
binding.  Increased slagging occasionally occurred during some testing
but the slagging  was usually associated with "unusual" damper positions
or  tilts. After  the correct operating procedures were  determined no
further slagging problems have developed. Heat  transfer in the boiler
appeared  to  be  unaffected  by  the modification  and there  was  no
significant change in economizer  exit gas temperatures.

Sampling  Locations
     Two  different  sampling  locations  were used  for the  emissions
testing on Valmont.  Baseline and guarantee testing were performed after
the  air heater  using  a fourteen  sample  point  matrix.  It  was later
determined  that testing   before  the   air  heater   would   be  more
advantageous  as the  data  can  be directly related the  combustion
process.  A nine point matrix was established  at the economizer outlet.

Baseline  Testing
     The  original NOx baseline testing was conducted over a period of
two weeks immediately before the outage.  Waiting until the final weeks
to  perform the  testing was  done  to  ensure  the data  would  be as
comparable as possible to post-installation values. The lower coal mill
was  taken off-line for major repairs  at the outset  of  the testing.
Using  the three available mills, testing was  completed at 80, 120, and
150MW. When operating with  three mills the lower auxiliary air and  fuel
dampers were approximately 25% open and the remaining  fuel/air  dampers
were equal at 40 to 60% open. Previous testing has shown that  a lower

                                  2-96

-------
baseline could have been achieved with the top mill out of service. The
coal mill repairs were rushed by working multiple shifts and the mill
was placed  in service two days  before  the outage began.  This  short
period  allowed baseline  testing at  full load;  however,   there  was
insufficient time for  four mill testing at the other loads. This proved
very important as the normal operation  of the  plant  is  four mills at
loads  above  120MW  and some  very  important baseline  data was  not
obtained.
     A series  of  ten  tests  were completed over a two week period to
define  the  NOx  emissions  across the  load  range.  Three  tests  were
conducted at reduced oxygen concentrations but the testing did not show
a strong O2  correlation. The baseline data will be  further discussed
in comparison to the LNCFS data presented below.

LNCFS Testing
     ESA conducted a series of 90 tests after the installation of the
LNCFS  in  order to document  fully  the changes  in boiler  operations.
ABB-CE  testing  personnel  organized  and  supervised  the  original
optimization  testing  and the guarantee  performance  testing while ESA
collected and organized  data.  After  the guarantees were met,  ESA
conducted all testing although ABB-CE advised and assisted throughout
the test program.

Overfire Air
     The  use  of  separated
overfire  air as  a  part of
the   LNCFS   can   provide
substantial  NOx  reduction.
Figure 3 documents the full
load testing. The amount of
overfire air is presented as
a damper position value and
not as an absolute air flow
value. The  unit  lacks flow
measurement    devices   to
measure actual air flow. All
concentric fire dampers were
closed  and   the  boiler was
operated with the auxiliary
air dampers  at approximately
60%  open  and  normal  O2 of  Figure 3
3.1%.  Note   that  with  all
overfire air dampers closed the  NOx  emissions were reduced  by 16% from
the baseline. This  reduction is  due to  air leakage  from the overfire
air  ports  and  the new  flame  holding  burners.  NOx emissions  were
gradually reduced as the dampers were opened until at  full opening, NOx
was reduced  by 36% from the original baseline.
Volmont 5 Overfire Air (§> Full Load
Aux 60%;CF 0%;02 3.1 %
1

CD 0-5
S
2

o °-4
z


"\-\.
"^^•^^i
\ ~ ~~i
^\.^J

i
I
0 10 20 30 40 50 60 70 80 90 100
Overfire Air Damper Position

I • Fuel 60% A Fuel 30%


                                 2-97

-------
     Overfire  air's  effectiveness was  substantially  increased  by
closing the fuel air  dampers to  30% at overfire air damper positions
greater than 70%.  Closing the  fuel air dampers  increase  the windbox
pressure and thus increase the air flow through the  overfire air ports.
This technique increased the
the original baseline.
     At   lower   load  the
overfire  air  is  generally
less   effective   as   the
windbox  pressure  reduction
decreases   the  amount  of
overfire  air.  This  can be
compensated by closing both
the  auxiliary  air and fuel
air  dampers   as   load  is
decreased.  Figure  4  shows
the NOx reduction  at a 150MW
load.  A  NOx reduction from
baseline of 38% was obtained
by opening  the overfire air
dampers   to  70%.  As  the
dampers   continue  to open
only a slight reduction is
achieved.
                             NOx reduction of  overfire air to 47% from
Valmont 5 Overfire Air @ Medium Load
0.6 >
i
0.55-
? 0.5-
5
2
\0.45-
° 0.4-
0.35-
0.3 H
C
Aux 30%; Fuel 10%; CF 0%; 02 3.4%
\
X^
^
\
\^_
™
i
10 20 30 40 50 60 70 80 90 100
Overlire Air Damper Posilion
                             Figure 4
Overfire Air Tilt
     Overfire  air  tilt was
originally held  at positive
tilts.   Testing  personnel
suspected  that keeping the
tilts  positive would  allow
for   increased   time  for
combustion     at    lower
stoichiometric ratios  thus
providing   for  lower  NOx
emissions.    Testing   was
completed  at both 100% and
66%   overfire   air   damper
positions  over a tilt range
of  -16 to  14  to determine
the  affect tilt had  on NOx
emissions.  Figure  5  shows
that  there  is minimal NOx  Figure 5
effect  for  various   tilts
although  it does show that minor  reductions occurred  in  -5 to  -10
range. The outlet sample matrix showed that neutral OFA tilts increased
NOx emission uniformity. This could explain the minor NOx reductions.
Valmont 5 OFA Tilt @ Full Load
Fue


5
.D
0 0.4-

, Aux 60%, CF 0%, 02 3%



^-~^^"'~


-20 -15


-10-5 0 5
Overfire Air Tilt
• 100% OFA + 66% OFA






10 15


                                  2-98

-------
     Overfire air tilt also affected superheat steam attemperation. The
amount of attemperation decreased at positive tilts. This is likely due
to the  decreased mixing  of the  furnace gas  that created  isolated
streams of cooler gas. The cooler gas streams would decrease superheat
heat transfer and thus lower required  attemperation.  During  the more
positive  tilt  testing,  one or  two of the  nine sample probes  would
experience significant carbon monoxide excursions though average CO was
not significantly increased. This  again  verifies less mixing  of the
flue gas.
Oxygen Concentration
     Figure 6 shows a series
of tests  that indicate the
variance of NOx emissions as
a function of boiler oxygen
concentration.   Sufficient
data for accurate comparison
was  only  available  for two
conditions   in   which  all
overfire  air dampers  were
fully  open.  In  one  case
there is a good correlation
with oxygen of approximately
78   ppm   NOx  per  percent
oxygen  change.  However,  in
the  second  test with the
fuel  air  dampers 30%  open,
there is no correlation with
oxygen .


D
2
5
_D
o035

2
-•- Fuel

Valmont 5 Oxygen @ Full Load
CF 0%, Aux 60%
,

*-" '"^^ ^—
^~r-^^
^^__^— + +
>-- •
6 2.8 3 32 3.4 3.6 3 S i
Percent Oxygen
307., OFA 67% •*• Fuel 30%, OFA 1007. * Fuel 0%. OFA 1007
60%, OFA 33% -A- Fuel 60%, OFA 83% !
i

Figure 6
Concentric Fire Air
     In much of the original
optimization   testing  for
overfire air,  the concentric
fire  (CF)  dampers remained
closed.   After   a    better
understanding of the correct
procedures for operating the
overfire air were obtained,
testing  began at different
levels  of  concentric  fire
air.  Figure   7  shows  the
affect on NOx  emissions for
different     levels    of
concentric  fire  air  when
operating at  full load and
100% overfire air.  The top





0

c


Valmont 5 Concentric Fire © Full Load
OFA 1 00%, Aux 60%


^^
• ~t— 	 '^
^m ^^n^^^- •

10 20 30 40 50 60 70
Percent Conccnlric Fire Air
• 02 3%, Fuel 0% + 02 3.6%, Fuel 30%

Figure 7
                                 2-99

-------
curve is at 3.6% oxygen with the  fuel  air  dampers open 30% while the
lower curve is at slightly reduced oxygen with less fuel air. Opening
the concentric  fire  dampers slightly  reduced NOx emissions  in both
cases. However,  NOx increased as the dampers are opened more than 20%.
     The increase  in NOx at high concentric fire damper  opening is
likely due to two  reasons. The first is  that as more air is added to
the main combustion zone the amount of  overfire  air is decreased. This
testing shows that "vertical" overfire air above the combustion zone
is more effective than the "horizontal" overfire air closer to the
combustion zone. The second reason is that furnace mixing was reduced
by the use of concentric  fire dampers. Figure 8 shows the uneven NOx
distribution at the boiler outlet  while using 66% concentric fire air.
Conversely, Figure 9  shows the distribution while using 22% concentric
fire air.  In this  case  the NOx is more evenly distributed across the
economizer outlet.
        Valmont 5 NOx Distribution
            Srtcrt [HE Modimujn BID L
  Figure 8
                                           Valmont 5 NOx Distribution
                                             Cone
                                    Figure 9
     The use of concentric fire air with  a high yaw angle does provide
 two  advantages. The  concentric  fire air  provides a  stream of high O2
 gas  next to  the boiler  wall  tubes. Substoichiometric combustion of a
 high sulfur coal  can cause  significant tube corrosion  due  to the
 formation  of  hydrogen sulfide gas. A  second  advantage of concentric
 fire air is the use of the high velocity  air to reduce  slagging of the
 wall tubes.  During  one  test at  the  Valmont site,  significant wall
 slagging occurred due to operation at substoichiometric  conditions. The
 slagging was  successfully removed by  closing the overfire air ports,
 directing  air to the  concentric fire air ports.

 Carbon Carryover and  Carbon  Monoxide
     Many  methods can be used to modify combustion to minimize NOx but
 a penalty  of increased carbon carryover and carbon monoxide emissions
 often exists. PSCC was very  concerned about increased  carbon monoxide
 and  fly ash carbon carryover. Guarantees were obtained to ensure that
 the  NOx reduction  would  be obtained at  minimum  operation penalty
 possible.
                                 2-100

-------
Valmont 5 Carbon Carryover

£
o
_Q
5 1 5-
T>
E 1 •
3 '
JD
c
D



.^^ •--.
--"^ Xj^

^^

80 90 100

110 120 130 HO 150 160 170 ISO
Nel Load (MW)
• LNCF •*• Baseline



     Fly ash  is sampled at
the FFDC hoppers to perform
weekly   boiler   efficiency
testing and the  Valmont unit
has  a  history  of  carbon
content    below   2%.    To
increase confidence  in the
carbon carryover values, it
was    decided   that    an
isokinetic  sample  would be
obtained from several points
in  the  duct to   obtain  a
representative  sample  for
this   testing.   Figure  10
shows  the  carbon  carryover
at various loads. In all but
the  minimum  load  testing,
carbon carryover is the same or lower after the LNCFS modification.
     Similar positive results occurred with carbon monoxide emissions.
Throughout  the  testing CO  emissions were below 30  ppm  and showed no
increase over pre-retrofit values. At conditions of high overfire air
tilt or  a  high  concentric  fire air damper openings, CO excursions in
a single sample point up to  1000 ppm did  occur.

Baseline' Comparison
     The guarantee test was completed early in the test schedule before
significant optimization testing  had  occurred.  A  summary of the test
results and conditions are shown in the table below. Also shown is an
optimized  test  that  provided  significant  NOx  reduction  at  better
operating  conditions.
Figure 10


NOx (Ib/MMBtu)
NOx Reduction
Boiler Efficiency
Unburned Carbon
Carbon Monoxide
Oxygen Concentration
OFA #3 Damper
OFA #2 Damper
OFA #1 Damper
OFA Tilt
Top Auxiliary Air
Fuel Air
Auxiliary Air
Concentric Fire Air
Lower Auxiliary Air

Baseline
0.664

86. 63
1. 6
<30
3.6%
NA
NA
NA
NA
50%
50%
50%
NA
50%
Guarantee
LNCFS
0.294
55.7%
86. 35
1.6
<30
3 . 6%
100%
100%
100%
+ 15°
100%
0%
50%
30%
100%
Optimized
LNCFS
0. 31
52.0%
NA
NA
< 0
3.6%
100%
100%
100%
-8°
50%
30%
50%
18%
0%
                                 2-101

-------
1
c
Valmont 5 Baseline LNCFS Comparison


3
E
]
D
£



^\
^\
"^\ "\J-__ _^A
~-^g^ B^^^--— *
"^a ^^^-^v



80 90 100 110 120 130 140 150 160 170 180
Nel Load (MVV)

• Orig 3 Mills •*• Orig 4 Mills a LNCF 3 Mills B LNCF 4 Mills


Figure 11
     A    comparison    of
baseline and optimized LNCFS
NOx  emissions  across  the
normal load range  is shown
in Figure 11.  Data is shown
for both three and four mill
operation. The post retrofit
values    show    that    a
significant  NOx  reduction
was  obtained  at the 120MW
load  by  removing  one  mill
from service.  This  is due to
the    increased    windbox
pressure   and   associated
increase  in  overfire  air
flow that occurs by removing
a  mill  from   service  and
closing  the associated  air register. Normal  operating  procedures at
Valmont are to remove one mill  from service between 120 and 130MW. Due
to the previously discussed problems with the  lower coal mill, baseline
data was  only obtained  with  four mill operation at  full  load.  It is
believed that the  four  mill  NOx baseline at  the 120  and  150MW loads
would  be  higher than the  three mill operation shown. Using the data
presented and the operating history of the unit,  it is projected that
NOx  emissions were  reduced by  the installations of  the LNCFS by 43%
across the load range.

Cherokee 4 Results
     The design  and  installation of the Cherokee 4 LNCFS conversion was
completed on schedule and without major incident. Due to the knowledge
gained from the Valmont installation,  it was  decided to  replace all
coal nozzles.  An outside engineering firm was  also retained to redesign
some coal piping hangers to  lessen pre-loadings  on  the  coal nozzles.
During initial operation several tilt shear pins were broken. The pins
likely broke due to either mechanical interferences with grating or due
to  a possible slag  buildup around the nozzles. Maintenance personnel
removed the interferences  and ABB Combustion Engineering Services, Inc
modified the  external drive  arms  to  reduce torque  on the  shear pins.
The modifications appear to have  corrected the problem.
     Unfortunately,  startup and  testing  of the unit  did not occur as
planned. Several startup problems not relating to the burners limited
generation  and  operation  of the  unit  while   firing  coal.  Three tube
leaks  in a short period resulted in unit outages to repair.  Due to
generation requirements the unit could not be  brought down immediately
for repair and  was operated on gas.
     When the unit was  operated  on coal,  a  significant  slag build up
occurred in the reheat section  of the boiler  within a  few days. It was
    2-102

-------
discovered  that  economizer  exit   temperatures   had  increased  by
approximately 100°F from the pre-installation values. A new economizer
was installed by a supplier other than ABB-CE during the outage.  Coal
sampling revealed that the ash fusion temperature of the coal had been
reduced  by  approximately  100°F from  the pre-installation  testing.
During the period of  slag  buildup the unit was  operated at full  load
for  long periods  to  accomplish  emissions  testing.  It  was  also
discovered that a third  of the  wall  soot  blowers were not operating.
Most blowers had  been out of  service for some time as  they  were not
required  to  control   slagging before   the  LNCFS  was  installed.
Significant wall  slagging was not  occurring but it is  possible  that
slag characteristics had changed enough to affect heat transfer to the
waterwalls.  The unit  was taken  off-line and the slag was removed by
manual means. Maintenance  efforts were also completed to get all the
wall blowers operational.
     The slag buildup  continued  to occur after  startup  and  the unit was
operated on gas until it could be brought  down for repairs. A dynamite
crew removed the  slag.  During this outage ABB-CE personnel installed
several thermocouples throughout the convection  section of the boiler
to define the  reasons  for temperature  increases.  It was decided to
limit unit generation to 90% and increase the oxygen to  approximately
4.5% to  minimize  flue  gas temperatures  in  the  reheat  section.  This
appears to have temporarily solved the slag  buildup problem.  The  unit
is currently operating without restriction or  significant slagging;
however, the unit is  on load  regulation  and is not operated at  full
load for long periods.
     It  is  currently  unknown  if  the  increased  economizer  exit
temperatures  and the slagging  problem  are  related to  the  burner
modification,   the   new   economizer,   the   change   in   ash  fusion
temperatures, or some operating  condition. Testing will continue until
this problem is resolved.

Sampling Locations
     Sample matrix  grids  were  installed at  the  economizer  and air
heater outlet.  Eight sample probes were installed at the outlet of the
air heater and  twenty sample  probes  were installed at the economizer
exit.  The  baseline  testing  showed  minimal  differences  in  emission
values between the two locations. It  was decided to use the economizer
sample location for all  testing.

Baseline Testing
     ESA conducted  the  original NOx  baseline testing over a ten day
period three weeks before the scheduled outage.  Testing was completed
to determine  the effect of  reductions  in  oxygen  concentration and
removing a  mill  from  service.  The  average NOx reduction from  a 1%
decrease in O2 was 56 ppm  (corrected to 3% O2).  The removal of an in-
service mill had  a  greater effect  as shown  in the  table  below. Mills

                                 2-103

-------
are identified by the letters A through E with mill A being routed to
the  lowest  burner  elevation.  The  mill(s)  removed  from  service  is
indicated in parenthesis.
Load
350
250
157
Mill Reduction
5->4 (C)
5->4 (A)
4(E)->3 (E,A)
NOx Reduction
15%
0%
33%
A reasonable NOx reduction is achieved by removing a mill from service
in all but the midload  test.  Testing at Valmont and previous testing
on  the  Cherokee  unit have  shown  significant  reductions  in  NOx at
midload by removing the top mill (E) verses the lower mill  (A).

LNCFS Testing
     Emission testing was not a  high priority after startup due to the
mentioned slagging problem.  ESA was on site for approximately two weeks
and  did collect  data  but  the  guarantee  performance  test  was  not
completed. Due to  the higher economizer exit  temperature,  the boiler
efficiency of the unit has been  notably decreased. The slagging caused
unusual gas  flow  and  carbon monoxide emissions  were higher than pre-
installation  values  with  significant  excursions.  ESA  is  currently
scheduled to begin  testing  in February  and will perform optimization
testing at lower loads.  The guarantee performance testing and full load
testing will be postponed until the slagging and temperature problems
are better understood.
Over/ire Air
     The  use  of  separated
overfire air has also proved
very effective  in combating
NOx  on the  Cherokee  unit.
Figure 12 presents data from
testing  at  350MW  with the
auxiliary dampers at 60% and
the fuel dampers  at 80%. On
the    Cherokee    unit   the
concentric fire air dampers
are automatically controlled
to  the same  opening as the
other auxiliary air dampers.
With   all   overfire   air
dampers   closed,   minimal
reduction from  the original
baseline occurred. A fairly  Figure 12
steep  drop  in  NOx occurred
Cherokee 4 Overfire Air <5> Full Load
0.55-
0 Si
?CU5-
m
\ 0 4 •
_a
| 0.35-
0.3-
0 25-
C
Fuel 80%
N*
\
X^




10 20 30 40 50 60 70 SO 90 100
Overfire Air Damper Position
                                 2-104

-------
until the overfire air dampers were opened 50%.  With all overfire air
dampers open,  a 47% reduction from the original baseline was obtained.

Oxygen Concentration
     Figure   13   shows  an
approximate 38 ppm reduction
in NOx for a 1% reduction in
02.    This    testing   was
completed with overfire air
dampers  fully open  and at
350MW.  This  data  compares
favorable with the baseline
testing   that   showed   a
similar   correlation  with
oxygen.    While    it   is
important  to   maintain low
excess     oxygen,     small
variances   do  not   greatly
affect NOx  emissions.
     This  data  also  shows
the effect of closing the  fuel air  dampers on NOx emissions. In nearly
every case  closing the fuel  air  dampers  increases  the NOx reduction.
This  is  likely due  to  an  increase furnace to windbox  pressure  that
increases the  overfire air flow.

Baseline Comparison
     The guarantee test has yet  to be completed  on Cherokee  4 due to
the  slagging  and economizer  exit  temperature problems.  All  current
indications show that the LNCFS can meet all guarantees  other  than
boiler  efficiency.   The baseline  values  are  compared  below  to  a
preliminary test of  the LNCFS  system  below.



m
2
2
> 0.275'
~Z-

Cherokee 4 Oxygen @ Full Load
OFA 100%; Aux & CF 60%


•tt ^ 	
*___4— — -— * """"" __^
A. 	 	


2.8 2.
9 3 3 1 3.2 3.3 3./I
Oxygen Concentration (%)
A Fuel 40% • fuel 60% ' Fuel 80%


3.5 3.6
Figure 13
Preliminary

NOx (Ib/MMBtu)
NOx Reduction
Boiler Efficiency
Unburned Carbon
Carbon Monoxide
Oxygen Concentration
OFA #3 Damper
OFA #2 Damper
OFA #1 Damper
OFA Tilt
Top Auxiliary Air
Fuel Air
Auxiliary Air
Concentric Fire Air
Lower Auxiliary Air
Baseline
0.533

88.87
2.2
<30
3.6%
NA
NA
NA
NA
78%
75%
75%
NA
100%
LNCFS
0.275
48.4%
NA
NA
<30
3.3%
100%
100%
100%
-10°
50%
80%
50%
50%
100%
                                 2-105

-------



3
CO
_D
0


1.
D

Cherokee 4- Baseline LNCFS Comparison
i
j
""\___ _^^^
*

i
\
i
c
0 200 250 300 !>-
Load (NMW)
Orig 3 Mills •* Orig 4 Mills • Orig 5 Mills s |_fCF 5 Mills


i
[



^

0


Figure 14
     Figure  14  presents  a
comparison of  the original
baseline  NOx emissions  by
load  compared  to  the data
that  is  available  for the
LNCFS  modification.   Data
obtained   with   one   mill
removed  from  services  is
also shown for the baseline
condition.   As   all  LNCFS
testing to date has occurred
at full load, no  LNCFS data
is available for  comparison
at  the  lower  loads.  The
approximate  reduction from
the  original  baseline was
48%.   Further   testing  is
planned in February at lower loads and with various overfire air damper
openings.

Summary

     Public  Service  Company  of  Colorado  installed  ABB  Combustion
Engineering  System's  Low NOx  Concentric  Firing  System to a 165 and a
350MW unit located in the Denver metropolitan area. The retrofits were
accomplished as part of a voluntary program to reduce the NOx emissions
of major metropolitan  area coal fired boilers by a minimum of  20%.
     The installation  and testing  of  LNCFS was  completed on schedule
and  without major difficulty  on  Valmont 5,   the  165MW unit.  NOx
reductions are greatest at  full  load  when the  separated overfire air
ports  are  most effective. As  load  is  decreased,  the overfire becomes
less  effective until  a mill  is  removed  from service.  The increase in
windbox pressure then restores effectiveness of  the overfire air ports.
Additional NOx reductions also can be obtained by partially closing the
fuel  air dampers  that decrease the stoichiometric ratio of combustion
and  thus  reduce  NOx.  Overfire  air tilt  did not greatly  affect NOx
emissions but did affect furnace mixing. The use of concentric fire air
did not significantly influence NOx emissions but did provide some help
in reducing  wall  slagging. NOx reductions  of over 50% are possible at
full  load  and  an  annual NOx  reduction of over  40% is expected due to
the LNCFS modification. No major changes  in boiler operation, unburned
carbon, carbon monoxide, or boiler  slagging have  occurred.
     The  installation of  LNCFS  on Cherokee 4,  the  350MW  unit,  was
completed without major difficulty but problems  with slagging and high
economizer  exit  temperatures have  limited testing.  It  is currently
unknown what has caused the  increase in  slagging and the higher boiler
exit  temperature. Although  the coal  source for  this unit  has not
   2-106

-------
changed, a  decrease in the  ash fusion temperature  of 150°F  and  an
increase in  the  ash content has occurred  over the  last  few  months.
Economizer outlet temperature comparisons  are  also more difficult  as
the economizer was replaced in the same outage as the  LNCFS. It is also
possible  that the  wall  slagging  pattern  has changed enough  that
currently  operating soot  blowers  are not  performing  effectively.
Although testing has  been limited, NOx  reductions  at  full load  are
about 48%. Additional  testing  is  planned on Cherokee  4 to  solve  the
slagging problems and gather more emission data.

Acknowledgements

     The authors would  like to thank Mr. Oliver Kruse, Valmont  Station
Manager,  and Mr. Jim  Stevens,  Cherokee Station  Manager, and  their
operating  and  engineering  staff  for  the  assistance and  patience
exhibited during the installation and testing of  these modifications.
Their assistance and flexibility in difficult situations were very much
appreciated.
     We would also  like to  thank  the personnel from  ABB-CE who have
reviewed  this paper  and contributed to  its content.  The  testing
personnel who spent long hours collecting and summarizing the data
presented,  often at  very inconvenient  times, are   also  gratefully
acknowledged.

References

Hawley  R.R.,  Collette  R.  J.  and  Grusha  J.  Public  Service  Co.  of
Colorado's NOx Reduction Program for Pulverized Coal Tangentially Fired
165 and 370MW Utility  Boilers.  Presented to  Power-Gen 1990,  Orlando,
Florida
                                2-107

-------
UPDATE 91 ON DESIGN AND APPLICATION OF LOW NOX COMBUSTION
       TECHNOLOGIES FOR COAL FIRED UTILITY BOILERS

                        T.  Uemura
                        S.  Morita
                        T.  Jimbo
                       K. Hodozuka
                        H.  Kuroda
            Kure Works  of Babcock-Hltachi  K.K.
               Kure,  Hiroshima,  737,  Japan

-------
                      UPDATE 91 ON DESIGN AND APPLICATION OF
          LOW NOx COMBUSTION TECHNOLOGIES FOR COAL FIRED UTILITY BOILERS
                                     T. Uemura
                                     S. Morita
                                      T. Jimbo
                                    K. Hodozuka
                                     H. Kuroda
                        Kure Works of Babcock-Hitachi K.K.
                            Kure, Hiroshima, 737, Japan
ABSTRACT

Babcock-Hitachi K.K.  (BHK)  has  been  using  the Hitachi-NR burner (HT-NR), which was
honored by the  Japan  Mechanical Engineering  Society  Award  in  1986, to solve prob-
lems of low-NOx emissions in coal fired boilers.
The  results   in  Japan  have  been  favorable.    Moreover,   BHK  has  contributed  to
European projects  which modified  existing coal  fired  boilers to  overcome future
stringent regulations by granting licenses to European boiler manufacturers.  These
results were  also successful.
Furthermore,  RHK plans to convert the HT-NR into it's second  generation  type  (HT-NR2)
as an extremely low-NOx boiler for the future.
In this paper the following topics will be introduced.

    1)     HT-NR for the newest 1000 MUe boiler

          --  Hatsuura power station unit-1 for the Electric Power Development
             Co., Ltd. (Japan)

    2)     HT-NR for the retrofits of existing boilers

          --  Plem Buggenum Maascentrale power station unit-5 for EPZ (The
             Netherlands)

          --  PGEM Njimegen power station unit-13 for EPON (The Netherlands)

          --  Inkoo power station unit-4 for Imatran Voima Oy (Finland)

    3)     Development test results of the HT-NR2 using a BHK 90 - 100 million
          Btu single burner testing furnace.
                                       2-111

-------
INTRODUCTION

Recently, Japan's low-NOx combustion technologies  have  achieved  remarkable  progress
in the field of utility boilers based on stringent  environmental  protection regula-
tions made by the government.
The main  counter-measures  are the combination  of  the low-NOx burners  and Two  Stage
Combustion  (TSC)  as  shown in  Figure  1.   In the  case  of most conventional  low-NOx
burners  which  simply lengthen  the  flame by means  of  delayed combustion,  however,
the combustion  efficiency  slows  down  and it becomes extremely difficult  to recover
the "trade-off" defect  between  NOx  reduction and increasing unburned  carbon  in  ash
(UBC).
Two Stage Combustion (TSC),  increasing the  residence time of combustion  gas between
the  sub-stoichiometric  burner  zone  and  After-Air  injection  point, can  accelerate
the  post flame  NOx  decay  and consequently  augment NOx  reduction.   In  general,
however, UBC tends to increase  and/or scatter, if  the residence  time of  the combus-
tion  gas  between  After-Air  injection  point  and furnace exit is  too short.  In  other
words, if TSC is  applied to  an  existing  boiler and  UBC  is required  to  be  kept  as  it
is, reliable NOx  control cannot be obtained  as shown in  Figure ?..
Therefore, the  above techniques  may  be  useful  only for  newly designed  boilers with
larger furnace  volumes  and  with high  capacities  (good  pulverizing  fineness)  in  the
coal  mills.
flethodol ogy

In  early  1981  Babcock-Hitachi  K.K.  (BHK)  started  further  development work on  NOx
reduction  burners  to  minimize the risks of  NOx-UBC "trade-off" caused by  the  con-
ventional  technologies, mainly  focusing  on  the In-Flame NOx Reduction  based on  the
principle  of High Temperature NOx Reduction, HT-NR.
Figure  3  shows  the flame structure  of  the  new low-NOx pulverized coal burner,  the
HT-NR burner, based on this concept.
The  volatile  NOx,  which  is  produced via volatile-N,  has  extremely large  chemical
reaction  rates  on  the  flame  front.   However, under  fuel  rich conditions, in  line
with  the  rapid  progress  of  02  consumption,  excessive  (overshooted) hydrocarbon
intermediates  (hydrocarbon  radicals, etc.)  come into  play and  contribute to  the
decomposition of ingredients.   It  has  been  confirmed, during  the  development work,
that  a   rapid  ignition and  higher   temperature  reducing  flame  can  accelerate  the
reactions  of  the NOx  decomposition.   The HT-NR  Burner  has the following  features
which can  achieve such conditions.

     •      Rapid  ignition  by the flame stabilizing ring with  ceramic  parts

     •      Separation of external air by the guide sleeve

     •      Promotion of the air-chars mixing in the  post flame  zone due to
           higher swirling of external air


APPLICATION

The  low  NOx combustion technology of HT-NR burner is  now widely  employed  to only in
Japan but  also in Europe  and other nations.


New  Boiler Unit  with HT-NR Burners

flatuura  P.S.  lu  1000  Mile for  Electric  Power  Development  Co., Ltd.  the  newest  and
largest  capacity unit  in  Japan,  started  its'  commercial operation in  June  of  1990.
(See Figure 4)


                                        2-112

-------
This  boiler,  manufactured  by BHK,  is equipped  with a  low-NOx combustion  system
(HT-NR burners + TSC).   (See  Table 2)
NOx emission from the stack is minimized by the DeNOx  equipped  after  the  Air  Heater
Planning coals have a wide  range  of properties as  shown  in  Table  2.
In  general,  the  higher  the Nitrogen  content  (N)  in coal  and  the  higher  the  Fuel
Ratio  (= Fixed Carbon/Volatile flatter), the more difficult  it  is  to achieve low-NOx
combustion.  (See Figure  5)
Though this unit uses coal  with the highest Nitrogen  content  and  highest  Fuel  Ratio
among  all  imported  coals  usually  used in  power plants in Japan,  excellent  low-NOx/
NOx-UBC  combustion  performance has been continuously maintained since its  commis-
sioning.   (See Figure 6)
Operation  techniques  of  combustion  control   are  also important  for  achieving  and
maintaining stable  low-NOx  combustion  performance.
In  this  unit,  an operating-assistance system, the computer aided combustion  moni-
toring system, is installed to support plant  supervisors.
For example, the data from  multi-eye  flame detectors  and  automatic traverse exhaust
gas analyzer installed  in the economizer outlet are  very  useful  for achieving  opti-
mum  fuel/air  distribution  in  the furnace  which prevents  NOx-UBC from scattering.
Figure  7 shows  a  CO  profile  leaving  economizer   obtained  by  the automatic  "grid"
sampli ng.
Fig.  8  is  a  picture of  the flames in a 1000  MUe boiler  operating at  the  guaranteed
minimum  load condition.   Stable combustion flames  were confirmed.
Generally, in  the  minimum-load operation of  the  boiler, it becomes  more difficult
to  get  stable  combustion  because  the  amount  of heat  radiation  to the cloud of  pul
verized  coal  particles  is reduced.  Moreover,  in  the burner's  low load zone, reduced
C/A ratio  (coal/primary  air)  causes unstable  combustion.  Since  the HT-NR Burner  is
equipped  with  a "flame-stabilizing  ring"  at  the  tip of the  fuel  nozzle, we  can
reduce the exclusive coal firing  minimum load without  adding  extra equipments.


Low-NOx  Retrofit Using  HT-NR  Burners

There  are several  cases  of   low-NOx  retrofits  for  existing  boilers  using  HT-NR
burners  as shown in Figure  9.
The low  NOx  combustion  systems  of existing  boilers may  be  classified into  the  fol-
lowing three types.

    •     Non    "Low-NOx"

    •     Conventional  Low-NOx Burner

    •     Conventional  Low-NOx Burner with TSC


In  Japan,  most  existing  P.C. fired  units  already  operate with both  conventional
low-NOx  burners  and TSC  systems.
From  1984  to 1987,  BHK  has  replaced  existing  low-NOx burners  with Hitachi-NR  burn-
ers in several  units.
In Europe, on the other hand, most of the existing  coal-fired  combustion  units  were
not the  "low-NOx" type.
From  1988  to  1989, Stork  and Tampellar, Dutch   and  Finnish  boiler manufacturers
respectively, modified  their  three existing non-low-NOx  burners  into  HT-NR  burners.
Two  of  these  units  also  incorporated  two-stage  combustion.   Results  of  these
modifications are shown  in  Table  3.   In  this record,  the results of  plants A and B
in Japan are the same as  introduced at the symposium  in  1987  and  1988.
Of the  three modified  units  in Europe, the  Inkoo  PS boiler No. 4 of Imatran  Voima
Oy (IVO) Power Company  of Finland is  shown below.
                                       2-113

-------
A low-NOx modification of an Inkoo PS boiler  No. 4

    •     This is a Benson type boiler with an output  of  265  MUe  ;  a  two-stage
          combustion system of the HT-NR burner and  overfire  air  were employed
          in 1989 for the low-NOx combustion  system.

    •     A total of 16 HT-NR burners were arrayed in  four  stages  and four
          rows in the furnace rear wall  and eight after-air ports  were placed
          in the front wall and rear wall, four pieces  each.

    •     The two-stage combustion system was designed  to feed  0  to 25 %  of
          the air  into  the after-air ports.    Therefore, the  excess  air  factor  of
          the burner unit could be varied in  the range  of 0.95  to  1.25.


 In August 1989, two months after the start of operation,  the  performance  was  tested
 by using  two  types of coal, and  the NOx  level  and  unburnt content  in ash were  as
 shown  in  Fig.  10.  As compared with  the  operation  before modification, there was  a
 reduction of about 50 %.  The slagging was unchanged.


 Design and Estimation

 To achieve  extremely  low-NOx  combustion,  the combination  of TSC  and  HT-NR burners
 is most effective and useful.
 The basic equation  in  Figure  11,  which  is  used  for  design  and  estimation, explains
 that the  Total  Fixed  Nitrogen (= No +  HCN  + NH3 +  N  in char) should  be  minimized
 before After-Air injection.

    •     Higher Heat Release Rate in burner  zone (BHR) raises  thermal  NOx
          formation.

    •     Extremely low stoichiometric ratio  of the  burner  zone (SRg)  could
          produce slagging and/or corrosion especially  with  fusible and high
          sulfur coal/ash.

    •     Higher In-Flame NOx Reduction efficiency (TINR)  ^s  obtained  by using
          higher volatile coals, and rapid ignition  flame conditions  are  pro-
          moted by finer pulverizing.

    •     TSC effect is advanced by  lengthening the  residence time  of the
          reducing gas between burner zone and after-air  injection.


 Development of Hitachi-NR2 Burner

 At BHK, we  are  now developing  a  super  low-NOx burner  (HT-NR2 burner)  aimed at  even
 lower  NOx to  cope  with  the needs  of the coal  fired  thermal  power  of  the  next gene-
 ration.   The  new  burner  based  on  the  principles of the  HT-NR burner, intensifies
 the ignition and expands the reducing flame.  A structural  diagram  of HT-NR2  burner
 is shown in Fig. 12.

     Features
     •    The basic principles are the same as in HT-NR burner

     •    The following mechanisms 1. and 2. have been added.

          1. Intensification of ignition


                                       2-114

-------
             Formation  of  stable  combustion  and  elevation  of  temperature  by
             primary air velocity  control  shell  (pulverized  coal  concentration
             regulator).   The  pulverized  coal  around  the  velocity control
             shell  is supplied  into  the  flame  stabilizing  ring  by inertia,
             while  primary  air  flow  is diffused  at  the  front  of the  velocity
             control shell.  Therefore,  a  flow of  pulverized  coal  of  high  con-
             concentration  is  led  to the  flame stabilizing  ring,  intensifying
             the  ignition.

          2. Expansion  of  reducing  region

             Expansion  of  the  reducing region  by tertiary  air separator
             By widening the aparture of  the tertiary air  feed  port  and
             separating of  the  secondary  and tertiary air  feed  ports,  the
             reducing region in the  flame  is expanded.


Figure  13  shows  the  results  of measurement of  behavior  of  gas  concentration  near
the burner,  using HT-NR burner and NR2 burner,  in  a  combustion furnace with  a  coal
corn-combustion  capacity  of 500 kg/h.  Uhile maintaining high combustion  efficiency
by  promotion  of ignition,   the  NOx  decomposition reaction in the  flame is  promoted
also.   Figure  14  shows  the combustion  test   records of  a  full-scale burner  (coal
combustion  capacity 4000  kg/h)  using  a   large-sized  combustion test furnace.   The
NOx and  unburnt  carbon  characteristics  of the HT-NR2 burner  have  been proved to be
more outstanding  than those of  the  NR burner.


CONCLUSION S RECOMMENDATIONS

Babcock-Hitachi  K.K.  has  developed  and  installed  extremely  low-NOx burners, HT-NR
Burners, with successful results.
This burner concept of  "In-Flame  NOx Reduction"  is  highly  effective  both  for  exist-
ing boilers and future  boilers.
Use  of   In-flame  NOx  reduction  technology will  play  a  major  role  in   successful
retrofitting of existing old furnaces.
On  the   other  hand, it  will  be  applied  to new  boilers  suitable  for present  and
future needs.
Furthermore,  in  order  to meet  the  potential  needs of the next-generation  of coal-
fired power plants, BHK is  working  toward  further  NOx reduction and  is developing  a
burner with  extremely  low  NOx  emissions  (HT-NR2 Burner).   Advanced performance of
the  HT-NR2  Burner has  been  confirmed   by  using  a large   scale  combustion  test
faci1ity.


REFERENCES

1.   S.  Morita et al., "Update on Coal  Combustion Technologies", the  Hitachi  Hyoron,
     vol. 72 - No.  6, 1990.

2.   S.   Morita  et  al .,  "Design  Methods  for  Low-NOx Retrofits of Pulverized  Coal
     Fired Utility  Boilers",  EPA/EPRI  Joint Symposium on Stationary Combustion  NOx
     Control, 1989.

3.   S.   Morita,   "Low-NOx  Combustion Technology  of Pulverized Coal   Fired  Utility
     Boilers", Journal of  the Japan  Boiler  Association, No.  231-10,  1988.

4.   I.   Eknan  et  al .,   "Desul phurization  by Limestone  Injection  combined  with
     Low-NOx Combustion",  GEN-UPGRADE 90.
                                        2-115

-------
UBC    NOx
                Combust i on
                Completion
                      Zone
                                    Technical  Point
                                  •Mixing of After-Ai r
                                   Longer Res idence Time
                       ig.1   Concept  of low-NOx Furnace
          Design Condition
             • The Same Coal Property
             •The Same Pulverrizing
             •The Same UBC Requi rement
 New Technology


Dual/2rows
  After-Air Port

 Advanced TSC
                                   In-Flame NOx Reduction     Hitachi-NR Burner
UBC
NOx

^
^
SRBNR
                                                   150ppm
                                                   V


                                                  200ppm
                                                  •^

                                               N0x = 300ppm
                            RT1- NOx
                   fig.2  NOx-UBC  "Trade-Off"
                              2-116

-------
iic j n       High Swirl Re;
Wind Box  ^

                 Guide Sleeve
Fl ame
Stabilizing
Ring
                                                         Vo I at i I e
                                                                       + 0;
NO
                                                     Axial  Distance  from  Burner
                           Fig.3   Hitachi-NR Burner
                                  2-117

-------
          SECONDARY  „
          SUPER HEATER
                 \
                                               REHEATER
                                               PRIMARY
                                                SUPER  HEATER
                                                ECONOMIZER
                                     PR I MARY l| 11 SECONDARY
                                 ir—AIR HEATER lAIR HEATER
                                         "^
              Turbine  Output               1, 000  MWe
              Evaporat ion                  3, 170  t/h
              Super  Heater  Out let  Press.   25  MPa
              Super  Heater  Outlet  Temp.    543/5691)
              Fuel                         Coal
Fig.4  Matsuura  P.S.T  (Electric Power  Development  Co.,Ltd)
         *)  FN(Relative  NOx Emission Facter)

                                   FN=1.2        Higher  NOx
                          FN = 1.
            *)
            ^k \        C-Coal
                                                  PIanning coaI
           Lower  NOx
                     NOx  ^  200ppm
                     UBC  ^  5%
                      FR
        2
 Fixed  Carbon
Volatile  Matter
   Fiy.5  A  Relative NOx  Emission  (Matsuura  1", lOOOMWe)
                        2-118

-------
      Eco  02  = 3 — 3. 5  (%. dry)
   C-Coal  /
FR = 1. 9. N = 0. n
Ash=9S

D-Coal
   A-Coa I
FR = 2. 6. N = 2.
Ash=15X
0
  FR=I. 0. N=1. 5%
  Ash=8°/o         FR = 2. 2. N = l. 8%
  LA^-  _         Ash=]X
       100
                               150
               iler Out I et  NOx (ppm. 6%02)
                                                     'Guaranteed  Point
                                                     FR  -    Fixed Carbon     ,_,
                                                           Volati le  Matter   [~'
                                                     N       :  dry a sh  f re e
                                                     Ash     :  dry base
                                                     Eco  02 :  02  Leaving
                                                                      Economizer
             200
     Fig.5  Operation Results of Guaranteed Coals  (Matsuura r. IDOOWtVe)
 05A-5
                                                    90/02/27 19:30    «»MU
                                tJ.'tl It (MCtt  H UI1U 90/02/08 19:04:00 1001MU
                   90/02/08  17:53-18:45
                   1001 tlW
                      9 0 \~ff.t.  |  ->9'>J h-f,
        Cl  C2  C3  C4  C5   C6   C7  C8  C9  C[0
 CO
         Fig.7   A  Graphic  CRT of  "Grid-Measurement"  Leaving Econimizer
                                   2-119

-------
Flame  Condition  at Minimum Load
Operat ion in  the  1000 MWe Boiler
        2-120

-------
 —I  Cond it ion  }	
  Same Furnace Volume
  No  Derat i ng
-t —
 Ci rcular  Register Burner
(Non-"Low-NOx")
Dual  Register  Burner
(Conventional "Low-NOx")
Hitachi-NR  Burner
After-Air  (OFA)  Port
<
< >
< >
< >-
\x
600 MWe (1989)
.
\ 180 MWe (1988)
<
* jj
i 3-
{ 1-
V

 80  t/h   (1984)
 80  t/h   (1985)
200  MWe   (1985)
200  MWe   (1986)
350  MWe   (1987)
                  265 MWe   (1989)
           Fig.9   Menu of Retrofits  with  Hitachi-NR Burners
                                             100% boiler load
                                             three upper  burner levels  in  service
                                             tota I air factor 1. 22


                                               __^          Col umbi an coal
                                               o           A before retrofit
                                               «=>           A af ter retrofit
                                               e
                                               o.
                                               —'           Amerlean coal
                                               5          •  before retrofit
                                               z          O after retrof it
        0.9        1.0       1.1       1.2
       Stoichiometric  ratio  on  burnur  level
          Fig.10   Example  of  Results of NOx/UBC Performance
                               2-121

-------
              = k-exp(BHR)-f(SRB)-Ndaf-(1-  7NR)-exp(-RTi)
  Total  Fixed  Nitrogen    Heat release  j  Fuel-N
(Before Injection After-Air)    R3te       i
                                   Sto i ch i ometry
     -- AAPs	
     -• BNRs '-
                                                    Gas  Residence
                                                        T i me
                  NOx.Fi nal
                           Re-formation ?NR  = rj  (VM)-(TRF)
                              RT1
                                    T
I n-F I ame
Re due t ion
   by
 HT-NR
 Temperature
    of
Reducing Flame
                          TFN
                                                  Volatile  Matter
         Fig.11   Key Equation for   Low-NOx Combustion  System
                          Wi nd Box
VeIoc i ty Control

      Space Creator
          FIame Stab i
 Oil  Atomizer
              Pry Air+P. C.
                                                     Space Creator
                                 Ve loc i ty  Control She I I
              Fig.12  Features of  Hitachi-NR2  Burner
                           2-122

-------
CD
,uuu
900
800
700
600
500
400
300
200
100
0

•— • — '"^
-Hilddii-NR2 XT |mprovement of
,*/* combustion efficiency
I // — -— --1 	 n 	 <^— 	 ->
- \.¥^ ^R
- , A A- <> •—
\/ M IMK promotion of
^ / \^^ NOx reduction"
i » i i
D 1 2 3 4
j»» m —
sduct ion of Char-NOx
S3
Reduction of
boi ler outlet NOx
,/r—^l
6
-
Furnace f
IUU
90 5
80 X
o
70 g
60 o
50 £
40 c
o
30 -M
00
20 _g
10 S
0
;xi t
                  Distance  from burner
    Fig.13  Behavior  of NOx Combustion  Efficiency in  the Furnace

            by Using  a  Post-Period Super Low-NOx Burner
             10
         to
         ro
         c
         O
T3


E   2
D

C
               60
                                      Coal:Newlands

                                           (FR=2.3)
                                      Hitachi-NR
                                           Hitachi-NR2
                   100     120    140    160  180


                   NOx (ppm,6%02)
    Fig.14  NOx and Onburned  Carbon Performance  of Hitachi-NR2  Burner
                             2-123

-------
Table 1   Specifications of P. C.  Combustion  Equipment
                                 (Matsuura  1U. lOOOMWe)
Mil 1
Burner
AAP
Type
Quant i ty
Capac ity
Loading Pressure
Classifier
Type
Quant ity
Heat Input
Arrangement
Type
Arrangeme nt
MPS-118 (B&W Type)
7 (1 for Stand-By)
95.3 t/h.Mill (HGI=50)
12. 4 MPa (Oil Press.)
F i xed Stationary Type
Hitachi-NR
70
152. 8 x 10
10 bnrs x
5 bnrs x
Dual Type
10 AAPs x
Burner
(10 for
6 kJ/h
3 rows, Opposed
1 row, Opposed
1 row, Opposed
Stand-By)
(Top Row)

Table 2   Planning Coal  (Matsuura  P. S.  1U. lOOOMWe)
1 ten
GCV
Proximate IM
VM
FC
Ash
S(Total)
F R
N
IDT
Base Unit
A. D J/kg
A. D %
A. D %
A. D %
A. D %
A. D %

d. a. f %
Oxi. °C
Desi gn Coa 1
> 25. 1



< 20
< 1. 0
< 2. 4
< 2. 1
> 1200
PI anni
Min
24.2
1. 5
23. 9
38. 5
3. 5
0.2
1. 1
0.8
1210
ng Coal
Max
29.0
9.6
40. 6
59. 1
26.0
1. 3
2.4
2. 1
> 1500
 Table  3   Low-NOx  Retrofit  of  Existing  Coal  Fired  Boilers
Plant
A
B
C
0
E
Capac ity
200MWe
350MWe
180MWe
265MWe
600MWe
Volati le
Content (%, daf)
34-
30-
23-
35-
30-
56
40
40
40
40
NOx Reduct ion
(Commerc i a 1
35-
25-
40-
40-
30-
Efficiency (%)
Operat ion Base)
45
30
50
50
45
Remarks
1985-86
Japan
1987
Japan
1988
Europe
1989
Europe
1989
Europe
                        2-124

-------
             Session 3




 LARGE SCALE COAL COMBUSTION
Chair:  D. Eskinazi, EPRi and R. Mali, EPA

-------
DEMONSTRATION OF LOW NOX COMBUSTION CONTROL TECHNOLOGIES
             ON A 500 MWe COAL-FIRED UTILITY BOILER

                          Steve M. Wilson
                           John N. Sorge
                      Southern Company Services

                          Lowell L. Smith
                          Larry L. Larsen
                  Energy Technology Consultants, Inc.

-------
                  Demonstration of Low NOx Combustion Control Technologies
                             on a 500 MWe Coal-Fired Utility Boiler
                                        Steve M. Wilson
                                         John N. Sorge
                                   Southern Company Services

                                        Lowell L. Smith
                                        Larry L. Larsen
                               Energy Technology Consultants, Inc.
A DOE Innovative Clean Coal Project (ICCT) Project was awarded in 1989 to demonstrate retrofit
technologies to control NOx emissions on a 500 MWe wall-fired boiler. The primary objective of the
project is to demonstrate the control effectiveness of Advanced Overfire Air (AOFA), Low NOx Burners
(LNB) and the combination of these technologies under short-term controlled and long-term load dispatch
conditions. The project involves four test evaluation phases  Baseline, AOFA, LNB and LNB with
AOFA.  Each phase will evaluate NOx control effectiveness and the impact of the technologies on boiler
operation and backend cleanup equipment.

This paper provides an overview of the program test plan and instrumentation for the four phase program.
The test results from the Baseline and Overfire Air Port retrofit phases of the program are presented.
Comparisons are made between the short-term controlled test results and the long-term load control test
results for these phases.  Comparisons are also made between the Baseline and AOFA long-term retrofit
results to establish the NOx control effectiveness for this technology.
                                             3-3

-------
INTRODUCTION

This paper describes the technical progress of a U.S. Department of Energy (DOE) Innovative Clean Coal
Technology (ICCT) Project demonstrating advanced wall-fired combustion techniques for the reduction
of nitrogen oxide (NOx) emissions from coal-fired boilers. The project is being conducted at Georgia
Power Company's Plant Hammond Unit 4 near Rome, Georgia.

The project is being managed by Southern Company Services, Inc. (SCS) on behalf of the project co-
funders: The Southern electric system, the U.S.  Department of Energy (DOE), and the Electric Power
Research Institute. In addition to SCS, The Southern electric system includes five electric operating
companies: Alabama  Power, Georgia Power, Gulf Power, Mississippi Power, and Savannah Electric and
Power.  SCS provides engineering and research services to the Southern electric system.

The Innovative Clean Coal Technology Program is a jointly funded effort between government and
industry to move the most promising advanced coal-based technologies from the research and
development stage to the commercial marketplace. The clean coal effort sponsors projects which are
different from traditional research and development programs sponsored by the DOE. The traditional
projects focused on long range, high risk,  high payoff technologies with the DOE providing the majority
of the funding. In contrast, the Clean Coal project objective is to demonstrate commercially feasible
advanced coal-based technologies which have already reached the "proof-of-concept" stage. As a result,
the clean coal projects are jointly funded endeavors between the government and the private sector which
are conducted as Cooperative Agreements in which the industrial participant contributes at least fifty
percent of the total project cost.

The primary objective of the Plant Hammond demonstration is to determine the long- term effects of
commercially available wall-fired low NOx combustion technologies on NOx emissions and boiler
performance. Short-term tests of each technology are also being performed to provide engineering
information about emissions and performance tends. A target of achieving fifty percent NOx reduction
using combustion modifications has been established for the project. The project seeks to address the
following objectives:

       1)    Demonstrate in a logical stepwise fashion the short-term NOx reduction capabilities of the
            following advanced low NOx combustion technologies:

              a) Advanced Overfire Air (AOFA),
              b) Low NOx Burners (LNB),
                                             3-4

-------
              c) LNB with AGFA.

       2)    Determine the dynamic long-term emissions characteristics of each of these combustion
            NOx reduction methods using sophisticated statistical techniques.

       3)    Evaluate the progressive cost effectiveness (i.e., dollars per ton NOx removed) of the low
            NOx combustion techniques tested.

       4)    Determine the effects on other combustion parameters (e.g., CO production, carbon
            carryover, paniculate emission characteristics) of applying the NOx reduction methods listed
            above.

PROJECT DESCRIPTION

The stepwise approach to evaluating the NOx control technologies requires that two plant outages be used
to successively install (1) the advanced overfire air ports and ducting and (2) the Foster Wheeler
Controlled Flow/Split Flame (CF/SF) low NOx Burners. These outages were scheduled to  coincide with
existing plant outages in the spring of 1990 and the spring of 1991.  The final LNB retrofit  outage will be
completed by late April 1991.

Following each major retrofit outage, a series of four groups of tests are performed  (1) diagnostic,  (2)
performance, (3) long-term and (4) verification. The diagnostic, performance and verification tests
consist of  short-term data collection under carefully established steady-state operating conditions.  The
diagnostic tests are designed to map the effects of changes in boiler operation on NOx emissions and
establish NOx  trends. The performance tests are used to evaluate a  more comprehensive set of boiler and
combustion performance indicators including paniculate characteristics, boiler efficiency, and boiler
outlet emissions. Mill performance and air flow distribution are also established during the performance
testing.  The verification tests are used to establish whether any changes in NOx emission trends might
have occurred during the long-term test phase.

One of the major objectives of this demonstration project is  to collect long-term, statistically significant
quantities of data under normal operating conditions with and without the various NOx reduction
technologies. Earlier demonstrations of combustion emission control technologies have relied solely on
data from a matrix of carefully established short-term (one to four hour) steady-state tests.  Utility boilers
seldom operate in this steady-state manner considering the dynamic nature of the plant equipment
operation needs and economic dispatch strategies employed. Due to this dynamic mode of  operation,
                                              3-5

-------
statistical methods have been developed for use with long-term emission and operational data that allow
determination of the achievable emissions limit (Ref.  1) or emission tonnage of a control technology (Ref.
2). These analytic methods have been developed over the past fifteen years by the Utility Air Regulatory
Group (UARG) - Control Technology Committee. Data collection criteria used in these methodologies
are now accepted as benchmarks for establishing  the achievable SC>2 and NOx emission limits.  These
criteria along with other criteria established by EPA will be used to determine the achievable NOx
emission level in each of the four operating conditions for Hammond Unit 4 - Baseline, AGFA, LNB and
LNB with AGFA.

The major emphasis of this paper is on description of the NOx Characteristics resulting from the Baseline
and AGFA retrofit short-term and long-term test efforts. Both short-term and long-term testing and
analysis have been completed for the Baseline configuration and have been thoroughly documented in
Reference 3. The short-term and long-term test efforts in the retrofit AGFA configuration have been
completed, however, detailed analysis of the data is still in progress as of the date of publication of this
paper.  Consequently, information provided in  this paper relative to the AGFA retrofit is preliminary and
may be revised upon publication of the final DOE Interim Test Report.

BOILER AND AGFA DESCRIPTION

Boiler Description. Hammond Unit 4 is a balanced draft Foster Wheeler Energy Corporation (FWEC)
opposed wall-fired boiler rated at 500 gross MWe with design steam conditions of 2500 psig and
1000/1000 °F superheat/reheat temperatures, respectively. Six FWEC Planetary Roller and Table type
MB-21.5 mills provide pulverized eastern bituminous coal (12,900 BTU/lb, 33% VM, 53% FC, 1.7% S,
1.4% N) to 24  Intervane burners. The burners are arranged in a matrix of 12 burners (4W x  3H) on
opposing walls with each mill supplying coal to four burners in an elevation. The unit is equipped with a
coldside ESP and utilizes two Ljungstrom air preheaters.

AGFA Description. Figure 1  schematically illustrates the AOFA retrofit on Hammond Unit 4. The
AGFA system consists of ductwork on each side of the boiler extending  from immediately downstream of
each secondary air venturi to a separate overfire air windbox located above the burner windbox. Eight
FWEC can-in-can overfire air ports supply preheated  air directly above each burner column on opposing
walls. The ductwork and OFA ports were designed to provide improved overfire air penetration across
the furnace.  In addition, numerous dampers were provided to optimize the flow distribution of the
system.  The AOFA system incorporates four sets of OFA flow control devices, 1) windbox/AOFA flow
proportioning dampers, 2) AOFA guillotine shutoff dampers, 3) AOFA flow control dampers and  4) can-
in-can distribution dampers. The windbox/AOFA proportioning dampers are set in one fixed position at
                                             3-6

-------
commissioning. The guillotine shutoff dampers are used only for isolating the AGFA system from the
secondary air supply. The AOFA flow control dampers and the can-in-can dampers are used to modulate
the OFA flow and the distribution of the air across the furnace. In addition to these control dampers,
curtain air was supplied to protect the furnace walls.

TEST INSTRUMENTATION DESCRIPTION

A complete data acquisition system (DAS) was installed during the fall of 1989. This custom designed
micro-computer based system is used to collect, format, calculate, store, and transmit data derived from
power plant mechanical, thermal, and fluid processes. The extensive process data selected for input to the
DAS has in common a relationship with either boiler performance or boiler exhaust gas properties.

The DAS includes a continuous emissions monitoring system (NOx, SC>2, C>2, THC, CO) with a multi-
point flue gas sampling and conditioning system, an acoustic pyrometry and thermal mapping system,
furnace tube heat flux transducers, and boiler efficiency instrumentation. The instrumentation system is
designed to provide data collection flexibility to meet the schedule and needs of the various testing efforts
throughout the demonstration program.  A discussion of the various instrumentation follows.

Extractive Continuous Emissions Monitor fECEMI.  An underlying objective of the ICCT project is to
evaluate the long term effectiveness  of retrofit NOx control technologies.  The Extractive Continuous
Emission Monitor (ECEM) provides the means of extracting gas samples for automatic chemical analysis
from sample points at strategic locations in the boiler exhaust ducts. The system quantitatively analyzes
gas samples for NOx, SO2, CO, O2, and total hydrocarbons (THC). The results from the five analyses,
along with the status of the ECEM, are continuously transmitted to the DAS computer where the data is
processed and stored. The ECEM comprises sample probes and lines, a sample control system consisting
of valves and distribution manifolds, pumps, sample conditioners (filters, condenser/dryer, pressure
regulation  and a moisture detector), flowmeters, gas analyzers and an automatic calibration system.
Automatic or manual calibration is achieved by sequentially introducing certified gases of known zero
and span value for each analyzer.

Acoustic pyrometer. The acoustic pyrometer is a micro-computer controlled system that transmits and
receives sonic signals through the hot furnace gas above the fireball from multiple locations around the
girth of the boiler. The acoustic pyrometer provides average temperature data for straight line paths
between any two transceivers not located on the same furnace wall. The acoustic pyrometer provides a
means of analyzing the variations in  the combustion process. The velocity of the sonic pulses is used to
compute an average path temperature between two transceivers which, when combined with the other
                                            3-7

-------
path temperatures, allows computation of isotherms at the plane of acoustic pyrometer transceivers.  At
Plant Hammond Unit 4, the horizontal plane of the transceivers is approximately 15 feet above the
uppermost elevation of burners.

Fluxdome Heat Flux Sensors.  The DAS instrumentation includes heat flux sensors that detect the heat
absorption into the boiler's furnace wall tubes at strategic locations in the furnace. These flux
measurements are intended to provide an indication of both the furnace combustion gas temperature and
the condition of the wall ash deposits in the near-burner zone. Comparisons of the flux measurements
during the various phases of retrofit may indicate whether any beneficial or undesirable effects on the
furnace wall tubing is associated with the low-NOx technologies.

The heat flux sensors consist of small metal cylinders welded to the fire side surface of a boiler tube. The
shape, size, and weld specifications of the cylinder are carefully controlled to assure exact dimensions in
order to provide a specific heat path from the furnace/tube interface into the boiler tube.  Two type K
thermocouples are embedded in each cylinder at prescribed depths. The temperature gradient (typically
0-70 °C) detected by the thermocouples  is proportional to the heat flux at the point of measurement.

Flue Gas Ch Instrumentation.  A flue gas oxygen analyzing system is installed in the boiler exhaust ducts
at the economizer and air heater outlets.  This system provides an accurate reading of C>2 in the flue gas
and allows for detection of air leakage at the air heater seals. The measurement system uses in-situ
zirconium oxide measuring cells located in the flue gas paths. This method eliminates many of the
repetitive maintenance problems found in extractive systems.  Zirconium oxide probes are commonly
used in power plant applications and provide an accuracy of ± 0.25 percent. This system  undergoes an
automatic calibration at frequent intervals.

Hammond Unit 4 has two probes located in each of the two economizer outlet ducts and  two probes  in
each of the two air heater outlet ducts. These probes are approximately positioned in each duct to obtain
a representative flow weighted average.  Outputs from the probes are continuously transmitted to the
DAS.
                                              3-8

-------
BASELINE NOx CHARACTERIZATION

During the Baseline series of test, the unit operated, for the most part, as a base loaded unit which only
reduced load at night to shed slag or accommodate system load requirements.  As a consequence, most of
the short-term and long-term testing was performed at loads in the range of 400 to 480 MWe. A total of
62 short-term tests were completed in the baseline phase and continuous long-term data was gathered at
five minute averages from December 1989 through early April 1990.

Short-Term NQx Characteristics. The objective of the short-term testing was to establish the NOx trends
for the major parameters that influence emissions on this unit, i.e., excess oxygen, mill pattern and load.
The major premise behind the short-term data collection effort was that due to the potentially high
variability of the data, relatively representative trends could be established during short-term testing;
however, an accurate estimate of the absolute NOx level could be best determined through use of long-
term data. Testing was performed in such a manner as to eliminate some of the variability by establishing
trends at one boiler configuration (same test day). The following NOx characterizations reflect this test
philosophy.

At the high load condition of 480 MWe, characterization of the NOx over the excess oxygen range was
complicated by design constraints which limited the range to ± 0.75 percent about the nominal 2.7
percent O2 operating point. Figure 2 illustrates the trends over this excess oxygen range (solid lines
represent data collected on the same day under the same configuration). The data show significant
variability in not only the NOx levels at a given O2 set point but also with respect to the slope of NOx
versus O2- In general the slope of NOx versus O2 averaged approximately 110 ppm/%O2 with a
variability in measured NOx of ± 6 percent ( - 120 ppm band).

Figure 3 illustrates the NOx characteristics at a slightly reduced load of 400 MWe where mills can  be
taken out of service.  The nominal operating excess oxygen level at this load is 3.2 percent.  These  data
show the same general trends as those for the 480 MWe condition for the two mill patterns tested.  The
average NOx versus O2 slope of 60 ppm/%O2 was considerably lower than for the 480 MWe condition.
It is evident from Figure 3 that the "B" mill out of service pattern yields higher absolute NOx emission
levels. For these mill patterns the variability in absolute NOx was in the order of ± 9 percent (160  ppm
band). This influence of mill pattern was evident for other patterns at this load and complicates the
ability to ascertain the representative NOx emissions at loads where various mill patterns are possible.

The load range NOx characteristics in the Baseline configuration are shown in Figure 4 for the tests
performed over the excess oxygen excursions.  Based upon  the average conditions tested, the slope of
                                              3-9

-------
NOx versus load was determined to be approximately 1.2 ppm/MW.  As can be seen, however, the NOx
can vary by as much as ± 25 percent about the mean for the normal operating excess oxygen level.

The type of data collected is adequate for characterizing trends in NOx, however, as can be seen by the
relatively large variability, is would be difficult to establish precise characteristics (absolute NOx  levels)
without significantly more short-term data.  The long-term data collection portion of this demonstration
project had as its goal the establishment of the mean NOx characteristics with estimates of the
uncertainties. These characterizations can be  used to statistically establish the achievable NOx emission
level according to EPA criteria.

Long-Term NOx Characteristics. The long-term data collected in the Baseline testing allowed
determination of the statistical characteristics  of the data such as the mean emission level, the 95 percent
confidence interval and the autocorrelation coefficient (Ref. 1).  These statistical characteristics are
necessary for establishing the achievable emission level as well as the true dynamic load-following NOx
characteristics. Figure 5 illustrates the differences between the short- and long-term data results at 480
MWe. The long-term data demonstrates a mean NOx level of 870 ppm at the nominal 2.7 percent excess
oxygen operating condition while the short-term test results show a mean level of 970 ppm (12 percent
difference).  The short-term data generally fits within the 95 percent confidence band, however, all of the
data is above the mean level that would normally be experienced during uncontrolled operation. The
explanation for this disparity most likely is a result of such variables as coal variability, minor unit
operating changes (air register settings, etc.) and possibly weather conditions affecting the coal grinding
(wet coal) as  well as the fact that long-term data includes transients in operating ©2 level which may  be
greater than the steady load excursions.  The important point is that these  normal excursions can influence
the short-term data taken at one  point in time  but are essentially averaged out during normal long-term
operation.

Figure 6 shows a comparison of the short-term to long-term data which illustrates that, in this case, that
the long-term mean is less  than the short-term mean NOx level.  The trend for the short- and long-term
test results was consistent over the normal operating load range. The short-term mean NOx level
consistendy falls within  the 95 percent confidence interval for the long-term data. This indicates that the
short-term data is a subset of the long-term data and is therefore representative of the operating
characteristics.

Using the long-term data, statistical procedures were employed to estimate the NOx emission  levels that
could be achieved based upon EPA criteria for 30-day rolling average compliance.  The achievable
emission level is dependent upon the degree of autocorrelation, the mean emission level and the relative
                                             3-10

-------
standard deviation or variability of the daily average data.  Based upon 52 days of long-term data that
satisfies an underlying data collection criteria, for the load  scenario experienced during the period
between December 1989 and April 1990, the achievable emission level was determined to be 1.24
Ib/MMBtu (Mean = 1.16, p = 0.54, a = 0.11).

ADVANCED OVERFIRE AIR NOx CHARACTERIZATION

Short-Term NOx Characteristics.  Early in the commissioning of the AGFA system, the can-in-can OFA
port dampers overheated in the closed position requiring an outage to permanently fix them in the full
open position. This prevented obtaining any information on the influence of the retrofit on the Baseline
NOx emissions at the 0 percent OFA port opening due to the requirement for leakage cooling air for the
ports. The minimum OFA port setting was approximately  5 percent open.  Control of the air flow to the
AOFA ports was accomplished using the flow control dampers.

An acceptable setting for the windbox/AOFA proportioning dampers was first established based upon the
minimum design windbox to furnace pressure differential.  Subsequent to this, exploratory tests were
performed to assess the effectiveness of the AOFA ports with respect to the relative opening position of
the flow control damper.  Figure 7 provides  a summary of a number of high load tests performed  prior to
establishing a final operating configuration for the AOFA system. These data illustrate the same level of
variability that was experienced during the Baseline test series which necessitated establishing the trends
on the same day, rather than at random, to eliminate the influence of unidentified biasing parameters.  The
data at 480 MWe indicated that the NOx reduction at full load averaged 160 ppm or 20 to 25 percent
overall reduction between the 5 and 50 percent AOFA port open  position. Further, indications were that
the reduction diminished significantly above the 50 percent open position.  For this and other operational
considerations, the 50 percent AOFA flow control damper open position was chosen as the nominal
setting. Testing at reduced load proved this to be satisfactory down to a load of 300 MWe. At this setting,
the measured AOFA flow at full-load was determined to be approximately 20 percent of the total
combustion air flow. Below this load the AOFA flow control damper was closed to the indicated 25
percent open position.

A series of tests was performed at high load  with the AOFA ports set at the 50 percent open setting to
determine the NOx characteristics with respect to excess oxygen  excursions. Figures 8 and 9 show the
results of these characterizations at the 480 and 400 MWe load points. These data show somewhat less
variability than that for the Baseline test series. A potential reason for this is the fact that a long standing
practice of secondary air register modulation for steam temperature and flame appearance control was
eliminated after Baseline testing when it was discovered that this contributed to the NOx variability.  At
                                            3-11

-------
the full load condition the NOx was 67 ppm/%O2 compared to 110 pm/%C>2 in the baseline
configuration, or considerably less sensitive. The mean short-term NOx level at 2.7 percent C>2 was
approximately 550 ppm compared to 970 ppm in the baseline configuration (indicated 40+ % reduction).
As will be explained below, the operating excess oxygen level had to be raised above the Baseline levels
due to relatively high CO emissions on one side of the boiler.

The short-term NOx characteristics over the load range are shown in Figure 10 for all of the excess
oxygen levels tested.  The short-term data indicated a minimum NOx at 400 MWe due possibly to the
higher operating excess oxygen levels used at the 300 MWe load point (nominally 4.5 percent).

Long-Term NQx Characteristics.  Long-term testing is to be completed in early March 1991.  As a result,
at the date of publishing of this paper only a limited amount of long-term data was available for analysis.
These limited data are used only to illustrate the differences between short- and long-term data results.

At the end of the short-term testing it was determined that significant CO emissions were emanating from
one side of the boiler during AGFA operation.  As a result of this undesirable operating condition, the
recommended operating excess oxygen levels were increased at loads between 300 and 480 MWe as
shown in Figure  11. This was not an unexpected finding based upon the characteristics of previous OFA
retrofits.

Based upon the one week of long-term data available, a comparison is made between the short- and long-
term data for the AOFA  retrofit. Figure 12 shows that the short-term mean NOx emissions  are at least 100
ppm below the long-term mean.  With only one week of long-term data it  is difficult to establish if this is
the long term trend for all of the long-term data (in excess of 9 weeks). One observation is that the NOx
versus load trends for both data sets are consistent.
                                             3-12

-------
COMPARISON OF SHORT-TERM AND LONG-TERM NOx CHARACTERISTICS

If the long-term trend illustrated in Figure 12 holds for the entire 9 weeks of long-term data, the
characteristics shown in Figure 13 will establish the true characteristics of the AGFA retrofit
effectiveness. As can be seen from this figure, the short-term data demonstrates a 40 percent reduction at
full load while the long-term data indicates a reduction of only 20 percent. However, increases in
combustibles loss-on-ignition (LOI) values were experienced. As shown in Figure 14, post AGFA
retrofit LOI values are  approximately twice the pre-retrofit levels. The near-isokinetic CEGRTT samples
are obtained continuously from two, single point, sampling probes located in the gas path immediately
following the economizer. The mass train samples are obtained isokinetically at a grid located at the
electrostatic  precipitator inlet. The balance of the AGFA results will be reported in a Phase II Interim
Test Report that will be issued late in 1991.

FUTURE PROJECT  ACTIVITIES

Retrofit of the Foster Wheeler CF/SF Low NOx Burners will be completed in early May 1991.
Subsequent to shakedown tests, a series of tests will be performed to establish the effectiveness of the
burners with the AGFA ports closed. These tests are expected to be complete in early October 1991.
These tests will be followed by testing of the boiler with the AGFA ports open to the nominal position.
This testing is scheduled to be completed in late March 1992. Interim test reports for each of these phases
will be issued shortly after completion of the analysis of the data.

ACKNOWLEDGEMENTS

The authors wish to gratefully acknowledge the support and dedication of the following personnel for
their work at the wall-fired site: Mr Ernie Padgett, Georgia Power Company and Mike Nelson, Southern
Company Services, for their coordination of the design and retrofit efforts and Mr. Jose Perez, full-time
Instrumentation Specialist from Spectrum Systems, Inc. We also thank Messrs Jim Witt and Jimmy
Horton of Southern Company Services for their work coordinating the procurement and installation of the
instrumentation. We would like to recognize the following companies for their outstanding testing and
data analysis efforts at  Plant Hammond: Flame Refractories, Inc., Southern Research Institute, W. S. Pitts
Consulting and Radian  Corporation. Finally, the support from Mr. Art Baldwin, DOE ICCT Project
Manager and Mr David Eskinazi, EPR1 Project Manager, is greatly appreciated.
                                            3-13

-------
REFERENCES

1.    R. E. Rush and L. L. Smith,  "Long-Term Versus Short-Term Data Analysis Methodologies -
     Impact on the Prediction of NOx Emission Compliance'1  EPA/EPRI  Joint Symposium on
     Stationary Combustion NOx Control, New Orleans, Louisiana, March 1987.

2.    W. S. Pitts and L. L. Smith,  "Analysis of NOx Emissions Data for Prediction of Compliance with
     NOx Emissions Standards".  AWMA Combustion in the Environment Conference, Seattle,
     Washington, March 1989.

3.    Advanced-Wall-Fired Low NOx Combustion Demonstration - Phase 1 Baseline Tests.  U.S. DOE
     ICCT n Demonstration Project, Interim Report (Draft Report), Southern Company Services,
     November 1990.
                                          3-14

-------
      CombuMkxi Mr
                                                                           Flu* Gai to
                                                                           Air PrBheater
                                            Boundary Air Ports
Automated Data Collection Systerr
o Continuous Emission Monitor
o Acoustic PyromatBf
o Haat Flux TransOuccm
o Control Room Data
          FIGURE 1  Modifications of the Plant Hammond Unit 4 Boiler
   1200
   1100
52.
E
   1000
O
o5  900
en
LU
X
§  800
    700
              480 MWe NOMINAL LOAD
              ALL. BURNERS-IN-SERVICE:
                        2.5                3                3.5
                                  EXCESS OXYGEN , %
                 4.0
        FIGURE 2 Baseline Short-Term NOx Characterization at 480 MWe
                                      3-15

-------
   1100
 CM
 O
   1000
 g-  900
 co

 LLJ
 x
 O  700
    600
           400 MWe NOMINAL LOAD

          •— E-MILLOUT OF SERVICE

          fr-- B-MILL OUT OF SERVICE
                              3           4

                           EXCESS OXYGEN , %
    FIGURE 3  Baseline Short-Term NOx Characterization at 400 MWe
   1100
c7 100°
O
Q.
Q.

CO

O
CO
CO
5
LU
x
O
   900
   800
   700
   600
                  Average
                                                      I
      150     200     250     300     350     400     450     500     550

                              LOAD , MWe

     FIGURE 4  Baseline Short-Term Load Range NOx Characteristics
                               3-16

-------
        1100
      Q.
      a  900
      CO

      o
      CO
      co

      Lu  800
      x
      O
         700
         650
                4SO MWe NOMINAL LOAD

                ALL MILLS IN SERVICE
                                    SHORT-TERM DATA
            UPPER 98% a
               MEAN
              LOWER 5% Cl
                                                  LONG-TERM DATA
                     1.5        2         2.5         3

                            EXCESS OXYGEN , Percent
                                                          3.5
       FIGURE 5  Comparison of Baseline Short- and Long-Term O2 Trends
    1200
    1100
 O 1000
 S?
 n
     900
  Q.
  Q.
g  800
O
co
CO  700
 LU
     600
     500
     400
                      SHORT-TERM DATA
          UPPER 95% Cl   m
             MEAN "'
         . LOWER 5% Cl
                                                  LONG-TERM DATA
       250         300        350        400        450

                              GROSS LOAD, MWe
                                                            500
                                                                       550
FIGURE 6  Comparison of Baseline Short- and Long-Term Load Range Trends
                                    3-17

-------
   1100
   1000
eg
O
5?
O.

E
Q.
Q.

CO

g
co
co

HI
x
O
    900
    800
y   700
    600
    500
    400
                                             Each Symbol Type
                                             Represents Different
                                             Test Day	
              10      20      30     40      50

                         NOMINAL OFA POSITION
                                                  60
                                                         70
                                                                 80
        FIGURE 7  AOFA Port Opening Characterization at 480 MWe
OIAJ

c\T
O 700
CO
a
a.
CO 600
O
CO
CO
LU
O 500
Z
400




480 MWe NOMINAL LOAD
ALL MILLS IN SERVICE
50 % AOFA PORT OPENING

" 28
DAY 28




43
2 ^f

38
s^**





43

^













                     3              4

                       EXCESS OXYGEN , Percent
      FIGURE 8  AOFA Short-Term NOx Characterization at 480 MWe
                               3-18

-------
NOx EMISSIONS . ppm (3% O2)

400 MWe NOMINAL LOAD
50 % AOFA PORT OPENING




^J
>-^a
DAY 47 » _j
"-^""If
M *° 39


/ —
s
^^
V




-(All Mills in Service j


"E" Mill Out of Service!

 23456
                EXCESS OXYGEN , Percent

  FIGURE 9 AOFA Short-Term NOx Characterization at 400 MWe
ouu
OTnn
/ uu
Q.
E
a
a.
W ft/in
DUU
o
co
co
55

m
X enn
Q OUU
z
^.nn


1
o
a
c
A





O
o




0


*•»,
O *"~ -











(
c
(
c
t
— — . _ —

o

c








o
k ^ . -*"
--J
^tl






o

8
0 °
o
A
-^ &
o
0
9
0


0

















250 300 350 400 450 500 55
                  GROSS LOAD , MWe

FIGURE 10  AOFA Short-Term Load Range NOx Characteristics
                         3-19

-------
5
Percent
.&.
z
LU
O
o
EXCESS
CO
2
2









"• x





RECOI
RECOMMEh
\
-. ^^^^
-. ^x



MENOEO MINIMUM C
DED MINIMUM O2 U
/
7
— L_
•«,s
4 ""-,
/
ULEVELFORSASEl
VELFOHAOFAOPEF


" 	 	 	 .


""--._
NE OPERATION
AT1ON







50 300 350 400 450 500 55
                          GROSS LOAD , MWe

     FIGURE 11 Baseline and AOFA Operating Excess Oxygen Curves
OJ
O

3?
CO
   Q.
   Q.
     1100
     1000
      900
      BOO
  g
  CO
  52   700
  O   600
      500
      400
          LONG-TERM DATA FOR WEEK 2/10 TO 2/16/91
                                         LONG-TERM DATA
          LOWER 5% Cl oo
                                         SHORT-TERM DATA
     250        300       350       400       450

                          GROSS LOAD, MWe
                                                     500
                                                               550
FIGURE 12  Comparison of AOFA Short- and Long-Term Load Range Trends
                                  3-20

-------
   1000
    900
CO
O

5?
o.

E   eoo
Q.
Q.

C/3

O
W   700
c/2

LU
X

Z   600
    500
                     LONG-TERM AGFA
                        __ SHORT-TERM AGFA  __
      250
                300
                                                   20 %    I

                                                         40%
                                                        500
                          350        400        450

                           GROSS LOAD, MWe


          FIGURE 13  Comparison of Baseline and AGFA Operation
                                                                  550
Oio

vO
^9

O
     c
     o
    n

    1*

    "E 3
     o
    O
       2
                Mass Train Samples
              CEGRIT Samples
                                                         Baseline
        250
                 300
                           350
                                     400
                                              450
                                                        500
                                 Unit Load (MWe)

        FIGURE 14  Comparison of Baseline and AGFA LOI Values
                                3-21

-------
     REBURN TECHNOLOGY FOR NO
 CONTROL ON A CYCLONE-FIRED BOILER
R.W.  Borio, R.D. Lewis, M.B. Keough
    ABB Combustion Engineering
       R. C. Booth

       R. E. Hall

       R. A. Lott

       A. Kokkinos

       D. F. Gyorke

       S. Durrani
Energy System Associates

U.S. EPA

GRI

EPRI

DOE-PETC

Ohio Edison
       H. J. Johnson   OCDO

       J. J. Kienle    East Ohio Gas

-------
ABSTRACT

Cyclone-fired boilers have typically produced higher NO  than other
types of coal-fired utility boilers.  Cyclone-fired bolters are
generally not amenable to in-furnace NO  reduction technologies;
reburning represents an in-furnace NO  reduction technology that is
well suited to cyclone boilers.  The Environmental Protection
Agency, Gas Research Institute, Electric Power Research Institute,
Department of Energy, and Ohio Coal Development Office have
cosponsored a program conducted by ABB Combustion Engineering to
demonstrate natural gas reburning on a cyclone-fired boiler at Ohio
Edison's Niles Station.  Ohio Edison and East Ohio Gas have both
provided in-kind financial contributions to the program.

The paper provides a preliminary summary of results from recent
parametric testing of the reburn system which was installed and
commissioned during the third quarter of 1990.  Key variables
evaluated during reburn testing included excess air, natural gas
flow rates, recirculated flue gas flow rates, and additional air
flow rates.  Nitrogen oxide reductions were shown to be strongly
influenced by reburn zone stoichiometry.  The effect of reburning on
boiler thermal performance was evaluated; changes in waterwall heat
absorption and convective pass heat absorption are presented along
with changes in boiler efficiency.  Electrostatic precipitator
performance is compared for base case coal firing and the reburning.
Finally, mention is made of thicker ash deposits on the back wall of
the secondary furnace since installation of the reburn system.
INTRODUCTION

Recent passage of the 1990 Clean Air Act Amendments has underscored
the need for establishing commercially acceptable technologies for
reducing power plant emissions, especially sulfur dioxide (SO ) and
nitrogen oxides (NO ).  NO  and sulfur oxides (SO ) lead to formation
of acid rain by comoining with moisture in the atmosphere to produce
nitric and sulfuric acids (1,2,3).  NO  also contributes to the
formation of "ground level" ozone.  Ozone is a factor in the creation
of smog, leads to forest damage, and contributes to poor visibility.
                                                                   x
Electric utility power plants account for about one-third of the NO
and two-thirds of the SO  emissions in the U.S.  Cyclone-fired
boilers, while representing about 9% of the U.S. coal-fired generating
capacity, emit about 14% of the NO  that utility boilers produce.

Given this backgroud, the Environmental Protection Agency (EPA), the
Gas Research Institute (GRI), the Electric Power Research Institute
(EPRI), the Department of Energy   Pittsburgh Energy Technology Center
(DOE-PETC), and the Ohio Coal Development Office (OCDO) have sponsored
a program led by ABB Combustion Engineering (ABB-CE), to demonstrate
reburning on a cyclone-fired boiler.  Ohio Edison is providing Unit
No. 1 at their Niles Station for the reburn demonstration along with
                                 3-25

-------
financial  assistance.  The Consolidated Natural Gas Company (CNG),
specifically East Ohio Gas, has financially shared in the program.
Ohio Edison and East Ohio Gas are both sharing a portion of the
differential cost of natural gas.

Unit No. 1 went into commercial operation in 1955 and is a 108 MWe
(net) natural circulation reheat boiler operating with a pressurized
furnace.  Steam conditions are 1485/374 psig* and 1000/1000°F.
Working with ABB-CE are Energy Systems Associates (ESA), Physical
Sciences Inc. Technology (PSIT), and Mitsubishi Heavy Industries
(MHI).

Reburn technology involves creating a second combustion or "reburn"
zone downstream from the main burners in a boiler.  Combustion gases
that result from burning a fossil fuel in the main combustion zone,
move to the "reburn" zone where additional fuel, in this case natural
gas, is injected.  The injection of additional fuel creates a
fuel-rich zone in which the NO  formed in the main combustion zone are
converted to molecular nitrogen and water vapor which occur naturally
in the atmosphere.  Any unburned fuel leaving the reburn zone is
subsequently burned to completion in a downstream burnout zone where
additional  air is injected.   Further details of the reburning process
can  be found in the literature (4,5,6,7).  Reburning is especially
attractive  for cyclone-fired boilers and other wet-bottom boilers
since low NO  burners and most other low NO  combustion technologies
used on conventional boilers are not applicable to cyclone-fired and
wet-bottom  boilers.  The overall goal of the program is to
successfully demonstrate a 50% reduction in NO  emissions from a
cyclone-fired boiler employing reburning technology.  Figure 1 shows
the  overall project scope and schedule.

The  engineering design of the reburn system has been completed and
reported previously  (8).  This paper presents results of the
parametric  testing of the reburn system installed during the summer of
1990.
REBURN SYSTEM

DESCRIPTION

Viewed in terms of its components, the reburn system is composed of
equipment/materials which are familiar to operators of utility power
plants.  The reburn system is relatively compact, requiring a small
amount of space when compared with tail-end treatment systems; this
could be an advantage for utilities where indoor and/or outdoor space
is limited.
       Readers more familiar with metric units may use the conversion
       table at the end of this paper.
                                  3-26

-------
The reburn system can be described in terms of equipment necessary for
the creation of the reburn zone and for the burnout zone.  Key
components for the reburn zone are the flue gas recirculation (FGR)
fan, ductwork/associated control dampers, natural gas
pipeline/associated control valves, and the windboxes and nozzle
assemblies where the natural gas and flue gas are mixed shortly before
injection into the lower part of the secondary furnace of the boiler.
A mixture of flue gas and natural gas is injected through five windbox
nozzle assemblies, referred to as Upper Fuel Injectors (UFIs)
(Figure 2), along the back wall of the secondary furnace.

Key components for the burnout zone are the ductwork/associated
control dampers, and the windboxes and nozzle assemblies where
combustion air, referred to as Additional Air (AA), is injected into
the upper part of the secondary furnace.  Greater detail on the design
and operation of the UFIs and AA nozzles was provided in an earlier
paper  (8).

The reburn control system uses an Allen Bradley programmable
controller to operate the reburn system in an automatic,
load-following mode.  Natural gas flow, at a predetermined percentage
of unit heat input, and recirculated flue gas flows are based on coal
flow demand input.  The additional air flow is based on natural gas
flow with the final excess oxygen designed to be slightly lower than
the normal cyclone excess oxygen level.

The reburn system has been tied into the main boiler control system
for safety and control purposes.  The natural gas reburn fuel controls
have been set up in a last-in-service/first-out-of-service logic.  The
FGR system remains in service independent of the reburn natural gas,
except for loss of control power.  All system dampers/valves fail shut
except for the natural gas vent valves which fail open.   Flame
scanners are not used in conjunction with the UFIs since there is no
visible flame in the reburn zone.

The use of combustible gas measurement as a system safety input will
be evaluated with data that have been collected during parametric
testing.  Operation of the reburn system has not required an increase
in operating personnel, an advantage from the utility's point of view.
INSTALLATION

The reburn system was installed with minimal disruption to normal
power plant operation.  The four key phases of reburn system
installation involved: (1) procurement of material  and delivery on
site, (2) pre-outage activities, (3) outage activities, and (4) post-
outage activities.  A key consideration was the installation of all
direct boiler-related equipment/materials during the utility's normal
4-week boiler outage.
                                3-27

-------
Major items obtained during the procurement period included the UFI
and AA windboxes, their waterwall tube panel inserts, FGR fan,
recirculated flue gas and AA ductwork, and the control system.  The
FGR fan and the control system represented those items requiring the
longest lead time at about 26 weeks.

Pre-outage work included demolition of the old FGR fan and associated
ductwork along with required asbestos abatement.  The natural gas
pipeline was installed up to the point where it would connect to the
windboxes.  Structural steel was reinforced in those areas where the
recirculated flue gas ductwork would be installed and some minor
revamping of access stairs and platforms was done to accommodate
installation of the new ductwork.

At the commencement of the boiler outage on May 21,  1990, boiler
casing and refractory, at the locations for the UFI  and AA windboxes,
were  removed exposing the straight sections of waterwall tubes which
would  be cut out.  Dimensions of waterwall sections  removed to
accommodate the prefabricated UFI and AA tube panels were about 3  ft
wide  by  15 ft.  After welding in the tube panels, the windboxes were
welded to flanges provided as part of the tube panel structure, and
seal  boxes were built around each windbox and tube panel to prevent
any  furnace leakage  (this is a pressurized furnace).  Windboxes were
tied  into the  previously installed ductwork by the installation of
expansion joints which allowed for growth of the boiler versus the
stationary ductwork.  The boiler was hydrostatically tested,  followed
by the installation of refractory in the seal boxes  and seal-welding
of all outer casing.  Following  an air pressure test to locate and
seal-weld  any  remaining furnace  casing leaks, the boiler was  fired up
 (to  allow  for  chemical cleaning  and curing the refractory) and
returned to service on June 25,  1990.

CHECKOUT/START-UP

A key activity during the post-outage time frame was checkout and
start-up of the  reburn system, the objective being to verify  that  all
components worked as designed.   During the outage all mechanical and
electrical  subsystems were verified to be operational.  During system
start-up the various  subsystem interactions and sequencing were
verified.  Minor changes to the  control system programming and
adjusting  of the time delays based on actual device  responses were
also  completed. The gas reburn system was designed to operate in
either a reburn mode  (natural gas being injected) or a non-reburn  mode
 (no  natural gas  being  injected).  In the non-reburn  mode some minimal
amount of  cooling FGR or air is  needed to maintain the integrity of
the  UFI  and AA nozzles; minimal  amounts of cooling FGR or air were
determined during the post-outage time frame.

Reburn system  operation was initially simulated without the use of
natural  gas to verify operation  of the comprehensive control  system
safety related permissives.  Natural gas was injected in small
quantities for the first time on August 29, 1990.  Full-load  automatic
operation with 19% natural gas was achieved on September 21,  1990.
                                 3-28

-------
PARAMETRIC TEST PROGRAM

OBJECTIVES AND SCOPE

The reburn system, as installed at Ohio Edison Niles Unit No. 1, by
design, incorporated a large amount of operational flexibility to be
able to examine and optimize reburning and boiler operations.  The
primary objective of the parametric testing program was to determine
an operational mode which would result in low NO  (not necessarily
lowest NO ) while minimizing other potentially detrimental effects on
boiler performance.  These other effects included:

1. Minimizing other gaseous combustible and particulate emissions;
2. Minimizing fuel and auxiliary power costs;
3. Minimizing degradations in boiler performance (e.g., decreases in
   boiler efficiency, use of reheat attemperator spray, or
   excessive superheater or reheat steam or tube metal
   temperatures).

A secondary objective of the parametric testing was to establish a
reburning data base which could be used to evaluate reburning for
other boilers.

During the parametric testing, approximately 150 test points were
completed to examine 13 existing boiler and reburn system operational
variables.  The operational variables examined included:

        Baseline Test Variables

        t  Cyclone Excess Air
        •  Cyclones in Service
        t  Boiler Load

        Reburn Test Variables

        •  Reburn Zone
             Natural Gas Flow
             Flue Gas Recirculation Flow/Compartment Bias
             UFI Tilt/Yaw
             UFI Horizontal Bias
        •  Burnout Zone
             Air Flow
             AA Tilt/Yaw
             AA and UFI Tilt Combination

Because of the large number of independent test variables, it was not
possible to examine every permutation and combination.  The parametric
testing was set up and conducted to "step-through" the variables in a
decreasing priority sequence for each of the three key boiler "zones"
                                 3-29

-------
(cyclones,  reburn zone, and burnout zone).  Initially, nominal
operating conditions were selected for each variable, then, once a
variable had been examined, it was reset to a "near optimum" condition
for subsequent tests.  "Near optimum" conditions were selected based
on the test strategy previously described.

In addition, to ensure comparability of the results, many tests were
repeated as well as conducting tests examining a single variable on
the same day.  This was necessary because, even though significant
effort was expended, it was difficult to replicate cyclone operating
conditions on a day-to-day basis.  This difficulty and its
implications are further discussed in "Test Results."

During the parametric testing, a limited number of more comprehensive
tests were completed and were referred to as "maxi" tests.  Maxi tests
were run at generator loads of 108 and 86 MWe (net) at baseline (100%
coal firing) and 18% natural gas reburn conditions utilizing the
reburn configuration found to represent an optimum during the
parametric investigations.  Purposes of the "maxi" tests were to:

•  Determine the effect of reburn system operation on the furnace
   temperature entering the reburn zone and the convective pass.

•  Assess the effect of reburning on the flue gas conditions entering
   the electrostatic precipitator (ESP).

•  Measure the size distribution and mass loading of the
   particulates entering the ESP.

t  Evaluate the effect of reburn on the collection efficiency of
   the ESP.
BOILER PERFORMANCE AND EMISSIONS MONITORING SYSTEMS

During the parametric testing, in addition to normal control room
board data, most important boiler performance operational variables
were electronically recorded in a personal computer.  These data
included flows, temperatures, and pressures for boiler water, steam,
air, and fuel .

Oxygen (02) concentrations of the flue gas leaving the four  cyclones
were measured using the four existing water-cooled probes and
instrumentation.  These probes are located on the rear wall near the
bottom of the secondary furnace with each of the   probes
approximately lined up with one of the cyclone exhaust streams.

Gaseous emissions of NO , (L, CO, C0?, S0?, and THC (total
hydrocarbon) were measured at the bofler 5xit/air heater  inlet via  10
sampling probes spaced across the boiler outlet duct.  (See Figure 3.)
During testing samples were sequentially drawn from each  of the 10
probes to be able to assess gaseous emissions profiles.
                                 3-30

-------
ESP performance was assessed by particulate loading measurements
before and after; and particle size distribution and moisture content
were determined using EPA Method 5 isokinetic traverses of the ducting
using a cascade impactor.  Flyash resistivity and dewpoint
measurements were made at the ESP inlet using a Wahlco resistivity
probe and a Land dewpoint probe, respectively.  Flyash samples were
also taken at the ESP inlet duct test point using high volume sampling
techniques to ascertain carbon content.  Boiler bottom ash was sampled
from the slag tanks below the wet-bottom slag taps.

Temperatures and velocities in the boiler were sampled with in-furnace
traverses using a water-cooled probe.  Temperatures were measured
using a shielded high velocity suction pyrometer and velocities were
measured using a five-hole pitot tube.  (Furnace velocity data have
not currently been analyzed and will be reported later.)  The
in-furnace sampling on this pressurized furnace was limited to two
planes in the secondary furnace. Traverses of the first plane,
representing the inlet to the secondary furnace, were made using three
of the four ports that the plant has for measuring cyclone oxygen
levels. (The fourth port was not accessible.)  Measurements made in
the second plane, at the secondary furnace outlet immediately below
the superheater surface, were carried out using an aspirated test port
located in one of the furnace sidewalls.
TEST RESULTS

General

Previous testing by others has shown that one of the key variables
affecting NO  emissions was reburn zone stoichiometry (5,6,7,8,9,10).
Stoichiometry is defined as the ratio of actual air supplied compared
to the theoretical amount of air required to completely combust the
available fuel.  It is important to understand the methodology used in
establishing this value.  First, the oxygen content of the flue gas
effluent from each of the cyclones was measured to ascertain cyclone
stoichiometry.  Second, the accurately measured reburn natural gas
flow rate was compared to a corrected boiler coal flow (corrected coal
flow was based on indicated coal flow, boiler efficiency, and plant
heat rate) to determine natural gas and coal fuel fractions on a heat
input basis.  Third, the reburn zone stoichiometry was computed by
summing the mathematical products of the stoichiometry and fuel (heat
input) fractions for the cyclones and the reburn fuel flows.

While variations to cyclone excess oxygen level were evaluated and
documented, relative to its impact on NO , it is not a variable which
can be used to optimize NO  emissions at this unit.  Altering the
cyclone excess oxygen from the normal 2.0-2.5% 02 (10-13% excess air)
level for an extended period may have detrimental effects on cyclone
tube life or slag tapping if the oxygen level is lowered or raised,
respectively, beyond the normal range.
                                3-31

-------
The Ohio Edison Miles No. 1 unit was originally designed as a 125 MWe
(net) unit; however, except for short periods, it has operated at 108
MWe (net) for the past 30 years.  There are several causes for this
derating including fan limitations early in the boiler history and
other more recent operational  problems (principally, primary furnace
and cyclone tube wastage).  Since the employment of reburning
essentially decreases the firing rate in the cyclones, and because
slag tapping is eventually affected as cyclone loading decreases, it
is important that the history of a unit be known if an application of
reburning is contemplated.  Boiler derating, as is the case on the
subject boiler, and cyclone-firing configuration are two factors that
can affect the application of reburning technology.

NO  emissions for the subject cyclone-fired boiler at 108 MWe (net)
averaged approximately 705 ppm (all NO  emissions reported have been
corrected to a 3% excess 02 basis).  Tnis emissions level was
representative of normal operation with a mean cyclone excess oxygen
level of 2.0-2.5% 02 (10.6-13.6% excess air).  Slight variations in
individual cyclone operation resulted in day-to-day data scatter of
approximately +25 ppm.

Changing the cyclone excess oxygen level changed the NO  emissions
slightly.  For example, a 1% decrease in cyclone excess oxygen, from 3
to 2% 0-, decreased NO  emissions by approximately 15 ppm.

Reducing the cyclone-firing rate also reduced NO  emissions.  At
86 MWe  (net), a 20% decrease in boiler load, NO  emissions under
normal  operating conditions were approximately 630 ppm, a 75 ppm or
10% decrease in emissions from normal full load operation.  At reduced
boiler  load a similar trend of decreasing NO  for decreasing cyclone
excess  oxygen was seen.  Baseline NO  emissions results showing the
effects  of boiler load and 0? are shown in Figure 4.

Carbon  monoxide (CO) emissions in the baseline mode of operation were
typically very low, under 30 ppm.  Baseline SO  emissions varied
between  2400 and 2700 ppm due to slight variations in coal sulfur
content.  Negligible THC gaseous emissions were observed during
baseline and reburn testing.

Coal Variability

The  Eastern bituminous coal fired at the Niles plant arrives by truck
from approximately 15 supply mines located in the Ohio, Pennsylvania,
and West Virginia area.  No one mine supplies more than about 10% of
the total coal supply used.  Initially there was some concern that
coal variability at the Niles plant could add uncertainty to the
results  and conclusions drawn from those results.

However, frequent samples and subsequent analysis of the coal have
shown the fuel composition to be very consistent.  Table 1 presents a
composite coal analysis based on analysis of 21 coal samples.  Some
statistical data showing the good consistency of the analyses are also
shown.
                                  3-32

-------
                                TABLE 1
                    OHIO EDISON MILES UNIT NO. 1
             Composite Coal Analysis   As Received Basis
Proximate Analysis
Deviation
                                Average
Maximum
 Value
Minimum
 Value
                                                              Standard
                                                      31.
                                                      45.
                                                      11.
% Moisture (Total)                 7.8       9.3       6.6
% Volatile Matter                 32.2      33.7
% Fixed Carbon (By Difference)    47.3      49.3
% Ash                             12.6      13.6

  HHV Btu/lb                     11576     11870     11277
  Ib Ash/10  Btu                  10.9      12.0       9.6

Ultimate Analysis
% Moisture                         7.8       9.3       6.6
% Hydrogen                         4.4       4.5       4.3
% Carbon                          63.4      65.3      61.8
% Sulfur                           3.3       4.1       3.0
% Nitrogen                         1.4       1.5       1.3
% Oxygen (By Difference)           7.1       8.4       5.5
% Ash                             12.6      13.6      11.4
  Total                          100.0
                      0.76
                      0.62
                      0.94
                      0.62

                       176
                      0.63
                                                                0.76
                                                                0.06
                                                                1.11
                                                                0.31
                                                                0.05
                                                                0.73
                                                                0.62
Based upon the consistency of the coal composition, coal variability
should have negligible effect on the test program results.

NO  Emissions as Function of Key Variables

Reburn Zone Stoichiometry

As expected, it was found that some of the test variables had a
pronounced effect on NO  emissions and other variables had little or
no effect on NO .  Reburn zone stoichiometry was found to be the key
parameter affecting NO  emissions.  Figure 5 shows the relationship of
reburn zone stoichiometry with NO  emissions.  The reburn zone
stoichiometry was varied by adjusting either the reburn natural gas
flow rate or the cyclone excess air level.  For the full-load tests
the reburn zone stoichiometry was varied from 0.88 to 1.06.

NO  emissions were shown to be linearly related to reburn zone
stoichiometry (for the test range) and decreased by approximately 180
ppm per 0.10 (or 10%) decrease in reburn zone stoichiometry.  For a
constant cyclone excess oxygen level an approximate 10% decrease in
reburn zone stoichiometry would result from a 9% increase in reburn
natural  gas fuel fraction.  For example, with the normal cyclone
excess oxygen level  of 2.5% 0, (13.6% excess air), increasing the
reburn natural  gas fuel fraction from 9 to 18% would result in a
decrease to the reburn zone stoichiometry from approximately 1.03 to
0.93 and decrease the NO  emissions from approximately 480 to 300 ppm
(+25 ppm).              X
                                  3-33

-------
Reburn natural  gas flow (Figure 6) presents the NO  emissions data
versus the amount of reburn natural gas fired.  Two interesting
results shown include: (1) the linearity of the NO  reduction with
increasing natural gas flow for a given cyclone excess oxygen level;
and (2) for a given reburn zone stoichiometry, the NO  emissions
results were similar regardless of whether the stoichTometry was
achieved by changing the reburn natural gas flow rate or by changing
the cyclone excess oxygen level.
Recirculated Flue Gas Flow

The purpose of flue gas recirculation (FGR) in the reburn system is to
assist in the penetration of the reburn fuel and promote mixing of the
reburn fuel with the bulk furnace gases without significantly
increasing the oxygen content or stoichiometry in the reburn zone as
would happen if air were used instead of FGR.  Pilot scale research
(10) has also shown a small incremental NO  reduction with increasing
levels of FGR.  Figure 7 presents the results of tests where the FGR
flow rate was varied from approximately 3 to 11% of the total flue gas
flow.  Both baseline (no natural gas) and 18% natural gas reburn test
series are shown.  FGR had no effect on NO  emissions with or without
reburning.

The lack of any effect of FGR on NO  during the baseline tests was
likely due to:  (1) coal combustion s being essentially completed (no
further nitrogen release); and  (2) changes in thermal NO 's being not
measurably affected because of the relatively low thermaT dilution
created by introducing FGR (previously measured temperatures showed
approximately 2300-2400°F for the reburn zone inlet).  For the reburn
tests, varying FGR had no effect on NO ; this was likely due to the
good mixing that occurred regardless of the FGR flow rate.  Earlier
flow modelling (8) had shown that cyclone effluent gases tend to hug
the rear wall where the reburn jets were placed.  The importance of
FGR flow is likely to be very unit specific; e.g., a large open
furnace where reburn fuel penetration is required.

After determining the sensitivity of NO  reduction to FGR flow rate,
it was decided to operate at a reduced level (about 5%) with the FGR
fan inlet dampers nearly closed.  This was advantageous since lower
levels of FGR minimized changes in boiler steam-side performance
(discussed later) and decreased auxiliary power usage.  At the Niles
unit, safety requires the use of FGR.


Other Reburn System Variables

Somewhat surprisingly, NO  emissions were essentially not affected by
the other reburn system operating variables including upper fuel
injector (UFI) tilt, yaw, or flow bias or additional air (AA) injector
tilt, yaw, or flow bias.  It is interesting that burnout air (AA) did
not change NO  emissions; and that NO  has not reformed in the burnout
zone.  This mSy be due to the cool furnace gas temperatures which do
not promote thermal NO  formation or due to  the lack of fuel bound
nitrogen in the reburn fuel.
                                 3-34

-------
Boiler Load

Testing to establish the effect of boiler load on NO  emissions was
completed during the parametric test.  The partial load testing was
conducted at 86 MWe (net) since it represents the approximate load at
which the fourth cyclone is removed from (or placed into) service as
boiler load is decreased (increased) to ensure adequate slag liquidity
for removal from the furnace bottom.  Also, the cyclone loading is
nearly equivalent for the 86 MWe (net) baseline non-reburn tests and
the 108 MWe (net) 18% natural gas reburn tests.  Figure 8 compares the
NO  emissions at these two loads as a function of reburn zone
stoichiometry.  At the reduced load the NO  values were lower for all
conditions.  The decrease in NO  emissions for a 10% change in reburn
stoichiometry at 86 MWe  (net) was approximately 130 ppm compared to
the 180 ppm reduction rate observed for an equivalent change at full
load. However, because of lower baseline non-reburn NO  levels the
percent NO  reduction was nearly equivalent for the two test boiler
loads.    x

Other Gaseous Emissions

NO  emissions reduction was essentially linear with increasing natural
gas flow and did not significantly change with other reburn system
variables.  The selection of an "optimum" natural gas flow  to be used
during the forthcoming long-term tests will be based more on
minimizing other gaseous pollutants and changes in boiler performance.
During the shakedown period high levels of CO emissions were observed,
especially during high reburn fuel flow.  These CO emissions were
subsequently decreased to typically below 100 ppm as the reburn system
operating variables were optimized.  It was found that the vertical
tilt position of the UFIs and AAs were principal factors when high CO
emissions were observed.  CO was minimized by downward tilting of both
the DPI and AA nozzles.  A -17° (from horizontal) was selected as best
for the UFI nozzles and  -10° for the AA nozzles (Figure 9).  In
addition to lowering the average CO emission level a more uniform CO
and Op profile was generated across the boiler exit duct.  This trend
can be seen by comparing Figures 10 and 11.

Higher CO emissions in the center of the duct were frequently observed
during the early testing and were probably due to unoptimized
additional air (AA) injector adjustments leading to insufficient
penetration and mixing.  Variability in cyclone 0^ concentrations was
also a contributing factor.  It was also observed that creation of a
tangential swirl in the furnace, by yawing the nozzles on one wall in
one direction and those on the other wall in the opposite direction,
further reduced CO emissions.  The minimum exit CO resulted with the
UFI tilts at -17°, the AA tilts at -10°, and the AA yaw set up to
impart a clockwise swirl (viewed from above).

Emission of S02 decreased with increasing natural gas flow as
expected.  On average the SO,, decrease was inversely proportional to
the reburn fuel flow; however, there was a significant amount of
scatter (+ 10%) due to coal sulfur variations.  Gaseous THC emissions
were negligible for all tests.
                                 3-35

-------
Boiler Thermal  Performance

The main impact that reburning had on boiler thermal performance was a
shift in the heat absorption from the waterwalls to the convective
pass sections.   The Miles unit does not have an economizer; therefore,
the increase in convective pass absorption was observed in the
superheater and reheater with slight increases to temperatures and
attemperator spray flows.  The changes in boiler thermal performance
were due to:

   1.   Decreased cyclone loading with natural gas reburning in the
        secondary furnace; and
   2.   FGR used with reburning to inject the natural gas.

Table 2 presents selected boiler thermal  performance data for a
baseline (coal  only) test and a nominal reburn test.  These two tests
had equivalent boiler load and excess air at 115 MWe gross (108 MWe
net) and 15%, respectively.  The reburn test had 17.2% natural gas
reburn fuel, on a heat input basis.  For the boiler superheater, it
can be seen that the attemperator spray flow increased from 1.3 to
4.5% of main steam flow due to reburning.  The primary superheater,
which is located prior to the attemperator, had a 20°F increase in
steam temperature at its outlet.  The secondary superheater inlet,
which is just after the attemperator, was 15°F lower for reburning
(showing the higher spray flow).  The final superheat steam
temperature was  slightly higher for reburning.

The reheat  steam section showed an attemperator spray flow of
approximately 3% of total reheat steam flow for reburning.  This spray
flow represents leakage by the closed control valve when the block
valves were open.  The final reheat steam temperature was also 12°F
higher with reburning which raised it to the design point of 1000°F-
The control dampers in the split boiler rear pass were set differently
for the baseline and reburn tests.  For the baseline test the
superheat dampers were closed and the reheat dampers were open to
increase the flue gas flow and hence reheat steam temperature.  For
the reburn  test, the superheat dampers were open and the reheat
dampers were closed to limit the reheat absorption and attemperator
spray requirements.  Note that the "closed" damper positions are still
approximately 20° open and still have a significant amount of flue gas
flow passing through them.

With reburning, boiler heat absorption in the waterwall decreased by
approximately 5% and convective section heat absorption increased by
approximately 5%.  The decrease in waterwall absorption is due to
decreased cyclone loading.  The increase in convective pass absorption
is due to increased gas temperatures (calculated to be 30°F at the
furnace outlet plane) and increased flue gas weight (due to FGR) with
reburning.
                                 3-36

-------
Reheater absorption increased by only 4% while superheater absorption
increased by 6% due to the adjustment of the backpass flow control
dampers.  Steam temperature profiles were also monitored during this
program.  Thermocouples were installed on approximately every fourth
tube element at the primary and secondary superheater outlet headers.
Negligible changes were observed in primary or secondary superheat
profiles between baseline and reburn tests.

The boiler efficiency with natural  gas reburning decreased by 0.7%.
Table 3 presents a breakdown of the efficiency and a comparison
between baseline and reburn.  The primary reason was a 1% higher loss
due to a higher moisture in the flue gas in the reburn test due to the
higher hydrogen content of the natural gas versus the hydrogen content
in the coal.  This loss was somewhat offset by a lower ash pit loss
and a lower carbon heat loss due to less coal being fired when
reburning.

Overall, the boiler performance did not change appreciably with
natural gas reburning.  It was observed that, because of the lower
cyclone loading with reburning, it was necessary to monitor cyclone
excess oxygen levels more closely at reduced load to maintain slag
temperatures and viscosity for effective molten slag removal.

Carbon in Ash

Carbon loss in flyash was not significantly affected by reburning.
Flyash samples were taken and analyzed for approximately two-thirds of
the tests completed.  Bottom ash samples were taken once per day.
Full-load flyash carbon levels of approximately 30-35% have typically
been shown with a range of 25-45% carbon in the flyash.  The flyash
carbon level has been compared to reburn natural gas flow and cyclone
excess air, variables which may normally have correlated to flyash
carbon level, and no relationship was found.  The bottom ash carbon
levels have typically been less than 1%.  Thus for a 12.6% ash coal
and a baseline flyash/bottom ash ratio of 30:70, the baseline carbon
heat loss was approximately 1.2 to 1.4% and for reburning, with a
reduced coal flow and hence flyash loading, the carbon heat loss was
approximately 1.0 to 1.2%.

Reasons for the high unburned carbon under baseline and reburn
conditions are unclear.  Possible causes include coal properties,  coal
particle size distribution and cyclone aerodynamics (greater expulsion
of coal fines).  During reduced load operation the average flyash
carbon content decreased to approximately 20% carbon in flyash.  This
would be expected with more residence time, decreased cyclone loading,
and decreased expulsion of particulate from the cyclones.

Furnace Gas Temperatures

Reburn Zone Inlet Gas Temperature

Figures 12 and 13 show the results of the flue gas temperature
traverses that were made at the inlet to the reburn zone.  The furnace
depth at the traverse locations was 13 ft and the maximum traverse
depth was physically limited to about 10 ft.  At 108 MWe  (net) the
                                 3-37

-------
baseline average gas temperature was 120°F higher than with 18%
reburn.   The tests at 86 MWe (net) showed a similar trend:  the
baseline gas temperature averaged approximately 100°F higher than with
reburning.   For both the baseline and reburn tests there was a 200 to
300°F decrease in flue gas temperature from the rear wall to the
division wall.  The temperature profiles for baseline and reburn at 86
MWe paralleled one another while the baseline and reburn temperatures
at 108 MWe showed that they were considerably different near the back
wall  but began to approach the same value as the probe was moved
towards its maximum insertion depth.  Comparison of the average
temperatures and profiles measured during the 108 MWe reburn test with
the 86 MWe baseline test show very similar results.  Note that with
18% reburn at 108 MWe, the coal loading to the cyclones is only
slightly higher than at 86 MWe with 100% coal.  Therefore, it would
not be unusual for the average reburn zone inlet temperatures from
these two configurations to be the same.

Furnace Outlet

Figure 14 shows the results of the temperature traverses at the
furnace outlet plane.  The traverse depth represents approximately one
third of the boiler width.  The furnace outlet temperature with reburn
averaged 130°F higher at 108 MWe than the base case; i.e., 100% coal.
At 86 MWe the average temperature with 18% reburn was about 65°F lower
than the baseline temperature.  This difference, though generally
corroborated by the boiler thermal performance evaluation, is not
fully understood and will be examined along with other data that have
not currently been analyzed.

Electrostatic Precipitator Performance

Electrostatic precipitators (ESPs) replaced mechanical collectors in
the early 1980s to improve particulate collection efficiency.  The ESP
was si^ed quite liberally with a specific collection area (SCA) of
278 ft /ACFM; it is normally operated with only three of its five
fields energized, and operated in this mode an opacity of 2.5% was
routinely achieved during parametric testing involving both baseline
and reburn testing.

ESP collection efficiency was determined by sampling at the inlet to
the ESP and in the stack using EPA Method 5.  Ammonia injection is
used to control acid smut emissions at the unit; the ammonia injection
point is upstream of the ESP inlet sampling port.  No attempt was made
to optimize the ammonia injection operation to account for changes in
flue gas composition (such as S03) during reburn system operation.

Results of particulate sampling the stack showed that particulate
loading increased with reburn compared to baseline for both full load
and partial load cases primarily as a result of the flue gas
conditioning ammonia injection.  At 108 MWe the particulate loading
was 0.032 lb/10 Btu for 100% coal firing and 0.043 lb/106Btu for the
reburn case where 18% natural gas was used, on a total heat input
basis.  The trend was the same for the partial load test (86 MWe),
where6100% coal firing gave 0.022 lb/10 Btu compared with 0.027
lb/10 Btu
                                 3-38

-------
when 18% natural gas was used in the reburn test.  Despite the small
increase in particulate loading in the reburnfitests the actual
particul'ate loadings of 0.043 and 0.027 lb/10 Btu for the full load
and partial load reburn casesg respectively, are well below the
regulating limit of 0.1 lb/10 Btu.  Duplicating the 100% coal  firing
particulate loading levels leaving the stack should be feasible by
optimizing the flue gas conditioning system (ammonia injection).

Flyash resistivity ranged,from approximately 10   to 10   ohm-cm for
reburning and 10    to 10   ohm-cm for 100% coal firing over a load
range from 86 to 108 MWe.  This change was due to less than optimal
SO, levels with reburning due to ammonia conditioning.   The net
effect of a lower inlet loading with the higher flyash resistivity
resulted in a reduction in ESP efficiency.  ESP efficiency at full
load was 99.3% with 100% coal and 98.0% with 18% reburning.  This
trend was similar at partial load.

At the inlet to the ESP particle size distribution (PSD) tests were
conducted and it was found that PSD was a function of cyclone loading.
Full load PSD test results are shown in Figure 15.  The bulk of the
particulate is above 10 micrometer diameter in size.  There is a
significant decrease in the amount of particulate above 8 micrometer
diameter when reburning was used during full load tests.  The size
distribution results shown in Figure 15 reflect  the impact that
reducing the cyclone coal-loading has on lowering the amount of
particulate carryover.

The testing at 86 MWe (Figure 16) shows that between the 5 to 10
micrometer particle diameter size range there was an increase of
particulates as a result of reburn.  Above 10 micrometer there was
virtually no change in particulates.  The coal loading to the cyclones
is sufficiently low that further reductions in cyclone loading (i.e.,
reburning at partial load) do not impact particle carryover and PSD.
Note that the 108 MWe reburn test and the 86 MWe baseline tests have
nearly identical cyclone firing conditions (coal flow and excess Op)
and very similar particle size distributions.

ASH DEPOSITION IN THE SECONDARY FURNACE

Observations of the secondary furnace, particularly the back wall,
during the 6 months following installation of the reburn system
indicated a greater buildup of ash deposits than had previously been
seen.  Following shutdown of the boiler during the planned year-end
outage in December 1990, the presence of thicker ash deposits on the
back wall of the secondary furnace was verified.

The presence of the thicker deposits did not affect operation of the
reburn system insofar as NO  reduction is concerned and did not appear
to affect boiler thermal performance.  However, due to the short
duration of boiler operation with the reburn system in place, the
effect of deposition on boiler thermal performance is not conclusive.
                                  3-39

-------
It has been estimated that the reburn system was operated with natural
gas only about 20% of the time from June 25 to the late December 1990
outage.  Boiler deslagging was completed during the year-end 1990
outage; one month later it was observed that thicker ash deposits
again resulted with absolutely no natural gas having been injected.
Given that the sister unit at Miles Station (Unit No. 2) burns the
same coal under the same conditions of load and excess air as No. 1,
it can be reasonably concluded that heavier ash is depositing  under
the non-operational mode of the reburn system; i.e., with nozzle
cooling flue gas only.  This is not to say that heavier ash would not
have also deposited if the reburn system was in operation, although
some arguments can be made to suggest that deposits in the reburn zone
might well be thinner during reburn system operation.

Although the root cause of the heavier ash deposition has not been
firmly established at this time, it has been hypothesized that the
heavier ash deposition on the back wall of the secondary furnace is
being caused primarily by the recirculated flue gas forming a cooler
boundary layer along the back wall; other contributing factors could
also be the particulate in the recirculated flue gas and the new studs
that have been installed on each of the five new panels on the back
wall.  A plan is being formulated to confirm or refute the hypothesis
after which a solution will be determined.

SUMMARY

A  reburn system was installed by the end of June 1990 on Unit No. 1 at
Ohio Edison's Miles Station.  Following system shakedown, parametric
testing was carried out during the last 2\ months of 1990.  Key
findings from a preliminary review of the data collected are as
follows:

•  NO  reductions ranged from 30 to 70% during parametric
   testing.

i  NO  reductions in the 50 to 60% range are possible with
   acceptable boiler operation and CO emissions.

0  Waterwall heat absorption decreased by approximately 5% and
   convective pass heat absorption increased by 5% with 18%
   natural gas reburning.

•  Boiler efficiency decreased by 0.6% with 18% natural gas
   reburning due principally to higher latent heat of vaporization
   losses because of fuel moisture formation.

•  Furnace outlet gas temperature increased slightly with reburning
   (0°F change to 130°F) increase at various boiler loads and
   conditions.

c  ESP collection efficiency was lowered slightly with the reburn
   system in operation due to lower ESP inlet loading with similar
   outlet loading and a non-optimized flue gas conditioning system.
                                 3-40

-------
•  Thicker ash deposits have formed on the secondary furnace back
   wall since reburn system installation.

The unit is currently being operated in a baseline mode, the primary
purpose of which is to collect tube wastage data.  Following a planned
mid-year outage the unit will be operated continuously in a reburn
mode for 6 months, nominally ending in February 1992.
                                 3-41

-------
                           ACKNOWLEDGEMENTS
The Natural Gas Cyclone-Fired Reburn Demonstration program  is
sponsored by a number of organizations with significant contributions
by many participating organizations.  The authors would like to
gratefully acknowledge the high levels of cooperation, excellent
technical advice, and support in a number of specific areas from  the
following people.

Ohio Edison:
   J. Dulovich
   S. Brown
   R. Bolli
   R. Rook
   Plant Operators

EPA:
   J. Ford

Energy Systems Associates:
   B. Breen
   G. Dusatko
   J. Bionda
   R. Glickert

Physical Sciences Inc., Technology:
   S. Johnson

Research Triangle Institute:
   G. Tatsch

North Carolina State University:
   L. Stefanski

University of North Carolina:
   R. Ledbetter

ABB Combustion Engineering:
   A. Kwasnik
   R. LaFlesh
   P. Jennings
   A. Ingui
                                 3-42

-------
                              REFERENCES
1.  R. I. Bruck, (1987), "Decline of Boreal Montane Forest Ecosystems
    in Central Europe and the Eastern North America   Links to Air
    Pollution and the Deposition of Nitrogen Compounds," Proceedings:
    1987 Joint Symposium on Stationary Source Combustion NO  Control,
    Volume 1, EPA-600/9-88-026a (NTIS PB89-139695).         x

2.  C. Hakkarinen,  (1987), "An Overview of Environmental Issues
    Related to Nitrogen Oxides in the Atmosphere," Proceedings:  1987
    Joint Symposium on Stationary Source Combustion NO  Control,
    Volume 1, EPA-600/9-88-026a (NTIS PB89-139695).    x

3.  A. H. Johnson,  T. G. Siccama, (1983), "Acid Deposition and Forest
    Decline," Environmental Science Technology, 17:294a-305a.

4.  J. Kramlich, T. Lester, J. Wendt, (1987), "Mechanisms of Fixed
    Nitrogen Reduction in Pulverized Coal Flames," Proceedings:  1987
    Joint Symposium on Stationary Combustion NO  Control, Volume 2,
    EPA-600/9-88-026b (NTIS PB89-139703).      x

5.  C. Kruger, G. Haussmann, S. Krewson, (1987), "The Interplay
    Between Chemistry and Fluid Mechanics in the Oxidation of Fuel
    Nitrogen from Pulverized Coal," Proceedings:  1987 Joint Symposium
    on Stationary Source Combustion NO  Control, Volume 2,
    EPA-600/9-88-026b (NTIS PB89-139703).

6.  M. Toqan, et al., (1987), "Reduction of NO  by Fuel  Staging,"
    Proceedings:  1987 Joint Symposium on
    Stationary Source Combustion NO  Control, Volume 2,
    EPA-600/9-88-026b (NTIS PB89-139703).

7.  J. Freihaut, W. Proscia, D. Seery, (1987), "Fuel Bound Nitrogen
    Evolution During the Devolatilization and Pyrolysis of Coals of
    Varying Rank,"  Proceedings:  1987 Joint Symposium on Stationary
    Source Combustion NO  Control, Volume 2, EPA-600/9-88-026b (NTIS
    PB89-139703).       X

8.  R. W. Borio, et al., (1989), "Application of Reburning to a
    Cyclone Fired Boiler," Proceedings:   1989 Joint Symposium on
    Stationary Combustion NO  Control, San Francisco, CA, Volume 1,
    EPA-600/9-89-062a (NTIS PB89-220529).

9.  Y. Takahashi, et al., (1982).  "Development of 'MACT' In-Furnace
    NO  Removal  Process for Steam Generators," Proceedings of the
    19&2 Joint Symposium on Stationary Combustion NO  Control, Volume
    1, EPA-600/9-85-022a (NTIS PB85-235604).        x

10.  H. Farzan, et al., (1989), "Pilot Evaluation of Reburning for
    Cyclone Boiler  NO  Control," Proceedings:  1989 Joint Symposium on
    Stationary Combustion NO  Control,"  San Francisco, CA, Volume  1,
    EPA-600/9-89-062a (NTIS PB89-220529).
                                 3-43

-------
                              UNIT CONVERSION TABLE
  To Convert From


British Thermal  Units

British Thermal  Units/
Hour/Square Feet

British Thermal  Units/Pound

Degrees Fahrenheit

Feet

Inches

Pounds/Square Inch
Pounds/100 British
Thermal Units

PPM at 3% 0,,
 PPM at 3% 0,
Square Feet/(Actual
Cubic Feet/Minute)
  To


Joules

Watts/Square Meter


Joules/Kilogram

Degrees Celsius

Meters

Centimeters

Kilograms/Square
Centimeter

Kilograms/Joule
Milligrams/Cubic Meter
at 6% 02

Pounds/106 British
Thermal Units *
Multiply By


 9.478xlO"4

 0.3171


 2.326xl03

 (T-32J/1.8

 0.3048

 2.54

 14.223


 2.326xl03
 1.70
 1.306x10
                                                                      -3
Square Meter/                0.305
(Actual Cubic Meter/Minute)
   *  This conversion factor is based on composite fuel  analysis  with 18%,  natural
gas reburn fuel and 82% coal.  For different fuel fractions  different conversion
factors would be required.
                                       3-44

-------
    Table 2: Ohio Edison Miles Unit No. 1,
          Boiler Performance Data
Net Load (MWe)
Fuel Type (%)
Coal
Natural Gas
Excess Oxygen (Air) (%)
Flue Gas Recirculation (%)
Flows (Ib/hr)
Main Steam
Superheat Spray
Reheat Spray
Temperatures (°F)
Feedwater
Primary Superheater Outlet
Secondary Superheater Inlet
Secondary Superheater Outlet
Reheat Inlet
Reheat Outlet
Flue Gas-Air Heater Inlet
Flue Gas-Air Heater Outlet
Air-Air Heater Inlet
Air-Air Heater Outlet
Heal Absorption (106 Btu/hr)
Primary Superheater
Secondary Superheater
Reheater
Waterwalls
Baseline
108
100
0
2.8(15.0)
,) 1.3
845,200
10,260
0
483
et 736
ilet 718
)utlet 997
682
988
680
5t 251
120
575
u/hr)
124.5
160.7
123.2
590.3
Reburn
108
82.8
17.2
2.7(14.2)
4.5
843,700
37,760
2,320
484
758
703
1000
685
1000
685
250
118
581
132.6
170.7
127.9
563.0
   Table 3: Ohio Edison Niles Unit No. 1,
             Boiler Efficiency
 PERCENT COAL
 PERCENT NATURAL GAS

 HEAT LOSSES (PERCENT)
 DRY GAS LOSS
 MOISTURE IN FUEL LOSS
 MOISTURE IN AIR LOSS
 RADIATION LOSS
 ASH PIT LOSS
 HEAT IN FLYASH LOSS
 PYRITE REJECTION LOSS
 CARBON LOSS
 TOTAL LOSSES %
 EFFICIENCY %
BASELINE

   100
    0
REBURN

  82.8
  17.2
2.70
4.35
0.06
0.24
0.54
0.01
0.00
1.39
9.29
90.71
2.66
5.34
0.06
0.24
0.44
0.01
0.00
1.15
9.90
90 10
                   3-45

-------
                                              1987     1988
          1989         1990         1991          1992
TASK NO. TASK TITLE || |
1 TECHNOLOGY REVIEW BBBHi
2 BASELINE CHARACTERIZATION TESTS

3 REBURN SYSTEM DESIGN ••••
4 SYSTEM FABRICATION / INSTALLATION / STARTUP

5 PARAMETRIC GAS REBURN AND BASELINE TEST

6 LONG TERM PERFORMANCE TEST

7 PERFORMANCE ANALYSIS ,' TECHNOLOGY TRANSFER
8 SITE RESTORATION (OPTIONAL)

I I
•
...
A
••1
m








I I I I I I



™
mmm








I I I I I I I




••«
B
BB
c





III III I

..^
D





•••

••••••••


M I II








•
E
•
—

           SIGNIFICANT EVENT KEV
A HWnt datn for rctxjrn system design tesl program @ Miles (2/24/88-3/11/88)
R Outage to install rebum system / UT mapping (5/20/90-6/19/90)
C Outaue twUT mapping (12/P6/90 1/3/91)
D Outage tor UT mapping (Date to coincide with planned annual outage)
E Oulagefar UTmappIng (12/26/91-1/5/92)
                               Figure 1: Overall Project Scope and Schedule of Gas
                                    Reburn Project at Ohio Edison Miles Unit 1
                                     Figure 2: Upper Fuel Injector Windboxes
                                                         3-46

-------
                                                                                  SUPERHEAT/REHEAT
                                                                                  CONVECTIVE PASSES
                                                              CBS '
                                                           RECIRCULflTIOH
                                                              FHH
                             Figure 3: Boiler Exit Gaseous Emissions Sample Matrix
•O  MO

I
*
8
                                      A 86 MWe Net

                                      a 108 MWe Net
                      CYCLONE 02,%
                                                                                           o No Natural Gas

                                                                                           a 10% Natural Gas

                                                                                           a 14% Natural Gas

                                                                                           + 18% Natural Gas
                                                            O.DO   0.65   0.90  0.95   1.00   1.05   1.10   1.16   1.20

                                                                      REBURN ZONE STOICHIOMETRY
      Figure 4: Baseline NOx Versus Cyclone O2
             108 MWe and 86 MWe Net
                                                        Figure 5: NOx Versus Reburn Zone Stoichiometry at
                                                                     Various Gas Flow Rates
                                                                       108 MWe, 10%FGR
                                                  3-47

-------
                                   a1.5-2.0CydoneO2

                                   02.0-2.5 Cyclone Oz

                                   & 2.5-3.0 Cyclone Oz
                  NATURAL GAS FLOW,%
    Figure 6: NOx Versus Cyclone O2 at Various
                 Gas Fbw Rales
               108MWe,  10% FOR
800-
TOQ-
£ »-
o.
<•> 400-
1 «-
200-
100-





a Baseline
n 18% Natural Gas
|
4 6 8 10 12 14
FLUE GAS RECIRCULATION FLOW,%
 Figure 7: NOx Versus Percent Flue Gas Recirculation
                       (FGR)
         IDS MW» Met

         o No Natural Gas

         a 10% Natural Gas

         + 14% Natural Gas

         a 18% Natural Gas
                                   86 MW« Net

                                   • No Natural Gas

                                   • 9 4% Natural Gas

                                   • 183% Natural Gas
             o.e       i.o       1.1       1.2

               REBURN ZONE STOICHIOMETRY
 Figure 8: NOx Versus Reburn Zone Stochiometry at
              Various Gas Flow Rates
           108 MWe and 86 MWe Loads
1000O-


1000-
0.
0
o
100-

-3

° UFITilt@-17°
0 UFITilt@0°
A UFITilt@-25°

^^_ 	 	 —•*-
^-~_ _D____^^-^^<^
^-^___ __^-—~-^~^
D
0 -20 -10 0 10 20 30
ADDITIONAL AIR VERTICAL TILT,degree8 from horizontal
Figure 9: CO Versus Additional Air Tilt at Several Upper
                  Fuel Injector Tilts
      108 MWe Net, 5% FGR, 17.5% Natural Gas
o
o
                                 CO, ppm
                                          Flgur* 3 «how*
                                        0  mtmple location
      DISTANCE FROM DUCT RIGHT SIDEWALL.ft
Figure 10: 02 and CO Versus Boiler Exit Duct Sample
                     Location
             Non-Optimized Operation
                                                               17.1 % Natural CM
                                                               2.B* Cyolon* 07
                                                               -17" UFl Tilt
                                                               -10° AA TIH
                                                               Clockwtaa AA Yaw
                                 - CO, ppm
                                                                                                   Ffgur« 3 ahowt
                                                                                                 o •ampla location
       I   3  «  9  12  1$ 18 21  24  27  30  33  38

        DISTANCE FROM DUCT RIGHT SIDEWALL.ft
 Figure 11: O2 and CO Versus Boiler Exit Duct Sample
                      Location
                 Optimized Operation
                                                   3-48

-------
3000
2800
°o 2600
1
E 2400
CD
1-
2200
2000

CYCLONE A CYCLONE C CYCLONE D
\
\ "^~\ Baseline
... \ . \ \ 0
'" \ ^
	 Baseline
	 Reburn
246810 246810 246810
Distance from Rear Wall, ft




2600
u_
°v 2400
13
E 2200
1—
2000
1800
CYCLONE A CYCLONE C CYCLONE D
. \ 	 x\ TT\
'"~~-- ~Xv-- x\
\
	 Reburn
246810 2468 10 246810
Distance from Rear Wall, ft
Figure 12: Flue Gas Temperatures
Reburn Zone Inlet (108 MWe Net)
Figure 13: Flue Gas Temperatures
 Reburn Zone Inlet (86 MWe Net)
opnn, 	
C.C.\)\J
2100

2000
,, 1900
5-

8" 1800
3
$
1 1600
1500
1400
1300
1 200
108 MWe Net 86 MWe Net

/ ""•"'
,/ 	 ^ 	
' / ~~ / ^

,' / s — / '
/ 1
/ I . •'
f
1
	 Baseline
	 Reburn
, 	 	














2 4 6 8 10 12 2 4 6 8 10 12
Distance from Left Side Wall, ft


















E 15
3 14
? 13
° 12
f- 11
a 10
0)
E 9
D 8
o _
i e
1 5
1 3
% 2
« ,
"E 1






•
r\
\
A
i \\
o 1 00% Coal I i VA
& 18% Reburn 1 I \
I i \
/\< \
^0, ^Ck •
1
0)
3j
g
to
1
LLJ

15
14
13
12
11
10

9
8
7



5
4
3
2
1
Q














f\
o1 00% Coal I \
\
a 18% Reburn A \
/ i \
t^'' / \M,,


















0.01 0.10 1.00 10.00 100.00
Aerodynamic Diameter, IJ/TI
                          Figure 16: Particle Size Distribution
                                   (86 MWe Net)
                                       3-49

-------
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.  L.  King
                            Babcock Energy Limited
                               Porterfield Road
                          Renfrew, PA4 8DJ,  Scotland


                       J. Macphail  Babcock  Energy Limited
                               Technology Centre
                                  High Street
                           Renfrew PA4 8UW, Scotland

-------
                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.L. King
                              Babcock Energy Limited
                                 Porterfield Road
                            Renfrew, PA4 8DJ, Scotland.
                                    J. Macphail
                              Babcock Energy Limited
                                 Technology Centre
                                    High Street
                            Renfrew, PA4 8UW, Scotland.
ABSTRACT

In June 1987 after two years of development, sixty 37 MW(t) Mark I Low NOx Axial
Swirl Burners were retrofitted to Drax Unit 6.  This is a highly rated opposed
wall pulverised fuel fired boiler, firing a typical UK bituminous coal.  Baseline
preconversion NOx levels were 830 ppm at 3% 0 .   Subsequent to the retrofit, NOx
reductions of 25 to 30% were achieved, but could not be maintained due to ash
deposits local to the burner quarl interfering with the desired near burner
aerodynamic flow pattern.  A detailed investigation on the plant, using on line in
furnace video probing led to modification of the throat refractory arrangement, a
modification which resulted in deposit elimination.


After testing and demonstration at full scale in the Babcock Energy Large Scale
Test Facility an improved burner design was retrofitted to Drax Unit 6.  NOx
levels in the range 350 to 390 ppm at 3% 0  have been achieved, a reduction of
over 50%, with no significant change in the overall boiler efficiency.  Quarl
slagging has been eliminated on Unit 6 and plans are in hand to retrofit further
burners at Drax in 1991/92.


In addition to describing the results and experience obtained on Drax Unit 6,
results are also presented for a 48 MW(t) Low NOx Axial Swirl Burner fired in the
Babcock Energy Large Scale Test Facility with a range of coals which represent the
extremes of NOx related bituminous coal properties traded on a world wide basis
for utility boilers.
                                       3-53

-------
INTRODUCTION

In the mid 1980's the then CEGB initiated a series of three year trials on
operating boilers to establish the NOx reduction that could be achieved by
combustion modification, the effect on boiler operation of these modifications and
any increased operating costs that might result.  Drax with 6 off 660 MW opposed
wall fired boilers of similar design represented both a major proportion of the
coal fired generation capacity and had a size of burner (37 MW(t)) which had
application to other boilers.  In June 1987, during a planned outage, Drax Unit 6
was retrofitted with 60 off Babcock Energy Mark I Low NOx Axial Swirl Burners,
which had been developed by Babcock Energy during 1985 and 1986.  Prior to the
retrofit in March 1987, the performance and NOx output of the unmodified plant was
determined in a  'preconversion'  test, the condition of the unmodified plant being
judged to be similar to that of a boiler returned to service after overhaul as
less than 6000 hours of operation had been accumulated.  Subsequent to the
retrofit, a further series of characterisation trials were to be performed to
determine the efficiency of the conversion.
PLANT DESCRIPTION

There are six 660 MW(e) units at Drax Power Station.  The boilers are of Babcock
Energy design, Units 1 to 3 being ordered in 1966, with Units 4 to 6, which are
essentially of similar design, being ordered in 1978.   Operational experience with
all six units now covers over 400,000 hours service.  The boilers operate at a
superheater outlet pressure of 165.5 bar, and 568 C steam temperature.  The
furnace design is very highly rated, being designed for maximum combustion
efficiency, having a burner belt heat release rate of 1.6 MW/m    The furnace
chamber of each boiler is divided by a partial central division wall, which cannot
be sootblown.  Thirty standard Babcock Energy circular turbulent burners, supplied
by five mill groups are arranged in five horizontal rows on the furnace front
wall, and thirty on the furnace rear wall.  Each burner row is fed from one mill,
there being 10 Babcock Energy 10E vertical spindle mills in total.  The full
specified range of coals can be covered at MCR with nine mills; for the typical
design coal MCR can be achieved with 7 or 8 mills in service.  (See Table 1 for
the original plant fuel specification).


Air supply to each mill group of burners is controlled by individual dampers to
each windbox/mill group.  Each burner has a central oil lightup burner, of Spectus
tip shut off design on Units 4, 5 and 6, and of Babcock Energy Y-jet design on
Units 1, 2 and 3.  An integral core air fan to provide stoichiometric combustion
air for the oil burners is installed on Units 4, 5 and 6.  The oil burner is rated
at 0.25 kg/sec of Class 'G' residual fuel oil, (equivalent to 20% of the total
boiler heat input) and is used for boiler warm up and coal ignition/stabilisation
duties.


Preconversion NOx levels at 100% boiler load are summarised in Figure 1.  In all
cases, NOx is stated corrected to 3% 0  dry at the ID fan outlet.  At 3% 0  at the
economiser outlet, total NOx emission levels were 832 ppm with the eight top mills
firing,  and 747 ppm with the eight bottom mills firing.  Combustion efficiency
loss was typically 0.3 to 0.4% GCV, which corresponds to approximately 1% carbon
in ash.   Fuel characteristics associated with the preconversion test, which are
typical of the fuel normally fired at Drax, are presented in Table 2.
                                      3-54

-------
LOW NOx BURNER DESIGN

Initial Development

The initial development of the Low NOx Axial Swirl Burner has been presented
elsewhere (1) but in summary the programme involved development at 12 MW scale,
followed by the installation of two 37 MW(t) burners in a 500 MW front wall fired
plant.  The process performance of these burners was assessed by using inflame
probing technigues, and the burner mechanical integrity was also demonstrated.


Mark I Low NOx Axial Swirl Burner Design Features

In the Low NOx Axial Swirl Burner, the combustion is staged within the burner, all
the combustion air being passed through the burner throat opening, and the mixing
of the combustion air with the fuel being controlled by burner aerodynamics.  The
Mark I burner has the following features, Figure 2:-

    i)    A light up oil burner with an integral combustion air (referred to
          as core air) supply.  The light up burner, which is used for
          furnace heating up and low load support for the coal flame, has a
          heat input of 20% of the main coal burner.  The air supply for the
          oil light up burner is supplied from a core air fan mounted on the
          windbox end of the core air tube, sufficient air being supplied
          for near stoichiometric combustion of the oil.  This approach
          ensures excellent oil light up capability and improved control of
          the oil flame under cold furnace conditions.  Under coal firing
          conditions, there is no reguirement for the core air fan to be in
          operation.

   ii)    Pulverised fuel enters the burner through an elbow and then
          passes through an annular pulverised fuel pipe, surrounding the
          core air tube.  The elbow is designed so as to decelerate the flow
          entering it, the deceleration resulting in a redistribution of the
          pulverised fuel.  Subsequently, as it exits the elbow, the
          pulverised fuel is re-entrained uniformly into the pulverised fuel
          pipe.  The fuel elbow is lined with a wear resistant chromium iron.

  iii)    The combustion air is subdivided into two streams, referred to as
          secondary and tertiary air, the secondary air flow being controlled
          by a secondary air damper.  A swirling motion is applied to the
          secondary air stream by an axial swirl generator, Figure 3.  This
          is a more efficient swirl generation process than the conventional
          radial swirl generator, and mechanically less complex.  The level
          of swirl imparted to the secondary air stream is controlled by the
          axial position of the swirl generator in the conical section of the
          secondary air barrel.

          Minimum swirl is obtained with the secondary air swirl generator
          retracted out of the conical section, a proportion of the air
          bypassing the swirl generator.  Maximum swirl is obtained with the
          swirl generator fully inserted into the conical section, thereby
          allowing no air bypass.
                                     3-55

-------
          The  tertiary air stream is subdivided into four distinct streams
          by the  use  of 'splitters'  in the tertiary air annulus.   No
          swirling motion is  applied to the tertiary air.

   iv)     Due  to  the  high incident heat fluxes on the burner (greater
          than 500 kW/m ), burner components nearest the furnace  are
          manufactured from heat resisting material.

    v)     Two  flame monitors  are fitted,  one for the oil light up burner
          flame and one for the coal flame.  Both monitors are of the
          Babcock dual signal type.
Retrofit Requirements

The Mark I burner was installed in Drax Unit 6 with the minimum of retrofit
modifications.   No modifications were made to the windbox or pf pipework
positions, and the existing flame monitors were reused, as was the oil light up
burner assembly.  Modifications were necessary to the burner throat profile to
allow the slightly larger throat diameter of the Low NOx burner to be
accommodated.  These modifications did not involve any tube alterations, the
refractory tiles which are used to form the throat opening being reduced in
thickness.  The burners were supplied from the Babcock Renfrew Manufacturing
facility fully assembled, Figure 4, to enable ready insertion into the appropriate
windbox, and were retrofitted to the boiler during a 35 day outage in June 1987.
MARK I LOW NOx AXIAL SWIRL BURNER PLANT RESULTS

NOx Emission Levels

Following installation of the Mark I burners during a period of routine operation
set aside to allow the burners to be optimised, overall NOx reduction levels of
some 25 to 30% were recorded.  However these readings were not maintained and with
time NOx levels gradually rose and stabilised at a level some 10 to 15% below the
preconversion value.  Alteration of burner settings had no significant effect on
the NOx levels being obtained.  A rapid boiler shutdown led to significant ash
deposits being shed from the furnace chamber, as judged by the amount of bottom
ash to be cleared.  On return to load, a NOx reduction of some 25 to 30% was
achieved, but then subsequently stabilised over a period of days to a level 10 to
15% below the preconversion level.


Relationship between NOx Levels and Burner/Ash Deposit Accumulation

The foregoing observations indicated that a reasonable level of NOx reduction was
achieved when the boiler was free of deposit accumulation, and that the burner NOx
reduction apparently fell away after a period of boiler operation.  It was not
clear at this stage as to whether the increase in NOx levels was due to an
increase in thermal NOx as the furnace chamber became progressively dirtier, or
due to the effect of deposits on burner operation and in particular interference
with the desired near burner aerodynamic flow patterns.
                                        3-56

-------
Figure 5 is a photograph of a typical deposit pattern in the burner belt area
immediately after the unit came off load.  Whilst the deposits are probably of no
greater magnitude than those associated with the standard circular burner, the
topographic structure of the deposit is different.  The structure associated with
the low NOx burner illustrates a clover leaf type pattern, as compared to the more
conical frustum pattern associated with the standard circular burner.  Closer
examination shows that the indentations in the clover leaf pattern associated with
the low NOx burner correspond to the location of the tertiary air splitters.


Simulation of these deposits on a 12 MW test burner indicated that the extent of
the deposit was sufficient to interfere with the near burner aerodynamics, and
significantly to increase NOx levels.


Quarl Deposit Mechanisms

In order to provide a better understanding of the mechanism of quarl deposition, a
water cooled video probe developed by the CEGB was used (2).   The probe was
extremely useful in being able to identify the various mechanisms involved in
quarl slagging as they occurred.  These mechanisms can be summarised as follows:-

    i)    Partially sintered ash particles, up to 25 mm in size, which had
          initially been deposited on the upper furnace walls, roll slowly
          down the furnace wall, the aggregates maintaining their basic
          shape and form as they roll down the wall.

   ii)    When the aggregate encounters a refractory region (e.g. the
          front face of a refractory throat tile) its motion is arrested
          and the aggregate adheres to the refractory surface.

  iii)    Deposit build up continues in all directions with successive
          aggregates adhering both to the refractory surface and to other
          already adhered aggregates.

   iv)    As the deposit build up increases in size then the deposit
          surface becomes progressively more molten and fused, and in
          addition to attracting aggregates the deposit grows further due
          to the collection of airborne ash from the furnace chamber.
          Build up of deposit generally occurred more rapidly on out of
          service burners.

    v)    The aerodynamic effects of the tertiary air splitters
          result in ash being drawn back into the divergent section
          of the burner, producing the typical clover leaf pattern.


To overcome the deposition on the tile face, the quarl tube assembly on 4 burners
was modified to replace the tile face with a furnace tube, so that there was
minimal refractory surface available for ash adhesion.  Subsequent operation
showed that the modified quarls remained free from deposits, with major reductions
in the extent of 'eyebrow' deposition.
                                       3-57

-------
There was however a tendency for ash to build up downstream of the tertiary air
splitters, even after additional ventilation of the splitters was introduced.  It
was therefore concluded that:-

    i)    Quarl deposit formation could be virtually eliminated by
          attention to the detail of quarl design.

   ii)    Whilst the principle of the tertiary air splitter design was
          satisfactory, their application in a high temperature coal ash
          situation was unacceptable and an alternative approach was
          necessary.


Consequently an alternative burner design was developed and tested on the Babcock
Energy Large Scale Test Facility at Renfrew, prior to modification of the Mark I
burners  in Drax Unit 6.
MARK III LOW NOx AXIAL SWIRL BURNER

Burner Development

The development of the Mark III burner design has been fully described elsewhere
(1), the final Mark III design, Figure 6, evolving through an intermediate Mark II
design.  The Mark II design differed from the Mark I design in that:-

    i)    Tertiary air swirl was introduced to provide improved
          aerodynamic characteristics compared to the aerodynamic effect
          of the tertiary air splitters.

   ii)    Controlled fuel distribution within the burner was introduced,
          as a result of further development testing at 12 MW (1).


Concern over the scaling criteria of several processes simultaneously from the
12 MW test burner to the 40 MW plant burner led Babcock Energy to invest in a
Large Scale Test Facility in Renfrew (Figures 7 and 8).  Prior to demonstrating
the Mark II burner in the Test Facility, the facility was calibrated using a
standard circular turbulent burner.  The coal quality and fineness was similar to
that used in the Drax Unit 6 preconversion tests.  The results from the circular
turbulent burner are presented in Figure 9, together with the results from the
Mark II design and an improved Mark III design.  The Mark III design differed from
the Mark II design in that the position of the ends of the burner tubes were
optimised, and means were introduced to improve the fuel ignition characteristics.
At 3% operating 0 , NOx emissions are reduced from 722 ppm to 300 ppm, an overall
reduction of 58%.  The Mark III burner NOx emission levels correspond to a 60%
reduction in fuel NOx levels, a figure considered to be close to the maximum
reduction that can be achieved with an internally staged low NOx pulverised fuel
burner.
Following the demonstration of the full scale Mark III burners in the Large  Scale
Test Facility, 60 Mark III burners were installed in Drax Unit in August  1989.
Prior to the installation of the Mark III burners, all sixty burner quarls had
been modified to the design developed during the Mark I burner investigation, and
                                        3-58

-------
the Mark I burners modified to Mark II design.  The satisfactory operating
experience with the Mark II design has been presented elsewhere (1), the enhanced
NOx reduction capability of the Mark III design (Figure 9) making it an attractive
retrofit to the Mark II design.


Mark III Plant Burner Optimisation

Experience from the Mark III burner development tests had shown that the NOx/
unburned carbon characteristics of the burner could be  'optimised1 by adjustment
of the secondary air swirl level (via the axial movement of the secondary air
swirler) and the secondary air flow rate (controlled by the secondary damper).


The results of the series of burner optimisation trials with the Mark III burner
on Drax Unit 6 are summarised in Figure 10.  The figure clearly demonstrates how
NOx and unburned carbon can be optimised for the Drax situation by adjustment of
the burner settings, the results obtained on the plant reproducing the
characteristics of the burner in the test facility.  Accordingly for the
demonstration of the Mark III burner in Unit 6 the burner settings were those
corresponding to Test 3, i.e. the minimum carbon in ash settings.
Burner Demonstration

The results of the demonstration tests performed with the Mark III burner design
in September 1989, approximately three weeks after installation, are presented in
Figure 11, whilst Table 3 summarises the overall boiler performance for both the
preconversion case and with the Mark III burners installed.  Essentially:-

    i)    Overall NOx levels are reduced from 832 to 381 ppm, a reduction
          greater than 50%, the figure obtained with eight top mills in
          service being lower than the current EC directive for new boiler
          plant.

   ii)    Carbon in ash levels have increased from 1 to 1.5% preconversion
          to around 2.5% with the Mark III design.  However, greater furnace
          heat absorption and a consequent reduction in gas temperature
          leaving the airheaters offset this efficiency loss.


Other Effects of the Low NOx Modification

Furnace Performance.   A significant result of the installation of the Mark III
burners has been the reduction of arch level flue gas temperatures by 50 to 100 C
compared to the preconversion situation.  This reduction in exit gas temperatures
has not posed any problems in operation of the plant, or in maintenance of the
required steam conditions, there now being less attemperator spray water flow than
preconversion.   The reduction is attributed to the furnace water walls being
significantly cleaner with the Mark III burner design compared to the standard
circular turbulent burner, as a direct result of the improved control of fuel and
air flow within the burner required for efficient low NOx combustion.  Flame light
off,  burner stability and turndown is excellent.
                                      3-59

-------
Side Wall Wastage.   The improved furnace performance would suggest that local wall
reducing conditions are not present, and this has been corroborated by subsequent
furnace tube thickness measurement which showed very low wastage rates.


Quarl Slagging.  The modifications carried out on the burner quarls have resulted
in the quarls remaining free of significant ash deposition during service, with
any deposition in out of service burners being rapidly cleared away on the burners
return to service (Figure 12).


Long Term Performance

Performance tests have been performed on Unit 6 at roughly six monthly intervals
since the Mark III retrofit in August 1989.  These tests show that the low NOx
characteristics of Unit 6 have been maintained, with no problems associated with
quarl or furnace deposition.  Inspection of the burner flameholder after 6 months
service showed some limited damage to the ceramic segments of the flameholder,
mainly associated with the fixing method.  Wear on the pulverised fuel side of the
burner, the components having seen over 18 months service, is very low, and no
significant problems are anticipated with those or any other components meeting
the specified burner component life of 38 months (= 25,000 hours).


Plans are in hand to retrofit more Mark III Low Nox Axial Swirl burners at Drax in
1991/92, and further orders have been received from both the UK and the Far East
for Mark III burner retrofits.
THE EFFECT OF COAL QUALITY ON LOW NOx BURNER PERFORMANCE

Introduction

As noted in the previous section, the performance of the Mark III Low NOx Axial
Swirl Burner has been proven in a highly rated opposed wall fired boiler
environment, firing a typical UK bituminous coal of sensibly constant fuel
properties.  In order to demonstrate the performance of the burner firing
bituminous coals whose NOx related properties differ significantly, a series of
tests were performed on a 48 MW(t) Low NOx Axial Swirl Burner installed in the
Babcock Energy Large Scale Test Facility in Renfrew.


Coal Properties

Three coals were selected, Table 4, whose properties represent virtually the
extreme, from a NOx emission point of view, of bituminous coals fired on utility
boilers and traded on the world market.  The essential coal properties can be
summarised as follows:-

    i)    Indonesian coal, with a high volatile matter content and a high
          inherent moisture level and nitrogen content.

   ii)    A UK coal, similar to that fired on Drax, and chosen to provide
          a link between the 37 MW and the 48 MW burner data in the
          Test Facility.
                                        3-60

-------
  ill)    A South African coal, with a low volatile matter content and a
          high nitrogen content.


The properties listed in Table 4 are on an as fired basis i.e. after the coal had
been pulverised off site and delivered to the test facility.


Results Obtained

Figure 13 shows the variation of NOx emissions with operating oxygen for the three
coals with the burner firing at 45 MW.  At 3% operating oxygen, NOx emissions
range from 250 ppm with the Indonesian (low fuel ratio) coal, to 435 ppm with the
South African (high fuel ratio) coal.  A value of 380 ppm is obtained with the UK
coal.  This is higher than that obtained with the 37 MW burner design due to two
factors i.e. different burner settings and increased thermal NOx in the test
facility.  For all three coals, CO levels and carbon in ash levels are typical of
those obtained in the Test Facility, being at 3% operating oxygen, typically 400
ppm and 4 to 6%.  Carbon in ash levels with the South African low volatile coal
tend to be slightly higher than those for the high volatile coal, as might perhaps
be expected.
CONCLUSIONS

The Mark III Babcock Energy Low NOx Axial Swirl burner has been in operation on
Drax Unit 6 since 1989, and has been producing a consistent reduction in NOx
levels of 50 to 55%.  This is an excellent achievement for a highly rated plant
such as Drax, which is now capable of operating below the EC directives.
Operationally advantageous changes in the heat transfer pattern in the boiler have
been measured, with the furnace chamber running significantly cleaner and guarl
slagging virtually eliminated.  Further burner testing in the Babcock Energy Large
Scale Test Facility has demonstrated the effect of coal quality on NOx emissions
and that a wide range of coals can be burned in an efficient and stable manner.


The Babcock Energy Low NOx Axial Swirl burner can therefore be considered as a
proven combustion technique for the clean efficient combustion of pulverised
coals.  In parallel with two stage combustion techniques and appropriate furnace
design, NOx levels commensurate with those specified in most worldwide
legislation can be achieved.  In the retrofit situation, dependent on coal quality
and furnace design, NOx levels lower than 0.5 Ib/mBtu can generally be achieved.
ACKNOWLEDGEMENTS

The authors wish to express their gratitude to the Station Management and Staff at
Drax Power Station for their assistance in all stages of the execution of this
project.


This paper is published with the permission of Babcock Energy Limited.
                                        3-61

-------
REFERENCES
     A.R. Jones, G.S.  Riley and B.M.  Downer  "The Development and Use of Special
     Probes for Investigating the Effects of Ash on Furnace Operation."  2nd
     Conference on the Effects of Coal Quality on Power Plants .   St. Louis
     Missouri, 19   21 September 1990.

     R.M. Clapp, J.L.  King and J. Macphail  "Development and Application of an
     Advanced Pulverised Fuel Low NOx Burner."  1990 International Joint Power
     Generation Conference.  Boston,  21 - 25 October 1990.
                                       3-62

-------
      900
      800
      700
      600
  NOx
P.P.M.
   AT
 3%  02
 x  8 TOP MILLS
 o  8 BOTTOM MILLS
                                I  .0
                                0.5
 GCV
LOSS
                12345
        ECONOMISER OUTLET OXYGEN  (% DRY)
               Figure 1.  Preconversion Test Results
                     Source:  Drax Unit 6.
 IBdlARY AR-i
8ECONCARY MR ,— 6ECONQART DAMPER
  3WBLER
        -lERTMRT
         aPUTTH)   MMPER
           Figure 2.  Mark I Low NOx Axial Swirl Burner
                           3-63

-------
      Figure 3.   Axial  Swirl  Generator
Figure 4.   Mark I Low NOx Axial Swirl Burner
                    3-64

-------
                 Figure 5.  Typical Deposit Pattern
MAFK HI
           Figure 6.  Mark III Low NOx Axial  Swirl  Burner
                                 3-65

-------
                                                                   AIR BLOWER
 FURNACE
GAS PROBE
                                 TO WASTE HEAT BOILERS
                                     OR ATMOSPHERE
                                                               LOSS IN
                                                            WEIGHT FEEDER
      WATER
     COOLING
       TANK
FEEDER/
EJECTOR
           FURNACE WITH
            STEAM HOOD
                              SECONDARY AIR
   PF BURNER
   OIL BURNER
                                         KEROSINE
                                           TANK
                   STEAM
                  SUPPLY
                                                         FLOW CONTROL DAMPER
 PUMP
MAKE UP
 WATER
  TANK
          Figure 7.   Schematic of Large Scale Burner Test Facility
            Figure  8.   Babcock  Large  Scale  Burner  Test  Facility
                                    3-66

-------
     cr
     Q
      C\J
     o
     0.
     Q.
     X
     o
900
800
700
600
500
400
300
200
100
STANDARD  BURNER
                             MK  II LNASB
MK  III LNASB
                 12345
                    %  02  (DRY)
              Figure 9. Test Facility Results
 MEAN
CARBON 4
IN ASH
        2
                                  x TEST  I
                                  o TEST  2
                                    TEST  3
                           TEST  4
                                 I	I	I	|	L
        320       350             400       430
                   NOx P.P.M. AT  3% 02
        Figure 10.  Summary of Optimisation Test Results
                       3-67

-------
        500

  NOx
P.P.M.
  AT    400
 3% 02


        300
        200
o 8  TOP MILLS
° 8  BOTTOM  MILLS
                  I      2      3     4.      5

            ECONOMISER OUTLET OXYGEN  (% DRY)

            Figure 11.  Demonstration Test Results
               Figure 12.  Quarl In Service
                        3-68

-------
       500
       400
       300

  NOx
P.P.M.
  DRY  200
  AT
 3% 02

        100
  SOUTH
 AFRICAN
  COAL
                                          U.K.  COAL
INDONESIAN
   COAL
                 12345

                  OUTLET  OXYGEN  (%  DRY)

         Figure 13.  Effect of Coal Quality on NOx Emissions
                         3-69

-------
                                    TABLE 1
                        DRAX COAL ANALYSIS SPECIFICATION
        Properties
Higher Heating Value KJ/Kg
Moisture                 %
Ash                      %
Volatile                 %
Sulphur                  %
Chlorine                 %
Ash Initial             °C
Deformation Temperature
Design
Basic
24 400
8
20
28
2.0
0.4
1200
Coals
Range
17 680-27 910
4-24
3-40
20-32
0.5-4.0
0.03-1.0
1000-1345
Coals Burned
to Date
17 840-29 120
3.8-16.0
5.5-33.7

1.0-2.38
0.06-0.44
1060-1350
                                    TABLE 2
                TYPICAL COAL PROPERTIES FOR THE PRECONVERSION TEST
             Coal Analysis
                            Ash Analysis
     Moisture
     Volatile Matter
     Fixed Carbon
     Ash
     Nitrogen
     FC/VM
     Nitrogen (daf)
     GCV (MJ/kg)

     Pf Fineness
     %  <   75 micron
     %  <  150 micron
     %  <  300 micron
11.0
28.8
44.2
16.0
 1.18
 1.59
 1.64
24.96
66.2 - 69.3
92.5   93.8
99.6   99.8
Silica
Alumina
Iron Oxide
Calcium Oxide
Magnesium Oxide
Titanium Oxide
Potassium Oxide
Phosphorus
Sulphur
57.8
24.1
 8.9
 1.4
 1.8
 0.92
 3.09
 0.25
 0.56
                                     3-70

-------
                                    TABLE 3

            COMPARISON OF PRECONVERSION AND MARK III BURNER BOILER DATA


                           ^reconversion                   Post Conversion
TEST NO:
DATE:
TIMES:
Unit Load (MWe)
Mills in Service
Economiser Outlet
02 (% dry)
Arch Level Temperature
FEGT
NOx corrected to 3%
02 (dry)
1
18.03.87
09.30
11.30
662
8 Top
3.75
°C 1492
°C 1130
ppm
2
18.03.87
13.00
15.00
660
8 Top
2.76
1508
1129
823
3
18.03.87
16.30
18.30
661
8 Top
2.42
1529
1141
783
Al
20.09.89
09.30
11.45
660
8 Top
4.65
1402
1075
479
                                                           A2       A3       A4

                                                         20.09.89  20.09.89 20.09.89
                                                          12.30    16.00    19.00
                                                          14.30    18.00    21.00

                                                           655      652      654

                                                          8 Top    8 Top    8 Top

                                                          3.59     2.85     3.68

                                                          1413     1447     1430

                                                          1066     1081     1088
                                   TABLE 4

                          ANALYSIS OF  COALS TESTED


                       Indonesian      United Kingdom       South African

GCV  MJ/kg                26.55              27.13                  27.12
H20    %                  10.5                3.4                    3.1
VM     %                  40.6               31.6                   25.3
FC     %                  44.7               46.8                   56.4
Ash    %                   4.2               18.2                   15.2
                                    3-71

-------
UPDATE ON COAL REBURNING TECHNOLOGY FOR REDUCING NOX IN CYCLONE BOILERS

                              A.S. Yagiela
                     Cyclone Reburn Project Manager
                            Babcock & Wilcox
                            Barberton,  Ohio

                              G.J. Maringo
                 Combustion Systems Development Engineer
                            Babcock & Wilcox
                            Barberton,  Ohio

                              R.J. Newell
                      Supervisor Plant Performance
                      Wisconsin Power & Light Co.
                          Cassville, Wisconsin

                                H.  Farzan
                       Alliance Research Division
                        Senior Research Engineer
                            Babcock & Wilcox
                             Alliance, Ohio

-------
                 UPDATE ON COAL REBURNING TECHNOLOGY
                  FOR REDUCING NOx IN CYCLONE BOILERS
                             A.  S.  YAGIELA
                    Cyclone Reburn Project Manager
                           Babcock & Wilcox
                            Barberton,  Ohio

                             G.  J.  MARINGO
                Combustion Systems  Development Engineer
                           Babcock & Wilcox
                            Barberton,  Ohio

                             R. J.  NEWELL
                     Supervisor Plant Performance
                    Wisconsin Power & Light Company
                         Cassville, Wisconsin

                               H. FARZAN
                      Alliance Research Division
                       Senior Research Engineer
                           Babcock & Wilcox
                            Alliance, Ohio
ABSTRACT

Encouraging results have been obtained  from engineering feasibility and
pilot-scale  proof-of-concept studies  of coal  reburning  for  cyclone
boiler  NOx  control.   Accordingly,  Babcock  & Wilcox  (B&W)  completed
negotiations for a Clean Coal cooperative agreement with the Department
of Energy  (DOE)  to demonstrate coal reburning  technology for cyclone
boilers.  The host site for the demonstration is the Wisconsin Power &
Light (WP&L) Company's 100 MWe Nelson Dewey Station.

Reburning involves the  injection  of  a  supplemental  fuel (natural gas,
oil,   or  coal)   into  the  main  furnace to   produce  locally  reduced
stoichiometric conditions which convert the NOx to molecular nitrogen,
thereby reducing  overall  NOx emissions.  Currently,  no commercially-
demonstrated  combustion  modification   technigues  exist   for  cyclone
boilers to  reduce NOx  emissions.   The  emerging  reburning technology
should  offer cyclone  boiler  operators  a   promising alternative  to
expensive flue gas cleanup technigues for NOx emission reduction.

This  paper reviews baseline testing results at the Nelson Dewey Station
and  pilot-scale  results  simulating  Nelson  Dewey  operation  using
pulverized  coal  (PC)  as  the reburning  fuel.   Outcomes  of  the model
studies as well as the full-scale demonstration design are discussed.

                             INTRODUCTION

The Department of Energy (DOE) under its Clean Coal II solicitation is
sponsoring  B&W and  WP&L to perform a  full-scale  demonstration of the
reburning technology for  cyclone  boiler NOx  emissions  control.   This

                                  3-75

-------
full-scale  evaluation  is  justified  via  a  previous  Electric Power
Research   Institute-sponsored    (Project   RP-1402-30)   engineering
feasibility  study  and  EPRI/GRI  (EPRI  RP-2154-11;  GRI:5087-254-1471)
pilot-scale evaluation of reburning for cyclone boilers performed by B&W
(1)(2).  The feasibility study revealed that the majority of cyclone-
equipped boilers  could successfully  apply this  technology  to  reduce
their NOx emission levels by  approximately 50%-70%(1).  The pilot tests
evaluated the potential of natural gas, oil, and  coal as the reburning
fuel in reducing NOx emissions.   The data  obtained from  the pilot-scale
project substantiated  the results predicted by the feasibility  study.
Though oil/gas  reburning could  play a  role  in reducing NOx emissions
from cyclone boilers,  B&W coal reburning  research has also shown  that
coal performs nearly as well as gas/oil without deleterious effects on
combustion efficiency.  This  means that boilers using  reburning  for NOx
control can maintain 100% coal usage instead of switching to 20%  gas/oil
for reburning.   As a result of the coal  reburning  research performed to
date, the technology has advanced to the point  where demonstration  on a
commercial scale is imminent.

Currently,  105  operating,   cyclone-equipped  utility  boilers   exist,
representing approximately 15% of pre-New Source  Performance Standards
(NSPS) coal-fired generating capacity  (over 26,000 MW).  However, these
units  contribute  approximately  21%  of the NOx  emitted  since their
inherent turbulent, high-temperature combustion process  is conducive to
NOx formation.  Although the majority  of the cyclone units are 20 to 30
years old, utilities plan to  operate many of these units for at least an
additional  10  to  20  years.    These  units  (located  primarily  in  the
Midwest) have been targeted for Phase  II Federal Acid  Rain NOx emission
limitations.

The coal reburning demonstration project for cyclone boiler NOx  control
will  be carried out  at WP&L's  Nelson  Dewey Station,  Unit  No. 2, in
Cassville,  Wisconsin.   The  unit is  a  B&W RB-type boiler  with three
cyclone  furnaces.    Unit  No. 2  is  small  (nominal  100 MWe)  to limit
project costs, but large enough  to assure  that  the reburning technology
can  be  successfully  applied  to   the cyclone-fired  utility   boiler
population.   As part  of  the project,  B&W's  six-million  Btu/hr Small
Boiler  Simulator  (SBS) pilot facility was utilized  to duplicate the
operating practices of WP&L's Nelson Dewey Unit No. 2.   The coal  that is
fired at Nelson Dewey was fired in the SBS cyclone and  also was utilized
as  the reburn  fuel.    During the  field  test phase   at  Nelson Dewey
Station, emission  and performance data was acquired  and analyzed before
the coal reburn conversion to determine the NOx reduction  and  impact on
boiler  performance.   Combining  these combustion  test  results  with
physical  and numerical modeling  of  the technology as applied to Dewey
Unit No. 2 provides a  comprehensive test program not only for successful
application  of  WP&L's  unit,  but  for the cyclone population as a whole.

From WP&L's perspective, involvement in this project was undertaken for
several reasons.   The  State  of Wisconsin  enacted  acid rain legislation
in  1986,  which will be fully implemented in  1993.   Federal acid  rain
legislation  will  require  NOx reductions from cyclone fired  boilers
beginning in 1995.  The state law requires significant reduction of SO2
emissions  and  the study  of potential  reduction  of  NOx emissions.
Approximately  50%  of WP&L's  coal-fired  capacity  is  generated  from
cyclone  boilers installed between  1952 and 1969.   These boilers are
vital to meeting the electricity  needs of WP&L's customers.  However, of
concern to WP&L  is that these cyclone boilers  produce about 75% of the
NOx emitted  within the WP&L system.   Environmental concerns have  been


                                  3-76

-------
complicated by the fact that no commercial combustion technologies exist
for controlling NOx emissions from cyclone boilers.  Based upon  WP&L's
internal  analyses of  several  advanced technologies,  coal  reburning
surfaced as the least-cost retrofit alternative.  With these reasons and
a desire to promote cost-effective emission reduction technologies, WP&L
accepted B&W's offer to participate and host this project.

               BACKGROUND/REBURNING PROCESS DESCRIPTION

The  cyclone  furnace   consists  of  a  cyclone  burner  connected to  a
horizontal  water-cooled cylinder,  commonly referred to as the cyclone
barrel.  Air and crushed coal are introduced through the cyclone  burner
into the cyclone  barrel.   The larger coal particles are thrust  out to
the barrel walls where they are captured and burned in the molten slag
layer  which is formed; the  finer particles burn in suspension.  The
mineral  matter melts,  exits  the  cyclone  furnace  from  a tap  at  the
cyclone throat, and is dropped into a water-filled slag tank.  The flue
gases  and remaining ash leave the cyclone and enter the main furnace.

No commercially-demonstrated combustion modifications have significantly
reduced  NOx emissions without  adversely  affecting  cyclone operation.
Past tests  with combustion air  staging achieved 15 to 30% reductions.
Cyclone tube corrosion concerns  due to the resulting  reducing conditions
were not fully addressed  because of the short duration of these  tests.
Further investigation of staging for cyclone NOx control was halted due
to  utility  corrosion  concern.    Additionally, since  no  mandatory
federal/state  NOx emission  regulation was  enforced,  no  alternative
technologies were  pursued.

The use of selective catalytic reduction (SCR)  technology offers promise
of controlling NOx emissions from these units, but  at high capital and
operating costs.   Reburning  is  therefore  a promising alternative NOx-
reduction  approach for  cyclone-eguipped  units  with  more  reasonable
capital and operating  costs.

Reburning is a process by which NOx produced in  the cyclone is reduced
(decomposed to molecular  nitrogen) in  the main furnace by injection of
a secondary fuel.   The secondary (or reburning)  fuel creates an oxygen-
deficient (reducing) region which accomplishes decomposition of the NOx.
Since  reburning  can  be applied  while the cyclone  operates  under  its
normal oxidizing  condition, its  effects on cyclone  performance  can be
minimized.

The reburning process  employs  multiple combustion zones in the furnace,
defined as the main combustion,  reburn, and burnout zones, as shown in
Figure  1.    The   main combustion  zone   is  operated  at  a  reduced
stoichiometry and  has  the majority of the fuel  input  (70  to 80% heat
input).  Most past investigations  on natural gas-/oil-/coal-fired units
have shown  that  the  main combustion  zone of  the   furnace  should be
operated at a stoichiometry of less than 1.0.   This operating criteria
is  impractical  for  cyclone units  due  to the  potential  for   highly
corrosive conditions,  since many  cyclones  burn high-sulfur, high-iron
content bituminous coals.   To avoid this  situation and its potential
consequences, the  cyclone main combustion  zone was determined  to be
operated at a stoichiometry of no less than 1.1  (2% excess 02) .

The balance of fuel (20 to 30%)  is  introduced above  the main combustion
zone (cyclones)  in the  reburn  zone  through  reburning  burners.   To
protect the tubes  around  the reburning burners  in  the reburning zone


                                  3-77

-------
from fireside corrosion, air is introduced through these burners.  The
burners are  operated in  a  similar  fashion  to a  standard wall-fired
burner except that they are fired at extremely low stoichiometries  (less
than 0.6).  The furnace reburning zone is operated at stoichiometries in
the range  of  0.85  to 0.95  in  order to  achieve maximum NOx reduction
based on laboratory/actual  boiler application results.   A sufficient
furnace residence time within the reburn zone is required for flue gas
mixing and NOx reduction kinetics to occur.

The balance of the required  combustion  air—totaling  15 to 20% excess
air at the  economizer outlet—is  introduced through over-fire air  (OFA)
ports.  As with the  reburn  zone,  a satisfactory residence time within
this  burnout  zone  is required for complete combustion.   These  ports
should be  designed  with  adjustable air  velocity  controls  to enable
optimization of mixing for complete  fuel burnout prior to exiting the
furnace.

                          PROJECT DESCRIPTION

The  objective  of  the  cyclone  demonstration  is  to  evaluate  the
applicability  of  the  coal  reburning   technology  for  reducing NOx
emissions in  full-scale cyclone-equipped boilers. The performance  goals
are:

     1)   Provide  a   technically  and  economically  feasible  low-NOx
          alternative for cyclone boilers to  achieve a greater than 50%
          NOx reduction where one currently does not exist.

     2)   Show significant reductions in  emission  levels of oxides of
          nitrogen achieved  at a low capital and very low operating cost
          (compared to the SCR technology).

      3)   Show that  there is no  need for a supplemental fuel.  Reburn
          will be  carried out  using the  present boiler fuel which is
          coal.

      4)   Provide  a   system that  will  maintain   boiler reliability,
          operability, and steam production performance after retrofit.

To  meet  the  above stated goals,  the coal reburn  project consists of
three separate phases:

          PHASE I -  Design and Permitting

          The  coal  reburn  system will  be designed based  upon  B&W's
          pilot-scale  combustion  tests,  physical  and  numerical flow
          modeling tests, past experience, and knowledge of full-scale
          burner/OFA port/control system retrofits.  Baseline emissions
          and performance data will be collected on WP&L's Nelson  Dewey
          Unit No. 2.

          PHASE II - Procurement,  Construction, and Start-up

               A.   Long Lead-Time Item  Procurement

                    For  schedule  purposes,  long  lead-time equipment
                    will  be  ordered during  the design and permitting
                    phase.
                                  3-78

-------
               B.   Construction and Start-up

                    The  coal  reburn system  will  be  fabricated and
                    installed at Nelson Dewey No. 2 and started up.

          PHASE III - Operation and Disposition

          Parametric/optimization  and performance  tests  will  assess
          emission reductions and boiler performance capability of the
          technology  at both  full-load  and reduced-load operation.
          Readiness for commercialization will be determined from both
          a technical and economic viewpoint.

The overall project duration is 43 months and was initiated on October
1, 1989,  with  construction  to  start in June  1991  for a November 1991
operation.  Figure 2 shows the overall program schedule.

A  summary of the overall project  organization of  participants  is as
follows:

                         Project Organization

     Department of Energy - 50% funding co-sponsor
     B&W  - Prime contractor and project manager
     WP&L - Host site utility and funding co-sponsor
     State of Illinois - funding co-sponsor
     Utility funding co-sponsors
     Acurex Corporation - testing subcontractor
     Sargent & Lundy - architect engineer subcontractor

The utility funding co-sponsors are:

1)   Allegheny Power System
2)   Atlantic Electric
3)   Associated Electric Co-op, Inc.
4)   Baltimore Gas & Electric
5)   Iowa Electric Light & Power Co.
6)   Iowa Public Service
7)   Missouri Public Service
8)   Kansas City Power & Light
9)   Northern Indiana Public Service Company
10)  Tampa Electric Company

                   SBS PILOT-SCALE SIMULATION TESTS

Technical Objectives

The technical  objectives  of the pilot-scale  combustion  tests  were to
demonstrate NOx  reductions  of nominally  50 to 60%  while maintaining
acceptable  cyclone/boiler   operating  conditions.   B&W's  six-million
Btu/hr Small Boiler Simulator  (SBS)  pilot  facility (see Figure 3) was
utilized to duplicate the operating practices of  WP&L's Nelson Dewey No.
2.  Baseline and coal reburning pilot tests were performed to evaluate
the potential applicability of this technology.   The majority of these
tests were done while firing the project's demonstration coal (Lamar -
a medium  sulfur,  1.87%,  bituminous coal from Indiana).   The numerous
parameters which are varied to help determine the technology's potential
are as follows:   main cyclone/reburning burner fuel splits, reburn coal
                                 3-79

-------
type, furnace zone stoichiometries and furnace zone residence times.  In
addition to determining NOx reduction potential, other variables such as
mixing,  corrosion,  fireside  deposition,   combustion  efficiency  and
precipitator performance are evaluated.

Research Facility

B&W's six-million Btu/hr  SBS utilized to perform the pilot-scale cyclone
coal reburning tests has been described elsewhere  (1)(2).

Baseline Test Results

The  baseline  tests  are  performed under  normal  cyclone  operating
conditions  and identify the  benchmark data to which  the  subsequent
reburning test results are compared.  The key parameters  measured during
the baseline tests included NOx emissions,  Furnace  Exit  Gas Temperature
(FEGT), unburned  carbon, CO, H2S, fly ash resistivity,  and particulate
loading/deposition.

Figure  4  illustrates  the  NOx  emission   levels  obtained  during  the
baseline tests.  Operating the cyclone at six-million Btu/hr resulted in
baseline NOx levels of  950 to 1070 ppm (corrected to 3% O2)  while varying
excess O2% from 2 to 3.75%, respectively.   Since operating at 3% excess
O2 is considered typical,  the baseline NOx level utilized to compare with
reburning conditions is 1025  ppm.  Reducing the SBS load to 75%  of  rated
capacity  (4.5 million  Btu/hr) resulted in NOx emissions of 915 to 1000
ppm  (corrected  3%  O2)  when  varying  excess   O2  from  2.4 to  4.3%,
respectively.  The  NOx emission level  while operating  at a  typical  3%
excess O2 was 950 ppm.

Baseline  FEGT's  at  full   and  75%  loads  were   2175°F   and   1975°F,
respectively,  at an  excess  O2%, of  3%.    Fly ash samples collected
throughout  the long-term deposition test phase  resulted in an  unburned
carbon  (UBC)  level of 3.5% or  an associated combustion  efficiency  of
99.99%.   During  short-term tests a  discrepancy in UBC results,  which
will be discussed later, was  observed  with UBC levels  of less than  1%
corresponding  to an  associated combustion  efficiency  of essentially
100%.  Stack CO  emission levels and measured H2S concentrations in the
lower  furnace were  low.  CO  (ppm) levels throughout the baseline  tests
were  less than 50  ppm and  no H2S  was detected.   Fly  ash resistivity
measurements were collected  at  the  simulated precipitator inlet.   The
measured  resistivity was 4.5 x  1010  ohm-cm.

Reburning Test Results

Lamar  coal  was fired in the cyclone as it was operated  at 65 to 80% of
total  load  under excess air conditions.  Reburning  coal  firing  provided
the  remaining  20 to 35% heat  input.   In  order to obtain various in-
furnace  reburning zone stoichiometries (0.85  to  0.95),  the reburning
burners were operated  at substoichiometric conditions.  The  balance of
air  was  then  introduced through two OFA  ports located  in the  upper
furnace rear wall.

Major  comparisons between the SBS baseline/reburning tests given  below
include  the following:  NOx  reduction, Furnace  Exit  Gas Temperature
(FEGT),  combustion efficiency,  fireside  corrosion,  and  precipitator
performance.
                                  3-80

-------
NOx Emissions

A 43 to  75%  NOx reduction (from the  baseline  NOx level) was achieved
during coal reburning under various test conditions.   These results are
reported as overall reductions and  consist of basically two components:

     NOx reduction via lower heat input at the cyclone burner
     NOx destruction via the reburning process

The following results are  based  upon  the overall NOx reductions while
varying reburning zone stoichiometries and burner flue gas recirculation
addition.

Reburning Zone Stoichiometries.  Incorporating coal reburning operation
during full load and 0% Flue Gas Recirculation (FGR) conditions revealed
NOx reductions on the order of 49 to 73% from baseline when varying the
reburn zone stoichiometry from 0.95 to 0.85, respectively.  Maintaining
the  cyclone  stoichiometry at  1.1 throughout  the  test  sequence  is
critical  due to the  potential  corrosion/operating concerns  of  the
cyclone.    Thus,  while  maintaining  the  cyclone stoichiometry,  the
reburning zone stoichiometry is varied by increasing  the  amount  of heat
input diverted to the reburn burners (while also  maintaining a constant
reburn burner stoichiometry).  To  obtain the  0.95 to 0.85 reburn zone
stoichiometries, the corresponding  cyclone/reburn burner coal splits are
80/20 and 66/34, respectively.  Figure 5 shows NOx emissions  (at 3%  02)
versus reburning zone stoichiometry at six million Btu/hr.  This figure
also shows  the  NOx emission  consequences  while  varying the reburning
coal fraction (standard, medium, and  fine grind  corresponds to  63, 78,
and  90%  through  200  mesh,  respectively) .   The  data  show  a  good
correlation between all tested grind  sizes versus NOx.

Reducing load to 75% of rated boiler capacity (0% FGR) and utilizing the
fine grind coal  size revealed NOx reductions on  the  order of 43 to 63%
when varying reburn zone stoichiometry from 0.95 to  0.86.

Flue Gas Recirculation.  Adding FGR to the reburning burners secondary
air  zone increases  the  mass  flow rate  through the burner  and thus
results  in  higher burner velocities  with  increased  pressure  drop and
turbulence.  Comparing the NOx reduction capability during FGR addition
with the baseline 0% FGR case reveals that NOx reductions of 53 to 75%
were achieved while varying the reburn zone stoichiometry from  0.95 to
0.85.  Thus,  a slight improvement in NOx reduction was observed when FGR
was utilized.

Furnace Exit Gas Temperature

Furnace  exit gas  temperature did not  change  significantly   between
baseline and reburning operation.  Baseline FEGTs at full load  and 75%
load (0% FGR) were 2175"F  and  1975"F,  respectively, at a stack O2 of 3%.
Incorporating reburning  at  full  load  revealed  minimal  FEGT  effects
within a range of  about ±40"F.   The addition  of FGR actually caused a
temperature  quenching/tempering  phenomenon  to   occur.    Lower  load
operation showed a potential FEGT  increase up to about 100°F.  This
increase at lower loads would not be a problem.  Thus, based upon this
data (in conjunction with past pilot  test/engineering  studies),  no
significant reburning impacts  on FEGT would be predicted.  Additional
investigations  incorporated  higher  FEGT  values within  the  boiler
performance models  for  the Nelson  Dewey No.   2  demonstration  site to
determine performance changes (superheat and reheat  attemperator spray
quantities) with higher FEGTs.

                                 3-81

-------
Combustion Efficiency

Reviewing  unburned  carbon  data  showed  that the  pilot-scale testing
predicts an increase in unburned carbon when operating under  reburning
conditions.   Reducing  the  reburn  coal  grind  size  does  reduce the
magnitude of this increase.  Thus, the full-scale reburn retrofit  design
will  incorporate the  capability  to  change the  coal  particle  size
distribution.

Since the  total  ash content to the furnace  is  low in cyclone boilers
(due  to  the  slag tapping  capability),  increases  in  unburned  carbon
levels in  the fly ash should also be  evaluated based upon changes  in
combustion efficiency.  Fly ash  samples were collected throughout the
long-term  deposition  test  phase  (48  hour  continuous runs)  for  both
baseline  and  reburning cases.   The unburned  carbon  (UBC)/combustion
efficiency results are as follows:

          Baseline  - 3.5% UBC/99.99%
                      Combustion Efficiency
          Reburning - 5.1% UBC/99.94%
                      Combustion Efficiency

Short-term baseline unburned carbon levels were  also measured and  they
were  found to  be low  (less than  1%).    Also  during  the short-term
testing, operation in the  reburn  mode at about  a 0.9 (approximately 27%
reburn fuel)  reburn zone stoichiometry,  a fly  ash carbon content  of
approximately  5-6%  was observed  when   utilizing the  fine grind  (90%
through 200 mesh) coal size.

Two items of interest become apparent.   The first is that  over a  longer
period of time, the  baseline UBC increased to about 3.5% versus the  less
than  1% reported during the short-term  tests.  The  second  item pertains
to the small  change in  combustion efficiency observed (0.05% decrease
during  reburning).    Thus,  although  the  pilot  scale   tests  have
highlighted  unburned  carbon as  a  potential issue,  minimal  impact  on
combustion efficiency should result.

Corrosion Potential

H2S measurements within the reburning zone were taken in order to  help
assess potential  corrosiveness when applying reburn technology.  While
firing Lamar coal (Indiana -  1.87% sulfur), baseline and reburning cases
showed H2S concentrations  of 0 ppm and  0-200 ppm, respectively.   During
reburning operation, H2S levels measured near the boiler side walls  were
low.  The maximum H2S levels were found  between the flames  of the  reburn
burners.  Thus, minimum H2S  contact with the boiler walls was  observed,
which is a desired effect.

Precipitator  Performance

No change in measured fly  ash resistivity  was observed between baseline
and reburning conditions.   Via this parameter,  no  loss in  precipitator
efficiency  would be  predicted.    However,  higher  particulate  mass
loadings were observed due to the injection of pulverized  coal into the
furnace.  This predicts possible  increases in precipitator  outlet grain
loading.
                                  3-82

-------
                      WP&L BASELINE TEST RESULTS

Baseline  tests  were  performed  at Nelson  Dewey  Unit  No.  2  prior to
installation  of the  coal  reburning  system in  order to  provide the
benchmark data to which subsequent reburning results will be compared.
The test  sequence included  collecting data at three load conditions—
100%,  75%,  and  50%—and  at  different  excess  air  and  flue  gas
recirculation levels.   Thus, the baseline characterization  not only
identified   normal    or    typical     conditions    for    boiler
operations/performance,  emissions characteristics,  and electrostatic
precipitator performance, but the  test matrix was structured to identify
changes in these parameters when  excess air and flue gas recirculation
rates are  varied.   This will provide  future  background data for coal
reburning operation.

NOx and Percent Loss on Ignition  (unburned carbon) Emission Levels

Figures 6 and 7  show the full load (110 MW) baseline  stack NOx emission
levels  (ppm  corrected to 3%  O2)  and  percent loss on  ignition  (LOI),
respectively, as measured by the  Acurex Testing Company versus various
excess oxygen contents as measured at  the economizer.  Figure 6 reveals
NOx levels ranging from approximately 640 ppm to  700 ppm  (corrected to
3% 02) when  economizer  outlet O2% was varied between about 2  and 4%,
respectively.  Since operating at 3%  economizer outlet O2 is considered
typical,  the normal  baseline  NOx  level is  662 ppm at 3% O2.  Figure 7
shows  percent LOI  varied  from  approximately  18%  down  to 9%  while
increasing excess O2% from 2 to  4%, respectively.

Additionally, Figures 8 and 9 show the relationship between  NOx (ppm at
3% 02) and percent  LOI  versus boiler load  (MW)  during  typical boiler
operation (3% economizer outlet O2) .   As shown in  Figure 8, varying the
load from 55 MW to  110  MW resulted in NOx  levels of approximately 550
ppm to 662 ppm (at 3% O2), respectively.  Figure 9  reveals that percent
LOI remained fairly  constant  over the load range (approximately 16 to
17% LOI).

                  REBURN SYSTEM DESIGN CONSIDERATIONS

The demonstration boiler host site at  WP&L's Nelson Dewey Unit No. 2 is
shown in  Figure  10  and pertinent boiler information is summarized in
Table 1.

The reburning system design considerations included utilizing physical
and numerical modeling  activities along with  B&W low NOx burner/over-
fire air  port  design experience.  The size,  number, and location of
reburn burners and OFA ports were  determined.  The goal—to  obtain good
mixing at the reburn  burner  elevation and  OFA ports—is essential for
NOx reduction  and combustible  burn-out,  respectively.    In addition,
penetration of the reburn burners  fuel streams into the cyclone hot flue
gas is  of concern  since  over-penetration  or  under-penetration would
cause tube wastage  in the  boiler, along  with  potential  burner flame
instability problems.

Simultaneous modeling of the  cyclone, reburn  burners  and  OFA ports
within one  system is a  new and  unique procedure.  Development of a
modeling  methodology to  assess  mixing  and  penetration  results was
required.  The  following plan was developed to  meet the above  stated
goals:
                                  3-83

-------
          Develop a procedure to simulate cyclone boiler flue gas flow
          in cold flow and numerical models.  Compare (validate) these
          results with actual  baseline flow measurements  obtained at
          Nelson Dewey.

          Utilize validated  cold  flow model to  simulate  the  reburn
          system  conditions   using  fundamental   laws  of  aerodynamic
          similarity.

          Utilize validated  numerical  model to  simulate  the  reburn
          system conditions using B&W's FORCE and CYCLONE model computer
          codes.

Physical Flow Modeling.   Table 2 shows  a summary of the test matrix for
the physical  flow  modeling of WP&L's  baseline  and  mixing (reburning)
tests.  In  addition,  it  shows the actual field test flow measurement
conditions  which  were performed at  Nelson  Dewey during  the baseline
field tests.

For  the Nelson  Dewey field test  baseline condition,  flow  profile
measurements were obtained during cold  (C) air flow conditions near the
proposed reburn burner elevation of  666 ft.  Also,  hot  flow (50% oil
firing) velocity profiles were obtained at the  cyclone exit and at the
above stated  666 ft.  elevation.  Baseline testing in the l/12th scale
model was performed using cold air and measurements  were  obtained at
three elevations:

     1)   666 ft (near the proposed reburn burner elevation);
     2)   681 ft (within the reburn zone just prior to the OFA ports);
          and
     3)   700 ft (furnace exit).

Comparing the field and  the l/12th model measurements  from elevation
666ft.showed good gualitative agreement between the collected data.This
agreement showed that  high gas flow was concentrated at the boiler rear
wall  with  some  negative  (recirculation)/low turbulent flow  near the
front (target) wall.

The mixing (reburning) tests were performed in the  l/12th scale model by
utilizing  a  different temperature air  stream  at  the various  inlet
locations  and  then  evaluating  mixing  potential  by  measuring  the
resultant  temperature downstream.    For example,  to  evaluate  reburn
burner  mixing effectiveness,  ambient  air  was  introduced  through the
cyclones and  heated (H)  air  introduced through  the  reburn burners and
the resulting mixing temperature measured at the OFA inlet elevation of
681  ft.    Figure  11  shows  a  comparison  between  various  operating
conditions for this particular case.  The example compares:  four reburn
burners; 25%  reburn fuel  split and 5%  flue  gas  recirculation (FGR) to
the burners  at  full load versus  four  reburn burners;  30%  reburn fuel
split;  and  7.5% FGR.   The resulting  normalized  temperature profiles
(where 1.0  is equal to an ideal mix)  show that  improved mixing is seen
with the latter set of conditions.   Thus,  flexibility exists with the
reburn  system-mixing  potential since  the  latter case  conditions are
being incorporated into the Nelson Dewey design.

Numerical Flow Modeling.    Based upon an  initial disparity between the
numerical   and   physical  flow  models,  reburn   burner  penetration
capabilities, the numerical  model  was  used  to  actually  model the
                                 3-84

-------
physical  flow  model.    Since  the physical  model set-up  criteria was
conservative, the numerical model needed to predict those physical model
results prior to having confidence in the numerical model.  A detailed
1/12th scale numerical flow prediction was completed for the four reburn
burners,  25%  reburn fuel split,  and 5%  FGR cases.   The predictions
showed good agreement with observed physical flow model smoke patterns
and measured mixing performance at the  681  ft elevation.   Thus, these
predictions  benchmarked  the   numerical  model   for scale-up  of  the
reburning system design.

Table 3 shows a portion of the numerical  flow modeling test matrix which
has been  performed.   The conditions  which were  varied  included load
(MW) , fuel split, FGR rate, reburn burner  parameters (number, size, and
side spacing) and OFA port parameters  (number, size, and side spacing).
Summarizing all  the various cases which were reviewed revealed that case
"4a"  (4  reburn  burners, 5% FGR,  and 25% reburn  fuel  split)  provided
comparable results to  the predicted SBS  mixing  test  conditions which
showed approximately 80% mass flow with stoichiometric ratio (SR) less
than  1.0,  (which was  our  target  level)   within  the   reburn  zone.
Obtaining  the  percent  mass  flow  with  SR  less  than  1.0,  which  was
predicted  during the  pilot  SBS  test  phase,  provided NOx  reduction
prediction confidence for  the Nelson Dewey retrofit.  In addition, since
none of the cases using  three-reburn burners  approached this level, the
decision to proceed with the four-reburn burner design was made.

Figure  12  compares  the results of  the  three- versus  the four-reburn
burner cases with and without FGR.  The figure shows  mean stoichiometric
ratio versus furnace elevation  as well as percent mass flow with SR less
than  1.0  versus furnace  elevation.   Minimum differences between the
three- and four-burner  cases are observed  in  the no-FGR conditions, but
the percent mass  flow with SR  less than 1.0  does not  approach the 80%
target level.  Utilizing FGR distinguishes that four burners should be
used to provide maximum flexibility such that the 80%  SR  less than 1.0
could be obtained.

                               SUMMARY

The conclusions/recommendations  of the  physical and  numerical  model
activities are as follows:

          Qualitative agreement between  physical flow and numerical flow
          results.

               Baseline configuration
               Reburning configuration

          Numerical model can  be used  for  qualitative evaluation and
          scale-up of reburning system.

          Four  reburners  and  OFA  ports   provide the  best  mixing
          performance.

          Include the capability to add  5 to 10%  flue gas  recirculation
          to the reburn burners.

          Maintain the 25 to 30% fuel split  capability-

Detail Design Considerations.  Utilizing  the conclusions/recommendations
from  the  physical/numerical  modeling along  with  B&W's low NOx system


                                  3-85

-------
design experience, the following  reburn system design was determined.
An isometric view of the overall  system design is shown in Figure 13.
The reburn system includes the following list of major components:

          Four B&W S-type coal firing burners.

          Four B&W dual air zone over-fire air  (OFA) ports.

          Furnace wall panels to accommodate  the burners  and OFA ports.

          One B&W MPS 67 pulverizer with associated components including
          coal piping.

          One hot primary air fan and motor.

          One gravimetric feeder.

          One coal silo, 150 ton capacity.

          Numerous flues and ducts to transport air/flue  gas to various
          system components.

          A  new  enclosure  to house  the  pulverizer and its associated
          components.

          A new motor control center  and transformer to power the reburn
          system.

          Numerous dampers and  drives to control flows to the various
          system components.

          One  seal  air  fan  and motor to  provide seal  air  to the
          pulverizer/feeder/hot PA fan.

          New reburn  system  microprocessor control system.

 Figure 13 shows a general overview  of  the reburning system and  how  it
 compares to  the existing  boiler arrangement.    The  pulverizer  (and
 associated  equipment)  will  be  located in  a  new building  enclosure
 between column rows "12"  (existing  building)  to "14" and "C" to "G" -
 The hot primary air (PA) is taken off  the left  side of  the air  heater
 and ducted to the PA fan inlet.   Tempering  air  is fed to the PA prior to
 the PA fan inlet in order to control pulverizer air inlet temperatures.
 Automatic dampers  will be  available  in each of  these ducts.    In
 addition, an isolation damper (automatic) will be located just prior to
 the PA fan inlet to  allow maintenance on the  fan/pulverizer when the
 boiler is operating.  An air monitor will also be located just prior to
 the PA fan  inlet to  measure  total air flow to  the pulverizer.

 Secondary air to the reburning burners will also be supplied from an air
 heater outlet takeoff point located at the  center bottom  point  of the
 air heater.   An  automatic damper and air monitor will be located within
 this line in order to control and measure total secondary air  flow to
 the burners.   Gas recirculation can  be introduced into this  system (if
 necessary)  such that the  total mass  flow through the burners  can  be
 varied.    The  gas  recirculation  (GR) takeoff  is  located  after the
 existing system's GR fans and is tied into the secondary air duct prior
 to the burner splits.  An automatic  damper (tight shut-off) and monitor
 are available in this  flue to  control and  measure flow.   Finally, this


                                  3-86

-------
air/gas  to the  burner  subsystem contains  four  manually adjustable
dampers in each of the lines leading to the individual burners.  These
dampers will be utilized during system commissioning to balance  flows to
each burner in case an imbalance exists.

The OFA  system obtains  its  feed from  the existing  boiler's  hot air
recirculation system.  The hot  air recirculation  system is  available to
take air from the air heater outlet to the FD fan discharge (basically
an air preheat system originally designed to help protect  against cold
end air heater corrosion) .  The OFA takeoff is prior to a booster fan in
this system.  The duct work which leads to the four OFA ports includes
an automatic damper/air monitor to control and measure total air flow to
the OFA system.

The basis  of  the reburning  technology  is  the  range of in-furnace
operating  stoichiometries along  with reaction  times.    In order  to
accurately control the  process, additions to the existing control system
have to be made in order  to control the  fuel and  air splits between the
cyclones, reburning burners, and OFA ports.  The existing control system
at  Nelson  Dewey  is  the  Bailey  Network  90™,  a  state-of-the-art
microprocessor system.  Additions to the microprocessor are  possible due
to the existing system's flexibility.


                        FUTURE WORK/CONCLUSION

The focus of this demonstration project will be to determine maximum NOx
reduction capabilities without adversely  impacting plant  performance,
operation, and maintenance.  In  particular,-  the  prototype evaluations
will confirm and expand the results of the pilot-cyclone test programs.
Both  steady-state and  transient operation  will  be  evaluated.    The
following summarizes the specific items to be evaluated:

     1)   Major reburn process parameters on NOx reduction capability;
     2)   Combustion  efficiency  (based  on  unburned carbon  and  CO
          emissions);
     3)   Boiler thermal efficiency;
     4)   Furnace temperature and heat absorption profiles;
     5)   Slaging and fouling;
     6)   Corrosion potential;
     7)   Gaseous and particulate emissions; and
     8)   Electrostatic precipitator operation.

In addition to completing the detail design of the reburn system, future
work includes  investigating local heat transfer with boiler performance
models to  help predict  changes   in  furnace heat  flux distributions.
Unburned carbon models  will also  be utilized to help predict changes in
UBC levels during reburning.

Another initial concern of the  coal reburning  is  the amount and control
of  the  additional  ash  loading  to  the  boiler  and  electrostatic
precipitator.    Based upon a  preliminary study  by B&W and EPRI, the
existing precipitator  has  sufficient margin  to  adequately control the
increased fly ash added by the coal  reburn process.   In addition, the
ash  mean  particle  size   is  expected  to  increase  and  aid  in the
precipitator collection efficiency.

Finally,  to investigate fireside corrosion—a potential side effect in
retrofit low NOx technologies—tube  sections in the reburn  zone will be


                                  3-87

-------
replaced with new  tubes  during  the  installation  phase.   The  new  tube
samples will  be  taken  out after the long-term  performance  tests,  and
analyzed for potential  corrosion  effects.   Ultrasonic measurements  will
also  be  taken before  and  after the  long-term performance  tests  to
evaluate furnace  tube wastage.  In addition, in-furnace  H2S measurements
will be taken and corresponding corrosion rates will be  predicted.   This
information will  be used to  investigate  potential problems  and provide
recommendations for preventing such  phenomena if necessary.

In conclusion,  the modifications  discussed in this paper constitute the
retrofit of a  feasible  coal reburning system to the  selected host unit.
Thorough testing  of this  system along with obtaining information on the
boiler's  baseline  operating  performance  will  provide   a  complete
evaluation  of  the  usefulness  of  coal  reburning  as  a NOx  reduction
technology  for cyclone-fired  boilers.     All  the   work  to  date   has
substantiated that the  goals  of  this project  are attainable.

                              ACKNOWLEDGEMENTS

The authors extend  their appreciation to  the following B&W personnel  for
their   help   in    the  performance  of    the   SBS   testing   and    the
numerical/physical flow modeling activities:   Hamid  Sarv,  Rick Wessel,
Vince  Belovich, Ray Kim,  and  George Watson.

                                  REFERENCES

1.    Maringo,  et al.,  "Feasibility of  Reburning  for  Cyclone  Boiler  NOx
      Control", 1987  EPA/EPRI Joint Symposium  on Stationary  Combustion
      NOx Control,  New Orleans,  Louisiana,  March  23-27,  1987.

2.    Farzan,  et al.,  "Pilot  Evaluation of Reburning  for Cyclone Boiler
      NOx  Control",   1989   EPA/EPRI   Joint   Symposium   on   Stationary
      Combustion NOx Control,  San Francisco, California,  March 6-9, 1989.
               Legal Notice:

               The Babcock  &  Wilcox Company  pursuant  to a cooperative  agreement
               partially  funded by the U.S. Department  of Energy (DOE)  and a grant
               agreement  with IDENR for the DOE and IDENR and neither the  Babcock &
               Wilcox  Company,  DOE, IDENR,  nor  Southern  Illinois  University  at
               Carbondale, nor any person acting on their behalf:

               a.  Makes any  warranty or representation,  express or implied, with
                  respect to  the accuracy,  completeness,  or  usefulness of  the
                  information contained  in this report, or  that  the use of  any
                  information, apparatus, method, or process disclosed in this report
                  may not infringe privately-owned rights; or

               b.  Assumes any liabilities with respect  to the use of, or for damages
                  resulting from the use of,  any information, apparatus,  method or
                  process disclosed in this report.

               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 U.S. Department  of Energy.  The views and opinions of
               authors expressed  herein  do not  necessarily state or reflect those of
               the U.S. Department of Energy.
                                      3-E

-------
                         Table 1
                     Boiler Information
                   Nelson Dewey, Unit 2
Name Plate Rate:
Type:
Primary Fuel:
Operation Date:
Boiler ID:
Boiler Capacity:
Boiler General Condition:
Boiler Manufacturer:
Boiler Type:
Reburning Demonstration
 Fuel:
Burners:

Particulate Control:
Boiler Availability:
100 MWe
Steam Turbine
Bituminous and Sub-Bituminous Coal
October 1962 - Unit No. 2
B&W RB-369
Nominal 11 MWe
Good
Babcock & Wilcox
Cyclone Fired RB Boiler, Pressurized
Indiana (Lamar) Bituminous Coal,
Medium Sulfur (1.87%)
Three B&W Vortex-Type Burners,
 Single-Wall Fired
Research Cottrell ESP
90% Availability
Table 2
Matrix of WP&L Baseline and Mixing Tests
Test
Type
Baseline
Test
Mixing Test
Test
Facility
WP&L Boiler
1/12 Model
1/12 Model
Gas Temperature
Cycl
C
H
C
C
C
Rbm


C
H
C
OFA


C

H
Measurement Plane
Cycl Exit

X



666
X
X
X
X

681


X


700


X

X
                   700' —
                   681' —
                   666'-
               Cyclones
                           Burnout
                             Zone
            I OFA
                            Reburn
                           \ Zone
                                      Reburner
                          3-89

-------
Case
No.
1
2
3
3a
4
4a
4b
5
5a
Load
100
100
110
110
110
110
110
110
110
MWe
MWe
MWe
MWe
MWe
MWe
MWe
MWe
MWe
Numerical Fl<
Fuel Reburne
Split FGR
30%
30%
25%
25%
25%
25%
25%
25%
25%
0%
0%
0%
8.7%
0%
5%
8.7%
0%
8.7%
Table 3
>w Modeling Test Matrix
r Reburneis
No.
3
4
3
3
4
4
4
3
3
Size
20"
18"
20"
20"
18"
18"
18"
20"
20"
Spacing
117.5"
7'9"
10'
10'
6'8"
6'8"
6'8"
10'
10'
No.
3
4
4
4
4
4
4
3
3
OFA Ports
Size
28"
24"
22"
22"
22"
22"
22"
26"
26"
Spacing
117.5"
7'9"
6'8"
6'8"
6'8"
6'8"
6'8"
10'
10'
  In furnace
  0.85 - 0.95
Stoichiometry


  Coal
 Bunker
 Cyclone
 Furnace
70%-80% Heat Input
   Crushed Coal  I   v^
 1.1 Stoichiometry |      JN
 Main Combustion        O
       Zone        B&W Boiler
                    RB-369
                100 MW capacity
        Sec.
        Air
       Heater
                                    Primary
                                   Air Heater
 Overfire Air Ports
 Balance of Air
 1.15-1.20 Overall
 Stoichiometry

^ Reburning Pulverized
x Coal Burners
 20% - 30% Heat Input
 Pulverized Coal
 0.4 - 0.5 Stoichiometry
                                                             Precipitator
                           J
                                                    Stack
Flyash Handling
 and Disposal
                     Figure l--Coal reburn project - system layout, cyclone firing.
                                              3-90

-------
Phase  I:    Design & Permitting
           • Modeling &  Pilot  Scale
             Testing
           • Baseline Characterization
           • System Design
Phase  HA:  Equipment Procurement
        B:  Construction & Start-Up
           • Fabrication
           • Installation
Phase  III:  Operation & Disposition
           • Optimization
           • Long-Term Performance
           • Reporting & Disposition
                                        1989
     1990
12 Mo.
10 Mo.
      H
       9 Mo.
         5  Mo
        2.5  Mo.
1991
                    7  Mo.
                          -i
   5 Mo.
1992
                                     9 Mo.
                                       11  Mo.
993
              Figure 2--Coal reburning for cyclone boiler NOX control project schedule.
                  STACK
                                 STEAM
                                                      SUPERHEATER
                                                      FOULING TUBE
                                                      DEPOSITION PROBE
                                                                    FURNACE ARCH
                                                                   PRIMARY AIR
                                                                   AND COAL
                                                                        TERTIARY AIR
                                                                         SECONDARY AIR
             FLUE GAS
             RECIRCULATION
                                            SLAG TAP
                                                               MOLTEN SLAG
                                                                  SLAG COLLECTOR
                                                                  AND FURNACE
                                                                  WATER SEAL
                              Figure 3--Small boiler simulator (SBS).
                                              3-91

-------
                1100


                1000-


                 900-
       NOX
Corrected to 3% 02,
      (ppm)
                 800-
                 700-
                 600-
                  500-
                                                 D Lamar Coal @ 6 x 106 Btu/hr.
                                                 + Lamar Coal @ 4.5 x 106 Btu/hr.
                                           Excess Oxygen, (%)
                      Figure 4--SBS baseline NO* emission levels.
                 1000-
                  800-
      Stack NOX
      corrected to
        3% 02     600 -|
        (ppm)
                  400-
                  200
    Legend
    Baseline conditions,
    no reburning
Reburning Conditions
  A Standard grind coal
  • Medium grind coal
  * Fine grind coal
                       Boiler Conditions
                       • Cyclone @ 10% excess air
                       • 0% Flue gas recirculation
                       • 6 x 106 Btu/hr load
                                    0.9             1.0             1.1
                                      Reburning Zone Stoichiometry
                                                   1.2
                    Figure 5--SBS NOX emissions with coal reburning.
                                        3-92

-------
             800
             700-
 Stack NOX,
Corrected to
  3% 02l
  (ppm)
              600-
              500
                                             Boiler Load - 110 MW (Full Load)
                                             Test Procedure - EPA Method 7E
                                234

                                 Economizer Outlet Excess 02, (%)
     Figure 6--Baseline NOX emission levels vs. excess 02. Nelson Dewey  Unit 2.
        19.0
        17.0
     o 15.0
        13.0
        11.0
         9.0
           0.0
                    Boiler Load - 110 MW (Full Load)
1.0          2.0          3.0
           Economizer Outlet
              Excess 02%
4.0
5.00
   Figure 7--Baseline %LOI emission levels vs. excess 02 %. Nelson Dewey - Unit 2.
                                     3-93

-------
         700 i
          600-


 Stack NOX
corrected to
  3% 02
  (ppm)   500 -
          400
             30
50
            Excess 02% at
            Economizer Outlet - 3% 02 (Typical)
            Test Procedure - EPA Method 7E
70          90
  Load (MW's)
                                                           110
                                              130
     Figure 8--Baseline NOX emission levels vs. load. Nelson Dewey - Unit 2.
    20.0
     18.0
 _op

 8
     12.0
    10.0
         50      60       70       80       90      100      110       120

                                   Load (MW)


    Figure 9--Baseline %LOI emission levels vs. load. Nelson Dewey - Unit 2.
                                   3-94

-------
 Figure 10-Wisconsin Power & Light Company, Nelson Dewey - Unit 2 (RB-369): cyclone reburn
 project.
                                                                Rear Wall
    700'-
    681'-
 Cold,—
 Air—*
     1_

Cyclone
                 To Fan
           Burnout
            Zone
IOFA
            Reburn
           .  Zone
                 4 Reburners
                25% Reburner
                 5% FGR Flow
         Momentum Ratio Base

               T Tc
                                 TR-T,
         Front Wai I
WP&L Mixing Test 31 - Elev. 681'

          Rear Wall
                     Heated
  Reburner
                 4 Reburners
               30% Reburner
               7.5% FGR Flow
         Momentum Ratio Base
                                                                Front Wall
                                                       WP&L Mixing Test 33 - Elev. 681'
                          Figure ll--Examples of mixing test results.
                                            3-95

-------
       700
        690
    o   680
        670
        660
                     With FGR
                                                 700
                                                 690
   680


§
OJ
   670
                                                 660
             3 Rbnrs, 3 OFA Ports
                                                           4 Rbnrs, 4 OFA Ports
              0.9     1      1.1    1.2


             Mean Stoichiometric Ratio
      0    20  40   60   80  100



        % Mass Flow with SR <1
Figure 12--Effect of number of reburners and OFA ports on reburning system mixing performance.





                                        3-96

-------
          Arrangement of Equipment for the Cyclone Reburn
          Project  at  Wisconsin  Power  & Light Company's
                  Nelson Dewey Station, Unit No. 2
                    Partial Section
                     of Air Heater
                                         Overfire Air Port
Figure 13--Equipment arrangement for the coal reburn project.
                       3-97

-------
DEMONSTRATION OF LOW NOX COMBUSTION TECHNIQUES AT
  THE  COAL/GAS-FIRED MAAS  POWER  STATION  UNIT  5

                J.  Van  der Kooij
     Sep,  Dutch Elecricity Generating  Board
  Ulrechtseweg 310, 6812 AR Arnhem, Netherlands

                    H.K. Hwee
                    A.  Spaans
                  Stork Boilers
 Industriestraat 1, 7553 CK Hengelo, Netherlands

                    J.J. Puts
                    N.V. Epz
  Begijnenhof 1,  5611 EK Eindhoven, Netherlands

                   J.G.  Witkamp
                    N.V. Kema
  Utrechtseweg 310, 6812 AR Arnhem, Netherlands

-------
         DEMONSTRATION OF LOW NOx COMBUSTION TECHNIQUES
        AT THE COAL/GAS-FIRED MAAS POWER STATION UNIT  5
                        J.  van der Kooij
            Sep, Dutch Electricity Generating Board
         Utrechtseweg 310, 6812 AR Arnhem, Netherlands

                           H.K. Kwee
                           A. Spaans
                         Stork Boilers
        Industriestraat 1, 7553 CK Hengelo, Netherlands

                           J.J. Puts
                            N.V.  EPZ
         Begijnenhof 1, 5611 EK Eindhoven, Netherlands

                          J.G.  Witkamp
                           N.V. KEMA
         Utrechtseweg 310, 6812 AR Arnhem, Netherlands
ABSTRACT

Unit  5  of the  Maas Power  Station is  a  coal/gas-fired  boiler
with  horizontally  opposed firing burners.  In  this boiler  HTNR
low-NOx  burners  and after  air  ports  have been  installed  to
demonstrate the viability of  low-NOx combustion techniques.  The
aim was  to prove  that  in new installations NOX  concentrations
of  less  than  400 mg/mj  are  feasible  for  a large  variety  of
coals,  as  well  as  to  determine  the impact  of  MO^   control
technology on  boiler  operation, performance  and  maintenance.
Prior  to  the  retrofit  of   the  boiler  the  HTNR  burner  was
modified to accomodate both coal and gas-firing.
Results are presented on:

*     NOX emission  with  natural gas  firing
*     a parameter  research on NOX  emission and  burnout:  this
      programme  included burner setting,  stoichiometry at  the
      burners,  boiler load and excess air for three coal types;
*     corrosion tests with reference materials;
*     slagging tests with special probes;
*     fly ash quality.
                          3-101

-------
INTRODUCTION

Since 1974  diversification of the  fuel  package for  generation
of electricity has been one of the  main objectives  in  the  Dutch
energy policy. The  reasons for preventing excessive  dependence
on  a  few  fuel   types  are  to   maintain  the  reliability  of
electricity supply and to  stabilize electricity  tariffs. As the
decision on  extending the use of  nuclear energy was  postponed
and  natural  gas was  reserved  for  high-quality  applications
only, a temporary switch  to fuel oil  was made in 1978.  For the
longer term it was decided to make  increasing  use of  coal,  thus
reducing the  role  of  fuel  oil  to  a  minimum.  This process  was
carried  out  in   various  phases.   First,  the  existing  power
stations,   which  used  to  be  coal-fired,  were  once again  made
suitable  for  coal-firing.  Subsequently,  two  new  coal-fired
power  stations were  built and  four   existing gas-fired  power
stations were converted  into coal-fired  ones.  In  1988 a  coal
capacity of 3900  MW thus became available.

Approximately  40%  of  electricity  production  is now based  on
coal-firing.  This  percentage will  be  maintained  during  the
nineties.   Old power  stations will be closed  down,  while  new
coal-fired  power stations  will  be commissioned.  Three  600 MW
coal-fired  units  will   be   built  at Amer,  Hemweg   and  the
Maasvlakte. They are  scheduled  for completion by the  middle of
the years 1993, 1994 and  1997 respectively.

As  can be  seen  from Table  1,   the   standards  for  coal-fired
stations are  tightened gradually.  An   important  development was
the  formulation of national standards  in  the Decree on Emission
Standards   for  Large  Combustion   Installations  (1987).   It
specifies  maximum  emission  concentrations  for  SO,,  NOX  and
particulates   for   coal-,  gas-,  and  oil-fired   plants.   The
combustion  installations  in  question  range  from steam  boilers
to process furnaces, stationary engines and gas  turbines.

For  coal-fired  plants  for  which  a  license  was   or  would  be
granted  after August  1,  1988,   the  national  NO.  standard  is
400 mg/mj   (STP,   dry   at  6%   02 ) .   For  installations   with
horizontally  opposed  and  front-wall   firing  the experience  in
the   Netherlands  with  modern  combustion  modifications  was
considered  to be  insufficient  in relation   to  the  standard.
Therefore  the decision was taken  to  carry out  a  demonstration
project at the Maas Power Station unit 5.

The  project is performed  in  the  framework of the Concerted NO -
Abatement   Programme   of   the   Dutch  electricity   generating
companies.   In the programme  low-NOx  techniques  are applied  in
new  and  existing  installations,   while  new  technologies  are
demonstrated   to   assess   their  applicability.    Full-scale
demonstration  is regarded  as an essential  step in  confirming
the  viability of the  low-NO   technology.  The main  thrust of the
demonstration  is  not  only to quantify potential NO  reductions
in  actual  boilers,  but also  to  determine the exact  nature and
the  impact of  NOX controls on boiler  operation,  performance and
maintenance.
                           3-102

-------
The demonstration  of  advanced low-NCL combustion techniques  at
the  coal/gas-fired  Maas  Power  Station  unit  5  is   a   joint
activity of many organizations.
EPZ, the electricity generating company  in  the Southern part  of
the Netherlands placed the power station at  the disposal  of  the
demonstration  project.  They  commissioned   Stork   Boilers   to
convert the  power  station  and operated  it  in the low-NOx  mode.
Sep, the Dutch  Electricity  Generating Board  funded  the project
and was  responsible for coordination  of the  activities  within
the  Concerted  NOX  Abatement  Programme.  The  Dutch Government
expressed  its  interest in  the project  and participated  in  the
project  via  NOVEM,   the  Dutch  Association  for   Energy  and
Environment,  in the  framework of  the  National  Coal   Research
Programme.   Also,   the Commission  of  the  European Communites
subsidized  the  project  in   the  framework  of  their  Energy
Demonstration Programme: Gasification,  Liquefaction and  use  of
solid  fuels.   The  total   amount   of  national  and   European
subsidies  add up  to approximately 10  million Dfl,  i.e.  45%  of
the project  costs.  After conversion  of  the power  station  EPZ,
Sep,  Stork  Boilers  and NOVEM  jointly  performed  a   research
programme,   together with KEMA,  the   research institute of  the
Dutch  utilities.   REMA  participated  as  a  consultant  in  the
project  and  was   responsible  for   a   number  of   special
measurements and the scientific survey.

OBJECTIVES OF THE DEMONSTRATION PROJECT

The aim  of  the project is  to  prove that  for new coal-fired
power  plants  a  NO^ emission level  of  400 mg/rr^ can be  attained
without  adverse side  effects.  In  the  retrofit situation,  the
goal  of  400 mg/rr^  is considered to  be  reached for given  coal
types when NOX emission is  below 470  mg/rn^ and the carbon  in  ash
content  is  below  2.6%.  The  relaxation  of  the  standard  by
70 rng/m^  compensates for  the thermal   load and the dimensions  of
the furnace  in  the retrofit situation in comparison with  a  new
coal-fired boiler.
This  goal  was  verified  during  guarantee  measurements.   For
Drayton  coal  430 mg  N0x/m3  was  measured  at  6%  Oj  and  1.4%
unburnt carbon in ash.

The scope of the demonstration programme was  wider.
In order to  investigate  the  impact of N0x-control technology  on
boiler  operation,  performance  and  maintenance  the  operating
experience was evaluated and several  possible side  effects were
studied in a research programme.
With  respect to   the  operating performance  of  the  unit  the
emphasis was on:
*     safe and reliable operation of  the boiler
*     rapid response to load changes
*     continuation of unit efficiency.
In the course of  the  research programme the  impact  of  the coal
quality was  studied.  Three coal types  were  tested. These coal
types  are   different  with   respect   to   combustibility,  NO
formation  and  burnout, but  are  well within the range of coal
types that  are fired at the Maas power station. For each  coal
                          3-103

-------
emission  measurements  were  performed  and the  fly ash  quality
and slagging tendency was investigated.
Moreover,  a research  project was  carried  out  to  investigate
whether  the combustion  modifications  give  rise  to  increased
fireside corrosion.

DESIGN OF THE LOW-NOx COMBUSTION SYSTEM

To  achieve  the  NOX  and   UBC   performance   specified  in  the
objective  of  the demonstration project,  low-NOx burners  (HTNR)
were  applied  in  combination with  two-stage  combustion.  The
boiler  and the  combustion  system  are  shown  scematically  in
Figure 1.

Low-N0x Burners (HTNR)

The original  HTNR burner was developed  by Babcock Hitachi  for
firing pulverized coal with fuel oil as a supporting fuel.
For the  specific conditions  in  the  Netherlands,  where  natural
gas is  used as a  secondary  fuel  up  to full load,  modification
of the original HTNR burners was necessary.
The existing mill  system at  the Maas power station, unit  5 with
a  relative  high  coal/air ratio in excess  of 0.7 also led to  a
minor  modification  of  the  internals  of  the   original  HTNR
burners which were designed for a coal/air ratio of 0.5.

The above-mentioned modification implied the following:
*     implementation of  a  number of  gas  spuds  surrounding  the
      pulverized coal nozzle;
*     implementation  of  a  gas  igniter  at  the centre  of  the
      burner;
*     modification of  the  design  of  the flame stabilizing ring
      for a coal/air ratio of 0.7.
The modified  HTNR burner for coal-  and  gas-firing is shown  in
Figure 2.

Low-NO,, Combustion System
      A             	

To enhance  NOX  reduction,  a  two-stage combustion technique  is
applied.  One  row  of  after air  ports  (AAP)  was added at  both
sides  of  the boiler.  In  order to optimize  the mixing  between
the after air  and the combustion  gases  of  the  main  combustion
zone, both  the momentum and  the swirl  of  the after air could  be
adjusted  (Figure 3).

Test Programme

Prior to  the retrofit of the  boiler, the  burner  and the  lay out
of the furnace were tested by two trials:
*     combustion   test  with  a  scaled   down  version  of  the
      modified HTNR  burner  (1:7)  in  the 4 MWth  test  furnace  of
      Babcock Hitachi;
*     water  flow  simulation  of  the  gas/air  mixing  in  the
      furnace. Scale based on geometry 1:30.
                           3-104

-------
Combustion  Test   The  combustion test  was carried  out  in  two
steps:
*     coal/gas combustion test with the following objectives:
      -  assessment of flame stability and NOX performance;
         verification of burner design;
         assessment  of  the  ignition  behaviour of  the  gas  and
         coal flame;
*     confirmation  of   the   coal  combustion  test  with   the
      following objectives:
         evaluation of NO.  and UBC performance;
      -  prediction  of  NOX  and UBC performance  for Maas power
         station, unit 5.

During  the  combustion  tests Blair Athol  and Wambo coal  were
used and a  parameter study  was carried  out to  evaluate   the
influence  of  burner setting,  stoichiometric  ratio,  primary/
secondary/tertiary air ratios and position of  the gas spuds.

Some data  from the  combustion  test are presented  in Figure  4
for Wambo coal.
From these  results  a 15% higher NOX emission  was estimated  for
the  modified  burner  in  comparison   with  the   original   HTNR
burner. The outlook for UBC performance was unchanged.
It  is   not   known  whether  this  phenomenon   is  caused  by   the
increased space  between the secondary  and tertiary air and/or
by the higher coal/air ratio of the modified burner.

Water Flow Simulation  One of  the conditions  for achieving  good
combustion  performance  for  two-stage  combustion  is  intimate
mixing  in  the  furnace  of  the  AAP  combustion  air and   the
combustion gas coming  from the burner zones.  Special attention
is paid  to  the design of  the  AAP' s due to the limited  furnace
height available  and the  staggered arrangement  of   the  burners
of Maas Power Station, unit  5.

The design  of  the AAPs was  verified by  means of  a  test  with  a
three-dimensional water flow simulation (Figure 5).
The  flow  analysis   is  performed  using  the  Image Processing
Technique. The flow  is visualized using polystyrene  as a tracer
in  combination  with  the   tomography  technique   (slit  light
illumination).
Figure 6 presents a typical result.

MODIFICATION OF THE BOILER

Maas Power  Station  unit 5 is  a  coal-  and gas-fired unit rated
at  177 MWe  and  it  was  commissioned  mid-1966.   The  boiler
manufactured by  Stork  Boilers is  a  Benson  type  with  divided
second  pass  for  steam   temperature  control.  The technical
specification of the  boiler  and the combustion system is given
in Table 2.
In  the summer  of  1988  the  boiler  was modified  to a  low-NOx
combustion system.
                           3-105

-------
The modification involved the following main elements:

*     Replacement  of  the  16  single  register  burners  by  16
      advanced low-NC^  HTNR burners.
*     Replacement  of  the burner  throats  while maintaining  the
      same  staggered arrangement  of  the  horizontally  opposed
      burners as before the retrofit.
*     For  two-stage combustion  8 after  air  ports  (AAP)  were
      placed above the upper rows of the burners.
*     Modification  of  the  combustion  air  ducts  in  order  to
      accomodate for two-stage combustion.
*     Upgrading  of the  control  system  of  the  boiler and  the
      flame monitoring system.
*     Extension of the flue gas analysing equipment.

As mentioned previously,  the HTNR burner  shown schematically in
Figure  2  is a modified HTNR burner,  suitable for coal and  gas
firing.   Due  to   the  limited   available   furnace   height   and
limitation  of space surrounding  the  furnace a  special  after  air
port as shown schematically in Figure 3 was  designed.
The  total  modification  took  less  than  three months  and  the
boiler  with  advanced  low-NOx  combustion  was  commissioned  in
August 1988.

DEMONSTRATION PROGRAMME

To  establish the  results  of  the combustion  modifications,  a
comprehensive measuring programme was set up.
The  measuring programme  consisted  of  a  pre-retrofit  baseline
test and  a  post-retrofit  test series.
A brief summary of the test series is given  below.

Pre-Retrofit Baseline Tests

The  objective  of  these  baseline tests  was  to  establish  the
basis  for  evaluation  of  the  boiler  and  emission  performance
before and  after the modification of the combustion  system.

Data    for    coal    and    gas    firing   were    collected   for
characterization of:
*     NO. emission  and  unburned  carbon  (UBC)  as a  function of
      boiler  load  and excess air;
*     fly ash quality and slagging behaviour;
*     boiler  efficiency.
These tests were carried  out in March and May  1988.

Post-Retrofit Tests

The  post-retrofit  tests  that followed the  commissioning  of the
boiler  with  the  new  low-NOx  combustion   system  involved  an
extensive and systematic  variation of  the boiler and combustion
system operating parameters.
The  post-retrofit  tests  consisted of  two  test series,  i.e.  test
series for  gas  firing and coal firing  respectively.
                           3-106

-------
Gas Firing  The post-retrofit tests for gas firing were  carried
out for two weeks in November/December 1988.
Similar to the baseline test series, data were compiled  for  the
characterization of NOX and CO as a function of boiler load  and
excess air.
To  identify  the capability  of  the combustion  system,   several
combustion operating modes were tested:
*     conventional combustion. In this case the windbox  dampers
      of the after  air ports were  closed. An appropriate flow
      of  purging air  is  necessary to protect  the  after  air
      ports against excessive temperatures;
*     two-stage  combustion mode with  variation of  the  burner
      stoichiometric ratio.

Coal Firing   A comprehensive post-retrofit  test  programme  for
coal firing  was set  up to  compile data for  evaluation. This
programme involved the following:

*     Parameter study to develop data for NOX emissions  and  UBC
      as a function of coal properties, boiler load,  combustion
      mode, excess  air, burner  and AAP adjustment  and  burner
      stoichiometry.
*     Corrosion test by means of test tubes in the furnace wall
      and gas analysis near the furnace wall.
*     Study of the slagging conditions at  the furnace wall.
*     Characterization  of  the industrial applicability  of  the
      fly ash.
*     Recording  the  boiler  operation  under  normal   combustion
      conditions.

Extensive  tests  were  carried  out  in  the  period  from  January
1989 to  June  1990.  During this test period  the  demonstration
programme was interrupted due to the following problems:

*     Slagging at the burner and AAP throat
      Soon  after   the  beginning  of   coal  firing  in   the
      conventional combustion mode  heavy  slagging was observed
      at the AAP  throat.  This was  attributed to  the excessive
      refractory of the throat.
      This problem  was solved by  removing  some of  the  throat
      refractory. A  few months  later  the  same problem occurred
      at the burner throat. It was solved in  a similar way.
*     Damage to the flame stability ring
      After one year's service, damage to the flame  stabilizing
      ring  was  observed.  This  damage  was  more   serious than
      expected.   Inspection of  the  ring  showed  that  this  was
      caused by  extremely high  temperatures.  These  conditions
      occurred during gas firing and when the burner  was out  of
      service.
      To avoid high temperatures of the flame stabilizing ring,
      the  burners  were  modified.  This modification  involved
      lengthening of the gas spuds  and providing the  pulverized
      coal nozzle with purging air.
                           3-107

-------
In  spite  of  the  above-mentioned  problems,  comprehensive  data
have been developed for three coal types.
For each coal type a continuous  test  period varying  from  one  to
two months has been taken.

RESULTS OF THE COMBUSTION TESTS

Gas Firing

The post-retrofit  tests showed  that  a remarkable NO  reduction
was achieved after the boiler modification.
A summary of the test results is illustrated in Figure 7.

Before  retrofit  NO  concentration  was  slightly  higher  than
500 mg/mL   After  the   modification  NO  concentrations  were
between 150  and  250 mg/rn^, dependent  on  the burner  setting.  For
two-stage combustion the  influence of burner setting  was small
and a  NO  concentration of 100 mg/mj was  measured,  corresponding
to  80% reduction in comparison with before retrofit.
During  the  post  retrofit test a larger  excess  air  flow  (2%  02
at  air  heater  inlet)  is  necessary  to  achieve  low  carbon
monoxide concentrations.
Optimization  of  combustion   to  achieve  CO-free  combustion  at
lower  excess  air  was  not  an  objective  in  the  demonstration
project.

Coal Firing

In  this  section a  brief  overview  is  presented  of  the  data
obtained  after modification  and  a comparison  is made  with the
performance  before  retrofit.  A  summary  of the  coal properties
is  given  in  Table 3.

Effect  of Boiler Load   Typical NO  emissions  as a  function  of
boiler   load  selected   from   the   post-retrofit  tests   are
illustrated  in Figure  8.
In  general  it  is  apparent  that  NO  increases  with load.  A
similar trend  is observed for the  pre-retrofit results.
Concerning  the UBC analysis  of  the  fly ash UBC  appeared to  be
almost  boiler-load-independent.  The effect of excess  air on  NOX
emission  is  also shown  in this figure.

Effect  of Combustion Mode and excess Air  A comparison  of the
NO  and UBC  performance under pre-  and post-retrofit conditions
was made  for the coal  types investigated.
NO  and UBC as a  function  of  the  excess  air are  illustrated in
Figures  9,  10  and 11 for  three coal  types.
These  figures also  show the effect of the combustion mode, i.e.
conventional  and two-stage combustion.
For both combustion modes a  similar trend is apparent  for the
effect  of excess air on NO and UBC performance.
A  comparison  between  the post-retrofit  tests  and  the  pre-
retrofit  baseline tests gives the  following results:
                           3-108

-------
 *     For  the  conventional  combustion  mode approximately  30%
      NO.  reduction  is  achieved  after  modification  of  the
      boiler.
      At  low   excess   air   a   higher  UBC  is   measured  after
      modification  of  the  boiler.  This can  be explained as  the
      amount  of excess air  is  effectively  lower, because  2-4%
      of  the combustion  air is  supplied through  the  AAPs  as
      cooling air.  Moreover, the burners  are designed  for  two
      stage  combustion.   The   design  is   not   optimized   for
      conventional  combustion.
 *     In  two-stage  combustion  mode,  approximately   50%   NO
      reduction is  achieved  after modification.
      Concerning  UBC performance no  significant difference  in
      UBC performance  is  measured for 02-concentrations  at  the
      air  heater inlet higher  than   4%.  The measured  UBC  is
      within the limit  of 5%.

 Off-stoichiometric   Combustion     Tests  were   performed   to
 investigate  the  combined  effect  of   two-stage  combustion  and
 off-stoichiometric   combustion.   In    this   way   the   burner
 stoichiometry is  different  for  the upper  and the lower burner
 levels.  Off-stoichiometric  combustion  can  be  implemented  in
 several ways:
 *     by  controlling the  pulverized  coal flow  to  each  burner
      level  (mill operation) and/or
 *     by  controlling  the  combustion   air  flow  to  each  burner
      level  (windbox damper  operation).

 Due  to  the   limitations of  the mill  system  only one  test  with
 biased  fuel  distribution was  carried out.  As  shown  in  Figure
 12, the effect  on NOX was  small.

 Windbox damper  operation  offered more potential to  control  the
 burner  stoichiometry of each burner level. The results of  these
 tests are illustrated  in Figure  12.
 By reducing  the combustion air flow to the  lower burner levels
 by  20%  and  feeding this  to the  upper burners  more  than  20%
 extra NOX  reduction  was obtained, while UBC was  kept  constant.

 Effect  of After Air Adjustment   As mentioned before, the mixing
 of  the  after  air  with the combustion  products of  the  main
 combustion zone is  of paramount  importance.  Experiments  were
 performed  to optimize the  settings  of  the after  air  ports,
 including the rotation  and the momentum of the after air.
 The results of these tests are illustrated in Figure 13.
As shown  in  this figure,  it is apparent  that particularly  the
 secondary  air  register adjustment,   which  defines  a  certain
 swirl level, has an important  influence on NOX emission  and  UBC
 performance.
 Less  swirl  of  the  after  air  led to   lower  UBC and  higher NOX
emission.  For high  swirl  levels NOX  also  increases. The reason
 is  that for high  swirl   levels of  the  after   air  ports  the
distribution  of the combustion  air   between burners  and  after
air  ports   could   not  be   maintained.   These   data  therefore
represent a higher burner stoichiometry.
                           3-109

-------
SIDE EFFECTS OF COMBUSTION MODIFICATIONS

In the framework of the demonstration project KEMA  investigated
the possible  consequences  of  the  combustion  modifications  on
fireside corrosion,  slagging and  fly ash  quality.  Before  and
after  the  retrofit  samples were  taken and  measurements  were
performed.  The results and conclusions are reported below.

Fireside Corrosion

When two-stage combustion is applied, reducing atmospheres  will
occur  in  the furnace  locally.  When  such  atmospheres exist  at
the  surface  of  boiler  tubes,   fireside   corrosion   might  be
accelerated.  In  order  to  establish  the risk  of   increased
fireside corrosion for the evaporation tubes, eleven  test  tubes
were welded  in  the  furnace.  Eight  test  tubes are 5 metres  long
and consist  of  three  different  materials:  13CrMo44,  15 Mo3 and
310. To measure wall thickness accurately  30 test surfaces  were
marked. By means of ultrasonic measurement  the thickness  of all
test surfaces was determined.
The test tubes  were welded in the  front  and  side walls of the
furnace  in  positions  ranging  from  below  burner level  row  to
above the after air ports.
After exposure during 17 months the  test tubes were removed and
examined. It  appeared  that the thickness  of  only eight out  of
273 test  surfaces had changed  by  more  than 0.05  mm, which  is
the accuracy  of the measurement method.  Four  of these  samples
were not reliable. As for the remaining four samples  a decrease
was measured in two samples and an increase in the other two.

The  general  conclusion   was  drawn  that  after  17  months  of
exposure no  decrease of  wall thickness  was established  greater
than 0.05  mm.  This means  that the  decrease  of wall  thickness
over a  period  of  twenty  years under comparable conditions  will
be  less  than 1 mm.  It should  be  mentioned that the  conditions
near the furnace walls are  not extremely reducing.  Measurements
of  the gas composition  showed  that  some  Oj was  available and
that the concentrations of CO and H2S were  not  extremely  high.

Slagging Tests

Slagging is considered to  occur in  two   stages. In   the  first
stage  deposition  occurs  in the  liquid  state.  These  components
may  adhere  to  the  evaporation   tubes.   The  most   important
components  that  are  in  the  liquid  state   at  ±  450 °C  are
alkali/iron sulphates and partly oxidized pyrite.
After  the  formation of a  liquid layer other  materials can  be
deposited  due  to  the  sticky  nature   of the   material.   The
deposited materials  insulate the surface  from the evaporation
tubes,   so  that the temperature  increases  and gradually  higher
melting materials can be deposited.
The first stage of slagging was investigated with an  air-cooled
metal probe  at  425°C.  For the  second stage a ceramic  probe was
used at the local furnace temperature.
                           3-110

-------
The  results of  the slagging  tests with  the metal  probe  are
summarized  in  Table 4.  They  showed  that  the  deposition  rate
before  retrofit   varied  between  0.60-0.78  mg/cm  ,h.   After
retrofit  the  values  differ  over  a   wide  range   (0.04-10.24
mg/cm2,h) .
Compared with the composition of the coal ash the deposition  is
enriched in Fe and  to  some  extent  also  in Na, K and  S. The  low
iron content  in  the deposition  before  retrofit  was  remarkable
(2.2%). After retrofit values higher than 30% were measured  for
the  same  coal type. Comparable concentrations were  found  for
other coal  types. This indicates that after the retrofit  partly
oxidized  pyrite  has a  better  chance  to  start  the  slagging
process.
In the  same way  examination of the test tubes i.e. the surface
of  material 13CrMo44  with  SEM/EDS  indicated the  presence  of
alkali iron sulphates and possibly also sulphides.
In  agreement  with  the  above-mentioned  theory on  staged  slag
formation,  there was no  relationship  between the melting  point
of the  coal ash and the deposition of materials  on the  metal
surface. Also, the  influence of  the additive  copper oxychloride
was negligible.
The composition  of  the  deposits on the ceramic probe,  however,
is comparable to that of the coal ash.
In the deposits  sampled  when.Cerrejon coal  was fired  the  effect
of the  additive  CuOCl  is  clearly visible  in the  structure  of
the  slag.  SEM  graphs show  larger  pores in  comparison with  the
situation  without   additive.  For  Illawara  and  ANR  coal  this
effect  is  not  visible.  The structure of  these  samples is  more
brittle  than   the  Cerrejon  sample,  which  was  heavily  melted
through.

Fly Ash Quality

Lab tests were carried out  in  order to  assess the  applicability
of the  fly ash for  use  in  concrete or  cement.  The experiments
included  amongst other  things  the  determination  of  particle
size distribution,  compression  strength  and  surface area.  All
the  investigated fly  ashes met  the standards  for  industrial
application.

OPERATIONAL EXPERIENCE

During  the  day the  boiler  is  operated  at full load.  Two-stage
combustion  is applied as the normal firing  mode. For  the  start-
up procedure  and at night,  when the boiler is operated  at  low
load, the after air  ports are  closed and  conventional firing is
applied.    Although   from    visual    observation    combustion
performance   is   still  good   at   low   load  with   two-stage
combustion, the flame signals decrease  for  some burners.  It  was
shown that  the problem can  be  solved by adjusting of the flame
scanners.

In  general the  operational experience  is  positive  after  two
years  of  low-NOx operation. There  were no  adverse  effects  on
boiler efficiency.   In general  the  low-NO, combustion  system did
                           3-111

-------
not effect the  dynamic  behaviour of the boiler, except  for  the
mentioned change in firing mode at  low  loads.  The  problems with
slagging on the  burners  and after air ports were  solved.  There
is  no  significant  difference  in furnace  wall and  superheater
slagging behaviour before and after retrofit.

Bottom-ash Hopper Explosions

In  September  1989,   about  one  year  after  commissioning,   an
explosion in  the bottom-ash hopper  occurred  during  combustion
trials with ANR  coal  and two  doors of the hopper  were  damaged.
The  bottom-ash  hopper  is  filled with  water  and  consists  of
four,  partly  connected,  compartments  from which the  ash  is
removed at regular intervals through  de-ashing doors. After  the
first  time  a series  of  explosions,  some  more  violent  than
others, followed. At  first  the explosions occurred  during  two-
stage  combustion   but   later  on   also  during   conventional
combustion. It was  concluded  that the nature  of the  phenomenon
was a steam explosion and not a gas explosion:
*     CO-measurements  just  above the  bottom-ash hopper  showed
      only traces (maximum 300 ppm);
*     the increase  in  furnace  pressure during  an  explosion  was
      small and  no  damage was found  in the boiler  itself;  the
      damaged  doors  indicate  a  shock-wave  inside  the  water
      pool.

Before the  boiler  modifications no problems  of this kind were
experienced,  although  in the past  small  pressure  waves  in  the
boiler  were  measured  when  big  lumps  of  slag fell  into  the
hopper.  Therefore,  before  retrofit  measures  were  taken   to
reduce the  slagging propensity by adding copper oxychloride to
the  coal  in  order  to obtain  a  more  crushable slag.  Although
regular  observations   did  not show  an  increase   in  slagging
propensity  after the  retrofit, a  possible explanation  for  the
explosions was  the  formation  of  much more porous  and  crushable
slag   than    in  the    past,    which    disintegrates    rather
instantaneously  on  impact  with  the water  filled  hopper.  This
would  lead  to a very  fast heat  exchange and  subsequent  steam
formation. The following measures were taken:
*     a change in the burner settings  in  order to  obtain a  more
      slender  flame with less  impingement on  the  side  walls of
      the boiler;
*     no more addition of copper oxychloride;
*     as  it   was   experienced  that  sometimes   an   explosion
      occurred  after  the refilling  of the  hopper with  water,
      more gradual refilling was applied;
*     safety  valves were applied on  the  doors to  avoid damage
      in case of an explosion.

Some  bottom ashes were  analyzed to  get  an  impression of  the
porosity  by measuring  the density.  Comparison with  the  data
before retrofit  did not  show  great  differences. Also  no  clear
relation  was  found  between  coal type  and the  occurrence  of
explosions. Until now no  satisfying  solution  has been  found  and
small  explosions  continue  to  occur,  but  they  do  not  cause
serious damage and the problem is manageable.
                           3-112

-------
Emission from Day to Day

It is  interesting  to  compare emission levels day by  day  during
normal  operation  of  the   boiler   with   the   results  of   the
demonstration   tests.   Therefore,   each   day   during   stable
conditions,  the NO  concentration  is measured  in  combination
with several important parameters.  In  Figure 14  the  NO  emission
in mg/rn^ at 6%  Oj  and  the amount of  unburnt  carbon in the  fly
ash  is  presented for the month  September 1990.  In  this  period
the  boiler was  operated near full load. All data represent  two-
stage  combustion.  Steam production  and 02 concentration  before
the  air heater  are also plotted  in  the figure.  Three  coal  types
were fired  in  this month: Mingo  Logan,  Hobet  and Anker  Blend.
For  three  days  coal blends  were  fired.  The composition of  the
coal is given in Table 5.

The  average  values  for  one   month were:  steam   production
550  t/h; NO concentration  540  mg/mj  at 6% Oj and unburnt  carbon
1.7%.  Further   evaluation  of the  emission data during  normal
operation  of  the boiler  is  still  in  progress,  but  these  data
indicate that the  NO  figures for normal   operation are  close  to
those obtained  in the tests; the fly ash  quality was good.

CONCLUSIONS

The  HTNR burner  developed  by  Babcock Hitachi  for  pulverized
coal  firing  was  modified  to   accomodate  both  coal  and  gas
firing.  The  dual  fuel  burners   and after   air   ports   were
installed  in the Maas Power Station  unit 5 to  demonstrate  the
viability of low-NO.  combustion  techniques.
In the  framework or  the demonstration programme the  combustion
system  was optimized  and extensive  tests were conducted  for
three coal types and natural gas.
For  natural  gas the NOX  reduction amounts to values  between  50
and  70%  in the  conventional combustion   mode and  80%  for  two-
stage combustion. For coal  firing typical reduction  percentages
are  30%  and  50% respectively.   Adjustment of  the  burners  was
generally  effective  for  NOjj and UBC. Adjustment  of the  after
air  ports appeared to be critical for UBC.
Typical results  obtained  for Cerrejon coal are  480  mg  NO/m^ (6%
02 )   at  95%  load  and  25%  excess  air.   For  ANR  coal  the  NO
concentration was  500 mg/mj at 6% Oj. Combination of  two-  stage
and   off-stoichiometric  combustion   resulted  in   15%   more
reduction without  impairment of  UBC.  For Illawara  coal,  which
is characterized by  a  high fuel ratio,  the  NO  concentration
amounts to 600 mg/m^ at  6%  Oj.
In the  framework of the  demonstration programme the impact  of
combustion    modifications   on    fireside    corrosion    was
investigated. It  appears that under the conditions  prevailing
in the furnace  of Maas Power Station  Unit 5 there  is no risk of
increased fireside corrosion.
During the programme  severe  slagging problems have  occurred on
the  burners  and after  air  ports.  These  problems  were  solved.
Measurements with  a slagging probe  indicate that  there is  an
increased  tendency  to  form  a  liquid   deposition  layer  in
comparison with the situation before the  retrofit.  In
                           3-113

-------
contradistinction  to  these  slagging  tests,  it  appears  that
there  is  no   significant   difference  in  furnace  wall   and
superheater slagging behaviour before and after retrofit.
In  general the  operational  experience  is  positive  after  two
years  of  low-NOx  operation.  There  were no  adverse effects  on
boiler efficiency.  In  general  the  low-NOx  combustion  did  not
effect the dynamic  behaviour  of  the boiler, with  the exception
of  the  change  in  firing  more   at low  loads.  Until  now  no
satisfying solution has been  found  to  prevent  bottom-ash  hopper
explosions,  but  they  do   not  cause  serious  damage  and  the
problem is manageable.
                          3-114

-------
Air
Air
     -0-
AAF
                       BURNER
                               BURNEI
                               Air
                                                       Air
              I
              1	.
                           MILL 1
                   -0--


                   -0"
          -0-	
                           MILL 2
         FIGURE 1. SCHEMATIC PRESENTATION OF THE FURNACE
                  AND COMBUSTION SYSTEM
                                                                laniter
  FIGURE 2.  MODIFIED  HTNR-BURNER FOR  COAL AND GAS FIRING
                            3-115

-------
                                   Primary Air
  FIGURE  3. AFTER AIR PORT  (AAP)
           300         400   t       500          600
                            -Checkpoint for NOx
                         Coal Feed Rate  (Kg/h)

FIGURE 4i  RESULTS OF COMBUSTION TEST   4 MW TEST BURNER
          COAL TYPE: WAMBO
                    3-116

-------
                     ilLj_
              FIGURE 5.  WATER MODEL FOR FLOW SIMULATION
               Sf = Sr
               Ff  =  Fr      Sf 
-------

1
Q
? --.-.
S
Q

5

^
i»-i
»-i

0 6

-^
•^"
^--'

O 7
^--
^-"
^--^


o a
o L Le r

^^
r-^


0 9
.oad ("
0

y


O i

• 3* 02
4 4.3» O2


DO
                                                   2H 02

                                                   
          FIGURE B. EFFECT OF BOILER LOAD  ON  NO-EMISSION
                    COAL TYPE:  CERREJON
                             3-118

-------
14OO
eg
O
S! 100O
if
1
en 800
£ 600
§
40O








/
/ ^
<^-




X
^-

»






CONV.
A PostRelrofit
CCXJV
O PostRolrofit
TSC
                                                                    O   BOO
               O2 Inlet Alf Heater <%)
                                                                                  O2 njet Au  Hetatwr (-/J
    b   8
    1
\
                     Pre-Relrofit
                     CONV
                     PoslRctrofit
                     CONV
                     PostR«trofil
                     TSC
                                                                                    \
PoilHk,
rsc
               O2 Inlet Ar Heater 1%)
                                                                                  O2 inlbt  Au  Heotej  ("A,)
           FIGURE  9.  EFFECT OF  EXCESS  AIR
                       CERREJON AT 951 MCR
                                                        FIGURE  10. EFFECT OF EXCESS AIR
                                                                    ANR AT 1001 MCR
O   800
#
               246

              O2 lnle( Aif Healer 1%)
                                             Post  Heiio-
                                             fit CONV.
                                             Post  Reuo-
                                             fit TSC
                                             Post Reuo -
                                             fit CCTJV.
                                             Post Reuo-
                                             ht TSC
              2468

             O2  irilet  Aw Hea itif IW


         FIGURE  11.  EFFECT OF EXCESS AIR
                       ILLAWARA AT  95Z MCR
                                                                                                                   3   O
                                                                               0,ee-1.O4  076-1.14  1.04-O.67   1.O4-O.9
                                                                                 Sfl txjner   Bottom - U|»«-
                                        FIGURE 12.  EFFECT OF  BURNER  OFP-STOICHIOMETRY
                                                     OH TWO STAGE COMBOSTIOH
                                                      3-119

-------

0 500
(0
1

1
E;
Q 400
A


9 ^


O
~. _





-









x





/:
/ I
/


_ " ^


'' .

3


-


A prim air
k O prim air
damper 30%
V prim air
- damper 50%
L D prim air
1 damper 75%
O prirn air
damper 1OO%
                   0    20    4O   60    8O   10O

                      Secondary Air Register AAP (%)

o
7
A- -








I




N
X
s
ss!




s
N
",
S

^ prim air
damper 1O%
O prim air
damper 3O%
7 prim air
damper 5O%
D prim air
damper 75%
° prim air
damper 1OO%
                   0    2O    4O    6O    80    1OO

                      Secondary Air Reaister AAP (%)



FIGURE 13.   EFFECT  OF AAP ADJUSTMENT  CERREJON AT 95% MCR
    steamprod.     —-  O2-conc bef  	NO—cone  in    —  • unbunt
    t/hr                airheater           rng/hin3            carbon
ouu

-------
                             TABLE  1
      NOX  EMISSION STANDARDS FOR COAL-FIRED POWER STATIONS
Power station
    Commissioning   Standard
                  Remarks
Maas no . 6
Maasvlakte no. 2,
Borssele no. 12
Maasvlakte no. 1
Amer no. 9
Hemweg no . 8

Maasvlakte no. 3
1986
1987
1987
1988
1993
1994

1997

270 g/GJ
* 750 mg/m3

400 mg/mj
300 mg/rn^

200 mg/m]
test value
190 g/GJ * 530 mg/m^
as a criterion for
low-NO. burners
&
commitment to pursue
200 mg/rn^

                             TABLE  2
                   MAAS  POWER  STATION UNIT 5
Unit capacity
Boiler type
Boiler manufacturer
Commercial operation
Steam production

Steam pressure/temperature
at superheater outlet
Number of burners
Burner heat capacity
Number of mills
Mill type
Coal consumption
Coal/air ratio
              177 MWe
              Benson
              Stork Boilers
              1966
              580 t/h
              188 bar  /  540°C
              16 (2x2x4, opposed)
              30 MW
              2
              tube mill
              64.5 t/h
              0.725
                             TABLE 3
                         COAL PROPERTIES
LCV/GCV

Moisture
Ash
Volatile
HGI
Fuel ratio
Ndaf
MJ/kg
Cerrejon

26.3/27.5

  12.4
   5.4
  33.0
  48
   1.49
   1.60
Illawara

26.9/27.8

   5.0
  16.6
  19.1
  78
   3.12
   1.55
 ANR

28.5/29.6

  6.6
  7.7
 31.8
 50
  1.69
  1.70
                           3-121

-------
                            TABLE 4
               DEPOSITION ON METAL SLAGGING PROBE

Coal type          Additive          Deposition rate
                    CuOCl               (mg/cm2.h)

ANR*                  +                 0.60-0.78
Cerrejon              +                 0.06-0.57
Cerrejon              -                 0.04-0.30
Illawara              +                 0.30-1.26
Illawara              -                 2.02-3.13
ANR                   +                 0.40-5.45
ANR                   -                 0.40 - 10.25
   pre-retrofit
                            TABLE 5
               COAL  COMPOSITION  IN  SEPTEMBER  1990

                Mingo Logan     Hobet     Anker Blend
                   "A            B            C

Moisture (%)        7.5          6.9          9.3
Ash (%)            10.5         11.2         11.6
Volatile (%)       28.0         30.7         28.9
LCV (MJ/kg)        27.6         27.4         26.5
N-daf (%)           1.5          1.6          1.6
                          3-122

-------
    Three-Stage Combustion (Reburning)
         on a Full Scale Operating
          Boiler in the U.S.S.R.
                    By
R.C. LaFlesh, R.D. Lewis, and O.K. Anderson
       Combustion Engineering, Inc.
1000 Prospect Hill Road, Windsor, CT 06095

              Robert E. Hall
  U.S. Environmental Protection Agency
 Air and Energy Engineering Research Laboratory
        Research Triangle Park, NC 27711

                V.R. Kotler
All Union Heat Engineering Institute (VTI)
                Moscow, U.S.S.R.

-------
                               ABSTRACT

This paper  presents the  results of  a program  to  complete  the  preliminary
design of a three-stage combustion  (reburn)  system  for nitrogen oxides (NO )
emissions control on an operating 300 MWe coal  fired  boiler  in the U.S.S.R.
This project was sponsored by  the U.S.  Environmental  Protection Agency (EPA)
in support of  the protocol  of the Eleventh Meeting of the  Stationary Source
Air Pollution  Control  Technology Working Group, Moscow, U.S.S.R.,  November
1988.

The program to design  the reburn  system was  composed  of five  major tasks:  1)
visiting  the  host  site  in  the  Ukraine   to  exchange  design  and  operating
information;   2)  translating  Soviet   design   documents   into  English;   3)
performing process calculations;  4)  conducting physical flow modeling; and 5)
developing a  preliminary system design  which  included  general  arrangement
drawings and furnace performance analyses.

The overall preliminary reburn system  design was completed and was presented
to and accepted  by  Soviet representatives during a June  1989 meeting at the
EPA's  Air and  Energy  Engineering  Research Laboratory  (AEERL)in  Research
Triangle  Park,  NC.   The  Soviets are  currently  completing the  final detail
design and are targeting  completion  of hardware  fabrication and installation
by the  fourth  quarter  of 1991.   All indications to date are that reburning
will be  a viable NO   reduction  technology for  the type  of boiler (opposed-
wall-fired, wet bottom) that the host steam generating unit represents.

BACKGROUND

A joint U.S./U.S.S.R. committee for cooperation in the field of environmental
protection has sponsored meetings of a working group on stationary source air
pollution control technology over the  past 13 years.   The U.S. Environmental
Protection Agency's Air and  Energy  Engineering  Research  Laboratory has  been
responsible for  technical information  exchange  activities under this program
and, as  of the  Eleventh Working Group meeting in Moscow, November 1988,  has
sponsored the first major joint  U.S./U.S.S.R. air  pollution control research
project with the objective of  implementing NO   control technology  on a large
coal fired boiler in the Soviet Union.

The Soviet Union has substantial interest  in controlling air pollution and is
currently  developing   a  program  for  legislating  NO   emission levels  from
electrical utility  boilers.    The current plan  calls  for  the  following NO
emission rules to be implemented  for  all  new boilers  having greater than 463
tons/hr (420 metric tons/hr)  steam flow:

     Natural Gas   0.08 lb/106 Btu (125 mg/Nm  @ 3% 0  )
     Fuel Oil   0.12 lb/10  Btu (185 mg/Nm  @ 3% 0 )
     Coal   0.18 lb/10  Btu (225 mg/Nm  @  6% Q^)

NO  emissions  from  existing  utility boilers are also to be  regulated under
the proposed  legislation.   Again, the  current  plan calls  for the following
NO  emission rules:
  x
                                   3-125

-------
     Natural Gas   0.15 lb/10  Btu (250 mg/Nm  @ 3% 0 )
     Fuel Oil   0.19 lb/10  Btu (290 mg/Nm  @ 3% 0 )
     Coal (brown)   0.28 lb/10  Btu (340 mg/Nm  @ 6% 0 )
     Coal (bituminous)   0.33 lb/10  Btu (400 mg/Nm  @ 6% 0 )

The  joint  U.S./U.S.S.R.  program  called  for  the  U.S.  side  (Combustion
Engineering, Inc. under contract to U.S. EPA) to provide the Soviet side with
a  preliminary  design for  an in-furnace  NO  control  system for  a specific
Soviet  boiler  in anticipation  of  meeting  impending  NO   legislation.   The
Soviet  side's  participation  in the  project includes  final detail  design,
fabrication,  installation,  and  testing of the  system.    The  U.S.  side  is
assisting in  an  evaluation of the system's  effectiveness  in controlling NO
by  providing NO  monitoring instrumentation  and  technical support  during
testing.

Both  sides  focused  on the use  of  reburning as  the technology  of choice for
this  project.   Reburning  (referred  to  by  the  Soviets  as   "three-stage
combustion")  is  an  attractive  alternative  for  in-furnace  NO   control where
either  physical  or  operational modifications  to  the  existing  fuel  firing
system  (as  required with most low  NO   retrofits)  would be problematic from a
boiler  operating standpoint.   Application  of  reburn  technology does  not
require  any configuration  operational changes  to a boiler's  existing firing
system.  The Soviet side selected  a wet bottom  (slagging)  unit  as the design
case; any change in the existing firing system to reduce NO  could negatively
impact  on slag  management  in the boiler.  As a  result,  reburning technology
was  chosen  for implementation  on  the Soviet boiler.   The boiler  chosen to
represent the  design  case is located at the Ladyzhinskaya  Power Station in
the  Ukrainian  city  of Vinnitsa.  (Figure  1.)   The power  station  consists of
six   300  MWe  coal   fired boilers   of  the  Soviet   type  Tnn-312.    These
supercritical  (3625  psig  or   255   kg/cm )  steam  pressure  units   employ
swirl-stabilized opposed-wall-fired coal burners (16 per  boiler) which burn
locally available high volatile bituminous coals of relatively high (35%) ash
content.  The boilers  operate under slagging conditions;  that  is,  a  portion
(-30% by weight)  of  the ash is  retrieved as wet  slag  at  the furnace  bottom,
and  the  remaining 70% by weight  of ash input into the boiler is collected at
the  furnace outlet by electrostatic precipitators.

According to  Soviet  data,  baseline NO   emissions from  these  units typically
range between 0.5 and  1.0 lb/10  BtuX (650 and  1300  mg/Nm  @ 6%  0  ).   The
objective  of the  project  described  herein is   to  demonstrate  that  reburn
technology  can  reduce  NO   emissions   to a   level  consistent  with  the
aforementioned proposed  NO  rules on  a  Soviet  type TTTTr-312 boiler.   If the
demonstration is successful, the Soviets may extend the use of the technology
to  the  other units at Ladyzhinskaya as well as  300  other units  of this class
located elsewhere in the Soviet Union.

REBURN  PROCESS OVERVIEW

The  concept  of  reburning  or   three-stage combustion  and  the  postulated
chemical reactions which account for  the reduction  of NO   in the reburn zone
have been addressed elsewhere in the literature and will 'not be reiterated in
detail  here (1,2,3,4).   Briefly,  reburning is  an  in-furnace  technique for
reducing  NO   by  creating  a  slightly  reducing  (substoichiometric)  zone
downstream  of the primary combustor as shown schematically in Figure  2.  The
                                   3-126

-------
reducing zone  is  created by introducing  fuel  into a zone  with insufficient
oxygen available to fully combust the  fuel.  The  presence  of a reducing zone
creates  intermediate   nitrogen-containing  species   (e.g.,  NH.,  HCN)  which
subsequently react with  previously formed  NO   to form the  desired product,
molecular  nitrogen.   Any   unburned  fuel  leaving  the  reburn   zone   is
subsequently  burned  to   completion  in   the  burnout  zone  when  additional
combustion air  is added.  Reburning can  be used on  all types  of  fossil fuel
fired boiler configurations using coal, oil, or gas  as  primary fuels and,  in
fact, has been  successfully  employed  on  a number of large  utility  oil fired
boilers  in  Japan  where   oil  has been used  as  the reburn  fuel  (5) .   The
technology is  particularly  adaptable  to   slagging  furnaces  employing cyclone
combustors  (6)  or  swirl stabilized  burners  similar  to  those  used  in  the
Ladyzhinskaya  units.   Since these  combustors  may  not be  able  to tolerate
significant changes to their operation,   such  as  lower excess  air  or staged
air  injection,  without the possibility of creating other  problems  (such  as
slag tapping:  the removal of coal-ash slag through  the furnace bottom while
still in its molten state),  they are limited to  an  in-furnace NO  reduction
technology that does not depend on significant  changes  to  its present mode of
operation.    Reburning does  not  require  that any  significant  operational
changes be made to the primary  combustor  or burners.   The  key requirement is
that  the fuel feed   rate  be  reduced  in the   primary  combustor  with  an
equivalent amount  (on a  Btu basis)  of fuel being  injected into the reburn
zone, usually  not more  than  20% of  the  total fuel  input.  The excess  air
(hence the air/fuel  stoichiometry within  the  main  burner zone)  can  be held
constant, thereby avoiding the potential  for operational problems.

SYSTEM PROCESS DESIGN

The project was initiated by an exchange  of boiler design information between
the  U.S.  and  Soviet   sides.  Meetings  were held both  in the  U.S.   and  at
Ladyzhinskaya  in  order  to  facilitate the  transfer  of information.   Reburn
system  process design was  then  initiated based on  the   following  overall
process objectives:

     1)   Meet key design criteria for effective NO  reduction while
          minimizing any impact on normal boiler operation.

     2)   Incorporate operational flexibility within the design to permit
          optimized host unit performance.

Key  design criteria commensurate  with the  overall design objectives  for the
reburn   system  were   initially  established.    The  criteria   consist   of
theoretical  criteria  for  effective  NO  reduction, obtained  through the open
literature,   and  practical,   commercial   considerations  for  reburn  system
design,  installation,  and operation.

Key design criteria for the reburn zone were determined to be:

•    Inject  reburn fuel into as high a temperature zone as possible.
•    Maintain average  stoichiometry between 0.90 and 0.99.
•    A small amount of 0. should be present to  promote formation of OH and H
     radicals.
•    Maintain a minimum furnace gas residence time of 0.5  sec.
•    Maximize entrainment, mixing, and dispersion of reburn fuel.
                                  3-127

-------
•    Avoid direct fuel impingement on boiler walls.
•    Minimize the number of required boiler penetrations.
•    Locate fuel injection nozzles to minimize boiler/structural steel
     modifications.
•    Provide for maximum flexibility of reburn fuel jet direction and flow
     rates.

Key design criteria for the burnout zone were determined to be:

•    Inject burnout air in as low a temperature zone as possible commensurate
     with obtaining fuel burnout before entering the superheater surface.
•    Provide for rapid mixing of air to minimize pockets of unburned fuel.
•    Avoid direct air impingement on furnace walls.
•    Minimize final excess oxygen commensurate with obtaining good fuel
     burnout.
•    Maintain a minimum furnace gas residence time of 0.6 sec.
•    Minimize the number of required boiler penetrations commensurate with
     obtaining good mixing.
•    Locate burnout air injectors to minimize boiler structural modifications
     while providing good mixing.
•    Provide for maximum flexibility of air jet direction and flow.

With the above in mind, as well as the Soviet utility's requirements that the
reburn  system not  adversely affect  slag  tapping, not  increase  tube  metal
temperatures  beyond  design limits,  and  not affect general  slagging/fouling
characteristics,  CE  initiated  the  preliminary  design by  performing  mass
balance and combustion calculations on the overall process.  Existing process
flows were used  in this calculation,  along  with  the Soviet coal analysis, an
estimate  of  flue gas  recirculation  (FGR)  mass  flow  necessary for  the  fuel
injectors  (natural  gas used  as  reburn  fuel,  FGR as  a transport medium to
enhance  mixing),  furnace   dimensions,  and   Soviet   supplied  furnace   gas
temperature  information.   A  proprietary CE  computer code  was employed to
calculate stolchiometric ratios and gas residence times in boiler zone 1  (the
furnace bottom  to the reburn  zone  start or main  burner zone) , zone  2   (the
reburn  fuel   injection  position  to  the burnout air  injection position or
reburn  zone),  and  zone   3   (the burnout  air  injection  position  to  the
horizontal furnace outlet  plane,  or  burnout zone).  Estimates for  the reburn
fuel and burnout  air  injection position  (elevation) were input into  the  code
and then  iteratively  adjusted  to achieve reburn/burnout zone stoichiometries
and furnace gas  residence  times  consistent  with  the key design criteria.  An
example of output from the process design calculations  for  the final design
case for the reburn zone is shown in Table 1.

The above process calculations set the preliminary elevations for  the reburn
fuel  and  burnout air  injectors,  based  on  calculated furnace  gas residence
time of 1.59,  1.00,  and 0.92 sec  for zones 1, 2, and  3,  respectively.   The
preliminary reburn fuel elevation was set at 66.6 ft (20.3 m) and the burnout
air  injector elevation at  95.8  ft  (29.2  m)   consistent  with  the above gas
residence time calculations.

PHYSICAL FLOW MODELING

Isothermal flow  modeling  studies were conducted  as  part of  this  program in
order to optimize the number, location, configuration, and operating
                                  3-128

-------
parameters  for  the reburn  fuel  and burnout  air injector  system.   Baseline
furnace aerodynamics were evaluated as well as the performance of a number of
candidate reburn system configurations.

Total mass  flow rates for  the reburn  fuel,  carrier gas  recirculation,  and
burnout air,  as  well as injection  elevation  in the furnace,  were specified
from the  process flow calculations  described previously.   Candidate  reburn
fuel  and  burnout  air   injector  configurations  were  identified  based  on
previous experience.

These  configurations  were  then  screened  by  calculating   expected  jet
trajectories  for each  configuration's  reburn  fuel  and  burner   air  nozzle
arrangement.  Jet velocities  investigated  ranged from 146  to  275  ft/sec (45
to  84  m/sec) .  It  was  assumed that   all  the  reburn  fuel  and burnout  air
injectors  would  be located  on  the   boiler's  front  and  rear walls.   The
boiler's  side walls  were   not  considered  for  potential   injector  locations
because of equipment interferences on  the actual unit.

The jet  trajectory  screening  process   indicated that reburn fuel  and burnout
air jet velocities  in the 146-229 ft/sec (45-72 m/sec)  range were promising,
as were  configurations which  located  six injectors on the  front  and/or rear
walls  for  both  reburn fuel/burnout air cases.   The trajectory calculations
also  concluded  that jet   penetration and  potential  for  mixing could  be
enhanced if both  the reburn fuel/burnout air injectors could  be  tilted down
as much as 30  from horizontal.

With consideration  of the  above,  physical flow  modeling was  initiated.   A
1/16 scale  geometrically similar  isothermal  flow  model  of a Ladyzhinskaya
boiler was fabricated.  This model,  shown in Figure 3,  encompasses the boiler
from the furnace bottom to  the inlet of the economizer section.

Particular  attention  was paid to  the design of  the existing  main  burners,
reburn and burnout  air nozzles, and existing  gas  recirculation nozzles (used
to  control steam temperatures).    The main  burner "free"  exit   areas  were
adjusted  in  accordance  with Thring-Newby  modeling  criteria (7) in  order to
account  for the  combustion  process  expanding those gases  exiting  the burner
and reducing  jet  momentum  flux.   Jet  penetration and  dispersion,  as related
to  the  reburn fuel,  FGR, and burnout  air jets,  are modeled  by  maintaining
equivalency between  the  inlet jet to  bulk furnace  gas  mass ratio  while
simulating   jet   trajectories.    These   modeling  procedures  have   been
successfully applied in over 30 years of CE modeling experience. (7)

The first  series of  tests  performed  in the  cold  flow model  was  a  baseline
evaluation  of  the  Ladyzhinskaya  furnace's  aerodynamics  with  the  model
simulating normal (non-reburn) operation.   A qualitative  evaluation  of the
flow field  was  performed using  smoke to  trace jet penetration  and mixing;
limited quantitative tests were also performed to establish three-dimensional
furnace gas velocity profiles under baseline conditions.

The  baseline  characterization highlighted  the  fact  that the   flow  field
exiting the main burner zone was reasonably uniform in terms of direction and
velocity,  suggesting that  the reburn  system injector/burnout air  injector
system would not have to overcome major flow maldistributions  in the existing
furnace aerodynamic flow field in order to  provide for adequate gas mixing.
                                  3-129

-------
Following the baseline  tests,  simulated reburn fuel and  burnout air nozzles
were fabricated and  installed  on the furnace flow model.   It was decided to
model  two  compartments at  each reburn  fuel/burnout  air  injector position.
Based  on   previous   experience,   this   arrangement  would   both  maximize
penetration and dispersion of  the  natural  gas  reburn fuel, recirculated flue
gas  transport  media, and burnout  air streams,  and provide  for operational
flexibility.  A  "pant  leg"  nozzle configuration  for  the  upper compartment
(Figure  4)  enhances  lateral  jet   dispersion;  while  the  lower compartment
employed a  conventional single nozzle arrangement.  All  compartments  in the
model had the ability to tilt up or down to investigate the effect of tilt on
jet mixing.

The total number of  injectors  investigated  in  the  model  ranged from 12 to 32
(inclusive  of  both  the reburn  fuel  injectors and  burnout  air injectors);
these  injectors  were  assumed  to  be  symmetrically  located on  the  boiler's
front and rear walls.

Reburn system  configurations were  tested  in  two  phases.   Phase 1 evaluated
the  performance  of  the reburn fuel injection  system  without a burnout air
system  in  operation.   Phase  2  evaluated  the  performance  of   burnout  air
injection systems with  the best  performing  reburn  fuel  system from the Phase
1  efforts.   Nozzle  free  areas and mass  flow  rates were  the key variables
investigated.

REBURN FUEL SYSTEM FLOW TESTING

Figure 5  presents the  jet  trajectories  of three  reburn  fuel  injection jet
velocities  for the  configuration where 12  individual  injection  nozzles were
located on both the  front and rear walls for a total of 24  injection nozzles.
Injection vertical tilt angle  was 0 .  The trajectory  lines  shown represent
the  leading  edge  of  the visualized jet.   For  the  sake of clarity,  only the
trajectories from the  front  wall have  been  shown.   The trajectories from the
rear wall  were symmetric  to  the front wall trajectories.   It  can be seen in
Figure 5 that 146  ft/sec  (45 m/sec)  injection  velocity is sufficient to have
the  reburn fuel  jet penetrate  slightly past  the centerline   of  the  unit.
Remembering that  there  is  another  set of injectors located on the rear wall
injecting  at  the  same  time,  penetration  of slightly past  the  centerline is
considered  ideal  in  terms   of  providing  a uniform  fuel  distribution.   The
other  injection velocities over-penetrated  and  impacted  the opposite wall of
the  furnace.  The 183 ft/sec (56 m/sec) jet impacted at a point  just slightly
below  the  burnout air  injection elevation.  The  229 ft/sec  (70 m/sec) jets
Impacted the opposite wall just slightly above their point  of injection.

Figure 6 shows the dispersion pattern (flow visualization using  smoke tracer)
for  the   146  ft/sec (45  m/sec)  reburn  fuel  jet  at  0° tilt.   This  jet
penetrates almost horizontally to  the  center of the  unit before it turns up.
Dispersion was found to be very good  in  all  but a small  area located along
the  wall  just  downstream of the  injection point.   This  particular  area was
devoid  of  injected  material.   The   jets  quickly  dispersed   in  both  the
side-to-side  and  front-to-back  directions.   As   a  result  of  the  physical
modeling, the reburn injector jet velocity design target was established at  a
nominal velocity of  146 ft/sec (45 m/sec).
                                  3-130

-------
For all  configurations and  injection velocities,  tilting  the  nozzles  down
significantly increased the  amount of dispersion but not the  cross furnace
penetration.  At an injection velocity of 146 ft/sec (45 m/sec), for example,
the penetration did not appear  to  increase more  than a few  percent.  Angling
the injection nozzles down did, however,  result  in some  of  the reburn fuel's
being intermittently  recirculated  into  the  main burner  zone  of the furnace.
This observation is shown in  Figure  7.   Down-tilting did result in increased
downward penetration  and  thus  increased  the  residence  time  of  the injected
fuel.   Tilting  the  injection nozzles down,  therefore, was recommended  as a
means of increasing the overall flexibility  of  the field installed injection
system.   Downward  injection  may be  useful  as  a method  of  optimizing  the
reburn process  if  the furnace  load should  change and if  the  temperature
profile within the injection zone should change.

For all injection scenarios,  the side-to-side dispersion of injected material
was found to  be  adequate,  Figure 8.   Material injected  from  the two nozzles
located adjacent to the side walls experienced strong jet attachment to these
walls.  This phenomenon is detailed  in Figure 9.  Because this jet attachment
is unwanted, it is recommended  that  the  side wall injection  nozzles be given
yaw capability  in  order to direct flow  away from  the  wall.   Alternatively,
they should be designed at a fixed yaw angle to  inject at least 18  away from
the walls.

BURNOUT AIR SYSTEM FLOW TESTING

The performance  of  the burnout  air  system was  almost identical to  that
observed for  the reburn fuel  system.  The major  difference  in the injection
performance was  in the penetration  of  the jets  for a  given  velocity.   Two
reasons  for this have  been  identified.   First,  the  mass  flow rate  of  the
burnout air is  70%  higher than that  for the reburn fuel.  This,  in itself,
under identical  cross flow conditions,  would account  for a  30%  increase  in
the penetration  of the  jet.    Second,  the  aerodynamics  in  the burnout  air
injection zone are more conducive  to  higher  levels of penetration because  of
the change in flow direction caused by the turn at the top of the unit.

Figure 10 presents  the leading edge  trajectories of three burnout  air  jets
simulating  146,  183,  and  229  ft/sec   (45,  56,  and  70 m/sec)  from  flow
visualization tests.   For this  case,  there were  6 windbox locations  on  both
the front  and  rear  walls  for  a  total   of  12  burnout  air   locations.   The
trajectories from the jets located on the rear wall were omitted on Figure  10
for the sake of clarity.  Injection  at 229 ft/sec (70  m/sec)  resulted in the
jet's impacting the opposite wall of the unit almost directly across from the
point of injection.   At 183  ft/sec (56  m/sec)  the jets  missed the rear  wall
and exited  the  unit  low above  the arch.  At 146 ft/sec  (45  m/sec)  the  jets
quickly  turned  upward and did  not  provide  sufficient penetration  into  the
center of the  furnace.   Rear wall jets  did  not penetrate as  effectively  as
front wall  jets because  they  were   injecting  into  a  flow  field  that  was
approaching counterflow.

It was  recommended from  the  physical  flow  modeling  phase  that  the  design
injection velocity  range  of  the burnout air system be between 183  and 229
ft/sec (56  and 70 m/sec).   A  burnout air  injection  system  having velocity
capability  in  this  range  would enhance  rapid  burnout air's  mixing into the
                                  3-131

-------
bulk  furnace  gases  issuing  from  the  reburn zone,  thus  ensuring  complete
burnout of combustible products.

PRELIMINARY REBURN SYSTEM DESIGN

The  approach  used  to  develop  the  preliminary  reburn  system  design first
involved  establishing initial  locations  and  configurations  for  the  reburn
fuel/FGR  and burnout  air subsystems, including shape, flow rate,  and nozzle
velocities.   These  selections were based on  the previously described reburn
system  design  criteria,   mass  flow  balance/stoichiometric  calculations,
physical  flow modeling,  a physical  inspection of the  boiler,  and a review of
existing  equipment arrangement drawings provided by the Soviet side.

As previously mentioned,  reburn system modification  analysis  was  limited to
the  front and rear  walls  of the  boiler;  the  side  walls  of the  unit  were
essentially inaccessible  due  to equipment  interferences.   The Ladyzhinskaya
boiler's  waterwalls  are made  up  of lower,  midwall,  and upper  waterwall
sections.   Distribution headers  are located between each  pair  of sections.
For  each  section  the waterwall tubes  were  arranged  in a  closely  packed "S"
shaped configuration.  To minimize  the  impact of a reburn system retrofit on
water circulation the  proposed  reburn fuel  and burnout air windbox locations
were  matched to  coincide  with these  "S"  shaped  tube   configurations.   An
example of this is shown in Figure 11 for the reburn  fuel injector system.

For  the initial reburn fuel and burnout air injector  locations, six locations
across the  front  and rear  walls  were selected,  primarily  to  enhance mixing
processes  and   avoid  extensive  waterwall   circuit modifications.    This
arrangement  is   illustrated in  Figure  12.   The  reburn  injector would  be
located at an elevation  of  66.6 ft (20.3 m).

Based upon  the  flow  modeling it  is recommended  that both the reburn  fuel
injector  and burnout  air   injector windboxes  be   multicompartmented  with
individual control  dampers  on each compartment.  This is because,  consistent
with the  physical flow  model  results,  one large  diameter jet and compartment
provided  for  better  deep  furnace penetration and dispersion while  smaller
diameter  angled  jets  and  compartments  provided  for  better  near  field jet
mixing  and  dispersion.   The individual  compartment control capability  is
recommended for  performance optimization during initial  system  start-up and
normal load-following boiler operation.

The  reburn fuel  injector windboxes were designed for 10%  FGR for  the reburn
fuel  transport  media,  20%  natural  gas  (percent  of  total fuel heat  input)
reburn fuel,  and a  nozzle  exit jet velocity of  146 ft/sec  (45  m/s) .   The
general reburn windbox  nozzle configuration is presented in Figure 13.  Each
windbox is segmented  into three vertical compartments.  FGR can be introduced
Into all  three compartments,  natural gas  is introduced through the upper and
lower  compartments,   with   the  center  compartment designed   for  future
capability of using  fuel oil  or pulverized coal  as reburn fuels.  The reburn
fuel  (natural gas)  would be premixed with the FGR  at the  nozzle exit with  a
natural gas  spud  located at the nozzle  tilt axis  centerline  at the  rear of
the  nozzle.   The  pant-leg  arrangement  is  clearly  shown in Figure 13.  Note
also that the bottom natural gas nozzle has a capability to yaw approximately
                                  3-132

-------
The  physical  flow  modeling  suggested  the  desirability of  vertical  tilt
capability for  the  reburn injectors over  a + 30  from horizontal  range,  so
this capability was  recommended.

The burnout  air windbox  and  compartment  arrangement was  similar.   However,
only two  compartments  were required  and  yaw adjustment  capability  has  been
added to  the  two  smaller upper compartment  nozzle  tips.   This configuration
is presented in Figure 14.

As  in  the case  of  the  reburn  fuel  injectors,  it is  desirable to have  12
injectors, 6 on the front wall  and 6 on the  rear  wall.   The injectors would
be  located  at  a  boiler  elevation  of  95.8  ft  (29.2  m) .   For  operational
flexibility, it was  recommended that both  the  upper and  lower  nozzles  have
independent yaw capability,  as  well as vertical  tilt capability (+  or  -30
from horizontal).

FURNACE THERMAL PERFORMANCE

The  objective  of  this  phase  of  the  preliminary  design  study  was  to
investigate  the potential  impact  of retrofitting  a  reburn  NO   reduction
system  on  furnace  performance  for  the   Ladyzhinskaya  Power  Station  host
boiler.  Objectives  were  to determine if  changes would  occur in:  the furnace
exit  gas   temperature,   the   furnace  hopper   gas  temperature,   and   the
distribution of and total heat absorption in the furnace with and without the
reburn NO  reduction system.
         x            J

Other objectives in the  analysis  were to estimate any  changes  in combustion
efficiency  and/or  flyash  carbon   content  when  the reburn system  was  in
operation.  No assessment was made of convective pass performance (superheat,
reheat, or economizer  sections) with  the  reburn system  due  to  any changes in
flue gas  weights  (mass flows),  total boiler heat  absorption (as  compared to
furnace  heat  absorption),  or  changes  in  boiler  exit  gas  temperature  (as
compared  to  furnace   exit  gas  temperatures)  in  this   thermal  performance
analysis.

The  furnace   thermal  performance   analysis  was  completed   utilizing  a
proprietary Combustion Engineering developed computer code (Figure 15).   The
function  of  the program was  to determine,  through a series of  heat balance
calculations,   the  heat   transfer   from   the  combustion  products   to   the
waterwalls,   the  corresponding  gas  temperatures,  and   the  furnace  outlet
temperature  of   the   combustion   products.   The   combustion   history   and
combustion products were determined based  on a  fuel analysis,  fuel  and air
mass flow rates and injection locations,  fuel particle size distribution, and
a set of fuel char combustion kinetics.

The  computer  program  was  first set  up  to  emulate  the  current  as-found or
baseline  conditions  using  information provided  by  the  Soviet side.   The
program was then calibrated for the Ladyzhinskaya unit based on a furnace gas
temperature profile  provided by the Soviets.

The furnace thermal performance with  reburning was then determined using the
design locations and  flow rates for the reburn fuel, recirculated  flue  gas,
and burnout air and the  revised main  burner fuel and air flows and the upper
furnace flue gas recirculation flow.
                                   3-133

-------
Both furnace  gas  temperature profiles and  furnace waterwall/heat absorption
profiles were generated as a result of the thermal analysis.  These are shown
in Figures 16 and 17, respectively.

Unit performance predictions based on the thermal analysis can be summarized
as follows:

•    Furnace  heat  absorption  will  be  up  to  4%  lower  when  natural  gas
     reburning is employed

•    Furnace  exit  gas temperature  (FEGT)  will be  up to  60  F  (35 C) higher
     when natural gas reburning is  employed

•    Furnace bottom  gas temperature will  be  up to 80 F (45 C) lower with the
     reburn system in operation

•    Overall   furnace   waterwall   heat  absorption   profiles   will  not  be
     significantly altered with reburning

•    Carbon  heat  loss  will decrease  with  natural  gas  reburn,  predicted
     carbon-in-flyash: 2.3% baseline,  0.6% natural gas reburn fuel

•    Overall  boiler efficiency  will  be  up to  1.0%  less with  natural  gas
     reburn due to moisture heat loss

Since  reburn  system operation  (with  natural gas) was projected to  increase
FEGT   by  60°F  (35°C),   the   Soviets  requested  that   additional  furnace
performance analyses be conducted  in  an attempt to decrease or eliminate this
FEGT  increase.   It was also  requested  that an analysis  be  conducted on the
effect of  the  upper  furnace  FGR nozzles  on FEGT with the goal of eliminating
these  additional  waterwall  penetrations.   The Soviet side  stated that,  for
this  supplemental analysis  and for  the  later  detailed  design,  if  it  was
demonstrated  advantageous,   it  may be  possible  to  alter the  burner (first
stage)  zone  excess  air quantity    For the  preliminary  reburn  system design
the  burner zone  excess  air level was  maintained  constant  both  with  and
without reburning to minimize potential effects on ash slag tapping.

In support of the  above,  a  sensitivity  analysis (using  the  CE proprietary
code  previously  described)  was  conducted to  assess  changes in  1)  FEGT,  2)
cumulative  furnace  heat  absorption,  and  3)  flyash  carbon   content  with
permutations  in  A)  total  FGR flow rate,  B)  upper furnace  FGR  flow rate,  C)
main  burner  zone  excess air  (stoichiometry) ,  D)  reburn  nozzle/zone FGR,  and
E)  reburn  fuel ratio.

Key  parameters varied during the  sensitivity study  were  main  burner excess
air,  total FGR rate,  upper  furnace FGR rate, FGR quantity introduced through
the reburn fuel injectors, and  total  quantity  of  natural  gas as  reburn fuel.
Table  2 summarizes the major conclusions reached  from the  sensitivity study.

It can be  seen that  decreasing  the  main burner excess air  was predicted to be
beneficial  in meeting   the  sensitivity  study  objectives  of   decreasing
 (lowering)  the furnace  exit gas   temperature  and increasing  (raising)  the
waterwall  absorption.  A  slight increase  was predicted for the  flyash-carbon
content.   Thus,  it  is  recommended that,  for  optimized  reburning,  the main
                                   3-134

-------
burner  excess  air  level  be  decreased to  the extent  possible while  still
maintaining conditions appropriate for slag removal.

The two  parameters  analyzed pertaining  to  the objective  of  eliminating the
upper  furnace  FGR  ports both  showed  that  this  elimination  would not  be
beneficial in  regard  to  furnace performance.  Eliminating  the  upper furnace
FGR ports by maintaining  the FGR mass flow but changing  its  location to the
reburn elevation resulted in a  predicted increase in FEGT  of 13  F (7 C) and
lower waterwall absorption  of 8x10   Btu/hr  (2x10   kcal/hr).   Eliminating the
upper  furnace  FGR  ports by decreasing  the  FGR flow  rate  resulted   in  a
predicted increase in FEGT of 29 F (16 C).   Therefore, it is recommended that
the upper furnace  FGR ports not be  eliminated as part of  the  reburn system
design;  in fact,  the  sensitivity study showed that  increasing  the amount of
upper furnace FGR reduced the FEGT.

The  sensitivity  analyses also  examined  the  possibility  of decreasing the
amount of FGR  used  with  the reburn fuel injectors by 2.5%  from 10% to  7.5%.
For the  sensitivity study,  the reburn  FGR  was decreased  with  an equivalent
increase  in  upper furnace  FGR.   From a furnace  performance  standpoint this
change was predicted  to  be  beneficial with a  predicted decrease  in FEGT and
an increase in waterwall absorption.

Finally,  the  sensitivity study examined the  possibility of  decreasing the
total  reburn  fuel  flow  rate.   This  was  done with  the assumption  that the
previously  recommended   decrease  in  main   burner  excess   air  will  be
incorporated into  the optimized  reburn  system design.   This  assumption was
necessary to maintain the reburn  zone stoichiometry in  the desired range for
NO  reduction.
  x

Decreasing the reburn fuel  ratio was beneficial with respect to both lowering
the FEGT  and raising  the  waterwall  absorption.  Thus,  it  is recommended that
the  reburn fuel  flow  rate be  decreased  for the  optimized  reburn  system
design.

Table  3  summarizes the process  flows and  the predicted  furnace  performance
results  for:

     1.   Baseline As Found Operation
     2.   Preliminary Reburn Operation
     3.   Optimized Preliminary Reburn Operation

As shown  in Table  3,  with the optimized preliminary reburn design conditions
the predicted  furnace exit  gas temperature was  18 F (10 C) lower  than the
predicted current baseline  as found during  operation.  Also the furnace  total
waterwall heat absorption was  essentially  equal  at 625 x  10   Btu/hr (157 x
10  kcal/hr).  And  finally,   the flyash carbon  content was predicted  to remain
at a very low  level.

FINAL SYSTEM DESIGN/PROJECT  STATUS

Agreement  has  recently  (November  1990)  been reached  on the  final reburn
system  design  based on  several technical  meetings  held  over  the past year
between  the  U.S.  and Soviet  sides.   The Soviet  side,  responsible for  final
detailed  design,  fabrication,  installation,  and  operation of the  system, has
                                   3-135

-------
agreed to  proceed  with the project, basing  the final design  largely on the
preliminary design provided by the U.S. side.

The Soviets have recommended certain deviations from the preliminary design.
These deviations  are highlighted  in Figure  18.   Most  of the  changes  were
recommended by the  Soviet  side in order to  minimize  retrofit  complexity and
cost.  It  has been jointly  agreed that  the  majority  of  these  deviations
between the  preliminary  design and  final  design will have  an insignificant
impact on  optimum system performance.   There are,  however, three significant
differences between  the  preliminary and final  designs.    First,  the Soviets
had recently  decided  to  retrofit  an aerodynamic "nose"  (Figure  18)  into the
Ladyzhinskaya boiler;  this  modification will be implemented to  improve  heat
transfer in the boiler's convection  section  during the same outage scheduled
for  the   reburn  system  installation.    The   preliminary  reburn  design  was
completed by  the U.S. side prior to notification of the nose modification; it
is the joint  opinion of both  the  U.S. and  Soviet sides  that the  nose  will
significantly affect  upper  furnace aerodynamics but  will likely not have  a
significant negative impact on reburn system effectiveness.

The  Soviets  also recommended  combining each of the  two  innermost  reburn and
burnout air windboxes into a single windbox,  in order to facilitate retrofit.
Both  sides  agree  that  this   configuration  shouldn't negatively  impact  on
reburn  system performance.   As  a  result,   there  will be five reburn  fuel
windboxes  on  both the  front and rear walls and  five  burnout air windboxes on
both  the  front  and  rear  walls,  as  opposed  to six  each  in the  preliminary
design.

The  Soviets have also recommended  that the variable yaw  and tilt capability
defined  in the preliminary design  be  eliminated;  i.e.,  the  nozzles  would
operate  with fixed  tilt  and  yaw  vertical and  horizontal  angles.   This
decision was  taken by  the Soviet side  to minimize  system complexity and cost
as well as to maintain the  project  retrofit  schedule.  Variable  tilt and yaw
in   the  U.S.  side's  experience   enhances   the   field   installed  system's
flexibility  in  optimizing  reburn  fuel and  burnout  air  mixing  in  the  bulk
furnace  gases,  allowing the  system to  be   tuned  in the  field  in  order  to
optimize NO   reduction.   Discussions between the U.S. and Soviet  sides  have
been held  to  jointly agree on the fixed yaw and tilt angles; the decision has
been made  to  fix all horizontal yaw angles  at 0  and the vertical tilt angles
at both  the  reburn fuel and burnout air locations at -15°  from horizontal.
Both sides believe that these fixed angles  offer the best opportunity for the
Ladyzhinskaya installation  to  meet  its reburn  system  performance  objectives
in lieu of the availability of a variable angle nozzle arrangement.

The current schedule  calls  for the  Soviet side  to  complete detail design and
fabrication drawings  by April  1991.   The Soviets plan to  install  the reburn
system during a 120-day outage which is  scheduled  to  start in  May 1991.   The
U.S.  and  Soviet   sides  will  then  jointly  plan  a   test program   for  the
Ladyzhinskaya power  station.   The  U.S.  will  then  support  the Soviets  in
quantifying reburn  system performance;  these tests are  to occur  between the
fourth quarter 1991 and the second quarter  1992.
                                  3-136

-------
REFERENCES

1.   J. Kramlich, T. Lester, J. Wendt, (1987), "Mechanisms of Fixed
     Nitrogen Reduction in  Pulverized  Coal  Flames,"  Proceedings:   1987 Joint
     Symposium on  Stationary  Source Combustion  NO   Control, Volume  2,  EPA-
     600/9-88-026b (NTIS PB89-139703).            X

2.   C. Kruger,  G.  Haussmann,  S.  Krewson,   (1987),  "The Interplay  Between
     Chemistry and  Fluid Mechanics  in the Oxidation  of Fuel Nitrogen  from
     Pulverized  Coal,"   Proceedings:   1987   Joint  Symposium  on  Stationary
     Source  Combustion  NO    Control,  Volume   2,   EPA-600/9-88-026b  (NTIS
     PB89-139703).         x

3.   M. Toqan, J.  Tears,  J  Beer,  L.  Radak,  A.  Weir,  (1987),  "Reduction of
     NO  by Fuel  Staging,"  Proceedings:   1987 Joint  Symposium  on Stationary
     Source  Combustion  NO    Control,  Volume   2,   EPA-600/9-88-026b  (NTIS
     PB89-139703).         X

4.   J.  Freihaut,  W.  Proscia,  D.  Seery,  (1987),  "Fuel  Bound  Nitrogen
     Evolution During the Devolatilization  and Pyrolysis  of  Coals of Varying
     Rank,"   Proceedings:    1987  Joint   Symposium  on  Stationary   Source
     Combustion NO  Control, Volume 2, EPA-600/9-88-026b (NTIS PB89-139703).

5.   Y. Takahashi,   et  al.   (1982),   "Development  of  'MACT'  In-Furnace  NO
     Removal  Process  for Steam Generators,"  Proceedings of the  1982  Joint
     Symposium   on   Stationary    Combustion   NO     Control,    Volume    I,
     EPA-600/9-85-022a (NTIS PB85-235604).         X

6.   R. Borio,  R.  LaFlesh, R.  Lewis, R.  Hall, R.  Lott,   A.  Kokkinos,  S.
     Durrani,  "Reburn Technology  for Boiler NO  Control," Sixth  Annual  Coal
     Preparation,   Utilization,   and  Environmental   Control   Contractors
     Conference,  August 6-9, 1990, Pittsburgh, Pa.

7.   D. Anderson,  J.  Bianca,   J.  McGowan,  (1986),  "Recent   Developments  in
     Physical Flow Modeling of Utility Scale  Furnace,"  Industrial Combustion
     Technologies, American Society for Metals.
                                   3-137

-------
                                Figure 1: Ladyzhmskaya Power Station
   Diverting Portions of the Fuel and Combustion Air Streams from
   the Main Burner(s) for Injection into the Post Flame Gases
  Mechanistic Model for
     NOx Destruction
NOx Formation Inhibited
Due to Fuel Rich Conditions in
Reburn Zone: NOx Destruction
is Promoted Due to Secondary
Flame Radical Attack on NO
Produced in Primary Zone to
Form Molecular Nitrogen
 Staging Air-
Reburn Fuel-
    Primary
   Fuel-Air"
       Hypothesized NOx Destruction Mechanism
             CH
                   HCN
                        OH.H
                                    OH,H
       Figure 2: Basic Reburn Process Description
                                                               Figure 3: Ladyzhmskaya Flow Model
                                                 3-138

-------
       I      I
      mm
                  TOP LEVEL OF NOZZLES
                     SECTION "AA"

                 BOTTOM LEVEL OF NOZZLES
                      SECTION "BB"
       II   II   II  II   I I  II
                 COMBINED REBURN NOZZLE
                       ASSEMBLY
                    FRONT ELEVATION
   Figure 4: Schematics of Model Reburn Fuel/
             Burnout Air Injector Bank
                                                      Gas  Recirculation
                                                          Nozzles
 Figure 5: Jet Penetration - Reburn Fuel Injectors
Gas Recirculation
    Nozzles
Figure 6: Jet Penetration/Dispersion - Reburn Fuel
                     Injectors
Gas Recirculation
    Nozzles
                                                          Figure 7: Effect of Injection Downtilt on Jet
                                                                   Penetration/Dispersion
                                                 3-139

-------
     Gas  Recirculation
         Nozzles

        Burnoul Air
         Injectors
        Reburn  Fuel
        Injectors
•no	a	o-
                          Main Burner Zone
   Figure 8: Lateral Jet Dispersion - Reburn Fuel
                      Injectors (Middle Nozzles)
                           Gas  Reclrculatlon
                               Nozzles

                              Burnout Air
                               Injectors
Reburn Fuel
 Injectors
                                                E3   -BED	E3	El--
                                                                              Main Burner Zone
                         Figure 9: Lateral Jet Dispersion - Reburn Fuel
                                            Injectors (Sidewall  Nozzles)
Gas  Recirculation
    Nozzles
                                                     Figure 11: Example of Tube Modification for Reburn
                                                                    Fuel Injector Installation
 Figure 10: Jet Penetration - Burnout Air Injectors
                                                 3-140

-------
                                      Burnout Air
                                    / Injectors
                                      Reburn Fuel
                                        Injectors
Figure 12: Preliminary Reburn System Design
                                                                                    Upper Compartment

                                                                                      [FGR » CH.)
Middle Compartment

  (FGR)

  (FUTURE OIL OR COAL)
   Lower Compartment

     (FGR + CH,)
                                                         Figure 13: Reburn Fuel Injector Windbox
                                                                       Arrangement
                                                                Upper Compartment
                                                                Lower Compartment
                               Figure 14: Burnout Air Injector Windbox
                                            Arrangement
                                               3-141

-------
Inputs
Fuel Information
• Particle Size Distribution (R)
• Apparent Density (Pf)
« Chemical Characteristics
• Ash Characteristics
Drop Tube Furnace
System Information
• Char Activation Energy (E)
• Char Frequency Factor (A)
• Fuel Swelling Factor (a)
e Fuel Volatile Matter (VM DTFS)
Boiler Information
• Design Parameters
• Operating Conditions
l\
/lathematical
Model
Proprietary
Computer
Code
-
Outputs
• Temperature /Time
History
• Overall Fuel
Combustion Efficiency
• Percent Carbon in
Flyash
• Percent Carbon
Heat Loss
• Heat Release/Heat
Absorption Profile
Figure 15: Flow Diagram for Boiler Combustion Performance Model Simulation
     Horizontal Furnace
          Outlet Plane

             Upper GR


           Burnout Air



              Reburn



          Main Burner
           Main Burner
Gas Reburn
Predicted
w/o Reburn
Soviet
Measurements
                   1500
                                  2000            2500
                                Furnace Gas Temperature (F°)
                     Figure 16: Predicted Furnace Gas Temperature
                                        Profile
                                (("F -32) x 5/9 = °C)
                                                                 3000
        Horizontal Furnace
             Outlet Plane
     Gas Reburn
     Predicted
     w/o Reburn
              Main Burner  —
              Main Burner  —
                      20000   30000  40000  50000   60000   70000
                             Waterwall Heat Absorbtion Rate (Btu/Hr/Ft2)
                         Figure 17: Waterwall Heat Absorption Profile

                            (Btu/hr/ft2 + 13273.0  = Cal/sec/cm2)
              80000
                                 3-142

-------
35.1m -
29.2m -
         Tilt
0
20.3m — <3±25°   £> --
          Tilt
            <£ Burner

            <£ Burner
                         -FGR Nozzles (6)
   Tertiary Air
• (Burnout) Nozzles
 (6 Front, 6 Rear)
          Reburn Fuel and
           FGR Injectors
          (6 Front, 6 Rear)
                                           31.0m —
              -15°.
26.8m • ••-15'  Fixed
        Fixed
                  20.3m
                    12.00m
    8.75m
            >Main Coal Burners
            (8 Front, 8 Rear)
        .-15°  -15°.
        Fixed  Fixed
                                      (£ Burner
                                        Burner
                                                                      FGR Nozzles (5)
i	27.5m
   Burnout Air
    Nozzles
 (5 Front, 5 Rear)
                                       -*	20.3m
                                        Reburn Fuel and
                                         FGR Injectors
                                        (5 Front, 5 Rear)
                                                               12.00m
                                              • 8.75m
                                               >Maln Coal Burners
                                               (8 Front, 8 Rear)
5.90m
                                            5.90m
U.S. Side Preliminary Proposal               Final Design Arrangement

                             Figure 18: Design Arrangements
                    Reburn Heat  Input (20%  of Total)  (Ib/hr)   24314
                    Percent by Weight FGR  (%)                   14.7
                    Flue Gas Recirculation  (Ib/hr)            410407
                    Gas Residence Time  (Zone  2)  (sec)           1.00
   Volume Flow  (ft3/sec)  Wet
      Gases Out  of  Reburn Zone

   Reburn Zone  Stoichiometry
   Ash Flow Rate  (Ib/hr)
   Gas Compositions at Outlet
      of Reburn  Zone
    (% by Volume Dry)
                    Gas Compositions at Outlet
                      of Reburn Zone
                    (% by Volume Dry)
                                                           CO,  =
                                           SOZ =
                                           H20 =
                                           CH,, =
                                           CO, =
                                                           S02
                                                           H20
                                                                 47084
                                         0.97
                                       102965
                                        17.20
                                         0.00
                                        82.22
                                         0.286
                                         0.00
                                         0.29
                                        17.11
                                         0.00
                                        81.78
                                         0.285
                                         0.53
                                         0.29
                    Table 1: Reburn Zone Process  Calculations
                                         3-143

-------
       Parameter Changed


Decreasing Main Burner Excess
Air 20% to 5%
Changing Upper FGR Elevation
(35.1m) to Reburn Elevation (20.3m)
Decreasing Total FGR by 3.2%
by Eliminating Upper FGR
Decreasing Reburn FGR by 2.5%
by Increasing Upper Furnace FGR
Increasing Total FGR by 3% by
Increasing Upper FGR
Decreasing Total Reburn  Fuel by 8%
with Main Burner Excess Air @ 5%
Furnace
Exit Gas
Temperature
(°F)
Lowered 20
Raised 13
Raised 29
Lowered 10
Lowered 20
Lowered 21
Furnace
Waterwall
Absorption
(10s BTU/hr)
Raised 6
Lowered 8
Raised 2
Raised 6
Lowered 2
Raised 9
Carbon
In
Flyash
(%)
Raises 0.4
Raises 0.1
No Effect
No Effect
No Effect
No Effect
                 Table 2:  Furnace Performance Sensitivity Analyses
                             ((°F -32) x 5/9 = °C)
                              (Btu/hr)/4 = kcal/hr
                         70

                         %
Performance Variables
Reburn Fuel Ratio
Total Excess Air
Burner Zone Excess Air %
Total FGR              %
Reburn FGR            %
Upper Furnace FGR     %
Furnace Exit Gas     °F(°C)
Temperature
Furnace Heat Absorption
x106 Btu/hr (x106 kcal/hr)
Flyash Carbon Content  %
Baseline
as Found
N.A.
20
20
18
N.A.
13.2
1967(1075)
Preliminary
Reburn Case
20
20
20
18
10
3.2
2028(1109)
Optimum
Reburn Case
12
20
5
21
7.5
8.7
1949(1065)
                                 626(158)
                                    2.3
606(153)
   0.6
625(157)
   1.2
                     Table 3: Furnace Performance Summary
                                  3-144

-------
           Session 4A
COMBUSTION NOX DEVELOPMENTS I
 Chair: W. Linak and D. Drehmel, EPA

-------
AN ADVANCED LOW-NOX COMBUSTION
 SYSTEM  FOR  GAS AND OIL  FIRING
        R.A. Lisauskas
        C.A. Penterson
   Riley Stoker Corporation
   Worcester, Massachusetts

-------
                          AN ADVANCED  LOW-NOX  COMBUSTION

                           SYSTEM FOR GAS AND OIL FIRING

                                  R. A. Lisauskas
                                  C. A. Penterson

                             Riley  Stoker  Corporation
                             Worcester,  Massachusetts
ABSTRACT
A new low-NOx combustion system for gas and oil-fired industrial  and utility
boilers is discussed.  The system consists of an advanced Riley low-NOx STS burner
used in conjunction with overfire air and recirculated flue qas.   One of the
distinctive features of the low-NOx STS burner is the use
to form a separation layer between the primary and secondary
                                gas.
                              of recirculated
                                 f 1 ame
                                                                          flue gas
                                                                   zones.
This advanced
boilers in Western
              low-NOx combustion
                   Europe.
results are summarized for two recent retrofit
are presented for both gas and oil-firing.   NO.
have been demonstrated on a natural  gas-fired <
tion of this advanced system to  U.  S. gas  and
discussed.
     system has been implemented on several  power
Combustion system modifications and emission test
                                               applications.   Field emission data
                                                emission levels of less than 50 ppm
                                               20 MWe utility boiler.   The applica-
                                               oil  wall-fired boilers  is also
INTRODUCTION


Environmental  concern over power plant stack emissions has grown steadily over the
past decade.  In spite of this concern,  the 1980's saw little change in U. S. NOX
regulations.  However, recent passage of new federal  Clean Air amendments and
proposed new state regulations make it likely that U.  S.  industry will  soon be
required to meet revised emission standards on both new and existing boilers.

Unlike the United States, Europe and Japan did impose new emission regulations
during the 1980's.  In 1984,  German legislators recommended stringent NOX emissions
standards for new and existing boilers(l).  These standards defined new emission
limits for all  large combustion systems  firing gas, oil and coal.  As a result,
large numbers of industrial  and utility  boilers in Germany and other European
countries have  been retrofitted with low-NOx systems.   We believe this recent
European low-NOx retrofit experience is  of  particular interest to the U. S. power
industry.  This paper focuses on some of this experience  applied to gas and oil
fired systems.
                                       4A-1

-------
Deutsche Babcock, the parent company of Riley Stoker Corporation, has had
considerable experience in supplying combustion systems to meet the demands of
German and European air pollution codes.  Low-N0x combustion systems have been
implemented by Deutsche Babcock on a wide variety of industrial and utility
boilers.  Since 1984, Deutsche Babcock has supplied low-NOx systems to over 110
liquid and gas-fired boilers.  More than 520 low-NOx burners have been retrofitted
to a variety of boiler configurations.  In order to meet stringent emission limits
many of these retrofit applications incorporate combustion modification
techniques, such as flue gas recirculation and overfire air, in combination with
new low-NOx burners(£).  NCL reductions of over 80% have been demonstrated with
these new systems.  New fuel injectors have also been developed in response to the
changing quality of heavy fuel oils.  This technology and experience is now
available to the U. S. power industry through Riley Stoker.

One new burner system   the Swirl Tertiary Separation (STS) burner   is
particularly well suited to U. S. wall-fired boiler retrofit applications.  This
new burner system is the subject of this paper.  In addition to presenting
operating results from recent European retrofit installations, we will also discuss
the application of this new combustion system to U. S. boiler design
configurations.

DESCRIPTION OF LOW-NOX BURNER SYSTEM

New STS burner systems have been recently retrofitted on gas and oil wall-fired
boilers in both Germany and Sweden.  In addition to reducing NOX, the burners were
designed to both minimize boiler pressure part changes and maintain acceptable
combustion conditions.

Figure  1 is an illustration of the STS burner equipped with swirl control.  As is
typical in many European boiler designs, combustion air is controlled individually
to each burner.  A spiral box, or scroll (shown in Figure 1) is used to supply the
combustion air to the  burner.  The scroll is divided between primary and  secondary
air passages with control dampers and flow metering installed immediately upstream.
Total air flow to the  burner is divided between the primary and secondary air
passages.  The exact distribution of primary and secondary air can be adjusted
depending on the level of internal burner staging required for NOX control and
overall combustion performance.

The ability to independently control swirl imparted to the primary and secondary
air streams provides great flexibility in controlling flame length and shape.  It
also  ensures flame stability under low-NOx firing conditions.  Adjustable air vanes
within  the scroll are  used to control the degree of swirl and subsequent  fuel air
mixing.  Between these two swirling air streams a separate recirculated flue gas
stream  can be  introduced forming a distinct separation layer between the  primary
and secondary  air.

The introduction of  this separating layer of inert flue gas acts to delay the
combustion process and reduces NOX in the following manner:

    •      Peak flame temperatures, particularly on the surface of
           the  primary  combustion zone, are reduced by a surrounding
           blanket of inert flue gas.

    •     The  rapid  mixing of secondary air is prevented; thereby,
          reducing the oxygen concentration in the primary combustion zone.
                                        4A-2

-------
Unlike flue gas mixed with the primary or secondary air streams, the flue gas
separation stream is unswirled and concentrated.  This serves to delay secondary
air mixing until  after first stage oxygen has been consumed and the flame has
cooled.  The intent of the separation layer, therefore, is to control both thermal
NOX formation and NOX produced from nitrogen contained in the fuel.


Additional NOX reduction is achieved through staged combustion.  A portion of the
total combustion air can be introduced through overfire air ports above the burners
to provide external air staging.  This overfire air is controlled and metered
independently of the combustion air to the burners.  Low-N0x burners combined with
flue gas recirculation and overfire air offer an integrated approach for maximizing
the reduction of NOX emissions on gas as well as oil firing.

As shown in Figure 1, oil is burned using a centrally located steam or mechanically
atomized oil gun.  Natural gas is burned using spuds or canes located within the
primary core of the burner.


FIELD RESULTS


Arzberg Power Station


Low-N0x STS burners have been installed at Arzberg Power Station Unit NO. 6 in
Arzberg, West Germany.  The boiler, shown in Figure 2, is a once-through Benson
boiler rated at 1.58 million Ibs steam per hour and generates 220 MW of
electricity.  The  unit is currently equipped to fire natural gas or light NO. 2
oil.   In  1988, the boiler was retrofitted with sixteen low-NOx  burners, each rated
at 153 million Btu/hr heat input.  Burners are arranged horizontally for opposed
firing on  four levels.  As stated earlier, the STS burner design was selected to
fit within existing burner openings.

NOX emission limits for this retrofit project were 50 ppm*  for  natural gas firing
and 75 ppm for light oil.  The retrofit combustion system was designed with the
flexibility of introducing recirculated flue gas through either the burner zone
separation annulus or having it mixed directly with the combustion air to the
burners.   One tertiary air port was also installed in close proximity to each
burner but was later found to be ineffective for NOX control.   A level of overfire
air ports  was added on both front and rear waterwalfs above the burner array for
staged or  off-stoichiometric firing.  As shown schematically in Figure 3, all flows
including  primary, secondary, tertiary and recirculated flue gas were independently
controlled and metered.

Prior  to  the retrofit, NOX emissions from natural gas firing averaged 300
ppm.   Testing was  conducted following the retrofit to optimize  the operation and  to
commission the boiler.  Figure 4 illustrates the effect of  mixing  flue gas
recirculation into the combustion air on NOX emissions for  natural gas firing.  NOX
is reduced with  increasing amounts of flue gas  recirculation (FGR) flow.  With  20%
FGR and 10% OFA  flow, NOX emissions were reduced to 75 ppm.  By increasing  the
amount of  recirculated ffue gas to 30%, NOY decreased to 50 ppm.
   All NOX and CO concentrations  are  dry  and  referenced  to  3% 02-
                                        4A-3

-------
Additional  testing was then performed to evaluate the effect of introducing FGR
flow through the burner annul us for NCL control.  The total amount of FGR flow
remained at 30% with 10% OFA.  Figure 5 illustrates the effect of introducing
increasing percentages of FGR flow through the burner annulus or separation layer.
Uhen more than 50% of the total FGR flow was introduced through the separation
layer (the- remaining amount being mixed in with the combustion air) NOX decreased
significantly.  The lowest measured NOX emission approached 25 ppm when nearly all
of the FGR flow was passing through the burner annulus.  CO emissions remained less
than the 15 ppm throughout this testing and flame stability or scannability was not
a problem.

A limited amount of testing was performed on NO. 2 fuel oil.  Data were collected
while operating at 15% FGR and 15% OFA flow rates.  NOX emissions of 75 ppm were
achieved at full load and decreased to approximately 60 ppm at 50% boiler load.  CO
emissions remained below 25 ppm for all test conditions.


Va'rtan Power Station

An advanced STS burner system has also been retrofitted at the Va'rtan Power Station
in Stockholm, Sweden.  The Va'rtan unit, commissioned in 1976, is rated at 250 MW.
It is a once-through Benson style boiler designed for heavy oil firing.  As shown
in Figure 6, the burners are mounted on a single wall in a 4 X 4 array.  Each
burner is supplied individually with air and is equipped with a Deutsche Babcock
oil pressure/steam pressure atomizer.  In addition to STS burners, the retrofit
combustion system includes both OFA and FGR.  The existing FGR system was modified
to supply flue gas to each burner as well as the lower furnace.

The post-retrofit NOX guarantee limit for the Va'rtan unit is 0.27 lb/10^ Btu or
approximately 210 ppm.  NOX emissions measured during recent commissioning tests
are shown in Figure 7.  Emission levels (at high load) for the new system are 30 to
40% lower than the guarantee value.  The data spread is due to differences in
operating conditions and varying fuel oil nitrogen content.  Average fuel oil
nitrogen content is 0.3%.  During the recent tests, high load excess oxygen
measured 1.3-1.4% upstream of the air heater corresponding to an excess air level
of less than 7%.  CO emissions were less than 40 ppm.  These results were achieved
with  10-11% OFA and 15% FGR.  Approximately one third of the flue gas was
introduced through the burners.  The remaining flue gas was introduced to the lower
furnace for steam temperature control.

APPLICATION TO U. S. BOILERS


The STS burner design has been adapted by Riley Stoker to U.S. wall-fired boiler
firing systems.  Contrary to the European practice of individual burner air
supplies, 1). S. wall-fired boilers are equipped with common windbox/multiple burner
arrangements.  Because of this, the burner inlet scroll, described in Figure 1, has
been  replaced by primary and secondary air swirl vane registers surrounded by flow
control shrouds.  All other burner components remain the same.  As shown in Figure
8, the movable shrouds are operated by single actuators and can be automated with
boiler load.  The shrouds control the primary/secondary air flow split
independently of swirl vane position.  Flow measurement devices are positioned
between the burner barrels to provide a relative flow indication between the
burners.
                                        4A-4

-------
A prototype 85 million Btu/hr STS burner designed for windbox applications (Figure
8) is currently being tested in Riley Stoker's large pilot combustion test facility
located at the Riley Research Center in Worcester, Massachusetts.  This facility is
designed to simulate the near field combustion conditions of full scale
furnaces(3).  Test variables include firing rate, flow biasing ratios, the amount
of flue gas recirculation and injection method, level of burner staging, swirl
setting, excess air and oil/gun positions.  The test program has several
objectives:  (1) to fully characterize the burner's low-NOx capability under U. S.
boiler operating conditions, and (2) to evaluate the sensitivity and trade-off of
various burner adjustments on NOX control and other combustion operating parameters
such as flame shape and particulate emissions.  The prototype burner is being
tested on natural  gas and NO. 6 fuel oil.  The fuel oil selected for the test
program is a 2% sulfur oil with an asphaltene content of approximately 10%.  Test
results will be available within the next several months.


SUMMARY


Advanced STS burners have been successfully retrofitted on several gas and oil
fired power boilers in Western Europe.  These retrofits have been achieved within
existing burner openings.  STS burners in combination with overfire air and flue
gas recirculation have exceeded their emission goals.  NOX levels of less than 0.06
                                                    .     X
       Btu on natural gas and less than 0.2 lb/10^ Btu on heavy oil have been
demonstrated.

The introduction of a separate flue gas stream, or dividing layer through the
burner throat has been shown to be effective in reducing NOX on natural gas.
Additional testing is required to evaluate the effectiveness of this separation
layer during heavy oil combustion.

STS burner designs have been developed for U. S. wall -fired boiler burner/windbox
arrangements.  Prototype burner tests are being carried out to ensure that European
experience is duplicated under U. S. boiler operating conditions.


REFERENCES


(1)  P.W. Dacey, "An Overview of International NOX Control Regulations,"
     Proceedings:  1985 Symposium on Stationary Combustion NOX Control, Vol. 1,
     EPRI CS-4360, January 1986.

(2)  R.  Oppenberg, "Primary Measures Reducing NQX Levels on Oil- and Gas-Fired
     Water Tube Boilers," Conference of the Association of German Engineers,
     Duisberg,  FGR,  September 26, 1986.

(3)  R.  Lisauskas, et al_., "Experimental  Investigation of Retrofit Low-N0x
     Combustion Systems," Proceedings:  1985 Symposium on Stationary Combustion
     NO  Control, Vol. 1, EPRI CS-4060, January 1986.
                                        4A-5

-------
            Primary  air      Secondary  air
   Fu« I  oil
   a ~t o m i z e n
F u « L  o I I
    EL
                Nat ura I
                   QQ9
                             FIue  gas
       Figure 1.  Low-N0x STS  Burner  Equipped  for  Gas  and Oil  Firing and Individual Air Supply

-------
Figure 2.   Arzberg Power Plant Unit NO.  6
                   4A-7

-------
CO
                 A Sta
ck
                                               F.D.fan

                                               Steam/air
                                               preheater
               FLue gas/air  preheater
               Furnace
                              Vort ex burner
FLue  gas
   fan
                                                                                    F.D.fan
                                                                              FLue  gas/air  preheaier
                 a.   Original  System
                                                  b.  Low-N0x Retrofit System
                                  Figure 3.   Low-N0x  Combustion System at Arzberg Power Station

-------
                    80
                 Csl
                 o
                    40
-p..
>
CO
                 O-
                 CL
                    20
                          OverfireAir K)X
                       15            20            25

                                   Flue Gas Recirculation

                                Rate into Combustion Air, %
30
                        80
                                                                                        60
                     CsJ
                     O
                        40
                      E
                      Q_
                      Q-
                      >T
                     O
                                                                                        20
Overfire Air

Flue Gas Red
                                      culotion 30%
       25
50
75
                                       Flue Gas Percentage

                                     in the Separation Flow, %
                    Figure  4.   NOX  as a  Function of  FGR into
                                the  Combustion  Air -
                                Natural Gas Operation
                      Figure  5.   Nox  as a Function  of FGR  into
                                  the  Annulus  -
                                  Natural  Gas  Operation

-------
                 Figure  6.   Vartan  Power Station
        150
   eg   100
   o
        50
0.8        1.0        1.2         1.4         1.6


                 Steam Flow, million Ib/hr
                                                              1.8
Figure  7.   NOX Versus Boiler  Load Post-Retrofit Heavy Oil  Firing




                                4 A-10

-------
                                         Shroud Actuators
Burner Front  Plate
F.G.R. Inlet
Connection
 Register Turning  Vanes
                                                                          Primary Air Shroud
                                                                                      Secondary Air Shroud
                                                                                                 Furnace Woterwall
                                                                                               Oil  Gun With Diffuser
                                                                                               Cos  Cones
             Figure  8.  Riley Low-N0x STS Burner  (Model  90) for Windbox Burner Arrangements

-------
  NOX REDUCTION AND CONTROL
USING AN EXPERT SYSTEM ADVISOR

      G. Michael  Trivett
   Monenco Consultants Ltd.
       Calgary, Alberta
           T2P 3W3

-------
             NOx  REDUCTION AND CONTROL USING AN EXPERT SYSTEM ADVISOR
ABSTRACT

In a  continuing  effort to  reduce emissions  of  NOX from  their coal  fired  power
units, TransAlta  Utilities  has  undertaken a program of combustion optimization for
low-NOx operation.   In addition  to testing  and tuning  these  units,  Monenco  is
developing  an on-line  Expert System  to enable operators  to  continuously maintain
low-NOx emissions.

Characterization  and  optimization  tests for low-NOx operation were conducted at the
Sundance Generating  Station on  the 375 MWe tangentially fired Unit  #6.   The tests
produced an extensive database which will  be  incorporated  into  the  Expert System.
The tests  confirmed  that a  reduction  of NOX emission of 5 to 15% could be achieved
by improved control  procedures.

The Expert  System advisor  will  incorporate real-time  input  from sensors  such  as
oxygen analysers,  temperature indicators,   NOX  analysers,  CO  analysers,
burner/pulverizer status,  etc.   A graphical  computer  interface will  show current
readings  and a  message board  will  provide  recommended corrective actions  to
minimize NOX emissions.
                                       4 A-15

-------
                                   INTRODUCTION
NOX emissions are  one  of the precursors  of acid rain  for  which coal fired  power
plants  are  a significant contributor-    During tests conducted  by Monenco at
TransAHa Utilities Sundance Plant Unit #6, modest NOX  reductions of  between  5 and
15 percent from  baseline  levels  were achieved.   Reduction of  NOX  emissions  on  a
continuous basis requires considerable diligence  by  the operator in  balancing the
competing requirements of  steam  temperature control,  ash  slagging conditions  and
low-NOx  operations.   An expert  system can  be  developed to  review  the  available
operating data and by use of a knowledge base make recommendations  to the operator
to minimize NOX  emissions while maintaining optimum unit performance.

For a moderate NOX reduction of 10%, the expected development costs for  the  expert
system represent a NOX avoided cost of  $60/tonne  based  on a  35 year  operating life.
By  comparison a  Selective  Catalytic  Reactor  (SCR)  with an  80% NOX  reduction
efficiency represents a cost of over $2000/tonne.  Therefore the cost  effectiveness
of developing the NOX control  expert system (NOXPERT) which  could be applied  to all
of TransAlta's coal fired units,  appears  very attractive.

The potential benefit  of the application of an  expert  system  to Sundance Unit #6
would be  a  reduction in  NOX emissions from  existing  levels at  a reasonable  cost.
On  an  annual  basis this  amounts  to  a  reduction  of   between  210  to  630  tonnes
N0x/year  from an  estimated  baseline rate  of 4200 tonnes N0x/year.  In addition to
the total  NOX reduction, a  proportional  decrease in N0£  should result  which may
reduce the contribution to the  brown plume effect.

The general  approach to  be taken using an  expert system would  be  to minimize NOX
formation  by combustion  modification or  in-furnace  techniques.   Typical  factors
that may  be modified to reduce NOX  emissions include reduced excess air  operation,
increased fuel fineness,  increased  furnace wall sootblowing frequency and fuel/air
staging in the combustion zone.
                                        4 A-16

-------
Each of the above factors can be optimized.  However the interrelationship  between
factors requires some degree of  compromise  to achieve a  reasonable  low-NOx emission
with acceptable unit performance.  These interrelationships were developed  as part
of the NOX testing  program  and  the results of the  tests  form the majority of the
support data  for the expert  system  development.

Expert systems  have been developed  as  an extension of research  in  the  field of
artificial  intelligence (AI).  An expert  system is a computer  program  that combines
concepts, procedures and  techniques  derived  from AI.   These  techniques allow the
design  and development  of  computer systems  that use  knowledge and  inference
techniques to analyze and solve  problems  in a way similar to human  reasoning.

This specific  application of  an  on-line  or real-time  expert  system to control NOX
emissions is novel.   The application of an expert  system  to  real-time control is
becoming more  advanced  with chemical process companies applying  systems  to wider
and  wider  functions.    However,  the real-time  aspect, particularly with future
closed loop  control action,  remains as leading edge  technology.   Expert system
"shells" or development tools are available from  only a  short  list  of potential
suppliers.    Consequently  the  development  of the  NOX  control  expert  system
represents leading  edge  development  in  both  hardware/software and steam generator
control applications.
                          NOX EMISSION CHARACTERIZATION
TransAlta  Utilities  Sundance  Plant has  a total  gross  generating capacity of
2100 HWe.  Units I and 2 are each 300 MWe and Units  3,  4,  5  and 6 are each 375  HWe.
All  6  units  are  Combustion  Engineering  design with  forced circulation utilizing
tangential  firing with  a  dual  furnace arrangement.   Units  3, 4, 5 and  6 are  also
equipped with manual  tilting overfire air nozzles.   These  units are supplied with a
low  sulphur  sub-bituminous  "C"  coal  from  the  nearby Highvale Mine.  Unit  #6, in
particular,  is equipped  with  5 pulverizers,  feeding coal  to  5 elevations of
standard  C-E  tangentially  fired coal  nozzles  on  eight corners.    The  burner
arrangement is non-Low NOX except for the overfire air  nozzles which are positioned
within the  burner  windbox setting.

The  NOX tests were  conducted on  Unit #6  at Sundance  and  provided detailed
information on the relationships between  NOX emissions and the control  actions.  A
                                       4 A-17

-------
test  matrix  was  arranged  to evaluate  the response on  NOX  emissions  by the
following:

     •    Burner tilt position;
     •    Fuel-air settings;
     0    Auxiliary air settings;
     •    Overfire air tilt position;
     0    Increased coal fineness;
     0    Reduced excess oxygen; and
     0    Continuous operations.

The  relationships  between  these control  variables  plus  furnace  cleanliness  or
sootblowing frequency  and  unit load are depicted in Figure 1.   Fuel NOX, which is
dependent on  fuel  nitrogen content, is shown with a dashed  line since its control
requires  primarily fuel switching to a lower nitrogen content coal or combustion at
sub-stoichiometric conditions.  Thermal NOX can be reduced by adjusting the control
parameters, essentially reducing the flame temperature.

A change  in one of these  control  parameters may reduce NOX,  however,  it may also
result  in an  unwanted  change, for example, to performance.   The objective then is
to strike an  acceptable balance between NOX reduction and overall unit performance.

Some  guidelines were set regarding  acceptable performance for the NOX tests.   These
 included  the  requirements  not  to: jeopardize safety of operations; increase the fly
 ash carbon  content above  0.7% by weight;  and  increase  the  average carbon monoxide
 emissions  above  50 ppm by volume.  Also normal  steam  outlet  temperatures from the
 superheater and reheater would be maintained within their acceptable ranges.

 One of  the major objectives of the NOX test  program was to determine the baseline
 or existing emissions  on  a continuous  basis.   An initial test series was  conducted
 on a twenty-four  hour basis  over  four  days.   All  operating  conditions were set
 as-is and  in automatic  mode  where  required.   The  data  from  these  tests are
 presented in  Figure 2 and  show the high,  low  and  mean for  each twenty-four hour
 period.   The  descriptor on  the x-axis  represents both baseline and series  G  results
 for days  1  to 4.   The  series  G data, or optimized settings following the  parametric
 test  matrix,  are  also  shown for a direct before  and after comparison.

The overall  reduction  in  NOX  emissions achieved was between  5 and  15 percent.   In
addition  the  peak NOX emissions were  reduced  in the  final  test series as compared
to the  baseline levels.   The target settings for excess  oxygen, burner  tilts and
                                        4 A-18

-------
furnace cleanliness  were achieved  on  Day 3  of  the  final  test  series.   Consequently,
this test day produced  the lowest  mean  NOX  emissions  and the  lowest peak  values.

Excess oxygen  (02)  has an  impact  on NOX  emissions  as  shown  in  Figure  2.    Large
ranges of  excess  oxygen  during the  tests  contributed  to the large  range  in NOX
emissions,  and  the  high  oxygen  peaks  directly contributed  to the  high NOX  peak
values.  A reduction in  excess oxygen  during  the  optimized test series  as seen  in
Figure 2 aided in  reducing the mean NOX emissions.

The  effect of  burner  tilt  position,  either positive or  negative  degrees  from
horizontal, has an effect on NOX emissions.   Figure 3 shows  the high  and low  range
for burner tilts for both the baseline data and series  G  data.  The test days  with
a large  range  in  tilt  position,  had a  corresponding  large range in  NOX emissions
and the highest peak levels.

As  noted  previously,  changes  to  some  control parameters  may  result  in  an
undesirable  effect  such  as  reduced  performance.    A decrease  in  performance  is
typically  indicated  by  two  factors:  the amount of unburned  carbon  in the fly ash
and the  amount  of carbon monoxide in the  flue gas.   Figure  4 shows  the  effect  of
increased  fineness  and  excess oxygen on NOX  emission and the resulting   carbon  in
ash content.   The data represented by test number A was  conducted essentially  with
baseline coal  fineness and  fixed  burner  tilts at  the  horizontal position.   The
remaining  data  points,  B,  C, D  and  E were tested with  an increased  coal fineness
and for  tests C, D  and  E decreasing  excess 02.   As shown, an  increase in fineness
results  in a reduction  in the carbon in ash  content.   By  increasing  fineness  with
normal excess oxygen,  an expected  increase  in  NOX  resulted.   However,  the increased
fineness also allows the  excess oxygen  to  be  reduced, reducing NOX emissions.   As
shown in Figure 4, as  excess oxygen is  reduced, carbon in ash increases.   From this
information  an optimum  excess  oxygen  level,  without  degrading performance  as
indicated by the carbon in ash level, was  determined  for  these tests  at  a level  of
2.5%  62-   Reduced excess  oxygen  also  has the  benefit  of  increased thermal
efficiency due to  reduced heat loss,  reduced fan power and draft  losses.

The changes made to pulverizer/classifier  settings to  achieve a finer  coal  grind
were  minor and consequently had little  effect  on  the stack  opacity.   Greater
changes in fineness, however, may  impact the stack opacity and consequently  opacity
data should be included for the expert  system  to review.

The overall  effects on  performance  between   the  baseline data  and  the  optimized
settings data  as  reflected  in the carbon in fly ash  levels and carbon monoxide
                                        4 A-19

-------
emissions are  shown in  Figure 5.    Since  the  fineness  in the  baseline or  as-is
condition  was  considered  to  be  out of  specification,  increased  fineness  was
expected to give an improvement in unburned carbon  levels  with  lower  CO emissions.
As noted the increased fineness, as reflected  in the reduced carbon content,  allows
the excess oxygen to be reduced without degrading performance.

Overfire air  (OFA),  which  is  considered  to be  a  primary NOX control  technology,
allows  separation  of  the  fuel   and combustion  air  resulting 1n  lower  flame
temperatures and  lower thermal NOX  formation.   The angle that  the overfire  air
enters  the  furnace, relative  to  the  horizontal  position, was tested  and as  the
angle approaches the horizontal position, or closer  relative to the flame, the  NOX
emissions  increase.   Consequently  the  greatest separation  angle,  or  30 degrees
above horizontal,  results  in  the  lowest NOX  levels.   In  addition  this OFA  tilt
position should remain fixed regardless of the  automatic  positioning  of the  burner
tilts.

The combination  of  both controllable  and  uncontrollable  parameters in  coal  fired
steam generator  operation results in  a  dynamic, complex  system  for NOX  control.
Changes  in  burner tilt  position in response to  steam temperature  control  demands,
furnace  wall  slagging, burner  to burner  coal  flow imbalances  and variations  in
excess  oxygen  distribution  from furnace to furnace  result in fluctuations in  NOX
emissions.

The amount  of  change  or range  in  the control  parameters  noted gives a comparable
range in the resulting NOX  emissions.  Figures  2 and 3 show this  effect for  excess
oxygen  and  burner tilt position, respectively.   In the optimized  test series  G  the
movement  of burner tilts  was  limited to  within +  or  -  10  degrees with manual
adjustments and  overall both the  tilt range and excess oxygen ranges are  smaller.
This resulted  in a  reduced  peak for  NOX emissions and an  overall  average reduction
between the baseline and series G emissions.

The NOX control expert system will  review each of the noted parameters and based on
desired settings   and  expected  relationships, should   produce a  reduction  in
emissions similar to the difference between the baseline and optimized test data.
                            INPUT DATA FOR THE NOXPERT
The  input  data  required for the  NOX control  expert  system will  be  supplied  from
three sources.   The  first  source will be the flue  gas analysis data.   The second

                                        4A-20

-------
will  be panel board data with steam generator  condition  information,  and the third
source will  be the operator/user.

The input data for the flue gas  analysis will  include the following:

     •    Nitrogen Oxides (NOX);
     •    Carbon Monoxide (CO);
     •    Excess Oxygen (02); and
     •    Stack opacity.

The oxygen and  carbon  monoxide  monitors,  two  in each of the two  ducts  leaving the
economizer,  should  provide sufficient data  for averaging  and  fault  diagnosis.
Suspect data  could  be detected  if  outside  the normal   expected  range  and
recalibration might be suggested by the expert advisor.   Also,  the opacity monitors
will  be  used  to detect high excursions  possibly  caused  by too fine  a  coal  grind.
The units currently are not equipped with NOX  analysers  and separate  monitors would
be required for each of the dual furnaces.

The panel  board will  provide input to  determine  what the steam  generator  current
operating conditions are corresponding  to the flue gas analysis.   These data will
include the following:

     •    Boiler/Turbine Load (MWe or steam flow);
     •    Burner tilt position (+  or - from horizontal);
     •    Superheater/Reheater outlet steam temperatures;
     •    Windbox to furnace pressure differential;
     •    Fuel Air damper position (% open for all 5 levels);
     •    Auxiliary Air damper position (% open for intermediate levels);
     •    Bottom air damper position (% open);
     •    Top air damper position  (% open);
     •    Overfire air damper position (OFA %  open);  and
     t    Burner level  status,  mill  coal feeder status (on/off).

The first two sources of data will  be on-line  or  real-time at  a specific recording
frequency.   The third  source of information will  be  input by  the operator/user as
required by the expert  system.  This data will include  information  available only
at infrequent periods such  as the  following:

     •    overfire air tilt position (manual control, degrees  from horizontal);
     •    carbon content of the  fly ash;
                                        4A-21

-------
     •    sootblower status and frequency;
     •    coal  fineness  for all  5 mills;  and
     •    instrument calibration information.

In addition,  the  operator/user will  respond to  specific enquiries  made by  the
expert system for updates on the available data.   For example  time  limits  could be
set (time-stamping) for  validity periods  after which updates  should  be made.

The real-time data will  also be recorded  at a  specified  frequency so as to  maintain
the latest information and status available.  This  frequency,  however,  is  expected
to be in the order  of minutes  for  updates  due to the large dead times  as  a  result
of the  boiler  system thermal  inertia.  As a  result,  the rules  for  low-NOx  firing
can be reviewed and acted upon over a reasonably long  time period.

The  input  data mentioned  above  is  depicted  as  a  schematic diagram  showing
communication  with  the  expert system  in Figure 6.    The  input  data  plus  the
recommended control  action  to  minimize NOX emissions  would be echoed  on-screen as
shown in Figure 7.
                                  LOW NOX RULES
The NOX control expert system is currently being  developed  by Monenco for TransAlta
for application on  Unit #6 at the Sundance  Plant.   The development  plan includes
various stages from prototype to a fully fielded  and implemented advisory system 1n
a one year time frame.  Once fielded,  a closed loop  control  scheme will  be reviewed
for operation within a distributed control  system (DCS).

The following  details  some of the specific concerns in developing this  system  and
outlines some generic features of an  expert system.

An expert system comprises three parts:

     •    facts;
     •    rules; and
     •    inference engine.

The facts  describe aspects  of the domain,  for  example,  that  the  furnace  excess
oxygen  is  at  3.2% by  volume or  that the  top  elevation  of  burners are  out-of-
                                        4A-22

-------
service.   Rules describe  what an  expert might  do with  the facts  to reach  the
objective,  for example reduce the oxygen  to  reduce the NOX emission.   In  the same
way in which a human operator infers a solution to a problem based on the available
facts and previous  experience with a similar  problem,  the inference engine combines
rules and facts in  the knowledge base to reach a conclusion.

By definition rules  are  English-like sentences used for defining  the  knowledge  of
an expert.   Rules can  be grouped  as subsets  which can  be  activated  or deactivated
independently.  The  intent  is to control which  rules  are used (fired)  when  a  new
fact is introduced  (asserted) into the knowledge base.

The  facts that  describe  the domain  are  determined on-line from the various  input
sources such as the excess oxygen sensors.   Once  the  facts are known by the expert
system, given constraints such as time validity,  then the  rules can  be searched  to
determine an appropriate  decision  or recommendation.

Rules  are  defined  using an  IF-THEN syntax   that logically  connects  one  or  more
antecedent (or premise)  clause with  one or more  consequent (or conclusion) clause.
A rule  says that if the  antecedents are true, then the consequents  are also  true.
The  antecedents and consequents of rules refer to specific facts  that  describe  the
state of the  domain.   Therefore each fact describes some  particular aspect of  the
domain's state.  Together,  the rules and facts make up the knowledge base.

The  inference  engine analyzes the  rules  and facts for  any rule  antecedents  that
match existing facts.  The process   of matching  is finding a rule clause  with  the
same pattern  of words  (in  the same  order) as  a  fact  in the  knowledge  base.   When
all of the antecedent clauses in a rule have a  corresponding  fact in the knowledge
base, the inference engine can assert the consequent of the corresponding rule into
the knowledge base  as a new fact.

The  inference engine  consists of a  generalized  computer program  that  knows  about
reasoning strategies and  various ways to combine rules and facts,  but knows nothing
about  any  particular  application.   The  knowledge  base of rules  and  facts  is
nonprocedural, while the inference  engine is  highly  procedural.   In  other words,
rules  and   facts  represent  what  the  knowledge  is,  but  the  inference  engine
determines  how that knowledge should be analyzed.

The  following examples illustrate rules which  would  comprise the knowledge  base.
The order or sequence of  the rules is not important since the inference engine will
use or fire  the rule that satisfies  the current facts.
                                        4A-23

-------
    •    If ppm NOX is greater than NOX Limit
            ppm CO is less than CO Limit
            Excess 02 is greater than 2.5%
            Load  is 100% MCR
         Then reduce Excess 02 to 2.5%

    •    If ppm NOX is greater than NOX Limit
            ppm CO is greater than CO Limit
            C in  Ash is greater than C Limit
            Any Fineness less than 65% thru 200 mesh
            Any Fineness greater than 1.5% on 50 mesh
         Then  Recommend  Increase in that  Fineness  to 65% thru  200  mesh and less
         than 1.5% on 50 mesh

     •     If ppm NOX is greater than NOX Limit
            ppm CO is less than CO Limit
            Excess 02 is 2.5%
            C in  Ash is less than C in Ash Limit
            Burner tilt position less than 0 degree
         Then Sequence wall sootblowers

Additional  rules  have  been developed to set  initial  conditions,  such as equipment
availability  (i.e. Overfire Air), optimal  excess oxygen  as  a function of steaming
rate  required  and  hierarchical  rules  to  dictate  priorities such  as  steam
temperature control or  carbon monoxide  limit override.   The  rules for  the NOX
control  aspect  of  the  NOX PERT are invoked only under steady load cases  and during
transitional periods will be overridden in order to maintain  safe  operations.

The development of the rules,  based on the example decision tree  shown in Figure 8,
consists of logical  If-Then clauses  satisfying  the  various branches.    Once the
complexity of  the rules  are  fully defined,  the expert  system "Shell"  will  be
selected and a  prototype NOX PERT will be assembled.
                         CONCLUSIONS AND RECOMMENDATIONS
There are  three  major techniques available  for the reduction  of nitrogen  oxides
(N0x).   They are  classified as:
                                        4A-24

-------
     •    operational  control,
     •    combustion  control, and
     •    post-combustion  control.

In each case there is an associated economic cost with a corresponding increase in
equipment complexity.   The least expensive cost is to alter operational control of
existing burner equipment by means  of the subject expert  system.   The next least
expensive is a retrofit of  the burner equipment to  the  low-NOx style of burners.
And  finally,  the most  expensive control  alternative  is  by  post-combustion
techniques such as a  selective  catalytic reactor.

The costs  associated  with each technique  and  the level  of NOX reduction offered
ranges from an estimated $60/tonne of NOX removed for an Expert System, $150/tonne
for  low-NOx burners  with overfire air  and  $2,000/tonne for  an  SCR.   Utility
companies in Germany  who have installed SCR's also recommend that NOX be reduced by
combustion  first to  minimize   the cost  of an  SCR both  in  capital  and  annual
operating costs.

Based on the NOX test results from Sundance Unit  #6,  manipulating some key controls
can reduce NOX emissions by  up  to  15 percent from baseline  emissions.  By uniformly
applying rules for maintaining low-NOx operations,   from operator  shift  to shift,
an overall reduction  on both monthly and annual bases would  be expected.   Conse-
quently, an expert system  that  advises the operator using a consistent set of rules
could provide a NOX reduction up to and possibly  surpassing the  15 percent from the
Sundance tests.

Based on these findings,  an  expert system to reduce NOX  emissions  from a utility
coal-fired steam generator represents a least  cost  approach  with good probability
of success.    Further,  a  successfully fielded expert  system  could be  offered
commercially to other Utilities with similarly designed and equipped units.
                                 ACKNOWLEDGEMENTS
The author would  like to acknowledge  the support offered  by TransAlta Utilities
Corporation   for  funding  the  work  reported  here.   Also special  thanks  to
Richard  Bane,  Malcolm  McDonald  and M1ke Blakely of TransAlta for technical guidance
through  the various  stages of this work.
                                       4A-25

-------
Figure 1.  NOx Emission Contributing Factors
ouu-

250-

CM~ 200-
0
5?
(3)
~ 150-
D.
CL.
X
0
z 100-

50-


0-








-


'.



_



^^






!.



^









r~


u


~^~- — ^_^^
^""~^

~






-^__










J

^^








NOX






















n


\







LU LJ











J

O2






"^^^









-H

^^





n

y
T
^
^^












[


"


















[




IU
-9

-8
-7

-6

-5


-4


-3
-2


-1

BS-1 BS-2 BS-3 BS-4 G-1 G-2 G-3 G-4
Baseline
Test Name Ootlmlzed
                                                    CM


                                                    O
                                                    c
                                                    0)
                                                    D)

                                                    S1
                                                    o
                                                    x
                                                    LU
   Figure 2.  NOx versus Excess Oxygen
                  4A-26

-------
  300-
  250-
   200-
co

©


Q.
Q_
  150-
  100-
   50-
                               NOX
                                         Burner Tilt
        BS-1
                 BS-2

               Baseline
BS-3
BS-4      G-1

  Test Name
G-2     G-3

   Optimized
                 Figure 3.  NOx versus Burner Tilt Position
Q.
_Q.

X
O
                                                                       100
                                                                      -80
                                                                       60  2

                                                                           £
                                                                           D)
                                                                           
-------
   500-

   450-
   400-
c\j
O
*  350-
E  300-
a

.g  250-

o  200-
 c
 o
_a
O
   150-

   100-
    50---
                      617 ppm
         BS-1
                                                   C in Ash
BS-2
Baseline
                          BS-3
BS-4     G-1
  Test Name
                                                             -r
G-2      G-3
    Optimized
                                                                      G-4
                                                         -rO.9

                                                          -0.8

                                                          -0.7


                                                          -0.6

                                                          -0.5

                                                          -0.4

                                                          -0.3

                                                          -0.2

                                                         .-0.1
                                                                                .c
                                                                                O
                  Figure 5.   Carbon Monoxide and %C in Ash
                      Figure 6.  Input Data for NOxPERT
                                     4A-28

-------
r
CD
                          NOx Control Expert System
                              Reheater T= Normal
                             Superheater T=Normal
                             OF A Tilt= + 30 Deg
                             Burner Tilt= -10 Deg
                               Coal Fineness
Most Recent Data:
  Jan 17,1991
Burner Status
Level A= ON
Level B= ON
Level C= ON
Level D= ON
Level E= OFF
Fuel Air Pos'n
10098
10098
10098
10095
5098
co =
12 ppm
                                                                                               Message Board
                                                           Date: Jan 31, 1991
                                                           Time: 09 :45:33
                                                           Recommended Actions
NOX Level Higher Than 110 ng/j
* Suggest Reduce 02 by 0.6 98
* Verify Carbon in Ash= 0.698
* If NOX Doesn't Lower-
  Then check coal fineness
                                                     Figure 7.  Graphical Display for NOxPERT

-------
CO
o
                                          Up Ex02
                                                                       N. n«w    |




                                                                       '^'  J
L.4 - 3<«4V
-
CinfiSH >
CinfiSH
1*4.4 =
jicady
-
PPM NOX >
NOX Limii
-
Ex02 >
Ex02 3<


-------
NOX PAPER FOR THE 1991 EPRI/EPA JOINT SYMPOSIUM ON A STATIONARY
            COMBUSTION NOX  CONTROL  MARCH  25-28,  1991
                       THE  CAPITAL  HILTON
                        WASHINGTON D.C.
                        JANUARY  17,  1991

          AN R&D  EVALUATION OF  LOW-NOX  OIL/GAS  BURNERS
            FOR SALEM  HARBOR AND BRAYTON  POINT  UNITS

                         Rui F. Afonso
                         Nino M.  Molino
               New England Power Service Company
                    Westboro, Massachusetts

                        John J.  Marshall
                    Riley Stoker Corporation
                   Worchester,  Massachusetts

-------
                 NOX  Paper  for  the  1991  EPRI/EPA  Joint  Symposium
                      on Stationary Combustion NOX Control
                                March 25-28,1991
                               The Capital Hilton
                                Washington, D.C.
                                 January 17,1991
                  AN R&D EVALUATION OF LOW-NOX OIL/GAS BURNERS
                    FOR SALEM HARBOR AND BRAYTON POINT UNITS
                                  Rui  F.  Afonso
                                 Nino M. Molino
                        New England  Power Service  Company
                             Westboro,  Massachusetts

                                John J. Marshall
                            Riley Stoker Corporation
                            Worcester, Massachusetts
                                    ABSTRACT



A thorough R&D program  to  evaluate  low NOX,  high efficiency oil/gas burners was

developed and  conducted by  New  England Power  Service  Co.   In  anticipation of

retrofitting  Brayton  Point  and   Salem Harbor  Units 4,  the  burner  evaluation

project  involved  a  series  of  evaluations  designed to  progressively  identify

burner  technologies  most  likely   to  meet   New   England  Power  performance

requirements.     Detailed  characterization  of  atomization  quality   using  an

Aerometrics   Phase/Doppler   particle  analyzer  was  performed  for  five  of the

initial eight candidate burners.   Pilot scale combustion tests at  Riley Research

test  facility,  including  oil, gas  and  dual  (oil/gas)   firing  conditions,  were

conducted for  the  baseline and  final  candidate  burners.   Retrofit  and  cost

impact  studies  were  conducted  for  the  selected  burners to  ensure  the  most

efficient/economic application.   This paper focuses  on  combustion  test  results

at the Riley Research test  facility.
                                      4A-33

-------
                                  INTRODUCTION

New England Power Service  Company  (NEPSCo)  and Riley Stoker  Corporation  (Riley)
evaluated state of  the  art oil  and gas  burners  for New England  Power  Company's
Salem Harbor,  Unit 4 (Salem Harbor)  and Brayton Point, Unit 4 (Brayton  Point).

As originally  designed  by Riley for  cycling/intermediate  load operation,  these
units typically  operate with a  capacity factor  between  45% and  65%.    Each  of
these wall  fired  units  produce  3,250,000 Ib/hr  of  steam  at 955°F  superheat  and
reheat, generating  about  440  MWe net.   Figure 1  shows a cross  sectional  view  of
the  units.    Twenty  four rear  wall  mounted  burners  fire  into  a  pressurized
furnace.  Figure 2  illustrates the general windbox  arrangement.   Two  half depth,
vertical  platen  walls  divide   the  furnace  into   three   cells.     Horizontal,
drainable  superheater  and reheater  surfaces  allow  for  timely  start  up  and
shutdown in cycling duty.

The  existing   Rodenhuis  and  Verloop  TTL7  burners   are  nominally  rated  at  200
million  BTU/hr.   Designed before 1970,  they represented burner technology prior
to the  1971 New Source  Performance Standards  (pre-NSPS).   Currently,  NOX  control
regulations  do not  apply at Salem  Harbor.    Brayton Point,  designed in  1969,
adheres  to  a  0.3 Ib/million  BTU (~234ppm)  NOX   emission  limit  imposed  by  the
Massachusetts  Department  of  Environmental  Protection.   Bias firing and Flue  Gas
Recirculation  (FGR)  enable  Brayton  Point   to  achieve this  level  of  control.
Particulate  emission limits  of 0.12  and  0.05  Ib/million  BTU  apply  to  Salem
Harbor  and Brayton  Point, respectively,  and are  achieved  with the  installed
electrostatic  precipitators.   Salem Harbor  operates  near  10% excess air  with a
boiler  efficiency  of  about  87%.   Brayton  Point  uses up  to  15%  FGR with  10%
excess  air.

Performance  goals for  the new  burners  include  0.3 Ib/MBTU  and  0.2  Ib/MBTU  of
NOX,  respectively,  for oil and  gas  firing  at 5%  excess  air,  while  maintaining
carbon  in  the  flyash  (oil firing)  to  less than  20%.   This  will  reduce  total
particulate  emissions  and improve  precipitator  performance.   Typically,  both
units average  30-50% carbon in the flyash.

A  Rodenhuis  & Verloop TTL5 burner served  as the baseline burner for  testing  in
the  Combustion Burner  Test Facility (CBTF) at the  Riley Stoker Research  Center.
This  80 million  BTU/hr  burner  is  similar  to  the 200  million  BTU/hr  burners
installed  at  both Salem Harbor  and  Brayton Point.   Field  data  from Salem Harbor
                                       4A-34

-------
was compared  to the baseline  burner test results  in the  CBTF.   This  analysis
provides  guidelines for  predicting  the  potential   field  performance  for  the
candidate burners tested in the CBTF.

Comparative  data  illustrates   improvements  achieved  with  the  latest   design.
Additionally,  results   are  provided  for  natural  gas  firing  and  simultaneous
firing of both natural  gas and residual fuel  oil in various ratios.
THE R&D PROJECT

Previous  papers  (1)  describe  the  background  and  technical  approach  to  the
selection of the candidate burners for Brayton Point and Salem Harbor.

In summary, the program consisted of five primary tasks:
    •    Evaluation of vendor proposals
    •    Evaluation of atomization performance
    •    Baseline burner combustion tests
    •    Candidate burners combustion tests
    •    Engineering and economic evaluation of burner retrofits

The  evaluation  process  started  with  the  submittal   of  proposals  by  eight
participating   manufacturers,   and   evolved   through   detailed    atomization
performance testing  (five  candidates), to  pilot  scale  combustion tests  for  the
baseline and two final candidates.

Atomization  testing  was   performed  using  surrogate  fluids,  air/water  and
air/water-glycol  to  simulate steam/oil  at  various  viscosities  and  operating
conditions.     Spray  droplet   size  distributions   (spatial   and   temporal),
mass/volume fluxes  and  atomization efficiency  were  compared  for  each  burner.
Review  of  these results combined  with the  initial  proposal  evaluation,  led  to
the selection of the two burners for combustion testing.

This  paper focuses  on  a  comparison  of  the  results   of the  baseline  burner
(R&V TTL5) and  a  new generation Rodenhuis  & Verloop  TTL 22.5 burner capable of
residual fuel  oil and natural gas firing.
                                       4A-35

-------
PERFORMANCE CRITERIA

NEPSCo established  the  following  full  load  performance  goals and  operational
characteristics for candidate burners:

Full Load Performance

         NO., emissions less than 0.3 lb/106 Btu for 0.5% wt. nitrogen fuel;
         0.2 Ib/Mbtu for natural gas
    •    Particulate emissions, less than 0.05 lb/106 Btu after the
         precipitator (precipitator efficiency = 60%)
    •    CO emissions, less than 100 ppm
    •    Flue gas oxygen content, less than 1% by volume (5% excess air)

Operational Characteristics

    •    Turndown, 6:1 to 10:1
    •    Dual fluid atomizer using air or steam
    •    Fuel oil pressure, less than 300 psig at the gun inlet
    •    Fuel oil viscosity at the burner 100-150 SSL)
    •    Fuel flow:  21 gpm #6 oil, 220 KCFH natural gas

In  addition  to  the  performance criteria,  the following concerns were considered
for the evaluation:

    •    Operability and reliability
    •    Mechanical design
    •    Materials of construction
    •    Cost
    •    Related experience of the various burner technologies


DESCRIPTION OF TEST BURNERS

Baseline Test Burner  - TTL5:   Over twenty years ago  the  Rodenhuis  & Verloop B.V.
company  of  Holland  developed the TTL5 burner.   Compared to the typical  utility
or  industrial burners using pressure  or  steam  assisted  atomizers the TTL5 burner
was unique  in  that low  pressure  primary air, at  a  nominal 35"WG,  atomizes tne
                                      4A-36

-------
oil.   No registers are available to control or shut off the secondary  combustion
air flow to individual burners.

The TTL5,  rated  at  80 million BTU/hr,  simulates  the  combustion aerodynamics  of
the geometrically similar 200 million  BTU/hr  TTL7 installed at Salem  Harbor  and
Brayton Point.   Figure 3 illustrates  the TTL5 burner  arrangement installed  in
the CBTF test  rig.   Further details of the atomizing  bullet  assembly are  shown
in Figure  4.  The spiral  atomizer  head and the  two stages  (one  adjustable)  of
swirling primary air are indicated.

Primary air  flow is  about  7% of  the  total  combustion  air flow  at  full  load.
During turndown the primary air flow stays essentially constant.   In the  case  of
the Brayton  Point  and Salem Harbor installations primary  air  is  taken from  the
upstream side  of the air heater.   Booster fans distribute air to the burners.
The first  stage of  primary air supports and boosts  the  initial  swirl  of  oil
issuing from the spiral atomizer to  produce a conical  shaped  thin film sheet  of
oil.    The second  stage  primary  air  rotates in the  opposite direction.    The
interaction  between  these  counter-rotating flow fields reduces the  oil   film  to
fine droplets  needed  for  combustion.   An  adjusting rod  allows the second  stage
primary air  swirl to  be changed.   This is  done  by  varying the second stage  air
proportions through  radial  and  tangential slots.   Both  extremes  of 100% radial
to 100%  tangential  can be  achieved.   At  full  load oil pressures  operate  in  a
range of 50 to 80 psig depending on oil type  and viscosity.

The flame  stabilizer  is a conical  annular  diffuser ring  mounted  on the  primary
air tube.  Just  upstream  of the  flame  stabilizer an annular ring with a  venturi
shape at the  exit,  helps  to distribute the air  uniformly  to  the  diffuser.    The
entire  assembly penetrates just  less  than   halfway  into  the  venturi  shaped
refractory throat.

Advanced Burner  -  TTL/MG22.5:   Figure 5  depicts  the  new generation Rodenhuis  &
Verloop TTL/MG22.5 selected  for  installation  in  the CBTF  test  rig.  Rated  at 80
million BTU/hr,  the TTL/MG22.5  is  similar in  design and combustion  aerodynamics
to the 200 million BTU/hr TTL/MG50 proposed for field retrofit.

The advanced burner  incorporates many  design  changes  over the  baseline TTL5.   A
sliding  shroud  register  allows  secondary  combustion  air   flow  control  and
                                      4A-37

-------
shutoff.   No  swirl  of secondary  air  occurs  at the  register  of this radial  air
entry,  axial  flow burner.

The  atomizer  bullet  assembly  retains  the   operating  parameters   and   design
philosophy of the  TTL5  burner  but  new  hardware  designs enhance  and  improve
atomization.   Two  stages of primary air  surround  the  spiral  oil  atomizer.   The
assembly  shown  in  Figure 5  illustrates  the  design  without  an adjustable  slide
for changing  the second  stage  swirl.   The assembly is available with or  without
this  capability.     In   this   program,   testing  included  a  temporary  slide
arrangement to evaluate the impact of the additional hardware.

A  multivane   diffuser  replaces the  conical  bluffbody  flame   stabilizer  of  the
TTL5.   This  changes the combustion aerodynamic  patterns  for  mixing  the  primary
air/oil flows with  secondary combustion air.   The  vaned diffuser provides a flow
that  is   counter  in rotation  to  the   second  stage swirl  of  primary  air.    The
interaction  of  these  flows  increases the  degree  of  mixing  between streams.
Improved  flame  stability  and   turndown  characteristics  provide   operational
advantages at low excess air operation.

The  refractory  throat area  increases about  15% over  the  TTL5.    Shape  of  the
throat changes  to  smaller entrance and exit angles  with  a longer axial  section
(Figure  5).    The  combination of  increased  throat  area and  a  5% excess  air
operating level reduces  the secondary air velocity nearly 20%.

Natural gas  enters the  burner through an annul us created  by a concentric tube
surrounding the  primary  air  tube.   A  set  of  nozzles distributes the  natural  gas
at  the  inlet  to the flame stabilizer.  Pilot  holes  discharge a small  amount of
fuel at the hub of  the stabilizer directly to the primary flame zone.
 TEST  FUEL CHARACTERISTICS

 To maintain  a  relatively  consistent  quality of residual fuel oil  a  storage  tank
 at  Brayton  Point was set  aside  to  hold fuel  throughout  the test program.   The
 tank  is rigid  roof  with  heating  but no  mixing  equipment.   Though capable  of
 storing  over  two  hundred thousand  barrels,  less  than  forty  thousand  barrels
 (1,680,000  gallons)  were  in  the tank  at  the start  of the  program.   Over  the
 course  of 18 months  in the test  program about  120,000  gallons were fired.   Truck
                                      4A-38

-------
shipments of  6000  to 8000  gallons  were made  from Brayton  Point  to the  10,000
gallon heated underground storage tank at Riley.

Each truck delivery was sampled as part of a quality plan. This assured that  any
significant changes in fuel  quality  would  not  complicate the evaluation of test
results.  Some minor  variability was expected  due  to  the  design  of the storage
tank, lack of mixing  equipment,  interrupted  heating,  the long period of testing
and  the  relatively small  size  of a delivery.   Table  1 lists  the typical fuel
analysis  obtained  from  truck  deliveries.   This  fuel  is   consistent  with   the
residual fuel oil fired at Salem Harbor and Brayton Point.


RESIDUAL OIL COMBUSTION TESTS

Baseline Burner And Performance Targets

The  Combustion  Burner Test  Facility  (CBTF)   at  Riley  accommodated  all   burner
testing  described  here.    Design details  of  this  nominally rated  100 million
BTU/hr coal, oil  and gas fired facility appear  in other  publications (1,2).   The
results  of  the baseline  Rodenhuis & Verloop TTL5  burner testing  in January  and
February  of  1989 have already  been  presented   (1).  Additional  baseline  burner
testing  in  April  1990 expanded  and verified  the  earlier  results.  One  of  the
objectives of the baseline test was to establish correlations to field data from
Salem Harbor (3,4,5).

Due  to  various burner adjustments,  two  stable flame  characteristics were found
for  the TTL5.   A long  narrow  flame  and a   shorter  wide  flaring  flame were
obtained  by  changing the   location  of  the   spiral   oil   atomizer along with
adjustments to the primary  first and  second  stage air  ratio.   The long  narrow
flame is  not  acceptable  as  representative of  field operations  due  to  potential
flame impingement.   To  establish the baseline correlation  only  the  TTL5 wide
flame test results were used.

Currently,  at  full  load  oil  firing  and  less  than  5%  excess air,  Brayton  Point
NOX  levels must not exceed 0.3  Ib/million BTU  (-234 ppm  NOX  at 3%  Oxygen).  This
is  achieved  in  practice  by  utilizing  flue gas  recirculation  (FGR) through  the
TTL7  burners along  with  bias  firing.    Salem Harbor  is  not  subject  to  NOV
                                                                                A
regulatory limits.
                                      4A-39

-------
The performance goal  for carbon content  in  flyash is 20%  or  less.    Full  scale
field  testing  indicates  a  direct  relationship  between  flyash   carbon  and
asphaltene content of the  residual  fuel  oil  (5).   Currently at Salem Harbor,  an
oil/water emulsification system helps  control flyash carbon content.

Figures  6  and 7 respectively  show  the  correlations  for  NOX and carbon  content
reductions  desired  in  the  CBTF and  field results.   Burner  area  heat  release
(BAHR) correlates the  NOX  data for the CBTF  and  the  field.  This technique  has
been  successfully  applied  in  previous  investigations  for  the  Environmental
Protection Agency (6).  Asphaltene  levels in  the  residual fuel  oil correlate  the
carbon content results  in the CBTF  and the field.

Figure 6  illustrates  the  analysis  used  to develop potential NOX  targets  for  new
burners  applied  in  Salem Harbor or Brayton Point.  The  Salem Harbor  field data
is  scattered  at  about 350 ppm NOX  at a  BAHR  of 350,000 to  363,000  BTU/hr-sqft.
Results  from  baseline  burner  tests  in the CBTF range  from 210 to 230 ppm NOX  at
a  nearly constant BAHR  of 75,000   BTU/hr-sqft.   Line LI connects  the CBTF  and
field data  for similar  burner operating and flame  conditions and  establishes  the
methodology for projecting NOX emissions.

To  determine  the target NOX levels  in the CBTF line L2 is drawn parallel  to line
LI  from  the field requirement of 0.3 ID/million  BTU  (234 ppm NOX).   The values
of  line  L2  represents a NOX reduction of  about 35%.   Accordingly, the NOX target
in  the  CBTF  is  104  ppm  NOX.    Ideally the  need  for FGR,  bias  firing  and
emulsification will  be  reduced  or  eliminated  with the new  burner, while  meeting
the NOX  emission goals.

Figure  7 shows the  analysis  to develop  carbon  loss  reduction  targets  for  new
burners  applied  in  the field  and  firing  neat oil.   During the  field  test with
asphaltene  levels in  the  oil  at  6.1% and 14.7% the carbon  content in the flyash
was 21.7%  and   34%   respectively   (5).     Considering carbon,  ash  and  sulfate
contents  in the flyash  and  the  amount  of carbon and ash  in  the as-fired fuel
oil, the  carbon loss  was 0.037% and 0.084% respectively.  For  the same amount of
ash and  sulfates in  the flyash,  the flyash  carbon  goal of 20%, translates  to
carbon loss values of 0.033%  for the 6.1% asphaltene  oil  and 0.04%  for the 14.1%
asphaltene  oil.   Figure 7  shows  the actual  field  results and  the equivalent
points    needed    to    achieve     20%    carbon    content    in   the    flyash.
                                      4A-40

-------
This analysis  indicates  reduction requirements  in  carbon loss  from  11% to 52%
for  the  range  of  oils  tested   in   the   field.    The   following  description
exemplifies this procedure:   during TTL5  burner tests  the asphaltene content of
the as-fired fuel  was  11.2%;  from Figure 7 the  target  reduction in carbon loss
at this level of asphaltenes  is about  40%;   at full load conditions in the CBTF
the TTL5 burner produced  a  carbon  loss value  of 1.84%;  therefore, 40% reduction
from this level equates  to  1.10%  carbon  loss.   The actual percent reduction for
new burners  will  depend  on the as  fired  content of asphaltenes  at  the  time of
the test.
TTL5 And TTL/MG 22.5 Performance

A number of  independent  parameters  were varied during testing:  burner hardware
settings, fuel  oil  viscosity,  excess air, overfire air  and  load turndown.   The
burner  hardware adjustments  included  variations  in  the  relative  location and
settings of burner components.  Items such as  atomizer and diffuser position and
primary air staging ratios were varied from their design points.

In comparing the NOX performance of the TTL5 with the TTL/MG22.5, excess air and
overfire air provided greater impact than viscosity.  Viscosity  varied from 21.0
to 31.5 cStokes  (100 to  150  SSU).   The TTL5 NOX remained relatively constant at
220  ppm  over this  range.   The  TTL/MG22.5  NOX  decreased  about 14% to  174 ppm
while the viscosity  increased but  the  appearance  of  the flame  at  the root and
the end did not support operating at the higher viscosity levels.

The ability of  the TTL/MG22.5  to  operate  efficiently  at low excess air provides
its  greatest  advantage over  the  TTL5.   At  10% excess  air  (2.0%  02)  the TTL5
produces about  220  ppm NOX.   The TTL/MG22.5  produces  slightly higher NOX  (-230
ppm) at  10%  excess  air.   For the 5% excess  air design operating condition, the
TTL/MG22.5 produces 163 ppm NOX.   This represents nearly a 25%  reduction in the
level of  NOX  at  design  conditions.    Figure  8 shows  NOX versus  exit oxygen.
Although the TTL5 burner operated stably at lower oxygen, the  flame increasingly
contained  dark  areas  and  smoke.     The   TTL/MG22.5   produced   good   flame
characteristics down to 0.2% oxygen and perhaps  lower but there  was no practical
reason to pursue  testing below this  point.   The level  of  CO  was constant over
the range of 1.0% to 2.2% oxygen  for each burner,  with the TTL5 producing  about
48 ppm while the TTL/MG22.5 yielded a lower 38 ppm.  Down to 0.2% the  TTL/MG22.5
                                      4A-41

-------
CO level  increased  to about 43  ppm,  still  lower than  the  TTL5 at 1.0%  oxygen.
CO is  expected  to be  lower  in the  field  due to tnulti  burner interactions  and
longer residence time.

After  establishing  the N0x/exit oxygen  relation,  most tests  for  the TTL5  were
conducted at  about  2.2% oxygen which is typical  of field operating  conditions.
The TTL/MG22.5 tests  stayed  near 0.8% oxygen.   Figure 9 shows the  impact  of OFA
on NOX  emissions.   The change  in excess  air accounts for most of  the  difference
between  the  two  burners.  At  25%  OFA,   the  TTL5 NOX  levels  are  lowered  by  43%
while  the TTL/MG22.5  levels dropped 37%.    The TTL/MG22.5  achieves  the  target
level  of NOY, 104  ppm, at  20% OFA.   CO  levels  during  OFA  testing showed  no
            A
significant changes from the unstaged condition.

Figure  10 summarizes  the NOX data in  a comparison to the targets developed using
Figure  6.   From  this  data  the  application  of the new generation TTL/MG50  burner
is   projected to decrease  NOX  about  17%,  (see,  L3  in  Figure 10).   Staging  with
10%  OFA would reduce  NOX by 26% (L4).   Using  20% OFA, (Figure 9)   would  achieve
the  target NOX emission  (L2).

The  carbon  loss  reductions  obtained  with  the  TTL/MG22.5  reached  the  goal
established  from the  TTL5 burner  and the guideline  given  in  Figure  7.    Three
sets of firing  conditions  were evaluated  for  the  TTL/MG22.5 burner;   low  excess
air, low excess  air with OFA  and  low excess air with  a  combination  of OFA  and
flue gas  recirculation (FGR).   The  use  of  FGR had little impact on reducing  NOV
                                                                                A
during  residual  fuel  oil firing.  This is consistent with the  findings  of  others
(8).   As  illustrated  in Figure 11,  the  TTL/MG22.5 indicates a 47%  to  57%  carbon
loss reduction compared to the TTL5.

Natural  Gas and  Natural Gas/Residual Oil  Dual-Firing

Brayton  Point plans  to add natural  gas  capability  to  in  the future.    Since  the
unit has  never  fired  natural gas,  field data  to develop  a  NOY correlation  with
                                                              A
the  CBTF  are  unavailable.  The combustion testing of the TTL/MG22.5 with  natural
gas  in  the CBTF  provided information  on the response of NOV  emissions  to  various
                                                          A
hardware  and operating parameter adjustments.
At  design  operating  conditions  of  1%  02,  NOX  emissions  are  about  60  ppm.
Variations in oxygen between 0.8% and 1.1% resulted  in negligible changes  in NOY
                                      4A-42

-------
emission.    CO  did  not  vary  for  this range  of excess  air testing  (12  ppm).
Staging from 0% to 15% reduced  NOX from 61  ppm to 54 ppm indicating a very weak
response to staging.   Again, the CO levels remained near 12 ppm.

Simultaneous firing of natural gas and residual oil  in the TTL/MG22.5 burner was
tested in  the  CBTF.   In the  field Brayton  Point wants the  option  to fire both
fuels  simultaneously,  either  through  the  same burner, dual-firing,  or  through
separate burners,  co-firing.  As a single burner  installation,  the CBTF test rig
allows the  evaluation  of only  dual-firing  conditions.   The  state  NOX emission
regulations for new  sources  are specific for gas or oil  firing,  but not  stated
for  simultaneous  firing.   For  these  conditions,  a  linear  extrapolation  may  be
used as a first cut guideline by state environmental agencies.

Dual-firing tests were setup  to evaluate  NOX emissions  at varying ratios  of gas
and  oil  heat  input.   This  was done  with  and  without OFA.   The  results are
summarized  in   Figure  12.    As  natural  gas is  introduced  into  the  burner   a
significant increase in NOX emissions  occurs,  line  12.   This trend stops  near  a
ratio of 80% oil/ 20% gas.  Then the NOX decreases linearly to the level for 100%
gas.   A  similar situation  occurs  with  15%  OFA but  the effect  is diminished  at
these  lower  NOX levels.    In  this test program  similar trends were  found in  a
more conventional  steam  atomized low-NOx burner  except the  initial  rise  in NOX
did not diminish with the use of 15% OFA.

SUMMARY

NEPSCo  has supported  a program  that  will  assist  in  meeting  commitments   to
retrofit units  at Salem Harbor  and Brayton  Point.  The  goals of the program aim
for improving unit efficiency,  operation and fuel flexibility and simultaneously
reducing  NOX  emissions,  carbon loss  and   particulate  emissions.    The  program
evaluated several high efficiency,  low-NOx  burner technologies.  Communications
with  other users,  burner  proposal   evaluations,  atomizer  tests,   large  scale
combustion  tests,  retrofit impact analysis  and  economic  evaluations  have all
played a roll  in  advancing  the  program.   The selected technology, Rodenhuis and
Verloop  TTL/MG50,  is  scheduled  for  installation  at  Brayton  Point  and  Salem
Harbor in 1991,  Full scale test results should be available in 1992.
                                      4A-43

-------
REFERENCES

1.)    R.  F. Afonso, N.  M.  Molino,  D. C.  Itse,  J.  J.  Marshall, J. F.  Hurley and
      S.  Lindeman,  Evaluation  of Low-N0x,  High Efficiency  Oil  and Gas  Burners
      for  Retrofit  to  Utility  Boilers.  Presented at:  American  Flame  Research
      Committee  1989   International   Symposium  on  Combustion   in   Industrial
      Furnaces and Boilers, Short Hills,  New Jersey September  25-27,1989  and the
      EPRI Fuel Oil Utilization  Workshop, Clearwater  Beach, Florida November 1-
      2,  1989, EPRI GS-6919, July 1990.

2.)    R.  A. Lisauskas,  Experimental   Investigation  of  Retrofit  Low-No^  Combustion
      Systems. Proceedings  of  the  1985 Symposium on  Stationary Combustion  NOX
      Control, Vol. 1, EPRI CS-4360, January 1986.

3.)    D.  V. Giovanni  and T.  Sonnichsen, Gaseous and Particulate  Emissions  Tests
      at  Salem Harbor  Unit  4.   Report  submitted  to  New  England Power  Service
      Company by KVB,  Inc., NY,  December  1972.

4.)    G.  Dusatko,  Salem Harbor  Unit NOX  Reduction Program. Report submitted to
      New  England  Power Service  Company  by KVB,  Inc., NY,  September 1973  and
      supplemental report in July 1974.

5.)    N.   M.   Molino,  G.  Dusatko,  Field  Test   of a  Processed  and  Emulsified
      Residual  Oil  at  Salem Harbor Station    Unit  No.   4.  Published  by  the
      American Society  of Mechanical Engineers,  345, East  47th ST., New York, NY
      Ref. No. 87-JPGC-FACT-B.

6.)    C.   C.   Masser,  R.  A.  Lisauskas,  D.  C.  Itse.  Extrapolation  of  Burner
      Performance  From  Single  Burner  Tests  to  Field  Operations. Presented  at:
      1985  Joint  EPRI/EPA  Symposium  on  Stationary  Combustion  NOY  Control,
      Boston, MA, May 6-9,  1985.

7.)    P.   N.  Garay. Add  low-nitrogen,  emulsified  oils  to   list  of emerging  NO
      controls.  Power Magazine, October  1990.

      K.   M.  Bentley,  S.  F.  Jelinek,  NOX control  technology  for boilers  fired
      with natural gas or oil.   TAPPI Journal,  April 1989.
                                      4A-44

-------
 Figure 1.  Brayton Point Unit  4
               ©
                               ©-
©  ©       ©  ©


               ©  ©
               0  ©       ©  e-A
                                            OMAT
                                            OIL
                                            IGfUTCII
                                               CII-1
                                             LILII1NCIIS
                                             EL. 55'
                                            BURNERS
                                            FOHNEY
                                            trrE
                                            TTL
                                                       =n-
                                                        ^Z>)
                                                 	ID'-o   I
Figure  2.   General Burner Arrangement
                          4A-45

-------
        FLAME  STABILIZER


   PRIMARY AIR BULLET

           OIL GUN
                                                       REFRACTORY
                                                          THROAT
Figure 3.   Rodenhuis and Verloop  TTL5 Burner Arrangement
  Figure  ^.   Rodenhuis and  Verloop TTL5 Atomizing Bullet
                              4A-46

-------
Burner assembly
1/ Combustion air
21 Atomizing air
3/ Oil supply
4/ Gas supply
5/ Oil burner
 6/ Ignitor
 II Flame detector
 8/ First stage atomizing air
 9/ Second stage atomizing air
10/ Cylindrical air damper with drive
 Oil and  gas  burner
             Figure 5.  Rodenhuis and Verloop TTL/MG Burner
                      and Atomizing Bullet Assembly
                               4A-47

-------
         400:
              o Salem Harbor

              * CBTF TTL5
xLl
         250-
         200^
          1501
           5Q-.
             0   50   100  150   200  250  300  350  400  450  500


                   Burner Area  Heat Release,  103 Btu/hr-ftz
              Figure 6.   Correlation of CBTF  NOx
                          Emission to Full Scale
oo
en
o
o
CD
o
u.iuu -



A f"\7 c '
0 A7H "
n np;^ :
n n^n :
n HSR "
n nsn "*
n OAR ~
n nin ~
n nis =
n nin ~
n nos "
n n9n
n ois
n run
n nns
0 QQO
0




















0 1




















0 2




















0 3




















.0 4




















ill!
.0 5




















| 1 I! i
0 6












I""
4 	 -1






0 7










s


: 	 —






0 8









/
/









0 9








^
^


— —







.0 10







/




. 	







,111
.0 1






/s
^




	







.0 12





/






*~^*—







1 1 1 1 1
.0 i:




/*















Mil
.0 14



/*
















1 1 1 1
.0 15
                            ASPHALTENES,  %
                      • -ACTUAL             X-GOAL

          Figure  7.   Salem Harbor Unit  Performance
                      (Asphaltenes Versus  Carbon
                              4A-48

-------
 300
 200 -
 260 -
 240 -
 220 -
 200 -
 100 -
 150 -
 UO -
 120 -
 100 -
  00 -
  60 -
  40 -
  20 -
   0 -
                                               TTL5
              TTL/HG 22.5
                                         TARGET IN CBTF (101 ppm)
                             EXIT OXYGEN, %

  Figure  8.   Oil  Firing    Effect  of Exity  Oxygen  on  MOx
000
200 -
260 -
210 -
220 -
200 -
100 -
160 -
UO -
120 -
100 -
 00 -
 60 -
 40 -
 20 -
  0 -
TARGET IN CBTF (104 ppm)
                              12
                                      16
                                                        24
                                                                20
                                or A, i
           Figure 9.   Oil  Firing    Effect of  OFA  on NOx
                                  4A-49

-------
to

o
o
QQ
Cd.
            400
            350 -I
            300 -_
            250 -;
         =   200
             100
LI  Baseline
L2  Limit  0.3 Ib/HMBTU
L3  TTLMG/22.5   0%OFA
L4  TTLMG/22.5   10% OFA
                 0    50   100   150  200  250  JOO   350  400  450  500

                     BURNER AREA HEAT RELEASE, 103 BTU/HR FT2



         Figure  10.  Projected  NOx Emissions  to Full  Scale

                       CUlF-PAimCIILAlG SiAMI'l INQ TEST CONDITIONS AND CAmiON LOSS
2.4 :
9 n -
1 fi
1.2 -.
n H
0.4 -;
0.0 :
FACIUTY
cnrF
CBTF
CUFF
CBTF
CBTF







BUnNETI
DESIGN
TTI 0
TADGirr
TTUMG22 5-GTU. BULLT
TTL/MG22.5-STD. EULLT
TTL/MG22 5-STD GUI.LT



































TEST
NO.
^2~
-
91
90
101









OIL
ASPI 1ALT
11.2
11 2
O.D
0.0
9.0








I I I I
OIL
VISCOSITY
SBU
150
-
105
103
103





t


I I I I





38
34
loi

I I I I
EXIT
O2
2.2
00
0.7
0.8
0.7








II I I


SI





I I I
OFA
Va
0
-
0
11
9
Fen
0
-
0
0
9








I I

CARBON
LOSS
"Mi
1.01
1.09
0 92
0.9Q
079








I I I








I I I I
         0   I    2
        4567
                                              9   10   II   12   13    14   15
                               ASPHALTENES, %
                     0-TTL5:*-GOAL:-|--R&VTTL/MG

                   Figure  11.   Performance Comparison
                                 Asphaltenes vs.  Carbon  Loss
                                      4A-50

-------
    300

    220 -

    ZSQ -

    2W -
ro


-------
DEVELOPMENT OF AN ULTRA-LOW NOX PULVERIZED COAL BURNER

                        Joel Vatsky, Director
                  Timothy W. Sweeney, Supervisor
               Combustion and Environmental Systems
                  Foster Wheeler Energy Corporation
                      Perryville Corporate Park
                   Clinton, New Jersey 08809-4000

-------
        DEVELOPMENT OF AN ULTRA-LOW NOX PULVERIZED COAL BURNER
                                    Joel Vatsky, Director
                                Timothy W. Sweeney, Supervisor
                            Combustion and Environmental Systems
                              Foster Wheeler Energy Corporation
                                  Perryville Corporate Park
                                Clinton, New Jersey 08809-4000
ABSTRACT

Foster Wheeler has been utilizing the Controlled Flow/Split-Flame Low NOX burner for both new and
retrofit applications  since 1979. This Internally Staged    burner attains 50-60% NOX reduction, as
compared to pre-NSPS turbulent burners, without utilizing any staging ports. A new burner has been
developed which combines  the  internal staging concept with  another patented Foster Wheeler
technology: fuel staging. This new design, which is defined as Internal Fuel Staging  , is consistently
achieving NOX levels of 0.25 lb/10  Btu with bituminous coals containing 22-35% volatile matter and
fuel nitrogen of 1.8%. This represents at least 75% reduction from  turbulent burner levels. This paper
discusses the results of comparative tests between the standard CF/SF Low NOX burner and the new
IPS design.
                                          4A-55

-------
       DEVELOPMENT OF AN ULTRA-LOW NOx PULVERIZED COAL BURNER
INTRODUCTION

An advanced low NOX burner has been developed which achieves NOX emission levels as low as 0.25
lb/106Btu; equivalent to reductions of up to 75% from turbulent burner levels. The Internal Fuel
Staged™ design is based upon concepts developed and patented by Foster Wheeler in the late 1970's.
Development of the IPS design was not undertaken at that time in favor of proceeding with the
Controlled Flow/Split-flame low NOX design. The latter is Foster Wheeler's standard low NOX burner
which as been in utility and industrial service, in both new unit and retrofit applications, since 1979.

The CF/SF burner was developed to achieve at least 50% NOX reduction in retrofit applications and
meet the 1979 New Source Performance Standard, of 0.5 lb/106 Btu for sub-bituminous and 0.6 lb/106
Btu for bituminous coals, without the simultaneous use of supplementary NOX controls such as overfire
air.

The IPS design, having a NOX control capability significantly greater than that of the CF/SF, was
delayed in development simply because there was no commercial market for a system with such low
NOX capabilities. By 1989 it became apparent that new source requirements were tending toward 0.3
lb/10  Btu and retrofit requirements toward 0.5 or lower.

Although the CF/SF design operates  below the  0.5 level,  in both new units and retrofits without
overfire air, and below 0.3 with overfire air, Foster Wheeler decided to advance the technological
capability by developing the IFS. Commenced in May 1990, the IPS development was completed in
September 1990 and has been offered commercially, with considerable success, since that time.

Development testing was done on Foster Wheeler's 80 million Btu/hr Combustion and Environment
Test Facility. Since substantial testing of the CF/SF burner had been performed on the CETF and
correlated with utility boiler data, and there are over 5,000  MW of retrofitted  and over 7,000 MW of
new units with this design, all development testing of the IFS was comparative with CF/SF data.

Typical NOX emissions from the CF/SF burner, on  theCETF, are about  0.4 lb/106 Btu. Typical IFS
emissions are 0.25-028 lb/106 Btu, at least one-third lower; both without overfire air.

It should be noted that the IFS design differs from the CF/SF in only a single component: the fuel injector's
nozzle.  Consequently, it  represents  only  a minor change  in overall design  since all  other burner
components, and the operating method, are identical to the CF/SF design. The IFS coal nozzle is, therefore,
easily retrofittable to CF/SF burners currently in operation, thereby converting it to the EPS.

CONTROLLED FLOW/SPLIT-FLAME LOW NOX BURNER

The CF/SF design shown in Fig. 1 is based upon the principle of Internal Staging™ of the flame. This
principle was developed and defined by Foster Wheeler in the 1970's.


                                           4A-56

-------
INTERNAL STAGING™ is defined as:

A low NOX burner design which two-stages the secondary air flow and stages the primary air/fuel
flows within the burner's throat while maintaining classical turbulent burner flame patterns and low
pressure drop: < 4.0" H20.

The Controlled Flow /Split-flame name is derived from the operating functions of the burner:

           Controlled How for the dual register design which provides for the control of inner
           and outer swirl zones along with a sleeve damper which allows independent control
           of the quantity of secondary air flow to each burner.

           Split-Flame for the coal injection nozzle which develops a split-flame pattern for
           obtaining low NOX emissions.

Key criteria within the overall design philosophy are summarized as follows:

           Mechanical reliability to be such that after long term operation movable components
           would still operate.

           Combustion air flow and swirl to each burner to be independently controllable.

           Adjustable primary air/coal velocity to ensure optimum relation between primary
           and secondary air streams.

           No increase in primary or secondary air pressure drop so that existing PA and FD
           fans can be used.

      •     Burner  capacity  to  cover the  complete  range of  industrial  and  utility use:
           approximately 30 to 300 million Btu/hr.

           Plug-in   retrofitability,  i.e.,  no  pressure  part changes,  no  burner  piping
           rearrangement  and no major  windbox modifications when installed on most
           existing wall-fired boilers.

The CF/SF low NOX burner's components and their functions are described below:

           Perforated Plate with Sleeve Damper: used to control secondary air flow on a per
           burner basis. By measuring the pressure drop across the perforated plate an index of
           air flow  is obtained. The  air  distribution,  vertically and horizontally, within the
           windbox is thus optimized by adjusting the sleeve dampers to obtain equal burner
           stoichiometries. This is a one time optimization after which the "open" position is
           fixed. The sleeve damper has  "closed", "ignite" and "open" positions and is used,
           instead of the main radial  vane register, to shut off the air flow when the burner is
           out of service. It is controlled by an electrically  operated linear drive, but is not
           modulated with load.

           Dual Series  Registers: provide improved flame shape control by two-staging  the
           secondary air. A key mechanical reliability feature of this configuration is that  the
           blades and drive mechanisms, set back from the furnace wall, are well "shaded"
           from direct  flame  radiation.  Consequently, the registers operate  at windbox


                                           4A-57

-------
           temperature and do not overheat, warp or bind. Additionally, once the flame is
           optimized for proper shape, the registers are fixed. They remain in their optimum
           position and are not modulated with load or closed when the burner is taken out of
           service since the sleeve damper performs the latter function. The burner essentially
           becomes a fixed register type with the option for adjusting the register if a major
           fuel change occurs; the drive mechanisms being manual.

           Split-Flame Nozzle: segregates coal into four concentrated steams. The result is that
           the volatiles in the coal are driven  out and are burned  under more reducing
           conditions than otherwise would occur without the split flame nozzles. Combustion
           under these conditions converts the nitrogen species contained in the volatiles to N2,
           substantially reducing NOX formation.

           Adjustable Coal Nozzle: allows primary air/coal velocity to be optimized without
           changing primary air flow. The proper relationship between primary and secondary
           air is important for both good combustion and flame shaping.  Once optimized no
           further adjustment is required.

Succinctly, only the sleeve damper, used to shut off the secondary air flow, is moved when the burners
are taken in or out of service. Thus, after optimization, the burners become fixed register types.

Mechanical reliability of the design concepts, materials and operational methodology has been fully
confirmed by commercial experience.

ADVANCED OVERFIRE AIR SYSTEM

NOX emissions from the burning of pulverized coal have three (2) sources: Thermal NOx generated
from thermal fixation of atmospheric nitrogen (Nz) at high flame temperatures, conversion of bound
nitrogen in  the coal's volatile fraction and conversion of bound nitrogen in the coal's char fraction. The
latter being the most difficult part of the emission to control.

Foster Wheeler's NOX control philosophy has been based upon using the CF/SF low NOx burner which
achieves a high degree of thermal and volatile-fraction fuel NOX reduction; with char-fraction fuel NOX
reduction to a lesser degree. This low NOX burner's effectiveness is uniformly consistent in reducing
NOX emissions by 50-60% from uncontrolled levels.  When  lower levels  are required  an advanced
overfire air system (AGFA) can be  incorporated to increase NOX control to the 70-80% range. These
results are valid for both new steam generators and retrofittable existing units.

Figure 2 schematically illustrates a typical AOFA system. It is characterized by a set of overfire air ports
placed well above the top burner level to provide relatively long residence time between the top burner
level and the overfire air port  level.  One port is located above each burner column with an additional
port near each sidewall. This is illustrated in Figure 2 where a four-burner wide arrangement uses six
overfire air ports on each firing wall.

Older overfire air  arrangements, which used fewer ports and shorter residence time, can achieve only
about 20-30% NOX reduction from the low NOX burner's level. This AOFA system increases the NOX
reduction capability to the 40-50% range. Foster Wheeler demonstrated  this system's capability in  the
early  1980's  when  two (2) new utility boilers,  a  275  MW front-wall fired  unit and a 550 MW
opposed-fired unit, were tested. The combination of the CF/SF low NOX burner and the above noted
AOFA principles  enabled both units to  operate at NOX  levels of about  0.2 lb/106 Btu.  During
subsequent commercial operation the overfire air ports were closed since  the NOX regulations to be


                                          4A-58

-------
attained were only those of the 1977 NSPS. Without overfire air both units operated at about 0.41b/106
Btu.

The AGFA System was not commercialized by Foster Wheeler until the late 1980's since there was no
regulatory need to achieve emission levels below 0.3 lb/10  Btu. This system is now being incorporated
on several new steam generators, ranging in size from 65MW to 550 MW. In early 1990 it was retrofitted
to a 500 MW unit, having the arrangement shown in Figure 2, where NOX emissions were reduced 78%
to about 0.26 lb/106 Btu.

FIELD EXPERIENCE: CF/SF AND AGFA

Since its introduction in 1979, when retrofitted to the 36Q MW San Juan No. 1 unit of Public Service
New Mexico, the CF/SF low NOX burner has been  successfully retrofitted to a total of ten (10) utility
boilers. Table I  summarizes this  experience, which totals 5,135 MW. The average NOX reduction
attained on these units is nearly 60% without overfire air; and nearly 80% with AOFA.

Table II is a listing of the  projects underway for 1991: eight (8) units totalling 3,635 MW. Options in
these projects add another 2,380 MW for a total of over 11,000 MW and over 600 burners.

The NOx control capability and results are summarized on Figure 3, which contains data from four of
the ten utility units,  two industrial units and Foster Wheeler's Combustion and Environmental Test
Facility. Table in lists the range  of fuel properties in these retrofit applications. The NOX control
capability is consistent.

Figure 3  graphically illustrates  the  NOX  control  effectiveness  of  the  low NOX  Controlled
Flow/Split-Flame burner.  The plot is of total NOX emission  against Burner zone Liberation Rate (a
measure of heat input to the burner zone; the higher this number the hotter the lower furnace), for
turbulent burners, CF/SF low NOX burner and CF/SF in combination with AOFA. The data show both
industrial and utility units  with burner capacities ranging from 30 to 300 million Bru/hr. The curves are
not load curves but,  rather, represent  the  full  load NOX emission for each value of Burner Zone
Liberation Rate.

Conclusions drawn from the summary information contained in this figure are:

      1.    NOX control is independent of burner capacity:  large burners and small burners
           achieve the same degree of NOX reduction.

     2.    The low  NOX burner (lower two curves) is much  less sensitive in the thermal
           environment than is the turbulent burner. There is a much smaller slope of the NOX
           vs BZLR  low  NOX curves than for  the turbulent burner curves indicating a very
           small amount of thermal NOX is emitted by the low NOX burner (due to its lower
           flame temperature).

     3.    NOX reduction in  the higher temperature units  is somewhat greater than in the
           lower  temperature (lower BZLR) units due to the substantial decrease in thermal
           NOX in the former.

     4.    Uncontrolled  NOX  emissions from single wall-fired  units is higher than from
           opposed-fired units, yet this difference is eliminated by the CF/SF burner.
                                           4A-59

-------
     5.    Foster Wheeler's Advanced  Overfire Air  System provides  an additional  NOX
           reduction of 40-50% (to total reduction of 70-80%) below levels emitted by the low
           NOX burner only.

Foster  Wheeler  has  retrofitted  wall-fired steam   generators  from  all  major  domestic boiler
manufacturers, including an 800 MW opposed-fired unit originally equipped with three-nozzle "cell"
burners. The effectiveness of the CF/SF low NOX burner is the same regardless of the boiler in which it
is installed.

Of utmost importance to steam generator operation are its performance and efficiency. In  none of the
low NOX burner  retrofits performed by Foster Wheeler have either boiler performance of efficiency
been deteriorated from pre-retrofit conditions. Typically efficiency is improved due to:

           Reduced excess air operation yielding lower stack losses and reduced forced  draft
           and induced draft fan power.

           Lower burner pressure drop yielding further F.D. fan power savings.

           Cleaner furnace walls (reduced slagging).

           Unburned carbon levels equal to, or lower than, original equipment burners.

Succinctly, unit operations are equal to, or better than, pre-retrofit while NOx is reduced at least 50%
without overfire air.

Also shown on Figure 3 is the effectiveness of AOFA on three utility units and the CETF. All results,
without and with AOFA, are uniformly consistent in terms of NOx control.

Two projects are of particular interest: Units 1 and 7 on Figure 3. Unit 1 is an 800 MW boiler originally
equipped with 18-3 nozzle cell burners and was retrofitted with 48 CF/SF burners in early 1989. Unit 7
is a 500 MW  boiler retrofitted with CF/SF burners and AOFA in early 1990. Since the results of these
two retrofits are typical of all others they will be briefly discussed below.

      1.    Four Corners Unit No. 4: Cell Burner Boiler - Arizona Public Service

           This boiler is a Babcock & Wilcox opposed-fired, supercritical steam generator with
           a maximum continuous rating (MCR)  of  5,446,000 Ib/hr  main steam  flow at
            1000/1000F and 3590 psig. The unit was fired by 18, 3 nozzle cell burners (54
           throats). Turbine rating is 820 MWG (780 MW net). Unit 4, along with its sister Unit
           5, was built in the late 1960's; and went into commercial operation in 1969 and  1970,
           respectively. In 1971  the State of New Mexico instituted a retroactive NOX emission
           limit of 0.7 lb/10 Btu for coal-fired units  constructed  prior to 1971. Over the years
           Arizona Public Service conducted several test and evaluation programs to arrive at
           an acceptable means of  achieving  the NOX  limit without  degrading boiler
           operability, performance or efficiency. In  1988 Foster Wheeler was contracted with
           to provide a low NOX conversion with a guarantee lower than the State limit (0.65 vs
           0.7).  The unit was experiencing severe slagging, wide spacial variation of Furnace
           Exit Gas Temperature and resultant high superheater outlet metal temperatures.

           Consequently, Arizona Public Service  chose to proceed with a complete
           revision  of  the lower furnace. This consisted  of replacing both front and


                                           4A-60

-------
rear firing walls with a new burner arrangement, six (6) burners wide and four (4)
burners high.

Foster Wheeler's scope of supply consisted of 48 CF/SF low NOX burners, twelve
burner wall panels (each containing four throats), new igniters and scanners and
miscellaneous equipment. Arizona Public Service had already contracted with B&W
to supply nine new MPS-89N pulverizers for each unit (Unit 5 mills were previously
installed and Unit 4 mills were installed during the low NOX conversion). Due to the
new firing configuration the  total number of mills purchased  for both units was
reduced from 18 to 16.

It should be  emphasized that this  conversion  represents the most  extensive
modification that can be performed on a cell-burner boiler's firing system. However,
the method chosen was  designed not  only to reduce NOX emissions but also  to
improve boiler operability. In particular to produce a more uniform Furnace Exit
Gas Temperature (convert to one mill supplying a burner level) and reduce slagging
(vertically spreading the burners). Slag reduction also occurs because of the lower
flame temperature inherent in the CF/SF burner.

Other cell burner  equipped boilers may not require  modification as extensive  as
Four Corners 4 & 5. In many  units only a direct low NOx burner replacement may
be acceptable.

Also, of the total of 23 units either completed, underway or as contractual options,
the two Four Corners units are the only ones requiring panel replacement. The vast
majority  of  boilers,  regardless  of  original  equipment  manufacturer,  can be
retrofitted with Foster  Wheeler  low  NOx  burners without  pressure  part
modification.

The results of this retrofit are summarized on Table IV which compares pre and
post-conversion conditions.

Significantly, burner pressure drop has been reduced 2-2.5" BtzO which will yield
substantial forced draft fan power savings.

During the past nearly two years since start-up the unit has operated according to
system load requirements, almost continually at full load. The following results
have been observed:

- - There is no flame impingement on any heat transfer surface and no extension of
   flames.

- - Slagging and clinkering have been nearly eliminated on Unit 4. The furnace walls
   are now the cleanest they have been since initial unit start-up 20 years ago.  In
   contrast the unconverted  Unit 5 continues to have the same  level of slag
   accumulation that Unit 4 had prior to conversion.

- - There has been no increase in FEGT, instead it appears  to have decreased as a
   result of increased furnace absorption caused by the clean walls.
                                 4A-61

-------
     - - Combustion efficiency has not been decreased, as indicated by CO and unburned
        carbon levels.

     In summary, NOX emissions, have been reduced about 60% with no degradation in
     boiler performance or efficiency. Unit operability has been significantly improved
     due to the near elimination of the slagging/clinkering problems and elimination of
     excessive superheat tube metal temperatures.

     Unit 5 is currently undergoing retrofit.

2.    Hsin-Ta #1: Taiwan Power  Corp:  This is a  500  MW Foster Wheeler natural
      circulation unit in commercial operation since  1978. Originally equipped with an
      early  non-split flame low NOX burner (FW Controlled How design) and a basic
      1970's vintage overfire air system, it was designed to meet a NOX regulation of 0.7
      lb/106 Btu. More recently, the NOX regulations  in Taiwan have been reduced. As a
      result TPC purchased 24  CF/SF low NOX burners and the AOFA system for both
      Foster Wheeler units on that site.

      Unit 1 was converted in early 1990,  and has been operating for approximately ten
      (10) months.  Because of the simultaneous  use  of CF/SF burners and AOFA this
      unit's operating data will be presented in  more detail than Four Corners #4. The
      firing system and geometrical arrangement are as shown schematically in Figure 2.

      After start-up and low NOX optimization a series of performance tests were done
      over a period of several weeks. Since Taiwan imports most of its coal, fuels from the
      US, Australia and South Africa were fired as part of normal plant operations. Table
      V lists the typical range of coals fired at this station. Fuels varied on a day-to-day
      basis. As shown on the table volatile matter varied  from 22.5% to 36.1%  with fuel
      nitrogen contents covering the relatively high range of 1.85% to 2.28%.

      However, the low NOX systems effectiveness  is such  that there is no significant
      affect of these fuel properties on NOX emissions. Boiler performance and efficiency
      variations were only those normally expected due to fuel properties affecting gas
      weights and moisture content.

      Figure 4 compares NOX emissions as a function of load for the CF/SF burner with
      AOFA closed and open. Load was reduced in  the manner normal for that station
      with pulverizers taken out of service at the same loads they were prior to the low
      NOX retrofit. The  unit achieved full load with all mills in service or any single mill
      out of service. Consequently full load testing was performed with all mills in or one
      top or bottom mill out of service.

      NOX  emissions decrease  monotonically from  full load values  average 0.464 and
      0.266 lb/10  Btu with AOFA ports closed and open, respectively. For simplicity of
      graphical illustration only seven (7) data points  are shown, out of a much larger set.
      The  temporal and fuel-property related variations in emissions cover  a narrow
      range: at not time did NOX emissions exceed 0.5 or 0.3  with AOFA closed or open,
      respectively.

      Figure 5 is a similar plot of windbox-to-furnace pressure drop as a function of load.
      These results are  also typical of the Foster Wheeler low NOX system. At full load


                                      4A-62

-------
           with AOFA closed and all mills in service AP does not exceed 3.0" H2O. When
           AOFA ports are opened AP decreases to less than 1.5" HiO at full load. In both cases
           AP decreases monotonically as load decreases.

           Figure 6  illustrates  the change  in variability of  the NOX emission  as  more
           sophisticated NOX controls are used at full load. From a mean (X) emission of 1.23
           lb/10 Btu and standard deviation (S) of 0.06 in the uncontrolled case, X and S are
           decreased to 0.464 and 0.032 with CF/SF burner and no overfire air. When AOFA
           ports are opened X = 0.266 and S = 0.012. The data includes all miH combinations
           and the range of fuels listed in Table V.

           The following summarizes the results of Figures 4,5 and 6.

          - - NOX does not increase as load is decreased.

          - - Full load pressure drop is no greater than for typical turbulent burner values, less
             than 3.5" HiO at full load.

          - - High burner pressure drop is not needed to attain high NOX reduction and low
             absolute emission level.

          - - NOX emission variability due to operating condition and fuel properties decreases
             with the use of more'sophisticated NOX controls.

           Although the above results  and  conclusions are  presented  for  two  specific
           retrofitted units, they are typical of data obtained on all other retrofits as well as
           from thousands  of additional MW's of  new steam  generators utilizing the  same
           equipment.

INTERNAL FUEL STAGED™ LOW NOX BURNER DEVELOPMENT

The  development of the IPS low  NOX burner has  been  performed  on the Combustion  and
Environmental Test Facility (CETF) located at Foster Wheeler's manufacturing facility in Dansville, NY.

Among the types of test work being performed at the CETF are  burner development, low NOx furnace
design evaluations,  sulfur dioxide  control by  dry  sorbent injection,  fuels evaluations, integrated
combustion/emissions control testing and customer support and problem analysis.

The CETF's furnace (Fig. 7) is arranged to produce conditions which closely match those of commercial
equipment, for example:

           Furnace residence time is limited to about 2 sec. max. between the burner centerline
           and furnace exit.

           Furnace  Exit Gas Temperature  (FEGT) is about  2200 °F and can  be varied, by
           adjusting total furnace absorption.

           Bumer/fumace aerodynamics are similar to  those of commercial equipment as are furnace
           mixing patterns, due to both overall geometry and the two-burner-high arrangement.
                                            4A-63

-------
The  arch configuration, which is  required to provide proper conditions for low volatile fuel
combustion, is also important to produce the desired velocities/residence times and FEGT with
horizontal firing. The reduced upper furnace  cross-section is  responsible for the increased
velocity  and resultant decreased residence time. Decreased upper furnace heat transfer surface
reduces  absorption above the burner zone, thereby increasing FEGT to levels normally achieved in
commercial equipment.

Although not  currently equipped with  a superheater,  the unit has  screen tubes at the furnace exit
followed by a  water-cooled economizer which yield realistic gas quenching rates. In this way all flue
gas constituents, gaseous and particulate, are comparable to field equipment. Commercial practices,
used throughout the CETF system  design, have been  mated  with  research-oriented considerations
wherever practicable to maximize the usefulness and flexibility of the system.

The  CETF utilizes a direct-fired system (i.e., hot-primary-air-swept ball mill) feeding either a single 75
million Btu per  hour  arch-fired  twin-cycle burner  assembly  or  two 40  million Btu  per hour
horizontally-fired burners (shown) which fire into a refractory-lined water-jacketed furnace. The water
jacket operates under water-head pressure and utilizes  natural circulation, producing steam which is
vented to the air through a steam drum above the furnace. Combustion gases leaving the furnace flow
over horizontal, convection tube surfaces (economizer)  cooled by forced-water circulation. The gases
then pass through a two-stage  air heater (a tubular air heater  followed by a heat pipe air heater), a
baghouse dust collector, an induced-draft fan, and  then  the  stack. The level of  sulfur dioxide is
controlled by injecting a sodium-based sorbent into the gas stream prior to the baghouse. In addition to
back-end sulfur dioxide cleanup, the  CETF has the capability of  furnace injection of calcium-based
sorbents for evaluation of in-situ sulfur dioxide control.

IPS DESCRIPTION

The IFS  concept, as noted in the introduction, was conceived in the  1970's but not  developed at that
time. Internal Fuel Staging is defined as:

            A low NOX burner which two-staged the secondary air  flow and internally stages
            the fuel flow and primary air flow such that co-axial flames are developed within
            the burner's throat while maintaining classical turbulent  burner flame patterns and
            low pressure drop:  < 4.0" H2O.

Figure 8 schematically illustrates the IFS design. The philosophy of this development was to utilize as
much of the commercially-proven hardware of the Controlled How /Split-Flame low NOX burner as
possible. As can be seen in Fig. 8, the IFS differs from the CF/SF in only one respect: the coal injector's
nozzle.

AH other mechanisms and functions are identical to those of the CF/SF, including operating method.
Externally, the two burners appear to be identical. However, internally the fuel injection's nozzle has
been redesigned to produce split-flames surrounding  a co-axial  internal flame. The result is NOX
emissions approximately one-third lower than those attained with the  CF/SF design.

Table VI is a comparison of the key features of the CF/SF and the IFS. The similarity is so great that an operator
would note no difference between the two designs, other than the lower emission from the IFS.

The IFS burner is intended for use in both new and retrofit applications. Also, if units which are already equipped
with the CF/SF design require lower emissions all that is necessary is to replace the fuel injector.
                                            4A-64

-------
IPS PERFORMANCE RESULTS

Due to extensive performance and emission data available for the CF/SF low  burner, from the CETF
and new and retrofit field applications, IFS testing on the CETF is compared to CF/SF data.

Boiler performance results demonstrate no adverse difference in either furnace absorption or Furnace
Exit Gas Temperature on the CETF between the two low NOX burner designs. Since extensive field data
already exits which shows no change in boiler performance when the CF/SF is retrofitted to boilers
initially equipped with turbulent burners, it can be inferred the IPS design will, similarly, not adversely
affect boiler performance.

A wide range of fuels, sub-bituminous and bituminous have been tested on the CETF using the CF/SF
burner. For the EPS development testing two baseline bituminous coals have been used: a low volatile
and several high volatile coals.

Table VII compares NOX and CO data at full load for  the two burner designs with the high  and low
volatile coals. NOX is reduced 33-35% below CF/SF levels using the IFS. For all data, overfire air ports
are closed and excess C>2 is 3.4-3.6%. Note that for both fuels the variability in NOx emission decreases
significantly, to about the same standard deviation.  Typical coal analyses are listed in Table VIE.

Figure 9 compares NOX and air-to-coal  ratio (A/C) to load. Over a turndown ratio of nearly 3:1 NOX
decreases  monotonically with load from about 0.27 to about 0.21 lb/10  Btu. A/C increases as load
decreases from about 2.1:1 to about 2.6:1, covering the range of typical vertical roller pulverizers. This is
a significant result in that reduced load operation  does not deteriorate  the IFS burner's performance:
NOX does not increase as load decreases.

It is also instructive to compare the CF/SF NOX data from the 500 MW boiler, without and with
overfire, to the CF/SF  and IFS data from the CETF without overfire air. Figure 10 presents such a
comparison. The mean uncontrolled emissions for  the 500 MW and the  CETF are 1.23 and 1.04 lb/10
Btu for the fuel ranges listed in Tables V and Vin respectively. The 500  MW unit's emission in higher
than that of the CETF because it is a hotter unit and  was firing higher nitrogen fuels than the CETF.

On each unit, NOX is reduced by over 60% with the CF/SF low NOX burner without overfire air. Note
that both  the uncontrolled data  and  CF/SF data have nearly identical  standard deviations
approximately 0.06 and 0.03, respectively. When overfire air ports are open on the 500 MW unit NOX is
further decreased to a mean value of 0.266 lb/10  Btu: a total reduction of over 78%.

On the CETF the IPS low NOX burner yields a mean NO>c emission of 0.269 lb/106 Btu without overfire
air: a total reduction of nearly 75%. Note that the standard deviations on the 500 MW unit with overfire
air ports open is about the same as that of the CETF when the IFS is used with overfire air ports closed.
Again illustrating that more sophisticated NOX controls become  less sensitive to operating conditions
and fuel parameters.

A further comparison from this data is the NOX reduction from CF/SF levels due to AOFA or IFS. On
the 500 MW unit NOX reduced 42.7%  when the OFA ports are open.  On the CETF, the IFS burner
reduces NOX  33.6%  below CF/SF levels. The IFS low  NOx  burner is  achieving a NOX reduction
capability that is 80% of the additional reduction AOFA attains beyond CF/SF levels.
The relationship of the IFS results to field data are graphically illustrated on Figure 11, which is a simplified version
of Figure 3. Clearly, many commercial boilers will be able to operate at levels below 0.4 lb/10  Btu without AOFA
Where lower levels are required the IFS low burner can be supplemented with the AOFA System.
                                           4A-65

-------
SUMMARY

Foster Wheeler's standard Controlled Flow/Split-flame NOX burner has been successfully retrofitted
to ten (10) utility steam generators ranging in size from 225 MW to 800 MW, with an additional 6000
MW either underway or as options to existing contracts. In 1991 3635 MW will be retrofitted. Of these
three units, two B&W and one Foster Wheeler, will be receiving the new Internal Fuel Staged   low
NOX burner design.

The EPS burner development has been successfully completed and it has been offered commercially for
new steam generator and retrofit use. The full range of emissions and boiler performance guarantees
are offered. In addition to the above noted retrofits, the IPS will be installed on the following new steam
generators:

                           2 x 65 MW; NOX guarantee:         0.32

                           2 x 150 MW*;NOx guarantee:       0.27

                           1 x 550 MW; NOX guarantee:        0.32
      *
            These Units will also be equipped with Foster Wheeler's SCR system to achieve 0.1
            lb/106Btuatthe Stack

The IPS design achieves  NOX levels one-third lower than those attained by the standard CF/SF low
NOX burner, when tested on the Combustion and Environmental Test Facility. NOX levels below 0.27
lb/10  Bru without simultaneous  use of overfire air,  are routinely attained with a wide variety of
bituminous  coals with  NOX  emissions showing little  sensitivity  to  operating  mode  or  fuel
characteristics. These performance results are obtained at low burner pressure drop and with short
flames which do not cause adverse  changes  to furnace  absorption  rates  or Furnace Exit Gas
Temperature.

It is fully expected that these  results will be duplicated in the new unit and  retrofit applications
currently underway.
                                            4A-66

-------
Year

1979
1980
1981
1986
1987
1989
1990
1990
1990
1990
Total
MW
360
525
275
650
500
800 (Cell)
800
500
500
225
5135 MW
0.95
*•
X-
1.0
1.15
1.25
1.10
1.23
1.15
1.15

             Table I
        Low NOX Burner
       Retrofit Experience

        NOX db/106 Btu)

Before            After
                                 0.42
                                 0.40
                                 0.40
                                 0.41
                                 0.55
                                 0.50
                                 0.45
                                 0.27+
                                 0.50
                                 0.50
                                                                Reduction
      OEM
56
—
—
59
52
60
59
78
56
56
FW
FW
FW
FW
FW-UK
B&W
FW
FW
FW UK
FW
  New Boiler Converted during erection         + With Advanced Overfire Air System

  NOTE: Boiler ages range from 10 to 25 years
        Unit Size (MW)

        800 MW (Cell-Burner)
        150 MW
        200 MW (3 x 200)*
        650 MW
        500 MW
        525 MW
        650 MW (2 x 650)
        160 MW (3x160)
  Total 3635 MW
*
 (     ) = Contractual Options
                           Table II
                       Low NOX Burner
                  Retrofits Underway of 1991

                        No. of Burners

                            48
                            16
                            18(54)
                            24
                            24
                            24
                            24(48)
                            16(48)
                           194
Boiler OEM

   B&W
   B&W
   B&W
    FW
    FW
    FW
    FW
    FW
                                          4A-67

-------
      HHV
      N2
      Si
      Ash
      Average NOX Level
      NOX Reduction
          Table III
   Range of Fuel Properties

                  8000
                   1.0
                   0.5
                   5.0
                    13.500 Btu/lb
                    2.3%
                    3.5%
                    2.5%
                            0.451b/10 Btu
                          55 - 60 % (No OFA))
Parameter
No. Mill in Service
Excess Air (%)
NOx(lb/106Btu)
CO (ppm)
Burner A P ("HzO)
Carbon Loss (% Eff.)
Results:
Performance and efficiency not degraded
Eliminated slagging and clinkering
Significant power savings
           Table IV
     Four Corners # 4 Data
    800 MW Cell-Burner Unit
           Pre-Conversion
               All (9)
                 18
                1.27
                <40
                5-6
               Post-Conversion
                    All (8)
                    13-15
                  0.48 - 0.52
                    20-40
                     3-3.5
                                   Table V
                     500 MW Advanced Low NOX Retrofit
                             Range of Fuels Fired
                                (As Received)
VM (%)
EC (%)
Ash (%)
H2O (%)

C
H
O
N
S
HHV (Btu/lb)
22.51
54.05
15.55
 7.49

64.49
 3.79
 5.48
 2.07
 0.73
10,909
28.07
52.63
12.45
 6.85

67.39
 4.11
 6.27
 2.28
 0.65
11,605
36.15
41.47
12.01
10.37

61.99
 4.79
 8.70
 1.85
 0.59
11,018
                                  4A-68

-------
Desgn Parameter

Registers
   Adjustments
Sec. Air Flow Control
Fuel Injector
Primary Air Velocity Control
NOX Reduction
NOX Level
                             Table VI
                   Low NOX Burner Comparison

                              CF/SF

                               Dual
                             Manual
                        Elec. Sleeve Damper
                            Split-Flame
                               Yes
                             55 - 60%
                          0.41b/106 Btu
                                            IFS

                                           Dual
                                          Manual
                                    Elec. Sleeve Damper
                                        Split-Flame
                                           Plus
                                   Coaxial Internal Flame
                                            Yes
                                          70 - 75%
                                      0.27 lb/106 Btu
                                        Table VII
                                 CETF Comparative Data
                                       CF/SF vs IFS
Fuel V.M. (%)
      N2 (%)
XS02(%)
Burner Type

NOX (lb/106 Btu): Avg.
            Std. Dev.
CO
   Avg.
Std. Dev.
0.41
0.033

54
13



IFS
0.273
0.014
59
11
33
1.59
3.5 - 3.6
CF/SF
0.40
0.035
70
15



ffS
0.260
0.011
75
17
NOX Reduction (%)
                                  33.4
                                               35.0
                                         4A-69

-------
                                          Table VIII
                                    CETF Fuels: IPS Testing

VM                                         32.70                              23.49
FC                                          48.16                              59.41
ASH                                         9.11                               4.34
H2O                                        10.03                               7.86

C                                           67.34                              77.37
H                                           4.50                               4.68
O                                           5.60                               3.19
N                                           1.59                               1.87
S                                            1.85                               0.79

HHV                                       11,935                             13,877
NOX Reduction                              79.5%                             73%
NOX Level                                  0.266                              0.273

No significant difference in NOX emissions due to coal characteristics
                                           4A-70

-------
 ELECTRIC SLEEVE
  DMPER DRIVE

 MANUAL OUTER
 REGISTER DRIVE
                                           PERFORATED PLATE AIR HOOD

                                                      MOVABLE SLEEVt
                                                         DAMPER
                                                                  01 SUN
                                                              SPLIT FLAME
                                                             ICOAL NOZZLE

                                                            ADJUSTABLE
                                                            INNER SLEEVE
        Fig 1. Controllled Flow Split Flame Burner
LOWER
FUBNACE
AIRPOBI!
                                               IYPICAI
                                               OVEBFIRi
                                               AIRPORT
                                               DANFEt!
                  SECONDARY
                  AIR DUCT
                  VEN1UB
                                                              AIR DJC1!
    Fig 2. Typical Advanced Overfire Airport System
                             4A-71

-------
IV)
            (1)   800  MW Four Corners #4
            (2)   626  MW Pleasants #2
                 275  MW Front Wall Fired
            (3)    360 MW Front Wall Fired
             •    525  MW Opposed Fired
                   1.2
           Lb/M Btu
                   0.8
                   0.6
                   0.4
                   0.2
                           Single Wall Fired Units
Pre-NSPS  Burner
                                    (4)  110,000 Lb/Hr 4 Burner
                                    (5)  125,000 Lb/Hr. 4 Burner
                                    (6)  CETF
                                    (7)  500 MW Opposed Fired
                                                                      Opposed Fired Units
                               (5)
                                   (4)
                                              (6)
                                                         (3)
                                     (7)
                                                                 (2)
                                              (1)
                                                                    CF/SF Low NOx Burner
                                                        CF/SF  H- Advanced OFA
                           50    100   150   200   250   300   350   400   450
                               BURNER ZONE LIBERATION RATE
                                    (10 3Btu/Hr-Ft
                                                 Foster Wheeler Energy Corp.
                                            Combustion & Environmental Systems
                                         Fig. 3. NO^Reduction Summary

-------
   NOX
lb/106Btu
                                 KEY


                         O  ALL MILLS IN SERVICE

                            ONE MILL OUT OF SERVICE

                            TWO MILLS OUT OF SERVICE
                 300       350       400
                             LOAD (MW)
               450
                                                          500
             Fig. 4. Advanced Low NOx Retrofit 500 MW Boiler NO^ vs Load
  w  - -.
     X4 0
     —T • W
'Z'CXJ*1
£L   3.5

w   3.0
O
<   2.5
Z
a:   po
     D^> w
n,
O   1-5

X   1-0
O
m   0.5
                 0  ALL MILLS IN SERVICE

                    ONE MILL OUT OF SERVICE

                    TWO MILLS OUT OF SERVICE
          QVERFiREASR
                     300
350
                                            400
450
                                                                  OVERFIREAIR
                                                                      OPEN
500
          Fig. 5. Advanced Low NOx Retrofit 500 MW Unit Pressure Drop vs Load
                                     4A-73

-------
NOx
lb/106Btu













1.4 -
1.3 -
1.2 -
1.1 -
1.0 -
0|*S
.9
0.8-
0.7-
0.6-
0.5-

0.4-
0.3-
0.2-
0.1 -
UNCONTROLLED
_ sggssssgg £ =1 ,.,3
~ wxsxsas S = 0.06








REDUCTION
62.5% 78.6%


w
— CF/SF JL™«
LNB ^^•••i* ^ = 0.464








^
ONLY *KMKWKW g 0032 ^f
LNB JL^
AOFA"""^^ '
o — U.Ul^i
          CF/SF =  Controlled Flow/Split-FIame
          LNB  =  Low NOX Burner
          AOFA - Advanced Overflre Air
          LOAD = 500 MW; 02 = 3.5%

         Fig. 6. 500 MW Boiler NOx Reduction
Fig. 7. Combustion and Enviornmental Test Facility

                   4A-74

-------
;     0.40
Stu) 0 35 -
     0.30-
     0.25-
     0.20-
     0.15-
     0.10-
     0.05-
     0.00
            ELECTRIC SLEEVE
             DAMPER DRIVE
            MANUAL OUTER
            UGISTER DRIVE
                  Fig. 8. Internal Fuel Staged Low NO^ Burner7
               10
                    20
                          30
                               40   50   60
                                               70
                                                    SO
                                                          90   100
                                    LOAD (% MCR)
                    For all test points:    CO < 50 ppm ; O2 = 3.4 - 3.6%
                                         AP = 1.5" H2O Maximum
                                         No Overfire Air
               Fig. 9. CETF Testing Internal Fuel Staged Burner
 3.0   NG
 2.9
 2.8
 2.7
 2.6
-2.5
-2.4
-2.3
-2.2

 2.0
                                  4A-75

-------
NOx 1-4~
, c •! o
lb/10b Btu 1-3
1.2 -
1 1
i . i
1.0 -
.9
00
.0
0-7
. /
Or*
.O
Or-
.O
0.4"
0.3-
0.2 -
0.1 -



UNCONTROLLED
SS&89Kf&KK6
7 = 1 .23
S = 0.06
78.6% 62 5%



:>
CF/SF
NO OFA
»BS5«55e«BW^
flU
CF/SF
OFA -• OP
74.
S
— 0.464
S = 0.032
>
£ = 0.266 "
S = 0.012
EN
UNCONTROLLED
KMWWMQMOK ••
61%
6% 1
I CF/SF
^ NO OFA
^W 7 ri /i r\c
awOTMSBSKwaas W ~ U.fUO
v^^^^ S = 0.034
*= 7 =0.269
IPS S = °'015
NO OFA

500 MW 80 M Btu hr
Steam Generator Test Facility
Fig. 10. Low NOx Burner Comparative Data Controlled Flow/Split Flame vs Internal Fuel Staged
                                        4A-76

-------
-si
                       KEY:
Uncontrolled Levels
CF/SF - with AGFA
O  CF/SF - no OFA
.  IFS - no OFA
                                                                    CF/SF Low NOx Burner
                                                         CF/SF + Advanced OFA
                           50   100   150   200   250   300   350   400   450
                               BURNER  ZONE LIBERATION RATE
                                     (103 Btu/Hr-Ft2)
                                       Foster Wheeler Energy Corp.
                                  Combustion & Environmental Systems
                                      Fig. 11. NO, Reduction Summary IFS vs. CF/SF

-------
REDUCTION OF NITROGEN OXIDES EMISSIONS BY COMBUSTION PROCESS
       MODIFICATION IN NATURAL GAS AND FUEL OIL FLAMES:
            FUNDAMENTALS OF LOW NOV BURNER DESIGN
                                      A

                   M.A. Toqan, L. Berg, and J.M. Beer

                   Massachusetts Institute of Technology
                        Cambridge, MA 02139

                   A. Marotta, A. Beretta and A.  Testa
                             Eniricerche
                                Italy

-------
                     Reduction of Nitrogen Oxides Emissions by
        Combustion Process Modification in Natural Gas and Fuel Oil Rames:
                     Fundamentals of Low NOX Burner Design

                        M.A. Toqan, L. Berg, and J.M. Beer

                       Massachusetts Institute of Technology
                              Cambridge, MA 02139
                        A. Marotta, A. Beretta and A. Testa

                                   Eniricerche
                                       Italy
Abstract
    Increasingly tight environmental regulations for NOT emission from coal-, oil- and gas-
fired utility boilers are forcing utility  and industrial users of fossil fuels to pay greater
attention to control of NOX in oil, coal and even gas-fired units.  Effective control of NOX
emissions requires the application  of  one or a combination of methods of combustion
process modification  including staged  air and staged fuel injection, the use of low-NOT
burners, and possibly even post-combustion clean-up such as NH3 injection into combustion
gases.

      To date, the degree of NOX reduction achieved by these  technologies has been
observed to vary widely and to depend on the combustion system in question.  Primarily
this wide variation in performance of staged systems and of low NOX burners is due to lack
of understanding of the overlapping processes of the  nitrogen-hydrocarbon chemistry and
the mixing/temperature histories of the fuel in the flame.

      To address the variation of performance of low-NOx burners a theoretical and an
experimental  investigation is being  carried out at  MIT which is focused  on the
fundamentals of low-NO^ burner design,  applied to natural gas and fuel oil combustion.

      In the experimental investigation a multi-annular burner developed by Massachusetts
Institute of Technology is used. Flame studies are carried out in the 1.2 m x 1.2 m x 4.5 m
test section of the MIT Combustion Research Facility  (CRF).  The CRF is a pilot plant
scale combustion  tunnel, having a  3 MW^ multi-fuel firing capability and  designed to
facilitate detailed investigation of industrial type turbulent diffusion flames. The burner is
equipped with a fuel gun surrounded by primary, secondary and tertiary air supplies.  Mass
flow rates for each of the three air supplies external to  the fuel gun can be independently
controlled, and for each supply the swirl can be adjusted over a wide range by means of an
independent  moveable block swirler.   A shroud-diffuser is used to maintain physical
separation of the secondary and tertiary air jets entering the combustion chamber.

      The results from this investigation  are pertinent to the design principles of low-NOx
burners including  scaling criteria.
                                      4A-81

-------
INTRODUCTION

      The most widely used design strategy for NOX reduction is staged combustion. The
creation of fuel-rich and -lean combustion zones in flames by means of staging the input
of either air or fuel is a successful method of NOX emission control.  Nevertheless, the
degree of NOX reduction achieved by these technologies has been observed to vary widely
and to depend on the combustion system in question (1,2,3).

      While  the principles  of staged  combustion control of NOX emission  are well
established  (4,5,6), their practical realization  is hampered by lack of information on the
overlapping processes of the nitrogen-hydrocarbon chemistry and the mixing-temperature
history of the fuel in the flame. The problems are especially difficult in the case of the so
called "internal staging" process, in which the fuel-rich and -lean combustion zones must
be produced  by  appropriate  fuel-air  mixing in a single low-NOx burner, rather than by
producing fuel- rich and -lean combustion zones in the combustion chamber using overfire
air.

      This problem is addressed in an experimental  research project at MIT focused on
the principles of low-NOx burner design as applied to natural gas and oil combustion.  This
paper reports results obtained with both natural gas and No. 2 and No. 6 oil as fuels. The
experimental  RSFC (Radially Stratified Flame Core) burner was developed at MIT based
on the patented design of a multiannular burner (7). The  RSFC burner is attached to the
flame tunnel (3 MW^, 1.2 x 1.2 x 4.5 m) of the MIT Combustion Research Facility (CRF).
Parallel with  the experiments a mathematical  modeling study is carried out;  the progress
of combustion along the flame is computed for the effects of design and operating variables
of the burner, using the "Fluent"  fluid  dynamics code.  In this paper, the  relationships
between burner input parameters and emissions of CO  and NOX are reported.

EXPERIMENTAL

      The  MIT  Combustion Research Facility was designed to permit detailed in-flame
measurements of the flow field and of spatial distributions of temperature and chemical
species concentrations to be made. Variable heat extraction along the flame - by the use
of completely and partially water cooled furnace sections - is used to closely simulate large
scale flame systems. Access to the  flame for optical or probe measurements is provided by
a 1.0  m  long slot at the burner and by instrument ports at every 30 cm length  further
downstream along the flame  tunnel.

      The  experimental RSFC burner  the concept of radial flame stratification

      The  burner consists of three concentric annuli with each of the annular nozzles at
a larger radial position extending further in the axial direction (Fig. 1). Fuel  is introduced
in the center  through a fuel gun. The three sections of the burner can be axially  adjusted
as may be required to maintain  optimal geometry at turn down. Additional features of the
burner include independently variable swirl control in each annular air nozzle by means of
IFRF moveable block  swirlers.
                                      4A-82

-------
CO
00
           TERTIARY AIR


          SECONDARY AIR-
                                  3?
                  11
                                      r
                                                                     BURNER QUARL
T
AIR
                       STEAM
                                FLUE GAS
                              PRIMARY AIR
                                               MOVABLE SHROUD


                                                 ^
                                                                    BURNER QUARL
                                               SWIRL GENERATOR
                     Figure 1. Schematic of Low-NOx Radially Stratified Flame Core Burner

-------
      The  operation of the burner  is based on the principle that a combination of a
positive radial density gradient and rotating flow field has a stratifying effect on the flame
by virtue  of the damping of turbulence at the interface of the central flame zone and the
colder  air flow radially surrounding the flame  (8).  This feature was chosen to give the
name:  Radially Stratified Flame Core (RSFC) Burner. The flame so produced would have
mixing zones in the opposite sequence to the  Flame type 1. used in the EFRF terminology,
as it would start with a narrow fuel jet flame of low air entrapment followed by an
internally recirculating flame region in which combustion is completed (Fig. 2).

      The  flexible design  of the experimental burner permitted the variation of input
parameters  over wide ranges. The burner parameters were:

       Fuel jet velocity and angle
       Fuel gun position (relative to the face of the burner)
       Air distribution to Primary, Secondary and Tertiary air flows
       Radial distribution of the swirl velocity hi the air flow
       Axial separation of the Primary, Secondary and Tertiary air flows

       Experimental Matrix

       Using natural gas and No. 2 and No.  6  oil and preheated combustion air (450°F)
parametric flame experiments were carried out with the RSFC burner. In the experiments,
burner input parameters were varied to determine their effect upon NOj and CO emissions.

       The ranges of variables adopted were:

            Fuel jet velocity:                50 - 600 ft/sec.
            Fuel jet angle:                  0° - 25°
            Fuel gun position:               -45 - 0 cm
            Primary air flow rate:            0 - 100%
            Secondary air flow rate:          0 - 100%
            Tertiary air flow rate:            0 - 100%
            Swirl number of primary air:     0 - 2.79
            Swirl number of secondary air:    0 - 1.90
            Swirl number of tertiary air:      0 - 1.39

       Elemental analyses of the No. 2 and No. 6 fuel oils are listed in Table 1.

       Measurements

      Temperature, and gaseous concentrations of CO, CO2, NOX and O2 were measured
at the exit of the combustion tunnel.
                                      4A-84

-------
     TERTIARY AIR
 FUEL GUN-:
 PRIMARY AIR

     SECONDARY AIR
                     EXTERNAL RECIRCULATION
                     ZONE
FLAME ENVELOPE

INTERNAL

RECIRCULATION
ZONE
Figure 2.  Schematic of a Radially Stratified Low NOX Flame
                                4A-85

-------
                                         Table 1
Ultimate
Analysis
C
H
N
Ash
H2O
Asphaltene
Content
Heating value,
(Btu/lb)
Weight %
No. 2
86.80
12.44
0.14
0.01
—
—
19,640
No. 6
86.46
9.67
.53
.08
.4
9.5
18,236
EXPERIMENTAL RESULTS

       Natural Gas Combustion

       In  the N.G. tests,  98 flames were investigated for the effects of burner input
variables upon NO^ and CO emissions from the combustion tunnel.  The input variables
found to have effect upon  NOX and  CO emissions are:

            type of fuel nozzle
            fuel gun position within burner
            primary air fraction
            radial displacement of swirl  from flame axis

       Type of Fuel Nozzle

       Two parameters,  the exit velocity of the fuel jet from the gun and the angle of the
jet relative to the flame axis, were considered in the design of the fuel nozzles.   Several
nozzles were built to allow  the velocity of the fuel to range from 50 ft/sec, to 600 ft/sec, and
the angle  to vary from 0° to 25°. Results obtained from the combustion  tests with these
nozzles are shown in Fgures 3 and 4. It is noteworthy that while CO emission levels were
very low for all cases, they increased  slightly with increasing  fuel jet velocity. On the other
hand, NOX emission levels were more influenced by the fuel jet angle: i.e., increasing the
fuel jet angle from 18 to 25° increased NO^ concentration at the exit by ~ 25 %.
                                      4A-86

-------
             NATURAL GAS
E
a.
a.
cc
i-
z
UJ
O
Z
o
o

o
o
1 UU"
90-
80-
70-
60-
50-

40-
HU
30-

20-
10-
n

m^_____ 	 •*
	 -^^ NOX




» »
CO r — — —
% Swirl No.
Prim, air low high
Sec. air zero
Tert. air high high
- i uu
-90
-80
-70
-60
-50

-40
HU
-30

-20
-10
.n


Q.
•z.
o
h-
CC
1-
Z
UJ
o
z
o
o
X


    20  30  40
50  60  70  80  90  100  110 120 130

FUEL JET VELOCITY (FT/SEC)
EFFECT OF FUEL JET VELOCITY ON NOx AND CO

        EMISSIONS (O2 at exit = 1.5 %)
                   Figure 3.
                    4A-87

-------
  100-

— 90-

a 80

z 70-
O
H 60
DC
f-
Z
LU
O
Z
o
o
o
o
50-

40-

30-

20-

10-
   0
            NATURAL GAS
% Swirl No.
Prim, air low high
Sec. air zero
Tert. air high high N ^)
y\

^ 	 *" 	 *
H 	 — •


I i I
_1 	 1 P

CO


eL.\J\J
-180
-160

-140
-120
-100

-80

-60

-40
-20
n
E*
Q.
Q.

Z
g
j^
QC
K-
Z
01

O
O
X
O
z

         0
                 15

             FUEL JET ANGLE
25
EFFECT OF FUEL JET ANGLE ON NOx AND CO

       EMISSIONS (O2 at exit = 1.5 %)



                  Figure 4.
                    4A-88

-------
      Fuel Gun Position

      The axial position at which the fuel is introduced within the burner is known to be
important in determining the flame structure.  Fluid dynamically it affects the interaction
of the axial fuel jet and the swirling annular  air flow.  To investigate the effect of this
parameter upon NOX and CO emissions, several flames were investigated in which the
location of injection  of fuel  within the  burner was  varied.   Figures 5 &  6  illustrate the
effect of this variable for the cases of strongly and weakly swirling primary air.  The
negative values of the fuel gun positions shown in Figures 5 & 6 indicate the distance
between the end of the burner face and  the fuel gun tip.  A negative value implies that the
gun has been retracted into  the burner throat It can be seen from Figures 5 and 6 that
the fuel gun position has little effect on  NOX concentration.  However, CO emissions were
observed to increase  dramatically when the fuel gun was moved  in the burner for certain
burner configurations.

      Primary air fraction

      The conditions represented in Fig. 5 with 51% of the air supplied as primary air give
high  NOj values,  ranging  from  110 to  135  ppm, while  CO  concentrations  are
understandably low because  of the early aeration of the fuel.  It is the case illustrated in
Fig. 6 that deserves further discussion.  With the low primary fraction, NOX levels are La
the range of 75 to 85 ppm which shows  that even at a low level of swirl in the primary air,
fuel/air mixing is damped in  the near field.  However, as the primary air  fraction is raised
as illustrated in Figures 7 and 8, NOX emission levels increase due to the early mixing of
the fuel with the combustion  air.   It is noteworthy that for the cases  which have  low
primary air fraction, the lean stage mixing further downstream is inefficient without strong
swirl in the tertiary air.  For the  condition of high  swirl degree  of the  primary air, NO
concentration is mainly  dependent upon the  primary air fraction.  The CO emissions,
however, are more dependent upon the swirl degree of the secondary/tertiary air.  For the
cases in Fig.  8 the CO  concentration  remains virtually constant over the full range of
primary air flow fraction as long as the tertiary air has a high degree of swirl. An optimum
flame was found in which the burner input conditions reflect the above trends: low primary
air fraction, with high swirl, high secondary mass flow fraction with over critical degree of
swirl, and low tertiary air flow with no swirl, ( NO emission at 3%  O2: 70 ppm; CO: 56 ppm
and the O2 concentration in the exhaust  1.85%).

      Similarly favorable conditions were obtained with low primary, low secondary and
high tertiary air flows as long as swirl was imparted both  to the primary and the tertiary air
flows.

      Radial displacement of swirl from flame axis

      With the multi-annular burner  it is possible to produce a wide range of different
types of swirling  flows.  The two extreme cases  are the  free and  forced vortex flows.
Assuming a uniform  axial velocity profile, free vortex flow is obtained by high swirl in the
primary air, low swirl in the secondary air and no swirl in the tertiary air.  A forced vortex
                                       4A-89

-------
           NATURAL GAS
^uu-
-^ 350-
Q.
3 300-
Z
2 250-
|E 200-
z
0 150-
z
o
0 100-
o
0 50-
n-
% Swirl No.
Prim, air medium high
Sec. air zero
Tert. air medium high

H ~ 	 	 ~~* 	 —————4.
NOx


CO »_— _ »— — 	 	 "


-------
  E
  Q.
  Q.
 O
 DC
 h-
 UJ
 O
 Z
 o
 o
 o
 o
                NATURAL GAS
    1000
900-
800-
700-
600-
500-
400-
300-
200-
100-
      %
Prim.air low
Sec. air low
Tert air high
Swirl No.
  low
  low
 zero
      0
      -50
       -45   -40  -35   -30   -25   -20
             FUEL GUN POSITION (CM)
          -15
                                          100
-90
-80
-70
-60
-50
-40
-30
-20
-10
                    E
                    Q.
                    a
g
h-
QC
Z
01
O
O
O
o
        0
      -10
EFFECT OF FUEL GUN POSITION ON NOx AND CO
         EMISSIONS (O2 at exit = 1.5 %)

                       Figure 6.
                         4A-91

-------
             NATURAL GAS
E
Q.
_O.
Z
O
H
<
rr
i_
r^
z
01
o
0
o




iuu-
90-
80-
70-

60-
50-
40-

30-
20-

10-
o-
NOX

^ — """^
^


CO ^^-^^
^~"^\^^
% Swirl No.
Prim, air - high
Sec. air - high
Tert. air zero

i | | [- 	 1 	
-
-------
E
Q.
CL
Ol
O
-z.
O
O
O
O
  100
80-
p  60-
40-
20-
    0
    0.0
                NATURAL GAS
                          PRIM AIR
                          SEC AIR
                          TERT AIR
              0.4           0.8
            FRACTION OF PRIMARY AIR
Swirl No.
HIGH
HIGH
 200
-180
-160
-140
-120
-100
-80
-60
 40
 20
 0
1.2
     E
     Q.
     a.
     o
     (J
     x
     O
EFFECT OF (PRIM./TERT.) AIR RATIO ON NOx AND
           EMISSIONS (O2 at exit = 1.5%)
                                                CO
                      Figure 8.
                        4A-93

-------
swirling flow is obtained by high swirl in the tertiary air, a lower swirl in the secondary air
and a no swirl in the primary air.  In a Rankine type vortex a forced vortex in the core of
the rotating flow combines with a free vortex on the outside. A Rankine vortex can be also
produced by the appropriate adjustment of the radial distribution  of the swirl velocity in
the RSFC burner.  To investigate the  effect of the radial displacement of the peak swirl
velocity (tangential velocity) component in the combustion air from the flame axis, several
flames were generated by imparting varying swirl degrees to the primary, secondary and
tertiary air jets.  The effect of this parameter on NOX concentration is illustrated in
Figure 9. It can be seen that lowest NOX emission was obtained with a Rankine vortex type
swirl velocity distribution.

       No. 2 and No. 6 Oil Combustion

       Based on the  experience gained  from natural gas combustion, 41 flames were
investigated for the effect of burner input parameters upon NOX and CO emission levels.
The parameters varied included:

       (1)    Type of fuel nozzle
       (2)    Primary air fraction
       (3)    Radial displacement of swirl from flame axis

       Type of fuel nozzle

       Six hole  "Y" jet twin fluid atomizers were used and the angle between the six fuel
jets and the axis was varied to range from axial (0°) to 25° half angle. NOX concentrations
obtained using these atomizers for several primary/secondary/tertiary air ratios and several
swirl  numbers  are  shown in Figures  10 and 11.   The experimental  results  show  the
importance of the fuel jet angle upon NOX emission levels. For No.  2 oil, changing the fuel
jet angle from axial (0°) to 25° raises NOX emission levels from 54  to 90 ppm for the best
cases  examined and from 90 to 250 for the worst cases.  For No. 6 oil, using a 25° atomizer
instead of an axial  (0°) nozzle has the effect of raising NOT concentration levels at the exit
of the combustion  tunnel from 97 ppm to ~ 170 ppm  for the best cases and from ~ 240
to ~ 300 ppm for the worst cases.

       Primary Air Fraction

       The effect of the primary fraction ratio upon NOT emissions is shown in Figure 12.
It is noteworthy that for both fuel nozzle types when the primary  air  fraction  constitutes
~ 0.5  of the total combustion air or higher, NOX emissions are high i.e., ranging from 86
to 140 ppm for No. 2 oil and ranging from 235 to 330 for No. 6 oil.  As the combustion air
is diverted towards the tertiary air port, je concentration of NOX measured at the exit for
the axial (0°) nozzle  drops to 59 ppm and 105 ppm for No. 2 and No. 6 oils, respectively.
A similar trend is observed with the 25° fuel jet atomizers.  However, the lowest emission
level  achieved with the latter could not match that  obtained with the narrow angle fuel
nozzle.  These results are  similar to those obtained with  natural  gas. It is concluded that
only a small fraction  of the  combustion air has to be  supplied  around the fuel spray to
                                       4A-94

-------
130
  0
          NATURAL GAS
          1            2
         OVERALL SWIRL NUMBER

EFFECT OF TYPE OF VORTEX AT BURNER EXIT ON
          NOx EMISSIONS
                Figure 9.
                 4A-95

-------
                No.2 OIL
  250
                2           3
                  CASE NUMBER
EFFECT OF FUEL JET ANGLE UPON NOx EMISSIONS
                  Figure 10.
                    4A-96

-------
                N0.6OIL
  350
             1
                  CASE NUMBER



EFFECT OF FUEL JET ANGLE UPON NOx EMISSIONS







                  Figure 11.
                    4A-97

-------
                               NOx
CO
   g

   <  200H
   UJ
   o
   z
   O
   O
 Prim, air frac.  Low  High  Low  High  Low High  Low  High
Fuel Jet Angle
                            25V
         25°
   EFFECT OF PRIMARY AIR FRACTION ON NOx AND  CO EMISSIONS

                      FROM  OIL COMBUSTION
                             Figure 12.
                              4A-98

-------
ignite the fuel and stratify the flame.  The remaining fraction can then be introduced as
tertiary air.

      Radial displacement of swirl from flame axis

      Using narrow angle fuel sprays, combustion experiments were carried out with No.
2 fuel oil to investigate the effect of radial displacement of peak tangential velocity of the
air upon NOX emissions.

      In Figures  13 and  14 the effects of the swirl degree,  the radial distribution of the
swirl velocity of the combustion air and the angle of the fuel  spray upon NOX emission are
illustrated.  NOX emission is seen to decrease with increasing swirl number and "RanMne"
vortex is favored for the air flow  and a narrow angle spray for the fuel oil. Minimum NOX
levels were 97 ppm for No. 6 fuel oil and 54 ppm for No. 2  fuel oil.

      NOX emissions as a function of combustion air swirl vary differently in gas and oil
flames with the RSFC burner. In contrast to the gas flames  in which NOX emissions were
increasing as the  swirl  degree was raised above its critical value for vortex breakdown
(S ~  0.6), they continued to decrease in both No. 2 and No. 6 oil flames for much higher
values of the combustion  air swirl.  This difference is thought to be due to the lower
entrainment rate and higher penetration depth of the narrow angle fuel spray compared
to the gas jet.  As a result of this,  radial stratification of the flow can be maintained to a
higher value of the swirl number in the oil flames than in the gas flames.

      The  lowest levels  of NOX  and CO emissions  obtained  by purely  combustion
aerodynamic means are shown in Fig. 15.  In recent experiments in which gas recirculation
and steam injection were used, significant reductions were reached in NOX emission levels
while maintaining  low CO concentration at the exit of the combustion tunnel.  These
results will be reported in our paper prepared for the 1991 Fall Meeting of the American
Flame Research Committee.

CONCLUSIONS

      An experimental  investigation  has been  carried out with  a flexible experimental
low-NOx burner of novel design. The burner is designed to achieve staged combustion by
a combination of radial flow stratification and axial air staging in the flame. Fuel/air mixing
is suppressed by radial flow stratification close to the burner but is then promoted by a
toroidal recirculating  flow further downstream of the burner.

      Parametric experimental  studies  carried  out in the  flame tunnel  of  the  MIT
Combustion Research Facility (CRF) permitted optimization  of the burner for low-NOx and
CO emissions by determining favorable conditions for the radial distributions of the air flow
and the swirl velocity at the exit from the burner and for the central fuel injection velocity
and angle.  The  results showed that for several operational modes of the burner input
variables, highly  stable flames with low-NOx and  CO emission levels were attainable.
Minimum  values  of NO. and CO  emissions  obtained  by the optimization  of  the
                                      4A-99

-------
                   No.2 OIL
  100-



   90-



o  aoH

5?
   70-
S
Q.
0.
o  6°H
z

   50-
   40
                       RANKINE VORTEX
     0   0.2   0.4  0.6   0.8   1    1.2   1.4   1.6   1.8

                    SWIRL NUMBER
   100-
   90-
S  8o^
I  ^
O  60-
   50-
   40
                          FORCED VORTEX
     0   0.2   0.4  0.6   0.8   1   1.2   1.4  1.6   1.8

                     SWIRL NUMBER
   EFFECT OF TYPE OF VORTEX ON NOx EMISSION
                    Figure 13.
                      4 A-100

-------
  400


  350


(7 300'
O

n 2501

Q.
t 200

s
z ISO-


  100-
50
 0.5
                FREE VORTEX
                         10 degrees
               1.5      2
              SWIRL NUMBER
                             2.5
                                                              FORCED VORTEX
                         25 degrees

                         10 degrees

                          0 degrees
0  0.2 0.4  0.6  08  1   1.2  1.4  1.S  1.8

             SWIRL NUMBER
                     400
                     350-
                   fj 3001
                   O

                   o 2501

                   a.
                   £. 200

                   O
                   z 150
                     100


                     50
                            0.5
                                         RANKINE VORTEX
    25 degrees

     0 degrees

    10 degrees
                                   1     1.5     2
                                    SWIRL NUMBER
                                                   2.5
Figure  14.  Effect of Type of Vortex Upon  NOX  Emission
                                    4A-101

-------
                 No. 6 Oil
                     NOx BB CO
E
Q_
Q.
O
cc
LU
O
•z.
O
O
  120
  100-
   80-
   60-
40-
   20-
            N.G.        No.2        No.6

    NOx AND CO EMISSIONS FOR AN OPTIMUM

          BURNER CONFIGURATION
                     Figure 15.
                      4 A-102

-------
aerodynamic variables (staging by flame stratification) were 70 ppm NOX and 56 ppm CO
for natural  gas, 54 ppm NOX and 23 ppm CO for No. 2 fuel oil and 97 ppm NO^ and
32 ppm CO for No. 6 fuel oil.

      The data reported here  are initial results only; recent experiments presently under
evaluation show that further significant reductions of NOX emission can be obtained with
the new radiaUy stratified flame core (RSFC) burner by means of the controlled admixing
of small amounts of recirculated  flue gas and steam.

ACKNOWLEDGEMENTS

      Financial support from  a consortium of utility companies:  Empire State Electric
Energy Research Corporation (ESEERCO), Eniricerche, Electric Power Research Institute
(EPRI), Ente  Nazionale  Per  L'Energia Elettrica  (ENEL),  Florida  Power and Light
Company and Southern California Edison Company is gratefully acknowledged. We thank
ABB-CE  for their expression of  readiness in producing a commercial burner should our
experiments be successful.  The authors are  indebted  to Ms. Bonnie Caputo  for the
production of the many drafts and the final form of the paper.

REFERENCES

1.  Morita, S., Kiyama, K., limbo, T., Hodozuka, K_, and Mine, K, "Design Methods for
    Low-NO,,  Retrofits  of Pulverized  Coal  Fired  Utility  Boilers",  EPRI/EPA Joint
    Symposium on Stationary Combustion NOX Control, Mar, 1989.

2.  Lisauskas,  R.A., Reicker, E.L., and Davis, T., "Status of NOX Control Technology at
    Riley Stoker", EPRI/EPA  Joint Symposium on Stationary Combustion NOX Control,
    Mar, 1989.

3.  Thompson, R.E., Shiomoto,  G.H., Shore, D.E., McDannel, M.D., and Eskinazi, D., "
    NOX  Emissions Results for a  Low-NOx  PM Burner  Retrofit" EPRI/EPA Joint
    Symposium on Stationary Combustion NOX Control, Mar, 1989.

4.  Sarofim, A.F., Pohl, and  Taylor,  B.R., "Strategies for Controlling  Nitrogen  Oxide
    Emissions During Combustion of Nitrogen Bearing Fuels, "AIChE Symposium Series
    No. 175. 74, 67, (1978).

5.  Farmayan,  W.F., M.Sc. Thesis, "The Control of Nitrogen Oxides Emission by Staged
    Combustion,"  Department  of Chemical  Engineering,  Massachusetts  Institute of
    Technology, Cambridge, MA, April  1980.

6.  Beer, J.M., Jacques, M.T., Farmayan, W.F., Gupta, A.K., Hanson, S., Rovesti, W.C.,
    "Reduction of NOX and Solid Emissions by Staged Combustion of Coal Liquid Fuels."
    Nineteenth Symposium on Coal Liquid, Haifa, Israel, Aug. 1983.
                                    4 A-103

-------
7.   Beer, J.M. British Patent No. 1099 959.  Jan. 17, 1968; U.S. Patent: "Low NO^ Rich-
    Lean Combustor Especially Useful in Gas Turbines", Jul 11, 1989.

8.   Beer,  J.M.,  Chigier, N.A.,  Davies,  T.N. and  Bassindale,  K.: "Laminarization  of
    Turbulent Flames in Rotating Flow Environments" Combustion and Flame, Vol. 16.
    pp 39.45 (1971).
                                      4 A-104

-------
    DEVELOPMENT OF LOW NOX GAS BURNERS

      Shyh-Ching Yang, John H. Pohl, Steven J. Bortz,
            Robert J. Yang, Wen-Chen Chang

            Energy & Resources Laboratories
         Industrial Technology Research Institute
               Hsin Chu, Taiwan, R.O.C.
        W. J. Schafer Associates, Irvine, CA 92718
R-C Enviornmental Service & Technologies, Irvine CA 92718

-------
                 Development of  Low  NOx Gas  Burners

              Shyh-Ching Yang1, John H. Pohl2, Steven J. Bortz3,
                      Robert J. Yang1, Wen-Chen Chang1

                       Energy  & Resources  Laboratories
                  Industrial Technology  Research  Institute
                           Hsin Chu, Taiwan, R.O.C.
                 2W. J.  Schafer Associates, Irvine,  CA 92718
        3R-C Environmental Service &  Technologies, Irvine, CA 92718

ABSTRACT

The Energy  Commission, Ministry  of Economic Affair(EC, MOEA),
Republic of  China, has  a  program  to develop  2.5MW low  NOx  gas
burners.   This paper reports  the  results on  premixed and
nonpremixed  swirl  burners.

The  versatile premixed and nonpremixed  swirl  burners were  designed,
fabricated and  tested in Energy and Resources  Laboratories'(ERL)  10
xlO^  Btu/hr furnace.  It is shows that the NOX  can  be controlled to
levels  of  less than  15ppm.  The peak  flame temperatures  required to
maintain required NOX levels  were achieved by mixing  sufficient flue
gas,  and/or  other  diluents and  well control the  mixing  rate.

The ERL's  test  furnace is designed to  simulate  single  burner  industrial
boilers.  The furnace  is  1600mm square inside,  and is lined with
500mm thick refractory for the first 2500mm.  The last 3840mm of the
furnace is water  cooled.   The  furnace is  built of 250mm segments and
operated with residence time  at 2-3 sec  and 5-10  mm water  positive.

The results  showed  that dilution of the  premixed gas  by air or flue gas
could  reduce NOX emission to  12.5ppm(dry,  3% 02) and is relatively
temperature  insensitive for premixed  burner.   The  results  of non-
premixed  swirl burner  showed  that turndowns of 5/1,  with CO
emissions  less than 50ppm, O2  1   2.5%, NOX emissions  about
10-12ppm were  achieved   with  the  conditions of flue gas  recirculation
of 15%,  and primary zone stoichiometry of 0.7.
                                     4 A-107

-------
INTRODUCTION
Recently,  international  attention has increased to the role of  nitrogen
oxides in lake acidification,  oxidant formation  and forest  damage.
Stringent  regulations to reduce  the  allowable  emissions of nitrogen
oxides are being  promulgated in many industrial area of  the  world.
Under the clean  air  regislations, the combustion  industry  is being  faced
with  the  necessary of  having to reduce nitrogen  oxides  from  its existent
units.   Combustion  modifications(i.e. low-NOx  burners,  reburning  or
staged combustion)  generally  afford the  least  capital for  achieving
reductions(l).

This  paper  summerizes  the  results  of  a program  funded by the Energy
Commission Ministry of Economic Affairs(EC,MOEA) Republic of China to
develope  and to  determine  the  ability  of burner designs  and  operating
conditions to  achieve the  low levels required and expected for NOx
emissions.

To achieve this  objectives,  chemical reaction  and  aerodynamics control
of the flame  were used for burner  design (2,3,5,6,7).  In  this  program,
four  burner  designs, premixed  (1),  partially premixed  (4),  nonpremixed
fuel  jet and  swirl burner  (3), were  tested in Energy  and Resources
Laboratories'  10  x 106  Btu/hr furnace.
EXPERIMENTAL
The Energy  and Resources  Laboratories'  test furnace is nominally  rated
for  lOxlO6  Btu/hr  and is  designed to simulate single  burner industrial
boilers.  A  schematic of the furnace is shown in Figure  1.   The furnace is
1600mm square inside,  and is lined  with 500  mm thick refractory  for
the first 2500 mm.  The last  3480 mm of the furnace is  water cooled.
                                     4 A-108

-------
The  furnace  is  built  of 250 segments.   Each segment has a rectangular
port  for veiwing  and probing the flame. The  furnace is  operated  with
force draft at 5    10 mm water  positive.

The  furnace  is  equipped with  a  rear  port  for flame  photography  and
measuring  the exit concentration  of  the  flue gas.   The  wall  are
instrumented with thermocouples  to monitor the  change in wall
temperature  and to  measure the  heat extraction  in  the  wall-cooled
sections.   The furnace is also  equipped with a number  of  sheathed
thermocouples  on  axis  to  monitor  the  relative  flame  temperatures.

Flue gas is drawn by an induced  draft fan  from  the stack.  The  flue gas
is cooled in  a direct spray  tower, monitored by  an orifices plate and fed
to the burner.   The  direct spray  tower cools the flue  gas to!50°C
(300°F). Use of the direct  spray  will increase the  water  content of the
heat capacity, and to a limited extent heat which can  be extracted from
the flame  by the  flue gas.

Gases  are  drawn  from  the  furnace  through a water-cooled probe  for
continuous analysis.  The gases are  analyzed for  CO and  CO2 with non-
dispersive  infrared analyzer, for  NO  and  NO2 with  chemiluminescence
analyzer, and for oxygen  with  an electrochemical  analyzer.  A suction
pyrometer  was  used  to  measure  the  temperatures  in the flame and
sheathed thermocouples were used  to continuously track changes in the
lower  temperature regions  of  the furnace.

The  versatile premixed, partially premixed, nonpremixed fuel  jet  and
swirl flow burners were designed, fabricated and  tested  in E&RL's
furnace.  Those configurations  are shown in Figures  2  and 3.

      • Premixed Burner (Figures 2 and 3a)
In this  configuration nature gas is introduced  from  the  bottom,  is  mixed
                                     4 A-109

-------
in a chamber  with air. Flue gas,  if used, is drawn by a fan from the exit
of the furnace,  cooled to  150°C  by a  water  spray, measured with  an
orifice meter, and fed radially into  the chamber.   The rear nature  gas
chamber is  isolated from  the  air  chamber by two  layers of  1mm
opening  screen  to  help prevent flash backs.   The  face  of the  burner is
covered  with  two  layers of 1mm opening  screens  to  prevent flash back
from  the  furnace.  The burner is  set within  a cast refractory quarl.  The
quarl  is  equipped  with a  pilot flame and a flame  detector.

      • Partially  Premixed Burner (Figures 2 and  3c)

In this  configuration,  nature  gas is  introduced  from the  bottom, is mixed
in chamber  I  with air. The mixed air  and flue  gas is fed to the chamber
II through  a perforated  shroud.

      • Nonpremixed  Fuel Jet Burner (Figures  2 and 3b)
In this  configuration, nature gas is  introduced  from the bottom.   Flue
gas,  if used, is  mixed in chamber I  with  nature gas.  The air is fed to the
chamber  II  through a  perforated  shroud.

        •  Swirl Burner (Figure 3d)

In this  configuration, nature gas is  introduced  in  the  central gas  gun.
The  air  is  drawn through  a swirl  generator to generate a  low pressure
drop, low turbulence swirling flow  to  achieve the desired  ignition, flame
geometry and burnout  characteristics for  a  given fuel.   The  flue  gas
recirculation and  air staging were  used in  this testing.
RESULTS
These versatile burners were  tested  in  the  E&RL's  10x10^  Btu/hr
                                      4A-110

-------
furnace.   Results  from the tests are reported in this paper.
Measurements were made of gas composition (NOX, CO, CO2, and 02) and
flame temperature.   The  temperature  at  37.5mm  from the quarl was
measured using  a suction pyrometer;  other temperatures were
measured using sheathed  termocouples.   Data  was taken  under the
follow conditions:

      • Premixed and  Nonpremixed Fuel Jet Burners

              premixed flame  and  diffusion flame.
              load 3.9-6.6xl06  Btu/hr
              flue gas O2 2.0-9.0%(dry)
              flue gas  recirculation 0-10%
              mixing factor, M=l  for  premixed burner
              mixing factor, M=0  for  nonpremixed fuel jet burner

The results  of NOX emissions  and  flame  temperatures  are  shown in
Figure 4. The emissions of NOX decreased  with increasing or decreasing
oxygen  in case  of without flue gas recirculation.  Where ^ is the
percentage of the  flue gas recirculation,   ®r  is  the total  stoichiometric
and M is the mixing factor.   The mixing  factor  is  defined as  the  ratio of
air mixed with fuel  per total air introduced at  a given   ®r
            M= air  mixed with fuel/total  air introduced at  a given ®T
                                                                  (1)

For  premixed  burner, M=l  whereas  for  partially  premixed  burner,
The  influence of a "flame temperature"  is also shown  in  Figure 4.  The
"flame temperature"  used is  the value measured by  the suction
pyrometer  at 375mm from the quarl.   The NOx emissions decreased
with decreasing "flame  temperature".  For  load  of 3.9x10^ Btu/hr,  the
                                      4A-111

-------
NOx emissions  of  0.02  Ib/MM  Btu corresponds the flame  temperature
of 1245°C. The NOx emissions increases to 0.07 Ib/MMBtu  at  "flame
temperature"  increases to  1500°C.  The track of the NOX emissions  and
"flame  temperature"  are  similar.

The  relative  "flame temperature",   9 = T/TBAS,  and relative NOX
emissions, ^ = NOx/NOxBAS of %  =1.10-1.45  and   5  =0-10%
for premixed  and nonpremixed fuel jet burners are shown in Figure 5.
The  emissions of NOX and "flame  temperature"  decreased  with
increasing  the  amount of flue  gas recirculation.  The relations of  ^  and
& versus  ^   are
                       ^  = 1- 0.075 5               (2)
           and
                       6=1- 0.006 £              (3)
Where  ^  is in percent.  These results give the agreement of the  'H  and
© approach to 1 as  ^ equal  to  zero.

            • Partially Premixed Burner

              premixed  flame
              load 3.9-6.6xl06Btu/hr
              flue gas O2  1.5-4%(dry)
              flue  gas recirculation 0-10%
              mixing  factor; 0
-------
concentrations of a  given  flue gas  recirculation rate.   These graphical
solutions can  tell us the phenomena of NOx emissions of the partially
premixed  burner and  the  design  and  operation methodology  for  this
type of burners.   For example, the conditions of O2=2.1%, ^=4%;
O2=3.1%, 5 =6%;  and O2=6%, 5 =6% will give the same NOX emissions  of
0.03 Ib/MM Btu for the  load of  3.9xlQ6 Btu/hr.  Figures  8 and 9 give
the comparison  between  prediction  and experiment results for those
various burners.   The  agreement of these  values at various operation
conditions  is  very well.
It is  shown  that these burners have  been developed  to  meet the strict
emissions regulations,  NOX<0.03 Ib/MM  Btu, mandated by  the South
Coast Air Quality Management  District(SCAQMD).

            • Nonpremixed  Swirl  Flow burner
                    diffusion flame
                    load  5.9x106-10.0x106 Btu/hr
                   flue gas  O2  1.0-5.0%(dry)
                    flue gas recirculation  0-20%

The NOX emissions of the  swirl burner with  the  load  of  10.0 x 106
Btu/hr and flue gas recirculation  rate  of  0-10%  are shown in Figure 10.
The total stoichiometric,  ®T  is 1.05-1.2.  The emissions of NOX decreased
with  decreasing oxygen and/or increasing the  amount of flue  gas
recirculated.  The  NOX emission reduction can  be defined as
(4)
      RE(%) = (1 - XTX   ) x 100%
       S        NOXBAS
After regression  these data,  which gives the reduction  of NOX  emissions
versus flue gas  recirculation rate  is

      Rt(%) = £ (%)/( 133 0T- 1.15)        (5)
                                      4A-113

-------
The  NOx emissions at  the  load  of  5.9x10^ Btu/hr, without flue gas
recirculation  , are  shown in Figures 11  and  12. With  extrapolation  of
equation (5)  at each  recirculation rate,   ^ 5-15%, it  shows  that the
agreement  of prediction and  experiment of NOX  emissions is well.  The
NOx  emissions at  the  load of 6.6xl06 Btu/hr  and 8.7xl06 Btu/hr  with
air  staging of primary  stoichiometric 0.6-0.64  and with  4-5%  flue  gas
recirculation  are  shown in  Figure  13.   The total  stoichiometric  is 1.05-
1.3.  The NOX emissions for unstaged  are  0.03-0.048  Ib/MMBtu. The air
staging and  flue gas recirculation  can  provide a significant  further
reduction  of  the NOX emissions.   The minimum level of NOx  which could
be  achieved  was  10-12 ppm  .


CONCLUSIONS

Four versatile premixed,  partially  premixed,   nonpremixed fuel jet and
nonpremixed swirl burners were  designed,  fabricated and tested  in
Energy and  Resources Laboratories'  lOxlO^Btu/hr furnace.   Results
show  that 12.5ppm(3% O2)  can be achieved  either  by diluting the flame
with  air or  flue  gas to lower the  maximum flame temperature for  the
premixed flame  or delay  the mixing  and  to   lower  the  flame temperature
for the diffusion  flame.

This  paper has mentioned  the  minimum NOX levels for various
premixed and  diffusion flame burners.  Some results provide  the
concept of NOX  emissions reduction methodologies.    The  agreement
between the prediction and experiment  of NOx emissions for  various
burner and  operation  conditions are very well.

Use of excess air  and/or flue gas recirculation  to reduce NOX will  depend
on  individual circumstances of each boiler.  Increased of excess air in  a
premixed burner will  reduce NOx  emission,  will  not  require  installation
of  ducts  and high temperature  fans but will  reduce  boiler efficiency.
Use of flue  gas  recirculation and air staging of a swirl  burner  will reduce
NOx will  require  ducts and  high  temperature  fans, but will  not
drastically influence the boiler  efficiency.
                                      4A-114

-------
REFERENCES
1. Pohl, J.H.,  S.-C. Yang, C.-H. Chen and R.Yang,  "The Performance of Low
NOx Gas  Burner Configurations:  I. Premixed," October, 1990, American
Flame  Research Committee International  Symposium,  NOx Control,
Waste  Incinerators and  Oxygen Enriched  Combustion,  San Francisco,  CA,
U.S.A.

2. Pohl, J.H.,A.W. Bell, S.-C. Yang,  C.-H. Chen, and R.Yang," Development of
a Full Sized Ultra  Low NOx Industrial  Burner," 1990,April presented at
the American Flame Research Committee Members  Meeting, Tuscon,
Arizona U.S.A.

3. Bortz, S.J. and S.-C. Yang,"Development of a  Generalized Burner  Design
Procedure" October,  1990, American Flame  Research  Committee
International  Symposium,  NOx  control, Waste Incinerators  and Oxygen
Enriched Combustion, San  Francisco, CA,  U.S.A.

4. Pohl, J.H.  S.-C. Yang, R. Chang and R.  Yang, "The Influence of Burner
Geometry and  Operation on NOx Emissions:  II.  Partially
Premixed",January,  1991,ASME Third Fossil  Fuels  Combustion
Symposium  Houston, Texas, U.S.A.

5. Yang S.-C., R.-S. Juang,  W.-C.  Chang, and J.-S. Chen,  "Velocity
Measurements  and  Energy Distribution  for  Isothermal,  suddenly-
Expanding,  Swirling  Flow in  an Industrial Burner with Bluff-body",
Energy, 1_5,  NO.11,  pp.1015-1021,1990.

6.R.-S. Juang, S.-C. Yang,  W.-C. Chang, and J.-S.Chen, "Flow Characteristics
on  Isothermal  Sudden Expending  Swirling Flow in  an Industrial  Burner
with Bluff Body,"(Accepted by  J. of Chem. Eng.  of Japan).

7. Yang S.-C.,  et al., "Isothermal Swirling  Flow in the Expanding Quad of
Industrial  Burner with  a Bluff  Body", November  1989,Proceeding  of the
1989  International  Gas Research  Conference, Tokyo,  Japan.
                                     4A-115

-------
                   PREMIXED  BURNER
>

CD
                            UJ
                            £
                                                     AIR
                                                     FLUE GAS ,   -1-
                                                                                           250 MM
Figure  1   Energy and Resources Laboratories
         lOxlO6 Btu/hr  Test Furnace.
        Figure 2 Versatile Natural Gas  Burner

-------
                  PREMIXED BURNER
 FUEL'
               FLUE GAS
AIR
      Figure 3a Versatile Premixed  Gas  Burner
        NONPREM/XED   FUEL  JET  BURNER
FUEL
               FLUE GAS
AIR
     Figure 3b Versatile Nonpremixed  Fuel Jet Burner
                       4A-117

-------
            PARTIALLY PREMIXED BURNER
FUEL
                  AIR       AIR FLUE GAS
   Figure  3c  Versatile Partially Premixed Burner
          NONPREMIXED   SWIRL  BURNER
    Figure 3d Versatile  Nonpremixed Swirl Burner
                       4A-118

-------
   0.08  -
   0.07  -
   0.06
DQ 0.05
CD
 ^0.03
 00
   0.02
   0.01
                                                             1500
• O 3.9 X/06
A A 6.6 X/06
• El 3.9 XIO6
• O6.6X/06
        (.05
           (.15
                                     1.25
1.35
                                    0
                                                             1400
                                                                   0
                                                                   o
                                                                   CO

                     I3OO
                                                                        1200
                                                     1.45
                                      T
         Figure  4 NOx  Emissions  and "Flame Temperatures"
                   of the Premixed  and Fuel  Jet Burners.
                                  4A-119

-------
   1.0



   0.9


 oo

 QQ 0.8
 X
O
2
\

 x
O
0.7
0.6
0.5
0.4
0.3
    0
                  O
                  A
                  D
                  O
BTU/HR


 3.9X I06
 6.6X I06

 3.9XIQ6

 6.6XI06
                                     M
                                               0T  = I./O-
 1.45
                                        0

                                        0
                                        =   / -0.006
                                      =  1-0.075
                                        10
                                                  12
14
                                       (6
                                   (  %)


       Figure 5  Relative NOX Emissions and  Relative
                 "Flame Temperatures"  Versus  Flue Gas
                 Recirculation  Rate
               1.0


               0.98


               096   ^
                     00
                     h-
               0.94   \
                     h-

               0.92   »
                                                                     0.90
                                 4 A-120

-------
    0.07
    0.06
    Q05
m   a04
 X  0.03
O
    0.02
    0.01
              3.9X/ObBTU/HR
    0
               I       I
     PREDICTION


          M=0 —
I	I
        0
7      8
                                (0
                                       02
          Figure 6 Prediction of NOx Emissions for Partially

                   Premixed  Burner
                                   4A-121

-------
            6.6 XI06BTU/HR
                                                      PREDICTION
   0.07
   0.06
   0.05
QD
\ O04

QD
-J


 x 0.03

O
2

   0.02
   0-01
   0
       0
                                   I
456


     02  (%)
8
10
             Figure  7 Prediction of NOx Emissions for Partially

                      Premixed  Burner
                                    4A-122

-------

O
   0.09  -
   0.08  -
   0.07  -
   0.06  -
Q 0.05  -
0.
><  0.04
   0.03  -
   0.02  -
   0.01   -
        0
                  3.9XIO°BTU/HR 02(%)
                               0.69
                               0.66
                               0.69
                               0.66
                               0.73
                               0.69
                               0.68
                               0.70
                               0.69
O.O/    0.02   0.03    O.04    0.05   0.06   0.07   0.08

              NOX   (EXPERIMENT), (LB/MMBTU)
                                                                   0.09
            Figure 8  Comparison  Between  the  Prediction and
                      Experimental Results of NOX Emissions
                                     4 A-123

-------
   0.09
   0.08

   0.06
§ Q°5

O

S 0.04
Ct
CL


 Xo.03
   0.02
   001
               6.6XI06BTU/HR
             0.01    0.02   0.03   0.04  0.05   0.06   0.07


                    NOX (EXPERIMENT); (LB/MMBTU)
0.08   0.09
             Figure 9  Comparison  Between  the  Prediction and

                       Experimental Results of NOX Emissions
                                   4 A-124

-------
   003
   0.02
  .
 X
o
   o.o /
   o
1.0
                   IO.OXI06 BTU/HR ;

               O      f = 0 %
                D
10
                      NOX REDUCTION
             I       I
 NOX,BAS

I      I
                                   -JX/00%= g /(I.33&T-U5)
1.05
     1. 10
                                             I./5
1.20
         Figure  10 NOX Emissions of Swirl  Burner
                                 4 A-125

-------
     0.04
     0.03
te
\
QD
O
     0.02
     0.01
                    5.9 XI06  BTU/HR;  fi =
                  O
        NOX ,   , PREDICTION
                                  = I-0.01 £ /n.330T-/./5)
     0
                J	I
1.0
1.2
                                                1.3
1.4
                                  0T
               Figure  11  NOX  Emissions of Swirl Burner
                                  4 A-126

-------
   0.04 -
i
o
I-
o

a
   0.02  -
 X
Q  0.01
   0
                   0.01
                                0.02
0.03
0.04
                  NOX  | (EXPERIMENT);  (LB/MMBTU)





            Figure 12 Comparison  Between Prediction  and


                      Experimental Results of NOX Emissions
                                  4 A-127

-------
0.05
0.04
6.6X106 BUT/HR
   D UNSTAGED
   •PRIMARY STOiCH=o,60-o.64
   EPRIMARYSTOICH=0.60-0.64, 4-5% FGR


 I	—C0<20  PPM
                                             8.7X/06 BTU/HRI
                                             O UNSTAGED
0.01
0
    1.05
                 1.15
1.2
1.25
1.3
                             0
                               T
        Figure  13 NOX Emissions  of  Swirl Burner with Air

                 Staging and Flue Gas  Recirculation
                               4 A-128

-------
          Session 4B




LARGE SCALE SCR APPLICATIONS
   Chair:  E, Cichanowicz, EPRI

-------
UNDERSTANDING THE GERMAN AND JAPANESE COAL-FIRED SCR EXPERIENCE
                      Dr. Phillip A. Lowe
                          INTECH Inc.
                        11316 Rouen Dr.
                     Potomac,  MD 20854-3126
                      Mr. William Ellison
                      Ellison Consultants
                       4966 Tall Oaks Dr.
                       Monrovia, MD 21770
                     Mr. Michael Perlsweig
                   U.S. Department of Energy
                    Office of Fossil Energy
                      Washington,  DC  20854

-------
          UNDERSTANDING THE GERMAN AND JAPANESE COAL-FIRED SCR EXPERIENCE
ABSTRACT
This paper examines the Japanese low sulfur and the German  medium sulfur, coal-fired
selective  catalytic  reduction  (SCR)  costs  and  operating  experiences.    Their
implications for high sulfur,  U.S.  coal-fired applications is also presented.  It has
been observed that the costs  are more strongly related to  the  application size than
they are to the location of the SCR catalyst section.  However, operating costs and
issues are more strongly related to the location of the  SCR system.  The German Tail
Gas configuration  technology,  in which  the SCR system is located downstream of the
flue gas desulfurization  system, should be more easily transferred to high sulfur,
U.S.,  coal-fired   applications,  and  if newer  low  temperature  catalysts  or  less
expensive flue gas reheat  designs  are developed, it could  become the configuration
of choice.  Otherwise,  site specific  conditions, such as retrofit difficulty,  will
probably  dominate  the  selection  process  for  applying High  Dust  or Tail  Gas SCR
designs.  A new issue  not  addressed in the German and Japanese SCR experience  will
be how the spent catalysts  are controlled, since they may be classified as a hazardous
waste in the U.S..

INTRODUCTION & BACKGROUND

In 1970, the Japanese initiated the use of SCR technology for NOx control on large,
electric utility boilers, including coal, oil,  and gas service.  The Japanese locate
their  SCR  reactor   before  the   particulate   collection  equipment  (High  Dust
configuration) , or when they have  a high temperature particulate control system they
locate it after  that equipment (Low Dust  configuration) .   There  were some initial
problems, principally with the formation of deposits on the  catalysts and down stream
equipment.   Studies  showed  that  sulfur  bisulfates  were   forming,  and  they  were
depositing  on  the  equipment   or  mixing  with  ash  particles  and  the mixture  was
depositing  on  the  equipment.   Considerable  additional  product  development  was
undertaken,  and  they  eventually  developed  a  long  life,  reliable SCR  system for
reducing NOx emissions.  Their  approach included:  reformulating the catalyst to reduce
its potential  to  form  sulfur trioxide; during  operations reducing  the  amount  of
ammonia used by about  10-15%  from  the design specification and further controlling
the ammonia injection,  if required, to  assure that  the ammonia leakage past the SCR
reactor (e.g.,  ammonia  slip) is less than 10 ppm and preferably less than 5 ppm;  only
treating about one-third of the boiler's uncontrolled NOx with the SCR system  by using
combustion modifications to control  the other two-thirds of the uncontrolled NOx; only
operating the  SCR during steady plant operations  (it can,  however, be operated during
slow transients but it  is not  operated  during start  up or  shut down); and operating
with low sulfur, low ash content coals  (less than 1% sulfur and 10% ash).  Because
of continuing problems with instrumentation, control of  the process has been based
upon using  calculated  values  of  NOx  and ammonia,  and  the  measured  values are
principally used as a  trim signal  for  the  control setting.  Through 1990 they have
installed 40 SCR systems on 10,852  MWe  of  coal-fired utility  service.

In the early 1980s,  the German  utilities  began an extensive pilot plant evaluation
of Japanese  SCR designs for use at German electric utility power plants.  Ultimately,


                                        4B-3

-------
they conducted about 70  different  pilot  plant experimental  studies.  Since 1985, 129
SCR systems have been installed on 30,625 Me of coal-fired utility service.  Their
coal quality is quite different from  the Japanese:  they use low to medium sulfur and
ash coals  (0.7 to 1.2%  sulfur  with low  ash  coals  but  on  occasion up to 1.5% sulfur
and 25-30% ash,  although such ballast coal is usually blended with the cleaner coal);
and their plant operations include intermediate duty (e.g.,  daily  cycling and usually
shut down  during  the weekends).  Also,  the Germans retrofit their SCR  systems to
existing plants, whereas the Japanese have both  retrofit and new plant applications.
The Germans  have  both  wet  bottom firing  (e.g.,  slagging  boilers) and  dry bottom
firing, whereas the Japanese use dry  bottom  firing.  The principal German SCR design
is either the High  Dust system or a  Tail Gas system where  the SCR  is  located after
the flue gas desulfurization (FGD) equipment.  Initially, the Tail Gas configuration
was specified for wet bottom firing when the fly ash was recirculated to the furnace
section to be slagged (recycling produced high  arsenic levels in the flue gas which
poisoned the catalyst).  It  is important  to note that the  Germans  also do not operate
the SCR during  start up or  shut down periods  (ammonia injection  is typically keyed
to having the catalyst temperature at 554 °F or  greater).

TECHNICAL IMPACTS ON SYSTEM COSTS

Typically,  the  Japanese  report   SCR   costs   at   $35-80/kW"",   depending  on  the
application.  The initial German  experience"'2'11 indicated  that  costs  of  $60-189/kW
can be  expected,  depending  on  the boiler size  and  firing  conditions,  and  the coal
used.   U.S.  cost  estimates  range  from  $80-100/kW  for new  plant  installations"''1".
A key reason for the greater reported costs in the  German  literature  is that they are
for retrofit  applications  at  older,  more  congested  plant sites,  compared  to  the
Japanese and the new plant basis for  the U.S. estimates.  German studies prepared in
1985 and 1986  compared the expected costs for the High Dust and Tail Gas applications.
Figures 1  and  2 are typical of the German results  reported"'31-  Often  the capital
costs reported exclude the cost of the  initial catalyst charge.

These figures and the results  from the initial  SCR installations (from 1985 through
1987) appear to be used  by many authors  to identify that the High Dust configuration
is less expensive than the Tail Gas configuration.   However, if the entire data base
of  applications  through  1990 are  considered,   it  appears  that  the  Tail  Gas
configuration  is  no more  costly and   may  even   be  less  expensive.    High  Dust
applications require long plant shutdown periods  (during which the  lost  power must
be purchased from other  sources) to allow for the connection of  the  SCR system to the
plant.  On the other hand, the  Tail Gas system requires a flue  gas  reheat system which
can consume as much as 3-4% of the fuel costs.   Of course,  costs are controlled, to
a very large degree,  by the local site conditions, and thus either configuration could
be  the  least  expensive at  any  specific site.   The High  Dust system  (in  retrofit
applications) sometimes require penetrating the boiler primary pressure boundary to
remove or bypass part of the economizer  so that  the plant can continue to operate at
part power while maintaining the required SCR reactor  inlet tempertures.  This is an
expensive modification  that is generally not identified in reported  costs such as
provided in Figures 1  and  2.   On  the  other hand,  the Tail Gas  systems  requires a
second heat exchanger or auxiliary firing to bring  the  flue  gas  up to the SCR reactor
operating  temperature   (and  to recover the excess heat  before the  flue  gas  is
discharged through  the stack).   The heat exchanger  equipment can easily increase the
Tail Gas system's capital costs by 20%,  which is significant.

In considering the capital costs, it should  be recalled that Figures 1 and  2  are based
upon analysis assumptions and are not plots  of data. The early capital cost estimates
were important for establishing budgets,  and thus cost  assumptions that increased the
estimated cost were often used  to conservatively predict the expected costs.  In 1988
Jung"1 examined the reported costs for  19 German plants and reported that the total
                                       4B-4

-------
 capital  expenditure had an arithmetic mean of  about DM 66.3/kW,.   More  importantly,
 his  analysis showed that  there  was  no statistical  difference  between  wet and dry
 bottom firing (and hence High Dust and Tail Gas SCR configuration) ,  but that the data
 is represented  by:
                               C =  0.358  x MW,0-"1

       where C is the total capital cost in millions of DM,  expressed in January 1988
       DM;  and MW, is the power plant net thermal output in megawatts

 The uncertainty factor needed to account for having a 90  percent  confidence that all
 of the actual cost data is represented by this expression can be expressed  by applying
 a multiplier on the  coefficient  0.358:

                         C  = 0.358  x  (1.65 or  0.65)  x MW/-"1

       where  the multiplier 1.65  is  used to  define  the  upper bound of the cost data,
       and  the multiplier 0.65  is used  to define the  lower  bound  of the cost data.

 Since  the uncertainty can be  accounted  for  in the coefficient,  the 0.775 exponent can
 be thought of as  the  factor that  accounts for the effects of size or  scale.  The range
 in  the  coefficient  0.358   (+65%  to -35%)  indicates that  there  is  considerable
 variability in the cost data base.  This suggests  that more detailed analyses of the
 reasons  for  the cost differences would be  useful.

 However,  the  equation  is important  since  it explains the large difference  in  the
 reported cost/kW  for High Dust  and Tail  Gas  SCR applications  as  being primarily
 related  to the size of  the system being retrofit.   It is also  important to note that
 the inflation in Germany between  1985  and  1990 has been small enough that this cost
 expression can be used to account for  the  DM in any  year of convenience.   Others''1'1
 have  examined the   reported capital  costs and  they  report similar  exponential
 relationships between total  costs and  the  size of  the  SCR  application.

 There  is some distortion in  the  capital cost data that needs  to be understood,  and
 it also needs to be recognized that  the above expression does not account for those
 distortions.  First, the catalyst costs for the early  SCR  applications were 40,000-
 60,000 DM/m3.  By 1989 competition and the shrinking market  for catalysts reduced the
 costs  to 20-25,000 DM/m], and  in one case  the  catalyst cost was  reported at  17,000
 DM/ms.  Thus, as time went on  the  total capital costs were being reduced significantly
 because of the reduced  charge for the catalyst.  To obtain  a more realistic estimate
 of the present cost for  an SCR system,  the data that was used to generate the above
 equation  should  be  normalized  to  assume  the  same  unit  costs  for  the  catalyst.
 However,  that is not easy to accomplish because  the  German utilities do  not  always
 identify the catalyst component of the total costs, nor do  they use a standard chart
 of accounts.  Thus,  the reported capital costs can vary from  plant to plant because
 significantly different  items are included in the total  cost number.  Also, regulatory
 concern  about the safety of  anhydrous  ammonia storage and preparation systems has
 increased significantly  during  the application period.  Now the  regulators may require
 remote location of the storage vessels,  double wall vessels  and piping, and increased
 instrumentation and monitoring.  These  differences can increase the  total plant costs
by an amount approximately equal to  the changes accounted  for in the cost decreases
that have occurred for the catalyst  charge.

By 1987 the installed number  of High Dust systems  was about twice the number of Tail
Gas systems,  suggesting that  the market agreed with the interpretation  that the High
Dust system was the  least expensive.  However,  by  the end of 1990 the number of Tail
Gas systems was slightly more than the  number of High Dust systems""-  This suggests
                                       4B-5

-------
that price (and technical) considerations have shown  that  the  Tail  Gas is at least
competitive if not  less  expensive.   However,  just  as the cost data  is  not always
clear,  there is other evidence that  High  Dust  systems are  preferred.   For example,
the company Steag has reported that  in 1989 the High Dust systems accounted for 80%
of the dry bottom boiler  applications and 16% of the wet bottom boiler applications
(for a total of 18,213 MW  of capacity) , while the Tail Gas accounted for  the remaining
units (for a  total 9,997 MW of capacity).  Thus,  in 1989 and 1990,  more  smaller units
applied the Tail  Gas system.   Since  the  cost/kW is greater for  smaller  units, on the
surface the raw cost data would imply that  the  Tail Gas is more expensive.   Jung's
analysis, shows that this is not the case.

The earlier cost  estimates assumed  the  same  catalyst  design  for both  the High Dust
and the Tail Gas configurations.  Actual  experience*141 for honeycomb catalysts  is that
the High Dust systems have used a catalyst  pitch  of 6-7.5  mm with a space velocity
of  2,000-3,000/hr.   Tail  Gas  designs  have  used  a  3.7-4.2  mm  pitch  with space
velocities of  4,000-6,500/hr., and they  can also  use a  more  reactive  catalyst
formulation since the flue gas is cleaner than that which is  treated  by  the High Dust
system.  It has been observed in Germany and Japan  that sticky,  very  small sized dust
particles can cause  more serious catalyst deactivation problems than does arsenic (see
below for a further discussion about arsenic) .  Also, the greater dust accumulation
on the equipment  in High  Dust systems has contributed to SCR fires in  Japanese oil-
fired plants.  The  High  Dust  systems often employ  a  dummy leading  edge  to  control
catalyst erosion.   These  physical conditions  mean  that  a smaller catalyst  and SCR
reactor  can  be specified for  the  Tail Gas  system  (it  can  be 50-60%  smaller1111) ,
reducing its system costs.  Some of this  impact is mitigated by the  fact  that the flue
gas saturation at the Tail Gas location means  that  the SCR  reactor  must process  up
to 20% more flue  gas volume than would a High Dust system on  that plant, causing its
costs  to  increase  (assuming  that  both  designs  operate  at the same  catalyst
temperature).

Examination of the assumptions that were used to prepare Figure  2 also indicates that
some of  them  have biased the results to  indicate that High Dust systems are less
expensive to operate.  For example,   the figure  is  based upon assuming the catalyst
life is  3 years  for the High  Dust system and  5 years for the  Tail  Gas  system.
Experience indicates that  3-4 years  (with occasional 5 years life)  is consistent with
the High Dust catalyst  design.  However, some Tail Gas system operators report that
they are not experiencing  catalyst degradation,  and  that  they expect the catalyst to
last for up to 80,000 operating hours (more than ten years).   The Japanese on clean
flue gas,  but not  in  coal  service, have  had catalysts last  more  than  10 years.
Catalyst life is  an  important item,  others1" have reported that  catalyst lifetime can
dominate  the  estimate  of the levelized  cost  for  SCR  systems.  Figure  2 was also
prepared with the Tail  Gas system being  charged for the cost of the flue gas reheat
system when a good part of that  cost would have been incurred whether or not the SCR
system were present  because  some reheat is needed  to add buoyancy  to  the flue gas
after FGD treatment.  In  reality, only  the  incremental costs beyond those required
for the FGD reheat system  should be  charged  to the Tail Gas system.  The figure also
assumes that all  the reheat energy costs  must be charged to the Tail Gas  system costs.
The High Dust  system requires that  an economizer  by-pass be  installed to allow the
SCR  system  to operate  when   the plant  is  at part  power.    In  the  assumed 4,000
hours/year duty cycle that was used  to generate Figure 2, the plant will be  at part
power for a significant fraction of the operating time.  Under  those conditions bypass
of the economizer would lower the overall  power plant efficiency  by 2-4%, but that
cost impact was not  included in the  analysis used to generate  the figure.  The higher
assumed  initial   capital  costs  for  the Tail  Gas  configuration  require a larger
levelized capital recovery factor for that  design.   All  of these assumptions  tend to
make the operating cost estimate for  the High Dust  system less expensive and the Tail
Gas system more expensive.
                                        4B-6

-------
 Another  factor that can be very important to account for  is the fact that the  retrofit
 of  a  High Dust system  may be very difficult  and more  expensive  due  to  the congested
 available space, the need to reinforce the boiler building substructure and supports,
 the requirement to penetrate the boiler wall and to remove heating surfaces, and the
 need  to relocate soot blowers.  Those costs and the  plant down time (and the resulting
 costs for replacement  power while  the  plant is shut down)  should be charged to the
 High  Dust  system initial costs.  On the other hand, the  Tail  Gas system can be sited
 where it  is convenient and the final hook up to the main plant can be performed more
 quickly  (typically in a few  weeks).  However, the more remote and accessible the Tail
 Gas site,  the  potentially longer the flue  duct runs  and  the higher would be their
 corresponding costs.   The greater  dust  loadings,  the potential  for fires involving
 the SCR system (several fires were  reported in Japan in  1989), the risk  during plant
 upset conditions for ammonium  bisulfate  deposits to form on downstream  equipment, and
 the need  for washing the downstream equipment to remove  deposits and then to control
 and treat the  wash water so  it can be released  to  the environment  all  suggest  a
 greater down-side risk is associated with the High Dust system during the lifetime
 of  the plant.   This  risk should be  reflected as  a  cost,  but  it  is  normally  not
 considered in the analyses  used  to  prepare  figures such as Figure 2.

 The levelized cost assumptions do not include  costs for control and disposal of the
 spent catalysts.  The  shorter  the  catalyst lifetime the greater  this problem could
 be  over  the life  of  the plant,  and the  greater  the  expense  associated  with  the
 problem.   In Germany  and Japan it is  assumed that  the catalyst supplier  will  be
 responsible for the spent catalyst, and in Germany  the typical contract  requires the
 supplier  to receive the spent catalyst.  The experience in Japan  (there has not yet
 been  enough experience in Germany with  spent catalysts)  is that  it is not economical
 to  recover the catalyst  materials,  and  they are simply  disposed  of by the catalyst
 supplier.  Thus,  the assumption  that the disposal  cost  are negligible is reasonable
 for German and Japanese applications.   The  assumption  needs  to be  checked  for  its
 validity  at local U.S.  power plant  sites.

 U.S.  cost  estimates for  new plants  can  also be used to  help calibrate the reported
 costs  from German and  Japanese  plants.   In 1984m  the costs  for a new 500 MW plant
 were  estimated  at  $70-80/kW.    In  1989(1)  similar  applications  were  estimated  at
 $101/kW.   The 1989 costs actually represent  about  a 40% reduction in the SCR system
 costs, compared  to the  1984  estimates,  as is indicated when  the 1984  costs  are
 escalated to the same basis as the reported 1989 costs'61.  However,  the  1989 estimates
 also include  the 50% reduction in catalyst costs as  reported in the  German literature.
 With retrofit cost factors added, the U.S. estimates compare  favorably  to the German
 reported costs;  they are, however,   significantly greater than the reported Japanese
 costs  for  new plants.

 OPERATIONS AND MAINTENANCE

 Another important  lesson learned  deals with the receipt, storage,  use, and measurement
 of ammonia.  It has been established110'11'"1 that if  the  SCR design is to provide  a high
 level  of  NOx control,  extensive flue gas  flow modeling and  flow straightening are
 needed to guide the design of the ammonia injection  system and to assure that  the flue
 gas NOx-ammonia mixture is uniformly distributed across  the catalyst cross sectional
 area.   The ammonia  injection  designs  in Germany  typically  employ 30-40 injection
 points per square meter of flow  area.   Each  injection nozzle may  have  an individual
 flow control  (or each flow tube may  have  the  controls if several nozzles are installed
on a single flow tube) , so that the  system can be optimized during the plant shakedown
period (optimization is the  attempt  to develop  a uniform NH,/NOx  ratio  throughout the
entire inlet cross section of the SCR reactor).  However, the use of many injection
points can lead to  a false  sense of security.  It  has been observed  (see Figure  3)
 in  at  least  one design  that  used   multiple injection nozzles  that  dust  and other
                                        4B-7

-------
deposits collected inside  the ammonia  flow  tube, indicating that conditions occurred
which caused the  flue gas to enter some of the nozzles rather than for ammonia  to exit
those nozzles.   Since the  resulting  flow  maldistribution built up  slowly  during
operations, it was not detected  by  the  available  monitoring  instrumentation.  But,
the hoped for NOx control  that was supposed to be achieved by having a large number
of "tuned" ammonia injection nozzles was not achieved.

Much of the plant shake down and acceptance period is spent qualifying the valving,
motors, and  switches  in  the ammonia distribution system  (and  in  assuring that the
desired flow injection pattern is achieved).  Typically,  full  flue gas velocity, NOx,
and temperature profiles  are taken  and a less extensive ammonia distribution profile
is taken.  A 10% variation in the NHs/NOx  ratio would be an excellent optimization,
a 25-30% variation would not be  unusual.  Often two measurement instruments are used
so that a average value can be determined.

The Germans  initially  used,  in  their  High  Dust  applications,  the Japanese criteria
of designing for an ammonia slip of  less than 5 ppm, but during operations they have
found that they must  limit  the slip to less than 1  to  3 ppm(11)  in  order to provide fly
ash  that  can be used  in other  commercial  processes.  This  also  requires that the
ammonia injection be  such that the ratio of  NH,/NOx be less than 0.85  or the slip will
exceed the limit after a  short operating period'81.  The amount of acceptable ammonia
slip introduces  a  significant  uncertainty  in the life  expectancy of  the catalyst.
That is, later in its  lifetime the catalyst may be  able to  provide for the design NOx
reduction with a 5 ppm ammonia slip, but not the needed 1 to 3 ppm slip.  This item
remains to be evaluated,  based upon  actual,  long term, operating experience.  In the
Tail Gas configuration, because  of the very clean nature of the  gas,  the ammonia slip
can be set by air emissions criteria, and slips as  large as 20-30 ppm should not lead
to operational problems (ammonia odor and plumes are troublesome at 50 ppm or greater
ammonia concentrations).

As the German  regulators dealt  with  additional  SCR installations,  they have  become
more concerned about  the  safety implications of  anhydrous  ammonia,  and increasing
design  restrictions  reflect  those concerns.   In  Germany,  anhydrous  ammonia  is
primarily delivered by rail,  truck  transport  is prohibited except  for volumes less
than  500  liters.  The Germans  generally store on  site  a 15  to 30 day  supply  of
anhydrous ammonia.  Often the ammonia is diluted to a mixture of 8% or less ammonia
before it is introduced into the flue gas.   Normally, clean,  cold,  fresh air is the
diluent in order to help  keep the ammonia nozzles  clean.   Two storage tanks are used
to allow for uninterrupted operation.   They also increase  the inherent safety margin
if a tank accident should occur.  Many of the systems use double wall  tanks  and double
wall piping  from the point of receipt to the exit from the ammonia vaporizer.  Warm
water  heating  (compared  to  electrical heating  used  at  some  U.S.  gas  turbine
installations) is used to vaporize the ammonia.   However, the safety requirements are
locally developed, and some  installations  use  single wall systems,  some facilities
bury their  tanks,  others were allowed  to  install  above  ground tanks.  Increasingly
restrictive  safety criteria have caused the ammonia system costs to increase almost
as much as the SCR  catalyst costs have decreased"1.  Figure 4 is  a presentation  of the
purity  specifications  used for  the purchase of  anhydrous ammonia.   The  water and
oxygen are controlled such that the delivered ammonia is  in region A of the figure.
Equipment corrosion is possible in region B, and corrosion will occur if the oxygen
and water content  are  such  that  the  ammonia is  in region  C.   However,  corrosion by
water alone is unknown, and the  storage  of  hydrous ammonia with 25% or greater water
content is  common.

Actual plant operations in  Germany  have  shown  that some of the catalysts can store
ammonia.  The storage has no effect during steady state operation, but this  ammonia
inventory is released  from  or added to  the catalyst  during transients such as load
                                         4B-8

-------
 changes  or shut  downs.   This phenomenon must  be recognized in  the  design of  the
 ammonia  injection control system, as the desorption  process has been found, in  at
 least  one  plant,  to  take up to eight hours.  For example, if the ammonia  injection
 control  does not  account  for  desorption during  shut down,  excess  ammonia will be  in
 the flue gas during low temperature operations,  causing too high an  ammonia slip  and
 possibly forming and depositing bisulfate salts.  Since ammonia desorption/absorption
 can produce a 30 minute lag in the NOx control1", the selection  of the  averaging time
 for meeting the NOx emission  control requirement  is very  important  for establishing
 if the plant operation can remain within the  permitting requirements. Since  bisulfate
 formation  is not controlled  by  the  ammonia  slip  as much  as  it is by the  SOS
 concentration, the lack of  a  continuous ammonia monitor and  control by ammonia slip
 levels is  not serious  to  the  overall satisfactory operation  of  the  SCR system.

 Other  design areas  that  were  not  addressed during the initial German installations
 but which have been indicated by operating experience are:  1) sulfur trioxide and acid
 attack of the duct liners  downstream of the SCR has occurred in some cases, indicating
 that  the  oxidation  potential of  the  catalyst  (for  converting  sulfur dioxide  to
 trioxide)  will be an important consideration when high sulfur coal is  fired.  2) the
 gas to gas heat  exchanger for  reheating  the  flue gas for the Tail Gas system has been
 required to be redesigned to make it less complicated  and  to  produce less leakage of
 the untreated gas into the treated gas.  Leakages as  high as 7% are reported, but 1.5-
 3% appears to be  more common.  It is possible to design for zero leakage,  and some
 plants  have such heat  exchangers.   This  suggests  that the  use  of rotary  heat
 exchangers should be  replaced with the  use  of non fluid mixing heat exchangers for
 reheating  the  flue gas  to  the Tail Gas  system operating conditions.   3) control
 problems under load swing conditions have indicated  that the ammonia instrumentation
 remains  an issue as well  as  the impact of ammonia  absorption/desorption  from  the
 catalysts.  This could be even a more pronounced problem in high  sulfur coal service.

 The initial  reason  for developing  the  Tail Gas  configuration  was  to overcome  the
 catalyst poisoning that was experienced in a few  pilot plants that serviced wet bottom
 boilers that also recirculated the fly ash back to the  boiler to  slag  the ash.  That
 operation  was found to increase  the arsenic concentration in the flue gas  by 10 to
 100 times  from  the  level  found when  no  fly  ash  recirculation was used.   That
 information was initially used to  conclude that the Tail Gas design was required when
 wet bottom firing with  full  flue gas  recirculation  was  used at  the power plant.
 Gutberlet'111 has since examined the arsenic concentration in 14 wet bottom plants that
 had varying amounts of fly ash recirculation.   Figure 5(1" presents  data from the 14
 separate boilers.  The upper plot  presents the relative amount of arsenic in the flue
 gas before the air preheater. The lower plot shows the relative amount  of arsenic
 in the coal being fired.  The data was plotted so that the amount of  arsenic in the
 flue gas would  increase as the data  is viewed from left to  right.   The  figure clearly
 shows that arsenic content in the coal  and  operation with  100% fly ash recirculation
 back to the boiler can cause both high  or  low  arsenic concentrations in the flue gas
 (Gutberlet  did  not  identify  the  actual  arsenic concentrations).    Thus,   fly  ash
 recirculation by itself  is  not  a sufficient  parameter  to determine if arsenic
 poisoning  would be a  problem  for  a High Dust  SCR application.   Gutberlet  concluded
 that the composition  of  the fly ash itself was  a significant  factor, and  that  the
 greater the amount of  calcium oxide in  the fly ash  the less arsenic would be found
 in gaseous form in the flue gas.   He  recommended that the calcium oxide in the  fly
 ash be at  least  3% and  preferably greater than  5%;  that  fly  ash recirculation be
 restricted in general if  possible  and especially during boiler soot blowing; and that
 coals  with low arsenic content be  used (most German coals have arsenic  levels of 5-25
mg/kg,  and the  term  "low arsenic  content" was  not  defined or  compared  to those
 levels).    Separate Japanese'111 studies  (the coal  arsenic  levels were  not  reported)
have shown that   calcium  content  in the  fly  ash is  a primary  factor  for causing
catalyst deactivation  in  High Dust SCR configurations.   This  further supports  the
                                         4B-9

-------
German data that fine dust can be a greater catalyst "poison" than is arsenic.  The
Japanese data  is  for dry bottom  firing.   The  Japanese  experience is  that  if  the
calcium oxide content is  less  than 1%  the catalyst will not have to be replaced even
after 38,000 hours of operation.   Five to eight  percent calcium oxide in the fly ash
can cause a comparable catalyst deactivation in 20,000 hours, and catalyst replacement
will be required before  25,000 hours.   They  also  found  that  high sulfur coals that
produce a higher ratio of  gypsum in the calcium compounds in the  fly  ash are less
likely to cause rapid catalyst deterioration.   Thus,  it  appears  that  if there is a
species such as arsenic  in the flue gas, it  can react with  the  calcium and is thus
removed as  an active  poison.  However, if there is  no species  available  to react with
the calcium,  the  calcium itself  can blind  the  catalyst.  The desirability of some
calcium content in the flue gas needs  to be  carefully studied  for U.S. high sulfur
coal applications.   Such a study  will  also have to  evaluate  the  type of FGD design
used if Tail Gas SCR systems are being considered.

IMPLICATIONS FOR U.S. SCR OPERATIONS

Extensive overseas use of  SCR  has established design and cost criteria as well as
credibility for the use of this technology in  low sulfur fuel applications, including
coal, oil,  or gas.   An important  consideration  in this  commercial success for High
Dust applications has been the  recognition  that  the ammonia:sulfur trioxide reaction
to  form ammonium  bisulfite and bisulfate can be controlled  by  limiting the ammonia
slip.


Low Sulfur Coal

The  large  U.S.  population  of low sulfur,  coal-fired,  electric  utility  boilers,
primarily those firing western  subbituminous  coals, should be  able  to readily utilize
the commercial  design and  operating retrofit experience from Japan  and Germany to
achieve NOx  emission levels,  when  and  if  required, as  low  as  100 ppm  or  0.17  Ib
NOj/million Btu.  The German  operating  experience  is more important for such U.S.
applications because many of  the  U.S.  installations  are  on peaking utility boilers,
whereas Japan coal-fired service  has been essentially all in more simple, base load
boiler cases.  An appropriate  SCR  system design  strategy based on German practice on
low sulfur, coal-fired,  dry bottom boilers  would call for the retrofit of combustion
modifications including  low NOx burners  to  reduce the gross emission to the range of
325 to 400  ppm(m-  This greatly mitigates the  overall cost and ammonia slip problems,
since the resulting SCR  removal efficiency  requirement is no more than 70 to 75% to
achieve a 100 ppm stack  emission.   Such  an  approach,  if it achieved only half of the
four million  annual  ton  of NOx emission inventory reduction in  the U.S.,  which is
implicit in Title IV of  the new Clean Air  Act,  would result  in nearly  100,00 MW of
retrofit SCR system capacity.   That would be  more  than the entire present worldwide
population of existing SCR facilities.

Key design premises for  this  coal-fired SCR service  are  tied to the use of a design
ammonia slip of 3-5 ppm.  The  suppliers and  users in Japan'201 have identified them as:

      •     Vertical downward gas flow reactors to prevent ash accumulation.

      •     Linear gas velocities  of  16-20 ft/sec  (5-6  m/s) at  maximum continuous
            rating to prevent ash accumulation  and erosion.

      •     Use of a  grid-shaped catalyst with a channel  spacing (pitch) of 0.275-0.3
            inches (7-7.5 mm)  to prevent ash accumulation and erosion.

      •     Catalyst  layers  formed  without   seams   along  the  gas flow  direction
                                       4B-10

-------
             (including  optional  use  of a sacrificial initial stage) to prevent  ash
             accumulation  and  erosion.

      •      Ash  deposition  removal by  intermittent  vacuuming or soot blowing.

Technical literature issued by Japanese SCR system suppliers emphasizes  the  following
key elements  for a  successful SCR installation on peaking  utility boilers:

      •      Reliable  ammonia  feed control.

      •      Adequate  ammonia  feed distribution  across  the  cross  section  gas flow
             area.

      •      Flue gas  duct designs that ensure good  mixing of flue gas and ammonia
             feed.

      •      Provision for suitable control of  ammonia feed at part load.

In Germany,  important new SCR experience has  been  gained  on slagging  (wet bottom)
boilers, which produce different flue gas and fly ash characteristics that can impact
the SCR catalyst'21'.  Full-scale experience since 1986 in the use of the  High Dust SCR
designs for  some wet  bottom boilers  has confirmed excessive  impacts on the catalyst
life  and  has lead  to their  reworking  approximately 3,000 MW of  such  retrofits to
convert them to  Tail  Gas  designs.  It  is clear that  such boilers  are best served by
the Tail Gas  design.


High  Sulfur  Coal

The potential  impacts of  the much higher sulfur  dioxide and especially  the higher
sulfur  trioxide  concentrations in the flue  gas from U.S.  high sulfur,  coal-fired
plants will  be an  item  of special concern.   It may  cause  a new optimum of  catalyst
reactivity,  linear  velocity,  ammonia  slip,  and operating  temperature  window  to be
established.  Since much of  the impacts of ammonia absorption/desorption, sulfate and
bisulfate  formation,  poisoning  of the catalyst,  and  blockage or  blinding of  the
catalyst are  surface  effects, it is  not possible to scale the existing  German and
Japanese High Dust  SCR results with  confidence without the benefit  of prior testing
at the expected  operating conditions.   Although  the Tail  Gas results  can be scaled
with  somewhat  more  confidence,  questions about calcium poisoning or  trace element
carry over from  the FGD system  indicates  that  these designs also need  to be tested
prior to final design selection.

Patterning its  work after the major pilot  plant  test  program  in Germany,  EPRI is
carrying forward a  $15 million collaborative bench and pilot-scale  research program
to include  definition of  costs and technical feasibility  for  the use of  SCR in
domestic medium and high sulfur coal  service'"1-  Testing will be conducted at as many
as 14  separate  facilities  over  a four year period  to assess  SCR process design,
catalyst life, instrumentation and controls, and plant integration.

Application of the Tail  Gas  system design is an important and possibly vital approach
for using  SCR technology in high sulfur  coal service.  Such process applications would
avoid the  potentially  excessive rate of  air preheater  fouling and catalyst degradation
from high sulfur service,   High  Dust  SCR design,  even with  an ammonia slip value as
low as 1 ppm.

Cyclone-fired wet bottom boilers, for high  sulfur  service and with their typical NOx
emissions of 0.8  to  1.8 Ib N02  million Btu  (500 to 1,100  ppm, represent a major market
                                        4B-11

-------
for the Tail Gas  SCR  design.  There are presently 105 operating cyclone units.  They
represent about 14% of  the  pre-NSPS coal-fired generating capacity  (over 26,000 MW).
However, these  units contribute about 21%  of the NOx emitted by pre-NSPS units because
their combustion design is  conducive to NOx formation, and their firing design is not
conducive  for  low NOx burner  technology.    Furthermore,  other  conventional  NOx
reduction techniques  such as two-stage combustion cannot be applied to the full extent
due to associated operational concerns with  corrosion.   Although many of  the units
are 20  to  30 years of  age, many of the utility owners plan to operate  them for an
additional 10 to 20 years.  Since the majority of these units are  located in the acid
rain  emission   control area  (the  Midwest) ,  cyclone boilers   may  be  especially
appropriate for NOx control.  With the well  established  use of Tail Gas SCR systems
on wet bottom  boilers  in Germany, this class  of  boilers  appears  to be a key target
market sector  for Tail Gas SCR designs.

REFERENCES

 1.   NOx Task Force,  Economic Commission for Europe, Technologies for  Controlling
      NOx Emissions from Stationary Sources. July 1986.

 2.   "Development & Application of NOx-Flue  Gas Treatment  in  the Federal Republic
      of  Germany."  NATO/CCMS Meeting,  Control  of  Air  Pollution From  Combustion
      Systems, Duesseldorf, October 1988.

 3.   B. Schaerer, N.  Haug, and  J-H.  Oels.  "Cost of  Retrofitting Denitrification."
      Proceedings of  the Workshop on Emission Control  Cost,  Esslinger  am Neckar, FGR,
      1987.

 4.   "Selective  Catalytic Reduction for Coal  Fired  Power Plants."  EPRI CS-3603,
      October  1984.

 5.   C. P. Robie, P. A. Ireland, and J.  E.  Cichanowicz.  "Technical Feasibility and
      Economics  of  SCR  NOx Control  In  Utility Applications." Proceedings  1989
      Symposium on Stationary Combustion  Nitrogen Oxide  Control.  EPRI GS-6423, July
      1989.

 6.   P. A. Lowe. Selective Catalytic Technology. Burns  and Roe Service Corporation
      report submitted to  the U.S. Department of Energy,  December 1989.

 7.   J.  Jung.  "Capital  Expenditures in  S02-  and  NOx  Reduction  in  the  German
      Electricity Industry." VGB Kraftwerkstechnik,  Vol.  68, No.   2,  February 1988.

 8.   P. A. Lowe,  and M. Perlsweig. "Recent Experiences for  SCR Systems at Coal-Fired
      Utility  Boilers." Proceedings of the American Power Conference, March  1990.

 9.   J. E. Cichanowicz, and G.  R. Offen, "Applicability  of European SCR Experience
       to U.S.  Utility  Operation," 1987 Symposium on  Stationary Combustion Nitrogen
      Oxide Control, EPRI  CS-5361, August 1987.

 10.  P.  Necker.  "Operating  Experience  with  the  SCR  DeNOx  Plant  in  Unit  5  of
      Altbach/Deizisau Power Station." 1987  Joint Symposium on Stationary Combustion
      NOx  Control, EPRI CS-5361, August 1987.

 11.  M. Novak,  and  H. G.  Rych.  "Design  & Operation  of  the  SCR-Type NOx-Reduction
      Plants at the Duernohr Power  Station in  Austria." 1989 Symposium on  Stationary
      Combustion Nitrogen  Oxide  Control,  EPRI GS-6423, July 1989.
                                       4B-12

-------
12.  J. Ando.,  "NOx  Abatement  for Stationary Sources  in  Japan,"  EPA-600/7-83-027
     (PB83-207639),  May 1983.

13.  P. Necker.  "Experience Gained  by Neckarwerke  From Operation  of SCR  DeNOx
     Units." 1989 Symposium on Stationary Combustion NOx Control, EPRI GS-6423, July
     1989.

14.  K. Staebler,  et.  al..  "Secondary Measures  for NOx Reduction  Experience with
     Pilot- and Commercial-Scale  Plants."  VGB Kraftwerkstechnik,   68,  No.  7,  July
     1988.

15.  H. Gutberlet.  "Influence of  Furnace  Type on Poisoning of DENOX  Catalysts  by
     Arsenic." VGB Kraftwerkstechnik, Vol 68, No 3, March 1988.


16.  S. Nagayama,  et.  al.. "SCR Application  for NOx Control  in Coal  Fired  Utility
     Boilers." Proceedings of the 7th Pittsburgh Coal Conference,  September 1990.

17.  A. Kinoshita.  "Trends   in  Environmental Control  Costs  for  Coal-Fired  Power
     Plants in Japan."  Proceedings of  the 7th Pittsburgh Coal Conference,  September
     1990.

18.  M. Hildebrand.  "Status  of  the Art of  Flue-Gas  Purification  at E.S.C.  Power
     Stations S02/N0x-Reduction  (in German)." Elektrizitaetswirtschaft,  Jg. 89, Heft
     9, 1990.

19.  K. Leikert, H. Reidick,  and H. Schuster. "Low NOx-Emission  Combustion of Fossil
     Fuels in the Federal Republic of  Germany." Proceedings of the Fourth  Seminar
     on the Control  of Sulphur and Nitrogen Oxides  From Stationary Sources,  Graz,
     Austria, 1986.

20.  P. A. Lowe. "Utility Operating  Experience with  Selective  Catalytic  Reduction
     of Flue Gas NOx.", Proceedings  of the Second  International Conference  on Acid
     Rain, Washington,  B.C., 1985.

21.  G. R. Offen,  et. al..  "Stationary Combustion  NOx Control." Journal of  the Air
     Pollution Control Association,  July 1987.

22.  Electric  Power  Research Institute,  "Selective Catalytic Reduction for  NOx
     Control, A Proposed Collaborative Bench- and Pilot-Scale Research  Program,"
     April 1988.
                                      4B-13

-------
 a:
 ca
 Uj
 a:
 Uj
     100-

      90-

      80-

      70-

      60-

      50-

      40-

      30-

      20-

      10-
0 VET BOTTOM BOILER CORL-FIRED

S QRT BOTTOM BOILER CORL-FIREO

[ID OIL-  FIND GRS-FIRED BOILER

80'/ NO*  REMOVffL
              100    200    300    400    500
                        CflPPCITY [MV*1]
                                                 600
                                         700
   FIGURE  1.   GERMAN SCR  CAPITAL  COSTS VERSUS CAPACITY
    J.5-
s:
Q
o
o
*-J

-------
FIGURE  3.    DUST  DEPOSITS  IN  AN  AMMONIA  FLOW  TUBE
                          FIGURE  4
            -GERMAN OXYGEN AND WATER PURITY SPECIFICATION.
                	  FOR ANRYROUS AMMONIA
                               O, [ppm]
                               48-15

-------
                        ARSENIC IN FLUE GAS
  l.O
  0.1
 0.0 1
O.001
  1.0
  0 5
        • 100% ash recirculation
        ® partial ash recirculation
        ° no ash recirculation
                            e
                            o
                            o
                            o
ARSENIC  IN COAL (dry)
        o
        !   2   3   4  5   6   7  8   9   1011121314
                         TEST NUMBER
   FIGURE 5.  RELATIVE CONCENTRATION  OF ARSENIC  IN
               THE FLUE GAS  AND IN THE COAL
                        4B-16

-------
OPERATING EXPERIENCE WITH TAIL-END AND HIGH-DUST
 DENOX-TECHNICS AT THE POWER PLANT OF HEILBRONN
                  Dr. H. Maier
         Energie-Versorgung Schwaben AG
               KriegsbergstraBe  32
                7000  Stuttgart  10
                     Germany
                     P.  Dahl
         Energie-Versorgung Schwaben AG
               KriegsbergstraBe  32
                7000  Stuttgart  10
                     Germany

-------
                  OPERATING EXPERIENCE WITH TAIL-END AND HIGH-DUST
                   DENOX-TECHNICS AT THE POWER PLANT OF HEILBRONN
ABSTRACT
At the Heilbronn power station two different SCR-DENOX technics are installed.
The high dust SCR of the unit Heilbronn 7 has  been  started  up  in September  1986 and
operates till now appr. 25.000 hours.
It is a two line arrangement, each of it denitrificating  50 %  of the total  flue gas
amount.
The loss of catalysts activity and the ammonia slip have been measured  in dependence
of operating hours  and  the  results  compared with  expected or theoretical  calcula-
tions.

Decrease of catalysts activity, fly ash plugging, irregular distributions in the SCR-
reactor  (velocity,  temperature,  NH3/NOx-ratio) and  side  reactions (acid particles
formation and emission, ammonia sulfates) have been investigated  in detail.

In the units Heilbronn 3   6 the tail-end configuration is installed.
The flue gases of 4  slag tap boilers are desulfurized downstream  the electrostatic
precipitators in one FGD plant (wet  limestone  process yielding gypsum as byproduct)
with a capacity of  approximately 1.800.000 m3/h.  The tail-end SCR denitrificates the
flue gases, which  are free  of dust  and desulfurized,  with two  lines,  each of  it
treating 50 % of the total flue gas amount.
The first line started up in the middle of  1988 and operates till now appr.  14.000
hours, the second one in October  1990.  In contrast  to high  dust DENOX-plants,  which
had been proofen in Japan for several years, there was no experience with  tail-end
arrangements.
Therefore different start up times resulted from the demand, to research  and  optimize
such a configuration at the first line.
                                       4B-19

-------
                  OPERATING EXPERIENCE WITH TAIL-END AND HIGH-DUST
                   DENOX-TECHNICS AT THE POWER PLANT OF HEILBRONN
                          (Dr.-Ing. Herwig Maier, Paul Dahl)
1. INTRODUCTION
At Heilbronn the EVS  operates  following combined heat and- power units:
    •      Unit  3-6,  with four bituminous coal fired boilers and a slag tap
                      firing process.
                      Technical parameters as shown in fig. 1.  (1st slide)
    •      Unit  7,      with one bituminous coal fired boiler and dry firing
                      process.
                      Technical parameters can be seen in fig. 2 (2nd slide).

In the Federal Republic  of  Germany  the  N0x-emissions from bituminous  coal fired
boilers with more than 300  MW  thermal output  are  limited to less than 200 mg/m3.

This usually needs  post  combustion  flue  gas cleaning technics.  A well known and
widely used technology is the  selective  catalytic reduction or  SCR-process, where
ammonia reduces with  the aid of catalysts nitrogen oxides to  nitrogen and water at
temperatures between  300 °C and 400  °C.  (Fig. 3,  3rd slide).
The SCR-process can be installed between boiler outlet and air  preheater (the  so
called high dust arrangement)  or downstream of the electrostatic precipitator  and
the flue  gas desulphurization  plant  (the so called tail end configuration).
At Heilbronn power  station  both possibilities are realized.

2. THE HIGH DUST DENOX OF HEILBRONN  7

2.1 General Informations
During start of erection (1982) and  trial run (end of 1985) environmental protec-
tion laws have been actualised several  times, of  course each  time they changed to
lower limit values  for SOX-  and N0x-emissions. Until 1982 the  N0x-reduction by post
                                       4B-20

-------
combustion flue gas technics has not been a severely discussed, feasible solution.
But in January 1984 the N0x-emission  limit  for  new coal  fired plants was reduced
from 800 mg/m3  to  200  mg/m3. So, during the phase  of erection of Heilbronn 7 we  were
forced to realize a SCR-plant as soon as possible. The erection of the DENOX-plant
started in the last quarter of 1985. One year  later, in September 1986 the SCR was
put into operation.
Fig. 4 (4th slide) shows the high dust arrangement of the SCR-plant at Heilbronn 7.

2.2 Operating Experience
Since beginning of operation the loss of catalysts activity and resulting ammonia
slip is measured periodically and compared with expected respectively theoretical
calculations.
Fig. 5 (5th slide) shows, that after 12.000 operating hours the status of the SCR-
plant was much better than expected (and guaranteed). But,  as one can see in fig. 5
too, during the first months of 1990 (after appr. 18.000 operating hours) we
recognized increasing amounts of ammonia in the fly ash (up to 100 ppm) and
consequently we got problems with our waste water, where the ammonia concentration
is limited to 10 mg/1. Therefore we were forced to set measures immediately:
    •      ordering  an  access  layer  of catalysts
    •      manually  cleaning of  the  DENOX  plant
    •      measuring  of catalysts  activity
    •      measuring  the  ammonia  slip  before and after  cleaning.

The results,  that partially can be recognized again in fig. 5, were:

    •      The relative activity  Kt/K0 of  the catalysts  was 0.77,  after 18.500
          operating  hours  a really  sufficient result
    •      the cleaning of  the  plant  decreased the  ammonia slip  from  appr.
          2 vpm  to  0.5 vpm
    •      optically  we estimated  a  loss  of  active  surface of  appr. 22 %,
          caused by  fly  ash plugging  (fig.  6, 6th slide).

These results caused  (beside manual cleaning) further steps, to bring the plant
back to optimal operation conditions.
                                       4B-21

-------
Fig. 7 (7th slide) shows, how the need of access catalyst volume or,  at  the  other
hand, the loss of volume needed for good operation, is influenced by  irregular
distributions of
    •     NH3/NOx-ratio
    •     temperature
    •     velocity

It can be seen clearly, that optimizing the NH3/NOx-distribution is most  effective.
Therefore, measures to be done are:

    •     control  of  the  NH3/NOx-distribution and,  if necessary,  optimize the
          NH3-feeding
    •     we  recommend generally  and  ordered for  Heilbronn  7  sootblowers
          between  all  catalyst  layers,  including  the dummy, to keep  the  plant
          free from  surface  losses  caused  by fly-ash plugging.  The  sootblowers
          will be  installed  this  summer
    t     to  keep  further  troubles  far from us  during  that  time, we  replaced
          the first  layer  by a  new  one.

In fig. 8 (8th slide)  we tried to estimate  the consequences of replacing  one
catalyst  layer in dependence of fly-ash plugging. You can see clearly, how  impor-
tant the  sootblowers  are regarding the life time.

2.3  Side  Reactions
The  most  important and well known side reaction in  high-dust plants  is the cataly-
tic  oxidation of  S02  to S03  and, depending  on flue gas humidity and temperature,  the
formation of  sulfuric acid. This can cause troubles in two ways:

The  Formation of  Ammonia Sulfates like (NHJ, SO,  or NH^HSO,,.  These compounds  can
cause pluggings on the gas-gas-reheater,  therefore  needs washing of  it  and  creates
ammonia-loaded waste  water.
According to  the  very  low ammonia slip,  which we try to run at Heilbronn 7  because
of other  reasons  (saling of high-quality fly-ash),  this is no problem till  now.  The
ammonia concentration behind SCR is not sufficient  to form ammonia sulfates  at
amounts,  which could  hurt.
                                       4B-22

-------
Formation and Emission of Acid Particles.  Sulfuric acid condensates at the cold

raw-gas side of the gas-gas-reheater and is transported to the clean-gas side.
There an acid film forms at the walls of the flue gas ducts and the stack and is

adsorbed by fly ash or gypsum particles. That means, particles are created with an

access of free sulfuric acid. Especially during starts after a few days stop the

acid particles are solved from the walls by thermal expansion, transported to the

stack and, after reaching full load and corresponding gas velocity, emitted within
a few minutes.

This phenomenon causes severe troubles  if attacking the dope of cars nearby the

power plant.

To fight these emissions, we have the possibility


    •      to  reduce the  efficiency of  the ESP  and  increase  the  amount of
          neutralizing fly  ash  (this we  are  doing  at the  moment)

    •      to  set  constructive measures  in the  clean  gas duct  like  especially
          precipitators  (we are testing  this)

    •      in  future,  when catalysts  have to  be renewed, to  use  types with  low
          conversion  efficiency.


2.4 Conclusions

After appr. 25.000 hours of operating the SCR-Plant at Heilbronn 7 we made follo-

wing experiences:


    •      The decrease of catalysts  activity is far  from  expected  values  and
          much better than  assumed.

    •      Fly ash  plugging  can  cause  severe  problems,  so  we recommend to
          install  sootblowers between  each  layer.

    •      Irregular distributions  of  velocity,  temperature  and  NH3/NOx-ratio
          can not  be  avoided.  The  most  effective measure  at relatively  low
          costs  is  to optimize  the distribution of NH3/NOXI  i.  e.  optimizing
          NHj-injection under respect of not avoidable irregular distributi-
          ons.

    •      Formation and  emission of  sulfuric acid  and  acid  particles makes  a
          lot of  troubles especially  at  plants with  much  starts and shutdowns.
          Measures  to solve this problem are possible  and we  try to realize
          them at  the moment.
                                       4B-23

-------
   t     Formation of  ammonia  sulfates  and  consequences  out  of  this  is  no
         problem. This  follows  from the very good  activity-slope  and conse-
         quently  low NH3-sl ip.

3.  THE TAIL-END DENOX OF HEILBRONN 3 -  6
3.1 General Informations
The start up of these elder units at Heilbronn was between 1958 and  1966. As
mentioned before, also for these plants the N0x-emissions were limited to 200 mg/m3
in the middle of the 1980's.
That means again, that additional to primary measures post combustion technics  had
to be  installed.
But in contrast to the situation at unit 7, the well proofen  high dust  technic  is
often  not feasible for slag tap boilers.
Problems arise due to  interference with the compact design of elder  plants and  to
relatively high content of catalyst poisening gaseous trace elements  (like arsenic)
in the flue gas.
These  were the reasons, why we  decided  to erect a so called tail-end  SCR plant
behind electrostatic precipitators and  flue gas desulphurization. This  configura-
tion  is shown  in figure 9  (9th  slide).
Because of no experience with such an arrangement, we realized this  project within
three  steps. The first was a pilot plant, which is not in discussion  here. Just  let
me say, that the principal feasability  could be proofed.
The second step was the commercial scale tail end plant for one half  of the total
flue gas amount of the units 3-6, appr. 900.000 m3/h.  This  first line  started up
in the middle of 1988 and operates till now more than 14.000  hours.  An  intensive
research project was running with this  line from January until December 1990. The
first  two steps have been partially sponsored by the EEC. The third  step conse-
quently was DENOXing the second half of the flue gases of unit 3   6. This second
line started its operation in December  1990 with the trial run.

3.2 Operating Experience
Catalysts activity has been measured after 3.700 and 6.500 hours of  operation with
the result of no activity  loss. At the  moment we expect the results  after 13.000
hours. The ammonia slip measured at the DENOX-plant additionally after  those times
was below 1 ppm.
                                       4B-24

-------
Because of the very low dust- and S02-concentration with the tail-end configuration,
the design of the catalysts is different to that of the high dust arrangement, as
one can see in figure 10 (10th slide). Smaller pitches cause  in spite of the  low
dust concentrations some pluggings, so we are cleaning the plant from time to time
manually and we think about the installation of sootblowers.

In contrast, to the high dust plant the flue gas temperatures in front of the tail-
end reactor is appr. 50 °C (behind FGD). Therefore we have to operate a reheating
system (9th slide), which is a combination of regenerative gas/gas preheater  (GAVO)
and gas burner. The last one we need to compensate the hot side temperature
approach of the GAVO, which is below 30 °C.
Up to now the GAVO operates very well. The total leakage is  about 3.7 % at full
load, that means an increase of the NOX  clean  gas  concentration  of  appr.  25 mg/m3
caused by the GAVO. To prevent an increase of differential pressure according to
GAVO plugging, we clean it once per shift using sootblowers, but we don't wash the
GAVO.

3.3 Side Reactions
In principal we have to discuss the same side reactions as before,  but obviously it
is expected to cause less problems because

    •      the  S02-concentration behind FGD is very  low
    •      the  conversion  rate  depends  on  the temperature (it  decreases with
          decreasing  temperature  rapidly)  and we have  significantly  lower
          temperatures with  the  tail-end  configuration  (290  °C    320  °C)
    t      also  the  ammonia  slip  is  far from dangerous  concentrations.

And indeed till now there are no problems with side reactions at the tail-end
plant.

3.4 Conclusions
After appr.  14.000 hours of operating the first line of the  tail-end DENOX plant at
Heilbronn 3-6,  and after three months with the second line we made the experi-
ence,  that this configuration is a suitable one for slag tap fired boilers:
                                       4B-25

-------
The operating characteristics are shown in figure 11 (11th slide).
          Again  the  decrease  of  catalysts  activity is  far  from expected
          values.

          Fly  ash  plugging  can be  handled  by  manual  cleaning,  but  sootblowers
          should be  more  effective.

          The  danger of formation  and  emission  of  acid particles  is  less  than
          at the high dust  plant and we  don't expect troubles  from this side
          in the future.

          Catalyst poisening  caused  by gaseous  trace elements  can  be  avoided,
          if using a tail-end arrangement.
                                       4B-26

-------
Flue gas flow rate     Nm3/h                   1,8 x 106
Output                MWnet                         426
NOx-concentration

-  inlet of SCR         mg/m3                   <   1.000
-  outlet  of SCR        mg/m3                   <     200
NOx-removal              %                    >      80


SOx-concentration

-  inlet  FGD            mg/m3                 1.600 - 3.400
-  outlet FGD           mg/m3               <   160  -  340

SOx-removal           %                       >      90

Dust concentration

-  inlet  electrostatic     mg/m3                       6.000
  precipitator

-  outlet electrostatic   mg/m3                   <     200
  precipitator

-  outlet FGD           mg/m3                   <      50


Operation time

DENOX 1              h                          14.000
DENOX 2              h                           2.000



            Figure 1.   Technical Parameters of Heilbronn 3-6
                               4B-27

-------
Flue gas flow rate     Nm3/h                     2,3 x 106
Output                MWnet                          700
NOx-concentration

-  inlet of SCR         mg/m3                   <       800
-  outlet  of SCR        mg/m3                   <       200
NOx-removal              %                      75 -   80


SOx-concentration

- inlet  FGD            mg/m3                 1.500 -  3.200
- outlet FGD           mg/m3               <  200  -   400


Dust concentration

- inlet  electrostatic     mg/m3                 7.000 - 12.000
  precipitator

- outlet electrostatic   mg/m3                  <        50
  precipitator

Start up of DENOX-system                  September 1986


Operation time        h                            25.000



            Figure 2.  Technical Parameters  of Heilbronn 7
                               4B-28

-------
CD

ro
CD
                                                                    catalyst

                                                                 32CTC-40OX
 m tf~?~\
4     +
                                                               \    catalyst    r\ /<^\   **
                                                               .) 320.c_400.c  -3 Qy + 6
                                              Figure 3.   Principals of the  SCR-Process

-------
                                                      flue gas
                                                       duct
DO

CO
O
                                         DENOX-
                                         reactor
                                    Figure  4.  High dust SCR-arrangement  at Heilbronn 7

-------
CD
CO
         rel.
       activity
         K/K0
                              2500    4600    6300   8500         12000
                                                     operating hours [h]
18500
           NH3-slip
            [vpm]

            5

            4

            3

            2

            1
                  Figure  5.  Heilbronn power station unit 7;  DENOX-plant.  Loss of activity and ammonia slip

-------
00

CO
-.
• V.,.  ' A • • V •

                                 Figure 6.  Fly-ash  plugging  in  the  DENOX-reactor of Heilbronn 7

-------
                          NH3/NOX



vP
0s
0
E
3
0
CO
0)
O
O
CO

50 -
45 -
40 -
35 -


30 -


25 -

20 -


15 -
10 —
 5--
    0     5     10     15    20     25    30

           irregular distribution %
   Figure  7.  Heilbronn power station
   unit 7; DENOX-plant. Need of access
   volume  vs  irregular distributions of
   temperature, velocity and NH3/NOX
                 4B-33

-------
CD

CO
          o
          CO

          0)

          '•?
          JB

          
CO
to

JZ
CO
CO
I
o
o
£
                        Figure 8.   Heilbronn power station  unit  7; DENOX-plant. Loss of activity and NH;
                                   in fly-ash in dependence of plugging; replacing of the 1st layer

-------
CD
CO
cn

                                    NH?-
                                  injection
                                               T  T
DENOX-
 burner
                                                              \
G

A

V

O
                                                              regenerative
                                                             gas preheater
                                                                                         | stack

                                                                                         I   bypass-burner
                                                                                                        DENOX
                          Figure 9.   Tail end arrangement of the DENOX plant  at Heilbronn unit  3-6

-------
                              high-dust
                              system
                          tail end
                          system
CD
CO
                    7.4 - 7.0
                    430 - 470
                    1.0
                    320 - 400
                    100
pitch
surface
activity
temperature
pressure loss
  mm
m2/m3
           4.2
           750
       1.0-1.2
(270) 300 - 320
    -180-250
                             Figure 10. SCR catalysts for slag-tap firing systems

-------
Temperature                        290 °C -  320 °C
NOX5 behind SCR                    150 mg/m3
NH3-slip                            <  1 ppm
Pressure drop
   total  plant                       37  mbar
   reactor                           7,5 mbar
   GAVO                           14,0 mbar

Start up- /  Shut down-times
   heat  up  from 20 °C to 320 °C     51/2 hours
   cool  down from 320 °C to 20 °C   24    hours

Operational values
   natural gas consumption          1.300 ms/h
   NH3 consumption                   170  kg/h
   electrical power consumption         4 MW

          Figure 11.  Operating  Characteristics of DENOX 3-6
                             4B-37

-------
        SO3 GENERATION -
JEOPARDIZING CATALYST OPERATION?
          R. Jaerschky
             A.  Merz
           J. Mylonas
       Isar-Amperwerke AG
       Brienner StraBe 40
     8000  Munchen  2,  Germany

-------
                           S03  GENERATION  -
                   JEOPARDIZING CATALYST OPERATION?
ABSTRACT

Isar-Amperwerke AG's modern,  hard-coal-fired combined power and  dis-
trict heating plant in Zolling went into operation at  the  end of 1985.
To ensure conformance  with  the applicable emission limits, the power
plant was  initially equipped  with  primary  NOx  reduction measures,  a
high-efficiency electrostatic precipitator and a desulfurization plant.
With the introduction  of a  more  stringent limit for NOX emissions,  it
became necessary  to retrofit  a  DeNOx  plant.  This DeNOx plant, which
functions  on the  principle  of  selective catalytic  reduction using
ammonia,  went into operation at the beginning of 1988  and  achieved the
required separation efficiency without difficulty.

After a brief period of operation, however, acidic particles started to
be emitted.  Extensive investigations revealed that these emissions  were
the result of the  catalysts'  having  a very high SO2/SO3  conversion rate.

On the basis of  the  investigations results, steps were taken which
reduced the  emission  of acidic  particles to an absolute minimum.  It
became apparent, though,  that a permanent solution to the problem would
require replacing the  catalysts.

For this reason,  the  catalysts were replaced mid-1990 after approxi-
mately 12,000 hours of operation by a  new type  with a  much  lower  con-
version rate.

This paper  reports on the  operating  results  obtained  with  the  old
catalysts,  the investigations carried out regarding SO2/SO3 conversion,
and first experiences  gained with the new fill.
                                 4B-41

-------
ZOLLING POWER PLANT

Zolling Power Plant - formerly known as Leiningerwerk Power  Plant  Unit
5 - benefits from a number of environmental protection features:

         High efficiency thanks to supercritical steam conditions

         Extraction of district heat

         Highly efficient flue gas cleaning with DeNOx,  dust removal
         and desulfurization

         A liquid waste processing system with ammonia stripper

         A pleasant architectural design (Fig.  1)
                    Figure 1. Zolling Power Plant
                                4B-42

-------
The technical data of the Zolling Power Plant are listed in Table 1:



                               Table 1

                            TECHNICAL DATA
    Max. thermal  output  of  furnace                    1,144 MW
    Main steam mass  flow                               384 kg/s
    Main steam temperature                              540 °C
    Main steam pressure                                 274 bar
    Reheat temperature                                  540 °C
    Reheat pressure                                      55 bar
    Gross output  at  terminals                           450 MW
    Net output at terminals                            420 MW
    Max. district heat extraction                       270 MW
     (at a gross output at terminals of:                 392 MW)
The power plant burns a mixture of German hard coals with a low ash and
sulfur content (see Table 2) . It operates primarily in the lower inter-
mediate load range and  is frequently shut down at weekends and at night.
                               Table 2

                    COAL ANALYSIS (AVERAGE VALUES)
    Water                                              8.51 %
    Ash  (free  of water)                                7.52 %
    Volatile substances  (free  of water and ash)       28.44 %
    C content                                         73.46 %
    H content                                          4.32 %
    N content                                          1.54 %
    S content                                          0.96 %
    Net  calorific value                               28.99 MJ/kg
FLUE GAS CLEANING FACILITIES

Table 3 correlates the applicalbe emission limits in Germany with  the
flue gas values at the boiler outlet and at the  stack of  Zolling Power
Plant:
                                 4B-43

-------
                                Table 3
             Dimension         Raw  gas      Limit      Clean gas

           mg/mN3  (6%  O2)          650         200          190

           mg/mN3  (6%  02)        1,900         400*'        200

     Dust  mg/mN3  (6%  O2)        7,100          50            5

     )  but at least 85% SO2 separation
NO
SO
To  ensure conformance with the applicable limits,  the power plant is
equipped with a high-dust DeNOx plant,  an electrostatic precipitator and
a  flue gas desulfurization plant  (FGD) ,  the arrangement  of which is
shown  in  Fig.  2.
                    Figure 2.  Flue Gas Flow Diagram

The following  in-furnace N0x controls are  also  incorporated into the
boiler:

         Dynamic classifier for finer pulverization

         Low-NOx  burners for quasi-stoichiometric combustion

         Air staging with over fire air
                                 4B-44

-------
These primary measures result in a flue gas N0x content of about 650 mg
NO2/mN3 (6% O2)  at  the boiler  outlet,  together with  a low  nonburned
residue content of the fly-ash  (less then 3%)  [1, 2].
DENOx PLANT

Zolling power plant initially went into operation without a DeNOx plant,
but  a  retrofit was rendered  necessary by increasingly  stringent NOx
limits.

With a  view  to linking the boiler  outlet  and the air preheaters,  we
selected the high-dust configuration (see Fig. 2), which is more favor-
able from the energy point of view.

We  decided on  plate-type catalytic  converters on  account of  their
superior corrosion resistance and lower pressure  losses. The main design
data of the DeNOx plant are summarized in Table  4.
                               Table 4
                      DESIGN DATA OF DENOx PLANT
     Flue  gas  volumetric  flow (damp)                     400 mN3/s
     Flue  gas  temperature min./max.              300 °C/ 400 °C
     NOX reduction                                        70 %
     Max.  NH3 slip after 12,000 hours of operation         5 ppm
     Volume  of catalytic  material                        522 m3
     Specific  surface  area of catalytic  material         330 m2/m3
OPERATING RESULTS

The objective of 70% NOx reduction was reached without difficulty throu-
ghout the service life  of the  catalysts  (approximately  12,000 hours).
Activity checks  carried out on the  catalysts  showed  that deactivation
took place much  more  slowly than had been postulated in their design
(Fig. 3) . After 12,000 hours of operation,  the  remaining activity of the
catalysts was  still  94% of the  original  value.  This minimal  drop in
catalyst activity  resulted in only a small  increase  in the  ammonia
content of the fly-ash from about 4  mg NH3/kg at the start to around 10
mg/kg prior to the catalyst  replacement. Figure 4 shows the increase in
ammonia content  of  the fly-ash  in  the first quarter  of 1990;  these
values point to  an NH3 slip  of  approximately 0.5  ppm.
                                 4B-45

-------
       110
       100
     .
     CO
     0)
     CC
        90
        80
                                          Actual activity deterioration

                                         Designed activity  deterioration
                                                                            no
                                                                            100
                                                                            90
                                                                            80
                              5 000
                                                   10 000
                                                                            70
                                                                        15 000
                                  Hours of operation [h]
    Figure  3.  Activity Deterioration of  Catalyst at  Zolling Power Plant
        30
        20
      E
      Q.
      a..

      m'
15
      o
        10
             Ammonia concentration
                    of fly-ash

                 Linear regression
                                                                             30
                                                                     25
                                                                             20
                                                                     15
                                                                             10
         "  	I     I      I      I	I"""	I	I	
                     01.01.  15,01.   31.01.   15.02. 28.02.  15.03.  31.03.  15,04.  30.04.
                                 Spot-check analyses 1990


Figure 4.  Ammonia Concentration  in Fly-Ash  Prior to Catalyst  Replacement
                                       4B-46

-------
Throghout the entire duration of operation, no problems were encountered
with fly-ash deposits on the catalytic material; at the full-load duty
point the pressure  loss across all the catalyst layers  (htotal =  2m) re-
mained virtually constant  at  3  mbar for  the entire 12,000  hours  of
operation.

After only a short period of operation all  these positive aspects were,
however, overshadowed by the emission of  acidic particles, which caused
damage to  the paintwork of  cars  parked  in the vicinity  of the power
plant and jeopardized the  good reputation  of our  power  plant.
S03 GENERATION

After extensive investigations and measurements it became apparent that
the emission  of acidic particles  was  the result of oxidation of SO2 to
S03 in  the DeNOx catalytic converters. This sulfur trioxid subsequently
reacts with the steam, of which there  is plenty in the flue gas,  to form
gaseous sulfuric  acid.

The S02 conversion rate K (S02) is applied to quantify the oxidation of
S02 to  S03
K (S02) =   - — — - -   X 100%
                           2^
              '   a' downstream catalyst     '   3' upstream catalyst

                             upstream catalyst


with the C  (SO2)  and C (SO3)  concentrations entered in ppm.
The following  parameters have a significant  effect on the conversion
rate  (cf.  [3]) :

         The chemical  composition  of the  catalytic  substance  -
         particularly the vanadium pentoxide  content

         The ratio of the surface area of the catalytic material to
         the flue gas volumetric flow

         The flue gas conditions, particularly the  temperature
Figure  5  shows, as  a function  of  the flue  gas temperature,  the S03
concentration C (SO3)  measured upstream and downstream of  the  catalytic
converters in the course  of  the  investigations.
                                 4B-47

-------
     30
     20 -
    E
    CL
    Q.
    1__
    CO
    O
    05
    O
15 -
     10 -
      5 -
           upstream of catalyst

          downstream  of catalyst
                    O
                                                                  30
                                                            - 25
                                                                 - 20
                                                            - 15
                                                                 -  10
                                                                 - 5
              A /-^
                JO
                ' O O
       300
                  320
                            340        360
                            Flue Gas Temperature [°C]
                                                380
                                                          400
              Figure 5. SO3 Concentration in the Flue Gas

 It became apparent that the SO3 concentration in the flue gas was raised
 considerably by the DeNOx catalytic converters,  and that there was  a
 noticeable  dependence on  the flue  gas  temperature. At  a  flue gas SO2
 content  of  665 ppm,  the SO2 conversion  rate calculated for  375  °C was
 1.2%  and for 405  °C  3.3%.

 At low load  (120 MW) and temperatures of around 315 °C, the SO3 content
 was lower downstream of the DeNOx plant than upstream, i.e. at this duty
 point the catalyst  stores S03.
S03 CONCENTRATION IN FLUE GAS PATH

In order to better understand the mechanisms  involved in the formation
of acidic particles,  several series of measurements were taken to deter-
mine the SO3 concentrations at a given  time at various points along the
flue gas  path downstream  of the DeNOx plant.  Figure 6  plots the SO3
concentration downstream of  the catalytic  converter,  downstream of the
air preheater,  and in the stack (i.e.  downstream of  the regenerative
gas reheater); this series of measurements begins at  low load (110 MW;
flue gas temperature 315 °C) , with a rapid increase to full load between
6.00 and 7.00 a.m., causing  the temperature to rise to 375  °C.
                                 4B-48

-------
  30
  25 -
   20 -
 Q_
 a.
 O
 CO,
 o
   10 -
   5 -
   downstream of catalyst
           	A	

downstream of air preheater
           	e	

        in  the  stack
                           110MW
                           315°C
                                       450 MW
                                       375 °C
                        n           r
                        6           8
                          Time
                                                       10
                                                                   30
                                                                 - 25
                                                                 - 20
                                                                 - 15
                                                                 - 10
                                                                 - 5
Figure  6.  SO3  Concentration  in  the Flue Gas as  a Function of  Time
GU —
25 -



20 -

Q.
£;
co 15 -
O
CO,
O
10 -

5 -


















0 	


downstream of catalyst
downstream of air preheater
	 9- 	

in the stack
	 B- 	



250 MW
345 °C


^ 	 L




\ \
12 14





/
/
1
.
1
{





4
/
l
I
1
/
i
l









k"" X-X.
X
	
""A- — -
~~ — A



450 MW
390 °C


Activation of steam air preheater
I
1 .--.
* ,'
-+''''* 	 V"


- 25



- 20



- 15



- 10

- 5
-
	 1 	 1 	 1 	 •- o
16 18 20
                                  Time

Figure  7.  SO3  Concentration  in the Flue Gas  as  a Function  of Time
                                4B-49

-------
A similar experiment  is shown  in Figure 7, but  in this  instance there
was no soot  blowing,  which meant that the flue gas temperatures were
higher. Starting from 250  MW and tFG =  345  °C, the plant was  run up to
full load between 15.30 and 16.00,  causing the flue  gas  temperature to
rise to 390 °C. At  18.00 the flue gas temperature downstream of the air
preheater was raised  from  130 °C to 150 °C by activating the  steam air
preheaters.

In both cases  the  SO3 concentration downsream of the  DeNOx plant rose
sharply on load increase and stabilized after about 3 to 4 hours at a
somewhat lower level; this phenomenon is the  result  of thermal desorp-
tion of stored S03  (see  Fig.  5) .

The SO3 concentrations downstream of the air preheater were considerably
lower, which can be accounted for by  the fact that the temperatures in
the  region of the  air  preheater are below  the  acid  dew point.  This
results in sulfuric acid precipitating on the fly-ash,  of which there
are  large quantities  (sulfur content of the  fly-ash 0.5%).

The  amount  of sulfuric  acid  which precipitates depends  to a  large
degree on the temperature  in the air preheater, as is clearly  illustra-
ted  in Figure 7. Activating the steam air  preheaters,  thus causing the
flue gas temperature  downstream of the air preheaters  to rise from 130
°C to  150  °C,  resulted  in  a  decrease  in the  precipitation of SO3 from
24 ppm at  130  °C to 15 ppm at  150  °C.

The measurements also revealed that almost all of the SO3 still present
in the flue  gas downstream of the air preheater precipitates  in  the
regenerative  gas  reheater;  in all experiments  the concentration  of
gaseous  SO3  in the stack  was  close  to the  minimum detectable  level
(approximately 0.5 ppm). As there is virtually no fly-ash in the  region
of regenerative gas reheater to absorb the precipitated  acid  fraction,
further investigations were carried out to determine what was  happening
to the sulfuric acid.

After  one week's operation, a  plate was removed  from  the regenerative
gas reheater. On the cold  side  of the heat exchanger plate there was an
oily deposit of concentrated sulfuric acid which, extrapolated for the
entire heat transfer surface of the regenerative gas  reheater, amounted
to a stored volume of 600  kg of sulfuric acid [3].

Solids such  as residual dust,  iron oxides and gypsum  are deposited in
this acid film. The rotation of  the heat  transfer plates allows these
acid-soaked particles to pass to the  treated  gas side  of the  regenera-
tive gas reheater,  where they can break away and, given a sufficient gas
velocity, be carried  out of the stack.

Acidic particles were  emitted primarily on  rapid load increase, particu-
larly on plant run-up after weekend shutdown.

                                 4B-50

-------
MEASURES TO PREVENT ACIDIC EMISSIONS

On the basis ot the investigations, the following steps were  taken  to
prevent the emission of acidic particles:

         Soot blowing was stepped up to ensure  that the flue gas
         temperature at boiler outlet remains below 380  °C, even at
         full load. As Figure 5 shows, this  halves the SO3 concentra-
         tion downstream of the air preheaters,  i.e.  it reduces the
         SO2 conversion rate from 3.3 % at 400 °C to 1.2%. As Figures
         6 and 7 show, most of this SO3 fraction can be seperated
         out in the air preheater.

         The flue gas temperature downstream of the air preheater
         was lowered,  but at temperatures  of less than 130 °C the
         pressure losses  began to increase considerably, which was
         apparently the result of ammonia  bisulfate deposits. The
         temperature was  therefore  raised  again to 135  °C.

         Baffles were fitted to  the floor of the  flue  gas ducts
         upstream and downstream of the regenerative air reheater
         for separation of the large particles  (see Figure 8).
         Figure 8.  Baffles for Separation of Large Particles
                                 4B-51

-------
         During each weekend  shutdown  the flue gas ducts  in the
         region of  the  regenerative gas  reheater were  freed of
         particles.

         The blowing procedure of the regenerative  gas reheater was
         modified;  the hot and cold untreated gas  sides are blown
         simultaneously so that the loosened particles remain on the
         untreated  gas side.

         The plastic heat exchanger plates were removed  from the
         regenerative gas reheater in order to reduce  cooling and
         reheating  and to achieve a smaller separator  surface.

         The facility for drying the treated gas from the FGD plant
         using hot, SO3-laden untreated gas was taken out of service.

         During each weekend shutdown the regenerative gas reheater
         was flushed out with large quantities of low-pressure water
         (10 bar).   After  each flushing cycle  (duration  3  hours,
         volume of water around 150 m3) ,  the pH of the water dischar-
         ged was measured;  the flushing process was terminated as
         soon as a  virtually  neutral pH (around 6)  was reached.

As a result of  applying these measures,  no emmisions of  acidic particles
have been  detected  since April 1989. However, the  effort involved and
the damage done to the plant components  by the sulfuric  acid  (corrosion,
shortening  of  service life,   etc.)  are tremendous. It  was therefore
decided that,  in the long term,  the catalyst  would have to  be replaced
by a type with a lower conversion rate.
THE NEW CATALYST

Our Japanese suppliers, like all well-known manufacturers  of catalytic
converters,  invested  a lot of  effort  in  developing a low-conversion
catalyst.

Since the  reduction of NO  by  NH3  takes  place on the  surface of  the
catalyst, while the oxidation  of SO2 to SO3  is a volumetric reaction,
i.e. it increases linearly with  the volume of the catalytic material,
the plate  thickness  of  the catalytic  converters was  reduced,  thus
decreasing the volume while maintaining the  surface  area for the DeNO
process.

A further significant reduction in the  conversion  rate was achieved by
refraining from adding vanadium pentoxide to the catalytic material.
However,  this heavy metal promotes DeNOx activity, particularly in the
300 -  370 °C  temperature range,  this measure  resulted in a  10%  increase
in the necessary catalyst volume to 574 m3.

                                 4B-52

-------
We were assured that this new catalyst would,  at  the  same NOx reduction
efficiency  of  70%,  have a maximum ammonia slip  of 5 ppm after  16,000
hours of operation. An SO2 conversion rate of  0.9% at  400  °C was antici-
pated;  inservice  measurements  under  normal  power  plant  conditions
yielded values of around 0.7%. Right  from the start,  however, a notice-
ably higher NH3 slip of about 1.5  ppm was registered.
FIRST OPERATING RESULTS WITH THE NEW CATALYST

The  catalyst was  replaced during the  unit outage  in  June/July  1990.
Acceptance  testing was performed in September  1990 with the following
results:

The  catalyst achieves the required N0x  reduction without the increased
NH3 slip anticipated on the basis of the experimental measurements. Spot
measurements showed the maximum to be  0.5 ppm NH3 and the average  0.3
ppm. These values were  confirmed by the fact that the NH3 content of  the
fly-ash after catalyst replacement (Figure 9) is similar to that prior
to replacement (Figure 4).

Figure  10 presents  the results of  the SO3 measurements  upstream  and
downstream  of the catalytic converter.
      30
      25 -
      20
    E
    Q.
    a.
    co"
    Z^
    o
15
      10
          Ammonia  concentration
                 of  fly—ash
              Linear regression
       15.08.  31.08.  15.09,  30.09. 15,10.   31.10. 15.11.  30.11.  15.12.  31.12.  15.01.  31.01.
                          Spot-check analyses 1990/1991

Figure 9.  NH3  Concentration in the Fly-Ash after  Catalyst Replacement

                                  4B-53

-------
    30
    25 -
     20 -
   E
   Q_
   Q.
   (FT
   o
   c/^
   O
15 -
     10 -
           upstream of catalyst
          downstream of catalyst
                                                                 30
                                                                - 20
                                                           - 15
                                                                - 10
                                                                - 5
       300
                 320
                           340
                                     ~
                                     360
                                               380
                                                         400
                           O-*>_>        O'-'VJ
                           Flue Gas Temperature [°C]

             Figure 10. SO3 Concentration in the Flue Gas

At  a  constant  SO3 content  upstream of  the catalytic  converter,  the
values downstream of the DeNOx plant are considerably lower than those
shown in Figure 5. At 394 °C, only 7 ppm were measured  (old catalyst >15
ppm) , thus confirming the SO2 conversion rate of 0.7% at 400  °C  in actual
power plant operation.

As was the case with the old  catalyst, storage of SO3 occured in the low
load range (tFG = 315  °C) .  Figure 11 reveals  the surprising fact that no
detectable release of the stored SO3 was observed on load increase from
200 to 450 MW.

Figure 11 also  shows that  the SO3  concentration  downstream of the air
preheater is close to the minimum detectable level  of  about 0.3 ppm. It
should be added that,  under all load  conditions,  the gaseous  SO3 and
sulfuric acid aerosols  detected in  the  flue  gas ducts  downstream of the
air preheater were always  in the minimum detectable  range.

As anticipated,  the total  S03  fraction  can  therefore be separated out
in the air preheater by reducing the temperatures to values  below the
acid dew point.  The fly-ash analyses confirm that the  S03 content of the
flue gas upstream of the  air preheater  is  now  considerably  lower; at
otherwise constant values,  the  sulfur content dropped from 0.5% to 0.3%.
                                 4B-54

-------
    Q.
    CL
    o
    CO.
    o

5 -
0 -

0 -
5 -


upstream of catalyst
	 A 	
downstream of catalyst
	 e —
downstream of air preheater
	 E 	
200 MW
310°C





450 MW
365 °C


__ _-e"''
i
2 4

- 25
- 20
- 15
- 10
- 5
III u
6 8 10 12
                                  Time

Figure 11.  SO3 Concentration  in the Flue Gas Path as a Function  of Time

 Since catalyst replacement, there has apparently been no further preci-
 pitation of sulfuric acid in the  regenerative gas reheater. On flushing
 the regenerative gas reheater with LP water, the first water discharged
 was found to be approximately neutral  (pH > 4.5),  so it was decided to
 extend the flushing  intervals.

 To  summarize,  the SO2  conversion rate is one of the  most significant
 criteria  to  be considered when  selecting a high-dust  catalytic con-
 verter.

 Our experience with a low-conversion  catalyst has shown that a trouble-
 free  operation of the DeNOx plant is possible, without risk  of  the
 emission of acidic particles.
REFERENCES

1.   G. Musset, U. Schroder, E. Swoboda and D. Kiefer. "Operating Experi-
     ence with the Low-NOx Firing Concept in Unit 5 of the Leiningerwerk
     Power Plant  of Isar-Amperwerke AG"-VGB-Kraftwerkstechnik. Vol. 69,
     No.  4,  April 1989.

2.   R.  Jaerschky and A. Merz.  "NOx  Reduction  at  Zolling Power Station
     Pre-Combustion  and In-Furnace Measures,  SCR  Catalyst Equipment".
     ASME Paper  90-JPGC/FACT.  October 1990.
                                 4B-55

-------
H. Gutberlet, A. Dieckmann, A. Merz and L.  Schreiber. "SO2 Konver-
sionsrate von DeNOx-Katalysatoren - Messung und Auswirkung auf nach-
geschaltete Anlagenteile"- Chemie  im Kraftwerk 1989. pp. 86-96.
                             4B-56

-------
SCR OPERATING EXPERIENCE ON COAL-FIRED BOILERS

              AND RECENT PROGRESS
               Edward S.  Behrens
      Joy Environmental Equipment Company
             Monrovia, California
                 Senichi Ikeda
      Electric Power Development Company
                 Tokyo, Japan
                Teruo Yamashita
              Idemitsu Kosan KK.
                 Chita, Japan
            Gunther Mittelbach,  PhD
          Deutsche Babcock Anlagen AG
               Krefeld, Germany
               Makoto Yanai,  PhD
        Kawasaki Heavy Industries,  Ltd.
                  Kobe,  Japan

-------
1.0 - INTRODUCTION

Selective Catalytic Reduction (SCR) technology development traces
its roots back to the early 1970's.  Its use  for NOX reduction in
oil- and gas-fired  energy-conversion plants  has  long been widely
accepted as Best Available Control Technology (BACT) by many U.S.
regulatory bodies.   Its  acceptance on  coal-fired plants overseas
has demonstrated its ability to reduce NOX from these fuels.  But
its acceptance  in the U.S.  for coal-fired power  generation has
lagged  somewhat  due,  in part,   to perceived problems  revolving
around the typically hightr sulfur coals found here.  Nevertheless
over 50 commercial coal-fired plants overseas are  proving that with
proper physical and chemical catalyst designs and operation these
perceived problems  can  be overcome.  This paper will  detail the
operating experience of  three coal-fired  commercial power plants
using this technology to  successfully control their  NOX emissions
and present an up-to-date review of SCR technology.
2.0 - COAL-FIRED SCR OPERATING EXPERIENCE

Detailed below are the design  specifications and operating results
of SCRs  in thr^^  -oal-fired  power  plants  in Japan  and Germany
including  other  items  of  operational  interest  with  respect to
experience with the SCRs installed.


2.1 - Takehara Power Station

Electric Power Development Company's Takehara Power Station, Unit
1, in Hiroshima,  Japan is a 250-MW, coal-fired boiler burning 2.3-
to  2.5-percent  sulfur  coal.    It  uses hot-side  low-dust  SCR
arrangements in  two  parallel  SCR reactors A+B,  each  handling 50
percent of the flue gas.  The reactor B SCR was placed in service
in 1981,  and the  current catalyst charge has been in service since
1985 using Type  470  catalyst  elements having a 7 mm  pitch.   The
SCRs  are  located   downstream   of  the  hot-side  electrostatic
precipitator (ESP) and upstream of the  air preheater.   Flue-gas
temperature is 658'!  (348'C)  with a  NOX removal  efficiency of 80
percent at full load.

Figure 1  is a photo of the downflow reactor at Takehara, and Figure
2 is a sectional  elevation of  the SCR reactor showing three layers
of catalyst and the vertical vanes used to assure proper gas-flow
distribution.  Figure 3 is a computer generated diagram of the SCR
gas-flow showing  even gas distribution across the catalyst modules.
Catalyst modules are loaded from the side of the reactor by means
of a fork-lift track assembly.
                               4B-59

-------
The Takehara SCR unit 1-B
         FIGURE #1
                                                      Sectional Elevation of The
                                                       Takehara SCR unit 1-B
                                                          FIGURE  #2
              Gass Flow Analysis of
              Takehara Reactor Unit 1-B
                                               I
                         ;;;iiS%|£&nr  Vv
                            FIGURE #3
                                4B-60

-------
Figure  4  shows  fine  dust deposition  from  the  ESP  on  the  top
catalyst  module  after 34,000  hours of  operation  using the  current
large openings,  Type 470,  7 mm catalyst type which is a thin-wall
catalyst.   Flyash from the ESP is very fine and is highly adhesive,
and  is  one  of the  reasons a  high opening type  470  catalyst  was
selected.   Despite  the high SO2 levels  (1800 ppm)  entering  the SCR
no  plugging  of  the air  preheater  by ammonium  salts  has  been
observed.    No additional preheater washings  have  been necessary.
This  is  accomplished by  maintaining  low NH3 slip levels.

                                   Typical ash deposit with no sool blowing
                                   after 34,000 hr. operation at Takthara.
            FIGURE #4
Table  1 details design data  and Table  2  shows  actual performance
results from  annual performance tests  of  the SCR at the  Takehara
Power  Station.
                       Lo«d fMW« par SCR)

                       C*» Flow (Bcfat

                       T«*p«r«tur«  (T)

                       NO, In (ppwvd)'

                       SO, In (ppmvd)

                       SOj conwrvion (%)

                       fly A*h (gr/icf) dry

                       D*HO, Efficiency (%)

                       HH, Slip  Ippwvd)'

                       HO, Out ( ppWd )

                       CltalysC  d«lt* P tin K,0)

                       Cacjiyvc  contact ar«« |MJ/MJ)
  125

253,435

  659

  300

  1150

  0.5

  0. 04

  80
                          Corr»ct»d to 6%
                                   b»for« C»t»ly«t
                                  4B-61

-------
                               TABLE 2

                   TAKEHARA UNIT 1-B SCR PERFORMANCE TEST RESULT

Cat vcliiM
G»« TeKMratur*
NO, 1=
NO, Out
D»NO, t'flciency
Slip ST
Slip KB.

x 1,
•r
pp»
PP«
*
ppa
PP»

000 SCFM

(«% Oj)
(6* Oj)
(6» Oj)
(design
basis)
DEC.
1985
246
649
306
61
80
0.1
0.1
MOV .
1986
256
669
283
60
79
0.1
0. 1
OCT.
1987
235
649
251
45
82
0.2
0.2
FEB.
1988
246
644
249
52
79
0.1
0.1
OCT.
1988
239
651
270
57
79
0. 1
0.1
OCT.
1989
246
653
315
64
90
0.2
0.2
JUNE
1990
235
660
279
51
82
0.2
0. 2
      SO, Ir.
      SO, 1=
      SO, Out
      so,
ppm
ppm
ppm
1,490  1,130
 5.2   4.9
 7.6   7.4
 0.18   0.21
1,340
 3.2
 3.2
 0.08
1,210
 .8
4.6
0.15
 956
2.2
2.9
0.08
1,040  1,170
2.1   1.8
3.3   2.8
0.12   0.09
    Catalyse >u raplaod in th« periodic inspection becvwn Sept«m£«r and November, 1985.
2.2 - HKW  Reuter West.  Boiler E/D

The po^er  station Reuter West of the Berliner  Kraft- und Licht AG
in Berlin, Germany does  include 2  x 284  MW coal-fired power plants
burning various  coals with sulfur  contents up to 1.2 percent.  The
boilers D  and E of those  two  plants went on the  line in 1988 and
has  operated about 15,000  hours  till  November  1990.   It  is an
example of the hot-side, high-dust SCR arrangement. The SCR reactor
is located between the  economizer and the  air  preheater, upstream
of  any  particulate-removal  or  flue  gas  desulfurization   (FGD)
equipment. The  flue gas temperature entering the  SCR depends upon
coal  burned,  firing  rate,   and   the   condition  of  the  boiler
(dirty/clean);  the  normal  average  value  is  about  360"  C. NOX
removal  is more  than 85% depending  on  removal requirements; the
average ammonia  slip is about  1.5  ppmvd and the SO2 conversion to
SO3 about 0.5 percent.

Figure 5 is a photo of a  model  of  the SCR unit  at Reuter West, and
Figure 6 is a sectional elevation of one  of the half-capacity SCR
reactors of  one  boiler.    It  shows space  for  three  and one-half
layers of catalyst.   Layer No.  1 uses only one  module while  layers
No. 2  and 3  use  two modules  stacked  on top  of  each other;  the
fourth layer  is  a spare and  is currently  not  used.   It is  held
vacant  for use  in  future catalyst  replacement  programs  or,  if
needed, additional NO removal. Stacking modules facilitates
                                4B-62

-------
interchangeability of  the  modules  and  reflects  the  fact  that
catalyst elements  can only be extruded  up to  about 1 m in length.
The typical level of potential catalyst  poisons  found in  fly ash of
the Ruhr coal  (K20 -  4  to  5%;  CaO - 6.3%; MgO -  1.5%;  P205 - 0.6%)
have   not   appeared  to   significantly   accelerate   catalyst
deterioration.  Table 3 shows examples of measured operation dates
of the SCR of  one  of  the boilers at HKW Reuter  West.
        The SCR at Reuter West
           FIGURE #5
                                      	im_ ^ 	 	t-
                                                            Sectional
                                                            Elevation
                                                            of the SCR at
                                                            Rente- "'e_
                                                 FIGURE
                                 4B-63

-------
                                     TABLE 3
HVTK Reuter West
Boiler E
Dewag
EXAMPLES
OF MEASUBEB
OPERATION DATA
230
680
5.5
350
865
0.5
5.11
85
1. 1-2. 1
52
f/rf) 427
165
662
6.3
350
865
86
. 7-1.6
50
                  Load (MHe)
                  Temperature (F)
                  Avg. 02t
                  NO, In  (ppmvd)
                  SO2 In  (ppravd)
                  SO2 conversion (%)
                  Fly Ash (gr/scf) dry21
                  DeNO, Efficiency (%)
                  KH, Slip (ppravd)
                  NO, Out (ppmvd)
                  Catalyst contact area

                     "  Corrected to 6* O2
                     !)  Design Condition
                     31  Measured at 725 "F

The operation experiences  are good.   The activity  losses  of  the
catalyst  which   were  controlled  after  about  10,000   hours  of
operation were much  less than  expected.  Due to the relatively low
SO2-conversion rate  and  low  NH3-slippage no  plugging  at the  air
preheater has occurred and  no washing  of it was necessary since the
start up of the plant.  The soot  blowing of the catalyst layers is
done only one time per week.


2.3 - Aichi Refinery

Idemitsu  Kosan KK's   Aichi  Refinery  Boiler No.  4 power unit  in
Aichi,  Japan is  a 40-MW,  coal-fired boiler burning 0.4-percent
sulfur coal.  It was placed in  service in  1986 and  uses a hot-side,
high-dust SCR arrangement.   The SCR was designed for  an  efficiency
of about  59  percejit producing  an  outlet NOX consistently below 90
ppm.   Inlet  NOX  generally  ranges  between 200 and  250 ppm and the
SCR sometimes operates at >60% NOX removal, maintaining outlet NO
below 90 ppm.

The air preheater is a Lungstron type with  coirten elements.  This
air preheater was designed to  permit  NH4SO4  deposition at the cold
side of the  hot-side  element.   Since  start-up,  there has been no
need for washing  between the scheduled annual outages,  nor has any
element been  replaced  or repacked.
                               4B-64

-------
Aichi No. 4 power unit  is the  earliest  SCR system equipped with a
baghouse  downstream.    In  general,  it  has  been predicted  that
baghouses downstream of an SCR would experience increased pressure
drop due to NH3 leakage and SO3 increase by the SCR.   However, the
baghouse at Aichi continues to operate  without increased pressure
loss, because of the low NH3 slip and low S02 to SO3  conversion.


Figure  7  shows the  Aichi boiler  and  SCR  unit.   Figure 8  is  a
computer-generated  gas-flow diagram  for this  SCR  reactor  which
utilizes a turning vane to improve distribution minimize gas eddies
and reduce pressure  drop,  across the catalyst module  surface.  Once
per day,  soot  blowers  operate  to  clean the  first catalyst  layer
only.    There  has  been  no observation  of catalyst plugging  or
increased pressure  drop.    NH3  in  the  ash  was 28 ppm  after  5000
hours of operation.
       The SCR at the Aichi Refinery
Gas Dow analysis of the Aichi Reactor
           FIGURE  #7
        FIGURE #8
                               4B-65

-------
Activity of the  catalyst  has been tested annually and the activity
loss  was  very  small  compared with other high-dust  systems  (see
figure 10).  The major cause of deterioration  is deposition of Ca,
and Si on the  surface  of  the  catalyst.

As  a   test,  KHI  enclosed several   catalyst  elements of Type 555
geometry, among  the usual Type 470 catalyst in the Aichi SCR.  This
allows actual  exposure of  the test catalyst  to the  flue  gas for
extended periods.   Inspections of  the Type  555  over a  three-year
period showed  very  little plugging.  Similar to  Type 470 adjacent
to test elements.

Table 4 lists  design conditions  of the  SCR installed at  the Aichi
Refinery, Boiler 4.
                                  TXBLE «

                                Aichi Refinery
                                   Unit 4
                              Ideaitsu Kohsan Co.

                                          Design Ccnditiona


                    Load (HW«)                      40

                    Ga« Flow (scfn)               115,965

                    Temperature (F)                 716

                    Avq. Oj*                      3.56

                    HO, In  (ppnvd)'                  220

                    SO, In  (ppmvd)                  429

                    SO, conversion (%)               1.2

                    Fly Ash (gr/scf) dry        8.73  15.28

                    DeKO, Efficiency (t)               59

                    HH, Slip (ppavd),                3-5

                    HO, Out (ppwvd)'                  90

                    Catalyst d«lta P (in H20)          3.1

                    Catalyst contact ar«« (H2/)*3)       470



                    1  corr«ct«d to 6% Oz

                    • Kaximua allowable bafor* Catalyst Replacement
                                  4B-66

-------
2.4 - Present and Possible Future Catalyst  Geometries.

Present and possible future catalyst  element  geometries  are  shown
in  Table  5.   The  thin-wall  elements  contemplated  for future
installation exhibit improved NOX reduction and reduce S02 to SO3
conversion (see 3.1).  The thin-wall geometry  is also less prone to
fly ash plugging.  Since Type 555 has been  successfully  tested in
the Aichi reactor, and Type  572, which has a larger void/opening
ratio than Type  555, its future application looks promising.
                           TABLE 5

            Examples of grid honeycomb geometry
Cells per Side
Slaius
Georr... .]
Identification
Wall
Thickness (mm,
Opening
rale (%)
Application
Gas Firing
Boiler / Turbine
Oil Firing Boiler
Diesel
Coal firing Boiler
Municipal Waste
Incinerator

20x20
Conventional
Type 427
1.35
64







21x21
High Opening
Type 470
1.00
71







25)
Conventional
Type 555
1.00
68







c25
Hxjtl Opening
Type 572
0.80
73







35 >
Conventional
Type 751
0.80
64







L35
High Opening
Type 816
0.55
76







                               4B-67

-------
2.5 - NH3 Injection.

Proper mixing  of NOX  and NH3 is important  to  insure required  NOX
reduction.   Figure  9  shows the NHj injection  systems  employed at
Takehara   and   Aichi   which  provides   adjustability   for  gas
distribution  through  header  regulation and  individual  orifice
replacement.  During start-up of these NH3 injection systems, it is
necessary to fine tune  the AIG to  ensure performance.   However in
smaller  plants,  like  Aichi,   it  has been  demonstrated  that  the
standard deviation  of NH3 distribution is usually within allowable
limits.    However,   if  fine   tuning  is  needed  to optimize   NH3
distribution,  it is easily accomplished.
                   Typical SCR Ammonia Injection Grid
                                          to Reactor
Flow equalizer
 (Pipe grid)
                                Detail of NH, injection nozzle
                             FIGURE  #9
2.6 - Observed  Catalyst  Operation

At both Takehara and Aichi, catalyst physical and activity changes
over the life of the catalyst have been carefully monitored.  This
is done  as part of catalyst  management program.  Tests  are made
annually.  Since Takehara  is  a  low-dust installation,  erosion, of
course,  has  not been  experienced.   At  Aichi,  some slight erosion
has been noted  in  the top layer of catalyst,  but  it has  not been
severe   and   will   not  compromise  anticipated  catalyst  life.
Deterioration  is  relatively low for a  high  dust system.   Primary
cause of deactivation is deposition of Ca^i on catalyst surface.
Figure 10  shows the relative catalyst  activity reduction for the
three  coal-fired  plants  discussed versus  cumulative  operating
hours.  All  three  plants exhibit better than expected  catalyst
                               4B-68

-------
activity  after 20,000+ hours of operation.   Figure 11 shows the  KHI
catalyst  activity measurement  facility,  and Figure  12  shows their
erosion simulation  facility.
                            CATALYST ACTIVITY DETERIORATION
                              FIGURE  #10
            KHI Catalyst Activity
                Test Facility
             *—  A ^Tcr*-.  . ito'flE
KHI Catalyst Erosion
    Test Facility
                FIGURE  #11
       FIGURE  #12
                                 4B-69

-------
3.0- MODERN SCR TECHNOLOGY

The fundamental basis  of  SCR  technology  is  based on the catalyzed
reduction of NOX  (mixtures  of NO2  and  NO) with ammonia (NH,)  into
innocuous water  (H2O) and nitrogen N2)  in two general reactions:

     NO + NH3 + 1/402 -» N2 + 3/2H2O (main reaction)
     NO + NH2 + 2NH3  -"• 2N2 +  3 H2O

NH3 (in the  form of liquid anhydrous or aqueous ammonia), which has
been vaporized it is  diluted with air.   The  mixture  is injected
into the  flue-gas stream.    The  NH3 is  injected upstream  of the
catalyst  appropriate   to  NOX  removal   requirements   through  a
distribution grid.


3.1 -  Catalyst Properties

Ceramic,  homogeneous,  honeycombed  catalyst  elements  measuring
approximately  6-in.  x  6-in.  square are extruded up to  39-in.  long
(150  mm x  150 mm x  1  m) .    Titanium  oxide  (TiO3)  as the  base
material and is used to disperse  and support the vanadium pentoside
(V2O5) .   Tungstein  oxide (WO3)  provides thermal  and  mechanical
stability.  This titanium-based catalyst has been proven to provide
the highest durability and  excellent reactivity.  By changing the
mixing ratio of th--> active components,  the catalyst can be tailored
to meet specific  flue-gas requirements.

The vanadium content controls the reactivity of the catalyst.   But
it also catalyzes  the  oxidation  of SO2 to SO3.   Therefore in high-
sulfur applications,  it  is  necessary to  minimize the vanadium
content.    Through  homogenous  distribution  of V2O5   throughout
catalyst elements, activity reduction of possibly low V2O5 catalyst
is minimized.

Within the  honeycombed  catalyst  elements,  the incoming  NOX/NH3
mixture enters micropores on the catalyst's surface and diffuses
back out  after the chemical  reactions  have  taken place within the
catalyst material  itself.   Therefore one of the  goals  in catalyst
development has  been to  attain  a good mixture  of  macro-pores  to
support gas diffusion and micro-pores  to  support the reaction
itself.

Honeycomb   pitch   (flue-gas   passages)   can also  be   varied  to
accommodate a range of  flue-gas  dust  loadings.   The  sectional
geometry  of a few  of  the  flue-gas-flow patterns  used in these
catalyst  elements are  shown  in  Figure  13.   Since the  effective
depth   of  catalyst  for  NOX  reduction  occurs   near  the  surface
 (approximately 0.1 mm  deep) it is possible to reduce the catalyst
                               4B-70

-------
volume if the catalyst surface area is increased by increasing the
number of cells.   This produces  a  smaller cell  pitch with thinner
walls.   Alternatively, S02 oxidation  occurs  more  slowly and takes
place deep inside the catalyst material.  Therefore,  the SO, to S03
conversion  can be  decreased by decreasing  inner-wall  thickness
without  reducing NOx-removal activity.
                     Honeycomb Catalyst Elements
                            FIGURE #13
All  of these  catalyst optimization  features must,  however,  be
tempered in coal-fired applications by the fact that  large-passage
honeycomb  patterns  permit freer flow  of  the dirtier flue gasses
typically encountered, but they also present less  catalytic surface
area  for  reduction  of NOX.   In  general,  erosion-proof catalysts
demonstrate lower activity compared to catalysts used in  low-dust
environments  such  as gas/oil-fired  applications.    Erosion-proof
catalysts  inevitably have a  smaller volume  of  micro-pores, the
major  cause  of their  lower  activity.  Recent  developments have
succeeded  in  increasing   the  activity  of  these  lower-activity
catalysts to near optimum  levels.


3.2 - Catalyst Modules and Reactor Design

The honeycombed catalyst  elements  are assembled into steel-cased
modules of the required size,  Figure  14,  for  ease of handling and
installation.   Modules can be inserted into the  SCR reactor on
rollers  or by  an  overhead  crane.    Modules  are  then  stacked
horizontally or vertically within the  SCR  reactor  on engineered
support structures.   In coal-fired power plants, flue-gas flow is
vertically downward.  Figure 15, to facilitate the passage of fly
ash  through  the   catalyst   elements  with  minimum  drop  out.
Horizontal dimensions of the SCR unit are set to optimize gas
                               4B-71

-------
velocity pressure  drop and distribution through the  SCR  catalyst
elements; the number of vertical layers of modules  (usually 2 to 4)
is  determined  by  the desired  NOX  removal  efficiency  and  the
temperature of the flue gas.  NH3 is introduced at the inlet of the
SCR reactor through an Ammonia Injection Grid  (AIG)  system which
mixes the NH3  thoroughly with  the incoming flue gas before entering
catalyst-module  array.
                           Catalyst Module
                            FIGURE #14
                      Typical Downflow SCR System
                             FIGURE  #15
                                4B-72

-------
3.3 - SCR Catalyst Operating Conditions

Optimum flue-gas temperature for SCR reaction is between 500°F and
800°F  (260°C  to 427°C).   Below  this temperature  range,  chemical
reactivity is impaired, and above it, physical catalyst damage can
occur through sintering.  Catalyst degradation can also result from
poisoning through chemical reactions which  tend to neutralize the
catalyst's  reactivity and masking   caused  by ammonium  bisulfate
(NH4HSO4) , ammonium sulfate ((NH4)2SO4) and fly-ash deposition.  Any
of these adverse effects can cause NH3 leakage to increase through
the SCR which is generally limited to <5 ppm for high-sulfur coals.
This occurs  because  the catalyst has become less  reactive,  thus
requiring more NH3 to achieve the same NOX removal.  Figure 16 shows
NH3 slip  increases  with  time  with  constant DeNOx. While  design
criteria calls  for  slightly  over two year's operation before NH3
slip increases to the 5-ppm level,  a predicted mean operating life
of over three  years  is expected.   On the  other  hand,  measured
operating unit  experience shows that catalyst life of  well  over
four years can be anticipated.   This is an example of the catalyst
maintenance program used at Aichi.
                             Aichi SCR Catalyst
                                       'De s igne d
                                           Predicted Me an
                       . ;i-;^kr1^^---e	"l-ioasured Hctual
                           B   12   IB  2B   24   JB
                          Cumulative Operation Hours (XI008 hour
                    InlBl N0«:220 ppm, Outlet N0«:90 ppm, lype-170,
                            FIGURE #16
Figure 17 shows that the mole relationship between NOX removed and
NH3  consumed is nearly linear except in the high efficiency range.
In order to achieve high removal efficiency with low ammonia slip,
the catalyst volume must be increased.  This results in higher SO2
                              4B-73

-------
to S03 conversion.   Increasing temperature also increases SO2 to S03
oxidation.   This is,  of course, undesirable and  should be held  to
<3 percent  since SO3 promotes the formation of  ammonium sulfate and
ammonium bisulfate  which  can plug  downstream  heat  transfer and
emission-control equipment.   For a given  catalyst volume a typical
interrelationship  of  temperature,
reduction is shown  in Figure 18.
SO2 to  SO3 conversion  and NOX
                      NO, reduction >nd Nil, slip vj. Nil, / NO, mole ratio

                      100
                           04    06    08    10
                             NH-,/NOx Mole Ratio
                              FIGURE #17
                      NO, mlucllon and [Mrcrnl SO, cnrnTnlon »!. SCR Icmprrlure
£ 90
u
£ 85
u
UJ
s 80
3
•o -_
o> 75
IT
X
o
Z 70

.
/
/
High sullur /
calalysl /
*

/
/ -**
Lr -.---:''' . '
—
s
to
a
§
U
20 o
1 5 ^
o
1 0 ~.
05 °

                          600  650  700  750  800
                              Temperalure (°F)
                             FIGURE  #18
                                 4B-74

-------
Location of the SCR unit within a coal-fired power plant's flue-gas
stream is critical.  Three options are  available  here,  Figure .19;
(1)  a hot-side, high-dust  system is located after  the  economizer
and  upstream  of  the  air  preheater  and  any  emission-control
equipment such as an electrostatic precipitator  (ESP),  baghouse, or
FGD; and  (2) A hot-side,  low-dust ESP unit; and  (3)  A  cold-side,
low-dust system is located after the air  preheater and emission-
control equipment.
                  BOILER .	NH
                   Hot Side High Dust System
                  Hot Side Low Dust System
                   Cold Side Low Dust System
                            FIGURE #19
In a hot-side,  high-dust reactor,  a large-pitch catalyst  must  be
used  to  accommodate  the  heavy  dust   loading.     Furthermore,
reactivity of the catalyst is reduced to minimize oxidation of S02
to SO3.   Conversely, in a cold-side, low-dust  system,  most of the
particulates  and  SO2  have   been   removed  from  the  flue  gas.
Therefore, a small-pitch, high-surface area  catalyst  can be used.
Generally this  requires less catalyst volume,  and a  more active
catalyst can be used s-nne there is little concern for oxidation of
S02.  But the primary advantage  of  this  system as well as the hot-
side low dust  system  is the  reduced deterioration  rate  of  the
catalyst.   However,  reduced  operating  temperatures   can  require
reheating the flue gas.  In most operating coal-fired  power plants
the costs associated  with  reheating the flue  gas outweigh  the
savings from reduced catalyst volume and maintenance  of the cold-
side SCR.
                               4B-75

-------
3.4 - Catalyst Module  Replacement Schedules

Fifty to  60 percent  of the  cost of  an SCR  system lies  in the
catalyst.   Therefore  catalyst  life  and  associated  replacement
schedules have a  significant impact on  the  economics  of the SCR.
The life limiting factor for these catalyst elements is increasing
NH3 slip with time which should be held  to  <5  ppm,  for coal-fired
hot-side SCRs.
Figure 20 is a typical  SCR configuration,  and Figure 21 shows two
possible catalyst module  replacement programs that can be used.
                     Cathalyst Module Arrangement
                        in a typical SCR reactor
                                     NH3 INJECTION CONTROL
                 SOOT BLOWS
                                       CHTRLYST MODULE
                            FIGURE #20
In  the  first example shown, three layers of  catalyst modules are
used  initially with provision  for  a fourth  layer.   When the NH3
slip  has  reached  5  ppm  after  approximately  24,000  hours  of
operation,  a new layer of  modules  is added  in the  vacant bottom
position.  Then after an additional 16,000 hour of operation or so
when  the NH3 slip has again increased,  a  new  top  layer of moduler
is  installed,  replacing the original  top  layer.  When the  NH3 slip
again increases after another  16,000  hours  of  operation,  a new
intermediate  layer  is  installed,   again  replacing  one  of  the
original module layers.  In subsequent replacements,  a new  layer is
added,  and  the oldest  layer is  removed.

In  the  second  example shown in Figure 21, the full catalyst charge
is  replaced every  24,000  hours.   This  results  in  a  70  percent
                               4B-76

-------
 increase in catalyst consumption over life of the plant.  But, if
 the additional spare layer module  replacement program is chosen,
 due consideration must be paid to the capability of the ID fan to
 accommodate the  resultant  increased pressure loss  caused  by the
 addition of another catalyst layer.
                       16   2432  40   49  56   64   7   80
                         CUHULRTIVE OPERRTION (1000 hour*)
                    Predicted Progr tm of Ctt•1y11 R«oI «c t ng In High-Dust SCR
                            FIGURE #21
 3.5 - Catalyst Regeneration

 During  operation,  fine  dust  partical deposit  on the  catalyst's
 surface  causing the  micropores to  plug  and  reducing  activity.
 Generally, these particles can be removed,  so that the catalyst can
 be reused.  However,  regeneration methods do not fully restore the
 original  catalytic activity  because of such factors  as  sintering
 due to heat, but they do approach  it.  Currently,  it  appears  that
 the best  method of catalyst  regeneration involves sandblasting,
 using  a  sand  grain  size  of  O.lmm which  is  blown  through  the
 deteriorated catalyst's passages.  However, further development of
 this  technique is required  before  full  commercial  practice  is
 available.

 4.0 - CONCLUSION

This operating data on coal-fired  commercial power plants burning
medium sulfur coal indicates  that  SCR is an effective technology
 for reducing NOX emission and are not presenting  abnormal  operating
difficulties due to fly  ash or SO2/SO3 in the flue gas,  excessive
NH3  slip,  nor high  SO2 to SO3  conversion.
                              4B-77

-------
   TECHNICAL FEASIBILITY AND COST
             OF SCR FOR
      U.S.  UTILITY APPLICATION

             C.P.  ROBIE

            P.A.  IRELAND

UNITED ENGINEERS & CONSTRUCTORS INC
         WESTERN OPERATIONS

                AND

          J.E. CICHANOWICZ
  ELECTRIC  POWER RESEARCH  INSTITUTE

-------
                         TECHNICAL  FEASIBILITY AND COST
                                    OF SCR FOR
                            U.S. UTILITY APPLICATION
                                   C. P.  ROBIE

                                  P. A. IRELAND

                      UNITED  ENGINEERS & CONSTRUCTORS  INC.
                               WESTERN OPERATIONS

                                       AND

                                J. E. CICHANOWICZ
                        ELECTRIC POWER RESEARCH INSTITUTE
ABSTRACT


The cost of utilizing Selective Catalytic Reduction (SCR) for NOX reduction  in
both new and retrofit applications is presented.  Retrofit cases include hot-side
SCR technology applied to both PC and cyclone-fired units and post-FGD SCR
technology applied to a PC-fired unit.  Technology status is assessed based
primarily on recent European experience.  The impact of operational  effects and
resultaot modifications on downstream equipment are included in the  analysis.
The hot-side capital costs (December 1989 dollars) range from $78 to $87/kW for
the new PC-fired case, $125 to $140/kW for the retrofit cyclone case, $96 to
$105/kW for the retrofit PC case.  The single post-FGD SCR case evaluated is
estimated at $140/kW.  The hot-side levelized costs range from 5.3 to 5.9
mills/kWh for the new case, 8.2 to 9.1 mills/kWh for the retrofit cyclone   fired
case, and 5.9 to 6.5 mill/kWh for the retrofit PC-fired case.  The levelized cost
for the single post-FGD SCR case presented is 6.8 mills/kWh.

INTRODUCTION

The feasibility and cost of applying ammonia-based selective catalytic reduction
(SCR) to control nitrogen oxide (NOJ  emissions  from power plants firing U.S.
coals is of considerable current interest.  Although the NOX control  requirements
of the 1990 Clean Air Act Amendments (CAAA) focus on low NOX burner  technology
and other forms of combustion control, other factors   such as the CAAA NOX
emissions averaging provision, and strict NOX control  requirements considered  by
various state and local regulatory agencies provide the prospect of SCR
application in the U.S.  In fact, applications for several low sulfur coal-fired
facilities developed by independent power producers in selected northeastern
states either require SCR, or a detailed, factual accounting of the  feasibility
of SCR for the site.  The considerable extent of SCR application in  Japan and
Europe for low sulfur fuels has been a significant factor in promoting the
application of this technology in the U.S.

This paper completes the presentation of data from an EPRI-funded activity to
evaluate the feasibility and cost for various potential applications of SCR.


                                      4B-81

-------
This study addresses the following six applications, proposed as representing the
range of potential SCR  applications:

    1.   New Plant   low sulfur  coal
    2.   New Plant   high  sulfur coal
    3.   Retrofit    cyclone  boiler,  high  sulfur  coal
    4.   Retrofit    conventional  (wall  or T-fired)  boiler,  high  sulfur  coal
    5.   Retrofit    post-FGD (e.g.  reactor following  S02  scrubber)
    6.   Retrofit    oil-fired boiler

Results  for cases  4  and 6  were reported at the 1989 Symposium (8).   This paper
summarizes results  for  cases 2,  3,  and  5, with limited case 4 results repeated
for comparison.

DESIGN  PREMISES

Key design Assumptions.  SCR costs  are  significantly  influenced  by several  key
design  assumptions.   The most important design variables  used in this study are:

     1.    Catalyst  life    Several coal-fired European  SCR  installations  have
          operated  for over two years without catalyst replacement  and only
          moderate  measured loss in  activity.  A  catalyst  life of four years for
          coal-fired hot-side SCR applications and four years  for post-FGD SCR
          applications  has  been used in  this evaluation.

     2.    Catalyst  cost    Catalyst costs in Europe have decreased since  1985 by  a
          factor of approximately 2.5,  primarily  due to a  very competitive supply
          situation.   Accordingly, this  evaluation covers  catalyst  costs from
          $330/ft3  to $660/ft3, covering the  range seen in Europe.

     3.    Ammonia slip   Ammonia slip in European SCR  installations  is typically
          specified at 5 ppm, while  some utilities recommend even lower  levels (2
          ppm).   For several  coal cases  in this study,  both  5  ppm and  2  ppm  slips
          have been evaluated.

     4.    Space Velocity   Advances  have been made in  catalyst formulation to
          minimize S02  to S03 conversion,  to  develop smaller pitches and to
          provide resistance to fouling  by trace  elements.   These various advances
          are reflected in the space velocities used for the cases  evaluated.


 Case Definition.  In order to develop representative  costs  for both the new and
 retrofit SCR study cases,  typical power plant layouts and design conditions were
 selected.  In the case of the retrofits,  actual  U.S.  power  plant layouts provided
 the basis for design conditions selected.  For the new plant  application, design
 conditions and layout were selected based on similar  EPRI studies  evaluating the
 cost of flue gas desulfurization processes.  Six study cases  were  evaluated in
 this study,  however, only Cases 2 to 5  are the subject of this paper and are
 described in Table 1.   General arrangement drawings for the study  cases evaluated
 in this paper are provided in Figures 1 to 4.

 For the conventional hot-side SCR applications (reactor between  economizer  exit
 and air heater inlet),  the reactors were located above the  particulate  collection
 device.  In the post-FGD application,  a wet FGD  system precedes  the reactors
 which were placed above the heat recovery units  (Gas-Gas-Heaters).
                                        4B-82

-------
SCR Process Design.   To obtain budgetary SCR system costs,  a performance
specification for the catalyst and reactor was developed for each case.  The
specifications were  developed using fuel analyses, plant performance and
emissions data, and  desired control of NOX,  residual  NH3, and byproduct  S03.
Included in the specification was variation in certain process  variables such as
NOX removal  and ammonia slip  for  selected  cases.   Three  SCR system  suppliers
provided quotations  to these specifications.

The design basis and vendor supplied design data for each of the cases evaluated
in this paper are shown in Table 2.  Sensitivity analyses are provided for the
new plant, hot-side  design (Case 2) and the cyclone-fired,  hot-side retrofit
design (Case 3), to  show the cost and performance impacts of reducing the ammonia
slip from 5 to 2 ppmvd.  For retrofit of hot-side SCR to a conventional
pulverized coal-fired boiler (Case 4), the effect of reducing the uncontrolled
NOX emission rate (by adding  combustion controls)  while  still meeting the same
NOX emission limit is evaluated;  specifically,  lowering  uncontrolled NOX  emission
rate from 0.60 to 0.40 Ib NOX/MM  Btu reduces the  SCR NOX  removal  from 80% to  70%.


Consistent with typical practice, one reactor per air heater was used as the
design basis; the cyclone-fired retrofit (Case 3) uses a single reactor  (1 x 100%
tubular air neater), while the other hot-side SCR cases  utilize two reactors (2 x
50% trisector air heaters).  The post-FGD case utilizes  twin reactors because two
(2 x 50%) Ljungstrom heat recovery units were utilized.

The hot-side applications utilize downflow reactors, with additional capacity to
add a spare catalyst layer.  Also, steam sootblowers are employed in the design
along with ash hoppers and ash transfer equipment.  In the post-FGD application
the reactor is also  designed as a downflow unit with capacity to add a spare
layer.  The post-FGD reactor design does not require sootblowers and ash
collection hoppers.

The hot-side cases employ a catalyst with a 7.07 mm pitch (20 x 20 grid) while
the post-FGD case employs a catalyst with a 4.2 MM pitch (35 x 35 grid).  The
lower pitch (higher specific area) and higher activity (per unit volume) of the
post-FGD catalyst allows a space velocity considerably higher than required for
the hot-side cases.

The ammonia storage  and supply systems were designed using a truck unloading
station and a storage  island providing seven days storage at an MCR rating.
Steam vaporizers are utilized for ammonia vaporization and dilution air  is
provided from the discharge of the primary air fans in the hot-side cases, while
the post-FGD case utilizes separate dilution air fans.

SCR PROCESS IMPACT

The hot-side SCR process, because of its location directly downstream of the
boiler and upstream of the air heater, impacts every component of the flue  gas
train and the boiler itself through its effect on the air heater (and in some
cases the economizer).  The degree of  impact varies with power plant
configuration, environmental control components, type of fuel,  and  emission
control requirements.  The post-FGD SCR process impact is much less  severe
because of its location at the end of the flue gas train.

Hot-Side SCR (coal)  Impact

The impacts of hot-side SCR in coal-fired applications are summarized on Figure
5.  The principal impacts are on the boiler, air heater  and  ID fan.  Other


                                       4B-83

-------
impacts are on the particulate collection device (ESP), wet limestone flue gas
desulfurization (FGD) process, FGD reheat system, waste disposal system and water
treatment system.

Boiler.  The principal effects of hot-side SCR on the boiler will be the loss of
overall thermal efficiency, and additional operations and control complexity,
particularly for cycling units.  Also, auxiliary power consumed  by the SCR
process will reduce the net generating capacity.

Loss of thermal efficiency results from  air  heater modifications and  an
economizer bypass which will  result  in higher air heater  flue  gas exit
temperatures.  The result will be loss in the net generating capacity for the
same quantity of fuel consumption.

Air Heater.  The potential for formation of  ammonium  sulfates  and bisulfates
coupled with the presence of  fly ash  necessitates air heater modifications  in the
hot-side SCR cases.   Modifications to the air heaters in  the PC  boiler cases
include adding high pressure  steam soot  blowers  at both the cold and  hot ends,
adding high pressure  water wash capability,  replacing 24  gage  heat transfer
surface material with 18 gage, replacing intermediate and cold end double
undulating  (DU) heat  transfer surface with notched flat (NF) surface, and adding
bypasses and dampers  for on-line washing capability.

In the cyclone-fired  boiler case, to reduce  the rate  of ammonium compound
deposition and build-up, all  the existing 2" diameter tubes in the cold end, and
25% of the tubes in the hot end were replaced with 3" diameter tubes.  Also, a
steam  soot blowing system utilizing medium pressure superheated  steam at both the
hot and cold ends was added to reduce the rate of deposits.

In this case,  it is expected  that some residual  ammonia may be captured by the
FGD system resulting  in a build-up of ammonium species in the FGD liquor.
Although this may complicate  scrubber sludge reuse or disposal,  no cost impact
has been assigned.

Stack.  The increase  in the flue gas S03  concentration across  the SCR could
result in  increased opacity of the flue gas  plume.  Recent data  from an EPRI
sponsored  study with  a member utility shows  a direct correlation between stack
opacity and sulfuric  acid concentration.   To reduce opacity control  measures may
be required to reduce the S03 concentration.   A  typical  method  of reducing  S03  in
the flue gas would be to inject NH3 upstream of  the  ESP.   The  specific impacts or
costs  associated with this effect have not been  evaluated in this study, however.

ID Fan.  To overcome  additional pressure drop (up to  11" we) associated with the
hot-side SCR, the existing ID fans were modified.  For the retrofit cases it was
assumed that new, larger diameter wheels could be placed into the existing fan
housing to overcome the additional  static pressure drop.  The modifications
included replacing the fan wheel, shaft,  bearings and motor.

ESP.   SCR  effects on  the ESP  include higher  volumetric flowrate, higher negative
operating  pressure, higher S03 concentration, higher flue  gas  temperature and
precipitation of ammonium compounds on fly ash.

Higher flue gas volume (an increase of up to 9.4% in the PC-fired cases) results
from  higher flue gas  temperature (20°F),  lower flue gas static pressure, and
increased  mass flow  (the latter due to increases air heater leakage and dilution
air),  and  will have a significant impact on  ESP operation.  The  increase in flue
gas volume will effectively reduce the Specific  Collecting Area  (SCA) and the
                                       4B-84

-------
concentration of particulate in the flue gas.  The result will be that the ESP
may require additional power to deliver the same particulate removal efficiency.

Greater negative operating pressure could require re-enforcement of the ESP.
This effect was not considered in the capital cost analysis.

In the high sulfur coal applications, the S03 concentration in the flue gas is
estimated to increase by 18 ppm across the SCR.  Typically, an increase in S03
would be expected to reduce the fly ash resistivity significantly.  However,  the
increase in the flue gas temperature in the PC-fired cases (to keep the flue  gas
above the acid dew point) may counteract the effect of the S03 increase,  possibly
producing little net change.

Ammonium compound precipitation on the fly ash typically has a beneficial  impact
on ESP performance by helping the fly ash agglomerate, preventing reentrainment.

The cumulative effects of all the above could be significant on an ESP; a pilot
test program would be required to determine actual design and operations impacts.
In this study case it was assumed that the only net effect on the ESP operation
was an increase in power consumption by about 12%.

In the cyclone-fired boiler case the flue gas volume increase is expected to  be
3.8%.  This result is lower than the PC cases because of a negligible increase in
the leakage rate across the tubular air heater and only an 8°F flue gas
temperature increase at the air heater exit.  Only a slight increase in the ESP
power consumption was assumed in this case.

Ash Disposal/Reuse.  Ammonium compound content in the fly ash can have an impact
on waste disposal or marketing practices; for example, these compounds decompose
and release ammonia at elevated pH.  While Eastern U.S. coals are not alkaline in
nature and ammonia would not be expected to gas off upon wetting, fixation with
alkaline species could result in an ammonia odor problem.

Similarly, reuse options for fly ash contaminated with ammonium compounds may be
limited.  Direct use as an admixture in cement manufacturing may be jeopardized
if the ammonium compound content is too high.

FGD/Reheat.  The chief effect on the FGD system is an increase in the water
evaporation rate and steam reheat requirement.  The higher inlet temperature  and
higher mass flow rate will result in an increase in water evaporation in the
absorber, as well as a significant increase in steam use by the FGD reheat system
(SOT reheat assumed).

A slight increase in power consumption could occur from having to increase the
FGD liquor recirculation rate in order to maintain the same S02 removal
efficiency.  The higher liquor recirculation rate might be required as a result
of dilution of S02 in the flue gas,  and higher flue gas volumetric flow rate
(saturated gas flowrate).  This effect was not considered in this analysis.

FD Fan.  In the PC-fired boiler cases (e.g. employing Ljungstrom air heaters) the
FD fan will consume slightly more power to account for a higher mass flow rate.
The mass flow increase results from an expected higher air heater leakage rate.

Water Treatment.  Introduction of nitrogen species into the air heater wash water
requires additional water treatment equipment.  Nitrogen species are introduced
into the wash water as ammonium bisulfate and sulfates.  A biological treatment
                                      4B-85

-------
process is utilized to convert the nitrogen species to free nitrogen.  The
effluent is assumed to be discharged to the existing on-site water treatment
equipment.

Post-FGD SCR Process  Impact

The impact of post-FGD SCR on  power  plant  operations  and  equipment  is  less
significant than that expected with  hot-side  SCR,  as  the  SCR  reactor and
ancillary equipment follow all major process  equipment.   The  impacts are  shown  by
Figure 6.

Boiler.  The boiler is affected  only insofar  as auxiliary power consumption is
increased.  The  increase  in  the  auxiliary  power consumption (reduction in the net
generating capacity)  will  increase the Net Plant Heat Rate.  Natural gas  consumed
in elevating the SCR  inlet gas temperature will also  increase the NPHR.

ID Fan/Booster  Fan.   The  increase in the  flue gas pressure drop associated  with
the post-FGD SCR process  is  estimated at  14.5 in w.c.   The pressure losses  are
principally across the inlet and outlet of the Gas-Gas-Heater (GGH)  and the SCR
reactor.  Addition of a booster  fan  into  the  flue gas  train will  increase the
complexity in flow and pressure  control.   In  this case the booster  fans are
located  upstream of the stack; one booster fan is supplied for  each  SCR reactor
train.

Water  Treatment.  Nitrogen species will be introduced  into  the  air heater wash
water  as  a result of  ammonium  bisulfate deposition on  heat  transfer  surface.
With relatively  little S03 capture expected within the FGD system, some
additional S03 generation across  the catalyst, and the absence of fly ash, the
rate of  chemical deposition  on the GGH  equipment  is expected to be quite
significant.  A biological treatment  process was  included to treat the
wastewater.

FGD.   The SCR process affects  the  FGD system only indirectly.   Because of the
location  of the GGH,  FGD system  mist eliminator operation will be critical.
Excessive mist  carryover could result in loss  of heat recovery  (resulting  in
increased natural gas consumption) and  an  increase flue gas pressure drop,
possibly  limiting generation capacity in addition to detracting from plant heat
rate.

Stack.   Retrofit of the post-FGD  SCR process will almost certainly have an impact
on the stack.   If the original plant design included a wet stack, the 225°F GGH
exit gas  temperature  will require  liner replacement.  In this  design case  it was
assumed  that the original design  included  steam reheat (50°F)  and that the stack
was designed for approximately 180°F.  The effect of the increase in the flue gas
temperature to  225°F  was considered negligible.

Higher S03 concentration  in the flue gas may result from oxidation of S02  across
the catalyst.  While  some of the  S03 is likely to form ammonium/sulfur compounds
and deposit on the GGH surface,  there may  be a net increase in the SO,
concentration which could increase plume opacity.

COST DEVELOPMENT

To develop total process capital  costs, physical layouts of the ductwork and SCR
reactors were developed.   From these drawings, lengths of ductwork and structural
requirements were estimated.   All  costs are presented in December 1989 dollars
                                      4B-86
J3

-------
The operating and capital cost impact of SCR on other plant components was also
estimated.  For major pieces of equipment, such as the air heaters, ammonia
storage system and ID fans, vendors were consulted in developing the cost of the
modifications.  For smaller equipment items and piping runs, UE&C utilized in-
house data to arrive at equipment costs.

EPRI's Technical Assessment Guide (TAG) provided the basis to estimate fixed
operating and maintenance costs.   Variable operating costs were determined by
calculating utility and raw material consumption rates.  Considered in the
variable operating costs were the following:

         SCR catalyst replacement
         Ammonia consumption
         Ammonia vaporization steam
         Incremental Sootblowing steam
         Incremental ID/Booster fan horsepower consumption
         Incremental FD fan horsepower consumption
         Incremental ESP power consumption
         Water treatment chemicals
         Air heater efficiency loss
         Incremental FGD reheat steam consumption
         SCR catalyst disposal
         Incremental fly ash disposal cost
         Natural gas consumption

RESULTS

Selected results from this study are summarized in Figures 7 to 10, while
sensitivity of results to catalyst cost and life are provided in Figures 11 to
14.  Highlights are discussed as follows:

Capital Costs.  Total capital requirement (TCR) for each of the cases is
presented, indicating the contribution of the reactor/catalyst, structural
modifications and/or support equipment, air heater, ductwork, NH3 injection,  flue
gas handling, and contingencies.   Figure 7 shows capital cost is least for new
units, due to the absence of retrofit considerations, and reduced catalyst
quantity from lower boiler exit NOX emissions.   These same factors,  retrofit
considerations and boiler exit NOX emissions,  are  responsible for the cyclone
boiler having the highest cost for the hot-side application.  Post-FGD capital
cost is high due to the GGH, which adds significantly more cost than is saved
through simplifying reactor design and reduced catalyst quantity.

Decreasing the ammonia slip from 5 to 2 ppm (shown for both cases 2 and 3) is
expected to increase the TCR by about 12% due to a larger catalyst volume
requirement.

The cost impact on the SCR of reducing the boiler NOX emission rate from 0.60 to
0.40 Ib NOX/MM Btu  is shown by Case 4.0 and  4.1.   Reduction of the boiler NOX
emission rate (through combustion modifications),  while meeting the same emission
limit of 0.12 Ib NOX /MM Btu,  reduces the SCR capital cost by $9.4/kW.
(Level ized costs reflecting both capital and operating costs must be compared to
judge the full benefit.)

The catalyst and reactor cost represents about 40-50% of the TCR  in the hot-side
SCR cases.  In the post-FGD SCR case, the catalyst cost represents only about 17%
of the TCR.  The largest cost item in the post-FGD SCR case are the twin GGH's
used for heat recovery.
                                      4B-87

-------
The contingency ranges from 14.4% to 18.2%.  The highest contingency is assigned
to Case 3 due to uncertainties  in high sulfur coal applications, coupled with
tubular air heaters and  a very  high boiler NOX emission rate.

Levelized Cost.  Figure  8 presents  levelized costs for  the  same design  cases,
depicting generally the  same  trends between costs  for  new units, cyclone  boilers,
conventional  PC boiler,  and post-FGD  application.  The  data shows  that  variable
operating costs and fixed charges  represent about  50%  of total  levelized  cost  for
the hot-side  application.  The  most significant component of fixed charge is the
recovery of capital for  the reactor and  catalyst.   Similarly, the  most
significant component  for variable  O&M is catalyst replacement cost.  Comparison
of cases 4.0  and 4.1  shows the  benefit of adding combustion controls to reduce
the NOX reduction requirement of the SCR;  the results  indicate  that  the SCR  cost
can be"reduced  from 6.54 to  5.88 mills/kWh by  reducing the  boiler  emission rate
from  0.60 to  0.40  Ib  NOX/MM Btu.   In the  case of the post-FGD SCR  process, fixed
charges  represent  about  65%  of  the  total  levelized cost. Note that the results
consider a  0.93 mills/kWh credit for  a 50°F  steam  reheat system that is no longer
required upon retrofit of the post-FGD SCR process.   This credit would, of
course,  not  apply  for  units that employ  wet  stack  operation.

The levelized costs for  Case  3.0,  the  cyclone boiler,  are significantly higher
than  the costs  expected  with  retrofit  to  a PC-fired  boiler.   This  is due  both  to
higher capital  requirement and  catalyst  replacement  cost due to the  large  volume
of catalyst  required  in  this  application.

Figure 9  shows  levelized costs  in  terms of $/ton NOX removed.  Primarily,  the
data  shows  the  impact  of the  boiler NOX emission rate on the cost to remove a ton
of NOX.  The  cyclone-fired boiler (Case 3.0)  shows the lowest levelized  cost
 (about $l,100/ton  NOX  ).  Although the cost of SCR for application  to cyclone
boilers is  significant,  the high uncontrolled boiler NOX emissions  reduce  costs
on a  per ton  basis.

The highest  levelized  cost is shown by Case 4.1  where combustion controls were
 added to reduce  the SCR  NOX reduction requirement from 80%  to 70%.  Lowering the
 boiler exit  NOX emission rate correspondingly increased costs on a  per  ton basis.

 Figure 10  provides  a  more detailed cost comparison between  a  post-FGD and hot-
 side  SCR process  in terms of  levelized costs (mills/kWh).  The power plant,  fuel,
 and NOX  reduction performance is identical for both cases.   The levelized  costs
 for the two  process options are comparable, however, as  described  earlier, the
 reheat credit of  0.93  mills/kWh for the post-FGD process may  not be  applicable  to
 specific sites  if  a wet  stack is used.   Also, note that a 4-year  catalyst life
 was used in  the  post-FGD cost development, six years is  closer to  the currently
 expected life.   Catalyst replacement is the most significant O&M cost item for
 the hot-side  process,  while natural gas cost (and heat rate  penalty) is the  most
 significant  O&M  cost  item for the cold-side process.

 Effect of  Catalyst  Life  and Unit Costs.  Sensitivities of the cost results to
 both  catalyst cost  and life are provided by Figures 11 to 14.  Base  case
 economics were  developed assuming a four year catalyst life  for both hot-side  and
 post-FGD SCR  processes;  a six year catalyst life for the post-FGD  SCR is now
being predicted.   Base case catalyst cost of $660/ft3 was utilized; this cost
reflected budgetary quotations  from the primary  U.S. SCR catalyst  vendors with
coal-fired  experience.   It is possible that catalyst costs will approach those  in
Europe ($400-450/ft )  due to world market competition.
                                       4B-88

-------
The figures show that the SCR applications which require the largest quantity of
catalyst are most sensitive to both catalyst life and cost.   The post-FGD process
(Case 5) is the least sensitive due to its relatively small  catalyst charge.

CONCLUSIONS

Conclusions developed from this study are:

    •    The capital  cost of SCR in 500 MW (nominal)  size U.S.  plants is expected
         to be:
         A.   $96   $105/kW for hot-side retrofits to conventional  (tangential  or
              wall) coal-fired power plants.
         B.   $125   $140/kW for hot-side retrofits to cyclone-fired boilers.
         C.   $78-87/kW in new plant hot-side applications.
         D.   $140/kW for post-FGD retrofits.

    t    The levelized cost of SCR in U.S. coal-fired power  plants  (500 MW size
         range) is expected to be:
         A.   5.3-5.9 mills/kWh for new hot-side power plant applications.
         B.   5.9 to 6.5 mills/kWh for hot-side retrofits to conventional-fired
              units.
         C.   8.2 to 9.1 mills/kWh for hot-side retrofits to cyclone-fired units.

         D.   Approximately 6.8 mills/kWh for post-FGD retrofits to con-
              ventional units assuming a credit for reheat (0.93 mills/kWh).

    •    The levelized cost of removing a ton of NOX  utilizing  SCR  is  expected  to
         range as follows:
         A.   $3,300   $3,800/ton NOX for new coal-fired  plant  hot-side
              applications.
         B.   $1,100   $l,250/ton NOX for coal-fired  cyclone  boiler hot-side
              retrofits.
         C.   $2,750   $4,250/ton NOX for coal-fired  conventional boiler hot-side
              retrofits.
         D.   $2,850/ton NOX for post-FGD SCR retrofit  to a  conventional  boiler.

    •    The levelized cost of removing a ton of NOX  is  lowest  with high NOX
         emission rates.  The levelized cost of removing a ton  of NOX for  a
         cyclone-fired boiler with a 1.80 Ib NOX/MM Btu  NOX emission rate  is
         estimated at $l,100/ton NOX.

    •    The SCR capital cost in a new power plant application  is substantially
         less than in a retrofit application.  The cost of a new plant SCR is
         expected to be about 34% lower than a retrofit,  the lower cost is due
         largely to new boilers having lower NOX emission rates and an attendant
         reduced catalyst requirement, and the absence of costly existing
         equipment modifications required in SCR retrofit applications.

    •    SCR capital  costs are higher for cyclone-fired boilers because of their
         high NOX  emission rate.   The SCR capital  cost  for cyclone-fired units  is
         expected to  be about 45% higher than that expected  for conventionally-
         fired power  plants.

    •    Catalyst life and catalyst unit cost significantly  affect  levelized
         process costs.  For most hot-side SCR applications,  an increase in
         catalyst life from 2 to 4 years reduces levelized cost by  30%.  A
         reduction in catalyst unit cost from $660/ft3  to $450/ft  (for  cases
         assuming a four year catalyst life)  reduces  levelized  costs by 15%.

                                      4B-89

-------
    s    The levelized cost of NOX removal  for  both  hot-side  and  post-FGD SCR
         processes is similar, but the components of the cost vary significantly.
         Compared to hot-side SCR, post-FGD applications requires 30% more
         capital, but feature lower catalyst replacement costs.

REFERENCES
1.   Bauer, T. K., Spendle, R. G., "Selective Catalytic Reduction for Coal-Fired
    Power Plants:  Feasibility and Economics," Stearns-Roger Inc., EPRI  CS-3603,
    October 1984.
2.   Cichanowicz, J. E., Offen, G. P., "Applicability of European SCR Experience
    to U.S. Utility Operation," Proceedings:  1987 Joint Symposium on Stationary
    NO. Control,  EPA/EPRI, New Orleans,  1987.
3.   Cichanowicz, J. E. et. al., "Technical Feasibility and Economics of  SCR NOX
    Control in Utility Applications," Proceedings:  1989 Joint Symposium  on
    Stationary Combustion NOX Control,  EPA/EPRI,  March 1989.
4.   Electric Power Research Institute (EPRI),  TAG   Technical Assessment  Guide,
    Volume I:  Electricity Supply - 1986, EPRI P-4463-SR,  December 1986.
5.   Ellison, W., "Assessment  of S02 and  NOX  Emission Control  Technology  in
    Europe," EPA-600/2-88-013, February 1988.
6.   Nakabayashi, Y., Abe, R., "Current Status  of SCR in Japan," Proceedings: 1987
    Joint Symposium on Stationary NQX Control,  EPA/EPRI, New  Orleans,  1987.
7.   Necker, P.,  "Operating Experience with the SCR DeNOx Plant  in  Unit 5  of
    Altbach/Deizisau Power Station," Proceedings:  1987 Joint Symposium  on
    Stationary NOX Control,  EPA/EPRI,  New Orleans,  1987.
8.   Osborn, H. H., "The Effect of Ammonia SCR  DeNOx  Systems on  Ljungstrom Air
    Preheaters," C-E Air  Preheater, EPRI RP 835-2,  June 1979.
Table 1
Case Definition
PLANT DESCRIPTION
Case
Retrofit
Capacity, MW (gross)
Boiler Type
Air Heaters
Participate Control
S02 Control
Reheat
Gross Plant Heat Rate, Btu/kUh
Capacity Factor, %
Remaining Life, years
SITE CONDITIONS
Location
Seismic Zone
Urban Site
FUEL
Type
Area
Higher Heating Value, Btu/lb
Sulfur Content, wt. %
Ash Content, wt. %

2.0
No
546,600
PC
Ljungstrom
Baghouse
Wet FGD
yes
9,137
65
30

Kenosha, WI
I
No

Coal
III i no is No. 6
10,533
3.74
9.51

3.0
Yes
536,000
Cyclone
Tubular
ESP
None
No
9,974
65
20

Kenosha, WI
I
No

Coal
1 1 1 inois No. 6
10,533
3.74
9.51

4.0
Yes
536,000
PC
Ljungstrom
ESP
Wet FGD
Yes
9,197
65
20

Kenosha, WI
I
No

Coal
Appalachian
13,100
2.60
9.10

5.0
Yes
536,000
PC
Ljungstrom
ESP
Wet FGD
Yes
9,197
65
20

Kenosha, WI
j
No

Coal
Appalachian
13,100
2.60
9.10
                                      4B-90

-------
                                                                                       Table 2
                                                                                   SCR Process Design
CD
CD
CASE NUMBER
DESCRIPTION

SCR DESIGN BASIS
Boi ler Type
Economizer Outlet Temp. 3HCR, °F
Economizer Excess Air, %
Boiler NOx Emission Rate, Ib/MM Btu
NOx Concentration, ppmv (actual)
NOx Emission Limit, Ib/MM Btu
NOx Reduction Rate, %
NH3 Slip Rate, ppmvd (3 3% 02)
Guaranteed Catalyst Life, years
Reactor Configuration

Ammonia Storage, days
SCR DESIGN
Space Velocity, SCF*/ft3-hr
Linear Velocity, actual fps
Operating Temperature, °F
S02 Oxidation rate, %
Catalyst Geometry
Surface Area, m2/m3
Pitch, mm
Catalyst Layers (active + spare)
Soot Blowers
Ammonia Consumption, Ib/hr
Gas-Gas-Heater (GGH)
-Number
-Untreated Gas In/Out, °F
-Treated Gas In/Out, °F
SCR COST DEVELOPMENT
Catalyst Cost, $/ft3
Expected Catalyst Life, years
Ammonia Cost, $/ton
Natural Gas Cost, $/MM Btu
Plant Life, years
Capacity Factor, %
2.0
New,
hot-side

PC
725
24
0.40
364
0.08
80
5
2
Twin,
Vertical
7

2,750
18.2
725
1.10
Grid
470
7.07
4 + 1
Yes
941.4

NA
NA
NA

660
4
145
NA
30
65
2.1
New,
hot-side

PC
725
24
0.40
364
0.08
80
2
2
Twin,
Vertical
7

2,300
18.2
725
1.20
Grid
470
7.07
4 + 1
Yes
932.3

NA
NA
NA

660
4
145
NA
30
65
3.0
Retrofit,
hot -side

Cyclone
682
20
1.80
1700
0.36
80
5
2
Single,
Vertical
7

1,800
18.2
682
1.10
Grid
470
7.07
6 + 1
Yes
4,478

NA
NA
NA

660
4
145
NA
20
65
3.1
Retrofit,
hot-side

Cyclone
682
20
1.80
1700
0.36
80
2
2
Single,
Vertical
7

1,500
18.2
682
1.10
Grid
470
7.07
6 + 1
Yes
4,468

NA
NA
NA

660
4
145
NA
20
65
4.0
Retrofit,
hot -side

PC
725
24
0.60
572
0.12
80
5
2
Twin,
Vertical
7

2,530
18.2
725
1.20
Grid
470
7.07
4 + 1
Yes
1,383

NA
NA
NA

660
4
145
NA
20
65
4.1
Retrofit,
hot -side

PC
725
24
0.40
381
0.12
70
5
2
Twin,
Vertical
7

2,960
18.2
725
1.20
Grid
470
7.07
4 + 1
Yes
807

NA
NA
NA

660
4
145
NA
20
65
5.0
Retrofit,
cold-side

PC
NA
24
0.60
428
0.12
80
5
2
Twin,
Vertical
7

6,000
22.0
625
0.39
Grid
795
4.2
2 + 1
No
1,383

2 X 50%
129/550
625/226

660
4
145
2.98
20
65
               * SCF 3 32°F

-------
   ^BOILER
                                PLAN
  TQS, EL22Q'-Q'
                          ELEVATION
                                              rSS
Figure  1.   Case 2 Plan  and Elevation  General Arrangements.
                             4B-92

-------
                       BOILER ROOM
                        UNIT NO I
BOILER ROOM
 UNIT NO 2
                                            : SCR
                                            - REACTOR
                                           60'-O
                      /IAN
PHECIPITATOR NO 1

\ /
PRECIPITATOR NO 2
.
                                STACK
                                                 A
                                 ELEVATION LOOKING WEST
Figure 2.   Case  3 Plan and  Elevation General  Arrangements.
                              4B-93

-------
                   ELEVATION
Figure 3.  Case 4 Elevation General Arrangement.
    Figure 4.  Case 5 Plan General Arrangement,
                        4B-94

-------
03
CD
Oi
                 AIR HEATER
             Ammonium Blsulfate     •
             Fouling              ,
             Higher Exit Gas Temp.   ,
             Higher Leakage
             Higher AP
             Higher Steam Sootblow Rate
             Higher Water Wash Rate
             Higher Steam pressure &
             Superheat
             Additional  Dampers For
             On-Llne Wash
   BOILER
NPHR Increase
Temp. Bypass
Reduced KW
      FD FAN
• Higher Mass Flow      •
• Provide Dilution Air     •
• Higher Hp Consumption •
AMMONIA STORAGE   •
• Operator Training      •
  & safety
WATER TREATMENT     •
• Treat AH Wash For     •
  Nitrogen             .
      ESP
Higher Inlet Gas Volume
Higher Gas Temp.
SO3 NH3 Conditioning
Higher AP
Resistivity Affected
   FLY ASH
Marketability Impact
Odor Problems
Additional Equipment
For SCR
   ID FAN
Higher Mass Flow       •
Higher Volumetric Flow   •
Higher AP
   REHEAT
Higher Mass Flow
Increased Steam Usage
       FGD
Volume Increase
Higher Inlet Temp.
Increase In  H2O Evap.
SOj Concentration Dilution
FGD Wastewater Treatment
For NH3
    STACK
Increased Opacity
Higher SO3
                    FD FANS
ChD
                                   WATER TREATMENT
               TO EXISTING
            W & WM SYSTEM
                                                                                   ID FAN
                                                                                                          J	»  WASTE TO
                                                                                                               DEWATERING
                                      Figure  5.   Hot-side SCR  Design/Operations  Impact.

-------
                    PLANT
           • NPHR Increase
           • Reduced Kw
           • Natural Gas Supply
             Required
           • Additional Plant
             Complexity
 WATER TREATMENT
• Treat GGH Water Wash
  tor Nitrogen
  Compounds
      FGD
Mist Eliminator
Operation Critical
   AMMONIA STORAGE
• Operator Training & Storage
       STACK
• Higher SO3
• Increased Temperature
• Increased Opacity
• Increased Volume
CD
CD
                                                                                                                                       PILUTIQN
                                                                                                                                       AIR FAN
                   FD FAN
                                                                                                                       TO EXISTING
                                                                                                                     W & WM SYSTEM
                                                                                              WATER TREATMENT
                                          Figure  6.   Post-FGD  SCR Design/Operations Impact.

-------
160
    Total Capital Requirement, $/kW
10
   Mllls/kWh
       5.91
   5.32
                                             COST ITEMS
                                             KUl  CONTINGENCY
                                                  FACILITIES,ENG.,FEE
                                             HZ]  ID FAN.WiWM.REHEAT
                                             dl  AIR HEATER/GGH
                                             ^M  STRUCTURAL
                                             CU  DUCTWORK
                                             KS  NH3 STORAGE
                                             ••  REACTOR/CATALYST
     2.0 2.1    3.0 3.1     4.0 4.1
                    CASE
5.0
               Figure 7. Total Capital Requirement.
                  9.08
              8.17
                         6.54
                                     6.76
                              5.88
         COST ITEMS
         HZ] FIXED CHARGES
         •^ VARIABLE O&M
         •I FIXED O&M
   2.0 2.1     3.0  3.1     4.0 4.1     5.0
                    CASE
              Figure 8. Levelized  Cost (mills/kWh).
                               4B-97

-------
$/ton NOx Removed (Thousands)
                         4255
              1227
 2.0  2.1    3.0 3.1     4.0 4.1
                CASE
                                        COST ITEMS
                                        I   I FIXED CHARGES
                                        •Jj VARIABLE O&M
                                        •I FIXED O&M
       Figure 9. Levelized Cost ($/ton NOx  removed).
  Mills/kWh
u
8

6
4

0


n

o




it
B







6.54


Hi









i
1
^


1

7 68





««f



-0.9C






H

M
1
3













COST ITEMS
EFFICIENCY LOSS/GAIN
1 i W&WM, STEAM
CD NATURAL GAS
POWER
1 1 CATALYST
,™™




      CASE 4.0        CASE 5.0
    Figure 10. Hot-Side vs Post-FGD SCR Cost Comparison.
                          4B-98

-------
10
  $/ton NOx Removed (thousands)
                 23456
                      Catalyst Life, years
— Case 2
-*— Case 3
~*~ Case 4
-B- Case 5
        Figure 11. Levelized $/ton NOx versus Catalyst Life.
20
15
10
   Levelized Mills/kWh
                 23456
                      Catalyst Life, years
            Case 2
Case 3
Case 4
Case 5
        Figure 12. Levelized Mills/kWh versus Catalyst Life.
                            4B-99

-------
 Levelized Mills/kWh
300
400           500            600
     Catalyst Cost, $/ft3
                                                          700
           Case 2
         Case 3
Case 4
Case 5
      Figure 13. Levelized Mills/kWh versus Catalyst Cost.
   Total Capital Requirement, $/kW
 300
               400           500           600
                    Catalyst Cost, $/ft3
                                           700
      r
           Case 2
         Case 3
 Case 4
                                                 Case 5
    Figure 14. Total Capital Requirement versus Catalyst Cost.
                          4B-100

-------
  APPLICATION OF COMPOSITE NOX SCR CATALYSTS IN
                COMMERCIAL SYSTEMS

B.K. Speronello, J.M. Chen, M. Durilla, R.M. Heck

              Engelhard Corporation
              101 Wood Avenue South
                 Iselin,  NJ 08830

-------
Application Of Composite NOx SCR Catalysts In Commercial Systems

         B.K. Speronello, J.M. Chen, M. Durilla, R.M. Heck

                       Engelhard Corporation
                       101 Wood Avenue South
                         Iselin, NJ  08830


Abstract

Composite NOx SCR catalysts have been installed in a variety of
commercial NOx control systems, including coal fired power plants,
gas turbines, stationary engines, and chemical plants.  This paper
reviews how such catalysts performed in these systems, and it relates
key features of the composite catalyst design to catalyst
performance.  These data illustrate how composite SCR catalysts can
cut catalyst volume and reactor size by over 50% (relative to
conventional SCR catalysts), with no loss of NOx removal efficiency.
Background

The term composite honeycomb catalyst refers to a catalyst design
strategy where a layer of catalytic material is bonded to a strong,
thin-walled honeycomb support.  This design has been used for years
to treat exhaust from a variety of sources, including:  stationary
internal combustion engines1, gas turbines^, chemical
processes3, and, most notably, automobiles4-  In 1984, Engelhard
developed a composite catalyst for the selective catalytic reduction
(SCR) of NOx by ammonia^.  This catalyst contains a catalytic layer
of V205/Ti02 (V/Ti) supported on a cordierite ceramic honeycomb.
Recently, a composite zeolite SCR catalyst was developed to extend
the maximum operating temperature for the SCR reaction from nominally
450°C up to 600°C.

To date, composite catalysts have been demonstrated in 10 pilot tests
(see Table I) and 12 commercial installations (see Table II).  In
addition, Table II lists another 12 commercial installations in
varying stages of design and construction.  Overall, the treated
flows range from 3 Ib/sec for a small chemical process to 650 Ib/sec
for a 50 MW gas turbine.  This paper discusses some of the reasons
why the composite catalyst design was chosen for SCR, and summarizes
some of the experience that has been developed with composite SCR
catalyst.  In particular it focuses on factors pertinent to the
application of these catalyst to the exhaust from power generating
installations.
                               4B-103

-------
Discussion

Figure 1 illustrates the basic structure of a composite V/Ti SCR
catalyst, and compares it with a conventional SCR catalyst made by_
extruding catalytic material into the honeycomb shape.  The composite
design is appropriate whenever mass transfer factors, such as
boundary layer diffusion or pore diffusion, limit the penetration of
reactant gases to a thin layer at the catalytic surface.  This is the
case with NOx SCR.

The benefits of the composite design include:

     1.    High strength due to the strength of the underlying
          structural ceramic;
     2.    Thin walls made possible by the support's high
          strength;
     3.    High geometric surface area at constant pressure drop
          due to the thin walls;
     4.    Excellent abrasion resistance due to the hardness of
          the ceramic support;
     5 .    High activity due to high geometric surface area and
          greater porosity within the catalytic layer;
     6.    Wider temperature range of operation due to better NOx
          mass transfer characteristics;
     7.    Inherently low SC>2 oxidation activity, and
     8.    Contains  85% less heavy metals.


 Composite SCR catalysts can be made with walls as thin as 12
 thousandths  of an  inch (0.30 mm) compared to ca. 27 thousandths  (0.65
 mm)  for  conventional extruded SCR catalysts.  Because they have
 thinner  walls, composite SCR catalysts can have much smaller openings
 (and consequently  higher geometric surface area)  than conventional
 catalysts.   Figure  2 illustrates the benefits of thinner walls and
 increased geometric surface area for increasing catalyst activity and
 reducing catalyst  volume and reactor size.  Figure 2 compares the the
 amount  of SCR catalyst that would be needed to achieve either 80% or
 90%  NOx  reduction  using catalysts of different cell densities (ie.
 channel  opening  sizes measured in cells per sguare inch, CPSI).  For
 example, in  a design to achieve 80% NOx reduction at a pressure drop
 below 3  inches water column, a conventional 40 CPSI catalyst would
 require  a relative  catalyst volume of 3.3.  In comparison, a 200 CPSI
 composite catalyst  provides the same performance with 65% less
 catalyst (a  relative catalyst volume of only 1.2).

 Figure  3 illustrates another benefit of smaller channel openings,
 wider range  of operating temperature.   It compares the curve of NOx
 conversion  vs.   temperature for a 200 CPSI composite SCR catalyst
 with a  catalyst  having a more conventional cell density, 25 CPSI.
 The  200  CPSI composite catalyst was tested at much higher flow rate
 than the 25  CPSI  catalyst  so that both provided similar NOx
 conversion  at temperatures below 350°C  (660°F).  As temperature was
 increased,  both  catalysts  exhibit the conversion peak and subsequent
 drop off that is  characteristic of the  onset of excessive ammonia
 oxidation in V/Ti  SCR catalysts.  However, the peak for the composite
 catalyst is  over  5  conversion points and  40°C higher than the 25 CPSI
                                4B-104

-------
catalyst.  In addition, at 450°C (840°F), the normal maximum
continuous operating temperature for V/Ti, the 200 CPSI catalyst
still provided over 85% NOx conversion, while the 25 CPSI catalyst
had fallen to only 70%.

Figure 4 explains why composite catalysts demonstrate inherently low
SC>2 oxidation activity.  It is a graph of SC>2 conversion vs.
relative catalyst loading (in gram catalyst/in^ of catalyst, relative
to a base loading).   A medium cell density composite SCR catalyst has
a relative catalyst loading of IX,  while a conventional extruded SCR
catalyst has a typical loading of 10X on this scale.  Because of the
low S02 oxidation activity of composite SCR catalysts, this test was
run under exceptionally severe conditions to increase oxidation
sufficiently to allow for precise measurement.   SC>2 oxidation in
commercial operation would be significantly lower than shown in
Figure 4.

At conversion levels of < ca. 20%,  the rate of SC>2 conversion is
controlled by the number and potency of the active sites for S02
oxidation.  Under this rate limiting condition the extent of 502
oxidation to SO^ increases in proportion to the amount of catalytic
material in the catalyst.  As a result, composite catalysts have
inherently low S02 oxidation activity, because they contain only
about 10% as much V/Ti as extruded SCR catalysts.  To compensate for
this deficiency, the SC>2 oxidation activity of extruded catalysts is
suppressed by incorporating SC>2 oxidation demoters into the catalyst
formulation.  These additives, however, are also reported to suppress
NOx removal activity^.

The relative hardness of the ceramic honeycomb support results in
excellent erosion resistance for composite catalysts in high dust
environments.  This characteristic, and catalytic performance in
several pilot and commercial installations are discussed in the
following sections.


Coal Fired Power Plants

High Dust, Hot Side:

The first experience with composite SCR catalyst for NOx control in a
high dust, hot side coal fired power plant environment began in 1988.
Small portions of the conventional SCR catalyst beds of two operating
coal fired power plants in Germany were replaced with composite
catalyst.  One was on a dry bottom boiler, and the other a wet bottom
boiler.  Pertinent operating characteristics are included in Table
III.  Of particular interest were the dust loadings, which ranged up
to 15 g/Nm3 (ca. 7 grain/DSCF).  Each SCR system contained 2 layers
of catalyst; each 1 meter deep.  Several sleeves of composite SCR
catalyst (25 CPSI, 5.1 mm cell pitch) were mounted in each catalyst
layer.

A 6 inch long block of catalyst was mounted at the inlet of each
layer to test for possible erosion.  These inlet blocks were removed
after 18 months and returned to Engelhard for erosion analysis.  The
remaining catalyst is continuing in operation.
                               4B-105

-------
Two types of erosion were evaluated; axial (ie. length) erosion of
the catalyst block and wall erosion (ie.  thinning of the catalytic
layer).  Neither type was detected.

Axially, there was no change in the length of the catalyst blocks,
nor was there evidence of honeycomb wear to the sharp edges at the
inlet face of the catalyst.  The hard ceramic support completely
resisted abrasion in this environment.

Wall erosion was measured by electron microscopy.  Three samples were
cut from each aged catalyst block plus a fresh control.  One sample
was taken from the center of the block, and the other two were taken
from spots located ca. 1 inch in from each end.  These were mounted,
polished to reveal a cross section of each wall, and 5
photomicrographs were taken of each sample.  These showed that there
was no evidence of erosion of the catalytic layer.  There was no
difference in wall thickness between the aged and fresh catalysts,
and no correlation between depth of the catalytic layer and sample
location within a block.  Figure 5 shows representative
photomicrographs comparing aged catalyst with the fresh control.  The
lighter layer near the center of each photo is the catalytic V/Ti,
and the darker material beneath it is the ceramic support.  The depth
of the catalytic layer was unaffected by aging.

These  results are consistent with prior observations showing that
flow straighteners and hardening of the catalyst inlet face could
control erosion of the SCR catalyst, and there was no report of wall
thinning away from the inlet face7.

Analysis at Engelhard indicates that catalyst erosion is limited to a
transition zone from turbulent to laminar flow at the inlet to the
honeycomb channel.  The depth of this transition zone can be affected
by  factors such as gas velocity and angle of incidence of the gas
with respect to the catalyst channel, but once the flue gas develops
into full laminar flow  (always within millimeters of the inlet face
of  the catalyst) there is  little interaction between abrasive
particles and the catalyst wall.  Conseguently there is negligible
wall thinning within the block.  With composite catalysts, the hard
ceramic  substrate resists  erosion at the inlet, so, in addition to no
wall thinning, there is also no axial erosion.

While  the main purpose of  this experiment was to study catalyst
erosion  in the high dust environment, the SCR performance of these
aged inlet samples were also measured.  The results are shown in
Figure 6.  It is a graph of NOx conversion as a function of tem-
perature for aged catalyst taken from the SCR beds of both power
plants.  Considering that  the samples represent the most contaminated
and deactivated portion of the bed, the inlet 6 inches, it is notable
that both provide over  80% NOx reduction and little or no ammonia
slip at  a typical commercial flow rate.


Medium Dust, Cold Side:

Composite SCR catalysts have been, and are continuing in pilot plant
tests  on several coal fired power plants downstream of the flue gas
desulfurization systems.   Figure 7 shows the results for a pilot
                                4B-106

-------
reactor system that used a 200 CPSI V/Ti composite catalyst to
determine if catalyst with a pitch of less than 2 mm could operate in
that environment without plugging.  Table IV summarizes pertinent
operating conditions.  The test included a total of 2500 on-stream
hours (including 40 pilot plant start-ups and shutdowns)8.  At
20,000 1/hr VHSV and 350°C, NOx conversion averaged 92% with <5 ppm
ammonia slip.  For this application, the particulate concentration
(primarily gypsum, CaCC>3 and SiC>2)  averaged about 50 mg/Nm3 (0.02
grains/DSCF)  with peaks of up to about 120 mg/Nm3 (0.05
grains/DSCF).  The inlet NOx concentration varied between 350 and 420
ppm.  No apparent decline of catalyst performance was seen throughout
the test period.

Since the channel size of the 200 CPSI catalyst was significantly
smaller than typically used for this dust level, provision was made
for soot blowers to prevent catalyst plugging.  Initially the soot
blowing frequency was every 2 hours.  Over the first 800 hours of
operation the soot blowing frequency was steadily reduced with no
increase in pressure drop across the catalyst.  After 800 hours, soot
blowing was stopped completely.  Pressure drop rose slightly and
stabilized at a 15% increase from the original level.  While every
installation is different, these results indicate that it is possible
to operate high cell density SCR catalysts at moderate dust levels
without plugging.

Figure 8 shows on-line NOx reduction efficiency and ammonia slip
versus feed NH3/NOx ratio during this pilot test.  It shows that by
using a 200 CPSI catalyst it was possible to achieve >92% NOx
conversion with <5 ppm NH3 slip at with relatively little catalyst
(20,000 1/hr VHSV).

Aging of these catalyst modules plus several others containing lower
cell density catalyst was continued in the SCR reactor downstream of
the FGD unit of another coal fired boiler.  Dust levels averaged 10
mg/Nm3 (0.005 grains/DSCF), and inlet NOx ranged between 550 and 600
ppm.  To date the catalysts have accumulated a total of 15,800 hours
of operation.  Laboratory activity tests made at 60,000 1/hr VHSV on
core samples taken at 0, 2500 and 7540 hours showed no decline of
catalyst activity from the fresh level (see Figure 9).  Testing after
nearly 16,000 hours shows what may be a slight decline in conversion,
but even these results are within the range of normal test
variability.

These data demonstrate that this composite SCR catalyst with a high
cell density can be applied successfully to boiler exhaust even where
the presence of particulates is significant.  This can represent a
several fold reduction in catalyst volume for achieving the same high
NOx performance efficiency as conventional catalyst designs of lower
cell density.


Stationary Engines

Composite V/Ti based SCR catalyst technology has also been applied to
a wide range of stationary engine applications.  Engine sizes have
ranged from 210 to 3900 horsepower.  Fuels have been natural gas,
digester gas, and #2 diesel fuel.  Ammonia control strategies have
                                4B-107

-------
ranged from manual adjustment to continuous monitoring with control
by an in-plant host computer.  In each application, the required NOx
conversion and NH3 slip performance limits were achieved.

Of particular interest are the results from one application which
utilized #2 diesel fuel.  This application was operated on 300 CPSI
(1.5 mm pitch) composite catalyst as a demonstration unit by the EPA
at their Air and Energy Research Laboratory at Research Triangle Park
in North Carolina9.  The demonstration was run for 4000 hours in
approximately 100 hour increments.  Particulate levels in the exhaust
ranged from 27 mg/Nm3 (0.013 grains/DSCF)  during steady state
operation to up to 100 mg/Nm3 (0.05 grains/DSCF) during each start-
up.  There was occasional evidence of increased pressure drop due to
accumulation of wet soot on the ultra-fine pitch catalyst during
periods of frequent repeated cold starts.  As a result, the catalyst
was manually air lanced four times during the demonstration.  After
each cleaning, 85% NOx conversion was maintained.  There was no
evidence of particulate buildup during periods of continuous
operation.

This test demonstrated that ultra-fine pitch composite SCR catalyst
could provide continuous steady state NOx emission control on a
diesel engine operating on #2 fuel at a substantial reduction in
catalyst volume relative to conventional catalysts.
 Conclusions

 The  composite catalyst design strategy using ceramic supports offers
 several unique advantages for NOx selective catalytic reduction.
 They include greater mechanical strength, exceptionally high
 activity,  excellent erosion resistance, and inherently low S02
 oxidation  activity.  These benefits have been demonstrated in both
 extensive  pilot scale testing and commercial installations.


                          ACKNOWLEDGEMENT

 The  information included in this paper was generated in collaboration
 with several Engelhard colleagues, including:  Dr.  J.W.  Byrne, Mrs.
 C.Hirt, Mr. J.Hansell, and Mr. M.Tiller.
                                4B-108

-------
K.Burns, M.Collins, R.M.Heck, Catalytic Control Of NOx
Emissions From Stationary Rich-Burning Natural Gas Engines,
ASME 83-DGP-12, 1983

J.M.Chen, R.M.Heck, K.R.Burns, M.F.Collins, Development Of
Oxidation Catalyst For Gas Turbine Cogeneration Applications,
82nd Annual Meeting And Exposition Of The Air Pollution
Control Association, June 1989, Anaheim, California

R.M.Heck, M.Durilla, A.G.Bouney, J.M.Chen, Ten Years
Operating Experience With Commercial Catalyst Regeneration,
81st Annual Meeting And Exposition Of The Air Pollution
Control Association, June 1988, Dallas, Texas

J.J.Mooney, C.E.Thompson, J.C.Dettling, Three Way Conversion
Catalyst - Part Of A New Emission Control System,  SAE 77-
0365, 1977

R.M.Heck, J.C.Bonacci, J.M.Chen, Catalytic Air Pollution
Controls - Commercial Development Of A Catalyst For The
Selective Catalytic Reduction Of NOx, 80tn Annual Meeting And
Exposition Of The Air Pollution Control Association, June
1987, New York, New York

J.Ando,  "NOx Abatement For Stationary Sources In Japan", EPA
600/7-83-027(1983)

L.Balling, D. Hein, DeNOx Catalytic Converters For Various
Types Of Furnaces And Fuels - Development, Testing,
Operation, 1989 EPA/EPRI Joint Symposium On Stationary
Combustion NOx Control, March 6-9, 1989, San Francisco,
California

D.Dreschler, presented at the German VGB Conference, Feb.
1986

J.H.Wasser, R.B.Perry, Diesel Engine NOx Control With SCR,
EPA/EPRI 1985 Joint Symposium On Stationary NOx Control, May
1985, Boston, Mass.
                         4B-109

-------
Figure 1:
  Two Different  NOx SCR  Catalyst Designs
     Composite
                           Extruded Vanadia/Titania
       Strong,  -
       Thin Walled
       Ceramic
       Support

  Catalytic
  Layer Of
  Vanadia/Titania

 Figure 2:
        Catalyst Volume Can Be Cut 50 - 65% By
            Increasing Geometric Surface Area

  Pressure Drop, Inches Water Column
                    100 CPSI
        200 CPSI
                           ;40 CPSI
                                            111 "CPSI j
  o-—     --  -   -  --  -  	
   1.0                        10.0
                     Relative Catalyst Volume
350 deg. C, 20 ft/sec Velocity (@ T)
                                  NOx SCR Efficiency

                                      90% Conversion

                                      80% Conversion

                                    -- — ._ ._!_  J... . J_ I
                                                      100.0
                         4B-110

-------
CD
              Figure 3:
                Increasing SCR Catalyst Geometric Area
              Widens  The Operating Temperature Window
100
90

80

70

60
50
40
2£
NOx Conversion, %
D
D — 	 	
\ \
\ \
\
\
\
\
\
\
\
\
D 200CPSI, 60.000VHSV
+ 25CPSI, 12.000VHSV
; i









50 300 350 400 450 500 550
                              Temperature, deg C
    Figure 4:
        Composite SCR Catalysts Exhibit
      Inherently Low SO2 Oxidation Activity
                                                                  10
                                                                    SO2 Conversion,
                                                                   o
                                                                                   Extruded Catalyst
                                                                                     Equivalent To
                                                                                       5X - 15X
            1/1 NH3/IMOX Ratio
   O.OX     0.5X     1.0X     1.5X     2.OX     2.5X
        Relative Catalyst Loading, (g/in3)/base

100 CPSI, 350 deg. C, 450 ppm SO2

-------
                                  figure1 5
                                        Composite SCR Catalyst Aged In High Dust, Coal Fired Boiler
                                        Exhaust Shows No Evidence Of Wall Erosion
CD
r\j
                                                         Fresh Catalyst
              Catalyst Aged In Dry Bottom Boiler Exhaust
Catalyst Aged In Wet Bottom Boiler Exhaust

-------
  Figure 6:
          Inlet 6" Of Composite SCR Catalysts From
             High Dust SCR Beds Of Coal Boilers
      NOx Conversion, %
   100 i	7	
   80
    60
    40
    20
                          Ammonia Slip, ppmv
                        	100
        —" Dry Bottom Boiler

        -**- Wet Bottom Boiler
                                           80
                                         - 60
                                           40
                                         -  20

     300
325       350       375       400

     Temperature, degrees C
                                                          o
                                                        425
3000 hr-1 VHSV, 1/1 NH3/NO*
      Figure 7:
          200 CPSI Composite SCR Catalyst Did Not
             Plug In A Moderate Dust Environment
   100
    80
    60
    40
    20
     0L
      NOx Conversion, %
                    Change In Pressure Drop, %
                    	 	 -100
                 24
           No Soot Blowing
         Soot Blowing
       Frequency, hours
                                                          80
                                                          60
                                                          40
                                                          20
      0    400    800    1,200  1,600   2,000   2,400   2,800

                       On-Stream Time, hours

50-120n mg/Nm3 dust
                             4B-113

-------
      Figure 8:
      NOx Conversion And Ammonia Slip Over
      Composite SCR Catalyst After Coal FGD
                                             Figure 9:
                                                Composite SCR Catalyst Unaffected By
                                                 2  Years In Coal Fired Utility  Exhaust
CD

4^
      NOx Conversion, %
   100	
    95
    90
    85 -
                       NH3 Slip, ppm
    0.70
0.80     0.90    1.00     1.10
     Inlet NHS/NOx Ratio
425 ppm NOx, 20,000 1/hr VHSV,
200 CPSI, 50-120 mg/Nm3 Dust
                                                          100
                                                                             NOx Conversion, %
NH3 Slip, ppm
                                                          80 F
                                              60 |_  Catalyst Condition
                                                 I I  n  Aged 2500 Hrs
                                                    -I-  Aged 7450 Hrs
                                                    ''•>  Aged 15800 Hrs
                                                       Fresh Catalyst
                                                          40
                                                                         1.0     1.1     1.2
                                                                          NHS/NOx Ratio
                                           400 ppm NOx, 320 deg C
                                           60,000 1/hr VHSV, 200 CPSI
                                                                                                             1.3
                                                                                                      100
                                                                                        -80
                                                                                                     60
                                                                                                     40
                                                                                                    J20
           1.4

-------
Table I:
   Engelhard Composite SCR  Catalyst Pilot  Tests
    Location

    United  States
    Germany
    Germany
    Germany
    Germany
    Germany
    United  States
    United  States
    United  States
    United  States
Type

Diesel  Engine
Cold Side,  Coal  Fired  Heating Plant
Cold Side,  Coal  Fired  Power Plant
Cold Side,  Coal  Fired  Power Plant
Hot Side,  Coal,  Dry  Bottom Boiler
Hot Side,  Coal,  Wet Bottom  Boiler
Gas  Turbine
Natural Gas Boiler
Gas  Turbine  -  Simple  Cycle (980  F)
Gas  Turbine  -  Simple  Cycle (1085  F
Table II:
  Engelhard's  25 Composite SCR Catalyst Systems

Location
California
4 Natural
Alabama
California
California

California
California
California
California
Texas
New Jersey
California
New Jersey

Application -Catalyst*
6 Engines-VNX
Gas and 2 Digester Gas
Chemical Process-ZNX
Industrial Boiler-VNX
Refinery Heater-VNX
Refinery Heater-ZNX/VNX
50 MW Gas Turbine-VNX
1 MW Gas Turbine-ZNX
Annealing Furnace-VNX
5-Refinery Heaters-VNX
Chemical Plant-VNX
3-Dual Fuel Engine-ZNX
Refinery Heater-VNX
2-50MW Gas Turbines-VNX
Approx.
Startup
1984

8/90
10/90
10/90
10/90
11/90
2/91

3/91
3/91
5/91
6/91
10/92

Flow,#/sec
200 to
4000 HP
3
17
84
46
650
23
29
16-31
99
5
39
844
*VNX(tm) & ZNX(tm) are V/Ti and Zeolite SCR Cat.
                         4B-115

-------
Table
  Operating  Conditions Of Hot Side SCR Systems
  Boiler Type
  Coal Type
  Ash Content, %
  Sulfur Content, %
Dry Bottom
Wet Bottom
German/Foreign  Ruhr/Saarland
6 - 12
< 1.5
  Temperature, deg. C  360

  Inlet NOx, mg/Nm3   800
  Inlet SO2, ppmv     1000
  Particulates, g/Nm3  10 - 15
  Soot Blowing
Monthly
4-7
<  1.5

360

1300 - 1500
1000
2

Every 10 Days
Table IV:
      Moderate Dust SCR Pilot Plant Conditions
      Location

      Flow Rate, scfm
      Temperature, deg. C
      Space Velocity, hr-1

      Inlet NOx, ppmv
      Inlet SO2, ppmv
      Particulates, mg/Nm3

      Coal Type
      Sulfur Content, %
      Ash Content, %
       Slipstream Off FGD

       2060
       350
       20,000

       350 - 420
       50
       50 - 120

       Bituminous
       0.95 - 1.18
       5.0  - 6.5
                       4B-116

-------
 SCR  CATALYST  DEVELOPMENTS  FOR THE  U.S. MARKET

by    T. R. Gouker
      Research Division
      W. R. Grace & Co.-Conn.
      7379 Route 32
      Columbia, MD  21044

      C. P- Brundrett
      Davison Chemical Division
      W. R. Grace & Co.-Conn.
      10 E. Baltimore Street
      Baltimore, MD  21202

-------
                    SCR  CATALYST  DEVELOPMENTS  FOR  THE  U.S.  MARKET
                                      ABSTRACT

      This paper reviews SCR catalyst development from its invention in the U.S.
through power plant applications of the technology in Japan and West Germany.
Building on this experience, the requirements for adaption of the SCR process  to
U.S.  high-sulfur coal  applications are discussed.  Grace's SCR catalyst development,
SYNOX, is then reviewed for its application to U.S.  boilers firing high sulfur coal.


JAPANESE SCR DEVELOPMENT

      SCR was originally invented and patented by a U.S. company  in 19591,  but its
use was limited to a few industrial  applications, such as pollution control from
nitric acid plants.  It wasn't until the 1970's  that SCR gained application to power
plant NOX emissions.   The first utility applications took place in Japan.   The
Japanese identified SCR as  a suitable approach for controlling NOX and began a
stepwise application of the technology to all three types of fossil-fueled boilers:
gas,  oil and coal.  By  1985, there were more than 200 commercial  installations
operating in Japan2.

      Application of SCR to power plant exhaust  was not a simple matter.  A number
of problems were encountered during  development.  These included:  (1) catalyst
poisoning by sulfur species in the fluegas;  (2)  ammonium bisulfate deposition in the
catalyst and on downstream  equipment; and (3) equipment corrosion due to increased
S03 levels in the flue gas.  The single most important contribution the Japanese
made  to the development of  SCR was to switch from noble metals to base metal oxides
for the catalyst3.   The use of titanium dioxide supports with mixtures of vanadium
and tungsten oxides as catalysts, solved the major problems associated with oil and
gas-fired utility fluegas applications.

      Additional developments were required, however, to address the  issues of
flyash plugging and erosion for coal-fired service.  In 1978, pellet  catalysts were
given-up in favor of parallel-flow honeycomb or  plate catalysts.  The low conversion
targets (40-60%) coupled with the relatively low fly ash content of Japanese boilers
and the fact that most units were relatively new, allowed SCR application go
smoothly.  Continued development led to ceramic  honeycomb and plate-type catalyst
configurations which provided high geometric surface area with low tendency for
flyash plugging.  In the early 1980's the focus  of Japanese work on SCR was on
optimizing surface geometry and avoiding flyash  plugging.  This optimization reduced
the size (and cost) of SCR  reactors  by a factor  of 2 and greatly  improved the
economics of the SCR process.  At this stage of  development, ceramic  honeycomb-
based catalysts had a pitch of 13 (wall thickness of 2 mm and channel opening of  11
mm) in 1978.  By 1982, catalysts with a pitch of 7.5 (wall thickness  of 1.4 mm and
channel opening of 6.1 mm)  had been  demonstrated.  Since the mid-1980's, catalyst
research in Japan has principally focused on understanding the deactivation
mechanisms of SCR catalysts*.   Mechanisms have been proposed for for the effects of
alkalis, alkaline earths, and heavy  metals such  as arsenic.  Research on SCR
continues in Japan, but due to the more demanding commercial environment, the lead
role in SCR technology development shifted to West Germany in the mid-1980's.

WEST GERMAN SCR DEVELOPMENT

      Driven by tough regulatory standards, West German utilities made rapid


                                        4B-119

-------
progress, lowering their coal-fired power plant NOX emissions to below 0.16
Ibs/million Btu5,  about half that of the Japanese national standard.  To do this,
the SCR process could not run at the 40-60% conversion  levels as  in Japan,  but
instead systems were designed for up to 90% NOx removal.

      The first SCR facility to be retrofited  in West Germany went  into  operation  in
1985.  By the end of 1990, more than 23 GWe of power plant capacity had  SCR
installed to control NOX emissions5  (Figure  1).  In support of this intensive
emission control program, about 60 pilot plants were constructed  and  operated.   The
pilot plants were required to address the design issues which surfaced  in  the more
demanding environments of West German boilers.  In addition  to  dry  bottom  units, as
in Japan, SCR was retrofitted on slagtap boilers.  Since  few German utilities has
low  NOX burners, NOX concentrations  were much  higher  than  in  Japan.   Coal sulfur and
ash  contents were also higher than in Japan, leading to higher  SOX and flyash
concentrations  in the  fluegas.  The SCR pilot  program resolved  several  issues
involving SCR process  technology.  The  effect  of erosion, poisoning and  fouling  on
catalyst lifetime were quantified.  The potential  for ammonia slip was  also
determined,  along with the resulting effect of ammonium bisulfate deposition.

       In general, the  process experience with  dry  ash boilers in West Germany has
been similar to the  results  experienced by coal-fired utilities in Japan.   The
majority of the loss  in  activity was due to interactions  of  the catalyst with the
flyash.   The flyash  had  several types of effects on the monoliths:  physical
 fouling, poison transfer, and bulk plugging.   Sub-micron  ash particles accumulated
on the  surfaces of  the elements and blocked the pores of  the catalyst.  This
 physical fouling  prevents the NOX and ammonia  from reaching the active sites of the
 catalyst and leads  to  a  reduction in catalytic performance.  Figures 2 and  3 compare
 the  ESCA analysis of fresh  and aged samples typical of dry ash  boilers6.  This data
demonstrates the  buildup of  the wide variety of flyash elements near the surface.
The  reduction  in  the intensity of the titanium signal further demonstrates  the
 covering of the surface  by  flyash.

       Flyash was  also  found  to contribute poisons  to the  catalyst surface.  The
 transfer of alkali  metals from the flyash lowered  the activity  of the catalyst.
 This was found  to occur  primarily due to the leaching action of moisture on the
 flyash  during  start-up and  shutdown of  the units7.   Since these metal  salts are
 soluble in  water, the  presence of moisture can promote the redistribution  of these
 elements.   In  laboratory studies, alkali salts have been  shown  to be a catalyst
 poison  due  to  the formation  of inactive complexes  with the vanadium and  tungsten.

       In some  cases,  bulk plugging of the channels by accumulations of dust was
 observed.   Dust plugging occurred when  flyash  "flaked" off of upstream equipment.
Wire screens were installed  on most units to break up these  flakes as they
 approached  the  catalyst.  Soot blowers  were also installed to blow the flakes back
 and  break them  up during pilot plant start-up  and  shutdown.

       In addition to  occasional plugging problems, severe erosion problems due to
fluegas  flow maldistributions have occurred.   The  use of  flow straightening vanes
and/or  "dummy"  catalyst  layers has reduced gas flow distribution  problems  and
associated  erosion.

      While the experience  in installing SCR on dry ash boilers was generally
similar  in  both Japan  and West Germany, there  has  been  a  significant  shift in
expectation of  average catalyst life between SCR on the two  continents.   Perhaps
because  of  improvements  in catalyst technology or  improvements  in process
installations,  SCR  catalysts in West Germany are experiencing better  activity
maintenance than the Japanese high-dust, coal-fired SCR catalysts experience
                                        4B-120

-------
suggested.  Figure 4 shows a graph of relative catalyst life versus operating time5.
This figure illustrates a 5-10% higher level of activity per year for a number of
West German installations when compared to the curve based on Japanese experience.
This has turned out to be a surprising development for the industry in that most
organizations expected SCR to experience higher rates of activity loss in West
Germany.  In planning for potential activity problems, a number of new catalyst
formulations were investigated for the West German market but nor commercialized.
These developments put the United States in a position to reap the benefits of
recent SCR catalyst advances as the technology is introduced on U.S. boilers.

U. S. SCR ISSUES

      As mentioned earlier, SCR technology was originally invented in the United
States.  While extensively installed abroad, U.S. SCR application has expanded from
its beginnings on nitric acid plants.  SCR is in service on an increasing number of
gas and oil-fired boilers, turbines, and industrial applications such as refinery
process heaters.  Grace has been offering SCR systems into this growing market under
the trade name Camet  for a number of years.  Systems have been designed to provide
up to 90% NOX removal.   Due to the low sulfur and low fly ash concentrations,
catalyst pitch for these applications can be greatly reduced.  With conventional
ceramic plate type catalyst, pitch can only be reduced to 3.0 mm.  Camet catalysts
are supported on a thin metal substrate (2.5 mils) and can be manufactured with
openings equivalent to a 0.2 pitch catalyst, greatly reducing the volume of an SCR
reactor.  An example of a Camet SCR system is shown in Figure 5.  The system is part
of the Santa Maria Cogeneration Project in California.  It is owned and operated by
Bonneville Pacific Corporation of Salt Lake City, Utah.

      Except for pilot testing by the EPA and EPRI8'9, SCR has not been implemented
on coal-fired boilers in the U.S.  The boiler types in the U.S. are similar to those
in West Germany, therefore a great deal of information can be transferred from the
German experience.  Several issues must be addressed, however, before SCR will gain
wide spread acceptance in coal service in the United States.  An assessment must be
made of the impact of higher sulfur content of Eastern Bituminous coals on SCR life
and performance.  The effects of differing fly ash constituents must be identified
and quantified.  In addition, the possibility exists for some as yet unrecognized
boiler conditions or coal characteristics to impact SCR performance.

      For SCR to be installed on U.S. facilities, especially those burning high-
sulfur U. S. coal, risks must be reduced to an acceptable level.  To do this,
engineering companies and catalyst vendors will have to develop information on a
number of key areas including:

      1.    The proper space velocity, linear velocity and reaction
            temperature to minimize ammonia slip at required NOX removal
            levels.

      2.    The tolerable level of ammonia slip under high S02 and S03
            conditions.

      3.    The performance of catalyst and its deactivation rate in
            flue gas and fly ash from U. S. coals.

      4.    Performance of the air preheater when exposed to high S03
            levels and subsequently high levels of NHi,S04.

      5.    Adhesion characteristics of fly ash from U. S. coals to SCR
            catalyst.  The effectiveness of soot blowing and the effects
                                       4B-121

-------
            of residual fly ash on catalyst activity.

      The most efficient way for catalyst vendors to assist  in the development  and
demonstration of SCR technology on high-sulfur U. S. coals  is through  the design  of
catalysts tailored to meet the above needs in a cost-efficient manner.   High  levels
of denox activity while minimizing ammonia slip can be met  by catalysts  designed  to
have higher NOX conversion activity.   The effects of NHAS04  deposition  can  be
minimized by designing catalysts to have a low S02 to S03 oxidation activity.  The
detrimental effects of fly ash and deactivation can be counteracted  by catalysts
designed to be more durable, and poison resistant.

U.  S.  SCR DEVELOPMENT

       The design  requirements  and the  recognition that a major portion of  SCR
process  costs  are still  associated with the catalyst has led researchers to
vigorously  pursue improved catalyst designs.  Grace's approach has been  to  focus  on
 increasing  catalyst  activity and life, thus reducing catalyst volume requirements
while  extending  the  useful life and consequently reducing SCR process  costs.   To
determine  the  potential  for major improvements, Grace undertook  a  fundamental  study
 of the limits  of SCR catalyst  performance, developing a mathematical model  that
 expressed  catalyst performance (NOX conversion,  ammonia conversion, S02 oxidation)
 as a function  of the properties of the catalyst.

       The  model  accounts for the key design parameters of the catalyst,  including
 its composition,  monolith channel shape and dimensions, monolith wall  thickness,
 pore structure,  overall  volume, and aspect ratio.  The model has been  found to
 correlate  well  with  Grace's commercial data base, adjusting only surface-kinetic
 rate parameters.   Modeling results indicated  that a substantial  reconfiguration of
 the pore structure of the catalyst could  increase NOX conversion by about 50%, while
 simultaneously increasing resistance to poisoning and thereby extending  catalyst
 life.   The  model  indicated that the NOX conversion improvement would be  selective
 with respect  to  S02  conversion.  That  is, the undesired S02  oxidation reaction would
 not be enhanced  by the pore structure  reconfiguration.  This is  due to the  fact that
 the reduction  of NO  is diffusion limited whereas the rate of oxidation of S02 is
 kinetically controlled.

       The  model  showed that an optimum balance between surface catalytic activity
 and diffusivity  could be provided by a bimodal pore structure with a substantial
 percentage  of macropores (Figure 6)  .   However,  such  pore  structures could not be
 attained using titania as the  catalyst support.

       To solve this  problem, Grace succeeded  in engineering  silica to  provide the
 necessary  macropores for a new catalyst pore  structure while maintaining the
 necessary  intrinsic  denox activity.  Grace researchers developed a preparative  route
 to deposit  titania within the  silica to produce a novel catalyst support (Figure  7).
When extruded  and impregnated  with vanadia, the new catalyst, trade  named  SYNOX™,
 resulted in the  anticipated 50% improvement in activity11.   We  have demonstrated
excellent  hydrothermal  stability through near 3000 hours in  simulated  fluegas with
no noticeable  decrease in activity (Figure 8) in testing of laboratory-scale pieces.
 Proposals  have been  made to expose these new  monoliths to power  plant  stack gas side
 streams.   Such  pilot plant tests have  been planned by the Electric  Power Research
 Institute  and  Southern Company Services  (for  the U.S. Department of  Energy).
 Independent of the pilot plant tests,  Grace has initiated a slipstream test of its
own.

       A  slipstream reactor has been set up at the TVA Shawnee Steam  Plant  in
Paducah, KY,  to  study  the SYNOX catalyst  in fluegas from a  100 MU  boiler burning
                                       4B-122

-------
high-sulfur coal.  The effects of flyash,  alkali salts, and other constituents will
be studied and compared to their effects on a conventional titania catalyst.  The
boiler effluent contains about 300 ppm NOX,  3000 ppm S02,  and a flyash  loading of
about 3 grains per standard cubic foot,  the Grace slipstream reactor is located
downstream from the mechanical separators  and air preheater,  as shown in Figure  9.
The fluegas temperature entering the unit  is about 150°C and  must be reheated to
about 350°C before passing over the catalyst.

      The reactor contains 3 catalyst baskets stacked one on  top  of the other and a
dummy basket at the top.  The dummy basket contains inactive  catalyst support and is
used as a flow straightener.  Each catalyst basket contains 16  pieces of 2.5 cm  x
2.5 cm x 23