Proceedings1989 Joint Symposium on Stationary Combustion Nox Control San Francisco CA March 6-9 1989 Volume 2


iiir'/s- buu / y- yy-ubiib
June 1989

PROCEEDINGS;

1989 JOINT SYMPOSIUM ON STATIONARY COMBUSTION NOx CONTROL
San Francisco, CA, March 6-9, 1989
Volume 2

Compiled by

Claudia Runge
Electric Power Research Institute
3412 Hillview Avenue
Palo Alto, CA 93404

EPA Project Officer: ¦	EPRI Project Manager:

William P. Linak	David Eskinazi

U.S. Environmental Protection Agency	Electric Power Research Institute

Air and Energy Engineering Research Laboratory	3412 Hillview Avenue

Research Triangle Park, NC 27711	Palo Alto, CA 93404

AIR AND ENERGY ENGINEERING RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NC 27711 •

-------
TECHNICAL REPORT DATA

/Please read Inunctions on fhe reverse before completing)

1, REPORT NO. [2.

EPA/600/9-89/062b |

'WSTO im

4, TITLE and subtitle

Proceedings: 1989 Joint Symposium on Stationary Com-
bustion NCx Control, San Francisco, CA, March 6~9,
1989, Volume 2



6. PERFORMING ORGANIZATION CODE

7. AUTHOR(S)

Claudia Runge, Compiler

B. PERFORMING ORGANIZATION REPORT NO.

9, PERFORMING ORGANIZATION NAME AND ADDRESS

Electric Power Research Institute

3412 Hillview Avenue

Palo Alto, California 93404

10. PROGRAM ELEMENT NO.

11. CONTRACT/GRANT NO.

NA (Inhouse)

12. SPONSORING AGENCY NAME AND ADDRESS

EPA, Office of Research and Development

Air and Energy Engineering Research Laboratory
Research Triangle Park, NC 27711

13. TYPE OF REPORT AND PERIOD COVERED

Proceedings; 3/87-3/89

14, SPONSORING AGENCY CODE

EPA/600/13

is. supplementary notes project officer is William P. Linak, Mail Drop 65, 919/
541-5792.

V

is. ABSj^RACT^rp^ proceecj^ngS document presentations at the 1989 Joint Symposium on
Stationary Combustion NCx Control, held March 6-9, 1989, in San Francisco, C/.
The symposium, sponsored by the I). S. EPA and EPRI, was the fifth in a series de-
voted solely to the discussion of control of NCx emissions from stationary sources.
Topics discussed included'low-NCx combustion developments such as burner design
modifications and reburning; coal-, oil-, and gas-fired boiler applications; flue gas
treatment processes; fundamental combustion studies; and industrial and commer-
cial applications. Also presented were manufacturers' updates of commercially
available technology and an overview of environmental issues involving NCx control.,
The 4"day meeting was attended by persons from 14 nations. More than 50 papers *
were presented by EPA and EPRI staff members, utility company representatives,
boiler and related equipment manufacturers, research and. development groups, and
university representatives. The proceedings are in two volumes. Volume 1 includes
background, combustion NOx developments I and II, manufacturer's update, advan-
ced combustion technology, and incineration-.- Volume 2 includes SCR coal applica-
tions, fundamental combustion research, post combustion NCx control development,
fundamental combustion research, new developments, and oil and gas combustion.^

17. KEY WORDS AND DOCUMENT ANALYSIS

a. DESCRIPTORS

b, IDENTI F 1ERS/OPEN ENDED TERMS

c, COSAT1 Field/Group

Pollution Toxicity

Nitrogen Oxides Measurement

Combustion Electric Power Plants

Fossil Fuels Boilers

Catalysis Research
Incinerators Development

Wastes

Pollution Control

Stationary Sources

Hazardous Waste

Municipal Waste

Research and Develop-
ment

13B 06T
07B

21B 10B
21D 13A
07D 14F

14G

19, DISTRIBUTION statement

Release to Public

18. SECURITY CLASS (Thit Report)

Unclassified

21. NO. OF PAGES

¦ ^ i " ~~|

20. SECURITY CLASS (This page)

Unclassified

22. PRICE

EPA Form 2220-1 {S*7J>

-------
ABSTRACT

The 1989 Joint Symposium on Stationary Combustion NOx Control was held in San
Francisco, California, March 6-9, 1989. This Symposium, jointly sponsored by EPRI
and EPA, had as its objective the exchange of information regarding recent
technological and regulatory developments pertaining to stationary combustion N0X
control in the United States and abroad. Topics covered during the Symposium
included; low-NOx combustion development; coal-, oil-, and gas-fired boiler
applications; flue gas treatment; fundamental combustion studies; and environmental
and regulatory issues. The Symposium Proceedings is published in two volumes.

DISCLAIMER

The work described in the papers in this volume was not funded by the
U.S. Environmental Protection Agency. The contents do not necessarily
reflect the views of the Agency and no official endorsement should be
inferred.

-------
PREFACE

The 1989 Joint Symposium on Stationary Combustion N0X Control was held March 6-9,
1989, in Sari Francisco, California. This symposium, jointly sponsored by EPRI and
EPA, was the fifth of its kind devoted solely to the discussion of control of N0X
emissions from stationary sources. Specific topics discussed included the low-NOx
combustion developments such as burner design modifications and reburning, coal-,
oil-, and gas-fired boilers applications, flue gas treatment processes, fundamental
combustion studies, and industrial and commercial applications. Also presented were
manufacturers' updates of commercially available technology and an overview of
environmental issues involving N0X control.

The four-day meeting was attended by persons from 14 nations. Over fifty papers
were presented by EPRI and EPA staff members, utility company representatives,
boiler and related equipment manufacturers, research and development groups, and
university representatives.

Symposium cochairpersons were David Eskinazi, Project Manager in the Air Quality
Control Program of EPRI1s Generation and Storage Division; and Dr. William P. tinak.
Chemical Engineer in the Combustion Research Branch of the EPA's Air and Energy
Engineering Research Laboratory. Each of the cochairpersons made brief introductory
remarks, Mr. David Finnigan, Counsel for the House Energy and Commerce Committee
delivered the keynote address. Written manuscripts were not prepared for the
introductory remarks or keynote address; therefore, they are not published herein.

The Proceedings of the 1989 Joint Symposium have been compiled in two volumes.

Volume 1 contains papers from the following sessions;

•	Session 1:	Background

•	Session 2:	Combustion NQX Developments I

•	Session 3:	Combustion NQX Development II

•	Session 4:	Manufacturer's Update

•	Session 5A:	Advanced Combustion Technology

•	Session 5B:	Incineration

iil

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

•	Session 6A: SCR Coal Applications

•	Session 6B: Fundamental Combustion Research

•	Session 7A: Post Combustion N0X Control Development

•	Session 7B: Fundamental Combustion Research

•	Session 8: New Developments

•	Session 9: Oil and Gas Combustion Applications

Also included in both volumes is an Appendix listing the Symposium attendees.

iv

-------
CONTENTS
VOLUME 1

Paper	Page

SESSION 1:	BACKGROUND

Chairman: I. Torrens, EPRI

"Innovative Clean Coal Technology NOY Control Technology,"J. Temchin

and H. Feibus	X	1-1

"Policies for N0X Control in Europe," A.-K. Hjalmarsson and J. Vernon	1-9

"Environmental Effects of Nitrogen Oxides," R. Perhac	1-29

"NpO Emissions from Fossil Fuel Combustion," W. Linak, J. McSorley,

R. Hall, J. Ryan, R. Srivastava, J, Wendt, and J. Mereb	1-37

"Measurement of ^0 from Combustion Sources," L. Muzio, M. Teague

T. Montgomery, G. Samuel sen, J. Kramlich, and R. Lyon	1-55

SESSION 2;	COMBUSTION N0„ DEVELOPMENTS I

Chairman: D. Eskinazi, EPRI

"The Application of Combustion Modifications for NOx-Reduction to Low-Rank
Coal-Fired Boilers," K. Hein	2-1

"Predicting Boiler and Emissions Performance by Comparative Turbulent/Low

N0X Burner Testing on a Large Test Facility,"	2-23

J. Vatsky and C. Allen

"Reduction of N0X Emissions from a 500MW Front Wall Fired Boiler," P. Beard,
W. Brooks, K. Johnson, K. Matthews, P. Wells, and J. Vatsky	2-53

"NOy Emissions Results for a Low-NOx PM Burner Retrofit," R. Thompson,

G. Shiomoto, D. Shore, M. McDannel, and 0. Eskinazi	2-67

"Retrofit and Boiler Performance Evaluation of the Low N0„ PM Firing System
at Kansas Power & Light," R. Lewis, A. Kwasnik, R. LaFlesh, and D. Eskinazi 2-87

"ENEL's Ongoing and Planned N0X Control Activities," A, Benanti,D. Bonolis,
R, Tarli, A, Baldacci, A. Piantanida, and A. Zennaro	2-109

v

-------
Paper

SESSION 3:

COMBUSTION NO DEVELOPMENT II
Chairman: R. Hall, EPA

"Pilot Evaluation of Reburning for Cyclone Boiler N0X Control," H. Farzan,
I. Rodgers, G. Maringo, A. Kokkinos, and J. Pratapas

"Application of Reburning to a Cyclone Fired Boiler," R. Borio, A. Kwasnlk,
D. Anderson, D. Kirchgessner, R. Lott, A. Kokkinos, and S. Durrani

"Design Methods for Low-NOx Retrofits of Pulverized Coal-Fired Utility
Boilers," S. Morita, K. Kiyama, T, Jimbo, K. Hodozuka, and K. Mine

"New Approach to N0X Control Optimization of N0X and Unburnt Carbon
Losses," M. Kinoshita, T. Kawamura, S. Kaneko, and M. Sakai

"The XCL Burner - Latest Developments and Operating Experience," A. LaRue

"Application of Low N0X Combustion Technologies to a Low Volatile Coal
Firing Boiler," S. Miyamae, T. Kiga, and K. Makino, and K. Suzuki

"N0X Control: The Foster Wheeler Approach," J. Vatsky
"N0X Control Update - 1989," A. LaRue, and P. Cioffi

"1989 Update on N0X Emission Control Technologies at Combustion Engineering,
R. Donais, M. Cohen, and M. McCartney

"Status of N0X Control Technology at Riley Stoker," R. Lisauskas,

E, Reicker, and T. Davis

"The Conversion of Fuel-Nitrogen to N0X in Circulating Fluidized Bed
Combustion," Y. Lee, and M, HiItunen

"N0X Control in Coal Gasification Combined Cycle (IGCC) Systems," N. Holt,
E. clark, and A. Cohn

SESSION 4:

MANUFACTURER'S UPDATE
Chairman: G. Offen, EPRI

SESSION 5A:

ADVANCED COMBUSTION TECHNOLOGY
Chairman: N. Holt, EPRI

vi

-------
Paper	Page

"Development of NCL Control Technologies for Coal-Fueled Stationary Diesel

Power Plants," S. Johnson, C. Senior, C. Katz, and R. Wilson, Jr.	5A-29

SESSION 5B:	INCINERATION

Chairman: W. Linak, EPA

"Reduction of NCL Emissions from MSW Combustion Using Gas Reburning,"

C. Penterson, D. Itse, H. Abbasi, V. Wakamura, and D. Linz	5B-1

"Application of Low N0„ Precorrbustor Technology to the Incineration of

Nitrogenated Wastes," R. Srivastava, J. Ryan, W. Linak, R. Hall,

0. McSorley, and J. Mulholland	5B-23

"The Effect of Fuel Nitrogen on NOx Emissions from a Rotary-Kiln

Incinerator," J, Lighty, D. Gordon, D. Pershing, W. Owens, V. Cundy, and

C. Leger	5B-45

VOLUME 2

SESSION 6A:	SCR COAL APPLICATIONS

Chairman: E. Cichanowicz, EPRI

"Reduction of Nitrogen Oxides from Coal-Fired Power Plants by Using the SCR
Process. Experiences in the Federal Republic of Germany with Pilot and
Commercial Scale Denox Plants," i, Schonbucher	6A-1

"Experience Gained by Neckarwerke from Operation of SCR DeN0„ Units,"

P. Necker	6A-19

"Recent Developments in the SCR System and Its Operational Experiences,"

H. Kuroda, I. Morita, T. Nurataka, F, Nakajima, Y. Kato, and A. Kato	6A-39

"The First De-NO* Installation in the Netherlands, a Demonstration Project at
EPON - Nijmegen Power Station," J. Koppius-Odink, W. Weier, and W. Prins 6A-57

"Operating Experience of SCR Systems at EPDC's Coal-Fired Power Station,"

T. Mori, and N. Shimizu	6A-85

vii

-------
/

Paper	Page

"Technical Feasibility and Economics of SCR N0X Control in Utility
Applications," C. Robie, P. Ireland, and J. Cichanowicz	6A-105

SESSION 6B:	FUNDAMENTAL COMBUSTION RESEARCH

Chairman: B. Martin, EPA

"Prediction of Fuel and Thermal NO in Advanced Combustion Systems,"

R. Boardman, and L. Smoot	68-1

"NQX Reduction in Fuel-Rich Natural Gas and Methanol Turbulent Diffusion
Flames," M, Toqan, J. Teare, J. Beer, A. Weir, Jr., and L, Radak	6B-21

"Reduction of Fuel-NO by Increased Operating Pressure in a laboratory-

Scale Coal Gasifier," K. Nichols, P. Hedman, and A. Blackham	68-39

"Evolution and Reaction of Fuel Nitrogen During Rapid Coal Pyrolysis and
Combustion," G. Haussmann, and C, Kruger	6B-61

"The Effect of Process Variables on NO* and Nitrogen Species Reduction in

Coal Fuel Staging," K. Knill, and M. Morgan	6B-75,

"N0„ Emissions in a Pilot Scale Circulating Fluidized Bed Combustor,"

J. Zhao, J. Grace, C. Lim, C. Brereton, R. Legros, and E. Anthony	6B-93

SESSION 7A:	POST COMBUSTION N0X CONTROL DEVELOPMENT

Chairman; C, Sedman, EPA

"Design and Operation of the SCR-Type NOx-Reduction Plants at the Diirnrohr

Power Station in Austria," M, Novak, and H. Rych	7A-1

"Denox Catalytic Converters for Various Types of Furnaces and Fuels -
Development, Testing, Operation," L. Balling, and D. Hein	7A-27

"Assessment of Japanese SCR Technology for Oil-Fired Boilers and Its
Applicability in the U.S.A.," P. Lowe, W, Ellison, and L. Radak	7A-41

"N0X Control in a Brown Coal-Fired Utilty Boiler," J. Hoffmann,

J. von Bergmann, D. B'dkenbrink, and K. Hein	7A-53

SESSION 7B:	FUNDAMENTAL COMBUSTION RESEARCH

Chairman: W. Linak, EPA

"Formation of Nitrous Oxide from NO and SO? During Solid Fuel Combustion,"

G. deSoete	78-1

viii

-------
Paper	Page

"Fuel Nitrogen Mechanisms Governing N0X Abatement for Low and High Bank

Coals," J, Wendt, A. Bose, and K. Hein	78-14

"Methanol Injection - A New Method for Nox and SO3 Control," R. Lyon	7B-29

SESSION 8:	NEW DEVELOPMENTS

Chairman; E. Plyler, EPA

"Catalyst Poisoning in the Selective Catalytic Reduction Reaction,"

R, Yang, 0, Chen, M, Buzanowski, and J. Cichanowicz	8-1

"Catalytic Filter Bags," M. Kalinowski, and P. Aubourg	8-9

"Advanced In-furnace N0X Control in New European Coal-Fired Power Plant,"
K. Bendixen, and J. Pedersen	8-15

"Two-Stage DeNOx Process Test Data from Switzerland's Largest Incineration

Plant," 0, Jones, L. Muzio, E. Stocker, P. Nliesch, S. Negrea,

G. Lautenschlager, E. Wachter, and G. Rose	8-21

"Effects of Catalyst Developments of the Economics of Selective Catalytic
Reduction," T. Gouker, 0. Solar, and C, Brundrett	8-27

"Full Scale Demonstration of Additive NO? Reduction^ with Dry Sodium

Desulfurization," V. Bland	8-33

"Shell Process for Low-Temperature NOx Control," F. Goudriaan, C. Hesters,

and R. Samson	8-39

"N0X Reduction of Waste Incineration Flue Gas," B. Herriander	8-45

"Reburning and Repowering for NOx Control on Large Utility Boilers,"

S. Chen, E. Holler, D. Pershing, and A. Walters	8-51

SESSION 9:	OIL AND GAS COMBUSTION APPLICATIONS

Chairman: A. Kokkinos, EPRI

"Retrofit of an Advanced Low-N0x Combustion System at Hawaiian Electric's

Oil-Fired Kahe Generating Station," J, Yee, R. Freitas, D. Giovanni,

S. Kerho, and M. McElroy	9-1

"Gas Turbine Nitrogen Oxide (NOJ Control Current Technologies and

Operating Experiences," L. Angetlo, and P. Lowe	9-19

"Demonstration of an Automated Urea Injection System at Encina Unit 2,"

J. Nylander, M. Mansour, and D. Brown	9-35

ix

-------
Paper

"Retrofit Implications of a Low N0X Burner System on a 230-MW Oil- and
Gas-Fired Boiler," J. Carnevale, and J. Klueger	9-57

"Retrofitting Low-NOx Burners for Gas and Oil Firing," J, Gerdes, Jr.,

R, Waibel, and L. Raaak	9-75

"Engineering Evaluation of Combined N0x/S0£ Removal Processes: Interim
Report," to. DePriest, J. Jarvis, and J. Cichanowicz	9-89

APPENDIX A:	LIST OF ATTENDEES

x

-------
Session 6A
SCR COAL APPLICATIONS

Chairman: E. Ciehanowicz, EPRI

6A-i

-------
REDUCTION OF NITROGEN OXIDES FROM COAL-FIRED POWER PLANTS
BY USING THE SCR PROCESS. EXPERIENCES IN THE FEDERAL REPUBLIC OF GERMANY
WITH PILOT AND COMMERCIAL SCALE DENOX PLANTS

Dr. B. Schcnbucher
Energie-Versorgung Schwaben AG
P.O. Box 10 12 43
7000 Stuttgart 10
Federal Republic of Germany

ABSTRACT

According to legal requirements in the Federal Republic of Germany (FRG) the N0X
concentration of flue gas from coal-fired power plants has to be reduced to

3

< 200 mg/m (s.t.p., dry) until 1990 latest. Hard coal-fired power plants with an
electric capacity of a total of about 28 200 MW are retrofitted with DENOX plants
mainly using the selective catalytic reduction process (SCR process). In hard
coal-fired power plants with a dry bottom boiler the DENOX plants are mainly
operating as high dust systems. Power plants with wet bottom boiler have the DENOX
plant appended to the flue gas desulphurisation plant (FGD plant) as a tail end
system. The decision for these two SCR arrangements based on several years of
experiences with pilot plants. The essential knowledge out of the operation of the,
pilot plants and experiences with commercial scale DENOX plants in the FRG will be
presented.

1. INTRODUCTION

New N0X emission standards for flue gas from power stations have been established
in the FRG in April 1984. According to the legal requirements the N0x concen-
tration of flue gas from coal-fired power plants have to be reduced to below
3

200 mg/m (s.t.p., dry) until 1990 latest. Hereof affected are coal-fired power
plants with an electric capacity of a total of 37 500 MW.

In power plants.with an electric capacity of 9 300 MW domestic lignite is fired.
For these power plants primary measures such as off-stoichiometric combustion
(biased burner firing, burner out of service and overfire air) and flue gas
recirculation in combination with low excess air firing will be used to reduce the
N0x concentration below 200 mg/m (s.t.p., dry). For some of the lignite-fired
power plants the application of the selective non-catalytic reduction (SNCR)
process is planned as an additional measure/1,2/.

6A-1

-------
For hard coal-fired power plants with an electric capacity of a total of 28 200 MW
mainly the so called selective-catalytic reduction process is applied. The SCR
process was developed in Japan and used there on a commercial basis for the first
time. In order to test this process under conditions of the German power plants
more than 70 pilot plants have been operated during the last 4 years. Since
December 1985 the first commercial scale DENOX plant according to the SCR process
for a hard coal-fired power plant is in operation. At present 28 hard coal-fired
power plants with an electric capacity of more than 6 400 MW are retrofitted with
a DENOX plant using the SCR process.

The first part of this report shows the essential results from the pilot plant.
The second part will report on experiences with commercial scale plants for hard
coal-fired power plants in the FRG.

2. HARD COAL-FIRED POWER PLANTS IN THE FRG

In the FRG hard coal-fired power plants are mainly operated'in the middle load
range with 3 000 to 5 000 hours of operation per year. Daily start-up and shutdown
as well as week end shutdown between April and September are the usual operating
conditions. At present 90 million tons per year hard coal are used for the
generation of current of which 85 % are domestic hard coal. Table 1 shows-essential
datas of German hard coal.

Table 2 shows the flue gas data at the exit of the economizer. The N0x concen-
tration depends on the layout of the boilerroom, the quality of coal, the tempera-
ture of air for combustion and the excess air as well as on the type of firing
(wet or dry firing) and on primary measures for the N0x reduction. About half of
the hard coal-fired power plants is equipped with wet bottom boilers while the
other half has dry bottom boilers.

Complementary in Table 3 the essential compounds of fly ash out of hard coal-
fired power plants are summarized.

Since the middle of the 70ties the FRG started with considerable efforts to reduce
the NO concentration by optimizing firing conditions (e.g. using low NO burners

*	X

and over fire air ports). Normally, the N0V concentration of flue gas can be

3	3

reduced to between 650 mg/m and 800 mg/m for dry bottom boilers by using primary

measures. For wet bottom boilers the NO concentration mainly is between 1 300 and
•j	x

2 0C0 mg/m (s.t.p., dry).

6A-2

-------
3

This means that the NO emission standard of 200 mg/m (s.t.p., dry) in hard

A

coal-fired power plants can not be met with primary measures only. The application
of secondary N0x reduction measures is necessary.

3.	SCR PROCESS

In the FRG most pilot- and commercial scale plants-for NQX reduction are based on
the selective catalytic reduction process (SCR process) developed in Japan
(Figure 1). With the SCR process, a catalyst is used to reduce the flue gas
components NO and NOg selectively to nitrogen and water by adding ammonia (NH^).
The catalysts differ in their chemical composition and their shape.

The honeycomb and metal carrier plate catalysts consist of surface active titanium
dioxide (Ti02) as carrier material. Vanadium pentoxide	and tungsten

trioxide (WOj) are the most important active components 111. Also suitable for the
NO reduction is activated carbon /4/. The optimal operating temperature for

A

catalysts on basis TiOg is- between 280 °C and 4Q0 °c. Activated carbon is pre-
ferred for temperatures of around 100 °C.

In the FRG the SCR process was tested in various configurations (Figure 2);

upstream the air preheater ("high dust system")

between the electrostatic precipitator and the flue gas
desulphurization plant (FGD plant)

downstream the FGD plant ("tail end system").

In the FRG most FGD systems for power plants work on the wet-scrubber process
using lime or limestone as absorbent and high purity gypsum as byproduct. In some
power plants the spray-dryer process is used for 50^ reduction.

4.	RESULTS FROM DENCX PILOT PLANTS USING THE SCR PROCESS
4.1. DENOX pilot plants in the FRG

The most DENOX pilot plants	used the SCR process /5/. In the majority of cases the
SCR process was tested as a

high dust system for

-- hard coal-fired power plants with dry or wet bottom boilers

5A-3

-------
-- lignite-fired power plants and as a

tail end system for hard coal-fired plants with wet bottom boilers.
4.2. Catalyst properties

The main object of tests in the pilot plants was to determine the properties of
catalyst. These inculde (Table 4):

The N0x removal efficiency (tj nq )

The NQX removal efficiency is influenced by a number of parameters like:

-- the NQX concentration at inlet into reactor and
exit from reactor (C^jJ , C^q ),

X	X

-- the flue gas temperature (tFG5'

-- the molar ratio NH3/N0X (X),

-- the catalyst capacity. This is normally defined as area velocity
(AV: quotient of flue gas volume stream Vpg [rn^/hj and
catalyst surface AMt [m2]),

-- oxygen concentration Cn and

2

-- chemical and physical data of the catalyst.

The catalyst activity (kt)

Corresponding to the Eley Rideal mechanism NH^ is chemically absorbed in a
fast running reaction step and reacts with NO from the gas phase. This,
reaction can be considered as an apparent first order reaction /6/, The
various catalysts which differ with respect to structure and composition
can be characterized by this k value. With the first admission of flue gas
a desactivation of the catalyst begins. The decrease of activity mainly is
influenced by the flue gas and ash composition.

The NH3-slip

In order to keep the M0x removal efficiency at a constant level during
catalyst life-time, the NH^ volume flow must be increased from time to
time. By this way loss of catalyst activity can be compensated. However,
the NH3 concentration in the flue gas downstream the DENOX plant thereby
rises (NH^ slip). As a result of unavoidable secondary reactions of
ammonia with other flue gas components such as SC^ or HC1 the formation of

6A-4

-------
ammonia salts increases. On the basis of operating results obtained from
Japan the admissible NH^ slip was limited to below 5 vppm (= 3.8 mg/m )
by the installation suppliers in order to limit negative effects in the
plant equipment downstream the DENOX plant such as air preheater or FGD
plant. An other point is to avoid an unallowable contamination of the
fly ash with NHj /7-10/.

The SOg conversion rate

Finally the SC^ conversion rate is another important charcteristic of
the catalyst. It must be considered together with the NHg slip.

The investigations in pilot' plants have shown- that with a NH^/NC^ molar ratio
higher than 0.85 to 0.95 a measurable slip will occur within a relatively
short operating period. If the molar ratio and thus the N0X removal efficiency is
raised even further there will be a greater risk that the specified limit for the
NH^ slip will soon be reached or exceeded.

It is therefore important to determine how long the NO removal efficiency can be

A

maintained on a certain level without exceeding the slip limit of 5 vppm. The
results for the progress of activity with various types of firing, fuels and
configurations are shown in Figure 3 /11/. Kg is the activity coefficient of fresh
catalyst and k. after an operation period of t. Characteristic is that the
decrease of the activity in relationship to the operating time is relativly small
for high dust SCR systems downstream hard coal-fired boilers with dry firing and
for tail end SCR systems. These results are in good correlation with Japanese
results for hard coal-fired power plants.

In case of the high dust SCR system downstream hard coal-fired boilers with wet
firing and 100 % ash recirculation the activity decline runs very steep /11,12/.
Investigations in the FRG indicate that this decline mainly is caused by gaseous:
arsenic compounds in the flue gas /13,14/. Besides the arsenic content of the coal
(with domestic coal mostly < 5 vppm, partly up to 20 - 25 vppm) the type of firing
and the CaO content of the fly ash are of great influence on the arsenic concen-
tration in flue gas.

The high temperatures in the combustion chambers of slag tap furnaces lead to a
increased volatilization of arsenic compounds compared to dry bottom firing
systems. Experiences showed that for wet bottom boilers with 100 % ash recircu-
lation the concentration of arsenic compounds in the flue gas (upstream air
preheater) can reach values up to 1 mg/m . The arsenic concentration in the flue

6A-5

-------
gas also depends on the CaO content in fly ash. Arsenicoxid reacts with CaC at
temperatures above 500 °C to form calciumarsenit (Ca^AsO^):

3 CaO + ASO2 + O2	Ca^tmsO^)2

By using the tail end SCR system especially for hard coal-fired boilers with wet

bottom firing the catalyst is neither subjected to the full dust content of flue

gas nor with components like calcium-, magnesium-, potassiume-, sodium- or

arsenicoxid. Experiences have shown that with the tail end system it is possible

to use catalysts with considerably higher specific geometric surfaces than in the

case of the high dust system (Figure 4). The main reason for this is that the dust

3

concentration of the flue gas downstream the FGD plant normally is below 50 mg/im
(s.t.p., dry) so that pitch of the catalyst can be reduced. With the high dust
system the pitch is between 7.4 and 7.0 mm. With the tail end system a pitch of
4.2 mm normally is the optimum compromise between specific geometric surface and
pressure loss /15,16/.

Moreover by modifying the composition of the catalyst material and/or the manu-
facturing process, activity can be increased. Compared to high dust systems the
comparable catalyst volume with low dust systems can be 50 to 60 % less. At the
same time the reaction temperature can be lowerd to about 300 °C. Guaranteed
life-time for catalysts is the case of the tail end system is 24000 hours of
operation and 5 years in maximum. Guaranteed catalyst life-time for high dust
systems downstream boilers with slag-tap firing and 100 % ash recirculation is not
more than 12000 hours or 2 years in maximum.

The disadvantages of tail end systems are a markedly higher pressure loss, an
expensive reheating system and the cost of fuel needed to cover the temperature
loss in the gas preheater /17/.

5. RESULTS FROM COMMERCIAL SCALE DEMQX PLANTS USING THE SCR PROCESS

At present 28 hard coal-fired power plants with an electric capacity of a total of
7 470 MW are equipped with DENGX plants using the SCR process. All 28 DENOX plants
have been retrofitted. 18 DENOX plants (5 300 MW) were retrofitted as a high dust
configuration. Some of these DENOX plants have been in operation since 3 years
/11,14,18-26/. 11 DENOX plants ( 1 070 MW) are erected as tail end systems down-
stream of the FGD plant. Only two of the power plants with wet bottom boilers are
retrofitted with a high dust SCR system.

6A-6

-------
The first DENQX. plant as a tail end SCR system is in operation since September
1987 /27/, DENQX plants for dry bottom boilers are designed tor a NO reduction

X

efficiency of 75 to 80 I. With a space velocity (SV) of 2 650 to 3 000 [1/hJ
(nominal pitch: 7.0 - 7.4) guaranteed catalyst life-time is between 12 000 and
16 000 operating hours and a maximum of 3 to 4 years (with a NHj. slip of < 5 vppm
at the end of the warranty period). The DENOX plants for the two power plants with
a wet bottom boiler are designed for a N0x reduction efficiency of 90 %. The SV in
these cases is in the range of 2 000 [l/h] . The expected life-time of the catalysts
for these both DENOX plants is about 12 000 operating hours. In both plants a
100 % ash recirculation is not realized and therefore, the arsenic problem should
not be so considerable.

The operation of the DENOX plants so far has shown that the expected high M0X
reduction efficiency will be reached. The experiences also show that espescially
in case of high dust configurations the flow to the first catalyst layer is of
great importance.

The catalyst can only then be used at an optimum when a uniform velocity, tempera-
ture and NHj/NQx distribution exists upstream from the first catalyst layer and no
dust stratification occurs, which locally leads to erosion or plugging of the
catalyst. This aim is often very difficult to achieve particularly with retrofits.
The best way is to carry out flow tests on' scale models in order to receive
conclusions for large scale plants. By installation of a flow rectifier layer
(dummy) upstream from the first catalyst layer, an optimum design of the reactor
hood contour and the positioning of guide plates (baffles) a good velocity
distribution with a deviation of 10 - 12 % from the average value can be achieved.
Because of the expected unequal distribution of the NO and flue gas quantities
upstream from the ammonia injection position an injection grid with many individual
nozzles (30 - 40 nozzles per square metre) is necessary.

The example of the DENOX plant for Unit 7 of our He:Ibronr power station shows
some characteristic operation results. Unit 7 has an electric output of 700 MW. It
is fired with domestic hard coal and is equipped with a dry bottom boiler and a
tangential firing system. The unit is in operation since 1985 with approx.
4 000 operating hours per year. The DENOX plant was retrofitted and is in operation
since September 1986. F-igure 5 shows essential design data of the DENOX plant
/18/.

6A-7

-------
The NHj slip did not increase after 8 500 operating hours. This also applies for
the SO2/SO2 conversion rate. Above the first catalyst layer soot blowers are
installed which are in operation every two weeks. So far the pressure loss
Increase is lower than expected.

The NO concentration at the exit of the DENOX plant works this way that at

3

constant boiler load the value is about 160 - 170 mg/m (.s.t.p., dry). In case of
load changes the control system ensures that the N0x limiting value does not
exceed 200 mg/m"* (s.t.p., dry).

In Figure 6 the NO decline across the three catalyst layers is shown. This figure

A

gives an example how the main conversion zone for N0x is at first clearly in the
first catalyst layer. The third layer apparently only serves the adjustment of the
NH^ slip. This tendancy remains principally with used catalysts, however, the N0X
conversion shifts increasingly into the second and third layer.

Even though few DENOX plants have more than 10 000 - 12 000 operating hours, it
can be assumed that the guaranteed catalyst life-time will be reached. The SCR
process as high dust system proves to be a reliable technology for reduction of
NO concentration to < 200 mg/m (s.t.p., dry) for dry bottom boilers.

A

Furthermore, the experiences show that erosion or plugging of the catalyst is not
to be expected as long as the ash content of the flue gas is below 10 to 12 g/m .
But the use of soot blowers is necessary. In case of hard coal with relative high
ash content erosion and plugging is possible. Because of this it was necessary to
exchange the catalyst earlier in some plants.

At present the NH^ slip normally is determined on wet-chemical methods. Continuous-
ly operating measurement instruments are under going tests /23,28/.

The experiences show that.with a NH^ slip of more than 2 vppm fly ash cannot used
by the cement industry anymore. This especially applies for hard coal with low ash
content (< 5 - 8 %) /29/.

Till now the operation of tail end SCR DENOX plants is without problems. The
activity loss of the catalyst is very low as this was already found out in the
pilot plants.

At present more than 50 DENOX plants for hard coal-fired power plants with wet
bottom boilers and an electric capacity of a total of more than 5 300 MW as the
tail end configuration are in operation or under erection.

6A-8

-------
Essential reasons therefor are the short life-tine for catalysts in the high dust
system downstream wet bottom boilers and space problems when placing a DENOX plant
upstream of the air preheater.

Under these conditions the tail end system is very often the cheaper solution even
though considerable additional energy costs must be accepted for reheating the
flue gas /17/.

We hope that modified catalysts which have been employed recently in experimental
plants will have a longer life-time. However, it is advisable to wait for longterm
experiences before reliable statements can be made about the life-time of these
improves catalysts /30/.

6. References

/I/ Kallmeyer, D. and Konig, J.: Einfuhrung von NO -Minderungstechniken in mit

A

rheinischer Braunkohle gefeuerten Kraftwerken.. BWK Brennstcff-Warme-Kraft 39
(1987), issue 5, pp. 249 - 253)

12/ Reidick, H.: Primare NQX-Minderung bei Braunkohlefeuerungen. EVT-Register,
issue 46 ( 1987), pp. 32 - 36 (EVT: Energie- und Verfahrenstechnik GmbH,
P.O. Box 10 12 43, Jchannesstr. 37-45, 7000 Stuttgart 10, FRG]

ill Biffar, W., Drews, R., Hess, K., Lehrert, R. and Scheidsteger, D.: Ent-
stickung von Rauchgasen mit SCR-Katalysatoren. BWK Brennstoff-Warme-Kraft 38
(1986), issue 5, pp. 211-216)

/4/ Sodec, F.: Einsatz von Aktivkoksen zur NO -Minderung. Lecture in "Haus der

A

Technik", Essen, June 23/24, 1987.

/5/ Stand der Stickoxidminderung durch SekundarmaBnahmen. VGB Kraftwerkstechnik
87(1988), issue 1, pp. 60-68

/6/ Weber, E. and HObner, K.: Ergebnisse reaktionskinetischer Untersuchungen zur
katalytischen N0x-Reduktion mit Ammoniak. In: VGB-Handbuch "N0x-Bildung und
NOx-Minderung bei Dampferzeugern fur fossile Brennstoffe". Essen 1986/88

/7/ Bubjak, W. and Gutberlet, H.: Diskontinuierliche Ermittlung des NHgSchlupfes
durch Untersuchung des Flugstaubes. In: VGB-Handbuch /6/.

/8/ NO^-Symposium Karlsruhe 1985 - Internationale Betriebserfahrungen.

University of Karlsruhe, february 1985

6A-9

-------
19/ Hem, K.R.G., Konig, J. and Sals, M.; Problematic des Ammoniakschlupfes bei
Anlagen zur NOx-M:nderung. VGB Kraftwerkstechnik 67 (1987), issue 9pp. 861-865

/10/ Bubjak, W. and Gutberlet, H.: Untersuchungen zum Eintrag von NH^ aus DEN0X-
Anlagen in REA-Abwasser. VGB Kraftwerkstechnik 67 (1987), issue 9, pp.
876-883

/11/ Stabler, K., Schonbucher, B. and Bilger, H.; NOx-Minderung durch Sekundir-
maBnahmen. VGB Kraftwerkstechnik 68 (1988), issue 7, pp. 735-743 and english
issue: VGB Kraftwerkstechnik 68(1988), issue 7, pp. 652-659

/12/ Schailert, B.; Erfahrungen aus einem zweijahrigen Betrieb m:t DENOX-Pi iot-
Anlagen. VGB Kraftwerkstechnik 88 (1988), issue 4, pp. 432-440

/13/ Gutberlet, H.: EinfluB der Feuerungsart auf die Vergiftung von DENOX-Kataly-
satoren durch Arsen. VGB Kraftwerkstechnik 68 (1988), issue 3, pp. 287-293

/14/ Schumacher, B.: Bisherige Erfahrungen mit SCR-DENOX-Anlagen zur Stickoxid-
minderung. EVT-Register, issue 47 (1988), pp 27-38

/15/ Schonbucher, B., Tilch, G., Quittek, Ch.: SCR-Versuchsanlagen hinter einer
NaB-REA. VGB Handbuch /6/

/16/ Hannes, K. and Eichholtz, A.: Erfahrungen der Steag AG mit Versuchsanlagen
zur Stickoxidminderung. Anwendungsreport Rauchgasreinigung. VDI-Verlag 1986,
pp. 84-89

/17/ Schonbucher, 3. and Quittek, Ch.: Nachrtistung von DENOX-Anlagen in bestehende
Kraftwerksanlagen. Sammelband VGB-Konferenz "Kraftwerk und Umwelt", pp.
151-160

/18/ Schonbucher, B. and Fritz, P.: Auslegung, Anordnung und Funktion der DENQX-
Anlage fur Block 7 lm Kraftwerk Heilbronn. VGB Kraftwerkstechnik 67 (1987),
issue 3, pp. 245-252

/19/ Schonbucher, B., Fritz, P. and Quittek, Ch.: Betriebserfahrungen mit den
DENOX-Anlagen in Kraftwerk Heilbronn. Sammelband VGB-Konfernz "Kraftwerk und
Umwelt 1987" pp. 65-71

/20/ Schonbucher, B.: Betriebserfahrungen m.it der DENOX-Anlage des Blockes 7 lm
Heizkraftwerk Heilbronn der E.nergie-Versorgung Schwaben AG. VDI-Special:
NOx-Minderung in Rauchgasen 10 (1987), issue 10, pp. 57-61

6A-10

-------
!2M Kister, D. and Rieche, 6.: Demonstrationsanlage zur Rauchgasentstickung des
Schmelzkammerkessels C im Kraftwerk Knepper C. BWK Brennstoff-Warme-Kraft 39
(1987), issue 7, pp. 93-98

/22/ Necker, P. and Becker, J.: Auslegung, Anordnung und Funktion der DENGX-
Anlage, Kraftwerk Altbach/Deizisau, Block 5, einschlieBlich Betriebserfah-
rungen, VGB Kraftwerkstechnik 67 (1987), issue 4, pp. 368-377

/23/ Necker, F.: Betriebserfahrungen mit der Entsti ckung im Kraftwerk Altbach/
Deizisau, Block 5. VDI-Special /20/, pp. 35-42

/24/ Necker, P.: Weitere Erfahrungen mit der DENOX-Anlage in Block 5, Kraftwerk
Altbach/Deizisau, nach mehr als einjahriger Betriebszeit: In: VGB-Handbuch /6/

/25/ Wahl, D.-J.: and Seibel, G.: Verringerung der Stickoxidemissionen in den
kohle- und ofbefeuerten Kraftwerksblocken der VKR mit katalytischem und
nicht-katalytischem Verfahren. VDI-Special /20/, pp. 44-56

/26/ Wahl, D.-J.: Neueste Erfahrungen mit den Entstickungsanlagen der VKR. VGB
Kraftwerkstechnik 67 (1987), issue 12, pp. 1198-1203

/27/ Becker, J.: High-Tech fOr low-dust. Energie-Spektrum (1988), issue 9,
pp. 18 - 28

/28/ Barnet-Wiemer, H., Heidendael, M., Karanatsios, I. and Vijgen, H.: Analytik
und Regeltechrik an einer Versuchsanlage zur katalytischen NO^-Emissions-
minderung. In; VGB Handbuch /6/.

/29/ Schallert, B., Gutfaerlet. H. and Wahl, D.-J.: Die Aiterung von DENOX-Anlager,
als LeitgroBe fur die Auswahl und Einsatzbedingungen in groBtechnischen
Anlagen. In: VGB-Handbuch /6/

/30/ Probst, K.: NO -Minderung in den Versuchsanlagen im Kraftwerk Franken II. In

A

VGB-Handbuch /6/.

6A-11

-------
manufacturers

4 NG+4 NH, + 0,

2 NO,+ 4 NH3 + Oj
(100) 250
-------
3 « $

operating hours

8 !> 8 1391390

type of fifing

SCfl System

ary/wat battcm

tad em

(fry Mttoni

high dust

wet bottom

high dust

<100% fly as* recircutetian)

¦JE3

dry bottom'

high dust

Japanase experiences

Figure 3. DENOX pilot plants in the FRG.
catalyst activity

high dust tail end

system system

fsi

pitch

430-470

surface

mVnfi3

750

1.0

activity*

1.0-1.2

100

volume

% '

40-50

320-400

temperature

'•C

(270)300-320

100

pressure loss

-180-250

12000

lifetime"

24 000

•basis: V2/0,/Ti0j , **NH3 slip - 5 vppm {» 3,8 mo/m3)

Figure 4. SCR catalysts for wet bottom boilers

6A-13

-------
• 2 • IWOWrf/hdtd.dry)

• 2M mg/ms EOOOtv SysasinnwIiTMiifjiaraiitee)

1S33D -18 OBO nfexpsctetiS
NH, stp • -S vppm{« a a mg/Bi3)

Figure 5. Heilbronn power station, unit 7,

DENOX plant (high dust SCR system)

«• 50

t ^

«0 3

D T

50 1

-1

operating tours

0 2S08 4800 SMO » MOO

2500 *m

U-J LIT

sioo n ma

due gas

catalyst
layer

Z50Q

4800

6300 ft 8500

«¦ so
m

; catalyst
: layer

Figure 6, hard coal-fired boiler. NQX decline across the
catalyst layers for a high dust SCR system.

6A-14

-------
Table 1

properties of german hard coal

healing value (net) MJ/kg 1B - 30

humidity % 6-12

ash % 8-12(30)

volatile % 17-30(40)

sulphur % 0.8-1,5

nitrogen % 1.2-10

chlorine % 0.1-0.3

Table 2

hard coal-fired boilers> flue gas composition (behind economizer)

temperature

¦c

300

400(420)

water

COj

S02

mg/m3

1500

3500

so,

mg/rn3

NO*

mg/mJ

650

O
CO

1200

2000"'

dust

g/m3

-12(40)

• it* bottom tmg*r*tm
«* wot bottom Ma
-------
Table 3

hard coal-fired boilers* fly ash composition

»2 % 42 - 55

F8j03 % S-12

AljOj % 24-33

CaO % 0.5-a

MgO % 0.6-4

Na^} % 0,2-1.2

KjO % 1-5.5

C % *5

Table 4

SCR process, catalyst properties
*N0X removal efficiency

1
-------
Tables

Hefflsronn power station. Unit 7. DENOX plant
change of catalyst caracteristics

hous of operation NH3 sfip SO/SO3 pressure loss

vppm

conversion
I

(reactor)

0.4

5.7

2500

0.6

5.S

MOO

0.5

5.8

6300

0.5

6.3

8500

0,6

6.5

6A-17

-------
Experience Gained by Neckarwerke from Operation of SCR DeNOx Units

Dr.-Ing. Peter Necker
Esslingen

Abstract

Neckarwerke installed two SCR DeNOx units as a retrofit measure for existing
power stations. Since December 1985, a high-dust DeNOx unit has been in ope-
ration in the Altbaeh/Deizisau Power Station, Unit 5, downstream of a boiler
with dry ash removal (operating time in December 1988: 18,400 hours). In the
Walheim Power Station, Unit 2 (150 MW), an SCR DeNOx unit has been operated
since November 1987, also as a high-clust system downstream of a slag tap
fired boiler. The operating experience gained with this DeN'Ox unit is dis-
cussed in this paper. The measures resulting from the operating experience
are presented, in particular in the field of plant engineering arid behavior
of the catalytic converter, Two effects in connection with the DeNOx units -
the NHs freight of the flue dust and the air preheater loading - which have
gained particular importance for operation, are described.

1. Introduction

Neckarwerke are a regional power utility company (7.5 billion kWh/a,
peak load of network 1370 MW, 6 coal-fired units, 2 interests in
nuclear power stations). They operate 2 coal-fired units with SCR
units to reduce the NOx emissions.

In the Altbaeh/Deizisau power station, an SCR unit has been operated
on Unit 5 (420 MW) downstream of a boiler with dry ash removal since
1985. At the end of December 1988, this SGR unit has been in opera-
tion for 18,400 hours. In the Walheim power station, also an SCR De-
MOx unit has been operated on Unit 2 (150 MW) downstream of a slag
tap fired boiler since October 1987. At the end of December 1988,
this unit had 7,553 operating hours. Both units are of the high-dust
SCR type. The emission limit value to be met by both units is 200 mg
NOx/m3 (Unit 5: 6% Oz, Unit 2: 5% 02).

Because both units are demonstration objects for introduction of the
SCR technology in Germany, they are sponsored by the German Federal
Environment Agency (UBA) in Berlin.

2. The DeNOx Units of Neckarwerke

Both the Altbaeh/Deizisau Unit 5 and Walheim Unit 2 systems were re-
trofitted in existing stations.

Initially, the DeNOx unit in Unit 5 had been designed so that only
80% of the flue gases, were to be purified in the catalytic converter

6A-19

-------
at full boiler load (SCR partial-flow system). Fig. 1, left side,
shows this circuit. A paper on the DeNOx unit /1/ and the first oper-
ating experience gained with the SCR system - which were altogether
positive - (up to approx. 6,000 operating hours) was presented at the
EPA/EPRI NCx Symposium 1987 in New Orleans /2/. Further operating ex-
perience has shown that the NO* contant downstream of the boiler,
i.e. upstream of the catalytic converter, depends on the coal quality
used to an extremely great extent. Fig. 2, left side, illustrates
this great dependence; the concentrations at the inlet into the SCR
unit are between 400 rag NQx/rrr1 and almost 800 mg NOx/in3 . Since this
partial-flow system had been designed for an NOx inlet concentration
between 550 and 600 mg NOx/m3, the SCR unit was clearly overstrained
with these high inlet concentrations with acceptable NK3 leakage, or
unrestricted use of all coal types available on the German market was
not possible. Pig. 3 gives a comparison between tne required separa-
tion level of a partial-flow system and that of a full-flow system as
a function of the NOx inlet concentration. If it is assumed that a
separation level of 75 to 80% is the level that can be attained with-
out special measures, this level is already reached at just under
500 mg N0x/m3 with a partial-flow system while a full-flow system
comes up to this level at 800 mg NOx/ni3 and above. In order to be
more flexible with regard to the purchasing of coal and operation of
the boiler unit, Neckarverke therefore decided to convert the par-
tial-flow system into a full-flow system in summer 1988. Fig. 1,
right side, shows the new plant design. This configuration, with 1
start-up bypass and two reactors, has been operated since November
1988. Fig. 4 illustrates the installation layout in the boiler house.
The alteration work made it necessary to shut down the unit in May
1988 for 3 weeks (regular inspection) and for 1 week in November
1988.

Fig. 5 shows the installation layout of the DeNOx unit at Unit 2 in
the Walheim power station. The holler - which originally had 2 flues
- was altered so that the flue gases are routed upwards in a newly
built flue gas duct and from top to bottom across the catalytic con-
verter. Due to the .existing two air preheaters, the entire DeNOx
system has been designed with 2 ducts. NHa injection takes place at
the beginning of the upward duct section. Of the 4 planned catalytic
converter levels, 3 levels are currently equipped. In the air prehea-
ters, the existing metal sheets were replaced by sheets which are
suitahle for DeNOx (enameled and easy to clean). Due to the altera-
tion in spring 1987, the unit had to he shut down for 6 months.

Operating Experience

This report on the operating experience places less emphasis on re-
sults which are produced and explained with the highest degree of
scientific meticulousness, but is intended to provide information on
experience gained in "normal" operation, with all its constraints and
shortcomings. The measures derived from that are to be presented and
an indication given of how the problems can be better coped with.

6A-20

-------
3.1 Experience with Process Engineering

The SOx reduction in SCR systems is based on the addition of NH3 to
NO* containing flue gas in such a manner that the appropriate NH'j
molecule is available to each NO* molecule to the largest possible
extent. The NQx/NHs mixture reacts at the catalytic converter to form
nitrogen and water vapor. For this, the process must be designed so
that these requirements can be safely met in large-scale plants under
all operating conditions. This results in the following design cri-
teria for the system:

- Adjustable NHa dosing (according to the required N0\ volume)

- Intimate .mixing

- Satisfying the reaction conditions at the catalytic converter

- Measures to protect the catalytic converter against harmful flue
gas components (e.g. dust).

Table 1 is a compilation of the chemical/physical requirements, supp-
lemented by the possible technical measures and the interfering ef-
fects. The need to meet the requirements to an ever increasing degree
rises with increasing NOx separation level. Unit 5, the first SCR
DeNOx unit in Germany on a large-scale industrial basis, can be char-
acterized in this respect as follows:

- Difficult dosing conditions

- Extremely short mixing section.

- Mixing effects by eddy tlow at the shutoff dampers

- Flow homogenization by duct extension

- Favorable dust load- (6 to 8 gr/m3).

The extremely difficult conditions were able to be handled satisfac-
torily to a certain degree at the main reactor for separation levels
of up to 7 5%, due to

- previous model experiments,

- two separate NH3 injection systems, and

- some modifications and adaptations of the NH3 nozzle diameters
which were corrected several times.

For installation of the secondary catalytic converter, the conditions
for dosing and mixing are solved far better - due .to the experience
gained from the primary catalytic converter (highly turbulent dosing
range, baffling, mixing section, section for flow steadying). Figures
6 and 7 present the results from fiH3 leakage measurements downstream
of the primary and secondary catalytic converters, for a separation
degree of approx. 71%. The distribution of the NHs leakage across the

6A-21

-------
flue gas duct cross section downstream of the catalytic converter can
be used as a measure to indicate that the requirements have been met.
The better conditions at the secondary converter result in the fact
that the NHs leakage downstream of the catalytic converter is distri-
buted far more evenly, A criterion for this is the standard deviation
from average. The results were obtained by a specific large-scale
test program.

Downstream of the secondary catalytic converter (Fig. 7), the stand-
ard deviation from the average value is 36%, and downstream of the
primary catalytic converter it is 691. It must be pointed out here
that the absolute NH3 leakage values are very low, with approx.
0-6 nig NH3/1113 downstream of the primary catalytic converter, and
approx. -0.15 Kg NH3 downstream of the secondary converter (see Chap-
ter 3.2).

The Walheim Station, Unit 2, is equipped with a slag tap fired boi-
ler. 30% of the German boiler plants have this type of firing which
produces very high NO* emissions. These are generally between 1200
and 2400 mg NGx/m3. Before the NOx reduction measures were installed
in 1987, Unit 2 in Walheim had emission values between 1900 and 2200
mg N0x/m3 - depending on the coal type used (Pig. 2, right side). For
this reason, primary and secondary measures had to be taken to attain
the limit value of 200 mg NOx/m3. By installing multi-stage mixing
burners supplied by Steinmueller as a primary measure, the NOx values
downstream of the boiler were reduced to 1100 to 1600 mg NOx/m3 - de-
pending on the coal type. This meant that a separation level of 82 -
87% was required for the secondary measure! For such extreme separa-
tion levels, special emphasis had to be placed on an optimum NH3
dosing (20 nozzle systems with separate external adjustment feature),
long mixing section, baffles and even flow distribution. Fig. 5 shows
the solution found for installation from compliance with these re-
quirements. Special mention is made of the constant pressure distri-
bution upstream of the rectifier level. The necessarv separation
levels of approx. 85% are attained by these measures with extremely
low leakage rates. The NH3 leakage rates are 0.15 rig NHa/m3 after
approx. 6,600 hours. NOx and NH3 profile measurements across the flue
gas cross section resulted in standard deviations of 30% for NOx arid
19% for NHa - referred to the average value. At the present time,
only 3 catalytic converter levels are equipped.

The experience gained from operation of Unit 5 was fully taken into
account in the design of the DeNOx system in unit 2 of the Walheim
power station, and it thus created the conditions for troublefree
operation of Unit 2.

3.2 NHa Concentration at Flue Dust

The NHa. deposits at the flue dust have proven to be the crucial prob-
lem for the operating characteristics of the DeNOx system at Unit 5
in the Altbach/Deizisau power station. As the further measures can be
better understood when this operating problem is known, this item is
dealt with first in the description of the operating experience.

6A-22

-------
In the order placed for the DeNQx system, the SHs leakage downstream
of the catalytic converter was specified at max. 5 vpm (= 3.8 mg
NH3/1113) as the guaranteed value during the catalytic converter life,
This procedure was based on information, recommendations and "exper-
ience" of the licensors of the German plant suppliers. Own operating
experience and findings were not yet available in Germany. The accep-
tance of these "guaranteed values" has proven to be an error from the
point of view of the plant operator; this value is too favorable for
the suppliers.

For disposal of the fly ash separated off in the electric filter,
Neckarwerke have entered into an agreement with a large cement mill.
Since the fly ash has pozzolanic properties, a share of appro;. 22%
fly ash can be added to'the fired cement'clinker; thus producing a
fly ash cement with specific properties. If the NH3 concentration in
the flue dust exceeds a defined value, this may result m a release
of NH3 during concrete mixing (i.e. wetting) due to the alkaline
characteristics of the cement and, thus, to a noticeable smell.
Depending on the intended purpose and type of use, we believe that
100' mg NHa/kg fly ash - without special treatment of the fly ash - is
a certain limit value of the NH3 concentration. For 60 mg NHa/kg and
above, special conditions apply for disposal. If the SH.3 content of
the fly ash jeopardizes the utilisation of the flue dust and, thus,
waste disposal of a power station, and. this is caused by the DeNOx
system, this problem must be carefully studied. Own tests have shown
that the concrete properties are not altered by these effects.

Neckarwerke use relatively high-quality hard coal types, with an ash
percentage of 6 to 8%. Initial information from the years 1984/85 and
1986 /4/ on the path the NH3 leakage takes showed the distribution
given in Fig, 8, top; i.e. approx. 70% of the NH3 leakage are emitted
via the fly ash. Mathematically speaking, on this basis the use of a
coal with 7% ash and an NHa leakage of 1.4 vpm (= 1.0 mg/ni3) results
in approx. 100 mg NHs/kg, and a leakage of 5 vpm (= 3.8 mg/m3) in
approx. 350 mg NH3/kg.

During 10 test series that were carried out in the meantime and took
approx. 1 week each, an NH3 depositing rate on the flue dust of 70%
was not able to be reproduced. The vast majority of measured values
is clearly below this. Fig- 9 illustrates the results obtained by the
10 test series. The low deposition rates on the flue dust are partly
combined with very high separation levels and the corresponding high-
er leakage.

The results presented on the bottom of Fig. 8 were found by a test
series that was specially aimed at pinpointing the NIli leakage. A
clearly measurable NH3 volume flow of up to 50% of the NHi leakage
could be detected in the heated combustion and mill air flow /6/. The
NH3 deposition rate on the flue dust was around 50%. Further, an NH3
deposition on the air preheater of 10% could not be verified. Based'
on long-term observation of the air preheater in connection with air
preheater washing, values of clearly < 5% must be assumed (see Chap-
ter 3.4), With these test results, the incongruity /3,5/ in the ana-
lysis of the NH3 leakage which could be recognized since 1986 was
able to be explained by the NH3 transfer fiom the flue gas side to

6A-23

-------
the air side in the air preheater. It must however be emphasized at
this point that these statements are subject to relatively high
tolerances because the measurements were carried out. on a large-scale
plant- It is extremely difficult to maintain all parameters of in-
fluence constant over the extremely long test period required for
such tests (desirable are > 8 hours). Fig. 10 shows an example of the
effect reproduced repeatedly during these test series, i.e. that the
NH3 concentration on the flue dust increases over the test period
under constant load (increase by approx. 15 to 30%) /7/,

On the basis of the above experience, we now design the catalytic
converters for an NH3 leakage of 2 vpm (= 1,5 rag NHa/m-). However, we
wish to warn against an oversizing of the catalytic converters, be-
cause high converter volumes (= low Sv velocities) lead to an in-
crease in the SO2/SO3 conversion rate, thus creating fresh problems.

Concerning Unit 2 in Walheim, generally speaking it can be stated
that the leakage values are still very low, due to the process opti-
mization for mixing and the operating time of 6,000 hours. The NH3
values on the flue dust are therefore between "0 and 20 mg NHa/kg.

While parameters "such as boiler load, ash percentage and calorific
value of the coal were constant or corrected in the above correla-
tions between NH3 leakage and JJH3 concentrations on the flue dust,
Fig. 11 gives a result of the NHa concentration on the flue dust as
obtained in "normal" operation of Unit 5 since its startup in Decem-
ber 1985, subject to all the ups and downs of everyday operation.
During the first six months (up to about June '86), the NTI3 content
of the flue dust was not given the importance it now has; rather the
optimization and improvement measures on the plant played the main
role. This was followed by the phase of introduction of analytic che-
mistry in the plant laboratorv which is shown by some uncertainty of
the test results (July to August 1986). If the results of the test
campaigns, for which extreme situations were created intentionaj1y
for the plant, are not taken into account, the measurement results of
the NH3 concentration on the f!ue dust can be put down to some trend
curves: first the increase in NH3 concentration between September
1986 and April 1987, then the phase during which the NHa dosing was
reduced to limit the NH3 concentration in the flue dust (May to De-
cember 1987), followed by two phases with rising NH3 concentrations •
in the flue dust. These four phases exactly illustrate the history of
the Unit 5 DeNOx system, the operating experience with the catalytic
converter and the conclusions drawn.

3.3 Catalytic Converter Behavior

3.3.1 Catalytic Converter Behavior in Unit 5

Part of the catalytic converters in the DeNOx system of Unit 5 have
been in operation for 18,400 hours till December 1988, Fig. 12 gives
the activity curve of these catalytic converters. It can be clearly
recognized that the actual activity decrease was below the value on

6A-24

-------
which the design of the system was based. Dust deposits, blockage or
erosion phenomena could not be found. Therefore, the function of the
catalytic converter can be fully confirmed.

A decreasing activity is characterized by. the fact that the NHs leak-
age increases with constant separation efficiency. This continuous
increase in the NH3 leakage over the period of operation also becomes
apparent by the increase m. the NHa concentration on the flue dust in
Fig. 11, This trend manifests itself in different ramps. Until Hay
1987, a first increase takes place - the result of the decrease in
activity which can be recognized in Fig. 12 up to approx. 8,000 oper-
ating hours. It began to show that the critical NHo concentrations on
the flue dust would be reached. It must be pointed out that the NH;i
leakage of 2 mg Nte/m3 downstream of the catalytic converter was
still far below the guaranteed value at this point in time. On the
basis of the load forecasts for 1988, we decided in autumn 1987 to

- equip the third layer with a new catalytic converter in December
1987, and to

- convert the system into a full-flow DeMGx system in May.

According to the operating'experience after approx. 10,000 hours in
autumn 1987 - which gave no indication of problems with dust erosion,
blockage and deposits -, we decided to use catalytic converters vich
a reduced web thickness (instead of 1.4 mm now 1.05 mm; increase in
inner surface + 16%) for the new equipment. This improvement resulted
in a reduction of the NH3 concentration in the flue dust, at the same
time maintaining the limit value of 200s mg N0x/m3 (since January
'88). After this, again an increase in the NH3 concentration in the
flue dust due to the activity decrease in the caf a i vt i c: converter was
found. The catalytic converters with the smaller subdivision have
stood the test.

In May 1988, the reactor bypass ducts were shut off during the annual
Unit inspection so that the secondary catalytic converter could be
installed. From Kay to November 1988, the flue gas was passed soJeiy
through the reactor. Thus, the reactor was overloaded by 25% at 100%
boiler load when compared with the original design. During the shut-
down in May, the rectifier level installed in the reactor was equip-
ped with a catalytic converter level which consisted of the catalyst •
bars which were removed in December and split into halves over their
length. With this measure, the remaining residual activity of the re-
moved catalytic converters was utilized and the activity of the reac-
tor increased with regard to the overload operation. The success of
this measure manifests itself in the decrease of the NH3 concentra-
tion in June. The activity decrease of the catalytic converter, in
turn, can be discerned by the rising NH3 concentration in the flue
dust. Visual inspections of the catalytic converter level used as a;
rectifier level have justified this risky enterprise. The erosion
effects on the "rectifier catalytic converters" are extremely small.
In November, the ducts to the newly installed secondaiy reactor were
opened, thus reaching the ultimate stage. The first and second levels
of the secondary reactor are equipped with the remaining catalytic
converters from the third layer of the main reactor which was removed

6A-25

-------
in December 1987, and the third level with new catalytic converter
modules. The success of this measure manifests itself in the NH3 con-
centration values of the flue dust since December 1988.

Fig. 13 presents a summary of the effects of the different measures
on the available relative overall activity of the catalytic conver-
ter, referred to the initial activity and inner specific surface. In
order to understand the relationship between the measure taken and
the influence on the NHa flue dust concentration better, the ordinate
was drawn at the same scale as in Fig. 11 on top. The figure shows:
at the end of 1987 activity increase by new catalytic converter layer
with larger inner surface; in May 1988 increase by installation of
half the used catalytic converter layer; in November 1988 increase by
opening the secondary catalytic converter.

During the activity tests on the catalytic converter modules, a
strong divergence was found in the results of the different test in-
stitutes (Fig. 12, at 11,200 hours). Standard tests with 300 mm long
sections from the start and end of the 1 m long catalytic converter
bar give different results than the test of f.ive 20 cm long sections.
Pig. 14 illustrates this relationship and indicates that the proce-
dure for the activity measurement must be defined very carefully'- in
particular because it is a matter of guaranteed values.

The operating experience gained from the 18,400 operating hours of
the catalytic converters at Unit 5 and the resultant measures can be
summarized as follows:

- The activity decrease of the catalytic converters is smaller than
the value that formed the basis of the design.

- No erosion problems arose at the catalytic converters; therefore,
the web thickness was reduced during the first catalytic converter
replacement.

- The replacement of the rectifier level by catalytic converter
modules did not result in an additional erosion problem with the
dust load occurring and the given flow conditions; therefore, the
rectifier level can be eliminated.

Comprehensive measurements on the formation of S2O gave values of

less than 5 mg/m3 upstream of the catalytic converter and 3 nig/m3

downstream of it. The effect of a reduction of the NzO concentration

downstream of the catalytic converter could be proven in both DeNOx
units.

3.3.2 Catalytic Converter Behavior in Unit 2

Unit 2 of the Walheim Power Station is equipped with a slag tap fired
boiler. Before the NO* reducing measures were introduced, the entire
flue dust caught in the E filter was fed back into the furnace, mol-
ten in the furnace body and drawn off from the furnace as granulated
material (Fig. 15). The coal contains a multitude of trace elements
which are released during the combustion process, among others As. As

6A-26

-------
is identified as a catalytic poison. Since As evaporates at the tem-
peratures in the boiler furnace body and is therefore not removed to-
gether with the granulated material, but condenses at the flue dust
in the flue gas path when the. temperatures decrease, the As is con-
centrated in the flue gas due to the flue dust.feedback to the boiler
in the case of slag tap fired boilers, As becomes a catalytic poison,
in particular, when it is in the gaseous pnase in the fiue gas. with
the high-dust arrangement of SCR systems directly downstream' of the.
boiler, there is a risk of the catalytic converter becoming desacti-
vated prematurely by the As. This clanger to the catalytic converters
can be countered in two different ways:

- Reducing the As concentration by interrupting the full flue dust
feedback

- Moving the SCR system to downstream of the flue gas desulphuriza-
tion plant.

In /8/, the installation of the SCR system downstream of flue gas
desulphurization plants is described in detail. At Unit 2 in the
Walheim power station, Neckarwerke decided on the high-dust arrange-
ment for the SCR system. This decision included the equipment for
flue dust removal. Detailed tests have shown that this measure allows
the As content of 4 mg As/m3 (gaseous) - which is very low in this
system - to be reduced down to 1 mg As/m3. The overall As concentra-
tions (gaseous and Bound to fine dust) are in the range of 365 mg
As/m3 without removal (corresponding to a concentration in the dust
of 4,900 mg As/kg), and 87 mg As/m3 (corresponding to 1,000 mg As/kg)
with removal. Further, it could be determined by measurements that
the highest possible removal during soot blowing is particularly ef-
fective for a reduction of the As content in the flue gas (As concen-
trations up to 2,000 mg As/m3). These measurements were verified by
As analyses of the flue dust deposits in the boiler (As concentra-
tions 5,000 up to max. 20,000 mg As/kg).

These measures (approx. 10% removal during normal operation; 100%
removal during soot blowing) apparently prevented the suspected rapid
catalytic converter poisoning. Fig. 16 shows the course of the acti-
vity reduction in comparison to the design values. Here, too, the
actual activity reduction is smaller than was assumed in the design.
It must be noted that - contrary' to the catalytic converters at Unit
5, the activity reduction at the catalytic converters in the Kalheim
power station is different at the various levels. Here, the activity
reduction on level 1 is clearly stronger than on levels 2 and 3; a
certain As catcher effect is attributed tc the first level.

3.4 _ Air Preheater Washing

In earlier reports /t,8/, only test washing cycles at minor sectors
of the air preheater were described; in 1988, 2 full washing cycles
of the air preheater became necessary due to the rising pressure
drop. During the washing operations, the Nils that was stored in the
air preheater over the operating period comes out again in concentra-
ted form. Air preheater sheets (easy to clean, middle and end layer

6A-27

-------
enameled) that are suitable for DeNGx are installed in both systems.
The installed air preheater blowing device is operated once a day.

Pig. 17 gives the course of the NH3 concentration during the indivi-
dual washing operations for the test and full washing cycles perform-
ed. The presentation of the test washes clearly shows that the NH.s
settles in the air preheater to an increasing extent over the operat-
ing period. The rapid decrease in concentration indicates that the
ammonium salts are easily soluble. For the full washes, the washing
cycles could no longer be distingu i s tied as strictly, but distinction
was only made into preliminary washing, main washing and rinsing.
Here, too, it can be seen that the waste waters with 3 high NK3
charge can be separated off at the beginning for disposal and that
the higher volume of washing water only has a small charge. Approx.
1200 to 1400 m3 of washing water were needed for the full washes.
The NH3 quantities contained could be estimated from the UH3 concen-
trations and washing water volumes. The quantities are given in Fig.
17 with the corresponding operating times. From these figures, the
deposit figures for the air preheater used in Chapter 3.2, Fig. 8,
can be estimated. Concerning the second full wash, it can be stated
that this was not caused by a defective DeNOx unit (relatively small
NH3 deposit), but that the fault must he blamed on the operation of
the calorifer or air preheater blowing device.

Analyses of the air preheater washing water resulted in the fact that
the ion balance between the sulphate ions and the cations found is
not balanced. This leads to the conclusion that a large part of the
SO3 anions contained in the flue gas is directly absorbed by the flue
dust.

The following measures were based on the operating experience with
the air preheater:

- Daily blowing of the air preheater, under careful observation of
the pressure drop and the function of the bloving device

- For necessary air preheater washes:

- Separation of the rinsing water into a batch with high iTHs concen-
tration and one with low NHa concentration

- Provision of appropriate storage volumes for collection of the
rinsing waters.

4. Summary

4.1 In the Altbach/Deizisau Power Station, Unit 5 (420 MVinet, boiler with
dry ash removal), the high-dust SCR DeNOx system was converted from a
partial-flow system into a full-flow system.

In the Walheim Power Station, Unit 2 (150 >!',¦;), also a high-dust SCR
system was installed downstream of a slag tap fired boiler. In the
extension of the Unit 5 system and in the new* building of Unit 2, the

6A-28

-------
operating experience gained with the partial-flow system in Unit 5,
i.e.; .

- finely adjustable Nils closing in an optimized flue gas flow,

- well designed long mixing sections with baffles and constant
pressure distributors,

- careful design of the flow upstream of the' catalytic converter,
without the formation of dust streaks,

was taken into account by design measures.

4.2 The NH.i charge of the flue dust increasing over the operating time:
has proven to be an important factor for operation of the power sta-
tion in connection with the DeNOx system. Based on this knowledge and
the strong dependence of the NOx concentration upon the coal types
used, the system of Unit 5 was extended into a full-flow system. For
this, a special catalytic converter replacement method was employed.
This measure allowed the NHa charge of the flue dust to be reduced to
values between 20 and 30 mg NHa/kg. The low flue dust concentration
of approx. 6 to 8 g/m3 at Unit 5 must be noted.

4.3 In December 198B, the catalytic converters at Unit 5 have been in
operation for 18,400 hours and those at Unit 2 for approx. 7,500
hours. The decrease in activity was lower in both cases than that
used as the basis for their design. The As poisoning of the catalytic
converter expected due to the slag tap furnace in Unit 2 did not oc-
cur, due to the removal of the fly ash (in particular during soot
blowing).

No erosion or blocking effects due to the fly ash were found in
either system.

4.4 After approx. 16,800 operating hours, the air preheator in Unit 5 had
to be washed for the first time. The deposits could be easily removed
with water. The first washing water batches had a high NHs content.

4.5 Altogether, it can be stated that SCR DeNOx systems can also lie oper-
ated at limit values of 200 mg N0x/m3, if the following is observed;

- Design requirements in the plant design;

- Correct interpretation of the supervision results; and

- Early taking of appropriate measures.

However, the efforts and measures are considerable.

6A-29

-------
List of References

/!/ Meeker, P. and J. Becker;

Auslegung, Anordnung unci Funkt.ion tier DeNQx-Anlage Block 5' ein-
schliefllich Betriebserfahrungen: VGB-Sondertagung "Neue Steinkohle-
blocke am Neckar", Stuttgart 1986; VGB Essen 1986

/2/ Necker, P.:

Operating Experience with the SCR DeNO*Plant in Unit' 5 of the Alt-
bach/Deizisau Power Station; Section No, 5A-Z; Joint Symposium on
Stationary Combustion NO* Control; EPRI/EPA, New Orleans, 1987

/3/ Necker, P.;

Weitere Erfahrungen mit der DeNOx-Anlage im Block 5, Kraftwerk Alt-
bach/Deizisau nach raehr als einjahriger Betriebszeit; Sammelband
VGB-Konferenz "Kraftwerk und Umwelt 1987"; VGB Essen 1987

/¦'»/ Reidick:

DeNOx-Anlagen, Mitteilungen der EVT Stuttgart 1986, and some Japanese
Sources, 1985 and 1986

/5/ Bujak, W and H. Gutberlet:

Untersuchungen zum Eintrag 'von NHa aus DeNOx-Anlagen in REA-Abwasser;
Sammelband VGB-Konferenz "Kraftwerk und Umwelt 1987"; VGB Essen 1987

/6/ Wurz, 0.:

Bericht NH3/N0x-Messungen am Block 5 der Neckarverke Elektrizitats
versorgungs-AG Esslingen, Dezember 1988

/7/ Lehmann:

Messungen an der DeNOx-Anlage Block 5; Studienarbeit; Institut fiir
Industriebetriebslehre und Industrielle Produktion; (miversitat
Karlsruhe; Karlsruhe 1989

/8/ Schonbucber, B.:

Reduction of Nitrogen Oxides from Coal-Fired Power Plants by Using
the SCR Process - Experiences in the FRG with Pilot and Commercial
DeNOx Plants; EPA/EPRI 1989 Joint Symposium on Stationary Combustion
Control; San Francisco 1989

/9/ Necker, P.:

Betriebserfahrungen mit der DeNOx-Anlage im Kraftwerk Altbaeh/Dei
zisau Block 5; VDI-GET-Iagung, Februar 1988, Hannover; VDI-Verlag
Essen

6A-30

-------
S hematic drawing of flue gas path

of the SCR De NOx system

Alt bach/Deizi sau Power Station, Unit 5

Fig.1

NO* tals NO!)

2300
2200
2 100

2000
1 900
1 800
1 700
1 600
1 500
1 UX
1 300
1 200
1100
1 000

coal from Hw Scoriand

0 l—i—

0 jO

Fig. 2

90 too 1%1
boiler load

WQihfi* Power Station, Unit 2 15D MWel

Influence of coal type on rre NQX
concentration at the boiler outlet

NQxtalsNO?)

(mg/m3* N.tr j
( S 02 I

1100
1000
90G
800
700
600
500
400
3CC
200

100
0

coal from Hie Ruhr
District

cool fram the
Sacrland

mixed cool

0 10 20 B0 W 50 60 7 0 80 90 100 (%]

boiler load

AWbach/Oeizisau Power Station,Unit 5 460MWel

6A-31

-------
800 1000 1200 1400

NOx inlet concentration (mg

Fig, 3

Seperation efficience by NOx outlet concentration 200mg/m3
partial flow system / full flow system

View„X"

varw east

SsM *»*« tfc

SSSSSl gy^l fTTHHTH
5mc*or- fccnow>z#r- QeNO* -

Byposs bypass reocfcr

nr--"
-------
mg/cbra

mi Kit v m u u m w xissxii m g m n % a

Kti m iq

NECKRRHERKE
RBT. UMWELT

OeNOx -Melip rcg r arr.m

NH3 measurements; test on

Dec.14,1988 Primary catalytic
converter

Kosssl : 33 X Kat; 33 's.
ETfl-Kit: 7 1 *

Sttichf,: .SB Sasrkoh\s .
R™ .SSL S-* .45 (56^) ung.
Pngaben in mg/cdm i . N .

Fig.6

1 mq/cbm

N2 . 2

N2 , 3

: NECKRRWERKE

RBT. UfiWELT

DeNOx—Mefip roo rmmm

NH3 measurements; test on
Dec.14,1988 Secondary
catalytic converter

Kes38 ! ; 93 Kat: 93 *

ETP-K|t: 71 X

Sttich'f . : , S & SiarkoHle

5?- '.151 S- . 0567 c 3 7 >:) ung

Rngiben in mg >' cbff> > i , N .

Fig, 7

6A-33

-------
nh3

an* preheat'W'

D*NO«
(¦•actor

100%

electro 5 hi'ic prccijMfafor

10%

Ot SO* - Plant

35% • 15%

€o®bust,ien air

SQ%t W%

Fig. 8

Balance NHj leakage

IQQl

BD-
S? ?D"

f &o-

o SO-
&>

40-
w

* 30-
20-
10-

0eS0*-PI V 'jurwM ' DtcSI ' Jonas ' Dec 68 1
seriss 4 $9fm% 6 series? swesfl series 9 series 1Q

Fig, 9

NH3 deposit on flue dusf.m per cent of the NH3 leakage

1i« U«« «,« •<>» 14« 17« Tint

Fig, 10

NH3 concentration in flue dust over the daily
operating period with constant boiler load

6A-34

-------
400
JSC

360
3<0
JJO
300
280

^50

e220

"2Q0
*

gl 80

clS2

i 4Q
1 20
i 00
SO

so
40

frc

ft*

30f

1 ar

got

)bs

¦or

lin

iir

—

)0r

J0-,

i/t

<91

I
*

A
M

X-.

ft*

jfc

g*
*

¦w'

gjj

*?%r

-4

1 -

L-*
*

«Q
u.

-»

a.
u

O-

a
2

z ' c
< UJ

t£
<

Q
Z

z
<

-5

UJ
u.

3
<

0
z

0! /KU- 10-01/'

WElLUNG UHWELT

NH3 concentration in flue dust over the
operating period of the SCR QeNQv system

I load > 90 % )

Fig, 11

Average specific catalytic converter activity
Overall reactor. Unit 5

Fig.13

(referred to catalytic converter at outtot I

6A-35

-------
K/K0

• **

u
c

0
D

a
£

~.
l!

y
Q

D
D

r
c

5
£

¦ ?

ISJ

i
<

0
>

0 1000 2000 3000 4000 5000 6000 7000 6000 9000 10000 11000 12000 13000 U0Q0 15000 16000 17000

1580 3600 5600 3200 11200 18000

17 200

operating hours

Altbach-Deizisau Power Station,Unit 5

Results of catalytic converter activity measurements

Fig 12

0,1-

o' 1 — i 1 1 1 ""

o 0,2 0,4 0,6 0,8 1,0

Results of catalytic converter activity measurements over the bar length
Fig. 14 Altbach/Deizisau Power Station, Unit 5

6A-36

-------
electrostatic

boiler precipitator chimney

slag separation

Flue dust recycling during operation
Fig. 15 of slag tap fired boilers

0 1000 2000 3000 iOOO 5000 6000 7000 MOO 9000 10000 »0C0 12000 13000 * 000 15000 16000 17000
4500

m— operating hours

Walheim Power Station, Unit 2

Results of catalytic converter activity measurements

Fig. 16

6A-37

-------
cn

tbD-

laa-
3Uh
"JOO
?6G-
220-

operaHon hour 5609

oparatton hour 8200

Eh iso-
£

o*
E

sto-
ma-

h!>-

J ti-
ll-

Dec.1986

Jin

fcftQ-
I./0
)W~
ikO-
IUO

260-
22&-
iao-
KO-

60-

s-.fjl «Y» tp

1 2 3 4 5 6 7 ft 9 10
flushing

operation hour 14 200

June 1988

fulS washing cycte

HoEL

4.60-

UQ-

380-

3t 0

300-

no-

\m-

HO-

100-

20-

June 1987

1 2 3

f^Lrji rp

».60-

—

U0-

K0-

1DB-

2bQ-

n 0-

0* """

6,u»n

10a-

20-

operation hour 11 200

Dec 1987

QJlaa

T 1 *—* j • | "|—'1* «[• 'I" ¦|'-*T

1. 2 3 I 5 6 7 8 9 10 11 12. 13

flushing

WU-i(Mdf*li Of U» tlr pr«&*|t«r

*?0 W ^linl with m «lr.grtftt*Ur at l
-------
RECENT DEVELOPMENTS IN THE SCR SYSTEM AND
ITS OPERATIONAL EXPERIENCES

H. Kuroda

I. Morita
T. Murataka

Kure Works of Babcock-Hitachi K.K.

Kure, Hiroshima, 737, Japan
Or. F. Nakajima
Dr. Y. Kato

Kure Research Laboratory of Babcock-Hitachi K.K.

Kure, Hiroshima, 737, Japan
Dr. A. Kato
Hitachi Research Laboratory of Hitachi Ltd.

Hitachi, Ibaragi , 319-12, Japan

ABSTRACT

In the previous EPRI/EPA Joint Symposia held in Denver, Colorado in 1978 and 1980,
we were honored to present our NOx removal system.

Since then, regulations, for NOx emissions have become more stringent in the light
of growing industrialization and its environs, bringing about great numbers of
Selective Catalytic Reduction (SCR) systems in Japan, Europe and the USA, Having
been awarded 30 plants for gas firing, 27 for oil firing, 20 for coal firing and 5
miscellaneous plants, approx. 80 % of which were awarded within the last 8 years.
These SCR plants were not only for dry bottom boilers but also wet bottom boilers,
fluidized bed boilers, . heat recovery steam generators in cogeneration systems,
diesel engines, iron & steel plants etc.

At this 1989 Joint Symposium we wish to introduce the progress and development of
technology on SCR systems during these 8 years and then briefly present one of the
newly developed technologies, namely the catalyst resistant to Arsenic.

6A-39

-------
INTRODUCTION

In Japan stringent Ambient Air Quality Standards and Emission Regulations were put
in force in the 1970s because air pollution went from bad to worse due to rapid
expansion of industrialization.

The emission regulations stipulated for the large stationary sources, such as big.
factories and thermal power stations, were so stringent that the pollution control
systems have been installed one after another in both retrofit and new plants.
The SCR system were firstly developed in Japan, after extensive research and put
into worldwide commercial usage.

In the other industrialized countries such as the Federal Republic of Germany
(FRG), the USA etc. the air pollution situation has become serious and many SCR
systems came to be installed as the emission regulations tightened in such
countries.

In the USA the first demonstration plant of a SCR system for a boiler plant was
installed in 1982. However, other installations in boiler plants did not follow
simultaneously.

Several years later the increasing demand for SCR systems started in a different
field from that of conventional boilers.

Owing to PURPA (the Public Utility Regulatory Policies Act) the construction of a
cogeneration plant which produces electricity and steam for multipurpose utiliza-
tion instead of conventional boilers has become commonplace.

Moreover, many SCR systems for cogeneration projects have been taken up in the
state of California, where NOx emission regulations appear to be particularly
stringent in the USA.

In FRG when the NOx emission regulation was stipulated, the installation of SCR
systems started with coal fired boiler plants, because coal is the predominant fuel
in thermal power plants.

Before its installation a delegation{13 people) from the state of Baden Wurttemberg
first visited Japan in October, 1983 for an investigation of SCR systems and a lot
of delegations followed, resulting in certain license or cooperation agreements
between the companies in both countries.

In general the SCR technology developed in Japan has been applied in the FRG
without much difficulty. Only one matter which we had not experienced in Japan and
had to develop was the NOx removal from flue gas from a slag tap boiler, a kind of
wet bottom boiler.

Afterwards, this will be described more as an example of one of our developments.

SA-40

-------
HISTORY OF DEVELOPMENT AND OPERATIONAL EXPERIENCES

History of development

Fig. 1 shows the development of the Babcock-Hitachi DeNOx Process.

We commenced studies for NOx removal from exhaust gases in 1963, conducted hundreds
of catalyst screening tests, life tests and optimum process studies, then began to
Install pilot plants in the field in 1973. At this early stage of development we
had conducted a great deal of pilot tests together with laboratory tests.

As we already introduced the history of our early development at the previous
EPRI/EPA Joiny Symposia (1), (2) held in Denver in 1978 and 1980, we hereby
describe the further development since then.

The needs of SCR plants for retrofit utility boiler plants have rapidly increased
in Japan and almost all SCR plants for retrofit boilers were installed during the
first half of 1980s. Since then its demands are mainly for new utility boiler-
plants.

In the USA the first SCR system for a cogeneration plant which was supplied by
Babcock & Wilcox Co. under the license of Babcock-Hitachi K.K. has been
successfully operating since early 1986.

In Europe we have conducted about 20 pilot tests in FRS, Austria, Denmark and
Sweden since 1984 in cooperation with our licensees, Uhde/Lentjes, Babcock Energy
Ltd., Flakt, and Voest Alpine as a consortium partner through which our catalyst
and SCR systems were demonstrated and proven applicable for commercial use.
However, while the SCR systems developed in Japan were applied to boiler plants in
FRG, we encountered one problem, which had not been experienced in Japan.

Rapid deactivation of DeNOx catalysts was observed only in a slag tap boiler with
100 % ash recirculation which was not installed in Japan, and it became apparent
that it was strongly related to the type of firing carried out.

The new catalyst was developed and tested in the pilot plants, resulting in the
affirmation of its durability and its reliability for commercial use.

During the last 8 years we have increased the application of our SCR systems to
various kinds of plants not only in Japan, but also Europe and the USA.

Me are confident that if unexpected problems should occur in the future, our expe-
riences and accumulated know how will enable us to correct any conceivable problems.

Operational experience

Having been awarded 30 plants for gas firing, 27 for oil firing, 28 for coal

firing and 5 miscellaneous plants. Among these plants, in the field of utility

operating are 10 plants for gas firing, 18 for oil firing and 19 for coal firing

6A-41

-------
as of January 1989 (Others in this field are under design or construction).

Typical operational experience (utility) of SCR systems for three types of fuel
gas, oil, and coal is summarized in Fig. 2.

It can be seen generally that the DeNQx catalyst deteriorates comparatively quickly
during short periods from the start up, then stabilizes. Although the catalyst
life is greatly dependent upon its operational conditions, the longest operation
without catalyst replacement or addition among the above mentioned plants is about
9 years for gas firing, about 8 years for oil firing and about 6 years for coal
firing,

Countermeasures to Deterioration of Catalyst

The deterioration of the catalyst due to aging is inevitable over a long period of
time in the operation of a SCR system, therefore preventative countermeasures are
an important subject. However the deterioration rate of the catalyst is relatively
slow, and there are economical means to make most of the remaining catalyst
activity.

The dependence of the volume of catalyst required with operating time is compared
for the additional method and the replacement method in Fig. 3.

It is clear that the addition method is more advantageous, as it utilizes the
remaining activity of the catalyst more effectively.

The plate type catalyst is particularly advantageous if the addition method is
used, because additional catalyst units can be stacked directly on the initially
loaded catalyst in the vertical gas flow arrangement (4).

The first commercial SCR system treating whole flue gas from the boiler was commis-
sioned in 1978 at Chita No. 5 and No. 6 of Chube Electric Power Co. (respectively
700 MW gas firing).

They had operated for approx. 9 years without a catalyst addition or replacement.
Then, the partial replacement followed.

Concerning SCR systems for coal fired boilers, the largest SCR plant for coal fired
boilers in Japan was commissioned in 1983 at Takehara Power Station No. 3 of
Electric Power Development Company (700 MW) and in 1986 at Duernrohr Power Station
No. 1' and NO. 2 of VKG/EVN in Austria (405 MW and 352 MW respectively) went into
commercial operation.

As for combined cycle/eogeneraton plants the first SCR plant was commissioned in
1981 at Kawasaki P.S. of Japanese Railway (141 MW). It operated for about 8 years
without a replacement or addition of catalyst (3), (5).

Thereafter, a replacement was carried out and it has continued operating.

6A-42

-------
The first NOx removal system for a cogeneration plant in the USA was commissioned
in early 1986 at Willammette Industries Inc. in California in the USA (8).

It has been operating satisfactorily for about 25,000 hours since its start up.
We would like to introduce photographs with some specifications of the above men-
tioned plants.

• Fig. 4 : SCR plants at Chita No. 5 and No. 6 of Chubu E.P.Co.

700 MW gas firing

First plant treating whole gas exhausted from Japanese
utility boilers

• Fig. 5 : SCR plant at Takehara No. 3 of Electric Power Development

Company (EPDC) 700 MW coal firing
The largest coal fired plant in Japan

• Fig. 6 : SCR plants at Ouernrohr No. 1 and No. 2 of VKG/EVN

405 MW and 352 MW coal firing

First plants treating whole exhaust gas from boilers in
Europe

a Fig. 7 : SCR plant at Kawasaki No. 1 of Japanese Railway 141 MW
combined cycle plant

First plant in the world treating exhaust gas from a com-
bined cycle

• Fig. 8 ; SCR plant at Willammette Industries Inc. in California

22 MW cogeneration plant

First plant treating exhaust gas from cogeneration plant
in the USA

Next, the catalyst resistant to As which is newly developed is outlined.

Mechanism of Catalyst Poisoning by As

Since 1984, we conducted about 20 pilot tests in Europe as described before for the
demonstration of the applicability of the catalyst developed by us. . During these
pilot tests, we found a certain decrease in activity of the catalyst in a few
plants. After extensive investigations, it was finally discovered that the trace
element Arsenic (As) in flue gas caused its deterioration.

Fig. 9 shows representative results, DeNOx efficiency and the concentration of As
on the surface of the catalyst, obtained in our laboratory from the analyses of
the sample catalysts taken out from the pilot plants after 6,000 - 9,000 hours
operation. In this figure, DeNOx efficiency was measured under our standard test
conditions.

It is apparent that in the case whereby the catalyst was used in the flue gas from
the wet bottom boiler with 100 % ash recirculation, a considerable decrease in

6A-43

-------
activity of the catalyst occurs simultaneously with a rapid increase of As.
The research work such as the analysis of elements adhered to the DeNOx catalyst
was done to find the causes and mechanisms of the deterioration of the catalyst in
our research laboratories.

Fig. 10 illustrates the general concept of catalyst poisoning by As which was
obtained from the research.

In the case whereby As is not contained in the flue gas, NH3 is adsorbed on the
active site of the catalyst and then reacts with NOx to form Ng and H2O.

However, when As exists in the flue gas, it is adsorbed on the catalyst active
site. Consequently, the adsorption of NH3 is obstructed, resulting in the deacti-
vation of the DeNOx catalyst.

Enrichment of As in the Boiler System

According to Ray et al ' s paper (6), the trace elements contained in coal are
roughly classified into the following three groups depending upon their behavior as
shown in Fig. 11.

« Group A : Concentration in bottom ash is almost the same as that in
fly ash

• Group B ; Selectively condensed on fly ash

• Group C : Emitted to stack as gaseous phase

Among the above three groups, the trace elements in Group B such as 6a, Cu, Pb, As
etc. take upon a special behavior in the coal fired wet bottom boilers with ash
recirculation.

Fig. 12 shows the comparison of the behavior of the trace element As "without ash
recirculation" and "with ash recirculation".

Because of the condensation effect of As, the total amount of As in the flue gas
from the boiler with 100 % ash recirculation becomes about 8 to 11 times higher
than that without ash recirculation. And according to Gutberlet (7), the gaseous
concentration of As in the flue gas before an air preheater becomes several hundred
ug/Nm3 in the case of 100 % ash recirculation plants, although it is about 5-20
ug/Nm^ for the wet bottom boilers without ash recirculation and less than 1 yg/Nm^
for the dry bottom boilers.

This condensation effect is closely related to catalyst deterioration by As
described above.

Development of New Catalyst Resistant to As

Through careful investigations and painstaking studies it was found that the dete-
rioration of the catalyst is caused by the accumulation of As on the catalyst and

6A-44

-------
the development of New Catalyst resistant to As started to solve this problem.
Some of the approaches for the solution were shown in Fig. 13.

Firstly we tried to reduce the pore-sizes of catalyst

When the As compound in flue gas is AS4O5, its diameter would be approx. 7-8 A,
larger than that of NO and NH3, approx. 3-4 ft. If the size of pore-diameter is
controlled approx. 5-6 ft with technology of catalyst production, As compound is
shut off from the pores, which will result in the prevention of deterioration by
As.

Another approach was the investigation and study for the ingredients which have
less adsorption ability of As.

As is adsorbed on the active site of NH3 and obscures the NH3 adsorption. If the
affinity of active site with NH3 is stronger than As, the catalyst activity will be
able to be kept high.

Furthermore, we investigated the improvement of active site by the special
treatment in order to protect the active site from As.

Several hundreds of catalysts were produced, and tested for the evaluation of
resistance to As in the laboratory. Then the selected catalysts were installed in
the pilot plants for the demonstration of their applicability. Finally, we came to
the conclusion that the TiOg based catalyst in which the active site is improved by
the special treatment is the most effective countermeasures.

Fig. 14 shows the test results of the New Catalysts developed for wet bottom which
were.taken out from the pilot plants in comparison with the ordinary catalysts for
dry bottom. The results were satisfactory and the New Catalyst durability was
ascertained.

The New Catalyst have already supplied to the commercial plants.

SUMMARY

1. For the last 8 years in the 1980s the application of our SCR systems has been
expanded from dry bottom boilers to various kinds of plants such as wet bottom
boilers, fluidized bed boilers, heat recovery steam generators, diesel engines,
iron and steel plants etc., and not only in Japan, but also Europe and the USA.

2. In FRG we experienced a certain decrease in activity of the catalyst due to As
adsorption in a few plants which were operated in slag tap boilers with 100 %
ash recirculation.

3. The investigation and studies for a catalyst resistant to As were sucessfully
performed and we developed a New Catalyst for Commercial utilization.

6A-45

-------
4. Me are confident that if unexpected problems should occur in the future, our
experiences and accumulated know how will enable us to correct any conceivable
problems.

ACKNOWLEDGMENT

This paper contains data which were obtained in cooperation with the domestic or
overseas electric power companies and with our lincensees of this process.
The authors wish to express their thanks to them.

REFERENCES

(1) H. Kuroda, F. Nakajima "Some Experience of NOx Removal in Pilots and Utility
Boilers" EPRI FP-1109-SR Semi nor proceedings, July 1979.

(2) T. Narita, H. Kuroda et al. "Babcack-Hitachi NOx Removal Process for Flue Gases
form Coal-fired Boilers" IERL-RTP-1084

The EPRI/EPA joint Symposium on Stationary Combustion NOx Control Denver
Colorado USA, Oct., 1980

(3) T. Narita, T. Kumura "DeNOx Equipment of Combined Cycle Power Plant" CI MAC
1983.

(4) H. Kuroda, T. Masai "Babcock Hitachi NOx Abatement Technology" NOx Symposium
Karlsruhe, West Germany Feb., 1986.

(5) H. Kuroda, T. Ishikawa, T. Murataka "Latest DeNOx System for Combined Cycle
Plant" CIMAC 1987.

(6) Ray, S.S. and Parker, F.G. : Characterization of Ash From Coal-Fired Power
Plants. Springfield, VA : National Technical Inforamtion Services, January
1977.

(7) H. Gutberlet "Influence of Furnace Type on Poisoning of DeNOx Catalyst by
Arsenic" VGB Kraftwerks Technik Vol. 68 No. 3 March 1988 (English Issue).

(8) M.G. Radin, B. Boyles "Turbine Exhaust Gas DeNOx Using Selective Catalytic
Reduction" American Power Conference, April 1987.

6A-46

-------
an
>

SELECTIVE CATALYTIC
REDUCTION

H 63 S W3 I 74 I 75 I 76 1 77 | 78 [ 79 | 80 | 81 I 82 1 83 I 84 I 85 I 86 1 87 I 88 I 89 I 90 1

R&D

NON-CATALYTIC
REDUCTION

NH, INJECTION

SIMULTANEOUS
DeSOx/DeNOx

WET ABSORPTION

LAB- TEST

LAB, TEST

PILOT
COMMERCIAL

)

Fig. 1

DEVELOPMENT OF BABCOCK-HITACHI DeNOx PROCESS

-------
>
H

I-
U
<

1.2
1.0

0.8
0.6

LNG

o A- 1 PLANT
• A - 2

• 11 i 1

o B PLANT

3 1-2

o.sr*

0.6

COAL/HIGH DUST SYSTEM

-•o-

-o-

O c -1 PLANT
• C - 2

I—

< 0.6

1.2
1.0

0.8

COAL/LOW DUST SYSTEM

-9* Qa

o D-1 PLANT

• D - 2

* D - 3

•uo

OPERATING TIME (YEAR)

Fig. 2 CATALYST ACTIVITY WITH OPERATING TIME

6A-48

-------
SPARE SPACE

REACTOR

CATALYST
BLOCK

Oos

m > ¦£.

5 «t 2

OH h-
uj « <
KUC

2001
0

FIRST
(17%)

SECOND
*

THIRD
9

50%

-a * a

OPERATING TIME (YEAR)

ADDITION METHOD

CATA. TO
BE ADDED

CATA, ORIGINALLY
INSTALLED

LU — O-

C£ ^

5^0 400

2^'S 200

CE u Q£

w <

< 0E

ripCT

(1GQ%)SECONDTHIRD fourth
V 9 v v

^-r- ^

OPERATING TIME (YEAR)
INSTALLED REPLACEMENT METHOD

r 400%

CATA. TO
BE REPLACED

Fig. 3 COMPARISON OF THE ADDITION METHOD WITH THE REPLACEMENT METHOD

6A-49

-------
11
10

u. £

o 5

-------
Btproduced from
best available copy.

CUSTOMER

BOILER FUEL

PROCESS

GAS QUANTITY

INLET NO* CONC.
DeNOx EFF,
OPERATION

CHUBU ELECTRIC POWER CO.
CHITA P.S. No.5 & No,6 UNITS

LNG

SELECTIVE CATALYTIC REDUCTION

1,910.000 Nm3/h
(EQV. 700 MW)

50 ppm(S% 02)

80 % or MORE

MAR. 1978 (No.5!

MAY. 1978 {No.6)

| Fig. 4 SCR PLANTS AT CHITA P.S.
OF CHUBU E.P.C0.

CUSTOMER

BOILER FUEL

PROCESS

GAS QUANTITY

INLET NO* CONC,
DeNOx EFF.
OPERATION

E.P.D.C,

TAKEHAftA P.S, NO.3 UNIT
COAL

SELECTIVE CATALYTIC REDUCTION

2,320,000 Nm3/h
(EQV, 700 MW)

300 ppm

80 % or MORE

MAR. 1983

Fig. 5 SCR PLANT AT TAKEHARA P.S.
¦ OF E.P.D.C.

6A-51

-------
CUSTOMER

BOILER FUEL
PROCESS

GAS QUANTITY

INLET NOx CONC.

DeNOx EFF.
OPERATION

VKG

DUERNROHR P.S.
No.1 UNIT
COAL

SELECTIVE CATALYTIC
REDUCTION

1,210,000 Nm3/h
(EQV. 405 MW)

800 mg/Nm3 (6% Oj)
(390 PPM)

80 % or MORE

MAR. 1987

EVN

DUERNROHR P.S.
No.2 UNIT
COAL

SELECTIVE CATALYTIC
REDUCTION

1,038,000 Nm3/H
(EQV. 352 MW)

650 mg/Nm3 (6% 02)
(317 PPM)

80 % • or MORE

MAR. 1987

Fig. 6 SCR PLANTS AT DUERNROHR P.S. OF VKG & EVN

6A-52

-------
CUSTOMER
FUEL
PROCESS
GAS QUANTITY

DeNOx EfF.
OPERATION

JAPANESE RAILWAY
KAWASAKI P.S. No.1 UNIT
KEROSENE

SELECTIVE CATALYTIC REDUCTION

1,024,000 Nm3/h
IrQv 141 MW;MS9001B)

80 % or MORE

FEB.1981

Fig. 7 SCR PLANT AT KAWASAKI P.S,
OF JAPANESE RAILWAY

CUSTOMER
FUEL

PROCESS

GAS QUANTITY

INLET NO* CONC-
DeNO* EFF.
OPERATION

WILLAMETTE INDUSTRIES, INC
NATURAL GAS

SELECTIVE CATALYTIC REDUCTION

195,000 Nm3/h

(EQV. 22 MW , GE LM5000)

75 ppm (15% O2!

80% or MORE

MAY. 1986

Fig. 8 SCR PLANT AT WILLAMETTE
INDUSTRIES INC.

6A-53

-------
GROUP A

CONCENTRATION

IN BOTTOM ASH = CONCENTRATION IN FLY ASH

Ce Fe La Rb Sm Th

Co Hf Mg Sc Sr Ti

Eu K Mil Si Ta

GROUP B

SELECTIVELY CONDENSED ON FLY ASH

As Ga Sb
Cd Mo SE
Cu Pb Zn

GROUP C

EMITTED TO STACK AS GASEOUS PHASE
Hg CI Br

Fiy. 11 BEHAVIOR OF THE TRACE ELEMENTS IN COAL

WITHOUT ASH RECIRCULATION

WITH ASH RECIRCULATION

Fig. 12 ENRICHMENT Of ARSENIC (As)

6A-54

-------
REDUCTION OF PORE SIZE OF
CATALYST

CATALYST RESISTANT TO As

PREVENTION OF As-
ADSORPTION TO ACTIVE SITE

INCREASE OF CATALYST
ACTIVITY

NEW SUBSTRATE OTHER
THAN Ti02

NEW ACTIVE COMPOUND
RESISTANT TO As

Fig. 13 APPROACHES FOR THE DEVELOPMENT OF
CATALYST RESISTANT TO As

K = AVx Ln 1/(1 -n)}

AV ; AREA VELOCITY (m/h)
n; DeNOx EFFICIENCY {-)

OPERATING HOURS (h)

. 14 OPERATING HOURS VS K IN WET BOTTOM BOILER (EXAMPLE)

6A-55

-------
6A-56

-------
THE FIRST DE-NOx INSTALLATION IN THE NETHERLANDS
A DEMONSTRATION PROJECT AT EPON - NIJMEGEN. POWER STATION

Ir. J.M. Koppius-Odink
N.V. EPON
8025 AZ ZWOLLE
The Netherlands

Ir. W.L.C. Weier

N.V. KEMA
6812 AR ARNHEM
The Netherlands

Dr. Ir. W.L. Prins
ESTS BV
1970 AL. VELSEN-NOORD
The Netherlands

ABSTRACT

In order to demonstrate the feasibility of the Japanese SCR technology in 1985 a
demonstration installation has been installed by ESTS BV for denitrification of
the fluegases from the coal fired boiler No. 12 at EPON, Centrale Gelderland,
Nijmegen. The DeNox plant handles 50% of the fluegases. This national demo
project is financed by the Government through the N.V. NOVEM. Initial start-up
was in 1987.

From this demo project experience is reported on:

* Operating performance of the installation, performance of the individual
catalyst layers, experience on catalyst consumption and inspection results.

* Consequences of ammonia slip: experience with possible fouling of air
preheater, E-filter, ducts and fan. Determining the effect of residual NH3
in the flyash application.

* Operating experience with:

NHj injection: control quality, distribution of NH^ injection over the

catalyst, fouling of ammonia-injection nozzles;

a specially developed by-pass gas mixer;

plant instrumentation for monitoring NH3, N0X and SO,;

soot blowing;

maintenance.

t^ishexperience^wflfaSInSaSifeien (19895 an eeansraie evaluation based on

6A-57

-------
I

INTRODUCTION

EPON installed at the location Nijmegen (The Netherlands) a demonstration
installation for denitrification of the flue gases from the coal fired boiler
No. 12. This project was financed by the Government, through the Netherlands
Agency- for Energy and the Environment (NOVEM), in a joint effort by the The
Ministry of Environmental Affairs and The Ministry of Economic Affairs.
The purpose of this national demo project is to demonstrate the feasibility of
the Japanese SCR technology in the Netherlands. The plant has been constructed
by ESTS BV in 1986. KEMA, the joint organisation of the Dutch Utilities, is
primarily involved in the execution of the measurements. In the future this
project will give a contribution to a practical construction, operation and
exploitation of this kind of technology in the Netherlands.

After explanation of the targets of the project and a description of the denox
system, experience will be reported going step by step through the denox instal-
lation:

eco bypass, temperature control

{bypass} fluegas mixer

ammonia supply, injection and mixing

reactor and catalyst

soot blowing

flyash transport system

down stream equipment: air preheater, E.S.P. and I.D. Fan
fly ash application

control system and monitoring equipment

maintenance

conclusions

THE DEMO PROJECT

Relevant items for demonstration in this project are:

* ¦ Operating"performance of the installation, performance of the individual

catalyst layers, experience on catalyst consumption and inspection results,

* Consequences of ammonia slip: experience with possible fouling of air
preheater, E-filter, ducts and fan. Determining the effect of residual NH^ in
the fly ash application.

*" Operating experience with:

NH3 injection: control quality, distribution of NH^ injection over the
catalyst cross-section, fouling of ammonia-injection nozzles;
a specially developed by-pass gas mixer;
plant instrumentation for monitoring NH3, N0x and S03;

~ soot blowing;
maintenance.

At the end of this demonstration project (1989) an economic evaluation based on

this experience will be made.

SA-58

-------
THE DENOX INSTALLATION

The retrofit denox installation is of the high dust, high temperature, type.
The main data for the Electricity Generation Unit are:

Utility
Unit

Location
MW el

Type of boiler
Fuel type

EPON, Zwolle

Gelderland, Unit 12, in operation since 1966
Nijmegen, The Netherlands
125 MW
U-firing

imported bituminous coal (heavy fuel oil)

The data for the SCR plant are:

Fluegas flow ;
Operating fluegas

temperature :

NOx content (inlet) ;

Removal efficiency :

NH3 slip :

Load variation :

Catalyst type :
Catalyst manufacturer:
Number of catalyst

layers :

Dummy layer :
Total volume catalyst:

L x B x H of reactor :

Guarantee period :

210,000 m3/h (502 of the total fluegas flow)

320-400"C (Eco outlet)

509 ppm vol. at 3% 02

> 80%

< 5 PPm
40 - 100£

Honeycomb
MHI

95 m3

9,1 x 5,1 x 25 m

2 years

In the design of the system a special problem was the little place available for
the reactor. The space between the E.S.P's has been utilised.

Only one of the two fluegas streams {50%) is treated by the denox installation.
This makes it possible to compare flyash from the denox plant and similar flyash
from the stream without denox. A separate flyash silo for the denox flyash was
installed'. In this way both flyash sorts can be tested for applications.

In order to compensate the additional pressure drop from the reactor a new induced
draft Fan has been installed.

Figure 1 gives an overview on the installation. Figure 2 gives a scheme of the
system,

After the economiser-the fluegas stream is split into two ducts leading to the two
air preheaters. In one of the two ducts the SCR reactor is installed. A bypass
over the economiser is installed with a- fluegas flow control valve. In this way
the fluegas temperature in the reactor can be controlled. The eco-bypassed
fluegases are mixed in a specially, for this purpose developed, mixer.

ECONOMISER BYPASS; TEMPERATURE CONTROL OF INLET FLUEGAS

After start-up in December 1986 it appeared that the inlet temperature was too

lew to allow the Denox-r^aetgr fep epeeafee with ammonia Injfeelien. To prevent

aaunoniumbisulrate formation in tne catalyst the minimum continuous inlet tempera-
ture at design S03 level should be 3(+0°C. Only short operation between 320-3^0°C
is allowed if it can be followed by "regeneration" by controlling the inlet

6A-59

-------
Even with open bypass, measurements showed that the inlet temperature could not
be raised sufficiently. Therefore early 1987 the economiser was modified, with
the result that inlet temperatures were obtained as described above.

In practice the eco-bypass is operated now always completely open. At partial load
(below 80>) however, the inlet temperature is still too low (below ^20°C). Ammonia
injection is stopped in that situation*

the bypass fluegas mixer

In order to get an uniform temperature distribution at the inlet of the reactor
and specially since lower temperatures are not allowed in the catalyst, a fluegas
mixer was installed to mix the appr. k50°C Eco-bypass flow with the 300"C Eco-
outlet flow• A specially for this application developed fluegas mixer (see
fig, 3) is installed.

The velocities of different adjacent layers of fluegases are approx. 22 a/s. and
5 xn/s. and are optimal for Eddy diffusivity with large turbulence. Good nixing
(i.e. rel. standard dev. £10%) is obtained after 3 x diameter of the duct.
Advantages of this mixer are:

1. Fluegas at high velocity is not deviated in velocity direction; the low
velocity fluegas is led around a bend.

2. The high .velocity gas "pulls" the low velocity, gas saving fan capacity.
The system is free from flyash accumulation.

AMMONIA SUPPLY, INJECTION AND MIXING

One NH3 storage tank is installed for this demo project, including an evaporator
vessel and buffer vessel at the beginning of the NH^ (vapour) transmission pipe
of approx.' 300 m. This system needed only minor attention like seal replacement
and control valve cleaning.

The ammonia injection system is shown in figure 4, The distribution system has
valves that can be used to influence the distribution in the fluegas. In case
of fluegas velocity gradient in the duct ammonia injection should be adjusted to
get homogenous concentrations of NH3 in the fluegas. Already very good distribut-
ion was obtained with all valves in open position.

The injection is done by means of 16 injection pipes, each having 16 nozzles.
During 11.000 operating hours only one pipe had to be cleaned.

In the fluegas duct ahead of the reactor a static mixer is installed to unifoi'iily
diffuse NH3 through the fluegas. The static mixer consists of a grid of vertical
pipes (for horizontal nixing) and horizontal pipes (for vertical mixing).

REACTOR AND CATALYST

The vertical flow reactor is formed by three catalyst layers, a dummy layer on top
and soot blowers cleaning the top side of the catalyst layers.

Important eharaeterisfcies fee.the.feasts? eoetjatlBn arei

homogeneous temperature distribution ana NH3 concentration

homogeneous gas velocity
catalyst activity

6A-60

-------
TEMPERATURE AND NH3 CONCENTRATION DISTRIBUTIONS

Homogeneous temperature and NH3-concentration distributions at the inlet of the
reactor are very important. Even with the precautions taken the temperature
distribution at the inlet of the reactor was just acceptable. At the boiler side
we found the lower temperatures, appro*, 30 C below the average' temperature.

This underlines the necessity to give much attention to the mixing of the bypass
flue gas stream with the main stream.

So far we could not obtain sufficient reliable values of the NH3-concentration.
In the meantime the measurement procedures are improved and we expect to get an
impression of the NH--concentration distribution at the inlet shortly.

GAS VELOCITY DISTRIBUTION

After start up of the boiler during inspections severe flyash build up was
established on the dummy layer at the boiler side, the inside of the corner of
the reactor inlet. See figure 5* In figure 6 fluegas flow simulation result is
given. It confirmes the observations of flyash build-up. Also the flow pattern
was confirmed visually on the walls of the reactor. See figure 7-

From figure 6 also the effect of the dummy layer can be seen. The first layer of
catalyst'obtains a homogeneous gasflow by the operation of the dummy layer. Main
purpose of the dummy layer is to direct the velocity of the flyash vertically.
In future installations guide vanes should be considered above the dummy layer in
an attempt to minimise the flyash build-up. In this installation static soot
blowers were installed above the dummy layer on the locations with flyash build-
up. They operate once a shift. As a result the flyash build up was decreased to
acceptable levels for catalyst operation.

CATALYST LAYERS, ACTIVITY

During 11.000 hours boiler operation almost no flyash build-up was observed on
top of the catalyst layers. Internal plugging of the honeycomb "channels" is
only slowly progressing and is now in total a few percent of the catalyst area.

The denox removal efficiency of the catalyst layers is of course mainly determined
by the molar ratio NH3/N0x at the inlet of the reactor. As is shown in figure 8-14
easily a N0V removal efficiencies was obtained of 85# with the catalyst at start
up. Long term operation at removal efficiencies far above 80% removal is not
allowed.

Potential concentration gradients of N0X and NH3, or gas velocity gradients may
cause NH, slip through the reactor, which leads to fouling of catalyst and
downstream equipment by ammonium bisulfate.

The space velocity at design flow in the reactor is 2225 h"1. The performance
after 2500 operating hours of the individual layers can be seen from; figures
8-11. The performance after 4500 hours is given in the figures 12-14.

From the figures can be seen that before the catalyst the NQx and NO distribution
is within the accuracy of the measurement method and the stability of the
soeratinR BaFBBefcers in the rsaetor. _

After the First catalyst layer the concentrations of N0X equals the NO-concen-
tration of meaning that all N02 is converted after one layer.

The conversion of N0x is almost complete after the first layer. The concentrations

6A-61

-------
It is planned to take 3 catalyst samples each year and to test these samples
accurately for activity in a bench-scale unit to "overcome" the inaccuracies in
the measurements in the denox-plant.

The total denox-efficiency was approx. 15% during these measurements, since the
control signal for NH,-supply from the NQX monitor at the inlet gave temporarily
a lower molar ratio NH3/N0X.

After the three catalyst layers there is a profile in N0x concentration {maximum
concentration deviation is approx. from the average).

AMMONIA SLIP

At the startup of the denox installation a low ammonia slip (smaller than 2 ppa)
was determined. Also on this item measurements after the start-up had such an
inaccuracy that until now no significant increase could be established.

SOOT BLOWING

Soot blowing of the catalyst layers appeared to be necessary in this installation.
The frequency is once per day. This can perhaps be reduced. Because of the flyash
build up on the dummy layer, this installation is not representative For soot
blowing in future installations. It may be that in future installations the
frequency of soot blowing can be reduced considerably.

The pressure drop over the reactor was maintained stable as shown in table 2.

FLYASH TRANSPORT SYSTEM

There were no special problems. Some wear of pipes for.pneumatic transport
occured.

DOWNSTREAM EQUIPMENT: AIR PREHEATER, E.S.P. AND I.D. FAN
AIR PREHEATER

Initially some very small white spots of a few mm's on the renewed lamella
packages were observed. But no increase in amount was established in later
inspections.

ELECTROSTATIC PRECIPITATOR

Until now visual inspection gave no difference in fouling behaviour in the EP's.
In both cases the electrodes and plates were covered by a homogeneous layer of
approx. 2 mm thickness.

I.D. FAN

On the blades,us daeeslbs were sbeepyedi Hewever¦ en fcha awia seme depesfsts wen
observed resulting in increase In unbalance. Installation or a soot Blower solvei

this problem.

6A-62

-------
FLYASH APPLICATION

In the Netherlands flyaah is used for several applications. If flyash is con-
taminated with ammonia, it could lead to problems when used for light-weight
aggregate production,, for fabrication of building stones and for production of
concrete.

See for a detailed analysis for the coal and flyash table 3- There is no sig-
nificant increase in NR, content in the flyash from the Denox system compared
to the flyash from the Fluegas without denox system. The relative short operating
period does not allow conclusions on this item (the catalyst did not drop
measurable in activity).

Tests planned to check the denox flyash in its application has therefore not yet
been carried out; first a measurable increase in NH4 + in flyash has to be
es t abl i she^Z.

In case the ammonia slip-remains at this low levels, a test is considered by
temporary increase of the NH2/NOx molar ratio to artificially establish an
ammonia slip of a few ppn {design is < 5 ppm) and then collect flyash for analysis
and possible testing in its applications.

THE CONTROL SYSTEM; MONITORING EQUIPMENT
BOILER CONTROL SYSTEM

At installation of the new induct draft fan with its higher operating pressure
head and different characteristic a new fan control system had to be incorporated
into the existing boiler control system.

Two turbine type fluegas flow meters were installed (see figure) at symmetrical
positions in the gasstream in the high dust environment. The original control
signal- (boiler pressure) is going directly to the existing inlet vane control
motor and via an arithmic function (a "curve) to the1 new inlet vane control motor.

The experience with this additional control system is that it was relative
sensitive. Specially when the boiler did go up in load after a stop the system
needed attention for complete stability. It must always be assured that the
instrument air supply to the insertion turbine meter bearings is on.

The turbine meters function well; only an electronic's failure has been
experienced until now.

NOx REMOVAL CONTROL SYSTEM

The fluegas flow signal shown in figure 15 forms together with the inlet N0X
concentration signal and the NH3/NOx molar ratio set at 0,8 the signal to the
NH3 feed control loop.

The outlet N0X concentration signal is (together with the inlet N0x signal) used
to derive a signal for N0x removal efficiency. This gives a correction to the
llficiency^of eettfcpel Isep in arder tee keep e eenafeanfe N0|-removal

Boiler load fluctuations and fluctuations of inlet N0X concentrations are
accounted for by this control loop, keeping the N0x removal at the required

6A-63

-------
Due to problems with the NO monitoring equipment (see below) an operation of this
control loop has not been possible during a major part of the operating period.
The NH3 flow control loop was operated manually. In this situation the dynamic's
of the control system could not be tested so far.

MONITORING EQUIPMENT

The analytical equipment is sampling in a high dust environment causing com-
plications.

NO equipment : In the infrared cells optics were fouled by As and Se,

together with flyash. Thus As and Se are bound to fine
flyash, therefore in the NO monitors after EP's this forms
no problem,

In the sampling line an EP and special filter have been
installed, improving the reliability of the system. Still a
systematic difference of 16-20"% with another method is
existing. It is being investigated.

SOj equipment : The installed equipment is not suitable for continuous

measurement.. Calibration is very time consuming. After
calibration the equipment is reliable for only a few hours.

NH3 equipment : The equipment {range 0-10 ppm NH-jS is with the present

catalyst activity operating near its detection limit.
Laboratory methods applied however, give not yet consistent
results.

MAINTENANCE

In figure 16 the maintenance hours by the electric/control department and the

mechanical department are shown per section of the Denox plant.

CONCLUSIONS

The main conslusions in this first part of the demo project are:

* The Denox system is operating well; catalyst activity reduction could not
yet be measured after 4500 hours.

* Only minor fouling was observed- in the downstream equipment after the Denox
reactor.

* For new Denox-systems in the design fase data on operating temperatures
after economiser and required inlet temperatures have to be checked carefully.
This implies in particular for retrofit situations.

* Special attention in future systems for fluegas flow distribution at inlet of
reactor is neeesaary, Guide vanes are in meet eases neeessary,

* Monitoring equipment (e.g. for N0x} for future systems needs much attention
to give sufficient reliable results. Maintenance attention is high.

6A-64

-------
Figure 1

6A-65

-------
Figure 2. Scheme of the Denox Installation

6A-66

-------
ECO-bypass

Flueqas from ECO-cutlet

Figure 3• Schematic drawing of fluegas mixer

6A-67

-------
Figure 4.

Ammonia injection, system.

6A-68

-------
69-V9

•uot^nqxaasTp motjssS ps^Birtoieo *9 ©jnSxj

isKv 1 ^sAiBq.B3

jisA-ej Auiuina

11 >

~» .
tf ,« i,

,« ,» ,» '» 't 'i t* '•

't ''

S i» 'r i' % tr ,'
'r ff t* '* tf

rf Ff ff rf /
't if 'j rf h l' 't

rriM

't 't tf '¦

f! tj J* '[ »' '»
11,

«. ft

it ^ ft t,

t' i

1 S •¦

'1 1' *1 ^ ^ S '» •.

1' 0 0 % o s % .-
% o 0 W \ v * •

' ~ ~ " "Vk**.1? .

•v^v'vrKf
' VC~<- *T

^vV*f^ j -1" »1,1

,4-#-

-------
Figure 7- Gas velocity distribution

6A-70

-------
BOILERSIDE

7 6 5^321 nozzle

729

729'

750

640

698

¦J29

4,3

4,4

4,2

4,0

719

729

770

1930

698

739

4,0

4,6

4,3

3.8

729

739

760

1930

709

719

709

739

4,0

4.4

4,4

3,8

640

739

750

719

729

4,3

4,4

4,3

4,0

N0X

NO } concentrations in ng/mj
02

SUMMARY

N0X

nsg/m^

mg/mj*

min.

719

698

3,8

av.

739

714

4,2

max.

770

739

4,6

0,8

Figure 8. Nox , NO and 02 concentrations before 1st catalyst layer.

6A-71

-------
BOILERSIPE

7 6 5 ^321 nozzle No.

191

150

191

158

640

191

150

191

158

3.0

2,8

3,2

201

168

195

158

1930

201

168

195

158

2,9

3,5

3,0

3,1

160

181

158

1930

160

181

158

2,7

3,6

3,0

2,8

160

168

160

6^0

160

168

160

3,1

2,8

2,6

N0X

NO } concentrations in rag/m^

SUMMARY

NO,

Bg/m3

mg/m3

min.

144

2,6

av.

168

3.0

max.

201

3,6

1,0

Figure 9- Nox, NO and 02 concentrations between 1st and 2nd catalyst layer.

6A-72

-------
BOILERSIDE

7 6 5 4 3 2 1 nozzle

640
1930

distance
in mm

1930
640

N0x

NO } concentrations in mg/m3

SUMMARY

NO*

mg/in3

mg/m3

' %

min.

3.4

av.

Hit

114

3,6

max.

160

3,7

0,3

Figure 10. Nox, NO and 02 concentrations between 2nd and 3rd catalyst layer.

154

109

103

160

. 154

109

103

160

3.5

3,6

3.6

160

129

160

129

3.4

3.7

3,7

3.6

150

123

150

123

3,7

3.6

3,7

3.6

123

ill

123

117

3,5

3.^

3.5

6A-73

-------
BOILERSIDE

236

201

175

226

236

201

175

226

4.8

5,0

4.8

4,6

226

181

164

195

226

181

164

195

5,2

5,1

4,9

5.0

2 ,

212

189

170

201

212

189

170

201

5.0

5,0

4.8

4,8

640 1930 1930 640

NOx

NO } concentrations in mg/mjj

SUMMARY

N0X

mg/m3

min.

164

4,6

av.

198

4.9

max.

236

5,2

0.6

Figure 11. Nox, NO and 02 concentrations downstream 3rd catalyst layer.

6A-74

-------
BOILERSIDE

7 6 5 4 3 2 1 nozzle

631

633

645

640

589

592

604

4,8

5,1

5.3

631

645

663

1930

589

604

622

4,9

5,2

5,3

645

657

653

1930

604

616

612

4,9

4,6

4,5

645

657

651

640

604

616

610

5,0

4,4

4,6

E- »Co„ce„t„tlo„8 in ^

SUMMARY

N0X

mg/n|

mg/m3'

min.

631

589

4,4

av.

646

605

4,9

max.

663

622

5,3

0,9

Figure 12, Hox, NO and 0Z concentrations before 1st catalyst layer.

6A-75

-------
BOILERSIDE

7 6 5^321 nozzle

210

187

154

64o

210

187

154

5.4

5.1

210

177

144

1930

210

177

' 144

5.4

5.6

5.0

195

164 .

144

1930

195

164

144

5.5

5.2

5.0

203

191

156

640

203

191

156

5.4

5.2

5.0

NOx

NO } concentrations in mg/m^
02

SUMMARY

N0X

mg/m3

min.

144

5.0

av.

178

5.3

max.

210

5.6

0,6

Figure 13. Nox, NO and 02 concentrations between 1st and 2nd catalyst layer

6A-76

-------
BOILERSIDE

7 6 5^321 nozzle

183

144

138

640

183

144

138

5.0

5,3

4,8

173

142

107

1930

173

142

10?

5.2

5,4

5,1

152

121

1930

148

121

5.0

4,8

^ ,5

148

115

640

148

115

5,5

4,8

NO } concentrations In mg/tn^

SUMMARY

mg/n>3

mg/in3

min.

4,5

av.

132

5,0

max.

183

5,5

101

1,0

Figure 14. Nox, NO and 02 concentrations between 2nd and 3rd catalyst layer.

6A-77

-------
Figure 15. Integrated control scheme of Dencx with boiler.

6A-78

-------
Electric/Control

Mechanical

Figure 16. Survey maintenance activities.

6A-79

-------
0 1 2 .

years —>

Figure 17- Calculated decrease of catalyst activity.

6A-80

-------
Table 1

TYPICAL OPERATING DATA
(taken froa plant monitoring equipment)

lowest

average highes'

value

1. Fluegas after fluegas mixer

347

355

364

2. Fluegas before reactor

339

346

355

3. Fluegas after reactor

336

347

355

4. Fluegas after air preheater

136

14?

156

5. Mixture air ammonia

tte

pressure/pressuredifference

978

996

1. Inlet fluegas pressure

mbar

957

2. Pressure fluegas after reactor

mbar

951

972

990

3. p reactor (dummy+cat layers)

mbar

5,0

6,1

7,8

4. p NH^ dosing equipment

mbar

55,1

56,8

59.6

5. p air preheater'

. mbar

6,5

7,4

8,6

Flows

1, Fluegas

n|/h

194406

204268

213875

2. Ammonia

m*/h

27,2

25,5

53,9

3- Air

m3 /h

2455

2508

2613

Gasanalyses

1. 02 )

2,1

3,3

5,6

2. S02 ) before the reactor

ppm

288

516

653

3- NO )

ppm

203

348

401

5. NO )

ppm

119

6. NQ2 ) after the reactor

ppm

7- S03 )

ppm

6,5

12,3

14,5

Vibration ID fan

mm/s

0,1

1,8

2,6

Power consumption

1, ID fan

238

259

288 •

2. Ammonia system

3- Air fan

1. NOx removal

74,8

84,7

100,0

2. Total power consumption

253

276

309

Ammenia eensyopiisn

kg/h

20,31

33.75

40,42

6A-81

-------
Table 2

A P OVER CATALYST LAYERS

Operating hours Reactor layer ap in mbar

2500 dummy layer 0,2

2500 cat. layer 1 2,0

2500 cat. layer 2 1,7

2500 cat. layer 3 2,0

4500 cat. layer 1 1,5

cat. layer 2 1,9

cat. layer 3 ¦ 1.^

6A-82

-------
Table 3

ANALYSIS OF FLYASH WITHOUT AND WITH DENOX

Ji&L& 2L

-HIE-

Component

PH NK^ S A12°3 Ca0 Pft2°3 K2° Mg0 H&2° P2°5 Si02 Ti02 P

Fly&sh without

denox 9.t 8,8 0,12 32.1 2,36 6,11 1.37 0,01 0,36 ,0,82 50,2 1,24 0,13*0,01

Fly&sh after

tenox 7,5 7.0 0,2? 33,? 2,18 5,72 1,58 0,89 0,48 1,24 51,2 1,23 0,18*0,01

Flyash without

ienex 9,3 4.2 0,16 27,5 2,03 6,93 2,26 1,30 0,35 0,32 52,1 1,02 0.14*0,01

Fly&sh after

dene* 6,2 6,2 0,26 28,0 1.91 6,66 2,40 1,26 0,44 0,43 50,9 1,03 0,22*0,01

Average sample
frets? silo

without denox 9,4 3,0 0,16 26,6 2.12 6,92 1,62 1,26 0,60 0,55 55»6 0.93 0,10*0,01

Average sample
from silo

after aenox 9,2 3,8 0,18 23,5 2,13 6>?9 1,46 1,10 0,46 0.7D 53,7 1*00 0,10*0,01

Coal analysis

Bound water
Total water-
Ash

Volatiles
Heat value

Heat value (calculated)
S

Average sample

25-01 - 29-01-86

1.6 %

7,9 *

9.0 %

29.0 %

28.5 MJ/kg

27,4 MJ/kg

0.7 %

70.2 %

4.5 %

1.4 *

0.09 %

6,2 %

28 jig/g

6A-83

-------
OPERATING EXPERIENCE OF SCR SYSTEMS
AT EPDC'S COAL FIRED POWER STATION

by: Tatsushi Mori and Noriyuki Shlmizu
Thermal Power Department
Electric Power Development Co., Ltd.
Tokyo, Japan 104

ABSTRACT

Selective catalytic NOx reduction (SCR) system has been installed at Takehara coal
fired No. 1 Unit (250 MW) and No. 3 Unit (700 MW) since July 1981 and March 1983
respectively, and is being installed at Matsuura coal fired No. 1 Unit (1,000 MW)
which will be commissioned in July 1990. The SCR systems at Takehara No. 1 and
No. 3 Units have been operated in good condition through catalyst addition and/or
replacement. This paper refers* to the current operating experience and inspection
results of the SCR systems at Takehara No, 1 and No. 3 Units. The procedure of
catalyst addition and replacement is also described. In addition outline of
Matsuura No. 1 SCR system is introduced. Furthermore, SCR cost analysis is
reviewed based on the latest information. In the final section, EPOC s R&D
activity on SCR catalyst is presented.

1. INTRODUCTION

Regulations currently in force on power stations in Japan as to air pollutant
emissions include national laws (Air Pollution Control Law), prefectural
ordinances and individual agreements made with concerned municipalities
(agreements on environmental preservation). Among these three kinds of
regulations, prefectural regulations are generally stricter than the national
laws, and regulations specified in agreements with concerned municipalities
are stricter than prefectural regulations. This means that emission levels
are practically regulated by agreements with concerned municipalities.
Historically speaking, in addition, the laws have been revised several times
for stricter regulations, and regulations agreed in agreements have tended to
be stricter one after the other although local characteristics affect such-
agreements much.

As for NOx emission, one of the air pollutants emitted from power stations,
combustion methods were Improved at Initial stages by using the two stage
combustion method, low NOx burners and their improved versions. It was
generally recognized, however, that such methods as suppression of NOx
generation would be inadequate when emission standards got stricter than the
100 to 200 ppm level. Technical development of flue gas denitrification
processes then started. Currently, SCR process, which is one of the dry flue
gas denitrification processes, is used widely at thermal power stations in
Japan mainly because of its advantages that the system is simple and that 1t

Preceding page ianfc i

-------
produces no byproducts. As of March 1988, the SCR system is used at thermal
power stations of which generation capacity amounts to 36.4 GW, corresponding
to 37% of the total thermal power generation capacity of 99.7 GW in Japan. As
for coal fired power stations among them, the SCR process is employed at 22
units, which corresponds to 4656 of all coal fired units of 48. Counting by
the power generation capacity, the SCR process 1s used for facilities of 6.2
GW,, which corresponds to 54% of the total coal fired power generation capacity
of 11.4 GW.

The SCR system is classifiable into a low dust and a high dust systems as seen
in Figure 1. This distinction comes from the location of the SCR, which is
downstream or upstream of ESP in the flue gas flow. The ESP therefore is a
hot side ESP for the low dust SCR system, and a cold side ESP for the high
dust SCR system. Either process has its own merits and demerits, but both of
them have been employed actually for coal fired power generation, and they are
operating well. EPDC employs the low dust system for pulverized coal firing
power plants. Note that the systems shown in Figure 1 include a flue gas
desulfuri2ation system which is in general, use in Japan.

Ash of imported coals of low sulfur content used in Japan shows high electric
resistances of more than 1012 ohm-cm at temperatures around 150°C
corresponding to the temperature range of the cold side ESP. The dust
collection efficiency of the ESP is deteriorated in such cases, and a larger
dust collection area, i.e., a larger-ESP is required to compensate for the low
efficiency. The electric resistance of such ash, however, lowers to 1010
ohm-cm or lower around 350°C, which corresponds to the temperature range of
the hot side ESP, and a favorable dust collection efficiency is assured. This
means that the use of a hot side ESP,. I.e., the use of the low dust SCR
system, allows to use a wider variety of imported coals, and it provides a
higher flexibility in procurement of overseas coals.

EPDC employed the low dust system in the SCR demonstration test from the
reason given above and considering the results of the pilot plant tests which
EPDC carried out in order to study SCR systems and catalysts at its power
plants from 1976 to 1982 In cooperation with five representative SCR
manufacturers.

Fully funded by the Government, EPDC carried out an SCR demonstration test at
its Takehara No. 1 Unit (250 MW, PCF) for a year from July 1981. The
demonstration plant was the first SCR system which treated all flue gas of a
coal fired power unit. The system consisted of two trains of SCR, and.
catalysts supplied by two SCR manufacturers were used as selected in the
preceding pilot plant test. The demonstration test ended with fruitful
results. The SCR system of Takehara No. 1 Unit is operating in good condition
since then as a commercial system with addition and replacement of catalyst.
Results of the demonstration test and findings obtained through operational
experiences thereafter were applied effectively to construction of Takehara
No. 3 Unit (700 MW, PCF; full flue gas treatment by SCR). The SCR system of
Takehara No. 3 Unit is operating well since its start of operation in March
1983 with one addition of catalyst since then. Those experiences are further
reflected on Matsuura No. 1 Unit (1,000 MW, PCF; full flue gas treatment by
SCR) which is currently under construction. The SCR system of the unit to be
completed In July 1990 will be the largest in the world as an SCR system for
coal fired power plant.

6A-86

-------
2. EXPERIENCES IN SCR OPERATION AT TAKEHARA NO. 1 AND NO. 3 UNITS

2.1 Outline of SCR'

Figure 2 shows a .view of the SCR system of Takehara No. 3 Unit (700MW). The
SCR system of Takehara No. 1 Unit (250MW) is similar in shape to that of
Takehara No. 3 Unit of vertical downflow type.

Table I shows current SCR specifications of Takehara No. 1 and No. 3 Units.
The SCR system of Takehara No. 1 Unit consists of two trains, "A" and "B",
The catalyst for A train is of plate type, and that for B train is honeycomb
type. The SCR system of No. 3 Unit also consists of two trains, but both of
them use the same plate type catalyst as that for A train of the No. I Unit.
Although NOx concentrations at SCR inlet and outlet are different between No.
1 and No. 3 Units, NOx removal efficiency is the same value of 80%,

2.2 SCR Operation

Figure 3. shows records obtained at start up operation of Takehara No. 3 Unit.
After the start of the boiler, ammonia is not injected until the flue gas
temperature exceeds 300°C. Below that temperature, denitrification reaction
rate is very low, and ammonia, if injected, reacts with SO3 in flue gas and
produces ammonium sulfate "compounds. Such ammonium sulfate compounds deposit
on catalyst surfaces deteriorating the catalyst activity. When the flue gas
temperature exceeds 300°C, denitrification reactions occur by injection of
ammonia, and the NOx level lowers. Although the time is short, the level of
NOx emission is high until service-in of the SCR system. To be able to start
the SCR as soon as possible in the plant start up operation is very important.
Takehara No. 3 Unit is equipped with a boiler economizer bypass system for
this purpose. The damper opening, position of this system is programmed as a
function of the generator output, assuring to make the SCR inlet gas
temperature high quickly for quick SCR service-in even at low generator output
by introducing a part of the flue gas directly into the SCR bypassing the
economizer.

Figure 4 shows the NOx control system of Takehara No. 3 Unit. Under normal
operations, the amount of ammonia injection is controlled so that the NOx
concentration at SCR outlet is 50 ppm. Figure 5 shows operation records
obtained during a load change. Note that the outlet NOx concentration has
been set lower in order to supress peak NOx concentrations during load change
below the operational guide line of 60 ppm. It is also possible to control
the NQx removal efficiency by setting NH3/NOX mole ratio.

EPDC carries out SCR performance tests before and/or after the periodical
inspections which are conducted annually as specified by laws. Tables 2, 3
and 4 show SCR operational data obtained from the recent performance tests.
The SCR systems have operated in stable condition as seen from the data.

2.3 SCR Performance Change in Time

The denitrification efficiency depends also on the NH3/NOX mole ratio. If the
catalyst activity deteriorates therefore, the denitrification efficiency
lowers, unreacted ammonia increases even for ammonia injection of the same NH3
/NOx mole ratio. It is possible therefore to check the changes in SCR
performance by measuring unreacted ammonia concentration under the same flue
gas conditions and the same denitrification efficiency. Figures 6, 7 and 8
show changes in unreacted ammonia concentration measured in the SCR

6A-87

-------
performance tests carried out at the Takehara No, 1A, IB and No. 3 SCRs.
Figures indicate that the slip ammonia increases gradually due to the ,
performance deterioration of SCR caused by gradual deactivation of catalyst in
time and that the slip ammonia drops due to the improved performance of "SCR by
addition or replacement of catalyst.

Ammonium sulfate compounds which come from unreacted ammonia through SCR cause
AH clogging and fan vibration. EPDC sets the maximum operating concentration
of slip ammonia to be 5 ppm for Takehara No. 1 Unit and 4 ppm for Takehara No.
3 Unit from experiences in their operation. Figure 9 shows relations between
the concentration of slip ammonia and the AH washing interval. Correlations
between them are obvious, and. AH washing frequency is roughly once per six
months for each maximum operating concentration of slip ammonia. This means
that one washing in a year between periodical inspections is adequate for AH
maintenance. One thing to be noted on this matter is the shorter AH washing
interval of the No. 3 Unit for the same concentration of slip ammonia compared
with the No. 1 Unit. It may come from the remodeling of the AH of No. 3 Unit
which was made after one year from the start of operation in order to burn
high moisture content coal. The temperature distribution of AH elements along
gas flow changed by the remodeling and the temperature range where ammonium
sulfate compounds deposit moved toward upstream of gas flow, that is, to the
intermediate portion of AH. The soot blower capacity was intensified to cope
with this situation, but its effect seems still weak relatively.

The reason for using not the catalyst activity itself but the concentration of
slip ammonia as the control value for SCR operation comes from the fact that
the decrease in NOx removal efficiency, i.e., increase in NQx at SCR outlet,
which comes from deterioration of catalyst activity in time can practically be
taken care of by increasing the amount of ammonia injected. Therefore the
concentration of slip ammonia is controlled on the SCR to prevent cloggings
and vibrations of downstream facilities which are more serious for the plant
operation.

EPDC periodically measures changes in the catalyst activity of actually
installed sample catalyst as an aid for SCR performance management in addition
to the control of slip ammonia described above. Figure 10 shows such an
example of changes in activity of sample catalysts. The figure indicates that
the deterioration of catalyst activity is rather slow after the initial
period. Note that direct comparison among activity test results of sample
catalysts for Takehara No. 1 A, No. 1 B and No. 3 SCRs is not appropriate due
to different test conditions.

2.4 Catalyst Addition and Replacement

Catalyst additions and replacements have been made on Takehara No. 1A, IB and
Takehara No, 3 SCRs.

For Takehara No. 1A SCR, catalyst was added at the periodical inspections of
1983 and 1985, i.e., two years and four years after the start of operation of
July 1981' respectively. The amount of catalyst added at each addition was
16.5% of the original volume. Such additions were made because it was
estimated from the trend in the concentration of slip ammonia that the AH
washing frequency of more than once a year would be required if the SCR
performance were not improved at such occasions. As shown in Figure 10, the
rate of deterioration in catalyst activity is slow, and EPDC decided not to
replace old catalyst but to add new catalyst to effectively utilize the
activity of existing catalyst, thus achieving a higher SCR economy.

6A-88

-------
Catalyst amounting 12.5% of the original volume was added also for Takehara
No. 3 SCR In the similar procedure at the periodical inspection of 1987, i.e.,
four years after commissioning in-March 1983, because the primary AH washing
frequency had increased to once per three months. Figure 11 and Photograph 1
show mounting of additional catalyst.

By such additions of catalyst, the rate of slip ammonia decreased as shown in
Figs. 6 and 8, and the AH washing frequencies have decreased to levels where
AH washing 1s made only at every periodical inspection for the No. 1A and once
at every periodical inspection and once between inspections for the Ho, 3
Unit.

For the No. IB SCR, on the other hand, a catalyst addition amounting 30% was
carried out one year after commissioning In July 1981. The SO3 conversion
rate of the catalyst was rather high resulting in a high frequency of AH
washing, and the tube type catalyst was replaced completely with honeycomb
type catalyst at the periodical Inspection of 1985, i.e., four years after
commissioning. As shown in Figure 7, the concentration of slip ammonia is
extremely low since the replacement, and AH washing is made only at each
Inspection.

EPOC manages catalyst addition or replacement schedule based on the trend of
slip ammonia concentration obtained by the SCR performance tests. And the
result of performance tests on the sample, catalyst is considered also for the
management. Figure 12 shows schematic diagram of the SCR performance
management.

2.5 inspection of SCR

The SCRs of Takehara No. 1 and No. 3 Units have been subjected to inspections
at each periodical inspection of the plants. At the SCR inspection,
accumulation of dust on catalyst layers, support beams and baffle plates,
wear, deformation and corrosion of catalyst, and condition of ammonia
injection nozzles are checked visually. Photograph 2 and Figure 13 show a
state of dust accumulation on catalyst layers observed during inspection of
the No. 1 SCRs. It has been confirmed that the accumulation of dust of the
level shown in the photographs and the figure does not affect the SCR
operation. Such dust, however, is removed during inspection. No major
problems have been found in recent inspections, and SCR systems are operating
well continuously without requiring any repairs except those which are made
during periodical inspections.

2.6 Facilities Downstream SCR

Flue gas from boilers contains a small amount of SO3 generated during
combustion. The concentration of such SO3 is on the order of 1% or less of
the SO2 concentration although it depends on the boiler type. In addition,
the catalyst of the SCR converts a part of SO2 into SO3. A part of ammonia
injected in the SCR, on the other hand, passes the SCR unreacted with NQx,

Such slip ammonia and SO3 in the flue gas react and cause deposits of ammonium
hydrogen sulfate and ash compounds on downstream facilities existing in low
temperature ranges such as AH and fans. These deposits then cause troubles
such as clogging or vibration. The use of low SO3 conversion rate catalyst
and suppression of slip ammonia therefore are important factors for
suppression of adverse effects due to ammonium sulfate compounds on downstream
facilities.

6A-89

-------
For low dust SCR systems, catalyst with low SO3 conversion rate of 0,5% or
less is generally used and SCRs are operated with the maximum value of around
5 ppm for slip ammonia. These measures have been effective for reducing AH
clogging at Takehara power station as well.

Table 5 shows differential pressure change before and after AH washing. The
table indicates that washing of AH is effective for cleaning the clogged AH.
As seen in the case of Takehara No. 3 Unit, however, the washing of a
secondary AH requires more than 10 hours, about 1,000 m-3 of washing water and
waste water treatment, resulting decreased plant availability and increased
maintenance costs. The selection of AH element shapes which prevent deposit
accumulation and the selection of soot blower type and its operation schedule
become important considerations to be given in the design and .operation of the
AH. As shown in Table 6, these considerations have been taken into account
for Takehara No. 1 and No. 3 Units.

As for the BUF, which is a boost up fan for FGD system, among fans existing in
downstream of the SCR, those of Takehara No. 1A and IB have been cleaned
frequently in recent years to cope with actual and possible vibration. The
cause of the increase in vibration is unbalanced deposits on the impeller of
BUFs. The relation between the deposit and the slip ammonia is not clear now,
and it would be a future study item. The BUFs of Takehara No. 3 Unit have
different type impeller and no specific problems exist without dust deposits.

For other downstream facilities of SCR such as IDF, FGD, Gas/Gas heat
exchanger and FGD wastewater treatment system, no problems specifically
ascribable to SCR have occurred.

3. CONSTRUCTION OF LARGE-SCALE SCR

EPOC has employed the low dust SCR as the denitrification system for Matsuura
No. 1 Unit {1.000MW) which is under construction and scheduled to start its
commercial operation in July 1990. The decision comes from the use of the hot
side ESP for collecting fly ash effectively for a wide variety of overseas
coals. It goes without saying that the decision was made based on the huge
data obtained through operation of hot side ESPs at Matsushima No, 1 and No. 2
Units and Takehara No. 3 Unit covering a time span of more than 8 years,
experiences in the operation of low dust SCRs at Takehara No. 1 and No. 3
Units and results of economical evaluation of the whole flue gas treatment
system for Matsuura No. 1 Unit including ESP, SCR, FGD and AH as well.

An overview of Matsuura No. 1 SCR is given in Figure 14 and Table 7.

Points of considerations taken into account in the design of Matsuura No. 1
SCR from experiences in operation of Takehara No. 1 and No. 3 SCRs include the
following:

• Selection of optimum space velocity

• Additional catalyst space allowing for catalyst volume of, 25%

• Suppression of dust deposits

Distance between catalyst layers, distance between beam and catalyst
layer, shape of baffle plate and employment of guide plate

6A-90

-------
• Improvements in the control system

The SCR plant of Matsuura No. 1 Unit will be the world's largest one when
completed.

4. SCR COST

It has been a general understanding in Europe and the U.S. that the SCR system
1s rather costly. EPDC, however, has demonstrated that the SCR is rather
attractive in cost through experiences in construction and operation of SCR
systems for Takehara No. 1 and No. 3 Units.

An example of economic evaluation for a new 350MW SCR system is presented
here. Table 8 shows premises employed for the economic evaluation. The unit
is assumed to be of middle size for base load operation. The coal
characteristics are those typical for coals used at imported coal fired power
stations in Japan. The SCR is a low dust type according to experiences at
EPDC, and NOx concentrations at SCR inlet and outlet are based on those of
Matsuura No. 1 Unit. As for the catalyst life, this example assumes that the
whole catalyst is replaced at a cycle of 4 years, which has a large margin on
the safety side compared with the catalyst life experienced at Takehara No. 1A
and Takehara No. 3 SCRs. As for the economic indices, the exchange rate of
130 yen/S is used, and the discount rate of 10%, which is typical for Japanese
electric utilities, is used. As for the plant life, it is assumed to be 20
years, which is more practical than the depreciation period of 7 years under
taxation laws in Japan.

Table 9 shows the construction cost and the leverized bus-bar cost calculated
from the premises described above. The construction cost and the bus-bar cost
are as attractive as S71/kW and 3.5 mills/kWh respectively. As described in
INTRODUCTION, SCR systems are widely used in Japan. Information which have
been accumulated through actual construction and operation of SCR plants are
dependable and uncertainty factors which may lead to a large contingency fund
are minimal. In addition, the catalyst performance has been improved
gradually as seen from an example of the honeycomb type catalyst used for the
replacement at Takehara No. IB SCR. Further improvements in economy are
expected for the future through technical innovations and competitions among
manufacturers.

5. R&D ON SCR CATALYST

The catalyst life management 1s an important factor for stable operation of
SCR plants. EPDC has been conducting various tests for evaluation of catalyst
performance and clarification of factors for catalyst activity deterioration
at its Engineering & Research Institute using catalyst test equipments since
1985, and the catalyst life management is being carried out taking results of
such tests into account in addition to the results of performance tests
conducted at actual SCR plants.

The catalyst activity test equipment is outlined in Figure 15. Various
catalyst samples such as plate type and honeycomb type catalyst samples from
actual SCR systems are tested using both this, equipment and other analysis
devices in order to determine the catalyst activity, establish the method for
catalyst life forecast and clarifiy the factors which cause the catalyst
activity deterioration. These studies are summarized in Table 10.

6A-91

-------
Results in the R&D activities up to now are as follows. In the field of
determination of sample catalyst activities, the selection of standard
conditions such as gas temperature, gas composition, NH3/NOX mole ratio and
area velocity have been decided in order to measure the NOx removal efficiency
and the SO3 conversion rate by the catalyst activity test equipment. Based on
such achievements, we plan to continue investigations to establish the method
for catalyst life forecast.

As for the factors which cause catalyst activity deterioration, followings
have been found.

• The reduction 1n specific surface area leads to catalyst activity
deterioration.

• As to the change in pore size distribution in low dust plate type
catalysts, pores of size around 50A in diameter decrease while the ratio
of pores of intermediate sizes around 100A in diameter increases.

• From the adherent material analysis no attachment of poisoning alkali
metal elements such as Na and K are recognizable on the surface of sample
catalysts as far as being analyzed by a electron probe micro analyzer
(EPMA) although dust has deposited on the surface. Forced poisoning tests
on catalyst samples have revealed, in addition, that the dust is somewhat
responsible for catalyst activity deterioration and'that the effect of
potassium is large among alkali metals.

• As for the surface conditions, the attachment of dust to catalyst surface
depends on the catalyst type and the place of catalyst installation in SCR
reactor.

• As for the crystal structure, no change has been recognizable by X-ray
analysis.

From the test results above mentioned, the change in catalyst surface
conditions such as specific surface area and pore size distribution have close
relation to catalyst activity deterioration in the case of SCR catalysts of
Takehara Power Station.

6. CONCLUSION

EPOC has been playing pioneering roles among electric power utilities in Japan
in the development and utilization of the SCR process for coal fired power
plants since 1976 by its steady accumulation of experiences represented by the
pilot tests, demonstration tests and more than 6 years of commercial
operations of SCR plants. The SCR process is now used widely at about half of
all coal fired power stations in Japan, and the SCR system now belongs to a
well-matured technology in Japan. EPOC envisions for larger SCR systems, and
a world's largest SCR unit will be completed in 1990. It has been
demonstrated that SCR system is attractive in cost, and the system is getting
more economical through technical innovations and competition among SCR
manufacturers. EPOC continues its efforts on improvement of the SCR system
including the establishment of a method for catalyst life management and
further improvement of system economy.

6A-92

-------
REFERENCES

1. "NOx Removal from Flue Gas", Y. Nakabayashi, Conference on "The Role of Coal
in the Swedish Energy System, Possibilities and Limitations", Sodertalje,
Sweden, October 1988

2. "Flue Gas Treatment Facilities Operation Experience", T. Miyasaka, the same
conference as Reference 1

3. "Design of Flue Gas Treatment Facilities", T. Kainuma, the-same conference as
Reference 1

4. "Current Status of SCR in Japan", Y. Nakabayashi and R. Abe, 1987 Joint
Symposium on Stationary Combustion NOx Control, New Orleans, U. S. A., March
1987

5. "Updated Technical and Economic Review of Selective Catalytic NOx Reduction
Systems", J. E. Damon and D. V. Giovanni, the same joint symposium as
Reference 4

6. "The- Operation Results of DeNOx System (SCR) for Unit No. 3 (700 MM) at
Takehara Thermal Power Station, Japan", J. Watanabe, NOx Symposium Karlsruhe
1985, Karlsruhe, West Germany, February 1985

6A-93

-------
BOILER HOT SiDS AH 10 FAN GGH SU FAN FGD STACK

esp n

SCR

KHK>H

|— l<

j-+L

LOW DUST SCR SYSTEM

BOILER AH COLD SIDE !D FAN GGH BU FAN FGD STACK

NM, ESP _

—»

SCR

wv" j— I1*

HIGH DUST SCR SYSTEM

Fig. 1 Type of SCR System

Injection Norile

Fig. 2 Takehara #3 SCR Reactor

6A-94

-------
nfN/H

ISO

. 300

tnr

NH? Injection
Sl3'l

Gas Temperature

r 6" ]' 8*^ 9* Iff-
ot ra

Boiler Ignition PsraW in

I r \Y 14' 15' IF

Source: Reference 6

Fig. 3 Start Up Record of Takehara #3

Gas Flow Intel NQx

*G3

-*EB

NH 3 Row

SCR Out
NO* Sc!

Mete Rati© Set

i>0—

Source; References N Ha Flow Control Valve

Fig, 4. NOx Control System at Takehara #3

* O

8 m
% 8

3 §

U
w

300

00;

Load

£ o

NOx m

a o

3 m

1——r—-t—

SCR inlet NOx
roo

Plant Load
-------
Conditions

Flue gas volume 407,500 m' N/ti
inlet NO* (as 8% 0») 3D0 ppm
Outlet NOx (as 6%Oi)

DeNOx Ed.

. Gas temperature

18,5% addiiion

Resouna: B«t«enc« t

16 i> 32 «

Operation Time (hours)

60 ppm

' %
348 "C

Fig. 6. Slip NHa Trend at Unit No, 1 A-train

- 30* C»BIy»t •CrtMn

Conditions

Rut gas volume 407,500 re? Nrti>
Inlet NOx (as 6% Oj) 300 ppm
Outlet NO* (as 7% O,) 60 ppm
DeNOx EH. 80 ppm

Gas temperature 348 °C ,

rAU Catalyst layer

Rsplacerrent
iP ip# lyp® !e

Honeycomb type)

27Sm° —p. 134 rt?

a IS 34 32 jo *a

Soweer Reierencs 1 Operation Tims (hours)

Fig. 7. Slip NHa Trend at Unit No. 1 B-train

fi I.COO)

3 .

.9.

« 5

Conditions

Rue gas volume 2,230,000 rrf N/h
Inlet NOx (as 6% 0«) 250 ppm
Outlet NOx (as 6% O i) 50 ppm
DeNOx Eff. 80%

Gas temperature 360 5C

12.5% addition
(Tolal J ,060 m')

B4 A.},

& & &

BSA.f. - 06 At, "07 A i.

<18 (X 1,000)

soufta: Botarsnea i Operation Time (hours)

Fig. 8. Slip NH3 Trend at Unit No. 3

6A-96

-------
I

Takehara #1

Slip NH3 (pprrs)

Takehara #3

1?y AH

Slip NH, (ppmj

Fig. 9 Slip NH3 vs. AH Washing Interval

0,9
0 B
0.7
0,6

Gas Temp. 350 C
NIVNO 1.2
AV 51 mm

Gas Composition

16 ZA 32 AO

Takehara #1 Q-Traln

Takehara #3

Gas TCmp, 3^0'C

m3mo 1 ,o

AV 25m/h

Gas Composition

MO
50,
Or
H,Q

300ppm
BOOppm

Gas Temp, 350"C
NH,/NO- 1.2
A¥ 51m/h

Gas Compos lion

,so2
o,

Ha0

200ppm
5Q0ppm
3%

12%

Operation Time (hour)

Fig. 10 Activity Deterioration of Sample Catalyst

Note: Direct comparison among above three graphs is inappropriate

(x 1.000)

6A-97

-------
j 10900 J

Fig. 11 Additional Catalyst at Takehara #3

Fig. 12. SCR Catalyst Performance Control Flow Diagram

Source: Reference 2

Fig, 11 Dust Deposits on Catalyst layers at #1-A Train

6A-98

-------
NHj Injection Nozzle

Fig. 14 Matsuura # 1 SCR Reactor

Fig. 15 Catalyst Activity Test Equipment

6A-99

-------
Table 1, SCR Specifications of Takehara #1 and #3

Train Takehara #1A/B Takehara #3

Gas Volume
Gas Temperature

m3N/h
°C

407,500 1407,500
329 - 348

2,230,000
360 - 380

NOx In
NO* Out

DeNOx Efficiency

ppm (6 % O2)
ppm (6 % O2)

300
60
80

250
50
80

SOx In

SO3 Conversion

ppm
%

1,850
0,5 ma*.

1,000
0.5 max.

Dust in

nig/m^N

100

Slip NH3

ppm (6 %C^)

3.7/4.0

3.6

Catalyst Type

...

Plate (10mm pitch)
/Honeycomb (6mm opening)

Plate (10mm pitch)

Catalyst Volume

232/134

1,060

Space Velocity
Area Velocity
Linear Velocity

1/h

m/h
m/s

1,760/3,037

8.7/6.5

4.8/5.2

2,110
10.4
6.9

Reactor Dimension

8.5W * 8.5D x 10.8H

10.9W * 12D x 13H

Catalyst Stage
Train

—

4/3

1/1

4.5
2

Table 2. Takehara #1-A SCR Performance Test Result

Aug. 1985

Dec. 1985

Nov. 1986

Oct. 1987

Feb. 1988

Gas Volume

x 1,000 m3 N/h

404

389

399

396

398

Gas Temperature

°C

343

344

352

341

348

NOxIn

ppm (6 % 02)

266

287.

271

238

252

NO* Out

ppm (6 % O,)

DeNOx Eificiincy

Slip NH3

ppm (6 % Os )

7.2

2.3

¦ 1.3

2.8

2,5

Slip nh3

ppm (design basis)

* 6.6

1.9

2.9

2.7

3.0

SOx In

ppm

1,520

1,420

1,240

1,420

1,050

SOa In

ppm

4,0

4.4

5.3

2,1

3.4

SOa Out

ppm

5,0

4.8

5.7

2,5

3,8

SOa Conversion

0.07

0.04

0,03

0.04

* : calculated on the design basis described in Figure 6

Catalyst was added In the periodical inspection between September and November, 1905.

6A-100

-------
Table 3. Takehara #1-B SCR Performance Test Result

Aug. 1985 Dec. 1985 Nov. 1986 pet. 1987 Feb. 1988

Gas Volume

x 1,000 m3 N/h

300

396

411

377

395

Gas Temperature

°c

333

343

354

343

340

NOx!n

ppm (6 % O2 )

274

306

283

251

249

NO* Out

ppm (6 % 0 2 >

DeNOx Efficiency

Slip nh3

ppm (8 % O2 )

8,1

0.1

0.2

0.1

Slip NH3

ppm (design basis)"

6.8

0.1

0.2

0.1

SO* In

ppm

1,440

1,490

1,180

1,340

1,210

SOa In

ppm

4.3

5.2

4.9

3.2

2.8

SOa Out

ppm

19.0

7.6

7.4

3.2

4.S

SOa Conversion

0.75

0.18

0.21

0.08

0.15

*: calculated an the design basis described In Figure 7

Catalyst was replaced In the periodical inspection between September and November, 1985.

t* i i ¦ « > jin a »—«p-% r* r «r ¦ r—\ i •

Table 4. Takehara #3 SCR Performance Test Result

Sept. 1905 Aug. 1986 Aug. 1SB7 Dec. 1987

Gas Volume

x 1,000 m3N/h

2,140

2,318

2,346

2,364

Gas Temperature

°C

360

366

365

360

NOxIn

ppm (6 % Oa)

246

228

245

226

NOx Cut

ppm (6 % 0,)

DeNOx Efficiency

Slip NH3

ppm (6 % 03 )

2.6

2,4

6.2

2.0

Slip NH3

ppm (design basis)*

3.9

4.5

5.3

2.3

SOx In

ppm

221

205

405

200

SO 3 In

ppm

1.3

1.0

2.1

2.6

SO 3 Out

ppm

1.7

1.5

3.6

3.4

SOa Conversion

0.15

0.24

0.37

0.4

*: calculated on the design basis described in Figure 8

Catalyst was added In the periodical inspection between September and November 1987.

Table 5. Pressure Recovery by AH Washing

Differential Pressure cf AH (mntAc)

Before Washing

Mi-Washing

Takehara #1

A-AH

B-AH

Takehara #3

X
<

114

2ry A-ah

2ry 0-AH

6A-101

-------
Table 6. Basic Consideration for AH Clogging

«AH Element NF type layer in possible deposit areas

* Soot Blower Reinforced blower such as retractable, twin or multi nozzle types

» Blowing Schedule Increased blowing frequency, ex. 6 times/day

Table 7. SCR Specification of Matsuura #1

Design. Basis Max. Gas Vol. / Max. Inlet NO*

Gas Volume

x 1,000 ma N/h

3,100/3,029

Gas Tempeature

370

NO* In

ppm (6 % Oi)

200 1 250

NOx Out

ppm (6 % O z )

DeNOx Effic»roy

75/80

SOx In

ppm

200- 1,000

SOj In

ppm

3.S max.

SO] Conversion

0.5 max.

Dust In

too

Slip NHa

ppm

2/3.S

Catalyst Type

—

Plate (10 mm pitch)

Catalyst Volume

1,541

Space Velocity

1/h

2.010

Area Velocity

w/h

Linear Velocity

nVs

6.5

Reactor Dimension

12.1WX 15.60 X 14.5H

Catalyst Stage

—

Train

—

Table 8. Premises for Economic Evaluation

Plant Generating Capacity
Plant Efficiency
Plant Use Factor
Annual Capacity Factor
Electricity Transmitted

Coal Calorific Value
Sulfur Content
Nitrogen Content
Ash Content

SCR Type
NOx In/Out
Catalyst Life
Train

Ammonia Cost

Discount Rate
Escalation Rate
Plant Life

350 MW x 1, new unit
38.3 % (net)

8.5%

70%

1,961 GWh/year

6,200 kcal/kg JIS ad

1.2%

1.8%

15 %

Low Dust System

250/50 ppm (80% DeNOx Efficiency)
4 years
2

3850/t

Currency Exchange Rate Yen 13Q/$

10 %
0 % pa
20 years

6A-102

-------
Table 9. SCR Cost

350 MW x 1, rtaw unit,
1389 value, 1$ =» 130 Yen

CONSTRUCTION COST

Reactor and Ducting

$23 tm

Catalyst

Ammonia Facilities

Others (Electrical Equipmant,

C & I, Civil Works)

Gn-Sita Engineering &

Support Facilities

interest during Construction

Home Office Fee

TOTAL CAPITAL REQUIREMENT

$71' m

UEVERIZED BUS-BAR COST

Capital

1.5 mills/kWh

Operation and Maintenance

0,4

Catalyst Exchange

0.9

Ammonia

0.5

Other Utilities

0.1

Tax

0.1

TOTAL 3US-8AR COST

3.S millsAWh

Table 10. R&D on SCR Catalyst

Sample Catalyst Activity Test

Catalyst Life Forecast

Study on Factors in Catalyst Deactivation

«Specific surface area N 7 adsorption (BET)

• Pore size distribution Mercury method

• Adherent material analysis Atomic absorption analysis

«Surface stale SEM

«Crystal structure X-ray Analysis

5A-103

-------
Reproduced from

best available copy.

Lightly Deposited Area Heavily Deposited Area

Takehara#"!-A Train

Lightly Deposited Area Heavily Deposited Area

Takehara #1 - B Train

Photo. 2 Ash Deposits at Takehara =1 SCR

6A-104

-------
TECHNICAL FEASIBILITY AND ECONOMICS
OF SCR NOx CONTROL
IN UTILITY APPLICATIONS

C.P. ROBIE
P.A. IRELAND
UNITED ENGINEERS & CONSTRUCTORS INC.

STEARNS-ROGER DIVISION
AND

J.E. CICHANOWICZ

ELECTRIC POWER RESEARCH INSTITUTE

ABSTRACT

An evaluation is presented of the technical feasibility and economics of
retrofitting selective catalytic reduction (SCR) technology to one oil and one
coal fired power plant. The current status of the technology is assessed based
primarily on recent European experience. Costs are included in the analysis for
operational effects on downstream equipment. The capital cost ranges from 73 to
$94/kW for the oil case and 78 to $101/kW for the coal case; the levelized busbar
cost ranges from 3.5 to 12.2 mills/kWh for the oil case and 5.2 to 6.6 mil 1 sVkVlh
for the coal case.

INTRODUCTION

Recently proposed acid rain legislation in Congress has focused attention on the
cost of SCR NOx control. In addition, European experience with SCR has expanded
the understanding of the cost of this technology. This paper presents an
evaluation of the technical feasibility and cost of retrofitting the SCR process
for N0X control to one typical oil-fired and one coal-fired U.S. power plant.

The cases examined in this paper are the first two of six to be examined in a
larger study performed for EPRI, assessing the cost and feasibility of SCR. The
six cases are defined as follows:

1. Retrofit - oil/gas fired plant

2. Retrofit - tangential, high sulfur coal-fired boiler

3. Retrofit - cyclone, high sulfur coal-fired boiler ,

4. New plant - low sulfur coal

5. New plant - high sulfur coal

6. Retrofit - high sulfur coal-fired boiler implementing "cold-side" SCR
downstream of flue gas desulfurization (FGD) equipment

6A-105

-------
Complete results for all cases will be presented in an EPRI report anticipated in
the fall of 1989.

Regulations in Europe have led to significant application of SCR N0X control.
By 1992 application to over 30,000 MW of capacity is anticipated. Currently,
some SCR installations have been in operation for about three years. In
addition, numerous European SCR pilot plants have increased the understanding of
this technology. Based on this European experience, the following design
assumptions were selected for the evaluation:

1. Catalyst life - Several coal-fired European SCR installations have operated
for over two years without catalyst replacement and only moderate measured
loss in reactivity. A catalyst life of four years for coal-fired plants and
six years for oil-fired plants has been used in this evaluation.

2. Catalyst cost - Catalyst costs in Europe have decreased over the past four
years 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 the coal cases, a 5 ppm slip only has been evaluated. A slip
of 10 ppm for the oil case was utilized.

4. Space Velocity - Advances have been made in catalyst formulation to minimize
SO2 to SO3 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 two cases in this evaluation.

CASE DEFINITION

In order to develop a representative cost of retrofitting SCR, a typical existing
power plant layout and typical design conditions were selected for each case.
Minor modifications were made to provide study cases with broader industry
applicability. Definition of the two study cases are summarized in Table 1.

In both cases, the SCR reactor was located above the existing air heaters because
of space constraints. The oil case required a high width to length ratio to
allow placement of the reactor between the boiler and the stack. Vertical
(downward gas flow) reactor configurations were selected for both cases; the coal
case because of ash in the flue gas and the oil case due to space constraints.
Plans and elevation sketches of the SCR reactors are shown on Figures 1 and 2,

SCR PROCESS DESIGN

To obtain budgetary SCR system costs, a performance specification for the
catalyst and reactor supply was,developed for each case. , The specifications were
developed using fuel analyses, plant performance and emissions data from the
actual power plant sites. Additionally, emissions performance and ammonia slip
requirements were specified for each case as well as certain sensitivities of
interest. Three SCR system suppliers provided quotations to these specifications.

6A-106

-------
The SCR design basis and resultant vendor* supplied design data are shown in Table
2. Two cases each for both the oil and coal plants were developed to show the
effect of changing the uncontrolled N0X emission rate (by combustion control
techniques) while still meeting the same N0X emission limit. The N0X
emission limit of 0.03 lb N0X/ MM Btu for the oil case requires 90% N0X
reduction for case A and 80% for case B. Similarly, the N0X emission limit of
0.12 lb N0x/MM Btu for the coal case requires 801 N0X reduction for case A,
and 70% for case B.

A single, vertical reactor was specified for the oil case, while twin (2 x 501),
vertical reactors were specified for the coal case. Both reactor designs include
the capacity to add an additional catalyst layer. Typically, the fifth catalyst
layer would be added after two years (coal case) or three years (oil case) of
process operation. Yearly replacement of a single catalyst layer would begin
after four years for the coal case and six years for the oil case.

The reactors were designed with steam soot blowers between each catalyst layer.
The soot blowers are recommended in the oil case because of the peaking duty
criteria and in the coal case because of fly ash content in the flue gas.

The ammonia storage and supply systems for both cases we re 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 coal case and the
Forced' Draft fans in the oil case. In the oil case, an urban site is assumed and
the NH3 storage island is located below grade.

The low sulfur and ash content of the fuel in the oil case allows selection of a
catalyst with a high specific area and high activity which results in a space
velocity considerably higher than the coal case.

SCR PROCESS IMPACT

The SCR process, because of its location directly downstream of the boiler and
upstream of the air heater, affects every component of the flue gas train and the
boiler itself through its effect on the air heater. The degree of its effect
varies with power plant configuration, environmental control components, type of
fuel, and emission control requirements.

Oil Case Impact

For the oil case, the SCR design impact is summarized on Figure 3. The principal
effects are on the boiler, air heater and the requirement for new ID fans. These
and other effects are discussed below:

Boiler. The oil-fired power plant selected for analysis utilizes a pressure
boiler. That is, forced draft (FD) rather than induced draft (ID) fans are
utilized. To accommodate SCR several modifications were implemented.

To maintain the flue gas temperature above about 600"F (minimum recommended SCR
operating temperature) over all load conditions, an•economizer bypass has been
added. The bypass flow will be automatic, controlled off the SCR inlet
temperature.

To prevent possible implosion of the boiler, pressure control bypasses have been
added to the new ID Fans. The ID Fans will complicate pressure control in the
boiler.

6A-107

-------
Modifications to the air* heater will result in a higher flue gas exit
temperature, reduced air inlet temperature and reduced plant efficiency. The
result will be slightly less generating capacity at MCR conditions.

Air Heater. Formation of ammonium bisulfate (from SCR operation) and its
deposition on air heater surfaces requires major modification to the twin,
horizontal air heaters.

The cold-end baskets will be replaced with NF (notched, flat) surface. This will
result in a higher flue gas exit temperature of approximately 5°F and a loss in
boiler efficiency. The cold end basket will be combined with the intermediate NF
basket to provide a continuous surface to facilitate cleaning.

High pressure steam soot blowers will be added to both the hot and cold sides of
the air heater. The high pressure steam will require replacement of the 24 gage
air heater heat transfer surface with 18 gage material.

To provide continuous operation, on-line water washing capability is.provided;
the water wash is required to clean ammonium bisulfate deposits off heat transfer
surfaces. On-line water washing requires adding high pressure water wash
assemblies upstream and downstream, air bypasses (from air side to flue gas side)
and dampers, and water wash pumps and piping.

To ensure proper flue gas distribution to the air heater, rectifying plates and
turning vanes have been added. In addition, the air heater outlet plenum and
ductwork are modified to accommodate new ID fans.

ID Fans. The existing FD fans are not capable of overcoming additional flue gas
pressure drop associated with the SCR process. Accordingly, new axial ID fans
have been added downstream of the existing air heaters. The axial fans provide
high efficiency service in the cycling operation specified for this application.
Total estimated pressure drop attributable to the SCR operation is 12 " W.C. A
contributing factor to the high pressure drop attributable to the SCR process is
the increase in the flue gas volumetric flow rate. The volumetric flowrate
(downstream of the air heater) increases as a result of higher flue gas
temperature, reduced pressure, and increased mass flow resulting from higher air
heater leakage and dilution air.

FD Fan. The FD fans will consume slightly more power to account for a higher
mass flow rate. The higher mass flow results from a higher air heater leakage
rate and in providing dilution for the injected ammonia.

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
nitrification and denitrification biological treatment process is utilized to
convert the nitrogen species to free nitrogen.

The discharge from the denitrification process is assumed to be discharged to the
existing on-site water treatment equipment.

Coal Case Impact

In the coal case, retrofitting SCR has similar impacts as discussed above for the
oil case and additional ones because of the greater complexity of environmental
controls. The impacts are summarized on Figure 4,

The principal impacts are on the boiler, air heater and ID fan. Other impacts

6A-108

-------
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. A tangential-fired boiler is utilized in the- coal study case. The
principal effects of SCR on the boiler will be the loss of overall efficiency and
additional control complexity.

Air heater modifications will result in a higher air heater flue gas exit
temperature resulting in a reduction of plant efficiency. The result will be a
loss in net generating capacity.

Air Heater. Formation of ammonium bisulfate coupled with the presence of fly
ash necessitates air heater modifications. The modification to the twin (2 x
50%) trisector air heaters 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.

ID Fan. To overcome additional pressure drop (11" W. C.) associated with the
SCR, the existing ID fans were modified. It was assumed that new, larger
diameter wheels .could be placed into the existing fan housings 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 SO^ concentration, higher flue gas
temperature and precipitation of ammonium bisulfate on fly ash.

Higher flue gas volume (9.4%) resulting from higher flue gas temperature (20°F),
lower flue gas static pressure and increased mass flow due to increased air
heater leakage and dilution air will have a significant impact on.ESP operation.
The increase in flue gas volume results in a reduction of the Specific Collecting
Area (SCA) and in a reduction in the concentration of particulate in the flue
gas. The result will be that the ESP will require additional power to effect 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.

The SO3 concentration'in the flue gas is estimated to increase-by 18 ppm across
the SCR. Typically, an increase in SO3 would be expected to reduce the fly ash
resistivity significantly. However, the increase in the flue gas temperature may
counteract the effect of the SO3 increase, possibly producing little net change.

An 'increase in the flue gas temperature by 20"F results from a loss of air heater
performance. The temperature increase will keep the flue gas temperature above
the acid dew point therefore not affecting fly ash resistivity significantly.

Ammonium bi sulfate precipitation on the fly ash can have a beneficial impact on
ESP performance by helping the fly ash agglomerate, preventing reentrai nment.

The cumulative effects of all the above could be significant on an ESP. As a
result, a test program would be recommended 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%.

6A-1Q9

-------
Ash Disposal« Ammonium bisulfate precipitation on the fly ash can have a
significant impact on waste disposal or marketing practices. In the case of an
Eastern coal with acidic constituents, offgassing of ammonia will not likely
occur during waste disposal. However, ammonium bisulfate will offgas ammonia
when the pH is raised, and as a result the fly ash may not be a marketable
commodity to the construction industries,

FGD/Reheat. The chief effect on the FED system in this study case 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. In this particular case, an increase in the water
evaporation rate of approximately 85 gpm will result.

The higher flue gas mass flow and Increase in water evaporation will result in a
significant increase in steam use by the FGD reheat system (50°F reheat
assumed). The increase in mass flow requires approximately 3.7 MM Btu/hr of heat
input.

A slight increase in power consumption could occur from having to increase the
FGD liquor recirculation rate in order to maintain the same SOj removal
efficiency. The higher liquor recirculation rate might be required as a result
of dilution of SO? in the flue gas, and higher flue gas volumetric flow rate
(saturated gas flowrate). This effect was not considered in this analysis.

In this case, it is expected that some residual ammonia may be captured by the
FGD system resulting in a build-up of ammonium sulfate 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 SO3 concentration across the SCR could
result in increased opacity of the flue gas plume. Recent data from an EFRI
sponsored study with a member utility shows a direct correlation between stack
opacity and sulfuric acid concentration. To reduce opacity increases, control
measures may be required to reduce the SO3 concentration. A typical method of
reducing SO3 in the flue gas would be to inject NH3 upstream of the ESP,

This impact has not been evaluated in this study, however.

COST DEVELOPMENT

To develop total process capital costs for the oil and coal case retrofits.,
physical layouts of the ductwork and SCR reactors were developed. From these '
drawings, lengths of ductwork and structural requirements were estimated.

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, S-R utilized
in-house data to arrive at equipment costs.

In developing operating costs, S-R utilized EPRI's Technical Assessment Guide
(TAG) to estimate fixed operating and maintenance costs. Variable operating
costs were determined by calculating utility and raw material consumption- rates.
Considered in developing the variable operating costs were the following:

* SCR catalyst replacement

* Ammonia consumption

* Ammonia vaporization steam

6A-11Q

-------
* Incremental Sootblowing steam

* Incremental ID fan horsepower consumption

* Incremental FD fan horsepower consumption
Incremental ESP power consumption

* Water treatment chemicals

* Air heater efficiency loss

* Increased air heater leakage

* Incremental FGD reheat steam consumption

* SCR catalyst disposal

* Incremental fly ash disposal cost

RESULTS

Figures 5 to 8 report the Oil Case cost results; Figures 5 and 6 report capital
costs, while Figures 7 and 8 report levelized costs.

* Figure 5 indicates that the Total Capital Requirement (TCR) varies from
$79/kW to $94/kW. The results indicate that the SCR capital cost
required to achieve 0.03 Ib/MM Btu is $7.6/kW less when boiler combustion
controls are added to reduce the boiler emissions from 0.30 to 0.15 Ib/MM
Btu. This results from more catalyst required to achieve the 0.03 Ib/MM
emission limit at the higher boiler emission rate; the 90%' case requires
a space velocity of 3370 hr"^ versus 3900 hr-"' for the 80% case.

* Figure 5 also shows the- impact of catalyst cost. Comparing Cases A and C
shows that the capital cost required to achieve 90% N0X reduction can

be reduced by $14.9/kW (16%) if the initial catalyst cost is reduced to
$330/ft3.

* Figure 6 shows the cost components that make up the TCR for Oil Case A.
Total Process Capital (TPC) comprises the largest component. Of the TPC
figure, approximately 50% can be attributed to catalyst and reactor
cost. It is interesting to note that for a catalyst cost of $330/ft3,
the TPC figure would be dominated by structural and ductwork costs.

* Figures 7 and 8 report the levelized cost of removing N0X. Figure 7
presents S/ton N0X removed, while Figure 8 presents busbar cost

(iril 1 s/kWh) associated with SCR N0X control. Both curves show the
strong effect of capacity factor on the cost as well as the effect of a
higher uncontrolled NQX emission' rate. Figures 7 and 3 indicate that
the lowest levelized SCR cost is for a high uncontrolled fJOx emission
rate at a high capacity factor. Cases A and C, at an uncontrolled N0X
rate of 0.30 Ib/MM Btu and a 65% capacity factor, exhibit the lowest
$/ton N0X removed cost of the cases presented.

Figures 9 to 12 report the coal case cost results; Figures 8 and 9 report capital
cost results while Figures 10 and 11 report levelized costs.

Figure 9 shows that the TCR varies from $101/kW to $78/kW. The results
show that the cost of achieving an emission limit of 0.12 lb: N0X/MM Btu
is about $9.1/kW less when combustion controls are applied to reduce the
boiler emission rate from 0.60 lb N0x/MM Btu to 0.40 lb N0X/MM Btu
(Cases A and B).

* The effect of catalyst cost is also shown on Figure 9. Case C reflects a
catalyst cost of $33G/ft3 while Case A results reflect a cost of
$660/ft3. The effect of the catalyst cost reduction is. a $23.2/kW

(23%) drop in TCR.

6A-111

-------
* Figure 10 shows the cost components that make up the TOR for Coal Case

A. The TPC makes up the largest component, of which the SCR Catalyst and
Reactor cost component comprises approximately 60%.

* Figures 11 and 12 report the levelized cost of removing N0X. Figure 11
shows that the lowest level ized SCR cost is for units with high
uncontrolled N0X emission rates. Comparison of the Case C and A
results show that reduction in the catalyst cost from $660/ft3 to
$330/ft^ would reduce the N0X removal cost by $61 5/ton N!GX removed.

* Figure 12 shows that a catalyst cost reduction from $660/ft^ reduces
the LBC by about 1.5 mil 1 s/kWh (Case A vs C).

* Figure 13 shows the significance of expected catalyst life on LBC for
both oil and coal. The expected catalyst life for the oil case is six
years, while the expected life of the coal case is four years. These
criteria show a dramatic decrease in the LBC when compared to previous
results based on expected catalyst life of one to two years.

CONCLUSIONS

The conclusions developed-in this study are:

1. The capital costs of retrofitting SCR to existing power plants in the U.S.
are expected to be:

A. $79 - 94/kW for oil-fired plants.

B. $78 - 101/kW for tangentially-fired coal power plants.

2. The levelized costs of retrofitting SCR to existing power plants in the U.S.
are expected to be:

A. 3.5 to 12.2 mils/kWh for oil-fired plants

B. 5.2 to 6.6 mils/kWh for tangentially-fired coal power plants.

3. The life cycle costs of SCR are driven primarily by capital costs for
oil-fired units, particularly for retrofit applications. The life cycle
costs of SCR for coal-fired units are equally effected by capital and
operating/maintenance costs. This conclusion differs from past evaluations
for new plants {EPRI CS-3603} in which high catalyst cost and a projected
catalyst lifetime of one year led to operating costs dominating life cycle
costs.

4. The installation of SCR N0X control is expected to have an effect on the
operation of downstream equipment including the air heater, I.D. fan, fly ash
handling and disposal, ESP, FGD, and wastewater treatment system.

5. Unresolved technical issues in the application of SCR control to coal-fired
power plants in the U.S. relate primarily to two issues:

A. High sulfur coals - The conversion of SO? to SO3, coupled with
already high levels of SO3 at the SCR inlet, can result in SO3
concentrations of 40 ppm or higher at the SCR outlet. These high SO3
levels could result in increased corrosion, air heater plugging, and
plume opacity problems.

6A-112

-------
B. Coal trace elements - The trace elements in ILS coals are significantly
higher, in many cases, than European or Japanese coals. These higher
trace element concentrations may significantly reduce the catalyst life
below the four years estimated in this study.

REFERENCES

1. Ando, J., "N0X Abatement for Stationary Sources in Japan",

EPA-60Q/7-79-2Q5, August 1979.

2. Babcock & Wilcox, "Steam/Its Generation and Use", The Babcock & Wilcox
Company, 1978.

3. Bauer, T.K., Spendle, R.G., "Selective Catalytic Reduction for Coal-Fired
Power Plants: Feasibility and Economics", Steams-Roger Inc., EPRI CS-3503,
October 1984.

4. Brown and Caldwell, "Process Design Manual for Nitrogen Control",
Environmental Protection Agency, EP 7.8/2:N59/975, October 1975.

5. Burke, J.M., Johnson, K.L., "Ammonium Sulfate and Bisulfate Formation in Air
Preheaters", Radian Corporation, EPA-6QQ/S7-82-025a, August 1982.

6. Cadrecha, M., "Preventing Acid Corrosion in Air Heaters", Power Engineering,
January 1980,

7. Cichanowicz, 0.E., Offen, G.P., "Applicability of European SCR Experience to
U.S. Utility Operation", Proceedings: 1987 Joint Symposium on Stationary
N0y Control, EPA/EPRI, New Orleans, 1987.

8. Electric Power Research Institute (EPRI), TAG - Technical Assessment Guide,
Volume I: Electricity Supply - 1986, EPRI P-4463-SR, December 1986.

9. Ellison, W,, "Assessment of SOg and M0X Emission Control Technology in
Europe", EPA-600/2-88-013, February 1988.

10. Goldschmidt, K., "VKR Full-Scale SCR Experience on Hardcoal Fired Boilers",
Proceedings: 1987 Joint Symposium on Stationary NO* Control, EPA/EPRI, New
Orleans, 1987.

11. Holcombe, L.J., Behrens, G.P. et al., "Manual for Management of Low Volume
Wastes from Fossil-Fuel-Fired Power Plants", Radian Corp., EPRI CS-5281, July
1987.

12. Howard, G. et al., "Proposed Rule 1135 - Emissions of Nitrogen Oxides from
Electric Power Generating Boilers", South Coast Air Quality Management
District staff report, August 1988.

13. Itoh, H., Kajibata, Y., "Countermeasures for Problems in N0X Removal
Process for Coal-Fired Boilers", Proceedings of the Joint Symposium on
Stationary Combustion N0y Control,'Volume II, OctoberT98Q.

14. Keeth, R.J., Balfour, D.A., "Utility Stack Opacity Troubleshooting
Guidelines", EPRI RP 2250-3, Draft Report, January 1989.

15. Metcalf and Eddy, Inc., "Wastewater Engineering: Treatment, Disposal,

Reuse", McGraw-Hill, 1979.

6A-113

-------
16. Nakabayashi, Y., Abe, R., "Current Status of SCR in Japan", Proceedings:
1987 Joint Symposium on Stationary N0X Control,, EPA/EPRI, New Orleans, 1987.

17. Necker, P., " Operating Experience with the SCR DeNOx Plant in Unit 5 of
Altbach/Deizisau Power Station", Proceedings: 1987 Joint Symposium on
Stationary N0X Control, EPA/EPRI, New Orleans, 1987.

13. Osborn, H.H., "The Effect of Ammonia SCR deNOx Systems on Ljungstrom Air
Preheaters", C-E Air Preheater, EPRI RP 835-2, June 1979.

19. Scheck, R.W., Damon, J.E., et al., "Technical and Economic Feasibility of
Anrnonia - Based Postcombustion N0X Control", EPRI CS-271 3, November 1 982.

20. Scheck, R.W., Mora, R.R., Belba, Y.H., Horney, F.A., "Economics of Fabric
Filters and Electrostatic Precipitators - 1984", Stearns Catalytic
Corporation, EPRI CS-4083, June 1985.

21. Singer, J.G., "Combustion: Fossil Power Systems", Combustion Engineering,
Inc., 1981.

22. Swann, D.R., Drissel, G.D., "Feasibility of Retrofitting Catalytic
Postcombustion N0X Controls on an 80-MW Coal-Fired Utility Boiler",
Stearns-Roger Incorporated, EPRI CS-1372, February 1980.

6A-114

-------
n/

: iiif&h

Figure 1. Oil Case: Plan and Elevation.

6A-115

-------
Figure 2, Coal Case: Plan and Elevation.

6A-1J5

-------
BOILER

NPHR Increase
Temp. Bypass
Pressure Control
Complexity
Reduced KW

FO PAN

• Higher Mass Flow

• Higher Hp Consumption

IP PAN

• New Fan For AP

• Bypass For Baiter
Pressure Control

AMMONIA STORAGE
• Operator Training
& safety

i NH3

AIR HEATER

• NH4 HSO4 Fouling

• N F Surface

• Water Wash Addition

• Sim, Sootblow
Addition

• Higher Leakage

• Higher Exit Gas Temp.
Higher AP

Additional Dampers for
On Line Wash

TO EXISTING
W & WM SYSTEM

WATER TREATMENT

Nitrogen Treating Capability

Figure 3, Oil Case: SCR Design/Operations Impact.

-------
AIR HEATER
Ammonium Bisulfate •
Fouling a

Higher Exit 6as Temp. .
Higher Leakage
Higher AP

Higher Steam Soolblow Rate
Higher Water Wash Rate
Higher Steam pressure &
Superheat

Additional Dampars For
On-Lino Wash

BOILER
NPHR Increase
Tenip. Bypass
Reduced KW

FD PAN

• 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

tO PAW

• Higher Mass Flow

• Higher Volumetric Flow

• Higher AP

REHEAT

• Higher Mass Flow

» Increased Steam Usage
FGP

• Volume Incroaso

• Higher Inlet Temp.

Increase In HjO Evap.
SO? Concentration Dilution
FGD Wastewater Treatment
For NH3

STACK
Increased Opacity
Higher S03

TO EXISTING
W&WM SYSTEM

WASTE TO
DEWATERING

Figure 4, Coal Case: SCR Design/Operations Impact.

-------

CATALYST COST, S/FT3

600

680

330

DENQX EFFICIENCY, *

BOILER OUT, #/MMBTU

0.30

0.18

0.30

Figure 5. Oil Case: Capital Cost Sensitivity,
1989 $.

100

OTHER jl.ai
CONTINGENCY (15 3)

ENGBQ. (e,B)

OENl. FACIL. (4.2)

TOTAL CAPITAL
REQUIREMENT

66.7

I other se e>
ID FAN (4.2)

* h moos.

NHS (3 8)

SCR
Catalyst
&

reactor

129.dj

TOTAL PROCESS
CAPITAL

Figure 6. Oil Case A: Capital Cost of SCR, 1989 $.

6A-119

-------
Thousands

M
I

L
L
S
I
K
W
H

20% 65%

C.F. C.F.

20% 66%

C.F. C.F
B

20% 66%

CJ. C.F.

ABC

CATALYST COST, $/FT"3 660 660 330
DENOX EFFICIENCY, % 90 80 90
BOILER OUT, LB/MMBTU 0.30 0.15 0.30

Figure 7. Oil Case: N0X Removal Cost,
Levelized 1989 $.

12.18

11.07

9.82

3.83

20% 65% 20%
c.r, c.f. c.r.
A g

65%

C.F.

20% 65%

C.F. C.F.

CATALYST COST, $/F"T3

660

330

DENOX EFFICIENCY, %

BOILER OUT, #/MMBTU 0.30 0.15 0.30

Figure 8. Oil Case: Busbar Cost Sensitivity,
Levelized 1989 $.

6A-120

-------
120
100
80

' #0

40
20

A B

CATALYST COST, $/FT'3

660

330

DENOX EFFICIENCY, %

BOILER OUT, #/MMBTU 0,60 0.40 0,60

Figure 9, Coal Case: Capital cost Sensitivity, 1989 $.

120

100

100.8

OTHER (2,8)
C0NTIN36NCV (18.7)

$ 80
/

W 60

SNQfla. I'M)
GHNl, F*CLL, (3,5)

70.7

total

PROCESS
CAPITAL
'JO.?)

TOTAL CAPITAL
REQUIREMENT

OTHER (0,8)
10 FAN [3 S)

A H MODS (7 6)

STRUCT STL.
I DUCT H3.2)

NH3 S3.2}

SOB
Catalyst
i

reactor

{*2,')

TOTAL PROCESS
CAPITAL

Figure 10. Coal Case A: Capital Cost of SCR, 1989 $.

6A-121

-------

CATALYST COST, $/FT3

660

330

DENOX EFFICIENCY, %

BOILER OUT, #/MMBTU

0,60

0.40

0.60

CAPACITY FACTOR, %

Figure- 11, Coal Case: N0X Removal Cost,
Levelized 1989 $.

8.61

CATALYST COST, $/FT"3

660

330

DENOX EFFICIENCY, %

BOILER OUT, #/MMBTU

0,60

0.40

0,60

CAPACITY FACTOR, %

Figure 12. Coal Case: Busbar Cost Sensitivity,
Levelized 1989 $.

6A-122

-------
20

L 10 -

o 2 4 6 8 10

CATALYST LIFE, years

Figure 13. Busbar Cost vs. Catalyst Life,
Levelized 1989 $.

6A-123

-------
Table 1
CASE DEFINITION

Plant Description
_Retrofit

° Capacity, MW (gross)

o Boiler Type

o A1r Heaters

o Particulate Control

o SO? Control

« Reheat

•> Gross Plant Heat Rate,

Btu/kKh

o Capacity Factor, 35

« Remaining Life, Years

Site Conditions

o Location

Fuel

Seismic Zone
Urban Site

Type

Higher Heating Value,
Sulfur Content, wt t

Btu/lb

Oil Case

Tes

400

Pressure

Horizontal, Twin

9004
20, 65
20

Southern CA

Yes

No. 6 Fuel Oi1

18,814

0.21

Coal Case

Tes

536

Tangential
Vertical, Twin
ESP

Wet Limestone
Steam

9197
65

Kenosha, Wl
I

Bit. Appalachian

13,100

2.6

Table 2
SCR PROCESS DESIGN

SCR Design Basis

tconomizer gut let Temp, 9 MCR, °F
Economizer Excess Air, %

Boiler N0X Emission Rate,

Ib/MM Btu
N0X concentration, ppmv (actual)
N0X Emission Limit, Ib/MM Btu
N0X Reduction Rate, %

NH3 Bleed Rate, ppmvd (& 316 Oj)
Guaranteed Catalyst life, years
Reactor Configuration

Anmonia Storage, days

SCR Design

Space velocity, SCF*/ft3-hr
Linear Velocity, actual fps
Surface area, nf/m-3
Pitch, imi

Catalyst layers (active + spare)
Soot Blowers

Ammonia Consumption, Ib/hr

SCR Cost Development
catalyst Cost, p/ft3
Expected Catalyst Life, years
Ammonia Cost, £/ton

Oil Study Case

Coal Study Case

654

725

12.5

0.30

0.15

0.60

0.40

312

161

572

381

0.03

0.12

Single,

Twi n

Twin

Vertical

3370

3900

2500

2960

21.0

18.2

555

470

5.94

7.07

4 + 1

yes

581

268

1,383

807

660

145

*SCF at 32"F

6A-124

-------
Session 6B
FUNDAMENTAL COMBUSTION RESEARCH

Chairman; B. Martin, EPA

\ 6B-i

-------
PREDICTION OF FUEL AND THERMAL NO
IN ADVANCED COMBUSTION SYSTEMS

R.D, Boardman and L.D. Smoot
Department of Chemical Engineering
350 CB, Brigham Young University
Prove, Utah, 84604

ABSTRACT

A computer model to predict in-situ and effluent nitric oxide (NO)
concentrations has been applied to advanced-concept, pulverized-coal
systems and evaluated by comparing model predictions with
experimental data. Specifically, the effects of pressure,
stoichiometric ratio, air stage location, and temperature were
predicted by the model and compared to experimental data". An
overview of the model theory and its approach is presented. The
basis for an expanded model including thermal NO and alternative fuel
NO mechanisms to explore the capabilities of predicting low rank coal
behavior and possibly fuel-staging (rebuming) is discussed. This
work is a continuation of previous model development and
evaluations.1,2

INTRODUCTION

A significant effort has been made at modeling gaseous pollutants,
including nitric oxides (NOx). For instance, NOx concentrations have
been predicted for simple laboratory-burner hydrogen, carbon
monoxide,1 and light hydrocarbon flames using well-defined
mechanisms.3'4'5 These approaches use scores of elementary reactions
to calculate the poo-1 of radical and stable oxy-hydroxyl,
carbonaceous and nitrogenous species. Figure 1 shows a partially
complete mechanism for the formation of NO from fuel nitrogen and
illustrates the radical species involved in the rate equations.
However, production of NOx in practical combustion devices is
significantly more complex and it would be extremely difficult to
employ such a comprehensive mechanism, particularly for prediction of
instantaneous radical species.

Consider a coal combustor in which nitrogen enters in the air and
also as a fuel-bound constituent. It is essential to- adequately
predict the governing transport phenomena (e.g. fluid mechanics, heat
transfer, and particle dispersion) and chemical kinetics involving
both heterogeneous and homogeneous reactions. An adequate description
of interactions: between the chemistry and turbulent fluid mechanics
also becomes imperative. Yet prediction of turbulent, coal-laden gas
flow through swirling jets using the best closure models available is
often insufficient. And even the most advanced devolatilization

6B-1

-------
models which describe the combined physical and chemical processes
are based on empirical correlations to macromolecular coal structure
and laboratory measurement of mass evolution. For example, the
FG/DVC {functional group/depolymerization-vaporization-cross-linking)
coal devolatilization model by Solomon et al.6 predicts species
evolution and tar formation, employing rank-independent kinetic
parameters, based on a number of general assumptions and coal
pyrolysis observations. Hence, the ability to accurately predict the
necessary spatia-lly—dependent , instantaneous radical species,
inherent in the complete kinetics mechanism, for turbulent coal
flames is currently not possible.

Because it is almost always impossible . to establish the exact
reaction network which is universally applicable to all combustion
configurations, it is desirable to reduce the nitrogen species
chemistry to a subset of only the precursors, intermediates and
products of greatest importance. The advantage of global kinetics,
based on only key and measurable intermediates, is the ability to
perform the calculations with greater efficiency and to account for
the effects of turbulence. Also, a common level of sophistication
between the coal conversion model and pollutant formation model is
maintained.

A method for predicting the conversion of coal-bound nitrogen to NO
was developed by Smith, Smoot, and Hill.7'8 This same theory has
also been adopted into other modeling codes.9'10 This N0X model is a
component of a general combustion code that provides theoretical
predictions for the temperature, velocity, major species, and other
properties at local points throughout turbulent combusting flow
fields. It assumes that a simplistic global kinetic model, explicit
only in HCN, O2, NO, and char surface area, is adequate to predict
the formation and destruction of NO. Sxtensive evaluation of this
model has been made and reported previously, 1'2 • 1 -1- Although a
promising degree of success has been demonstrated for this modelling
approach, efforts to improve and to generalize the NOx model for a
larger variety of'combustion applications are continuing.

MODEL DESCRIPTION AND THEORY

The strength of this approach to predict nitric oxide is that the
model is coupled to a comprehensive combustion model, PCGC-2
("Pulverized Coal Gasification and Combustion-2 dimensional"),
applicable to both two-dimensional laboratory and industrial-scale
practical combustors. PCGC-2 solves the set of general conservation
equations for a gas/particle-laden reacting axi-symmetric system.
Auxiliary equations are used to describe the chemical and physical
processes while the two equation k-£ turbulence model is used.

6B-2

-------
Simulations provide predictions for the mean gas field properties and
local particle properties for steady-state, turbulent flames. A
brief description of the model is given here to give general insight
into the comprehensive model.

The solution of the major field variables in PCGC-2 is based on the
"fast" chemistry assumption where all gaseous reactions go to
equilibrium locally, as soon as the reaetants are released to the gas
phase and mixed at a molecular level. Specification of the local
pressure, enthalpy, and elemental composition is sufficient to
determine the equilibrium composition, temperature, and fluid density
of the gas mixture. For turbulent flames, it can be reasonably
assumed that all species have equal turbulent diffusion coefficients.
Thus, the local elemental composition is determined throughout the
system by tracking the mixing of primary gas, secondary gas, and the
coal offgas. Two conserved scalars, f (primary mixture fraction) and
t] (coal offgas mixture, fraction), are used in PCGC-2 to track this
mixing. The energy equation is then solved to determine the
enthalpy. With knowledge of the instantaneous values of f, t\, and
enthalpy, the instantaneous values of species concentrations and
temperature are uniquely determined. A detailed physical and
mathematical description of the comprehensive model is given in a
number of references. ^ 3"?

In the existing N0X model theory of PCGC-2, fuel nitrogen release
from coal is assumed- to occur at a rate proportional to total coal
weight loss. The volatile nitrogen is assumed to be instantaneously
converted to HCN and is competitively reduced to N2 by reaction of NO
with HCN, The infinite kinetic rate assumption is no longer invoked.
In fact, random fluctuations in the temperature and species
composition dramatically influence the mean reaction rates. A method
for obtaining mean reaction rates from information provided by the
progress variables is possible, provided the instantaneous
temperature and species concentrations are unique functions of the
instantaneous progress variables. While PCGC-2 does not solve
directly for the instantaneous values of the progress variables, a
statistical distribution of the instantaneous values about the mean
can be hypothesized. Given the'mean value of the progress variable
(first moment), its variance about the mean {second moment) and an
assumed shape (called a probability density function - pdf) for its
distribution, the values of the instantaneous properties can be
statistically determined. Both the first and second moments are
provided by the turbulence model. Experimental'* and modeling
experience has provided information for choosing the shape of the
probability density function. Typicallv, a clipped Gaussian
distribution with Intermittency for pure primary and pure secondary
air is selected for the pdf. The mean rates are thus obtained by
integrating the instantaneous rates over the statistical probability
of the progress variables. When two progress variables are involved,
the integration is made over a double integral which expresses the
joint probability of each progress variable. The overall mechanism
is shown in Figure 2 and key model equations are listed in Table I.

68-3

-------
PREDICTIONS IN ADVANCED COMBUSTION SYSTEMS

Initial evaluation of the NOx model has demonstrated the ability of
the model to predict the'effects of increasing stoichiometric ratio
(S.R.) for both swirling and non-swirling diffusion flames in single-
stage, laboratory-scale and pilot-plant combustors,2 Experimental
NO data passed through a minimum at a swirl number of 2.0 - 3.0,
depending on the coal type. At a S.R. of 1.0, the model was also
shown to reliably predict the effects of particle size on NO
emissions and also the trend of increasing NO concentrations with
increasing coal moisture content.1

In order to further investigate the utility of the NQX model,
application and evaluation has been made for advanced-concept, coal-
combustion systems of entrained-flow gasification, fuel-rich and
staged combustxon, and C02—02 oxxdx^er combust x on. Specxfxcally, the
effects of pressure, stoichiometric ratio, air stage location, and
temperature were predicted by the model and compared to experimental
data. Experimental data sources for model comparisons were selected
for axi-symmetric combustion experiments where systematic variation
of key test variables (e.g., pressure or S.R.) were reported. Table
XI lists the combustion configurations, reactor dimensions, coal
types, firing rates, and other parameters. Comparisons of predicted
NO concentrations with experimental measurements were made after
acceptable agreement for the main stream variables was
demonstrated.13

NON-STAGED, AIR-COAL COMBUSTION

Cases were simulated for non-staged combustion at fuel-lean (S.R. =
1.20), near stoichiometric ratio {S.R. = 0.95) and moderately fuel-
rich {S.R. = 0.80) conditions. Figure 3 compares the theoretical and
experimental effluent NO concentrations over the range of
stoichiometric ratios shown. The measured data and axial profiles
are compared in Figure 4 for each case. For the fuel-lean case, a
high peak NO concentration is predicted initially but the profile
rapidly decays to approximately 8 percent above the experimental
profile.' A similar trend is predicted for S.R. = 0.95, except that
the NO profile decays to a level 10-15 percent below the measured
data. The predicted NO concentrations also follow this trend for the
SR = 0.80 case but are consistently lower by 15 to 30 percent along
the entire profile. The discrepancies between the measured and
predicted NO concentrations in the early region of the reactor may be
due to non-isokinetic sampling. The reactor was designed to
establish one-dimensional flow;14 however, it is possible that radial
stratification of the flame occurred in the vicinity of the burner.

6B-4

-------
In fact, the NO concentration profile predicted near the wall closely
matched the experimental data throughout the entire reactor,
including the early region. A separate explanation for this
discrepancy may be a result of assuming that fuel nitrogen is evolved
at a ' rate proportional to the total coal mass loss, Wendt15 found
that at low temperature, the early volatiles are nitrogen-free, If
the rate of nitrogen release is lower than the total coal mass loss,
then the model could overpredict the formation of NO in the near-
burner region.

AIR STAGED COMBUSTION

Staged combustion cases were accomplished by injecting additional air
downstream of the fuel-rich primary zones. Two simulations each were
completed for primary zone S.E. values of 0.95 and 0.80 respectively.
The secondary-zone air was injected into the reactor at axial
locations corresponding to residence times of 0.52 s and 0.90 s,
bringing the overall reactor S.R, to 1.20. Experimentally, it was
demonstrated that secondary-stage NO concentrations were independent
of primary zone NO levels14; however, effluent NO concentrations were
decreased as the first stage was lengthened because of the fast rate
of NO decline in the fuel-rich, high-temperature, primary zone.
Figure 5 shows the incremental reduction in NO at the exit for both
primary-zone stoichiometric ratios. A difference of ~200 ppm NO was
measured between the secondary-air stage locations of 0.52 s ana 0.90
s for each case. The predicted difference in NO concentration at the
sxxt closely matches the measured trend.

TEMPERATURE DEPENDENCE

The effects of temperature on fuel NO have been elucidated in a
number of investigations. From data.of fuel-lean (15% excess oxygen)
pulverized-coal diffusion flames, fuel NO was observed to be
essentially constant over a temperature range up to 2550 K for four
coal types.16 One of the coal types investigated was Colorado
bituminous coal. Model simulations have also been made for Colorado
bituminous coal, in a separate reactor but at similar operating
conditions using nitrogen substituted carrier gas at 15% excess
oxygen. A similar temperature insensitivity is shown in Figure 6 for
the model predictions which closely follows the trend measured by
Pershing and Wendt.16

ROLE OF HETEROGENEOUS DECAY

Predictions were examined to determine the- relative difference
between homogeneous and heterogeneous NO reduction. In most fuel-
rich locations, where concentrations of HCN were appreciable, the
heterogeneous decay was found to be insignificant compared to the
magnitude of homogeneous NO decay. This result is consistent with
the experimental observation17 that gas-phase destruction of NO is

6B-5

-------
the dominant NO reduction path in fuel-rich coal combustion.
Homogeneous decay was far more significant for the gasifier
predictions because of the high concentration of fuel-rich species.
However, for the non-staged combustion cases, homogeneous decay was
only initially more significant but became relatively less important
as the fuel-rich species were consumed. Thus, in the aft section of
the combustor, heterogeneous MO decay also became xrnportant.

ENTRAINED-FLOW GASIFICATION

Four simulations of three coal types were completed. Figures 7 and
8, respectively, compare the experimental and theoretical NO
concentration profiles for atmospheric and high pressure (5 atm)
gasification of Utah bituminous coal. Predicted and measured peak NO
concent rat xons are s x m x 1 ar xn Tnagn_tude and 1 ocat xon . Predx c t ed
concentration contours closely match the experimental maps for the
atmospheric case throughout the entire reactor while the high
pressure case also matches the peak NO value but decays somewhat
slower. A predicted exit concentration of 100 ppm is approximately
30 percent below the measured value for the atmospheric pressure case
while the predicted exit value of 4 ppm differs by only two ppm from
the measured value for the high pressure case.

Results for the simulation of Illinois bituminous coal were similar
to the Utah bituminous coal. However, for North Dakota lignite,
although the major species and flame structure were predicted
adequately, Figure 9 shows that both the location and magnitude of
peak NO concentration were incorrectly predicted. Instead, the
predictions resemble those of the bituminous coals. It has been
observed that lignites, relative to bituminous coals, produce larger
quantities of NH3 during gasification.It was also observed in
this study that NH3 concentration measurements for the North Dakota
Lignite were 200-300 percent higher than for the Utah and Illinois
bituminous coals. Evidence suggests that the initial release of NH3
and HCN provide an estimate of the amount, of nitrogen reduction
occurring in the vicinity of the reacting particles.Available
data suggest that the gas-phase fuel-nitrogen reaction sequence is
initiated by a rapid and nearly quantitative conversion of the parent
fuel nitrogen compounds to hydrogen cyanide and ammonia. HCN appears
to be the principal product when the fuel nitrogen is bound in a
aromatic ring while NH3 is the principal product when the fuel
nitrogen is bound with amines.20 As coal rank decreases, the number
of rings also decreases and hence the conversion of fuel nitrogen to
NH3 rather than HCN. Additionally, the rate of nitrogen release from
lignite coal likely does not occur at a rate proportional to total
coal mass loss. Since the existing NO model mechanism considers only
HCN, NO, and O2 ' species, the inclusion of NH3 in the global rate
expressions as well as an advanced devoiatilizaticr. model to predict
nitrogen species devolatilization may improve the model for low rank
coals simulations.

6B-6

-------
MODEL IMPROVEMENTS

From comparisons of model predictions to a diversity of experimental
data, a number of potential improvements to the N0X submodel have
been identified. It is recognized that the simplified mechanism will
not adequately account for the recycle of NO and KCN via reaction
with hydrocarbons and nitrogen intermediates as postulated by Bartok
and Folsom21 and Kramlich et ai.2 9 Second, since possible
explanations for the discrepancies noted above for the prediction of
lignite coal gasification suggest the importance of intermediate
species NHi, a more complete global fuel-NO mechanism is being
sought. Finally, the incorporation of thermal NO would further
improve the usefulness of model for predicting gas and fuel-oil
practical combustors. Improvements to the major combustion code,
PCGC-2, and the devolatilization model are being sought as part of
ongoing efforts to develop generalized combustion simulators.22

JOINT FUEL AND THERMAL NO MODEL

Recent modifications to the theory have included the addition of the
extended Zeldovich mechanism to the N0X model. Simplifying options
have been inel-uded in the code to allow investigation of techniques
to predict radical oxygen which appears in the rate equation.
Radical O will be estimated by assuming partial equilibrium
relationships among pertinent species according to the method
demonstrated by Saroflm and Fohl23. With provisions for thermal NO
formation incorporated in the model it is possible to predict the
joint formation of fuel and thermal NO. Figure 10 shows a comparison
of a model prediction with experimental data for gasification of Utah
bituminous'coal. Joint prediction of fuel and thermal NO increased
the peak centerline NO concentration by 25 percent over the value
predicted by fuel NO alone. And, although the predictions do not fit
every datum well, it is clear that the addition of thermal NO to the
theory improved the model prediction throughout the entire profile.

A study is planned to measure thermal NO in a:turbulent gaseous flame
reactor with heating rates equivalent to industrial combustors. This
effort should provide useful information for selecting suitable
partial equilibrium assumptions when predicting radical oxygen
concentrations.

ALTERNATIVE FUEL NO MECHANISM

Although a promising degree of success has been demonstrated with the
fuel NO mechanism postulated by Smith et al.7, for predicting
formation and destruction of NO in bituminous coal flames over
diverse conditions, predictions for low-rank coals have often lacked
good quantitative agreement as demonstrated earlier. One explanation
for this has been suggested to be the exclusion of NH3 as a species
in the global model of Figure 2. 'A review of NOx mechanism

6B-7

-------
investigations has been made with the purpose of identifying possible
alternative global fuel-NO mechanisms which include NH3 in the
correlated rate equations.

Fenimore24-2*', measured traverses of HCN, NH3, and NO throughout the
burnt gas of fuel-rich gaseous flames and postulated decay reactions
for each species. According to this mechanism, HCN and NO are
destroyed by the sequence:

HCN + OH —> NH3 + products (1)

NH2 + no -» N2 + H2O (2)

These global rates apply specifically to the decay of nitrogenous
species but do not account for either the formation of NO or HCN.
Investigators have applied the Fenimore reactions to coal systems to
determine better rate parameters .17' 29-31 jn a study of fuel-rich
combustion - of pulverized coal, Dannecker and Wendt (1984) determined
that although there is some variation from coal to coal, their data
support the hypothesis that N2 and NH3 formation and NO destruction
follow the Fenimore mechanism. The destruction of NO appeared to be
approximately first order which was consistent with Fenimore's
mechanism only if NH2 is constant and can be predicted,

Although variations of Fenimore kinetics are useful for predicting
the decay of NH3 and NO in the post-flame region, it is also
necessary to predict the formation rates of these species. From the
literature, two simplified mechanisms involving NH3 were identified
that could be potentially adapted to the existing N0X model. The
first mechanism is the reaction sequence shown in Figure 11
postulated by Miller et al.32 and includes the dominant reactions of
those shown in Figure 1 for converting HCN to NO and N2. This
simplified mechanism was determined by performing sensitivity and
rate-of-production analysis to match experimental data. Favorable
results have been shown for a one~dimensional, premixed, laminar
flame, a perfectly stirred reactor, and the time history of a
homogeneous reaction.3 Although this mechanism is based on well-
established elementary rate with known kinetic parameters,
unfortunately, even this simplified mechanism requires the
predictions of radical species H, 0, and OH. Currently, it has not
been shown possible to calculate these radicals a priori in turbulent
coal flames.

A global model has been presented by Mitchell and Tarbell33 involving
NH3, HCN, NO, and N2 as active centers. Figure 12 shows the global
mechanism formulated by Mitchell and Tarbell. By assuming the form
of the rate equations, which included formation of thermal NO
assuming partial equilibrium of radical oxygen with molecular oxygen,
Mitchell and Tarbell calculated the "best-fit" of relevant data to
obtain the kinetic parameters for fuel NO formation and destruction.
Extensive data of Muzio et al.,^4 covering a broad range of
temperatures and ammonia-NG stoichiometries in CH4/excess-air flames

6B-8

-------
in a plug flow reactor were correlated to rate equations having the
global form:

Since the rates were not each established through separate,
independent measurements, the applicability of the mechanism remains
questionable. It was determined to investigate the feasibility of
this alternative mechanism by comparing it to predictions of the
existing fuel NO mechanism of Hill, Smoot, and Smith.1 Table III
compares the two fuel NO mechanisms and lists specific advantages and
disadvantages of each. A program scheme has been developed to allow
the option of selecting between the fuel NO mechanism of Smoot,
Smith, and Hi.ll (References) and that of Mitchell and Tarbell.
Presently, simulations are being made to Investigate the later model.
Although it does provide a path for the recycle of NO back to HON, it
is still not considered to be completely sufficient for predicting
fuel-staging applications. It does include thermal NO contributions
assuming the extended Zeldovich mechanism with radical oxygen in
partial equilibrium with molecular oxygen only.

Figure 13 summarizes the revised' NO model which has been expanded to
include rates of thermal NO formation and an optional fuel NO
mechanism which is currently being investigated. The relationship to
the generalized combustion code is illustrated.

Given the complexity of chemical reactions in practical, turbulent
coal flames, it is difficult to apply comprehensive kinetic
mechanisms, involving undeterminable trace and radical intermediates,
to predict the formation and destruction of NO in a manner
universally applicable to combustion configurations. Hence a global
fuel NO mechanism has been used to predict NO concentrations in
pulverized coal flames. Simulation results have shown favorable
agreement with experimental measurements for a variety of combustion
configurations for bituminous and sub-bituminous coals. With the
exception of low rank coals and reburning configurations, the model
may be used to predict NO emissions and to investigate the impact of

ri V* a n ^ra e» 4 r\T%£i v a 4™ i in ft ffa 9 V\ 1 a e on as* a •¥• "5 <*¦" Vv 4 /"ntts £5 -1 v" ^ W. t * •v %-

iiy&!o a.xi ojpfsratiriy variciijiss oUwH c*o SLuicniorncuit i c i_9.Ti.iOf jdijiitri02.
swirl number, reactor pressure, and air staging. Efforts are being
made to further generalize the code by including the extended
Zeldovich mechanism to predict thermal NO formation and an expanded
global fuel NO mechanism into the code. The goal is to provide a
model that can be used to aid in optimizing N0X control strategies.

NH3 + O2 -» NO + H2O + 1/2H2
NH3 + NO -» N2 + H20 + 1/.2 H2

(3)

(4)

SUMMARY

6B-9

-------
ACKNOWLEDGEMENTS

This work is supported by Morgantown Energy Technology Center
(Contract No. DE-AC21-8SMC230 75 under subcontract from the Advanced
Fuel Research, Co. Financial support for the model evaluation came
from a Consortium of eleven organizations (Advanced Fuel Research,
Babcock & Wilcox, Combustion Engineering, Consolidation Coal Co.,
Electric Power and Research Institute, Empire State Electric Energy
Research Co., Foster Wheeler, Tennessee Valley Authority, U.S.
DOE/METC, U.S. DOE/PETC, Utah Power & Light).

REFERENCES

1. Hill, S.C., L.D. Smoot, and P.J. Smith, Prediction of Nitrogen Oxide Formation in Turbulent
Coal Flames," Twentieth Symposium (International) on Combustion, The Combustion Institute,
Pittsburgh, PA, 1391 (1984).

2. Smith, P.J., L.D. Smoot, and S.C. Hill, "Effects of Swirling Flow on Nitrogen Oxide
Concentration in Pulverized Coal Combustors," AlChE Journal, 1917 (1986).

3. Miller, J.A., and Bowman, C.T., "Mechanism and Modeling of Nitrogen Chemistry in
Combustion," Western States Section/Combustion Institute, Dana Point, CA (1988).

4. Peck, R.E., P. Glarborg, and J.E. Johnsson, "Kinetic Modeling of Nitrogen Oxide Formation in
Coal-Dust/Oxidizer Flat Flames,"Western States Section/Combustion Institute, Honolulu, HA
(1987).

5. Glarborg, P., J.A. Miller, and R.J. Kee, "Kinetic Modeling and Sensitivity Analysis of Nitgrogen
Oxide Formation in Well-Stirred Reactors," Combustion and Flame, 65,177 (1986).

6. Solomon, P.R., Hamblen, D.G.., Carangelo, R.M., Serio, M.A., and Deshpande, G.V., "General

Model of Coal Devolatilization," Energy & Fuels, 2, 405, (1988).

7. Smith, P.J., and S.C. Hill, and L.D. Smoot, "Theory for NO Formation in Turbulent Coal
Flames," Nineteenth Symposium (International) on Combustion, The Combustion Institute,
Pittsburgh, PA, 1263 (1982).

8. Hill, S.C., "Modeling of Nitrogen Pollutants In Turbulent Pulverized-Coal Flames," Ph.D.
Dissertation, Department of Chemical Engineering, Brigham Young University, Provo, UT
(1983).

9. Zinser, W. and U. Schnell, "Application of Mathematical Flame Modeling to NOx Emissions from
Coal Flames," Fundamentals of Physical Chemistry of Pulverized Coal: J.Lahaye edt., NATO ASI
Series, Martinus Nijhoff Publishers (1987).

10. Fiveland, W.A. and Wessel, R.A., "Numerical Model for Predicting the Performance of Three
Dimensional Fuel Fired Furnace," Eng. for Gas Turbines and Power, 110, No. 1, 117, (1988).

11. Boardman, R.B. and L.D. Smoot, "Prediction of Nitric Oxide in Advanced Combustion Systems,
"AIChE Journal, 34, No.9, 1573, (1988).

6B-10

-------
12. Smoot, L.D. arid P.J. Smith, Coal Combustion and Gasification. Plenum Press, New York, NY
(1985).

13. Boardman, R.D., "Further Evaluation of a Predictive Model for Nitric Oxide Formation During
Pulverized Coal Combustion," Master's Thesis, Brigham Young University, Provo, UT (1987).

14. Wendt, J.O.L, J.W. Lee, and D.W. Pershing, "Pollutant Control Through Staged Combustion of
Pulverized Coal," U.S.Department of Energy Technical Progress Report for Period January 1977 -
December 1977, FE-1817-4, (1978).

15. Wendt, J.O.L.,"Fundamental Coal Combustion Mechanisms and Pollutant Formation in
Furnaces," Prog. Energy Combustion Sci., 6, 201 (1980).

16. Pershing, D.W. and J.O.L. Wendt, "Pulverized Coal Combustion: The Influence of Flame
Temperature and Coal Composition on Thermal and Fuel NO," Sixteenth Symposium
(International on Combustion, The Combustion Institute, Pittsburgh, PA, 389 (1977).

17. Dannecker, K.M. and J.O.L. Wendt, "Fuel Nitrogen Mechanism During the Fuel Rich
Combustion of Pulverized Coal," paper 119c presented at the AIChE Annual Meeting, San
Francisco, CA (1984).

18. Freihaut, J.D., W.M. Proscia, and D.J, Seery, "Fuel Bound Nitrogen Evolution During the
Devolatilization and Pyrolysis of Coals of Varying Rank," 1987 Joint Symposium on Stationary
Combustion NOx Control, New Orleans, LA, March 23 - 26 (1987).

19. Kramlich, J.C., Lester, T.W. and Wendt, J.O.L., "Mechanisms of Fixed Nitrogen Reduction in
Pulverized Coal Flames,"1957 Joint Symposium on Stationary Combustion NOx Control, New
Orleans, LA, March 23 - 26 (1987).

20. Axworthy, A.E. and Dayan, V.H., "Chemical Reaction in the Conversion of Fuel Nitrogen to
NO*: Fuel Pyrolysis Studies," Second EPA Stationary Source Combustion Symposium, Sept.,
(1977).

21. Bartok, W. and Folsom, B.A., "Gas Reburning-Sorbent Injection - A Combined NOx/SOx
Control Technology ,'7957 Joint Symposium on Stationary Combustion NOx Control, New
Orleans, LA, March 23 - 26 (1987).

22. Solomon, P.R., Serio, M.A., Hamblen, D.G., Smoot, L.D., and Brewster, B.S., "Measurement
and Modeling of Advanced Coal Conversion Processes," 2nd Annual Report, D.O.E./M.T.C.E.
contract DE-AC21-86MC23075, (Sept, 1988) .

23. Sarofim, A.F., and J.H. Pohl, "Kinetics of Nitric Oxide Formation in Premixed Laminar Flames,
"Fourteenth Symposium (International) on Combustion, The Combustion Institute, Pittsburgh,
PA, 739 (1973).

24. Fenimore, C.P., "Studies of Fuel-Nitrogen Species in Rich Flame Gases," Seventeenth
Symposium (International) on Combustion, The Combustion Institute, Pittsburgh, PA, 661
(1979).

25. Fenimore, C.P., "Reactions of Fuel-Nitrogen in Rich Flame Gases," Combust, and Flame, 26,
249 (1976).

6B-11

-------
26. Fenimore, C.P., "Formation of Nitric Oxide from Fuel Nitrogen in Ethylene Flames," Combust,
and Flame, 19, 289, October, (1972).

27. Fenimore, C.P., "Formation of Nitric Oxide in Pre-mixed Hydrocarbon Flames," Thirteenth
Symposium (International) on Combustion, The Combustion Institute, Pittsburgh, PA, 373
(1971).

28. Fenimore, C.P. and H. A. Fraenkel, "Formation and Interconversion of Fixed Nitrogen Species in
Laminar Diffusion Flames," Eighteenth Symposium (International) on Combustion, The
Combustion Institute, Pittsburgh, PA, 143 (1981).

29. Glass, J.W. and J.O.L. Wendt, "Mechanisms Governing the Destruction of Nitrogenous Species
During the Fuel Rich Combustion of Pulverized-Coal," Western States Section/Combustion
Institute, University of Utah, Salt Lake City, UT (1982).

30. Bose, A.C., K.M. Dannecker, and J.O.L. Wendt, "Coal Compostition Effects on Mechanisms
Governing the Destruction of NO and Other Nitrogenous Species during Fuel-Rich Combustion,"
Energy and Fuels, 2, 301 (1988).

31. Bose, A.C. and J.O.L. Wendt, "Effect of Coal Compostition on Mechanisms Governing the
Destruction of Nitrogenous Species During Staged Combustion," Western States
Section/Combustion Institute, Honolulu, HA (1987).

32. Miller, J.A., M.C. Branch, W.J. McLean, D.W. Chandler, M.D. Smooke, and R.J. Kee. "The
Conversion of HCN to NO and N2 in H2-02-HCN-Ar Flames at Low Pressure,"Twentieth
Symposium (International) on Combustion, The Combustion Institute, Pittsburgh, PA, 673
(1984).

33. Mitchell, J.W., and J.M. Tarbeil, "A Kinetic Model of Nitric Oxide Formation During Pulverized
Coal Combustion," AIChE Journal, 28, 302 (1982).

34. Muzio, L.J., J.K. Arand, and D.P. Teixeira, "Gas Phase Decompostion of Nitric Oxide in
Combustion Products," Sixteenth Symposium (International) on Combustion, The Combustion
Institute, Pittsburgh, PA, 199 (1977).

35. de Soete, G.G., "Overall Reaction Rates of NO and N2 from Fuel Nitrogen," 15th (International)
on Combustion, The Combustion Institute, Pittsburgh, PA, 1093 (1975).

36. Zeldovich, Y.B., P.Y. Sadovnikov, and D.A. Frank-Kamentskii, "Oxidation of Nitrogen in
Combustion" (translated by Shelef), Academy of Sciences of USSR (1947).

37. Smoot, L.D., "Modeling of Coal-Combustion Processes," Prog, Energy Combust. Sci., 10, 229
(1984).

38. Levy L.M., Chan, L.K., Sarofim, A.F., and Beer, J.M., "NO/Char Reactions at Pulverized
Coal Flame Conditions," Twentieth Symposium (International) on Combustion, The Combustion
Institute, Pittsburgh, PA, 111 (1981).

39. Brown, B.W., "Effects of Coal Type on Entrained Gasification," Ph.D. Dissertaion, Brig ham
Young University, Provo, UT (1985).

40. Nichols, K.M., Hedman, P.O., and Smoot, L.D., "Release and Reactions of Fuel-Nitrogen in a
High-Pressure Entrained-Coal Gasifier," Fuel, 66, 1257 (1987).

6B-12

-------
Figure 1. Partially complete mechanism for conversion of fuel NO
to NO and N2. (Source: Miller and Bowman3)

Figure 2. Global fuel NO mechanism of Hill, Smith, and Smoot,1,2
using homogeneous rate expressions measured by
de Soete,35 and char reduction rate expression
measured by Levy et al.38

6B-13

-------
Stoichiometric Ratio

Figure- 3. Conparison of predicted effluent NO concentrations for
various reactor stoichiometrics with experimental data
of Wendt et al.

v_< ^

0) Q
5 X3
®

£ Q_

a '

^ i

zn \
c '

•2 "D

ra ®

c 5
o m

o Qj

c 5

o ^

O >

o *

¦g
'x
O

1500
1C0D
500
0

J 0 « % °o

00 o

[•

S.R.=1.20
, 1 1 L

__ _L, -

1,2

1200
800
400

° o 9

S.R.=0,95

1.2 1.4

.4 .6 ,8 1 1.2 1.4 1.6

Residence Time (s)

1 8

Figure 4. Comparison of predicted centerline NO. profiles with
experimental data of Wendt et al.H

6B-14

-------
900

a 600

1 300

0
0

Figure 5. Comparison of predicted, effluent NO concentrations for
air-staged combustion with measured data of Wendt et ¦
ai, 14 Secondary air injection located downstream from
the primary burner at axial positions corresponding
to residence times of either 0.52 sec. or 0.90 sec.

, , , r

Primary zone S.R. = 0.95
0 O Measured

n ¦ « —k»4

Primary zone S.R. ¦ 0.80
' ~ Measured

¦ Predicted

¦ 1 1 ¦ ¦ 1
.3 0.6 0.9

Residence Time of Stage Location

I. 700

| 600

f 700

o 600

Predicted

•

Experimental ~

1900 2200 2500 2800

Adiabatic Flame Temperature (K)

Figure 6. Temperature insensitivity of effluent NO

concentrations for fuel-lean combustion of Colorado
bituminous coal in CO2-O2 oxidizer. Measured data are
from Pershing and Wendt*®. Test conditions for
measured and predicted cases differed.

68-15

-------
Maasurad NO {ppm. dry basis)

140

20 40 60 BO 100 120 140 160 130

Gasifier Axial Distance, cm

Figure

Predicted NO
(ppm, dry basis)

0 20 40 60 30 100 120 140 160 180

Gasifier Axial Distance, cm

7, Comparison of predicted, and measured NO concentrations
during atmospheric gasification of Utah bituminous
coal. (Measured data of Brown39)

10
B
6
4

Measured NO
(ppm, dry basis)

i i i

0 1 20\ \40 60 80 100 120 140 160 130
3000 1000 500

Gasifier Axial Distance, cm

0/20 40 60 80 100 120 140 160 190
2000

Gasifier Axial Distance, cm

Figure 8, Comparison of predicted and measured NO concentrations
during pressurized (5 atm) gasification of Utah
bituminous coal. (Measured data of Nichols et al.40)

6B-16

-------
Gasirier Axiai Distance, cm

£ 10

s i a

^>5 6

—

-* C

U 0/ A

•MiJ *

m £

100

JL.

Predicted NO
(ppm, dry basis)

_i_

-L

20 40 60 80 100 120

Gasifier Axial Distance, cm

140

160'

1 80

Figure 9. Comparison of predicted and measured NO concentrations
during atmospheric gasification of North Dakota
lignite, (.Measured data of Brown^9)

£
a

1500

1200

900

1 , , 1 , 1 1-

i—'—r

~ •

•' Experimental N0X

O Predicted Thermal and Fuel NO
~ Predicted Fuel NO only

A .6 fl 1 1 2 1 .4 1 .6

Reactor Center-line Distance (m)

1 8

Figure 10.Comparison of predicted fuel NO and joint prediction
of thermal and fuel NO with experimental data,
(Measured data (circles) of Brown39)

6B-17

-------
+H.+OH,

HCN

NCO

Figure 1X,Simplified kinetic mechanism for conversion of fuel
"nitrogen to NO and N2. (Source: Miller and Bowman3)

Figure 12.Global fuel NO mechanism correlated by Mitchell and
Tarbeli33 to reactor data of Muzio et al,3^

6B-18

-------
I NO* MODEL I

Choose Fuel NO Mechanism.

Generalized Combustion Code
PCGC-2

/Specify Theimal NO Optooiu
. I. select partial cquilibciura method
far calculating mdkal [O]
£ select thermal NQ

* Pnmda velocium, npomin, aeruilibtiiai
gia phoe spades aaacajuuima thjoogbasu
tfaerauxardaauta.

* Pnowidef ptsicl® njectacy binary- taupeiiuttB
twawWi c-aiBpoiiaap, cq4 surface mk

* Provides uubuLeoce infonnmcis- mem cod var-
imm of prognss vniables used so track the gia.

Zeldovich Mechanism
forward rait* only

1. N2 ~ O"** NO + N

2. N + 02-*» NO* O

Swadtd ZtldgYiCh M«?h.

forward md rercriB rales

1.N2»0-»—KO + N

2. N + 02•••HO t O

3. OH NO+ H

Reproduced Irom
best available copy.

Mitshtll mi Tarbcll (19S2)

GlobaJ Reaction*

1.co«Liii230gaB'a>aia< HCN

2. HCN * 02 NM3 ~ ...

3. NH3 + 02 "** MO +...

4. NO + NH3 * N2+...
ch« + NO N2*...

€. NO + CmKn HCN+...
Therms NO Ratal
(not optional)

Modified Zaldovidj Modi.
Isewwdi «id nvne rues
paroil aquiUbrnim of [0]
«iib[02)

1. N2 + 0 HQ + N

2.N + 02 —•"•NO + O

3. OH + N NO + H

p^Ompiu HQ. NH3, HCN predicted profile*

Figure 13.Framework of an expanded fuel and thermal NO model
coupled with PCGC-2.

Table I. Equation set for NO formation and reduction model in
turbulent coal-laden, reacting flow processes.1'2

Mass balance For 02

(A)

^1™ '^NO

Species continuity

'^'hcn ~ "*&[ —C'y MHcn

Rate of nitrogen

(E)

» &>N5a/Mv

release from the

coai

Instantaneous

a?, - piix \Qtx)XHCN^^xpC-67.0 kcal/^n/^

reaction, rate lerms

s£>2 - IO'^Xhcn^no exp(-60 0 kcal//?T)/

Mean reaction rate

(H)

~ P !/ Jn P* d/ d?i

term by convolution'

over probability

density Functions

NO-Ghar reduction

wj-0^4,18 x I07Mfp^exp(-34 7 kcal/flf)

reaction

Instantaneous

X-, - Y,MjMt

mole fractions

Instantaneous

Y, * ir,- rf

mass fractions

10,

Deviation from

(L)

iTj - f/yf

fully reacted

mass fractions

6B-19

-------
Tabla II.

SUMMARY OF EXPERIMENTAL CASES SELECTED FOR MODEL EVALUATION

Combustion
Configuration

Gasification
L.1.89 m
D-0.2 m

Non-Staged

Air-Staged
1=2.2 m
D=0.15 m

Data Source
References

C02-02 Oxidizer Berry at at. (1986)
L=2.74 m
D=0.S m

Number Coal Feed
of Cases/ Rate
Figure NQ, (Ko/hr)

Brown (1988)

Nichols at at. (1987)
Azuhata etal. (1986)

Wendt era/. (1978)

Wendt era/. (1978) "

4/7,8,9

3/3,4
4/5

4/6

5
5

30-35

Reported NO
Concentrations

2-0 Profile

1-D Profile
1-D Profile

Effluent

Parameters
investigated

Pres. (1-5 atro),
Temp., Utah bit.,
III. subbil., N.D. Lig,

S.R. (0.8-1.2),
Kentucky bit.,
Secondary-Air
Stage Location
(0.52s - 0.90 s)

COj-Oj ratio
Colorado subbit.

L ¦ Reactor Length; D ¦ Reactor Diameter

Tabid XII. Advantages and Limitations of NQX model Fuel NO Mechanisms.
Hill, Smoot, and Smith Mechanism. 1,2

1.Global in intermediates, HCN, NO, 02, all which are measurable

2.Fundamental kinetic parameters obtained by parametric variation of species
in a well designed experiment35

3.Implicitly accounts for prompt NO formation
T.imil- at.inns

1.applicable only to systems where [HCN] » {[NO'] + [NH^] J

2.Requires fuel nitrogen to be converter to HCN before decaying to
intermediate and product nitrogen species

3.Recycle of NO back to HCN (e.g. reburning) not possible

4.Does not include thermal NO formation

Mitchell and Tarbell Mechanism il

My siH b 8

1.Global in intermediates, HCN, NH,, NO, and 0,

3 2

2.Accounts for dependence of reactions on NH3 concentrations

3.Accounts for recycle of NO to HCN through interaction with 2mHn.

4.Includes thermal NO in the rate equations

1.Kinetic parameters obtained by empirical correlation to relevant species

2.Dependence on CmHn sensitive to fuel type with coefficients m and n not
explicitly determined

3.Does not account for prompt NO formation either explicitly or implicitly

3.Disallows for thermal and fuel NO effects to be calculated separately

4.Thermal NO rates correlated by assuming partial equilibrium between atomic
and molecular oxygen only

6B-20

-------
N0X REDUCTION IN FUEL-RICH NATURAL GAS AND METHANOL TURBULENT DIFFUSION FLAMES

M.A. Toqan, J.D. Teare, and J.M. Beer

THE ENERGY LABORATORY
AND

DEPARTMENT OF CHEMICAL ENGINEERING
MASSACHUSETTS INSTITUTE OF TECHNOLOGY

A.J. Weir, Jr. and L.J. Radak
SOUTHERN CALIFORNIA EDISON COMPANY

ABSTRACT

The spatial distribution of NO concentration was determined in 1.0 MW.. input fuel -
rich methane/air and methanol/air f1ames in which the combustion air was doped with
nitric oxide; the NO concentration in the combustion air was varied between 0.0 to
3200 ppm.

The results show that when the input NO concentration in the combustion air was
1600 ppm the respective NO reduction efficiencies measured in the fuel-rich [4 =
1.3) natural gas and methanol flames were 93% and 76%. The reduction efficiency
was found to increase with increasing inlet NO concentration,* for an inlet NO
concentration of 3200 ppm the reduction efficiency in the natural gas flame was
96%.

In a theoretical study, a Sandia chemical kinetic code with kinetic parameters
developed by J.M. Levy and B.R. Taylor was used to predict the conversion of
nitrogeneous compounds to N2 in premixed, fuel-rich, plug flow systems as a
function of process variables such as initial NO concentration, gas temperature,
fuel equivalence ratio, residence time and fuel type. Results of the theoretical
computations show that the N0X reduction process can be considered to consist of
two major parts. In the first of these CH-j radicals react with NO to form HCN and
subsequently NHj species, and in the second part hydrogen abstraction of ammonia
species by hydrogen results in the formation of molecular nitrogen, N£. The
ultimate reduction level of N0X is therefore dependent on both the CH, radical and
H atom concentrations. At temperatures > 1700 K the calculated concentrations of
these radicals in the CH4 flame are higher than those predicted for the CH3OH
flame.

The experimental and theoretical results obtained provide information on the effect
of fuel type (CH4 vs CH3OH) on N0X reduction with relevance to the "reburn" method,
i.e., reduction of NO emission from the combustion of a primary fuel by the
injection into its burned gas a hydrocarbon fuel.

6B-21

-------
INTRODUCTION

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 N0X emission control.
The degree of conversion of fuel-bound nitrogen (FBN) to molecular nitrogen is
determined by the thermodynamics of the system (temperature, pressure and chemical
composition of the gas mixture), and the rates of the reactions of nitrogen
compounds in the fuel-rich flame zone (1,2,3,4,5). In practical combustion systems
the reactions of FBN and NO may be kinetically constrained and provision of
sufficient residence time in the fuel-rich zone is therefore essential for high
conversion to N2 (6,7,8). The nitrogenous compounds in the fuel-rich flame zones
may have different origins: they may be pyrolysis products of nitrogen-bearing
fuels, products of the fixation of atmospheric nitrogen by hydrocarbon fragments
early in the flame (9), or the result of reactions due to the secondary injection
of hydrocarbons into the NO-containing fuel-lean burned gas. The various pathways
for formation and destruction of NO are shown schematically in Figure 1. In the
combustion air staging technique, conversion of FBN to N2 rather than to NO is
mainly due to fuel-rich conditions prevailing in the primary flame zone near the
burner. In fuel staging, popularly termed "NO-reburn" by Vlendt et al., (10) and
Myerson(ll), hydrocarbon fragments such as CHg and CH react with NO to produce HCN
and NH-j compounds in a secondary fuel-rich flame zone, with the latter species
subsequently being reduced to N?. A detailed chemical kinetic model which
describes this route of NO destruction was extensively tested and validated by
Taylor (5) in a flat flame study.

Many experimental and theoretical investigations have been carried out to evaluate
the effects of fuel type, equivalence ratio, gas temperature of the reburn stage,
and the gas mixture residence time within that zone, on NOx reburning efficiency.
In one series of bench and pilot scale experiments, for example, Chen et al. (12)
concluded that NOx reduction efficiency was best for high N0X concentration at the
end of the primary zone. Residence times of at least 400 ms in the reburn zone
were required to favor molecular nitrogen formation. In another study, Lanier and
Mulholland (14) investigated the reduction of NOx by natural gas reburning in a
two-dimensional pilot-scale combustion system. An overall reaction order of 1.5
with respect to NO was observed. When the NO concentration in the primary lean
stage was low, the "reburn" efficiency was reduced due to the formation of "prompt"
NO in the fuel-rich region from the reactions of N2 with hydrocarbon radicals (as
shown in Figure 1). Other researchers (15) also found extensive reductions of
nitric oxide in laboratory-scale reactors when combustion products from a fuel-lean
primary stage were passed through a fuel-rich secondary stage.

6B-22

-------
Among the fuels used in "reburn" experiments, natural gas and propane have been
found to be better than carbon monoxide, hydrogen and coal. In a demonstration
test on a 50 MW utility boiler, Heir et al.(13) found that N0X emission levels were
reduced by more than 50% when methanol was used instead of natural gas as a reburn
fuel. These particular results were not readily explainable in terms that could be
generalized to other configurations, and thus they provided a motivation for
further study.

Motivation of present work

In the present work the primary research objective was to determine the critical
parameters which control N0X reduction efficiencies in turbulent fuel-rich
diffusion flames. This was attempted by means of an experimental investigation of
N0X reduction in the flame tunnel of the pilot-scale MIT Combustion Research
Facility. The fuel-rich flames were doped with varying amounts of NO which was
added to the burner air flow. Because of the utility-boiler results quoted above,
natural gas and methanol were chosen as experimental fuels.

In order to further elucidate the factors which control the relative effectiveness
of these two fuels in reducing NO to Nj, a parallel theoretical effort was mounted.
This effort was confined to plug flow geometries with instantaneous mixing of
reactants, but significant insight was obtained from parametric numerical
calculations carried out for a number of cases representative of staged combustor
conditions.

Direct numerical comparisons of the relative effectiveness of these two fuels are
straightforward, since temperature can be chosen to be the same in the cases being
compared. Similar comparisons based on experimental data must be viewed with more
caution, since several factors which influence the thermal balance within the flow
field can introduce some bias into the results. For instance, the lower calorific
value of methanol relative to methane implies that greater air preheat would be
necessary with the CH3OH fuel to yield equivalent temperature contours in the
flame; a compensative factor is the increased luminosity of the CH4 flame relative
to the blue flame observed with CH3OH - this induces greater cooling by radiative
heat transfer from within the CH4 flame.

The authors are aware that the doping of the burner air by NO creates conditions in
the fuel-rich hydrocarbon flame which are not strictly identical to those in NO
reburning. In the latter case the hydrocarbon fuel burns in combustion products
which contain CO2 and H2O in addition to NO and O2. There is a larger mass flow of
gas with depleted 0g concentration even though the fuel equivalence ratio is the
same as in the fuel-rich hydrocarbon-air flame. It is considered, however, that

6B-23

-------
the effects of these differences upon nitrogen compound interconversion are
calculable and that the results of the present study can be applied to NO-reburn
systems.

EXPERIMENTAL

The experiments were carried out using 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 scale combustion
tunnel equipped with a single, variable-swirl control burner of up to 3 MW^ multi-
fuel firing capacity, and designed to facilitate detailed investigation of
industrial type turbulent diffusion flames. In the experiments, a fraction of the
combustion air was introduced through the burner while the rest was injected
through eight injectors located 2.44 m downstream of the burner. Figure 2 shows
the configuration used in this study.

Experimental Fuels - Operating Conditions

The chemical composition of the natural gas fuel was: CH4 - 97 mole %; C02 - 0.5%;
C2H5 - 1.6%; C2H4 - 0.0%; C3H8 - 0.3%; N2 ~ 0.3%; and butane and pentane - 0.3%.
The methanol used was commercial grade, with impurity levels less than 0.5%.

In both flames the thermal input was 1 HW, with fuel flow rates established at 66.8
kg/hr of CH4, and 163.6 kg/hr of CH3OH. The burner fuel equivalence ratio and the
overall fuel equivalence ratio were maintained at - 1.3 and - 0.87 respectively;
the axial air velocity at burner exit was U0 - 19 m/s for natural gas and 14 m/s
for "methanol; the swirl number, S, was 0.75, and the combustion air temperature,
Ta, was.473 K. Flowfield similarity between the two fuel flames was maintained as
closely as possible by choice of fuel nozzle type, using a 70° (full angle) hollow-
cone multi-jet gas injector for the CH4 flame, and a 70° pressure jet hollow-cone
single-hole injector for the CH3OH. The combustion air in the experiments with
both fuel types was doped with NO ranging in concentration up to 3200 ppm.

Measurements

In-flame measurements of the spatial distributions of time average values of gas
velocity and gas temperature were made using water-cooled impact probes and suction
pyrometers respectively. To measure the light hydrocarbons concentrations, gas
samples were taken from different locations in the flame by means of water-cooled
sampling probes and then analyzed by gas chromatography for stable species
concentrations including H2 and to C4 hydrocarbons. N0X, C02, CO and 02
concentrations in the flame were measured by passing the flame gases through
standard chemiluminescent, IR and paramagnetic analyzers.

6B-24

-------
Experimental Results

The experimental flames were produced by a combination of fuel injection, in the
form of a 70s angle hollow-cone spray (methanol) or multi-jet injection of natural
gas with the jets subtending the same 70* angle, and a weakly swirling annular air
flow surrounding the fuel injectors. The fuel-rich stage created by this
configuration was followed by a lean stage in which secondary combustion air was
introduced to oxidize the incomplete combustion products such as soot, carbon
monoxide and hydrogen (see Figure 2).

To study the effect of inlet N0X concentration level upon the efficiency of its
reduction in the flame the incoming combustion air was doped with various levels of
nitric oxide. In the course of the experimental study radial traverses of gas
velocities, gas temperatures and chemical species concentrations, including those
of major stable species such as NOx, O2, CO2, CO, C1-C4 hydrocarbons and H2, were
measured at several axial stations in both the methanol and natural gas flames.
The main input variables of the experimental study were:

(1) fuel type - CH4 or CH3OH, and

(2) N0X concentration in the combustion air

Contours of N0X concentrations without and with doping of the combustion air (1600
ppm NO) and isotherms for the fuel-rich (0 = 1.3) natural gas and methanol flames
are shown in Figure 3. The results obtained show that destruction of NO in the
natural gas flame is more effective than in the methanol flame. In spite of the
attempts to maintain flow field similarity for methanol and natural gas flames,
there are significant differences observable. The temperature contours are
reasonably similar, but highest N0X concentrations are found close to the flame
axis in the methanol cases, and in an annular off-axis zone for natural gas flames.

In the case when the combustion air was not doped with nitric oxide, it is
noteworthy that the minimum N0X concentration measured at the end of the rich stage
was lower for the natural gas (- 25 ppm) than for the methanol flame (- 35 ppm).

The variations of CH4, and H2 concentrations along the natural gas and methanol
flames are shown in Fig. 4. In the methanol flame, the CH3OH rapidly decomposes
primarily to CO and H2; thus it is significant that CH4 concentration in the
natural gas flame remains substantially higher along most of the flame, by
approximately a factor of 10 on the axis, and by more than two orders of magnitude
at some off-axis locations. Higher CH4 concentrations entail also higher CH,
radical concentrations, and these latter are key to the initiation of the N0X
reduction process.

6B-25

-------
In further experiments the NO concentration in the combustion air was varied to
determine the effect of the inlet NO concentration upon the NO destruction
efficiency in the fuel-rich flame, with results presented in Fig. 5. Figure 5(A)
shows that for an NO concentration level of - 3200 ppm in the combustion air a 96%
reduction efficiency was attainable in the fuel-rich natural gas flame. As the
inlet nitric oxide concentration was reduced the NO reduction efficiency fell; it
was 85% for an initial NO concentration of 800 ppm. The measured N0X
concentrations at the exit of the fuel-rich stage are shown as the solid squares in
Fig. 5(B); if the secondary air could be admitted without further production of NOx
the dilution would yield lean stage exit concentrations shown as open circles.
However, some NO was formed as the secondary air was injected (about 20 ppm), as
indicated by the solid circles. This additional NO measured at the exit of the
combustor can be attributed to the oxidation of any unconverted NHj and HCN leaving
the rich stage, and/or to thermal NO formation.

THEORETICAL

NO in fuel-rich flames can be destroyed through two main pathways, either by
reacting with ammonia (NH-j,N) species to form molecular nitrogen, or by reacting
with hydrocarbon radicals such as CH and CHg to produce hydrogen cyanide which is
further converted to NH-j. The ammonia species thus formed can subsequently reduce
NO to N2-

In order to obtain information on the effect of process variables such as
temperature, fuel equivalence ratio and residence time upon the rate of NO
reduction, parametric numerical calculations were made assuming plug flow, with
instantaneous admixing of fuel into the NO-bearing burned gas.

NOv Numerical Analysis

A chemical kinetic code developed at Sandia National Laboratory (17) coupled with a;
differential equation solver developed at Lawrence Livermore Laboratory (18) was
used. The code solves the initial value problem for stiff or non-stiff systems of
first order differential equations. The equations considered in this study were
those for mass conservation of the species. By feeding a temperature profile, a
set of species initial mole fractions and a reaction set developed by Taylor(5)
(containing 183 reactions including interactions between nitrogenous and
hydrocarbon species) into the program, the species mass conservation equations were
solved for the different species considered in this study.

Results of Computations

A practical fuel-staged combustion system can comprise several zones, the first of

6B-26

-------
which would typically have fuel-lean stoichiometry. In the so-called "reburn"
stage, secondary fuel is injected into the NO-bearing burned gas to change the
stoichiometry to fuel-rich. Further downstream in a third stage combustion air is
introduced to burn the residual fuel completely; the fuel equivalence ratio is
changed to its lean exit value (# < 1.0).

The present calculations deal with the idealized second-stage of the above process,
in which the secondary fuel (CH4 or CH3OH) is assumed to be instantaneously mixed
with the first-stage combustion products. The latter correspond to the output of a
primary fuel-lean flame burning CH4 with ~ 0.90. The initial gas composition is
thus intrinsically fixed, except that both temperature and NO concentration are
treated as input parameters for the second-stage calculations, and are varied over
wide ranges. Other parameters to be varied in this "reburn" stage are fuel type
and fuel equivalence ratio; for each set of parametric choices the subsequent
changes in the chemical composition of the reacting mixture are evaluated
numerically under conditions of constant temperature.

Comparisons between the CH4 and CH3OH cases can be made for any arbitrarily-chosen
residence time. In general, the cases using CH4 as the "reburn" (secondary) fuel
approached steady state with a (calculated) residence time on the order of 400-600
ms» but longer times were required to achieve high NQX reduction levels in the
CH3OH cases. All comparisons discussed below use numerical data obtained for a
one-second residence time.

Effect of Temperature

Concentrations of selected species at the outlet of the fuel-rich stage (<£ = 1.3)
are plotted in Figure 6 for both natural gas and methanol as the "reburn" fuel.
The species chosen are NO, NH3 and HCN, but total fixed nitrogen (TFN) is also
plotted, this being the sum of all nitrogenous compounds exclusive of Ng. TFN is
important because any species such as NHi or HCN which are present at exit from the
"reburn" stage are likely to be oxidized (thus yielding additional HO) during
passage through the third stage.

The left-hand side of Figure 6 shows the above species as a function of temperature
for CH4 flames, with three levels of initial NO concentration (800, 100 and 40 ppm)
displayed parametrically. The right-hand side of Figure 6 shows similar plots for
CH3OH as the "reburn" fuel, with a single (800 ppm) level of Initial NO
concentration. For a high initial NO concentration (-800 ppm), and with the
reactor temperature maintained at - 1900 K, a TFN conversion to N2 of 97% is
obtained for the CH4 flame (See Figure 7). Both below and above this optimum
temperature of 1900 K the TFN yield (and hence the subsequent NOx emission)

6B-27

-------
increases; at 1400 K the estimated TFN conversion efficiency drops to 8% while
above 1900 K the thermal formation of N0X begins to play a significant role in
increasing the total fixed nitrogen leaving the fuel-rich stage in the methane
flame.

In the methanol flame the NQX conversion to N? increases continuously as the
temperature rises to 2000 K. However, the conversion efficiency reaches only 88%
at that temperature.

In Figure 8 the peak values of the sum of CH, CH2. CH3 species concentrations in
the plug flow reactor, together with the concentration of H atoms at residence
times corresponding to the positions of maximum NH3 concentration, are plotted
versus reactor temperature. The differences in NO-to-Nj conversion between the CH4
and CH3OH flames are due to the significantly higher CH, radical concentrations in
the CH4 flame; these cause an effective HO destruction even at relatively low
temperatures (1400 K). It can be concluded from Figure 8 that since the sum of CH-j
species concentration is high, the rate-limiting step in the N0-to-N2 conversion in
a CH4 flame is the reduction of NH-j to Ng by hydrogen abstraction. In the CH3OH
flame, on the other hand, the lower concentrations of CH-j provide relatively slow
rates of conversion of NO to NH-j even at high temperatures {- 2000 K); however,
increase in T is accompanied by increased concentrations of H and OH, and these
radicals cause depletion of NH-j species on an increasingly faster time scale,
thereby providing the chemical driving force for effective destruction of NO.

Effect of Fuel Equivalence Ratio

Solutions of the model equations for N0X, NH3, HCN and TFN concentrations
calculated at 1900 K as a function of fuel equivalence ratio are shown in Fig. 9.
The two pathways involving CH-j and NH-j species respectively can be seen to affect
the course of TFN reduction. The N0-to-N2 conversion is maximized when the exit
value of TFN reaches a minimum; for the: CH4 flame it is seen that with increasing
values of 4 the TFN concentration drops sharply to reach a minimum at 4 - 1.4
while a similar minimum value occurs at a more fuel-rich stoichiometry (4 ~ 1.6) in
the CH3OH flame. At a low value of 4 (- 1.2) the sum of CH, species concentrations
is around 500 ppm in the CH4 flame and it is less than 1 ppm in the CH3OH flame.
At high values of 4 (~ 2.0) the CH4 flame still has higher concentrations of CH-j
species, but the presence of high concentration levels of H and OH radicals in the
methanol flame again enables effective NO-N2 conversion, as in the T = 2000 K case
discussed above. It is noteworthy that at the optimum fuel equivalence ratios the
times required for achieving the plotted reduction levels of N0X are approximately
550 ms for both fuels.

6B-28

-------
CONCLUSIONS

• Experiments with NO-doped fuel-rich = 1.3) natural gas and
methanol flames showed high levels of NO reduction in both flame
types; for the initial NO concentration of ~ 1600 ppm, reduction
efficiencies were 93% and 76% in the natural gas and methanol
flames, respectively.

• Results of computations of the progress of NO conversion in fuel -
rich, premixed, plug flow systems show effects of process variables
such as "reburn" fuel type (CH4 vs CH3OH), fuel equivalence ratio,
temperature and residence time upon NO and TFN conversion to Ng.
The computations illustrate the need for high CHj concentrations in
the initial stages of the conversion to form HCN and amine species,
followed by high H and OH radicals concentrations to complete the
conversion to N2 by hydrogen abstraction reactions of amines.

• The higher CHj and Hg concentrations calculated and measured along
the CH4 flame are congruent with the higher NO reduction
efficiencies with this fuel type.

• Calculations indicate that as the fuel equivalence ratio is
increased beyond 4, - 1.6 (under conditions of constant flame
temperatures) the relative advantage of the CH4 flame diminishes,
with NO conversion efficiencies being similar to those obtained in
CH3OH flames.

ACKNOWLEDGEMENTS

Financial support from the Southern California Edison Company and the National

Institute of Environmental Health Sciences is gratefully acknowledged. Significant

contributions to the experimental research were made by students Tina Bahadori and

Tuomas Paloposki and by the CRF technical staff, Don Bash, William Mason and Mark

James. Special thanks to ~ Bonnie Caputo for preparing the manuscript.

REFERENCES

(1) Haynes, B.S., The Formation and Behavior of Nitrogen Species in Fuel-rich
Hydrocarbon Flames", Ph.D. Thesis, School of Chemical Engineering, University
of New South Wales, Australia, December (1975).

(2) Bowman, C.T., Progress in Energy and Combustion Science 1975. Vol. 1 pp. 33-
45. Pergamon Press. Oxford, New York,

(3) 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).

(4) Levy, J.M. Taylor, B.R. Longwell, J.P. and Sarofim, A.F., "CI and C2 Chemistry
in Rich Mixture, Ethylene-Air Flames", The Nineteenth Symposium on Combustion.
The Combustion Institute, Pittsburgh, PA p. 167, 1982.

(5) Taylor, B.R., "Reactions of Nitrogen Species In Fuel-Rich Flames", Sc.D.,
Department of Chemical Engineering, M.I.T., 1984.

6B-29

-------
(6) Farmayan, W.F., M.Sc. Thesis, "The Control of Nitrogen Oxides Emission by
Staged Combustion," Department of Chemical Engineering, Massachusetts
Institute of Technology, Cambridge, HA, April 1980.

(7) Beer, J.M., Jacques, M.T., Farmayan, W.F., and Taylor, B.R., Eighteenth
Symposium (International) on Combustion pp. 101-110. The Combustion
Institute, Pittsburgh, 1980.

(8) Be£r, J.M., M.T. Jacques, Farmayan, W.F., Gupta, A.K., Hanson, S., Rovesti,
W.C., "Reduction of N0X and Solid Emissions by Staged Combustion of Coal
Liquid Fuels." Nineteenth Symposium on Coal Liquid, Haifi, Israel, Aug. 1983.

(9) Fenimore, C.P., Thirteenth Symposium (International) on Combustion. The
Combustion Institute, 1971, p. 373.

(10) Wendt, J.O.L., Sternling C.V. and Matovich, M.A., Fourteenth Symposium
(Internat) on Combustion. Combustion Institute, Pittsburgh, 1971, pp 897-904.

(11) Myerson, A.L.: Fifteenth Symposium (International) on Combustion, p. 1805, The
Combustion Institute, 1975.

(12) Chen, S.L., McCarthy, J.M., Clark, W.D., Heap, M.P., Seeker, W.R., and
Pershing, D.W., Bench and Pilot Scale Process Evaluation of Reburning for In-
Furnace N0X Reduction, Twenty First Symposium (International) on Combustion,
The Combustion Institute, Pittsburgh, 1986.

(13) Weir, A., Jr., L.J. Radak, E.A. Danko R.A. Lewis, and H.W. Buchanan, "Methanol
Dual-Fuel Combustion", EPA and EPRI Cosponsored 1987 Joint Symposium on
Stationary Combustion N0X Control, March 23-26, 1987.

(14) Lanier, W.S., and Mulholland, J.A., Application of Reburning for NQX Control
to a Fine tube Package Boiler, Journal of Engineering for Gas Turbines and
Power, Vol. 107, 1985.

(15) Clark, W.D., Chen, S.L., Greene, S.B., Seeker, W.R., and Heap, M.P.,
"Reburning with Pulverized Coal for N0X Control", Presented at Western States
Section of The Combustion Institute Spring Meeting, University of Utah, April
5-6, 1982.

(16) Beer, J.M., W.F. Farmayan, J.D. Teare, M. Toqan, "Laboratory-Scale Study of
the Combustion of Coal-derived Liquid Fuels", EPRI Final Report, May 1985.

(17) Kee, R.J., Miller, J.A. and Jefferson, T.H., "CHEMKIN: A General-Purpose
Problem-Independent, Transportable, Fortran Chemical Kinetics Code Package",
Sandia Laboratories, New Mexico, Doc. No. SAND 80-8003, 1980.

(18) Hindmarsh, A.C., "ODEPACK, A Systematized Collection of ODE Solvers",
Numerical Methods for Scientific Computation, August (1982).

(19) Lanier, W.S., Mulholland, J.A., and Beard, J.T., Twenty First Symposium
(International) on Combustion, The Combustion Institute, Pittsburgh, 1986.

(20) Miyamae, S., et al., "Evaluation of In-Furnace NQX Reduction." Proceedings of
the 1985 Joint Symposium on Stationary Combustion NQV Control. May 7, 1985,
No. 4-4.

6B-30

-------
FIXATION OF ATMOSPHERIC CONVERSION

NITROGEN OFFUEL-N

NO RE8URN

AIR
NZ

HETEROCYCLIC
NITROGEN COMPOUNDS

FIXATION OF N2
BY HYDROCARBON
FRAGMENTS

ZELDGVICH
MECHANISM

CYANOGENS
(HCN, CN)

OXYCANOGENS

(OCN.HNCO)

AMMONIA SPECIES
(NH3, NH2, NH, N)

HYDROCARBON
FRAGMENTS

CH, CHZ

¦REDUCING ATMOSPHERE
OXIDIZING ATMOSPHERE

Figure 1. Formation arid Reduction of Nitrogen Oxides in Combustion; Mechanistic
Pathways

A - Ouar 1 and

Spray Nozzle
B - observation

W indow
C - Probe Port

Burner Experimental Cylindrical Cold-Wall Exhaust

Chamber Afterburner Chamber Section

Figure 2, Furnace Assembly and Air Staging System.

6B-31

-------
NATURAL CAS FUME

METHANOL FLAHE

sg. ami

6**X iKlj - UOO pp.)

l|fe^x0\ 22?

[KO] j - ISCC 7PB)

¦ISSO-o,

^ > ? ^5 il£
-2409/I2B8'

0J» SJW

JJW too

AXIAL DISTANCE FROM BURNER - (a)

Figure 3. Contours of N0X concentrations without and with doping of the combustion
air {1600 ppm NO) and isotherms for the fuel-rich ( ¦ 1.3) natural gas
and methanol flames.

6B-32

-------
0.150

e 0.100

o
a

0.050

0.000 —'—1

• •

i i f* r r i i

0 Natural Flame
¦ Methanol Flaae

ml I II I I,,, I I I I

0,00

0.50

1,00

1.50

2.00

55
O

10°

10"1

10"

§ 10~8

r—1—1—

¦¦¦ r

—r

1 1 1

—1—|—

1 r"

1—-r

I—'—

"i "» 'j

;

tural

Flame

Methanol

Flams 1

: 0

•

: 0

' O

r D

0 ;

1 .1;

. '

' >

0.00 0.50 1.00 1.50 2.00

AXIAL DISTANCE FROM FUEL NOZZLE (m)

Figure 4. H2 arid CH4 concentrations along the flame axis in fuel-rich (# ¦ 1.3)
natural gas and methanol flames.

6B-33

-------
1.00

>*
u
2
u

o
u

tl*

p
u
a
a
w

K
X

0.90

0.00

0 000 1800 2400 3300 4000

NOx CONCENTRATION AT INLET OF RICH STAGE (PPM*

120.0

£ 100.0

o
p

f-*

2
U

O
U

80.0

eo.o

40.0

200

0:0

1 r

/ ,¦
/ /

¦. MEASURED AT EXIT OF FUEL RICH STAGE
• MEASURED AT EXIT OF LEAN STAGE
° CALCULATED AT EXIT OF LEAN STAGE

(DILUTION EFFECTS ONLY)

0 000 1000 2400 3200 4000

NOx CONCENTRATION AT INLET OF RICH STAGE (ppm)

Figure 5. (a) Fractional reduction of NQX in fuel-rich natural gas flame as a
function of the initial NO concentration (introduced with
combustion air)

(B) N0X concentration measured at the end of the natural gas rich stage
and at the exit of the combustion tunnel as a function of N0X
concentration entering the fuel-rich stage

6B-34

-------
METHANE FLAMES

TEMPERATURE (K)

Figure 6. Effect of temperature on
fl ames.

METHANOL FLAMES

itr

10"

10'

« - 1.3
NOj,

800 ppm

JO"
10°

1400

1300

1800

2000

10"*

10'
10"
10"'
10"1
10*1
10"*
10'

10*

10'

1400

1800

2000

1400

isoo

1800

2000

ia°

1400

1600 IBOO 20DQ

TEMPERATURE (Kj

conversion in fuel-rich CH4 and CH3OH

-------
TEMPERATURE (K)

Figure 7. Predicted N0X reduction efficiency level in both CH4 and CH3OH flames as
a function of reactor temperature.

6B-36

-------
"Z.

o
p

o
a

t-j

io_

10"

lO"4

10"

to*

10"

a 10-1

10"

o HETIIANE FLAMES

g METHANOL FLAMES
_t i I

1400

1000

^ HETtlAWE FLAMES
q METHANOL FLAMES

1400

1600 1000

TEMPERATURE (K)

1000 2000

2000

Figure 8. Predictions of peak concentrations of s CH^ in the plug flow reactor
concentration of H atom at a residence time of peak NH3 concentration in
the reactor.

6B-37

-------
Q.
fX

p—«

o
2

103

10z

IQ1

10"

io3

1.0

Cx.

10"

10'

10°

<> METHANE FLAMES
METHANOL FLAMES

1.0

METHANE FLAMES

Q METHANOL FUVMES _

O METHANE FLAMES
Q METHANOL FLAMES

1.5

2.0

2,5

l ' 1 1 1 1 1 1 1 1

r 0

METHANE FLAMES ,

METHANOL FLAMES :

• • • • 1 ¦ . • ¦ !

1.5 2.0

FUEL EQUIVALENCE RATIO

2.5

Figure 9, Effect of fuel equivalence ratio upon N0X conversion in fuel-rich CH4
methanol flantes.

6B-38

-------
REDUCTION OF FUEL-NO
BY INCREASED OPERATING PRESSURE
IN A LABORATORY-SCALE COAL GASIFIER

K. M. Nichols
The Institute of Paper Chemistry
Chemical Sciences Division
P.O. Box 1039
Appleton, WI 54912

P. O. Hedman
Brigham Young University
Department of Chemical Engineering
350 CB
Provo, UT 84602

A. U. Blackham
Brigham Young University
Department of Chemistry
Eyring Science Center
Provo, UT 84602

ABSTRACT

Measurements of NO during laboratory-scale gasification of a Utah bituminous
coal verified that small increases in pressure (from 1 to 2 atm) at constant
residence time resulted in dramatic decreases in effluent NO levels. Tests were
conducted at 3 target levels of pressure (1, 2, and 4 atm) and 2 target levels of
residence time (450 and 900 ms). Oxygen-to-coal ratio for all tests was 0.90 (SR =
0.45). The dominant factor in causing lower effluent NO levels was the increased
kinetic rate of NO decay. Increased residence time in the fuel-rich gasifier
contributed to lower effluent NO levels, but was of minor importance when
compared to the effect of pressure on the decay rate. Concentrations of N2
appeared to be slightly increased and concentrations of TFN decreased as pressure
was increased. Neither TFN or N2 concentrations were affected by increasing
residence time. For all tests, nitrogen conversion exceeded carbon conversion by
about 10%. Neither nitrogen conversion nor carbon conversion was found to
increase with increasing pressure. Both increased slightly (4-5%) with increasing
residence time, evidence that most of the coal nitrogen and carbon was released
during devolatilization.

6B-39

-------
INTRODUCTION

It has been shown that increased pressure yields lower NO concentrations
during combustion. Nichols, et al. [1,2] measured NO levels during gasification of
Utah bituminous coal. At 1 atm operating pressure, NO levels were 30 to 300 ppm,
depending on oxygen-to-coal ratio, while NO levels at 5 and 10 atm were near zero
(less than 10 ppm) regardless of oxygen-to-coal ratio. Heberling and Boyd [3]
observed that in overventilated-Iaminar-diffusion-flames (H2 and CH2 fuel doped
with CH3NH2) the effluent NO levels decreased sharply as pressure was increased.
During fluidized bed combustion of a Pittsburgh bituminous coal, Vogel, et al. [4]
observed a dramatic decrease in NO concentration (from 1800 to 200 ppm) as
combustor pressure was increased (from 1 to 8 atm).

In these studies the observed decreases in NO concentrations were
explained by one or more of the following: 1) increased pressure causing a
decrease in the level of NO produced in or near the flame front, 2) increased
kinetic rate of NO decay due to' increased pressure, and 3) increased gas residence
time due to increased pressure. Nichols, et al. [1] compared the spatially resolved
NO measurements from 5 atm gasification of Utah bituminous coal to the
spatially resolved NO measurements from 1 atm gasification of Utah bituminous
coal. This comparison was particularly useful for showing that increased gasifier
pressure did not result in a decrease in the peak level of NO produced in or near
the flame. The observations that effluent NO decreased with increasing gasifier
pressure must therefore be explained by 1) the increased rate of NO decay at higher
pressure, or 2) the increased residence time at higher pressure, or a combination of
these two.

Both of these factors are expected to play an important role, based on the
commonly accepted theory of nitrogen chemistry in flames. The overall NO
behavior in combustion processes is usually modeled by an overall mechanism for
nitrogen speciation similar to that shown here [5-9],

+Ox

> NO

fuel-N —> HCN,CN —> NHj (1)

> N2

+NO

6B-40

-------
Initially, oxygen is readily available and the NO formation step (NHj + Ox —> NO)
dominates over the NO decaying step (NHj + NO —> N2) yielding a net NO
production rate which is rapid and positive. However, the rising NO
concentration tends to increase the importance of the NO decay step, while
simultaneously the decreasing oxygen concentration tends to decrease the
importance of the NO formation step. At some point, the rates of the two steps
become equivalent and a peak level of NO is reached. After this, the decay
reaction of NO dominates over the formation reaction, yielding a negative net NO
production rate (i.e., an NO decay region is reached).

The dominant pathway for the NO decay is generally thought to be of the'

form

-d[NO]/dt = k[NO][NHi] (2)

though the nature of the intermediate, NHj, is the subject of various opinions.
Dividing both sides of Equation 2 by [NO], and writing [NHj] as the mole fraction
of NHj times the total concentration (using the ideal gas law) gives

, -d(ln[NO])/dt = kYNH.P/(RT) (3)

Equation 3 suggests that for negligible changes in k, YNHj, and T, the decay rate of

the logarithm of NO concentration should increase approximately in proportion
to increasing pressure. Thus, small increases in pressure should cause large
increases in the rate of NO decay.

Increased residence time should also decrease effluent NO levels at a given
decay rate, by providing a greater length of time for the hot gases to remain in the
region of net NO decay. Experimental data are lacking to show which of these two
effects is more pronounced. Variations in these two variables (pressure and
residence time) were confounded in the previous experiments. That is, pressure
was not altered independently of residence time.

6B-41

-------
OBJECTIVE

The objective of this study was to further quantify the relative importance
between the effect of residence time on NO levels, and the effect of pressure on NO
levels. To5 accomplish this, a set of experiments were designed and executed with
the BYU gasifier using Utah bituminous coal. The goal of these experiments was
to obtain NO measurements under conditions where 1) pressure was varied while
residence time was maintained constant, and 2) residence time was varied while
pressure was maintained constant.

EXPERIMENTAL
Gasifier

The experiments were conducted in the BYU laboratory-scale entrained-coai
gasifier. The gasifier (see Figure 1) has been described in detail previously [10-12],
It was axisymmetric, fired downward and consisted of six flanged-sections of 12 in.
(30.5 cm) nominal diameter steel pipe, 30.5 cm each in length, and one flanged
section 15.2 cm in length. Each of the six, full length sections was equipped with
an access port which could be used either for insertion of a sample collection probe
or for attachment of a quartz sight-window. Castable ceramic (5.1 cm thickness)
lined the interior of each section, and the reactor top and bottom, and the access
ports. Inside reactor diameter and length were 20 cm and 189 cm respectively.

High purity oxygen (99.6 mol%, the balance being argon) was fed from a
bank of high pressure cylinders. The use of pure oxygen provided an
environment where the only source of nitrogen was the coal-nitrogen. The
pulverized coal was withdrawn from the pressurized feeder by an auger and
entrained in an argon stream which conveyed the coal into the injector. The coal
feed rate was accurately monitored by suspending the feed system on a balance
beam using a counter weight [13]. A load cell attached to the beam indicated the
weight (or weight change) in the hopper.

Sample Collection

A water quenched stainless steel probe (see Figure 2) was used to collect gas-
liquid-solid samples from the gasifier. It has been shown [14] by comparing the
results from sampling with quartz probes to the results from sampling with
stainless steel probes, that stainless steel does not significantly affect nitrogen
species sampled from combustion.

The sample separation system (see Figure 3) divided the sample for
chemical analyses. The quenched sample first passed through a gas-liquid

6B-42

-------
disengager. The gas was then passed through an ice bath and any further
condensate was removed by a second gas-liquid disengager. The remaining gas
was passed through a 2 mm filter which removed remaining particulate matter.
The liquid was collected, and also filtered with a 0.8 mm filter to remove
particulate matter.

Analytical Methods

The analytical procedures developed previously [15-17] were used to
determine nitrogen species concentrations (Ni, NO, NH3, and HCN) as well as
¦¦major gas species in the gas and liquid phases, and the elemental composition in
the solid phase. The liquid phase was analyzed for NH3 and HCN as NH3 and CN~
ion using specific gas and ion electrodes. The INb in the gas phase was determined
using a gas chromatograph equipped with a thermal conductivity detector while
the NO was determined using a chemiluminescence NO analyzer.

Elemental analyzers, which combust the solid sample were used to
determine the carbon, hydrogen, nitrogen, sulfur, and ash content of the raw coal
and char samples. The chemical analyses were performed as soon after each test as
possible to minimize any error caused by continued reaction or decay of species.
Steps were taken to chemically preserve each species.

Test Plan

To determine the relative importance of pressure and residence time on
effluent NO levels, a. test design was planned and performed which included 3
target levels of pressure (1, 2, and 4 atm) and 2 target levels of residence time (450
and 900 ms). The other parameters were maintained constant. The target oxygen-
to-coal mass ratio for all tests was 0.9 (equivalent to a stoichiometric ratio of near
0.4).

Concentrations of NO were measured at each of the combinations of
pressure and residence time. Replicate measurements were made to obtain a
measure of the reproducibility of the data. The residence time was controlled by
adjusting the combined oxygen-plus-coal feed rate. At a given pressure, for
instance, the residence time was doubled by halving the combined feed rate.

Test Coal

The coal for these gasification experiments was a high volatile Utah
bituminous coal obtained from the Valley Camp Mine. It was pulverized to a
mean particle size of near 50 |im. Size data, elemental composition and other
physical characteristics are shown in Table 1.

8B-43

-------
RESULTS AND DISCUSSION
NO Measurements

Figure 4 shows the effluent NO measurements from these tests. Because
the actual pressure and residence time values achieved varied somewhat from the
target levels, these NO measurements are plotted as a function of both variables;
residence time in Figure 4a, and pressure in Figure 4b. In Figure 4a, the actual
(measured) value for the pressure (in atm) is noted next to each data symbol. In
Figure 4b, the actual (estimated) value for the residence time (in ms) is noted next
to each symbol.

The residence time values were estimated from the following equation

t = [AP/(RG)](1 / a)ln[(ax + b)/b] (4)

where A was gasifier cross sectional area, P was gasifier pressure, R was the ideal
gas constant, G was the molar flow rate of gas, x was axial distance from injector
tip, and a and b were parameters in the temperature function T = ax + b. The
particular values used for a and b were -7,5 K/cm and 2800 K. The derivation of
Equation 4 is given by Nichols, et al. [1]. The assumption of a linear function for
temperature was evaluated by comparing the resulting residence time values to
the residence time values calculated by using more complex temperature
functions. Residence time values were not significantly different, so the more
simple linear temperature function was chosen. The assumption of constant gas
flow rate appears justified based on earlier observations that most of the coal mass
is released to the gas phase early in the flame, during devolatilization [18,19].

Levels of NO were dramatically influenced by pressure, but the influence of
residence time was secondary. Increasing pressure from 1 atm to approximately 2
atm caused effluent NO levels to decrease from an average of 230 to 12 ppm.
Increasing the pressure to 4 atm further reduced NO to about 2 ppm.

Two of the one atm data points near 400 ms in Figure 4a show a lower NO
concentration than the 1 atm data points at greater residence times. At first it
appears that NO concentration increased with increasing residence time, which
seems unlikely. This is, however, explained by the slightly greater than 1 atm
pressure (1.06 and 1.08 atm) corresponding to these data points. This strong
dependence of NO concentration on pressure is illustrated particularly well in
Figure 4b. Effluent NO concertration was extremely sensitive to pressure between
1 and 2 atm. This illustrates a potential NOx reduction method for any combined

6B-44

-------
cycle or staged combustion process, wherein the first fuel rich stage could be
slightly pressurized. It also illustrates a potential method for enhancing any NOx
reduction method which relies on the reaction of NO with another nitrogen
containing species to form N2,

A strong dependence of NO concentration on residence time was not seen
over the range of residence times of these experiments. The 1. atm NO
concentration values near the target residence time of 450 ms were 225 and 240
ppm (see Figure 4a), while the 1 atm NO values near 900 ms residence time were
210 ppm. This .slight decrease in NO concentration was probably due to the slight
increase in pressure, and not to the increased residence time. Even if the slight
decrease in NO concentration was attributed completely to increased residence
time, the effect of residence time would still be secondary or even negligible
compared to the effect of increased pressure on effluent NO concentration.
TPN and N? Measurements

The TFN (HCN + NH3 + NO) and N2 concentrations for these same tests are
reported in Figures 5 and 6 respectively. The concentrations of TFN appear to
have decreased with increasing pressure, while concentrations of N2 appear to
have increased. Residence time had no consistent effect on TFN or N2 levels.

Since the decay of NO is by reaction of NO with other nitrogen species, it is
expected that increasing pressure would have resulted in decreased TFN
concentrations and increased N2 concentrations. It is expected that if the
measurement errors were decreased, these trends would have been more clearly
evident.

Nitrogen Conversion

The measured nitrogen and carbon conversion values are shown in Figure
7. Nitrogen (or carbon) conversion is defined to be the percent of fuel nitrogen (or
carbon) which was converted from the coal to gas phase species. These values
were determined by measuring the nitrogen, carbon and ash in the raw coal, as
well as in the collected char samples, and using ash as an inert tracer. Azuhata, et
al. [201 used argon tracer gas, as well as ash measurements for determination of
carbon conversion during entrained flow gasification of Utah bituminous coal.
The carbon conversion values based on argon agreed well with those based on ash,
showing the validity of using ash as an inert tracer for determination of elemental
conversions during gasification.

For all tests, nitrogen conversion exceeded carbon conversion by 8-10%.
This is consistent with previous gasification measurements [1,16]. Malte and Rees

6B-45

-------
[21] reported the ratio of nitrogen conversion to carbon conversion to be between
1.3 arid 2.0 for coal gasification. Of course this ratio must approach unity as carbon
conversion approaches 100%.

An explanation for nitrogen conversion exceeding carbon conversion is the
behavior of coal nitrogen evolution with temperature. Carefully controlled coal
pyrolysis experiments [18,22] have shown that nitrogen evolution is more
sensitive to temperature than total mass evolution. For the gasifier of this study,
residence times varied from 400 to 1000 ms. Peak local temperatures in the gasifier
were probably near the maximum adiabatic flame temperature {at SR = 1.0), which
was calculated to be about 3200 K (at 1 atm). It appears that sufficiently high
temperatures existed for a sufficient length of time to cause the more temperature
sensitive nitrogen conversion to exceed the less temperature sensitive carbon
conversion.

Neither nitrogen conversion nor carbon conversion was found to depend
on pressure. Both increased only slightly with increasing residence time. This
provides evidence that most of the coal nitrogen and carbon are released during
devolatilization. If significant nitrogen were released during char oxidation, the
increased residence time would have had a greater impact on nitrogen conversion.
Temperature

The possibility that gasifier temperature was not constant among these tests
needs to be addressed. For adiabatic conditions, equilibruim calculations show that
flame temperatures are somewhat higher for higher values of pressure. For
pressures of 1 and 4 atm, adiabatic temperatures are 2560 K and 2600 K respectively
at the oxvgen-to-coal ratio of 0.90 (SR = 0.45) used for these gasifier tests, and
maximum adiabatic temperatures (at SR = 1.0) are 3070 K and 3250 K respectively.

The gasifier used for these experiments was, however, not adiabatic. The
heat loss for these tests was estimated to be approximately 20-25% of the heat
released from the coal. This was due to the relatively large ratio of surface area to
volume resulting from the relatively small scale of the equipment.

The dramatic decreases in NO concentrations occurred as pressure increased
from 1 to 2 atm. For this pressure increase, adiabatic calculations yielded an 11 K
temperature rise at SR = 0.45, and a 41 K temperature rise at SR = 1.0. These
temperature rises are relatively small even for adiabatic conditions, and would be
smaller yet for non-adiabatic conditions. The gasifier temperatures were more
likely controlled by factors other than pressure; factors such as oxygen-to-coal ratio
and heat loss. The observation in Figure 7 that carbon conversion was not affected

6B-46

-------
by pressure also suggests that temperature changes were negligible since carbon
conversion is known to be temperature dependent.

Temperature measurements were recorded from the gasifier exit, near the
gas sampling location, using a thermocouple. These uncorrected thermocouple
readings averaged approximately 1500 K, and showed no significant increase in
temperature with increasing pressure. These thermocouple measurements were,
however, not deemed very reliable due to the difficulty of determining the
convective and radiative effects on the thermocouple reading.

Major Gas Species

It is also important to consider what happened to the levels of the major gas
species (CO, H2, and CH4) as pressure and residence time were varied. This effect is
seen by calculating the energy content of the product gas, which results from the
heats of combustion of the primary combustible constituents, CO, H2, and CH4.
Figure 8 shows that the differences in energy content among the 3 pressure levels
were negligible. This suggests that increasing pressure from 1 to 4 atm had little
effect on the gasifier performance. At all 3 pressures, the energy content of the
product gas decreased approximately 5 kj per kg for each ms increase in residence
time.

CONCLUSIONS'

1) During gasification, small increases in pressure (i.e., from 1 atm to 2
atm operating pressure) resulted in dramatic decreases in effluent NO levels, at all
levels of residence time and coal feed rate tested.

2) The dominant factor in causing lower effluent NO levels with
increased pressure was the increased kinetic rate-of NO decay. Increased residence
time in the fuel rich gasifier contributed to lower effluent NO, but was of minor
importance when compared to the effect of pressure. The peak NO level reached
in or near the gasifier flame was not decreased by increasing pressure. On the
contrary, it was slightly increased, as shown by previous studies.

3) Concentrations of TFN (HCN + NH3 + NO) showed some decrease
with increasing pressure, while concentrations of N2 showed some increase with
increasing pressure.

6B-47

-------
4) In all cases, nitrogen conversion exceeded carbon conversion,
typically by about 10 percent,

5) Neither nitrogen conversion nor carbon conversion was found to
increase with increasing pressure. Both increased only slightly (4-5%) with
increasing residence time. This was evidence that the bulk of the coal nitrogen
and carbon was released during devolatilization.

6) Experimental temperature measurements did not show gasifier
temperature to vary with pressure or residence time. Equilibrium calculations
show that for a non-adiabatic reactor, increasing pressure from 1 to 2 atm will yield
insignificant temperature rises.

7) The energy content contained in the combustible species of the
product gas was not affected by pressure, but decreased with increasing residence
time, at about 5 kj/kg-ms.

ACKNOWLEDGEMENTS

This work was supported primarily by the U. S. Department of Energy,
Morgantown Energy Technology Center, Morgantown, West Virginia with Gary
Friggens and Leland Paulson as technical project officers. Financial Support of the
Research Division and the College of Engineering and Technology of Brigham
Young University is also acknowledged.

6B-48

-------
REFERENCES

1. Nichols, K. M., Hedman, P. O., and Smoot, L, D., "Release and Reactions of
Fuel-Nitrogen in a High-Pressure Entrained-Coal Gasifier," Fuel, Vol 66, (1987):
see also Nichols, K. M. 'The Effects of Elevated Pressure on the Formation of
Nitrogen and Sulfur Pollutants During Entrained Coal Gasification,' M. S. Thesis,
Department of Chemical Engineering, Brigham Young University, Provo, Utah
84602 (1985)

2. Nichols, K.M., 'Nitrogen Pollutant Formation in a High -Pressure
Entrained-Coal Gasifier,' Ph. D, Dissertation, Department of Chemical
Engineering, Brigham Young University, Provo, Utah 84602 (1987)

3. Heberling, Paul V., and Boyd, Michael G., 'The Effects of Pressure on NO
yield from Fuel-N in Laminar Diffusion Flames,' Comb, and Flame, Vol 41, p 331
(1981)

4. Vogel, G.J., Swift, W.M., Lenc, J.F., Cunningham, P.T., Wilson, W.I., Pank,
A.F., Teats, F.G-, and Jonke, A.A., 'Reduction of Atmospheric Pollution by the
Application of Fluidized-Bed Combustion and Regeneration of Sulfur Containing
Additives,' Report EPA-650/2-74-104 (1974)

5. Haynes, B.S., 'Reactions of Ammonia and Nitric Oxide in the Burnt Gases of
Fuel-Rich Hydrocarbon-Air Flames,' Comb, and Flame, Vol 28, p 81 (1977)

6. De Soete, G.G., 'Overall Reaction Rates of NO and N2 Formation from Fuel
Nitrogen,' Fifteenth Symposium (International) on Combustion, The Combustion
Institute, Pittsburgh, PA, p 1093 (1975)

7. Fenimore, CP., 'Studies of Fuel-Nitrogen Species in Rich Flame Gases/
Seventeenth Symposium (International) on Combustion, The Combustion
Institute, Pittsburgh, PA, p 661 (1979)

8. Miller, J.A., Branch, B.C., McLean, W.J., Chandler, D.W., Smooke, M.D., and
Kee, R.J., 'The Conversion of HCN to NO and NT2 in H2-O2-HCN-AR Flames at
Low Pressure,' Twentieth Symposium (International) on Combustion, The
Combustion Institute, Pittsburgh, PA, p 673 (1984)

9. Glass, J.W., and Wendt, J.O.L., 'Mechanisms Governing the Destruction of
Nitrogeneous Species During the Fuel Rich Combustion of Pulverized Coal;
Nineteenth Symposium (International) on Combustion, The Combustion
Institute, Pittsburgh, PA, p 1243 (1982)

6B-49

-------
10. Skinner, F.D., '1980 ASME Winter Annual Meeting,' New York, 1980; see
also Skinner, F.D., Ph.D. Dissertation, Department of Chemical Engineering,
Brigham Young University, Provo, Utah 84602 (1980)

11. Soelberg, N.R., Smoot, L.D., and Hedman, P.O., 'Entrained Flow Gasification
of Coal: 1. Evaluation of Mixing and Reaction Processes front Local
Measurements,' Fuel, vol 64, p 776 (1985): see also Soelberg, N.R., 'Local
Measurements in an Entrained Flow Gasifier,' M.S. Thesis, Department of
Chemical Engineering, Provo, Utah (1983)

12. Brown, B.B., Smoot, L.D., and Hedman, P.O., 'Effect of Coal Type on
Entrained Gasification,' Fuel, vol 65, p 673 (1986): see also Brown, B.B., 'Effect of
Coal Type on Entrained Gasification,' Ph.D. Dissertation, Department of Chemical
Engineering, Brigham Young University, Provo, Utah (985)

13. Cope, R.F., 'The Effects of Coal Rank in a Pressurized Entrained Coal
Gasifier,' M.S. Thesis, Department of Chemical Engineering, Brigham Young
University, Provo, Utah (1987)

14. Zabielski, M.F., Dodge, L.G., Colket, M.B., and Seery, D.J., 'The Optical and
Probe Measurement of NO: A Comparative Study,' Eighteenth Symposium
(International) on Combustion, The Combustion Institute, Pittsburgh, PA, p 1591
(1981)

15. Price T.D., Smoot, L.D., and Hedman, P.O., 'Measurement of Nitrogen and
Sulfur Pollutants in an Entrained-Coal Gasifier,' Irtd. Eng. Chem. Fundam., vol 22,
p 110 (1983): see also Price, T.D., 'Measurements of Pollutants in an Entrained-
Flow Coal Gasifier/ M.S. Thesis, Department of Chemical Engineering, Brigham
Young University, Provo, Utah (1981)

16. Highsmith, J.R., Soelberg, N.R., Hedman, P.O., Smoot, L.D., and Blackham,
A.U., 'Entrained Flow Gasification of Coal: 2. Fate of Nitrogen and Sulfphur
Pollutants as Assessed from Local Measurements,' Fuel, vol 64, p 782 (1985): see
also Highsmith, J.R., 'Fate of Fuel Nitrogen and Sulfur During Coal Gasification:
Detailed Local Pollutant Measurements in an Entrained Coal Gasifier,' M.S.
Thesis, Department of Chemical Engineering, Brigham Young University, Provo,
Utah (1982)

17. Burkinshaw, J.R., Smoot, L.D., and Hedman, P.O. and Blackham A.U.,
'Analysis of Sulfur Pollutants in Three-Phase Coal Combustion Effluent Samples,'
Ind. Eng. Chem. Fund., vol 22, p 292 (1983): see also Burkinshaw, J.R., "'The
Analysis of Sulfur Pollutants From An Entrained-Flow Coal Combustor,' M.S.
Thesis, Department of Chemical Engineering, Brigham Young University, Provo,
Utah (1981)

6B-50

-------
18. Pohl, J.H., and Sarofim, A.F., 'Devolatilization and Oxidation of Coal
Nitrogen/ Sixteenth Symposium (International) on Combustion, The
Combustion Institute, Pittsburgh, PA, p 491 (1977)

19. Peck, R.E.., Midkiff, K.C., and Altenkirch, R.AV 'The Evolution of Nitrogen
From Pulverized Subbituminous Coal Burnt in a One-Dimensional Flame,'
Twentieth Symposium (International) on Combustion, The Combustion Institute,
Pittsburgh, PA, p 1373 (1984)

20. Azuhata, S., Hedman, P.O. and Smoot, L.D. 'Carbon Conversion in an
Atmospheric-Pressure Entrained Coal Gasifier,' Fuel, vol 65, p 221 (1986)

21. Malte, P.D., and Rees, D.P., 'Mechanisms and Kinetics of Pollutant
Formation. During Reaction of Pulverized Coal,' chapter 11 of Pulverized Coal
Combustion and Gasification. Smoot, L.D., and Pratt, P.T., eds., Plenum Publishing
Company, NY (1979)

22. Blair, D.W., Wendt, J.O.L., and Bartok, W., 'Evolution of Nitrogen and
Other Species During Controlled Pyrolysis of Coal,' Sixteenth Symposium
(International) on Combustion, The Combustion Institute, Pittsburgh, PA, p 475
(1977)

6B-51

-------
entrained gasifier. [13]

68-52

-------
Figure 2. Schematic of traversing probe for gasifier.

6B-53

-------
Deionized Liquid/Solid Sample Gas Sample

Cooling/Quenching Collection Collection

Water

Figure 3. Gasifier sample collection system.

-------
300

Q. 5

Q. -D

C T3

O ©

"¦£* u

03 Z2

v_ —-

c c

03 .=

a Z.

c *-

o o

Q £

250
200
150 -
100
50

200

0.95

1.087O

1.06 (pressure, atmj

1.93

2.78

01.00
0.'

O 1 atm
~ 2 atm
A 4 atm

2.10

2.03

4.04

4.75

400 600 800 1000

Average Residence Time, ms

1200

300

E * 250

GL w

a. xj

c o

O flD

2 3

C O

o ©
O

2 >»

200

o 150

100

0 -

Q 588 (residence time, ms)
Q509
1009© 1002

543

832

2 3 4

Pressure, atm

1092

-A

Figure 4. Influence of a) pressure and influence of b) residence
time on effluent NO levels during gasification of
Utah bituminous coal,.

6B-55

-------
8000

E ~

"T £/)

g;_g 6000

3 4

Pressure, atm

Figure 5. Influence of pressure and residence time on
effluent TFN (HCN + NH3 + NO) levels during
gasification of Utah bituminous coal.

68-56

-------
8000

£ •-
a 2

Q. j§

c" "D
o ©

"¦IS "O
CO _3
¦*—* o

C c
CD —

" t

C CD
O c

2? 2000

^ T3

6000

4000

1009

>1002

OSBB

509

(?)407
Q373

919
409

^832

~ 543

~ 841 (residence time, ms)

2 3 4

Pressure, atm

092

O 1 atm
~ 2 atm
A 4 atm

Figure 6. Influence of pressure and residence time on
effluent N2 levels during gasification of Utah
bituminous coal.

6B-57

-------
100

V—

tc
>
c
o
Q

¦e

T>
c
co

c
d

nitrogen conversion D
_Oq—O-Pq

carbon conversion

O 1 atm
~ 2 atm
A 4 atm

200 300 400 500 600 700 800 900 1000 1100 1200
Average Residence Time, ms

Figure 7. Effluent nitrogen and carbon conversion during
gasification of Utah bituminous coal. Open
symbols (O) denote nitrogen conversion, closed
symbols (£) denote carbon conversion.

6B-58

-------
2000Q

S 1500C

600 800

Average Residence Time,

1000

1200

Figure 8. Calculated energy of product gas, in kJ per kg of wet
product gas.

6B-59

-------
Table 1
Utah Coal Properties

Elemental Analysis

Carbon 71.8 %

Hydrogen 5.0 %

Oxygen 12.6 %

Nitrogen 1.5 %

Sulfur 0.4 %

Ash 8.5 %

Mineral Analysis

Silicon Dioxide

59.4

Aluminum Dioxide

17.0

Iron Oxide

3.7

Calcium Oxide

8.7

Magnesium Oxide

2.1

Sodium Oxide

1.1

Potassium Oxide

0.7

Titanium Dioxide

1.3

Phosphorus Pentoxide

0.6

Sulfur Trioxide

5.0

Ash Fusion Temperature, K

Initial 1491

Softening • 1526
Hemispherical 1567
Fluid 1608

Proximate Analysis

Volatile Content 42.2 %

Fixed Carbon 47.3 %

Ash 8.4 %

Moisture 2.4 %

Particle Size Distribution

Range (p.nt)

wt. %

0.01 - 8.00

1.43

8.00 -10.1

2.72

10.1 -12.7

3.46

12.7-16.0

4.30

16.0 - 20.2

6.04

20.2 - 25.4

7.84

25.4-32.0

9.88

32.0 - 40.3

11.18

40.3 - 50.8

13.36

50.8 - 64.0

13.81

64.0 - 80.6

12.39

80.6 -102

8.19

102 -128

4.53

128 -161

0.87

Mass Mean Diam., 48.7 p.m

6B-60

-------
EVOLUTION AND REACTION OF FUEL NITROGEN DURING
RAPID COAL PYROLYSIS AND COMBUSTION

Gregory J. Haussmann and Charles H. Kruger
High Temperature Gasdynamics Laboratory
Department of Mechanical Engineering
Stanford University •

Stanford, CA 94305

ABSTRACT

The evolution and reaction of fuel nitrogen during the rapid pyrolysis and combustion of narrowly
size-graded pulverized coal has been studied in an entrained flow reactor. Data are reported and
interpreted for a 55 |im Montana Rosebud subbituminous coal and a 55 |xm Pittsburgh Seam #8
(PSOC-1451) HVA bituminous coal. The pulverized coal is rapidly mixed into a high temperature
gas stream, resulting in high heating rates (104-105 K/sec). Rapid pyrolysis is observed under these
conditions, with pyrolysis times ranging from 11 msec at 1750K to 50 msec at 1300K. Pyrolysis
products for both coals consist of CO, C02, light hydrocarbons, tar, and soot; nitrogen is released
primarily in tar and as HCN. During pyrolysis, nitrogen is released in approximate proportion to
carbon. Rapid combustion of most pyrolysis products occurs as they are released during fuel-lean
single particle combustion studies with the Rosebud coal, with virtually no hydrocarbons, tar, or
soot detected. During these dispersed particle oxidation studies, nitrogen is detected primarily as
NOx. This indicates that volatile nitrogen reduction to molecular nitrogen does not occur when
such chemistry is confined to the thin volatile boundary layer surrounding individual particles.
Pyrolysis studies using the Pittsburgh #8 coal reveal a significant tar yield, with most of the fuel-
nitrogen released with the tar. The tar then either decomposes into gas-phase products (with
nitrogen as HCN), or condenses into soot. At HOOK, 20% of the nitrogen released during pyrolysis
is incorporated into soot. This portion of the volatile nitrogen is removed from participation in any
nitrogen reduction chemistry which may be occurring, and thus may prove to be an important
pathway for fuel nitrogen conversion to NOx.

6B-61

-------
EXPERIMENTAL PROCEDURE

In the Stanford entrained-flow reactor (Fig. 1) narrowly size-graded pulverized coal is injected into
a high temperature gas flow, which is generated by an argon arc-jet. This allows for the
independent control of the gas temperature (1150 - 2000K) and the oxygen concentration (0 - 20%),
This high temperature gas stream flows through an electrically heated test section with interior
diameter 11.7 cm, at a velocity of 3 to 5 m/s. The Reynolds number based on diameter ranges from
1700 to 2000; in all cases the profiles of velocity and temperature are observed to be uniform both
axially and radially, except for boundary layers near the reactor walls. Coal is supplied to the
reactor from a fluidized-bed feeder, with particle densities ranging from 6 to 10 particles/cm3.
Particle are injected at the reactor centerline through a "sprinkler-head" nozzle, which injects
particles at a 45° angle to the bulk flow. This results in rapid mixing, with a particle mixing time on
the order of one msec. Residence times of up to 200 msec are achieved.

The extent of coal pyrolysis and combustion is investigated through the use of two classes of
diagnostics: sampling gas-phase products and analysis of partly-reacted particles. Visual
observation of reacting particles is also possible through two optical access ports. Gas and solid
samples are collected by means of a single probe with multiple inlets, which can be traversed to
sample at any location in the reactor. Gas-phase measurements include CO and C02 (non-
dispersive infrared spectroscopy), light hydrocarbons (gas chromatography), oxygen
(electrochemical analyzer), NO^ (chemiluminescent analyzer), NHL,, and HCN (ion-specific
electrodes). The extent of reaction is determined by following the conversion of carbon into gas-
phase products (CO, C02, and light hydrocarbons) and condensible pyrolysis products (tar and
soot). Gas-phase measurements are possible starting at a residence time of 5 msec, with an
estimated time resolution on the order of one msec.

Solid samples are collected through a quenching water spray. This technique is estimated to result
in a quenching time of about one msec. The collected material contains partly reacted particles and
condensable pyrolysis products (tar and soot). The tar/soot content of samples is determined
through extraction in tetrahydrofuran (THF), with repeated washings and agitation in an ultrasonic
bath. The THF/tar/soot mixture is decanted off and passed through an 0.2 fim Teflon filter, which

\*y .

has been found to trap soot particles, allowing the dissolved tar to pass through . The tar is
collected after room temperature evaporation of the THF. Micro-samples of partly reacted samples
and the separated tar, soot, and char fractions are analyzed for C,H,N, and ash. Overall sample
morphology is examined with a scanning electron microscope.

6B-62

-------
EXPERIMENTAL RESULTS AND DISCUSSION
Montana Rosebud Subbituminous

Experimental results for two coals are presented; Montana Rosebud subbituminous and Pittsburgh
Seam #8 HVA Bituminous (PSOC-1451). The proximate and ultimate analyses of these coals are
found in Table I. The overall pyrolysis and oxidation behavior of the Rosebud coal at a gas
temperature of 1750K is seen in Figure 2. The rate and yield of pyrolysis is not significantly
affected by the oxygen concentration, with 40% carbon conversion (corresponding to 55% DAF
mass loss) reached at a residence time of 11 msec. Pyrolysis products consist of CO, CO,, light
hydrocarbons, tar, and soot The light hydrocarbons evolved consist primarily of CH4 and Cj
hydrocarbons, with a typical ratio of CH4 / C^'s of 1 / 2.3 . Of the C2 hydrocarbons released, CjII^
is the dominant species, with lesser amounts of CjHj, and very low levels of CjHg. The relative
amounts of CjH, / / C2H6 detected range from 1 / 1.7 / 0.14 to 1 /10 / 0.03 . The condensible
pyrolysis product yield (tar plus soot), on a carbon basis, represents 35% of the total pyrolysis yield
at 1750K. The initial tar yield may be significantly higher than this, since the released tar rapidly
partitions into gaseous products and soot.

For the range of oxygen concentrations studied (2.0% -15%), pyrolysis and oxidation of the
pyrolysis products occur simultaneously at 1450K and 1750K, with all pyrolysis products oxidized
to CO or C02 during pyrolysis. Under oxidizing conditions, condensible pyrolysis products and
hydrocarbons are not detected in significant quantities at any time. This indicates either that tar
combustion is rapid enough to prevent soot formation, or that soot oxidation is rapid as well. Fuel-
nitrogen released during pyrolysis in an oxidizing environment is detected primarily as NO, with
only low levels of HCN, NH3, and tar/soot nitrogen detected. This high conversion of fuel-nitrogen
to NOx results from the fuel-lean conditions of the Rosebud studies (2.0% 02 represents 300%
excess air). At an oxygen concentration of 7.6%, nearly 100% of the original fuel-nitrogen is
converted to NOx, primarily as NO. This high level of NO, formation for particles burning in a
single particle combustion mode holds for gas temperatures ranging from 1150K to 1750K. In this
dispersed burning mode, each particle is surrounded by an oxygen environment, confining nitrogen
chemistry to a thin film surrounding each particle. Since the observed conversion to NOx is
virtually complete, only an insignificant amount of the volatile nitrogen has been reduced to N2
within these films. This suggests that low NOx conversion efficiencies can only be obtained when
released volatile nitrogen species from individual particle clouds aggregate in regions which are
deficient in oxygen, and rich in volatile nitrogen species.

Under oxidizing conditions, an abrupt decrease in the rate of carbon conversion is seen at the end of
pyrolysis, as the transition between pyrolysis and char combustion occurs (Figure 2). The carbon
conversion rates observed in the Stanford reactor during char combustion at 1750K are generally
lower than the diffusion-limited rate3. The char oxidation rate at 2.0% 02 is near the diffusion-

6B-63

-------
limited rate, with a transition from zone ID to zone II kinetics as the oxygen concentration rises
from 2.0% to 15%. The overall combustion rates and burn-out times observed at 1750K are similar
to those measured by Farzan and Essenhigh4 in a high intensity vertical tunnel furnace, where rapid
mixing was generated through impinging jet streams at the top of the furnace.

The DAF mass loss data for the present work is calculated from carbon conversion measurements
by following the evolution of the major elements in the parent coal (C,H,0). Carbon is used as a
tracer, with elemental analysis of the tar, soot, and char fractions allowing a determination of the
amount of hydrogen, oxygen, and nitrogen in each. Thus the percent release from the parent coal of
each component is known, and the corresponding DAF mass loss can be calculated. The correlation
between carbon conversion and DAF mass loss (Figure 3) reflects the initial loss of oxygen and
hydrogen at a faster rate than carbon.

The correlations for pyrolysis carbon, nitrogen, hydrogen, and oxygen loss vs. DAF mass loss are
plotted in Figure 4, representing gas temperatures ranging from 1300K to 1750K. The rapid loss of
oxygen at low mass loss levels indicates that oxygen containing functional groups, such as
relatively loosely bound carboxyl groups, are leaving early during pyrolysis. This observation is in
general agreement with the work of Serio et al5, where the functional group composition of
Montana Rosebud subbituminous was reported. Approximately half of the original oxygen was
reported as C02 (presumably as carboxyl groups, -COOH), with the remainder in ether groups
(—C—O—* ), primarily as "loose ether". These oxygen-rich functional groups would be expected to
evolve early during pyrolysis, resulting in the disproportionate loss of oxygen at early times. The
loss of oxygen and hydrogen vs DAF mass loss is almost identical, with nearly complete oxygen
and hydrogen loss during the devolatilization phase.

The evolution of fuel nitrogen during pyrolysis is fundamentally different from that of oxygen and
hydrogen, as seen in figure 4. A substantial delay exists between the initiation of mass loss and the
evolution of fuel nitrogen. A linear least-squares fit to the data has an x-intercept of 24% DAF
mass loss. After this period of no nitrogen release, the fractional loss of nitrogen is higher than that
of mass, with a curve fit slope of 1.36. This observation is consistent with the hypothesis of a
significant induction period during which essentially no fuel-bound nitrogen is released. Such an
induction period is thought to result from the relatively stable ring structures such as pyridines and
pyrroles (C5N and C4N rings) that comprise fuel-nitrogen6. During this induction period, pyrolysis
products evolve primarily through the elimination of peripheral groups. It is only later in pyrolysis
when significant bridge dissociation has occurred (and perhaps some ring rupture), that fuel-bound
nitrogen is released, either in tar or as HCN. The delay between the onset of mass loss and the
onset of nitrogen loss, as well as the higher fractional rate of nitrogen vs. mass loss, is similar to
results obtained by Pohl and Sarofim7. They also observed an induction period with no nitrogen
loss, followed by nitrogen release with a ratio of fractional nitrogen vs mass loss of 1.25 for a
lignite coal, and 1.5 for a bituminous coal.

6B-64

-------
Nitrogen Loss Correlation,

A simple first-order kinetics model has been applied to the evolution of fuel nitrogen from the
Rosebud coal. The model is outlined below.

dB/dt = -A exp[-E/RTp(t)] B(t) (1)

where:

B(t) = [N.-N(t)VN.

, = Ultimate Nitrogen Loss
Ea = Effective activation energy
A = Pre-exponential constant
Tp(t) = Particle temperature

To obtain the particle temperature as a function of residence time in the reactor, a simple heat
balance is performed, as below.

Q" = heat flux to particle = h(Tp - T<„) - ewoTw4 + eoTp4 (2a)

(4jtrp2)Q" = mpCp(dTp)/(dt) " (2b)

The parameters, and important assumptions, are below:

nip = particle mass, determined vs. time from experimental results
Tg = Gas temperature

Nu = hd/k = 2 + 0.3PrinRe06 (Note: Re decays rapidly after injection)

kg = Thermal conductivity of argon at Tfdm ~ +
ew = wall emissivity, - 1
ep = particle emissivity, - 1
AH^j = pyrolysis heat of reaction, - 0

Cp = Particle heat capacity, an increasing function of temperature

• Internal temperature gradients are small

• Stefan flow negligible

6B-65

-------
Equation (2b) is solved to give the particle temperature vs. residence time, and this data is input into
a computer model to calculate B(t) through numerical integration of eqn, (1). A least-squares best
fit for values of A and Ea for the data obtained in the Stanford entrained-flow reactor can then be
calculated. This transient temperature model for the particle temperature and nitrogen evolution
yields best-fit values of Ea = 22 kcal/mole, and A = 1.15xl06 The results of this calculation for
three gas temperatures can be seen in Fig, 5, where the value of B has been plotted vs. the transient
particle temperature. It can be seen that the evolution of fuel nitrogen occurs during the particle
heating stage, and that most nitrogen evolution occurs well before the particle temperature reaches
the gas temperature. The difference between the particle temperature and the gas temperature when
the loss of nitrogen reaches 50% of its ultimate value is: 90K at a gas temperature of 1300K, 240K
at 1500K, and 390K at 1750K. While at the lower temperatures of many pyrolysis experiments an
understanding of the particle temperature as a function of time may not be as crucial in evaluating
the kinetics ofnitrogen evolution, at gas temperatures and heating rates typical of pulverized coal
combustion (- 10s K/sec)9, the particle temperature history must be considered.

Pittsburgh Seam #8 HVA Bituminous

Recent experimental work in the Stanford entrained-flow reactor has focused on pyrolysis studies of
55 |im Pittsburgh Seam #8 HVA bituminous coal. The overall pyrolysis DAF mass loss for the
Pittsburgh coal at 14Q0K is seen in Figure 6. Pyrolysis is essentially complete at a residence time of
about 20 msec, reaching a DAF mass loss of 40%, with the distribution of pyrolysis products
continuing to shift until about 50 msec. The Pittsburgh #8 coal yields a significant amount of tar as
the primary pyrolysis product, with the initial gas-phase pyrolysis yield fairly low. This is
consistent with the results of Ballantyne et al10, where secondary reactions were eliminated through
laser heating in a cold gas. Under the conditions of their experiment, Pittsburgh #8 coal was found
to release only low levels of gaseous volatiles for particle temperatures up to 2100K at a residence
time of 30 msec. The current experiment reveals a significant initial yield of tar (Figure 6), with
secondary decomposition reactions and soot formation occurring simultaneously in the hot gas
environment. The tar yield drops off rapidly from a peak at the end of pyrolysis (20 msec), with
virtually no tar present at 50 msec. On a DAF mass basis, 60% of the tar present at 20 msec is
converted to gas-phase products, and 40% to sool The strong propensity of the tars evolved during
the pyrolysis of pulverized coals to form soot has been noted by several researchers1,2'11, although
only limited data exists as to the yield and nature of the soot produced. The pyrolysis product
distribution can be broken up further on a carbon basis, as seen in Figure 7. Gas-phase carbon
containing species consist primarily of CO and C02, with only a few-percent of light hydrocarbons
present. About 11% of the original coal carbon is ultimately converted into soot, representing about
30% of the total carbon released during pyrolysis.

68-66

-------
Size distributions of partly-reacted samples measured with a Coulter-eounter® are seen in Figure 8,
for the raw coal and for samples devolatilized at 1400K, collected at various residence times. The
solid curves represent the entire collected sample (char+tar+soot), while the dotted curves represent
the THF-insoluble fraction, i.e. char. As time goes on a significant peak at about 5 microns is
observed to develop in the whole sample size distribution, coinciding with the soot formation seen
in Figure 7. The size-distributions for the two sample fractions are nearly identical, with the peak
near 5 microns not present in the char fraction. Thus this peak is due to the soot/tar fraction of the
partly-reacted sample. Tar released during the pyroiysis of pulverized coal has been found to
condense primarily (-90%) onto the soot particles?. This peak at a particle diameter of 5 microns
represents a large number of individual soot particles cemented together by tar. Scanning electron
micrographs of the soot fraction show an agglomeration, with no identifiable structures down a size
of about 0.5 microns (5000 A), the resolution of the instrument used. Nenninger1, in a study of the
aerosols generated during coal pyroiysis in an entrained-flow reactor, found primary soot particles
ranging in size from 140 - 360 A.

The evolution of fuel-nitrogen is of particular interest, as the distribution of fuel-nitrogen into the
tar, soot, and gas fractions are not well understood. This distribution is represented in Figure 9, for
1400K pyroiysis of the Pittsburgh coal. Preliminary results from ion-specific electrode analysis of
the product gases indicate that HCN is the dominant nitrogen containing gas released, with only low
levels of NHj detected. The initial tar released contains a significant fraction of the evolved fuel
nitrogen. Tar nitrogen has been found to be bound in the same ring structures as in the parent
coal12, and thus is expected to be in relatively stable pyridine and pyrrole structures. The majority
of gas-phase nitrogen containing species result from secondary reaction of the released tar. A
significant portion of the fuel nitrogen, however, is incorporated into the soot fraction, with about
7% of the original fuel-nitrogen converted into soot. This represents 22% of the total fuel-nitrogen
released during pyroiysis ("volatile nitrogen"). Soot nitrogen appears to be relatively stable, with
essentially no loss of nitrogen from the soot fraction occurring from 50 msec to 120 msec.

SUMMARY AND CONCLUSIONS

The evolution and reaction of fuel nitrogen has been measured during the rapid pyroiysis and
combustion of pulverized Montana Rosebud subbituminous and Pittsburgh Seam #8 HVA
bituminous coals. The Rosebud coal exhibits rapid hydrogen and oxygen loss at high heating rates
(104-105 K/sec), with virtually complete loss of these elements during pyroiysis. A substantial
delay exists between the onset of mass loss and the onset of nitrogen evolution, representing a
significant induction period during which essentially no nitrogen loss occurs. This is thought to
result from the relatively thermally stable ring structures in which fuel nitrogen is contained.

During this induction period, pyroiysis products result primarily from the elimination of peripheral
groups (carboxyl, ether, methyl, etc.). It is only later in pyroiysis, after significant bridge

6B-67

-------
dissociation has occurred (and perhaps some ring rupture), that fuel nitrogen is released, either in tar
or as HCN.

A first order kinetics correlation for the evolution of fuel nitrogen has been applied to the Rosebud
data, coupled with a model for the transient particle temperature. The best fit parameters for
nitrogen loss during pyrolysis at gas temperatures ranging from BOOK to 1750K are an activation
energy of 22 kcal/mole, and a pre-exponential factor of 1.15xl06 sec"1. The rapid mixing of coal
particles into the hot gas stream in the current work produces higher heating rates than for the same
gas temperature in facilities with less rapid mixing. As the heating rate is increased, the
temperature range over which most nitrogen loss occurs widens and also moves farther from the
bulk gas temperature. In order to compare kinetic parameters from disparate experiments, an
accurate representation of the particle temperature during pyrolysis is essential.

Pyrolysis and oxidation of the pyrolysis products occur virtually simultaneously during single
particle combustion of Rosebud coal for oxygen concentrations ranging from 2.0% to 15%.
Complete conversion of fuel nitrogen to NOx is seen under oxidizing conditions for gas
temperatures ranging from 1150K to 1750K, Since each coal particle is surrounded by an oxygen
environment in this dispersed particle experiment, nitrogen reduction chemistry is confined to a thin
film surrounding each particle. Under such conditions in the Stanford reactor, only an insignificant
amount of the volatile nitrogen is converted to N2 in these films, indicating that low overall NO,
conversion efficiencies can only be obtained when volatile nitrogen species from individual particle
clouds can aggregate in regions which are oxygen poor, and rich in volatile nitrogen species.

The primary product released during the pyrolysis of a Pittsburgh Seam #8 HVA bituminous coal is
tar, which in the hot gas phase rapidly partitions either into gaseous products, or into soot. This
partitioning of tar into gaseous products and soot occurs quickly, and is complete by a residence
time of 50 msec at 1400K. During pyrolysis at 1400K, 20% of the volatile nitrogen is incorporated
into the soot fraction. Soot nitrogen is stable at this temperature, with no loss of nitrogen from the
soot observed. This portion of the volatile nitrogen is removed from any nitrogen reduction
chemistry which is occurring in the gas phase, and may prove to be a significant factor in the
overall pathways by which fuel nitrogen is ultimately converted into NOx.

ACKNOWLEDGMENT

This work has been supported by the Electric Power Research Institute through Grant # RP 8005-2,
under the direction of Dr. George Qffen.

6B-68

-------
REFERENCES

1. R.D. Nenninger, "Aerosols Produced from Coal Pyrolysis", Sc.D. Thesis, M.I.T. (1986).

2. MJ. Womat, A.F. Sarofim, J.P. Longwell, "Pyrolysis Induced Changes in the Ring Number
Composition of Polycyclic Aromatic Compounds from a High Volatile Bituminous Coal", 22nd
Symposium (International) on Combustion. The Combustion Institute, 1989.

3. G.J. Haussmann and C.H. Kruger, "Rapid Pyrolysis and Combustion of Pulverized Montana
Rosebud Subbituminous Coal", 22nd Symposium (International') on Combustion. The
Combustion Institute, 1989.

4. H. Farzan and R.H. Essenhigh, "High Intensity Combustion of Coal", 19th Symposium
("International) on Combustion, p. 1105-1111, The Combustion Institute, 1982.

5. M.A. Serio, D.G. Hamblen, J.R. Markham, and P.R. Solomon, "Kinetics of Volatile Product
Evolution in Coal Pyrolysis: Experiment and Theory", Journal of Energy and Fuels. Vol. 1,
1987, pp. 138-152.

6. P.R. Solomon and M.B. Colket, "Evolution of Fuel Nitrogen in Coal Devolatilization", Fuel.
Vol. 57, December 1978, pp. 749-755.

7. J.H. Pohl and A.F. Sarofim, "Devolatilization and Oxidation of Coal Nitrogen", 16th
Symposium (International) on Combustion. The Combustion Institute, 1976, pp. 491-501

8. D. Merrick, "Mathematical Models of the Thermal Decomposition of Coal: 2. Specific Heats and

Heats of Reaction", Fuel. Vol. 62, May 1983, pp. 540-546.

9. J.D. Freihaut, W.M. Proscia, and D.J. Seery, "Fuel Bound Nitrogen Evolution During the
Devolatilization and Pyrolysis of Coals of Varying Rank", Presented at the 1987 Joint
Symposium on Stationary Combustion NOx Control, New Orleans, LA.

10. A. Ballantyne, H. Chou, K. Neoh, N. Orozco, and D. Stickler, "Cold Atmospheric Pyrolysis of
Pulverized Coal Using 10.6 [Am Laser Heating", Presented at the American Chemical Society
New York City National Meeting, 1985.

11. WJ. McClean, D.R. Hardesty, and J.H. Pohl, "Direct Observations of Devolatilizaing
Pulverized Coal Particles in a Combustion Environment", 18th Symposium (International) on
Combustion. The Combustion Institute, 1981, pp. 1239-1248.

12. J.D. Freihaut, M.F. Zabielski, and D.J. Seery, "A Parametric Investigation of Tar Release in
Coal Devolatilization", 19th Symposium (International) on Combustion. The Combustion ,
Institute, 1982, pp. 1159-1167.

6B-69

-------
Argon Inlet

Arc Discharge

Flow
S1ralght«ner

Inaulaud
Guard Haalar

Optical
Access Ports

Coal
Fend

Gas! Particle
Sampling Probe

Exhaust
Fan

Figure 1. Stanford Entrained-Flow Reactor.

100

'to

o
>
c
o

c
o
n

CO
Q

CD
O

m
a.

Residence Time (msec)

Figure 2. Rosebud Carbon Conversion vs Residence Time.

6B-70

-------
100

01
>
c
o
o

c
o
a

u
o

0)
Q.

DAF Mass Loss

Figure 3. Rosebud Carbon Conversion vs DAF Mass Loss.

100

m
m
o

c
a>

a
a

i i i i i i—i—i—|—i—i—r-r

MONTANA ROSEBUD

1300K- 1750K

10 20 30 40 50

DAF Mass Loss

Figure 4, Rosebud Pyrolysis Elemental Loss vs DAF Mass Loss.

6B-71

-------
1.0

0.8

0.6

8 0.4

0.2

•! T | i—r -i—i—|—i—i—r

-r -|- i—i—r—i—|—i—r

T = 1750K I

Toas= 1500K -

T^gg = 1300K -

: \ . x

v. \

1 \\ \

: \ \

\ \ \

- \ \

: \ \

: \ '

\ \

\ —

= \

: i i ^

1 1 1 1 1 Tn. 1 !

1 1 t 1 I 1 1 1 1 1 1 1

800 1000 1200 1400 1600

Particle Temperature (Deg K)

Figure 5. Rosebud Pyrolysis Nitrogen Loss vs Transient Particle Temperature.

CTJ

U-
<
D

20 -

i i i i i i i r i i i i ! i i i i i i i i i i i i 1 i i i

PSOC-1451
55 (am
1400K

0.0% o„

3 20 40 60 80 100

Residence Time (msec)

Figure 6. Pittsburgh #8 Pyrolysis DAF Mass Loss vs Residence Time.

120

6B-72

-------
Residence Time (msec)

Figure 7. PSOC-1451 Pyrolysis Carbon Conversion vs Residence Time

Figure 8. Pittsburgh #8 Char Size Distributions vs Residence Time.

6B-73

-------
s

1—i—i—i—I—i—i—i—i—|—i—i—!—r

PSOC-1451
55 jim
1400K

0.0% 00

40 60 80

Residence Time (msec)

Figure 9, Pittsburgh #8 Pyrolysis Nitrogen Conversion vs
Residence Time.

-4

-6

1 1 I ¦ i ¦ t I i i i i e>

100 120

TABLE I - PROXIMATE AND ULTIMATE ANALYSES.

Proximate
Analysis
Dry Wt. %

Ultimate
Analysis
Dry Wt. %

Coal Type

BTU/Ib
(moist)

V.M.

ASH

Montana
Rosebud

10186

42.4

10.1

67.2

4.5

0.78

0.52

17.7

Pitt. #8
PSOC-14S1

14342

3.34

81.5

5.5 '

1.71

1.02

6.93

6B-74

-------
THE EFFECT OF PROCESS VARIABLES ON
NOx AMD NITROGEN SPECIES REDUCTION IN COAL FUEL STAGING

K J Knill and M E' Morgan
International Flame Research Foundation
P 0 Box 10.000, 1970 CA Umuiden
the Netherlands

ABSTRACT

An investigation of the effect of process variables on NOx reduction in the coal
returning zone was undertaken to resolve discrepancies noted in the literature.
High volatile bituminous coals were injected into a turbulent isothermal reactor
simulating a coal returning zone. The change in NOx, HCN and Nia concentrations
were measured over a wide range of process conditions.

The NOx concentration could be reduced from 200-600 ppm to less than 50 ppm under
optimum conditions. Reburn fuel fraction, primary stoichiometry, temperature and
residence time were interdependent .and significant parameters. However, inlet NOx
concentration, coal type and particle size were not significant at optimum
conditions. Reductions in NOx and HCN were accurately predicted by a published
model. The results provide insight into the processes occurring in the reburning
zone and guidelines for applying the results to furnace design are proposed.

INTRODUCTION

Fuel staging or "reburning" has been shown to be an effective method of NOx
emission reduction [1-3]. Reburning involves injecting fuel into a secondary
combustion zone downstream of the primary combustion' zone. . The reburning zone is
operated sub-stoichiometrically in order that hydrocarbon radicals from the
reburning fuel may reduce NO formed in the primary zone according to the reaction:

CHi + NO -> HCN + 0 (Rl)

Once the HCN has been formed, it may be reduced through NC0 and NHi to molecular
nitrogen according to the simplified scheme shown in Figure 1. This process occurs
rapidly in combustion environments with characteristic reduction times of
approximately 50 ms [4 - 6].

The variables of primary importance in fuel staging have been identified as:

i) Reburning fuel type: Several studies suggest that hydrocarbon gas is a more
effective reburning fuel than coal [7 - 11]. However, there is no general
agreement regarding the effectiveness of coal as a reburning fuel.

ii) Reburn fuel fraction: Sufficient hydrocarbon radicals must be available to
reduce NOx, but the guantity of coal required to produce'the radicals is not known.

6B-75

-------
iii) Reburn zone stoichiometric Sufficient oxygen must be available to convert HCN
(formed during hydrocarbon radical reaction with NO) to NCO. The optimum reburning
zone stoichiometry appears to be 0.65 to 0.95.

iv). Reburn zone residence time; A residence time greater than 400 ms is required
to reduce HCN to Nz before tertiary air injection. Some studies had concluded that
residence times as long as 1 s may be necessary. In certain applications, this
parameter may dictate the fuel staging configuration of the furnace.

v) leburn zone temperature: It has been shown that gas fuel staging reactions
are favoured at temperature greater than 1200°C. This is attributed to an increase
in the rate of the NO/CHj and HCN/0 reactions. The effect of temperature on coal
reburning effectiveness has not been resolved. It has been shown that, for some
coals, lower temperature may improve coal reburning effectiveness due to the
reaction of NH3 in the coal with NOx.

vi) Mixing: Mixing between primary NOx and reburning fuel must be rapid and
intense in order to minimize NOx emissions. Greater mixing may be attained by
increasing the reburn fuel injection momentum with additional secondary fuel
transport air. This may lead to early oxidation of reburning fuel hydrocarbon
volatiles, thereby reducing fuel staging effectiveness.

vii) Fuel nitrogen content: Reburning fuels containing nitrogen may have limited,
effectiveness in reducing NO emissions if the primary zone NOx level is less than
150 ppm. However, this laboratory scale observation has been refuted at the utility
boiler scale.

viii) Fuel reactivity: Coals with a higher volatile matter content and higher char
reactivity appear to be more effective reburning fuels. The higher volatile matter
content may result in a higher fraction of CHi radicals necessary to reduce NOx.
Highly reactive chars will require lower burnout times in the tertiary combustion
zone, but this has not been studied extensively in the literature.

A fundamental investigation of the processes occurring in the coal fired, fuel
staging zone was undertaken as a result of the uncertainties noted from the
literature. In this report, the results of a parametric study to determine the most
important variables in coal fuel staging are described. The change in NOx, HCN and
NHo concentrations in the reburning zone, measured over a wide range of process
conditions, was used to describe the optimum process conditions for coal -fuel
staging. Nitrogen species reduction was predicted using a published model.

METHOD

The trial was conducted using the isothermal plug flow reactor (IPFR) illustrated
in Figure 2. The reburning zone was produced by injecting coal into a preheated
gas stream which simulated primary zone coal combustion products. The preheated gas
was produced by burning natural gas in a mixture of air and nitrogen. The nitrogen
was added to moderate the gas temperature leaving the precombustor. The gas
composition along the reactor length could be measured to determine the
concentrations of NOx, HCN and NH3. The effect of the process variables, identified
in the introduction, on reburning effectiveness was tested.

Three bituminous coals were selected for this investigation and their ultimate and
proximate analyses are shown in Table 1. Bituminous coals were selected to study
the effect of volatile matter content on NOx reduction for coals of the same rank.

6B-76

-------
The coals had volatile matter contents ranging from 25% to 42% DAF, These coals
were air dried and pulverized to three particle size ranges: normal (75% < 15 um);
fine (93% < 75 urn); and micronized <100% < 40 urn),

The isothermal plug flow reactor consists of three sections; gas preheater, main
reactor and product collector. The gas preheater is a 30 cm diameter by 1 m long
refractory lined cylinder with a swirled gas burner at one end and a T-piece and
chimney at the other end. The IPFR main reactor consisted of a 117 mm ID by 2 or
5 m long alumina tube heated externally by graphite heating elements. The reactor
was connected to the preheater T-piece by a ceramic venturi. The preheated gas
flowed into the reactor through the venturi where it mixed rapidly with the coal,
injected through a water cooled feeder probe. A small amount of NO was added to the
transport gas in the feeder probe to obtain a predetermined NOx concentration. The
gas flowed out of the reactor through an aspirator and exhausted to vent.

Collection probes could be inserted into the reactor from the bottom. The NOx, CO,
CO2 and Ox concentrations were measured using a steam heated probe. The HCN and NHs
concentrations were measured using the IFRF HCN probe, described elsewhere [12].
The analysis of HCN and NHa were performed according to NEK 6489 and HEN 6472/6644
standards for measurement in waste water.

A Box-Behnkin experimental design [13] was used to determine the simultaneous
effect of the process variables on NOx, HCN and Nth concentrations. Each process
variable was studied at three levels in order to construct a response surface
curve. The conditions were:

RESULTS AND DISCUSSION

The effect of process variables on coal reburning were measured in 59 experimental
tests. Each test was conducted twice: once to measure NOx, CO, CO2 and 02
concentrations and once to measure the HCN and NH3 concentrations. During the HCN
and NHa samples, a probe located at the bottom of the reactor was used to measure
the NOx concentration to ensure repeatability.

Evolution of N-species from the Coals

The formation of N-species from the fine grind fractions of the three coals was
measured at 1250°C and residence tine of 50 and 100 ms. The amount of N-species
devolatilized as a fraction of the original coal nitrogen is listed in Table 2. HCN
was the predominate species, however, a relatively large fraction of NOx was
measured in these tests. It is improbable that NOx was a primary devolatilization
product. ! HCN has been found to be the main product of pyrolysis of N-heterocyclie
compounds and accounts for up to 95% of the primary gas phase N-products from coal
[14, 15]. The high concentration of NOx and NH3 measured in these experiments was
probably due to secondary reactions of nitrogenated tars with the preheated gas
[16, 17]. Few measurements were made in this investigation at residence time less
than 50 ras, and no attempt was made to differentiate between primary
devolatilization products and products of rapid secondary reactions.

1. coal type:

2. particle size:

3. reburn fuel fraction:

4. primary stoichiometry:

5. reburn temperature:

6. primary NOx concentration:

7. reburn residence time:

Obed Mountain, Scotts Branch, Hugo

normal, fine, micronized

10, 20, 30%

1.0, 1.075, 1.15

1100, 1250, 1400°C

200, 400, 600 ppm

0.1, 0.3, 1.0 s

6B-77

-------
Effect of Coal Type

In general, the bituminous coal type had a small effect on the NOx conversion in
the returning zone. The Scotts Branch and Obed Mountain coals had very similar
behaviour in most experiments and the Hugo coal was approximately 10% less
effective in reducing NOx (see Figure 3). The results indicated that there was
little difference between the returning performance of bituminous coals with
different volatile matter contents.

Effect of Particle Size Distribution

The significance of PSD on returning performance was very low. In comparative tests
for Hugo and Obed Mountain coals, the difference in NOx concentration between
micronized and normal grind coals was generally within experimental error. The
insensitivity of returning effectiveness to PSD implies that NOx reduction on the
external surface of char particles was not of primary importance. In practice,
particle size may be reduced.to improve carbon burnout without altering the
returning performance.

Effect of Reburn Fuel Fraction and Primary Stoichiometry

The effect of return fuel fraction and primary stoichiometry are displayed in
Figure 3. Each plot has a constant temperature of 1250°C, fine grind coal and
inlet NOx of 400 ppm. Although each of the variables had separate effects, NOx
reduction was more effectively described by the secondary stoichiometry which is a
function of both the primary stoichiometry and the reburn fuel fraction. In
Figure 3, the front, right-hand edge of each response surface represents a
secondary stoichiometry of 0.7 and the back left-hand edge represents a secondary
stoichiometry of 1.1. At 1250°C, the predicted NOx concentration decreased from
480 ppm to 140 - 210 ppm as the secondary stoichiometry decreased from 1.1 to 0.7.
Secondary stoichiometry had a similar effect on NOx reduction at gas temperature of
1400°C as illustrated in Figure 4. At 0.3 s staging residence time and 400 ppm
inlet NOx concentration, the measured NOx concentration decreased from 300 -
400 ppm to 30 - 100 ppm as secondary stoichiometry decreased from 1.0 to 0.73.

The maximum NOx reduction does not always coincide with the lowest TVFN
concentration. The effect of reburn fuel fraction and primary stoichiometry on TVFN
reduction is shown in Figure 5 for the same conditions used to create Figure 3. The
surfaces in Figure 5 have a saddle shape with a peak in TVFN concentration at
secondary stoichiometry of 0.7 and 1.1. The lowest concentration was predicted
along a line where the secondary stoichiometry was between 0.86 and 0.96. The
optimum secondary stoichiometry to minimize the TVFN concentration was coal
dependent. The minimum TVFN concentration for Hugo coal was at a secondary
stoichiometry of 0.86 compared to 0.96 for Obed Mountain coal. This was protably
due to the higher fraction of devolatilized N-species from Obed coal (see Table 2).

Effect of Temperature

The reburn zone temperature had the greatest influence on the NOx reduction
potential. The effect of temperature and secondary stoichiometry on NOx reduction,
for Scotts Branch coal at an inlet NOx concentration of 400 ppm and residence time
of 0.3 s, is illustrated in Figure 6. At a temperature.of 1100°C, the secondary
stoichiometry had little- effect on NOx reduction and a concentration of
260 - 300 ppm was measured. At 1400°C, the secondary stoichiometry had a strong
effect with NOx concentrations decreasing from 350 ppm to 50 ppm as the

6B-78

-------
stoichiometry decreased from 1.0 to 0.8. The fact that a reburning zone temperature
of 1400°C was necessary to obtain a NOx concentration of less than 100 ppm is in
agreement with conclusions made in the literature [18, 19].

The decrease in TVFN concentration at higher temperature was due to the reduction
of both NOx and HCN. The NH3 concentration tended to rise at higher temperature due
to NO and HCN destruction, however, with sufficient reburning residence time, the
NH3 would also be reduced. This is shown in Figure 7, which contains surface
response curves for the NOx and TVFN concentrations as a function of temperature
and time. The curves were generated for 20% reburning fuel using Scotts Branch coal
at a primary stoichiometry of 1.0 (secondary stoichiometry equal to 0.8) and an
inlet NOx concentration of 400, ppm. At 1100°C, the NOx and TVFN concentrations were
insensitive to residence time; the outlet NOx and TVFN concentrations were
approximately 280 and 500 ppm after 0.1 s with little reduction thereafter. This
suggests that the initial staging reactions were completed in less than 0.1 s and
additional residence time was not effective in reducing the NOx or TVFN
concentrations. At 1400°C the MOx and TVFN continued to decrease at longer
residence time. To achieve a NOx concentration of less than 100 ppm, residence time
greater than 0.2 s was necessary at a temperature of 1400°C and secondary
stoichiometry of 0.8.

Effect of Inlet NOx Concentration

The fractional conversion of NOx increased with higher inlet NOx, however,
reduction was only possible under selected conditions at 200 ppn inlet NOx
concentration. The effect of inlet NOx concentration on NOx reduction at different
temperatures is illustrated in Figure 8. The data are presented for Scotts Branch
coal at secondary stoichiometry of 0.8, - 0.9 and a residence time of 0,3 to 1.0 s.
At 1100°C, the outlet NOx concentration increased from 220 to 400 ppm as the inlet
NOx concentration increased from 250 to 600 ppm. At a temperature of 1400°C, the
outlet NOx concentration, was- approximately 50 ppm and was independent of inlet NOx
concentration. This indicated that the NOx may be almost completely reduced in
0.3 - 1.0 s for an inlet NOx concentration of up to 600'ppm.

For industrial application, this suggests that the reburning fuel should be
injected into a region of the flame where NOx is concentrated. This would maximize
the NH3 concentration early in the reburning zone resulting in the rapid reduction
of the remaining NOx. In a coal flame, large gradients in NOx concentration may
exist prior to tertiary air injection [20]. If these gradients also exist across a
furnace section, a detailed understanding of the NOx concentration profile could be
used tc determine the optimal reburning fuel injection location.

Effect of Residence Time'

Residence time had a strong influence on NOx and TVFN reduction at all process
conditions. A residence time greater than 0.3 s was necessary in order to obtain
absolute NOx concentrations less than 100 ppm (see Figure 7). Three tests were
selected for detailed measurement of N-species (DF1, DF2, DF3). These tests were
the most effective process conditions for TVFN reduction.

The rate of NOx reduction from the detailed flames was predicted using a model for
NOx decay in staged flames [21, 22]. The model makes no account for CHi/NO type
reactions but is based on the mechanism:

NHi + NO -> Nz + products (R2)

N + NO -) Nz + 0 (R3)

6B-79

-------
The concentrations of NHi species, where i may vary from 1 to 3 were assumed to be
in semi-equilibrium according to the reaction:

NHi + OH <=> NHi-i + HzO (34)

The OH radical concentration was also assumed to be in equilibrium.

The equations which describe the rate of NO and HCN decay in the furnace are:

1 d [NO]

=- [NHa] x 0i (OH,T) , (1)

[NO] dt

'1 d[HCN] [OH] * •'

= - fOH] x 03 (T) - X sk (T) (2)

[HCN] dt [HzO]

where , 0a and #« are functions of the OH radical concentration and temperature.

The change in N-species concentration was calculated using the measured HCN, NHa
and NOx concentrations from the detailed flames. The calculations began at 0,05 s
to separate the effects of devolatilization and the CHj/NO reaction (Reaction 1)
from Reactions 2-4. The HzO concentration was calculated from the inlet CO2
concentration and from the known ratio of CG2/H2O in the products of combustion for
natural gas.

Detailed Flame 1. The first detailed flame was measured at a temperature of 12504°C
and a secondary stoichiometry of 0.75. The inlet NOx concentration was set to
400 ppm and the O2 concentration to 0%.

The measured concentrations of nitrogen species and permanent gases are shown in
Figure 9. The NOx concentration decreased from 330 ppm at 0.05 s to 180 ppm at
1.0 s. The concentrations of HCN and NH3 rapidly increased in the first 0.1 s. The
rapid increase could be partially attributed to the devolatilization of nitrogen
species from the reburning fuel. Approximately 150 of the 200 ppm of HCN and 60 of
the 90 ppm of NH3 would be expected to come from the coal devolatilization based on
separate measurements (see Table 2). The remainder was probably attributed to fuel
staging reactions. After 0.1 s, the HCN decayed from 200 to 90 ppm and the NHa
increased from 90 to 165 ppm. The NH3 was formed from the conversion of HCN. Once
NH3 was formed, there was a very slow conversion to N2. Therefore, at 1250°C,
residence time greater than 1.0 s would be required to further reduce TVFN using
reburning. This may not be acceptable in industrial application.

The rate of NOx and HCN reduction in DF1 was underpredicted using the NOx decay
model. It may be possible that the OH radical concentration was not in equilibrium
at these conditions., Bose et al [21] noted that the OH radical concentration was
greater than equilibrium in rapidly cooled flames. Using the OH radical
concentration, suggested by Bose et al, led to overprediction of the rate of NOx
and HCN reduction [23].

Detailed Flame 2, The second detailed flame was measured at a reaction temperature
of 1400°C, primary stoichiometry of 1.0 and NOx concentration of 400 ppm. Secondary
stoichiometry was set to 0.75 by injecting 25% reburning fuel fraction. It was
foreseen that this condition may be applicable to industrial and utility boilers.

6B-80

-------
The concentration of nitrogen species and permanent gases as a function of
residence time is shown in Figure 10. The NOx concentration decreased from 400 to
30 ppm as the residence txme increased from ,0 to 1.0 s. It was estimated, fron
separate experiments, that the TVFN concentration from devolatilization was
340 ppm. The TVFN concentration was reduced to 490 ppm in 0.05 s and from 490 to
330 ppm up to 1.0 s. Rates of TVFN and the NOx reduction in the first 0.05 s were
an order of magnitude faster than after 0.05 s indicating that a different
mechanism was predominant during initial reduction. The HGN concentration increased
in the first 0.05 s to 150 ppm and .decreased to 25 ppm thereafter. Although 150 ppm
HCN was equal to the devolatilization yield from the coal, there was undoubtedly
HCN formed from the CIi/NO reaction (Reaction 1). The reduction of NOx to HCN was
probably occurring at a similar rate as the reduction of HCN to NCO. The NH3 yield
continued to increased to a final value of 290 ppm. After 0.3 s, NHa was the only
significant nitrogen species.

The model accurately predicted the NOx and HCN concentrations up to the furnace
position where their concentrations equalled 25 ppm. After this point, the model
predicted that the NOx and HCN concentrations tended to zero whereas a residual
concentration of 10-25 ppm was measured in the reactor. This suggests that the OH
radical concentration was in equilibrium at these conditions.

The results for DF2 suggest that this process configuration would be suitable for
application in a utility boiler. The reduction of NOx was rapid and complete in
0.3 s, although it would be necessary to reduce a fraction of the NHa to N? in the
tertiary combustion zone to minimize NOx reformation.

Detailed Flame 3. The third detailed flame was measured at a temperature of 1400°C,
inlet stoichiometry of 1.075 and1 400 ppm NOx. The secondary stoichiometry was set
to 0.85 by injecting 25% reburning fuel.

The concentrations of M-species and permanent gases are illustrated in Figure 11.
The NOx concentration decreased from 400 ppm to 20 ppm over 1.0 s. The rate of
reduction of NOx was slower than for DF1 or DF2. This was expected because the HCN
and NHa concentrations were only 19 and 15 ppm after 0.05 s resulting in a slower
rate of NO destruction. The HCN and NHa concentrations were much lower than
measured under devolatilization conditions suggesting that they were- rapidly
converted in the reactor. No oxygen was measured in the reactor at 0.05 s,
indicating that the process was rapid and complete. The NOx gradually decreased to
20 ppm with only a small increase in the concentration of NHa. This resulted in a
lower TVFN after 1.0 s {130 ppm). than was achieved in DF1 or DF2.

The model accurately predicted the NOx concentrations in DF3. The rate of NOx
reduction was slower than in DF2 since the concentration of NB3 was an order of
magnitude less. There was sufficient NHa to cause complete reduction of NOx in 1 s.

The success of this relatively simple model to predict NOx and HCN reduction is
useful in understanding reburning. It implies that the rate of NOx reduction may be
modelled with a knowledge of the NHa concentration and the reactor temperature.
Therefore, the CHi/NO reaction (Reaction 1) which dominated the reduction ¦ in the
first 0.05 s may be separated from the nitrogen reduction reactions which dominated
after 0.05 s. The reduction'scheme reported is applicable to both air and fuel
staging processes, and therefore, may be incorporated into a relatively simple NOx
reduction model for both systems.'

It may be postulated that a sufficient NH3 concentration was necessary to reduce
NOx in the reactor after 0.05 s. In coal fuel staging, this NH3 may be formed from
the coal volatile nitrogen products or from the reburning process itself. If the

68-81

-------
NOx and NHs concentrations- are-equal after 0.05 s, Reaction 2 would proceed rapidly
and the NOx and NHs' could be completely converted to Nz. In practice, balancing the
NOx and NH3 concentrations may be difficult to achieve. However, this concept
suggests guidelines for the implementation of fuel staging in furnaces.

CONCLUSIONS

The following conclusions may be made' regarding the importance of the measured
process variables on fuel staging.

Bituminous coal type had a small effect on NOx reduction. High volatile bituminous
coals could reduce NOx over a wider range of operating conditions. Hedium volatile
bituminous coal could be used to reduce NOx.

Particle size distribution was not a significant parameter affecting NOx or TVFN
reduction. This indicates that particle size may be varied without altering fuel
staging performance.

The effects of reburning fuel fraction and primary stoichiometry on NOx reduction
were strong. These two variables could be transformed to a single variable, the
secondary stoichiometry. • The concentration of NOx decreased with secondary
stoichiometry, however, TVFN was minimum at a secondary stoichiometry of 0.8 - 0.9.
Coals with a larger fraction of volatile matter required -a .higher secondary
stoichiometry to minimize TVFN concentration.

Temperature had a strong effect on NOx and TVFN reduction. At 1100°C. the NOx
concentration was reduced from 400 ppm to a- minimum of 260 ppir.. At 14G0°C, the NOx
concentration was reduced from 400 ppm to 30 ppm. Higher temperature promoted rapid
decay of HCN and NHs to N2.

The fractional conversion of NOx usually increased as inlet NOx concentration
increased. At 1400°C, the outlet NOx concentration was independent of the inlet NOx
concentration and NOx was reduced from 200 - 600 ppm to 50 ppm in 0.3 - 1.0 s. The
inlet NOx concentration had little effect on the concentrations of HON or NH3¦ It
was concluded that the NOx and NH3 could rapidly react together to form N2 at the
conditions studied.

Residence time greater than 0.3 s was necessary to obtain NOx concentrations less
than 100 ppm. The time to reduce NOx was dependent on the temperature and the
secondary stoichiometry. Less time was- required to reduce -the NOx at a higher
temperature and lower secondary stoichiometry.

The NOx and HCN reduction occurring after 0.05 s could be accurately predicted by a
simple model for NOx reduction in air staged flames. This suggested that the CHi/NO
reaction was complete before 0.05 s and the NHi/NO reaction dominated thereafter.

Application of fuel staging in burners or furnaces may be limited by the residence
time at high temperature necessary for complete NOx and TVFN reduction. Assuming
good mixing, a residence time of 1.0 s at 1400°C should be sufficient to completely
reduce NOx to less than 50 ppm and TVFN to less than 150 ppm. Lower temperature and
shorter residence times result in higher concentrations of both NOx and TVFN.

ACKNOWLEDGEMENTS

The IFEF would like to acknowledge that the work described in this report was
executed on behalf of and supported financially by organisations in Canada, the

6B-82

-------
Federal Republic of Germany and the Netherlands. The programme has been organised
under the auspices of the International Energy Agency and forms part of Annex II of
the implementing agreement for a programme of research, development and
demonstration of coal combustion sciences.

REFERENCES

1. Y. Takahashi et al. "Development of Mitsubishi "MACT" In-furnace NOx removal
process". Presented at US-Japan NOx Information Exchange, Tokyo, May 25 - 30,
1981.

2. R. Waibel and D. Nickeson. "Staged fuel burners for NOx control". Presented at
IFRF ¦ 8th Members Conference, Noordwijkerhout, the Netherlands, May 28 - 30,
1986.

3. J. k. Mulholland and R. E. Hall. "Fuel oil reburning application for NOx
control to firetube package boilers". Presented at the Joint ASME/IESS Power
Generation Conference, Milwaukee, October 20 - 24, 1985.

4. W. S. Lanier, J. A. Mulholland and J. T. Beard. "Reburning thermal and chemical

processes in a two-dimensional pilot scale system". Twenty-first Symposium

(International) on Combustion, The Combustion Institute, Pittsburgh, 1988 (in
press),

5. Y. H. Song, D. W. Blair, V. J. Siminski and W. Bartok. "Conversion of fixed
nitrogen to Nz in rich combustion". Eighteenth Symposium (International) on
Combustion, The Combustion Institute, Pittsburgh, 1981, pp 53 - 63.

6. Y. H. Song and V. Bartok. "Rate controlling reactions in fixed nitrogen
conversion to Nz". Nineteenth Symposium (International) on Combustion, The
Combustion Institute, 1982, pp 1291 - 1299.

7. S. L. Chen, J. M. McCarthny, ¥. D. Clark, M. P. Heap, V. p.. Seeker and
P. W. Pershing. "Bench and pilot scale process evaluation of reburning for

in-furnace NOx reduction". Twenty-first Symposium (International! . on

Combustion, The Combustion Institute, Pittsburgh, 1986.

8. J. A. Mulholland and R. E. Hall. "The effect of fuel nitrogen in reburning
application to a fire tube package boiler:. Proc of 1985 Joint Symp on
stationary combustion NOx control. Vol 1, No. 4, 1986.

9. W. J. Phelan. The effect of pulverised coal type and burner parameters when
staging air combustion for NOx reduction, International Flame Research
Foundation, IFRF Doc.No. K 20/a/180.

10. R. J. Roby and C. T. Bouwman. "Effects of hydrocarbons on nitric oxide removal
in rich premixed hydrogen oxygen flames". Presented at The Combustion
Institute, Canadian and Western States, Spring 1986 section meeting, Banff,
Canada, April 27 - 30, 1986. '

11. J. 0. 1<. Xendt, C.V. Sternling and M. A. Katovich. "Reduction of sulphur
trioxide and nitrogen oxides by secondary fuel injection". Presented at
Fourteenth Symposium (International) on Combustion, The Combustion Institute,
Pittsburgh, 1973, pp 897 - 904.

6B-83

-------
12. J. P. Smart and T. Maalman. An analytical procedure for the guantitative
determination of NH3 and HCP in combustion systems, International Flame
Research Foundation, IFRF Doc.No. F 72/a/10, October 1987.

13. G. E. P. Box and P. W. Behnkin. "Some new three level designs for the study of
quantitative variables". Technosnetrics, Vol 2, 1960.pp 455 - 475.

14. J. D. Freihaut, W. M. Proscia and D. J. Seery. "Fuel bound nitrogen evolution
during the devolatilization and pyrolysis of coals of varying rank". Presented

at 1987 Joint Symposium on stationary combustion NOx control, New Orleans,

March 23 - 26, 1987.

15. J. D. Freihaut and D. J. Seery. "Evolution of fuel bound nitrogen during the
vacuum thermal devolatilization of coal". ACS Preprints, Volume 26 No. 3,
New York, August 23 - 28, 1981.

16. R. E. Peck, R.A. Altenkirch and K. C. Midkiff. "Fuel nitrogen transformations
in one-dimensional coal-dust flames". Coabustion and flame. Volume 55, 1984,
pp 331 - 340.

17. S. I). Chen, M. P. Heap, D. V. Pershing and G. B. Martin. "Influence of coal
composition on the fate of volatile and char nitrogen during combustion".
Nineteenth Symposium (International) on Combustion, The Combustion ~ Institute,
1982, pp 1271 - 1280.

IE. h. R. Thorne, M. C. Branch, D. W. Chandler, R. J. Kee and J. A. Miller.
"Hydrocarbon/nitric oxide interactions in low pressure flames". Presented at

1986 spring meeting of the Canadian and Western _ St a t es sections of the

Combustion Institute, Banff, April 28 - 30, 1986.

19. V. F. Farnmayan, X. Togan, Yu. Tae-u, J. D. Tear and J. M. Beer. "Reduction of
NOx emission from natural gas flames by staged fuel injection". Proc 1985 Joint
Symp on stationary combustion NOx control, Boston, May 6-9, 1985.

20. R. Weber, J. P. Smart and W. J. Phelan. NOx reduction Kith coal firing by
application of both internal air staging and fuel rich precocbustors,
International Flame Research Foundation, IFRF Doc.No. F 037/a/16, February

1987.

21. A. C. Bose and J. O. L. Wendt, "Pulverized coal coabustion: Fuel nitrogen
mechanisms in the rich post-flame region". Twenty-second Symposium
(International) on Combustion, the Combustion Institute.

22. A. C, Bose, K. M. Dannecker and J. 0. L. Wendt. "Coal composition effects on
mechanisms governing the destruction of NO and other nitrogeneous species
during fuel rich combustion". Journal of Energy and Fuels.

23. K. K. Knill, J. S. A. Dekker and M. E. Morgan. Evaluation of the Effect of
Process Variables on NOx and Nitrogen Species Reduction in Coal Fuel Staging.
International Flame Research Foundation, IFRF Doc.No. F 037/a/20, November

1988.

6B-84

-------
C2H2

HCCO

CH2*—-^2—^HCNO

~ H

~ NO

HCN

~ NO

»H2

~ 02

>NO

-H

;nco

Fig. 1:Model for NOx reduction by fuel staging

* sampling system

Fig•2:Schematfc of the isothermal plug flow reactor

6B-85

-------
Fig-3=Effect of reburn fuel fraction and primary
stoichiometry on NOx concentration

6B-86

-------
450

.1 350
"5

I 250

150

50
0

Fiq-4=Effect of secondary stoichiometrv ond residence time on

NOx concentration

Scotts Branch coal

n /

•
. *

o y
/

/ /
s /

/ /

° /

•

Q 03 s

/ VI

/ /

•

/^0s

• ! :

0.7 G,8 09 1.0 1.1

secondary stoichiometrv

6B-87

-------
TVFN 700
concentration
(pprnl 600

Obed Mountain

Scotts Branch

Huoo

1.15 Xp

30 fnB

Fig• 5:Effect of reburn fuel fraction end primary_
stoichiometrv on TVFN concentration

6B-88

-------
Fig ¦ 6 Effect of secondary stofchiometry and temperature on

NOx concentration

Fig ¦ 7 : Effect of temperature and residence time on NOx and

TVFN concentrations

6B-89

-------
CSQ
"1350

1 250

£
o

1 150

50
0

Fig ¦ 8 : Effect of inlet NOx concentration on outlet NOx concentration

at different temperatures

400

concentration
tppm)

300-
200-
100-
0

0 0.5 residence time IsS 1.Q

Fig-9 '• N—species profiles for detailed flame 1

6B-90

inlet NOx concentration (pen)

~ p

rneas pred.

NOx d

HCN v

NH3 o

~ ~

*""" —

— — —«.

O
O

-------
4,00

concentration
IppmJ

300

200-

100'

meas- pred.

HCN
-------
Table 1

ULTIMATE AND PROXIMATE ANALYSES OF TEST COALS

ANALYSIS

ULTIMATE
-------
NO EMISSIONS IN A PILOT SCALE CIRCULATING FLUIDIZED BED COMBUSTOR
x

J. Zhao, J.R. Grace, C.J. Lira, C. Brereton and R. Legros
Department of Chemical Engineering
University of British Columbia
2216 Main Mall
Vancouver, Canada V6T 1W5
and E.J. Anthony
CANMET, Energy, Mines and Resources
555 Booth Street
Ottawa, Canada K1A 0G1

ABSTRACT

A variety of fuels, including woodwastes, several coals and pitch, have
been burned in a pilot scale circulating fluidized bed ccmbustor of
cross-section 0.15 m by 0.15 m and height 7.3 m. NO^ emissions were in the
range 80-220 ppm (40-110 mg/MJ), depending on the fuel type, and operating
conditions. These emissions increased with bed temperature, excess air
ratio and the feed rate of limestone used for in situ capture of sulphur.
For coals with similar nitrogen contents, N0^ emission levels are higher
for coals having higher, volatile contents. Preliminary measurements of radial
N0x concentration profiles indicate that there are substantial gradients
between the wall region and the centre of the combustor. This phenomenon must
be understood if results are to be scaled to large units.

INTRODUCTION

As an outgrowth of conventional fluidized bed (bubbling bed) combustion,
circulating fluidized bed combustion (CFBC) has found increasing applications
in the combustion of a wide variety of fuels, such as coal, peat, woodwaste and
oil (1). Circulating fluidized beds operate in the fast fluidization regime as
a relatively concentrated suspension of mean density typically 15-80 kg/m3 with
a superficial gas velocity of the order of 5-12 m/s. Solid particles (with a
mean diameter of 50-500 microns) are carried over from the top of a riser,
captured by cyclones or other gas/solid separators and returned to the base of
the riser through non-mechanical valves (e.g. L-valve) or seals.

The major advantages of CFBC compared with conventional fluidized beds are fuel

68-93

-------
flexibility, improved turndown, higher combustion efficiency and lower
pollutant emissions (2). Favourable NO^ emissions (typically 100-200 ppm)
have been reported (e.g. 3).

In spite of the increasing interest in CFB combustion, little information has

been reported concerning NO^ emissions and the factors which control these

emissions under the conditions of CF1C. By contrast, NO formation and

J x

reduction mechanisms in conventional fluidized bed combustion (F1C) have been
investigated fairly thoroughly.

It is generally accepted that formation of NO^ in bubbling fluidized bed
combustors is predominantly by oxidation of fuel-bound nitrogen. Temperature
has a strong effect on NO^ emission. In one study Gibbs et al. (4) reported
that in the range 650-850°C, NO increases linearly with temperature at a
rate of 2.6 ppm/°C. Thermal NO (I.e. NO derived by oxidation of nitrogen in
the air) is not detectable for temperatures of 800°C or below (5).

Excess air ratio is another major factor affecting NO^ formation. It is

clear that NO^ emissions increase with the excess air ratio. For example,

Hirama et al. (6) found that NO emissions almost doubled when the excess air

ratio was Increased from 1.0 to 1-25. Bed materials may have a catalytic
effect on NO formation. According to Hirama et al., among the bed materials
tested (silica sand, coal ash and limestone), limestone gave the highest NO ¦
Clearly the composition of the particular limestone may also play a role.

Detailed study of NO formation and reduction in bubbling fluidized bed
combustion of coal was conducted by Beer and colleagues (7). Axial profiles
of NO concentration were measured in an experimental unit. It was found that
NO is produced in the bubbling bed, then substantially reduced in the
freeboard due to the reduction of NO by char.

Staged combustion has been stated to be an effective way to reduce NO^
emission from fluidized bed combustion. However, Hirama et al. (8) reported
that staged combustion had little effect on NO^ emission- in CFB combustion,
unless fresh limestone is used.

Although CFBC and bubbling FBC have many similarities, the mechanism of NO^
emission may differ in important respects. The objective of this paper is to
show the experimental effects of operating variables on NO^ emissions based
on a well-instrumented pilot plant CFBC unit. Axial and radial concentration
profiles are presented in order to provide an improved understanding of the
mechanisms and factors affecting NO^ emissions.

6B-94

-------
EQUIPMENT AND INSTRUMENTATION

USC CFBC Facility

A schematic of the UBC CFBC facility appears in Fig. 1. The combustor is a
7.3m tall by 0.15 m square refractory-lined column. Mr is fed into the
combustor through a primary air distributor at the base. Secondary air is
introduced 0.9 m above the base either as opposed jets or with swirl. The
ratio of secondary to primary air can be varied over a wide range. During
start-up, the primary air is preheated by natural gas- through a gas burner.
The secondary air can be preheated during operation by the flue gas. A variety
of fuels can be fed by a screw feeder, by a rotary valve or by a gear pump (for
liquid fuels only). Heat can be removed from the riser by an instrumented
membrane wall cooling surface or by vertical internal stainless steel tubes.
Entrained solids leaving the top of the riser are separated by a
refractory-lined' primary cyclone and fall through a storage hopper into a
standpipe. Solids are then returned to the bottom' of the riser by an L-valve,
at rates which are controlled by air injection. Gas containing entrained fine
particles leaving the primary cyclone passes through a secondary cyclone, a
jacketed cooling section and a baghouse. Fine solids captured by the secondary
cyclone are recycled back to the bottom of the riser via an eductor'. The
combustor assembly is provided with a series of thermocouples, pressure taps,
sight glass and sampling ports. More detailed descriptions of the DBC CFBC
pilot plant are provided elsewhere (9,10).

Instrumentation and Gas Sampling System

During the experiments, temperatures, pressures and flue gas compositions are
recorded by a computer at five-minute intervals. Flue gas is monitored
continuously by a series of on-line gas analyzers, including non-dispersive
infrared (NDIR) analyzers for sulphur dioxide., hydrocarbon, carbon monoxide and
carbon dioxide, a chemiluminescent analyzer for nitric oxides (NO and N02)» and
a paramagnetic analyzer for oxygen. Gas samples'can also be withdrawn from a
sampling port between the primary and secondary cyclones. The combustor is
operated at small pressures, forcing the sample gas to pass through a stainless
steel filter, a cooling coil, a water trap (ice bath) and a drying tube to
remove moisture before reaching the gas analyzers.

In order to understand the mechanism of pollutant formation and provide useful
data for mathematical modelling, a multiposition sampling system was installed.
As shown in Fig. 2, sample gas can be withdrawn at five different heights along
the combustor. After passing through a filter, sample gas is cooled by a
water-cooled heat exchanger. Condensed water is drained from the bottom of the

5B-95

-------
exchanger. A drying tube filled with magnesium perchlorate (Mg(C10l( ) is used
to remove further moisture before the sample gas reaches the manifold on the
control panel. Air purges can be used to flush out solids from the filters
when samples are not being taken. The gas probe, made of stainless steel, is
connected to the filter by flexible stainless steel tubing allowing the gas
sampling probes to be moved to different radial positions. The five sampling
levels are shown in Fig. 2.

EXPERIMENTAL RESULTS AND DISCUSSION

Operating Conditions

For most of the combustion tests, bed temperatures were in the range 850-900°C.
Typical steady-state temperature distributions are given in Fig. 3 for several
fuels. Superficial gas velocities above the secondary air injection ports are
6-10 m/s. The secondary-to-primary air ratio is usually maintained in the
range 0.5-2.0. The excess air ratio is 1.17-1.24, giving 3-41 oxygen in the
flue gas. Approximately 100 kg of olivine sand (density 3066 kg/m3) with a
mean diameter of about 240 microns is used for start-up.

NO Emission from Combustion of Different Fuels
x

In early combustion tests, three different coals (Esso, Highvale and Minto),
two pitches and two woodwastes (sawdust and hogfuel) were burned. These runs
were performed without ash recycle from the secondary cyclone. The operating
conditions and NO^ emissions for some runs are given in Table 1, while fuel
analyses are provided in Table 2.

As indicated in Table 1, higher N0^ emissions occur when there is a higher
nitrogen content of the fuel. The net molar conversion of fuel nitrogen to NO
(moles N0x emitted/moles N in the fuel fed) were 0.17, 0.59 and 0.16 for Minto
coal, hogfuel and CANMET pitch, respectively. The ratios are in the range
0.05_0«2, suggested by Becker et al. (11) for coal and pitch. However, it
seems that the combustion characteristics of hogfuel are quite different as far
as NO^ emissions are concerned, giving much higher emissions when expressed on
this basis. This may be related to the high volatile content and low residence
time of the woodwastes.

In order to increase the operational flexibility and to improve the sulphur
capture and combustion efficiency of the CFB corabustor, a series of
modifications were made in early 1988 (10). The key changes which might have
an effect on emissions were the provision for recirculation of the fines
captured in the secondary cyclone, replacement of the bottom section of the

6B-96

-------
riser by a diverging section, and a switch from screw feeding Co pneumatic
feeding of both solid fuel and sorbent. Since these modifications, four
different coals, a Saskatchewan lignite, an Alberta low sulphur sub-bituminous
coal (Highvale), a New Brunswick high sulphur bituminous coal (Minto) and a
B.C. anthracite were burned. High combustion efficiency and sulphur capture
were achieved in these runs (10,12).. Analyses of these coals are also provided
in Table 2. Some preliminary N0x emission results follow.

Effect of Excess Air Ratio

Figure 4 shows the relationship between emission and excess air ratio for

Highvale coal. The apparent linear increase of NO^ level with excess air shows
the importance of controlling excess air in order to reduce NO^ emission.

Twenty percent or less excess air (oxygen in flue gas < 3.5%) allowed the NO

level to be maintained below 170 ppm for the conditions specified in Fig. 4.

Effect of Temperature

Although the effect of temperature on NO^ emission was found to be less strong
than in bubbling bed combusion (e.g. (4)), increasing temperature does lead to
higher NO^ emissions.. For Highvale coal at 760 and 890oC1 N0^ levels were 70
and 160 ppm for the same excess air. The rate of increase is about 0.8 ppm/°C.

Figure 5 shows the variation of temperature and N0^ after a step decrease in

the fuel feed rate was made during a combustion test with the lignite. The

temperature of the combustor gradually decreased from 850 to 725°C over a

period of two hours while other conditions remained unchanged. SO^ in the flue

gas decreased from 190 to 105 ppm. Note that the increase of CO level and

carbon loading due to the lowered temperature may have also contributed to the

reduction in NO emissions which occurred after the step decrease in the fuel
x

feed rate.

Effect of Staged Combustion

There was no clear correlation between NO emission and secondary to primary

air ratio for the conditions studied. Table 3 gives the NO emission for

combustion of the B.C. anthracite. The slight change of N0^ emission when the
secondary-to-primary air ratio was doubled shows that the effect of staged
combustion for this particular combustor is small. Similar conclusions can be
drawn also for other fuels, as seen in Table 1. In the current tests,
secondary air was added only 0.9 i above the primary air, and it may be
necessary for the secondary air entry to be much higher to have a noticeable
beneficial staging effect. 'As pointed out by Hirama et al. (8), staged

6B-97

-------
combustion in CFBC should be less effective than in conventional bubbling
fluid!zed bed combustion.

Effect of Ca/S Ratio

Some,

Similar to the findings fromAbubbling FBC units, e.g. (6), the influence of

limestone on NO^ emission was quite strong. Typical results appear in Table 4.

No emissions almost doubled when the molar Ca:S ratio increased from 0 to 3.0
x

for the conditions described.

Fuel Character and NO Emissions
x_

From our experience with different fuels it would appear that both carbon
content and the percentage volatile matter in the fuel may influence NO
emissions. For example, under similar combustion conditions, the unreactive
anthracite (Table 2), has a nitrogen content of 0.86%. Its emission is
only about 90 ppm compared to 160 ppm from Highvale coal which has less
nitrogen (0.71%). The low N0x emission for the anthracite is probably related
to its low volatile content (6.5%, compared to 31% for Highvale). A simple
correlation which works well for the various fuels investigated in this work
and for the data of Hirama et al. (8), all in the absence of limestone, is:

N0x (ppm) - 50 VN°-3 (1)

where is the product of the volatile matter from the proximate analysis and
the nitrogen content from ultimate analysis of same fuel (NX x Volatile^).

This correlation is only applicable for the conditions used in the two
combustors which produced the data, i.e. for temperature 850-900°C, gas
velocity 6-10 m/s, secondary to primary ratio 0-2.0, 15-30% excess air, abrupt
exit geometry and no addition of limestone. As shown in Fig. 6, Eqn (1) gives
a good fit for the wide range of nitrogen content (0.10-0.95%) and volatile
matter (6-611) covered by the fuels tested.

NO Concentration Profiles
x

To assist in the understanding of the mechanisms of NO formation and

destruction, gas samples were withdrawn at different levels of the riser. With
the system held at steady conditions, samples from different heights and
different radial positions could then be analyzed, one after another.

Preliminary results (see Fig. 7) show that the basic pattern of axial profiles
with samples taken at the wall is similar to those in bubbling beds, e.g. (7).
At the bottom of the riser (below the coal feed port), the N0^ concentration is
very low. However, a fast build-up of NO^ was found above the secondary air

6B-98

-------
entry. A maximum concentration occurred between 15 and 30% of the way up the

column, with the exact position depending on the operating conditions. NO^

levels then decrease continuously throughout the upper part of the riser. This

pattern can be explained by the release of fuel nitrogen and subsequent

reactions which may either enhance or abate the NO^ level. At the bottom, a

turbulent dense bed exists. Fuels fed into the riser need time to

devolatilize. Because of high gas velocity, typically 4-5 m/s below and 8-10

m/s above the secondary air injection port, the volatiles are released at

greater heights. Also, due to the reducing conditions at the bottom of the

riser, there is limited oxygen for N0^ to form. The catalytic effect of

limestone under reducing conditions tends to suppress NO^ formation (13).

Above the secondary air entry, the local oxygen concentration can be as high as

15%, and oxidation of fuel N can be catalyzed by the presence of limestone.

After a maximum NO concentration is reached, reduction of NO by char causes

x x

the NO^ concentration to decrease substantially in the wall region, as shown in

Fig. 7. Increasing the secondary-to-primary air ratio gives a higher N0x

concentration above the secondary air entry. However, N0^ levels converge and

are essentially identical by the top of the column.

Because of different hydrodynamics between bubbling and circulating fluidized

beds, the spatial concentration distribution of gas components may be very

different. The annulus-core flow pattern, in circulating fluidized beds may be

expected to cause radial gradients. Preliminary results from a recent

combustion test of Minto coal, in which NO concentrations at three radial

positions were measured, are shown in Fig. 8. The three radial position are

the axis of the riser, half-way between the axis and the wall and right at the

wall itself. Similar to the earlier findings shown in Fig. 7, the

concentration of N0x dropped rapidly in the upper part of the reactor for

samples taken at the wall. However,, the concentration of HO^ was found to be

almost uniform in the centre of the column. It is clear that local suspension

density (which determines carbon loading) affects strongly the profiles of XO,

as well as of other gaseous components. With high suspension density near the

wall, more carbon can react with N0^ and there is less oxygen present. On the

other hand, in the dilute core of the riser, the reduction of NO was less

effective. It is clear that the suspension density distribution has a direct

influence on NO formation and reduction in circulating fluidized bed
x

combustion. More comprehensive analysis of these results is under way.
CONCLUSIONS

A variety of fuels, including woodwastes, coals and pitches, have been burned
in a pilot scale circulating fluidized bed combustor. N0^ emissions were in
the range 80-220 ppm. Among the operating variables, temperature, excess air
ratio and limestone feed rate were found to have strong effects on NO
emissions. No clear influence of the secondary-to-primary air ratio was

6B-99

-------
found for the conditions investigated. NO^ emissions can be estimated from a
simple correlation in which both fuel nitrogen content and volatile matter are
used.

N0x concentration profiles at different heights and radial positions were
measured. Closely related to local suspension densities and oxygen
concentrations, greater reduction of NO^ was found near the wall, where solid
concentration and carbon loading were higher than in the dilute core of the
reactor. It is clear that these radial differences need to be accounted for in
modelling of nitrogen reactions in circulating fluidized bed corabustors.

ACKNOWLEDGEMENT

The authors are grateful to Energy, Mines and Resources, Canada which provided
the funds for this study. Assistance during the experiments from H. Aekroyd,
H. Li, R. Senior and R. Wu is gratefully acknowledged.

6B-100

-------
REFERENCES

1. A. Kullendorff and S. Andersson. A General Review on Combustion in
Circulating Fluidized Beds, in Circulating Fluidized Bed Technology, ed.
P. Basu, Perganon, Toronto, 83-96 (1986).

2. H. Kobro and C. Brereton. Control and Fuel Flexibility of Circulating
Fluidized Beds, in Ciculating Fluidized Bed Technology, ed. P. Basu,
Pergamon, Toronto, 263-272 (1986).

3. L. Reh et al. CFB Combustion: An Efficient Technology for Energy Supply
and Environmental Protection, Inst, of Energy Symp. Ser. No. 4, VI-2(l-ll)
(1980).

4. B.M. Gibbs et al. Coal Combustion and NO Formation in an Experimental
Fluidized Bed, Inst, of Fuel Symp. Ser. No; 1, London, D6(l-13) (1975).

5. M. Horio et al. Research and Development of NO Emission Abatement in a

Fluidized Bed Coal Combustor in Japan, Proceedings of the Sixth Intl. Conf.
on Fluidized Bed Combustion, Atlanta, 968-978 (1980).

6.. T. Hirama et al. A Two-staged Fluidized Bed Coal Combustion for Effective
Reduction of NO Emission, in "Fluidization", eds. D. Kunii and R. Toei,
Proceedings of the Fourth Intl. Conf. on Fluidization, 443-450 (1984).

7. J.M. Beer et al. NO Formation and Reduction in Fluidized Bed Combustion of
Coal, Proceedings of the Sixth Intl. Conf. on Fluidized Bed Combustion,
Atlanta, 942-955 (1980).

8. T. Hirama et al. Nitric Oxide Emission from Circulating Fluidized-bed Coal
Combustion, Proceedings of the'Ninth Intl. Conf. on Fluidized Bed
Combustion, Boston, 898-906 (1987).

9. J.R. Grace et al. Circulating Fluidized Bed Reactor Design and Operation,
Sadhana, JJD, 35-48 (1987).

10. R. Legros et al. Combustion Characteristics of Different Fuels in a Pilot
Scale Circulating Fluidized Bed Ccnbustor, Paper to be presented at the
Tenth Intl. Conf. on Fluidized Bed Combustion, San Francisco, 1989.

11. H.A. Becker et al. Pilot Scale Trial on Atmospheric Fluidized Bed
Combustion of a British Columbia Anthracite, Technical Report to EMR,
Canada, QFBC. TR. 87. 4, 1987.

12. C. Brereton et al. Final Report on Contract 52ss.23440-7-9136, to EMR,
Canada, 1989.

6B-101

-------
B. Leckner and L.E. Amand. Emission from a Circulating and a Stationary
Fluidized Bed Boiler: A Comparison, Proceedings of the Ninth Intl. Conf.
on Fluidized Bed Combustion, Boston, 891-897 (1987)•

secondary air

Figure 1: Simplified schematic diagram of circulating fluidized bed combustion facility
(UBC) do),

1. Reactor; 2. Windox; 3. Primary cyclone; 4. Secondary cyclone; 5. Re-
cycle hopper; 6. Standpipe; 7. Eductor; 8. Secondary air preheater; 9. Flue
gas coolers; 10. Baghouse; 11. Induced draught fan; 12. Fuel hopper; 13.
Sorbent hopper; 14. Rotary values; 15. Secondary air ports; 16. Membrane
wall; 17. Pneumatic feed line; 18. External burner; 19. Ventilation; 20.
Calorimetric section

6B-102

-------
Fig. 2 Gas sampling system and vertical sampling positions;

1. Sampling port; 2. Gas sampling tubeextending into reactor;
3. Gas as filter; 4. Flexible stainless steel tubing;

5. Heat exchanger; 6- Drying tube; 7. Control panel and manifold
Note that only one of the five sampling trains is shown.

6B-1Q3

-------
Typical temperature profiles for various combustion trials (10),

SB-104

-------
X

O 150
z

100

1.1

12 1-3

Excess Air Ratio

Fig. 4 NO emission vs. excess air ratio:

.0 m/s, T - 870°C, Secondary/primary air ratio =1.0

900

|700i-

600

500

400

300 .2

*Si

mo 0*

100

Time(min.)

Fig. 5 NO emission and bed temperature vs. time after step decrease in
fuel feed rate:

6.0 m/s, Oxygen

9.5, Secondary/primary air ratio = 1-6

6B-105

-------
Fig. 6 N(>x emission vs. (V^ = ^nitrogen x ^volatile In fuel)

*From reference (8) with conditions similar to those in this study.
Taiheiyo coal, N = 0.90%, volatile = 43.7%

Datong coal, N ¦ 0.90%, volatile ¦ 31.0%

Height (m)

Fig. 7 Axial NO concentrations at the wall for two secondary to primary
o x

air ratios for lignite combustion:

U = 8.5 m/s, T = 870°C, oxygen in flue gas = 3.0%

6B-106

-------
300

Fig. 8

3 4 5
Height (m)

6 7

NO^ concentration profiles for Minto coal combustion,
U = 7.6 m/s, T - 890°C, Secondary- to primary air ratio ¦ 1.07,
Ca/S = 3 •01, oxygen in flue gas - 3.0%

Table 1; N0X Emission from Combustion of Differen

Superficial gas
velocity (n/s)

Secondary/primary
air ratio

Bed temperature (C)

Ca/S molar ratio

Sulphur capture (%)

Flue gas 02 (%)

NO emission (ppm)

Minto Coal

7.0 7.8

1.0 1.0

865 870

1.36 1.36

72 70

2.0 3.0

130 138

Hogfuel

7.1 7.6

2.2 2.6
830 825

3.0 4.5
89 90

CANMET Pitch
7.5 7.8

0.56 1.07

847 895
1.7 1.5

3.8 3.4
165 145

6B-1Q7

-------
Table 2: Fuel Analyses (Wt%)

Proximate Analysis

Volatile
Fixed Carbon
Ash

Moisture

Ultimate Analysis

Carbon
Hydrogen
Nitrogen
Sulphur

Oxygen (by diff.)
Ash

High Heating
Value (MJ/kg)

Minto Coal Hogfuel CANMET Pitch

32.9 60.9 67.7

46.8 23-9 30.2

18.8 4.0 2.0

1.5 11.2 0.1

64.8 52.8 86-2

4.2 5.4 7.1

0.7 0.1 1.1

7.2 0.2 2.8

4.1 36.9 1.0

19.0 4.6 1.8

34.3 22.3 40.0

Highvale B.C. Anthracite Lignite

31.0 6.6 35.6

41.2 72.3 38.9

12.3 16.5 9.2
15.5 4.6 16.3

52.1 72.5 51.4

3.1 2.2 3.2

0.7 0.9 0.6

0.2 0.5 0.6

16.1 2.9 18.6

27.8 21.1 25.5

20.3 27.1 19.9

-------
Table 3:N0X Emission from Combustion of Anthracite
' (31 Oxygen)

Superficial Gas Bed Temperature Sec./Prim. Mr N0X (ppm)

Velocity (m/s) (C)

7.7 872 1.0 80

7.8 875 0.5 80

7.9 884 2.0 97
6.0 874 2.0 86
6.2 882 1.0 90

Table

4: NO Emission from

Combustion of Kinto Coal

(3% Oxygen, sec./prim, air

ratio = 1)

Superficial Gas

Bed Temperature

Ca/S

Sulphur Capture

Velocity (m/s)

(C)

<%)

8.3

880

8.4

900

1.8

87.0

8.3

882

3.0

99.2

N0X (ppm)

127
160
218

6B-109

-------
Session 7A
POST COMBUSTION NOx CONTROL DEVELOPMENT

Chairman: C, Sedman, EPA

/-*

j
j

^ TA-i /

-------
Design and Operation of the SCR-Type
NOx-Reduction Plants at the Durnrohr
Power Station in Austria.

H. Novak

Verbundkraft Elektrizitatswerke Ges. m. b. H
Am Hof 6A
A-1011 Vienna, Austria

H. G. Rych

EVN Energie-Versorgung Niederosterreich AktiengeselIschaft
A-2344 Maria Enzersdorf, Austria

ABSTRACT

In 1986 the two hard coal and / or natural gas fired units of Durnrohr Power
Station in Austria, which are equipped with SCR-plants and with dry scrubbers
started industrial operation.

The units have a power generating capacity of 405 MW and 352 MW. Construction
started in 1981, in 1985 the decision was made to add full stream SCR-plants
with 80 % removal efficiency. Until the end of 1988 the units have operated with
SCR-plants 2254 and 9846 hours under intermediate load conditions.
The SCR-plants are of high dust type and they are equipped with plate type
catalysts. The catalysts are situated upstream the Ljungstrom air heaters and
the dry scrubbers. The engineering and the catalyst were purchased from a

7A-1

-------
Japanese manufacturer. Great care was taken to ensure the necessary load
following abilities and to ensure conservation of the catalyst during the
relatively long annual standstill periods.

Various performance tests were carried out, this paper reports about the results
of two guarantee tests and two periodical tests. In addition sample catalysts
were taken out of the SCR-reactors in order to study the quality changes of
catalyst during operation time. The results of three sample catalyst tests
indicate the good condition of the catalyst.

INTRODUCTION

Two Austrian Utilities, the Verbundkraft Elektrizitatswerke Ges.m.b.H. (VKG) and
the EVN Energie-Versorgung Niederosterreich Aktiengesellschaft jointly erected
the hard coal fired Power Station DOrnrohr between 1981 and 1986. The Power
Station is located about 40 km west of Vienna. Each company owns and operates
one of the two units.

An overall view of the power station in the final stage of construction is shown
in figure 1.The boilers are single pass tower boilers, about 90 m high and they
are equipped with low-NOx tangential burners, have a dry bottom and an empty
downward fluegas duct between the economizer outlet and the airheaters. Hard
coal from Poland with an average calorific value of 24.3 MJ/kg and natural gas
can be used as fuels, both single or mixed.

Fluegas desulfurisation is performed by dry scrubbers with electrostatic
precipitators in 3 parallel streams. To achieve better flexibility in endproduct
sale, the flyash is separated by prefiIters upstream the FGD.

At the hearings for the approval of construction of the power station in 1981
the authorities required the installation of NOx reduction facilities as soon as
they were state of the art if ambient air conditions called for them.

After demonstration of the performance of the SCR-technology in Japanese power
plants, VKG and EVN decided to install such plants in both units of the Diirnrohr
Power Station in the beginning of 1985 after intensive preliminary studies. This
decision was made regardless of the low NOx concentrations in the environment of
the power station. The authorities required the SCR-plants to achieve a removal

7A-2

-------
efficiency of 80% at fluegas temperatures above 310° C arid the 8 hours average
of the outlet NOx concentration expressed as N02 to be lower than 200 mg/Nm3
(97.6 ppm). The NH3 slip was required to be below 5 ppm.

At the beginning of 1989 a new law came into force in Austria requiring by 1995
all existing power stations fired with coal must achieve NOx emissions lower
than 97,6 ppm and if fired with gas lower than 73,2 ppm. These values are
defined as 30 minutes average. The Dtirnrohr power station already fullfills all
these requirements today. - By the way, more than one third of the thermal
capacity installed in Austria, which is only 4000 MW, is already equipped with
SCR-plants.

Until January 1st, 1989 Unit 1 was operated 2254 hours and Unit 2 9846 hours
with the SCR-plants.

DESIGN AND LAYOUT OF THE SCR-PLANTS

The basic design parameters can be seen in table 1.

The ID-fans were already installed and would only be changable with significant
costs. Therefore it was necessary to limit the pressure loss of the SCR-system.
After intensive studies it was also decided not to change the airheater
plates, which were delivered already. In the tender specification, the suppliers
were asked to design the system for end of operating period flue gas
conditions. The safety margins gained this way in our opinion are an
important reason for the good system performance that was achieved. The contract
for the SCR-plants was awarded to a consortium consisting of Babcock Hitachi
Corporation (Japan) and Voest-Alpine (Austria). The plants are of High-Dust type
and they are equipped with Hitachi plate-type catalysts. The technical data can
be taken from table 2,

7 A-3

-------
Description of the SCR-System

Working temperatures of the SCR-system are between 310 and 420° C. Because there
was no space for a layout with reheaters downstream the FGD, the reactors were
installed above the existing airheaters in an altitude between 50 and 90 m above
plant zero. Due to the progress of construction the support structures of the
SCR-reactors were built in different ways for both units. At unit No.l, which
was almost finished, a separated support bridge" was built above the air heater
building while the housing of the airheaters of unit No.2 could be modified to
support the reactor. Figure 2 shows this arrangement.

The horizontal inlet duct of the SCR-reactor was connected to the bend of the
empty downward duct of the boiler. After a damper the ammonia injection grid and
a mixing grid are located in the inlet duct. The flue gas goes through the
reactor from top to bottom leaving at the outlet damper. The outlet duct is
connected to the flue gas duct above the air heaters. The flue gas duct and
SCR-reactor arrangement was modelled by Hitachi in a scale 1:20 to achieve a
good fluegas velocity distribution and low pressure losses.

Since the boilers and flue gas ducts are hanging free in the boiler house and
the SCR-reactors belong to a different static system, displacements of up to 180
mm may occur and the compensation and suspension system had to be sophisticated
to keep low forces in the connections .

Most parts of the SCR-reactors are housed, except for the inlet damper which in
due course gave trouble with the limit switches in winter.

Ammonia supply

Ammonia is supplied to the SCR-reactors from a common storage and vaporizing
system with a pressure of about 1.9 bars and is diluted with 9000 m3/h of air
extracted from the boiler house. The dilution air flow is constant over load.

The ammonia injection into the flue gas is carried out via 20 separately
adjustable nozzle systems into a flue gas duct of 12 x 5 m. The total number of
nozzles for each plant is 370. The most important part of initial startup was
the proper balancing of the NH3 distribution to achieve a uniform NOx
distribution at the outlet of the reactor.

7A-4

-------
Catalyst

The catalyst Is arranged in 2 layers of about 2.0 and 1.5 ni height in the
reactors. The reactors are designed to hold 140 % of the initial catalyst
volume. Before the exchange of layers additional catalyst will be loaded to
compensate the decrease of activity during operation.

The catalyst plates consist of a plasma coated steel grid with the ceramic
catalyst material applied to it. The plates are corrugated and assembled in
boxes like air heater elements. The boxes are put together in blocks weighing
about 1 or 2 tons. This block is the unit for handling, its size is about 1 x
2.3 x 1.3 m or 1 x 2.3 x 0.7 m. A system of hoists has been constructed for
transportation on site. Figure 3 shows catalyst blocks during the initial
loading of the reactors.

Above, between and below the catalyst layers the installation of soot blowers is
possible. So far the soot blowers have not been necessary, it seems they will
not be needed at all, because significant ash deposits on the surface of the
catalysts, erosion of the inlet edges of the plates or clogging have not been
observed. Only on some of the sample catalyst brackets, which are located about
1 to 2 meters above the layers" and protrude about 1 meter into the reactor, dust
deposits have been found, that clog the channels between the catalyst sample
plates.

Bypass

Into the existing bypass duct of the boilers a damper was installed between the
inlet and outlet ducts of the SCR-reactor. This damper is closed during
operation, but during startup and shut down operation it allows operation of the
boiler without gas flowing through the reactor. At startup the fluegas is led
through the reactors when its temperature exceeds 100 C. NH3 is injected above
300 C metal temperature into the reactor. The heating of the reactor does not
elongate boiler startup operation, but because of the 300 C limit for NH3
injection, the first 2 to 5 hours of boiler operation there is no NOx removal by
the SCR-system. The bypass helps to shorten the startup time of the plant, but
it does not seem to be necessary to install a bypass if designing a new plant.

7A-5

-------
Catalyst Conservation System

Since alkaline flyash traces on the surface of the plates can react with the
catalyst material in a wet environment, a dry conservation and heating system
was installed. As Austria is primarily a water power country (about 70 % of the
annual electric power production is made from hydro power stations), the thermal
power stations have almost no operation in Summer. During these long standstill
periods the air in the reactors is kept below 55 % relative humidity. The dryer
system extracts air from the top of the reactor and injects dried air at the
bottom. This system is equipped with a steam heater and is also used to heat up
the reactor to 50° C before startup of the plant. As startup fuel is natural
gas, this measure serves as a protection of the catalyst against condensation of
vapour on the catalyst surfaces. The dryer system has given some trouble,
because the dust in the reactor adheres to the dryer wheels and a high
efficiency bag filter has to be installed to ensure a sufficient lifetime of the
dryer wheels. The maintenance of the humidity control system and the hygrostatic
switches is a labour intensive job because the switches used in air condition
plants are affected by dust and humidity below 20 % which sometimes occurs.
During standstill temperature and humidity in the reactors are continuously
monitored.

Control System

The simplified line diagram of the control system of the DeNOx plants is shown
in figure 4. The NOx concentration in the flue gas is measured before the
reactor inlet and with a signal for the flue gas volume (calculated from the
amounts of fresh air and fuel) the amount of NOx is calculated. Based on the
given NH3/N0x molar ratio the NH3 injection is adjusted. In addition, due to the
changing NOx concentration in the flue gas, a feedback signal from the NOx
outlet concentration is superimposed which serves to adjust the molar ratio
value in case of deviation from the desired removal efficiency.

Step functions and limiting functions make sure that even in case of transient
conditions, for example mill failures, the injected NH3 amount does not increase
too much. It was decided to permit lower removal efficiencies for short periods
during transients to avoid breakthrough of NH3.

7A-6

-------
So far integration of the ammonia slip measurements in the control systems has
not appeared to be possible or useful.

In the beginning, due to the buffer behaviour of the catalyst, adjustment of the
control system was difficult. Finally it was possible to reach satisfactory
results inspite of the system's high inertia. However, deviations from the set
removal efficiency of 80% and outlet NOx concentration variations, for example
from 110 to 240 mg/Nm3 - based on 150 mg/Nm3 in undisturbed operation - cannot
be avoided, in case of a load increase from half to full load with simultaneous
start of one coal mill. Within approx, 1 1/2 hours the system settles for steady
values. If the control system is tuned more sensitive, it becomes instable.

Influence on Downstream Equipment

There was an intensive discussion about the deposition of sticky ammoniumsulfate
compounds on the air heater plates, but so far none of these were found. This
might be due to the low slip ammonia after the SCR. An abnormal increase of air
heater pressure loss did not happen either, only in two cases sulfatic compounds
on the cold end of the air heaters made washing necessary. These were caused by
flyash and S03 condensation due to very low inlet air temperatures.

The S03 in the fluegas is removed completely in the dry scrubber FGD, the air
heater outlet temperature is about 130 C, so far there were no signs of
corrosion in the flue gas ducts.

The fresh flyash, collected by the ESP before the FGD, contains less than 50 ppm
of ammonia and. is sold to the cement industry without problems.

The performance of the dry scrubbers was not influenced in any way.

7A-7

-------
PERFORMANCE TESTS

Introduction

In the contract with the supplier of the DeNOx-plants Voest Alpine/BHK two
guarantee tests beside the Taking Over tests are included. One test has to be in
the middle (approx. after 6000 hours of op.) and the other at the end of the
guarantee period of 12 000 hours of operation.

Beside the guarantee tests, periodical tests are carried out to check the
condition of DeNOx-plants periodically. At periodical tests a reduced number of
grid points are measured, compared to the guarantee tests. Periodical tests will
be carried out yearly at the end of each operation period after guarantee time
is over.

At the moment they give additional informations during guarantee time.

In addition to this, sample catalysts are taken out of the SCR-reactors in order
to study the quality changes of catalyst during operation time. At this moment
we can present the results of three sample catalyst tests which indicate the
good condition of the catalyst.

The following part reports about the test results of Unit 2, because of the
longer time of operation. The situation at unit l is similar.

Guarantee- and Periodical Tests

The time schedule in table 3 shows the dates and the related operational hours
of the DeNOx-plant at the Guarantee- and Periodical Tests.

At the Guarantee- and Periodical Tests the grid measurements had been conducted
at the inlet and outlet duct of the SCR-reactors. The positions of the related
cross sections are to be seen at figure 2.

There are 12 to 16 sampling nozzles at the side wall of the inlet and outlet
duct of the SCR-plants, depending on the dimensions of each cross section. In
order to get the actual grid of measurement points, 4 grid points at each
sampling nozzle have to be chosen according to the Austrian standards (principle
of area centre of gravity of rectangles equal in area).

7A-8

-------
For a well operating DeNQx-plant it is fundamentally necessary to have a good
adjustment of the NH3 injection. The distribution of the NH3 injected at the
inlet duct of the SCR-reactor has to be matched to the NOx and velocity
distribution at the cross section of injection.

Before adjusting the trim valves it is necessary to know the flue gas condition
at SCR-inlet and -outlet. Therefore the velocity-, temperature- and NOx
distribution is measured first. Based on these measurement results the trim
valves have to be adjusted until a uniform NOx distribution on the outlet of the
reactor is obtained. After each readjustment of injection valves NOx has to be
measured at selected grid points at the SCR-outlet.

In order to utilize fully the SCR-catalyst and the NH3, it is necessary to care
about the distribution of NH3 injection. Keeping in mind the demands of local
authorities and the necessity to protect the downstream components of the power
station, the NH3 slip distribution at the SCR-outlet has to be examined. In our
case the guarantee values, concerning NH3 slip, are less than 5 ppm for the
average NH3 value and. less than 6 ppm for each measuring point.

In order to calculate the S02/S03-conversion rate, S02 and S03 have to be
measured at selected points at SCR-inlet and -outlet.

Up to now the Taking Over Test in February 1987 and the 1st Guarantee Test in
April 1988 have been carried out. A detailed measuring programme was done
because of the importance of these tests.

Table 4 demonstrates the measured flue gas components, the number of sampling
points and the used measurement tecbniquesat Guarantee- and Periodical tests.

Table 5 gives a survey of the guarantee values out of contract and the related
measuring results of the.guarantee tests. The measuring results are corrected or
related to the design values of flue gas volume, flue gas temperature, NOx
concentration (inlet), 02-content, and mole ratio NH3/N0x. To check the
guarantee values the average values out of related grid measuring results are
used. It is obvious that there was no problem to keep the guarantee values and
there is no significant change in the DeNQx-plant condition, comparing the
results of Taking Over Test and of 1st Guarantee Test.

7A-9

-------
The distributions of relevant flue gas components at the corresponding cross
sections of flue gas ducts are an additional important criterion to estimate the
condition of a DeNOx-plant.

Figure 5 explains the NOx- and flue gas velocity distribution at the SCR-inlet
of the 1st Guarantee Test. It is remarkable to see the uniform NOx distribution
beside the relative uneven distribution of flue gas velocity at the NH3
injection point.

Figure 6 compares the NOx- and NH3 distribution at the SCR-outlet of the 1st
Guarantee Test. The well designed and adjusted DeNOx-plant is characterized by
the relatively uniform distribution of NOx concentration and NH3 slip at the
SCR-outlet.

To give a further description of the distribution of flue gas components, the
relative standard deviations (standard deviations related to the relevant
average values) of measured flue gas components at SCR-outlet are given in table
6.

The quality of NOx distribution at the SCR-outlet can be judged by comparing the
relative standard deviations of the different tests. The NH3 slip distribution,
depending on NOx distribution and adjusted NH3 injection, can be estimated as
satisfactory knowing about the difficulties to influence it and the low absolute
values.

Sample Catalyst Tests

At one wall of the reactor there are sample catalyst units, each three above,
between and under the two SCR-catalyst layers.

Each unit contains 40 pieces of sample catalysts. The dimension of one piece is
100 x 100 mm.

At this time already three times sample catalysts were taken out of the
DeNOx-plant Unit 2. Each time two pieces out of each sample catalyst unit were
taken out. These 18 pieces of sample catalysts were sent to Japan for
investigation at BHK laboratories. Therefore there are test results of three
samplings available beside the analysis of original catalysts out of production.

In table 7 the time schedule of sampled catalysts as well as the related hours
of operation of the DeNOx-plant of Unit 2 are given. The test items and methods
of analysing sample catalysts are given in table 8.

7A-10

-------
There are two different ways to describe DeNOx-efficency out of sample catalyst
test results.

One way is to compare directly laboratory test results of DeNOx-efficiency of
different samplings without a direct connection to the actual DeNOx-efficiency
of the real plant. To get comparable efficiency data, constant analysis
parameters must be supposed.

The other way is to look at the relative activity which is based on the activity
of the original catalyst. This is theoretically a good way to describe changes
of catalyst activity with time. One problem is to determine the activity of
original catalyst. It is possible to relate to the average value of efficiency
out of the original catalyst test result during catalyst production with a
relative high reproducibility. The other way is to store a sufficient amount of
sample catalysts outside of the DeNOx-reactor in a suitable manner, in order to
determine efficiency together with sampled catalysts in laboratory at each
sampling. If the unchanged condition of the stored original catalyst samples
with time is given, the described method to compare efficiency changes of
catalyst with time is preferable because of having no influence of different
analysing parameters.

In the Diirnrohr case the relative activity based on the average efficiency of
original catalysts has to be calculated, because there are no stored original
catalyst samples available. As described above we get uncertain values for
relative activity because of the high reproducibility of the original catalysts.
In some cases the value 1.0 is exceeded.

Figure 7 shows the efficiency graphs based on the sample catalyst test results
at different times. The efficiency was detected at three different temperatures
of the test reactor. There are no significant efficiency drops to be seen at all
three temperatures.

At table 9 the additional test results beside the DeNOx-efficiency at the
different samplings are listed. At the given measuring results there are no
significant changes in the case of S02 conversion, impact- and erosion
resistance between the original catatalyst and the results of sampled catalysts.
In the case of specific surface area, porosity and content of sodium and
potassium there are some changes to be seen due to the influence of operation.
The found measuring results are absolutly not critical due to the Japanese
experience .

7A-11

-------
CONCLUSION

The SCR-plarrts of the Durnrohr Power Station have already operated for two
years. As it was described in this paper, the SCR-plants are in an unchanged
good condition. There is no drop in removal efficiency and no significant
increase in slip ammonia detectible, the surface of the catalysts is dust free
and no erosion could be found.

The system operates together with the other components of the power plant
without making trouble or affecting the downstream equipment. We hope, that it
will stay that way.

7A-12

-------
I Reproduced from
best available copy.

Figure 1. Overall View, of Durnrohr Power Station

7A-13

-------
Figure 2. Durnrohr Power Station. SCR-Reactor of Unit 2.

7A-14

-------
Figure 3. Catalyst Blocks of the SCR-Plant
Diirnrohr Power Station.

Figure 4. NH3 Control System of Diirnrohr Power Station.
Simplified Line Diagram.

7A-15

-------
NOx-concentration at DeNOx-inlet
1st Guarantee Test

NOx-conc. [mg/m3]

1N 2M 3L 4K 6J 61 7H 8(3 9F 101 11D 12 C 13B 14A

number of nozzle

Flue gas velocity at DeNOx-inlet
1st Guarantee Test

velocity Im/s]

1N 2M 3L 4K 5J 61 7H 83 9F 10E f1D 12C 131 14A

number of nozzle

Figure 5. First Guarantee Test of Unit 2
Durnrohr Power Station
NOx Concentration and Flue'Gas Velocity Profiles
Inlet Duct of SCR-Reactor

7A-16

-------
NOx-concentration at DeNOx-outlet
1st Guarantee Test

NOx-conc. [mg/m3]

5H 8Q 7F 86 9D

number of nozzle

10C 11B 12 A

NH3-slip at DeNOx-outlet
1st Guarantee Test

NH3-conc. Img/m3]

2K 3J

6H tO 7F 8E 90 10C 118 12A

number of nozzle

Figure 6. First Guarantee Test of Unit 2
Diirnrohr Power Station
NOx-and NH3 Concentration Profiles
Outlet Duct of SCR-Reactor

7A-17

-------
History i i

Dumrohr P/S unit No. 2

'.'too'"

so-;

¦ >.,'
o

r&y
'¦

•sts-t"
©

»C

¦: zr.
••®:
Q

40i'

reaction
temp. 0

original
catalyst

sampia
catalyst

if 0 0

3 5 0
3 0 0

NH,/NO 1 ?
AV Sinvh
Gas ComDosinon

NO 200(lfOT
SO, SOOpfW
0, 3"i
H,a 12%

¦ o >. . ,:aooO').'^.

^ operatinghour(hr)

,,beginntn§of .. " '
GyarantB*Op«rattarr

4Q0Q

5310

Figure 7. Sample Catalyst Test of Un.it 2
Durnrohr Power Station
DeNOx Efficiency of Catalyst Samples

7A-18

-------
Table 1

DESIGN PARAMETERS OF THE SCR-PLANTS
Durnrohr Power Station

Unit 1 Unit 2

VKG EVN

Rated power MW 405 352

Coal consumption t/h 140 120

Nat.gas consumption m3/h i.N. 92 000 79 000

flue gas flow m3/h i.N 1210 000 1038 000

flue gas temp.

after ECO C 395 405

S02 IN max mg/m3 i.N 2300 2300

NOx IN {as N02) mg/m3 i.N 800 650

particulates IN mg/m3 i.N 16 000 16 000

delta p SCR max mbar 7 7.5

Cold starts per year 20 20

7A-19

-------
Table 2

TECHNICAL DATA OF THE SCR-PLANTS
Diirnrohr Power Station

Unit 1 Unit 2

Removal efficiency % 80

NH3 slip max ppm 5

S02/S03 conv,rate

at 375° C max % 1.5

No. of catalyst elements

(0,45 x 0,45 x 0,5 m) 5200 4800

Weight of catalyst t 411 380

SV h-1 2270

LV m/S 4.9

7A-20

-------
Table 3

SCHEDULE OF TESTS FOR UNIT 2
Diirnrohr Power Station

operational hours
of DeNQx

2011

1st

Guarantee Test

April

1988

4280

2nd

Periodical Test

Nov.

1988

6678

Kind date

of test of test

Taking Over Test Feb. 1987

1st Periodical Test Oct. 1987

7A-21

-------
t

Table 4

MEASURING PROGRAMME
Guarantee- and Periodical tests
at inlet (1) and outlet (2) of the DeNQx- reactor
of Durnrohr Power Station, Unit 2

flue gas number of points method of

component Guarantee T. Period. T. analysis

(1) (2) (1) (2)

NOx 56 48 28 48 chemiluin. and NDIR

02 56 48 28 48 paramagnetic

NH3 48 24 by wet process

502 6 12 6 12 NDIR

503 6 12 6 12 by wet process,

Japanese method

velocity 56 48 - - Prandtl tube

temperature 56 48 - - thermoelectric couple

7A-22

-------
Table 5

GUARANTEE VALUES AND MEASURING RESULTS
Diirnrohr Power Station Unit 2, coal firing

guarantee
items

unit

guarantee
value

Taking Over

Test
Feb.1987

1st Guarantee
Test
April 1988

DeNOx-
efficiency

80 or
more

81.5

NH3 slip

ppm

ave. 5
or less

0.32

0.63

S02-

convers ion
at 375 °C

1.5 or
less

1.11

1.40

pressure
loss

robar

7.5

5.5

5.4

7A-23

-------
Table 6

NOx- AND NH3 DISTRIBUTION
Outlet Duct of SCR-Reactor
DOrnrohr Power Station Unit 2, coal firing

Flue gas 1 relative standard deviation in % at:

component |Taking Over 1st Period. 1st Guarantee 2nd Period.

SCR-outlet |Test Test Test Test

NOx conc. | 8.2 10.2 9.5 15.3

NH3 slip ! 35.0 - 24.4

Table 7

SCHEDULE OF CATALYST SAMPLING
Diirnrohr Power Station, Unit 2

Date of sampling hours of operation

=a=s===s==3=a=s=s=s=ss=3=s=ss=s=s=aasssss==s=ssss2=,:s,=;.

1st sampled catalysts June 1987 811

2nd sampled catalysts May 1988 5310

3rd sampled catalysts Sept. 1988 5310

7A-24

-------
Table 8
SAMPLE CATALYST TESTS

item method

Catalyst activity A piece of catalyst is set in a reactor
(DeNOx efficiency) and the DeNOx-efficiency is measured,

using a mixed synthetic flue gas at
temperatures between 300 and 400 °C.

S02 conversion to S03 Three pieces of catalyst are set into a

reactor where a mixed gas passes across it
and S02 concentration is measured at inlet
and outlet of the reactor.

Specific surface area The surface, area of the catalyst is

measured with N2-adsorption method (BET
method).

Porosity

The pore volume of the catalyst is
measured with the method of filling
mercury into the pores by pressure.

Impact resistance

The weight difference of a piece of
catalyst is measured after it is dropped
ten times from 1 m height to a steel plate.

Analysis of accumu-
lated material

Analysis of alkaline metals (Na and K by
using AAS method).

Erosion resistance The weight difference of a piece of

catalyst is measured after 8kg of steel
grit are dropped under a slope of 45
degrees from 500 mm in height.

7A-25

-------
Table 9

RESULTS OF SAMPLE CATALYST TESTS
Durnrohr Power Station, Unit 2

Item original catalyst sampled catalysts

produced from Sept,85 1st 2nd 3rd

to March 86 June 87 May 88 Sept.88

SQ2 con-
version [%] av.:1.26 (0.62-2.20) 1.37 1.35 1.40

impact
resistance

[g/TP] av.:0.11 (0.05-0.35) 0.07 0.11 0.10

erosion
resistance

[g/TP] av.:0.13 (0.08-0.20) 0.17 0.16 0.14

specific
surface

area 1.0 0.929 0.921 0.911

porosity 1.0 1.036 1.184 1.114

sodium [%] 0.011 0.019 0.043 0.028

potassium [%] 0.002 0.011 0.051 0.023

7A-26

-------
DENOX CATALYTIC CONVERTERS FOR VARIOUS TYPES OF FURNACES AND FUELS
- DEVELOPMENT, TESTING, OPERATION -

L. Balling, D.. Hein
Siemens AG KWU Group
Main Department Combustion
Hammerbacherstr, 12 + 14
8520 Erlangen, FRG

ABSTRACT

Different types of furnaces, such as, dry bottom or wat-bottom furnaces, and
different fuels, such as gas, oil and coal, where a further distinction must
be made between hard coal, lignite, high-inerts coal and particularly sul-
phurous coal, call either for tailor-made catalytic converters or for a
special plant design. This paper deals with development, testing and operation
of honeycomb-type and plate-type catalytic converters in high-dust, low-dust
or tail gas configurations.

Siemens/KWU is the only manufacturer which produces both honeycomb-type and
plate-type catalytic converters and is thus able to look for the optimum tech-
nology for each individual application. In addition to catalytic converters
for dry-bottom furnaces, a catalytic converter which is particularly resistent
to poisoning by heavy metals was developed especially for wet-bottom furnaces.

This paper reports the most significant results obtained in laboratory tests
and power plant operation.

1. INTRODUCTION

To date, the selective catalytic reduction (SCR) process has consistently proven
itself for control of nitrogen oxides in power plant flue gases. Trouble-free
and economic operation of this system in various types of power plants with
various types of furnaces, fuels and operating conditions presupposes an opti-
mized plant fitted with catalytic converters matched to the individual plant.
To achieve maximum catalytic converter performance in the respective appli-
cation, it is furthermore imperative that the manufacturing process and cata-
lyst composition are tailored to plant requirements and, if necessary, suppor-
ted by plant systems engineering measures. This is only achievable via a syn-
thesis of knowledge of the catalyst behaviour which can only be provided in
detail by those who develop and manufacture catalysts, power plant mode of
operation experience (gathered by the utility and power plant contractor) and
the resultant influence on the DeNQx plant. Furthermore, an extensive base of
experience in process engineering is required, e.g. on those, measures for'
achieving uniform gas and flow distribution that are essential to optimize
DeNOx plant design and operation.

7A-27

-------
Power plant

/
1

— -¦

ES3

1 -

4 U in

Geometry

Type

Mech. behaviour

t.g,

e.fl.

e. a-

~ Honey comb type

~ HO-coal

~ Strength

a Plate type

~ ID-gas

~ Erosion

~ Geometrical surface

~ LD-afterFGD

(tail gas)

Modifications on catalyst

. _i

Design of SCR-plant

Plant specific measures

1 1

Type of firing

Fuel

Operation

e.g.

t.g.

~ Dry bottom boiler

~ Hard coal

~ intermediate load

~ Slag tap boiler

~ High inerls coal

operation

~ Cyclone tiring

~ an

~ Ash recirculation

~ High flue gas temp.

FIGURE 1: Design parameters for SCR-DeNOx-plants ¦

Figure 1 depicts these interconnections in simplified form; special power
plant requirements brought about by the type of furnace, fuel or mode of
operation employed must be met by adjusting catalyst composition and manufac-
turing techniques. Further catalyst modifications, e.g. to increase resistance
to poisoning, can increase the service life of catalysts, thereby substantially
improving cost-effectiveness.

The following discussion presents examples to illustrate the demands placed on
DeNOx catalytic converters and how problems can be effectively dealt with
through targeted experimental investigations.

2. SOLUTION ON PROBLEMS ILLUSTRATED BY EXAMPLES

DeNOx reactor design, as shown in simplified form in Figure 2, begins with cal-
culation of the catalyst volume V0 which would be required at the beginning of
operation from the characteristic data of the catalyst (e.g. the activity ks,
and operating conditions, which are defined primarily by flue gas volume and
composition. By allowing for the aging process, during which catalyst activity
drops from k0 to kt by the end of the guarantee period, Vt ideai the catalyst
volume required under ideal conditions can be calculated. Vt .jeai means that
the unavoidable deviations from the ideal associated with actual operation
have not yet been allowed for.

7A-28

-------
These factors may have diverse causes, as shown in summary form in Figure 3,
and must be allowed for by volume correction factors V in the design phase,
e.g.:

A: A non-uniform flow distribution, caused, for example, by unfavorable'duct
routing, reduces NOx conversion.

B: Combustion of high-inerts coal produces high fly ash content flue gas;
under such conditions, fly ash deposits and localized clogging are a
threat.

C: Wet-bottom furnaces with their high temperatures release increased concen-
trations of catalyst poisons and simultaneously alter the structure of the
ash particles such that fewer gaseous pollutants are bound to the ash.

This increases the threat of accelerated deactivation.

D: The overall service life of catalysts and, consequently, cost-effectiveness,
can be increased by plant design and, in particular by optimizing strate-
gies for loading extra catalysts and replacing spent catalysts.

Selected examples are referred to in this presentation to illustrate specific
problems, the solutions chosen and the results achieved.

7A-29

-------
— / 1

?) Distribution

~ of w, T, o, N0j(

D Fly ash

~ Deposition

~ Erosion

Catalytic material

~ Activity

~ Selectivity

~ Aging, poisoning

~ Geometry

~ Strength

~ Spare layer

~ Product control

FIGURE 3: Reasons for an increased demand of catalyst volume

2.1 Effects of Non-uniform Flow Distribution

If the design of an SCR reactor is based on one value for flue gas velocity,
flue gas temperature and stoichiometry, then it is equivalent to the ideal
reactor. In. reality, however, nonuniform velocity distributions occur, tempe-
rature striation is to be expected arid the NO concentration over the reactor
cross section may vary.

The deviations from uniform distribution always cause a reduction of NQx con-
version and must be compensated for by increasing catalyst volume to achieve
design performance.

Figure 4 shows a simplified reactor based on the assumption that three zones
of equal size are present, one with conditions precisely equal to the overall
average, a second with a positive deviation and a third with a corresponding
negative deviation of the respective parameter.

7A-30

-------
Required volume correction factors calculated on the basis of these assumptions
are plotted in Figure 4 .as a function of the magnitude of the deviation from'
the parameters flue gas temperature T, flow velocity w and stcichiometry
i.e. the ratio' NH3 to N0X at the reactor inlet. Design calculations perform
individual statistical superposition of these deviations together with NO dis-
tribution so that a total of = 81 partial reactors are allowed for. The
diagram clearly shows that the effect of temperature is relatively minor, and
that even oblique flow requires a significant increase in catalyst volume but
that large stoichiometry deviations cannot be accepted, owing to the magnitude
of the requisite volume correction factor. This means, for example that NO
distribution must be determined by field measurements during commissioning and
NH3 injection'adjusted accordingly.

The effects of non-uniform distribution are highly dependent on the degree
of NGX conversion. The diagram is based on a typical conversion rate of 85 %.
High conversion rates place stringent requirements on NH3 distribution. At a
conversion rate of 95 % and a stoichiometric deviation of + 6 %, for example,
NH3/N0x ratios > 1 occur. In such cases, local NH3 breakthrough would occur,
which could not even be prevented by vastly increasing the catalyst volume.

Due to their great economic importance, flow investigations are performed for
practically all plants. Figure 5 shows a test setup and test results. It can
be seen that the uniformity of velocity distribution can be significantly im-
proved by positioning guide baffles in the reactor entry hood and additional
flow straightener located upstream from, the first reactor layer. The findings
of distribution experiments on 1 : 15 scale models using air have proven fully
transferable.

7A-31

-------
FIGURE 5: Flow tests to optimize velocity distribution

The following point are to be kept in mmd when performing these design inves-
tigations:

- Simulation may not be restricted solely to the reactor, but rather must
include ductwork upstream from the reactor, since the causes of oblique
flow in the reactor are usually to be found here.

- Guide baffle arrangement must be such that no dust particle strictions
occur which could cause jet abrasion.

- Zero-flow and reverse-flow zones are to be avoided, since they lead to
buildup of fly ash deposits.

2.2 Prevention of Fly Ash Deposits and Erosion

.There is a close connection between flow patterns and the formation of ash
deposits on the catalyst surface. It has been debated whether a high fraction
of coarse fly ash, which allegedly has self-cleaning effect, or a much lower
concentration of fines causes partial clogging of the catalyst more rapidly.

Assumptions on the fraction of the catalyst cross section which will be
clogged - and therefore are no longer available for catalysis - by the end of
the service life must therefore be made during planning and design. This para-
meter is dependent on the structure and grain size spectrum of the fly ash and
ultimately also requires a design volume correction factor. It must also be
decided whether soot blowers are to be provided for cleaning the catalyst sur-
face and, if so, at what interval they are to be actuated.

7A-32

-------
Special investigations focussing on dust deposition were consequently performed

with the aim of

- clarifying the formation mechanisms of cornices (inlet side) and beards
(flow exit side),

- recognizing the parameters which affect buildup of these deposits,

- proof-testing design measures to prevent these deposits,

- proof-testing on-line cleaning equipment,

- determining the limits of the use of fine grid honeycomb catalysts with
regard to dust content and structure and

- investigating the differences in behaviour of honeycomb and plate-type
catalysts with regard to dust.

FIGURE 6: Deposition of fly ash

The test setup and typical results are shown In Figure 6. The most significant
insights gained in these investigations can be summarized as follows:

- Clogging of cells is caused primarily by portions of cornices and beards
which become detached.

- Coarse dust has a lower clogging tendency than fines (self-cleaning-effect).

- Rough catalyst surfaces are more conducive to build up deposits.

7A-33

-------
- The design is to minimize the size of stagnation zones, if necessary
prevent them by using peaked hoods or shrouds.

- Reverse flow zones are to be minimized by the use of curved surfaces,
shrouds and the like.

- Oblique inlet flow at angles > 15 0 are to be avoided due to the threat of
erosion.

- Erosion processes can be predicted and, for the conditions in an SCR plant,
can be controlled.

- The inherently flexible plate-type catalytic converter, as opposed to the
absolutely rigid honeycomb element, is less susceptible to dust and at
least must not be cleaned as frequently.

The results of these investigations provide guidelines which can be used for

making a decision on whether to install plate-type catalytic converters or

which honeycomb type can be selected without undue risk.

2.3 Deactivation by Poisoning

Catalysts age, that is, their activity drops during operation as a result of
irreversible processes in the catalyst. Usually high deactivation rates were,
however, measured in pilot plants in the high-dust region downstream from wet-
bottom furnaces. Investigations have shown that arsenic oxide is the cause.

SCR AH

ESP

FGD

Coal

Ash recirculation

FIGURE 7: Arsenic mass flow at a slag tap boiler with ash recirculation

7A-34

-------
By extensive measurements in power plants and balance calculations of cata- . .
lyst poison, we now know, that rapid deactivation is caused by arsenic oxide
buildup during ash recirculation, as shown in Figure 7. Whereas the catalysts
in a dry-bottom furnace plant are only exposed to the concentration contained
in the fuel, wet bottom furnace ash recirculation and melting of fly ash in
the combustion chamber increases the concentration to levels up to fifteen
times as high. It is the gaseous fraction of arsenic which destroys the cata-
lyst, and not arsenic bound to the ash.

The'solution to this problem may lie in any of the following;

- Preventing enrichment, e.g. by complete or partial removal, particularly of
highly laden fines

- Adsorption of gaseous AS2O3 on fly ash or an additive upstream from the
DeNOx reactor or

- Development of a catalyst resistant to arsenic.

Measurements of arsenic concentration in the individual sections of an electro-
static precipitator have shown that about 30 % of the total of arsenic agglo-
merated onto fly ash was precipitated bound to the fines fraction of fly ash,
which accounts for only 18 % of the total amount of dust, in the final section
of the electrostatic precipitatior. By partical extraction of only 18 % of the
fly ash, arsenic concentration can be reduced by a factor of 2 - 3. This
potential solution was not, however, pursued further, since the question of
what to do with these fines is still, open, rendering this concept unattractive.

Concentration "C" of
gaseous As

11000
Mfl/m3
750

500

250

Extraction
of ash

Metering
of

limestone

Relative activity

logt

Operating hours

FIGURE 8: Effect of arsenic on the deactiviation behaviour
of a special atalyst for slag tap boilers

7A-35

-------
Binding gaseous arsenic to a suitable additive supplied with the coal proved
useful for two reasons. It is a well-known fact that when the fraction of crys-
talline substances contained in the ash is high, only very low concentrations
of gaseous arsenic will be measured in the flue gas, even if the arsenic con-
tent of the coal was high. In a wet-bottom furnace plant where limestone was
added to the coal to enhance slag flowability, it was also observed that no
unusually high catalyst deactivation occured. For this reason, systematic
measurements have been performed to investigate the effects of limestone dosing
to bind arsenic. Results are given in Figure 8. It can be clearly seen that in
this special case, admixture of limestone to coal reduces the concentration of
flue gas arsenic from 1000 ug/m3 to less that 100 ug/m®. The"reduction of
arsenic deposited on catalyst specimens is particularly striking the effective-
ness of this measure.

In spite of these positive results, we do not consider this measure alone
sufficient, since the catalyst could suffer serious damage in a matter of
hours in the event that limestone dosing is not available. Activity losses of
50 % were measured in a pilot plant in just 300 hours.

For this reason it seemed imperative to develop a catalyst which

- adsorbes less arsenic, i.e. has a lower affinity for arsenic than'
conventional dry-bottom furnace catalysts

- is less susceptible to arsenic poisoning.

lime-
stone

Coal

Ash recirculation

Eilrastion

ofash

Relative As-content
in flue gas
In

0.5-

(?)

r~1 Slag tap i»Hef
ash recirculation

MM Extraction ol ash

i 1 Metering of limestone

1////A Dry bottom boiler

8 tS
~ Operating time

Catalyst r<* dry

fcDttom boiler

Mcdlfledcatalyst

FIGURE 9: Effect of extraction of ash metering on lime stone

7A-36

-------
The success of this development, which combines both measures, is illustrated
in Figure 9. This newly developed catalyst adsorbs only a limited amount of
arsenic up to a saturation level which means that poison-induced loss of
activity drops to a plateau. Normal aging occurs independent of this process,
and the effects are superposed.

Reliability and economy can be maximized by using this newly-developed catalyst
in conjunction with limestone dosing even when coal arsenic content, ash-
content and ash structure are unfavorable.

2.4 Effects of Design and Replacement Strategy Optimization

If the volume correction factors were minimized by the above mentioned
measures, economy can still be significantly influenced by optimized
strategies for addition of extra catalyst material and replacement of spent
catalyst. This is illustrated by the reactor in Figure 10 which is fitted with
catalyst elements in three layers. At time t-j (guarantee period) the total
activity of the reactor has dropped such that one layer must be replaced. The
layer of fresh catalyst material increases the activity potential, which then
decreases due to aging until a second layer must be replaced, and so on. In
this example, over a period of 100.000 hours a total catalyst volume of 1600 mJ
is required,, equivalent to four times the initial volume.

Activity

Specific
catalyst usage
per 1000 h:

16m3

20 - 40 60 h*10 100
—» Operating time •

Activity

120

20 40 60 hx1Q 100
—» Operating time

25%

Saving

12 m3

FIGURE 10: Improvement of catalyst Usage by strategy of change

7A-37

-------
Catalyst requirements are significantly more favourable when an additional
reserve layer is provided in the reactor. An extra layer of catalyst material
is loaded at the end of the guarantee period, which permits utilization of the
residual activity of the three initial layers. Catalyst, replacement is not
necessary until time t£. Over the same observation period, a catalyst volume
of only 1200 m3 is required, i.e. only three times the original volume.

After addition of the extra catalyst layer, reduced catalyst requirements are,
however, partially offset by increased auxiliary power needs due to the
greater pressure drop of flue gas flowing through four catalyst layers instead
of three. This presupposes that the induced-draught fan system is designed to
perform at this higher level. This example also demonstrates how inclusion of
all these considerations during the planning phase can increase the catalyst
usage factor.

Regular product monitoring employing catalyst specimens installed in the DeNQx
reactor and laboratory measurements can be used to determine the exact times
for addition of the extra catalyst layer or replacement in advance. This per-
mits coordination with scheduled power plant maintenance activities.

3. EXPERIENCE GAINED IN COMMERCIAL POWER PLANTS

Experience gained at two commercial power plants is described below to illus-
trate to what extent the plant and catalyst-specific measures developed on the
basis of the above design investigations have proven themselves under extreme
operating conditions:

- Power plant A, fired with high-inerts coal in which fly ash concentrations
as high as 50 g/m3 .occur, was fitted with catalytic converters on the high-
dust side.

- Power plant B, where low-volatile coal requires firing at high temperatures,
exhibits extremely high NQX concentrations ranging up to 2,000 mg/m3, there-
fore requiring a removal efficiency of about 95 %, was fitted with catalytic
converters on the tail gas side, i.e. downstream from the FGD plant.

In power plant A, with its extremely high fly ash content - up to 60 tonnes of
dust pass through the catalytic converter per hour - neither large deposits
nor plugging nor erosion occured to date. A plate-type catalytic converter
with increased pitch - and therefore greater flexibility - was installed.

Module and reactor structure were designed with a minimum of surfaces which
restrict flow, thereby minimizing the threat of cornice formation. Particular
attention was also devoted to the prevention of transverse flow patterns.

These plant design measures, and the selection of a highly erosion-resistant
modified binder for the catalyst material reduce the threat of erosion.

Initial inspections of the reactor after 1000 and 2500 hours of service
detected no serious ash deposits. As with laboratory investigations, slight
deposits in edge zones were found, however these are unavoidable and are of
negligible significance. Erosion damage has not occured to date. The very
positive results of commissioning measurements confirm the very favorable
chemical and mechanical properties of the plate-type catalytic converter
selected and the correctness of structural design measures implemented.

Power plant B ist characterized by an extremely high removal efficiency
requirement of about 95 %. This efficiency, as explained in Section 2.1, can
only be achieved without NH3 break through when the NH3/N0X molar ratio is

7A-38

-------
matched with extreme accuracy. For this reason permanent NO' sampling grids
were installed both upstream and downstream from the catalytic converter to
permit appropriate NH3 dosing even when NO concentrations fluctuate over a
broad range.

Honeycomb-type catalyst having a high specific surface area of 751 mz/m3 with
a pitch of 4,1 mm were installed. Catalyst activity was increased for example
by adapting the pore structure to these particular operating conditions using
a modified kneeding process (see Figure 2). Cell wall thickness was decreased
to reduce SO2-SO3 conversion and pressure drop. Flow straighteners made of
active catalyst material were installed above the first catalyst layer to
enhance the uniformity of flue gas flow.

Commissioning test measurements showed that removal efficiency not only met
but even exceeded the extremely high level required in a 14 x 20 m reactor
cross section without NH3 breakthrough. Measurements made downstream from the
reactor and in the stack have not detected any NH3 slip to date.

Even- such extreme operating,conditions, as demonstrated in the above plants,
pose no problems which cannot be solved by targeted research and development
focusing on matching the catalytic converter to the power plant.

4. CONCLUSIONS

When retrofitting power plants, the spectrum of requirements is by nature
particularly broad, since the catalytic converter design must be' tailored to
existing plant conditions. A survey of the spectrum of application conditions
is given in the list of references in Figure 11. References include nearly
equal numbers of high-dust and low-dust plants. High-dust plant catalytic con-
verters are subjected to flue gas temperatures as high as 430 °C. Wet-bottom
furnace power plants operate with full ash recirculation and gaseous arsenic
concentrations in the flue gas lie between 50 and 1000 pm/m3. A catalytic
converter installed downstream from a FGD plant which uses active char, however,
is exposed to extremely clean flue ga-s.

Faced with this broad spectrum of requirements, Siemens AG KWU Group opted for
manufacture of both types of catalytic converter, to be able to offer the best
possible solution, both from the technical and economic standpoint:

- honeycomb-type catalytic converters with a high specific surface area,
primarily for low-dust fluegas applications,

- plate-type catalytic converters with a lower suspectibi1ity to deposits
and erosion, mainly for applications with high-dust flue-gas.

The in-hcuse base of experience available to the Siemens KWU Group - gained in
the construction of power plants and design and development of catalytic con-
verters along with the associated instrumentation - is needed to be in the
position to offer complete package solutions, from fabrication to recycling,
for the most diverse applications.

7A-39

-------
Type

Number of

Power

Operating conditions

Catalyst type

applicat.

[MWe]

High dust

1450

High grade coal

Honey comb

(after boiler)

622

High Inerts coal

Plate

1560

High operating

Plate

temperature

1220

Slag tap boiler

Plate

Low dust

3005

After wet type FGD

Honeycomb

(after ESP

588

After spray absorption

Honeycomb

or FGD)

700

After active char FGD

Honeycomb

Refuse

Test facilities

Honey comb/

incinerator

Plate

industrial

After glass furnace

Honey comb/

application

(mobile test facilities)

Plate

FIGURE 11: Siemens-catalysts: references, operating conditions

7A-40

-------
ASSESSMENT OF JAPANESE SCR TECHNOLOGY FOR OIL-FIRED BOILERS AND ITS

APPLICABILITY IN THE U.S.A.

Dr. Phillip A. Lowe, PE
INTECH Inc.

Potomac, MD 20854

William Ellison, PE
Ellison Con su11 a n t s
Monrovia, MD 2177 0

Les Radak
Southern California Edison Co.
Rosemead, CA 91770

ABSTRACT

The Japanese have applied the Selective Catalytic Reduction (SCR) NOx
(nitrogen oxides) control technology to 74 oil and gas-fired electric
utility boilers. These applications have been made with specific
limitations and restrictions imposed through siting, design criteria,
and plant operating restrictions. Details of the Japanese experience
were assembled through field visits and an updated literature survey.
The results and conclusions developed during that study are presented
in this paper* In addition, the relevance of this technology to meeting
potential NOx emission objectives in the U.S.A. for oil-firing are
discussed. .

EXECUTIVE SUMMARY

It has been identified that in Japan the primary means for achieving.NOx
control for oil-fired utility boilers is by applying combustion
modifications and by firing low. nitrogen content fuels. Two thirds to
three fourths of the obtained NOx control is achieved with these
approaches. SCR is used as a.final "polishing" step to remove up to an
additional 70-105 ppm of NOx from--the flue gas, when that is necessary
to meet site-permitted conditions. SCR is thus a control designed to
remove 30% or less of the uncontrolled NOx levels from gas and oil-fired
utility boilers. Siting, design, and plant operating restrictions or
limitations are used in order to assure the continued operation of the
SCR system and a high availability of , the utility plant. Such
restrictions are imposed, even if they- subsequently limit the NOx
mitigation capacity of the installed SCR system.

Key Japanese design considerations include designing the SCR system for
a maximum of 80% removal of the NOx entering the SCR reactor (but also
planning to operate to achieve only 60-80% NOx reduction rates), and

7A-41

-------
basing"the SCR catalyst design on a titanium oxide formulation. Also,
the design provisions included allowing for the formation and deposition
of moderate amounts of ammonium bisulfates, and including in the design
the ability to conduct soot blowing or water washing to control those
deposits•

The siting philosophy appears to be that the SCR system should be
designed to fit the reasonably available space. This has resulted in
the retrofit of 10-30% NOx removal system at existing plants.

Operational restriction are perhaps the most important considerations.
The Japanese operated their SCR systems at reduced NOx removal rates
(typically 60-801) by restricting the amount of ammonia injected. This
appears to be done with the full concurrence of the regulators. This
approach further assures that operational problems and instrumentation
inaccuracies will not restrict the plant power capacity or availability.

In terms of the application of the technology to U.S. utility boiler
operations, important questions as to the impact of ammonia
contamination of the flue gas desulfurization waste waters and equipment
wash waters, and contamination of the boiler fly ash, and the selection
of monitoring and control instrumentation averaging times need to be
carefully considered because the Japanese experience indicates that
these are unresolved problem areas. The Japanese have partially solved
these problems by significantly restricting the amount of ammonia that
is used in operations, and by not requiring the SCR system to be
operated during low power, startup, and shutdown periods. Also, they
specify a fixed NOx emission limit, as contrasted to a percent NOx
reduction. At part load when the plant produces less NOx,, this means
that the SCR NOx removal rate can also be reduced.

INTRODUCTION AND BACKGROUND

"The NOx from- oil or gas-fired utility boilers is approximately 95%
nitric oxide {NO) and 5% nitrogen dioxide (NO2). The flue gas typically
contains 2-6% oxygen (Oz), depending upon the firing conditions. It is
known that a number of gases, such as hydrogen, carbon monoxide, and
methane can reduce the NOx to elemental nitrogen (Nz). The Japanese
developed the process using ammonia (NH3S because it selectively reacts
with the NOx while the other gases also readily react with oxygen. The
governing equations for the ammonia-based technology are:

4NH3 + 4N0 + O2 = 4N2 + 6HsO (1)

4NH3 + 2NOz + Oz = 3N2 + 6H2O (2)

Equation (1) has been verified in boiler operations, equation (2) is
still a hypothesis. Also, these reactions require at least 1% O2 to be
present. The optimum temperature for these reactions is about 1000°C.
Above this temperature a considerable portion of the ammonia is
converted to NO, below this temperature the reaction rate is too slow.
With a suitable catalyst present the reaction takes place at 300-400°C,
a temperature range that is available in the flue gas system of a
utility power plant.

If the reactions took place under ideal conditions, one mole of ammonia
would be required to remove one mole of nitric oxide. This is

7A-42

-------
stoichiometric chemistry. However, there are competing reactions and
some of the ammonia is oxidized to water and NOx; and maldistributions
of the NOx, oxygen, and ammonia mean that these components always are
not together in the correct amounts to stoichiometrically react as they
pass over the catalyst. Thus, in practice, one mole of ammonia reduces
0.80-0.97 moles of NOx.

Between 1977 and January, 1988, SCR systems were installed at 11
Japanese electric utility boilers fired with LNG and at 63 boilers fired
with oil*2 > . The aggregate capacity represented by these installations
is approximately 28,600 MW. A more meaningful measure of the impact of
the SCR technology is to account' for the installed NOx removal
efficiency and the NOx concentrations that are actually processed by the
SCR systems. In such a case, the aggregate capacity represented by
those installations is equivalent to the SCR removing about 76% of the
NOx from 100% of the flue gas from about 8,200 MW. Many of the retrofit
applications were designed for 10, 30, or 40% NOx reduction rates,
whereas applications at new plants are usually designed for 75-80% NOx
reduction rates.

It is also important to realize that the specified % NOx reduction is
for an "as new" plant. As the system ages during service the catalyst
becomes less reactive, causing a reduction in the NOx removal rate or
requiring a corresponding increase in the application of ammonia in
order to maintain the NOx removal rate (with a corresponding increase
in the inherent risks associated with not all of the ammonia reacting
to reduce the NOx).

The Japanese philosophy* *-• 2 >4 •11 >13 >17 > for implementing NOx control at
electric utility power plants is based upon applying extensive
combustion modification technologies and using fuels with low nitrogen
content. This accounts for removing 200-300 ppm of NOx and assures that
the NOx level that exits the boiler is in the 100-140 ppm (parts per
million by volume) range*2- 12> ,3> . The SCR technology is used to further
reduce the NOx concentrations by 10—80%, with the bulk of the
applications in the 65-80% range. This accounts for removing 70-100 ppm
of NOx. This important fact is often not clearly highlighted because
the SCR system performance is usually referenced as a percentage
reduction- of the NOx that enters the SCR reactor, and not as the total
amount of NOx removed or as a percentage reduction of the uncontrolled
NOx emissions.

Plants located in crowded urban sites often apply 10-30% NOx reduction,
as that is the most that can be achieved due to problems of physically
locating the catalysts.and duct work at the site. The 80% removal rate
has been clearly stated( 10•13*17) as being a realistic upper limit for SCR
system operation at an electric utility plant. This is the result of
the- need to limit ammonia discharge to the atmosphere, the difficulties
in obtaining greater reductions due to the inherent flow and
concentration maldistributions in such physically large systems, and
difficulties in measuring the discharge of the NOx and ammonia and those
measurements' impact upon the ability to reliably maintain control of
the plant.

Table 1 presents a summary of the Japanese boiler NOx emissions achieved
for various combinations of combustion modifications. Table 2 indicates
that SCR is only applied to about 15% of the system capacity at Tokyo
Electric Power Company, a company with predominantly urban-sited power

7A.-43

-------
generation facilities. Many of the Tokyo - Electric Power Company SCR
systems only achieve 10 or 30% NOx reduction. Similar situations exist
at the urban plants of Chubu and Kansai Electric Power Companies.
SCR SYSTEM DESIGN EXPERIENCE

Although the basic SCR system has remained relatively unchanged from the
initial designs, the catalyst itself has been the subject of continuing
improvement through research and development. Early utility operations
identified problems that were amenable to correction through design
changes. Such problems included: sulfur trioxide formation,* ammonium
bisulfate formation and deposition? catalyst poisoning by sulfur oxides
and trace elements in the fuel; erosion of the catalyst by "hard" fly
ash; and plugging of the catalyst by dust.

Table 1

Japanese Utility Boiler NOx Emissions For
Various Combustion Modification NOx Abatement Technologies*b>

FUEL

COAL

OIL

GAS

Fuel Nitrogen Content, %
TECHNOLOGY APPLIED
No Control

Flue Gas Recirculation (FGR)
Two-Stage Combustion {TSC)
FGR and TSC

FGR and TSC and Low-NOx Burner

0.7-3.0 " 0.1-0.5 0

NOx EMISSIONS < in ppm)(a>

500-800
400-600
300-500
250-400
150-300

250-400
180-300
180-300
120-250
60-150

200-300
140-200
150-200
100-150
40- SO

Table 2

Tokyo Electric Power Company Oil and Gas-Fired Boilers Applying
Combustion Modification and SCR NOx Abatement Technologies*b'

METHODS OF COMBUSTION MODIFICATION
GENERATING

{MW)

Flue Gas Recirculation (FGR)
FGR and Two-Stage Combustion (TSC)
FGR and Low-NOx Burners (LNB)
FGR and TSC and LNB

TOTAL

Boilers with SCR Also Added
Motes:

NUMBER OF
BOILERS

15
12
9
32

TOTAL
CAPACITY

1,860
5,290
2,710
13,610

23,470

3 ,540

(a) based upon 6% O2 for coal, 4% O2 for oil, and 5% O2
for gas. To correct the values to a 3% 02 reference,
multiply the values by 1,20 for coal, 1.059 for oil, or
1.126 for gas.

(b) from J. Ando, references 2 and 4.

7A-44

-------
The change to a titanium oxide-based catalyst resulted in a formulation
that solved most of "those problems in that it is resistant to sulfur
oxide poisoning, and it does not oxidize sulfur oxides to the trioxide
form. Without the trioxide present, ammonium bisulfate will not form.
Newer designs include titanium with molybdenum and tungsten oxides, and
they are reported to be even more resistant to the formation of sulfur
trioxide.

In addition, flow straighteners to reduce flow maldistributions at the
SCR reactor inlet, flue gas velocity limits in the catalyst region, and
the use of parallel flow geometries for the catalysts helped control
erosion and plugging problems. Extensive flow modeling, is used to guide
the design ¦ of the ammonia injection system. The designs include
multiple rows of catalyst sections, and the amount of ammonia injected
is. also restricted in acknowledgement that up to 2% of the flue gas can
bypass a row of catalyst blocks by passing through the spaces between
the blocks. The limiting of the amount of excess ammonia reduces the
duty on the catalyst and reduces its potential to produce bisulfates.
The result has been a more rugged and longer life catalyst, but that
trade off solution also has lowered, by 10-20%, the amount of NOx
reduction that can be achieved in actual operations.

These changes have been.made with the following basic design conditions
and approaches retained. The catalyst operation is still in the 300-
400°C range (recent designs have been in the 300-350°C range)', and the
catalyst high NOx reduction activity has been retained. The ammonia
injection and control system is stable and responsive. The injected
ammonia is vaporized and diluted with a carrier gas to a 5% ammonia
concentration before it is injected into the duct. This is well below
the ammonia flammability range in air of 16-25%. Blocks of catalyst are
stacked inside the SCR reactor, but room is left for the addition of
test coupons and additional catalyst blocks if that should prove
necessary. The test coupons are examined during the annual plant
inspections to help identify incipient problems, so that corrective
action can be taken before a difficulty becomes a major concern.
The Japanese design strategy assumes that a moderate amount of ammonium
bisulfate deposits will form and it recognizes that the deposit is
highly corrosive (and thus special materials should be selected for
those elements on which-it could deposit), but that it is removable by
soot blowing' and water washing. Where high particulate loads are
anticipated, they use a vertical downflow through the catalyst section.
Otherwise, they prefer horizontal flow through' the catalyst section
because that results in less expensive duct work.

SCR SYSTEM SITING EXPERIENCE

The SCR, system for a utility gas or oil-fired boiler is a large system,
about one half to two .thirds the size of a flue gas desulfurization unit
that might be applied to the plant. For a system designed to treat the
flue gas from a 250 MW plant, the basic SCR catalyst housing would be
a structure about 30 feet square and about 50 to 70 feet high. The
catalysts themselves could weigh as much as 400 tons, depending upon
their formulation and the base material selected.

Of special concern is the retrofit of an SCR system at a site that had
not initially planned for and provided space for its installation. The
Japanese regulatory philosophy'7} on retrofit siting is basically that

7A-45

-------
the SCR should be designed to fit into the reasonably available space.
At crowded urban utility sites, such as in Tokyo, this has resulted in
the installation of SCR reactors that are designed to remove 10-30% of
the NOx entering the catalyst reactor, as compared to the 80% removal
rate typical of new sites that can accommodate those larger facilities.
Also, SCR retrofits cost 40-60% raore(4 >10 > than do installations at new
electric utility plants.

In 1980<7> site visits were made to discuss and .to observe the SCR
systems installed by retrofit at: Chugoku Electric Power Corp.'s
Kudamatsu Station Unit 2 (375 MW) ; Chubu Electric Power Co.'s Chita
Station Unit 4 (700 MW) and Nishinagoya Station Unit 6 (500 MW) ,* and
Kansai Electric Power Co.'s Osaka Station Unit 4 (156 MW) and Sakaiko
Station Unit 6 (250 MW).

A- summary of the findings of that study is: 1. The plants all applied
combustion modification technologies in order to lower the boiler NOx
production. The SCR systems were designed to operate at about 80% NOx
reduction. 2. There was ample space available at each site for
positioning the SCR unit without the need to relocate major, equipment
items such as air preheaters, fans and the exhaust stack. 3. There was
a minimal need to modify existing foundations and steel structures to
support the weight of the SCR system components. 4. At part load
operating conditions the regulators required a constant NOx emission
limit (ppm) rather than a constant NOx removal percentage rate. This
effectively required less NOx reduction at part loads when the plant NOx
emissions are reduced compared to full load operation. 5. Minimal
relocation of underground utilities was required and new electrical
transformers were not required. 6. Boiler-structural modifications to
meet new pressure conditions were not required. 7. Flue gas reheat
systems were not required. 8. The design for boiler implosion
prevention was not required.

SCR SYSTEM OPERATING EXPERIENCE .

The early experience11» 2 •3 • * •a • 9 ¦14 •17«19 ¦20) with pellet (spherical shaped)
and tubular geometries used for forming the SCR catalyst sections, along
with the actual plant operations, identified a number of generic or
repetitive problems. These included: 1) ammonium bisulfate formation
and its subsequent deposition on the catalyst itself, termed "masking",
and on the downstream equipment. It has been found<2 >4'7 •11 >13¦17 > that:
ammonium bisulfate deposits occurred on the ammonia injection nozzles,
the SCR catalyst, the air preheater, the induced-draft booster fan, and
the gas-to-gas heat exchanger; 2) catalyst poisoning by sulfur oxides
and other constituents in the flue gas occurred? 3) catalyst plugging
by fine particle dust occurred,- 4) sulfur dioxide conversion to trioxide
followed by bisulfates formation occurred; 5) catalyst erosion by hard
particles in the fly ash occurred (this was principally a problem in
coal-fired plants) ,* and 6) NOx and ammonia measurement instrumentation
inaccuracy and low durability (or high maintainability) was experienced.

Extensive development activities for improvement of the catalyst
formulation and operational changes were taken in order to mitigate
those problems * The status of these efforts xs.

1) Bisulfate formation and deposition problems. The Japanese

attribute the design change to titanium- and vanadium-based

7A-46

-------
catalysts as the single most important catalyst development for
controlling this problem. However, this change does not remove
the sulfur, so additional operational changes were required. The
single most important operational change was attributed to being
the reduction of the amount of injected ammonia (even if it
reduced the overall NOx reduction). By keeping the excess ammonia
(e.g., the ammonia slip or bypass) to less than 5 ppm (and
preferably less than 3 ppm) the rate of bisulfate production is
small enough that the resulting deposits can be controlled by soot
blowing and water washing. Another operational change was to
reduce the catalyst operating temperature to less than 350°C to
reduce further its oxidation potential (but remaining above 300°C
to restrict bisulfate formation). The actual operating
temperature has been found5 to be a function of the inlet
concentrations of the sulfur oxides and ammonia' and the sulfur
dioxide oxidation potential of the specific catalyst being used.

2. The catalyst poisoning problems were corrected by converting
to the titanium-based catalyst material, which is more resistant
to poisoning by sulfur oxides. Operational changes such as
tighter controls on permissible fuel trace elements, such as
alkali and arsenic compounds, help eliminate these problems.
Catalyst samples are tested during the annual plant inspection.
Those test results can give an early warning of any significant
loss of catalyst activity.

3. Correction of the plugging of the catalyst pores with fine
particle dust (particles less than 1 micron in size) was
accomplished by soot control operational measures such as soot
blowing, and water washing. In addition, the use of parallel flow
catalyst geometries further reduced the opportunity for fine dust
to reach the catalyst surface. The utilities are restricted from
using "combustion enhancing additives" and cleaners because they
have been found^4' to contribute to the fine dust problem. The
covering or fouling of the catalyst surface with bisulfates is
similar to this problem, but the dust deposition is considered as
a "blinding" problem, while the gross covering by bisulfates is
considered as a "masking" problem. The difference is that soot
blowing, washing (some catalyst suppliers do not allow washing of
their catalysts), and volatilization by heating can remove the
masking, problem, but they are not effective operational mitigation
procedures for the blinding problem.

4. The steps taken to limit the formation and deposition of
bisulfate salts essentially corrected the problem of farming
sulfur trioxide.

5. Erosion control was obtained by a number of design changes
related to flow geometry and- flue gas velocity.

6. The instrumentation problems remain an open
issue< 2.6.8,11,13,15,16). in general, their long response times and
inability to obtain a reading in the ppm range that represents the
conditions in a duct that is meters in diameter are difficult
problems. The control of the SCR system is often not based upon
actual instrument readings of the NOx and ammonia levels, but is
based upon using calculated values. The instrument readings of
either of these concentrations can be used to trim the control
setting, but they may not be accurate enough and may have too slow

7A-47

-------
a response time for any significant control other than, steady
state 'or static operational conditions. Also, the maintenance
effort for the instrumentation is considerable56«8>.

The main SCR concern during startup and shutdown operations is the
potential for damaging the catalyst due to soot or moisture'11 > . The
Japanese do not operate their SCR systems during these periods and
during low power operations (e.g. power operation between 0-40%,
depending upon the plant being considered) .

The SCR system control is strongly tied to instrumentation issues. The
most obvious*11'12> 13> operational control strategy has been to design the
SCR system for operation at 80% NOx removal and to frequently operate
it at lower removal rates by injecting less than the design feed rate
of ammonia. This is done, of course, to limit the ammonia slip and to
thus assure that bisulfate problems do not occur. It also reduces the
dependency of the operations upon accurate emission measurements, since
all the ammonia should be consumed and the NOx emission level that
results appears to be acceptable to both the utility and the regulators.
Another way to look at this is that, in effect, additional catalyst
material, above that which is needed for operating conditions, is
installed. This contributes to the long catalyst lifetimes actually
experienced (in oil and gas service catalysts are warranted for 2 years,
yet they last for 4 to 8 years for oil and 7 to 11 or more years for
gas-firing, indicating considerably more catalyst charge than necessary
was present).. During site visits it was observed that, as a rule of
thumb, the systems were typically operated at 10% less NOx removal
efficiency than the specified design condition. Conversely, if 90-95%
NOx removal control were to be required and enforced, significantly
shorter catalyst lifetimes would be expected.

Actual operational control'4 > is based on "measuring" the flue gas flow
by measuring the fuel flow rate and calculating the resulting gas flow
rate that should be produced by burning that amount of fuel, or by
measuring the steam flow to the turbine and using that as a "measure"
of the amount of fuel being burned, and thus the _ combustion gas
produced. The NOx concentration is also "measured", But the lag time
(up to 10-15 minutes) of the measuring instrument means this measurement
is not an effective for control purposes except for non load-following
operations. It appeared that at several plants'13) the NOx level too was
a calculated number. Based upon a calculated flue gas flow rate and NOx
concentration, the controller logic specified an ammonia feed flow rate.
In some cases11" the ammonia slip measurement was used as a trim signal
to adjust the ammonia feed flow rate so that the slip stayed below 5

One of the reasons why ammonia monitoring is not used to provide a
control signal is that the ammonia meter often works on the principal
of splitting the measurement sample into two flow streams. In one
stream the ammonia is converted to nitrogen oxide and the total nitrogen
oxide is read by an instrument. In the other stream the nitrogen oxide
in the exhaust is read directly. The difference between these two
reading is proportional to the ammonia content. This is a difficult
measurement to make because both the NOx and the ammonia are present in
low concentrations, on the order of ppm, and the measurement is really
the difference between two nearly equal numbers. This can routinely
lead to significant accuracy problems.

7A-48

-------
As an example of the difficulty of relying upon the accuracy of the
operational instrumentation, the following is cited , but it occurred in a pilot plant operation. The important
point is not that coal ash eroded the catalyst, but that the monitoring
instrumentation had not identified that a significant' mal-operation and
a corresponding loss of substantial NOx control had occurred.

None of the plants visited during prior visits maintained a spare parts
inventory'13' for their SCR systems. However, all of the operating staff
had received specific training on operations and maintenance from the
SCR system supplier, and the system supplier also supplied engineering
staff to directly support the utility staff during the annual SCR system
inspection. The utility staff for operations and maintenance (the same
people perform both tasks.) varied from 6 to 14 per shift. The exact
size of the staff appeared, to be more a function of the utility
historical staffing patterns than a- function of the SCR system size.

Maintenance problems have contributed to the Japanese regulatory
authorities restricting the use. of instruments based upon
chemiluminescence principles'13*.. Even in cases where the regulators
have not restricted the use of such instruments, it was found that
utilities have replaced. them with designs based upon infrared and
ultraviolet adsorption principles. . Studies and the development of new
or more reliable instruments are being actively pursued!19) . A typical
NOx and ammonia instrumentation maintenance schedule is: daily
calibration; bi-weekly inspection and cleaning of the meters; tri-
monthly maintenance of meters and recorders; annual overhaul of
analyzers and recorders; and replacement of the entire instrumentation
system after 6 or 7 years.

The other significant maintenance activity is the actions taken to
control or remove soot and ammonium bisulfate deposits. Different
utilities use different but related processes. For example, at some
plants on-line soot blowing of the air preheater is done three times per
day for two hours at a time. Blowing is done from the hot and the cold
side of the unit. Two tons per hour of steam are used. This planned
schedule was based upon prior Japanese experience with other oil-fired
plants that injected ammonia for ash conditioning for hot electrostatic
precipitator operations, and was selected for the SCR system "just to
be safe". During low power operations, soot blowing often is performed
on a continuous basis.

One 700 MW boiler system uses 5,000 tons of wash water per washing. In
addition to cleaning the bisulfate material out of the wash water,
additional treatment is required to remove the high level of ammonia
contamination in the water. The washing cycle includes a period to cool
down the equipment, to wash, and then to dry the equipment. This takes
24 to about 40 hours, depending upon the equipment size, and is usually
performed during a weekend when the system power demand is low.

7A-49

-------
(I

In the case of the 700 MW unit sited above, air blowing was originally
used as the on-line soot removal method. They changed to steam blowing
in order to avoid cooling the hot side and further contributing to the
bisulfate problem through the impact of the cooled metal on the
bisulfate carrying flue gas. The low temperature side of the air
preheater is blown once per shift, the hot side once per day. A 200 psi
jet is used, and the blowing is continued for a two hour period. Even
with these measures, water washing such as discussed above, is required
once every three months.

Other utilities report that they employ water washing of the air
preheater at least twice a year. Soot blowing of the air preheater,
frequently the catalysts themselves, and other equipment items such as
the induced-draft fans occurs weekly and after each startup.

APPLICATION TO U.S. OIL-FIRED BOILERS

Since SCR suppliers in the U.S. have not undertaken pilot or
demonstration programs such as was done in Japan and Germany in order
to qualify the SCR designs for the specific fuels and operating
conditions that are anticipated, the level of risk at a U.S. application
is greater than that for a Japanese application. That is, the use of
SCR systems could place the boiler and plant at significant risk of
operational failure and subsequent plant unavailability if the SCR
technology application were to exceed currently proven operating
conditions.

To be consistent with the Japanese experience, the NOx level entering
the SCR system should not exceed about 150 ppm and the SCR system should
not be operated during relatively fast or frequent transients or load
changes. It is still possible to design SCR system for operations with
greater- than 150 ppm NOx at the reactor inlet, but then it can not be
claimed that the design has been validated through the Japanese
operating experience.

The ammonia injection system should be validated through design analysis
and or testing, to assure that it introduces a relatively uniform, fully
volatilized ammonia gas which has been diluted in a carrier gas such
that the ammonia concentration is less than 10%. An appropriate mixing
length to allow the ammonia to mix thoroughly within the flue,gas is
needed so that a uniform gas steam will enter 'the SCR reactor.

Ammonia contamination of wash waster and fly ash does not appear to have
yet been fully addressed'by U.S. regulators. Japanese experience has
shown that ammonia contamination of fly ash collected at the
electrostatic precipitator can be anticipated. This is a serious
problem for the Japanese because often they sell their ash to the cement
industry, and the ammonia odor makes it unsalable. In the U.S.,
contamination of fly ash with ammonia means that normal land-fill
disposal of the ash may not be possible' because of leaching of the
ammonia from the stored ash. If, for example, the ammonia slip were 5
ppm and 90% of that were retained by the ash normally present from oil-
firing, then the ammonia content in the fly ash would be about 0.14%.
However, ammonia leaching5 4} has been found to be a problem if the
ammonia content in the fly ash is 0.02% or greater.

Ammonia contamination occurs in the wash water used for removing soot

7A-50

-------
and bisulfates. This causes a water cleanup problem. For example, the
water wash at a 700 MW unit uses 5000 tons of water for each wash.
Ammonia contamination also occurs in the flue gas desulfurization (FGD)
system's waste water. This too adds the requirement for a water
treatment step before that waste water, can be disposed. It has been
suggested'8 > that ammonia contamination of the FGD waste water is
possibly the most significant adverse effect for SCR operations at U.S.
electric power plants.

The SCR system designers and the regulatory community need to recognize
the relative uncertainty in the measurement of NOx and ammonia, and the
controls and emission permit conditions also should recognize that. The
consequence of this is that it may be necessary to reduce the ammonia
injection rate, as is typically done in Japan, below the "design"
injection rate in order to assure that very low levels of ammonia bypass
the SCR reactor. This will be especially important if sulfur-bearing'
fuels are burned, and it may actually reduce the level of NOx control
that is obtained from the SCR system. However, this approach is
necessary if the availability of the entire plant is to be maintained
at acceptable levels.

In recognition of the uncertainty that will accompany the application
of SCR systems to large, oil-fired boilers-, extra space for additional
SCR catalyst charge should be designed into the SCR system, SCR catalyst
test specimens should be placed in the catalyst reactor and periodically
examined, and detailed inspections, of the SCR system should be planned
and conducted at least yearly. If the design basis assumes that soot
and ammonia bisulfate deposits will occur, and if specific control and
removal systems are designed into the SCR system, then the risk of
uncontrolled deposits will be significantly reduced.

Since there is a significant question about instrumentation accuracy and
durability, and since the Japanese experience is mostly for base load
operation, and since SCR operation is not required during low load,
startup, and shutdown, U.S. regulators should also consider similar
adjustments in their requirements for SCR operations. In addition, the
typical U.S. requirement for a 15' minute averaging time for emission
monitoring may not be realistic and a longer averaging times should be
considered.

REFERENCES

1. J. Ando. SCR for Existing Oil-Fired Boilers, August
1988.

2. J. Ando Draft. Report, SOz and NOx Control Technology
Developments in Japan, January 1987.

3. J. Ando. Recent Developments in SQa and NOx Abatement
Technology for Stationary Sources in Japan, EPA 600/7-
85-040, September 1985.

4. J. Ando. NOx Abatement for Stationary Sources in* Japan,
EPA 600/7-83-027, May 1983.

5. J. Ando. NOx Abatement for Stationary Sources in Japan,
SPA 600/7-79-1080, August 1979.

7A-51

-------
6. E. J. Cichanowicz and G. R. Of fen. Applicability of
European SCR Experience to U.S. Utilities, EPRX CS-5361,
Vol 2, August 1987.

7. W. Ellison, and T. A. ' Ritter. Survey of Japanese
Application of Selective Catalytic Reduction Systems,
November 1980.

8. W. Goldschmidt. VKR Full-Scale SCR Experiences in
Hardcoal Fired Boilers, EPRI' CS—5361, Vol -2, August
1987.

9. H. L. Faucett, et al. Technical Assessment of NOx
Removal Processes for Utility Application, SPA 600/7-
77-127, November 1977.

10. Jones, G.D., "Selective !Catalytic Reduction and NOx
Control in Japan, A Status Report", EPA 600/7-81-030,
January 1981.

11. P. A. Lowe. Report on the Status of Selective Catalytic
Reduction NOx Control as Applied to Gas Turbine Electric.
Power Generation Plants, NUS-4763, December 1985.

12. P. A. Lowe. Utility Operating Experience With Selective
Catalytic Reduction of Flue Gas NOx, Proceeding of the
2nd International Conference on Acid Rain, March 1985.

13. P. A. Lowe. Review of Japanese 'NOx Control Technology
Retrofitted to Coal-Fired Boilers, NUS 4465, January
1984.

14. J. D. Mobley. Assessment of NOx Flue Gas Treatment
Technology, Proceedings Joint Symposium on Stationary
Combustion NOx Control, IERL-RTP-1080, October 1980.

15. L. J. Muzio, et al. Control of Nitrogen Oxides:
Assessment of Needs and Options, Technical Support
Document, EPRI EA-2048, Vol 5, July 1983.

16. P. Necker. Operating Experiences with SCR DeNOx Plant
in Unit 5 of Altbach/Deizisau Power Station, EPRI CS-
5361, Vol 2, August 1987.

17. A. V. Slack. Applicability of Japanese NOx Control in
The U.S., DOE/MC/15091-1612, February 1980.

18. K. Suyama. The Improvement of NHs Injection Control
System For a Selective Catalytic NOx Removal System,
EPRI CS-5361, Vol 2, August 1987.

19. T. Suzuki. Development of Dry Type Flue Gas
Denitrification Equipment for Boilers (Japanese
language), IHI Engineering Review, May 1979.

20. Flue Gas De-NOx System Development of Catalyst for Dry
Type (Japanese language), Mitsubishi Juko Giho, July
1980.

7A-52

-------
NOx Control in a Brown Coal-Fired Utility Boiler

John E, Hofmann
Fuel Tech, Inc., Stamford, CT, USA

Joachim von Bergmann
Fuel Tech GmbH, Frankfurt, Germany

Dr. Dieter Bokenbrink
Professor Dr. Klaus Hein
Rheini sch-Westfalisches Elektrizitatswerk, Germany

ABSTRACT

Rheinfsch-Westfalisches Elektrizitatswerk (RWE), initiated a number of tests
during 1987 of the N0x0UT(R) Process for NOx reduction on a 150 MWe brown coal -
fired boiler. The NQxOUT Process uses urea and enhancer chemicals for the
reduction of NOx to molecular nitrogen, water vapor and carbon dioxide. The test
objectives of up to 50% NOx reduction and NH, slip of <5 ppm were met over a
range of operating conditions.

In 1988, a commercial NOxOUT system was subsequently installed on a 75 MWe brown
coal-fired boiler to comply with a controlled NOx level of 200 mg/Nm3
(approximately 100 ppm). NOx was first reduced to 300 mg/Nm3 (approximately 150
ppm) by the use of combustion modifications. Additional NOx reduction down to
180-195 mg/Nm3 (approximately 90-98 ppm) has been achieved using the NOxOUT
Process. NH, slip has been controlled to a level of <2 ppm through the
combination of enhancer chemicals plus selective injection.

INTRODUCTION

The Federal Republic of Germany is recognized as a world leader for the control
of sulfur and nitrogen oxides emitted from stationary combustion sources. Acid
rain arising from these pollutants is not only affecting human health but is also
considered to be one of the major contributors to the damage of the fauna, in
particular to forests, and to historical buildings (1, 2). Current legislation
and local agreements between industry and the state authorities require that
nitrogen oxides be reduced to a maximum of 200 mg/Nm3 (approximately 100 ppm) for
all boilers in excess of 300 MWth by the end of this decade.

Rheinisch-Westfalisches Elektrizitatswerk A.G, (RWE) has a total generating
capacity of more than 20,000 MW and is the largest utility in the Federal
Republic of Germany. One of the major fuels of RWE is brown coal which is burned
in power stations of about 10,000 MWe installed capacity, with individual units
sized up to 600 MWe.

The Rhinish brown coal differs distinctly from other fossil fuels such as
bituminous coals or lignites. A high moisture level of 55 - 62% on a raw coal
basis, a variable ash content of 2 - 20%, and a low nitrogen concentration of
0.3 - 0.4%, are typical for this fuel.

Consequently brown coal combustion requires specifically designed fuel
preparation circuits and boilers which in conjunction with the fuel properties
result in specific flue gas characteristics. The flue gas contains about 20%
water vapor resulting in maximum flame temperatures below 1200°C and fairly
constant NOx concentrations well below 800 mg/Nm3 (approximately 400 ppm).

7A-53

-------
It is noteworthy that this fuel is mostly burned in tangentially fired boilers
with jet burners in vertically arranged slits in the walls and/or comers of the
combustion chamber. Normal operation with one burner slit always out of service
at full load and more slits out of service at partial load differs from the
firing mode of higher quality coals and results in asymmetric horizontal
temperature distributions in the boiler, which has major consequences, as will
be discussed later.

In order to comply with the NOx emission requirements, after having completed
extensive large scale testing and economic evaluation of the selective catalytic
reduction (SCR) technique (3), and also having reached very promising results
with combustion modifications (4), RWE decided to retrofit 37 boilers with a
total capacity of 9,300 MWe with alterations to the combustion chamber only.
However, there were some indications that the final goal of a maximum NOx
emission of 200 mg/Nm3 (approximately 100 ppm) may not be achievable for all
boilers and under all operating conditions. Therefore, additional NOx removal
principals would be required. This led to large scale tests of various options
of injection systems.

One major option for RWE was the NOxOUT Process m. The NOxOUT Process licensed
through Fuel Tech, N.V. is a relatively low cost retrofit NOx control technology.
The NOxOUT Process can be installed with a short downtime and is applicable to
a very broad range of fuels and boilers. NOx reductions up to 80% are achievable
with this technology (5). The NOxOUT Process can be installed along with
combustion modifications to achieve even higher levels of reduction. Hence, the
NOxOUT Process is applicable as a primary retrofit application and can be applied
in conjunction with combustion modifications.

The NOxOUT Process is based upon the chemical reaction between nitrogen oxides
and urea.

2N0 + NH2C0NH2 + 1/2 02 > 2N2 + C02 + 2H20

The reaction takes place at a temperature of about 925 - 1,100°C. Research on
the urea-based process was initiated in 1976 under sponsorship of the Electric
Power Research Institute (6). Starting with the EPRI inventions (7, 8), Fuel
Tech has made a number of contributions to improve the application of the
technology (9, 11), including broadening and/or shifting of the optimum
temperature for reaction by the use of Fuel Tech proprietary chemical enhancers
(10, 12, 13, 14).

Fuel Tech patented enhancers have also been effective in the control of ammonia
formation (11). Ammonia (NH3), is a by-product of the reaction between NOx and
urea that can be formed under certain conditions. Ammonia production is
generally undesirable because of the possibility of forming sulfates and
bisulfates in the presence of sulfur trioxide and in the contamination of ash,
affecting ash disposal options. (Bisulfates have been known to cause fouling
in the air preheater area.) The sulfur content of RWE coal (0.2% by weight) and
the temperatures of the preheater area (<330°C) could provide a reactive
environment for the formation of bisulfates if there is ammonia present.

TEST PROCEDURES

The introduction of the NOxOUT Process to brown coal fired boilers was carried
out in various steps. The first preliminary tests had to show:

at least 30 to 50% NOx reduction is possible with NH, emissions, or
"si ip".below 3.8 mg/Nm3 (approximately 5 ppm)

7A-54

-------
the process 1s flexible enough to accommodate various combinations of
burners in operation

the process could cope with load variations down to partial loads of 60%

the process is also effective under simulated flue gas conditions which
were expected after the envisaged installation of combustion modifications.

These tests were carried out on a 150 MWe boiler (Unit Q, RWE Weisweiler power
station, figure 1} during the summer of 1987.

Temporary equipment was used for the D boiler tests which consisted of plastic
day tanks for chemical storage, a portable pumping skid, mixing headers for
combination of NOxOUT-A solution, enhancers and water, rubber hoses, and
distribution headers to split the flow between multiple wall injectors at each
of three levels of injection. NOxOUT-A is an aqueous solution of urea and
proprietary additives.

The first level of Injection was into the upper furnace region at 25 meters.

The second level was located immediately before the secondary superheater. The
third level was located at the top of the backpass after the primary superheater.
A number of different injectors were tried and the final arrangement consisted
of sixteen external mix injectors at level one (2 per penetration), nine pin jet
injectors at level two, and four internal mix injectors at level three. Steam
was used for cooling and atomization for Injectors at level one. Air was used
for cooling and atomization at levels two and three due to a limited availability
of steam.

After the successful completion of the preliminary tests a full scale
demonstration was conducted at the same power station on a 75 MWe boiler (Unit
C2, figure 2) during the summer of 1988. Combustion modifications, installed
during the spring of 1988, reduced the NOx emission from about 450 mg/Nm3
(approximately 225 ppra) to about 275-325 mg/Nm3 (approximately 140 - 165 ppm) at
full load. The major goal of the Unit C2 demonstration was the successful
performance of the NOxOUT Process for 4 weeks of continuous operation, during
which the NOx emission must not exceed the guaranteed level of 200 mg/Nm
(approximately 100 ppm). In addition, the process was required to meet this
emission limit while maintaining ammonia emissions below 5 ppm at all operating
conditions.

Based on experience from the D boiler, NOxOUT injectors were installed at three
levels; at 19 meters immediately above the burners but below the flue gas recycle
and overfire air ports, at 25 meters immediately before the secondary
superheater, and at 36 meters at the top of the backpass after the primary
superheater. There were 12 penetrations at level one, twelve at level two, and
four at level three. External mix injectors were used at all levels. Steam was
used as the cooling and atomizing fluid for all injectors.

TEST RESULTS - D BOILER

A. Temperature and Composition Profiles

Prior to the commencement of NOxOUT irtjection, temperatures were measured at each
of the three levels of injection. Temperature measurements were made using water
cooled suction pyrometers. Analyses for 02, CO and NOx were carried out at the
same time as the temperature measurements. Typical temperature measurements
for 100% load were found to be 1021°C (Level 1), 825°C (Level 2), and 580°C
(Level 3).

7A-55

-------
As expected, in the upper furnace region temperatures near the wall tended to
be lower than temperatures towards the center of the boiler. However, the
overall temperature profiles tended to be fairly uniform. The temperatures at
levels 1 and 2 were within the temperature window for enhanced urea chemicals.
The temperature at Level 3 however, was too cold for injection of chemicals
available at RUE.

Gas composition profiles for NOx, CO and 02 were recorded at the three levels of
injections. The NOx, CO and 0, concentrations differed at the three levels. NOx
values were highest at level 2, Higher NOx concentrations recorded at level 2
were associated with higher 02 concentrations at this level. The CO levels were
found to be significantly higher at level 1 than other levels. The higher
concentration of CO at level 1 indicates that complete burnout has not occurred
until after this point.

B. NOx Reduction. Level 1 Injection

Results obtained from injection of mixtures of NOxOUT-A and enhancer at level
1 in the D boiler are shown in Figure 3. The Normalized Stoichiometric Ratio
(NSR) is defined as the ratio of NHZ species injected divided by the initial
moles of NOx. Thus one mole of urea per mole of NOx represents an NSR of 2.0.
The initial NOx level for these series of experiments was approximately 550
mg/Nm3 (6% 02). The results show an essentially linear response to increasing
amounts of urea injection.

The effect of enhancer is to shift the temperature window for NOx reduction
activity to a lower temperature (5). Thus at a specific operating temperature,
the addition of enhancer could result in either increased NOx reduction,
decreased NOx reduction or no change in NOx reduction. A decrease in NOx
reduction implies that the reaction temperature (in this case approximately
1020°C) is such that the operation is on the high side of the reaction
temperature window. There is a benefit in operating on the high side of the
reaction temperature window in that by-product NH3 formation is suppressed while
NOx reduction is only slightly diminished (11).

The amount of NOx reduction achieved in these experiments implies a urea
utilization in the range of 27% with the low enhancer ratio shifting to 20% for
the high enhancer ratio (assuming that each urea molecule is theoretically
capable of reducing two molecules of NOx). As Figure 3 indicates, the response
of utilization to increasing NSR is essentially flat over the range of NSR
investigated. Other studies have shown that utilization will tend to decrease
with increasing NSR (5).

C. NOx Reduction. Two Level Injection

Urea and enhancer injection was performed simultaneously at levels 1 and 2 to
determine if multiple level injection would significantly (A) increase chemical
utilization, (B) reduce NH3 slip, and (C) increase the NOx reduction.

Results from injection of urea and enhancer at level 1 versus levels 1 and 2 at
comparable test conditions are contained in Table 1.

7A-56

-------
Generally, it can be concluded that urea utilization increases when urea is split
between two levels at the same overall NSR. These results were achieved despite
a slightly higher overall E/U ratio with two level injection. Although it has
previously been found that a higher E/U ratio leads to lower NH3 slip, NH3 slip
was as high with two level injection as with one. It is believed that NH, slip
from level 2 was the result of chemical impingement on the superheater platens
which were within 5 feet of the point of injection at Level 2. Efforts were made
to inject the chemicals between the platens, but it was virtually impossible to
completely eliminate direct chemical impingement.

D. Enhancer/Urea Ratio. Control of NH3 Slip

The enhancer/urea (E/U) ratio is one of the most significant process variables
in the NOxOUT Process. As previously noted, the enhancer/urea ratio can affect
the level of NOx reduction but it is particularly important in the control of
NH3 slip. Ammonia (NH,) is a by-product of the reaction between NOx and urea
that can be formed under certain conditions. The amount of NH, produced is a
function of various process variables, and tends to increase witn:

decreasing temperature

increasing the reagent to NOx mole ratio (increasing NOx reduction)
decreasing enhancer/urea ratio.

The amount of NH3 formed tends to increase with increases in NOx reduction and
therefore a trade off occurs when NOx reduction has to be forfeited to maintain
low NH3 levels.

Figure 4 shows a comparison of ammonia slip versus NOx reduction at two different
E/U ratios. At the higher E/U ratio less ammonia was generated per increment
of NOx reduction. The higher E/U ratio enables NOx reduction to be increased
from about 30% to about 50% at 5 ppm of NH3. NOx reduction can be further
increased to about 65% with an NH3 slip of less than 15 ppm.

E. Burners Out-of-Service

During the demonstration, extensive data collection was undertaken to determine
if differences in NOxOUT Process performance could be attributed to a specific
burner slit that was out-of-service. Temperature measurements were taken, as
previously described, prior to injection with different configirations of burner
slits in service. Only slight variations in daily average temperatures indicated
that there was no significant temperature difference with any specific burner
si it out-of-service.

In addition, process results obtained with the four different burner slits out-
of-service indicate no significant difference in process performance. The amount
of urea required to achieve a given level of NOx reduction and the level of NH3
slip were essentially unchanged.

F. Low Excess Air and Low Load

Two of the specific objectives of the demonstration were to determine the
flexibility of the NOxOUT Process to respond to reduced load and to low excess
air (simulated combustion modification).

Temperature profiles and gas composition analyses obtained at Level 1 revealed
significant differences for the modified conditions (Table 2).

7A-57

-------
Low excess air resulted in a higher temperature and a lower NQx content at level
1. Oxygen content was significantly lower and carbon monoxide (CO) was higher.
When load was reduced to 60% the temperature and NOx content dropped.
Substantial variation in CO content was observed from point to point at level
1 during the experiment with low excess air. These localized concentration
variations of CO could be the result of slow burnout of char. Process results
obtained with low excess air and low load are shown in Table 3 and compared with
data obtained at normal 100% load conditions.

During the experiments at higher NSR, enhancer alone was injected at level 2 to
provide an additional level of control over NHj slip.

Experiments at low excess air were initially carried out (exp. H) at the same
chemical flow rates as had previously been determined to be the optimum for 100%
load conditions (exp.G). The NSR increased from 2.4 to 3.0 because of the lower
baseline level of NOx and reduction declined from 49% to 45%.

At 60% load conditions (exp. I) the temperature at Level 1 is lower than at 100%
load (954°C vs. 1,021°C). At these conditions a three-fold increase in enhancer
rate was required to maintain the NH3 slip at 5 ppm. The fact that an NSR of
3.0 is required to achieve comparable NOx reductions for both low load and low
excess air conditions is most likely due to the lower initial NOx concentration
in the flue gas.

Overall, these data clearly demonstrate that the amount of enhancer required to
maintain NH3 slip below a target level depends upon the temperature at the point
of injection. The average temperature at level 1 varied over the range of 954
to 1,065°C. Within this temperature range, the amount of enhancer required to
meet NH, limits increased from 1 part enhancer to 30 parts urea at 1,065°C to 1
part enhancer to 2 parts urea at 954°C.

TEST RESULTS - C2 BOILER
A. Start-up Tests

Prior to the start-up of NOxOUT injection and soon after the C2 boiler returned
to stable conditions with the newly installed combustion modifications, Fuel Tech
measured baseline temperature, NOx, 0^ and CO, These measurements were carried
out in a grid at each level of injection.

Temperature and gas composition data gathered at 80 and 100% load are shown in
Table 4.

The overfire air which diverted a part of the combustion air to the ports above
19m had the effect of extending the primary combustion zone. The first injection
level was within this extended combustion zone. As a result of these conditions,
the values for both NOx and CO were extremely erratic and averages were
considered to be meaningless.

Although temperature levels at 19 meters would appear to be conducive to good
NOxOUT reagent utilization, the high CO levels existing at level 1 have the
impact of shifting the temperature window in which urea is effective (15). Early
injection tests at level 1 were not productive for NOx reduction. It was
apparent that the "effective" temperature at these high CO conditions was too
high for reasonable urea utilization. Further testing was based on injection
at the second (25m) level. The third level of injection (36m) was not required
because the total amount of NOx reduction required to achieve <200 mg/Nm3 proved
to be only about 35% which could be readily achieved at level 2.

7A-58

-------
B. Twentv-Five Meter Level Injection

NOx reductions in the range of 40-50% were repeatedly achieved at relatively low
NSR values from baseline NOx levels of 275-325 mg/Nm3. Figure 5 shows the
reductions and utilizations observed at varying NSR values. Although the
majority of the tests were performed with straight injectors, some comparative
tests were performed using angled injectors. NOx reduction and utilization were
not found to be sensitive to different types of injectors.

These tests were carried out at an E/U ratio in the range of 0.3 to 0.6. This
range was based on results previously determined to be optimum in the D boiler
tests. Specific tests to optimize the quantity of enhancer for the C2 boiler
remain to be undertaken.

Tests at 80% load with each one of the four separate burner banks out-of-service
showed that the process results were independent of the specific burner bank that
was out-of-service. Baseline NOx levels at 80% load were typically in the range
of 250-275 mg/Nm5 compared to 275-325 mg/Nm3 at 100% load. The NSR required to
achieve <200 mg/Nm3 was about 1.5 and reagent utilization is lower than that
achieved at 100% load. The reason for this is postulated to be the higher CO
concentration at 80% load. As previously noted, CO shifts the temperature window
to lower temperatures, and at the upper end of the window this results in
decreased reagent utilization.

C. Ammonia Slip

The Unit 02 contract between RWE and Fuel Tech specified that after the initial
optimization portion of the demonstration, the "recipe" for achieving the
necessary NOx reductions over a range of operating conditions would be turned
over to RWE to test its ability to operate continuously for four weeks. Ammonia
slip was a very important consideration for RWE. During the four week acceptance
period, 64 analyses of ammonia were to be undertaken to determine compliance with
the 5 ppm limit. Fuel Tech performed considerable ammonia sampling and computer
modelling to gather information on the potential for ammonia formation.

Samples of flue gas were withdrawn from four locations in the flue gas ducts
upstream of the air heater. The flue gas temperature at this point is
approximately 325°C. The baseline level of NH, was determined before the start
of injection and was found to be 0.7 ppm. There was no stratification of NH,
in the flue gas. Following the start of injection at the second level (25
meters) it was found that the overall specification of <5 ppm of NH, could be
maintained at both 100% and 80% load while also maintaining NOx at <200 mg/Nm3.

A computer model of the C2 boiler was utilized in an effort to get a better
interpretation of the flow and temperature regimes affecting ammonia formation
within the boiler at the second level of injection. Figure 6 shows a computer
generated temperature profile at 25 meters and the location of the 12 injectors.
The model showed that velocity and temperature are influenced by the boiler
throat which starts right above the 25 meter level. This throat shifts the swirl
in the flue gas which is characteristic of a tangentially fired boiler. This
shift combined with the location of the overfire air ports results in cooler
areas in the front right and back left corners away from the core of the swirl.
This suggested that injection into these cooler regions might be leading to NH3
si ip.

7A-59

-------
Elimination of five of the twelve injectors at level two combined with an
enhancer/urea ratio of about 0.6 led to the extremely low levels of NH3 slip.
A total of 64 individual analyses for NhL were carried out over a one month
period; 24 at 100% load and 40 at 80% load. The overall average of 1.5 ppm is
less than one ppm above baseline. The overall amount of slip is well below the
target of 5 ppm.

CONCLUDING REMARKS

The results of the above described projects at the two brown coal fired boilers
can be summarized as follows.

A NOx reduction of 40 to 50% can be achieved with less than 5 ppm of NH3
slip. Higher NOx reductions are possible depending on the ability to
achieve a good distribution of the chemicals and/or on the availability
of multiple levels for injection.

In particular, with a NOx emission of 300 mg/Nra3 (approximately 150 ppm)
or less, which was established by the application of combustion
modifications, the intended emission value of less than 200 mg/Nm3
(approximately 100 ppm) can be reached with an NH3 slip of well below 5
ppm.

The use of enhancer is effective in the control of NH3 slip over various
boiler loads especially when load changes result in temperature shifts
within the boiler.

The process is effective at various loads and operating conditions, without
having to change injectors.

Reagent utilization increases with multiple level injection of urea where
more than one level is available at appropriate temperatures.

The use of enhancer permits the chemical utilization to be optimized for
a specific reaction temperature.

The process can be easily retrofitted on an existing boiler with only small
disruption to on-going operations.

RWE's performance objectives for NOx reduction were met by the NQxOUT
Process. Further optimization of operating cost can be achieved as the
technology evolves. This will be done by optimizing chemical treatment
and injection parameters for specific boiler applications.

REFERENCES

1. Molrnen, V.A., "The Challenge of Acid Rain", Scientific American. Vol. 259.
No. 2, 30-38, August, 1988.

2. Bruck, R.I., "Decline of Boreal Forest Ecosystems in Central Europe and
the Eastern North America - Links to Air Pollution and the Deposition of
Nitrogen Compounds", 1987 Joint Symposium on Stationary Combustion NOx
Control, March 23-26, 1987, New Orleans, LA

3. Konig, J., Derichs, W., Hein, K., "Untersuchungen zum Einsatz der SCR-
Technik hinter Braunkohlefeuerungen", Sonderheft der Zeitschriften BWK,
Til und Umwelt im VDI-Verlag (1988), Heft 10

7A-60

-------
4. Hein, K., "The Application of Combustion Modifications for NOx Reduction
to Low-Rank Coal-Fired Boilers", 1989 Joint Symposium of Stationary
Combustion NOx Control, March 6-9, 1989, San Francisco, CA.

5. Epperly, W.R., Broderick, R.G., and Peter-Hoblyn, J.D., "Control of
Nitrogen Oxide Emissions from Stationary Sources", American Power
Conference, April 20, 1988, Chicago, IL,

6. Muzio, L.J., and Arand, J.K., "Homogeneous Gas Phase Decomposition of
Oxides of Nitrogen", EPRI Report No. FP-253, 1976.

7. Arand, J.K., and Muzio, L.J., "Urea Reduction of NOx in Combustion
Effluents", U.S. Patent #4,208,386 (1980).

8. Arand, O.K. and Muzio, L.J., "Urea Reduction of NOx in Fuel-Rich Combustion
Effluents", U.S. Patent #4,325,924 (1982).

9. Epperly, W.R., Peter-Hoblyn, J.D., Shulof, G.F, and Sullivan, J.C., "Multi-
Stage Process for Reducing the Concentration of Pollutants in an Effluent",
U.S. Patent #4,777,024 (1988).

10. Epperly, W.R., and Sullivan, J.C., "Process for the Reduction of Nitrogen
Oxides in an Effluent", U.S. Patent #4,770,863 (1988)

11. Epperly, W.R., O'Leary, J.H., and Sullivan, J.C., "Process for Nitrogen
Oxides Reduction and Minimization of the Production of Other Pollutants",
U.S. Patent #4,780,289 (1988).

12. Epperly, W.R. and Sullivan, J.C., "Process for the Reduction of Nitrogen
Oxides in an Effluent", U.S. Patent #4,803,059 (1989).

13. Bowers, W.E., "Reduction of Nitrogen-Based Pollutants Through the Use of
Urea Solutions Containing Oxygenated Hydrocarbon Solvents", U.S. Patent
#4,719,092 (1988).

14. Bowers, W.E., "Reduction of Nitrogen - and Carbon-Based Pollutants", U.S.
Patent 14,751,065 (1988).

15. Caton, J.A. and Siebers, D.L., "Comparison of Nitric Oxide Removal by
Cyanuric Acid and by Ammonia", Paper No. 88-67, Western States Section/The
Combustion Institute Fall Technical Meeting, Dana Point, CA, October, 1988.

7A-61

-------

TABLE 1

SINGLE VS. TWO LEVEL INJECTION

Experiment

No. of Levels

' 2-

1.:-

NSR (1)

1.2

1.0

2.7

2.6

3.0

3.3

E/U (2)

0.1

0.15

0.2

0.56

0.03

0.2

% Reduction
^Utilization
NH^, ppm

31.4
26.0
13.0

44.6
45.0
14.0

51.0
19.0
20.5

64.0
24.0

44.0
14.5

50.5
15.5
12.9

(1) NSR=ratio of NHi moles injected
divided by the Initial moles of NOx

(2) Weight ratio of Enhancer to Urea

TABLE 2

EFFECT OF BOOS ON TEMPERATURE AND GAS COMPOSITION

Conventional

Low Excess

Conventional

Excess Air

Air

Excess Air

Operation

100% Load

60% Load

Level 1

Avg. Tempf C

1,008

1,065

954

Avg. NOx mg/Nm^ (1)

460

321

362

Avg. CO, ppm

700

Avg. Oj, vol. %

5.2

2.2

5.7

Level 2

Avg. Temp?C

811

878

804

Avg. NOx mg/Nm (1)

513

328

347

Avg. CO, ppm

330

Avg. Oj, vol. %

6.3

3.1

5.5

Stack .

NOx, mg/Nm. (1)

584

403

427

CO, ppm

O2, vol. %

7.1

4.5

7.6

(1) Corrected to 6% ,

7A-62

-------

TABLE 3

EFFECT OF MODIFIED OPERATION ON NOxOUT PROCESS

Conventional

Low Excess

Conventional

Excess Air

Air

e xeess Air

Operation

100% Load

60% Load

Experiment

Baseline NOx

580

403

427

Level 1

NSR

2.4

3.0

E/U

0.14

0.03

0.5

Level 2

EWR (1)

0.33

0.13

0.42

% Reduction

% Utlization

NH3, ppm

4,5

3.5

5.0

(1) Weight ratio of enhancer to baseline NOx (as NC^).

C2 BOILER

TABLE 4

- TEMPERATURE AND GAS COMPOSITON

Load

100%

80%

Level

1(19 m)

2(25 m)

1(19 m)

2(25 m)

Temperature ,°C
NOx, mg/Nm
CO, ppm

O2, %

1040
250-400
500-10000
3.2

995
250-350
70-3000
4.1

992
150-400
10000-30000

1.2

980
150-400
250-10000
4.2

7A-63

-------
Fig. 1

K.W.E. D BOILER

Level 2 lil

Level 1 1161

|F Level 3 (4)

Fig. 2. R.W.E. C2 BOILER

7A-64

-------
Fig. 3. NOx REDUCTION AND UTILIZATION vs. NSR

R.W.E. UNIT D

100

BO -

70 -

«s

''"j
¦H

60 -

50 -

40 -

o ¦

30 -

20 -

Qr*

10 -

0 -

E/U Ratio = 0.1

E/U Ratio = 0.17

_o ^

0.0

i 1 r

0,3 0.6 O.B 1.1 1.4 1.7 2.0 2.3 3.5 2.B 3.1

NORMALIZED STOICHIOMETRIC RATIO (NSR)
E/U Ratio = Enhancer to Urea Weight Ratio
REDUCTION ^^ UTILIZATION ^ *

3.4

Fig. 4.

100
90 -\
80
70
60 H
50
40
30
20
10
0

AMMONIA vs. PERCENT NOx REDUCTION
R.W.E. - UNIT D

E/U Ratio = Enhancer to Urea Weight Ratio

Average E/U Ratio = 0.2

° Average E/U Ratio = 0.1

OOfiO

_—J ! ! , 1 , , | |

0.0 1.5 3.0 4.5 6.0 7.5 9.0 10.5 12.0 13.5 15.0

Ammonia (ppm)

7A-65

-------
Fig, 5,

¦I

«s

u
o

d
o
*3

a
3
¦d

S3
-------
Session 7B
FUNDAMENTAL COMBUSTION RESEARCH

Chairmap: W. Linak, EPA

7B-i

-------
FORMATION OF NITROUS OXIDE FROM NO AND S02
DURING SOLID FUEL COMBUSTION

S, G. de Soete
Institut Frangais du Petrole

1-4 Avenue de Bois-Preau
92506 Rueil-Malmaison Cedex
France

ABSTRACT

In recent months the formation of nitrous oxide as a product of reactions between
nitric oxide and sulfur dioxide or sulfite, has been highlighted in relation to
"artificial" N20 formation during storage of flue gas samples. Closer examination
shows, these reactions to bear a rather general character. The value of their over-
all rate constant as well as its temperature dependence also suggest, that they
may occur under typical heterogeneous combustion temperatures and residence times,
whenever sulfur compounds are present in the fuel, or whenever such sulfur species
are formed form e.g. lime stone, added for the sake of sulfur capture. The present
paper focusses on the formation of nitrous oxide due to the reactions of NO with
reduction products (CaSOR) of sulfates, in the temperature range of 1000 to T400
Kelvin. The reduction of sulfate in this range mainly occurs on bound carbon atoms
of the solid fuel. In the case of a coal char, the N20 formation from NO + CaSOR
interferes with N20 formation from char bound nitrogen. In order to avoid thi-s in-
terferes with N20 formation from char bound nitrogen. In order to avoid this inter-
ference, pure graphite has been- utilized as a fuel. Moreover, proper determination
of the rate of the reaction 2 NO + CaSOR—» N20 - CaS04, requires the knowledge
of (1) the reduction rates of the sulfate by bound carbon and by carbon monoxide
and (2) the concomitant heterogeneous reduction rates of N20 on bound carbon atoms;
these rates therefore have been determined separately in the same experimental
apparatus. The rate constant of reaction 12, as determined in the 1100-1400 K range,
shows to be identical (within the limits of precision) with the one formerly ob-
tained under ambient temperature conditions. Further study is in progress to check
(1) the role of specific sulfur compounds (sulfites, sulfides, S02) and (2) the
influence of oxidizing conditions on this type of N20 formation.

INTRODUCTION

Nitrous oxide is released as a reduction product of NO or NQ2 during the reactions

1 2

of the. latter with S02, either in the gas phase or in ionic solutions ( ' ). AT
ambient temperature, these1 reactions are' relatively.slow,, however fast enough, to
jeopardize seriously nitrous oxide analyses made on-flue gas samples, when stored
in containers for some hours.

In the, temperature range of 0 to 70 °C, an approximate value of the overall N20
formation rate has been obtained ( }, which, for example in the case of gas phase
formation of N20 from NO and. S02 in the absence of oxygen, may be expressed as ;

7B-1

-------
VN20~ 5 10"] exp(_1500 to 2000/T) P^Q (ngSQ2/citi3) (ng/h)

Boldly extrapolating this rate towards typical coal combustion conditions of tem-
perature, concentrations and residence times, it is suggested that reactions of that

kind might play a role in the N20 emissions from solid fuel combustors (especially of
the fluidized bed type), where they are likely to interfere with other N02 produ-
cing, homogeneous and heterogeneous mechanismes.

Scope of this study is to check the possible N20 formation due to heterogeneous
reactions of NO with sulfate decomposition and reduction products (such like sul-
fites, sulfides, S02, H2S). Some of these reduction products may be naturally pre-
sent in the coal or char; others will form from sulfates obtained during sulfur
capture techniques. The study aiming at the establishment of overall rate constants,
it is important to avoid any interference with N20 formation from fuel bound nitro-
gen atoms occuring during combustion, and to correct the global N20 yields for the,
unavoidable, heterogeneous reduction of the latter on fuel bound carbon atoms. This
is done by utilizing pure graphite as a fuel and by determining separately, on the
same apparatus, the heterogeneous reduction rates of N20 on graphitic carbon. Fur-
ther is it required to know the overall reduction rate of the sufate, i.e. the pro-
duction rate of sulfate reduction products due to the presence of reducing agents
such like bound carbon or CO; these reduction rates also have been determined sep-
arately on the same apparatus, in the absence of NO, and counterchecked in the pre-
sence of NO.

EXPERIMENTAL METHOD

A fixed bed method has been utilized. The reactor is a quartz tube of 11 mm i.d.,
placed vertically in an electrically heated oven. The solid phase, constituting a
bed of 8 mm height, is a mechanical mixture of high purity graphite (99.999 % purity,
25 microns size) and CaSQ4 (of 5 to 10 microns particle size); this solid phase,
situated in the center of the electrical oven, supported by a sintered quartz
plate, is further diluted by non porous, 100 microns sized quartz beads working as
a thermal buffer. The gas phase, flowing downward through the bed, is either NO +
argon, or N20+argon (for the study of heterogeneous reduction of nitrous oxide en
graphitic bound carbon; for these trials there is no sulfate mixed with the graph-
ite), or pure argon (for the study of CaS04 reduction by graphite), or CO+argon (for
the determination of CaS04 reduction rate by carbon monoxide, in the absence of gra-
phite).

A Nickelchrome/Nickel thermocouple, placed in the center of the bed, monitors the
instantaneous temperature of it. Outlet gases are continuously analysed for CO,
C02, NO and 02 (recording frequency = 0.5 Hertz), and di scontinuously for N20

7B-2

-------
(on-line gas chromatograph equipped with a back-flushed porapak column in series
with a molecular sieve column and an EC detector; carrier gas : 5% methane in argon;
detection level : less than 0.02 ppm N20; sampling and analysis every 4 minutes).-
All analyser and temperature outputs are recorded and treated by a microcomputer..
The reactor temperature is either kept constant or increased progressively using
a dT/d.t value of about 6 degree/minute, which is slow enough to ensure adsorption
desorption equilibrium.

EXPERIMENTAL OBSERVATIONS

We succesively consider (1 )¦ the phenomenon of the reduction of sulphate into its
reduction products (CaSO^) by either CO or bound carbon, and (2) the reduction of
NO into N20 on these sulfate reduction products.

Reduction of CaSQ4 by Carbon Monoxide and Bound Carbon

Reduction of Sulfate by CO. Upon heating sulfate in the absence of any reducing
agent, it decomposes into the oxide and S02 ;

CaS04-Uca0 + S02 + 1/2 02

For calcium sulfate, this decomposition is negligible below 1400 K, and therefore
the overall reaction I schould not be taken into account in the present study.

It is known form the literature ( ), that alkaline earth sulfites, such as CaS03»
also thermally decompose in the oxyde and S02, whereas alkali sulfites undergo auto-
oxireduction, according to the overall reactions :

CaSQ3 -i-Ca0 + S02
4 Na2S03.-^3 NaS04 + Na2S

In the presence of CO as a reductant, calciumsulfate is reduced already at 1000 K,
under formation of C02 and a, provisoriously not defined reduction product (CaSO^)

CaS04 + CO —CaS0R - C02

where CaSO^ designates a whatsoever sulfate reduction product stripped by one oxygen
atom with respect to sulfate; its concentration, (CaSO^) is expressed "per oxygen
atom stripped" and therefore is given by the sum of (4-n) times the concentration
of each particular CaSO^ species, e.g. ;

(CaS0R) =0 (CaS04) + (CaS03) - 2.{CaS02) + 4 (CaS) (1)

Figure 1 shows an example of reduction of CaSQ4 by CO; it clearly indicates the
oxygen/carbon balance of reaction 11 :

7B-3

-------
CaSOD formed = CO. - COni, = C02„, .
R in out out

(2)

Reduction of Sulfate by Graphite. The reduction of CaSQ4 by graphite bound carbon
yields CaS and mainly CQ2, some smaller quantities of CO being observed also. Desig-
nating by "a" the ratio of C0/CO2 in the outlet gases, one may write the overall
reaction as' ;¦

(2+a) CaSQ4 + (1+a) —^*!2-a) CaS0R + C02 + a CO

Figure 2 shows typical CQ/T and C02/T profiles measured during thermal treatment
under argon flow of an anhydrous CaS04/graphite mixture. At temperatures between
779 and 940 K, a first phase of partial sulfate reduction occurs, as shown by the
Fq/T curve. FQ designates the fraction of oxygen atoms removed from the sulfate mol-
ecule. During this first, L(ow) T(emperature) phase of reduction, only C02 appears
as a carbon oxidation product. At temperatures higher than 1130 K, a second, H(igher)
T(emperature) phase of reduction starts, releasing both CO and C02. In the trial
presented on figure 2, the final' temperature has been maintained for about one more
hour, until CO and C02 outlet concentrations had dropped to zero. At that final in-
stant, the fraction of carbon consumed (Fj.) had reached a value of about 0.38 and
the fraction Fq a value of about 0.43. The reduction of calcium sulfate by graphitic
bound carbon thus presents at least two successive and separate steps : a IT step,
yielding only C02, and a HT step, starting beyond 1130 K, yielding mainly C02 and
some CO and removing a much 1arger fraction of oxygen atoms form the sulfate.

In neither of these two reduction steps does the reeaction involve the total mass
of sulfate present. This is mainly due to the fact that the mechanical mixture of
graphite and sulfate particles does not provide an ideal reaction interface.

Reduction of NO into N20 by Sulfate Reduction Products

As an example, figure 3 shows,'the evolution of NO, N20, CO, C02, Fc and FQ with tem-
perature in a batch trial where NO is fed over a calcium sulfate/graphite mixture.
As far as the nitrogen species are concerned, one observes the following :

(a) At temperatures below about 1230 K, a slow decrease of NO and N20 occurs, which
is accounted for by the heterogeneous reduction of these two species on bound
carbon atoms, as may be seen from the comparison of the corresponding NO/T and
N20/T profiles obtained on pure graphite (i.e. in the absence of sulfate) with
identical NO/Ar or N2G/Ar mixtures (curves B).

The reduction of NO on graphite mainly produces N2 and CO, which may be ac-

,4,

counted for by the reactions ( ) :

NO + 2 (-C) -ip-(-CO) + (-C..-N)

7B-4

-------
(-CO) —*-CQ + free carbon site

and corresponds to the experimental overall mass balance ( 3 :

NO. - NO . = CO , (3)

in out out

whereas, in the reduction of N20 on graphite, tne main carbon oxidation prod-

uct is C02, according to the reactions ( 3 :

N20 + (-C) -1— N2 + (-CO)

N20 + (-CO) N2 + C02 .+ free carbon site
or : 2(-CO) —E-C02 + (-C) + free carbon site

corresponding to the mass balance ;

N20• - N20 . - 2 C02 (4)

in out out v '

(b) However, at temperatures beyond 1240 K, the NQ/T and N20/T evolution profiles,
measured in the presence of sulfate (curves A on figure 3), show a quite dif-
ferent shape than in the case of pure graphite (curves B).

The fast decrease of nitric oxide apparently cannot be accounted for by its
sole heterogeneous reduction on bound carbon. The N20 outlet concentration
starts increasing (instead of continuing to decrease as in the case where no
sulfate is present), then passes through a maximum before decreasing again,
suggesting that, in the given temperature range, it is the result of two com-
peting reactions : (1) a formation reaction, in which sulfate reduction prod-
ucts thus appear to play an important role and wnich we shall assume to cor-
respond to the overall reaction :

2 NO + CaS0RJL~ N20 + CaS04

and (2) a destruction reaction, most likely the already mentioned heterogeneous
reduction on bound carbon (reaction 73.

DETERMINATION OF THE RATE CONSTANTS

The rate constants k^, and have been determined from the experimental val-
ues of the time variations of N20 and CaS0R, respectively measured as ;

dN20 = VN20 = 2MND(XN20^out-XN20>tn)/vM (qHZQ N/s) (5)

dCaSOR = VCg0s =
dt R

'mCsulf't " ^Osulf't-At]''9CaS0R as 0/s)

where X is mole fraction, the atomic mass of N, molar volume at TPN and' D
the gas flow rate at TPN; nig ^ is the mass of oxygen removed from the sulfate,

7B-5

-------
obtained as :

^'Osulf M0°f '2&XC02 + ^XC0 " AXN0 " A XN2o'dt/vM ^

J 0

where Mq is the atomic ;nass of oxygen; Astands for in - Xj cut. The fraction
of oxygen atoms removed from the sulfate (Fg) is related to mQsu^f by the expres-
si on :

Fn = mn 1X M 1x/4Mn(m -,r) • (8)

0 Qsulf sulf u sulf i

where M ,, is the molar mass of CaS04 and (m ,,). the initial mass of sulfate

SUIT SU 1T 1

present in the reactor.

Determination of Constant

When NO is fed over a mixture of graphite and CaS04, one has :

v V _ V - V = V -V (with V' = V - V } (9)

N20 12 7 8 12 7 ^ 1 7 7 8; K '

One therefore has to determine the rate prior to abtain and k-^. Determi-
nation of the heterogeneous N20 reduction rate on bound carbon is made in the same
apparatus, feeding N20/argon mixtures over graphite. Reaction rate V'7 is found to

5 '

be first order in graphite mass and in partial N20 pressure ( ) :

V7 = k7 mC ^N20 %20 as N/s'

where mc is the instantaneous value of the mass of carbon present, equal to its
initial mass (m^) when multiplied by :

Fr - f \ (Xrn . + Xrn„ . )dt D M /m .v., (11)

C Jo CO,out C02,cut c ci M

and PN2q the mean value of the partial N2G pressure between reactor inlet and

outlet :

PN20 = P(XN20;in + XN20,out5/2 5125

\ 5

The rate constant has been determined by integral treatment of tne reactor ( )

k7 = 2V ln*XN2Q.in/'XN20,aut^f> VM + 8,0 exp{-16400/T) (9N/9C sPa) (13)
.V-j^ may now be determined from eq(9) as the sum t VN2q + v'^), and is found to be
approximately first order in the partial NO pressure and in CaS0R. It therefore
may be expressed as :

with

12 = k12 mSR ^NO (gN20 as N/s) (14)

^ = P (XN0,in + XN0,out)/2 {15)

7B-6

-------
P indicating total pressure. is the instantaneous value of the mass of sulfate
(expressed as mass of S) reduced into CaSO^ :

mSR = m0su 1 f VM0 4 F0 !Vmsu.lfVMsulf (16)

The value of k^, as obtained with CaS04, is given by (see figure 4) ;

k12 = 5.0 10"7 exp(-2000(+ 500)/T) (gN2Q N/gss Pa) (17)

This constant should be compared with that obtained at room temperature for the
N20 formation from NO and S(
the latter may be written as

N20 formation from NO and S02 ( ) ; when expressed in the same units as eq. (17)

k]2 = 1 to 4.8 10"7 exp(-1500 to -2000/T) (18)

The surprising similarity of these two overall rate constants as determined in the
1240-1350 Kelvin range and in the 273-340 Kelvin range, seems to indicate that
an analogous mechanism of N2Q formation is at work in both cases (see figure 5).

Determination of Rate Constant k-^

Feeding CO- over CaS34, the reduction rate of the latter is equal to the experi-
mental value of VCaSR, which in this particular case is equal to the CO consumption
rate :

V11 = VCaS0R = M0D(XC0,in " XC0,out)/vM (g0/s) (19)

Rate V-|-j is found experimentally to be first order in partial CO pressure and in
the mass of oxygen "available for reaction" in the sulfate. By "available for re-
action" one should understand that fraction of oxygen which actually can be strip-
ped from the sulfate, owing to the fact that, due to the non ideal solid/gas inter-
face, not all the oxygen atoms are able to react. This "available fraction" will
be designated by qav. which is the final value of obtained at the aid of a batch
trial,,when the CS2ou^. concentration has dropped to zero again and the CO ^ con-
centration has become again equal to C0^n (see figure 1). Rate constant k^ may
thus be determined from the expression. :

V11 = kll m0av^C0
where P^j is the mean partial pressure of CO in the reactor :

CO ~ p(XC0,in + XC0,out)/2 (21

The value of m0av may be calculated as a function of fq :

where

™0av =,H0(%lf»i(F0av 1221

lF0,at.ttF0,atUit"2 1231

7B-7

-------
According to eqs (19) arid (20), k^ is thus obtained in the units Pa_1s ^ . Values
of k-|-| determined in that way as a function of temperature are given on figure 4.
Its expression corresponds to :

kn = 2.8 105 exp(-33500/T) EPa"1s"1) (24)

Determination of rate constant

Rate constant k-^ has been determined both in the presence and in the absence of
NO, both methods yielding the same value.

In the Absence of NO. In the LT reduction phase, there is no CO production;
therefore is directly obtained as :

V10 = W - 2 V>(XC02,oUt]/vM (251

In a first approximation V^q shows to be first order with respect to both mass of
graphite and mass of sulfate; it therefore may be written as :

V1Q = k10m0avmC (90/s3 f26!

For LT reduction of Calcium sulfate, the following value of k^ has been determined

k 10.LT = 4-7 1C)6 exPf~1S200/T) fgc~V]) (27)

During the HT reduction phase, some CO is produced; therefore is to be deter-
mined as :

V10 = VCaS0R ~ V11 = k10m0avmC (g0/s) ' (28!

One obtains for in the HT reduction phase :

k1Q HT = 5.4 1011 exp(-40 000/T) (gc_1s"T 5 (29}

Values of k|g are compared to those of as a function of temperature on figure
4, k^ being given by curve A .

In the Presence of NO. For the determination of in the presence of NO, the ex-
perimental value of should be corrected for the existence of and V^ :

V10 = VCaS0D + M0 V12/2MN " V11 (gQ/s) (30!

The value determined for k^Q is almost identical to that obtained in the absence
of NO (see figure 4, curve B).

Comparison of eqs (27) and (29) suggests that, at LT one is dealing with a different
reduction type than at HT; for example : reduction of sulfate into sulfite at LT,
and reduction of sulfate into sulfide at HT. More work is in progress to elucidate

7B-8

-------
this question, using sulfites arid sulfides mixed with graphite.

CONCLUSIONS

1. Experimental evidence has been obtained of the existence of nitrous oxide for-
mation from NO and sulfur compounds under combustion temperature conditions.
A striking parallelism has been found with analogous reactions, recently stud-
ied at room temperature, showing an almost identical rate constant. The simi-
larity of the subjacent mechanism is obvious [see figure 6) :

(a) In the ambient temperature range, and more specially in the ionic state,
sulfur compounds may be either oxidized or reduced by NO. the Tatter play-
ing its double oxidant/reductant role (figure 6, B). .Oxidizing sulfite ions
or S02 into sulfates or S03, nitric oxide is reduced into N20, and NQ2 is

reduced into NO. Reducing in- turn the sulfate ion, nitric oxide oxidizes

¦ 2

into N02, thus completing tne cyclic mechanism ( ).

(b) In the combustion temperature ranges, contact times of NO with sulfate be-
ing very short, compared to usual room temperature situation, the sulfate
reduction step is more efficiently played by bound carbon-atoms or CO (fig-
ure 6, A). In the typical case of .calcium sulfate this reduction step re-
quires temperatures beyond 1200 K. The possibility of this reduction under
oxidizing conditions is now been checked (in progress),

i: A more practical question is the importance of N20 formation by this way. as
compared to other N20 production mechanisms at work during solid fuel combus-
tion, e.g. N20 formation from bound nitrogen atoms..Figure 7 shows a com-
parison, at different temperatures and concentrations of sulfur - N20 and
oxygen, between the overall reaction rates of N20 formation, either from N0+
sulfur species (V-^) or from bound nitrogen (V^) : it is clear from these data,
that Vl2/VN will become larger than one for either sufficiently high reduced
sulfur compound concentrations, or sufficiently low oxygen contents, or at suf-
ficiently low temperatures. This temperature condition situates reaction 12
as. a good candidate for N20 formation under fluisized bed conditions, the more
since other sulfates (like alkali sulfates) may be reduced at lower temperature
than calcium sulfate.

ACKNOWLEDGMENTS

The author wishes to thank Joseph McSorley and William Linak, from the US E.P.A.,
and Jost Wendt, from the University of Arizona, for their encouragement and con-
tinuous interest in the present research work.

7B-9

-------
REFERENCES

1. L.J., Muzio et alii, "Potential errors in grab sample measurements of N2Q from
combustion sources", 1988 Fall Meeting of the Western States Section of the Com-
bustion Institute, Dana Point, Ca, October 17-18, 1988, paper n° 88-70.

2. G. G,, de Soete, "Parametric study of N2Q formation from sulfur oxides and ni-
tric oxide during storage of flue gas samples", Report IFF N° 36 732, December
1988. (available on request).

3. J., Verhulst, J. C., Jungers, and A., Bruylants, "Algemene Scheikunde, Part ;
1 : Anorganische Chemie", Desoer Edit. Liege, 1953.

4. G. G., de Soete, EPA' Project Decade Monograph on Coal Combustion, Chapter 8,
"Reduction of nitric oxyde by solid particles" (to be published). .

5. G. G., de Soete, "Heterogeneous NO and N2Q formation from bound nitrogen during
char combustion", Joint Meeting of the British and French Sections of the Com-
bustion Institute, April 18-21 1989, Rouen, France.

Idem, "formation rieterogene de protoxyde d'azote a partir de 1'azote constitutif
de combustibles s solides", Report IFP, N° 36 752, March 1989. (available on
request).

1000 1200 1400

Figure 1. Reduction of CaSQ4 by CO,
954 pprn CO in argori on 500 mg CaS04

7B-10

-------
figure 2, Reduction of CaS04 by graphite, under
argon flow. 100 mg graphite + 5C0 rig CaS04

Figure 3. Formation of N20 from NO .and CaSOR. 1240 pprn
NO + 5 ppm N02 in argon on 0,1 g graphite + 0.5 g CaS04

7B-11

-------
Figure 4. Overall rate constants of CaSQ4 reduction by CO (K^) and by
graphite (k1Q), the latter determined in the absence (A) or in the pres-
ence (B) of NO. Overall rate constant of N20 foramt.ion from NO and CaSOR

Figure 5, Comparison of as determined at room
temperature range and in the 1220-1350 Kelvin range

7B-12

-------
A

Figure 6. Schematical comparison of the overall mechanism of N2Q formation from
NO and sulfur compounds; A : with (-C) as a reductant; 8 : with NO as a reductant

103/ T[K]——

Figure 7. Comparison of ana VN (from reference S); Cnar from low
volatite bituminous coal "Prosper"; P=Q. 1 rtPa; mj. = 1 g; 1000 ppm NO.

78-13

-------
FUEL NITROGEN MECHANISMS GOVERNING N0X ABATEMENT FOR LOW AND HIGH RANK COALS

Jose O.L. Vendt arid A run C. Bose,
Department of Chemical Engineering,
University of Arizona,
Tucson, A2 85721, USA

and

Klaus R.G. Hein,

RWE AG.,

Poscfach 1461,
5010 Bergheim, FRG.

ABSTRACT

Fuel nitrogen mechanisms governing the efficacy of staged combustion were
investigated for a low rank German Brown Coal and for a high rank'Utah Bituminous
Coal, burned in a downflow 2 kg/h laboratory combustor. Emphasis was on
determination of the effects of coal rank on the rates of destruction of
nitrogenous species in the fuel rich stags and on the subsequent conversion of
these species in Che burnout stage. Special experiments involving doped propane gas
as a fuel allowed isolation of the dominant: mechanisms in the coal runs.

This research revealed that reactions destroying NO and HCN were completely
controlled by known gas phase reactions, and not by heterogeneous reactions, as has
been hypothesized heretofore. The overall reaction is first order in NO and in MH3,
Furthermore, the late evolution of HCN into the bulk gas phase from both coals, ac
long time scales, was important in driving the gas phase reactions destroying NO
and other nitrogenous species.

INTRODUCTION

Although staged combustion is a mature technology, its efficacy and performance are
difficult to predict. Coal composition effects have been shown to be important, and
several mechanisms describing the destruction of nitrogenous species have been
proposed. Glass and Wendt ^ provided data on a single coal and showed that the
destruction of NO at stoichiometric ratios (SI) of 0.8 and 0.4 could be interpreted
using a simple homogeneous gas phase mechanism (based on Fenimore^-3) involving
reaction of NO with NHj together with partial equilibrium approximations for the
NH^ species and a global equilibrium assumption for OH. Mora recent work 4,5 has
indicated that these conclusions may not be generally valid, and that heterogeneous
reactions between NO and char may be important under coal combustion conditions,
and that these depend on coal composition. The situation is further complicated by
the fact that NO can be reduced on sooc, and this has been suggested ® as an
additional mechanism to be considered, especially under very fuel rich conditions.

The purpose of the research reported here was to help resolve some of the

7B-14

-------
Secondary Air

Sample/Utility J 55
Ports

Reproduced from
best avallnblo copy.

Exhaust

Figure 1. Experimental Combustor.

Jton aituffliMul 4>2 Rmb» 1,9,0

Figure 2. Temporal Profiles of Temperature and Nitrogenous
Species in the First and. Second Stages of a Staged Combustion
Configuration. Utah Bituminous Coal.

7B-21 I 7B-15 THROUGH 7B-20 ARE BLANK

-------
sa»a,a sf'os

Residence Time, s Residence Time, s

Figure 3. Base Case Profiles, Fuel Rich Scage: German Brown
Coal.

j*
m

S:.!" h,j2o

V- fhMZO

^ «¦ »

I-*- H,

^V,,, n ,

¦ Aw-f* MCNilO

r **"" 1 ' Wjiio

V"*NQ

0 I 2
Residence Time, s

"0 i 2
Residence Time, s

Figure 4. Base Case Profiles, Fuel Rich Stage: Utah Bicuxainous
Coal.

7B-22

-------
taoc >

£ 1400

iOOO,i2 p

<0 I-

0,06

* 04

| 02
GO. 100
250
SO
I 150

GS>

100
30
0

*un iO. SP«0

c.«

** 0 4

z a#

0,2
00

- / V * -5
/

* a .

/,*" v?

• % ¦-,* •,l

* *

«r ^ mil $— «»«!—1— l—l »¦ <»«»,»»»

Residence Tim#, l

Figure 6. Effect: o£ Temperature on Che Speciacion and
Distribution of Nitrogenous Species: German Srown Coal, SR-0,6

78-23

-------
Figure 7.Effect of Temperature on Speciation and Distribution
of Nitrogenous Species: Utah Bituminous Coal, SR-0,6,

* *

9 * Sff'O* Mm**
9 t S*>0«3
4 ft 1**0 y 3| {nwp
~ • 1I<9U —

*, ¦<«(*

a g Mtosi

**i immsn *m i
9 t V'dilN,

Oar*'*

« ±4* ^

;/%s*

Cmf ¦

fc** A

wrxo4!*-')

Figure 8, Comparison between Measured and Predicted NO
Destruction Mechanisms. Utah Bituminous #2 and Doped Propane
Gas Runs.

78-24

-------
/f« *f (.-I

Figure 9. Comparison between Measured and Predicted NO
Destruction Mechanisms. Seven Different Coals.

090*3 Pressor* Gas Run SR« 077

SOOOr Ttmmttihif!

-•wo

E
£

1200

1000

100

*00

200

2.0

-O- M«CK^«
-------
Germcn Brown Coal, Base Run SR = 0.64
JLQOOr Tgmpgrgsrufg Profit*

iQOG

aco

soo

400

2.0

NQ ,°re*>g

-O- SMC*U/*C
a frtflictM Empirical CM

A Prisierid EmswicoJOHi
(Noft^MKjr ReqriMienl
<> rrtiSicfM OH
(KuifhcJ

io io

R«id«net Tims, l

Figure 11. Measured and Predicted HO Decay Profiles: German
Brown Coal, SR-0.64,

German Brown Coal, 6as» Run Sft=O.03

fiesictnce Time, s

Figure 12. Measured and Predicted No Decay Profiles: German
Brown Goal, SR-0,83.

78-26

-------
Ufoh Bituminous Coal *2, Base Run SR = Q.6i

Residence Tim#, s

^ipai 4,2k. GaeeflwrWe '*a» Rjuarftnuai ON Moduli
MmM um pMiOBi MO vaktm.

Figure 13. Measured and Predicted NO Decay Profiles: Utah
Bituminous Coal, SR-0.61,

600

n
«

1 soo

I *00

s
i

| 300
€

& 200

100

Utah 8i>Mfw«wuf »2
O $S«Q*i e4MAmt
0 y"OG W, OduMan
+ 9| CnrcMmvt

~ sa-osa k, Q;m*»

# SR« 0,§2 ftaett«aMt dot* fa*
B Sfi»0fi2 Bom ftwi

Ocrman 3rown

o Sfl»Ofi3 ftsniftvn
0 §R«0.*« ftcrt«*un
A SH»0ii £nww*tt
~ Sft*QSfl Nt OJwfton. HrqfiUx*
9 S»»C.54 N, 3iUi**», U»UKKI

** °

0 (GO 200 300 «00 500 600
NO, JJpm -Pr*fisc?M by Nan-£»i3libfS!jjn Modct

Figure 14. Measurements versus Predictions: All Cases.

7B-27

-------
Table 1: Pulverized Coal Compositions

Proximate Analysis

Volatile Matter Fixed Carbon

Ash

Utah Bituminous #1

39.2

47.4

4.5

8.9

Utah Bituminous #2

46.6

43.1

1.8

9,3

Western Kentucky Bituminous

41.8

46.9

3,0

8,6

Texas Lignite

34.5

30.2

14.0

21.3

Beulah Lignite Low Na. #1

38.3

35.3

13,1

13.3

Beulah Lignite Low Na. #2

37.1

37,1

17,8

8.0

Beulah Lignite High Na.

36.3

33,9

22.5

7.3

RVE German Brown Coal

48.9

13.4

28.8

8.8

RWE German Brown Coal

38.8

30.7

15.3

15.2

Dried (from 74.4% moisture)

Ultimate Analysis

Oxygen

(Dry. with Ash)

Carbon

Hvdrozen

Nitroeen

Sulfur fDifferent

Utah Bituminous #1

70.87

5.11

1.64

1.03

12.08

Utah Bituminous #2

71.59

5.19

1.29

0,45

12.66*

Western Kentucky Bituminous

73.22

4.55

0.82

3.00

12.53

Texas Lignite

54.60

3.60

0. 85

3.39

16,21

Beulah Lignite Low Na. #1

62.34

3.07

0.92

3.48

16. 90

Beulah Lignite Low Na. #2

57.79

3.95

0.91

2.07

25.57

Beulah Lignite High Na,

59.80

3.90

1.08

1.02

24.78

RUE German Brown Coal

62.88

4.16

0.87

0.65

20.92*

RWE German Brown Coal

48.52

3.37

0.72

0.78

21.83*

Dried (from 74.4% moisture)

*Analytically determined

7B-28

-------
METHANOL INJECTION - A NEW METHOD FOR NQ,< AND SOa CONTROL

Richard K. Lyon
Energy and Environmental Research Corp
18 Mason Dr.

Irvine, CA, 92718

ABSTRACT

A theoretically predicted new homogeneous gas phase reaction has been
demonstrated experimentally both on the laboratory scale and in pilot plant
experiments. In this reaction the oxidation of trace quantities of methanol
provides a rapid and nearly quantitative conversion of NO to N0S at
temperatures greater than ?00°C and a rapid, nearly quantitative reduction of
SOa to SOe. This reaction, occurs under conditions which make its use in
boilers and furnaces entirely practical and hence it provides the basis of a
new process, the Methanol Injection process. Since the scrubbers now used to
control S0e emissions can also efficiently remove N0e, this new process may
allow wet scrubbers to control both S0E and NQx emissions while protecting them
from corrosion. Wet scrubbers necessarily operate at a temperature below the
acid dew point. Thus when flue gas containing S03 enters the scrubber a fine
mist of sulfuric acid is formed. Since the scrubber will typically only remove
50X of this acid mist, both the scrubber and the equipment downstream of it are
vulnerable to corrosion.. Since the Methanol Injection process selectively
reduces S03 to S0B, it can eliminate this problem.

The combination of Methanol Injection with either the Thermal DeNOx or Urea
Injection processes appears particularly attractive, since the combined
processes could achieve a high, degree of NQx control,and Methanol' Injection
would eliminate the NH<,HSQ& fouling problems which presently hamper application
of Thermal DeNOx and Urea injection.

INTRODUCTION

This paper describes a new technology that controls emissions of both NOx and
SOa. As has been discussed by Faucett, Maxwell, and Burnett1, there have been
a number of efforts to develop processes in which NO was oxidized to NDa, with
the N0e then being removed by a scrubber which also removed the SQe. These
efforts, however, all involved doing the NO to N0e conversion either with 03 or
C10e and the expense of these reagents tended to make the overall process
unattractive. More recently, Murakami and Izumi3 have suggested using methanol
oxidation to do the NO to N0a conversion. While this is potentially a much
more attractive method of converting NO, there is the problem that both the
examples and the claims of Murakami and Izumi are limited to temperatures below
7QO°C, i.e. limited to temperatures low enough so that N0e is not
thermodynamically unstable with respect to decomposition to NO. Corresponding
to this limitation to low temperature, the reaction times they show are in the
range of 1.5 to 5 seconds. Such long reaction times are not typically
available in most practical combustion systems.

7B-29

-------
While regulatory pressures to decrease NO* emissions are a relatively new
event, the problems that the presence of SC3 in flue gas can cause are of long
standing. The combustion of sulfur containing fuels produces a flue gas in
which typically 9B-99'/. of the sulfur oxides are S0e with 1-2'/. present as 30a.
While S03 is thus trivial in terms of total SCx emissions, it still represents
an important problem in terms of boiler and furnace operation. On cooling the
SO3 in the flue gas forms a sulfuric acid mist which can cause corrosion
problems or, if the flue gas contains NH3 from selective NO reduction, can form
NH„HSOfc with resultant fouling problems. The former can be an especially
severe problem in the operation of wet scrubbers. When S0e is removed from
flue gas in a wet scrubber the gas temperature is necessarily below the acid
dew point. On entering the scrubber the flue gas forms an acid mist. The mist
particles, having an extremely low diffusion coefficient are not removed
efficiently by the scrubber and tend to pass through it, causing corrosion
problems downstream of the scrubber.

While this represents a significant problem for SOx control technology, NHoHSQ*
formation is a problem for postcambustian NOx emission control. Several such
processes are available, Thermal DeNOx, Urea Injection and Selective Catalytic
Reduction, and all involve reducing NO in a manner that leaves traces of NH3 in
the flue gas. This can lead to NH«,HS0^ formation downstream of the NO
reduction zone with resultant fouling problems.

In this paper we report two discoveries: firstly that at temperatures greater
than 700°C methanol oxidation can provide virtually complete NO to NOs
conversion in reaction times as short as 0.05 seconds and secondly that while
the methanol oxidizes the NO to N0e, a second reaction occurs, a new
homogeneous gas phase reaction in which methanol selectively reduces S0a to
S0e. Both these rapid reactions are highly selective in that ppm
concentrations of S03 and NO are converted by quantities of methanol only
slightly greater than stoichiometric in the presence of massive amounts of De.
Prior to doing experiments the existence of this selective S0a reduction was
predicted by computer modeling (3) and subsequent experiments were in
reasonably good accord with the model's predictions.

These discoveries provide the basis for a new process for controlling both NOx
and S03 and in this paper results for this process from modeling calculations,
laboratory experiments and pilot plant tests will be reported. The potential
advantages and limitations of this new process will also be discussed.

EXPERIMENTAL

The bulk of our experiments were done using the flow system shown in Figure 1.
This flow systems consists of three subsystems: a system for blending fixed
gases and vaporized liquids, a reactor, and an analytical train.

The gas blending system is a set of rotameters using the dilution blending
method which allows preparing flowing gas mixtures of several components, the
amount1 of each component being accurately known and readily varied in the range
of several percent, down to a few ppm. Controlled amounts of CHaOH vapor were
added to the flowing gas mixture by passing Ne at a measured flow rate through
a saturator filled with methanol and kept at a controlled temperature (usually -
16°C) and mixing this methanol containing flow with the main flow.

7B-30

-------
As shown in Figure 1, the addition of Ha0 and SOsi to the flowing gas mixture is
done just-upstream of the reactor. A dilute aqueous solution of HeSCL, is
pumped' via a precision metering pump to a vaporizer. The rate of pumping is
chosen so as to provide the desired value of CHe01 in the-gas mixture. On
vaporization and heating above 300°C, HaSOtt is converted to He0 and SO* and the
concentration of HeSCL. in the solution is chosen so as to provide the desired
value of CS0a3 in the gas mixture going into the reactor. The construction of
this system is such that the acid and its vapors never see any materials other
than quartz and Teflon.

Once the appropriate gas mixture is prepared, it goes into the reactor, a
length of quartz tubing inside a three zone electric furnace. This quartz
tube reactor has three sections, an input leg of 1 mm ID, a center section of
10 mm ID and an outlet leg of 1 mm ID. The lengths of these three sections are
chosen in such a manner that the input and outlet legs are in the portion of
the electric furnace in which there is a temperature gradient and the center
section is in the portion of the furnace in which the temperature is uniform.
The' reactor thus provides an accurately defined time-at-temperature.

After leaving the reactor, the flowing gas mixture is analyzed. NO, CO, SOa
and 0e were measured, respectively by a Thermoelectron Chemiluminescent NOx
meter, a Thermoelectron Gas Filter Correlation CO analyzer, a Thermoelectron
Pulsed Fluorescence SD2 analyzer, and a Teledyne 02 analyzer. The first three
of these instruments were calibrated daily with calibration gas mixtures
purchased from Matheson, while laboratory air was used to calibrate the fourth.

Three different procedures were used to determine CS03] in the post reaction
gases. In procedure A a sample trap is used. The inlet end of this trap is
inside the three zone electric furnace so that the gases entering the trap are
at a temperature well above their dew point. The body of the trap was immersed
in boiling water. As discussed by Uendt et al.4* S03 may be quantitatively
recovered from flue gas by condensation as HeSO*. at 90 to 1QQCC. By using the
three way valve the entire gas flow was passed through this hot trap for a
predetermined time, after which all flows were stopped and the amount of HaSCU
collected was measured by the procedure of Fritz and Yamamura <51, i.e., by
titration with an BQ*/« alcohol /SO'/, water solution of barium perchlorate with
thorin indicator. This sampling procedure has the advantage of eliminating,
several potential sources of error, i.e. the fraction of S03 remaining in the
post reaction gas is simply the' amount of HaS0« collected in the trap divided
by. product of the rate at which the precision metering pump inputs acid and the
collection time.

In procedures i and C the extent to which the input 50* was converted to S0s
was determined for reaction mixtures in which S0e was not initially present by
measuring the amount of SQe produced by the reaction with the Pulsed
Florescence S0e meter. In procedure, B, then, the conversion of the S03 was
calculated as the ratio of S0B output measured by the S0e meter to S03 input
calculated from the known rate at which the metering pump delivered sulfuric
acid to the boiler.

In procedure C a converter was used, i.e. the setup shown in Figure 1 was
changed so that on exiting the reactor the flowing gas mixture could be sent
either to the S0B meter as shown in Figure 1, or sent via heated lines to a
converter, a tube packed with platinum catalyst and heated to 1000°C, and then

7B-31

-------
sent to the SOS meter. In this manner it was possible to measure both CS0e3 in
the post reaction gases and the sum of CS0e3 and [S03]. Also in procedure C
the metering pump and boiler shown in Figure. 1 were replaced.with a unit that
allowed the flowing gas mixture to pass over concentrated HaSOc» at elevated
temperatures. By controlling the temperature the

amount of acid vaporized into the gas mixture could be controlled. This
allowed doing experiments with.gas mixtures that contain no water vapor other
than that which was formed by reaction.

As discussed in detail below, computer modeling studies showed that when
methanol oxidizes in mixtures containing NO and S03 ,the NO is converted to N02
while the S03 is reduced to SOe, and that the optimum conditions for the two
reactions tended to be the same. Since the measurement of NO to N02 conversion
is much easier than S03 to SO* conversion, this fact was useful in choosing
experimental conditions for precise measurement of S03 reduction.

In addition to the above described laboratory scale experiments done with
synthetic gas mixtures, a limited set of experiments were done with a 15.2 cm
ID by 2.4 meter long refractory lined tunnel furnace fired with either natural
gas or coal at 15 kW (Figure 2). The temperature profile in this unit matches
the profiles typically found in uti-lity boilers.

RESULTS

Figure 3 shows the results of a series of experiments in which methanol was
oxidized at varying reaction times and temperatures' in the presence of NO while
Table 2 shows the resu-lts of a series of experiments in which various gas
mixtures were passed through the reactor, and the concentration of S02 in the
post reaction gas mixture'was measured by procedure A.

Two sets of experiments were done in which mixtures .containing both NO and S03
were reacted with CH30H and the amounts of NO converted to N0a and S03 reduced
to S0e were measured as a function of the input CH3OH concentration. One of
these sets of experiments was done using procedure B and the other with
procedure C. Figure h is a relative rate plot for the oxidation of NO to N0S
and the reduction of S03 to S0e, i.e. a log log plot of the NO and S03
conversions observed in these experiments.

Another set of experiments was done in which the NO to N0e conversion was
measured for a gas mixture containing 225 ppm NO, 4.6'/. 0a, varying amounts of
CHaOH, either 0 or 2600 ppm S02, and balance inert, reacting for 0.59 sec. at
601 °C. The presence or absence of the SOe was found to have no effect on the
NO to N0b conversion.

A qualitative observation is also to be reported: it was found under a variety
of conditions that the conversion of NO and S03 to N0e and S0= produces at
least one mole of CO per mole of NO and/or S03 converted.

Figure 5 shows the results of combustion tunnel experiments in which methanol
was injected in to flue gas produced by burning either natural gas, or natural
gas doped with SOe, or coal, and the conversion of NO to NQa was measured. It
is to be remembered that in these experiments methanol was used in 1/1 ratio to
NO, i.e. no methanol was supplied for the concurrent reduction of SQs. Thus it

7B-32

-------
is hardly surprising that the NO to NQE conversion was not as complete as was
observed in the natural gas experiments.

Figure 6 shows the results of computer modeling calculations for the oxidation
of NO' by CHaOH in the absence of S03 while figure 7 shows the computer model's
.predictions for the reduction of SOg by methanol in the absence of NO for
varying temperatures

and reaction times. Calculations were also done for a variety of mixtures and
conditions for mixtures containing both NO and SQs. Figure a, a log log plot of
SOS remaining versus NO remaining, shows data from these calculations.

DISCUSSION

Scientific Implications

The results above clearly show that it is possible to obtain nearly complete NO
to NOa conversions at elevated temperatures at which N0e is thermodynamics!Iv
unstable with respect- to decomposition to NO. Given the limitations of the
previous work this finding may be surprising, but it- is to be recognized that
this system is kinetically rather than thermodynamically controlled. As the
temperature is increased the time needed- for methanol oxidation to convert NO
to N0e decreases. Thus, if the reaction time is shortened as the temperature
is increased, it is possible to maintain conditions such that the initial
conversion of NO to N0e goes to near completion while the NQS reconversion back
to NO does not have sufficient time to occur.

The finding that the oxidation of methanol is capable of doing two apparently
opposite things, selectively and quantitatively oxidizing NO to NOa nhile at
the same time selectively and quantitatively reducing S03 to SO, /, may also
prove puzzling at first glance. These opposite behaviors, however, are both
results of the fact that methanol oxidation is an efficient method of producing
H0e free radicals. By acting as an oxygen atom donor HQe can oxidize NO via
the reaction NO + H0a = N02 + OH and by acting as a hydrogen atom donor HOa can
reduce S03 via the reactions, S03 + HOe ~ HSOs + 0a, HS03 + M = S0& + OH + M,

While the success of the computer modeling in predicting an entirely new
chemical phenomenon is gratifying., it's to.be realized that the agreement
between the model's predictions and the experimental results is only
qualitative. The model predicts that there will be a range of temperatures
about 200°C wide within which the NO to NOe and SOe to 50a conversion can occur
and that this "temperature window" will move to higher temperatures as the
reaction time is decreased. While' the experimental data confirm this, the
model shows the temperature window at 0.1 seconds reaction time to be centered
at 850®C while experimentally the 0.1 second window is centered at 725°,C.

Given that some of the rate constants used in the modeling have.significant
uncertainties small discrepancies of this kind are to be expected.

Similarly both the experimental data agrees with the model's prediction that a
log log plot of the fractions of NO and SO?, remaining should be a straight
line, but the empirical slope found is somewhat different from that predicted
by the model, again reflecting uncertainties in the rate constants used in the
modeling.

7B-33

-------
In most respects the laboratory data for the conversion of NO to N0a by
methanol are in agreement with the pilot plant data. However, it was observed
in the laboratory experiments that SOa had no effect on NO to NDe conversion
while the pilot plant data show an effect of having sulfur in the fuel. The
explanation for this seeming conflict is, of course, that during the pilot
plant experiments sulfur was present as both S0e and SOa. Since there tvas only
enough methanol present to reduce the NO, adding S03, which competes far the
H0e radicals, naturally lowers the NO to NOa conversion.

Practical Implications

The experimental results reported above show new chemistry that allows the
rapid conversion of NO to NOe (i.e. conversion on a time scale compatible with
use of this chemistry for flue gas treatment) and allows the rapid reduction of
S03 to SOe• The extent to which this new chemistry will be useful in solving
the emissions problems of boilers and furnaces depends on the answers to two
questions;, first is it practical to apply this new chemistry in boilers and
furnaces and second, if it can be applied, is what it does useful?

The answer to the latter question depends on the situation. In situations in
which only NOx control is needed, i.e. situations in which the flue gas is not
scrubbed prior to discharge to the atmosphere, Methanol Injection would not be
useful. If, however, the flue gas is to be scrubbed before discharge either
for SOe or dust control, then methanol injection has two benefits. It can
enable the scrubber to also remove NOx and it can protect the scrubber from
corrosion by sulfuric acid mist. It's inherent in the operation of wet
scrubbers that they cool the flue gas being scrubbed to a temperature below the
acid dew point. Thus the S03 in the flue gas forms an acid mist in the
scrubber, droplets in the submicron range. The efficiency of removing such
droplets is typically only 50%. Consequently they can cause significant
corrosion problems both in the scrubber and downstream of it. By reducing the
SOa to SOe Methanol Injection would eliminate this problem.

While enabling an SOa scrubber to also control NOx emissions and eliminating
corrosion due to sulfuric acid mist are potentially very useful, it is to be
acknowledged that Methanol Injection has the disadvantage of causing modest
emissions of CO. It is, however, important to keep this problem in
prospective. In most applications the use of Methanol Injection to control NOx
and prevent acid mist corrosion would cause' CO emissions of less than 300ppm
and those emissions would enter the atmosphere from the top of a tall stack.
If an automobile has a fuel economy of 25mile/galIon and has its CO emissions
controlled down to 2grams/mile, it is emitting 1300ppm CO at ground level.

One may tentatively answer the former question by comparing the chemistry of
Methanol Injection with that of the Thermal DeNOx process®. In the Thermal
DeNOx process one locates the point in a boiler or furnace at which' the flue
gas temperature is in a critical temperature range and at that point injects a
small¦amount of NHa into flue gas. This causes a rapid homogeneous gas phase
reaction to occur in which the NHa selectively reduces the NO to molecular
nitrogen and water. While the Thermal DeNOx process has been successful in
numerous commercial applications, there have been also situations in which it
couid not be applied. In some boilers and furnaces mechanical access is a

7B-34

-------
major difficulty: one cannot conveniently install an NHa injection grid at the
locatiorr in the unit required by the chemistry. In other units the problem is
nonunif ormi ty of the flue gas temperature: since Thermal DeNOx can only be
applied1 over a modest temperature range, it can happen that at the location in
a unit at which the average temperature is appropriate to Thermal DeNOx the
range over which the temperature vary exceeds the process's to-lerance. It is
also possible that the reaction time available in a given unit may be
inadequate: while the Thermal DeNOx chemistry is rapid it is not instantaneous
and the process is not appropriate to situations in which some minimum amount
of reaction time cannot be provided.

Analogous limitations are to be expected in a Methanol Injection process. In
some units it may be difficult to install a methanol injection grid at the
location required by the chemistry. It is, however, to be noted that since the
optimum temperature for methanol injection is different, from that for Thermal
DeNOx,. the required location of the injection grid is also different. Thus the
processes compliment each other in that units which cannot provide an
appropriate location for the injection grid of one process may still be able to
do so for the other.

If one'compares the data in .Figure 3 with laboratory data for the Thermal DeNOx
process®, one finds that for both processes the temperature range is roughly
200°C. Thus, like Thermal DeNOx, Methanol Injection will be inapplicable in
situations in which the flue gas temperature is highly inhomogeneous.

Continuing the comparison between Figure 3 and the data in reference S. one
notes that for the two processes each at its optimum temperature, the reaction
time necessary for a given level of conversion, e.g. BOX, is 0.1 sec for
Thermal DeNOx but only 0.02^ sec. for Methanol Injection. Thus while Methanol.
Injection, like Thermal DeNOx, is subject to a minimum reaction time
requirement, this requirement is significantly less severe for the former.

Overall, it seems fair to say that the chemistry of Methanol Injection involves
quite similar limitation to those of the Thermal DeNOx process, and one would
expect that this new process would, like Thermal DeNOx, be broadly but not
universally applicable.

The combination of Thermal DeNOx and Methanol Injection may be particularly
advantageous in situations in which high degrees of NOx control are needed.
Methanol Injection could provide a substantial increment of NOx control over
and above what is achieved by Thermal DeNOx, Moreover Thermal DeNOx has
occasional problems with fouling due to NH„S0«. formation, i.e. the reduction of
NO with NHa can leave small amounts of unreacted NHa in the flue gas.

Downstream of the Thermal DeNOx reaction zone the unreacted NHa can combine
with Ha0 and S03 to form NH«»S0«., a sticky and corrosive liquid which can foul
heat transfer surfaces. Since Methanol Injection provides S0a reduction it can
eliminate this problem, can solve the problem of byproduct production in the
Thermal DeNOx process. Similarly if Methanol Injection is used downstream of
Thermal DeNOx, the amount of methanol to be injected is greatly reduced, and
the problem of producing byproduct CO is also largely removed.

7B-35

-------
ACKNOyLEDGMENT

This work was supported by the National Science Foundation under a Phase I
Small Business Innovation Research grant, award number ISI-87&03Vt.

REFERENCES

1. H. L. Faucett, J. D. Maxwell, and T. A. Burnett, Technical Assessment of NQx
Removal Processes for Utility Application, EPA report EPA-600/7-77-127, p95 and
following

2 N. Murakami, and J. 12umi, U. S. Patent *#,350,669, 1992

3 R. K. Lyon, U.S. Patent ^,743,^36, 1988

k J. Wendt et.al., 14th Symposium [International] on Combustion,
p.890, The Combustion Institute, 1973

5 R. K. Lyon, U. S. Patent 3,900,559. 1975

7B-36

-------
Table 1

Demonstration of the Selective Reduction of S0a by CHa0H

T ,°C

t, sec.

Input

Gas Composition,

Balance N£,

Output

Gas

0®

Ha0

CHaOH

SOe

SOa

'/.

ppm

698

0,48

4.2

5.9

640

225

176

>0.1

698

0,48

4.2

5.9

840

2g5

176

>0, 1

700

0.4B

4.2

5.9

840

225

17B5

102'

6.8

699

0.48

4.2

5.9

B40

225

1705

176

12.9*

699

0.4B

4.2

5.9

B40

225

1765

176

9.6

11.1*

690

0.51

4.5

6.3

445

1705

176

5.6

698

0.55

4.8

445

1924

>0.1

697

0.50

5.1

2022

>0.1

697

0.55

4.B

445

1924

>0.1

593

0.5%

4.5

5.9

616

225

1705

176

3.7

593

0.54

4.2

' 5.9

840

225

1785

176

1.7

751

0.052

4.2

5.9

091

225

176

14. B

30.3

752

0.055

4.5

6.3

225

176

165*#

* In these runs in increase in CSOel«>u* was measured and found to be 180-190 ppm,
i.e. the sulfur balance was 101'/. based an total sulfur input and the differential
balance
-------
vent

Figure 1 Precision gas kinetics flow system.

7B-38

-------
""-J
CO

ta
\£>

Removable
Choke

Back-Fired
Heating
Channel

Removable
Cooling
Coi 1

Back-F1red
Heating
Channel

Removable
Cooling
Coil

~3J» -

h d

k $

Flue Gas
Sampling
Location

Back-Fired
Heating Channel

To Stack

Figure

Cross-sectional views of the control temperature

tower.

-------
Kinetics of NO oxidation by CH30H

5% 02r 400ppm CH30H, 215 ppm NO

Temperature, C

0.9 sec 0.1 sec 0.048 sec 0.024 sec

Figure 3 Kinetics of NO Oxidation

-------
K>

a
w

3
o

»»»

0.0

-0,1

•0.2

-0.3

y * -0.Q1Q + 02BZx R»QSB

* / *

/
/ o

/
/

•t.O -0.0 -0.6 -0.4 -0.2 -0.0

Log(NO)out/(NO)in

Figure 4 Experimental Data; log log plot of fraction NO
remaining against fraction SO, remaining, solid
points Method Bopen points method C.

7B-41

-------
CH30H/N0 -1

T, C

*Gas firing, A —+~ Gas firing, B ~^ Coal firing, G -

A = 240 ppm N0x, B = 240 ppm NO, C = 1080 ppm NO,

0 ppm SO 1100 ppm SO 730 ppm SO

X XX

Figure "5 NO conversion during gas and coal firing.

-------
Temperature, °C

Figure 6 Modeling Calculations for Oxidation of NO by CH.OH,
215 ppm MO, 400 ppm CH^OH, 5% Og, balance inert.

7B-43

-------
Temperature, °C

Figure 7 Predictions of the reduction of 30 ppm S0-
in the presence of methanol oxidation under
isothermal conditions. Reaction mixture
= 4X 0?, 2000 ppm S0_, 30 ppm SO-, 40 ppm
CH,0H, balance inert.

7B-44

-------
JO

¦0.5 -0.4 -0.3 -0.2 -0.1 -0.0

!•* NO/HCK

Figure 8 Data from computer modeling; log log plot of fraction
NO remaining against fraction SO3 remaining, data for
calculations at 600°C j^ith a reaction mixture of
10 ppm SO3, SOt) ppm NO, 800 ppm CH^OH,, balance inert.

7B-45

-------
Session 8
NEW DEVELOPMENTS

Chairman: E, Plyler, EPA

8-i

-------
CATALYST POISONING IN THE SELECTIVE CATALYTIC SEDUCTION REACTION

1. T, Yang, J. P. Chen and M. A. Buzanowski
Department of Chemical Engineering
State University of New York at Buffalo
Buffalo, NY 14260

and

J. E. Cichanowicz
Coal Combustion Systems Division
Electric Power Research Institute
Palo Alto, CA 94303

ABSTRACT

Activities of the V^O./TiO^ catalysts doped with various possible poisons have
been measured with a simulated flue gas under SCR conditions. The strongest
poisons are the alkali metal oxides; and the strength of the poison increases with
the basicity, i.e., Cs > lb > K. > Na > Li. Arsenic is also a poison, but is
weaker than all alkali metals except Li. The activities of all catalysts, both
undoped and doped,' are increased by the presence of SO2. The experimental results
indicate that Bronsted acid sites on V^O. are responsible for the SCR reaction,
and that the proton donicity of the V-G-H surface- group determines the activity.

INTRODUCTION

The V^O-ZTiO^ catalyst has gained acceptance for the selective catalytic reduction
(SCR) process, and has been tested commercially in Germany, Japan and other
countries. The catalyst life is an important factor in the economics of SCR.
Deactivation of the catalyst is caused mainly by the various "poisons" contained
in the hot flue gas, which in turn, depends on the type of coal used in the
combustion process. Very little is known about the poisons for the SCR catalyst.
The purpose of this study is to obtain a systematic and fundamental understanding
of the types of poison, as well as the poisoning mechanisms, with a specific
emphasis on the possible poisons contained in American coals.

8-1

-------
EXPERIMENTAL

The titanium dioxide powder (Degussa, P25) was mixed with distilled water at a

ratio of 1.75:1 by weight, stirred thoroughly, and dried in an oven at 60°C for 24

hrs., followed by 120°C for 72 hrs. The dried mass was crushed and sieved to

collect the 20-32 mesh fraction. This fraction was calcined at 600°C first in air

for 1 hr. followed by 6 hrs. in hellium at the same temperature. This support

gave a BET surface area of 30 ra /g.

The catalyst was prepared by incipient wetness impregnation. The TiC^ support was
impregnated with an oxalic acid solution of ammonium metavanadate, followed by
drying at 120°C for 3-4 hrs. and calcination at 500"C in 0^ for 3 hrs. The
loading was 5% in all samples.

To prepare the poison doped catalysts, the 57. V^O^/TiO^ catalyst was impregnated
with the corresponding poison precussor solutions. For Li20» NaK^O, Rb^O and
Cs^O doped catalysts, LiNO^, NaNO.^, KNO^, RbOOCCH^ and CsOOCCH^ were used as the
precussors. The impregnated, catalyst was dried at 120°C for 3-4 hrs, and their
calcined to decompose the nitrates and acetates.

A quartz tubular reactor was used for measuring the SCR catalyst activity. The
reactor (2 cm I.D. and 30 cm long) was equipped with a thermocouple well and gas
preheating section. The nitrogen was oxygen-free grade (< 0,5 ppm 0^)", the
was anhydrous grade (99.99% purity); the NO purity was 98.5% min.; SO^ purity was
99.92; and 0^ was ultrahigh purity, > 99.99%. The simulated flue gas consisted of
1,000 ppm NO, 1,000 ppm NH^» 2% 0^, 1% SO2 (when used) and the balance N^• The
space velocity was 1,500 hr \ The flow rates were controlled by FM 4575 mass
following a, gas blending system (Linde Div,:). The system is shown in Fig. 1.

The catalytic activity (NO^ conversion) was measured by

[NO ]. - [NO } „
x in x out

NO Conversion {%) = —i x 100%

x INO J.

x in

The NO and NO concentrations were measured by a chemiluminescent NO-NO--NO
x - 2 x

analyzer (Model 10, Thermo Electron Co.).

The rate equation for the SCR reaction is, as .reported in a number of
publications:

8-2

-------
1 NO = 1 0
W dt NO NH3

Tins fitstofdsir irsfcs coristsntj It} csin £s© cslcTLileitscl foT fclft0 plu^—flow itssctor byi

k - - p~ £a(l-X)

where C is concentration, F„- is molar feed rate of NO, is inlet NO

NO NO

concentration, W is weight of catalyst, X is conversion, and t is time. This rate
equation is applicable at O2 concentration > 1% and at above stoichiometric NH^/NO
conditions. The temperature programmed reaction was performed at a heating rate
of 10°C/rain.

RESULTS- AND DISCUSSION

Typical temperature-programmed reaction results are shown in Fig. 2 for the
standard 5% an<^ K^Q-doped catalysts. At temperatures higher than 25Q°C, the

oxidation of NH^ (forming NO) became significant thus reducing the NO conversion.
From these data, the first-order rate constants for the SCR reaction, k, can be'
calculated for temperatures below the peak temperature.

Figure 3 summarizes the activities' of the standard 5% catalyst doped with various
amounts of metal oxides as the poison.

The most prominent result is that alkali metal oxides are the most potent poisons.
Moreover, the strength of the poison follows the order of basicity, i.e.

Cs20 > Rb20 > K20 > Na20 > Li2Q

50% Deactivation of the standard catalyst occurred at the following alkali
metal/vanadium atomic ratioss

50% Deactivation; 0.07 (Cs/V);

. 0.31 (Na/V);

0.12 (Rb/V); 0.14 (K/V);
0.58 (Li/V)

8-3

-------
This result shows the long-range effect of the highly alkaline metals, e.g., each
Gs atom can deactivate (by 50%) approximately 14 V atoms. This result also
reveals the severity of the alkalinity of American coals in the SCR reaction.

The effect5of As on SCR has long been noted in the German experience. Our results
show that, although As is also a potent poison, its poisoning strength is weaker
than the alkali metals, except Li. Figure 3 also indicates that PbO is a strong
poison whereas CaO is a weak poison.

Effects of SO.,

It has been known from pilot plant experience as reported in the literature that
SC>2 deactivates the V^O^/TiO^ catalyst by forming NH^HSO^ deposit on the catalyst.
The results given in Table 1, however, showed a promoting effect of SO, on the
catalyst activity for both undoped and poison-doped catalysts. The promoting
effect of SO^ may be interpreted as the enhancement of the acidity of the catalyst
surface.

Table 1 also includes the activity data of CuSO^/A^O^. CuSQ^ is considered as a
solid acid with Bronsted acid sites. The activity of CuSO^ is similar to that of

at 300"C. However, it is much more active than at higher temperatures

(not shown in Table 1).

Active Sites and Possible Poison Mechanism

Extensive measurement on the chemisorption of NH^ on the' undoped and doped

catalysts have been undertaken in our laboratory. The measurements were done at

200'C. A nearly one-to-one correspondence between the amount of NH^ chemisorption

and the SCR activity was obtained for all catalysts included in Fig. 3. However,

° 2

a large amount of NH^ chemisorption (corresponding to 21 A /NH^ molecule) was also
measured for the Ti02 support, which has no SCR activity. These results indicate
that NH^ chemisorbs on all acid sites, including Lewis and Bronsted acid sites.
It is known, however, that the chemisorption of NH., on TiC^ is by Lewis acid and
hydrogen bonding, and that NH^ chemisorption on is on Bronsted acid sites.

Consequently, it appears that the Bronsted acid sites on are responsible for

the SCR reaction, which is likely in the form of -V-O-H groups. It follows that

8-4

-------
the donicity of H proton, i.e., the ease of breakage of the 0-H bond, is the rate-
limiting factor 'in the SCR reaction and determines the activity. The effect of
the alkali metals appears to be dontating electrons to the surface and
consequently strengthening the V-Q-H bond. And such an effect could be a long-
range one as indicated by the results with CS2O, which deactivates a large number
of active sites by each Cs atom.

Table 1

Effect of S02 on NO Conversion and late Constant (k) for SCI at 300»C.

(Other conditions are given in Figure 2).

Without S0„

With 1% SO.

Catalyst Conv. % kxlQ2, em3/g/s Conv. % kxlO2. cia3/g/s

5% V205/Ti02(A) 98.0

10.38

99,2

12.82

0.74% CaO/A

97.2

9.49

99.2

12.82

0,32% Li20/A

91.4

6.52

99.1

12.63

0.68% As203/A

96.7

9.09

99.2

12.82

7% CuS04/A1203

99.2

12.82

8-5

-------
Figure 1. Schematic of experimental system. TP: temperature program/control.
CAT; packed catalyst pellets.

Figure 2. Conversion of NO in the SCE reaction on K,0 doped catalyst. NHL = NO =

8-6

-------
Figure 3. Activity (as indicated by rat, constant) of 5% v^rtlOj doped with

different amounts of metal oxides where M/V » metal/vanadium atomic
ratio. Reaction condition: 300°C, NO = NH, = 1,000 ppm, 0^, = 2%,
N2 = balance, GH3V - 15,000 hr"1.

8-7

-------
\

(intentionally Blankf?

i $

•#

o •

fi
&
*4

#
It

rt:

«'

8-8

-------
CATALYTIC FILTER BAGS

Marie R. Kalinowski, Patrick F. Aubourg

Science and Technology
Owens-Corning Fiberglas Corporation
Technical Center
Granville, Ohio

ABSTRACT

Owens-Corning Fiberglas Corporation (OCF) has developed a catalyst-coated
fabric for baghouse applications to simultaneously control particulates and
NOx emissions. NOx is eliminated by catalytic reaction with ammonia. NOx
'reductions of. 901 with low ammonia slip (< 10 ppm) are achievable. Bench
scale tests conducted at OCF and at the University of North Dakota, Energy
and Mineral Research Center have shown the merit of this system; this data
will be presented. Eight hour tests were conducted on 8 X 11 inch coupons
using flue gas from a coal-fired boiler. Conversion efficiencies of up to
93% were measured. Pre!iminary tests have also been conducted on small bags.
Economic data will be presented.

Due to the very efficient use of the catalyst and combination of two
functions into one unit, cost will be significantly lower than use of a
commercial SCR (Selective Catalytic Reduction) unit and a particulate removal
unit.

Introduction

Many techniques exist for the reduction of NOx emissions from flue gas
streams of coal-fired boilers. Some methods, while relatively
cost-effective, are incapable of high (> 80%) NOx removal. Others, with the
potential for high NOx removal rates, result in large capital investments,
high operating costs, or operational difficulties, making the processes
unattractive.

Over the past several .years, Owens-Corning Fiberglas Corporation (OCF) has
developed a catalytic S-glass cloth capable of removing NOx at high
efficiencies. When sewn into a filter bag, the cloth will also effectively
filter particulate matter. There are several advantages to this concept; a
two-in-one pollution removal device, resulting in space and cost
minimizations; no undesirable by-products; an already-existent industry
familiarity with baghouse technology; and, of course, NOx removal
"efficiencies in excess of 90%.

Laboratory data on a simulated flue gas stream at OCF indicated that removal
efficiencies above 90% were indeed -feasible. In order to evaluate the
process in an operational flue-gas stream, OCF has worked with the Energy and
Mineral Research Center (EMRC) at the University of North Dakota. The
objectives of the work were to optimize the cloth's" efficiency for NOx

-------
removal and to examine the effects of various operational parameters
(stoichiometric ratio, coal type, etc.) on the removal efficiency.

How Does it Work?

In some respects, the OCF catalytic filter bag is similar to existing SCR
(selective catalytic reduction) technology. It employs ammonia in the
presence of a vanadia/titania catalyst for the reductions:

6 NO +4 NH3 5N2 + 6 H20
6 N02 + 8 NH3 --> 7N2 + 12 H20

Operating temperatures are also similar (650°F).

What makes the catalytic filter bag different is the form of the
vanadia/ti tania catalyst. Instead of being supported on a honeycomb-1 ike
structure, the catalyst is coated and bonded directly on an S-glass cloth
surface. Titanium isopropoxide and vanadium tri-n-propoxide oxide, sometimes
called "sol-gel" materials, are used for the coatings. After coating, the
sol-gels are air-dried, which hydrolyzes them, forming a highly porous layer.
Low temperature curing drives off any residual organics and partially
densifies the coating. The result is a catalytic coating--still highly
porous, and strongly bonded to the glass surface. It has been estimated(l)
that the OCF catalytic cloth would require only one-seventh the volume of a
traditional SCR catalyst to achieve equal levels of NOx removal.

Experimental

All tests were conducted on a slipstream from the coal burning Particulate
Test Combustor (PTC) at EMRC. Initially, filter bags were not available;
therefore, a special slipstream sampler was constructed. A fabric sample,
roughly 8 x 11 inches, was housed in an oven operating at 650 ± 25°F. After
the filter the gas flow was split, with 10 scfh going to a sample conditioner
and flue gas analyzers, and the balance of 20 scfh going to a gas pump and
dry gas meter for control of the total flow. Ammonia was injected into the
system well upstream of the sample point to ensure sufficient mixing into the
flue gas stream prior to sampling with the slipstream sampler. Baseline
operating conditions for the tests are listed in Table 1.

The filter bag trial used nine bags, each approximately 4 feet long and 7
inches in diameter. Glass cloth weight was approximately 14 ounces per
square yard. Experimental conditions were similar to those used for the
slipstream trial.

Results and Discussion

Conversion efficiencies for the first fifteen samples tested on the
slipstream sampler ranged between 70 and 93 percent. Differences in
conversion efficiencies arose from the amount of catalyst on the cloth,
curing differences, and cloth weight variations.

These tests on swatches being extremely encouraging, it was decided to
conduct tests on small filter bags.

Some problems were encountered during the early stages of the bag trial.
Conversion efficiencies were low, averaging 50 percent. The reasons for the
low conversion efficiencies were not immediately understood. Recent work has

8-10

-------
indicated that the silicones used as gasketing material attack the
vanadia-titania catalyst and decrease conversion efficiency. After
replacement of the original bags and regasketing the baghouse with Fiberfrax,
conversion efficiencies averaged nearly 80 percent. After two days of
operation, conversion efficiencies had not decreased and, equally important,
no holes or tears were found in the bags.

A second set of slipstream samples was tested using a different coal.
Baseline operating conditions are listed in Table 2. Conversion efficiencies
averaged 86 and 93 percent for the two samples. The stoichiometric ratio of
ammonia to NOx was also varied, and the ammonia slip was monitored. Results
are plotted in Figure 1. The conclusion was that high conversion
efficiencies could be obtained with minimal ammonia slip.

Economics 1

The key economic advantage of a catalytic fabric system lies in the fact that
major, new and unfamiliar equipment would not be required and that a single
device (a baghouse) could be utilized for the control of two major
environmental pollutants: NOx and particulate matter. The utility industry
has been required to control particulate matter for decades, and it has been
documented that fabric filtration is the control method of choice.

Once the fabric filtration method is chosen for particulate control, it can
be demonstrated, that the fabric cost is presently less than 1% of the
installed filtration system cost. By gaining both particulate and NOx
control via conversion to a catalytic fabric, significant economic advantage
is gained over the competing high efficiency NOx removal technologies. An
economic summary is listed here (2).

Installed Cost Operating Cost

$/KW im'Ts/kwh

Catalytic Filter

Baghouse (650°F System) 178 1-2

Competing High NOx
Removal Technologies
Adjusted to Include

Particulate Control 209-333 11-24-

The system modifications required to. allow for a high temperature baghouse
and installation of the required ammonia injection system are minor compared
to the hardware involved with the other competing high NOx removal
technologies.

Conclusions

Based upon preliminary, short-term testing, the following conclusions can be
made concerning catalytic filter bags.

1. High (:>9QI) NOx removal can be achieved.

2. Ammonia slip can be kept at low levels.

3. The technology is applicable to different coal types and the subsequent
differences in S03 levels, ash contents, etc.

8-11

-------
4. Cloth and catalyst durability are good.

5. Economics are attractive.

Acknowledgements

The following people were instrumental in the development and advancements
made on this project: Stanley J. Miller, Dennis L. Lauds!, Greg F. Weber,
and Michael L. Jones, all of the Energy and Mineral Research Center at the
University of North Dakota; Ken E. Carney and Dennis L. McGarry of
Owens-Corning Fiberglas; and Gary M. Nishioka, of H and N Instruments,

References

1. X. E. Carney, OCF communication to EPRI, December 11, 1986.

2. Economics on competing technologies were culled from several EPRI
reports. A bibliography is available upon request.

8-12

-------
Table 1. PTC Baseline Operating Conditions for First Slipstream
Sample- Trial

Temperature

NOx inlet concentration
S02 inlet concentration

% 02

% CO 2

Air-to-Cloth ratio

NHj/NOx

Test duration

Coal

Table 2. PTC Baseline Operating
Sample Trial

Temperature

NOx inlet concentration
S02 inlet concentration

% 02

Air-to-Cloth ratio
NH3/N0x
Test duration
Coal

650° F
800 ppm
300 ppm
51

15.51
2

1.0

11 hours

Velva, N.D. lignite

Conditions for Second Slipstream

650°F
900 ppm
1600 ppm
4.55SC

0.9

24 hours

Indiana bituminous

Indiana Bituminous Coal

^Conversion Efficiency (top) and NH3 Slip (bottom)

Versus NH3/NOx Ratio

100
80
60
40
20

j|" —'f if"—

NH.

Slip

, —i—¦—i—•—i—1—i—1—i—1—i—1—i—1—r
0.80 0.82 0.84 0.86 0.88 0.90 0.92 0.94 0.96

NH3 / N0X Ratio

Figure 1
8-13

-------
"5
%

(in t en t i onally;'fBl an k)

V %

i
I

, '

8-14 \

-------
ADVANCED IN-FURNACE NOx CONTROL IN
NEW EUROPEAN COAL-FIRED POWER PLANT

Knud Bendixen
Jar! Pedersen
Burmeister 1 Wain Energi A/S
23 Teknikerbyen
DK-2830 Virum
Denmark

ABSTRACT

NOx-control of coal-fired power plants is a legislative requirement in most com-
munities in the industrialized world. For more than a decade, Burmeister & Wain
Energi A/S develops Low-NOx combustion systems, among which some have been in-
stalled in power plants in Denmark, Sweden, and in the Federal Republic of Ger-
many, al1 countries with very strict regulations of the NOx-emissions from coal -
fired power plants. In the field of coal combustion in large-scale boilers, BWE
has concentrated its developments on tangentially fired boilers having circular
Lew-NOx burners and supplementary air staging in the furnace.

This paper describes the results obtained during the first 9 months of the oper-
ation of the new Low-NOx coal-fired 410 MWe cogeneration unit No. 9 at Walsum
Power Station in the Federal Republic of Germany. The proven NOx-content at fur-
nace outlet is less than 290 ppm (at 3% Og) at the same time as less than 3%
unburned in the fly ash is obtained.

INTRODUCTION

As designer and manufacturer of large-scale boilers, BWE has faced the new re-
quirement of Low-NOx combustion systems caused by the strict regulations of the
NOx-emission in Europe, and in compliance with the clients' needs, BWE has.de-
veloped Low-NOx combustion systems capable of meeting these new standards. To
FRG, a country of the world's most strict rules of NOx-emission (Ref. 1), BWE
has supplied a new Low-NOx coal-fired cogeneration boiler, Unit No. 9 at Walsum
Power Station, mainly used for base-load with a maximum generator rating of 410
MWe. The performance data are listed in Table 1.

The type of coal used is a domestic medium volatile bituminous coal. The coal
analysis is given in Table 2. The general behaviour of the coal is low-reactive,
therefore measures must be taken to ensure stable ignition at a burner turn-down
ratio of 1:2.5.

The requirement of the NOx-emission at stack outlet is 200 mg/Nrn^ (at 6% Oj, dry
gas) corresponding to 120 ppm (at 3% 02, dry gas), which is attained using a se-
lective catalytic reduction. The requirement of the NOx-emission at furnace out-
let is 500 mg/Nm^ (at 6% O2, dry gas) corresponding to 290 ppm (at 3% 0£, dry
gas) and must be attained within the combustion process.

8-15

-------
The fly ash must contain no more than 3% unburned carbon, as deposit of fly ash
is prohibited in the FRG, and, therefore, it has to be sold as an additive to
the cement industry instead.

INSTALLED LOW-NOX COMBUSTION SYSTEM

The boiler is designed for low specific heat release in the furnace, the conse-
quence of which is a furnace volume about twice as big as that of a convention-
ally designed furnace. At MCR, the specific furnace heat release is'0,105 MW/m3
(3.71 MW/ft^), giving a mean residence time of the coal particles of approxi-
mately 3.6 sec and a furnace gas outlet temperature of 1210 °C (2210 °F),

A tangential firing system with four levels of circular Low-NOx burners was
chosen. The air staging Low-NOx combustion system includes six different air
streams (Ref. 2).

The relatively low coal reactivity necessitates a fine grinding of the coal:

• max. 20% > 90 microns (170 mesh)

• max. 0.3% > 250 microns ( 60 mesh)

The coal is pulverized by four vertical roller mills with integrated rotary
separators, respectively feeding the four circular Low-NOx burners at the same
level.

The distribution of coal and air is within 5% of the mean value. This is
achieved for the air flow by separate venturi measurements for each burner, and
for the coal flow by using riffle-box coal splitters.

OPERATION AND MAINTENANCE

Ignition stability is strongly promoted by swirl of the secondary air, resulting
in a good flame detection and discrimination between individual flames, even at
turn-down to 40% burner load. Secondary air swirl has also proven to be favour-
able as regards minimizing of NOx and unburned in the fly ash.

Tertiary air co-axially around the flame acts as a protecting blanket against
corrosion and slag formation, protecting the walls in the nearby burner zone by
means of a relatively high Oj-content. The design of the firing system makes it
sufficient to operate the sootblowers only once every 24 hours in defiance of
the high ash content of the coal, yet providing optimum, reliable performance.

The vertical roller mills were proven capable of grinding the high-ash coal
to the required values using a relatively low specific power consumption,
11 kW/t/h,

RESULTS

During the commissioning period various parameter settings for the combustion
system were investigated in order to identify the best overall performance:

• NOx-content exit furnace .. < 290 ppm (at 3% O2, dry gas)

• Unburned in fly ash ¦ <3 %

• CO at furnace walls < 2000 ppm

t 05 at furnace walls >1 %

8-16

-------
CO and Oj were measured locally at a total of 284 locations, resulting in the
parameter setting of an air staging in the furnace as shown in Figure 1. This
relatively easy parameter setting has proven the NOx-content and unburned in the
fly ash to be within the required values with good margins.

As an example of the testing, Figure 2 shows the effect of over-burner air. The
feature of the over-burner air is to etablish a relatively fuel-rich Stage 2 of
combustion in the centre of the furnace. As can be seen from Figure 2» the NOx-
enri ssion is reduced by approximately 50 ppm, corresponding to 15% at full load
using over-burner air. Complete combustion and burn-out are established in
Stage 3 of combustion, using only 10% over-fire air.

In spite of the difficult type of coal and a newly developed design of the
firing system,, the take-over by the customer was on schedule.

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
services rendered to make this project possible;'

• Walsum Power Station
§ Steag ,

REFERENCES

1. 0. Rentz and R. Leibfritz. "Overview of Recent Developments in NOx-Control
in Europe". Paper presented at the 1987 Joint Symposium on Stationary
Combustion NOx Control, march 1987, New Orleans.

2. J. Pedersen. "BWE Low-NOx Combustion Systems".1 Paper presented at the
Swedish Flame Days 1987, September 1987, Studsvik.

The work described in this paper was not funded by the U.S. Environmental
Protection Agency and, therefore, the content does not necessarily reflect
the views of the Agency, and no official endorsement should be inferred.

8-17

-------
Over fire air

Over burner air
Circular Low NOx burner

Primary air + R F
Core air

Secondary air
Tertiary air

-A

—I—i—\ i' i 1—I—rri—I—i } i 1 j—!—

0,9 1.0 1,1 1,2

Stoichiometric ratio

Figure 1. Air Staging Principle

ppm
(at 3% 02)

400
300
200
100
0

0 200 400 600 800 1000 MWth
Furnace heat release

Figure 2. NOx-Content Exit Furnace

• Wit
o Wit

"i over burner'air
lout over burner air

Or-rrfT

8-18

-------
Table 1

WALSUM POWER STATION, FRG
Cogeneration Unit No. 9
Performance Data for Coal-Fired Boiler

Steam flow, outlet 335 kg/sec 2,650,000 lb/hr

Steam pressure, outlet 200 bar 2,900 psi

Steam temperature, outlet 535 °C 995 °F

Reheat pressure, outlet 42 bar 610 psi

Reheat temperature, outlet 532 °C 990 °F

Type of Boiler:
o One-pass Benson boiler
o Dry-bottom furnace
o Reheat

Maximum generator rating 410 MW_

Table 2

WALSUM POWER STATION, FRG
Cogeneration Unit No. 9
Coal Analysis, typical

Proximate, as

received:

Ultimate, as

received:

Moisture

8.70 %

Moisture

8.70 %

Ash

24.08 %

Ash

24.08 %

Volatiles

18,69 %

Carbon

57.47 %

Fixed carbon 48.53 %

Hydrogen

3.45 %

Sulphur

1.04 %

LHV

22.3 MJ/kg

Nitrogen

1.11 %

HHV

10,000 Btu/lb

Oxygen

4.15 %

General behaviour: low reactive

8-19

-------
\

(Intentionally Blank'

j-f
/y

>/•

¦T

$¦
X-

fi
U

8-20

-------
Two-Stage DeNOx Process Test Data
from Switzerland's Largest Incineration Plant

Dale G. Jones, Ph.D., P.E., EMCOTEK Corporation
Lawrence J. Muzio, Ph.D., Fossil Energy Research Corporation
Erhard Stocker. MHA Basel, City of Basel
Peter C. Nuesch, K+K Ofenbau AG
Stefan Negrea, P.E., KRC Umwelttechnik GmbH
Gert Lautenschlager, Ph.D., Noell GmbH
Erwin Wachter, K+K Ofenbau AG
Gerhard Rose, K+K Ofenbau GmbH

ABSTRACT

Preceding page blank |

-------
SUMMARY

A new process for NOx removal using two boiler zones for chemical injection has been
installed and tested at KVABasel, the largest municipal solid waste (MSW) incineration
plant in Switzerland. Urea and methanol are used in combination with proprietary
injection nozzles to provide NOx removal performance between 65% and 80% at flue gas
temperatures varying between 1500 and 1900 F. Methanol is used to provide
substantial reduction of ammonia slip and/or air preheater deposits. Between 1740
and 1830 F at 65% NOx removal, ammonia slip of 2.5 to 5.0 ppm (without methanol)
can be reduced to less than 1.0 ppm. Further commercial development Includes award
of a contract to install the Two-Stage DeNOx Process on a 325 MW coal-fired boiler in
West Germany. The startup date is in August, 1989, and we plan to provide additional
details in a forthcoming paper.

PLANT AND DeNOx PROCESS EQUIPMENT DESCRIPTION

There are four boilers at the KVA Basel incineration plant. Two are 46 years old (each
rated at 220 TPD) and two are 24 years old (each rated at 330 TPD). The Two-Stage
DeNOx Process was installed on one of the larger boilers. Figure 1 provides a furnace
system cross-section which shows the two zones of chemical injection, the flue gas flow
path, and the location of the temperature sensor used for process control. Urea is pro-
vided in bulk granular form and mixed with water prior to being pumped to proprietary
injection nozzles in the first zone. Methanol (or MeOH) is supplied by liquid pumps to
the second-zone injector nozzles. The flowrates of chemlcal(s) to the various injection
locations are controlled by the single flue gas temperature sensor. The flue gas flowrate
is 52.000 DSCFM at 11% 02 and the uncontrolled NOx level is 160 dppm. To provide
a high degree of flue gas mixing, injector system airflow totalling 1500 SCFM is used.

100

14001450 15001550 16001650 17001750 18001850 1900
Flue Gas Temperature. F
Temperature Window of Two-Stage DeNOx Process

Total Injection System Airflow = 1500 SCFM

Molar Ratios: (Urea/NOx) = 0.80 and (MeOH/NOx) = 0.14

Initial NOx = 160 dppm & Flue Gas Flow = 52.000 DSCFM (11% 02)

8-22

-------
WIDTH OF THE TEMPERATURE WINDOW

Due to wide variations in burning characteristics of the solid waste, the flue gas tem-
perature (measured by thermocouple upstream from the first injection zone) varies be-
tween about 1400 and 1900 F. The control system compensates for much of this tem-
perature fluctuation by changing the location(s) at which chemical(s) are injected. As
shown in Figure 2, using a molar ratio of urea/NOx of 0.80 plus methanol/NOx of 0.14
produces 80% NOx removal in the KVA Basel boiler at flue gas temperatures between
1600 and 1850 F. The data in Figure 2 was collected at 5-second intervals and subse-
quently averaged for a 24-hr period of operation. Scatter in the averaged values are due
to rapid temperature fluctuations in the KVA Basel furnace. The wide temperature win-
dow for high levels of DeNOx is due to (a) a high degree of flue gas mixing provided by
proprietary injector nozzles, (b) operation of the temperature control system to adjust
the location(s) where chemical(s) are injected, and (c) the drying time for urea droplets,
which provides some delay and allows better urea dispersion before DeNOx reactions
begin to occur. We have called this delay time phenomenon the "popcorn effect".

UREA/NOx MOLAR RATIO

Figure 3 shows the effect of
urea/NOx molar ratio on
percent NOx removal.

DeNOx levels of 78%, 65%,
and 48% are achieved at
urea/NOx molar ratios of
0.80, 0.60, and 0.40, re-
spectively. The corre-
sponding percentages of
urea utilization are 49%,

54%, and 60%. By defini-
tion, a urea/NOx molar
ratio of 0.50 equals 1.0
moles of NH2 per mole of
NOx. These results are
believed to be representa-
tive for electric utility boil-
ers where the flue gas tem-
perature at any given in-
jection location can be
maintained in a desired
range from about 1740 to
1830 F.

DATA ACQUISITION

Extreme care was taken to obtain repeatable and representative data during a two-
month testing program. For example, NO signals were obtained using two separate gas
analyzers: one chemiluminescent type and one NDIR type. In the same way, duplicate
signals were obtained for 02 and CO, and all signals, including boiler and process

8-23

Urea/NOx Molar Ratio
Effect of Urea/NOx Molar Ratio on Percent DeNOx

-------
operating parameters, were recorded using a computerized data-logging system.
Ammonia slip values were also determined by two different methods; (a) continuous
NDIR gas analyzer (operating at 500 F), and (b) 15-minute grab samples analyzed by
wet chemistry. It was found that both methods of ammonia sampling provided
consistent and representative results.

AMMONIA SLIP REDUCTION USING METHANOL

35"

One of the important features of the Two-Stage DeNOx Process is the use of methanol
to reduce ammonia slip and related concerns including air-preheater deposits. At KVA
Basel, relatively high levels of ammonia slip (i.e. 15 to 40 ppm at urea/NOx raUo of 0.80)
can be tolerated since the ESP system operates at 460 F and there is very little depo-
sition of ammonia salts, either on equipment surfaces or ESP ash. For coal-fired boiler
applications, however, such high levels of ammonia slip are unacceptable. Figure 4
shows ammonia slip as a function of flue gas temperature, both with and without
methanol injection at a
urea/NOx ratio of 0.62,
corresponding to 65%

NOx removal. As can be
seen, when methanol is
added, very large reduc-
tions in ammonia slip are
realized. Between 1740
and 1830 F at 65% NOx
removal, ammonia slip of
2.5 to 5.0 ppm (without
methanol) can be reduced
to less than 1.0 ppm. In
coal-fired boiler applica-
tions, this would be criti-
cally important for mini-
mizing air preheater de-
posits and/or, when re-
quired, ammonia con-
tamination of ESP ash.

The Two-Stage DeNOx
Process can thus be opti-
mized to meet site-spe-
cific requirements.

30-

i"50-
j*

5 45-

£ 40-

>
az

S5
¦d

¦d

6
i
<

25'

20"

15'

10'

5 "

i ,

NOTE: Initial NOx

1"" '

» 160 dp

pm (11% 02]

Figure

\ 1

[

\ Urea

Only
Dx) = 0.6

""¦"¦"V*—" (Urca/INV

d. *

.. j

v ] \

]

— h

, (M

lethanol
Urea/NO
sthanol/I

Added tXT 1 "
x) = 0.62 X-
VOx) = 0.30....Z^

L i

1250 1350 1450 1550 1650 1750 1850 1950
Flue Gas Temperature, F

Effect of Methanol Injection on NH3 Slip

OPERATIONAL CHARACTERISTICS AND DISCUSSION

Despite rapid flue gas temperature fluctuations of up to 100 F per minute observed at
KVA Basel, the Two-Stage DeNOx Process provides relatively uniform, reduced levels
of outlet NOx concentration. In Figure 5, typical strip chart records of gas analyzer NOx
signals show variations from 150 + 40 ppm (without chemical injection) to 40 ± 20 ppm
(with chemical injection).

The 160 dppm initial NOx level at KVA Basel is relatively low. At higher NOx levels, the
percentage NOx reduction increases for a given urea/NOx molar ratio. In higher NOx

8-24

-------
Operating Time, Hours

Operating Characteristics From Chart Records

level applications, the Two-Stage DeNOx Process is expected to operate with (a) lower
urea ratio, and/or (b) reduced injector system flowrate at a given DeNOx performance
level. It should also be noted from Figures 2,3, & 4 that ammonia slip increases sharply
when DeNOx is increased above about 65%. Whether or not this is an unavoidable
consequence of fundamental NHi-radical chemistry remains an open question.

In the KVA Basel boiler, there is a relatively high degree of stratified flue gas flow which
requires that high levels of chemical mixing be provided by the injection system. This
is why the injector airflow requirement is as high as about 3% of the total flue gas mass
flowrate. "typical electric utility boilers have a much higher level of natural turbulence,
so that the injector system flow requirement is less. For example, typical designs for
utility boilers show that the Two-Stage DeNOx Process achieves performance levels
similar to those observed at KVA Basel with an injector system flow requirement
between only 0.6% and 1.0% of the flue gas mass flowrate.

Proprietary injector nozzles are an important part of the Two-Stage DeNOx Process.
Boiler flow models and computer simulations are used to optimize the design of specific
injection systems. The Two-Stage DeNOx Process is based on simple engineering
principles which are applied to specific boiler configurations using sophistocated
design techniques. Proprietary chemicals are not used.

Further commercial development includes award of a contract to install the Two-Stage
DeNOx Process on a 325 MW coal-fired boiler in West Germany. The startup date is in
August, 1989. and we plan to provide additional details in a forthcoming paper.

8-25

-------
EFFECTS OF CATALYST DEVELOPMENTS OF THE ECONOMICS
OF SELECTIVE CATALYTIC REDUCTION

T. E. Gouker, J, P. Solar, and C, P. Brundrett*
W. R. Grace & Co.-Conn., Research Division
7379 Route 32
Columbia, Maryland 21044

ABSTRACT

This paper reports a quantitative analysis of the economics of SCR based on recent
commercial developments in West Germany. The costs of selective catalytic reduction
(SCR) have dropped significantly since the technology was introduced in Japan in the
1970's. The drop in SCR cost is shown to result from catalyst price reductions,
increases in catalyst lifetime, and improvements in catalyst activity. This analysis
suggests that any further significant cost reduction will have to come from
improvement in catalyst activity. A Grace proprietary formulation which has a 50%
increase in activity, resulting in a 20% decrease in both operating and capital
costs, will be described.

INTRODUCTION

Selective catalytic reduction (SCR) is now widely used for the removal of nitrogen
oxides from power plant stack gases in both Japan and West Germany and is a leading
candidate technology to meet potential future emission standards in the U.S. The
costs of implementing this control technology have been decreasing steadily since SCR
was first commercialized in the 1970's and costs have been significantly reduced in
the 1980's. Both-catalyst price and installation cost have dropped during the
installation of-SCR facilities in West Germany. In addition, improvements in
catalyst activity have- occurred, leading to additional reductions in both capital and
operating costs. Since the economics of SCR have not been addressed in the U.S.
since the early 1980's1, it is appropriate to reevaluate the economics of SCR prior
to its possible introduction in the U.S.

Process costs are shown to be highly sensitive to both- catalyst cost and performance.
Factors evaluated include reductions in .catalyst price, improvements in catalyst
activity, and increases in the useful lifetime of the catalyst. When combined, these
factors have decreased the current estimated operating cost of SCR by 50% over the
levels anticipated when legislation was enacted in West Germany. In addition, this
analysis suggests that research on additional improvements in catalyst activity,
rather than reductions in price alone, offers the largest potential opportunity to
further impact. SCR economics. Grace has identified a proprietary catalyst
formulation which has 50% higher activity. It has been calculated to decrease both
the capital and operating costs of SCR by 20%.

* W.R. Grace and Co.-Conn., Davison Chemical Division
10 East Baltimore Street
Baltimore, Maryland 21202

8-27

-------
BASE CASE

As a basis for comparative evaluation, the West German experience in SCR
retrofitting has been summarized and reflected for the case of a 500 MW, coal-fired
power plant. The boiler in this case has first been retrofit with a low NO burner
at a cost of 0.06 Pf/kWh. Thus, at an uncontrolled NO emission level of l¥00
rag/m3 (535 ppm), a low NO burner reduces emissions to roughly 800 mg/m3 (390 ppm).
Installation of SCR at an 80% removal efficiency results in a final NO level of
160 mg/m3 (78 ppm), which is below the regulated level of 200 mg/m3 (9^ ppm). For
the cases considered in this paper, the SCR unit is located in the high-dust
position under typical commercial conditions. At a flue gas flow rate of 1.55 nun'
Nm3/hr, a temperature of 38Q°G, and a sulfur concentration of 2000 mg/m3 (700 ppm),
ammonia slip is limited to 5 ppm. The catalyst used in the base case is a
titania-based 7.4 pitch honeycomb representative of those used in current West
German installations.

The base case, shown in the table, thus describes the averaged costs for
installation and operation of such a facility in West Germany. Although catalyst
prices of 45,000 DM/ma were common in the early 1980's when West German utility
companies first considered SCR installations, an average price based on recent
sales of 25,000 DM/m3 has been used in this case. In addition, the current
catalyst replacement strategy has been incorporated, resulting in an average
catalyst life of four years. The capital investment of 55 Mio DM and operating
costs of 0.56 Pf/kWh are consistent with other studies of West German
installations2'3.

CATALYST COST

The most significant impact on SCR economics during its introduction in West
Germany has been the large reduction in catalyst price. As shown in Figure 1, the
45% reduction experienced in catalyst price between 1984 and 1988 has led to about
a 25% reduction in the cost of operating an SCR facility. For a facility installed
today, there would be a corresponding reduction in capital costs for the initial
catalyst charge. The dramatic decrease in catalyst price can be attributed to two
factors: replacement of high-price Japanese suppliers with modern, efficient
manufacturing plants in Europe, and the formation of a competitive market. The
competitive market was driven by an over-supply of production capacity in the early
years of the West German market. It is not yet clear whether the current price of
25,000 DM/m3 used in the base case is at market equilibrium or whether the catalyst
producers can operate profitably at this level. However, further reductions in
price are unlikely due to the inherent, raw materials and manufacturing costs' of the
current titania-based catalysts in commercial use. Even if additional reductions
were possible, the impact of further price drops Is diminished relative to what has
already occurred. Figure 1 shows that an additional 10% cut to 22,500 DM/m3
results in a savings of 4% in total operating cost.

CATALYST LIFE

Significant extension of the average catalyst' life, which acts to reduce SCR
operating cost, has also occurred during the 1980's. Early assumptions of a two
year catalyst life have been increased to three or four years based on experience
in Japan. At a guarantee of three years of acceptable activity, innovations in
catalyst replacement strategy, through a staged replacement schedule of one layer
of the reactor inventory annually, allow for an average catalyst life of at least
four years. Figure 2 shows the cost impact of the catalyst life out to eight
years. Compared to the base case of four years, an average life of two years had a
38% higher operating cost. Further life extensions through six to eight years,
however, provide a diminishing return.

8-28

-------
CATALYST ACTIVITY

The third factor affecting the cost of SCR, catalyst activity, has received a great
deal of attention due to the reduction in installed catalyst volume possible with
increased activity. deduction of catalyst volume not only reduces o p e i" a t x ng costs,
but also capital requirements due, for example, to the use of a smaller reactor and
I.D. fan along with less catalyst. A number of improvements have been identified
recently and their impact on costs is shown in the cases presented in Figure 3.
Two approaches have been taken to increase the activity of a given reactor volume:
incorporation of higher geometric surface area, and improvement of catalytic
performance within the walls of the catalyst. Increased geometric surface area has
been obtained by reducing the pitch of the honeycomb from 7.4 to 6,7. This
increases the number of channels from 400 to 484 within each element. To prevent
dust plugging, the catalyst wall thickness has been reduced from 1,4 mm to 1.0 mm.
This change in catalyst pitch has resulted in a 10% reduction in both capital and
operating cost.

Improving the performance of catalytic materials also results in volume reductions
for an SCR reactor. Research has shown that, due to gas diffusion limitations,
only the ca. 100 microns nearest the catalyst surface are utilized in the current
catalysts. Through pore structure optimization, access to the catalyst interior
can bg made more effective without reducing the number of catalytically active
sites . Catalyst formulations with a 15% improvement in activity are being
demonstrated m West German pilot plants and Grace has identified formulations
which increase catalyst activity by 50% over the technology used in the base case.
These improvements translate into a 13% and 33% reductions, respectively, in
required catalyst volumes. As can be seen in Figure 3, activity improvement of up
to 50% result in additional reductions of 20% in both capital and operating cost.
The 50% activity improvement of Grace's new catalyst formulation also provides the
potential for further increases in catalyst life, providing for an additional
savings in operating cost.

ACKNOWLEDGEMENTS

The authors would like to acknowledge the assistance of Professor Otto Rentz for
his help in obtaining references to the West German SCR experience.

REFERENCES

1 Bauer, T.K., et al. Selective Catalytic Reduction for: Coal-Fired Power
Plants; Feasibility and Economics, Stearns-Roger Engineering Corporation,
Denver, Colorado, EPRI CS-3603, October, 1984.

2 Schonbucher, B, "Costs of a DENOX Plant on the Basis of the SCR Process,"
Proceedings of the Workshop on Emission Control Costs. Ed. 0. Rentz, et. al.
Karlsruhe; Institute for Industrial Production, University of Karlsruhe,
1987, pp. 279-296. ,

3 SchSrer, B., et al. "Costs of Retrofitting Denitrification." Proceedings of
the Workshop on Emission Control Costs. Ed. 0. Rentz, et. al. Karlsruhe:
Institute for Industrial Production, University of Karlsruhe, 1987, pp.
326-358.

4 Beeckman, J.W. and Hegedus, L.L. "Design of Monolith Catalysts for Power
Plant NO Emission Control." Proceedings of the 1988 Annual Meeting of the
American Institute of Chemical Engineers. Washington, D.C., December 1988,
72 e.

8-29

-------
FIGURE 1

Effect of Catalyst Price

20 ¦

Operating Cost, Mio DM/yr

10 '

piMgspf

•,ss 4

His

Pftiii

§:K;t

W XV *

« ' Xv

45,000

1984

ill

A ^ \J

ia«L

35,000 25,000

Catalyst Price, DM/m3

22,500

FIGURE 2

Effect of Catalyst Life

Relative Operating Cost

0 H

1 2 3 4 5 6 7 8

Average Catalyst Life, years

8-30

-------
FIGURE 3

Effect of Catalyst Improvements

7.4 Pitch 6,7 Pitch 15% Increase 50% Increase

Catalyst Improvement

HI Capita! HH Operating

TABLE 1
SUMMARY OF BASE CASE COSTS

Capital Investment Operating Oosts

Mio EM

Mio m/yr

Catalyst

Variable (^Derating Costs

Reactor/Controls

Catalyst

5.21

Ammonia system

Ammonia

0. 66

Cuct Work/Dampers/

Power/Steam

1.05

I.D. Fan

Subtotal

6.91

Air Preheator

Site and Steel Work

Fixed Operating Costs

Electrical

Operating labor

0.37

Subtotal

Maintenance -

Material arid labor

1.12

General Facilities

MnunistxatioryTax/

Engineering

Insurance

0.97

Construction Interest 2

Subtotal

2.46

Capital Servicing

Depreciation 2.80

Interest 1.50

Subtotal 4,30

Total Operating Cost 13.67

8-31

-------
\

jte

jAf

\ J?;

(Intentionally BlaM}

8-32

-------
(1)

FULL SCALE DEMONSTRATION OF ADDITIVE NQ2 REDUCTION
WITH DRY SODIUM DESLLFURIZATION

Verle V. Bland
KVB, Inc.

Denver, Colorado 80226

ABSTRACT

Recent work in the area of dry sodium injection for flue gas desulfun"zation, has
underscored the effectiveness of th'is technology for use in low-sulfur coal
applications in the power industry. Dry sodium injection, however, results in
increased concentrations of NO? in the flue gas exiting the scrubbing process.
Higher NO? concentrations at tne stack can result in an undesirable brown plume
coloration which could considerably reduce the desirability of the technology to
potential users.

Several additive compounds successful in controlling NQ2 formation from sodium
injection, have been identified in the laboratory and evaluated in a fabric
filter pilot plant system. This paper reports results of full scale testing of
three such additive compounds (urea, anhydrous ammonia, and ammonium bicarbonate)
at the City of Colorado Springs R.D. Nixon Generating Station Dry Sodium
Injection Demonstration¦Faci1ity.

INTRODUCTION

In Phase I of the Sodium Injection Project at the R. D. Nixon Power Plant
(completed in April, 1986), dry injection of sodium compounds into the flue gas
stream prior to the fabric filter, was shown to be an effective and economical
S02 removal technique. Only one major problem was revealed during the test work,
namely, the appearance (at the stack exit) of a- faint reddish-brown plume
whenever the SO2 control level exceeded 50%. At the time, it was theorized that
the cause of the visible plume was increased NOg emissions resulting from the dry
sodium desulfuri zation process.

This phenomenon had been previously observed during related E PR I research at the
Arapahoe pilot plant test facility. This research, conducted by KYB, had
identified several additive compounds that exhibited a high potential for
successfully controlling this NOj emissions problem. Therefore, the purpose of
Phase 2 of the R. D. Nixon Sodium Injection Demonstration Project was to further
evaluate the use of these NOg reducing additives on a full scale fabric filter
installation.

8-33

-------
PROCESS DESCRIPTION

In the sodium desulfurization process, sodium bearing crystalline compounds
(normally bicarbonate or sesquicarbonate) are pulverized and injected directly
into a coal combustion flue gas. Once the sorbent enters the gas stream, S0£
diffuses to the particle surface where it reacts to form sodium sulfate.

Lower NQX emissions also result from the dry injection of sodium compounds. The
mechanism by which this occurs is not well understood, but the demonstrated
capability of this process to reduce NCL levels could be significant. One
serious side-effect, however, does result from this N0X reduction. This side-
effect is an increase in NOg emissions across the dry scrubbing process; and
since NOg is a reddish-brown gas, its presence can create an undesirable
coloration to the power plant exhaust plume.

During previous pilot scale EPRI research, a number of additives were identified
as possessing potential in reducing this plume coloration problem. These
additives were;

• anhydrous ammonia CNH3)

• ammonium bicarbonate (NH4HCO3)

• urea (N^CQNHg)

Anhydrous ammonia appeared to aid in solution of the problem by reacting directly
with NOg in the gas stream, and, although not previously evaluated for this
purpose, it was assumed that ammonium bicarbonate would act in somewhat the same
manner. Urea, however, appeared to work not only by releasing ammonia to react
directly with NOg, but by inhibiting the actual process that produces NQo in the
first place. While the mechanisms are still not well understood and further
evaluation would be desirable, the Arapahoe results provided the incentive to
evaluate these compounds at the full scale.

SYSTEMS DESCRIPTION

The host facility for this demonstration was Unit No. 1 of the City of Colorado
Springs Ray D. Nixon power plant. The facility is a 222-MW pulverized, low-
sulfur coal unit employing two parallel fabric filter baghouses- for particulate
control.

The sodium injection system has a 100+ MW capability for treating flue gases
entering one of the two baghouses. The remaining baghouse flue gas was untreated
and as such served to provide valuable baseline information during the test
project.

Two NO2 control additive injection systems were installed at the Nixon plant.
One system supplied superheated ammonia vapor to the sodium desulfurization
process; the other system supplied either urea or ammonium bicarbonate as a
pulverized solid.

N02 REDUCTION ADDITIVE EVALUATION

Baseline tests were first completed with both sodium sesquicarbonate and sodium
bicarbonate to establish N02 emission levels (Figure 1). Then, tests with the

8-34

-------
three additive compounds (urea, ammonia, and ammonium bicarbonate) were conducted
over a range of additive feed rates, sorbent feed rates, and flue gas
temperatures. Although variable, nominal conditions for the majority of these
tests were:

flue gas temperature = 280°F

inlet S02 9 3% O2 = 400 ppm

inlet NO @ 3% 0g =350 ppm

sodium sesquicarbonate NSR = 1.5
sodium bicarbonate NSR = 0.9

To establish performance curves, additive feed rates ranging from 1 to 20 percent
by weight of the injected sodium sorbent were evaluated. Figures 2 and 3 show
these results. The maximum feed rate tested was different for each additive.
For urea, 100% N02 removal was attained at feed rates from 11 to 12 percent,
making additional injection unnecessary. For NH3, the maximum addition rate was
10% (75% M02 removal) due to a vaporizer system limitation. However, only a
marginal increase in NQ2 removal was experienced by increasing the ammonia
addition rate from 7 to 10 percent, indicating higher rates may not have resulted
in additional NO? reduction. Finally, the maximum ammonium bicarbonate (ABC)
rate of 20% was dictated by mechanical throughput limitations of the dry additive
feed system. N02 removal of only 41% was observed for ABC at this maximum feed
rate.

Since an N02 concentration of approximately 30 ppm was identified as the visual
threshold from the R. D. Nixon stack, reducing emissions below this level was
desirable. The visual threshold for NQ2, at other locations, would depend on a
number of factors and thus is site-specific. For S02 removals ranging from 70 to
90 pecent, the dry sodium scrubbing process resulted in an outlet NO,
concentration averaging 50 ppm. To be safely below the visual threshold, a
reduction in N02 of approximately 50% was required. The data in Figures 2 and 3
show that urea and ammonia were about equally effective in attaining this goal.
An addition rate of approximately 3.5% of either compound resulted in 50% NQ2
removal. ABC, however, was. not as effective. An ABC addition rate of 3.5% only
resulted in a 10% reduction of the NQ2 emissions.

Overall system NQX emissions were also affected by additive injection. N0X
removal typically, but not always, increased with the use of each additive.

Figure 4 shows comparable data for urea, ammonia, and ABC injection. For these
tests, an increase in NGX removal from 3 to 30% was observed.

CONCLUSIONS

Ttm project demonstrated, on a full scale utility boiler, that the use of
additives can successfully eliminate the brown plume problem associated with dry
sodium injection. An addition rate of approximately 3.5% (lb additive/lb
sorbent) of either urea or ammonia reduced the N0g from 50 ppm to 25 ppm (below
the visual threshold for the Nixon plant). Ammonium bicarbonate, however, was
not very effective. A 3.5% additive injection rate resulted in only 10% ri02
removal.

REFERENCE (1):

Bland, V. V., KVB, Inc., J. J. Hammond, Colorado Springs Dept. of Utilities,
R. G. Rhudy, EPRI, "Full Scale Demonstration of Additive NQ2 Reduction with Dry
Sodium'Desulfurization," presented at First Combined FGD and Dry S02 Control
Symposium, Cosponsored by EPA and EPRI, October 25-28, 1988, St. Louis, M0.

8-35

-------
no2 reduction additive evaluation

Baseline tests were first completed with both sodium sesquicarbonate and sodium
bicarbonate to establish NO2 emission levels (Figure 1). Then, tests with the
three additive compounds (urea, ammonia, and ammonium bicarbonate) were conducted
over a range of additive feed rates, sorbent feed rates, and flue gas
temperatures. Although variable, nominal conditions for the majority of these
tests were;

To establish performance curves, additive feed rates ranging from I to 20 percent
by weight of the injected sodium sorbent were evaluated. Figures 2 and 3 show
these results. The maximum feed rate tested was different for each additive.
For urea, 100% NOg removal was attained at feed rates from 11 to 12 percent,
making additional injection unnecessary. For Nl-U, the maximum, addition rate was
10% (75% N0£ removal) due to a vaporizer system limitation. However, only a
marginal increase in NO2 removal was experienced by increasing the ammonia
addition rate from 7 to 10 percent, indicating higher rates may not have resulted
in additional NO, reduction. Finally, the maximum ammonium bicarbonate (ABC)
rate of 20% was dictated by mechanical throughput limitations of the dry additive
feed system. NO2 removal of only 41% was observed for ABC at this maximum feed
rate.

Since an N0£ concentration of approximately 30 ppm was identified as the visual
threshold from the R. D. Nixon stack, reducing emissions below this level was
desirable. The visual threshold for NO2, at other locations, would depend on a
number of factors and thus is site-specific. For S02 removals ranging from 70 to
90 pecent, the dry sodium scrubbing process resulted in an outlet N0o
concentration averaging 50 ppm. To be safely below the visual threshold, a
reduction in NO2 of approximately 50% was required. The data in Figures 2 and 3
show that urea and ammonia were about equally effective in attaining this goal.
An addition rate of approximately 3.5% of either compound resulted in 50% N0£
removal. ABC, however, was not as effective. An ABC addition rate of 3.5% only
resulted in a 10% reduction of the N0g emissions.

Overall system N0X emissions were also affected by additive injection. N0X
removal typically, but not always, increased with the use of each additive.

Figure 4 shows comparable data for urea, ammonia, and ABC injection. For these
tests, an increase in NQX removal from 3 to 30% was observed.

This project demonstrated, on a full scale utility boiler, that the use of
additives, can successfully eliminate the brown plume problem associated with dry
sodium injection. An addition rate of approximately 3.5% (lb additive/lb
sorbent) of either urea or ammonia reduced the NOg from 50 ppm to 25 ppm (below
the visual threshold for the Nixon plant). Ammonium bicarbonate, however, was
not very effective, A 3.5% additive injection rate resulted in only 10% N0g
removal.

flue gas temperature
inlet S0n @ 3% 02
inlet NO @ 3% O2

sodium sesquicarbonate NSR = 1.5
sodium bicarbonate NSR = 0.9

= 280°F
= 400 ppm
= 350 ppm

CONCLUSIONS

8-36

-------
100

§S

00'

60-

40 •

¦ Present Project lasults
A Past Project Its tills

A ^

A £& §

0+-

—r-

T-

—r-

Figure

100

$02 Removal. %

Comparison of Phase 1 and Phase 2 Outlet N02 Emissions
Versus S02 Removal for Sodium Sesquicsrtxmate Injection,

"O#"

100

60-

o#
. »%

Ure®

40-

20-

• O
• 0

0.00

O Bicarb Data
• Sesqul Data

0,05 0.10 0.15

Additive/Sorbent (Lb/Lb)

Figure 2. NG2 removal as a function of additive to
sorbent mass ratio for urea.

0.20

8-37

-------
I

100

A Bicarb-NH3

A Sesiju(-NH3

O Bicarb-ABC

¦ Ses
-------
SHELL PROCESS FOR LOW-TEMPERATURE NOx CONTROL

Frans Goudriaan and Carl M.A.M. Masters*

Koni nklijke/Shel1-Laboratori um, Amsterdam
(Shell Research 8.V.)

Badhuisweg 3, 1031 CM Amsterdam, The Netherlands

and

Rene Samson

Shell Internationale Chemie Maatschappij B.V.
The Hague,' The Netherlands

ABSTRACT

This paper describes a process developed by Shell for the selective catalytic
reduction of nitrogen oxides.

A vanadium/titanium-on-silica catalyst has been developed for this process. The
stability of the catalyst is excellent.

The high catalyst activity allows operation at temperatures as low as 130 to
200 °C. The lower temperature limit is determined by the amount of sulphur dioxide
the flue gas.

To minimise pressure drop and to permit operation with dust-containing flue gases
improved version of the Parallel Passage .Reactor has been developed.

Several commercial projects are under way.

INTRODUCTION

NOx emission must be reduced drastically in the years to come, as an essential
contribution to a cleaner environment. In both the USA and Europe the legislation
on N0x levels in flue gases from stationary installations is tending to become
more restrictive. In most cases, primary reduction methods such as the use of
low-NOx burners will not in future suffice. Increasingly, therefore, it will be
necessary to adopt secondary methods, such as Selective Catalytic Reduction (SCR)
with ammonia.

This paper gives the status of the Shell SCR process for low-temperature deNOxing.

CATALYST; LABORATORY WORK

At the heart of the Shell SCR process is a highly active and selective catalyst.
In a previous publication (1) the development of an optimally impregnated vana-
dium/titani um-on-sil ica catalyst has been described. The carrier is Shell silica
with a surface area of 300 m2/g and a pore volume of 0.9 ml/g. It is shaped either

* Presently with Shell Development Co., Houston, Texas, USA.

8-39

-------
as spheres (S-995) or as granules (S-095), the size of which is adapted to the
application concerned.

Further development has resulted in a patented catalyst which is easier to
manufacture and has a higher stability, whilst the high activity is maintained.
The effect of the vanadium and titanium load is shown in Figure 1. The deNOxing
activity here is given as the percentage NOx conversion in a simple laboratory
catalyst ranking test, details of which are given in Table 1. Figure 1 shows
clearly the combined effects of titanium and vanadium content. At a constant V
load the activity can be greatly increased by increasing the Ti load and vice
versa.

An in-depth study of the structure of the catalyst has enabled us to pinpoint the
essential features of an active V-T1/Si02 deNOxing catalyst:

* Both Ti and V are very highly dispersed, as evidenced by XRD/SEM/TEM, Almost
no crystalline V205 or TiO, (anatase or rutile) are present.

* Both Ti and V are intimately mixed with Si02 (XPS analysis).

* Part of the Ti is present in a four-fold coordination with oxygen (i.e., not
six-fold as in anatase or rutile).

* Part of the V is present as V4+. These species are stabilised by titanium.

* The resulting surface area of the catalyst is relatively high (150 - 250 m2/g)

CATALYST PERFORMANCE IN FLUE GAS

A laboratory study has been conducted with S-995 and S-095 catalysts in fixed-bed
operation, the gas treated being a flue gas composition representative for
gas-fired installations:

At 160 °C and a GHSV of 25 000 Nm3/(m3.h) the NH3/N0^ ratio was varied between
0.3 and 0.9. The kinetics of the deNOxing reaction with NH3/N0x <1 are complicate
since each one of the components NO, NH3, 02 and H20 is involved. The reaction
order in NO is close to 1 while the order in NH3 appears to be lower than 1.

Water has a strong inhibiting effect. A typical result, demonstrating the high
activity of the catalyst, is that, at a NH3/N0x ratio of 0.5, the conversion
of ammonia was 58 %.

A semi-industrial test facility was constructed as a slip-stream unit in the flue
gas system of a commercial gas-fired ethylene cracker plant. A test run was
performed with some 300 1 of S-995 catalyst (3 mm spheres) in fixed-bed reactor
configuration. The run length was 6 months and excellent activity and stability
were observed. Typical performance data are:

The presence of sulphur dioxide in the flue gas is a major factor in determining
the lowest possible operation temperature. According to the literature, at a
given temperature and given water and ammonia concentrations it is the amount of
sulphur trioxide which determines whether deposition of solid ammonium (hydro)-
sulphate will occur. Apart from traces of SO, already present in the flue gas,
the main proportion of sulphate seems to be formed from SO, over the deNOxing
catalyst. Preliminary studies indicated the tolerable sulphur dioxide levels at

0, 2 %v

H,0 15 %v

CO, 7 %v

N2 76 %v

Temperature,
GHSV,

NH3/N0X ratio
NO conversion
NH3 slip,

°C 135 - 155

NmV(m3 .h) 2500 - 3500

0.6 - 0.8

% 60 - 80

ppmv (dry basis) < 5

8-40

-------
various temperatures. They are given in Figure 2. The poisoning of the catalyst
by ammonium (hydro)sulphate is reversible. The activity can be restored by opera-
tion at increased temperature.

REACTOR

For applications where dust is present in the flue gas and/or pressure drop must
be minimised we have developed a patented Parallel Passage Reactor. Its prin-
ciples are shown in Figure 3. It has been tested in both laboratory and semi-
industrial installations with low-dust flue gas.

A study of the mass transfer steps involved is in hand. This provides background
for the development of design principles for parallel passage reactors for
different applications. Design variables are the reactor dimensions, relative
catalyst and gas volumes in the reactor, the exchange surface area, catalyst
particle size, etc. They can be adapted to the specific application depending on
such factors as dust level, tolerable pressure drop, temperature, required
deNOxing conversion, and ammonia slip.

APPLICATIONS

The first commercial Shell deNOxing unit will be installed in an ethylene cracker
plant in West Germany, to come on stream in 1989. Here the low-temperature
deNOxing technology will be applied.

We intend to extend the applications to other fields. The Shell low-temperature
technology may be of particular interest where removal of dust and sulphur
dioxide is performed upstream of the deNOxing unit. Overall optimisation of
processes for the removal of sulphur, dust and N0X appears to be an attractive
possibility.

REFERENCE

1. M.J. Groeneveld, G. Boxhoorn, H.P.C.E. Kuipers, P.F.A. van Grinsven,

H. Gierman and P.I. Zuideveld. Proceedings of the 9th International Congress
on Catalysis, Calgary, Canada, 1988, Vol. 4, p. 1743 (paper 223-A).

GLOSSARY OF ABBREVIATIONS

GHSV, Nm3/{m3.h) Gas hourly space velocity

SCR

Selective Catalytic Reduction

SEM

Scanning electron microscopy

TEM

Transmission electron microscopy

XPS

X-ray photoelectron spectroscopy

XRD

X-ray diffraction

8-41

-------
Table 1

STANDARD ACTIVITY TEST FOR DENOXING CATALYSTS

Equipment: Fixed-bed tubular reactor;

catalyst volume 60 ml

Test gas composition:

Reaction conditions;

0, 5 %v
N, 76 %v
Cu, 13 %v
H20 6 %v

NO ~ 450 ppmv
NHj/NO ratio = 1.5(v/v)

Temperature 150 °C
GHSV 4000 Nm3/(m3.h)

(basis catalyst
volume)

Conversion, %

V content, %w

Figure 1. $-095 catalyst: effect of metal load
on deNOxing activity (% conversion) in standard
activity test

8-42

-------
1000

S02 in f lue gas, pprrw

100

z<£$

Deactivation

+ + + None
- Slight
Serious

150

200

GHSV = 4000 Nm3/{m3. h>;
H20 6 %v; 02 5%v;
NH3/NOx ratio 0,6- 1.5

I L_

2m 300

Temperature, °C

Figure 2. Deactivation of $-995 deNOxing catalyst by S02

Flue gas

Schematic view of a
vertical cross—section

Principles

- Stationary catalyst

Wire mesh - Gas + dust flows

through open channels

Catalyst beads
or granules

Reactants transported
radially to catalyst,
(mainly) by diffusion

Figure 3. Parallel-Passage Reactor

8-43

-------

. £

t,.

.¥

(Intentionally Blank)'*

•if

8-44

-------
NO Reduction of Waste Incineration Flue Gas
x

Bo Herrlander
Flakt Industriella Processer AB
3-351 87 VSxjO

ABSTRACT

o o

Catalysts for NCx reduction have been tested at 610 F and 480 F in a tail end
SCR-system after a waste incineration plant equipped with a CDAS-system for
acid gas removal. The catalysts operated at 610 F showed no deactivation after
2300s hours but the catalyst operated at 480 F showed a slight deactivation.
The ammonia injection control system was able to follow the very quick
concentration variations after waste incineration. No formation of dioxines or
laughgas over the catalysts was proven. The operation of the plant was trouble
free and there was no fouling of the heat exchanger surfaces,

BACKGROUND

Waste incineration with effective flue gas cleaning is a route increasingly
used to get rid of waste, mainly due to lack of space for landfill. Flue gas
cleaning has up till now focused on acidic gases and heavy metals as well as
dioxines removal. However, the NO emission from a waste incineration plant is
considerable. The specific emission (lb/Btu) from waste incineration, is in
the same range as for coal fired power plants. Further the operating time per
year is often very high. NO emissions contribute to the acid deposition as
well as photochemical smog.

INTRODUCTION

The waste incineration plant Munich South in Bavaria has to meet very string-
ent emissions. The two incinerators will take 40 tons/hour waste each, crea-
ting a flue gas flow of 141 260 scfrn. The energy released is used to produce
electricity as well as district heating. The plant is operated more than 6 500
h/year. The requirement for flue gas cleaning is HCL-removal 99.2%, SO^-
removal 95.2%, NO -removal 89% and particulate removal 99.95%. The correspon-
ding emissions are HCL less than 0.004 gr/scf, SO less than 0.011 gr/scf, NO
less than 0.031 gr/scf and particulates less than2Q.Q04 gr/scf. The flue gas X
cleaning is a retrofit and will be started January 1990. The acidic gases and
particulates will be removed in a conditioned' dry adsorption system (CDAS) and
the NO will be removed in a SCR-system (DSNIT), The CDAS-system is operated
with hydrated lime and consists of a fluidized bed reactor and a fabric
filter. The DENIT-system is connected after the CDAS-system to avoid catalyst
poisons.The DENIT-system performance is highly integrated with the catalyst
performance. In order to demonstrate an acceptable catalyst lifetime and
system performance especially the ammonia injection control, the Kempten pilot
plant was built.

I Preceding page blank j 8-45

-------
SCR PILOT

In unit No.3 in Kempten the waste is fired in a conventional stoker furnace
followed by a steam boiler where the flue gases are cooled down to about
390°F. The steam boiler is followed by the conditioned dr^ adsorption reactor
where the flue gases are further cooled down to about 280 F and fabric filter.
From this point a slip stream of ISOOscfm is diverted to the pilot plant. The
pilot plant cosists of a heat ^xchanger and 3 parallel reactor lines. Each
reactor has a capacity of 7 ft of catalyst. After the heat exchanger, an
electrical heater brings the temperature up to final reaction temperature.

Just after the1 electrical heater, before the reactor, ammonia is introduced in
a nozzle arrangement. A static mixer ensures a good mixing and even distribu-
tion of ammonia and NO . In the reactors the gas is evenly distributed over
the whole catalyst surface, by means of guidevanes. The reactor is equipped
with electrical heat tracing elements and good thermal insulation to avoid
heat losses and temperature gradients. The pilot plant was started February 2,
1988.

ANALYSER SYSTEM

Water vapour, nitrogen oxide, nitrogen- dioxide, sulphur dioxid and oxygen are
continuously monitored before and after the reactors.lt is also possible to
measure in between the catalyst layers. The sulphur trioxide and ammonia slip
are analyzed manually from gas samples.

PILOT TEST PROGRAM

Objectives

- To demonstrate acceptable catalyst life, i.e. deactivation rate.

- Reactor temperature optimization.

- Optimization of ammonia injection control to follow the very fast NO
concentration variations.

- Verify a trouble-free heat exchanger operation.

- Check for dioxine and N20 formation over the catalyst.

Conditions - technical data

Reactors in parallel
Nominal gas flow per reactor
Nominal reaction temperature
Inlet concentration at 10% 0,
NO

HcS

SO
SO^

Particulates

NO reduction degree
x

scfm

600

°p

480 - 610

ppir.

200-250

ppm

about 30

ppm

about 30

ppm

< 1

gr/scf

< 0.002

o,
"S

> 90

Catalyst activity is defined as k = - AVln {1-x)

AV = area velocity = fluegas flow/catalyst surface (ft"/ft

SV = space velocity (l/h) = AV • Y

Y = catalyst surface/volume

x = reduction degree

SR = stoichiometric ratio (NrL/NO 5

J x

(STJ/YJ^lnU-x) at SR = 1

8-46

-------
Catalysts

All tested catalysts are of parallel flow type, both honeycomb and plate
catalysts, have been tested in the pilot. The pitch varies between
1/8 - 1/6 inch. All tested catalysts are of high activity type.

TEST RESULTS

Catalyst response curves

Response curves for the three catalysts in the first test period are shown
below. Catalyst B shows an almost -theoretical dependency between mole ratio
and NO reduction at least up to a space velocity (SV) of 20 000 1/h. At low
SV-values the NH^-slip is very small but at a certain value the slip increases
very rapidly indicating a decrease in NO reduction from the theoretical
value. x

Catalyst A
t - 320 ° C

Mole Ratio (NH, / NO,)
+ SV = 18 000 - 20 000 1/h
Hsv < 12 000 1/h

Mole Ratio (NH3 / NO,)
~ SV < 12 000 1/h

Mole Ratio (NHa / NOJ
+ SV = 18 000 - 20 000 1/h
~ SV < 12 000 1/h

SV (1/h) (Thousands)
SR = 0.9 +/ - 0,05

SV (1/h) (Thousands)
SR = 0.9+/-0.05

SV (1/h) (Thousands)
SR - 0.9 +/ - 0.05

Catalyst deactivation

In the first test period the catalysts were operated with fixed SR-setting at
610 F for 2 300 hours. The catalysts were then removed and tested in a labora-
tory for determination of activity. The three catalysts did not show any
deactivation. In this application there is a fear for deactivation due to pore
condensation of salts, like ammonium sulphates and zink chloride. Also alkali
metals, mercury, arsenic, led etc. (1), may cause deactivation. It has been
shown at several plants, that the dry adsorption system'has a far better heavy

8-47

-------
metal removal than the wet scrubbing system (2). Fast catalyst deactivation
mainly from lead has been observed in test plants after wet scrubber systems.

In the second test period the temperature was lowered to 480°F. The catalysts
in this test run showed a minor deactivation loss. After 2 200 hours the
catalysts show between 98-99% of their virgin activity. It was however not
possible to fully modify the catalyst activity in order to compensate for the
lower operating temperature. The system economics for the lower operating
temperature alternative is therefore questionable.

Ammonia injection control system

It was possible to show that the ammonia injection control system was able to
follow the very quick concentration variations as well as load changes. This
is shown in the figure below. A load change up and down is simulated by a
momentare flow change. The concentration variation maybe as quick as 0.5 ppm/s
without any load changes. This change may continue for several hundred
seconds. If the load change is taken into consideration, the actual change
might be as high as 1 ppm/s. The catalyst will adsorb the ammonia. This
adsorption will be highest on a virgin catalyst and gradually decrease with
deactivation. The chosen control system consists of a feed-forward trim back
ammonia injection control system.

I13

%¦

412-

0 f-

* - * ^

—6

73;

I -«

100

¦3*

"V |~253^_r

\ N00y,
0-300 ppm

NH,
0-912 g/h

0-300 ppm

13-

12"

11-

Rue gas
Flow 0-5000 Nm'/h

Dioxin formation

Dioxins were measured before and after the catalyst. On both sides the concen-
trations wepe very low (calculated in dioxin equivalents according to Eadon
< 4.4 • 10" gr/scf). The conclusion is that there is no production of
dioxins over the catalyst (3). It is a well known fact that dioxins may be
reformed between 1060 F and 1710 F (4, 5), may be catalyzed by fly ash. In
this case, however, both fly ash and HCL is removed to a very large extent.

Laugh gas

Gas samples were taken both from upstream and downstream of the catalysts and
analyzed for laughgas, N-0. No laughgas was found in these samples. Used
methods detection limit Ts 5 ppm. Also by others no laugh-gas has been
detected (6).

8-48

-------
Heat exchanger operation

The heat exchanger operation was trouble-free. Every inspection showed clean
surfaces and no dust build-up. A slight corrosion was observed due to dew
point passage with the following condensation at every start-up. The heat
exchanger is of a non-leakage type with the gas to the reactor going on the
tube side and gas from the reactor on the shell side.

REFERENCES

1. Energi Spektruir. vol. 6, June 1987, p. 50 - 59
"Indizien process"

2. Energi ur avfall, Statens Energiverk, 1986:6, ISBN 91-38-09453-3

3. Energi Miljo Konsult Malmo; "Measurements of dioxine and other chlorinated
aromatics at Flakt SCR DeNox-plant, Kempten, Bavaria" Malm5 1988-09-28

4. JAPCA, No. 12, vol. 37, December 1987, p. 1451 - 1462

comparison of Dioxine, Furan and combustion _Gas data frorn Test programs
at three MSW incinerators ¦ '

5. MKS Studsvik: "Measurements of dioxine and other chlorinated aromatics"
1987-11-24

6. BWK Bd. 37 (1985) No. 12, December, p. 465, "Kein N^O durch jacar. ische
Entstichungskatalysatoren"

8-49

-------
\

V .
\

r %

(Intentionally BlankK

; \

X .

( 8-50

-------
REBURNING AND REPOWERING FOR NO* CONTROL
ON LARSE UTILITY BOILERS

S. L, Chen, E. C. Moller, D. W. Pershing
Energy and Environmental Research Corporation
Irvine, California

and

A. B. Walters
Florida Power 8 Light Company
Juno Beach, Florida

ABSTRACT

Tests conducted in a 3 MW^ Pilot scale facility indicate that N0X emissions can be
reduced by at least 50 percent through the application of natural gas reburning,
even in oil-fired boilers, such as the Florida Power & Light's 800 MWe units, with
relatively tight residence time limitations. Particulate emissions were found to
be strongly dependent on burnout zone stoichiometry; however, above 2 percent
excess O2 the emission levels were identical to those achieved without reburning.
SO3 emissions were found to increase with the application of reburning, but this
effect can be partially mitigated by distributing the addition of the burnout air.
Repowering existing steam generation units with combustion turbines decreases
thermal N0X formation due to reduced local flame temperatures and provides a
mechanism for incinerating a portion of the N0X generated in the combustion
turbine. Repowering alone produces 35 percent reductions in N0X emissions and
repowering in conjunction with reburning results in over 70 percent N0X reductions.

REBURNING '

Combustion of fuel oil produces N0X, SOx and particulate emissions. Combustion
conditions, such as excess air, which tend to promote high carbon burnout and good
flame stability also promote the formation of N0X and SO3. Conversely, the control
of NQX emissions through control of the combustion process may have an adverse
impact on carbon burnout. Reburning involves the injection of secondary fuel and
air downstream from the main combustion zone. Reburning is an attractive concept
because it allows the primary combustion zone to be operated, independently from
the secondary reburning zone, at conditions which optimize flame stability and
carbon burnout.

An experimental study was undertaken to optimize the use of natural gas
reburning for the control of NCX as well as unburned carbon and SO3 emissions from

| Preceding page blank j 8"51

I • !

-------
oil fired boilers. Parametric experiments were conducted in a 3.0 MWt pilot scale
facility configured to simulate the thermal and emissions characteristics of an 800
MWe boiler at FPL, particularly the short residence time (approximately 500ms)
available between the upper burner and superheater tubes.

OPTIMUM AIR DISTRIBUTION

Figure 1 summarizes the influence of the excess oxygen concentration in the burnout
zone. N0X, particulate, and CO emissions are shown as a function of excess 02 in
the exhaust for both baseline (uncontrolled) oil combustion and for an equivalent
level of oil firing with an additional 20% natural gas reburning. NO* reductions
in excess of 50% were achieved at all excess oxygen levels; above 2% exhaust 02.
the application of natural gas reburning had no detrimental effects on either
particulate or CO emissions. Below 2% 0g, the application of natural gas reburning
resulted in increased particulate and CO emissions, almost certainly due to the
presence of the fuel rich reburning zone which produced unburned carbonatious
fragments that were not completely burned out at these low NO* excess oxygen
levels. The Figure 1 data indicate that the final exhaust oxygen should remain at
approximately 2% excess O2 to minimize thermal losses and N0X emissions while still
achieving good carbon burnout. Thus, the recommended stoichiometry distribution
for the application of reburning in limited residence time, oil fired systems would
be: SR]_ = 1.1, SRg = 0.9, and SR3 - 1.1.

This study revealed a previously unknown complication; the application of reburning
with oil firing can produce a significant increase in SO3 emissions if the system
is not fully optimized* Enhanced SO3 formation appears to result from
homogeneous oxidation of SO2 by 0 atoms which are produced during the low
temperature oxidation of CO in the final burnout zone. This effect can be
significantly mitigated by distributing the addition of the burnout air to provide
at least two approximate steps in the stoichiometry change from SR2 35 0.9 to SR3 =
1.1. Splitting the burnout air into two streams has essentially no effect on any
of the other emissions.

REP0WERING

Repoweri ng existing steam turbine generation units with combustion turbines (CT's)
to create combined-cycle operation can produce heat rate reductions of up to 18
percent and provide additional unit capacity with low capital expenditure relative
to the cost of completely new power plant additions''*. Combustion turbine
repoweri ng with a supplemental boiler is attractive from an N0X emissions
standpoint because it decreases thermal M0X formation due to reduced local flame
temperatures, and provides a mechanism for incinerating a portion of the N0X
generator in the combustion turbine.

Tests were conducted in a 3.0 MWt pilot scale facility to evaluate natural gas
reburning for N0X control in a partially CT repowered 800 MWe boiler at FPL. The
combustion turbine exhaust was simulated using a natural gas fired preheater. The
initial studies focused on characterizing the differences between normal and
repowered operation. Subsequent experiments were conducted to evaluate the
applicability of standard reburning for N0X control in the repowered configuration
and to assess the potential of advanced reburning concepts using selective N0X
reduction.

8-52

-------
Influence of Repowering

The effect of repoweririg on NO* emissions is significant. As illustrated in Figure
2, repowering alone produced a 35-percent reduction in N0X emissions on a lbs N0X/
MBtu basis with a No. 6 oil primary and reductions in excess of 70-percent with a
natural gas fired primary. With both fuels, a major portion of the reduction is
associated with primary boiler incineration of the N0X generated in the CT. Both
theoretical calculations and experimental measurements suggest that the temperature
profile in the boiler should be approximately 200°F cooler with a repowered
configuration, and this caused a significant reduction in NO* levels with a natural
gas primary, but had little influence when heavy oil was the primary fuel. Data
obtained also suggest that the dramatic N0X reduction with repowering is achieved
without detrimental effects on SO3 emissions or carbon burnout at excess oxygen
levels above 2-percent. Increasing the N0X levels in the CT exhaust increased the
overall N0X emissions-, however, this effect was somewhat mitigated by incineration
of approximately 4G percent of the CT N0X as it passed through the primary flame of
the boiler.

Application of Reburning

Reburning was also found to be effective for N0X control in the repowered
configuration; reductions of approximately 50 percent were achieved under typical
FPL CT-repowering conditions'3). In addition, carbon burnout was not influenced by
repowering with reburning, but SO3 emissions were lower in the repowered
configuration compared to the reburning case with air as the oxidant {due to the
lower oxygen partial pressure and lower temperature of the burnout zone). Use of
an advanced reburning configuration'4' in conjunction with (NH^SC^ injection
produced N0X emissions below 100 ppm in the repowering configuration while good
carbon burnout and low SO3 levels were also maintained as illustrated in Figure 3.
Overall, an 80-percent reduction from the uncontrolled emission levels was obtained
by using 10- or 20-percent reburning in conjunction with selective NGX reduction
with (NH4)2 SO4 and repowering.

REFERENCES

1. Holler, E. C-, S. I. Chen, D. W. Pershing and A. B. Walters. "Optimization of
Natural Gas Reburning for Controlling Pollutant Emissions From Oil Fired
Boilers", presented at Spring Meeting, Western States Section/The Combustion
Institute, March, 1988.

2. Walters, A. B., J. R. Craig, R. J. Cullen, R. E. Herbster, and J. A. Van Laar.
"Gas Turbine Repowering Considerations", Feasibility Study for Florida Power
and Light Company, 1987.

3. Holler, E. C., S. L. Chen, D. W. Pershing and A. B. Walters. "Reburning For
N0X Control in Combustion Turbine Repowering", presented at Fall Meeting,
Western States Section/The Combustion Institute, October, 1988.

4. Heap, M. P., S. L. Chen, J. C. Kramlich, J. M. McCarthy and D. W. Pershing.
"An Advanced Selective Reduction-Process for N0X Control", Nature, Vol. 335,
Issue 6191, 1988.

8-53

-------
Open-Uncontrolled Shaded-Standard Reburning

600

500

-------
Figure 2. NO, mass emissions.

»>
V

500

400

300

200

100

Shaded Bars Represent
Repowerlng arid Returning
+ (NH4)2 SO^ (aq) Injection

W
V

Cl
1)

m
c

o e

£X Ob

o
cl

s»

ja
02

Csj

u
9

Figure 3. Effectiveness of advanced reburnlng.

8-55

-------
Session 9
OIL AND GAS COMBUSTION APPLICATIONS

Chairman: A. Kokkinos, EPRI

9~ i

-------
RETROFIT OF AN ADVANCED LOW-NOx COMBUSTION SYSTEM
AT HAWAIIAN ELECTRICS OIL-FIRED KAHE GENERATING STATION

u. L. 3. Yee
R. B. Freitas
Hawaiian- Electric Company, Inc.
Honolulu, Hawaii

D. V. Giovanni
S. E. Kerho
Electric Power Technologies, Inc.
Berkeley, California

M. W. McElroy*

Electric Power Research Institute
Palo, Alto, California

ABSTRACT

Hawaiian Electric Company, Inc. (HECO) completed major combustion system
modifications in mid-1988 on Kahe Unit 6, a Babcock s Wilcox (B&W) oil-fired
unit rated at 146 MW. The modifications were undertaken to reduce emissions of
oxides of nitrogen (NOx) and particulate matter (PM), and to restore operational
flexibility that had been impaired with burners-out-of-service operation
previously used for NOx control. Modifications included B&W primary gas dual
register burners (PG-DK3), front-wall and rear-wall overfire air ports,
extensive ductwork; for the overfire air and recirculated gas flows, revision of
the oil atomizing steam system, and enhancements to the automatic burner control
system. The installation of the PG-DRB low-NOx burners represents the first
application of this type of combustion system to a utility boiler in the United
States.

In conjunction with the modifications, a comprehensive testing program under the
joint sponsorship of the Electric Power Research Institute and HECO was
conducted to characterize emissions and operations before and after the
modifications1. The test program, test results, retrofit costs, and a
preliminary extrapolation of results to industry-wide application are
summarized. HECO's operating experience with the combustion system
modifications in attempting to reduce NOx and PM are also discussed,

* Now with Electric Power Technologies, Inc., Menlo Park, California

9-1

-------
INTRODUCTION

In January 1979, the -Hawaiian Electric Company was granted a Prevention of
Significant Deterioration {PSD) permit by the U.S. EPA for the construction of
an oil-fired steam-electric generating unit (Unit No. 6) at the Kahe Generating
Station. The permit specified. NOx emission limits of 0.23 lb/MBtu (180 ppm at
3% 02) and particulate matter (PM) emission limits of 0.03 lb/MBtu. Since the
commissioning of Kahe 6 in 1981, HECO has undertaken numerous emission reduction
test programs to achieve compliance with these limits. These efforts, involving
operational modifications and flue gas recirculation to the windbox, resulted in
optimized NOx and PM emission levels- for continuous operation of the boiler with
the installed combustion equipment. Optimization included operation in a
burners-out-of-service (BOOS) mode with 6 of 9 burners in operation, and with
flue gas recirculation to the windbox, resulting in typical NOx emissions of
0.28 Ib/MBtu and PM emissions of typically 0.06 to 0.08 lb/MBtu (with excursions
above 0.9 lb/MBtu).

In July 1987, HECO contracted with The Babcock & Wilcox Company to design,"
install, and commission state-of-the-art retrofit low-NOx combustion technology
at Kahe 6. The emission guarantees were 0.23 lb/MBtu NOx as NO2 (specified for
a fuel nitrogen content of 0.5%) and 0.1 lb/MBtu for PM emissions.
Additionally, it was intended that the retrofit technology would allow a return
to all-burners-in-service (ABIS) operation, thereby improving the operating
flexibility and generating capabilities of the unit which had been impaired with
BOOS operation. Specifically, a higher turndown was expected from improved
flame stability at low loads (the lowest load for dispatch was 95 MW with BOOS),
and a higher reliability in achieving full load was expected with the ability to
accommodate burner maintenance outages without load reduction (with BOOS,
shutdown of an additional burner would reduce generating capacity). This
retrofit, the subject of this paper, is the first installation in the United
States of the integrated application of low-NOx burners, flue gas recirculation
(FGR) and overfire air (OFA) to a heavy oil-fired utility boiler. The specific
combustion' system, designated "PG-DRB", is licensed by BSW from Babcock-Hitachi
(BHKJ who commercialized the technology in Japan.

HECO retained Electric Power Technologies, Inc. (EPT) to work in cooperation
with HECO Engineering, Environmental, and Production Departments for
implementation of the retrofit project. In this capacity, EPT participated in
the specification, design review, and startup of the equipment, and acted as
technical liaison between HECO and B&W. In addition, HECO and the Electric
Power Research Institute (EPRI) jointly sponsored an independent test program to
document the performance of the retrofit low-tJOx technology. EPT was retained
by HECO to coordinate the test program and was responsible for implementing the
tests at Kahe, reducing the data, and interpreting the results. The performance
of the original combustion system for both ABIS and BOOS operating modes were
characterized in a "baseline" test series that was conducted in early 1988,
immediately prior to the outage for installing the new equipment. Following
BSW's commissioning of the new equipment in September 1988, a comparative "post-
retrofit" test series was conducted.

This paper presents the results of the test program, including a discussion of
the baseline and post-retrofit emission results, a synopsis of the start-up and
operational experiences with the new combustion system, and a preliminary
assessment of the costs and applicability of the technology. A formal
documentation of this program will be provided in an EPRI final report to be
prepared later in the- year.

DESIGN AND OPERATION OF ORIGINAL EQUIPMENT

Kahe 6 is an oil-fired steam-electric unit rated at 146 MW (gross). The boiler
is a radiant reheat type, manufactured by B&W (refer to Figure 1 for an
elevation view). For NOx control, the boiler was originally equipped with nine
B&W dual register- burners arranged In a 3 x 3 array or. one wall, and flue gag

9-2

-------
recirculation (FGR) to the windbox which permitted up to 20% of the flue gas to
be mixed with, combustion air prior to the burners. The dual register burners
were intended- to provide delayed mixing of the oil and air which would result in
lower NOx. The design was based on B&W's dual register burners developed for
coal-firing. The FGR acts as a thermal diluent, suppressing peak flame
temperatures and reducing NOx formation. From 1971, when Kahe 6 was originally
designed, to 1978 when the design was processed by EPA for Best Available
Control Technology, this combustion system represented state-of—the—art
technology available from U.S. boiler manufacturers.

When the unit was commissioned to service with all nine burners in-service, the
NOx emissions were found to exceed the operating permit limit (i.e., the PSD
permit limit of 0.23 Ib/MBtu) by a considerable margin. In order to minimize
NOx, Kahe was operated with flue gas recirculation and only six active burners
in a BOOS mode. BOOS operation effectively stages the combustion -- the six
active burners operated fuel rich and the air entering through the idle burners
was mixed with the combustion, products later in the furnace to complete the
combustion process.

DESCRIPTION OF COMBUSTION SYSTEM MODIFICATIONS

The new low-NOx combustion system consists of seven main elements;

- PG-DRB burners,

Improved oil atomizers, and the use of constant differential pressure
steam atomization,

Utilization of existing flue gas recirculation (FGR) to the windbox
combustion air,

- "Primary gas" (PG) — flue gas recirculation directly tc the burners
(unmixed with, the combustion air),

- Qverfire air (OFA) system,

Upgrading of the Bailey 860 burner controls, and
Modification of the FGR and combustion controls.

The PG-DRB burner, shown schematically in Figure 2, is an improvement upon the
1st generation B&W dual register burner originally installed at Kahe 6, It
includes improved aerodynamics (changes in throat diameter, air distribution and
swirl, a new air register design, etc.) for enhanced NOx reduction, and the
introduction of pure recirculated flue gas through an annular sleeve between the
core air and secondary air exits of the- burner. In principle, PG serves to
"blanket" the base of the oil flame to reduce the availability of oxygen during
the initial phases of combustion (Ref. 1>. The burners were supplied with new
wall panel sections, flame scanners/ignitors, and complete burner front control
and piping assemblies. The burners also included unique on-line air flow rate
monitoring. capability which eliminated t.he need for windbox
compartmentalization. The firing system was intended to burn a variety of fuel
oils with sulfur content limited to 0.5%, and fuel nitrogen content. varying up
to 0.5%

To help ensure' minimum particulate emissions, HECO requested BSW to supply an
oil atomizer that would operate at constant atomizing steam-to-oil differential,
and provide a higher quality (smaller drop diameter) spray than that produced by
the existing Racer design (The Racers operated with a constant steam pressure as
the oil supply pressure varied with load) . , In response, a Y-jet steam
atomization system was originally supplied with the PG-DRB burner, which
provided atomizing steam at a higher flow rate than Racers.

The unit's existing FGR system was adequate for supply of FGR to the secondary
air and was left intact, except for modified controls. However, to provide flue
gas to the burner PG annuluses, additional piping (including a distribution
system within the windbox) was required to convey flue gas to each burner from

9-3

-------
the main PGR duct. The' piping was equipped with orifice meters and butterfly
valves to permit on-line monitoring and control for PG to each burner.

The overfire air system was designed to divert up to 30% of the total combustion
air to six OFA ports located on the front and rear boiler walls {three ports per
wall), approximately 10 feet above the top burner elevation. Each OFA port was
equipped with damper assemblies and air spin vanes to enable independent control
of air quantity, velocity and furnace penetration. A schematic showing the port
design is provided in Figure 3. Installation of the ports required elaborate
ductwork, windbox, and boiler water wall modifications. Like the burners, the
OFA ports were equipped with flow rate monitors, enabling accurate on-line
measurement of separate flows through the 3pin annulus and central core of each
overfire air port.

Flow modeling of the windbox and furnace was an integral part of the engineering
design effort by B&W. A l/8th scale model of the windbox was built and tested
to ensure that equal flow to the burners would be achieved with the internal
windbox changes caused by the new burners and PG distribution system1. A
separate l/8th scale model of the furnace was then constructed and tested to
establish placements and design for the OFA ports for optimum mixing. The
decision to use six ports (instead of three) and a 70:30 rear-to-front wall
distribution of overfire air was based on the modeling results. In addition,
individual burners and overfire air ports were flow calibrated at full scale
prior to shipment to facilitate air flow adjustment and balancing during
startup, commissioning, and emission testing.

DESCRIPTION OF EMISSION TEST' PROGRAM

The HECO/EPRI emission test program consisted of a pre-retrofit baseline test
series, and a post-retrofit test series that followed 3£W's start-up of the new
combustion equipment. The testing was also complemented by a documentation of
the retrofit costs, and an evaluation of the test results and costs as they may
pertain to the other utility oil-fired boilers. Specific objectives of the
testing were:

Baseline testa -- Establish a basis for comparing boiler and emission
performance before and after the combustion equipment modifications. Data were
developed to enable quantitative comparisons over the load range in terms of:
NOx, PM total mass, stack opacity, CO, and S03 emissions; excess air level and
02 balance; boiler efficiency; maximum load capability; turndown; achievement of
design steam conditions; fuel properties; and other operational factors. The
baseline tests included parametric testing during A3IS and BOOS operation with
variations in boiler load, excess 02 level, FGR rates, and fuel nitrogen
concentration. Particulate matter size distribution and composition analyses
were also performed at full load conditions.

Post-retrofit tests — The objectives were multi-fold: <1) Characterize boiler
and emission performance for the operating conditions prescribed by B&w and
compare them with the baseline conditions, (2) Systematically evaluate the
independent and combined effects of low-NOx burners, OFA, FGR, and PG on NOx
emissions, (3) Quantify changes in PM - emissions for operating conditions
corresponding to different NOx emissions, and (4) Identify potential
improvements in operation compared to operating modes resulting from startup and
commissioning.

Emissions testing utilized state-of-the-art extractive sampling equipment and
methodology, including multi-point gas traversing at the economizer outlet.
Testing was performed through the cooperative efforts of personnel from the HECO
Environmental and Production Departments, B&W, Tenerex, and Mullins
Environmental Testing Company, with overall testing being coordinated and
directed by 'EPT.

9-4

-------
STARTUP AND COMMISSIONING OF MODIFIED COMBUSTION SYSTEM

The new combustion system posed a large number of variables that could effect
the NCx and PM emissions of the unit;

1, Burner: a.. Outer spin vane- position

b. . Inner spin vane position

c. . Core air flow

d. . Secondary air flow

e.. Oil flow rate

£.. Axial position of the oil gun impeller relative to the
atomizer

g. , Axial position of the impeller relative to the burner
throat

2, Atomizer: a. Atomization approach (i.e., basic tip design)

b. Steam/oil pressure differential

c. Oil supply pressure

d. Oil viscosity (measured via on-line viscosity meter)

e. Spray cone angle

f. Number of exit holes and their arrangement

g. Steam-to-oil mass ratio

3, FGR:, a. Amount of FGR mixed with the combustion air

4. PG:

a. Amount of PG supplied to the individual burners

OFA: a. Percent of total air supplied to the CFA porta

b. Percent of total OFA supplied to front OFA ports

c. Percent of total OFA supplied to rear OFA ports

d. Air flow balance among front OFA ports

e. Air flow .balance among rear OFA ports

f. Air flow split between core and annulus of each OFA
port

g. Spin vane setting of each OFA port

As is apparent from the number of variables listed above, the PG-DRB system is
complicated and proper adjustment requires a comprehensive startup program.
Moreover, for wall-fired boilers, the emissions are significantly impacted by
the respective performance of individual burners. Therefore, proper tuning of
the boiler dictated that the combustion be optimized for each burner and the
system as a whole.

BSW performed the commissioning of the retrofit equipment in cooperation with
the HECC Production Department. The commissioning required 12 weeks (2
shifts/day, 7 days/week) , and consisted of the following elements:, startup;
evaluation of atomizers; parametric testing and optimization; and acceptance
testing.

During startup, it was immediately apparent that there was 'a significant trade-
off between NOx and opacity whenever staging or low excess air operation was
attempted. Achieving NOx below the level of 0.23 lb/MBtu simultaneously with
opacity below 20% proved to be infeasible. Despite the use of Y-Jet atomizers,
the major contribution to the problem was determined from field testing and
laboratory spray characterization tests to be atomization quality (i.e., the oil
drops were too large and the spray pattern was not uniform). This necessitated
an evaluation of several different B&W atomizer designs in the PG-DRB burner.

Several different B&W atomizers were tested including modified Racers with
increased steam rates, Y-Jets and T-Jets with different spray cone angles,
Ba'ocock-Hitachi "D-tips", and a developmental I-Jet design. These atomizer
types are characterized by their geometry, atomizing steam-to-oil mass flow

9-5

-------
rates, the the size of oil drops produced. For oil properties and operating
conditions at Kahe, the nominal Sauter Mean Diameter (SMD) of oil droplet size
distributions were 400, 300 and 225 microns for the Racer, Y-Jet and T-Jet,
respectively. The T-Jet was judged to be the best atomizer design available to
B&W for use with the PG-DRB burner, and parametric and optimization tests
performed with this tip demonstrated that emission guarantees <0.23 lb/KBtu NOx
and 0.1 lb/MBtu ?H) could be met. The atomizing steam consumption with the T-
Jet is 15 to 20 times that normally used with the original Racer tips.

The acceptance tests were performed under the following operating conditions
recommended by BSW: 101 FGR to the wir.cbcx, minimum (<21) PG flow to the
burners, 27% of the total air to the overfire air system (split 70% to the rear
porta and 30% to the front ports), and an average excess 02 at the economizer of
3%. Under these conditions while firing a 0.35% nitrogen fuel oil, NOx
emissions at a maximum load of 156 MW (steam valves wide open, slight
overpressurization) and at 62 MW (the guarantee 2.5:1 turndown point) were 0.207
and 0.175 lb/MBtu, respectively. Corresponding particulate emissions were 0.074
and 0.085 lb/MBtu, respectively. These results typified the best overall
emission performance achieved in the 12-week startup period.

PRE-RETROFIT (BASELINE) EMISSIONS

The discussion below provides a very brief overview of the baseline results.
More detailed presentations of key baseline test data are provided in the
following section where they are compared directly with post-retrofit data.

Typical NOx emissions for the original equipment at Kahe 6, when operated at
high load and with a nominal fuel nitrogen level of 0.3%, are illustrated in
Figure 4, For ABIS operation "without FGR, the NOx was typically 0.8 lb/MBtu
(600 ppm) . This NOx level is consistent with that observed for boilers of
similar size and vintage, firing heavy oil, and equipped with single register
burners. Hence, the burners at Kahe 6 did not produce low NOx that was expected
for the dual register design. The addition of FGR reduced the NOx by nominally
50%. The significant magnitude of this reduction with FGR suggests that the
burners were producing considerable "thermal™ NOx when 'operated without FGR.
This is a further indication that the dual register design was not effectively
delaying the mixing of oil and combustion air.

Operation in the BOOS mode resulted in substantial NOx reductions when applied
with or without FGR. With BOOS and 10% FGR the nominal NOx emission level was
0.28 lb/MBtu (approximately 220 ppm), or about a 65% reduction from the ABIS
operating condition without FGR. Increasing FGR above 10% did not provide
significant additional reductions in NOx, reflecting a relatively "steep"
dependence of NOx. on initial FGR rate.

Effects of Fuel Nitrogen on NOx Emissions

Fuel oils with nitrogen- content ranging from 0.24% to 0.45% were tested to
establish the influence of this parameter on NOx emissions. This range is
typical of the fuel nitrogen variations normally experienced at the Kahe
Station. Results at high load with FGR are compared in Figure 5 for both BOOS
and ABIS operation. As can be seen, the fuel nitrogen has a substantial effect
on NOx emissions. Under ABIS conditions, NOx emissions increased approximately
55 ppm for every 0.1% increase in fuel nitrogen. With BOOS, the sensitivity was
reduced to about 35 ppm/0.1% nitrogen. These results are consistent with data
in the literature which generally show a nearly linear NOx sensitivity to fuel
nitrogen, and lower conversion rates of fuel nitrogen to NOx under slower mixed
or staged combustion. The relatively high conversion rate, especially during
ABIS operation, further substantiates the belief that the dual register burner
design was not effectively producing delayed combustion.

9-6

-------
The post-retrofit tests were performed with, a relatively constant fuel nitrogen
content averaging 0,35%. Thus there was not an opportunity to develop data on
fuel nitrogen-.dependence for the PG-DRB system for comparison. Nevertheless,
the baseline results are believed, to constitute a unique set of data on the
dependency of NQx on fuel nitrogen for a single unit operated with and without
staged combustion. An analysis of the fuel oils used during the baseline and
post-retrofit testing is provided for reference in Table 1.

Particulate matter emissions were measured for both ABIS and BOOS operating
modes for each of two oils. The PM total mass emissions ranged from 0.06? to
0.092 lb/MBtu whxch xs. consistent wxth levels measured by HSOO durxng tne 1981
to 1986 time period for Kahe 6. There was no significant difference in PM
emissions between ABIS and BOOS; the above variation in emission levels' is
considered to be within the normal range of variation for this type of source
and accuracy of the PM measurement method. Stack opacity was typically 4-6%
during ABIS and BOOS operation.

Unburned carbon was the principal component of the PM. The carbon concentration
ranged from 70% to 77% of the PM total mass. This level of unburned carbon was
felt to be typical for a combustion system utilizing Racer atomizers and an SMD
of 400 microns. The sulfate component was approximately 2% which ia typical for
low-sulfur fuel oil.

POST-RETROFIT EMISSIONS

Similar to the baseline test series, the post-retrofit testing involved an
extensive, systematic variation of boiler and -combustion system operating
parameters. The discussion below focuses on the most significant findings: the
effectiveness- of FGR and PG in reducing NOx,. the effect of OFA, NQx dependence
on load, the trade-off between NOx and PM emissions, and the effect of the
modifications on excess 02.

Effaet-iveneas of Flue Gas Reel.rrulat.inn to the Winrtbox

The effect of FGR rate on NOx. is compared in Figure 6 for baseline and post-
retrofit operation. The FGR rate is defined as the amount of recirculated flue
gas divided by the sum of the total air and fuel flows. The total reduction in
NOx attributable to FGR with the new combustion system varies from 20% when OFA
is used, to about 30% when the OFA ports are closed. FGR rates higher than
approximately 7% are ineffective. The limited effectiveness of FGR as compared
to pre-retrofit conditions is apparently due to the already lower NOx formation
which occurs as a result., of using OFA and the PG-DRB burners

Effectiveness of Flue Gas Recirculation to the Burners (PG).

The effect of PG flow on NOx is shown in Figure 7. The PG rate shown is percent
flow, with 100% representing a PG flow equivalent to 5% FGR. PG has no effect
on NOx below the 50% demand point. Increasing PG to approximately 801 results
in a 9% decrease in NOx, with a corresponding increase in opacity from 13 to
161.

Effect of Ovarflre Air

The effect of OFA level on NOx and opacity is shown in Figure 8. A3 can be
seen, OFA is a very effective means of reducing NOx emissions. Use of 30%* of
the total air as OFA (the maximum possible for the system) results in a decrease
in NOx of over 50%. The burner air/fuel ratio at maximum OFA is approximately
81% of stoichiometric.

9-7

-------
The results in Figure 8 also illustrate the strong trade-off between NOx and
opacity {more on this topic later!. OFA levels of over 21% were required for
this fuel to reduce NOx to 0.23 Ib/MBtu (180 ppir.j , with resulting opacity levels
of between 10-15%. Opacity levels in this range tended to be unstable. For
example, at 304 OFA, the average value of opacity plotted was 17%, with peak-to-
peak oscillations of 12-22%.

These results correspond to the OFA flow bia3 recommended by B&w <701 of OFA to
rear wall, 30% to the front wall) . Further biasing of the OFA to a 9C%/10%
split improved opacity and enabled operation at slightly lower furnace excess
02. It was further demonstrated that with balanced burner-to-burner flows (as
verified by individual burner fuel and air flow measurements) and uniform air
flows to the OFA ports, a balanced distribution of excess 02 across the furnace
could be achieved. Thus, when the OFA and burner flows are accurately
determined (as was the case here), furnace 02 distribution can be controlled
quite well, - Rarely does conventional burner front instrumentation provide the
necessary measurements to accomplish this.Carbon monoxide emissions were
typically less than 50 ppm during all post-retrofit test conditions, and did not
correlate with opacity. Thus, they were not a prime consideration in
establishing operating excess 02 levels.

The NOx emissions over the normal load range of the unit are shown "in Figure 9,
comparing baseline and post-retrofit results. Under baseline conditions, NOx
increases with load up to the maximum load point. A similar trend is apparent
for the new combustion system, but it is much less pronounced, except for an
apparent rapid rise in NOx above 146 MW. The excess 02 at the economizer outlet
for all the data in Figure 9 varies with load from about 34 at high load to
about 6% at the.lowest-load.

The level of importance that the trade-off between NOx and opacity posed with
regard to system optimization (refer, to Figure 8) was unexpected from previous
experience reported by 3abcock-Hitachi (BHK) in Japan (over 5000 MW of PG-DRB is
operational there). It appears that this trade-off is a fundamental feature of
the PG-DRB system, when fired with heavy oils. Further assessment of the
Japanese experience in the light of the HECO results led to a conclusion that a
similar trade-off exists in Japanese installations, however, it is not an issue
there because the boilers are equipped with electrostatic precipitators for
particulate emission control. It is now suspected that the low NOx results
reported from Japan by BHK were obtained by operating the burners and OFA system
for low NOx, with resulting high: opacity levels controlled by the precipitators
which are installed on virtually all utility oil-fired boilers equipped with
low-NOx combustion systems. The PG-DRB system configured with Racer atomizers
per the Japanese experience could not meet the NSPS particulate standard {0.1
lb/MBtu) or opacity standard of 201 when operated in a low-NOx mode with >25%
OFA and >10% FGR.

Particulate mass emissions measured during B&W acceptance tests were comparable
to levels reported above for baseline conditions. However, it is very apparent
from opacity readings during other tests that higher PM emissions can occur when
the system is not operated under tuned conditions. Opacity levels are currently
10-15% as compared to 4-6% with the original combustion equipment.

9-8

-------
Effect of Modifications on Excess Oxygen Re qui renent s

Following the retrofit, slightly higher furnace excess oxygen levels compared to
previous BOOS operation (by about 2-3 tenth's of a percent) were required as
¦primarily dictated by opacity considerations. The unit is currently operated
with 3-3.5% excess oxygen at full load. The actual value varies from day to day
as fuel properties vary and the operators adjust furnace combustion conditions,
to meet NOx and opacity limits. The increase in excess air, although small, was
disappointing in that the retrofit was expected to result in reduced excess 02
operating requirements based on the Japanese experience and general improvements
in combustion balance and control systems.

K summary of the post-retrofit NOx results is provided in Figure 10, including
baseline data under similar fuel, load, and excess 02 conditions.
Significantly, a 75% reduction was achieved with all burners in service with the
FG-DRB system as compared to "uncontrolled" emissions with the original burners.

OPERATIONAL IMPACTS

The installation of the PG-DRB system led to a number of operational
improvements. Some of these can be attributed to the general upgrading of the
firing control system and the return to ABXS operation;

The boiler turndown range increased due to a reduction in the minimum
load from 95 MW to 50 MW that resulted from improved flame stability;
"the achievable rate of load change remained the same

Ignitor reliability and flame scanning improved over the load range

Full load operation can be maintained with up -to two burners out of
service for maintenance; Under previous BOOS operation, removing a
single burner from service for maintenance would require a reduction
in "maximum load.

Furnace excess 02 balance across the economizer outlet improved to
within <1.0% 02 as compared to a previous imbalance of >1.5% 02.

In addition, the ability to control steam temperatures was not affected by the
retrofit, and steam attemperation rates were virtually unchanged. This
indicates that heat absorption patterns in the boiler were not substantially
affected by the retrofit.

At equivalent loads the boiler is operating at 2 to 3 inches of water higher
windbcx pressure as compared to pre-modification operation. These higher levels;
are a consequence of more physical obstructions to flow with the PG-DRB burner
and the PG piping in the windbox. In addition, the burner throat diameters were
sized based on full load operation at 2% excess 02 or less. However, to control
opacity to below 20%, excess 02 levels are commonly near 31 and this higher air
flow rate through the dual registers of the PG-DRB burner results in increased
pressure differential between the furnace and windbox.

The higher opacity with the new system continues to occur, and necessitates
close operator attention to excess air, atomizer wear, fuel changes, and the
degree of staging with overfire air during normal operation of the unit.

9-9

-------
COST OF NOx CONTROL

The combustion system modification for Kahe 6 represents a total capital
investment by HECO of over 53.9 million, A preliminary breakdown of these costs
is as follows:

Process Capital — materials, fabrication and erection 52,775,000
costs for PG-DRB burners, ignitors, steam atomization,
overfire air ports, windbox & ductwork modifications,
firing system control modifications, and other misc. items

Engineering — HECO Production and Engineering labor, $080,000

3fiW engineering and flow modeling, and other outside
engineering services

^reproduction Costs — outside start-up and testing $101,000

services

Inventory Capital — spare parts $7,000

Allowance For•Funds During Construction — interest $14 3,000
paid on capital funding

Total Capital Investment: $3,906,000

These costs, equivalent to about $26/kW, are much higher than retrofit
combustion control costs estimates generally reported in the literature. There
are several reasons for this. First, the engineering and process capital costs
are higher than might be normally assumed due to; <1! the decision to
incorporate into the design a high degree of independent control and monitoring
of combustion system parameters, (2) windbox and furnace flow modeling to
support the system design, and (3) economy of size, 150 MW versus 500 MW for
paper studies. Secondly, the retrofit was far from a simple replacement of
existing equipment. As described above, the retrofit involved substantial
ductwork and furnace water wall modifications to accommodate the new burners, PG
flows, and OFA- ports. The latter required separate runs of duct supplying hot
air from the air heater directly to the OFA ports. Additionally, the new
combustion system involved a conversion of the steam atomization system and a
general upgrade of the Bailey firing control system. It should also be noted
that since the unit was already equipped with an FGR system, the costs that
would normally be incurred for the FGR fan and associated ductwork, controls,
and secondary air mixing equipment were avoided. Finally, the shipping costs to
Hawaii result in approximately 10% higher material costs than for a mainland
site.

The differential operating costs for the new combustion system are currently
being evaluated. Although the new system is not expected to result in higher
net operating and maintenance costs, several factors in this regard are being
evaluated: increased atomizing steam consumption for the new constant pressure
differential system, increases in operating excess 02, and potential changes in
FD and FGR fan power consumption.

APPLICABILITY OF RESULTS TO OTHER BOILERS

This project has provided the utility industry and E & W with a better
understanding of the retrofit cost, operational implications, and performance of
integrated dual-register burner, overfire air, and FGR low-NOx combustion
systems. Based on the Kahe retrofit experience and the test results, it can be
reasonably concluded that thi3 system is applicable to virtually all existing
wall-fired utility boilers firing residual fuels. In light of the
NOx/particulate trade-off, it appears that the ability to achieve desired NOx

9-10

-------
emission.' reductions with this technology will be especially sensitive to site
specific factors such as fuel nitrogen, particulate formation tendencies of the
oil, particulate/opacity regulations, and furnace 02 limitations. Since the
KQx/particulate trade-off appears to be a generic feature of the technology {and
may be relevant to other state-of-the-art combustion systems as well), more
attention to oil atomization as a possible avenue for reducing particulate
emissions is clearly warranted. In this regard an independent project funded by
EPRI, ESEERCO, and Consolidated Edison Company of New York. (Ref. 2 > is in
progress' to develop a methodology to predict particulate emissions for a given
fuel oil and boiler design, and to develop a design basis for improved oil
atomization.

Costs for the PG—DR3 system will vary, of course, depending on specific boiler
design and level of NGx control required. It appears that the PG system is not
effective and could be deleted from future designs, thereby avoiding its added
complexity, costs, and associated impact on windbox pressure drop. It is also
conceivable that FGR would not be required in some -instances where NOx
reductions less than the system's full capability are sought.

Finally, the design and engineering approach followed for this project —
involving flow modeling of critical flow distribution and mixing regions, flow,
calibration of individual components, and a high degree of independent control
and monitoring of combustion parameters — proved to be indispensable in
optimizing and operating the combustion system. Such an approach is recommended
for future retrofits..

CONCLUSIONS

Retrofit of the oil-fired-PG-DRB system at Hawaiian- Electric Company's Kaha 6
is the first application of this technology to a U. S. utility boiler. This
project has significantly enhanced the U.S. utility industry's understanding of
the performance and operation of this combustion system.

NOx and CO emissions were duplicative of the Japanese experience. However, it
was necessary to push beyond the Japanese experience to simultaneously control
NOx and particulate emissions. A trade-off between NOx and particulate
emissions appears generic to the technology and is a major factor in system
optimization. Further research to investigate better oil atoniization to
minimize particulate generation is recommended.

The NOx reduction goal of 0.23 lb/MBtu was achieved, notwithstanding the
competing requirements for particulate and opacity compliance. The demonstrated
NOx level of typically 0,2 lb/MBtu represents a 75% reduction from levels
documented with the original, burners when FGR was not used, a condition that
might be representative of many uncontrolled boilers.

Overfire air (OFA) and flue gas recirculation (FGR) were dominant influences on
NOx emission reductions with the PG-DRB system. Compared to the PG-DRB burners
operating alone, OFA and FGR each further reduced NOx by about 50% and 301,
respectively. With combined operation of OFA and FGR, the reduction was nearly
60%.

Recirculated flue gas injected directly.into the burner throat (designated "PG")
was only marginally effective in reducing NOx, and is not likely to be justified
for future retrofits.

The total retrofit cost of $26/kW reflects the relatively high complexity of the
PG-DRB retrofit; engineering and design efforts to ensure integrated functioning
of the burners, overfire air and flue gas recirculation systems; a high degree
of adjustability of NOx control parameters to facilitate system optimization;
the cost savings due to the use of an existing flue gas recirculation system;
and somewhat higher shipping and installation costs for a Hawaiian site. This

9-11

-------
retrofit coat is higher and more realistic than estimates generally published
for oil-fired boiler retrofit NOx combustion control.

ACKNOWLEDGMENT S

The authors wish to thank Babcock S Wilcox for providing teat equipment for the
HECO/EPRI tests. Additionally, the cooperation and advice from the Kahe Station
operating and maintenance personnel during the test program is gratefully
acknowledged.

REFERENCES

1. A. LaRue and P. Cioffi. "NOx Control Update - 1987." Proceedings
of the 1987 Symposium on Stationary Combustion Nitrogen Oxide
Control, vol.1, EPRI CS-5361, August 198,7.

2. ESEERCO Research Project EF85-15: "Heavy Oil Combustion", sponsored
by the Empire State Electric Energy Research Corporation (ESEERCO),
Consolidated Edison Company of New York, and EPRI. {in progress)

Table 1

KAHE 6 FUEL OIL ANALYSES

Range for Typical for
Basel ine Posf.-Rfttrnf i t

Composition {Wt) :

Carbon

Hydrogen

Nitrogen

Sulfur

Ash

Oxygen, by difference

86.2-87.4
6.3-11.6
0.20-0.46
0.34-0.47
0.01-0.03

87. 6
11.3
0.31
0.43
0 . 02

1.4-5.4

Conradson Carbon (%Wt)
Asphaltene (%wt)

7.6-10.3
1.7-3.2

6 . 7
1.3

Pour Point (degrees FS
API Gravity <860;F)
HHV (Btu/lb)

95-110
13-20
18560-19070

17
18860

9-12

-------
Figure 1. Kahe Unit #6 Elevation Drawing

9-13

-------
(SHOWN CLOSED!

AMD Q'JABRjUIT

Figure 2, Babcock & Wilcox PG-DRB Burner

Figure 3. Overfire Air Port with Velocity Controlled Throat

9-14

-------
N 800-/

136 MW, NOMINAL 02,0.31 FUEL N

0.23 LB/MBTU
= 180 PPM

BOOS

101F6R

Figure 4. Pre-retrofit "Baseline" NOx Summary

138 MW , 3% EXCESS 02, 10% FGR

a
O

$
m

>
tc
a

*
O
z

500

400

300

200

100

~ i' »

0.2 0 0,25

0.30 0.35 0.40
FUEL NITROGEN, % (WT.)

0.4S

0.50

Figure 5. NOx Dependence on Fuel Nitrogen Content

9-15

-------
01

>
EE

o.
a

O
z

700
600
500
400
300
200
100
0

—,—|—,—| , j—, j , ; , 1—.— .

O PRE-RETRO,ABIS,Q.47%N :
™ • PRE-RETRO,ABiS,0.23%N
+ PR E-R ETRO. BOOS. 0,3 2%N

¦ N

~ PG-DRB,NO QFA,0.3%N ;

35^

> ¦ PG-DRB,OFA,0.34%N

^—j -o | j ;

1 —I— 1

^hrh I ~f~ ;

•- i - • * i i > • * i ¦

6 8 10 12 14
FLUE GAS RECIRCULATION, %

1 6

1 8

2 0

Figure 6. Effectiveness of FGR in Reducing NOx Emissions, Comparing

Baseline and Post-retrofit Operation

DC
D

140 MW, FUEL N = .37%
7% FGR, 25% OFA

200

150

100

...u.Q

100% PG > 5% FGR

—NO

O * OPACITY

2 0

4 0 6 0

% PG PLOW

8 0 100

>
¥•

o
<

Figure 7. Effect of PG Rate on NOx

9-16

-------
fsf

>
cc
a

350-

140 MW, 0.3% FUEL N, 3.3% 02
GAS RECIRCULATION TO WINDBOX

300 ¦

250

200

150

100

INDICATES RANGE
.OF OPACITY
OBSERVED

2 0 3 0

0VERFIRE AIR, % OF TOTAL AIR

•25

4 0

Figure 8. Effect of 0FA on NOx and Opacity

>
CC
Q

a
o.

9 0 1 00 1 1 0 1 20 1 30 1 40 1 50 1 60 1 70
LOAD, MW (GROSS)

Figure 9. NOx versus Boiler Load for Baseline and Post-retrofit Operation

9-17

-------
CM

>¦

0.
Q.

x"
O

0.3% FUEL N, NOMINAL 02

180 ppm =
0.23 Lb/MBty

1 2 3

PRE-RETROFIT BASELINE
All Burners In Service
No Gas Recirculation
PRE-RETROFIT OPERATION
3/9 Burners Out Of Service
Gas Recirculation To WB

PG-DRB BURNERS

All Burners In Service
No Gas Recirculation To WB
4: PG-DRB BURNERS

Overfire Air
5: PG-DRB BURNERS

Gas Recirculation To WB
6; PG-DRB BURNERS
Overfire Air

Gas Recirculation To WB

Figure 10. Suirettary Comparison of Baseline and Post-retrofit NOx Emissions

9-18

-------
GAS TURBINE NITROGEN OXIDE (NOx) CONTROL
CURRENT TECHNOLOGIES AND OPERATING EXPERIENCES,

Leonard Angello
Electric Power Research Institute
3412 Hill view Avenue
Palo Alto, CA 94303

Phillip Lowe
InTech, Incorporated
11316 Rouen Drive
Potomac, MD 20854

ABSTRACT

National standards for NOx control for gas turbines are being evaluated and
modified throughout the world. " These standards are1 .often- supplemented by local or
site specific emission limits. Specific allowable emission levels range from
about 10 to 120 ppm (parts per million, volume basis) in the U.S., depending upon
the capacity and efficiency of the gas turbine application and the local environ-
mental needs for NOx control. In response to those requirements, the equipment'
manufacturers and plant operators have designed and installed one or combinations
of four control technologies: fuel switching; wet combustion controls; dry com-
bustion controls; and exhaust gas treatment. In addition, catalytic combustion
control systems are presently under development but are not commercially
significant at this time.

This paper presents a brief overview of the national NOx emission requirements of
Europe, Japan, and the United States, A description of the control processes
presently being applied at commercial facilities is provided and typical operating
results and experiences are also presented. Differences between the operating
results and design conditions are highlighted!. The advantages and disadvantages
of each control technology are summarized, and areas or topics of uncertainty or
high risk that could benefit from additional research are identified.

In this paper, all NOx values are referenced to dry, 15% oxygen exhaust conditions
unless otherwise stated. The units are volumetric, in parts per million (ppm)
unless otherwise stated.

Presented at the 1989 JOINT SYMPOSIUM ON STATIONARY COMBUSTION NCx CONTROL,
March 6-9, 1989, San Francisco, California.

9-19

-------
INTRODUCTION

When gas turbine power generation first entered the electric utility market it was
hailed as a clean, benign source of power. There were no giant cooling towers to
dominate the landscape; no coal delivery or ash removal; noise could be
effectively screened by suitable acoustic enclosures; and stack emissions were
barely visible-

Today, over 25 years later, the 140 MW unit with a turbine inlet temperature
approaching 2300"F is a reality. The inherently high potential for exhaust heat
recovery, and the greater efficiency of the combined cycle has insured a role for
the gas turbine in power production programs around the world.

But in parallel with this trend there has developed a growing awareness of the
need" to protect the environment. As gas turbine operating temperatures have risen
to provide greater output and efficiency, equipment manufacturers and plant
operators are confronted with increasingly strict air quality control standards -
particularly in regard to oxides of nitrogen (NOx).

NITROGEN OXIDE (NOx) EMISSIONS ANO STANDARDS

The major component of NOx is nitric oxide (NO) formed by oxidation of atmospheric
nitrogen at high temperatures. The reaction has been shown to be exponentially
dependent on temperature and linearly dependent of time. It has become an
increasing problem for designers with the introduction on new models with turbine
inlet temperatures over 1850°F.

Between 850°F and 1200°F the NO can be oxidized to NOg, a brown gas which can be
highly visible at part-load unless steps are taken to inhibit this reaction.
Together these two oxides of nitrogen are the principal components of NOx.

NOx forms in two ways. First, fuel-bound nitrogen can be oresent as a chemical
component of the fuel molecules. This reacts with combustin air as the fuel is
burned. The amount of NOx produced depends on the nitrogen content of the fuel.

Thermal NOx, on the other hand, is produced by oxidation of atmospheric nitrogen
at high firing' temperatures. Concern about NOx emissions has mounted with the
increase in firing temperatures and suppressing the formation of these gases is an
important facet of current gas turbine research.

Fuel type (natural gas or distillate fuel oil) as well as the fuel nitrogen
content has a major impact on NOx emissions. This is mainly due to the difference
in flame intensity (adabatic flame temperature) which different fuel produce when
they are burned.

Figure 1 illustrates the affect that fuel type and firing temperature have on
uncontrolled gas turbine NOx emissions. As shown, uncontrolled NOx emission range
from 90 to 140. ppm for a typical heavy-duty gas turbine fired, by natural gas over
the range of firing temperatures between 1800 to 2100"F. The same unit fired with
distillate fuel oil produces NOx emission of 140 to 210 ppm over the same range of
turbine inlet temperatures.

In the U.S., the requirement to control NOx and the limitations or emission levels
that must be achieved are presented in Federal and state regulations, with the
more severe requirements taking precedence. Thus, both requirements must be
examined for applicability.

9-20

-------
Table 1 presents a summary of selected national emission standards for NOx,
expressed as NCU. The table shows that the U.S. has the most-stringent national
standards for the industrialized countries examined.

Table 1

SELECTED NATIONAL NOX EMISSION STANDARDS
(For Application to Gas Turbine Facilities)

Turbine Emission Limits

Country Capacity, MWe ppm mg/Nm

FRG

60 &

larger

147

300

less

than 60

172

350

US ¦ •

10 6

larger

153

Japan (Natural gas)

10 &

larger

100

205

In addition to the national emission standards, local government authorities can
and do impose more stringent standards. In the U.S., for example, national
standards require something, in the range of 75-140 ppm' for electric utility gas
turbine emissions, depending upon the plant efficiency and fuel nitrogen content
as shown in Figure 2, However, several local authorities in California have
imposed standards for application to the smaller.turbines used in cogeneration
facilities, ranging from 15-42 ppm (some recent cogeneration facilities in
California have actually been permitted at 4.5 ppm), reflecting those regulators'
understanding of the lowest, achievable emissions rate (LAER) controls available.

As stated earlier, the type of fuel that is fired strongly impacts the NOx
emissions, and the emission standards generally recognized tins fact. Table 2
presents a comparison of the emission limits for oil and gas-firing, for selected
countries-.

Table 2

EMISSION LIMITS COMPARISON
FOR OIL AND GAS-FIRING

Country

Austria
Canada

FRG (after 1984)

Japan (after 1987)
Switzerland
U.S.

ECC (proposed up to 1995)

Ratio of Emissions Limits
Oil-Firing Gas-Firing

79
67
67
46

0.67
0.79

The numeric value of any emission limit is also directly influenced by the actual
NOx removal capacity of the specific technologies that can be applied to the
processes and equipment that will be governed by the emission standards. That is,

9-21

-------
emission limits can not be so restrictive that there is no technology available
that can be applied to meet the permit condition. However, as new technologies
are developed to meet the ever increasingly restictive emission limitations. They
become the standard by which the next round of emission limits setting is guided.

Current NOx control technologies for gas turbine applications employ fuel
switching wet combustion control (steam or water injection into the combustor), or
exhaust gas treatment (selective catalytic reduction technology) or combinations
of those controls.

Recently, dry combustion controls have been, or are being commercially, developed
by the gas turbine manufacturers, and catalytic combustion control processes are
being research. Each of these NOx control technologies is discussed in following
sections.

FUEL SWITCHING

The simplest way to reduce NOx is to reduce the nitrogen available for converstion
into NOx. Natural gas-firing will produce about two-thirds of the NOx that firing
with a light distillate oil will. Since the uncontrolled NOx emission from
natural gas-firing will not, in general, meet current emission limts, additional
controls are required. This imposes, the restriction that the subsequent control
technology must be compatible with the fuel. For example, high sulfur bearing or
fuels with alkali or arsnic compounds can poison the catalysts used in the
selective catalytic reduction (SCR) process. Thus, if that approach were to be
used, it would impose some constraints on the choice for fuel switching.

WET COMBUSTION CONTROLS

This is perhaps the most widely used, commercially available NOx control
technology. NOx reductions of up to 80% compared to the uncontrolled system are
possible.. Steam or water is injected into the combustion chamber at ratios of
fluid to fuel of up to 2.4 (by weight), although most systems have a practical
limit of injection ratios of 1.0 or less. The fluid reduces the combustion
intensity and temperature. When water is injected part of the fuel energy is used
to vaporize the water, which lowers the combustion intensity. Thus, water is more
effective than steam, on a weight basis. However, it can lead to more erosion and
wear problems.

Additional fuel can- be burned to counter the heat of vaporization effects, and the
turbine power output increases because of the additional mass throughput which
comes from the water flow. As the amount of water injection increases, the level
of unburned hydrocarbon and carbon monoxide emissions increase. Figure 3 gives a
typical example of the increase in CO as NOx is reduced. As the NOx control is
pushed to yield large percentage NOx reductions, significant CO and unburned
hydrocarbon (UHC) emissions can result, requiring the installation of additional
controls for those pollutants. These controls are often based upon the use of
catalysts and are very sensitive to temperature conditions. -This in turn can
impose strict limits on the part load operation of the gas turbine. A major
disadvantage of the wet combustion control processes is the cost associated with
the supply of the water or steam, and in some cases, the actual availability of
the water or steam itself.' The technical problems with fluid injection include
internal pressure fluctuations which increases in component erosion and wear.
Impurities in the fluid, especially in water, can significantly effect turbine
blade wear and erosion. The durability of liners and transition pieces have also
been impacted. Wet systems can yield 30 - 75 ppm-NOx emissions for gas firing and

9-22

-------
50 - 150 ppm for oil firing. Considering the point cf share increase in CO
emission as shown in Figure' 3, and the erosion and wear problems that occur with
water {or steam) injection, 40 ppm appears to be a practicable lower limit for
manufacturers to meet for gas firing and still warrant their equipment (although
some advertise the ability to achieve even greater control levels). Fifty to
sixty ppm would be the corresponding lower limit for distillate fuels.

The technical performance of wet combustion controls can be represented by data
similar to that given in Figures 4, 5 and 6. Figure 4 shows the NOx reduction
achievable as a function of the water injected into the combustor, expressed as
the ratio of the amount of Water to Fuel injected, on a mass basis. Similar
curves (Figure 5) can be anticipated for other turbines or for steam injection.
The data shows that between 60-80% NOx reduction (expressed as an R factor of 0,4
to 0.2 in Figure 4) is possible. For reductions much above 70% it can be seen
that the CO and UHC emission increase as such a rapid rate that an oxidizing cat-
alyst is usually be required to control those emissions.

Injecting water into the combustor increases the power output of the engine. This
is a function of the increased mass flow through The turbine if the fuel/air ratio
is maintained. The heat of vaporization of the water is lost, this results in
increased fuel requirements if operation is held to a. constant cycle temperature.
Figure 6 shows this impact for one turbine supplier. At a water to fuel injection
ratio of 0.8 the curve shows that 8% more fuel is required to maintain the turbine
output temperature'. At that condition, 6% more power output is achievable from
the engine.

The water injection system-will include manifolds to control the water flow and to
assure that water can not flow back into the fuel or air injection system. As the
control point is moved to more and more NOx reduction, the control of the amount
of water flow-, compared to the fuel flow, becomes critical. Very accurate flow
metering is required and very tight control on the water purity is also required
in order to eliminate or reduce erosion or wear problems. Some problems that have
occurred are nozzle clogging and corrosion, pressure fluctuations and wear due to
imbalances in water flow to specific nozzles within the combustor, and high
temperature differentials and flame instability, wnich also can lead to combustor
or turbine component wear and erosion problems. The water quality specification
should control particulate matter as well dissolved' chemicals and the pH of the
water supplied.

A summary of the disadvantages of this method of NOx control is that it will
increase the emissions of carbon monoxide and unburned hydrocarbons (and perhaps
require the use of an oxidizing catalyst for pollution control for these items),
it will increase the fuel consumption from 1 to 4% (or a decrease in the net power
output), the system capital costs and operating costs will increase, the engine
control and monitoring system complexity will increase, the control system will
require a large supply of water that can not be recovered (it is discharged as
steam from the exhaust stack), and the engine durability may decrease while the
system overhaul frequency Increases.

DRY COMBUSTION CONTROLS

This is a commercially available NOx control technology for some of the turbine
suppliers*, it is under development at others. Ory combustion controls show the
promise of significant cost reductions compared to wet combustion controls, and
perhaps a slight improvement in the magnitude of the control is also possible.
NOx reductions of up to about 60% have been achieved, and in special instances
reductions of over 80% have been reported (e.g. NOx discharges of 30 ppm). These

9-23

-------
technologies-are based upon reducing the flame turbulence and intensity, enhancing
the fuel/air mixing, and establishing fuel lean zones within the combustcr. They
generally rely on.a staged combustion with the first stage used as- a pilot burner
followed by a secondary stage of multiple fuel injection nozzles. The second
stage often employs burning a lean, premixed fuel/air mixture in order to assure a
uniform mixture and the avoidance of high temperature regions in the combustor.
Obtaining the uniform mixing is very important in achieving this goal.

The control of the two burner zones and the preparation and control of premix air
complicates the combustor control and often requires the use of sophisticated
mechanical control systems. The pilot or first burner is generally used to start
up and carry the initial engine load. The control system can be based upon using
variable geometry, variable air flow scheduling, or combinations thereof. At
higher loads the second stage picks up the load and the first stage is used to
provide preheat to the second stage. Figures 7 indicate how one supplier's
control system works. In Figure 7 the step change in the NOx level indicates when
the engine control was changed from relying on the first stage nozzle in the
combustor to the set of second stage nozzles. However, it also shows that the CO
limit could easily reach 500 ppm or larger, as the engine may have a very limited
operating region between having high NOx-low CO or high CO-low NOx emissions.
Manufacturers who use external silo-type combustors can more easily adapt their
designs to different sized turbine application than can those who use internal
combustors. A clear advantage of the dry control system is the elimination of the
costs of providing large quantities of clean water. But the impact of the dry
controls on component wear and erosion and the engine durability have- not yet been
demonstrated by large numbers of operating hours under U.S. utility operating
conditions. This remains an area of significant technical uncertainty, and
reliability testing in a field application will be a very important 'part of
establishing the actual reliability of these systems.

The technical performance of the dry combustion modification NOx controls is much
more a function of the manufacturer's equipment than is the corresponding perfor-
mance of wet controls. Thus, although Figure 4, 5 and 6 are typical of the wet
controls in general, the data of Figures 7 and 8 are really only representative of
the performance of one specific equipment supplier's dry, low NOx (OLNOx)
system. However, they are illustrative of the performance range possible for dry
controls. As indicated earlier, the dry controls usually have two stages of
combustion, and the NOx control performance has a step change in the NOx emissions
as the turbine load being carried is shifted from stage one to stage two. In
Japan, the turbine NOx emission typically is set at about 75 ppm, apparently on
the basis that an 80% efficient SCR control system also will be required. This
combustion of combustion control "and exhaust gas treatment will yield a 15 ppm
stack emission, thus tighter combustion controls are not needed. The actual
achievable control limit for the dry control technology could be less than 75 ppm,
but plant operating data to support a commercial decision to operate at lower NOx
emissions is not yet available. It is also noted that sc.ne U.S. and German
turbine suppliers have indicated that they may be able to achieve NOx emissions as
low as 30 ppm. Supporting data will be very important in evaluating these
claims. Also, the Japanese data show that the combustor control settings are
critical with respect to the emission of significant amounts of CO. The UHC
emissions are unreported, however, experience indicates that they follow the same
trend as the CO emissions. Hence, a report on very low levels of NOx emission
needs to be evaluated in the context of how that operational condition impacts CO
and UHC emissions.

The Japanese experience is that the combustion control needed to transfer load
between burner stages (e.g. to transfer from low to high load) is difficult to
achieve. Figures 7 and 8 show a step change with the NOx emissions at the load

9-24

-------
transfer point, indicating a step change in combustion conditions. Step changes
are, in general, not desirable from the point of view of engine stability and
durability. In one Japanese design a mechanical setting for bypass air flow is
required. Although it is stated that the control can be made while the system is
on line, reported plant startup problems suggest that establishing the control
setting is difficult. Also, that manufacturer is developing "improved" dry
controls, further indicating that the operation of the present control system 1s
difficult. It is noted that the ambient air condition can effect the control
system performance, and a 15 degree change in local temperatures can result in a
requirement to adjust the bypass air control setting. These conditions may have
helped contribute to the Japanese decision- to operate the turbines at 75% of full
load or higher, severely limiting load following capabilities.

In general, the turbine suppliers of dry NQx controls have not supplied data on CO
and UHC emissions, but it appears that they are operating the turbine in a manner
such that additional CO and UHC emission controls are not necessary (see Figures
7). One advantage of the dry control technology is the lack of the requirement
for additional diluent (e.g. water or steam) preparation and control facilities.
It may, however, be quite difficult to obtain a warranty and have the resultant
ability to operate the turbine on a backup or emergency liquid fuel, since the dry
combustion controls may not be able to accommodate the burning characteristics of
those distillate- fuels. Finally, long term reliability of the control process and
its long term impact on other engine components need to be established.

CATALYTIC COMBUSTION

This NOx control approach is not commercially available for combustors of a
physical size of general interest. Estimates of the remaining development time
range from 5 to 10 years-. Small combustors are available, and larger combustors
have been tested in laboratories. In some of those tests very low emissions, as
Tow as 1 to 5 ppm, have been reported; but in general higher emissions result when
the combustor is operated to obtain the high efficiencies needed for a gas turbine
applications. In actual application, the catalytic material is often placed upon
a substrate material such as a ceramic material. Pilot burners are used to
preheat the fuel/air mixture to the initial reaction temperature. Catalytic
combustion can produce combustor exit temperatures in the range of 2000°F,
suitable for further conditioning for gas turbine use. The major problems that
have been experienced deal with sealing the catalytic section against the
combustor shroud and in developing a rugged, durable catalytic section. Physical
failures such as cracks and the subsequent breakup of the combustor frequently
have been observed. Development needs to focus on producing durable combustors as
well- as controlling the combustion over an appreciable operating range.

EXHAUST GAS TREATMENT (Selective Catalytic Reduction (SCR))

There are control methods available and under development that treat the NOx in
the turbine exhaust gas to remove it. Under certain conditions reductions of NOx
of about 90% are possible, compared to the uncontrolled case. These methods can
be used independently or in conjunction with combustion modification controls.
Japan has been very active, cotroared to the U.S., in applying some of these
controls; but recent experience in Germany and earlier experiences in the U.S. in
coal-fired plants indicate that the technology must be modified if it is to be
successfully applied in U.S. operating systems and with U.S. operating
philosophies.

9-25

-------
Ammonia can be used as a reducing agent to react with nitrogen oxides and oxygen
to produce'nitrogen and water. Catalysts are used to promote the reaction at
lower temperatures. A series of titanium and vanadium oxide catalysts on metal or
ceramic substrates have been used extensively in Japanese utility facilities (and
other industrial facilities) to reduce NOx emissions. The catalysts are
formulated to optimize the conversion in the temperature range of 570 to 8505F,
with most operations focused near the 700°F temperature range. It takes one mole
of ammonia to reduce one mole of nitrogen oxide and two moles of ammonia to reduce
one mole of nitrogen dioxide. In commercial practice the catalyst is designed to
reduce 80 to 90 percent of the NOx emissions for ammonia injections, on a mole
ratio, equal to the amount of NOx present for the exhaust gas. In some cases it
has been advertised that the system was designed for 95^ NOx removal. In Japan
the operational practice (as contrasted to design conditions) is to operate the
system with reduced ammonia injection (reduced from the design injection
amount). This results in a NOx removal rate of 60 to 755$ with a corresponding
bypass of ammonia that passes through the reactor without reacting with the NOx,
thus slipping past the reactor and resulting as a low level pollutant emission. '

The Japanese also design their SCR systems for a space velocity (the volumetric
gas "flow divided by the catalyst volume) for gas-fired systems of about 4,000 to
10,000/hour (with a pressure drop of about 4-6 inches of water), with some
development being directed towards expanding that value to about 20,000/hour. The
reasons for the Japanese operational approach are varied. They appear to be a
result of their desire to limit ammonia emissions; a recognition of the difficulty
of controlling the system at the design condition; a recognition of the difficulty
of reliability measuring NOx and ammonia at the very small ppm operating
conditions; a recognition that such an operational strategy appears to materially
increase the durability and operational life of the catalyst and NOx reduction
system equipment ( a key factor 1n controlling system costs); and a recognition
that such operations reduce the problem in introducing ammonium sulfate and
bisulfate salts into the exhaust stream when fuels that contain small amounts of
sulfur are fired in the turbine. Most SCR applications in Japan are applied in
concert with some form of combustion modification NOx control.

In the U.S. there is less electric utility experience, and no gas turbine electric
utility experience with SCR NOx control. However, several cogenerators (mostly in
California) are using the control technology in conjunction with water or steam
injection combustion modifications for a total NOx control system designed to
limit the NOx exhaust emission to 8 to 15 ppm. The space velocity of the U.S.
systems has been on the order of 3C,000/hour (with a pressure drop of 2 - 4 inches
of water), with tests performed on systems with space velocities of up to
100,000/hour. In the U.S. the systems have been designed for an ammonia slip of
20 ppm or larger at ammonia to NOx injection ratios of about 1.0, and the systems
are operated for "90% or higher" NOx removal. Operating data is not readily
shared, so this high performance condition is not easily verified. There have
been mixed results with the new U.S. applications, several have come online
without apparent problems, others have had required extensive shakedown periods
and- do not appear to be operating as designed. Problems mentioned (but not
reported with any significant detail or with the release of experimental evidence)
include: the life and availability of the SCR catalyst itself; ammonia
consumption; ammonia injection controls .and monitoring instrumentation; and plant
operations, especially at throttled conditions. . It will take time to establish
the long term system reliability and availability and its impact on plant
economics.

In any SCR system the system operational temperature range is below the turbine
exhaust temperature, which is around 1000 - 1100°F. The SCR system is installed
in the heat recovery steam generator. The exact location is dependent upon

9-26

-------
whether the steam generator has auxiliary firing (or if duct firing is employed)
to increase the plant output. The SCR catalyst is usually located somewhere after
the superheater and before the economizer tubing. In small cogeneration systems
the ammonia injection, and hence the SCR operation, is terminated at part power
operation whenever the gas temperature entering the SCR reactor falls below that
required for efficient operation of the reactor. To continue operations at such
low temperatures would introduce excessive ammonia emissions into the exhaust
gas. In a large utility system it may be desirable to include physical
arrangements for bypassing the superheater in order to operate the SCR at lower
plant power levels. Such is the case for fired boiler operations in some Japanese
systems. Because of the.higher levels of ammonia bypass in the U.S. systems,
great care must be exercised when sulfur bearing fuels are burned in. the combustor
or the ammonium salts will be formed. One U.S. cogenerators has already
experienced problems because of the formation of ammonium bisulfate salts on
downstream equipment and the poisoning of the catalyst by the sulfur compounds in
the exhaust gas.

The SCR reactor also oxidizes sulfur dioxide to sulfur trioxide. This is another
cause for limiting the sulfur content in the fuel. The ultimate disposal of the
spent catalyst needs to be considered, since it contains heavy metals and is
classified as a hazardous waste material. In Japan the catalysts for gas turbine
operation have, been used for 4 to 7 years. In the U.S. they are being offered
with lifetime warrantees of 1 to 2 years.

Other problems that have occured in operating SCR systems include operational
problems with the ammonia injection system and the control and monitoring
instrumentation. The specification of the measurement instrumentation is still an
open question. In Japan the use of ehemiluminescence-based instruments have been
discontinued in favor of the use of infrared-based instrumentation, and
ultraviolet absorption-based instruments are under development. The
chemiluminescence-based instruments are still used in the U.S.

SUMMARY

The commercial status of gas turbine NQx control's to a large degree, will impact
the further development and demonstration of improved technologies. Wet
combustion control is commercially available and technically established. But
cost, its impact upon turbine wear and erosion and regulatory requirements of
lower NOx emission levels is causing this to become a less desirable technical
solution for NOx control.

Selective catalytic reduction technology is currently established in the
commercial market place. Technical issues persist, and the costs may therefore be
larger than those presently expected. Its use appears to be allocated for those
situations where the combustion control technology is insufficient to reach the
regulated NOx emissions limits. In those cases it is used in addition to the
combustion control- technology. It has not been applied as a stand-alone NOx
control system at a gas turbine facility, probably a further indication that it is
an expensive control system and that the control system reliability and its effect
upon the power generation system availability are not established at a confidence
level necessary to promote rapid commercial expansion of the technology. EPRI has
an ongoing research program that will evaluate the performance and reliability of
existing SCR/gas turbine systems.

Dry, low NOx combustion control is just entering the market place. It shows the
promise to be effective or even more effective as than wet NOx combustion
control. Questions remain as to its ability to allow load following operations,

9-27

-------
and its use may also require additional CO and UHC controls when very low NQx
emission levels are required. Because of its newness in the market place, there
is not sufficient information available to comment on the long term reliability of
the control system and its subsequent impacts on the rest of the engine. However,
EPRI anticipates developing a program of field tests to evaluate several of these
systems in actual utility settings.

WO,
(p«xm)

250

200

150

Mo. 2 FUiL OH.

NATURAL <2AS

too -

1000

ISO-HUMIDITY

1200 1400 1600 1800 2000
FWWKJ Tf W»tRATURf If)

2200

Figure 1. Effect of fuel type and firing
temperature on uncontrolled NOx emission for
a typical heavy-duty gas turbine.

9-28

-------
Smf* Cycto Rota (Bta/kWi)

* at 151 oxygen, on a dry bssis

Figure 2. New source performance standard limits
for nitrogen oxide as a function of simple cycle
heat rate and fuel-bound nitrogen content.

CO EMISSIONS (ppnw)

Figure 3. Relationship between nitrogen oxides
and carbon monoxide emissions.

9-29

-------
*' '* ' ' ' ' ' « '

_ _____ _____ — l.l ,.4 IMf"\
MUr/fial—mi

. HOx reduction by water Injection

Increase in hydrocarbons due to water injection Carbon monoxide Increase due to water Injection

Figure 4. Effect of water injection rate on
nitrogen oxide carbon monoxide and hydrocarbon
emissions for gas turbine base load operation
with natural gas.

9-30

-------
STEAM / FUEL MASS FU3W RATIO

Figure 5. Effect of steam injection rate
on NOx control efficiency.

9-31

-------
i«r

Figure 6. Effect of water injection rate
o gas turbine performance.

9-32

-------
200

FUEL

- NATURAL GAS

SYMBOL

•

o ; € ( o ' o

BYPASS

VALVE

POSITION

(CLOSE)

90
(OPEN)

'• !
4S 35 , 20

; (CLOSE)

FUEL
RATIO
PILOT/MAIN

100/0

30/70

I , :

30/70 ; 30/70 i 30/70

i I

TURBINE INLET TEMPERATURE, K

Figure 7. Dry low NOx (DLNOx coiubustor
emission characteristics.

9-33

-------
\Q
Q\

-------
DEMONSTRATION OF AM AUTOMATED UREA INJECTION
SYSTEM AT ENCINA UNIT 2

JAMES H. NYLANDER
SAM DIEGO GAS AND ELECTRIC COMPANY
POST OFFICE BOX 1831
, SAN DIEGO, CALIFORNIA 92112
AND

M.N. MANSOUR AND DOUGLAS R. BROWN
APFLIED UTILITY SYSTEMS, INC.
1140 EAST CHESTNUT AVENUE
SANTA ANA, CALIFORNIA 92701

ABSTRACT

The design of an automated urea injection system that can follow
changes in boiler load under load dispatch is discussed. The
effect of urea injector design on NOx removal is presented.
Parameters influencing NO removal -and NH3 emissions have been
identified and are reviewed. These include mixing, combustion gas
temperature, and urea to NOx mole ratio (NSR).'

The data show NOx removal to increase with the increase in NSR.
Over a broad boiler load range (50-100 MWJ using an NSR of 1.0,
NOk removal ranged between 43 and 53 percent with NH3 emissions less
than 25 ppm. The use of multiple levels for the injection of urea
improves NOx removal and provides more uniform removal over the
load range. It also results in sxgnxfxcantly lower Nemxssxons
at reduced loads.

9-35

-------
DEMONSTRATION OF AN AUTOMATED UREA INJECTION
SYSTEM AT ENCIKA UNIT 2

BACKGROUND

The use of urea injection for NOx emissions control was developed
in 1979 by KVB, Inc. (1) under Electric Power Research Institute
|EPRI) sponsorship. The process involves the injection of an
aqueous urea solution into the boiler furnace to mix with the
combustion gas, reducing NO to N2 and H20.

In late 1985, San Diego Gas & Electric (SDG&E) • initiated an
engineering evaluation of the urea process in Encina Unit 2, a 110
MW gas- and oil-fired boiler. Early tests in 1986 and 1987 focused
on characterizing important process parameters and on quantifying
the level of N0X removal that can be achieved (2). Results
indicated that urea injection is a viable NO,, control technology
for utility boilers.and other combustion sources.

Recent efforts, which are the subject of this paper, focused on
providing a clearer delineation of the maximum N0X reduction
capability of urea injection and on the development and
demonstration of a commercial urea injection system.

OBJECTIVES

The primary objective of this program was to demonstrate that urea
injection can be applxed commercially to control emxssxons

during the day-to-day operation of electric utility power plants.
Specific objectives consisted of the following:

• Develop an automated commercial urea injection system
that can follow changes in boiler load under normal
load dispatch conditions and maintain compliance with
specific NOx emission levelfs).

• Quantify the N0X removal capability of the process while
maintaining low NH3 emissions.-

9-36

-------
• Simplify the design of the urea injectors to improve
reliability and ease of maintenance.

• Quantify the composition and concentration of other
emissions (CO, amines, etc.} that may result from the
use of urea injection.

• Demonstrate commercial system reliability under long
term test conditions.

The work addressing the first three objectives has been completed
and is discussed in this paper, Work dealing with the remaining
objectives will be completed in the first half of 1989.

DESCRIPTION OF TEST BOILER

The demonstration of urea injection was performed on SDG&E Encina
Unit 2, a 110 MW gas- and oil-fired boiler built by Babcock and
.Wilcox (B&W). A side cross-section of the boiler is presented in
Figure 1. The boiler is fired with ten burners arranged in two
elevations with five burners- in each elevation. The unit operates
with a balanced draft furnace and is equipped with two vertical-
shaft Ljungstrom regenerative air preheaters.

The injection of urea was performed at two elevations. Elevation
A (Figure 1) contains nine ports, two on the side walls (ports 1A
and 9A) and seven on the front wall. Elevation B contains seven
ports, all of which are on the front wall. Urea injection at
Elevation A was performed, using ports 1A, 3A, 4A, 6A, 7A and 9A,
At Elevation B, ports 2B, 3B, 5B and 6B were used. Elevation B is
approximately 12 feet below Elevation A.

To avoid making penetrations into the boiler water wall, all ports
used for the injection of urea were existing view ports . Depending
on the boiler load, the location of these ports did not always
correspond to the optimum combustion gas temperature for urea
injection.

Testing of urea injection was performed while firing both natural
gas and residual fuel oil'. Due to current fuel prices and
availability, most of the testing was done on oil firing. Baseline
NO emissions for oil firing as a function of load ranged between
200 and 300 ppm (® 3% 02) as presented in Figure 2. All data
presented in this paper are for residual fuel oil firing.

9-37

-------
AUTOMATED INJECTION SYSTEM

A schematic of the urea injection system is presented in Figure 3.
The system was designed to receive 50 percent by weight
concentrated urea solution- in a storage tank (TK-1). The
concentrated urea is batch diluted to 35 percent to prevent
crystallization at low temperatures. Urea is transferred from the
storage tank (TK-1) to a day tank (TK-2) by a transfer pump (P-l),
The operation of the pump is controlled by two level switches
located in the day tank (TK-2). The switches transfer urea on
demand when the level in the day tank (TK-2) reaches a certain
minimum and stops the transfer pump (P—1) when the day tank (TK-2)
is full.

The urea is supplied to the boiler by an injection pump (P-2). A
water pump (P— 3) supplies water to dilute the urea to the final
concentration before injection into the boiler furnace. The urea
and water are mixed in a static mixer prior to injection.

The flows of urea and water are regulated by two flow control
valves. These- valves receive their signal from controllers which
determine the amount of flow. Two flow meters for the urea and
water provide flow indication to the controllers. The injection
system can operate with the water flow being directly proportional
to the urea flow or can be programmed to maintain a certain total
flow.

PROCESS PARAMETERS

Urea process parameters that influence N0X removal and NH3 emissions
are mixing, combustion gas temperature, urea to NO, mole ratio and
the initial NOx concentration in the combustion gas. A discussion
of each of these parameters is presented below.

Mixing

An obvious requirement for achieving high removal of NOx is to
uniformly mix the urea with the combustion gas. Mixing is affected
by the urea injector design, location, and operation and must take
into account bulk furnace fluid mechanics which will vary with the
boiler design, load and other operating conditions.

The injector design is by far the most important factor in
achieving good mixing. Uniform coverage of the furnace cross-
section is best provided using wide spray angle injectors and high
spray momentum. Atomizers producing coarse sprays at high

9-38

-------
injection velocity provide the best performance. This is because
the large drops with a high injection velocity have sufficient
lifetime and momentum to disperse and mix with the combustion gas
before the urea is released.

Combustion Gas Temperature

Combustion gas temperature is also an important process variable.
It affects both N0X removal and NH3 emissions.

Laboratory work sponsored by EPRI (1) has shown a strong dependence
of NQX removal on combustion gas temperature. The maximum N0X
removal was achieved in the laboratpry at temperatures in the range
of 1800-1900°?. Similar dependence on temperature has been
observed in the Encina evaluation. However, high removal of NOx
(>50 percent) was obtained at an average combustion gas temperature
ranging from 1460-2000°F.

In a practical application, wide variations in combustion gas
temperature exist at any given boiler load. These cause the urea
to be exposed to a range of temperatures when injected into the
furnace. The NOx removal achieved thus represents the composite
effect of these temperature variations.

Both laboratory and field data have shown NH3 emissions to be a
strong function of combustion gas temperature. Injecting urea at
temperatures of 1900°F and higher significantly reduces NH3
emissions-. Injection at lower temperatures (<18QQ°F) results in
a rapid increase in NH, emissions-. The laboratory data suggested
that maximum NOx removal and low NH3 emissions could be achieved by
injecting urea'at the optimum temperature of 1900°F.

Urea to NO. Mole Ratio

The total amount of urea injected is important,. It is' expressed
in this paper by the term Normalized Stoichiometric Ratio (NSR).
This term represents the ratio between the amount of urea injected'
and the theoretical amount required to reduce all of the N0X. An
NSR greater than 1.0 xnd^-cates that excess urea has been m^ected«

Increasing the NSR provides more urea to react with NO(. This
initially increases the level of NO, removal achieved. A point of
diminishing returns is reached, however, where a further increase
in NSR provides only a modest increase in NOx removal.

The NH3 emissions dependence on NSR is dominated by the effect of

9-39

-------
the combustion gas temperature. .If the combustion gas temperature
is relatively low (<1800°F), a rapid increase in NH3 emissions
results at high NSR. At high combustion gas temperature ( >1900°F) ,
the excess urea is oxidized to N0X, and thus the NSR has only a
minimal impact on increasing NH3 emissions.

Initial N0r Concentration

The effect of the initial N0X concentration in the combustion gas
on N0X removal has only been documented in the laboratory. These
data show that NO removal increases with increases in initial NOx.
The dependence of NQX removal on initial NO^ has not been observed
at Encina. This may be because the change in initial NOx was not
significant ( 200-300 ppm @ 3% 02) . Other process variables
pertaining to combustion gas temperature and mixing in the Encina
boiler appeared to influence more directly the level of NOx removal
achieved.

ATOMIZERS SCREENING

The effect of mixing on process performance was evaluated by
determining the N0X removal achieved by different types of urea
atomizers. The atomizers examined included several spray patterns
with different drop size distributions and were operated at a range
of pressures. A description of these atomizers and a discussion
of the NGX removal achieved by each is presented below.

Atomizers Tested

All atomizers tested represented commercially available mechanical
atomizers. The spray patterns for these atomizers included three
generic types that are labeled for the purpose of this paper as A,
B and C. The 'design details of these atomizers are considered
proprietary to the program sponsors.

The atomizers were mounted essentially flush with the furnace walls
with minimum penetration5 into the furnace cavity. This simplified
the maintenance and improved the reliability of the injection
system.

Atomizer Performance

NOk removals for the three types of atomizers are compared in
Figure 4. The figure presents the range of N0X removal achieved by

9-40

-------
each type of atomizer as a function of NSR. The data show that
the atomizer design (or urea mixing) can have a significant effect
on N0X removal. ¦

At an NSR of 1.0, NOx removal ranged from a minimum of 41 percent
to a maximum of 48 percent, and at an NSR of 2.0, the removal
ranged between, 55 and 67 percent. The achieved improvement in
process performance by selecting the appropriate atomizer design
was 17 percent for an NSR of 1.0 and 24 percent for an NSR of 2.0.

UREA PERFORMANCE - SINGLE LEVEL INJECTION

Laboratory work has shown that combustion gas temperature
influences both N0X removal and NH3 emissions. To investigate this
effect in the field, urea was injected over a range of boiler
loads. The urea was injected at either Elevation A or B (Figure
1). For a given boiler load and initial NOs concentration, this
permitted investigation of the effect of combustion gas temperature
on process performance.

The change in combustion gas temperature with boiler load was
documented on oil firing for Elevation A by performing temperature
traverses. A summary of these data is contained in Table 1. The
temperature measurements were made at the furnace exit and upstream
of the secondary superheater. Ports A2 through A8 were used to
make the temperature traverses at the furnace exit, and Ports A1
and A9 were used to make the measurements upstream-of the secondary
superheater. In general, the average combustion gas temperature
is lower at the furnace exit than upstream of the secondary
superheater. The data show that the average combustion gas
temperature increases from 1460°F to 2000°F as the boiler load is
increased from 50 to 110 KW, (Table 1).

No temperature measurements were made at Elevation B. However,
based on its position, the combustion gas temperature at this
elevation must be higher than at Elevation A because Elevation B
is closer to the flame zone. Injecting urea into Elevation B thus
provided a qualitative indication of the impact of higher
combustion gas temperature on process performance.

Testing was performed using an NSR of 1.0 and 2.0. The number of
injectors used at each elevation and their orientation were
optimized to obtain maximum NOx removal. Testing was performed
between boiler loads of 50 MW and 100 MW. Full load testing of 110
MW was not possible due to malfunction of a feedwater heater. The
dependence of process performance on boiler load (combustion gas
temperature) as identified during this testing is presented below.

9-41

-------
NO, Removal

The effect of boiler load (combustion gas temperature) on N0X
removal for an NSR of 1.0 and 2.0 is presented in Figure 5. The
figure shows NOx removal as a function of load using the boiler
load as an indirect indication of the combustion gas temperature.
The N0X removals achieved by injecting urea at Elevations A and B
are compared.

The data show that N0X removal is increased with the increase in
NSR. At an NSR of 1.0, NO removal ranged between 43 and 53
percent; at an NSR of 2.0, the removal ranged between 59 and 68
percent. N0X removal appeared less sensitive to combustion gas
temperature in the field testing. In the laboratory, at an initial
NOx concentration of 500 ppm, rsmova 1 dropped from 90 percent

to less than 10 percent when the combustion gas temperature was
reduced from 180O F to 1675°F. In the field testing, high removal
of NOx was obtained over a broad boiler load range (50-100 MW) .
This corresponded to an average combustion gas temperature between
1460°F and 2000°F. For this temperature range, the NOj, removal
varied between 59 and 68 percent for an NSR of 2.0. High NOx
removal was thus achieved in the field in spite of the relatively
low average combustion gas temperature at which urea was injected.

The moderate dependence of N0X removal on combustion gas
temperatures observed in the field testing is not fully understood.
One explanation pertains to the variations in combustion gas
temperature that-exist at any -given load. While an "average"
combustion gas temperature may be as 1ow as 1460 F, the variation
in temperature can spread over a 350°F range. The N0X removal
observed in the field' thus represents the composite effect of a
range of temperatures..

Injecting urea at Elevation A provided higher N0X removals than at
Elevation B throughout most of the load range. This trend is the
same for the two NSR values tested. The higher removal achieved
at Elevation A has been attributed to a more suitable combustion
gas temperature for the urea reduction of NCX.

A peak in NOx removal has been identified. This shows that the
process performance initially improves-1'when the load is increased.
Beyond a certain load level (depending on injection elevation and
NSR), a deterioration in performance occurs.

The initial improvement in performance with load is explained by
the increase in combustion gas temperature. The combustion gas
temperature at 50 MW, for example, is lower than optimum (average
temperature of 1460°F), and therefore an increase in boiler load

9-42

-------
(or combustion gas temperature) leads to an improvement in process
performance. That is why NQX removal at 50 MW is higher for
Elevation B than Elevation A. Being closer to the flame zone,
Elevation B offered a more suitable combustion gas temperature at
low load conditions.. Increasing boiler load beyond a certain level
increases the combustion gas temperature at both elevations.. This
results in a reduction in the level of NOx removal achieved.

It must be recognized that many operational variables, including
furnace wall cleanliness, furnace excess 0^^ burner firing pattern,
burner tip cleanliness (in the case of oil firing), and amount of
FGR can affect the combustion gas temperature. It is important to
be able to quantify the effect o'f these variables on the combustion
gas temperature in order to optimize the urea system performance.
Real tlme combustion gas temperature monitorxng during normal day*
to-day operation of the boiler would significantly improve the
ability to optimize process performance.

Fi111i8 b ions

The effect of boiler load (combustion gas temperature)' on NH3
emissions is presented for NSRs of 1.0 and 2.0 in Figure 6.
Similar to the N0X removal data, the combustion gas temperature is
represented in the figure by boiler- load.

The NH3 data are consistent with the N0X removal data and are also
¦in agreement with the data obtained in the laboratory. For both
NSR values, NH3 emissions are lower when urea is injected at
Elevation B. A reduction in boiler load or lower combustion gas
temperature results in higher NH. emissions for both elevations.
At very high boiler load (100 MW) , the level of NH3 emissions, did
not change with the injection elevation (A or B) because the
combustion gas temperature is relatively high at both elevations.
High NSRs, resulted in a corresponding increase in NH^ emissions,
suggesting NH3 is a byproduct of unreacted urea.

To avoid., excessive NH3, urea should be injected at a relatively
higher temperature. According to the laboratory data, the optimum
temperature that, would provide high NOx removal and low NH3
emissions is 1900°F. Again, in a field application, a uniform
combustion gas temperature is highly unlikely; therefore, the
process performance would reflect the composite effect of existing
temperature variations. Injecting urea at the upper end of the
temperature range would generally ensure low NH, emissions.

9-43

-------
UREA PERFORMANCE - BILEVEL INJECTION

Favorable results obtained at each injection level suggested that
enhanced process performance would be achieved with multiple
injection levels. In a multiple injection configuration, lower NSR
is provided at each injection elevation. This improves urea
utilization and, in turn, the overall N0X removal. The lower NS'Rs
also result in reduced NH3 emissions.

To quantify the potential benefits of this injection method,
testing was performed with urea simultaneously injected at both
elevations A and B. The split in urea flow to Elevations A and B
was optimized for each load to achieve maximum N0X removal. A
computer model based on single level injection data was developed
to define the optimum split in urea flow between the two
elevations. Urea flow to the top elevation varied from 40-70
percent of the total depending on the boiler load, with the balance
of the urea delivered to the bottom elevation. The results of this
bilevel injection:are discussed below.

NO, Removal

The effect of bilevel injection on N0X removal for an NSR of 1.0
and 2.0 is shown in Figure 7. The shaded area in the figure
designates, the range of, NO^ removals achieved by single level
injection |infection at Elevation A or B)• The plotted data show
the removal achieved using bilevel injection.

The bilevel injection results have shown that NOx removal achieved
by this method equals or exceeds removals obtained with single
level injection. More importantly, bilevel injection made NOx
removal essentially constant across the investigated boiler load
range. For an NSR of 1.0, the N0X removal ranged between 43 and 53
percent, and at an NSR of 2.0, the removal ranged between 59 and
68 percent.

NH^ tehH aalona

NH3 emissions with.-bilevel injection for an NSR of 1.0 and 2.0 are
presented in Figure 8. Again, the range for NH, emissions obtained
with single level injection is shown in the figure by the shaded
area. The bilevel injection data are plotted and show a
significant reduction in NH3 emissions at low load. This is the
result of reduced NSR at each injection elevation and the injection
of a portion of the urea at higher combustion gas temperatures
(Elevation B).

9-44

-------
PLANNED FUTURE WORK

Additional work is planned in the first half of 19 89 to complete
the remaining program objectives. The roost important part of this
work will be performing long term testing of the urea injection
system. The purpose of this effort is to verify the system
reliability during' normal day-to-day operation of the Encina
boiler. The intent of the test is to also quantify the long term
N0X emission reduction capability of the technology. Other work to
be performed will include the following:

• Assess the NO^ and NH3 concentration stratification in
the Encina boiler flue gas. Use these data to optimize
the urea injection rate to improve process performance;

• Determine the urea flow rate required to achieve
various specific NO, emissions target levels. Using
these data, program the urea controller to allow
operation in the automatic mode while maintaining
compliance with specific emission targets;

• Evaluate the automatic operation of the urea injection
system when boiler load is changed manually;

• Evaluate the automatic operation of the urea injection
system under normal load (dispatch controlled)
conditions;

•' Quantify the potential for other emissions ( CO^ amines ,
etc.).

CONCLUSIONS

Based on- the data' presented in this paper, the following
conclusions are reached:

• Variables that influence the performance of the urea
injection process are mixing, combustion gas
temperature,- and NSR. Based on laboratory data, a
combustion gas temperature of 1800-19 Q0°F offers' the
maximum N0X removal and minimum NHS emissions. Systems
should be designed to inject urea as close as possible
to this temperature range. Injecting urea at the upper
end of this .temperature range ensures low NH3 emissions;

9-45

-------
• N0X removal is increased with an increase in NSR, At
an NSR of 1.0, NO removal ranged between 43 and 53
percent, with NH3 emissions of less than 25 ppm;

•- Use of bilevel injection improves N0X removal at high
and low boiler loads, providing more uniform NOx removal
over the load range. It also results in low NH3
emissions at low loads;

• Boiler-specific combustion gas velocity and temperature
measurements are necessary for optimum urea injection
system design. The use of heat transfer modeling to
estimate the boiler combustion gas temperature where
measurements are limited also enhances' the optimum
system design;

• The atomizer design influences the mixing .of urea with
the combustion gas.

• Although not employed in this program, cold flow
modeling' may be helpful,in selecting the location and
orientation of urea injectors to ensure good mixing
with the combustion gas.

9-46

-------
ACKNOWLEDGMENT

This work was in part funded5 by Southern California Edison Company
(SCE), Electric Power Research Institute {EPRI), Los Angeles
Department of Water and Power (LADWP), Consolidated Edison Company
of New York (ConEd), and Empire State Electric Energy Research
Company (ESEERCO),

The financial and' technical contributions made by these
organizations are gratefully acknowledged. Special thanks are due
to David Eskinazi of EPRI and Lawrence Muzio of Fossil Energy
Research Corporation (FERCO) for their assistance in the
preparation of this paper. Thanks are also extended to Robert L.
Cravens and Andris R. Abele (1! of AUS for their support of the
field tests. Technical contributions made by Greg C. Quartucy
of KVB throughout this program are also gratefully acknowledged.

Presently with Environmental Energy Research.
Presently with FERCO.

REFERENCES

1. L.J. Muzio and J. K. Arand. "Homogeneous Gas Phase
Decomposition of Oxides of Nitrogen." EPRI Report No. FP-253—
1976.

2. M.N. Mansour, et al. "Full Scale Evaluation of Urea Injection
for NO, Removal." 1987 Joint Symposium on Stationary
Combustion N0X Control, New Orleans, Louisiana, March 1987.

9-47

-------
y—PORT IA

SCREEN
" TUBES

¦*—PORT 9A

SECTION

€)

PORT IB =
28:
3B -
4B =
5B :
6B :
7B:

SECTION

FIGURE

PORTS

IA{F.S)

9A(N.S.)

•Qy

PORTS

2A-BA

PORTS
»B-7B-

BOILER SIDE CROSS-SECTION

SCEMATIC OF ENCINA UNIT 2

-------
NOx, ppm (
-------
NO* K

SIGNAL 1 1/

STEAM

FLOW

SIGNAL

5°%

urea«=^>-

WATER

*=>-

TK-I

i r
—*QiV}

p-i

.—^LSH^

• —(lsl^

TK-2

CZKZZ3

P-2

CHCZZZ3-

P-3

v_y FCV.|

F E

STATIC
MIXER

FCV-2

TO BOILER
INJECTION
N022LES

FIGURE 3 UREA INJECTION SYSTEM

-------
NOx Removal, percent

70
65

50" />"vVX/

/¦ :X;XA#2

• VvV-

v.vy#
'//$¥
yW

0.5

1.5

NSR

9-51

-------
cn
r\>

NOx Removal, percent

6Q "

10 '

NSR - t.Q

-(- Upper Elevation ~ Lower Elevation

i 1 r

60 70 ao

Boiler Load, MW

100

NOx Removal, percent

NSR • 2.0

-j- Upper Elevation ~ Lower Elevation

~~r~

1 1 1—

70 HO 90

Boiler Load, MW

Figure 5. Effect of Combustion Gas Temperature on NOx Removal

-------
NH3, ppm

NH3, ppm

50-

-|- Upper Elevation f| Lower Elevation

NSR - 1.0

.J"

60 70 80

Boiler Load, MW

120

90 100

Upper Elevation Q Lower Elevation

NSR ¦ 2.0

70 80 90

Boiler Load, MW

100

Figure 6. Effect of Combustion Gas Temperature on NH3 Emissions

-------
NOx Removal, percent

80"

10"

NSR - 1.0

!¦' ¦¦! Single Level Range Hi Bitovel Injection

1 j

70 SO

Boiler Load, MW

100

NOx Removal, percent

NSR - 2.0

m Slnfile Level Range ~~Eh Bilavel Injection

60 80 70 ao SO

Boiler Load, MW

100

Figure 7. Effect of Bilevel Injection on NOx Removal

-------
50

20
10

60 SO TO 60 90 100

Boiler Load, MW

NH3, ppm

PI Single Levsl Range -Q- Bilsval Injection

NH3, ppm

140

o 1 i • -

SO 60 70 ao 90 too

Soiier Load, MW

Figure 8. Effect of Bilevel Injection on NH3 Emissions

-------
TABLE 1. COMBUSTION GAS TEMPERATURE DATA SUMMARY - OIL FIRING

Furnace Exit Upstream Secondary

Superheater (2)

Boiler Temp., Avg, Temp. Temp., Avg. Temp.

Load, °F Temp., Range, °F Temp., Range,

MW ^F °F

52 1270-1620 1462 350 1575-1680 1643 105

81 1430-1920 1710 490 1775-1930 1853 155

110 1580-2050 1881 470 1935-2080 2020 145

(1> Ports 2A-8A
(2> Ports 1A & 9A

9-56

-------
RETROFIT IMPLICATIONS OF A
LOW NOX BURNER SYSTEM ON A
230-MW OIL- AND GAS-FIRED BOILER

James J. Carnevale
Jeffrey A. Klueger
Los Angeles Department of Water and Power
Los Angeles, California 90051

ABSTRACT

The Low NOx Burners were selected as the most appropriate and cost-effective tech-
nology to demonstrate a 50-percent NOx reduction on our Haynes Generating Station
(HnGS), Unit 3, a Babcock and Wilcox (B&W) 230-MW opposed-fired, oil and gas unit..

The burners are based on Babcock-Hitachi Technology and were purchased from B&W
for $1,700,000 on a furnish only basis. The options for additional hardware and
engineering support by B&W brought the total contract to almost $2,300,000.

In addition to the burners, the Department anticipates over $9,000,000 in retrofit
costs including asbestos removal and disposal; compartment!ng the windbox; adding
a primary gas injection system; purchasing and installing new flue gas recircu-
lation fans and motors; adding ductwork to mix flue gas into the combustion air;
adding ductwork to supply air to the NOx ports; adding new controls; and solving
numerous interference and access problems. ¦

This paper will discuss some of the specifics of the retrofit.

BACKGROUND

About four years ago, the Los Angeles Department of Water and Power initiated a.
study to determine the current state of the art of the various NOx control tech-
nologies. One of the objectives of the study was to determine the most cost-
effective technology that could be retrofitted to our larger oil- and gas-fired
boilers to achieve at least a 50-percent NOx reduction. A second objective was
to select one of our boilers to serve as the demonstration unit for a full-scale
installation of the selected technology.

9-57

-------
The technology selected was the Low NOx Burners, and the demonstration unit
selected was our HnGS, Unit 3. The engineering go-ahead for the project was
received on January 20, 1987, and testing is scheduled to begin on May 30,
1990 {See schedule Figure 8),

The HnGS, Unit 3, boiler is a 230-MW (B&W), subcritical, single reheat unit that
was put into service in August 1964. The details of the boiler design are shown
in Table 1.

The HnGS, Unit 3, is currently limited to 125 ppm of NOx while burning natural gas
and 225 ppm of NOx while burning fuel oil (both limits at three-percent excess
oxygen). These limits are being met by a combination of combustion modifications
including four of the 12 burners out of service and overfire air ports (or NOx
ports). Gas recirculation is supplied to the hopper only for steam temperature
control and is not used for NOx control.

DISCUSSION

The Low NOx Burners are guaranteed to reduce NOx to 65 ppm- on gas and 110 ppm on
fuel oil containing 0.15-percent fuel nitrogen (both values at three-percent
excess oxygen). In order to achieve these guarantees, the following major modifi-
cations to the generating unit will be made:

1. Replace the 12 burners with Low NOx Burners of the primary gas-dual
register design.

2. Replace the open wirdbox with a fully compartmented windbox (including
asbestos removal).

3. Replace the four overfire air ports with dual register overfire air
ports.

4. Add ductwork to direct the flue gas from the discharge of the gas
recirculation fan to the mixing plenum where the flue gas is injected
through the airfoils into the combustion air.

5. Add a primary gas ductwork system to direct the flue gas to each burner.

6. Replace the two 800-HP (each) gas recirculation fans and motors with two
1,500-HP (each) fans and motors.

7. Replace, modify, and supplement most combustion-related controls and
instrumentation .

9-58

-------
8. Make numerous modifications to walkways, stairs, and piping systems.

9. Add crossover ducts to both the forced draft and gas recirculation
systems to allow unit operation with a forced draft or gas recircu-
lation fan out of service.

In order to better visualize some of the changes required, "Before" and "After"
figures are included. Figure 1 shows a cutaway view of the boiler in the current
or "Before" condition. Figure 2 is the companion flow schematic for Figure 1.
Figure 3 ("After") shows the additional ductwork required to compartment the' wind-
box, to provide preheated air to the overfire air ports, and to provide discrete
flue gas to each burner. Figure 4 is the companion flow schematic to Figure 3.

Each of the major modifications listed above is discussed below:

BURNER REPLACEMENT

The originally supplied Peabody burners will be replaced with B&W Primary Gas-Dual
Register Burners (PG-DRB). The PG-DRBs (Low NOx Burners) control NOx through the
use of delayed mixing, flue gas recirculation, and fuel staging. The Low NOx
Burners will allow up to 20 percent of the total flue gas to be recirculated to
the combustion process through several different methods.

At some operating conditions, all of the recirculated flue gas (up to 20 percent
of the total flue gas) will be mixed through specially modified airfoils into
the combustion air downstream of the air preheater. At other operating conditions,
the recirculated flue gas flow may be split so that 15 percent is mixed with the
combustion air and the remaining five percent will be injected as pure flue gas
(primary gas) to the burners. All ignitors and flame detectors will also be
replaced.

The existing burners use mechanically atomized fuel-oil guns. Fixed gas spuds
are used for natural gas firing. The Low NOx Burners will use steam-atomized
fuel-oil guns and adjustable gas spuds for natural gas firing. Provision has
been made for the mechanically atomized fuel-oil guns to fit and operate satis-
factorily in the Low NOx Burners, and some testing will be done with this con-
figuration. Emissions are not guaranteed for the mechanically atomized guns.

9-59

-------
CQMPARTMENTED WIND30X

Although the Low NOx Burners are of the dual register design, the registers are
used to control mixing and are not used to control flow. Therefore, in order to
properly control and balance flow, each Low NOx Burner requires that an upstream
flow-monitoring device and a flow-control device be provided. In order to meet
these requirements, our open windbox must be replaced with a compartmented windbox,
utilizing 12 separately monitored and controlled flow paths (one to each burner).
The additional ductwork required to achieve this is shown in Figure 5. A major
cost and schedule impact of this task is the asbestos removal and disposal from
the existing ductwork and piping.

OVERFIRE AIR PORTS

The unit is currently eouipped with four unsophisticated overfire air ports which
will be replaced with four-dual register overfire air ports supplied by B&W as
part of the burner contract. One damper and flowmeter upstream of each port will
control the total flow to each port. The penetration and mixing of the airflow
will be controlled by a sliding damper and spin vanes within each port.

Since the overfire air ports require pure heated air (without recirculated flue
gas), their air supply originates just downstream of the air preheater and up-
stream of the flue gas mixing airfoils. The additional ductwork required to
achieve this is shown in Figure 6.

FLUE GAS RECIRCULATION' DUCTWORK

This ductwork is new and directs the flue gas from each of the'gas recirculation
fan discharges to the mixing plenum sections of the combustion air duct. The
flue gas is injected through specially designed slots in the airfoils into the
combustion air. The air/flue gas mixture is then ducted to each burner. The
additional ductwork required to achieve this is shown in Figure 7.

9-60

-------
PRIMARY GAS DUCTWORK

This ductwork directs the pure flue gas directly to each burner and is a combi-
nation of the existing ductwork (no longer in use) and the new ductwork. This
system is shown in Figure 8 with the new ductwork having a circular cross
section.

GAS RECIRCULATION FANS

The existing gas recirculation fans provide gas recirculation to the hopper for
steam temperature control only. In order to provide gas recirculation to the
combustion air and primary gas to the burners, the fan discharge pressure had to
be increased. This required new fans and motors with, nearly double the existing
horsepower (800 each to 1500 each).

INSTRUMENTS AND CONTROLS

Because of the higher degree of control accuracy and speed, as well as the
numerous additional monitoring and control inputs required by the Low NOx
Burners, a distributed control system was selected to replace the existing
pneumatic system. In addition, the added flexibility of a distributed
system should simplify start-up.

MODIFICATIONS TO WALKWAYS, STAIRS AND PIPING SYSTEMS

Because of the massive amount of new ductwork required, numerous interferences
were created to walkways, stairs, and piping systems. A partial list of inter-
ferences is shown in 'able 2.

CROSSOVER DUCTS - FORCED DRAFT AND GAS RECIRCULATION FANS

With the existing open windbox, each forced draft fan is able to feed all 12
burners. With a compartmented windbox, the "A" fan feeds six burners, and
the "B" fan feeds the other six with no interconnection from side to side.

9-61

-------
In order to ensure stable boiler operation while one forced draft fan is out of
service, a crossover duct will be provided to connect the discharge of the two
forced draft fans. By the same reasoning, a crossover is required to connect the
discharge of the two gas recirculation fans.

ESTIMATED COSTS

Table 3 shows a breakdown of the estimated costs to complete this project.

9-52

-------
Figure 1
"Before" condition

9-63

-------
JJGEND

COkBUSHON MR

V7/////A C«S REC1RC

NOTE: 1 Ml S SCHEMMIC REPRESEIJIS
UNE-HM.F OF 1 He BOILER

MS RtCIOC F*N

FIGURE 2

flow schematic

COMBUSTION AIR & FLUE CAS DUCTS
HAYNES GENERATING STATION UNIT 3

-------
C\r \ IDF 7

riwUnL J

LOW NOx BURNER DUCTWORK MODIFICATIONS
UNIT 3

HAYNES GENERATING STATION

9-65

-------
cn
o\

CGMBUS1 ION AIR

CROSSOVER DUCT

LEGEND

COMBINATION AIR/CAS RECIRC MIXTURE 10 InDIYJOIIAl BURNERS
CCMBUS1 I Oti AiR

Y//////A «s Ktcmc

[X'XXXI OlSCRElt Cii RECIRC

mil: 1 MfS SCHEMATIC REPRESENTS
ONC-H4LF Of 1 HE SOIlES

FIGURE 4

LOW NOx BURNER SYSTEM FLOW SCHEMATIC
COMBUSTION AIR & FLUE GAS DUCTS
HAYNES GENERATING STATION UNIT 3

-------
FIGURE 5

LOW NOx BURNER DUCTWORK MODIFICATIONS
SECONDARY AIR SYSTEM DUCTWORK UNIT 3
HAYNES GENERATING STATION

9-67

-------
FIGURE 6

LOW N0X BURNER DUCTWORK MODIFICATIONS
OVERFIRE AIR SYSTEM DUCTWORK UNIT 3
HAYNES GENERATING STATION

9-68

-------
FIGURE 7

LOW NOx BURNER DUCTWORK MODIFICATIONS
GAS REC1RC SYSTEM UNIT 3
HAYNES GENERATING STATION

9-69

-------
FIGURE 8

LOW NOx BURNER DUCTWORK MODIFICATIONS
DISCRETE FLUE GAS SYSTEM UNIT 3
HAYNES GENERATING STATION

9-70

-------
4 A ? ft £ 3 UNIT 1

WORKING jCH^DUlF

RUN 6JANS3 13 <10

LOW NQX BURNER STSTEfl

!1QD£ O/FC

PR0JEC1V2 86A2

MAJOR EVENT SCHEDULE

SORT F START

i MTFRtf AL 1 MO?trHSS)

SCHFDUlE 3^ CHART

START engineering

|o I mTo"

!38fl

J F M A I M J J A S 0 N D

PURCHASE Lriat

23JJN6B

TpTjn [ a |h | j | J |T]1

[FJn](i

RETAIL OE51GN OF DUCTWORK ~

ESTABLISH LNflS CONTROL LOO IS 01 AM AM # 170CTS6
. MMAL DOCTUOHX LAYOUT DRAWINGS ^2i2CTS6

I >

AWARD ?:ew FlUE '/A3 RiCiSCUuATIDfi f- AN "J SDECSb

shop rAe^icAna.*i or ductwork

25SEPB9

3JAN5?

jj i A j L j-J h |D

i 2wJu'i_53
i

ALL HATFRIAL AND C UP¦ '¦ TO fiEi.0 2IJJL89

HATfiES 3

71? |m|a|h|j|"j|a|

SIR 35StT?09 L
FIElO COflSTRUCTlGri 25SEP09 Q

.7J1AR-30

START UP (CAS'MECHANICALLY ATCMIZtD OIL

I 25MAR90

ZSHARDO I I "i Aicmzroi ~ 3,!)f.<-n

I J I A

Schedule

Reproduced from
besi available copy.

-------
Table 1

Haynes Genera
Boiler

Manufacturer
Superheat Steam Conditions

Reheat Steam Conditions

Burners

Burner Arrangement
Fuel

Gas Recirculation Fans

ing Station, Unit 3
Design Data

Babcock & Wilcox

1,575,000 Ib/hr
1000°F
2050 psig

1,400,000 Ib/hr

1000°F

552 psig

12 Peabody Mechanically Atomized

Opposed-Fired, 4 per Deck

Low-Sulfur (0,251) Fuel Oil
and/or Natural Gas'

2 per unit, 800 HP each

9-72

-------
Table 2

Haynes Generating Station, Unit 3
Low NOx Burner System Retrofit Cost Summary
List of Interferences

A. Items interfering with Flue Gas Recirculation System.

1. Fire protection piping.

2. Heat tracing steam trap drain waste tank.

3. Horizontal run of five' pipes.

4. Walkway support beam.

5. Air heater wash piping valve station,

6. 12-inch seal air lines running north-south on both sides of unit.

B. Items interfering with the Primary Gas Ductwork.

1. Windbox drain valves and piping at both front and rear.

2. A seal air line above tempering air plenum.

C. Items interfering with the Compartmented Windbox.

1. Fire protection hose reels.

2. Burner cabinets and cabinet platforms.

3. All burner piping and control stations.

4. Seal air piping and booster fans.

5. Air heater wash drain piping and access platforms.

6. Two beams.

7. Attemperator valve stations.

8. Walkways.

9. Observation doors.

10. Stair tower.

11. TV cameras and wall openings.

9-73

-------
Table 3

Haynes Generating Station, Unit 3
Low NOx Burner System Retrofit Cost Summary

Sys tern

1. Low NOx Burners, Overfire
Air Ports, Dampers, Ignitors,
Flame Detectors, and Engineering
Support

2. Compartmented Windbox (Includes
asbestos removal)

3. Overfire Air Ports

4. Gas Recirculation Ductwork

5. Primary Gas Ductwork

6. Gas Recirculation Fans & Motors

7. Instruments and Controls

8. Modifications to Stairs,
Walkways, and Piping Systems

9. Crossover Ducts

Subtotals
Total

Engineering and Demolition and
Material Cost Erection Cost

52,300,000
$1,000,000

Included in 1
S 300,000
$ 200,000
S 700,000
$1,300,000

$ 700,000
$ 150,000
$6,650,000

$ 700,000
$2,150,000

Included in 1
$ 200,000
$ 50,000
$ 500,000
$ 900,000

$ 500,000
$ 150,000
$5,150,000

$11,800,000

9-74

-------
RETROFITTING LOW-NOx BURNERS FOR GAS AND OIL FIRING

J. Gerdes, Jr. & R, Waibel

John Zlnk Company
Tulsa, Oklahoma, USA

L. Radak

Southern California Edison Co.
Rosemead, California, USA

ABSTRACT

There is a growing necessity to consider environmental concerns in the operation of all
commercial combustion processes. Controls on emissions released to the environment are
being tightened across the nation and indeed throughout the world. However, performance
aspects- are as important as emissions, especially to the operators of the facilities. It is
important to understand both the capabilities of new emission control technology and how
that technology will affect system reliability and operating flexibility, especially when retrofitting
to existing equipment.

The John Zink Company has been engaged in a continuing program with the Southern
California Edison Company to apply Zink's low NO* staged fuel technology to utility boiler
burners. Development and testing has been conducted on various sized units ranging from
25 million Btu/hr to 122 million Btu/hr, Ambient air, preheated air and flue gas recirculation
tests have been conducted. The latest John Zink boiler burner development work involves full
scale testing of both a 95 million Btu/hr staged fuel AVC burner designed to operate with flue
gas recirculation and a 122 million Btu/hr staged fuel/staged air AVC burner. Special attention
has been paid to burner operating characteristics that affect stability and reliability.

There is a growing emphasis on controlling all types of emissions emanating from industrial
and commercial processes, not the least of which is combustion related emissions. Due to

INTRODUCTION

9-75

-------
recent and pending legislation, most utilities are faced with the need to reduce NO* emissions.
The Southern California South Coast Air Quality Management District has been a- leader in the
effort to control the amount of combustion products released to the atmosphere. Available
NOx control options tend to fall into one of two groups - control of the combustion process
(and thereby the production of NOx) or-chemical/catalytic treatment of the NOx laden exhaust
stream. The former, control of the combustion process, tends to be a less expensive
approach. Southern California Edison (SCE) has been engaged in a program with John Zink
Company to develop the burner retrofit technology to economically control the production of
NOx emissions through combustion system design. Although maintaining acceptable emis-
sion levels is critical from the environmental aspect, the affect any new NOx control technology
has on reliability and system operation is equally important and should be understood.

COMBUSTION SYSTEM NOx CONTROL METHODS

There are several methods of controlling NOx formation through burner design and combus-
tion system modifications. These include: burners out of service; low excess-air operation;
flue gas recirculation; air staging; and fuel staging. These techniques reduce NOx emissions
either by controlling the peak flame temperatures or by limiting oxygen availability in the region
of peak flame temperature. Before the discussion of this specific work, a brief description of
each of these techniques is given below.

FUEL STAGING

Staged fuel NO* control, developed and patented by John Zink Company, is accomplished
by injecting part of the fuel together with all the combustion air into the primary combustion
zone. Thermal NOx production in this region is limited by the low flame temperatures that
result when fuel is burned with high excess air levels. NOx production is low despite the high
oxygen concentrations because peak flame temperatures are low. The remainder of the fuel
is then injected into the tail of the flame emanating from the primary combustion zone. Partial
combustion of the fuel in the primary region provides inerts that reduce peak flame tempera-
tures in the secondary combustion zone while lowering the local concentration of oxygen.
Both factors limit the NOx production in the secondary combustion zone.

9-76

-------
Fuel staging is the most effective burner design technique of controlling gas fired thermal NOx.
Flame lengths for staged fuel flames are shorter than comparable staged air flames and, unlike
air staging, low excess air levels are possible. Excess air levels of 5% (1% excess O2) or less
are possible, which further enhance emission reductions.

AIR STAGING

For NOx control with fuels containing fuel bound nitrogen, such as No, 6 oil, the most effective
NOx control technique is air staging. NOx generated with such fuels comes from two sources
- thermal NOx and fuel NOx (that attributable to fuel bound nitrogen).

Air staging is accomplished by dividing the combustion air into two or more streams. A
sub-stoichiometric volume of the air is introduced in the primary zone to form a fuel rich
combustion region in which the fuel is only partially burned. A portion of the fuel bound nitrogen
decomposes to form molecular nitrogen thus reducing fuel NOx formation. The localized
sub-stoichiometric operation also lowers the peak flame temperature thereby reducing the
thermal NOx generated.

The rest of the air is injected down stream of this fuel rich primary combustion zone forming
a secondary bum-out region in which combustion is completed. NOx formation in this
secondary combustion zone is limited because primary zone products of combustion reduce
both the flame temperature and oxygen concentration.

BURNERS OUT-OF-SERVICE/OVER FIRE AIR

Operating with burners out-of-service is a rudimentary form of air staging. This technique
utilizes burners fired sub-stoichiometrically with the additional combustion air added elsewhere
in the firebox. Some systems provide this supplemental air through "Out-of-Service", upper
tier burners, while other systems have dedicated air ports located above the firing burners (
ie. "Over Fire Ports) which provide this air.

9-77

-------
LOW EXCESS AIR

Low excess air operation provides some control over production of both thermal and fuel NOx
by reducing the availability of oxygen during combustion. The level of control attainable,
however, is limited because combustion efficiency suffers as excess oxygen levels approach
zero. Operation at very low levels of excess oxygen results in increased carbon monoxide
and soot emissions.

FLUE GAS RECIRCULATION

Flue gas recirculation (FGR) combined with low excess air operation provides a reduction in
thermal NOx, although it has little effect on NOx caused by fuel bound nitrogen. The
recirculated- products of combustion injected into the combustion air stream provide inert
gases which lower the adiabatic flame temperature and, to some extent, the local oxygen
concentration. Beyond 15% FGR, the incremental effectiveness of added flue gas recirculation
diminishes. The maximum FGR level is generally limited by flame stability to about 25% to
30%.

EXPERIMENTAL SETUP

The John Zink Company International Research Center includes nine test furnaces of various
sizes and configurations that allow testing of burners ranging from less than 1 million Btu/hr
heat release to over 250 million Btu/hr. Burner development work is typically conducted on
full scale equipment, in this case burners rated for 95 million Btu/hr and 122 million Btu/hr.
Testing full sized equipment under conditions simulating the customer's field operating
conditions permits the direct interpretation of the data. It also permits testing of upset or startup
conditions in a controlled manner rather than on-line in the field.

The furnace used for this testing (Furnace #3) is a cylindrical, horizontally fired, water jacketed
chamber, 11 feet in diameter by 29 feet long. Integral to the furnace design are 15 sight ports

9-78

-------
providing an unobstructed view of the combustion process. Forced draft combustion air was
supplied through adjacent ductwork. The system can be configured for either ambient air or
preheated air (500°F - 600°F) operation at rates in excess of 122 million Btu/hr. In addition,
the system can be configured to simulate FGR operation (up to 25%) at various burner inlet
air temperatures.

The furnace draft and windbox pressures were measured using manometers. A suction
pyrometer measured stack gas temperatures. Stack emissions (O2, CO & NO*) were
continuously monitored using a water cooled probe to obtain exhaust gas samples. The
samples were dried and filtered using a Thermo Electron sample conditioning system. NOx
was measured with a Thermo Electron Chemiluminescent Model 10 analyzer; oxygen with a
Teledyne Model 326A analyzer; and carbon monoxide with a Thermo Electron Model 48 Gas
Filter Correlation CO Meter. All instruments were calibrated before and during testing using
certified span gases.

Typically the fuels tested are natural gas and fuel oil, ranging from diesel oil to #6 oil. However,
the John Zink test facility has the capability of using, customer supplied liquid fuels or blending
various gaseous components to simulate specific fuel composition. Gas firing rates are
determined using the burner pressure vs. heat release capacity curve after verification of the
capacity curve with a turbine meter. Oil flow rates are determined using inline flow meters.

BURNER DESCRIPTION

Figure.2.depicts John Zink's patented staged fuel burner design. It shows a cross-sectional
view of an AVC boiler burner. The John Zink AVC (Adjustable Vortex Control) series boiler
burner offers maximum flame pattern adjustability with a wide range of flame stability. The
AVC design allows operation at excess air levels of less than 0.5% excess oxygen. The AVC
burner is available in four basic designs; the standard emission series, the staged air series,
the staged fuel series, and the combined staged fuel/staged air series.

9-79

-------
The "Adjustable Vortex Control" assembly allows a full range of air contro! from axia! flow to
high swirl. The main burner air is divided into two streams designated primary and secondary
air, each with their own control register. On "Staged Air" burners, tertiary air registers are
added. This permits control of the air going to the secondary combustion zone.

The AVC burner design utilizes different NOx control mechanisms for oil and gas operation.
Tertiary air on fuel oil firing provides the air staging for NOx control. On natural gas, staged
fuel is used to control NOx emissions. Fuel staging is accomplished on gaseous fuels using
primary and secondary fuel injectors.

The results given herein are based on the testing of three burners, including comparisons to
previous results reported in "Fuel Staging Burners for NOx Control." The current work
compliments the earlier results with information obtained on larger capacity burners. The
burners tested were:

• Staged Fuel AVC rated for 95 million Btu/hr, gas fired with Flue Gas Recircula-
tion with 330°F

• Staged Air/Staged Fuel AVC rated for 122 million Btu/hr, No. 6 fuel oil and gas
fired with 620°F preheated air

The data from these tests are compared with data from previous tests on similar AVC burners
reported in "Fuel Staging Burners for NOx Control."

DESCRIPTION OF TESTS AND RESULTS

OIL FIRING

Although the bulk of the development work was concentrated on natural gas firing, limited
testing was done with a standard John Zink PM oil atomizer - a six-port, steam assisted Y-Jet
type oil gun. The fuel used was a heavy fuel oil with 0.3 wt percent fuel bound nitrogen. CO
emissions were low under all conditions. Most readings ranged from 0 ppm to 50 ppm, with

9-80

-------
the highest being 103 ppm. At the normal heat releases, the NOx ranged from 234 ppm to
250 ppm.

The oil flames were well defined. At high fire the oil flame tended to be straight and narrow.
At a firing rate of 122 million Btu/hr and 3% excess oxygen, the flame envelope was
approximately 10 feet in diameter and 26 feet long. As the oxygen was reduced to 2%,
combustion intensity decreased and the tip of the flame occasionally reached into the furnace
stack.

GAS FIRING

Utility boilers require burners with broad operating envelopes to ensure reliable operation even
during upset conditions. Multiple burner units need the capability of lighting both the pilot and
main fuel on out-of-service burners while the system is under load. For this reason it was
important to determine the stability of an individual burner at low fuel firing rates with the
combustion air at maximum flow.

The AVC gas burner was demonstrated to be stable under a variety of system upset conditions.
The primary gas injectors and the tile profile have been designed to provide a stable region to
burn the primary fuel. This serves a piloting function for fuel injected into the secondary
combustion zone. Specific upset conditions that were tested and were demonstrated to be
stable included:

• Pilot lightoff with full air flow

• Main burner lightoff at low fire with full air flow

• Full burner air flow with minimum gas flow

• Full burner air flow rapidly spiking fuel flow from low to high fire and back to low
fire

• Minimum fuel flow rapidly spiking air flow from low to high fire and back to low
fire

• Stable operation with cooling steam flowing through the oil gun

9-81

-------
Figure 3 shows stable operation points with full air flow. Fuel turndown with both the primary
and secondary fuel injectors in operation is approximately 4:1, limited by the stability of the
flame in the secondary combustion zone. As illustrated, turndown can exceed 20:1, even with
full air, if the secondary fuel injectors are turned off.

The NOx emissions for the 122 million Btu/hr AVC staged fuel/staged air burner were compared
with previous data on two 100 million Btu/hr AVC burners (one staged fuel burner, one
conventionally fired burner). Data from the most recent test correlate well with the previous
data. The results presented in Figure 4 show that under the test conditions, fuel staging
reduced NOx emissions by 50% compared to a conventionally fired burner with ambient air
and by nearly 70% with 600°F preheated air.

FLUE GAS RECIRCULATION

When very low NOx levels are required, flue gas recirculation (FGR) can be combined with fuel
staging to achieve exceedingly low NOx emissions. Figure 5 shows the results of two staged
fuel AVC burners tests - a 95 million Btu/hr burner operated at 330°F preheated air and a 50
million Btu/hr burner operated at 500°F preheated air. The data presented in Figure 5 indicate
the effect on NOx emissions with varying percentage of flue gas recirculation on staged fuel
burners. Greater than a 50% reduction was observed as the percentage of flue gas recircula-
tion was increased from 0% to 25% on both stage fuel burners. Note that the data show that
a burner operating with 500°F preheated air in a 2200°F combustion chamber provides NOx
emissions similar to a burner operating with 330°F preheated air in a 2400°F combustion
chamber. It is well known that NOx is strongly influenced by factors that affect the peak flame
temperatures. Some of the most important factors that influence flame temperature, in addition
to fuel composition, are air preheat temperature and combustion chamber operating tempera-
ture.

9-82

-------
CONCLUSIONS

There are two equally important aspects of utilizing low NOx technology to control process
emissions, The first is the level of emissions obtained with this technology. The second is the
affect on operations of implementing this technology. John Zink Company has developed
information In both areas using full scale equipment. Test results have been presented
indicating a reduction in NOx on the order of 50% or more with staged fuel over a conventionally
fired burner, depending on operating conditions. A further reduction of 50% is obtained when
FGR is used in conjunction with fuel staging.

The data from these recent tests show that it is possible to design a low NOx burner that can
also meet stringent stability requirements. In addition, the recent data correlate well with the
previous data. The1 data confirm the dependence of NOx generation on combustion chamber
and air preheat temperature. The data also confirm the ability of fuel staging combined with
FGR to produce very low NOx levels.

9-83

-------
GAS ANALYSIS

Figure 1. Staged Fuel Boiler Burner Test Installation

Figure 2. John Zink Staged Fuel/Staged Air LoNOx
AVC Boiler Burner

9-84

-------
Figure 3: Extended Burner Turndown
with Maximum Air Flow

AIR PftEHEAT TEMPER ATUflE,8 F

Figure 4. Comparison of NOx vs. Air
Preheat for Various Firing Configura-
tions at 2% Excess Oxygen

9-85

-------
120-

1 1 §«

100-

90*

ao-

\ \

u 80-
O

e 50-

a 40-

EL
0.

O 30-
z

20-

10"

0 50 M Btu/hr., SCWP
~ 95M8tu/hr.,330°F

5 10 <5 20 as

* FUJE GAS RECRCUUTION

Figure 5: Staged Fuel Burner:

NOx vs. FGR (2% O2, 33Q°F and 500°F)

9-86

-------
References

1. Waibel, R., et al., "Fuel Staging Burners for NOx Control", 1986 Symposium on Industrial
Combustion Technologies, Chicago, Illinois.

2, Martin, R. R., "Burner Design Parameters for Flue Gas NOx Control", John Zink Internal
Report, 1981.

3. Gerdes, Jr., J. H., "AVC Test with FGR", John Zink Internal R&D Report, July 1988.

4. Gerdes, Jr., J. H., "AVC Test with Preheated Air", John Zink Internal R&D Report, Decem-
ber 1988.

9-87

-------
ENGINEERING EVALUATION OF COMBINED NCySQ^
REMOVAL PROCESSES: INTERIM REPORT

W. DePriest
Project Manager
Sargent & Lundy
Chicago, Illinois

J. B. Jarvis

Senior Chemical Engineer
Radian Corporation
Austin, Texas

0. E. Cichanowicz
Project Manager
Electric Power Research

Institute
Palo Alto, California

ABSTRACT

In recent years, international pressures for strict control of N0X emissions have
rivaled those historically directed at S0X. Japan has required postcombustion
N0X control since the late 1970s and plans in Europe call for up to 30,000 MW of
N0X control capacity by 1990, Presently, the technology most often selected to
achieve this control is selective catalytic reduction (SCR). These international
developments may encourage U.S. regulatory agencies to consider some level of
additional N0X control for coal-, oil-, and gas-fired generation. This poten-
tial, in conjunction with the existing strict SO2 control requirements, has
renewed interest in the development of lower cost S0X and NQX control technolo-
gies.

Under Electric Power Research Institute (EPRI) Project RP 3004-1, Sargent & Lundy
along with its subcontractor Radian has been commissioned to perform an investi-
gation of combined N0x/S0x processes. The goal of this project is to identify
lower cost alternatives to the separate technologies of SCR for N0X control and
conventional flue gas desulfurization (FGO) far S0X control.

9-89

-------
Over 60 combined N0x/S0x processes were identified at the project outset as
potential candidates for application to the utility industry. These 60 processes
can be placed in the following six categories:

* solid adsorption/regeneration processes,

* irradiation of the flue gas processes,

* wet scrubbing processes,
gas/solid catalytic processes,

* electrochemical processes, and
alkali injection processes.

This paper represents an interim progress report of this investigation with
approximately 309£ of the project completed.

9-90

-------
INTRODUCTION

The increasing regulation of NQX emissions in Europe and the stringent regula-
tions in place in Japan for nearly 10 years may increase the pressure on U.S.
regulatory agencies to promulgate some, form of N0X control on fossil-fueled power
generation facilities here in the United States. Presently, selective catalytic
reduction1 (SCR) for NOx control and flue gas desulfurization for S02 control have
the most widespread use in both Europe and Japan. However, this combination has
some significant shortcomings, primarily the very limited experience with SCR on
high-sulfur fuels and the combined costs for both technologies.

An in-depth technical and economic understanding of potential alternatives to SCR
and FGD by the utility industry could identify a minimum cost control strategy
for N0X and S02 emissions., If the utility industry can identify the most cost-
effective alternate technologies and implement a development program to bring
these technologies to commercial feasibility, then they may be able to minimize
the capital investment and effect on plant efficiency typically associated with
controlling NQX and S02.

Another major motivation for this study is to reduce the process supplier's
uncertainty in applying their technology to the power industry. This study will
identify weaknesses in the processes to help focus future development expendi-
tures and ultimately reduce the risk to the suppliers and the users.

Over 60 combined NQx/S02 processes were identified at the project outset as
potential candidates for application in the utility industry. The goal of this
project is to select the most promising processes from these candidate processes
to meet specific criteria of interest to utility power generators. We will also
identify what additional development work will be necessary to reduce any per-
ceived risk associated with the selected process technologies.

DESIRABLE FEATURES

To facilitate the comparison of these combined M0x/S02 processes with SCR and
FGO, six desirable features have been identified as follows:

low/modest capital cost,

low/modest operating cost,

9-91

-------
reliability,

* capability to follow turndown and load swings accompanying cycling
plants,

* minimum long-term environmental risk, and

* broad applicability within the power industry,

PROJECT APPROACH

The project approach for evaluating, alternate technologies for controlling both
N0X and SOg has been organized into, six major tasks.

Screening Methodology (Task 1}

In this task a methodology was developed for selecting, from a population of over
60, approximately eight processes for detailed evaluation in later tasks. The
number of processes actually selected may vary depending on the outcome of this
screening methodology. This methodology has been designed to consider both the
desirable features of a process from a utility's operating perspective and the
applicability of each process for a given plant design and environmental control
requirements. This task has been completed, and details of the methodology are
presented in the Interim Results section of this paper.

Process Information (Task 2)

This task consisted of gathering the necessary data from the process designers to
support the questions asked by the screening methodology developed in Task 1.

This task is nearly complete with information continuing to be received from some
of the process suppliers.

Process Selection (Task 3)

This task involves the application of the Task 1 methodology using the Task 2
information. This task is currently underway, and the eight final process tech-
nologies should be identified by April 1989.

9-92

-------
Preliminary Assessment and Process Economics (Task 4)

The objective of this task is to develop specific designs for each of the eight
processes and use the results in a technical feasibility and cost analysis. An
important component of this task will be establishing critical design premises
that define a consistent and fair basis for evaluating and comparing the eight
processes. Based on these criteria, a conceptual engineering design (flow
sheets, P&IDs, major equipment 11st, consumables, and Q&M and capital costs) will
be developed and used to identify effects on plant heat rate, reliability, and
balance-of-plant costs for each process. This activity is scheduled to begin in
April 1989 and end in August 1989.

Sensitivity Analyses (Task 5)

The sensitivity of the individual-process costs to variations in the critical
design premises and economic parameters will be studied in this task scheduled
for the latter part of 1989.

Research and Development (Task 6)

In this task the development needs for each process will be assessed. These
development needs may be directed towards research required to bring"the process
to commercial viability or research that will simplify the process design and/or
improve the economics.

SCHEDULE

The duration of the project is 24 months with a start date of May 1988.

INTERIM RESULTS

Highlights of the activities completed to date are summarized as follows.
Preliminary Screening Methodology

The starting point for this project was a planning study conducted for EPRI by
Battelle that identified 58 processes as potential candidates for combined
N0x/S02 control. An additional seven processes were identified through other
sources. To reduce the number of processes to a more manageable level, a preli-
minary screening process was initiated. Each of the 65 process developers was

9-93

-------
contacted by telephone or letter. The objective was to determine the operation
of the process and obtain initial information about the operation of the
process. In order to continue in the evaluation, processes had to meet the
following minimum requirements:

* Some degree of experimental development had to have occurred on
the process. That is, purely conceptual processes were
eliminated from the evaluation.

Some level of development had to have occurred within the last
5 or so years. Processes inactive for longer than this were
considered abandoned.

Process developers had to be willing to provide additional
information on the process needed for evaluation.

Processes had to be applicable to controlling emissions from
coal-fired boilers. Some processes, either as a result of the
reagent used (e.g., high cost or limited' availability) or the
type of processing required, were judged to be better suited
for use in other industries or smaller applications and,
therefore, not applicable to the utility industry.

The above approach yielded 28 process developers engaged in research on various
N0x/SC'2 processes. These developers represent 18 clearly defined technologies
and 24 processes that have differences significant enough to warrant separate
evaluation.

Process Categorization

To facilitate the qualitative evaluation, processes meeting the preliminary
screening requirements were sorted into categories based on process type. Six
categories resulted, and the processes assigned to each category are listed in
Table 1. A brief description of the six categories, including the ultimate
"fate" of both the N0X and SC^ as well as some of the advantages and disadvan-
tages follows.

Solid Adsorption/Regeneration. The processes in this first category lead to the
following results:

* fate of SOg - sulfuric acid, sulfur, or liquified S02; and

* fate of N0X - reduction to nitrogen.

9-94

-------
TABLE 1

PROCESSES UNDER CONSIDERATION FOR COMBINED N0X/S02 CONTROL

Overall Process Group
Solid Adsorption/Regeneration

Process Name

Irradiation of the Flue Gas

Wet Scrubbing

Gas/Solid Catalytic Operations

Electrochemical Operations
Alkali Injection Operations

UOP/PETC Fluidized-Bed Copper Oxide Process
Rockwell Moving-Bed Copper Oxide Process
NOXSO Process

Mitsui/BF Activated Coke Process
Sumitomo/EPDC Activated Char Process
Sam"tech Nelsorbent S0x/N0x Control Process

Ebara E-Beam Process
Karlsruhe Electron Streaming Treatment
ENEL Pulse-Energization Process
CE Pulsed Streamer Corona Process
Concord UV S0x/N0x Control Process

Saarberg-Holter Iron Chelate Scrubbing

Argonne/Dravo ARGONNOX Process

Oow Electrochemical Regeneration Process

Dow Polychelant/UltrafTitration Process

PETC Electrodialysis Process

California (Berkeley) Ferrous Cysteine Process

Hal dor Topsoe WSA-SN0x Process

Degussa DESONOX Process

B&W S0x/N0x/R0x/B0 (SNRB) Process

Parsons Flue Gas Cleanup Process

Lehigh University Low-Temperature SCR Process

IGR/Helipump Solid-State Electrochemical Cell

Argonne High-Temperature Spray Drying Studies
PETC Mixed Alkali Spray Dryer Studies
Batelle ZnO Spray Dryer Process

9-95

-------
The processes in this category use a recirculating solid sorbent material to
remove SOg arid N0X from the flue gas. Sorbents used for this purpose include
activated carbon or coke, alumina-based sorbents with additives such as copper
oxide or sodium, and magnesium oxide-based sorbents.

In general, S02 is bound to the sorbent as sulfate and is released as either SO2,
or a mixture of SO2, ^S, or elemental sulfur, during regeneration in a reducing
atmosphere. Typical reducing gases can include natural gas, hydrogen, carbon
dioxide, or synthetic gas. The concentrated sulfur species released during
regeneration can be treated to produce byproducts, which include sulfuric acid,
liquified SO2, or elemental sulfur.

All of the processes in this category ultimately reduce N0X to nitrogen. How-
ever, the techniques to accomplish this vary. In some cases, N0X is catalytical-
ly reduced in the- adsorption step by adding a reducing agent such as ammonia to
the flue gas. This approach is used with activated carbon or coke sorbents and
in processes utilizing copper oxide. In other processes, N0X is absorbed from
the flue gas and reduced during sorbent regeneration. In one process, absorbed
NQX is released by heating the sorbent before adding the reducing gas. The
desorbed N0X 1s then recycled to the boiler where a portion of it reportedly will
be reduced to ^ within the burner flame zone. A steady-state N0X concentration
within the furnace is achieved that is somewhat higher than the original N0X con-
centration.

A contacting device is required to mix the flue gas and the sorbent. Typical
approaches involve the use of a fluidized or moving bed of the sorbent mate-
rial. Spent sorbent is continuously removed from the absorber, regenerated, and
then returned to the absorber vessel.

The processes differ in terms of the temperature regime in which they can be
applied. Copper oxide sorbents operate at about 700°F. Therefore, access to the
flue gas stream is required upstream of the air heater. Regeneration of the
copper oxide sorbent occurs at slightly higher temperatures. Other sorbents can
be used at lower temperatures (below 300°F) and can be applied downstream of the
air heater and/or particulate collection device. However, the required regenera-
tor temperature is typically much higher (above 1000'F).

9-96

-------
One of the major operating concerns for these processes is sorbent attrition.
Sorbent makeup costs can be high and are aggravated by the transport of solids
between and within the adsorption/regeneration steps.

Solid adsorption/regeneration processes have these primary advantages:

* most processes are capable of high SO2 and NQX removal levels;

• the processes produce no Targe waste streams or sludges;

* the processes can produce a potentially marketable byproduct
such as sulfuric acid, sulfur, or concentrated SO2; and

0 flue gas reheat is not typically required.

Solid adsorption/regeneration processes have these primary disadvantages:

~ sorbent recirculation of hot solids is a mechanically complex
operation;

• sorbent attrition (or consumption) may be a significant operat-
ing cost;

flue gas pressure drops through the reactor and subsequent
particulate collections device (if required) can be high; and

processes utilizing a fluidized bed or having a significant
temperature differential between the adsorption/regeneration
steps may have limited turndown capability.

Irradiation of-the Flue Gas. The processes in this second category lead to the
following results:

fate of SO? - ammonium sulfate or dry calcium sulfite/sulfate
solids, and

fate of N0X - ammonium nitrate or dry calcium nitrate solids.

Irradiation of the flue gas can be used to produce reactive intermediates that
subsequently oxidize SOg and NQX to their respective acids. In most cases,
ammonia is added upstream of the reactor so that ammonium sulfate and nitrate
solids are formed. These processes are applied following particulate control so
that the ammonium compounds can be used as fertilizers. A second particulate
control device is required to collect the dry ammonium salts.

9-97

-------
The most developed process in this group utilizes electron accelerators that pro-
vide high-energy electrons to the flue gas. There are, however, several other
techniques that have been investigated for irradiating the flue gas. For
example, an electrical pulse can be used to develop a corona discharge in a
reactor that has a configuration similar to an electrostatic precipitator
(ESP). The electrons generated during the pulse serve the same function as those
generated by an electron generator. Other techniques for irradiating the flue
gas include the use of microwave and ultraviolet radiation.

In general, irradiation processes operate in a temperature range between 150° and
200°F. Therefore, some cooling of the flue gas leaving the air heater is re-
quired to reach the optimum temperature range.

Processes employing irradiation of the flue gas have the following primary
advantages:

* SOo and NQX are controlled simultaneously in a single step for
which high removal levels can be obtained (at the expense of
higher power consumption);

processes'do not produce a high-volume waste stream;

* processes produce a potentially salable byproduct (ammonium
sulfate and nitrate), that could be used as a fertilizer;

* processes should have good load following characteristics; and

processes are suitable for retrofit application because access
to high-temperature flue gas is not required.

Processes employing irradiation of flue gas have the following primary disadvan-
tages:

process energy requirements are high;

* equipment for irradiating the flue gas is not yet available in
sizes required for application to utility boilers; and

the potential exists for emissions of ammonia, sulfuric acid,
and nitrous oxide (N20).

9-98

-------
Wet Scrubbing. The processes in this third category lead to the following
results:

fate of SO2 - calcium sulfite/sulfate sludge or ammonium
sulfate, and

* fate of NQX - conversion to nitrate and nitrogen-sulfur salts
or nitrogen.

Wet scrubbing processes can be used to absorb both S02 and N0X from flue gas.
Due to the low solubility of NO, however, special measures must be taken. These
include oxidation of NO to N02 in either the gas or liquid phase, or the use of
additives to increase the solubility of NO.

The first group of processes utilize an oxidant such as ozone, chlorine dioxide,
or hydrogen peroxide to oxidize NO to N0£. NO^ is more soluble than NO and
reacts to form nitrite and nitrate in the scrubbing liquor. Nitrite then reacts
with dissolved sulfite to form soluble sulfur-nitrogen compounds such as hydrox-
ylamine disulfonate (HADS). The principal drawbacks to this approach include the
high cost and hazardous nature of the oxidants and the generation of a waste
stream containing the soluble nitrogen salts (nitrate and sulfur-nitrogen
species). When chlorine dioxide is used as the oxidant, the waste stream also
contains chloride.

The second group of processes utilize an additive to increase the solubility of
NO. These additives include iron chelates such as iron EDTA, along with a
variety of other materials capable of forming complexes with NO. The iron-
chelate-NO complex can react with dissolved sulfite or other reducing agents to
form a regenerated iron chelate, nitrogen, and sulfur-nitrogen compounds such as
HADS. The drawbacks to this approach include oxidation of Fe to inactive Fe ,
generation of a waste stream containing soluble nitrogen-containing reaction
products, and high chelate consumption due to losses in the blowdown stream.

Development work on the iron-chelate processes has centered on techniques to
maintain iron in the active +2 form and to separate the iron chelate from the
dissolved sulfur and nitrogen salts in the blowdown stream. These methods
include electrochemical regeneration of the iron (+3), and electrodialysis and
filtration techniques to minimize losses of the iron-chelate.

9-99

-------
Both of the approaches described above are enhanced by elevated dissolved sulfite
concentrations in the absorption liquor. In the first case, sulfite helps to
absorb and react with NQ2. In the second case, sulfite is required to reduce the
iron-chelate-NO complex and also reduces a portion of the inactive iron (+3). In
order for sufficient sulfite to be present, the amount of SO2 absorbed from the
flue gas must exceed (by roughly a factor of two) the amount of N0X removed.
Therefore, these processes are better suited to high-sulfur fuels.

The fate of absorbed S02 depends on the type of chemistry employed in the
process. Since high dissolved sulfite is desirable, sodium-based (dual alkali),
magnesium-enhanced lime/limestone, or ammonia-based scrubbers could be uti-
lized. The lime/limestone and dual alkali chemistries produce a calcium
sulfite/calcium sulfate waste sludge unless gypsum is produced via forced oxida-
tion. An ammonia-based process, utilizing oxidation with ozone, could be used to
produce ammonium sulfate and nitrate. However, the product could be contaminated
with sulfur-nitrogen species, which might lessen its value as a fertilizer.

The primary advantages of wet scrubbing processes include the following:

processes could be applicable as retrofits to some existing wet
FGD systems;

S02/N0x removal can be obtained in a single vessel; and

* processes are suitable for high-sulfur systems.

The primary disadvantages of wet scrubbing processes include the following:

* processes use complex chemistries or employ hazardous reagents
(ozone or chlorine dioxide);

* processes produce a liquid waste containing soluble nitrogen
species (treatment of this stream will increase operating
costs); and

* high levels of NCL removal may not be obtainable using existing
FGD system contactor designs.

Gas/Solid Catalytic Operations. The processes in this fourth category lead to
the following results:

9-100

-------
* fate of S02 - sulfuric acid or sulfur, and
fate of N0X - reduction to nitrogen.

The processes in this category use a fixed catalyst bed {or beds) to remove S02
and N0X from the flue gas. Typically, the processes employ a catalyst bed that
has a design similar to that used for selective catalytic reduction (SCR). In
general, the processes operate at elevated temperatures and therefore require
access to flue gas before the air heater.

The fate of NQX is the same for all of these processes and involves the reduction
of N0X to nitrogen.. Reducing agents such as natural gas, ammonia, or synthetic
gas are used for this purpose. The fate of S02 varies, depending on the process,
with both sulfuric acid and sulfur as potential byproducts. No large volume
waste streams are generated In any of the processes.

Several of these processes use two catalytic reactors. In the first reactor, NGX
is reduced to nitrogen using ammonia injection and an SCR catalyst. In the
second reactor, SQ2 is catalytically oxidized to sulfuric acid. The sulfuric
acid is then condensed to give a salable byproduct. These processes are consi-
dered combined 502/N0x processes because the presence of the S02 oxidation cata-
lyst allows the SCR reactor higher levels of ammonia slip than in conventional
SCR systems. Ammonia leaving the SCR reactor is subsequently oxidized to N0X in
the second catalytic reactor. While this slightly decreases the net NQX removal
capability of the process, ammonia emissions are eliminated along with the poten-
tial for ammonium sulfate deposit formation downstream of the second reactor.

Another process is under development in which an SCR catalyst is located inside
the bags of a hot baghouse. Upstream of the baghouse, alkali is injected into
the flue-gas to remove S02. This generates a calcium- or sodium-based waste
stream. Ammonia is also injected upstream of the baghouse to promote NQX removal
in the SCR'catalyst. Prior removal of most of the S02, along with particulates,
may simplify the operation of. the SCR catalyst.

Development is proceeding on a process that uses a catalyst to reduce both N0X
and S02 (to nitrogen- and H2S). In this process, the H2S is recovered and con-
verted to sulfur.

9-101

-------
Finally, catalysts are being developed that may accomplish selective catalytic
reduction of N0X at lower temperatures. The goal is to develop a process that
can be applied more easily as a retrofit to existing boilers.

Gas/solid catalytic processes have these primary advantages:

• processes use fixed-bed reactors that do not require circula-
tion of solid sorbents;

high removal levels of SO2 and NQX are possible;

most of these processes can produce a potentially valuable by-
product with no large waste streams; and

stack gas reheat 1s not typically required.

Gas/solid catalytic processes have these primary disadvantages:

• most processes operate at elevated temperature (which in retro-
fit applications would require access to flue gas upstream of
the air heater);

the potential exists for secondary air emissions of species
such as ammonia, sulfuric acid mist, and ^S;

• some of these processes require the collection and concentra-
tion of dilute quantities of SO3 or HgS from the flue gas,
which increases the mechanical complexity of the process; and

- « catalysts may be deactivated by components in the flue gas,

which could necessitate periodic regeneration or replacement of
portions of the catalyst bed.

Electrochemical Operations. The processes in this fifth category lead to the
following results:

• fate of SO2 - sulfur or sulfuric acid,, and

• fate of N0X - reduction of nitrogen.

A number of electrochemical processes have been proposed for the removal of
S02. (Electrochemical potential can be used to either reduce S02 to sulfur or
oxidize it to sulfuric acid.) However, only one process has so far demonstrated
the capability for combined S0^/N0x removal. This process uses a solid-state
electrolytic cell with a ceramic electrolyte. In the cathodic zone of the cell,

9-102

-------
electrochemical reduction of SOg arid N0X yields sulfur, nitrogen, and oxygen ions
(which dissolve in the solid electrolyte). At the anode, oxide ions are con-
verted to oxygen gas. All three products—sulfur, nitrogen, and oxygen—remain
in the flue gas- Sulfur would be condensed downstream and sold as a byproduct.
The process is still in the early stages of development, and it is not yet clear
how the process would be applied at a coal-fired boiler. However, the cell
operating temperature appears to be in the range of 850°F to 950°F.

Processes employing electrochemical operations have the primary advantages:

* they are mechanically simple processes;

S02 and N0X are removed simultaneously;

no water, reagents, or other materials are required; and

elemental sulfur is produced as a byproduct with no other waste
streams being.generated.

Processes employing electrochemical operations have these disadvantages:

* sulfur recovery may be difficult because of the low sulfur con-
centration in the flue gas;

* . electrical requirements may be relatively high;

*¦ processes operate at high temperature, which complicates
retrofit applications; and

process is in early stage of development.

Alkali Injection. The processes in this sixth category lead to the following
results:

fate of S02 - dry calcium or sodium-based solid waste, and

fate of NQX - conversion to nitrogen and soluble nitrate or
sulfur-nitrogen compounds.

Several processes have been proposed that use alkali injection to remove SO^ from
flue gas. If an additive containing sodium is included, N0X removal can also be
obtained. Reportedly, sodium sulfite acts as a catalyst to oxidize NO to NC^,
which can react with sulfite to form nitrogen, nitrates, or sulfur-nitrogen com-
pounds.

9-103

-------
One process utilizing this approach involves the use of a spray dryer at a higher
than normal spray dryer temperature of around 20CPF. A sodium additive, such as
sodium hydroxide, is added to the lime slurry fed to the spray dryer, N0X
removal is optimized (at about 50% removal) at the higher operating tempera-
ture. However, SO^ removal decreases somewhat as the temperature is increased.
Test results show that most of the N0X removal occurs across the filter cake in
the baghouse.

A second approach, useful for lower sulfur applications, involves dry injection
of sodium compounds (e.g., sodium bicarbonate) into the flue gas leaving the air
heater. Due to NO oxidation,, however, NO2 concentrations may actually increase,
and can result in the formation of a visible, and therefore undesirable, brown
plume. Similar increases in NO;? levels have also been observed with the high-
temperature spray dryer approach. Test results have demonstrated that injection
of ammonia or urea with the sodium sorbent can reduce the tendency for plume
coloration.

Processes employing alkali injection have the following primary advantages:

« processes are suitable for relatively inexpensive retrofit
applications (e.g.-, existing spray dryers may be modified to
operate in the N0X removal mode); and

processes operate at relatively low temperatures (access to
flue gas upstream of the air heater is not required).

Processes employing alkali injection have the following primary disadvantages:

• N0X removals are relatively low (less than about 50%) and high
SO2 removals (greater than 90%) may be difficult to obtain when
N0X removal is maximized;

the potential exists for increased NO? emissions and the forma-
tion of an undesirable brown plume, although this problem may
be minimized by injection of ammonia or urea; and

« the dry solid waste stream contains soluble sodium compounds,
which could present some waste disposal problems and limit the
application of these processes at some existing utilities.

9-104

-------
Qualitative Screening Methodology

A detailed methodology has been developed to select the eight "best" processes
for more detailed evaluation.

A direct comparison of all 28 processes is not possible as each process differs
in its capability to remove N0X and SO2 and its applicability to existing plants
(retrofit) and new plants. Accordingly, three application categories for the
N0x/S02 processes have been defined to enable a direct comparison between and
among relevant processes.

These application categories are defined as follows:

* Near-term processes suitable for retrofit applications. Processes
in this category will be at or near a commercial status and will
be amenable to retrofitting within an existing plant site. As
these processes are primarily for retrofit application, extremely
high N0X and S0£ removal (e.g., 90%) may not be necessary, consis-
tent with the anticipated form of the possible acid rain legisla-
tion.

Near-term processes suitable for new plant applications.

Processes within this category will be at or near a commercial
status and will be suited for new plants where site constraints
are minimal. These processes will likely be required to meet high
N0X and SO? removal consistent with present and anticipated
revisions to NSPS requirements.

Long-term processes where significant research and development are
required' to achieve commercial status. Accordingly, these
processes may not be available for full scale use within the next
10 years. Processes in this category will most likely be limited
to those that can achieve the high removal necessary for NSPS.
However, processes that appear to be capable of achieving the
removal necessary to meet-anticipated acid rain legislation
requirements within, the anticipated time restraints will also be
included.

Since exact cost information is not available, a qualitative ranking will be used
to select the most promising processes. The qualitative ranking is based on the
desirable features listed in the beginning of this paper.

Several criteria were developed to qualitatively define these desirable
features and to allow a comparison to conventional state-of-the-art SCR and FED
technologies. For example, criteria that can be used to assess reliability

9-105

-------
include mechanical and chemical process complexity, tolerance to fly ash, and. the
presence of corrosive environments.

The processes being considered for combined S02/N0X control show considerable
variation in terms of these desirable features. With respect to costs, many of
the processes that produce a salable by-product have relatively high capital
costs due to the equipment required to collect, purify, and store the by-
product, However, these by-products can potentially be sold to offset operating
costs. Alternatively, while the wet scrubbing and alkali injection processes may
offer lower capital costs, these processes produce a sludge or solid waste
product that must be disposed of. Furthermore, these waste streams may
constitute a long-term environmental risk.

Similar trade-offs are evident for the other desirable features. For example,
the solid absorption/regeneration processes typically operate at a low tempera-
ture (which increases their applicability) and produce no large waste streams
(which minimizes long-term environmental risk). However, these processes require
circulation of the hot solid sorbent, which can adversely affect the reliability
and load-following capabilities of these processes. In contrast, the gas/solid
catalytic processes operate at higher temperatures, which decreases their general
applicability for retrofit applications. However, these processes are mechani-
cally less complex (which should result in better reliability) and may be better
able to follow load changes than the solid absorption/regeneration processes.

The process-developer-supplied information and engineering judgment on the part
of the project team will be used to give each process a score for each criterion
under each desirable feature. This score will represent the reviewer's opinion
on the impact that each criterion has on the desirable feature compared to
SCR/FGD,

For the cost-related desirable features, crude cost estimates will be prepared.
These crude cost estimates will, in all likelihood, be based on limited informa-
tion, However, a relative ranking among the processes will result.

The relative importance of each criterion will be considered by weighting each of
the criteria under each desirable feature. In addition, the relative importance

9-106

-------
of each desirable feature can be weighted. The resulting score for each process
will be determined by summing the weighted scores for each criterion under each
desirable feature. The framework for managing the qualitative evaluation has
been developed using a spreadsheet; however, the assignment of specific weighting
factors has been delayed until additional information on each process has been
obtained. The spreadsheet approach will allow sensitivity analyses to account
for the effects of alternative future utility environments.

SUMMARY

This project is approximately 30X complete and although no conclusions as to the
feasibility of these processes are clearly apparent, major steps have been taken
to achieve this goal. Research into the background and status of these processes
have allowed us to eliminate some of the processes and categorize the remaining
ones into similar process groups. We have also identified application categories
to encompass the forecasted emission scenarios for both new and retrofit power
generation facilities. These process categories and application categories will
facilitate matching the processes' respective capabilities (removal efficiency,
retrofitability, and status of development) with a range of applications and
ultimately facilitate a fair and equitable relative evaluation of the processes.

Future project activities will select and study in detail the eight most
promising processes capable of achieving reliable and low cost N0X and S02
control.

REFERENCES

Numerous sources (technical bulletins, product brochures, written correspondence,
telecons, etc.) of Information were used in the preparation of this interim
report on the evaluation of combined N0X/SC>2 remova^ processes. Readers of this
report should be aware that any conclusions presented in this report are
preliminary and subject to change as more definitive information is gathered from
the process developers. Specific reference information is available upon
request.

9-107

-------
Appendix A

EPRI/EPA Symposium on Stationary Combustion NOx Control
03/06/89-03/09/89
San Francisco, CA
Liat of Attendees

James R. Adams

Project Manager

New York State Electric & Gas

4500 Vestal Parkway, East

Binghamton, NY 13903

607/724-0849

Rui Afonso

Senior Research Engineer
New England Electric System
25 Research Drive
Westborough, MA 01582
508/366-9011

Ken Ahn
N/A

Coea Company
1510 Rollins Road
Burlingaae, CA 94010
415/697-0440

Charles D. Allen

Sr. Mechanical Consulting Engr.

Arizona Public Service Company

P.O. Box 53999

Phoenix, AZ 95072-3999

602/371-6750

5y B. Alpert
EPS I Fellow

Electric Power Research Institute
3412 Hillview Avenue
Palo Alto, CA 94304
415/855-2512

Leonard C. Angello
Project Manager

Electric Power Research Institute
3412 Hillview Avenue
Palo Alto, CA 94304
415/855-2873

James A. Armstrong

Vice President, Operations

ADA Technologies, Inc.

304 Inverness Way South, Suite 480

Englewood, CO 80112

303/792-5615

D. S. Arnold
Engineering Specialist
Kerr-McGee Corporation
P.O. Box 25861
Oklahoma City, OK 73125
405/270-2911

Wayne J. Aronson

Chief, Program Support Section

U.S. EPA

345 Court land Street
Atlanta, GA 30365
404/347-2864

Patrick F. Aubourg
Supervisor

Owens Corning Fiberglas
Tech. Center, P.O. Box 415
Granville, OH 43023
614/587-7604

A. Baldacci
Senior Engineer
ENEL

Via G. B. Martini 3
00198 Rome, ITALY
6/8509-2898

Lothar Balling
Dipl.-Ing.

Siemens AG, KWU Group
Hamnerbacherstrasse 12+14
0-8520 Erlangen
FEDERAL REPUBLIC OF GERMANY
9131/182350

A-1

-------
Joe Barsin

Manager, Industrial Projects
Babcock & Wilcox
20 S. Van Buren Avenue
Barberton, OH 44203
216/860-L552

Dwayne Bateson
Executive Assistant
TransAIta Utilities
110 12th Avenue, S.V.

Calgary, Alberta T2P 2M1
CANADA
403/267-7129

Nick Bayard de Volo
President

Energy Technology Consultants
2091 Business Canter Drive
Suite 100
Irvine, CA 92715
714/833-2522

Peter R. Beal

Manager, Business Development

NEI International Combustion, Ltd.

Sinfin Lane

Derby DE2 9GJ

UNITED KINGDOM

332/271111

Gordon Beales
Project Manager

Electric Power Research Institute
3412 Hillview Avenue
Palo Alto, CA 94304
415/855-2591

Rob Becker
President

Environmental Catalyst Consultants
P.O. Box 637
Spring House, PA 19477
215/628-4447

Joseph leer
Department Manager
Siemens AG, KVU Group
Hammerbacherstrasse 12+14
Dept. U3112, 8520 Erlangen
FEDERAL REPUBLIC OF GERMANY
09131/182932

Ted Behrens

Account Executive
Fuel Tech

61 Taylor Reed Place
Stanford, CT 06906
203/359-1320

Gerald R. Bemis
Engineer

California Energy Commission
R&D Office

1516 9th Street, MS-43
Sacramento, CA 95014-5512
916/324-3468

Angelo Benanti

Mgr., Thermal Equipment

ENEL

Via G. B. Martini 3
00198 Rone, ITALY
6/8509-2898

Knud Bendixen
Vice President
Burmeister & Wain Energy
23 Teknikerbyen
Virum, DENMARK 2830
45/2 857100

Mogens Berg

N/A

Elkraft Power Company
5 Lautruphoej
DK-2750 Ballerup, DENMARK
45/2 660022

Leif Bernergurd
Technical Officer
Environment Protection Board
Box 1302

S-17125 Solna, SWEDEN
8/799-1119

A-2

-------
Sisto Bertacchi
Senior Researcher
ENEL

Via A. Pisano 120
Pisa, ITALY 56100
050/535693

Gary L. Bisonett

Senior Engineer

Pacific Gas & Electric

245 Market Street, Room 434A

San Francisco, CA 94106

415/973-6950

Kerry Blaraire

Generation Studies

Nova Scotia Power Corporation

P.O. Box 910

Halifax, Nova Scotia

CANADA B3J 2W5

902/428-6655

Verle Bland

Manager, Applied Technology

KVB, Inc.

9342 Jeronimo

Irvine, CA 92718

714/587-2326

Richard Boardman
Student

Brigham Young University
184 CB

Provo, UT 84602
801/378-2076

Richard T. Bobick
Senior Process Engineer
Unocal Corp^jret Aojj
1201 Vest 5 th Street
P.O. Box 7600
Los Angeles, CA 90051
213/977-6435

Dieter Boekenbrink

Director

RWE

5180 Eschweiler, Box 1448
WEST GERMANY
02403-732112

Diego Bonolis

N/A

ENEL

Via G.B. Martini 3
00198 Rome, ITALY
06/85095192

Peter J. Booras
President

Yankee Energy Corporation
80 Boylston Street
Boston, MA 02116
617/542-0550

Richard W, Borio
Principal Consulting Engineer
Combustion Engineering
1000 Prospect Hill Road
Windsor, CT 06095
203/285-2229

Kerry W. Bowers
Senior Research Engineer
Southern Cornpany Services
P.O. Box 2625
Birmingham, AL 53202
205/870-6825

Wallace H. Bradley

V.P., Engineering

Austell Box Board Corporation

P.O. Box 157

Washington Street

Austell, GA 30001

404/948-3100

Jan Brandin

N/A

University of Lund
Chemical Engineering II
P.O. Box 124
S-221 00 Lund, SWEDEN
046/108284

Ron L. Bredehoft

Staff Planning Analyst

Chevron U.S.A.

324 W. El Segundo Boulevard

El Segundo, CA 90245

213/615-5000

A-3

-------
Bernard P. Breen
President

Energy Systems Associates
1840 Gateway Three
Pittsburgh, PA 15222
412/392-2380

Brent Brigham
Project Engineer
Unocal Corporation
P.O. Box 758
Wilmington, CA 90748
213/513-7600

R, G. Broderick
V.P.. Engineering
Fuel*Tech

61 Taylor Reed Place
Stanford, CT 06906
203/359-1320

W. J. Brooks
Senior Engineer
CEGB

Barnett Way, Barnwood
Gloucester GL4 7RS
UNITED KINGDOM
0452/652-306

Bert Brown

Applications Manager
Joy Technologies, Inc.
404 E. Huntington Drive
Monrovia, CA 91016-3633
818/301-1172

Douglas R. Brown
Engineer

Applied Utility Systems
1140 E. Cheatnut Avenue
Santa Ana, CA 92701
714/953-9922

Gordon Brown

Staff Engineer

Exxon Chemical Americas

5000 Bayway Drive

Baytown, IX 77522

713/425-5033

.A-4 ••

C. P. Brundrett
Manager

W. R. Grace & Company
10 E. Baltimore Street
Baltimore, MD 21202
301/659-9125

Mark R. Burkhardt

Research Chemist

ADA Technologies, Inc.

304 Inverness Way South, Suite 480

Englewood, CO 80112

303/792-5615

Kenneth R. Burns
Product Manager
Engelhard Corporation
Edison, NJ 08818
201/632-6640

Tony A. Burns

V.P. & Mgr., Env. Technology Group
S-Cubed, Division of Maxwell Labs
P.O. Box 1620
La Jolla, CA 92038
619/453-0060

John V. Byrne

Senior Research Chemist

Engelhard Corporation

Menlo Park

Edison, NJ 08818

201/321-5153

Scott Cameron

Sales Engineer

Babcock & Wilcox

17172 Abalone Lane, #207

Huntington Beach, CA 92649

714/846-0817

Gary A. Camody
Product Services Manager
Combustion Engineering
1000 Prospect Hill Road
Windsor, CT 06095
203/285-5039

-------
E. J, Carapobenedetto

Msrkdtm^

Babcock & Wilcox

20 S. Van Buren Avenue

Barberton, OH 44203

216/860-6762

Giovanni Caprioglio

Consultant

General Atomics

P.O. Box 85608

San Diego, CA 92138-5608

619/455-2918

E. J. Capriotti
Vice President, Sales
Fuel Tech

61 Taylor Reed Place
Stanford, CT 06906
203/359-1320

Janes Carnevale

Supervising Mechanical Engineer
Los Angeles Dept. of Water & Power
P.O. Box 111, Room 604
Los Angeles, CA 90051
213/481-4895

Baldwin K. Chan
Sr. Mechanical Engineer
Pacific Gas 5t Electric
77 Beale St., Room 2523
San Francisco, CA 94106
415/972-5236

Charles Chang

Assistant Supervisor, Air Quality
Los Angeles Dept. of Water & Power
111 N. Hop« Street
Los 1 es y 9^3012
213/481-3235

Shih-Ger Chang

Senior Scientist

Lawrence Berkeley Laboratory

University of California

Berkeley, CA 94720

415/486-5125

Wei-Yin Chen

Assistant Professor, Research
Louisiana State University
Department of Chemical Engineering
Baton Rouge, LA 70803
504/388-3059

Lisa J. Chrisman
Engineer

Fossil Energy Research Corporation
821 Hamilton Drive
Pleasant Hill, CA 94523
415/937-9007

Roger C. Christnan

Program Manager

Radian Corporation

13595 Dulles Technology Drive

Suite 200

Herndon, VA 22071
703/834-1500

Ed Cichanowicz
Project Manager

Electric Power Research Institute
3412 Hi11view Avenue
Palo Alto, CA 94304
415/855-2374

Leonard Clayton

Field Engineering

Bay Area Air Quality Mgmnt. Dist.

939 Ellis Street

San Francisco, CA 94109

415/771-6000

Robert J. Collette

Contract Manager

Combustion Engineering

1000 Prospect Hill Road, 5034-GC25

Windsor, CT 06095-0500

203/285-5687

Roger S. Cook

Environmental Engineer

Dept. for Environmental Protection

Kentucky Div. for Air Quality

18 Reilly Road

Frankfurt, KY 40601

502/S64-3382

-------
Frederick M. Coppersmith
Manager, Research & Development
Con. Edison Co. of New York
4 Irving Place
New York, NY 10003
212/460-3098

David A. Cowdrick
Senior Engineer
Tampa Electric Company
P.O. Box 111
Tampa, FL 33601
813/671-3361

Ed Cowle

Project Engineer
iechtel Power Corporation
12440 E. Imperial Highway
Norwalk, CA 90650
213/807-2423

Anthony Cowley

Mechanical Design Supervisor

B. C. Hydro

970 Burrard Street

Vancouver, British Columbia

V6Z 1Y3

604/663-2849

William R. Cress

Manager, Engineering Studies

Allegheny Power Service Corporation

Cabin Hill Drive

Greensburg, PA 15601

412/838-6721

Diane V. Croson

Mgr., LLW/NOx Process Dev.

Vestinghoustt Idaho Nuclear Company

P.O. Box 4000

Idaho Falls, ID 83403

208/526-3402

Michael Davidson

Mgr., New Product Development

Combustion Engineering

1000 Prospect Hill Road

Windsor, CT 06095

203/285-9005

Charles B. Davis
Senior Staff Engineer
Virginia Power
5000 Dominion Boulevard
Glen Allen, VA 23060
804/273-2619

Gerard G. De Soete

Dr. es Sciences

Institut Francais du Petrole

1-4, Avenue de Bois-Preau

Box N311, 92506 Rueil-Mailmaison

Cedex, FIANCE

331/47526145

Dennis D. Delaney
Senior Research Chemist
Unocal Corporation
376 S. Valencia Avenue
Brea, CA 92621
714/528-7201 X2184

Mukesh S. Desai
Engineering Specialist
Bechtel Power Corporation
15740 Shady Grove Road
Location 2A-12

Gaithersburg, MD 20877-1454
301/258-3150

Dilip Deshpende
Supervising Engineer
Alberta Power Company
10035 105th Street
Edmonton, Alberta
CANADA D5G 2V6
403/420-7177

William DePrlest
Project Engineer
Sargent & Lundy
55 E. Monroe Street
Chicago, IL 60603
312/269-6678

A-6

-------
Bill Dick

Design Specialist

Ontario Hydro

700 University Avenue

Toronto, Ontario

CANADA MSG 1X6

416/592-8593

Maurizio Didio
N/A

Snaraprogetti Spa

c/o Snamprogetti USA, Inc.

666 5th Floor

New York, NY 10103

212/399-1090

Richard Diffenbach
N/A

U.S. Department of Energy
P.O. Box 10940
Pittsburgh, PA 15236
412/892-6090

David J. Dodd

Mgr., Chemical Research

Ontario Hydro

800 Kipling Avenue

Toronto, Ontario

CANADA M8Z 5S4

416/231-4111 X6519

Roger Dodda

Air Quality Engineer

Wisconsin Electric Power Company

333 W. Everett Street

Milwaukee, VI 53201

414/221-2174

Zambelli Donato
Engineer

Azienda Servizi Municipal
Via. Lamarnora N. 230
Brescia, ITALY
030/3311-291

Yogesh P. Doshi

Senior Environmental Engineer

New Jersey Dept. of Env. Protection

401 E. State Street, 2nd Floor

Trenton, NJ 08625

609/633-7249

Dennis C, Drehmel
Technical Manager
U.S. EPA
AEERL -- MD-04

Research Triangle Park, NC 27711
919/541-7505

Joseph R. Dulovich
Plant Superintendent
Ohio Edison Company
1047 Belmont Avenue
Niles, OH 44446
216/544-7327

Michael K. Dunbar
Mechanical Engineer
Pacific Gas & Electric
77 Beale Street, Room 2547
San Francisco, CA 94106
415 972-0606

Richard A. Dye
Fossil Energy Department
U.S. Department of Energy
FE-4, Forrestal
Washington, DC 20585
202/252-6499

Owen W. Dykema
President

Dykema Engineering
5850 Canoga Avenue, Suite 400
Woodland Hills, CA 91367
818/712-0070

William C. Eddins

Dir., Division of Air Quality

Dept. for Environmental Protection

18 Reilly Road

Frankfurt, KY 40601

502/564-3382

Greg Eirschele
Environmental Engineer
Wisconsin Power & Light Company
222 W. Washington Avenue
Madison, WI 53701
608/252-3084

A-7

-------
John W. Eldridge

Professor of Chemical Engrng.

University of Massachusetts

39 Kendrick Place

Amherst, MA 01003

413/253-5991

Ton E. Emroel

Mgr., Combustion Engineering Dept.
Radian Corporation
P.O. Box 13000

Research Triangle Park, NC 27709
919/541-9100

Robert Epperly
Executive Vice President
Fuel Tech

61 Taylor Reed Place
Stamford, CT 06906'

203/359-1320

David Eskinazi
Project Manager

Electric Power Research Institute
3412 Hillview Avenue
Palo Alto, CA 94304
415/855-2918

Tony Facchiano
N/A

Coen Company
1510 Rollins Road
Burlingame, CA 94010
415/697-0440

Hamid Farzan

Senior Research Engineer
Babcock & Wilcox
1562 Season Street
Alliance, OH 44601
216/821-9110

Howard Feibus

Dir., Office of Coal Combustion
U.S. Department of Energy
F//-232, GTN
Washington, DC 20545
301/353-4348

W. D. Fellows

Senior Staff Engineer

Exxon Research & Engineering Co.

P.O. Box 101

Florham Park, NJ 07932

201/765-1250

P. G. Finlay

Head, Electric Power Section

Environment Canada

351 St. Joseph Blvd., 13th Floor

Ottawa, Ontario K1A 0H3

CANADA

819/953-1126

Dave Finnegan
Counsel , ,

Committee on Energy and Commerce
2125 Rayburn Bui Idl^^jc
Washington, DC 20515
202/225-2927

Kelly A. Fortin
Environmental Engineer
U.S. EPA -- Region 9
2921 Santos Lane, Apt, 2126
Walnut Creek, CA 94596
415/974-7043

Bill Fraser
Senior Vice President
TransAlta Utilities
P.O. Box 1900
Calgary, Alberta
CANADA T2P 2M1
403/267-7482

Steve Frey
N/A

U.S. EPA -- Region 9
215 Fremont Street
San Francisco, CA 94105
415/974-8071

Yuan Fu

Chemical Engineer
U.S. Department of Energy
P.O. Box 10940
Pittsburgh, PA 15236
412/892-4841

A-8

-------
Rajendra P. Gaikwad

Engineer

Ontario Hydro

800 Kipling Avenue

Toronto, Ontario

CANADA M5L 2Z6

416/231-4111

Michael Gamburg
Regional Sales Manager
Fuel Tech

750 17th Ave., Suite 102
San Francisco, CA 94121
415 221-6177

Gerald M. Gardetta
Environmental Affairs Admin.
Southern California Gas Company
810 S. Flower Street
Los Angeles, CA 90017
213/689-3365

James S. Geier
Public Health Engineer
Colorado Department of Health
4210 East 11th Avenue
Denver, CO 80220
303/331-8500

John Gerdes
Design Engineer
John Zink Company
P.O. Box 702220
Tulsa, 01 74170
918/592-4810

Marco Ghiringhelli
Liaison Engineer
Ansaldo -- Babcock & Vilcox
20 S. Van Boxen Avenue
Barberton, OR 44203
216/860-6029

Robert L« Gilbert
Program Manager
MK-Ferguson Company
One Erieview Plaza
Cleveland, OH 44114
216/523-6618

Alan F. Gillespie

Manager, Engineering Services

Foster Wheeler, Ltd.

P.O. Box 3007

St. Catharines, Ontario

L2R 7B7, CANADA

416/688-4434

Tom P. Gilmore

Product Line Manager

KT1 Corporation

1333 S. Mayflower Avenue

Monrovia, CA 91016-4099

818/303-4711

Dan V. Giovanni
Consultant

Electric Power Technologies, Inc.
P.O. Box 5560
Berkeley, Ca 94705
415 653-6422

Lisa Glatch
Process Engineer
Fluor Daniel

3333 Michelson Drive, MC-B1B
Irvine, CA 92730
714/975-3047

Frans Goudriaan

Senior Process Technologist

Kon./She11 Laboratory

Badhuisweg 3, Dept. KCP

1031 CM Amsterdam

HOLLAND

31/20 303957

Toby R. Gouker

Mgr., Stationary Emission Control
V. R. Grace & Company
7379 Route 32
Columbia, MD 21044
301/531-4131

John T. Graves

Environmental Superintendent
Minnkota Power Cooperative
P.O. Box 127
Center, ND 58530
701/794-8711

A-9

-------
Francois Grinard
N/A

Fuel Tech Europe, Ltd.
28a Cadogan Square
London, SWIX OJH
UNITED KINGDOM
01/581-2051

Michael Grimaberg
M. Sc.

University of Lund
Dept. Chem. Engineering
P.O. Box 124
S-221 00 Lund, SWEDEN
46/46 10 82 76

Charles V. Grinnell
V.P. and General Counsel
Fuel Tech

61 Taylor Reed Place
Stamford, CT 06906
203/359-1320

Robert R. Grinstead
Associate Scientist
Dow Chemical USA
2800 Mitchell Drive
Walnut Creek, CA 94598
415/944-2077

P. A. Gristwood

Head Combustion Technology .Group

Babcock Energy, Ltd.

165 Great Dover Street

London SE1 4YB

ENGLAND

01/407 8383 X2272
George Gunter

Corporate Engineering Specialist

New Brunswick Power

P.O. Box 2000

Fredericton, New Brunswick

CANADA E3B 4X1

506/458-4381

Vir V, Gupta
EPE IV
U.S. EPA

200 Churchill Road
P.O. Box 19276
Springfield, IL 62794-9276
217/782-2113

Jim Guthrie
Mechanical Engineer
California Air Resources Board
P.O. Box 2815
Sacramento, CA 95812
916/327-1508

Donald K. Hagar
President
Eagleair, Inc.

1150 Mauch Chunk Road
Bethlehem, PA 18018
215/868-1616

Leo E. Hakka
New Venture Manager
Union Carbide Canada, Ltd.
P.O. Box 700
Pointe-Aux-Trerables, Que
CANADA H1B 5KE
514/493-2617

Bob Hall

Chief, Combustion Research Branch
U.S. EPA
AEE1L — MD-65

Research Triangle Park, NC 27711
919/541-2477

David Ham

Vice President

PSI Technology Company

P.O. Box 3100

Andover, MA 01810

508/475-9030

Richard Haman
Principal Engineer
Detroit Edison
2000 Second Avenue
Detroit, MI 48226
313/897-0208

A-10

-------
Peter J. Handy

Engineering Sales Director

Rolls Soyce Canada

P.O. Box 544, Montreal AMF

Montreal, Quebec

CANADA H4Y 1B3

514/631-3541

Thomas Hansen

Chief, Mobile Source Section
U.S. EPA

345 Court land Street
Atlanta, GA 30365
404/347-2864

Stan Harding
Research Group Leader
Consolidation Coal Company
4000 Brownsville Road
Library, PA 15129
412/854-6735

Peter J. Hart

Supervisor, Emissions Comp.
Allison Gas Turbine Div. of GM
P.O. Box 420, SC T-01
Indianapolis, IN 46206
317 230-4186

Scott Hassett

Principal Engineer

Utah Power & Light Company

168 North, 1950 West

Salt Lake City, UT 84140

801/220-4839

Greg Haussoasn

Research Assistant

Stanford School of Engineering

Blgd. 520, Dept. of Kech. Engrng.

Stanford, CA 94305

415/723-1823

Edward Healy

Project Planning Engineer
Southern Company Services
P.O. Box 2625
Birmingham, AL 35202
205/868-5212

Dietmar He in
Professor

Siemens AG, KWU Group
Hammerbacherstrasse 12+14
D-8520 Erlangen
FEDERAL REPUBLIC OF GERMANY
9131/182350

Klaus Hein

Prof. Dr-Ing.

RWI AG, BV Fortuna

Postfach 1461

5010 Bergheim 4

FEDERAL REPUBLIC OF GERMANY

02271/584-2234

Todd Hellewell
Principal Engineer
Combustion Engineering
1000 Prospect Hill Road
Windsor, CT 06095
203/285-4919

Eric Hennen
Environmental Engineer
Dairyland Power Company
2615 East Avenue, So.
Lacrosse, WI 54601
608/787-1371

Bo Herrlander
Business Area Manager
Flakt Industres AB
S-35187 Vaxjo
SWEDEN

01146/470-87400

Steven A. Hickerson
President

Eacotek Corporation
8220 Doe Avenue
Visalia, CA 93291
209/651-2000

David Hickman
Senior Scientist
Corning Glass Works
SP-FR-6-1

Corning, NY 14831
607/974-3715

A-n

-------
M. B, Hilt

Mgr., Externally Funded Programs
General Electric
1 River Road
Schenectady, NY 1234S
518/385-9191

Anna Hinderson

Engineer

ABB Carbon

Box S61220

SWEDEN

N/A

Anna-Karin E. Hjalraarsson

Environment Group

IEA Coal Research

14-15 Lower Grosvenor Place

London SW1W OEX ENGLAND

01/828-4661

Lennart Hjalmarsson

Engineer

ABB Carbon

Box S61220

SWEDEN

N/A

Patrick A. Ho
Principal Engineer
Tampa Electric Company
702 N. Franklin Street
Tampa, FL 33602
813/228-4844

John E. Hofmann

Vice President, Development

Fuel Tech

61 Taylor Reed Place
Stamford, CT 06906
203/359-1320"

Bill Holt

Manager, Business Development

Allied Signal

P.O. Box 580970

Tulsa, OK 74158-0970

918/266-1400

Neville Holt

Senior Program Manager

Electric Power Research Institute

3412 Hillview Avenue

Palo Alto, Ca 94034

415/855-2503

Richard G. Hooper
Vice President
NaTec, Ltd.

216 Sixteenth Street Mall
Suite 1500

Denver, CO 80202-5129
303/820-5200

Hann S. Huang
Chemical Engineer
Argonne National Laboratory
9700 S. Cass Avenue
Argonne, IL 60439
312/972-3125

Allen Hubbard

Environmental Engineer

Wis. Dept. of Natural Resources

P.O. Box 7921

Madison, WI 53707

608/266-3450

Terry Hunt

Mechanical Engineer

Public Service Co. of Colorado

5900 East 39th Avenue

Denver, CO 80207

303/329-1113

Adrian Hyde
Engineer

UK Department of Energy
Thames House South, Millbank
London, SW11 4QT
UNITED KINGDOM
01/211-4547

Ivan V. Insua
Staff Engineer
Salt River Project
P.O. Box 52025
Phoenix, kZ 85072-2025
608/236-5240

A-1 2

-------
Paul Ireland

Project Engineering Manager
United Engineers & Constructors
P.O. Sox 5888
Denver, CO 80217
303/692-3420

Reda S. Iskandar
Project Manager
Corning Glass Works
MP-WX-01-5
Corning, NY 14831
607/974-7306

Sven Jaras
Engineer
Eka Novel AB
S-44S 01 Surte
SWEDEN

+46/303 98377

Jim Jarvis
Senior Engineer
Radian Corporation
8501 Mo-Pac
Austin, TX 78720
512/454-4797

Torgny Johansson
Manager Process Technology
Flakt Industres AB
S-35187 Vaxso
SWEDEN

01146 470-8700 0
David L. Johnson

Mgr., Production, Engrng, & Const.
Otter Tail Power Coopany
215 S. Cascade Street
Fergus Falls, MH 56537
218/739-8399

Keith Johnson
N/A

Foster Wheeler Power Products

0l^^upic Office Center

No. 8 Fulton Road

Wembly, Middlesex HA9 OJH, UK

01/900-2533

Kevin L. Johnson

Project Manager

Radian Corporation

3200 E. Chapel Hill Road

Research Triangle Park, NC 27709

919/541-9100

Steve Johnson

Mgr., Applied Combustion Tech.
PSI Technology Company
P.O. Box 3100
Andover, MA 01810
508/475-9030

Wayne L. Johnson

Sr. Process Engineering Specialist

ESSO Resources Canada, Ltd.

237 Fourth Avenue, S.W.

Calgary, Alberta

CANADA T2P 0N6

403/237-2721

Dale G. Jones
Vice President
Emcotek Corporation
8220 Doe Avenue
Visalia, CA 92391
209/651-2000

Gary D. Jones
Program Manager
Radian Corporation
P.O. Box 13000

Research Triangle Park, NC 27709
919 541-9100

Thomas E. Kaforey
Generating Plant Engineer
Ohio Edison Coopany
76 South Main Street
Akron, OH 44308
216/384-5974

Marie Kalinowski
Advanced Scientist
Owens Corning Fiberglas
Route 16

Granville, OH 43023
614/587-7620

A-13

-------
Dale A. Kanary
Engineer

Ohio Edison Company
76 South Main Street
Akron, OH 44308
216/384-5744

Hans T. KarIsson

Associate Professor

University of Lund

Dept. Cheto. Engineering Box 124

S-221 00 Lund, SWEDEN

46 46 10 82 44

Ari Karppinen

Research Scientist

Finnish Meteorological Institute

Sahaajankatu 22E

SF 00810 Helsinki

FINLAND

358/0 7581345

Robert F. Kaupp

Marketing Manager, Projects

Johnson Matthey

436 Devon Park Drive

Wayne, PA 19087

215/341-8502

Martin Kay

Supervising Air Quality Engineer
So. Coast Air Quality Mgmnt. Dist.
9150 Flair Drive
El Monte, CA 91731
818/572-6258

Robert T. Keen
Principal Engineer
Flchtner USA, Inc.

Overlook 1, Suits 360
2849 Paces Ferry Road, N.W.
Atlanta, SA 30339
404/432-6983

Robert L. Keller
Corporate Engineer
Dayton Power & Light Company
9200 Chautauqua Road
Miamisburgh, OH 45342
513/865-6234

Stephen E. Kerho
Consultant

Electric Power Technologies, Inc.
25108 Marguerite Pkwy., Ste 8239
Mission Viejo, CA 92692
714 380-7316

Fred Kessler

Project Manager

KTI Corporation

1333 S. Mayflower Avenue

Monrovia, CA 91016

818/303-4711

Lawrence P. King

Senior Marketing Specialist

Babcock & Wilcox

1562 Beeson Street

Alliance, OH 44601

216/829-7576

Masaaki Kinoshita
Assistant Manager
Mitsubishi Heavy Industries
1-1, Akunoura-Machi
Nagasaki, JAPAN
0958/28-6431

Larry H. Kirby

Assoc. Development Scientist
Dow Chemical USA

Texas Operations, B-1605 Building
Freeport, TX 77541
409/238-7721

Thomas Kirkey
Account Manager
Norton Company
2500 E. Ball load
Anaheim, CA 92806
714/635-4051

John B. Kitto, Jr.

Program Manager

Babcock & Wilcox

Research it Development Division

1562 Beeson Street

Alliance, OH 44601

216/829-7710

A-14

-------
J. S. Klingspor
Senior Engineer
Combustion Engineering
31 Inverness Center Parkway
Birmingham, AL 35243
205/991-2832

Jeff Klueger

Associate Mechanical Engineer
Los Angeles Dept. of Water & Power
111 N. Hope Street, Room 661
Los Angeles, CA 90051
213/481-4647

Kevin Knill

Research Engineer

Int'l Flame Research Foundation

P.O. Box 10000

1970 CA Ijmuiden

THE NETHERLANDS

31/2510-9364

Cathy L. Knowles
Fibers Department

E.I. du Pont de Nemours & Co., Inc.

Laurel Run Building

P.O. Box 80, 705

Wilmington, DE 19880-0705

302/999-3905

Kris W. Knudsen
Supervising Scientist
Duke Power Company
P.O. Box 33189
Charlotte, NC 28242
704/373-5104

Bernie Koch

Director, Project Development
Consolidation Coal Company
4000 Brownsville Road

Library, PA 15129
412/854-6612

A. Kok

Engineer

N0VEM

P.O. Box 17
6130 AA Sittard
THE NETHERLANDS
4490/95308

Angelos Kokkinos
Project Manager

Electric Power Research Institute
3412 Hillview Avenue
Palo Alto, CA 94304
415/855-2494

Johanna Koppius-Odink

Engineer

EP0N

P.O. Box 10087
8000 GB Zwolle
THE NETHERLANDS
038/272900

Richard S. Kostkas
Dir., Special Marketing Projects
The Peoples Natural Gas Company
625 Liberty Avenue
Pittsburgh, PA 15222
412/553-6608

Toshio Koyanagi
SCR Engineer

Mitsubishi Heavy Industry
2 Houston Center
Houston, TX 77010
713/652-9269

Donald E. Krider

Mgr., Thermal De-NOx Sales

Exxon Research & Engineering Co.

P.O. Box 390

Florhan Park, NJ 07932

201/765-2339

Hal M. ICruger
Engineer

Centerior Energy
P.O. Box 94661
Cleveland, OH 44101-4661
216/447-2233

H. Kulper

Manager, Fuel & Burner Department
Royal Schelde
P.O. Box 16, 4380AA
Vlissixigen, NETHERLANDS
311184 82583

A-15

-------
Robin L. Rump

Environmental Consultant

E. I. DuPont

P.O. Box 1089

Orange, TX 77631-1089

409/886-6239

Alfred Kurzenhauser
Technical Support Supervisor
LFC Power Systems Corporation
Building One, Suite 255
4000 Kruse Way Place
Lake Oswego, OR 97035
S03/697-0259

H. K. Kwee
Engineer

Stork Boilers B.V.

P.O. Box 20
7550 GB Hengelo
THE NETHERLANDS
31/74 401857

D. C. Langley

Regional Service Manager

Babcock & Wilcox

7401 W. Mansfield Avenue, #410

Lakewood, CO 80235

303/988-8203

Robert Langthorne
Manager

Canadian Energy Services, Ltd.
401 Salter Street
New Westminster, B.C.

CANADA V3M 5Y1
604/521-3322

Dorothy Lau
Design Specialist
Ontario Hydro

700 University Avenue, H10 F2
Toronto, Ontario MSG 1X6
CANADA
416/592-7306

Dennis Laudal ^

Research Engineer

Energy & Mineral Research Center

15 North 23rd Street

P.O. Box 8213

Grand Forks, ND 58201-8213
701/777-5138

Nick J. Lavxngia

Senior Project Engineer

Chevron U.S.A.

324 W. El Segundo Boulevard

El Segundo, CA 90245

213/615-5784

Peter Lawson
Supervising Engineer
Ontario Hydro
700 University Avenue
Toronto, Ontario
CANADA MSG 1X6
416/592-5393

Albert D. LaRue
Principal Engineer
Babcock & Wilcox
20 S. Van Buren Avenue
Barberton, OH
216/860-1493

Douglas J. Leadenham

Analyst

Electric Power Research Institute
3412 Hillview Avenue
Palo Alto, CA 94304
415/855-2954

G. K. Lee

Manager, Combustion Research
Energy Resources Canada
555 Booth Street
Ottawa, Ontario K2G 0N3
613/996-3179

George C. Lee
Engineering Specialist
Bechtel National, Inc.

50 Beale Street
Sao Francisco, CA 94105
415/768-1216

A-16

-------
Yam Y. Lee
Research Specialist
Pyropower Corporation
5120 Shoreham Place
San Diego, CA 92122
619/552-2312

George R. Lester

Senior Research Scientist

Allied Signal

P.O. Box 5016

Des Plaines, IL 60017

312/391-3314

Robert D. Lewis
Principal Engineer
Combustion Engineering
1000 Prospect Hill Road
Windsor, CT 06095
203/285-5968

F. Y. Liao

Senior Associate Engineer
Mobil R&D
P.O. Box 1026
Princeton, NJ 08540
609/737-4955

Vladimir Lifshits
N/A

Coen Company
1510 Rollins Road
Burlingame, CA 94010
415/697-0440

Sergio Ligasacchi

N/A

ENEL

Via A. Pisano 120

Pisa, ITALY 56100
050/535622

JoAnn S, Lighty
Assistant Professor
University of Utah
2254 MEB

Salt Lake City, UT 84112
801/581-5763

William P. Linak
Project Officer
U.S. EPA
AEERL -- MD-65

Research Triangle Park, NC 27711
919/541-5792

David G. Linz

Mgr., Land and Water Quality Res.
Gas Research Institute
8600 West Bryn Mawr Avenue
Chicago, IL 60631
312/399-8198

Robert A. Lisauskas

Director, Research & Development

Riley Stoker Corporation

Riley Research Center

45 McKeon Road

Worcester, MA 01610

508/792-4801

Steve Londerville

N/A

Coen Company
1510 Rollins Road
Burlingame, CA 94010
415/697-0440

Phillip A. Lowe
President
Intech, Inc.

11316 Raven Drive
Potomac, MD 20054
301/983-9367

Francis C. Luck

Research Engineer

Rhone-Poulenc France

14 Rue des Gardinoux

93308 Aubervilliers, FRANCE

33/147-396-361

Richard K. Lyon

Senior Scientist

Energy & Environmental Research

P.O. Box 189

Whitehouse, NJ 08888

201/534-5833

A-17

-------
0. L. Maaskant

N/A

Shell International Chemical Co,

P.O. Box 162

2501 AN The Hague

HOLLAND

070/773741

Denis Maftei

Sr. Process Engineer

Environment Ontario

880 Bay Street, 4th Floor

Toronto, Ontario

CANADA M5S 1Z8

416/965-3776

Mehdi Maibodi
Engineer

Radian Corporation
P.O. Box 13000

Research Triangle Park, NC 27709
919/541-9100

Jason Hakans i

Senior Technical Editor

Power Magazine

11 W. 19th Street, 2nd Floor

New York, NY 10011

212/337-4074

James W. Halone
Sales Engineer
Babcock & Wilcox
3734 Gundry Avenue
Long leach, CA 90807
213/595-8357

Aaron Maim.

N/A

Sierra Pacific Power Company
P.O. Box 10100
Reno, NV 89520
702/689-4011

W. N. Mansour
President

Applied Utility Systems
1140 E. Chestnut Avenue
Santa Ana, CA 92701
714/953-9922

Gerald J. Maringo
Development Engineer
Babcock & Wilcox
20 S. Van Buren Avenue
Barberton, OH 44203
216/860-6321

John L. Marion
Consulting Engineer
Combustion Engineering
1000 Prospect Hill Road
Windsor, CT 06095
203/285-4539

Bob Markoja

Manager, Environmental Services
Sierra Environmental Engineering
3505 Cadillac Avenue, N-3
Costa Mesa, CA 92626
714/432-0330

Gregory A. Marts1
Senior Engineer
Northeast Utilities
1866 River Road
Middletown, CT 06457
203/638-3175

Blair Martin
Deputy Director, AEERL
U.S. EPA
AEERL -- MD-60

Research Triangle Park, NC 27711
919/541-7504

Gianni Mascalze
Doctor

Howard Mason

Mgr., Energy Engineering Department

Acurex Corporation

485 Clyde Avenue

Mountain View, CA 94039

415/961-5700

A-l 8

-------
M, Massoudi

Project Manager

Ebasco Services, Inc.

One Parket Plaza, Spear Bldg.

Suite 600

San Francisco, CA 94105
415/777-3000

Mike Maxwell

Chief, Technology Dev. Branch
U.S. EPA
AIERL -- MD-04

Research Triangle Park, NC 27711
919/541-3091

Phil May
Project Director
Radian Corporation
P.O. Box 13000

Research Triangle Park, NC 27709
919/541-9100

Michael McCartney

Mgr., Firing Systems Engineering

Combustion Engineering

1000 Prospect Hill Road

Windsor, CT 06095

203/285-4677

Michael V. McElroy
Consultant

Electric Power Technologies, Inc.
P.O. Box 307
Menlo Park, CA 94026
415 322-0843

Marilyn McIIvane
Managing Editor
Mcllvane Coopany
2970 Maria Avenue

Northbrook, IL 60062
312/272-0010

Robert McIlvalue
President
Mcllvane Company
2970 Maria Avenue

Northbrook, IL 60062
312/272-0010

William M. McKinney

V.P., New Business Development
United Catalysts, Inc.

P.O. Box 32370
Louisville, KY 40232
502/634-7218

Ron Melton
Section Manager
Battelle Northwest
P.O. Box 999
Richland, WA 99352
509/375-2932

Jim Merrian

Sr. Environmental Licensing Engr.
Jersey Central Power & Light Co.
310 Madison Avenue
Morristown, NJ 07960
201/455-8563

Jean Mevel

Test Manager, Research & Dev.
Stein Industries
1921 Morane Saulnler Avenue
Velizy, Vellacoublay
FRANCE

1/34 65 44 48

Janes E. Meyers

Chemical Equipment Engineer

Detroit Edison

2000 Second Avenue

Detroit, MI 48226

313/897-0806

Paul G. Mikolaj
Consulting Engineer
Tosco Corporation
Avon Refinery
Martinez, CA 94553
415/228-1880

Thomas Mileus
Sales Manager
Flakt Industres AB
S-351 87 Vaxjo

SWEDEN

01146 470-87000

A-19

-------
Scott A. Miller
Staff Engineer

Dayton Power & Light Company
P.O. Box 468
Aberdeen, OH 45101
513/549-2641 X5127

Kenneth J. Mills
Market Manager
Norton Company
P.O. Box 350
Akron, OH 44309
216/673-5860

Ed Milone
Engineer

Con. Edison Co. of New York
4 Irving Place
New York, NY 10003
212/460-4887

Shigehiro Miyamae
Senior Engineer
Ishikawajima-Harima Heavy Ind.
2-16, Toyosu 3-chome, Koto-ku
Tokyo 135 JAPAN
03/534-4335

Cal Mock

District Manager

Babcock & Wilcox

1990 N. California Boulevard

Walnut Creek, CA 94596

415/947-1100

Hukam C. Mongia
Chief, Combustors R&D
Allison Gas Turbine Div. of GM
P.O. Box 420, Speed Code T14
Indianapolis» IN 46206-0420
317 230-5945

Curtis A. Moore
Counsel
U.S. Senate
Washington, DC 20510
202/224-5762

Dave Moore
Supervisor

Dayton Power & Light Company
Killen Station -- P.O. Box 147
Manchester, OH 45144
513/549-3911

K. A. Moore

N/A

Dykeraa Engineering
5850 Canoga Avenue, Suite 400
Woodland Hills, CA 91367
818/712-0070

Ton Morasky

Tech. Transfer Manager

Electric Power Research Institute

3412 Hillview Avenue

Palo Alto, CA 94034

415/855-2468

Akira Mori
Senior Engineer
Hitachi, Ltd.
c/o Hitachi America. Ltd.

950 Elm Avenue, Suite 195
San Bruno, CA 94066-3094
415/872-1902 X8131

Shigeki Morita

Senior Engineer, Kure Works

Babcock-Hitachi K. K.

c/o Hitachi America, Ltd.

950 Elm Avenue, Suite 195
San Bruno, CA 94066-3094
415/872-1902 X8131

Karen D. Morris
Mechanical Engineer
Pacific Gas & Electric
77 Beale Street, Room 2547
San Francisco, CA 94106
415 972-3384

J. K. Mueller
Project Manager
Unocal Corporation
1201 West Sth Street
Los Angeles, CA 90051
213/977-6419

A-20

-------
T. Murataka
Senior Engineer
Babcock-Hitachi K. K.
c/o Hitachi America, Ltd.

950 Elm Avenue, Suite 195
San Bruno, CA 94066-3094
415/872-1902

Hitoshi Murayama
Project Manager

Electric Power Research Institute
3412 Hillview Avenue
Palo Alto, CA 94304
415/855-2777

Paul Musser

Program Manager

U.S. Department of Energy

Germantown, MD 20874

301/353-2846

Lawrence J. Muzio
Vice President

Fossil Energy Research Corporation
23342 C South Points
Laguna Hills, CA 92653
714/859-4466

Peter Necker

Prokurist

Neckarwerke AG

Kuferstrasse 2

7300 Esslingen am Neckar

FEDERAL REPUBLIC OF GERMANY

0711/3190-3001

Stefan Negrea
N/A

KRC/Noell GnJbH

Alfred, Nobel Strasse 20

D-8700 tfurgburg

WEST GERMANY

N/A

Mike Nelson
Senior Engineer
Southern Company Services
P.O. Box 2625
Birmingham, AL 35202
205/870-6518

Kenneth M. Nichols

Asst. Prof., Engineering & Research

The Institute of Paper Chemistry

P.O. Box 1039

Appleton, VI 54912

414/738-3385

Per R. Nielsen
Senior Manager

Nordic Gas Technology Center

Dr. Neergaardsvej 5A

DK 2970 Hoersholia

DENMARK

45/276 6995

Ralph M. Nigro
Project Engineer
Delmarva Power
800 King Street
Wilmington, DE 19899
302/454-4872

Michael Novak
Engineer

Verbundkraft Elek.

AM HOF 6 A
Vienna A-1011 Wein
AUSTRIA

222/531 13 2590

James H. Nylander

Senior Engineer

San Diego Gas & Electric Co.

P.O. Box 1831

San Diego, CA 92112

619/931-7294

George Offen
Program Manager

Electric Power Research Institute
3412 Hillview Avenue
Palo Alto, CA 94304
415/855-8942

Yoshiya Ogasawara

Asst. Manager, Int'l Sales Div.

Hitachi, Ltd.

c/o Hitachi America, Ltd.

950 Elm Avenue, Suite 195

San Bruno, CA 94066-3094

415/872-1902 X8131

A—21

-------
Delbert M. Ottmers, Jr.
Vice President
Radian Corporation
P.O. Box 201088
Austin, TX 78720-1088
512/454-4797

Joseph H. Oxley
Program Manager
Battelle
505 King Avenue
Columbus, OH 43201-2693
614/424-7885

Y. S. Pan

Project Coordinator
U.S. Department of Energy
P.O. Box 10940
Pittsburgh, PA 15236
412/892-5727

Andy S. Paranj ape
Senior Project Engineer
Kerr-McGee Corporation
P.O. Box 25861
Oklahoma City, OK 73125
405/270-2906

Rasik Patel

Manager, Thermal Design

KTI Corporation

1333 S. Mayflower Avenue

Monrovia, CA 91016-4099

818/303-4711

Jarl Pedersen
Mgr., Combustion Systems
Burmeister & Vain Energy
23 Teknikerbyeo
Virum, DENMARK, 2830
45/285-2100 "

Craig Penterson
Staff Engineer, R&D
Riley Stoker Corporation
5 Neponset Street
Worcester, MA 01590
508/792-4829

Ralph Perhac

Department Director

Electric Power Research Institute

3412 Hillview Avenue

Palo Alto, CA 94304

415/855-2572

Michael Perlsweig

Flue Gas Cleanup Program Manager

U.S. Department of Energy

Fossil Energy, FI-23

Washington, DC 20545

301/353-4399

David W. Pershing
Dean, College of Engineering
University of Utah
2202 MEB

Salt Lake City, UT 84112
801/581-5057

Robert D, Petersen
Senior Mechanical Engineer
Burns & McDonnell
4800 East 63rd Street
Kansas City, MO 64141
816/333-4375

Ellen Petrill
Project Manager

Electric Power Research Institute
3412 Hillviav Avenue
Palo Alto, CA 94304
415/855-8989

James D. Phillip

Plant Chemist

Tenneco Minerals Company

P.O. Box 1167

Green River, WY 82935

307/872-6519

Larry A. Pierson
Project Manager
Babcock & Wilcox
20 S. Van Buren Avenue
Barberton, OH 44203
216/860-1103

A-22

-------
Everett L. Plyler
Dir., Combustion & Indoor Air Div.
U.S. EPA
AEERL -- MD-54

Research Triangle Park, NC 27711
919/541-2918

Gary Poe
Project Manager

Electric Power Research Institute
3412 Hi11view Avenue
Palo Alto, CA 94304
415/855-8969

John H. Pohl

Dir., Research & Development
Energy Systems Associates
15991 Red Hill
Tustin, CA 92680
714/259-9520

David Price
Project Engineer
Unocal Corporation
P.O. Box 758
Wilmington, CA 90748
213/513-7600

W. L. Prins

Manager, Process Technology
Tests Engineers & Consultants
P.O. Box 10.000
1970 CA Ijmuiden
THE NETHERLANDS
+312510 98603/92209

Walter Pullle

Technology Transfer Manager
Electric Pow«r Research Institute
3412 Hi11vie* Avenue
Palo Alto, CA 94304
415/855-2470

Suresh K. Punatar
Engineering Supervisor
Bechtel Power Corporation
50 Beale Street
San Francisco, CA 94119
415/768-0495

Brian Quil

Mechanical Engineer

Naval Energy & Environmental Suppt.

NEESA 11A

Port Hueneme, CA 93043
805/982-4748

Les Radak

Senior Research Engineer
Southern California Edison
2244 Walnut Grove Avenue
Rosemead, CA 91770
818/302-9746

Paul Radcliffe
Project Manager

Electric Power Research Institute
3412 Hillview Avenue
Palo Alto, CA 94304
415/855-2720

James M. Raines
Sr. Mechanical Engineer
Pacific Gas & Electric
77 Beale Street, Room 2553
San Francisco, CA 94106
415 972-2735

Jay Ratafia-Brown

Dir., Fossil Energy/Env. Projects

SAIC

1710 Goodridge Drive
P.O. Box 1303
McLean, VA 22102
703/448-6343

Janes L. Reese
Consulting Engineer
Applied Utility Systems
1140 E. Chestnut Avenue
Santa Ana, CA 92701
714/953-9922

Joseph N. Reeves
Research Manager
Southern California Edison
2244 Walnut Grove Avenue
Rosemead, CA 91770
818/302-9515

A-23

-------
Amir lehmat

Mgr., Utility Gas Development
Institute of Gas Technology
3424 South State Street
Chicago, IL 60616-3896
312/567-5899

Ernst Ringle

Delegate

Noell

199 Wells Avenue
Newton, MA 02159
617/332-6510

Alan P, Risch

Manager, Technical Marketing
UOP

39 Old Ridgebury Road
Danbury, CT 06817
208/794-5542

R. C. Rittenhouse
Senior Editor

Power Engineering Magazine
1250 S. Grove Avenue, Suite 302
Barrington, IL 60010
312/382-2450

Roger Robb
Senior Engineer

Northern Indiana Public Service Co.
5265 Hohman Avenue
Hammond, IN 46320
219/853-5821

Dick Robertson

Manager, Air Quality

TO Electric

400 North Olive, LB-81

Dallas, TX 75201

214/812-841#

Chris P. Robie

Process Engineer

United Engineers & Constructors

700 South Ash Street

P.O. Box 5888

Denver, CO 80217

303/692-2803

Tracy M. Robinson
Marketing Programs Manager
Fuel Tech

61 Taylor Reed Place
Stamford, CT 06906
203/359-1320

William Rogers
Sales Engineer
Engelhard Corporation
2000 Powell Street
Emeryville, CA 94608
415/596-1703

Bill Rovesti

Senior Project Manager

Electric Power Research Institute

3412 Hillview Avenue

Palo Alto, CA 94304

415/855-2519

Charlie Rowden

Lead Process Engineer

Fluor Daniel

3333 Michelson Drive

Irvine, CA 92730

714/975-3047

Richard E. Rozelle

Project Manager

Dow Chemical USA

Texas Operations, B-1605 Bldg.

Freeport, TX 77541

409/238-9846

Hans-Georg Rych

Electric Power Systems Division
EVN

Johana-Steinbock-Strasse 1

A-2344 Maria Enzersdorf am Gebirge

AUSTRIA

02236/83 611 526

N. C. Samish

Staff Research Engineer

Shell Development Company

P.O. Box 1380

Houston, TX 77251

713/493-7944

A-24

-------
Henry Sanematsu

Mechanical Engineer

Los Angeles Dept. of Water & Power

111 N. Hope Street

Los Angeles, CA 90012

213/481-4906

Charles E. Schlittler
Program Manager
Kerr-McGee Corporation
P.O. Box 25861
Oklahoma City, OK 7312S
405/270-3188

Herbert Schuster
Head, Combustion R&D Dept.

Deutsche Babcock Werke AG
Duisburger Str. 375
D-4200 Oberhausen 1
FEDERAL REPUBLIC OF GERMANY
0208/833-2853

Charlie Sedman
Project Officer
U.S. EPA
AEERL -- MD-62

Research Triangle Park, NC 27711
919/541-7700

Noriyuki Shimizu

Assistant Manager

Electric Power Development Co.

15-1, Ginza, 6-chome, Chuo-ku

Tokyo, JAPAN 104

3/546-9403

Gary H. Shioooto
Engineer

Fossil Inersr Research Corporation
23342 C South Pointa
Laguna Hills, CA 92653
714/859-4466

Dale E. Shore

Manager, Western legion

Energy Systems Associates

15991 Red Hill Avenue, Suite 110

Tustin, CA 92680

714/259-9520

Stefan P. Shoup
Senior Staff Engineer
Inland Steel Company
3210 Wat ling Street
East Chicago, IL 46312
219/399-5733

George Shulof
President & CEO
Fuel Tech

61 Taylor Reed Place
Stamford, CT 06906
203/359-1320

Sandy Simko
Project Manager
Radian Corporation
P.O. Box 201088
Austin, TX 78720-1088
512/454-4797 X5301

Robert E. Sistek
Environmental Met. Engineer
LTV Steel Company
3100 E. 45th Street
Cleveland, OH 44127
216/429-6477

Lowell L. Smith
Vice President
ETEC

2091 Business Center Drive #100
Irvine, CA 92715
714/833-2523

Richard C. Smith
Supervising Engineer
Union Electric Company
P.O. Box 149, MC-450
St. Louis, M0 63166
314/554-3529

Douglas Snoot

Dean, College of Engrng. h Tech.

Brigham Young University

270 Clyde Building

Provo, UT 84602

801/378-4327

A-25

-------
Gerry C. Snow
Senior Engineer
Radian Corporation
P.O. Box 13000

Research Triangle Park, NC 27709
919/541-9100

David W. South
Economist

Argonne National Laboratory
370 L1Enfant Promenade, S.W.

Suite 702

Washington, DC 20024
202/586-4758

Ruggero Spagliarisi

Engineer

Aeru -- Milano

Via S. Sisto

8-20123

Milano, ITALY

02/808966

Barry Speronello
Principal Scientist
Engelhard Corporation
Menlo Park
Edison, NJ 08818
201/321-5155

Dan Spisak

Project Engineer

Bechtel National, Inc.

50 Beale Street

San Francisco, CA 94105

415/768-1645

Richard T. Squires
Research Officer
CEGB

Marchwood Lab*

Marchwood, Southampton

UNITED KINGDOM
0703/663232

Ravi K. Srivastava
Project Engineer
Acurex Corporation
P.O. Box 13109

Research Triangle Park, NC 27709
919/541-2692

Susan J. Stamey
Mechanical Engineer
Tennessee Valley Authority
MR3S92B

Chattanooga, TN 37402
615/751-5324

Mark T. Staniulis
Senior Staff Chemist
UOP

Old Sawmill River Road
Tarrytown, NY 10591
914/789-3865

David L. Stavenger
N/A

Dow Chemical USA
2020 Willard H. Dow Center
Midland, MI 48674
517/636-5236

Harry Stecyk

Sr. Engineering Specialist
ESS0 Resources Canada, Ltd.
237 Fourth Avenue, S.W,

Calgary, Alberta
CANADA T2P 0H6
403/237-2721

Harvey Stenger

Professor

Lehigh University

Building A, MTC

Dept. of Chemical Engrng.

Bethlehem, PA 18015

215/758-4791

Donald J. Stepson

Marketing Director

Rolls-Royce, Inc.

16800 Greenspoint Park Drive

Suite 155S

Houston, TX 77060

713/875-2100

Dick D. Stern

Chief, Technology Applications Br.
U.S. EPA
AEERL — MD-63

Research Triangle Park, NC 27711
919/541-2973

A-26

-------
Jim E. Stresewski

Engineer

Black & Veatch

1500 Meadow Lake Parkway

Kansas City, MO 64114

913/339-2000

Y. P. Su

Principal Environmental Engineer
Stone & Webster Engineering Corp.
330 Barker Cypress
Houston, TX 77094
713/492-4334

Keith Summons

Senior Design Specialist

Ontario Hydro

700 University Avenue, H13 C10
Toronto, Ontario
CANADA MSG 1X6
416/592-5935

Kohei Suyama
Manager

Mitsubishi Heavy Ind. America
c/o Combustion Engineering
1000 Prospect Hill Road
Windsor, CT 06095
203/285-2491

Markus Tahtinen

Research Scientist

Technical Research Ctr. of Finland

P.O. Box 169

SF-00181 Helsinki

FINLAND

358/0 648931

Katharine K. Tait

Engineer

M0NENC0

400 Monenco Place, 801-6th Ave. SW

Calgary/Alta

CANADA T3C 1N1

403/298-4161

Yojiro Takahashi
Manager

Nippon Shokubai America, Inc.
101 East 52nd Street, 14th Floor
New York, NY 10022
212/759-7890

Harry S. Tang

Senior Research Engineer

Shell Development Company

P.O. Box 1380

Houston, TX 77251-1380

713/493-8424

Donald P. Teixeira
Technical Manager Fossil R&D
Pacific Gas & Electric
3400 Crow Canyon Road
San Ramon, CA 94583
415/866-5531

Robert Telesz
Marketing Manager
Babcock & Wilcox
20 S. Van Buren Avenue
Barberton, OH 44230
216/860-1103

Preston E. Tempero
Plant Manager
KPL Gas Service
Lawrence Energy Center
P.O. Box 249
Lawrence, KS 66044
913/843-8118

Jean Thomas Rose
Engineering Manager
Betz Industrial
Sooerton Road
Trevose, PA 19047
215/355-3300 X2365

T. R. Thomas
N/A

Westinghouse Idaho Nuclear Company
P.O. Box 4000
Idaho Falls, ID 83403
208/526-3086

A-27

-------
Bruce Thompson

District Manager

Nalco Chemical Company

4310 North 75th Street, Suite A

Scottsdale, AZ 85251

602/996-2238

Paul F. Thompson
President

Tenerax Corporation
303 Laurel

Friendswood, TX 77546
713/482-5801

Richard E. Thompson
President

Fossil Energy Research Corporation
23342 C South Fointe
Laguna Hills, CA 92653
714/859-4466

Richard S. Topor

Int'l Regional Manager, Licensing

Combustion Engineering

1000 Prospect Hill Road, 7110-1923

Windsor, CT 06095

203/285-9934

Majed A, Toqan
Program Manager, CRF
Massachusetts Inst, of Technology
60 Vassar Street, Bldg. 31-261
Cambridge, MA 02139
617/253-1721

Paul J. Torpey

Executive Director

Empire State Elec. Energy Res. Corp

1155 Avenue of the Americas

New York, NT 10036

212/302-1212

Ian Torrens

Department Director

Electric Power Research Institute

3412 Hillview Avenue

Palo Alto, CA 94304

415/855-2422

Sanh V. Tram
Senior Service Engineer
Babcock & Wilcox
12552 Tours Avenue
Garden Grove, CA
714/554-3844

Jay R, Turner

Research Assistant

Washington University

Chemical Engineering Department

Campus Box 1198, One Brookings Dr.

St. Louis, MO 63130

314-889-6042

James M. Valentine

V.P. Marketing & Business Dev.

Fuel Tech

61 Taylor Reed Place
Stamford, CT 06906
203/359-1320

Don Van Buren

Field Engineering

Bay Area Air Quality Mgrant. Dist.

939 Ellis Street

San Francisco, CA 94109

415/771-6000 X165

Simon Vanderheijden
Chief Chemist
Saskatchewan Power
2025 Victoria Avenue
Segina, Saskatchewan
CANADA S4P 0S1
306/566-3073

Galliano Varlali
Professor

University of Rome, Italy
Via ludossiana 18
Rome, ITALY 00184
396/4687321

Joel Vatsky

Dir., Combustion & Env. Systems
Foster Wheeler Energy Corporation
Perryville Corporate Park
Clinton, NJ 08809-4000
201/730-5450

A-28

-------
Gary R. Veerkamp
Sr. Mechanical Engineer
Pacific Gas & Electric
77 Beale Street, Room 2555
San Francisco, CA 94106
415/972-1576

Jan L. Vernon

Head, Environment Group

IEA Coal Research

14-15 Lower Grosvenor Place

London SW1W OEX ENGLAND

01/828-4661

John Verurink

Senior Combustion Engineer

Stork Boilers B.V.

P.O. Box 20

7550 BD, Hengelo

THE NETHERLANDS

074/401370

Denise M. Viola
Sales Manager
Engelhard Corporation
Menlo Park CN28
Edison, NJ 08818
201/321-5039

J. Von Bergmann
General Manager
Fuel Tech GmbH
Frankfurter Strasse 33-35
Eschbom 6236

FEDERAL REPUBLIC OF GERMANY
49/6196-48613

Paul Wagner
Project Engineer
Delmarva Power
1-95 & Route 273
P.O. Box 9239
Newark, DE 19714
302/454-4844

Ali A. Wali

Planning Analyst

Wisconsin Public Service Comm.

4802 Sheboygan Avenue

P.O. Box 7854

Madison, WI 53707

608/267-3595

Mats A. Wallin
M. Sc.

University of Lund
Dept. Chen. Engineering Box 124
S-221 00 Lund, SWEDEN
46 46-108244

Ed Warchol

Engineering Specialist
ESS0 Resources Canada, Ltd.
237 Fourth Avenue, S.W.

Calgary, Alberta
CANADA T2P 0N6
403/237-2721

Wendell Warnacut
Mech. Engineer
Tennessee Valley Authority
1101 Market St. MS-3S103C
Missionary Ridge Place
Chattanooga, TN 37402
615/751-6708

Andy M. Warren
Resident Service Engineer
Babcock & Wilcox

1990 N. California Blvd., Suite 400
Walnut Creek, CA 94596
415/947-1100

Larry Waterland
Program Manager
Acurex Corporation
485 Clyde Avenue
Mountain View, CA 94039
415/961-5700

A-29

-------
Greg Weber

Research Supervisor

Energy & Mineral Research Center

15 North 23rd Street

P.O. Box 8213

Grand Forks, ND 58202-8213
701/777-5222

V. L. Weier
N/A

m KEMA
P.O. Box 9035
6800 ET Arnhem
THE NETHERLANDS
085/566023

Roy J. Weiskircher
Sr. Env. Control Engineer
USS a Div. of USX Corporation
4000 Tech Center Drive
Monroeville, PA 15146
412/825-2824

Joat O.L. Wendt
Professor

University of Arizona
Dept. of Chemical Engineering
Building 11
Tucson, AZ 85721
602/621-6050

Arthur V. Vesa
Principal Engineer
The Detroit Edison Company
2000 Second Avenue
Detroit, MI 48226
313/237-5117

Phil Westia

Sr. EaviroraMDtal Engineer s
Arco; Product* Company

1900i-W,_ Crescent Avenue —
Anaheim, CA 92803
714/491-6854

Ed Wiken
Chief

Environment Canada
2067 Fairbanks Avenue
Ottawa, Ontario
CANADA
819/997-1090

Dale V. Wilhelm
Environmental Engineer
Tennessee Valley Authority
Environmental Quality Staff
Knoxville, TN 37902
615/632-6695

Ronald Wilkniss
Staff Engineer
Mobil Oil Corporation
3700 V. 190th Street
Torrance, CA 90509
213/212-4587

Steve M, Wilson
Research Engineer
Southern Company Services
P.O. Box 2625
Birmingham, AL 35202
205/877-7835

Phil Winegar

Senior Engineer

New York Power Authority

1633 Broadway

New York, NY 10019

212/468-6729

Johan Witkamp
Engineer
NV KEMA

Utzechtseweg 310
6900 ET Arnhem
THE NETHERLANDS
085/563625

Evan W. Wong
Engineer

Air Resources Board
P.O. Box 2815
Sacramento, CA 95812
916/445-5980

-------
Bernhard J. Wood
Senior Chemist
SRI International
333 Ravenswood Avenue
Menlo Park, CA 94025
415/859-2305

Ralph T, Yang
Professor

State University of New York
Dept, of Chemical Engineering
Buffalo, NY 14260
716/636-2519

Jerry L, Yee

Senior Mechanical Engineer
Hawaiian Electric Company
P.O. Box 2750
Honolulu, HI 96840
808/543-7708

Nancy Yee

Field Engineering

Bay Area Air Quality Mgmnt. Dist.

939 Ellis Street

San Francisco, CA 94109

415/771-6000

Cherif F. Youssef
Research Project manager
Southern California Gas Company
Box 3249 Terminal Annex, M.L. 73ID
Los Angeles, CA 90051
818/307-2695

Kenneth P. Zak

Marketing Research & Business Dev.
W. R. Grace fie Company
7379 Route 32
Columbia, MD 21044
301/531-4383

Jiansheng Zhao
M. Sc.

University of British Columbia
Chemical Engineering Department
2216 Mail Mall
Vancouver, B.C. V6T lWSmbia
604/224-8579

Melvin L. Zwillenberg
Principal Staff Engineer, R&D
Public Service Electric & Gas Co.
MC-T16H, 80 Park Plaza
Newark, NJ 07101
201/430-6636

A-31

-------