oEPA
United States Industrial Environmental Research EPA-600/7-79-050b
Environmental Protection Laboratory February 1979
Agency Research Triangle Park NC 27711
Proceedings of the Third
Stationary Source
Combustion Symposium;
Volume II.
Advanced Processes
and Special Topics
Interagency
Energy/Environment
R&D Program Report
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
. 3. Ecological Research
4. Environmental Monitoring
5. Socioecondmic Environmental Studies
»
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments of, and development of, control technologies for energy
systems; and integrated assessments of a wide'range of energy-related environ-
mental issues.
EPA REVIEW NOTICE
This report has been reviewed by the participating Federal Agencies, and approved
for publication. Approval does not signify that the contents necessarily reflect
the views and policies of the Government, nor does mention of trade names or
commercial products constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/7-79-050b
February 1979
Proceedings of the Third
Stationary Source Combustion
Symposium;
Volume II. Advanced Processes
and Special Topics
Joshua S. Bowen, Symposium Chairman,
and
Robert E. Hall, Symposium Vice-chairman
Environmental Protection Agency
Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, North Carolina 27711
Program Element No. EHE624
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
-------
PREFACE
These proceedings document more than 50 presentations and discussions
presented at the Third Symposium on Stationary Source Combustion held March
5-8, 1979 at the Sheraton Palace Hotel, San Francisco, California. Sponsored
by the Combustion Research Branch of the EPA's Industrial Environmental
Research Laboratory - Research Triangle Park, the symposium papers emphasized
recent results in the area of combustion modification for NOX control. In
addition, selected papers were also solicited on alternative methods for
NOX control, on environmental assessment, and on the impact of NOX control
on other pollutants.
Dr. Joshua S. Bowen, Chief, Combustion Research Branch, was Symposium
Chairman; Robert E. Hall, Combustion Research Branch, was Symposium Vice-
Chairman and Project Officer. The welcoming address was delivered by Clyde
B. Eller, Director, Enforcement Division, U.S. EPA, Region IX and the opening
Address was delivered by Dr. Norbert A. Jaworski, Deputy Director of IERL-RTP.
The symposium consisted of seven sessions:
Session I:
Session II:
Session III:
Session IV:
Session V:
Session VI:
Session VII:
Small Industrial, Commercial and Residential Systems
Robert E. Hall, Session Chairman
Utilities and Large Industrial Boilers
David G. Lachapelle, Session Chairman
Advanced Processes
G. Blair Martin, Session Chairman
Special Topics
Joshua S. Bowen, Session Chairman
Stationary Engines and Industrial Process Combustion
Systems
John H. Wasser, Session Chairman
Fundamental Combustion Research
W. Steven Lanier, Session Chairman
Environmental Assessment
Wade H. Ponder, Session Chairman
ii
-------
VOLUME II
Table of Contents
Session III: Advanced Processes
Page
"The Influence of Fuel Characteristics on Nitrogen Oxide
Formation - Bench-Scale Studies," M. P. Heap, D. PerShing, G. C.
England, J. W. Lee and S. L. Chen ............. ...... 3
"The Control of Pollutant Formation in Fuel Oil Flames -
The Influence of Oil Properties and Spray Characteristics,"
G. C. England, M. P. Heap, R. T. Horton, D. W. Pershing and
G. Flament ..... ....................... .. . 41
"The Generalization of Low Emission Coal Burner Technology,"
D. M. Fallen, R. Gershman, M. P. Heap and W. H. Nurick ........ 73
"Alternate Fuels and Low NO Tangential Burner Development
jL
Program," R. A. Brown ......................... Ill
"Pollutant Formation During Fixed-Bed and Suspension Coal
Combustion," D. W. Pershing, B. D. Beckstrom, P. L. Case
and G. P. Starley ............. .............. 147
"Advanced Combustion Concepts for Low BTU Gas Combustion,"
B. A. Folsom, C. W. Courtney, T. L. Corley and W. D. Clark ...... 163
"Catalytic Combustion System Development for Stationary
Source Application," J. P. Kesselring, W. V. Krill, E. K.
Chu, and R. M. Kendall ........................ 207
Session IV: Special Topics
"ERPI Low Combustion NO Research," D. P. Teixeira (Abstract)* .... 243
2w
"Flue Gas Treatment Technology for NO Control," J. D. Mobley ..... 245
j&
"Chemiluminescent Measurement of Nitric Oxide In Combustion
Products," B. A. Folsom and C. W. Courtney ....... . ...... 283
(*) See Volume V.
ill
-------
SESSION III:
ADVANCED PROCESSES
G. BLAIR MARTIN
CHAIRMAN
-------
THE INFLUENCE OF FUEL CHARACTERISTICS
ON NITROGEN OXIDE FORMATION
- BENCH-SCALE STUDIES
By:
M. P. Heap, D. W. Pershing, G. C. England, J. H. Lee, and S. L. Chen
Energy and Environmental Research Corporation
Irvine, California
-------
ABSTRACT
Results of two bench-scale experimental studies are presented showing
the effect of the physical and chemical characteristics of both liquid and
solid fuels on fuel nitrogen conversion. The formation of thermal NO was
prevented by the use of artificial oxidant mixtures (Ar/C02/02> free from
molecular nitrogen. Results are presented showing the effect of fuel nitro-
gen content, initial fuel/air mixing process, and droplet/particle size for
both staged and unstaged conditions. The effectiveness of staged combustion
as an NOX control technique appears to be dependent upon fjuel nitrogen
volatility.
-------
ACKNOWLEDGEMENT
The authors are pleased to acknowledge the assistance of several of
their colleagues, R. C. Horton, P. C. Lackey and J. Small in the conduct of
the experiments and to G. B. Martin and W. S. Lanier of the Environmental
Protection Agency for their support and encouragement.
-------
SECTION 1
INTRODUCTION
Fossil fuels contain trace specie which may form pollutants during the
combustion process, e.g., sulfur oxides, nitrogen oxides and fine particu-
late matter. The oxidation of chemically-bound nitrogen during the combus-
tion of fossil fuels provides a significant source of nitrogen oxides, and
sulfur emissions are directly related to the fuel sulfur content. Although
this paper is concerned with the fate of fuel nitrogen, the studies from
which the results are drawn include an assessment of fuel characteristics on
the formation of other pollutants, particularly the size distribution and
trace metal content of fine particulate emissions.
The combustion of coal and residual fuel oils accounts for the major
fraction of nitrogen oxide emissions from stationary combustion sources, and
since it is generally recognized that the oxidation of chemically-bound nitro-
gen contained in these fuels provides a significant fraction of the total
emissions, the factors affecting the fate of fuel nitrogen compounds are of
major importance in the development of low emission combustion systems. The
nitrogen content of coals and petroleum varies considerably. Typical values
range from 0.2 to 0.5 percent for residual fuel oils, and 1.1 to 1.7 percent
by weight for coals although for both fuels examples can be found outside
these ranges. Alternate liquid fuels derived from shale or coal have
higher nitrogen contents than conventional petroleum-derived liquid fuels
and thus, have the potential to produce significant nitrogen oxide emissions
if burned without the application of control techniques.
An effective method of controlling NO emissions from combustors fired
x
with nitrogen-containing fuels is to modify the combustion process in such
a way as to 'ensure that heat release occurs in two stages. Initially, the
fuel reacts under oxygen-deficient conditions which serves to maximize the
-------
production of N from the bound fuel nitrogen specie. Then, provided heat is
lost from the first stage, NO production during burnout in the second stage
is minimized. The optimum design of a staged combustor requires a knowledge
of the fate of the fuel-bound nitrogen as a function of the characteristics
of the fuel and the combustion process.
The effect of combustion on the fate of nitrogen contained in solid and
liquid fuels is illustrated by the simplified flowchart shown in Figure 1.
Before combustion can take place it is necessary that the fuel undergo some
physical transformation to enhance combustion rates by providing for effi-
cient fuel/air mixing. Coal will be pulverized or crushed external to the
combustor, and liquid fuels are injected through a nozzle which atomizes the
liquid, providing a spray of small droplets. The combustion of liquid or
solid fuels in turbulent diffusion flames can be conveniently divided into
three processes.
Devolatilization - As the fuel is heated, matter is driven from
the particle or droplet in the form of a gas or tar which may
undergo secondary pyrolysis reactions producing carbonaceous
solids.
Gaseous Combustion - The gaseous fuel components are burned when
they contact oxygen, provided conditions are suitable for ignition.
Solid Burnout - The char remaining after devolatilization and any
solid produced by pyrolysis reactions in the gas phase are
oxidized. The fraction of nitrogen contained in the fuel which is
converted to nitric oxide will depend upon the availability of
oxygen during the combustion of the gaseous and solid components.
It is generally recognized that fuel nitrogen can be considered in two
fractions: that which can be released from the fuel in the devolatiliztion
process (volatile fuel nitrogen), and that which remains with the solid
(refractory or char nitrogen). The relative amounts of volatile and char
nitrogen will depend upon the droplet/particle heating rate and their final
temperature. The division of fuel nitrogen into two components is compli-
cated by the fact that volatile nitrogen fractions may undergo pyrolysis
reactions where some of that nitrogen remains with the carbonaceous residue,
and is not converted to a simple nitrogenous gaseous molecule.
-------
The major factors governing the conversion of gaseous nitrogenous
specie to NO are well-known, whereas char nitrogen reactions are less well-
understood. However, it appears that in turbulent diffusion flames char
nitrogen conversion to NO is low. Staged combustion processes attempt to
provide a fuel-rich zone with optimum conditions to maximize the conver-
sion of fuel nitrogen compounds to molecular nitrogen because these specie
may form NO during the second stage burnout. An optimized rich first stage
for minimum NOX emissions will prevent char nitrogen from entering the burn-
out zone since conversion to NO in that zone will be small but finite, and
provide the appropriate gas phase stoichiometry and residence time to mini-
mize the content of gaseous nitrogenous specie. If the first stage is too
rich, then the formation of HCN and NH3 is favored; whereas if the first
stage is too lean NO itself will enter the burnout zone. Thus, the two
factors which will control the design of the rich first stage .of a low NOX
combustor are those which pertain to the partition of fuel nitrogen between
the volatile and char nitrogen fractions, and the kinetics (either homoge-
neous or heterogeneous) involved in the production of N2 from nitrogneous
species.
This paper describes a series of bench-scale experiments designed to
identify the influence of both the chemical and physical characteristics
of fuels on fuel nitrogen conversion to NO with the objective of building
a data base which will allow combustor designers to assess the effect of
fuel characteristics on pollutant emissions. The data base will be extended
to include development studies and field tests using the same fuels. Con-
sequently, a body of data will be assembled which will relate the influence
of scale, combustion condition and fuel characteristics to pollutant
emissions.
8
-------
SECTION 2
APPROACH
The combustion of fossil fuels allows the production of nitric oxide
from two sources. In order to assess the conversion of fuel nitrogen to
NO, it is necessary to eliminate or estimate the production of NO from
molecular nitrogen (thermal NO). Various investigators have used different
techniques to achieve this objective. The addition of cooled combustion
products (FGR) to the combustion air has been used as a method of eliminat-
ing thermal NO production with nitrogen-containing fuels (1), or emissions
from a non-nitrogen-containing fuel have been used to assess thermal NO
production (2). In this study thermal NO production was prevented by the
use of artificial oxidants which did not contain molecular nitrogen, thus
any NO appearing in the exhaust was formed either directly or indirectly
from the nitrogen contained in the fuel. The artificial oxidant consisted
of mixtures of argon, carbon dioxide, and oxygen to give the appropriate
oxygen concentration and adiabatic flame temperature. This method is not
free from criticism and has two major drawbacks:
1. The formation of fuel NO may be affected by the large excess
of nitrogen molecules present when air is used as the oxidant.
2. The addition of C02 which is used to balance flame temperatures
because of the different heat capacities of argon and nitrogen
may affect H, OH and 0 concentrations and this affect is most
likely to have a strong influence under fuel-rich conditions.
Although the use of artificial oxidants may not be ideal, it probably repre-
sents the most suitable method of assessing fuel nitrogen conversion during
the combustion of real fuels.
-------
EXPERIMENTAL SYSTEMS
Liquid and solid fuels were burned in two separate, but similar,
refractory down-fired tunnel furnaces whose design has been described in
detail elsewhere (3). The modular construction of both furnaces allowed
ready access for flame observation and staged air injection through side-
wall ports. Combustion air and artificial oxidant mixtures were supplied
from a high pressure source and metered by 600 mm length rotameters, and
they were preheated by electric circulation heaters. The furnace wall tem-
perature was monitored continuously and when not in use, both furnaces were
fired with propane to maintain their thermal equilibrium. Thermal NO pro-
duction was found to be very strongly dependent upon furnace wall tempera-
ture, and it was necessary to learn by experience to match propane heat
release patterns to those obtained with the test fuels to minimize experi-
mental scatter.
The burner used in the liquid fuel furnace is shown in Figure 2, and
several design iterations were necessary in order to provide a stable flame,
and yet prevent coke buildup at the nozzle for the wide range of fuels
investigated. Combustion air was introduced axially and its velocity could
be varied by the use of interchangeable sleeves inserted in the burner
throat. The interchangeable fuel injection system consisted of a 19 mm
diameter stainless steel tube which contained the atomizing air supply tube,
the fuel oil supply tube, a cartridge heater for final oil temperature con-
trol, and a chromal/alumel thermocouple positioned at the injector tip prior
to the nozzle for accurate oil temperature measurement. The major portion
of the experimental investigations carried out to date utilized a commercial
ultrasonic oil atomizer (capacity 0.55 cc/sec) because it provided adequate
atomization of the heavy fuel oils at relatively low flow rates. The tip of
the fuel nozzle was positioned at the beginning of the burner divergent
section, and in general, the visible flame front was displaced approximately
one nozzle diameter from the nozzle tip. The liquid fuels were preheated
electrically in a small supply tank drawn through a 60 micron filter into
a variable-speed Zenith metering pump. The pump outlet pressure was main-
tained constant during calibration, and operation by means of a micro-needle
valve. The oil flow rate was varied by direct mechanical adjustment of the
pump speed.
10
-------
Pulverized coal was supplied pneumatically from a hopper-fed screw
feeder to one of two burners which allowed the coal to be burned entirely
premixed or as a diffusion flame. Premixing was achieved by direct impinge-
ment of the coal jets with the main combustion air supply in a premixing
chamber which was separated from the combustion zone via a series of water-
cooled tubes which prevented flashback (see Figure 3). The coal/air mixture
burned in a plug flow mode with the ignition zone situated in the refractory
divergent. The coal, plus transport air could also be supplied to a variable
swirl burner and injected either axially or radially; thus, allowing experi-
ments to be conducted with very different fuel/air mixing rates. The variable
swirl was achieved by dividing the total oxidant flow into two streams, one
of which flowed axially around the fuel injector, and the other through fixed
swirl vanes providing a tangential velocity cpmponent to the mixed flow.
Details of the pulverized coal burners are presented in Figure 3.
The results presented in this paper all refer to exhaust measurements
made under excess air conditions, and the same sampling and analysis system
was used for both furnaces. It allowed for continuous monitoring of NO, NO ,
3C
CO, CO-, 0~, and S0_. Flue gas was withdrawn from the appropriate exhaust
duct through a water-cooled stainless steel probe. Sample conditioning
prior to the instrumentation consisted of an ice bath water condenser, quartz
wool filters, and a stainless/Teflon sampling pump. All sample lines were
Teflon and stainless steel.
FUEL SELECTION
The characteristics of the liquid and solid fuels tested to date are
presented in Tables I and II, respectively. The more significant fuel
selection criteria were:
Coals. Although the full range of coals selected has yet to
be investigated, -examples of each coal rank are to be tested,
as well as variations in nitrogen and sulfur content within in
each rank. Secondary criteria used in the selection process
were oxygen-to-nitrogen ratio and the form of the sulfur (either
organic or pyritic).
11
-------
Liquid fuels. The major selection criterion was the need to
include residual fuels from the major crude sources available
to U.S. consumers. Alternate fuels were used to extend the
nitrogen content range. Future experiments will use fuels of
widely different nitrogen mass evolution rates as a function
of temperature.
12
-------
SECTION 3
RESULTS
This section of the paper presents the results obtained to date for
both liquid and solid fuels under normal and staged combustion conditions.
The studies are far from complete, and therefore, only preliminary conclu-
sions can be drawn from the results presented below. Whenever possible,
liquid and solid fuels are presented in parallel in order to contrast the
fate of fuel nitrogen in liquid and solid fuels.
FUEL NITROGEN CONTENT, UNSTAGED
Figures 4 and 5 show the emission of fuel NO and the conversion of
fuel nitrogen to NO as a function of fuel nitrogen content for liquid and
solid fuels, respectively. Both sets of results refer to 5 percent overall
excess oxygen, and an air preheat level of 700°F. The results with coal
were obtained with the premixed burner and a particle size distribution of
70 percent through 200 mesh. Combustor heat release rates were similar for
all fuels. For liquid fuels, fuel NO emissions increased almost linearly
with fuel nitrogen content. However, for the coals tested to date there is
almost no effect of weight percent nitrogen on fuel NO emissions.
EFFECT OF FLAME ZONE TEMPERATURE, UNSTAGED
The flame zone temperature was varied by changing the carbon dioxide
content of the artificial oxidant, while maintaining the oxygen concentra-
tion constant. The results presented in Figure 6 indicate that flame zone
temperature has a much stronger influence on certain fuels than on others.
In general, fuel nitrogen conversion in solid fuels is more strongly depen-
dent upon the thermal environment than for liquid fuels, and certain coals
show a very strong dependence upon temperature. These results can be attri-
buted to the influence of heating rate and final temperature on the partition
of fuel nitrogen between the volatile and the char fractions.
13
-------
EFFECT OF INITIAL FUEL/AIR MIXING
Pulverized coal was burned in three modes in order to significantly
modify the fuel/air contacting process, and Figure 7 shows the effect of
the initial fuel/air mixing upon fuel NO formation for one coal as a function
of excess oxygen. It can be seen that premixed conditions lead to the maxi-
mum conversion of fuel nitrogen to fuel NO, and that minimum conversions are
obtained with an axial diffusion flame. The rate of mixing achieved in the
radial swirling flame is intermediate between these two limits, and the fuel
NO emissions are also between the premixed and the axial diffusion flame.
This result is not unexpected and is in agreement with several other investi-
gations (4,5), and can be attributed to the fact that as the coal/air mixing
varies from premixed to an axial diffusion flame there is less oxygen in
contact with the volatile fuel nitrogen fractions, and therefore, the forma-
tion of N£ from the volatile nitrogenous specie is maximized. Figure 7 also
shows that the effect of the initial fuel/air mixing conditions is also
dependent upon the coal composition, and not all coals are as sensitive to
fuel/air mixing rates as others which can be attributed to a variation in
the split between volatile and char nitrogen fractions with coal type. Fig-
ure 8 shows the influence of the initial fuel/air mixing upon emissions from
staged combustion systems with pulverized coal. These results are difficult
to interpret because the actual rich first stage residence time will depend
upon the fuel/air mixing type, but it appears that a backmixed first stage
gives lower final emissions than either the premixed or the axial diffusion
flame. These results were obtained using air as the oxidant because of the
potential influence of C0? under; fuel-rich, conditions/,
THE EFFECT OF PARTICLE/DROPLET SIZE
The influence of droplet size on the fate of fuel nitrogen in liquid
fuels has been investigated by varying the atomizing air pressure to the
ultrasonic nozzle, and by sieving the pulverized coal to burn narrow particle
size distributions. Figures 9 and 10 show the effect of mean drop size on
both fuel nitrogen conversion and thermal NO production for four liquid fuels.
It appears that the fuel nitrogen conversion for alternate fuels is much less
sensitive to liquid droplet size than for petroleum-derived fuels, where as
14
-------
drop size increases fuel nitrogen conversion decreases. Figure 10 shows
there is a very strong influence of drop size on thermal NO production
with an apparent maximum around a mean drop size of 30 microns. The ultra-
sonic atomizer was characterized under cold flow conditions using a laser
diffraction system to determine both the mean and the drop size distribution
of the spray as a function of liquid fuel flow rate, atomizing pressure, and
liquid viscosity. These tests were carried out in a purpose built test rig
and not with the actual oils used in the combustion tests (6).
Figure lla shows that under premixed conditions fuel NO production from
pulverized coal also decreases as particle size increases. However, as shown
in Figure lib, emissions under staged combustion conditions were lowest for
the smallest particles.
15
-------
SECTION 4
DISCUSSION
Heap (4) suggested that under rapid mixing unstaged combustion
conditions volatile fuel nitrogen compounds were mainly responsible for
the NO produced in pulverized coal flames, and that in order to minimize
emissions it would be necessary to maximize the evolution of these volatile
nitrogen compounds from the coal under fuel-rich conditions. It was sug-
gested that fuel nitrogen conversion was highest when the volatile fuel
nitrogen fractions were rapidly mixed with the available combustion air.
Thus, nitrogen volatility would be a major factor in fuel nitrogen conver-
sion to NO under normal combustion conditions. Figure 12 presents a com-
posite plot of all fuels tested to date showing the fractional conversion
of fuel nitrogen to NO as a function of weight percent nitrogen in the fuel.
If the nitrogen is expressed on a dry-ash-free basis, then the conversions
are much lower with coal than with any of the liquid fuels. This is entirely
consistent with the hypothesis presented above, since it is known that some
nitrogen remains with the char and there may not be an equivalence between
char fractions for liquid and solid fuels. Figure 12 also contains one data
point for a gaseous system (NH3 in methane) which represents the extreme in
nitrogen volatility. It can be seen that fuel nitrogen conversions are high.
Many of the results presented in the earlier section for coal can be
explained in terms of an effect upon fuel nitrogen volatility. If heating
rate and final temperature have a strong influence on the division of nitro-
gen between the refractory and volatile fractions, then either increase in
temperature or decreasing particle size would tend to maximize the volatile
nitrogen fraction, thus increasing fuel NO production under normal condi-
tions. Conversely, if in a practical system char nitrogen could not be pre-
vented from entering the second stage of a staged combustion system, then
16
-------
the use of fine particles would minimize NO emissions since the volatile
nitrogen fraction would be maximized.
The effect of fuel nitrogen volatility with liquid fuels is much less
pronounced under normal conditions. The Gulf Coast and the Alaskan oils
appear to have the largest mass of refractory nitrogen assessed by com-
paring nitrogen mass evolution rates obtained under vacuum distillation,
and yet their fractional fuel nitrogen conversion are somewhat higher than
other oils. Also, pyridine is an extremely volatile nitrogen compound, and
yet residual oils doped with pyridine to give the same nitrogen content as
undoped residual fuels give very similar NO emissions to the undoped fuel
(see Figure 13).
The normal combustion conditions used'for the liquid fuels may be such
that fuel volatility has little effect upon fuel NO conversion because the
spray and air distribution give a well-dispersed rapidly mixed heteroge-
neous system which is relatively insensitive to the rate at which the fuel
nitrogen compounds are released from the liquid drops. Under staged com-
bustion conditions, however, if all of the nitrogen has not been vaporized
in the rich stage then the effectiveness of staged combustion operation will
be minimized. Experiments have been carried out with a range of liquid fuels
which have different amounts of refractory nitrogen, i.e., that nitrogen
which remains in a residual fraction after vacuum distillation. Figure 14
compares the effect of staged combustion on total NO emissions from several
liquid fuels. Figure 15 compares the fractional reduction in NOX emissions
achieved under staged combustion conditions as a function of the fraction of
the original nitrogen remaining in the residue for two residence times, and
it can be seen that this parameter is remarkably effective in correlating
the fractional reduction measured in the small-scale furnace.
17
-------
SECTION 5
CONCLUSIONS
The results of the bench-scale studies on the fate of fuel nitrogen
during the combustion of solid and liquid fuels indicate that:
With liquid fuels, fuel nitrogen content is the primary
composition variable affecting fuel NO formation. With
coal, fuel nitrogen content may not be the primary variable
controlling fuel NO production.
Fuel NOX formation appears to be relatively insensitive to
thermal environment for both liquid and solid fuels burning
in a diffusion flame, whereas there is a strong dependence
of fuel NO production on flame temperature in premixed pulverized
coal flames.
Initial fuel/air contacting has a major impact upon fuel NO
production under staged and unstaged conditions.
With liquid fuels, minimum emissions under staged combustion
conditions depend upon the amount of refractory nitrogen in
the original fuel.
A major factor affecting the fate of fuel-bound nitrogen, and therefore,
fuel NO production during fossil fuels combustion appears to be the rela-
tive volatility of the fuel nitrogen compounds. Early evolution of nitrogen
compounds may maximize the benefits to be achieved from staged combustion,
and although alternate liquid fuels contain significant quantities of fuel-
bound nitrogen, NOX control may not be difficult to achieve because the
nitrogen compounds are volatile.
18
-------
REFERENCES
1. Turner, D. W. and C. W* Siegmund. Staged Combustion and Flue Gas Recir-
culation: Potential for Minimizing NO Emissions from Fuel Oil Combus-
tion. Presented at the American FlameXResearch Committee Flame Days,
Chicago, Illinois, September 6-7, 1972.
2. Appleton, J. P. and T. B. Heywood. The Effects of Imperfect Fuel-Air
Mixing in a Burner on NO Formation from Nitrogen in the Air and the Fuel.
Fourteenth Symposium on Combustion, published in The Combustion Institute,
Pittsburgh, Pennsylvania, 1973.
3. Pershing, D. W., J. E. Cichanowicz, G. C. England, M. P. Heap, and G. B.
Martin. The Influence of Fuel Composition and Flame Temperature on the
Formation of Thermal and Fuel NO in Residual Oil Flames. Presented at
"5C"
the Seventeenth Symposium (International) on Combustion, Leeds, England,
1978.
4. Heap, M. P., T. L. Lowes, R. Walmsley, H. Bartelds and P. LeVaguerese.
Burner Criteria for NO Control Volume I Influence of Burner Variables
on NO in Pulverized Coal Flames. EPA-600/2-76-061a, March, 1976.
J\.
5. Wendt, J. 0. L., J. W. Lee and D. W. Pershing. Pollutant Control Through
Staged Combustion of Pulverized Coal. U.S. DOE Report No. 1817-4, U.S.
DOE Technical Information Center, Oak Ridge, Tennessee, 1978.
6. England, G. C., M. P. Heap, R. C. Horton, D. W. Pershing and G. Flament.
The Effect of Fuel Properties and Atomizer Design on Emission Control
from Heavy Fuel Oil Fired Combustors. Paper presented at the Third
Stationary Source NO Symposium, San Francisco, California, March, 1979.
X
19
-------
Physical Transformation
Droplet/Par tide
Heatlng/Devolati1i zati on
Secondary
Reaction
Tar
Secondary
Reaction
NO
'N2
Figure 1. The Fate of Nitrogen in Solid or Liquid
Fuels During Combustion.
20
-------
Ultrasonic Twin-fluid Atomizer
Burner
Viewing Port Section
Thermocouple
Connection
Oil Heater
Connection
Combustion
Air \
JT 1L
Tunnel Furnace
Atomizing Air
* Oil Pressure
Tap
Oil Inlet
Viewing. Port
Ultrasonic
Nozzle
Burner Detail
Insulating
Block
Insulating
Refractory
High Temperature
Refractory
Flue
Furnace Cross-Section
Figure 2. Details of Liquid Fuel-Fired Furnace System.
21
-------
Insulating
Insulblock Refractory
High
Temperature
Refractory
Coal + Transport Air
a
/
i
i
i
i
i
k
^
o'l
0
o
0
0
o
0
o
_^B«^B-
^^"M
J
L
r
.^WH
3
r~
Main
Combustion
Air
Coal + Transport Air
Cooling
Water
Figure 3. Details of Pulverized Coal Furnace System and Burners.
22
-------
o
o
c
01
O)
o
M
r-
0)
3
Lt,
«*-
0
O
»r-
00
o>
c
o
o
>e
60
50
40
30
20
10
0
1 1 T 1 1 II II ||
-
Scranton A
- N' Dak- Utah ^
A Q Black Creek
A Western AL. O
~ Montana KY. ft ~
Savage Rga
- AL'
6
_ W. Va.
1 1 L 1 I 1 I I i i i
0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0
Weight % Nitrogen in Fuel (DAF)
1400
1300
^1200
° 1100
£ 1000
a
ll 90°
a.
o 800
^ 700
600
500
400
1 1 1 1 1 II II II
U?ah Black Cre^k .
- Scrfnton c ^ Western O _
N Dak Savage KY- Rosa AL.
- ' ' Mont. _
- O
W. Va.
~
I I 1 I I 1 i 1 1 I 1
0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1,8 1.9 2.0
% Nitrogen in Fuel (DAF)
Figure 4. Fuel NO and Percent Conversion of Fuel Nitrogen to
Fuel NO - Pulverized Coal Premixed.
23
-------
CVJ
o
X
o
Q.
1600
1400
1200
1000
800
600
400
200
90^
- 5% Excess 02
-
-
-
*
mm ^p
^* ^N^j
J^^1^
r 1 1 1 1
1 1 II
80
o
in
o
o
50
L£Q£NJ2
» ALASKAN DIESEL
, W. TEXAS DIESEL
CALIFORNIA NO. 2
ESSEX COUNTY
MIDDLE EAST
LOW SULFUR NO. 6
INDO/MALAYSIAN
DESULFURIZED VENEZUELAN
PENNSYLVANIA/AHARADA HESS
GULF COAST
VENEZUELAN
ALASKAN
CALIFORNIA #1
CALIFORNIA #2
CALIFORNIA #3
CALIFORNIA W
CRUDE SHALE AND BLEf4i>s
WITH O
SHALE-DERIVED DFM
SRC II
SYNTHOIL BLENDS
WITH 9
CHj, * NHj
cf
.4 .6 .8 1.0 1.2 1.4 1.6 1.8 2.0
Figure 5. Fuel NO and Percent Conversion of Fuel Nitrogen to
Fuel NO - Liquid Fuels, Ultrasonic Atomizer.
24
-------
700
CM
5 60°"
O
% 500 -
S 400
X
O
300
0)
if 200-
^ 100
I I
Wilmington
(1607
(1655) (17U°K)
(1514)
E>--
Alaskan Gulf^Coast
Venezuelan
East Coast
I
I
I
I
I
I
2100 2200 2300 2400 2500 2600°K
Theroetical "Flame Temperature
l_ O Utah
Savage, Montana
~O Western Kentucky
Scranton, N. Dak.
r-a W. Va.
O Black Creek, Al.
O Rosa, Al.
I I I
Volume % of C02 in Oxidizer
Figure 6. Effect of Flame Zone Temperature on
Fuel NO Formation.
25
-------
CM
O
1300
1200
1100
1000
900
U Ta.r = 6500F
. 800
£. 70°
§. 600
Q.
o 500
*5 400
"" 300
200
100
Utah Coal
70 x 103 Btu/hr
0.0
cT*
1400
1200
i
1000
8°°
600
400
200
2.0 4.0
02% in Stack
\<>
l
t
6.0
0
0
Total NO
5% 02
I
1.0 1.2 1.4 1.6 1.8
% Nitrogen in Fuel (DAF)
Figure 7. The Effect of Initial Fuel/Air Mixing - Coal.
26
-------
700 -
600 -
500 -
CVJ
o
o.
Q.
400
300
200
100
0
Utah Coal
0.4
0.6
Axial Diffusion
O
Radial Swirling
0.8
1.0
First Stage Stoichiometric Ratio
Figure 8. The Effect of Initial Fuel/Air Mixing - Staged, Coal.
27
-------
0>
o
o
tt>
70 -
65
60
55
50
45
A Indo/Malaysian
Q Alaskan
SRC II
Crude Shale
20 40 60 80 100 200
Rosin-Rammler Mean Dropsize (ym)
Figure 9. Effect of Droplet Size on Fuel Nitrogen Conversion.
28
-------
300
» 200
O)
Q.
a.
100
A Indo/Malaysian
D Alaskan
+ SRC II
Crude Shale
20 40 60 80 100 200
Rosin-Ranmler Mean Dropsize (urn)
Figure 10. Effect of Droplet Size in Thermal NO Formation.
A
29
-------
CM
o
c.
o
"ttJ
evj
o
130C -
1100-
900-
700-
500-
300
Normal Distribution
130C -
1100 -
900-
700
500
300
4.0
% in Stack
6.0
a) Premixed Unstaged
.4
.6 .8 1.0
First Stage Stoichiometric Ratio
1.2
b) Staged
Figure 11. The Effect of Coal Particle. Size on NO Formation.
30
-------
90 -
80 -
.X
o
2= 70 -
o
«r-
10
V
O)
60 -
50
40
30
20
O
O
I I
^ K k
1
I I I
OCH4 + NH3
O Distillate Oils
b Residual Oils
DAlternate Liquid Fuels
Coals
a
L
I
I
.2 .6 1.0 1.4 1.8
Weight % Nitrogen In Fuel (DAF)
Figure 12. Overall Composite of Fuel Nitrogen Conversion - Unstaged.
31
-------
700
600
CM
O
o 500
£400
0)
2 300
200
100
0.77% N
1.63% S
0.76% N
0.51% N
1.63% S
0.40% N
2.22% S
Pure Residual Oil
O Distillate Oil + Pyridine + Thiophene
A Residual Oil + Pyridine + Thiophene
I I I
_L
I
1
234
% Overall Excess 0,
Figure 13. Relative Volatility Effects - Liquid Fuels - tlnstaged.
32
-------
2000
1800
1600
1400
1200
§1000
Tiriii
Unrefined Shale Oil
" Alaskan Bunker C
O Gulf Coast No. 6
A Indo/Malayslan No. 6
Alaskan Diesel Oil
0.6 0.7 0.8 0.9 1.0 1.1
Primary Zone Stoichiometric Ratio
Figure 14. Effect of Liquid Fuel Composition on Emissions - Staged.
33
-------
0.24
0.22
0.20
0.18
a
4>
o>
S2 0.16
. « ^
o o
0.14
0.12
0.10
0.08
0.0$
I
I
I
I
I
0.3 0.4 0.5
%N in Residue \
% N in Fuel /
(
0.6 0.7 0.8
'% Residue Mass
100
0.9 1.0
(Fraction of Original Nitrogen Left in the Residue)
Figure 15. Importance of Refractory Nitrogen - Liquid Fuels - Staged.
34
-------
u>
en
Ultimate Analysis:
Carbon, %
Hydrogen, %
Nitrogen, %
Sulfur, %
Ash, %
Oxygen, % (by difference)
Conradson Carbon Residue, %
Asphalt ene, %
Flash Point, °F
Pour Point, °F
API Gravity at 60°F
Viscosity, SSU, at 140°F
at 210°F
Heat of Combustion:
Gross Btu/lb
Net Btu/lb
Calcium, ppm
Iron, ppm
Manganese, ppm
Magnesium, ppm
Nickel, ppm
Sodium, ppm
Vanadium,, ppm
Alaskan W. Texas California
Diesel Diesel Distillate
86.99 88.09 86.8
12.07 9.76 12.52
0.007 0.026 0.053
0.31 1.88 0.27
<.ooi <.ooi <.ooi
0.62 0.24 0.36
33.1 18.3 32.6
33.0 32.0 30.8
29.5 28/8 29.5
19,330
Essex
County
86.54
12.31
0.16
0.36
0.023
0.61
2.1
0.34
205
50
24.9
131.2
45
19,260
18,140
7.1
16
0.09
3.7
6.7
37
14
Middle E.
(Exxon)
86.78
11.95
0.18
0.67
0.012
0.41
6.0
3.24
350
48
19.8
490
131.8
19,070
17,980
1.2
2.6
0.02
0.08
13
0.98
25
Low S.*
No. 6 Oil
86.57
12.52
0.22
0.21
0.02
0.46
4.4
0.94
325
105
25.1
222.4
69.6
19,110
17,970
9.52
123.6
0.46
2.23
14.10
3.74
3.11
-------
CO
en
Ultimate Analysis:
Carbon, %
Hydrogen, %
Nitrogen, %
Sulfur, %
Ash, %
Oxygen, % (by difference)
Conradson Carbon Residue, %
Asphaltene, %
Flash Point, °F
Pour Point, °F
API Gravity at 60°F
Viscosity, SSU, at 140°F
at 210°F
Heat of Combustion:
Gross Btu/lb
Net Btu/lb
Calcium, ppm
Iron, ppm
Manganese, ppm
Magnesium, ppm
Nickel, ppm
Sodium, ppm
Vanadium, ppm
West
Texas
86.01
9.72
0.46
3.53
0.048
0.23
20.08
9.24
205
94
6.2
29,720
1,824
17,800
16,910
5,9
43
0.13
0.9
29
3.2
60
Alaskan
86.04
11.18
0.51
1.63
0.034
0.61
12.9
5.6
215
38
15.6
1,071
194
18,470
17,580
6.9
24
0.06
1.4
50
37
67
California California
#1 #2
85.75 85.75
11.83 11.44
0.62 0.77
1.05 1.63
0.038 0.043
0.71 0.71
8.72
5.18
38
19.5 15.4
246.1 854
70.00 129
18,470
17,430
21
73
0.8
5.1
65
21
44
California
#3
85.41
11.23
0.79
1.60
0.032
1.02
9.22
5.18
150
30
15.1
748.0
131.6
18,460
17,440
14
53
0.1
3.8
82
2.6
53
-------
TABLE I. LIQUID FUELS TESTED (CONTINUED)
Ultimate Analysis:
Carbon, %
Hydrogen, %
Nitrogen, %
Sulfur, %
Ash, %
Oxygen, % (by difference)
Conradson Carbon, Residue %
Asphaltene, %
Flash Point, °F
Pour Point, °F
API Gravity at 60°F
Viscosity, SSU, at 140°F
at 210°F
Heat of Combustion:
Gross Btu/lb
Net Btu/lb
Calcium, ppm
Iron, ppm
Manganese, ppm
Magnesium, ppm
Nickel , ppm
Sodium, ppm
Vanadium, ppm
Indo/
Malaysian
86.53
11.93
0.24
0.22
0.036
1.04
3.98
0.74
210
61
21.8
199
65
19,070
17,980
14
16
0.13
3.6
19
15
101
Venezuelan
Desulphurized
85.92
12.05
0.24
0.93
0.033
0.83
5.1
2.59
176
48
23.3
113.2
50.5
18,400
17,300
8.7
6.5
0.09
3.6
19
15
101
DFM Pennsylvania Gulf
(Shale) (Amarada Hess) Coast
86.18
13.00
0.24
0.51
0.003
1.07
4.1
0.036
182
40
33.1
36.1
30.7
19,430
18,240
0.13
6.3
0.06
0.43
0.09
<.l
84.82
11.21
0.34
2.26
0.067
1.3
12.4
4.04
275
66
15.4
1049
240
18,520
17,500
9.2
13.2
0.10
3.3
32.7
64.5
81.5
84.62
10.77
0.36
2.44
0.027
1.78
14.8
7.02
155
40
13.2
835
181
18,240
17,260
4.4
19
0.13
0.4
29
3.6
45
-------
CO
CO
Ultimate Analysis:
Carbon, %
Hydrogen, %
Nitrogen, %
Sulfur, %
Ash, %
Oxygen, % (by difference)
Conradson Carbon, Residue %
Asphaltene, %
Flash Point, °F
Pour Point, °F
API Gravity at 60°F
Viscosity, SSU, at 140°F
at 210°F
Heat of Combustion:
Gross Btu/lb
Net Btu/lb
Calcium, ppm
Iron, ppm
Manganese, ppm
Magnesium, ppm
Nickel, ppm
Sodium, ppm
Vana,diuro» PPm . , . , , . .
California
#4
86.66
10.44
0.86
0.99
0.20
0.85
15.2
8.62
180
42
12.6
720
200
18,230
17,280
90.6
77.2
0.87
31.4
88.0
22.3
66.2
Synthoil
86.30
7.44
1.36
0.80
1.56
2.54
23.9
16.55
210
80
S-1.14
10,880
575
16,480
15,800
1,670
109
6.2
170
2.6
148
6.5
Shale
(Crude)
84.6
11.3
2.08
0.63
.026
1.36
2.9
1.33
250
80
20.3
97
44.1
18,290
17,260
1.5
47.9
0.17
5.40
5.00
11.71
<.3
-------
TABLE II. SOLID FUELS INVESTIGATED
oo
vo
EER COAL DESIGNATION
Coal Source
Ultimate Analysis;
Carbon, %
Hydrogen, %
Nitrogen, %
Sulfur, %
Ash, %
Oxygen, %
(by difference)
Proximate Analysis:
Moisture, %
Volatiles, %
Ash, %
Fixed Carbon, %
Calorific Value
(Btu/lb)
Forms of Sulfur
Pyritic, %
Organic, %
Sulfate, %
B
West
Virginia
71.57
4.94
1.33
1.09
11.66
7.54
1.87
32.53
11.66
53.94
D
Utah
68.49
5.20
1.24
0.64
7.40
10.64
6.39
38.89
7.40
47.32
12,340
4
Scranton
N.D.
42.02
2.71
0.54
0.99
7.50
11.28
34.96
28.85
7.50
28.69
6,446
0.32
0.65
0.02
11
Montana
Savage #11
41.36
2.57
0.72
0.27
4.61
14.11
36.36
4.61
6,995
0.01
0.26
0.00
18
Rosa
Alabama
74.72
4.36
1.60
0.96
6.79
3.55
8.02
21.81
6.79
63.38
13,394
0.43
0.47
0.06
19
Black Creek
Alabama
80.27
5.09
1.75
0.74
4.45
5.45
2.25
28.28
4.45
65.02
14,284
0.07
0.67
0.00
20
Western
KY
69.50
4.76
1.33
2.95
7.81
8.85
4.80
36.10
7.80
51.1
12,450
-------
THE CONTROL OF POLLUTANT FORMATION IN FUEL OIL FLAMES
- THE INFLUENCE OF OIL PROPERTIES AND
SPRAY CHARACTERISTICS
By:
G. C. England, M. P. Heap, R. T. Horton, D. W. Pershing and G. Flament*
Energy and Environmental Research Corporation
Irvine, California 92715
International Flame Research Foundation, IJmuiden, Holland
41
-------
ABSTRACT
Pollutant production in liquid fuel flames is controlled by the
complex interaction of the liquid fuel spray and the combustion air flow
field. This paper presents results of a study to investigate the influ-
ence of atomizer design and fuel properties under both normal and staged
combustion conditions on nitrogen oxide formation. The sprays produced
from the various atomizers were characterized under isothermal conditions
by a method involving the diffraction of laser light.
42
-------
ACKNOWLEDGEMENTS
The work presented in this paper was carried out under EPA Contracts
68-02-2624 and 68-02-3125, and it is a pleasure to acknowledge the support
and assistance of Mr. G. B. Martin and Mr. W. S. Lanier, the Project Officers,
43
-------
SECTION 1
INTRODUCTION
Although the increased utilization of coal provides a partial solution
to the U. S. energy crisis, a balanced fuel economy necessitates that in the
future many industrial users will burn petroleum, coal or shale-derived
residual fuels. The relatively high nitrogen content and low hydrogen-to-
carbon ratios associated with these liquid fuels suggests that their use may
be accompanied by an increase in two pollutants unless preventive measures
are taken. The production of nitrogen oxides (NO ) and particulate matter
X
from a given fuel in turbulent diffusion flames is controlled by the fuel/air
contacting process, which is dependent upon the complex interaction of the
liquid fuel spray and the combustion air flow field. A major fraction of the
nitrogen oxides produced in residual fuel oil flames is formed by the oxidation
of nitrogen contained in the fuel. The fraction of fuel nitrogen which is
converted to fuel NO is almost completely dependent upon the oxygen avail-
ability during the thermal decomposition of the liquid droplets. Particulate
matter produced in residual fuel oil flames consists of soot, which is pro-
duced from the gas phase specie, and two other constituents, cenospheres and
coke, production of which is associated with the multi-component nature of the
fuel and the characteristics of droplet combustion. The formation of these
latter two components can be attributed to the occurrence of liquid phase
thermal cracking because of high droplet temperatures. Soot formation, which
occurs in three distinct phases: nucleation, surface growth, and coagulation,
is associated with the existence of fuel-rich zones within the flame. Thus,
local stoichiometry which is controlled by the turbulent mixing process and
is spray/flow field-dependent plays a dominant role in the formation of NO
and particulate matter.
Bench-scale studies (1) indicate that the total nitrogen content of
liquid fuels is the major parameter controlling fuel NO formation under
44
-------
overall fuel-lean conditions. The fuel nitrogen speciation appears to
be a secondary parameter unless the fuel is being burned under staged com-
bustion conditions, when fuel nitrogen volatility appears to limit NO
X
reduction achievable under staged combustion. Field studies (2) have investi-
gated the influence of fuel properties and atomizer design on both the pol-
lutant emissions from, and thermal efficiency of operating boilers. Although
these studies indicated that both of these parameters were important, the
nature of the tests precluded the identification of the controlling param-
eters. The results presented in this paper were obtained under conditions
which were typical of industrial practice, and yet allowed the influence of
atomizer type and fuel properties to be identified under comparable conditions,
The specific objectives of the research program were:
To relate the formation of nitrogen oxides in liquid fuel flames
to atomizer design and to fuel oil properties.
To investigate the interaction between liquid sprays and the
airflow pattern for conditions typical of package boilers under
normal and staged conditions.
The information generated by this program will provide for the generalization
of low-NO oil burner technology for application to package boilers.
45
-------
SECTION 2
EXPERIMENTAL SYSTEMS
The combustion experiments were conducted in an axisymmetric combustor
at a nominal firing rate of 3 x 10 Btu/hr. (.88 MW). This combustor, which
has been described in detail elsewhere (3), allowed the addition of cooled
combustion products to the combustion air supply and had the capability of
dividing the combustion air into two separate streams to allow operation
under staged combustion conditions. In these investigations the staged air
was injected through an axial boom inserted from the rear of the combustion
chamber. The distance of the air injector from the fuel atomizer could be
varied. Standard techniques and instrumentation were employed for the
measurement of flue gas composition.
BURNER AND ATOMIZERS
The commercial burners used in this investigation were fitted to an axial
oil gun in a simple burner system which consisted of:
a refractory divergent exit, 45 half angle
an annular air duct with interchangeable fixed-vane swirlers
generating an axial velocity of 100 ft/sec (31 m/sec) and swirl
numbers of 0.21, .45 and .79.
The oil gun was precisely positioned to locate all the atomizers at the throat
of the divergent and the combustion air supply was designed to give a uniform
air distribution across the throat. A steam/auxiliary electric heating system
was used to maintain the oil temperature (measured by a thermocouple) at the
oil nozzle constant. Table 1 lists the atomizers used in the study which were
of the twin fluid (air or steam assist) type.
SPRAY CHARACTERIZATION
The characteristics of the sprays produced by the commercial atomizers
used in the study were not available at the time this paper was written.
46
-------
Droplet size distribution will be determined under quiescent isothermal
conditions using a laser droplet size analyzer based on the Fraunhofer
Diffraction pattern of a laser beam by the particles moving through a cross-
section of the beam (4). The particle cloud of a spray causes a deflection
of a portion of the parallel beam giving a conical pattern of diffracted
light. There is a direct relationship between the diameter of the particle
(assumed spherical) and the angle of the light diffracted by this particle.
Therefore, it is possible to calculate the particle size distribution of a
cloud of particles by an evaluation of the diffracted light energy. The
conical diffracted light pattern is measured in the focal plane of a lens by
a photo detector element composed of 30 annular rings whose radii increase
from 0.148 mm to 14.56 mm. The instrument evaluates the drop-size distribution
and expresses it in a Rosin-Rammler form together with the weight percentage
and number of drops in each of 15 groups.
The advantages of this laser diffraction technique are that it is fast
in comparison to photographic techniques, requires no calibration for parti-
cles with diameters ranging from 10 to 500 ym and it does not interfere with
the spray. However, there may well be major limitations associated with the
use of this technique and a comparison of various optical techniques for spray
characterization has been initiated under the FCR program. The primary limi-
tation of the diffraction technique is that imposed by spray density which
arises due to multiple scattering. Other limitations include an undefined
influence of droplet velocity, a minimum drop size detectability of 10 ym and
the use of the Rosin-Rammler distribution.
Figure 1 provides an.illustration of the use of this instrument to
compare drop-size distribution from three atomizers:
- a pressure jet (mechanical)
a single lobe of a Y-jet (twin fluid)
- an ultrasonic atomizer (twin fluid)
It can be seen that the pressure jet and the Y-jet give very similar drop-
size distributions while the ultrasonic atomizer gives small droplets with a
narrow size distribution around 20 ym.
47
-------
FUELS
The fuels used during the investigation are listed in Table II, together
with selected properties. Figure 2 presents the results of a Differential
Thermal Analysis (DTA) and Figure 3 shows the mass evolution of nitrogen as
a function of distillation temperature. The DTA clearly shows the multi-
component nature of the petroleum-derived residual fuels. It can also be
seen that the nitrogen in the shale-derived fuels (Nos. 4 and 8) is evolved
at a much lower temperature than that from petroleum-based residual fuels.
Apparently the nitrogen species in the Gulf Coast and North Slope oils
(Nos. 5 and 6) appear to be refractory in nature since almost all the nitrogen
remains in the residuum.
48
-------
SECTION 3
RESULTS
The series of investigations planned to define the influence of atomizer
design and fuel characteristics on the emission of NO from residual fuel oil-
X
fired flames can most conveniently be described under two separate headings:
those concerned with normal excess air conditions, and those which pertain to
two-stage combustion in which the first stage was operated fuel-rich and air
was added to the rich combustion products downstream from the fuel nozzle to
complete combustion.
NORMAL EXCESS AIR CONDITIONS
The results presented in Figure 4 compare NO emissions as a function
X
of overall excess air level for a single fuel (No. 3) with a burner swirl
number of 0.45. It can be seen that in general the NO emissions increase
X
with increasing excess air level for each atomizer. However, there is a wide
range in emission level for different atomizers. The results presented in
Figure 5 show the influence of fuel type on NO emissions at two swirl levels,
X,
S = 0.45 and 0.79, with a single nozzle (B). These results clearly show the
impact of the flow field on NO formation.
x
It is generally recognized that NO emissions increase with increasing
fuel nitrogen content. Figure 6 presents a plot of NO emissions as a function
JV
of fuel nitrogen content for four different atomizers and two swirl levels.
For each atomizer there appears to be an almost linear relationship between
NO emission and fuel nitrogen content. However, this relationship is not
X
general and depends both upon atomizer design and the flow field conditions.
For instance, atomizer B gave very high NO emissions which were similar to
X
those produced by atomizer A at the low swirl level. However, at the higher
swirl condition emissions from atomizer A were almost unaffected by the swirl
level, whereas those from atomizer B were considerably reduced. At the inter-
mediate swirl level the four atomizers appear to fall into two categories,
whereas at the high swirl level three of the four atomizers behaved similarly.
49
-------
The data presented in Figure 6 illustrates some of the difficulties associated
with any general correlation relating fuel nitrogen content to emissions since
the nature of the combustion system has a significant influence.
STAGED COMBUSTION CONDITIONS
Staged combustion, i.e., the operation of a combustion system in which
the fuel originally burns under oxygen-deficient conditions, provides the
most cost-effective control technique for reducing fuel NO. The results pre-
sented in Figure 7 show the influence of atomizer design on the emission of
NO at an overall excess air level corresponding to three percent excess oxygen
as a function of the percentage of theoretical air in the fuel-rich primary
zone. The burner throat conditions were unmodified, and therefore, as the
percentage of theoretical air in the primary zone decreases, the burner throat
velocity also decreases. The second stage air was added at the same location
for each atomizer. The atomizer design has a significant influence upon the
minimum NO achieved under staged combustion conditions. All of the nozzles
show the same characteristic performance. As the primary zone becomes more
fuel-rich NO emissions decrease to a minimum and then increase. This increase
ii
is probably associated with the escape of fuel nitrogen specie from the primary
zone where they are oxidized to NO when the second staged air is added. Com-
paring the minimum emission levels with those reported for the bench-scale
furnace with the same fuels, it can be seen that the results shown in Figure 7
do not show the same degree of control. This can be attributed to two factors.
The bench-scale combustor has refractory walls, and therefore the initial fuel
rich stage will be at a higher temperature. Residence times in the bench-scale
combustor rich zone were also much longer than those achieved in the cold-wall
combustor.
The trade-off between the decrease in NO emissions and the increase in
X
smoke emissions is illustrated by the data presented in Figure 8 which shows NO
3
and Bacharach smoke number as a function of percent theoretical air in the
primary zone for three swirl conditions with a single nozzle. The NO emis-
sions appear to be relatively insensitive to primary zone swirl number for
this particular nozzle. However, it can be seen that the smoke emissions
differ considerably for the same range of swirl levels.
50
-------
A cylindrical refractory extension 36 inches (.91 meter) long was
added to the divergent burner exit to ascertain whether the reduced heat
loss in the first stage could account for the different emission levels
achieved in the combustor and the bench-scale experiments. The addition of
this refractory extension had almost no effect upon exhaust NO emissions.
Figures 9 and 10 present data comparing the performance of two atomizers
under staged combustion conditions for two different fuels and zero swirl in
the primary zone. One fuel was essentially free from nitrogen, and it can
be seen that the emission levels from the two nozzles are dependent upon fuel
nitrogen content. Also, at low theoretical air ratios in the primary zone
nozzle A gives higher smoke emissions than nozzle B for both fuels.
51
-------
SECTION 4
DISCUSSION
Data available from both field and development studies indicates that
the emission of nitrogen oxides from cold-wall combustors is a complex
function of fluid dynamics and fuel chemistry. The results presented in this
paper were part of an investigation planned to provide data under well-
controlled conditions to define the effect of fuel properties and atomizer
parameters on pollutant emissions from combustion systems typical of those
encountered in practice. At the time of writing the paper one section of
the investigation, that concerned with spray characterization, had not been
completed, and therefore the interpretation of the results can only be con-
sidered as preliminary.
The primary fuel parameter controlling NO formation in liquid fuel
X
flames is the fuel nitrogen content of the oil. NO emissions increase with
3C
increasing nitrogen content. However, the conversion of fuel nitrogen to NO
3C
is strongly dependent upon the mixing of fuel and air within the flame.
This interaction is illustrated by the results presented in Figure 11 which
shows the NO emissions as a function of fuel nitrogen content obtained with
atomizer B. The open data symbols refer to measurements made with pure com-
bustion air. The partially shaded data symbols refer to NO emissions mea-
sured when the combustion air was vitiated with approximately 20 percent
flue gas recirculation. It can be seen that with flue gas recirculation
emissions are almost independent of swirl level and very similar to those
obtained with unvitiated air at a swirl level of S = 0.79. This could be inter-
preted as meaning that the difference in emission levels between the two
swirl levels is due to thermal NO production. However, this is probably an
erroneous conclusion since the addition of flue gas recirculation will not
only lower flame temperatures due to the addition of an inert thermal ballast
but will also modify the mixing process due to the increased throat velocities.
Extrapolation of the two unvitiated lines to zero percent nitrogen indicates
52
-------
a thermal N0x production of approximately 170 ppm which is higher than that
achieved with low nitrogen diesel fuels. Figure 11 also allows a comparison
between the bench-scale results (1) and those obtained in this investigation.
The thermal characteristics of the bench-scale furnace are different from the
3,000,000 Btu/hr combustor, and therefore, only the fuel NO emissions have
x
been plotted. The same atomizer design was used for both investigations,
although the capacity was different. Apparently fuel nitrogen conversions
were higher in the bench-scale investigations than those obtained in the
combustor. This might well be associated with a change in drop size as the
atomizer capacity is increased, i.e., for a given atomizer design drop size
is not independent of capacity.
The influence of atomizer design on NO emissions is probably due to the
X.
combined effect of drop-size distribution, spray pattern, and the momentum of
the droplets. Since these parameters will control the heating rate, penetra-
tion of the droplets through recirculation zones and the local stoichiometry.
Figure 12 compares the performance of five nozzles with a single fuel as a
function of swirl level. As the swirl level increases the size and intensity
of the internal axial recirculation zone will increase, and the NO emissions
x
are for the most part, associated with the interaction of the spray and the
internal recirculation zone. It appears that different nozzle designs will
influence this interaction. For example, nozzle A shows a continual decrease
of emissions with increasing swirl level, whereas many of the other nozzles
show a minimum emission at the intermediate swirl level. Table III presents
the ratio of NO at swirl level equal to 0.5 and 0.79 for four atomizers and
2£
four fuels. This table indicates that there is also an influence of fuel
properties on the interaction of the spray and the flow field since for a
given atomizer increasing the swirl level may increase or decrease the nitrogen
oxide emission.
It was noted earlier that the NO reductions achieved in the bench-scale
x
experiments were higher than those obtained in the combustor. This can be
seen in Figure 13, which compares the emissions from a staged crude shale oil
containing over 2 percent nitrogen. In the bench-scale studies at the shortest
primary zone residence time of 400 msec minimum emissions were approximately
53
-------
200 ppm. In the combust or primary zone residence time was much shorter,
about 200 msec, and emissions were always greater than 300 ppm.
It can be seen that there is a very significant effect of atomizer
design upon NO emissions. The atomizer giving the lowest emissions under
X
staged conditions gave the highest value unstaged. Comparison with informa-
tion presented by Heap et al (1) would suggest that nozzle B has a much
smaller drop-size distribution than nozzle A, thus allowing for more rapid
vaporization of the fuel nitrogen compounds in the primary zone giving more
time for the formation of N., and thus, lower overall emissions.
54
-------
REFERENCES
1. Heap, M. P., D. W. Pershing, G. C. England, J. W. Lee, and S. L.
Chen. The Influence of Fuel Characteristics on Nitrogen Oxide
Formation - Bench-Scale Studies. Third Symposium on Stationary
Source Combustion, March 1979.
2. Cato, G. A., L. J. Muzio, and P. E. Shore. Field Testing: Application
of Combustion Modification to Control Pollutant Emissions from Industrial
Boilers, Phase II. EPA Report No. 600/l-76-086a, April 1976.
3. Muzio, L. J., and R. P. Wilson. Experimental Combustor for Development
of Package Boiler Emission Control Techniques. Phase I of III, EPA-R2-
73-292a (NTIS PB No. 224274/AS), July 1973.
4. Swithenbank, J., J. M. Beer, D. S. Taylor, D. Abbott, and G. C. McCreath.
A Laser Diagnostic for the Measurement of Droplet and Particulate Size
Distribution. University of Sheffield, Department of Chemical Engineering
and Fuel Technology. Report 1976.
55
-------
70
60
50
n
o
Pressure Oet
Ultrasonic
Y-Jet Single Hole
en
0.40
GO
30
20
10
Diameters (pm)
Figure 1. Dropsize Distribution for Three Atomizers.
-------
Indo/M»laysian
Middle Eastern
Gulf Coast
California
North Slope
North Slope Diesel
DFM
0.0
100
600
Temperature, C
Figure 2. DTA Analysis of Eight Fuel Oils.
-------
40
k 2 - Middle Eastern
A 3 - Indo/Malaysian
4 - DFM (Shale Derived
5 - Gulf Coast
D 6 * North Slope
O 7 - California
8 - Shale (Unrefined)
30
QJ
£ 20
10
500
800
900
1000
1100
Distillation Temperature (°F)
Figure 3. Vacuum Distallation Results - Nitrogen Evolution.
58
-------
260
220
CM
o
o
4->
10
180
140
100
D Atomizer A
O Atomizer B
Atomizer C
Atomizer D
I
_L
k Atomizer E
Q Atomizer F
Q Atomizer G
O Atomizer H
I
2.0
3.0 4.0
% Excess Oxygen (Stack)
5.0
Figure 4. Comparison of NO Emissions for Eight Commercial Atomizers
in a Common Flow rield.
59
-------
600
Swirl No = 0.45
500
400
01
o
Q.
Q.
300
200
_L
Swirl No = 0.79
O California
Q North Slope
^ Gulf Coast
£ Indo/Malays1an
k Middle Eastern
52
% Excess Air
Figure 5. Influence of Fuel Type on N0x Emission for a Single Atomizer in Two Swirl Levels.
-------
Atomizer B
Atomizer A
Atomizer F
Atomizer D
600
CM
o
o
4->
a
400
a.
a.
200
0
O Atomizer B
D Atomizer A
O Atomizer F
Atomizer D
0
0.2 0.4 0.6 0.8
Wt. % Nitrogen in Fu$l
1.0
0 0.2 0.4 0.6 0.8
Wt. % Nitrogen in Fuel
1.0
Figure 6. Influence of Atomizer Type at Two Swirl Levels.
-------
400
Swirl = .45
3% Excess Oxygen
Atomlzer
DA
A D
OB
300
01
no
CM
O
o
4->
IB
200
i
100
I
I
60
80 100
% Theoretical Air - Primary Zone
120
Figure 7.
Comparison of W Emission From Four Atomizers Under Staged Combustion.
-------
400
300
t
a
S 200
Q.
Q.
100
D 15° Swirl Vane Open Symbols - NO
O 30° Swirl Vane Solid Symbols - Smoke
A 45° Swirl Vane
0.6
0.7 0.8 0.9 1.0
Primary Zone Stoichiometric Ratio
1.1
1.2
8
s-
_ O)
6 -
-------
200
150
CTl
CM
O
100
a.
a.
50
i r
Alaskan Diesel 011
Zero Swirl
3% Excess 00
Atomi zer
Atomizer
A
A NOX
A Smoke
0.6
0.7
0.8 0.9 1.0
Primary Zone Stoichiometric Ratio
1.1
Figure 9. Influence of Atomizer Type on NO and Smoke Emissions Under
Staged Combustion - North Slope ftiesel Oil.
9
8
6
5
4
3
2
1
o
£
to
JC
u
to
CO
-------
400
300
CM
O
a.
a.
200
100
Alaskan RFO, 3%
Zero Swirl
O Atomizer B
A Atomizer A
I
Open Symbols - NO
Solid Symbols - Smoke
I
I
I
I
0,6
0.7 0.8 0.9 1.0
Primary Zone Stoichiometric Ratio.
1.1
1.2
9
8
7
6i
50
j*
4 <>
2
1C
3 tJ
J OQ
2
I
Figure 10. Influence of Atomizer Type on NO and Smoke Emissions Under
Staged Combustion - North Slope Bunker C.
-------
800
600
CM
o
o
4-»
0
400
200
0
Fuel NO
n
Bench Scale (1),
I
I
Swirl No.
O 0.45
D 0.79
0.0 0.2
0.4 0.6
Wt. % Nitrogen
0.8
1.0
1.2
Figure 11. Comparison of NO Emission From BenchScale and Cold-Wall
Combustor Experiments.
66
-------
T
320
300
280
260
CM
O
ox 240
o.
Q_
220
200
180
TIndo/MalayslanT
4% Excess O (Stack)
Atomizers
D A
O B
O c
.2
.4 . .6
Swirl Number'at Throat
.8
Figure 12. NO Emissions as a Function of Swirl Number for Five Atomizers.
JL
-------
Firetube Combustor
-a
1500
1400
1300
1200
1100
oo
0.6
Crude Shale |
3% Excess Op
No Swirl/Refractory Insert
O Atomizer B
D Atomizer A
£3 Bench Scale Experiments
I
0.8 0.9
Primary Zone Stoichiometric Ratio
1.1
Figure 13. NO Reduction - Comparison of Bench Scale Experiments With Cold Wall Combustor.
X
-------
TABLE I. LIST OF COMMERCIAL ATOMIZERS
ATOMIZER DESCRIPTION
A Internal mixing, air/steam assist, swirl
chamber and pintle
B Ultrasonic air assist
C Internal mix, prefilming, air/steam assist,
pintle
D Y-jet internal mixing, air/steam assist
E Y-jet internal mixing, air/steam assist
F Y-jet internal mixing, air/steam assist
G Y-jet internal mixing, air/steam assist
swirl cap
H Low pressure Y-jet internal mixing, swirl
cap air assist
69
-------
TABLE II. SELECTED FUEL OIL PROPERTIES
Property
Ultimate Analysis
Nitrogen %
Sulfur %
Asphalt ene %
API Gravity @ 60°F
Gross Heat of Combustion
N. Slope
Diesel
0.007
0.31
-
33.1
19,400
Middle
Eastern
0.18
0.67
3.24
19.8
19,070
Indo/
Malaysian
0.24
0.22
0.74
21.8
19,070
Diesel
Fuel
Marine
0.24
0.51
0.036
33.1
19,430
Gulf
Coast
0.36
2.44
7.02
13.2
18,240
N. Slope
Bunker C
0.51
1.63
5.6
15.6
18,470
California
0.77
1.63
5.18
15.4
18,470
Shale
Oil
2.08
0.63
1.33
20.3
18,290
Btu/lb
Vanadium ppm
25
101
45
67
44
<0.3
-------
TABLE III. RATIO OF NOV EMISSIONS AT INTERMEDIATE AND HIGH
A
SWIRL LEVELS FOR FOUR ATOMIZERS AND FOUR FUELS
ATOMIZER
FUEL A B C F
2 1.07 1.26 0.98 0.91
3 1.46 0.96 0.94 0.96
6 0.96 1.52 0.84 0.93
7 1.02 1.53 0.84 1.15
71
-------
THE GENERALIZATION OF LOW EMISSION
COAL BURNER TECHNOLOGY
By
D. M. Zallen, R. Gershman, M. P. Heap and W. H. Nurick
Energy and Environmental Research Corporation
Santa Ana, California 92705
73
-------
ABSTRACT
This paper describes the development of a low NO pulverized coal
jt
burner. Results are presented from a series of tests designed to define
optimum burner design parameters at three scales 10, 50 and 100 Btu/hr
heat input. NO emissions were found to be most sensitive to burner zone
x
stoichiometry and fuel injection parameters. Results are presented for
three bituminous coals. The use of dry sorbents for S0_ control is
discussed and results are presented which suggest that low NO coal burners
A
may provide optimum conditions for combined NO - SO control using flame
Jv Jv
zone sorbents.
74
-------
NOMENCLATURE
SR
T
SA
S
ppm
Primary Air
Secondary Air
Tertiary Air
(air/fuel) actual/
(air/fuel) stoichiometric
Stoichiometric Ratio
Temperature
Swirl angle
_ . , , G_ Angular momentum
Swirl number = -r^-r = . ? ,
G R Axial momentum x characteristic radius
Xt-
Parts per million by volume on a dry basis and reduced to
zero percent excess air
Air in fuel injector
Burner air in coannular passage around the fuel injector
Air used for staging
SUBSCRIPTS
Air with coal in fuel injector
Air in the coannular flow passage around fuel injector
Staging air injected from separate ports located
at some radial distance from the burner centerline
75
-------
ACKNOWLEDGMENTS
The work described in this paper was carried out during the conduct
of EPA Contract 68-02-2667. The authors are pleased to acknowledge the
considerable help and encouragement provided by G. B. Martin of the EPA
and of several of their EER colleagues C. McComis, M. Deming, R. Thomas,
R. Thalken and J. Lee during the course of this investigation.
76
-------
SECTION I
INTRODUCTION
An increase in the direct combustion of coal, either pulverized and
burned in suspension or crushed and burned in fixed or fluidized beds, repre-
sents the only near term solution to decrease our dependence upon the import
of foreign petroleum products because of increased utilization of coal. How-
ever, the increased use of coal enhances environmental problems associated
with its extraction, transportation and utilization. Coal contains several
trace compounds which have the potential to produce atmospheric pollutants
during the combustion process. This paper is concerned with the generaliza-
tion of technology which is controlling the emission of two pollutants,
nitrogen oxides (NO ) and oxides of sulfur (SO ) from pulverized coal-fired
2t j£
combustors. Direct coal combustion represents the major source of nitrogen
oxides emitted by stationary sources. Consequently, unless appropriate
emission control technologies are applied an increase in the utilization of
coal in direct fired combustors has the potential to significantly increase
NO emissions.
x
Historically, application of NO control to utility boilers represents
the most cost-effective approach to controlling total emissions from stationary
sources. Emissions from coal-fired utility boilers have been estimated to
account for 32 percent of all emissions from stationary sources. The EPA has
established an intensive program to demonstrate control technology for pul-
verized coal-fired boilers; industrial as well as utility, which will prevent
adverse impact on the environment in the event of increased coal firing in
water wall boilers. Technical goals have been established for 1985 which are
equivalent to 20 percent of the existing New Source Performance Standards
(NSPS) for NO emissions. These goals are to be met by combustion modifica-
2m
tion techniques (i.e., changing the process by which fuel and oxidant are
brought together then how reaction occurs and heat is released). These
techniques have been successful in allowing boilers to comply with NSPS.
77
-------
However, the 1985 technical goal might well represent the limit of their
application since the near term requirement dictates that the control
techniques be compatible with existing state-of-the-art boilers. Major modi-
fications to the boiler would involve an extensive development program and
necessarily the control technology would take time to be accepted by the
utility industry.
Two processes are involved in the emission of nitrogen oxide from coal-
fired boilers: 1) the oxidation of molecular nitrogen (thermal NO), and
2) the oxidation of nitrogen contained in the coal (fuel NO). Insofar as
bulk gas temperatures are low, and thermal NO formation is mainly restricted
to the heat release zone, combustion modifications which minimize peak flame
temperatures and maximize combustion product quench rates will control its
production. Fuel NO formation is determined by the conditions in the heat
release zone; fuel residence time under oxygen-deficient conditions being
the major factor in minimizing fuel NO production. Peak flame temperatures,
product quench rates and residence time in fuel-rich zones can be controlled
by burner design parameters (e.g., method of fuel and air injection, distri-
bution of axial and tangential velocity, burner geometry, etc.). Effective-
ness of these parameters as a means of controlling NO emissions from pul-
verized coal flames has been demonstrated in both pilot- and large-scale
systems. These investigations have led to the development of a burner which
can be applied to existing boiler designs to satisfy the technical NO emis-
JC
sion goals described above. This burner is based upon the concept of dis-
tributed mixing* which minimizes NO formation while satisfying other pro-
cess requirements (e.g., heat flux distribution, burnout, etc.).
The basic burner design features, the watertube simulators, and the
ancillary equipment used in this study have been described in detail else-
where (1). The results of earlier pilot-scale studies were used to develop
a prototypte burner whose performance has been demonstrated at a scale and
under conditions which are similar to actual watertube boilers. This paper
summarizes the results obtained to date with single burners at three scales
* Distributed Mixing Burner (DMB)
78
-------
with different coals, and discusses the implications associated with multi-
ple burner firing. In addition to NO control, the DMB has been used in a
series of investigations to assess the possibility of in-situ sulfur capture
using dry sorbents. The sorbents were added to the coal prior to pulveriza-
tion and it appears that the DMB has the potential to satisfy the process
requirements for both NO and SO control.
X X
79
-------
SECTION 2
LOW NO BURNER DEVELOPMENT
x
Pulverized coal combustion in burner stabilized, turbulent diffusion
flames is a complex phenomenon which cannot be described fully by a mathe-
matical model. Consequently, the development of the low NO DMB based
A
upon the original pilot-scale concept has progressed emperically. Two
different burner systems have been tested to date. These are:
A burner with a divided secondary throat which allows control
of axial and tangential velocity in two annular air streams.
Fuel is injected axially from an annulus surrounding a central
oil gun used for ignition.
A simple double concentric burner with fixed swirl vanes in both
the annular fuel and secondary air channels. Swirl is used in
the fuel injector to enhance flame stability. This burner design
was selected because of ease of construction in providing for
exact geometric and velocity similarity at three scales.
The velocity characteristics of the two burners are illustrated in Figure 1
for 50 x 10 Btu/hr burners operating at 4 percent overall excess air
as a function of the burner zone stoichiometric ratio. The influence of
several burner and operational parameters have been investigated with both
burners, and the composition of the three different coals which have been
used to date are given in Table 1.
The burner with the divided throat allows a wide variation in the
distribution of both axial and tangential velocity of the secondary air
stream and provides for considerable control over flame shape with the wide
angle burner exit (35 half angle). Three basic flow patterns were observed
in both cold and hot flow (see Figure 2). These were:
Type A with all axial flow except for an annular toroidal
recirculation zone which provided for flame stability.
80
-------
Type B with an axial toroidal recirculation zone.
Type C in which the flow attached to the combustion chamber
wall producing a wall jet.
The transition from A through C occurred as the swirl level in the outer
channel increased. Figure 2 also presents sketches of two general flame
types which were obtained with a narrower burner exit (25° half angle). The
shorter flame is much more compatible with existing boiler designs.
The low sufur Utah coal listed in Table 1 was used for the major
portion of the tests performed using the divided throat burner. The general
effects of overall excess air and burner zone stoichiometry were as expected.
NO decreased as the overall excess air and the burner zone stoichiometry was
Jv
decreased. Unlike more conventional burner designs, emissions were relatively
insensitive to variations in load for a given burner size. Various burner
exit geometries were tested and it was found that flame stability was improved
with a narrower burner exit whose length was equal to the throat diameter. A
schematic of the divided secondary throat burner is presented in Figure 3.
Figure 4 compares NO emissions for the divided secondary throat burner at
A
two scales. The burner was designed for constant air velocity and geometric
similarity. The larger burner yielded lower NO values. For the scaling
2v
principle used, the size of the fuel-rich burner zone would be proportional
to a linear burner dimension implying that the residence time in this region
is also proportional to burner scale for a given velocity. This provides one
simple explanation for the lower values obtained at larger scale. However,
other.parameters may well be important.
Further development studies have been carried out with the less com-
plex double concentric burner shown in Figure 5 at three design firing rates
of 12.5, 50 and 100 x 10 Btu/hr. The use of a single channel for the burner
throat provided for a more straightforward interpretation of the results, and
also allowed "exactly similar" burners to be produced at three scales. All
three burner scales employed similar distribution of tangential velocity and
similar locations for the outboard staged air injectors, and had a burner
exit half angle of 25°. The NO emission characteristics of this low N0x
DMB design have been evaluated and the influence of the following operational
81
-------
and design parameters assessed:
Burner zone stolchiometry. Both the overall burner zone stoi-
chiometry and the stoichiometry of the primary airflow.
Burner parameters including the degree of swirl in both the
fuel and secondary air channels. The location and velocity of
the outboard staged air injectors and the location of the fuel
injector.
Firing rate.
Fuel type.
Typical results obtained to date are presented below.
BURNER ZONE STOICHIOMETRY
The effect of burner zone stoichiometry for a range of excess air
levels is shown in Figure 6 for three different scales at comparable con-
ditions. The two large burners were tested in the large watertube simulator.
Burner zone stoichiometry is calculated based upon the total fuel flow and
the air being delivered with the primary and secondary streams (total air
minus outboard staged air). Comparison of the large and intermediate scale
burner test results (see Figure 6) show a moderate increase in NO with
2v
scale. However the results obtained with the smallest burner were highest
of all. It is possible that variations in combustor configuration may con-
tribute to this difference since the smallest burner was fired in a different
combustor than the two large burners but at a similar burner zone heat release
rate.
For a given fuel injector geometry variations in the amount of primary
air will not only change the local stoichiometry in the zone close to the
fuel injector it will also affect the primary velocity and the results pre-
sented in Figure 7 may be attributed to changes in either or both of these
properties. The data indicates that the primary stoichiometry for optimum
burner performance is dependent upon burner scale. This is probably
associated with the fact that the residence time in the fuel rich zone is
different for the three burner scales.
82
-------
BURNER PARAMETERS
The operation of the simple burner shown in Figure 5 involves a reduc-
tion of secondary airflow as the burner zone stoichiometry is reduced. This
reduces the tangential momentum of the secondary stream; thus, any balance
between primary and secondary momenta will be affected as the burner zone
stoichiometry is reduced at a given primary flow. It was found necessary
to provide the burner with fixed annular swirl vanes in the primary pipe to
ensure flame stability under staged conditions. This simple vane system
not only provides some tangential momentum, but also acts to divide the
coal stream into separate jets. Figure 8 compares NO emissions for various
primary swirl vane angles and it appears that a vane angle of 45 provides
the optimum performance. NO emissions were also found to be sensitive to
the level of swirl in the secondary stream. In general, NO emissions were
X
reduced as the swirl level increased, as can be seen in Figure 9.
The effect of both the location and velocity of the staged air jets
was investigated. Typical results are presented in Figure 10. Optimum per-
formance appears to be obtained with low velocity, and staged air injection
with the injectors located two burner-throat-diameters from the burner axis.
For the smallest burner tested the emissions and flame stability were found
to be sensitive to the location of the fuel injector. However, good per-
formance was obtained with the two larger burners when the fuel injector, was
located at the throat of the burner.
FUEL TYPE
It can be seen from the results presented in Figure 11 that the perfor-
mance of the low NO DMB is dependent upon fuel type for the 12 x 10 Btu/hr
J&
burner. There are considerable differences in the emission characteristics
for the same burner with different fuels, which is not necessarily associ-
ated with overall nitrogen content since under highly staged conditions
emissions from the three coals are comparable. The results obtained at
100 x 10 Btu/hr compared to that shown for 12 x 10 Btu/hr for the West
Virginia and Utah coals suggest that the effect of fuel type is also
associated with scale. For example at the smaller scale the Utah coal gives
a lower value of NO than the West Virginia coal. This characteristic is
x
reversed at the larger scale.
83
-------
SECTION 3
SULFUR CAPTURE
Control of oxides of sulfur can be achieved by one of three general methods:
1) the sulfur can be removed prior to combustion, either in a gasifier or
by physical or chemical cleaning; 2) sulfur may be absorbed by a solid during
the combustion process itself as in a fluidized bed combustor; and 3) the
sulfur oxides can be removed from the products of combustion. It is known
that certain compounds have the potential to produce sulfates, and certain
coals with high calcium and sodium ashes give high sulfate ash contents.
The use of sorbents such as limestone to remove sulfur oxides has been applied
unsuccessfully in several instances (2,3). Both small-scale studies and field
test have suggested that the retention efficiency of such sorbents can be
quite low and under these conditions the mass of sorbent required to achieve
the necessary level of S0_ control is excessive. However, sorbent utiliza-
tion efficiency has been shown to be sensitive to such parameters as tempera-
ture, stoichiometry, particle size, etc. A preliminary series of experiments
have been carried out with the low NO DMB with the intention of assessing
whether or not low NO operation enhances the possibility for in-situ sulfur
ji
capture , since the DMB provides a temperature/time stoichiometry history that
should maximize the sorbent utilization.
Preliminary tests were carried out using the Utah coal in a small
refractory tunnel furnace to assess the effect of sorbent dispersion, flame
temperature, the method of fuel/air contacting, and combustion air staging
upon sulfur capture. The effectiveness of the sorbent was assessed by
measuring the S0_ content of the combustion products. The two most signifi-
cant results of these small-scale studies were that the sulfur capture was
enhanced considerably when the sorbent was mixed with the coal prior to pul-
verization, and that increasing flame temperature enhanced rather than
inhibited the effectiveness of the sorbent. The influence of sorbent dis-
84
-------
persion can be readily understood since to achieve high sorbent efficiencies
it is necessary that the sorbent be uniformly mixed throughout the combustion
gases. The effect cf increasing temperature was somewhat surprising since
with limestone as the additive, it could be concluded that maximum flame
temperatures would produce dead-burning, and therefore, limit the activity
of the sorbent.
A series of preliminary exploratory investigations were also carried out
in the small watertube simulator with a 12 x 10 Btu/hr low NO DMB. The
x
objective of these preliminary studies was to assess the effect of sorbent
type, (calcium and sulfur compounds were used), sorbent ratio and coal type
upon sulfur capture, as well as to determine (1) whether there were signifi-
cant advantages of operating under low NO conditions, and (2) whether or not
X
sulfur capture influenced NO emissions. It is important to note that the DMB
design is such that under non-staging conditions the coal/air mixing profile
still provides for a slow mixing control core. Consequently even though
unstaged the DMB flame characteristics will be quite different from those
produced with typical commercial turbulent burners. The results presented in
Figures 12 and 13 for a low and high sulfur coal, respectively, are typical of
the results obtained in these preliminary experiments. Since exhaust S0_
concentration was used as the criterion for sulfur capture, care was taken in
the sampling procedure to prevent condensation in the sample lines. Also,
measurements were made of SO,, concentrations before and after tests with
additive to ensure that the sulfur dioxide levels were reproducible. With a
calcium-to-sulfur molar ratio of one, 50 percent of the sulfur was captured
by the additive for both the high and low sulfur coals, and with this parti-
cular calcium-to-sulfur ratio there appeared to be a significant effect of
operation under staged combustion conditions. In these preliminary investi-
gations no attempt was made to optimize the burner configuration for both NO
X
reduction and sulfur capture.
The results presented in Figure 14 show that there appears to be no
effect of sulfur capture upon NO emissions. From gas phase kinetic con-
siderations it might be expected that SO /NO would influence the overall
X X
NO emissions when the burner was operated staged. The results obtained to
X
date do not indicate that this is a significant factor.
85
-------
SECTION 4
DISCUSSION
The primary design objectives of the low NO DMB are:
x
- To maximize both the rate and total volatile evolution from
the coal particle. This does not only refer to nitrogen specie,
but also to fuel fragments since gas phase stoichiometry will
be dependent upon the volatile fuel fraction.
To provide an initial oxygen-deficient zone which minimizes NO
production, but then to add sufficient oxygen to the rich com-
bustion products to maximize the rate of decay of nitrogenous
specie (such as NH_, NO, HCN) to molecular nitrogen.
To provide optimum residence time and sufficient temperature
to maximize N_ production.
To maximize char residence time in the fuel-rich zone since
some nitrogen will remain in the solid with the potenital for
forming NO during burnout.
- To provide second stage air ensuring complete burnout.
- To produce an overall oxidizing envelope around the fuel-rich
core thereby minimizing the possibility for corrosion on the
combustor walls.
These objectives are obtained by providing for the optimum interaction
betwen the primary fuel jet and the swirl-stabilized recirculation zone,
together with delayed air addition from the outboard staged air injectors.
Figure 15 compares the NO levels achieved at three different scales for
a low NO DMB. Figure 16 presents the NO characteristics for the optimum
configuration at the same three scales. The differences between the various
scales may be associated with the variation in fuel residence time for
86
-------
different scales, or burner/chamber interaction affecting bulk gas tempera-
tures, and therefore, reaction zone temperatures.
In addition to the optimization of burner design parameters and the
development -of empirical scale criteria, the program to generalize low
emission coal burner technology has three other important objectives which
are:
1. Construct a fuel data base which will establish whether it is
necessary to vary the burner design as a function of coal
properties.
2. Determine the effects of burner/chamber interaction and
develop information for multiple burner configurations.
3. Provide direct comparison between prototype low NO burners
X
and commercially available burners.
4. To assess the effect of low NO operation on the emission of
X
other pollutants such as fine particulate matter and sulfur
oxides.
A comparison between the prototype and commercial burners will provide
information enabling the performance of the low NO DMB to be assessed when
X
operating under more practical conditions.
Three investigations are planned to assess the affect of burner/
chamber interactions on the performance of the low NO DMB. These include:
3C
Single-wall arrays to assess the effect of burner/burner
spacing, burner/sidewall spacing and the use of shared
staged air injectors.
Opposed-wall firing.
Off-axis firing. The minimum NO emissions achievable by the
X
DMB concept may provide excessive flame lengths. In order to
overcome flame impingement upon combustion chamber walls, four
burners will be tested firing off-axis in order that the inter-
action between the various long flames will prevent wall
impingement.
87
-------
The preliminary results involving the use of dry sorbents for SCL
capture are encouraging, and currently investigations are underway to pro-
vide confirmatory results by closing the sulfur balance after capturing the
solid. Figure 17 summarizes the results obtained to date with three additives,
two coals and various additive/sulfur ratios. It can be speculated that
the effectiveness of the DMB as a means of enhancing sulfur capture is
associated with the following:
Sorbent dispersion. Effective sorbent utilization requires that
its concentration must be high and that it is evenly distributed
in these regions where the sulfur specie are evolved from the
coal. In the tests described above the sorbent was mixed with
the coal before the pulverizer and therefore was evenly dispersed
in the form of small particles.
Sorbent activity. The sorbent is in contact with high sulfur
concentrations initially after calcination at the time it is
most reactive. Also the low NO DMB probably provides the
Ji
optimum time-temperature history for maximum sorbent activity
and minimal deadburning.
Peak flame temperatures occur under fuel rich conditions. The
sulfur and sorbent initially make contact under oxygen deficient
conditions and probably form a sulfide which is more stable at
higher temperatures than the sulfate. Consequently decomposition
of the "sulfur containing sorbent" will be minimized.
Reduced peak temperatures and low bulk gas temperatures. Low
NO operation requires that peak temperatures are minimized thus
ensuring maximum residual sorbent activity to capture sulfur in
the bulk gases after the heat release zone. The fact that the
sorbent is evenly dispersed will also tend to maximize sulfur
capture in this region.
These preliminary results suggest that the use of a flame zone sorbent plus
baghouse might well be a viable SO control technology for low sulfur Western
3C
coals or in addition to physical or chemical cleaning for high sulfur coals.
-------
SECTION 5
SUMMARY
The increased utilization of coal for power generation has the potential
for adverse environmental impact due to increased emissions of atmospheric
pollutants. One of the pollutants, NO can be controlled by modifying the
Ji
combustion process to maximize residence time under fuel rich conditions
in order to prevent the formation of NO from nitrogen contained in the coal.
The combustion of pulverized coal involves injection of the fuel and air
through a burner whose function is to ensure ignition stability and complete
combustion in the necessary volume. A burner has been developed which
controls the rate of pulverized coal air mixing that not only satisfies
normal process requirements but also minimizes the formation of nitrogen
oxide. This concept was originally evaluated at pilot scale but it has now
been demonstrated at a scale and in an environment which is typical of
waterwall boilers. The low NO DMB has been tested at three scales using
X
three bituminous coals and the results obtained to date can be summarized by:
Under optimum design conditions NO emissions are typically below
2£
200 ppm (dry at 0% 0-) for three burner scales (12, 50 and 100 x
106 Btu/hr).
' NO emissions are not independent of scale when a simple geometric
3C
scaling law is applied.
The emission characteristics of the low NO DMB appears to be
mildly dependent upon coal characteristics.
A preliminary series of investigations have been carried out to assess
the use of the DMB to control both NO and SO by adding dry sorbents to
3C X
the coal before the pulverizer. It was felt that the DMB would provide
ideal conditions to maximize sorbent efficiency. Limestone mixed with the
coal in calcium to sulfur molar ratios of 1, 2 and 3 gave 53, 73 and 88%
reduction in S0_ emissions respectively for a low sulfur coal. At low
calcium/sulfur ratios SO- reduction is enhanced by low NO operation.
^ . 3C
89
-------
It should be noted that the impact associated with the use of this
technology for SO control in field operating boilers requires a complete
X
assessment. The most immediate areas requiring process and economic analyses
are increased solids handling (both prior to and after combustion), mill
capacity, fly ash characteristics (size distribution and resistivity), slagging
and fouling in the boiler and the environmental problems associated with solids
disposal. Should the development and small scale studies continue to provide
encouraging results it is intended that the technology will be demonstrated
in an industrial boiler fitted with an EPA low NO DMB.
x
90
-------
REFERENCES
1. Gershman, R. E., M. P. Heap, and T. J. Tyson. Design and Scale Up
of Low Emission Burners for Industrial and Utility Boilers.
Proceedings of the Second Stationary Source Combustion Symposium,
Volume 5, Addendum EPA 600/7-77-073e, July, 1977.
2. Attig, R. C. and P. Sedor. Additive Injection for Sulfur Dioxide
Control - A Pilot Plant Study, National Air Pollution Control Adminis-
tration, Report No. 5460, March, 1970.
3. Gartrell, F. E. Full Scale Desulfurization of Stack Gas by Dry
Limestone Injection, Volume I, Tennessee Valley Authority, Chattanooga,
EPA 650/2-73-019A, (PB-228447), August, 1973.
91
-------
£ A») Divided Secondary Throat Design.
u
o
B.) Simple Double Concentric Burner Design
1.2 1.3
Figure 1. Air Velocity Characteristics for two 50 x 10 Btu/hr Coal Fired
Distributed Mixing Burners.
92
-------
Type A
Low Swirl
Type B
Medium Swirl
Type C
High Swirl
Flame Types
Low Swirl
S * 0.30
NO » 125 ppm
/\ "~
Medium Swirl
S * 0.44
B.) Low NO Flame Shapes
Figure 2. Flame Characteristics for DMB.
93
-------
Secondary
Air
Terti ary
Air
Furnace
Wall
Swirl Vanes
Coal &
Primary
Air
Secondary
Air
Retractable
Oil Nozzle
Carwrlc
Quarl
Tertiary
Air
Figure 3. Divided Secondary Throat Burner Design.
94
-------
Utah Coal 6.3% Stack
400
fr
a
CM
o
.300
I
a.
200
100
I I
Scaling:
Geometric/Constant Velocity
Primary Zone <* Burner Dimension
.°.
tres « Burner Scale
I
w 9 x 10° Btu/hr
O 50 x 106 Btu/hr
1.25
1.00
0.83
0.71
0.62
0.56
SR,
Figure A. Comparison of NO Emissions at Two Burner Scales
Divided Secondary Throat Design.
95
-------
c
60
H
CO
0)
Q
S-i
-------
[Utah Coal, SR1 = 0.25, SA1 = 45°, SA2 = 60°, Tert-Mid]
600
I
I
I
500
400
300
200
100
0
1*400
cvj
12 x 10° Btu/hr
SR2 = 1.0
200
100
T
I
I
50 x 10° Btu/hr
J L
400
300
200
100
n
- - I I 1 1 1
6
100 x 10° Btu/hr
^ *S*2
y>0.6
~~ O ^^^S*bV5
0.45
1 1 1 1 1
123456
1
, = 0.7
1
7
^~"
~
8
Excess 02 (% Dry)
Figure 6. Effect of Burner Zone Stoichiometry.
97
-------
500
400
300
200
100
0
500
1*400
[Utah Coal, SR2 = 0.65, SAj_ = 45°, Ter-M1d]
1 , !
12 x 105
0.350-
CM
300
200
100
500
400
300
200
100
0
I
50 x 10
1
100 x 10
1
1
SA2 = 45
1
= 0.15
I
1
SA2 = 60°
= 0.30
I
1
SA = 60
1
3456
Excess Q2 (% Dry)
8
Figure 7. Effect of Primary Stoichiometry.
98
-------
10
500
£400
o
*
CM
^200
100
12 x 10° Btu/hr
[Utah Coal, SRj = 0.25, SAg = 60°]
Tert-Far
O0.6
SA, -
.... SA:
60° _
45°
J
7 8
Excess 0,
50 x 10b Btu/hr
I
1 2
Dry)
Tert-Mid
SR2 =0.6
8
Figure 8. Effect of Primary Swirl.
-------
CM
o
8 e
500
400
300
200
100
I I
12 x 106 Btu/hr
I I
[Utah Coal, SRj = 0.25, SAj 45°, Tert-M1d]
500
SR2 = 0.6
SA2=45<
I I
8
400
300
200
100
I I
50 x 106 Btu/hr
I I
SR2 =0.7
60"
I I
I I
8
Excess 02 (% Dry)
Figure 9, Effect of Secondary .Swirl.
-------
[Utah Coal, SRj = 0.25, SR2 = 0.6, SAj = 45°, SA2 = 60°]
s;
o
A
CM
O
O
ft
Q.
Q.
X
o
500
400
300
200
100
0
II 1 1 1 1
12 x 105 Btu/hr
_
_ Mid/Low Vel
9 ^*?^ Tr Q
_ ^ Far/Low Vel _
_
1 1 1 I 1 1
L 2 3 4 5 6 7 £
I I
50 x 106 Btu/hr
Excess 02 (% Dry)
8
Figure 10. Effect of Tertiary Location and Velocity.
-------
700
r\> O
[SR, * 0.25, SA. = 45°, SA2 * 60°. Tert-Mid]
1 HI L
I I I
4.2% Excess 02
12 x lOpBtu/hr
J
I
I - L
Utah Coal
W.V. Coal .
W. KY. Coal
_l I
9 0.4 0.5 0.6 0.7 0.8 0.9
SRo
1.0 1.1 1.2
Figure 11. Fuels Performance.
-------
900
800
700
600
fc
Q
o
o
1400
a
CM
o
V)
300
200
100
0
Unstaged Utah
H Staged Utah
'O Unstaged Utah
+ Sorbent
Staged Utah + Sorbent
til /S^ /Zl
8
9 10
% Excess 0
Figure 12. The Effect of the Addition of Limestone to the Utah Coal Prior
to Pulverization on SO- Emissions (Ca/S - 1.0).
103
-------
3600
.3000
2500
o
«
o
1500
CVJ
o
in
1000
500
Urtstagad High Strifur Coat
Staged High Sulfur Coal
ISnitaged High Sulfur Coal * Sorbent
Staged High Sulfur Coal + Sorbent
Uasttged High Sulfur Coal After Sorber
8
10
Figure 13. The Effect of the Addition of Limestone to the Western Kentucky
Coal Prior to Pulverization on S0_ Emissions (Ca/S - 1.0).
104
-------
900
800
700
600
o
**
o
400
300
200
100
0
T
T
Unstaged Utah
Staged Utah
Unstaged Utah + Sorbent
Staged Utah + Sorbent
Staged Utah After Sorbent
4 5
% Excess 0,
8
10
Figure 14. The Effect of Sorbent on NO Emissions (Utah Coal Ca/S = 1)
105
-------
[utah Coal, SR, * 0.25, SAj = 45°, SA2 = 60°, M1d, 3 % Excess OJ
500 i 1 1 1 J
400
0^300
Q.
Q.
200
100
0
0.4
42 x 106 Btu/hr
100 x 10° Btu/hr
50 x 10° Btu/hr
0.6
SR,
0.8
Figure 15. Effect of Scale.
106
-------
700
[Utah Coal, 3% Excess 02]
600
500
Q 400
CM
O
300
ex
Q.
200
100
I I I
12 x 10° Btu/hr
I I I I
0.3 0.4 0.5 0.6 0.7
SR2
Figure 16. Performance at Optimum Configuration.
107
-------
100
CM
o
f
ce.
Mineral
Limestone +
Utah Low S.
Coal
^^i
53
^^^m
73
88
NaC03
Limestone +
High S. Coal
50
mmm^m
41
* NaHC03
80]
Pure
NaHCO,
52
o
123 1 24 2
Ca/S Ca/S N*/S Na/S
Figure 17. Summary of Sulfur Capture Results Using Various Sorbents
and Molar Ratios.
108
-------
TABLE I. COAL COMPOSITION
Proximate Analysis, %
(as received)
Moisture
Volatile
Ash
Fixed Carbon
Heating Value, Btu/lb
Ultimate Analysis, %
(DAP)
Carbon
Hydrogen
Nitrogen
Sulfur
Oxygen (by difference)
UTAH
6.39
38.89
7.4
47.32
12,340
79.45
6.03
1.44
0.74
12.34
W. VIRGINIA
1.29
31.01
13.76
53.94
12,500
83.92
5.66
1.55
2.06
6.81
W. KENTUCKY
3.68
38.19
17.90
40.23
10,640
77.57
5.7
1.63
3.8
11.3
-------
ALTERNATE FUELS AND LOW
NOX TANGENTIAL BURNER
DEVELOPMENT PROGRAM
BY:
Richard A. Brown
Acurex Corporation
Energy & Environmental Division
Mountain View, California 94042
111
-------
ABSTRACT
The EPA is continuing to explore control technology in the areas of alter-
nate fuels, waste fuels and control of NO emissions from tangentially fired
boilers in the government owned test facility located at Acurex Corporation.
Results from baseline and combustion control technology tests on coal oil
mixtures show that in general NO emissions of coal oil mixtures fall between
A
the NOY emissions of the parent oil and coal. Tests of NO control techniques
A A
including staging and a low NO burner design showed varying degrees of control
A.
depending on the particular fuel combination.
Baseline tests on refuse derived fuels (RDF) fired either with pulverized
coal or natural gas were determined. Emissions assessments made include NO ,
A
CO, particulate loading and size distribution, twelve trace metals and a cursory
search.for FCB and POMs. NO emissions decreased as the percent RDF increased
A
even though the available fuel nitrogen increased. Particulate loadings from
the RDF were concentrated in the less than 1 ]i size fraction.
An extensive research program has been initiated to develop a low NOY
A
burner design for coal fired tangential boilers. The initial tasks consist
of firing the pilot scale facility on natural gas, natural gas doped with
Nitrogen species, char, char plus natural gas, and coal to ascertain the
relative importance of the various flame regions, coal volatiles and nitrogen
evolution. In addition, a water model of the pilot scale unit is being con-
structed to ascertain the mixing patterns of the pilot scale unit as compared
to the full scale unit.
112
-------
SECTION 1
INTRODUCTION
The EPA multiburner multifuel test facility is continuing to explore
control technology for alternate fuels and waste fuels. In addition, an
extensive research and development program has recently been initiated to
control NOX emissions from tangentially fired utility boilers. This paper
will: (1) review the data from tests on coal-oil mixtures and cofiring of
refuse-derived fuels, and (2) present the test plans and objectives for the
NO tangential burner development program.
The EPA experimental multiburner, furnace facility (Figure 1) was
developed to study NO control technology problems associated with large - scale
A
utility and industrial boilers. Details on this facility design have been dis-
cussed in other papers (References 1 and 2).
The furnace fires from 293kw-thermal to 880kw-thermal (1 to 3 x 10 Btu/hr)
depending on the fuel and heat release per unit volume being simulated. The
facility may either be front-wall fired using one to five variable swirl block
burners,, or it may be corner-fired using four to eight tangentially fired burners
patterned after Combustion Engineering's design. A horizontal extension confi-
guration (Figure 2) with a single large variable swirl block burner was used to
conduct the coal-oil mixture tests. Radiant section cooling was simulated by
placing water cooling coils on the inside circumference of the refractory tunnel.
A convective section quenched the flue gases at the appropriate residence time.
The standard gaseous emissions (NO, CO, C0?, 02» S0_) were continuously
monitored throughout the test program.' Particulates were obtained using a high
volume EPA method 5 stack sampler; trace metals and organics sampling was
113
-------
obtained using the Source Assessment Sampling System (SASS) (Reference 3).
Coal-Oil Mixture Emissions
In many industrial and utility boilers, conversion to coal or partial coal
could be accelerated if only minor modification to the boiler were required.
Therefore, there is a growing interest in firing a mixture of coal and fuel oil
in existing units.
It is now necessary to determine if current control technology for coal
combustion is applicable to coal-oil systems or if further work is needed to
ensure that pollution standards can be met.
The coal-oil mixtures examined were prepared from parent fuels which
represent a broad range of fuel compositions. The compositions of the parent
oils, coals, and coal-oil mixtures are listed in Table I. The coals, pulverized
to 70 percent 200 mesh, were blended with the fuel oils and a suspension additive
supplied by Carbonoyl Company that constituted 3.75 percent by weight of the
mixtures. The mixtures were heated, thoroughly agitated by a variable speed
rotary pump, and then delivered to the burner nozzle through heat traced lines.
The two nozzle types used were an air atomized combustion swirl Delavan
Nozzle and an air atomized Sonicore Nozzle. Both of these nozzles underwent
erosion during the tests. The Delavan 440 hardened stainless steel nozzle
lasted approximately 3 to 4 hours on a coal-oil mixture of 30 percent coal
until significant erosion produced an unstable flame. The Sonicore nozzle
with a stellite tip lasted about 8 to 12 hours until performance deteriorated.
Other operational problems included the rapid deterioration of pump seals and
the gradual buildup and plugging of lines in the delivery system.
Figure 3 illustrates the results of the baseline emission tests. The NO
levels for the Chevron-based fuels fall in an intermediate range between the
parent fuels, while the data for the Amerada base mixtures are closer to the
parent oil. This difference may be caused by sulfur in the fuel, atomization
characteristics, and how the particular oil and coal volatilize.
114
-------
At this point there is insufficient data to determine which mechanism is
causing this effect; however, not all fuel combinations behave in the same
manner.
In this study, established combustion control technology for pulverized
fuel was applied to the coal-oil mixture, including staging the combustion air,
burner air distribution or "low NO burner" configurations and combined staging
, A
and low NO burner design. Only the combined low NO,, and staging tests will be
reported here. The low NOX burner involves axial injection of tertiary air at
the periphery of the burner diameter.
The curves in Figure 4 compare the results of applying burner air distri-
bution plus staging to straight burner air distribution. The purpose of this
comparision was to evaluate the mixture responses with an enriched flame zone.
For the western Kentucky/Amerada mixture, further enriching the flame zone re-
sulted in higher NO levels. This result may further illustrate the response of
the Amerada high sulfur parent oil to fuel-rich conditions. The Montana/Chevron
mixture results validate that each mixture responded favorably to the air distri-
bution, but that the composition of each mixture leads to a unique emission
curve.
The following conclusions can be drawn:
NO emissions from coal-oil combustion are affected by the composition
of the parent fuels which make up the mixture
Present conventional control technology used for pulverized fuel com-
bustion reduces NO emissions produced by coal-oil mixture combustion
Coal-oil mixture flames do not react consistently to some forms of
control technology, implying that the chemical properties of the fuel
and the physical processes which take place during combustion are both
important
Coal-oil mixture combustion is different from either pulverized coal
or residual oil combustion. Therefore, predictions of NO emissions
based on emission levels of the parent fuels often will not be valid
115
-------
In order to analyze and understand coal-oil mixture combustion, better
understanding of the combustion processes of each parent fuel must be obtained.
Also, the combustion of coal-oil mixture must be examined to determine if it is
merely a combination of the two individual processes or instead is a completely
different physical and chemical phenomena.
Future work should examine the role of fuel-bound nitrogen utilizing flue
gas recirculation, nitrogen evolution studies of the parent oils, and the effect
each fuel exerts on the other, such as shielding or physical separation in the
droplets.
116
-------
SECTION 2
REFUSE-DERIVED FUEL TESTING
There is condiderable heating value (4000 to 7000 Btu/lbm) associated with
municipal solid waste. If this resource could be used in steam boilers rather
than lost by incineration, a significant energy resource would be tapped.
Research studies have included using heat recovery incinerators, spreader
type stokers and suspension firing in large electric utility boilers. The EPA
has supported experiments in cofiring the refuse-derived fuel (RDF) in full
scale boilers in St. Louis, Missouri, Ames, Iowa, and Columbus, Ohio (References
4 through 7).
Although these full scale experiments are providing useful data, problems
associated with the many varieties of RDF need to be studied. Because the refuse
comes from local municipalities, there can be significant variations in the com-
bustion and environmental characteristics of the fuel from season to season or
from locale to locale.
There is little published data on the emissions from RDF when cofired with
other fuels. Kilgroe (Reference 8) reported that the St. Louis demonstration
site produced a moderate increase in chloride emissions but that the RDF did not
significantly affect the SO or NO emissions. Little information is available
i X
on the amount of trace metals and organics or the nature of the particulates.
The work performed here provides the initial data base to answer environmental
questions on four RDF's.
117
-------
The refuse-derived fuel tests were conducted in the main firebox in the
tangentially fired mode. First, a feed system (Figure 5) was designed to control
and measure from 10 to 60 Ibs/hr of varied refuse-derived materials. RDF is
delivered to the upper part of two diagonally opposed corner-fired burners.
Natural gas or coal is also delivered to these burners and to the other two
burners. The RDF feed system consists of a rotating drum hopper which deposits
the material on a conveyor belt. The conveyor delivers the material to a vertical
downcomer, where it is pushed through by a blast of air. Additional air at the
junction of the vertical downcomer and horizontal feed tube conveys the RDF into
the furnace through a horizontal water-cooled feed tube. The RDF feedrate is
controlled by a variable-speed drive on the feed belt; the drum is maintained at
constant optimum speed to keep the feed belt full.
The delivery tube is sized to prevent blockage while minimizing the trans-
port air. This sizing is critical for a small-scale facility where the minimum
pipe size is governed by the maximum particle size and the minimum conveyance
air needed to keep the material suspended. The received RDF material was
shredded from a nominal 2 to 4 inches down to 1 to 2 inches using a conventional
garden shredder to reduce the feed tube diameter and transport air flow to an
acceptable level.
The test program determined the gaseious, particulate trace metal and organic
emissions of refuse-derived fuel from San Diego, California; Richmond, California;
the Americology Facility in Milwaukee, Wisconsin; and Ames, Iowa.
All of these materials had gone through metals and glass separation and
a primary shredding. Table II shows the composition and heating value for each
fuel type.
Figure 6 shows the effect of NO versus excess air for the four fuels at
20 percent RDF and 80 percent natural gas. Although the NO levels are not par-
ticularly high, there was a definite difference between the fuels. The NO also
increases with both excess air and increases in RDF. Also, when the percent
RDF is increased and cofired with coal, the NO levels decrease (Figure 7) while
total fuel nitrogen increases. This is possibly the result of enriched fuel
118
-------
jets at the coal-refuse injection guns. Except at very low excess air (5 percent)
CO levels were always less than 100 ppm.
Results from the particulate, trace metal, and organics sampling also
provide some interesting preliminary information of RDF emissions. Table III
shows the results of particulate concentrations in the various size cuts for
four RDF materials cofired with natural gas. Although the total particulate
quantity was quite low in all cases, the majority of particles were smaller
than 1 U and collected only on the filter. Although the particulate may be
rather friable and break up in the sampling equipment during collection, it may
eventually end up in the respirable size range.
Table IV shows particulate loading in the same size cuts for two levels of
Richmond RDF cofired with coal and for coal alone. With the substitution of RDF,
the total grain loadings decreased with increasing RDF. However, in both cases,
adding RDF increased the percent of material in the. 1 u size cut over coal alone.
Thus, it appears that adding RDF may increase the grain loading in size cuts
less than 1 U. This result could produce problems for flyash collection equip-
nent. Percent combustibles in the particulate were generally less than 2 percent
except when the excess air levels were 10 percent or less. This result also
corresponds to generally low CO (< 100 ppm) and unburned hydrocarbon levels.
Table V lists the total trace metals in micrograms/Btu found in the parti-
culates and the condensable vapor for: 1) coal only, 2) coal plus 10 percent
BDF, and 3) gas plus 10 percent RDF. Increases in trace metal concentrations
varied among the three tests. In the coal only test, the lead concentration was
exceptionally high. In addition, no correlation was found as the percent of RDF
was increased.
It is difficult to draw any conclusions from this trace metal data.
Several factors may be contributing to the data variability:
The RDF material is nonhomogeneous and will vary from minute to minute,
hour to hour, and season to season
Metals from the furnace and sampling system could contribute to the
trace metal loadings
119
-------
Hold up in the convective section
Analytical error
These factors indicate the need for a broader data base to draw meaningful
conclusions on trace metal concentration when cofiring RDF with other fuels.
This data will contribute to that base, but a larger sample of data is needed
to statistically determine real trace metal effects.
In addition to trace metals, a limited search for organics in terms of
(PNA or PCBfs) was undertaken. A portion of the participate and the XAD-2
organic section resin of the SASS train were analyzed by liquid chromatography
according to EPA Level 1 procedures (Reference 9). Of the material divided into
the standard seven cuts, only cuts two and three were expected to contain PNA
and PCB. Therefore, these two fractions were combined for a single GC/MS
analysis. Five out of nine tests where organic samples were taken contained no
detectable compounds. Test conditions, and the PNA found in the remaining
samples, are listed in Table VII. No PCB was found in any of the test samples.
Little organic material, combustible CO, and unburned hydrocarbons were
found in the particulate and gaseous streams. It has been reported (Reference
8) that significant quantities of unburned material have been found in full-scale
tests. These pilot scale tests have higher combustion efficiency over full-scale
tests possibly because of the additional shredding and/or hot refractory walls
providing an improved ignition source.
In summary, up to 30 percent RDF may be cofired in a subscale test facility
without experiencing a reduction in combustion efficiency. Additional studies
are necessary to determine how this technology can be implemented in a full-scale
facility using the same degree of efficiency but a higher RDF percentage.
Furthermore, the flame's heat transfer characteristics must be clearly defined
to determine the effect on the boiler stearnside or to design a boiler specifically
for cofiring RDF. Finally, more data is needed to statistically determine the
trace metal and organic makeup of RDF cofired boilers emissions.
120
-------
SECTION 3
LOW NO TANGENTIAL BURNER DEVELOPMENT
A
INTRODUCTION
The RDF work is being followed by development of a low NO tangential
X
burner for utility boilers. The following sections describe the background and
program plans for this work.
Tangential coal-fired boilers produce approximately 10 percent of all
stationary source NO emissions and consume, in Btu's, approximately 9 percent
A.
of all fuel used in stationary sources. Reducing tangentially coal-fired boiler
NO emissions is necessary to maintain ambient air quality in the United States.
A.
The need to address the long-term capabilities of the corner-fired boiler is
becoming more apparent as recent developments in low NO burners for front wall-
A.
fired boilers have reached the 125 ppm range (Reference 10).
Therefore, the EPA has contracted with Acurex Corporation to conduct a
ith program to develop a low NOX burner concept for 1
boilers. The two principle goals of this program are to:
31-month program to develop a low NOX burner concept for tangentially fired
develop a better understanding of processes controlling NO formation
A
during combustion of pulverized coal in tangentially fired furnaces
develop low (50 to 100 ppm) NO combustion concepts for retrofitting
A.
or new designs of tangentially fired boilers
Figure 8 shows the flow patterns typically present in the tangentially
fired system. Air and fuel are distributed in a nonswirling slow mix character
through registers located in vertical orientation at each of the four corners.
Generally there are three to five fuel register levels. Fuel ignition is pro-
121
-------
vided by impingement of hot laterally adjacent streams and large-scale internal
recirculation of combusted gases. Thus, the ignition and combustion of each
corner jet serves to ignite and promote burning in the downstream adjacent jet.
A large-scale vortex is formed by the tangential nature of the air and fuel
injection. This vortex motion promotes mixing of fuel-rich regions, internal
recirculation flow patterns, and the slow-mix nature of the jets should inherently
make the tangentially fired system low in NO production. The pilot-scale
A.
facility models most of the significant parameters of the tangential system. For
example, the fuel and air are distributed in a manner similar to the full-scale
system. The bulk residence time and heat transfer rates are also maintained.
However, the residence time between adjacent jets is condiderably different
than in the full-scale system, and fuel is introduced only at one or two.
levels.
Although there is a three-dimensional aspect to the tangential boiler flow
patterns, our primary concern will be the flows encountered at a single elevation.
Figure 9 and the following paragraphs define three regions in a tangentially
fired boiler:
Near burner-early mixing
Intermediate jet flame
Recirculating fireball
Near Burner-Early Mixing Region
In this region, flow is predominantly an axial jet that entrains air and
small amounts of flue gas. The region consists of approximately the first five
flow path jet diameters where significant coal devolatilization occurs, including
the flame's ignition region. It is believed that a high peak NO is formed early
in this region.
Intermediate Zone
In this region, consisting of five to perhaps 30 fuel jet diameters, the
remaining air is mixed with the fuel, the majority of the fuel is burned, and
considerable flue products are entrained. In this region, one side of the flame
interacts with the fireball, and the opposite side interacts with the boiler
122
-------
wall. Locally rich combustion zones which have the potential to reduce entrained
burnt gas NO by flame processing can exist here.
A
Fireball
The fireball consists of the recirculating zone in the firebox center. It
is believed that this zone consists mostly of flue gas products and regions where
char is burned out. However, it is not known at this time if any additional
HO is formed here, or if NO formed early is reduced.
The proposed research program focuses on the physical and chemical pro-
cesses associated with NO formation in the near, intermediate and fireball
'' A
zones.
APPROACH
The proposed approach is deiplayed in Table VII.
Task A Tangentially Fired System Definition and Evaluation
The primary goal of Task A is to define the phenomena which control the
formation and reduction of both fuel and thermal NO in a classical tangentially
fired boiler. This task will isolate the critical chemical and physical processes
(e.g., early mixing, flame shielding, devolatization, etc.) and establish the
design parameters effect on these processes. Task A will also investigate the
uniqueness of tangential firing and attempt to definitively show why such a
system generally has lower NO emissions than wall-fired units. The majority of
A
this work will be done in the existing pilot-scale facility. In each region the
importance of the volatiles and char burnout will be investigated by doping char,
natural gas flames, and natural gas plus char with nitrogen and sulfur compounds.
We will also focus on hot sampling for NO and flame-flame interaction
studies. It is possible that NO formed in one flame can be reduced by passing
through an adjacent flame. Figure 10 shows parameters that will be investigated,
including the intersection point of the two flames, X, and the intersection angle
over a range of fuel richness in each flame.
Therefore, these subtasks will help focus subsequent tests on the those
chemical/physical processes offering the greatest potential for NOX reduction.
123
-------
At the completion of Task A, sufficient understanding and information on
tangential system NO., formation processes will be available to permit the design,
construction, and test of specialized low NO burner/firebox subcomponent
X.
hardware.
Task B Optimization of Near Burner and Intermediate Zone NOV Levels
A
The primary objective of this task is to follow through on the guidance
provided by Task A; and to design, construct, and test (in subscale), low NO
A
subcomponent concepts of near and intermediate zones. Sufficient flexibility
will be incorporated into the hardware to define optimal low NO conditions for
A
both the near burner and intermediate field zones. Parameters optimized during
these tests will be air and flue gas distribution, and mixing and temperature
history, or local cooling rates. The B tasks will utilize a new versatile
burner which fires into the horizontal furnace extension. In these initial
tasks, the near burner early mixing zone oxygen, flue gas availability and
temperature/cooling rate history will be optimized for low NO . Optimal and
A
off-optimal near burner configurations will then be tested with intermediate
zone air and flue gas distribution, and wall cooling. For these tests, air
and flue gas introduction ports, and wall cooling panels will be incorporated
into the horizontal firebox appendage.
Optimal near and intermediate zone low NOV concepts will emerge from the
A
Task B tests. These concepts will be applied in the pilot-scale tangentially
fired tests carried out in Task C.
Task C Optimization of Coal-Fired Tangential System Low NO Concepts
_ A
The primary objective of this task is to define and demonstrate in pilot
scale, low NOX tangentially fired systems for several coal types. This task
will adapt the low NO subcomponent concepts developed in Task B into the
tangential firebox and optimize the complete system for low NOY. The system
A
will be optimized for each coal type tested.
The optimized low NOX tangential system concept will be demonstrated on a
single characteristic coal. Sufficient parameter variation will be built into
124
-------
the hardware to optimize the system for the characteristic coal used in this
task or any other commonly used coal. Then the low NOV concept will be opti-
x
mlzed for several additional coal types.
Fireball Simulation Study
Throughout the 31-month program, a background study will determine how
to best simulate a tangentially fired system in subscale. The study will
identify and quantify the following parameters important to modeling:
Mixing rates
Temperature/time history
Fuel type and size
Burner firebox geometry
Previous hot and cold flow studies on tangential systems, such as those
by Juniper (Reference 11) will be studied.
To assess the overall flow simulation of the pilot scale versus the full
scale, a water model of each system is being constructed. Flow patterns will
be observed using a dye injection technique and recorded by still and motion
pictures. To assess the mixing patterns for the new concept in the tangential
mode, models will also be made of the low NOY burner design concepts.
A.
CONCLUSION
Completion of all proposed program tasks will result in a pilot-scale
demonstration of a low NO tangentially fired system for several coal types.
A
Following the program's completion, research will be conducted to either:
1) demonstrate the low N0_ concept in larger scale with possibly a greater
variety of fuels, or 2) carry out additional subscale tests. This will eventually
lead to an optimal low NO full-scale tangentially fired system capable of firing
A
several coal types.
125
-------
SECTION 4
SUMMARY
The paper presented three facets of EPA's program on emission character-
ization and control of alternate waste and conventional fuels. First the
facility simulated the combustion characteristics of a package boiler to test
emission characteristics of coal oil mixtures. Different fuel combinations of
a coal-oil mixture responded in a unique manner to conventional control techno-
logy. In addition, many operational problems (pumping characteristics, pump
seal wear, and nozzle erosion) were discovered.
Tests were also conducted to determine the feasibility of subscale
evaluation of cofiring refuse-derived fuels. A unique feed system was developed
to deliver a variety of EDFs from 10 to 60 Ibs/hr. However, it was found that
the test furnace was even more efficient in achieving complete combustion than
full-scale units.
Tests showed slight increases in NO and SO emissions over natural gas but
demonstrated a reduction in these levels when cofired with pulverized coal.
When cofired with natural gas, the particualte was found concentrated in less
than a 1 y range. The trace metal analysis showed no conclusive trends. Few
PNA organics emissions and no PCB materials were found.
The program plan to develop a low NOX tangential burner was presented.
This plan includes tests to: 1) determine formation and destruction of NO in
current designs; 2) optimize early and intermediate zone environmental conditions,
and 3) optimize low NOX concepts in the pilot-scale facility. A water modeling
126
-------
study will also be employed to study the mixing patterns in the pilot-scale and
full-scale systems.
127
-------
SECTION 5
REFERENCES
1. Brown, R. A. et al. Pilot Scale Investigation of Combustion Modification
Techniques for N<>x Control in Industrial and Utility Boilers. EPA-600/2-
76-152b. Proceedings of the Stationary Source Combustion Symposium,
Volume II, June 1976.
2. Brown, R. A. et al, Investigation of Staging Parameters for NO Control
A
in Both Wall and Tangentially Coal-Fired Boilers. EPA-600/7-77-073C.
Proceedings of the Second Stationary Source Combustion Symposium,
Volume III, Stationary Engine, Industrial Process Combustion Systems,
and Advanced Processes, July 1977.
3. Blake, D. Source Assessment System Design and Development, EPA 600/7-
78-018, August 1977.
4. Vaughan, D. A. and Associates. Report of First Year Research on
Environmental Effects of Utilizing Solid Waste as a Supplementary
Powerplant Fuel, Battelle Columbus Ohio Laboratories, EPA Research
Grant R-804008, June 1975.
5. Nydick S. E. and Hurley, J. R. Study Program to Investigate Use of
Solid Waste as a Supplementary Fuel in Industrial Boilers, Thermo-
Electron Corporation. EPA Contract No. 68-03-3005, January 1976.
6. Riley, B. T. Preliminary Assessment of the Feasibility of Utilizing
Densified Refuse Derived fuel (DRDF) as a Supplementary Fuel for
Stoker Fired Boilers, published report to EPA, 1975.
7. Vaughan, D. A., Krause, H. H., Hunt, J. F., Cover, P. W., Dickson,
J. D., and Boyd, W. K. Environmental Effects of Utilizing Solid
Waste as a Supplementary Powerplant Fuel. Seventh Quarterly Progress
Report, EPA Research Grant 804008-02-1, 1975.
128
-------
8. Kilgroe, J.D., Shannon, L.J., Gorinan, P. Environmental Studies on
the St. Louis Union Electirc Refuse Firing Demonstration.
9. IERL/RTP Procedures Manual: Level 1, Environmental Assessment,
Second Edition. June 1976.
10. New Coal Burner May Reduce Nitrogen Dioxide Emission 80 to 85 percent.
EPA/Environmental News Release, Office of Public Awareness (A-107),
Washington, D.C., September 22, 1978
11. Juniper, L. A. Flame Measurements in a Brown Coal Fired Furnace,
Combustion Institute European Symposium, Academic Press, 1973.
129
-------
Figure 1. Photograph of experimental multiburner furnace.
-------
IFRF Burner
,
i"i/i;J''*!
i'i i" i/A'/1!
"" 1r.il'5!
'!
Radiant
Section
I /v y\r
I
I
i]
Corrective Main Rrebox
Section
Figure 2. Package Boiler Configuration
131
-------
1400
Chevron Oil
Q.
1200
1000
800
600
400
200
1
10 20 30 40
% Excess Air
O Chevron No. 6
Q Virginia/Chevron 30%
& Montana/Chevron 30%
0 Montana Coal
o Virginia Coal
1400
1200
1000
800
600
400
200
Amerada Oil
10 20 30 40
% Excess Air
O Amerada No. 6
& Montana/Amerada
o W. Kty/Amerada
O W. Kentucky Coal
a Montana Coal
Figure 3. COM Baseline Emissions
132
-------
30% Virginia Coal/
Chevron Oil
30% W. Kty Coal/
Amerada Oil
30% Montana Coal/
Chevron Oil
,
I
800
600-
1200
0
.55 .65 .75 .85
D
O
800
^600
SW
&200
^8001
260*
o
Z
.55 .65 .75 .85
SRiA
.55 .65 .75 .85
SRi = 0.95 Staging + Low NOx Burner
SRt ^ 1.20 Low NOx Burner Only
Figure 4. Combustion Modification Results
133
-------
Tangential Burner
Main Firebox
Figure 5. RDF Feed System and Firebox
134
-------
20% Refuse/Natural Gas
200
CM
o
Q.
Q.
100
% NDMMF
1.26
o Richmond
A Americology
i° San Diego
o
Ames
10 20
EA%
30
Figure 6. Baseline NOX - Emissions - RDF & Gas
135
-------
Richmond Refuse/Pittsburg Coal
CM
O
Q.
600
500
400
300
200
100
0 10 20 30
% RDF (Heat Input)
40
Figure 7. Effect of Percent RDF Coal & RDF
136
-------
CO
Top View /air
bulk of combustion
air
Side View
air and fuel registers
Figure 8. Tangentially Fired System Flow Patterns
-------
Figure 9. Tangential Burner Flame Regions
138
-------
Short flame
Intersection distance
Figure 10. Intersecting Flame
139
-------
TABLE IA. COAL ANALYSIS
Proximate (Z WO^-
^ Coal
Moisture
Volatiles
Fixed Carbon
Ash
Rank
Ultimate (Z Wt)
Carbon
Hydrogen
Nitrogen
Oxygen
Sulfur
Ash
Heat of Combustion Btu/lb
Montana
21.23
35.16
34.27
9.34
Sub-bit. C.
53.26
3.35
0.87
11.16
0.78
9.34
8,972
Virginia
0.31
31.9
51.4
16.5
High-Vol. A
71.11
4.46
1.68
. 4.24
2.02
16.5
14,079
W. Kentucky
5.0
36.55
50.98
7.47
High-Vol. B
69.79
4.79
1.34
8.65
2.95
7.47
12,349
TABLE IB. OIL AND COM ANALYSIS
- -Jlixture 30Z (Wt)
Analysis """ ^JJoal
Carbon
Hydrogen
Nitrogen
Oxygen
Sulfur
Ash
Moisture
Heat of Combustion Btu/lb
API Gravity
Flashpoint COC°F
Viscosity, SSO at 100°F
V. Kentucky/
Amerada
80.23
9.00
0.63
3.92
2.44
2.26
1.52
17,600
Montana/
Amerada
75.27
8.57
0.49
4.67
1.79
2.82
6.39
16,600
Montana/
Chevron
75.88
8.37
. 0.83
4.77
0.89
2.87
6.37
15,500
Virginia/
Chevron
81.23
8.70
1.07
2.73
1.26
4.95
0.09
17,000
Amerada
Hess 16
84.71
10.75
0.36
1.93
2.22
0.93
19,867
15.3
. 204.0
2,500
Chevron
#6
85.57
10.52
0.81
2.08
0.93
18,292
12.3
182.0
4,900.0
*
140
-------
TABLE II. RDF FUEL ANALYSES
Ultimate Analysis*
Carbon %
Hydrogen %
Oxygen %
Nitrogen %
Sulfur %
Ash %
Moisture %
(as received)
Chlorine %
Heating Value
Btu/lb
Fuel Type
Pittsburg
No. 8 Coal
75.23
5.15
8.12
1.49
2.51
7.50
0.93
0.14
13,545
Richmond
Refuse
42.60
6.26
37.90
0.83
0.16
12.25
23.8
.46
7696
Ames
Refuse
40.49
6.01
30.04
0.73
0.35
22.38
15.2
.43
7831
Americology
Refuse
40.29
5.88
25.20
0.91
0.17
27.55
24.4
.72
7164
San Diego
Refuse
38.01
5.64
17.40
0.69
0.21
38.05
26.3
.79
7146
Dry Basis
141
-------
TABLE III. EFFECT OF RDF TYPE ON PARTICULATE SIZE DISTRIBUTION
Fuel
20% Ames
+ Nat Gas
20% Richmond
t- Nat Gas
20% Americo-
logy
f- Nat Gas
20% San Diego
f- Nat Gas
Filter
Qty (gr/ft3)
%
0.039
(69)
0.032
(91)
0.041
(86)
0.062
(80)
>10y
Qty (gr/ft3)
%
0.011
(19)
0.0004
(1.)
0.002
(5)
0.007
(9)
>3y
Qty (gr/ft3)
%
0.003
(6)
0.0004
(1)
0.002
(5.)
0.002
(3)
>iy
Qty (gr/ft3)
%
0.004
(6)
0.0024
(7)
0.002
(5)
0.006
(8)
142
-------
TABLE IV. EFFECT OF COAL & RDF CONCENTRATION ON PARTICIPATE SIZE DISTRIBUTION
m Richmond
Li RDF
ICoal
m Richmond
[ RDF
iCoal
bal Only
Filter
Qty (gr/ft3)
%
.026
(9.1)
.044
(7.6)
.021
(2.1)
>10y
Qty (gr/ft3)
%
.134
(47.3)
.269
(46.1)
.539
(56.3)
>3y
Qty (gr/ft3)
%
.107
(38.0)
.226
(38.8)
.344
(35.9)
>iy
3
Qty (gr/ft )
%
.016
(5.5)
.044
(7.6)
.055
(5.7)
143
-------
TABLE V. TRACE METAL CONCENTRATION FOR COAL VS 10% RDF + COAL VS 10% RDF + GAS
(yg/Btu)
ELEMENT
Cu
Zn
Mn
Pb
Cd
Be
Ti
Sb
Sn
Hg
As
Coal Only
Test #40*
1.9581
0.5294
0.1526
17.5319
0.0091
<0.0176
<1.7540
<0.0020
0.1300
<0.0015
<0.0881
10% RDF + Coal
Test #38a
<0.3319
0.8227
<0.2693
0.3091
0.0048
0.0013
<0.0587
<0.0090
<0.0913
<0.0173
<0.0323
10% RDF + Gas
Test #llba
0.3402
0.4468
0.0209
1.4996
0.0062
<0.0034
<0.0277
0.0333
<3.5033
<0.0009
<0.0184
20% excess air
144
-------
TABLE VI. ORGANICS FOUND
Test Condition
Gas Cofire
10% RDF
20% EA
Ames Fuel
Gas Cofire
10% RDF
20% EA
Richmond Fuel
Gas Cofire
10% RD?
20% EA
Americology Fi*el
Coal Cofire
10% RDF
20% EA
Organic
Fluoranthene
Pyrene
Phenanthrene
Fluoroanthene
Pyrene
Diphenyl Ether
Blphenyl Phenylether
Phenanthrene
Pyrene
Phenanthrene
Amount (ug/108 Btu)
10
332
64
160
576
3395
1697
59
104
98
145
-------
TABLE VII. PHASE IV LOW NO TANGENTIALLY FIRED SYSTEM
TASK A
Tangentially fired system definition and evaluation
chemical/physical processes defined
early mixing evaluated
flame processing of NO in intermediate and far zone evaluated
TASK B
Optimization of tangentially fired near and intermediate zone NO
early mixing, cooling, and entrainment of flue gas optimized
intermediate zone mixing, cooling, and entrainment of flue
gas optimized
TASK C
Optimization of low NOX tangentially fired systems for several coals
optimal burners in firebox with optimal fuel/air feeding of the
fireball
146
-------
POLLUTANT FORMATION DURING
FIXED-BED AND SUSPENSION COAL COMBUSTION
Bv:
D. W. Pershing
B. D. Beckstrom
P. L. Case
G. P. Starley
Department of Chemical Enaineerina
University of Utah
Salt Lake City, Utah 84112
147
-------
ABSTRACT
This paper summarizes the overall scope and the progress made during
the first five months of a grant to study the formation of pollutant species,
particularly nitrogen and sulfur oxides, during the fixed-bed and suspension
combustion of coal. The work will consider conditions typical of both
pulverized and stoker-fired coal systems with the major emphasis on spreader/
stoker fired boilers and furnaces. Specifically the program will consider:
1) the evolution and oxidation of fuel nitrogen and sulfur; 2) the retention
of sulfur oxides by the ash and/or solid chemical sorbents in suspension-
and fixed-bed burning; and, 3) the effectiveness of distributed air addition
for nitrogen oxide control. In addition, the study will attempt to quantify
the combustion process in a stoker environment and consider possible detri-
mental effects of control technology on boiler operation.
The approach is primarily experimental, utilizing a controlled mixing
history furnace and fixed-bed reactor to investigate suspension and fixed-
bed combustion, respectively. In addition, a model spreader/stoker system
will be fabricated, characterized, and used for evaluating advanced combustion
control concepts. The suspension furnace has now been designed and most
of the refractory casting is complete. The flow-control, system for this
furnace has beer, specified and essentially all of the required instrumentation
has been procured. Design of the fixed-bed furnace has just been initiated.
148
-------
ACKNOWLEDGMENTS
This research was supported by the U. S. Environmental Protection
Agency under Grant No. R-805899-01. The considerable help and advice of
G. Blair Martin, the EPA Project Officer, Robert D. Giammar, Battelle,
and Michael P. Heap, Vice President, Energy and Enviromental Systems is
gratefully acknowledged. In addition, thanks are due to Mrs. Colleen
Anderson for her help in completing the secretarial duties associated with
this project.
149
-------
PURPOSE AND SCOPE
This paper describes the progress of an investigation to study the
formation of air pollutant species, particularly NO, N02> S02, and S03> during
the combustion of coal particles with a view toward providing technology for
the design of high-efficiency, low-emission boilers and furnaces. The work
will consider conditions typical of both pulverized-coal and stoker-fired
systems, but the emphasis will be on the latter because they represent a
major source which has been largely overlooked previously.
At present stoker-fired boilers are significant in terms of both coal
consumption and environmental impact. In 1974 almost 20 percent of the
coal consumed in the United States was burned in stoker systems. Stoker-
fired, water-tube boilers are the single largest source of parti oil ate
emissions and the fourth largest source of SOX emissions, because neither
is in general controlled. They account for 13 percent of the NOX emissions
from all coal systems and their environmental impact may be even more
significant than indicated by the mass emissions, because stoker systems
are often located in congested metropolitan areas. Further, the energy
and environmental importance of stoker firing seems certain to grow with
the increase in industrial coal utilization that is projected. It is
important, therefore, that further knowledge be obtained on the formation
and control of nitrogen and sulfur oxides under conditions typical of stoker
firing.
In particular, this program will consider the following major research
areas:
1. The evolution and oxidation of fuel
nitrogen and sulfur;
2. The retention of sulfur oxides by ash
and/or solid-chemical sorbents in both
suspension- and fixed-bed burning;
3. The effectiveness of distributed air
addition for NOx control in both
pulverized- and stoker-fired coal
systems.
-------
In addition, the study will attempt to quantify the combustion process in
a stoker environment and consider possible detrimental effects of control
technology on boiler operation.
The approach is primarily experimental, utilizing a controlled mixing
history furnace and a fixed-bed reactor to investigate suspension and thick-
bed combustion, respectively. In addition, a model spreader-stoker system
will ultimately be fabricated, characterized, and used for evaluating advanced
combustion control concepts. The next section of the paper describes the
details of the approach. The third and fourth sections then discuss the
progress made during the first five months of the program.
151
-------
APPROACH
OVERVIEW
This study is scheduled for three years and is divided into five
separate tasks:
Task 1: Preparation of a detailed program plan.
Task 2: An experimental investigation of
suspension burning of bituminous coal.
Task 3: An experimental investigation of
fixed-bed burning of bituminous coal.
Task 4: An experimental and analytical optimi-
zation of combustion modification
technology using a model stoker system.
Task 5: Experimental studies to extend the
control-technology results to other
fuels.
Each task is discussed in detail in the following paragraphs. Tasks 1, 2, and
3 have been initiated. Task 4 will start in the second year of the program
and Task 5 will begin in the third year.
EXPERIMENTAL PLAN PREPARATION
Work is currently underway on a detailed program plan for the first
eighteen months of the study. It will include a description of the initial
experiments to be conducted in each task and detailed designs for the fixed-
bed and suspension experimental furnaces. To ensure the practicality
of the proposed experimentation, conversations will be held with stoker
manufacturers and users to define in detail the problems associated with
industrial stoker usage, the current trends, and future utilization of stoker
technology. This will include site visits to at least one fabricator, one
industrial user, and other researchers in the field.
152
-------
SUSPENSION BURNING
The purpose of the second task is to investigate the formation
of pollutant emissions during the suspension-burning phase of the
combustion process. These studies are intended to be relevant to both
pulverized-coal combustion and the suspension-burning phase of a spreader/
stoker system. Initial studies will focus on the relationship between
the combustion parameters and the evolution and subsequent oxidation
of nitrogen and sulfur species. The role of fuel-bound nitrogen in the
formation of NOX will be established, both for small particles that are
essentially completely burned in the suspension phase and for large
particles that only undergo partial oxidation. The possibility of gas-
phase sulfur capture will also be explored in this task.
Both the furnace and the flow-control system for the suspension-phase
studies have been designed as described in the next section of this paper.
Furnace fabrication is now more than 50 percent complete and essentially all
of the components for the flow-control system have been procured.
FIXED-BED COMBUSTION
The purpose of the third task is to utilize a simple, relatively inexpen-
sive experimental test facility to study the combustion of coal in a fixed-
bed environment. The experimental system will be designed to provide a
Lagrangian simulation of the time/temperature/environmental history seen by
a small section of i ^ stoker bed as it passes through an actual furnace.
Both thick- and thin-jed systems will be considered. The use of a simple,
fixed-bed coal furnace will allow evaluation of a large number of different
conditions, fuels, and additive materials without the complications associated
with moving stoker grates.
The design of the experimental system has not yet been resolved; however,
it is likely that the fixed-bed furnace will be a refractory chamber with a
removable bed section and auxiliary gas heating capability. Thermocouples and
sampling probes will be positioned both within the bed and above the bed.
Provisions will be made for independent control of both underfire and overfire
air. Present plans are to design the bed for firing rates up to 400,000
Btu/ft2/hr at a coal-feed rate of 20 Ibs/hr.
153
-------
The first set of experimental studies will focus on the relation
between fixed-bed combustion parameters and overall pollutant emissions.
The parameters of interest include:
. Bed composition
. Bed thickness
. Bed and gas-phase temperature
. Free-stream composition (stoichiometry)
. Particle size
. Bed-heating rate and amount
. Bed-residence time
. Amount and location of overfire air
Later testing will focus on the evolution and subsequent oxidation of fuel-
nitrogen and sulfur species both within and above the bed. Measurements of
nitrogen and sulfur intermediates (e.g. NHg.HCNjHgS, etc.) will be attempted.
Solid samples will be removed from the bed after various residence times and
analyzed for ultimate composition. Other studies will investigate the sulfur-
retention capabilities of the coal ash itself and the potential of solid
sorbent materials. Of particular interest will be the distribution of the
sorbent materials throughout the bed.
MODEL STOKER STUDIES
The purpose of the fourth task is to apply the pollutant reduction
concepts developed in Tasks II and III to an actual stoker
environment. An experimental furnace will be fabricated to directly simulate
both spreaker/stoker and fixed-bed systems. Initial experiments will be
conducted to define the interaction between the suspension burning and the
fixed-bed combustion since each of these will have been considered separately
in the previous tasks. Later testing will be directed toward defining and
optimizing new combustion control technology.
ALTERNATE FUELS
The purpose of the final four-month task is to study the impact of coal
composition on combustion control technology. It is anticipated that the fuels
to be studied here will include at least one desulfurized coal, a lignite, and
perhaps a markedly different alternate fuel such as wood chips.
154
-------
SUSPENSION BURNING
The suspension furnace was designed to meet the following criteria:
1. Capable of simulating the environment seen by a particle
in the suspension burning phase of spreader/stoker
combustion or in the burner zone of a PF furnace.
2. Independent wall-temperature control with high-
temperature capabi1i ty.
3. Adequate probing and visual access.
4. Variable residence time.
5. Well defined fluid dynamics.
The final furnace design utilizes three types of interchangeable furnace
sections. In each case the innercombustion chamber Js 6 inches in diameter and
the outer steel shell is 28-inches square. A cylindrical innerchamber was
chosen (in preference to a square one) because it is more amenable to analytical
modelling at some later date. A square outside surface was selected because it
reduced the cost of fabrication and provided a more suitable base for mounting
probes, thermocouples, etc.
The walls of the main furnace sections consist of an outer steel shell,
5 inches of 1900°F insulating block, 4 inches of 2500°F insulating refractory,
and 2 inches of 3400°F high-temperature castable refractory. Each of the mid-
sections contain 5 2-inch diameter ports for second-stage air injection and/or
insertion of species and temperature probes.
As presently conceived, the furnace can be operated in two modes:
The self-sustaining flame mode and the tubular-reactor, control-mixing history
mode. Figure 1 illustrates the first mode in which the pulverized coal and
at least part of the combustion air (or artificial atmosphere) enter the com-
bustion chamber at the top via a premixed burner. Current plans are to use the
premixed burner developed by Howard and Essenhigh (1). A modification of this
burner system has proved suitable for studying NOY formation in premixed, one-
/%
dimensional flows (2). This mode of operation will be used for characterizing the
nitrogen and sulfur evolution for particle sizes and size distributions smaller
than 50 percent through 100 mesh (i.e., pulverized coal and the fines in stoker-
coal system). In this mode, the flue gases will exit at the bottom of the
furnace through the horizontal section.
155
-------
In order to simulate the temperature and environment of the suspension-
phase burning in a spreader/stoker system, the second mode of operation will
be utilized (Fig. 2). In this case, a high-intensity gas burner will be
attached to the horizontal extension at the furnace bottom. The large, stoker-
sized coal particles will be fed from the top of the furnace and allowed to
fall downward into the stream of hot combustion products flowing vertically
upward. In this case, the premixed coal burner will be replaced with a
large, multihole coal injector and the majority of the combustion air will be
supplied from the bottom with the gas flame. In an actual spreader/stoker
system, the suspension phase combustion occurs with reduced oxygen concentra-
tions (less than 10 percent) and with both nitrogen and sulfur species already
present from the fixed-bed combustion. To simulate this, S0« and/or ammonia
will be added to the gas flame to produce the concentrations of SOg. S03, NO,
and NOp that would normally be present in the combustion gases leaving the
stoker bed. Both acetylene and methane will be used as fuels for the gas
burner to investigate potential XN/hydrocarbon interactions. In this mode,
the main flue gases will exit at the top of the furnace through the horizontal
section as shown (Fig. 2).
As shown in Fig. 3, the side walls have provision for auxiliary heating
(or cooling) by firing natural gas or passing cooling air through outer
channels. These channels are 2 inches in width and cover 50 percent of the
circumference. This will allow partial independent control of the.radient
heat transfer to the particles and more complete control of wall temperature.
The auxiliary firing sections are completely interchangeable so any portion
of the furnace can be heated (or cooled).
The flow-control system for the suspension furnace has been designed
and the required components procured. The flow system was selected to
provide precise metering of the streams entering the main combustion chamber
and adequate control of the other auxiliary gas and air streams. Flow
metering and control of the following subsystems are included:
156
-------
1. Main-chamber air.
2. Side-burner air.
3. Auxiliary-burner air.
4. Artificial oxidants
5. Side-burner gas.
6. Auxiliary-burner gas.
The combustion air is supplied by a 100 SCFM, 100 psi rotary vane compressor.
Primary, secondary, and tertiary air are metered with 600 mm high acuracy
Brooks rotameters. The second-stage air is metered with a large, 10-inch
Dwyer rotameter. All other air and gas flows are metered with 10-inch
Dwyer rotameters. The flow system also includes an artificial oxidant supply,
so that the combustion air can be completely replaced with an atmosphere
composed of argon, carbon dioxide, and oxygen in varying proportions. In
each case, the pure gases are supplied from high-pressure cylinders and
metered with rotameters.
157
-------
ANALYTICAL MEASUREMENTS
The furnace exhaust gases will be continuously monitored for NO, N02,
CO, C02, and 02. Using the instrumentation listed below:
Gases Instrument
NO, N02 Thermoelectron, Model 10AR,
chemiluminescent analyzer
CO, C02 Anarad, Model AR-600 NDIR
02 Beckman, Model 755, paramagnetic
oxygen analyzer
The other stable, gas-phase pollutants and intermediate species of
interest will be measured on a batch bases. These include hydrogen sulfide
(H2S), sulfur dioxide, and trioxide (S02, S03), carbonyl sulfide (COS),
carbon disulfide (CS2), ammonia (NH3), and hydrogen cyanide (HCN). S03 will
be measured using the standard ASTM condensation method. Measurement of
NH3 will be attempted using the Axworthy chemiluminescent technique. The
other species (S02, H2S, COS, CS2> and HCN) will be analyzed with a Hewlett-
Packard, Model 5830A, dual column, temperature programmable gas chromatograph.
The sulfur compounds will be sensed with a flame photometric detector. The
nitrogen containing species (e.g., HCN) will be measured with the rubidium-bead
nitrogen/phosphorus flame detector. Current plans are to use an acetone
washed 1/8-inch teflon tube, 4 feet long, packed with acetone washed Porapak
QS for separation of the sulfur compounds fcfter deSouza (3)).
158
-------
REFERENCES
1. Howard, J. B. and R. H. Essenhigh. Pyrolysis of Coal Particles in
Pulverized Fuel Flames. Ind. 'Eng. Chem., Process Design and Development,
6 (1), 1967.
2. Wendt, J. 0. L., J. W. Lee, and D. W. Pershing. Pollutant Control Through
Staged Combustion of Pulverized Coal. FE-1817-4, U. S. Department of
Energy, Washington, D.C., 1978. 158pp.
3. de Souza, T. L. C., D. C. Lane, and S. P. Bhatia. Analysis of Sulfur-
Containing Gases by Gas-Solid Chromatography on a Specially Treated
Porapak QS Column Packing. Analytical Chemistry, 47 (3), 1975.
159
-------
JL-L
en
o
TT
O O O O O
O O O O O
O O O O O
TT
c < x
(B fO fl)
_I-J Q.
OO N*
C fl>
-J Q.
Figure 1. Suspension Furnace - Self-sustaining Flame Mode
-------
\\\\\
SNNNX
\\\\\\NS\\N\N\\\\\\
\v\\\\\. \.\\vs.\\\\\
N\\\\S\\\\\\\\\\\\s
W W\v vv\ vv vvWVX
\\\\\\\\\\
\\\\\\\\\\\\K\\N\\'
\\\\\\\\S\\\K\\\\\'
Figure 2. Suspension Furnace - Tubular Reactor Mode
-------
IN3
\VAN\N\\\\\\\\\\
\\N\\\\\\\\\\\\\\
Figure 3. Detail of Auxiliary Heating (Cooling) System
-------
ADVANCED COMBUSTION CONCEPTS
FOR
LOW BTU GAS COMBUSTION
By:
B. A. Folsom
C. W. Courtney
T. L. Corley
W. D. Clark
Energy and Environmental Research Corporation
Santa Ana, California 92705
163
-------
ABSTRACT
The low Btu gas combined cycle power plant is an alternative to the
direct coal-fired steam cycle with the potential for low sulfur emissions and
high overall efficiency. However, low Btu gas contains ammonia which could
lead to high nitrogen oxide emissions.
This paper discusses the development of low NO combustor concepts for
X
this application. The thermodynamic performance of several alternative low
Btu gas-fired combined cycles is investigated and the combustor operating
conditions necessary to optimize thermodynamics performance are identified.
A kinetic mechanism describing fuel nitrogen conversion to NO is used
to analyze idealized combustors with these operating conditions and several
potential low NO combustor concepts are investigated.
X,
Synthetic low Btu gases with varying compositions were fired in three
atmospheric pressure flame reactors: diffusion flame reactor, premixed flat
flame reactor, and premixed catalytic reactor. NO emissions were found to
2t
be sensitive to the concentrations of ammonia and hydrocarbon gases in the
low Btu gas. Lowest NO emisssions were produced by the diffusion flame
Ji
reactor operating fuel-lean and the catalytic reactor operating fuel-rich.
164
-------
ACKNOWLEDGEMENTS
This paper is based upon work conducted under Contract No. 68-02-2196
with the-Environmental Protection Agency. The authors wish to express their
appreciation to Mr. G. B. Martin of the Environmental Protection Agency and
Messrs. J. Johnsen and J. Miltko of Energy and Environmental Research
Corporation for their assistance in various portions of the work.
165
-------
SECTION 1
INTRODUCTION
This paper discusses the development of low nitric oxide (NO) emission
combustors for low Btu gas (LEG)-fired combined cycle power systems. The
rising demand for electrical power coupled with the limited availability of
petroleum makes the construction of many alternate fuel-based power plants
by 1985-1990 very desirable. Since these plants will commence operation in
the future when environmental restrictions will probably be much more
restrictive than at present, the environmental design goals should be to
minimize emissions rather than to meet current New Source Performance Standards.
Several advanced coal-based power cycles are currently being developed.,for
this application (1,2,3,4). The integrated gasifier LBG-fired combined gas
turbine-steam turbine cycle is one of these alternatives. The primary advan-
tage of this cycle is the potential for low emission of sulfur products without
the economic and efficiency penalties associated with stack gas sulfur
removal (5,6). Since the coal gasification process operates fuel-rich, the
sulfur in the coal is converted mainly to hydrogen sulfide (H_S) in the low
Btu offgas which can be removed (potentially) more easily than sulfur dioxide
(S0_). Energy losses associated with the gasification and cleanup process can
amount to as much as 30 percent of the coal's heating value. However, if the
clean LBG is fired in a gas turbine combustor as part of a combined cycle power
plant the gains in overall power plant performance, when compared to conventional
steam cycles, may more than offset the losses in the gasification and cleanup
processes.
The minimization of NO emissions from LBG-fired combined cycles is an
important and at present unsolved problem. In the gasification process, some
of the nitrogen in the coal is converted into ammonia (NH~) in the LBG. Under
typical gas turbine combustor operating conditions a large fraction of the NH~
may be converted to NO resulting in significant emission.
166
-------
The concentration of NKL in the LEG depends upon gasifier design and
operating parameters and may be as high as 0.38 percent (6). The gas cleanup
system, which functions primarily to remove sulfur products and particulates
from the LEG, may remove a portion of NH.,, but the remainder will enter the
gas turbine combustor and may be oxidized to nitrogen oxides (NO ). A concen-
Jv
tration of 0.38 percent NH, in the LEG would produce approximately 1370 ng/J
6
(3.2 Ib N02/10 Btu) if all NH- is converted to N02. At present there are no
New Source Performance Standards (NSPS) for LEG fired combined cycle power
systems. However, it is reasonable to expect that when NSPS are promulgated
they will be at least as stringent as current NSPS for other power systems
such as gas turbines or gas fired steam generators. The current NSPS for
gaseous fossil fuel-fired steam generators with greater than 73.3 MW (250 x 10
Btu/hr) heat input is 86 ng/J (0.2 Ib N02/10 Btu) (7). For a combined cycle
firing an LEG with 0.38 percent NH- to meet this emission level, the overall
conversion of NH to NO would have to be less than 6.3 percent. It is well-
known that the conversion of fuel nitrogen compounds (such as NH.) to NO is
sensitive to combustor design and operating parameters. A recent analytical
study of combustion modification techniques applied to LBG-fired combined
cycle combustors has demonstrated the potential for significant reductions in
N0x emissions (8).
The development of low NO emission combustor concepts for LBG-fIred
combined cycle systems is the objective of a current Environmental Protection
Agency program (Contract No. 68-02-2196, G. Blair Martin, Project Officer).
This paper presents some preliminary analytical and experimental results from
this program. The following section discusses combustor design requirements
based on maximizing the thermodynamic performance of LBG-fired combined cycles.
The subsequent section outlines a kinetic mechanism for NO formation from NH_
and presents several low NO combustor concepts based upon kinetic analysis.
The next section presents the results of combustion experiments burning
synthetic low Btu gases with varying compositions in several types of com-
bustors. Discussion and conclusions follow.
167
-------
SECTION 2
DEFINITION OF COMBUSTOR REQUIREMENTS
Figure 1 is a simplified schematic diagram of an LBG-fired gas
turbine-steam turbine combined cycle power plant with integrated gasifier.
Only the major heat, work and mass flow paths have been shown for clarity.
The system consists of a gas turbine topping cycle with exhaust heat trans-
ferred to a steam turbine bottoming cycle through a waste heat recovery
boiler. Low Btu gas is produced in a gasifier supplied with coal, compressed
air from the gas turbine air compressor* and steam from the bottoming cycle.
This "integration" with other cycle components significantly reduces the
energy losses attributable to the gasification process.
"Raw" LB6 exits the gasifier at temperatures as high as 1370 K depend-
ing upon gasifier design and operation containing three pollutant precursors:
Sulfur compounds which can be oxidized to SO in the combustor,
a
9 Farticulates including carbon, tars and ash which may damage
turbine blades and may be emitted to the atmosphere, and
Ammonia which may be oxidized to NO in the combustion process.
4m
Raw LBG is processed through a gas cleanup system to reduce the concen-
trations of these materials prior to combustion. Several gas cleanup sys-
tems have been developed or are under development for this application (9) .
Hot gas cleanup systems process the LBG without significantly reducing its
temperature and remove a majority of sulfur products and particulates. Cold
gas cleanup systems require cooling the LBG to near ambient temperature (400 K)
and remove a majority of the sulfur products and particulates. They may also
remove a substantial portion of the ammonia. The sensible heat removed from
the LBG may be used for a variety of purposes, but it effectively bypasses
the gas turbine topping cycle and contributes to reduced efficiency. The
thermodynamic trade-offs in hot and cold gas cleanup processes have recently
168
-------
been investigated (8,10). Depending upon gasifier design and operating
parameters, the sensible heat in LEG offgas can amount to as much as 20 per-
cent of the total heat release and the losses due to intercooling the LEG
with a cold gas cleanup system can be substantial.
From a combustor design point of view, the choice of a hot gas cleanup
system as opposed to a cold system could well have two important effects:
Higher adiabatic flame temperature
Higher NH_ concentration
The sensible heat in the LEG contributes directly to the adiabatic flame
temperature. For typical LEG compositions, cooling the gas by 400 K will
decrease the adiabatic flame temperature at stoichiometric conditions by about
200 K, thus reducing the potential for thermal NO formation. If a low
A
temperature cleanup system is employed, a substantial portion of the NH_ may
be removed. Residual NH_ concentrations of 100 to 400 ppm have been predicted
for full-scale systems (3,4). These concentrations correspond to approximately
34 to 145 ng/J (0.08 to 0.34 Ib N02/106 Btu) for full conversion. While the
lower prediction meets current NSPS (neglecting the contribution from thermal
NO ), further NO control may be necessary to meet future standards.
x x
The term low Btu gas does not refer to a specific gas composition, but
rather to a family of fuels, produced by reforming coal with air and steam. The
3 3
heating value of LEG may range from 16,000 to 41,000 J/m (80 to 200 Btu/ft )
and the primary combustible specie are CO and H-. The ratio of CO to H con-
centrations ranges from 0.5 to 2. Hydrocarbon fuel gases (normally CH.) may
also be present in qua».-ities up to 10 percent, and nitrogen is the primary
diluent comprising 35 to 55 percent of the fuel gas. Other diluents include
CO., and HO. Trace amounts of H_S, COS and NH may also be present, since
£ £. f, j
these specie are not entirely removed by the gas cleanup system. The relation-
ship between LEG composition and combustor performance is currently under
investigation as part of this program and Section 4 discusses the performance
of combustor firing LBGs with a wide range of compositions.
Combustor operating conditions are dictated by thennodynamic and materials
considerations. High turbine inlet temperature (TIT) is desirable for improved
thermodynamic performance. Current materials and blade cooling technology
169
-------
limit TIT for stationary gas turbines to 1273 K to 1473 K. The combustor
pressure producing an optimum balance between cycle efficiency and specific
work output is usually about 12 atmospheres. The corresponding compressed
air temperature for high efficiency compressors is about 638 K. Improvements
in high temperature materials and turbine blade cooling are expected to
increase allowable TIT in the future. One proposed LBG-fired combined cycle
was designed around a 1644 K TIT (4).
If the exhaust heat from an optimized gas turbine cycle precisely matched
the heat requirements of an optimized steam turbine cycle there would be no
need for cycle variations. However, the gas turbine exhaust temperature is
usually too low for an optimized steam turbine cycle. The overall efficiency
of the combined cycle can be Improved by modifying the cycle arrangement to
reduce the mismatch between the energy content of the gas turbine exhaust and
energy requirements of the steam turbine cycle. This can be accomplished by
incorporating any of the cycle modifications listed below.
Direct fired turbine exhaust heater
High excess air supercharged boiler gas turbine combustor
Reheat gas turbine cycle
The thermodynamic performance of LBG-fired combined cycles incorporating
these cycle modifications was analyzed as part of this program and the analy-
tical details are discussed in references 8, 10, 11, 12 and 13. The results
are summarized in Table 1 which lists the overall efficiency and combustor
operating conditions for these cycles with two TITs: 1366 K typical of state-
of-the-art stationary gas turbines and 1700 R which may be achievable in the
future utilizing advanced turbine blade cooling techniques. These results
were calculated subject to the component efficiencies and other assumptions
listed in Reference 10. Since these assumptions may not necessarily be ful-
filled in future generation combined cycle systems, the efficiencies and
operating conditions listed should be considered as approximate.
The combustors listed in Table 1 can be grouped into two general types:
primary combustor where the oxidant is nonvitiated compressed air, and
secondary combustors where the oxidant is the partially vitiated combustion
products from the primary combustor. The primary combustors will operate at
170
-------
several atmospheres pressure with warm air (from the heat of compression)
and will generate a hot gas stream for the turbine. They will operate either
adiabatically or with only a small amount of heat (less than 20 percent) trans-
ferred to the bottoming cycle. The overall stoichiometry will be very fuel-
lean, greater than 250 percent theoretical air. The NO control goals for
X
primary combustors are to minimize conversion of NH_ in the LEG to NO (fuel
J Vt
NO) and minimize formation of NO from N (thermal NO ).
* x L x
The secondary combustors will operate at pressures less than the primary
combustors and may produce either a hot gas stream for a low pressure turbine
or a warm gas stream for a waste heat recovery boiler. The overall stoichi-
ometry of these combustors will be fuel-lean but richer than the primary com-
bustors. The amount of heat released in the secondary combustors will be
small compared to the primary combustors. The NO control goals for secondary
ji
combustors are the same as for primary combustors except that additional NO
control might be achieved by reducing some of the NO formed in the primary
X
combustor to N_.
171
-------
SECTION 3
NITRIC OXIDE FORMATION IN LBG-FIRED COMBUSTORS
Tyson et al. (14) have assembled a kinetic reaction set based upon the
work of many investigators which describes the formation of NO and N£ in com-
bustible mixtures of CO, H_ and CH,. Those reactions involving nitrogenous
species are presented in Table 2. Ammonia breakdown occurs via hydrogen
abstraction allowing nitric oxide formation by the oxidation of any of the
nitrogenous fragments. Under fuel-lean conditions, Reaction 4 predominates
and the major portion of the NH« is converted to NO. Under fuel-rich condi-
tions NO might still be formed via Reaction 4 if all the oxygen has not been
consumed, or by Reaction 6. Hydrogen cyanide has been observed by several
investigators, and for high temperature flames Morley (15) has shown that
regardless of their nature, nitrogen compounds are quantatively converted to
HCN in the reaction zone of premixed rich (less than 80% theoretical air)
hydrocarbon flames. Hydrogen cyanide formation from NO could be possible by
reactions such as 7 and 8. Hydrogen cyanide can also be produced from mole-
cular .nitrogen via reaction with hydrocarbon radicals (16). The subsequent
oxidation of HCN allows the interchange of XN specie since NH is one of the
specie formed as HCN is oxidized via the intermediary NCO. Once NH is pre-
sent then the reverse of Reactions 1, 2 and 3 allows the formation of ammonia.
Nitrogen formation can occur via the reverse of Reaction 5, or by reaction
with NO and other nitrogenous species. Although Reaction 13 is included in
Table 3, there is some question as to whether a reaction of this type can
take place. The reactions shown in Table 3 illustrate how ammonia in LEG
could be converted to either N_, NO, HCN or back to NH, in rich combustion
products. Since N. is thermodynamically favored at certain reactant stoichio-
metries and temperatures, the objective of the combustor designer is to
provide the most favorable temperature and stoichiometry histories to maxi-
mize N_ production and thus minimize total bound nitrogen species (EXN).
172
-------
Decoupling the .practical considerations of the fuel/air contacting pro-
cess, it would appear that several combustor configurations might minimize
NO emissions from LEG containing NH . Four such concepts were analyzed by
X j
applying the kinetic mechanism described above to idealized combustor models
consisting of various combinations of well-stirred and plug flow modules.
The simplest case considered was the rich/lean series staged combustor
shown in Figure 2. The ignition reactor provides the feedback of heat and
radical species for flame stabilization and the plug flow reactor, which
operates at an optimum stoichiometry, allows sufficient time for the reactions
listed in Table 2 to proceed towards equilibrium. Secondary air is added
uniformly over a short time period to burn out the remaining combustibles and
to dilute the combustion products to the desired TIT.
This combustor configuration was analyzed subject to the fuel composition
and combustor operating conditions listed in Reference 3. This is an adiabatic
gas generator operating at 12.0 atmospheres and 1589 K TIT fired with a LEG
containing 600 ppm NH3. Figure 3 shows the adiabatic equilibrium EXN concentra-
tion as a function of reactant stoichiometry. ZXN is strongly dependent upon
reactant stoichiometry, and the minimum occurs at 65 percent theoretical air.
On the rich side of this minimum the XN species are primarily NH_ and HCN,
whereas on the lean side the XN species are almost totally NO.
Figure 4 shows the results of kinetic calculations for a 1.0 msec ignition
zone followed by a plug flow section operating at 65 percent theoretical air.
Although £(XN) decays toward the equilibrium value of 2.9 ppm, the rate of
decay decreases with increasing residence time. For a finite primary zone
residence time,. both the decay rate and the equilibrium value are important.
Consequently, the stoichiometry which produces minimum Z(XN) at equilibrium
may not necessarily be the optimum stoichiometry for a combustor with limited
residence time. However calculations similar to those discussed above were
conducted for a range of stoichiometries and the minimum ZXN for all residence
times was found to occur with 65% theoretical air.
A secondary stage following a 65 percent theoretical air 50 msec primary
stage was analyzed for two air injection patterns.
Uniform air addition over 4.0 msec
Uniform air addition over 60.0 msec
173
-------
The results shown, in Figure 5 indicate that rapid air addition gives lower
emissions. Slow air addition allows the mixture to spend considerable time
near stoichiometric conditions where temperatures are high. Thus, a large
amount of thermal NO is formed. The decay after the peak is dilution. Full
conversion of the initial 18 ppm of NH would result in 8 ppm when fully
diluted. The rapid air addition case produces 14 ppm which can be attributed
to full conversion of the NH. emerging the secondary stage plus 10 ppm of
thermal NO . The total NO emissions are equivalent to 14 ng/J (0.033 Ib
f X X
N0_/10 Btu). Thus this combustor concept has the potential for very low NO
*~ X
emissions.
Three additional low NO combustor concepts are shown in Figure 6. These
2C
have been analyzed similar to the rich/lean series staged concept discussed
above and the results demonstrate the potential for low NO emissions and
X
certain improvements over the rich/lean series staged concept.
If the reduction of fuel XN species to N. is kinetically limited, the
rate of reduction can be accelerated by increasing the reactant temperature
(17). Concept A in Figure 6 is one way to achieve higher reactant tempera-
tures in an overall adiabatic system. Heat is removed from the hot rich
combustion products and transferred to the cooler reactants prior to ignition.
The results of kinetic calculations applied to this concept show that the
optimum stoichiometry for maximum XN decay shifts to richer conditions as the
amount of feedback heat transfer increases. For comparable overall residence
times, a combustor with 10 percent of the heat release fed back to the
reactants produces NO emissions 16 percent lower than the simple rich/lean
3E
series staged concept.
Kinetic calculations indicate that the decrease in XN decay rate as
residence time increases in a rich plug flow reactor is partially due to
depletion of the radical pool from the high superequilibrium concentrations
present immediately after ignition (14). By staging the air addition as in
Figure 6-B, the radical pool is periodically replenished leading to a more
rapid overall XN decay rate. The stoichiometry, air addition rate, and
residence times can be optimized for each rich stage to achieve a minimum
final XN concentration.
174
-------
Ammonia conversion to N. might well be maximized by contacting certain
species that would not normally occur in a sequential process by dividing the
total combustor into two parallel streams with different stoichiometries, and
then mixing the products of these two combustors prior to secondary burnout.
This concept referred to as parallel staging is shown in Figure 6-C.
Although in the limit, kinetic considerations dictate the ultimate con-
version of fuel nitrogen to N or NO, conversions in practical systems are
dictated by the realities of the fuel/air mixing process. The reactants can
be premixed to ensure that reaction takes place at a single stoichiometry or
the fuel and air may be supplied to the combustor separately, thus allowing
reaction to take place over a range of stoichiometries which will depend upon
the intensity of the mixing process. The following section describes the
experimental results obtained from firing LEG in several types of combustion
reactors.
175
-------
SECTION 4
LOW BTU GAS COMBUSTION EXPERIMENTS
The idealized fuel/air contacting assumed in the preceeding kinetic
analysis cannot be achieved in practice. The combustor designer must take
account of practical constraints such as combustor cost, maintainability, and
operability. Pressure drops must be minimized since they represent energy
losses, and combustor volumes (and therefore, residence times) must be kept
within bounds to reduce combustor cost and the potential for heat loss. The
approach of this study is to use the results of kinetic analysis to guide the
development of low NO combustion concepts primarily through experiments. This
JL
section discusses the results of a series of flame reactor experiments where
LBG containing various XN compounds was combusted over a wide range of fuel/air
contacting conditions. The objectives of these tests were to:
Evaluate the relationship between L'BG composition and XN
processing
* Investigate XN processing in LBG-fired reactors as a function
of fuel/air contacting method.
Three simple combustors were tested:
Laminar and turbulent diffusion flame
Premixed pseudo one-dimensional flame (flat flame)
Premised catalytically supported reactor
The nitrogen species included in these investigations are NH3, HCN and NO.
Hydrogen cyanide 'and NO were included because they are representative of the
XN species which might exit a rich combustion stage, thus providing informa-
tion on the optimum second stage reactor.
176
-------
EXPERIMENTAL SYSTEMS
The experimental equipment used in this investigation includes:
Reactant flow system
Interchangeable flame reactors
Analytical train
These components are shown schematically in Figure 7.
The LEG was synthesized by blending together high purity gases from
cylinders. All gases were high purity grade (99.97 percent or better) with
the exception of CO (99.0 percent). NH^, NO and HCN were supplied as custom
grade mixtures in nitrogen (+ 2 percent accuracy) to facilitate metering. The
oxidant was dry air. All gases were metered with sapphire jewel orifices
operated in the critical flow (sonic) regime. The pressures upstream of the
orifices were measured with high accuracy variable capacitance pressure trans-
f*
ducers which were calibrated periodically against a laboratory reference. The
pressures downstream of the orifices were maintained constant and the flow rate
through each orifice was calibrated by filling an evacuated tank. The esti-
mated total inaccuracy in each gas flow rate was 0.5 percent.
Variations in water vapor content of the LBG were achieved by metering
distilled water with a calibrated rotameter and prevaporizing before mixing
with the other reactant gases. The mixture was maintained well above the
dewpoint temperature to prevent subsequent condensation. Other design details
of the reactant flow system are discussed in Reference 18.
The diffusion flame reactor utilized in these experiments is shown in
Figure 8. Fuel gases entered through a stainless steel tube centered in a
sintered stainless steel disc. Air passed through the sintered disc and
mixed with the fuel gases by laminar or turbulent diffusion depending upon
flow conditions. A fully developed velocity profile was assured by a tube
length greater than 100 diameters. The surrounding tube was enclosed by a
water jacket to control heat losses. Combustion products exiting the reactor
were mixed by several rows of stainless steel water-cooled tubes normal to
the flow. Samples were withdrawn for analysis downstream of this mixing
section through a water-cooled stainless steel probe. All water cooled
177
-------
surfaces were maintained at 343 to 363 K.
The premised pseudo one-dimensional flame reactor is shown in Figure 9.
In this "flat flame" reactor the premixed reactants passed through a water
cooled stainless steel sintered disc. Once the reactants were ignited a thin
planar flame stabilized approximately 1.0 mm above the disc. The flat flame
reactor was enclosed by a quartz tube and combustion products were sampled
with stationary probes. The flat flame reactor was moved axially to obtain
combustion product samples at any desired distance downstream from the flame
front.
The premixed catalytically supported reactor is shown in Figure 10. The
catalyst used in these tests was a platinum coated graded cell monolith
supplied by the Acurex Corporation. The desirable features of the graded cell
catalyst have been discussed previously (18). The maximum recommended tempera-
ture for this catalyst is. 1588 K (2400 F). Type K thermocouples were cemented
in two of the cells in the downstream segment to monitor maximum monolith
temperature. To minimize heat losses and approximate adiabatic conditions, the
catalyst was mounted in a refractory tube and electrically backheated. The
reactants passed through a sintered stainless steel disc immediately upstream
of the catalyst and combustion products were sampled with a stainless steel
water cooled probe. Additional details of this reactor's construction and
operation are documented in Reference 13.
The same sampling and analysis systems were used for all flame reactor
studies. Combustion products were analyzed for 09, CO, C09, NO, NO , NH and
£ & Jt .3
HCN. The sample train components were constructed entirely of stainless steel,
glass and Teflon. Ammonia and HCN were trapped by bubbling known volumes of
combustion products through three water baths in series. The absorbed NH-
and HCN concentrations were then measured by ion specific electrodes. The
other product species were monitored continuously. The sample was dried to
a dewpoint of 273 K in a cyclone water trap located close to the sample
probe to minimize residence time between probe and trap. Nitric oxide and
NO were measured with a Thermal Electron Corporation Model 10A chemilumines-
Jt
cent analyzer operated as discussed in Reference 19. NO concentrations
JL
were measured after the sample had passed through a stainless steel converter
operated at 1073 K. This converter could not be operated under oxygen-
deficient conditions because of the well-known reduction of NO.
178
-------
EFFECT OF LEG COMPOSITION ON FUEL NITROGEN CONVERSION
The experiments involving the effect of LEG composition on the forma-
tion of fuel NO were carried out in the diffusion flame and flat flame
3t
reactors. The experimental procedure was to fire the reactors with and with-
out an XN dopant and to calculate the percentage conversion of the dopant to
IN as follows:
(XN measured\ / XN measured \
(with dopant I - I without dopant)
/ j\/ \ / j \ /
(ppm dry) / \ (ppm dry) /
nopant co AIM i = " Trrr'- : rrr * '
K ,. I /XN calculated for full\
/ I conversion of dopant j
\ (ppm dry) /
For experiments at fuel-lean stoichiometries, only NO and NO were measured.
Ji
For experiments at fuel-rich stoichiometries, NO, NH_ and HCN were measured
and EXN was calculated as the sum. For rich experiments the XN measured with-
out the dopant was negligible and for lean experiments, thermal NO was
Ji
typically 50 ppm while full conversion of the dopants was usually a factor
of 20 higher.
Figures 11, 12 and 13 show the results for a laminar diffusion flame. All
data refer to a constant fuel volumetric flow rate, 366 K fuel and air preheat
and 150 percent theoretical air. The presence of hydrocarbon fuel in the LEG
was found to have the most significant effect upon the conversion of NH» to
fuel NO (see Figure 11). Without methane (CH,), percentage conversions were
X. "
typically 5 percent of the dopant. This rapidly increased to greater than 20
percent when the fuel gas contained 5 percent CH.. Above this concentra-
tion further increases in CH, content did not have a significant effect
upon fuel fuel nitrogen conversion. The influence of ethylene (C-H.) as the
hydrocarbon fuel is similar to that of CH,; however,wmaximum conversions
are lower typically of the order of 20 percent for an LEG containing 10 per-
cent C_H,. It can also be seen in Figure 11 that the influence of CO/H2
ratio was minimal compared to that of the presence of a hydrocarbon. Similar
experiments varying the diluent composition to include N«, CO^ and H_0 showed
diluent composition effects to be on the same order as CO/H_ ratio effects (12),
The partial products of combustion of the first stage in a staged heat
179
-------
release combustor can contain several nitrogenous species (NO, NH_ and
HCN) as well as CO, H and a hydrocarbon fuel. The results in Figure 12
show that the conversion of a given nitrogenous specie in an LBG fuel is not
only dependent upon the hydrocarbon fuel content, but also on the nature of
the fuel nitrogen specie. High conversions of small quantities of HCN and
NO were measured for small dopant concentrations, and these conversions were
almost independent of hydrocarbon content. However, at the same dopant level
the conversion of NR- is strongly affected by hydrocarbon content.
Figure 13 shows the effects of hydrocarbon fuel concentration and dopant
leval on NH_ and NO conversion. With the exception of NO as the dopant and
zero percent CH,, percentage conversion of dopants decreased as the dopant
levels increased. However, this decrease in conversion was insufficient to
counter the increase in XN available for conversion and total NO emissions
x
increased with dopant level. At NO dopant levels greater than 500 ppm there
was a significant influence of hydrocarbon content upon fuel nitrogen conver-
sion (NO to NO ). This suggests that if the effluents of a rich primary con-
tained a hydrocarbon fuel gas and NO, as the only bound nitrogen species,
significant reductions in emissions might be achieved in a lean secondary burn-
out stage.
Figure 14 shows the results of similar experiments varying CH, concen-
tration in LBG-.fired flat flames. The conversion of NH3 to NO is essentially
the same with and without CH, except at very rich conditions where some
is formed in the case with CH,.
HCN
The effect of a hydrocarbon fuel on the conversion of NH_ to NO can
be explained by consideration of the reactions shown in Table 2. In a lami-
nar diffusion flame NH_ breakdown occurs on the fuel-rich side of the heat
release zone, and N_ formation can occur by reaction of the nitrogenous
fragments. Any NO that is produced can react with these nitrogen fragments
producing N_. When a hydrocarbon fuel is present hydrogen cyanide can be
produced from NO via reactions such as 7 and 8, thus effectively removing
nitrogen specie from the zone in which they can be more easily converted to
N_. Once HCN is produced, then the XN specie exchange reactions allow the
regeneration of NO as 0_ diffuses into the reaction zone.
180
-------
In the flat flame the reactants are premixed and NH breakdown occurs
in the presence of 0 . Thus NH_ and the. resulting fragments compete with
the fuel species for 0-. Reduction reactions are not favored. At very rich
stoichiometries much less 0_ is 'available and the effects of hydrocarbon
content become more pronounced.
It is well-known that the fuel/air contacting conditions in a diffusion
flame are dependent upon fuel and oxidant flow rates. For a simple coaxial
system such as the one tested here, a laminar flame is produced at low flow
rates. As the flow rate is increased there is a transition to a turbulent
flame with noticeable audible noise and a brush-liek appearance. As the
flow rate is increased further, the flame lifts off the port allowing partial
premixing of fuel and air beneath the reaction zone. Figure 15 shows the
conversion of NH, to NO in LBG-fired diffusion flames operated in all three
fuel/air contacting conditions.
Under laminar conditions, the hydrocarbon content is an important variable
as discussed previously. Under attached turbulent conditions the conversion
increases somewhat but the strong effect of CH, concentration is similar to
the laminar case. However, in the lifted condition the effect of CH,
disappears and conversion is nearly complete. Thus from a fuel nitrogen per-
spective the attached laminar and turbulent flame produces fuel nitrogen
conversion similar to a premixed system.
EFFECT OF COMBUSTOR CHARACTERISTICS ON FUEL NO FORMATION
The effect of combustor characteristics on fuel NO formation was
2C
measured by firing an NH_ doped LBG in the three reactors described previously
over a range of reactant stoichiometries. The LBG composition synthesized for
these tests is listed below:
CO
H2
CH4
N2
NH3
20 percent
20 percent
5 percent
55 percent
471 ppm
33 3
The heating value of this gas is 37 x 10 J/m (182 Btu/ft ).
181
-------
Measured concentrations of NO, NH. and HCN are presented in Figure 16 for
the diffusion flame, the premixed catalytic reactor and the premixed flat
flame reactor operating fuel-rich. For the diffusion flame and catalytic
reactors, combustion products were sampled downstream of the mixing device
described previously. For the flat flame reactor combustion products were
sampled on the axis approximately 45 msec from the primary reaction zone.
Axial probing traverses showed no significant variation in these concentra-
tions over a product residence time range of 2 to 50 msec.
The results in Figure 16 are similar for all reactors: as the percentage
of theoretical air in the reactants decreased, the concentration of NO in the
reactants decreased, the concentration of NO in the products decreased while
the NH_ and HCN concentrations increased. The diffusion flame utilized for
these tests incorporated a water-cooled housing. As the percent theoretical
air was reduced from overall lean conditions the flame expanded toward the
cooled walls. Under rich conditions the base of the diffusion flame operated
in the normal fashion with a near stoichiometrie flame front, but the upper
portion was quenched along the cold walls. This accounts for the high NH.
throughput under rich conditions.
For the premixed catalytic reactor the HCN concentration was virtually
zero down to 70 percent theoretical air. As the percent theoretical air
was further decreased, HCN increased rapidly until at 45 percent theoretical
air the measured NH. and HCN concentrations were the same. In the premixed
flat flame NH. concentrations were very low, and the HCN concentration
increased as the percentage theoretical air decreased. The NO concentrations
in the premixed flat flame were generally higher than those in either of the
other two reactors.
182
-------
SECTION 5
DISCUSSION AND CONCLUSIONS
The results presented in this paper represent the initial portion of a
study to develop combustor concepts to minimize NO emissions from LEG fired
2£
combined cycle power systems. Low Btu fuel gases may contain up to 3800 ppm
NIL and this could result in high NO emissions under typical gas turbine
J X
combustor operating conditions. However it is known that the conversion of
NIL. to NO can be restricted by modifying the combustion process in such a way
as to ensure that the NH reacts initially under fuel-rich conditions.
Kinetic analysis of idealized combustors has shown the potential for very low
NO emissions even with high NH_ concentration in the LEG.
* J
Before the development of bench-scale combustors can commence it is
necessary to ascertain the parameters controlling fuel NO formation in the
LEGfired systems. LEG fuel gases will vary in composition depending upon the
gasifier design and operating parameters. Thus, if the combustor concepts are
to be universally applicable then it is necessary to establish whether the
combustor must be tailored to the particular fuel gas being produced. The
results indicate that by far the most significant effect of LEG composition
on fuel NO formation is the presence of a hydrocarbon gas.
2C
Kinetic considerations suggest that optimum temperature/stoichiometry
histories exist to maximize N- production. However, practical considerations
require that the fuel and oxidant be mixed either before or after injection
into the reaction chamber, thus providing the potential for different types
of combustors based upon the fuel/air contacting process in the primary zone.
A series of simple reactor experiments have been carried out with synthesized
LBG to determine the influence of combustion characteristics on the processing
of NH_ in LEG. Figure 17 compares the EXN (measured rich) and NO (measured
j X
lean) for the three reactors firing the baseline LEG doped with NH and similar
183
-------
data for the diffusion flame reactor firing an LEG with no hydrocarbons. For
the baseline LEG containing CH,, the diffusion flame and catalytic reactors
produce comparable minimum conversions and the flat flame reactor produces
much higher conversion over the range of stoichiometries tested. (It could
not be operated richer than 60 percent theoretical air due to flame
instability.) It is difficult to compare the three reactor types since their
operational characteristics are different. The maximum operating temperature
of the catalyst was 1588 K, and the undiluted reactants could not be burned
between 40 and 230 percent theoretical air. Therefore, to obtain the results
shown in Figures 16 and 17 it was necessary to dilute the reactants with
nitrogen to ensure that this maximum temperature was never exceeded. In a
real system this would require recirculation of cooled combustion products or
heat loss from the primary section. Fenlmore et al. (20) have carried out
similar .diffusion flame experiments and have shown that as the.flame becomes
turbulent there is a pronounced increase in fuel nitrogen conversion to NO .
3t
In the experiments reported here, the transition from laminar to turbulent
flames did not produce comparable results. This can be explained only if it
is accepted that at these turbulence levels the turbulent flame front pro-
vides a continuous reaction zone around the fuel stream. If turbulent levels
were increased to such levels that the reaction zone were stretched and
extinguished then fuel oxidant mixing could take place without reaction and
the subsequent formation of fuel NO would increase.
184
-------
REFERENCES
1. Evaluation of Phase 2 Conceptual Designs and Implementation Assessment
Resulting from the Energy Conversion Alternatives Study (EGAS), pre-
pared under Interagency Agreement E(49-18)-1751, NASA Report No.
TM X-73515.
2. Shaw, H., A. E. Cerkanowicz, and S. E. Tung. Environmental Assessment
of Advanced Energy Conversion Technologies - Interim Report, Vol. 1,
State-of-the-Art. Contract No. 68-02-2146, U. S. Environmental Pro-
tection Agency, Cincinnati, Ohio, 1977.
3. Harris, L. P., and R. P. Shah. Energy Conversion Alternatives Study
(EGAS), General Electric Phase II Final Report: Vol. II, Advanced
Energy Conversion Systems - Conceptual Designs: Part 3, Open Cycle
Gas Turbines and Open Cycle MHD, General Electric Report No. SRD-76-
064-2, NASA Report No. NASA CR-134949, 1976.
4. Beecher, D. T., et al. Energy Conversion Alternative Study (EGAS),
Westinghouse Phase II Final Report: Vol. II - Combined Gas-Steam
Turbine Plant Using Coal-Derived Fuel, Westinghouse Report No. 76-
9E9-ECAS-R2v.2, NASA Report No. NASA CR-134942, 1976.
5. Fluor Engineers and Constructors, Inc. Gasification-Combined-Cycle
Power Plants. EPRI Journal, July/August, 1978. p. 43.
6. Robson, F. L., W. A. Blecher, and A. J. Giramonti. Combined-Cycle
Power Sytems. EPA-600/2-76-149, U. S. Environmental Protection
Agency, Washington, DC, 1976. p. 359
7. Environmental Protection Agency Title 40, Chapter 1, Subcharter C,
Part 60-Standards of Performance for New Stationary Sources, Federal
Register, Vol. 36, No. 247, December 23, 1971.
8. Tyson, T. J., M. P. Heap, C. J. Kau, B. A. Folsom, and N. D. Brown.
Low NOX Combustion Concepts for Advanced Power Generation Systems
Firing Low Btu Gas. EPA-600/2-77-235, U. S. Environmental Protection
Agency, Washington, DC, 1977.
9. Dravo Corporation, Handbook of Gasifiers and Gas Treatment Systems,
NTIS Report No. FE-1772-11, Feburary 1976.
10. Folsom, B.A., T. L. Corley, M. H. Lobell, C. J. Kau, M. P. Heap,
and T. J. Tyson. Evaluation of Combustor Design Concepts for Advanced
Energy Conversion Systems. In: Proceedings of the Second Stationary
Source Combustion Symposium, Vol. 5, Addendum, EPA-600/7-77-073,
July 1977.
185
-------
11. Heap, M. P., T. J. Tyson, J. E. Cichanovicz, R. Gershman, and C. J. Kau.
Environmental Aspects of Low Btu Gas Combustion. In: Sixteenth Sym-
posium on Combustion, The Combustion Institute, 1977. p. 535.
12. Folsom, B. A., C. W. Courtney, M. F. Heap, and G. B. Martin. The Effect
of LBG Composition and Combustor Characteristics on Fuel NO Formation.
In: Fourteenth Annual International Gas Turbine Conference, A.S.M.E.,
Gas Turbine Division, March 1979.
13. Folsom, B. A., C. W. Courtney, and M. P. Heap. Environmental Aspects
of Low Btu Gas-Fired Catalytic Combustion. Proceedings of the Third
Workshop on Catalytic Combustion, October 3-4, 1978.
14. Tyson, T. J., et al. Fundamental Combustion Research Applied to Pollu-
tant Control. First Annual Report, EPA Contract 68-02-2631. In prepa-
ration.
15. Morley, C. The Formation and Destruction of Hydrogen Cyanide from
Atmospheric and Fuel Nitrogen in Rich Atmospheric-Pressure Flames.
Combustion and Flame, Vol. 27, 1976. ppi 189-204.
16. Myerson, A. L. The Reduction of Nitric Oxide in Simulated Combustion
Effluents by Hydrocarbon-Oxygen Mixtures. Fifteenth Symposium (Inter-
national) on Combustion, The Combustion Institute, Pittsburgh,
Pennsylvania, 1975.
17. Sarofim, A. F., J. H. Pohl, and B. R. Taylor. Mechanisms and Kinetics
of NO Formation Recent Developments. Paper presented at 69th Annual
Meeting AIChE, Chicago, November 1976.
18. Kesselring, J. P., W. Y. Krill, and R. M. Kendall. Design Criteria
for Stationary Source Catalytic Combustors. EPA-600/7-77-073c,
U. S. Environmental Protection Agency, Washington, DC, July 1977.
p. 193.
19. Folsom, B. A., and C. W. Courtney. Chemiluminescent Measurement of
Nitric Oxide in Combustion Products. In Proceedings of the Third
Stationary Source Combustion Symposium, March 1977.
20. Fenimore, C. P. Effects of Diluents and Mixing on Nitric Oxide from
Fuel Nitrogen Species in Diffusion Flames. Seventeenth Combustion
Symposium, published by the Combustion Institute, 1978. p. 1065.
186
-------
oo
Coal Input
Steam
Pressurized
Gaslfler And
Cleanup System
Air
Inlet
Compressor
LBG
Adiabatlc
Gas Generator
Turbine
Steam Turbine
Generator
Condenser
AWi
i
Cooling Water
Feed-
Water
Pump
-3-
Generator
Waste Heat
Boiler
Figure 1. Simplified Schematic of Basic LBG-Fired Combined Cycle.
-------
Air
620K
CO
00
LBG
Fuel
422K
A1r
620K
Well
Stirred
Ignition
Plug Flow
Secondary A1r
Plug Flow
To Turbine
- 1589K
218% TA
Rich Primary
Secondary Burnout
Stage
Figure 2. Limit Case Schematic Diagram of a
Rich/Lean Series Staged Combustor.
-------
1000.0 -
100.0 -
to
r-
CO
O)
t
Q.
10.0 -
65 100
300
400 500
Stoichiometry (% Theoretical Air)
Figure 3. Adiabatic Equilibrium £XN for LEG.
189
-------
0
80 100
Residence Time (MSEC)
Figure 4. Kinetic Analysis of 65% Theoretical
Air Rich Primary.
-------
10
0
I
Secondary Air Added Over 60 MSEC
Secondary Air Added Over 4 MSEC
< 100% Conversion of XN Entering Secondary to NO (Zero Thermal NO )
A A
I
I
I
I
J_
I
I
0 10 30 50 70 90 110 130 150.
Secondary Stage Residence Time, MSEC
170
190
210 230
Figure 5. Second Stage Air Addition.
-------
Fuel
Air
Heat
Addition
Rich Com-
bustion
Zone
Heat
Removal
To Second
Stage
Burnout
A. Heat Feedback
ieL /^ ^\
In taMMMM^^M
i *R- / ^
r_ \ 1
r
L u j
I
*,
*1
^^^
i .
I *"
I
i h
^^
j ,
i
1 ,
1 '3
i
M t
To Second
Staae
Bumout
B. Distributed Air Addition
Mixing Zone
C. Parallel Staging . f
To Second
Stage
Burnout
Figure 6. Idealized Low NO Combustor Concepts.
X
192
-------
ID
CO
Exhaust
Stack
High Purity Gas
Cylinders
\/ent
Gas Bubbler
Train for Wet
Sample
8 Channel
Critical Orifice
Flow Metering
System
In Situ
Calibration
System
Water Trap
Sample
Probe
Vaporizer
Calibrated
Rotameter
Zero and Span Gases
O O
Reactant
Preheater
500°C
Max.
Rotameters
NO/NO
Chemi luminescent
Analyzer
CO
NDIR
co2
NDIR
°2
Paramagnetic
Bypass and
Wet Sample
Flowrate
Measurement
Figure 7. Experimental Apparatus.
-------
Sample Probe:
Stainless
Baffle Plate
Cooling Water
Circulation
Pump
Water Cooled
Exhaust Stack
Stainless Steel
Water Cooled
Mixing Section
Stainless Steel
Water Cooled
Flame Tube
Sintered
Stainless Steel
Disc lOOy
30xidizer Plenum (Heated)
0.63 cm
0.55 cm
0.250 in
0.218 in
O.D.
I.D.
Oxidizer
Fuel
Figure 8. Diffusion Flame Reactor.
194
-------
Movable
Burner
Assembly
Stack
1 '
y
91 c
(36
Water C(
Stainle;
Probe
:m
in
1
^V^B
M^BB
^a
^^m
1
4.8 cm
(1.9 in,
OOOOQ
\
OO
I
^
Reactants u n
Inlet H2°
Ih
Thermo
coupl
.1
Sample
=3
Ul K
tl
In Out
Flat Flame
Sintered
Stainless
Steel disc (20 y)
Cooling Coil
7.6 cm (3.0 in)
Quartz Tube
Stainless
Steel Weldment
Perforated
Plate
Out
Figure 9. Flat Flame Burner.
195
-------
Exhaust
Stack
Water-Cooled
S.S. Sample Probe
Graded Cell
Catalyst
Ceramic Wool
Insulation
Electrical
Heating
Element
1500 w.
WO y
Sintered
S.S. Disc
Castable
Refractory Cylinder
0)
Q.
Figure 10. Catalytic Reactor and Housing.
196
-------
30
20
O
-P
c
O
r*
to
0)
O
O
_ NH3 = 3059 ppm
Diluent = 55% N,
150% T.A.
CH4 Hydrocarbon
0
O
O CO/Hg =1.0
& CO/H2 = 2.0
Hydrocarbon
O CO/H2 = 1.0
0
0
10
Fuel Hydrocarbon Content
Figure 11. Laminar Diffusion Flame Processing
of NH
197
-------
100 ppm NO
100 ppm HCN
100 ppm NH3
CO/H2 = 1.0
Diluent = 55% N?
150% Theoretical Air
Fuel CH4 Content (%)
10
Figure 12. Diffusion Flame Processing of XN
at Low Concentrations.
198
-------
10
1000
Dopant NO
O 0 % CH
- D 5 %
A 10% CH
A 10%
I
2000
3000 0 1000
Dopant in LBG (ppm)
2000
3000
Figure 13. Laminar Diffusion Flame Process of NH~ as Function
of Dopant and Hydrocarbon Concentrations.
-------
0% CH,
5% CH,
Full Conversion
o
o
X
o
Qi
1
o
100
90
80
70
60
50
40
30
20
I
120 140
% Theoretical Air
O EXN (rich) or NO -
(lean)
O
D
NO
NH3
HCN
120
140
Figure 14. Flat Flame: Effects of Fuel Hydrocarbon Content
on NH3 Conversion to XN.
-------
LBG Diffusion Flame
(NH3) = 500 ppm
150% Theroetical Air
CO/H2 -1.0
(N2) = 55%
Full Conversion
O Q% CH^
D 5% CH,
ro
O
100
80
X
o
o
I 60
10
S-
O
^.co 40
z.
20
0
I
II
I
Attached
Laminar
Flame
Attached Turbulent Flame
1
Lifted Turbulent Flame
1
1
0
2000
4000
6000 8000
10000 12000
14000
Figure 15. Diffusion Flame: Effects of Flow Characteristics
on NH. Conversion to XN at 150% Theoretical Air.
-------
\ Diffusion Flame
N471 ppm NH, 1n LBG
N J
300
20°
o
£
+j
0)
u
o
o
X
100
TIIir
Flat Flame
471 ppm NH- 1n LBG
Thermal NOJ
IIIIII
Catalyst
1n LBG
ZXN
D NH3
A HCN
O NO
Diluted with N,
to 1477 K *
\ Equilibrium
AcMabatic Flame
Temperature
I
0 20 40 60 80 100 120 0 20 40 60
Theoretical Air (%)
Figure 16. Comparison of XN Emissions from Three Reactors.
-------
Low Btu Gas
CO/H2 =1.0
55%
NH
2 = 471 ppm
O Diffusion Flame (5% CH4)
Diffusion Flame (0% CH4)
D Catalyst (5% CH4)
A Flat Flame (5% CH4)
100
ro
o
co
s-
o
o
4J
c
o
I
I/)
0)
c
o
o
50
1
i
1
20
40
60
80 100 120
% Theoretical Air
140
Diffusion (5%) ~
""^V J
160
180
200
Figure 17. NH« Conversion to ZXN for Three Reactors,
-------
TABLE I. OVERALL EFFICIENCY AND COMBUSTOR OPERATING CONDITIONS
FROM THERMODYNAMIC ANALYSIS
CYCLE/TIT (K)
Basic/ 1366
Basic/ 1700
Duct Heater /I 366
Duct Heater/1700
Supercharged n ,,.
Boiler /1JD
Supercharged/1700
Boiler
Reheat/1366
OVERALL
EFFICIENCY
39.1
46.3
40.1
46.8
41.0
47.3
44.6
COMBUSTOR OPERATING CONDITIONS
TYPE*
AGG
AGG
AGG.
ADH
AGG
ADH
SB
SB
AGG
ARC
PRESSURE
(ATM)
16.0
30.0
16.0
1.1
30.0
1.1
16.0
30.0
20.0
6.9
OUTLET
TEMP,
(K)
1366
1700
1366
914
1700
914
1366
1700
1366
1366
OXIDANT AIR
STOICH.
(% TA)
NV**
NV
NV
377
NV
286
NV '
NV
NV
425
TEMP.
(K)
697
844
697
762
844
840
697
844
536
1078
STOICH.
(% TA)
377
286
377
314
286
268
309
265
425
283
HEAT***
0
0
79.3
20.7
91.0
9.0
16.3
6.9
62.3
37.7
ro
2
* AGG - Adiabatic Gas Generator; ADH - Adiabatic Duct Heater; SB - Supercharged Boiler;
ARC = Adiabatic Reheat Combustor
** NV = Nonvitiated Air
*** For duct heater and reheat combustor, Heat % is % of total heat liberated in combustor.
For supercharged boiler it is % of total heat transferred to bottom cycle.
-------
TABLE II. MAJOR FEATURES OF KINETIC SCHEME DESCRIBING
THE FATE OF AMMONIA DURING COMBUSTION
Ammonia Breakdown
NH3 + H -» NH2 + H2 1
NH_ + H -> NH + H 2
NH + H + N + H 3
Nitric Oxide Formation
N + 0 -* NO + 0 4
N + 0 -» NO + N 5
N + OH -» NO + H 6
Hydrogen Cyanide Formation
via CH NO
CH, + NO , no «mn ' HCN + . . . 7
3 and CH NOH
CH + NO ^ HCN + 0 8
CH -t- N + HCN + N 9
XN Specie Exchange
HCN + OH -» CN + H20 10
CN + OH * NCO + H 11
NCO + H -*- NH + CO 12
NH + H2-*NH2-fH 2R
NH2 + H2 * NH3 + H 1 R
Nitrogen Formation
N + NO + N+0 5R
NH + NO -> N + H0 13
205
-------
CATALYTIC COMBUSTION SYSTEM DEVELOPMENT
FOR STATIONARY SOURCE APPLICATIONS
By:
J. P. Kesselring, W. V. Krill, E. K. Chu, and R. M. Kendall
Acurex Corporation
Mountain View, California 94042
207
-------
ABSTRACT
An experimental program has been conducted for the Environmental Pro-
tection Agency to develop design criteria for catalytic combustors as applied
to stationary systems. The program included catalyst screening tests from
which the graded cell concept was developed. The graded cell catalyst exhibits
greatly enhanced combustion characteristics in terms of Increased maximum
throughput. Advanced testing of the graded cell catalysts showed high heat
release rate capabilities with low emissions of CO, HC, and thermal NO .
j£,
Operation of the catalysts under fuel-rich conditions showed capability to
control fuel NO emissions. Additional criteria for system scaleup and opera-
X.
tion under varying preheat and pressure conditions were also generated.
Catalysts developed during the program were incorporated into three small-
scale systems with heat extraction. A radiative catalyst/watertube system,
utilizing direct heat removal from the catalyst, was devised and tested. The
concept has potential application to watertube boilers. A model gas turbine
combustor was tested at pressures between 0.101 and 1.01 MPa (1 to 10 atmos-
pheres) to investigate operating characteristics and fuel nitrogen conver-
sion to NO . The final system, a two-stage combustor, was constructed to
2k
utilize fuel-rich first stage combustion for fuel NO control. Measured
A
emission results make the concept attractive for a variety of future system
applications.
208
-------
INTRODUCTION
Catalytic combustors have shown promise in reducing the levels of CO,
HC, and NO emissions over those of conventional burners in laboratory tests.
X.
In order to develop and later demonstrate catalytic combustors for commer-
cially viable systems, two related development activities were performed.
The first activity involved the development and characterization of catalyst
systems, while the second activity focused gn the analytical and experi-
mental evaluation of three system concepts employing catalytic combustors.
The following section of this paper presents data on a variety of
catalyst models with varying substrate, washcoat, and catalyst materials.
This data indicates the catalyst system heat release capabilities, preheat
requirements, temperature limitations, and operational life. The subse-
quent section presents three subscale designs and operational results for
watertube boiler, gas turbine combustor, and two-stage combustion systems
using catalytic combustors. The concepts, designs, analyses, and data
presented here were generated under EPA Contracts 68-02-2116 and 68-02-2611,
Task 11.
209
-------
CATALYST DEVELOPMENT
In order to develop catalyst models for subsequent system application,
a number of desirable catalyst characteristics were identified. These
characteristics include:
Low ignition temperature
Low preheat requirements for sustained combustion
Combustion uniformity throughout the bed
High heat release capability
High combustion efficiency
Low pollutant emissions
High operating temperature
Fuel flexibility
Long life
Catalyst models tested included variations in substrate, washcoat, and catalyst
materials as well as substrate geometry and preparation techniques.
As reported at the Second Stationary Source Combustion Symposium (Ref-
erence 1), an initial series of combustion tests were performed at the Jet
Propulsion Laboratory and at Acurex. Results from those tests indicated
that catalyst performance was improved by:
1. Increased catalyst loading, resulting in lower initial lightoff
temperatures, higher mass throughputs, and increased lifetime at
1367K (2000°F).
2. Increased cell size, allowing higher possible mass throughputs at
the expense of increased hydrocarbon emissions.
3. Heavier hydrocarbon fuels which promote lightoff at lower ignition
temperatures.
4. Hydrogen sulfide (H^S) fixation of platinum catalysts to promote
retention of platinum surface area.
210
-------
5. Presintering of catalyst washcoats which reduces burying of active
catalyst below the surface during combustion.
6. Stabilization of y-AlJO,. washcoats with cesium oxide (Cs20) up to
5 weight percent increased surface area. Stabilization of alumina
washcoats with ceria up to 5 weight percent had a negative effect
on surface area.
7. Decreased cell size to significantly reduce unburned hydrocarbon
and carbon monoxide emissions.
8. Catalyst beds of combined large cell and small cell monoliths also
significantly increase throughput (at a given preheat temperature)
and overall catalyst life with low emissions.
9. Bed temperature uniformity was increased by operation at higher
temperatures.
Catalysts with combined large and small cell segments represent the best
concept developed by the test series. This graded cell concept (three seg-
ments shown in Figure 1) was further developed through predictions of the
PROF-HET catalytic combustion computer code (Reference 2). Finally, prelimi-
nary data with ammonia-doped natural gas indicated a potential for control of
fuel nitrogen compounds to NO under lean conditions.
Ji
GRADED CELL CATALYST TESTS
Additional catalyst screening tests were conducted at Acurex on the
graded cell configuration. The primary objective of these tests was to
identify the best catalysts for system application. Catalysts were also
tested for high temperature capability, system scaleup design criteria, and
conversion data of fuel nitrogen to nitrogen oxides under varying operating
conditions.
Screening catalysts were obtained from six sources, including W. R.
Grace and Company, Universal Oil Products, Inc., William Pfefferle (a
private consultant), Matthey Bishop, Inc., Johnson Matthey, and Acurex.
Substrate materials were either DuPont alumina or Corning zirconia spinel.
Washcoats varied from proprietary preparations with high pretest surface
211
-------
area to no washcoat with low pretest surface area. Catalysts were either
precious metal, metal oxide, or mixtures. The catalyst loadings and fuels
used are listed in Table I.
To support combustion test results, pre- and post-test catalyst physical
measurements were made. These measurements included catalyst surface area
and dispersion performed in the EPA/Acurex catalyst characterization labora-
tory. Additional scanning electron microscopy (SEM) and energy dispersive
analysis by X-ray (EDAX) tests were performed at the Jet Propulsion Labora-
tory as required. Table II is presented for reference, and summarizes all
surface area and dispersion measurements.
Catalysts .were aged for 10 hours and compared in maximum throughput,
varying atoichiometry, and minimum preheat tests. Some catalysts, primarily
precious .metals with low loadings, exhibited lifetimes of less.than 10 hours,
precluding extensive data evaluation. Others were tested in a complete 20
to 30 hours test sequence. In addition, some catalysts showed preferential
activity for either fuel-rich or fuel-lean combustion or for operation only
at higher preheats of 700K to 811K (800°F to 1000°F).
The maximum throughput (volumetric heat release rate) capabilities were
compared for various graded cell catalysts to indicate relative activity.
Catalyst data selected from Table I are compared below.
MAXIMUM THROUGHPUT COMPARISON DATA**
Catalyst Type
UOP Proprietary
Pt-Ir/Al203
Stab. Pt/Al203
JM Proprietary
Spinel
Noble Metal/
Manufacturer
UOP 2
W. R. Grace
Matthey
Bishop
Johnson
Matthey
Pfefferle 3
Acurex 3
Bed
K
1522
1256-
1589
1611
1589
1617
1644
SV, 1/hr
343,900
327,400
348,000
603,300
443,100
661,400
in J
q »hr-Pa-m3
3.0 x 106
2.8 x 106
5.2 x 106
7.2 x 106*
5.1 x 10s*
7.8 x 106*
Bed
Uniformity
Excellent
Ragged
Ragged
Excellent
Excellent
Excellent
1756 2,070,000
Acurex 4
Proprietary
**A11 results with natural gas/methane fuel
*Bed not blown out
tMultiply entry by 2.7 for Btu/hr-atm-ft3
38.9 x 10
6*
Excellent
-------
The first three catalysts, either precious metal or proprietary materials,
exhibited volumetric heat release rates (q111) of approximately 3.7 x 106
J/hr-Pa-m3 (10 x 106 Btu/hr-atm-ft3) at blowout conditions. The next three,
either proprietary or cobalt oxide catalysts, reached approximately 7.4 x 106
J/hr-Pa-m3 (20 x 106 Btu/hr-atm-ft3) without blowing out. This heat release
rate represents the maximum facility capability in the screening test con-
figuration. It is apparent that the cobalt oxide catalysts are generally
capable of higher heat release rates than the precious metals in the graded
cell configuration. The oxides required operation at higher temperature,
however, to produce combustion efficiencies comparable to the precious metals
at lower temperatures. The last catalyst, noble metal on a proprietary
substrate, was tested to the limits of the test facility without blowing out.
During graded cell catalyst testing, operating temperatures were varied
from 1367K to over 1978K (2000 to 3100°F). Thermal NO data were obtained
2k
with natural gas as the primary test fuel. A number of the test models are
compared in Figure 2. From 1367 to 1644K (2000 to 2500°F), little variation
among the NO emissions with catalyst type is apparent with NO levels below
Jv
20 ppm. At 1756 to 1867K (2700 to 2900°F), the NO production rate begins to
3C
increase, as with conventional flame combustion.
In Figure 2, two different catalyst geometries are shown for fuel-lean
operation in the 1644 to 1978K (2500 to 3100°F) range. The upper curve is
for a two-segment graded cell model (A-029) that acted primarily as a flame
holder for downstream gas phase reactions. The lower curve represents data
for a three-segment catalyst (A-030) where significantly increased combustion
occurs on the surface of the catalyst bed. It is apparent that maximizing
the amount of surface reaction occurring in a catalytic combustor minimizes
the production of thermal NO . These results indicate the importance of the
graded cell configuration, particularly for thermal NO control in high excess
X
air gas turbine configurations.
In addition to baseline catalyst screening tests, selected catalysts
were tested to characterize conversion of fuel-bound nitrogen to nitrogen
oxides with catalytic combustion. Ammonia was added to natural gas to simu-
late fuels of varying nitrogen content. Exhaust gas analyses for NO by
213
-------
chemiluminescent analyzer and for ammonia (NH«) and cyanide (HCN) by specific
ion electrode were performed.
A nickel oxide/platinum catalyst was prepared at Acurex and tested over
a range of stoichiometries from 55 to 200 percent theoretical air at a nominal
bed temperature of 2400°F. Fuel dopant concentration ranged from 2500 to
10,000 ppmv NH- in the fuel. The results are shown in Figure 3 as the percent-
age of the incoming NH, converted to NH,, HCN, and NO. The NH3 conversion to
NO increased from zero levels under very fuel-rich conditions to better than
90 percent on the fuel-lean side. NH. conversion to HCN showed the opposite
trend high under rich conditions and decreasing to zero on the lean side.
Unconverted ammonia was highest below 70 percent theoretical air and remained
at low levels under lean combustion.
The total of these three curves (dashed line and cross symbols) repre-
sents all assumed NO precursor species for the NiO/Pt catalytic combustor.
A distinct minimum occurs between 70 and 80 percent theoretical air, where
only 20 percent of the fuel nitrogen is converted to NO precursors.
A second graded cell catalyst (cobalt oxide/platinum) was tested with
simulated fuel nitrogen. The fuel nitrogen conversion is shown in Figure 4.
The ammonia conversion to nitric oxide provided the samp, characteristic curve
as the nickel oxide/platinum catalyst. Differences in the HCN and NH- species
measured, however, resulted in lower total NO precursor (NO + NH^ + HCN)
levels under fuel-rich conditions. The m-tn-tmim occurred at a lower value of
theoretical air (60 percent) than that of the previous nickel oxide catalyst
(75 percent). The cobalt oxide catalyst could thus be operated fuel rich
without dilution to achieve low conversion of fuel-bound nitrogen to nitrogen
oxides.
The low fuel nitrogen conversion of these two catalysts at 60 to 80
percent theoretical air has important system implications. Combustors which
could operate fuel-rich, possibly with staging, have potential for fuel NO
control. This characteristic of the catalytic combustor under fuel-rich
conditions is the basis for the two-stage catalytic combustor developed
under this program.
214
-------
Because of the significance of controlling fuel NO with catalytic
A.
combustors, additional testing of NH~-doped propane and methane was con-
ducted. It should be noted that fuel N conversion in rich catalytic com-
bustion is achieved for much shorter residence times than those associated
with conventional rich-lean combustion systems. Tests covering a range of
stoichiometries from rich to lean were run to determine the effect of theo-
retical air, bed temperature, mass throughput, fuel type, and fuel-bound
nitrogen type and concentration on the conversion of fuel-bound nitrogen to
NO or NO precursors. In addition, an oxygen/argon mixture was used as the
X 3t
oxidant in a series of tests to perform a nitrogen balance across the cata-
lytic combustor. Test results indicate that operating the combustor at
lower bed temperatures (1367K compared to 1478K) minimizes the formation of
fuel NO under lean conditions but does not affect the formation of NO
X X
precursors under rich conditions. Figure 5 shows the results for a proprie-
tary UOP catalyst tested with ammonia-doped natural gas fuel. The lower
bed temperature is postulated to minimize gas phase reactions, thereby mini-
mizing NO formation under lean conditions. Under fuel-rich conditions,
however, the stoichiometry has a much greater effect than bed operating
temperature.
The effect of fuel nitrogen concentration on NO formation under lean
2v
conditions is shown in Figure 6. Using the proprietary UOP catalyst at 1478K
bed temperature, the conversion of NH- to NO was seen to decrease as nitrogen
J 3t
weight percent in the fuel increased.
The results of nitrogen mass balance investigations using oxygen/argon as
oxidizer instead of air are shown in Table III. As the table indicates, a
maximum deviation of ±12 percent between the input fuel nitrogen content and
the measured nitrogen species in the combustion products was obtained, except
for two data points. N_0 was not detected for any test condition. Under
fuel-rich conditions, as expected, Nj was the dominant product of the chemi-
cally bound nitrogen conversion process.
Design criteria for the graded cell catalyst concept also included scaleup,
pressure, and preheat characteristics. Catalyst blowout was again selected to
compare activity. Based on the results of small-scale testing, a Universal
Oil Products catalyst was selected for scaleup. Analytical modeling predicted
215
-------
that combustion throughput capability would scale proportionately to bed
frontal area. Therefore, the bed diameter was increased to provide a 2.7
increase in frontal area over the small-scale model. Bed length remained
fixed at 7.6 cm (3.0 inches).
Test data showed that maximum throughput (at blowout) did scale approxi-
mately with frontal area. Maximum throughput reported for the small-scale
catalyst was 258.5 MJ/hr (245,000 Btu/hr) and 3.42 x 106 J/hr-Pa-m3 (9.3 x
106 Btu/hr-atm-ft3) volumetric heat release rate. This compares to a volumet-
ric heat release of 4.38 x 106 J/hr-Pa-m3 (11.9 x 10s Btu/hr-atm-ft3) at
926.3 MJ/hr (878,000 Btu/hr) for the scaleup catalyst at 672K preheat.
A series of blowout tests were then conducted to determine the operational
mass throughput limit of the catalyst for varying preheat and pressure condi-
tions. The blowout points used are shown below.
BLOWOUT DATA CATALYST A-041
Data
Point
1
2
3
4
Bed Temp
K (°F)
1588
1588
1588
1588
(2400)
(2400)
(2400)
(2400)
Preheat Temp
K (°F)
608
478
389
603
(635)
(400)
(240)
(625)
Max. Fuel Flowrate
Kg/hr (Ibm/hr)
15
10
8
22
.0
.5
.4
.2
(33
(23
(18
(49
.0)
.1)
.5)
.0)
Pressure
MPa (atm)
.195
.140
.134
.301
(1.
(1.
(1.
(2.
93)
39)
33)
98)
Two things are evident from the data:
o Blowout scales linearly with pressure
fuel max atm mfuel max, 1 atm)
o Blowout for catalyst A-041 was exponential in preheat temperature,
although a relatively weak exponential factor was determined.
Operation of the catalyst was, of course, possible at any combination of
preheat and fuel flowrate below the blowout limits.
The exponential behavior of preheat temperature and its effect on blowout
was shown in PROF-HET code predictions (Reference 2). The blowout data ob-
tained support this prediction and can be employed as design criteria for
operation of catalyst A-041 under varying conditions.
216
-------
The graded cell catalyst screening tests identified the design criteria
required for incorporating the graded cell configuration into system appli-
cations. Specifically, these criteria included mass throughput and heat
release capabilities, emissions under varying operating conditions and with
nitrogen-containing fuels, lightoff requirements, and current lifetime capa-
bilities. In addition, the state of the art in catalyst development has been
evaluated, including an understanding of preparation techniques, material
capabilities, and material interactions.
SYSTEM CONCEPT TESTING
The design criteria generated for graded cell catalyst configurations
were used in the specification of three small-scale systems incorporating heat
extraction techniques. A radiative catalyst/watertube system exhibited a
stoichiometric, water-cooled combustor for boiler application. A model gas
turbine combustor utilized high excess air for catalyst and exhaust gas tem-
perature control. Finally, a two-stage combustor was constructed to evalu-
ate fuel nitrogen control with application to either boiler, furnace, or
turbine equipment.
RADIATIVE CATALYST/WATERTUBE SYSTEM
The radiative catalyst/watertube concept is shown in Figure 7. A
stoichiometric fuel/air mixture is fed to the radiative section which contains
a close-packed array of catalyst elements and watertubes. The mixture is par-
tially combusted by the catalyst which is kept at a low surface temperature by
radiation heat loss to the watertubes. The combustion products and remaining
unburned fuel and air are then passed to a downstream catalytic adiabatic com-
bustor to complete combustion reactions. A final convective section extracts
energy from the fully combusted gases. Both catalyst sections operate well
below the maximum use temperature of the catalyst supports the radiative
section by radiative cooling and the adiabatic section by dilution of the fuel/
air mixture with exhaust products from the radiative section. The radiative
section was constructed and tested independently of the downstream adiabatic
combustor and convective sections.
217
-------
An initial test series was conducted using a platinum catalyst on Coors
alumina cylinders. Stoichiometry was varied from 50 to 219 percent theoreti-
cal air and fuel mass flowrate from 2.1 to 6.7 kg/hr (4.7 to 14.8 Ibm/hr) of
natural gas. Preheat conditions were also varied. Figure 8 shows the energy
extracted by the cooling tubes out of the total available energy at the bed
inlet as a function of stoichiometrie ratio.; The total available energy in-
cludes the fuel heating value and the sensible preheat energy. Thermal
input to the catalyst cylinders is primarily controlled by the adiabatic
flame temperature of the fuel/air mixture. This temperature peaks near unit
Stoichiometry, and as a consequence, the tube temperatures have a correspond-
ing maximum. As theoretical air percentage increases above 100 percent,
catalyst surface temperature begins to decrease, decreasing the radiant
exchange to the watertubes. The higher total mass throughput, however,
also increases convective heating of the watertubes such that at fixed fuel
flowrate the energy exchange does not fall off rapidly. Overall combustion
efficiency at 100 percent theoretical air was calculated as approximately
37 percent from the data. Significant emissions of CO and HC were measured
for the radiative section due to the incomplete combustion. NO levels
Zb
were consistently below 2 ppm as measured.
The radiative section was also tested to evaluate fuel nitrogen conver-
sion characteristics of the system. For those tests, natural gas was doped
with ammonia and Stoichiometry was varied from 52 to 120 percent theoretical
air.
Figure 9 shows the fuel nitrogen conversion characteristics of the
radiative system for natural gas doped with 2000 ppm of ammonia. Low NO and
3E
high NH. values above 100 percent theoretical air are consistent with the
incomplete combustion characteristics of the radiative section. The low point
in the total NO precursor curve (NH, + HCN + NO ) at 60 percent theoretical
A J X
air is similar to those obtained for metal oxide graded cell catalysts. It
should be noted that this low level of conversion is attained even though the
fuel is not fully combusted.
The radiant section as tested is not fully suited for complete system
development. The addition of the downstream adiabatic catalytic combustor
218
-------
would result in too high a temperature in that region at stoichiometric con-
ditions. This occurs since combustion efficiency in the first stage was not
quite as high as expected at the nominal 4.3 kg/hr (9.5 Ibm/hr) design con-
dition. The"radiative catalyst/watertube section did exhibit excellent per-
formance at stoichiometric conditions with very low levels of NO . The
3v
potential for control of fuel nitrogen conversion and the extremely stable
operation experienced under all test conditions make it attractive for future
boiler applications.
MODEL GAS TURBINE COMBUSTOR
Since the graded cell catalyst was demonstrated to have the low preheat,
high heat release, and pressure capabilities required for gas turbine applica-
tions, a 1055.1 HJ/hr (106 Btu/hr) model turbine can and fuel injection system
were constructed. The system and catalyst are shown in Figure 10. Testing was
performed at Acurex and Pratt and Whitney Aircraft (West Palm Beach, Florida)
facilities. Acurex and UOP catalysts were prepared on both duPont alumina
and Corning zirconia spinel support materials of varying configurations.
The model gas turbine combustor was first tested at Acurex with propane
at pressures between 0.101 and 0.354 MPa (1 and 3.5 atmospheres). A heat re-
lease rate of 263.8 MJ/hr (250,000 Btu/hr) at 1478K (2200°F) bed temperature
was run as the nominal test condition. No significant emissions of NO or CO
JL
were obtained.
Pratt and Whitney test data were obtained with propane, No. 2 oil, and
No. 2 oil with 0.5 weight percent nitrogen as fuels. Heat release rates to
844 MJ/hr (800,000 Btu/hr) were achieved with low NO emissions for both pro-
Jv
pane and No. 2 oil. Some difficulty was encountered with flashback and flame-
holding on the fuel nozzles when running No. 2 oil. High CO and unburned
hydrocarbon emissions resulted from operating at the low bed temperatures
(near the breakthrough limit) required to avoid flashback. Variations in
pressure under lean conditions were not found to affect emission levels.
Tests run with pyridine-doped No. 2 fuel oil, however, increased the NOX emis-
sions levels, representing percentage conversions of fuel nitrogen to NO of
2£
100, 61, and 55 percent for test pressures of 0.303, 0.505, and 0.707 MPa,
219
-------
respectively. Subsequent inspection of the test hardware showed that low
liquid fuel inlet velocities due to a fabrication error causing all liquid
fuel to be introduced through the large diameter gaseous fuel injection ports
were responsible for the flashback and flameholding.
A final test series was conducted at Acurex (after the test hardware
had been reworked) with natural gas, natural gas doped with ammonia, and No. 2
oil. Pressures from 0.101 to 0.808 MPa (1 to 8 atm) with natural gas and 0.101
to 0.505 MPa (1 to 5 atm) with fuel oil were included. A decrease in ammonia
converted to NO with pressure was observed. These results are consistent
with those obtained at Pratt and Whitney with pyridine-doped No. 2 fuel oil.
Maximum throughput for the catalyst at 0.303 MPa, 1244K (3 atm, 2100°F) pres-
sure and bed temperature, and 561K (550°F) preheat temperature was also in-
vestigated. At space velocities near 200,000 per hour, the catalyst began to
break through with increasing CO and unburned hydrocarbon emissions. Nitrogen
oxide emissions remained at near zero levels throughout the test. Full blow-
out was not achieved as control of catalyst temperature during breakthrough
produced difficulties in system control. The maximum heat release obtained
was 615 MJ/hr (583,000 Btu/hr).
Final tests were conducted with diesel fuel to compare emissions with
those from natural gas. An increased bed temperature was maintained for the
oil tests to maintain uniform bed conditions and suppress soot formation.
The NO, levels were higher than for natural gas (15 ppm compared to 3 ppm) due
primarily to the small amount of fuel nitrogen in diesel fuel.
The results of the model gas turbine testing demonstrated the application
of the graded cell concept in a system similar to current turbine combustor
designs. High mass flowrates were achieved in a relatively small volume com-
bustor. Overall pressure drop for the combustor and fuel injector were meas-
ured at less than 1 percent at 0.303 MPa (3 atmospheres) test pressure.
TWO-STAGE COMBUSTOR
The two-stage catalytic combustor is attractive for two reasons. First,
it allows control of bed temperatures to those compatible with the support
material without large excess air requirements. Second, the first stage can
be operated fuel-rich, which has been shown to be advantageous for reduced
220
-------
conversion of fuel nitrogen to nitrogen oxides. A two-stage combustor was
designed and constructed to exhibit these concepts.
The two-stage combustor is shown schematically in Figure 11. A fuel-
rich mixture is introduced into the primary stage which contains a graded
cr"1.! catalyst bed. The fuel is partially combusted, and the released energy
is removed by an interstage heat exchanger. Sufficient secondary air is
then injected into the combustion products to complete combustion of the
remaining fuel in the second stage. The full system combustor would also
include a second heat exchanger to remove the combustion energy released in
the second stage.
The two-stage combustor containing two cobalt oxide catalysts was tested
with natural gas at 0.101 MPa and 0.202 MPa pressures (1 and 2 atmospheres).
Lightoff and steady-state operation presented no unusual control problems.
The combustor was tested at an overall stoichiometry varying from 70 to 150
percent theoretical air at a nominal fuel flowrate equivalent to 211 MJ/hr
(200,000 Btu/hr) heat release rate. The first stage stoichiometry was varied
from 40 to 70 percent theoretical air. Ammonia was added to the natural gas
fuel at a rate of 0.2 to 0.4 weight percent.
Bed temperatures ranged from 1256K to 1660K depending on theoretical air,
for a relatively constant preheat of 617K (650°F). The energy extracted in
the interstage heat exchanger represents 50 to 60 percent of the combustion
energy generated in the first stage.
The results of vu2 fuel nitrogen conversion data are shown in Figure 12
as a function of overall combustor stoichiometry. The data show that when
operating above 100 percent theoretical air, only nitrogen oxides are normally
present. Under overall fuel-rich conditions, fractions of ammonia and cyanide
are also present. These results are consistent with fuel nitrogen data ob-
tained on the single cobalt oxide catalyst (model A-037, Figure 4). The data
in Figure 12 show a nominal 30 percent conversion rate of fuel nitrogen to NO
X
precursors with a value of approximately 27 percent near overall stoichiometric
conditions. A slight decrease in conversion was noted at 0.202 MPa (2 atm)
pressure.
The data shown in Figure 12 at approximately 10 percent conversion levels
varied in test conditions from the other data in two respects:
221
-------
1. Most significantly, the first stage was operated at higher values
of theoretical air (60 and 70 percent) compared with 50 percent for
the initial data, and
2. The first stage catalyst had experienced some sooting by later
test times when the data were taken, causing the catalyst to
operate at lower temperatures with incomplete combustion.
The first stage sooting of the cobalt catalyst proved to be a limiting factor
in the test life of the system. The incomplete combustion occurring at later
test times was evident by increasing measured carbon monoxide levels.
The demonstration of the two-stage combustor showed a number of important
results:
1. The two-stage combustor is effective in controlling conversion of
fuel nitrogen to nitrogen oxides under stoichiometric and fuel-
lean conditions.
2. A slight decrease in nitrogen conversion was found at 0.202 MPa
(2 atmospheres) pressure.
3. The variation of first stage stoichiometry impacts overall fuel
nitrogen conversion.
4. First stage sooting of the cobalt oxide catalyst was a limiting
factor in combustor operating life.
Application of the concept of both boiler and turbine systems is possible.
Convective heat exchangers downstream of each catalyst stage would provide
for steam raising in boiler applications. Interstage cooling would not be
required for gas turbine systems where high excess air in the second stage
could be used to control the catalyst and exhaust gas temperatures. Further
work is required for optimization of the system and catalyst elements for
specific applications.
CONCLUSIONS
As a result of this research and development program, significant progress
has been made toward developing a practical catalytic combustion system. Be-
fore the step to demonstration can be taken, however, additional work relating
222
-------
to the integration of the catalytic combustor into the total combustion system
must be performed.
Based upon the analysis and test results of this program, the design,
fabrication, and operation of catalytic combustors with high volumetric heat
release rates and low emissions have been demonstrated. Both precious metal
and oxide catalysts have been tested over a wide operating temperature range.
The precious metal catalysts should be limited to temperatures below 1589K
(2400°F) for catalyst life considerations, while oxide catalysts can be oper-
ated for long periods at temperatures above 1644K (2500°F). Catalyst perfor-
mance has been greatly enhanced through the use of graded cell monoliths and
higher catalyst loadings.
Catalytic combustors have been shown to be effective in controlling
both thermal and fue} NO emissions. The thermal NO control appears to re-
X X
suit from maximizing surface reactions in the combustor, while fuel NO can
3t
be minimized by operating at a rich fuel/air ratio which minimizes the forma-
tion of HE-, HCN, and NO, with complete combustion of CO and HC at a later
time.
The maximum throughput of a catalytic combustor is a linear function of
pressure and an exponential function of preheat. Thus, for a given preheat,
the catalyst is face velocity limited in throughput ability.
Small-scale catalytic combustion system configurations have been tested
and show the feasibility of direct radiative removal of bed heat for
temperature control, two-stage catalytic combustion for temperature and fuel
NO control, simulated exhaust gas recirculation through the use of nitrogen
X
diluent for temperature control, and high excess air operation. The combus-
tion system concepts that have been operated show that it is possible to
operate near stoichiometric conditions with less than 10 ppm NO and CO in a
A
natural gas-fired catalytic combustor.
A number of areas in catalytic combustion need to be addressed to capital-
ize on the progress to date. Additional testing of simple and mixed oxide
catalysts for combustion and fuel nitrogen conversion abilities is needed,
along with life testing of selected catalysts to 1000 hours at various
pressures.
223
-------
Exploratory work with heavy fuel oils (Nos. 4, 5, and 6) and pulverized
coal should be conducted to determine system feasibility and fuel preparation
problems. The potential of catalytic combustion in controlling NO emissions
from the combustion of these fuels is great and needs early experimental
verification.
Development of auxiliary systems required to interface with the catalytic
combustor is also needed. This includes lightoff systems, temperature control
systems, and fuel and air introduction systems. In addition, further testing
of the radiative catalyst/watertube, two-stage combustor, and gas turbine com-
bustor systems is needed to more thoroughly define operating ranges with a
variety of fuels.
Finally, the design, fabrication, and operation of a demonstration unit
should be undertaken when the above work is completed. The demonstration unit
would be operated as a laboratory device for several months prior to the initi-
ation of field demonstration tests.
224
-------
REFERENCES
1. Kesselring, J. P., ^t &L., "Design CritgEia_for Stationary Source Cata-
lytic Combustors," published in Proceedings of the Second Stationary
Source Combustion Symposium, Volume III, EPA-600/7-77-073c, July 1977,
pp. 193-228.
2. Kelly, J. T., et al.., "Development and Application of the PROF-HET
Catalytic Combustor Code," Paper No. 77-33 presented at the Western
States Section of the Combustion Institute, October 1977.
225
-------
ro
ro
*******
*.******
*******
*******
********
*******
mmmmm*l
mmmmmmm
Figure 1. Corning square-celled extruded monolith structures.
-------
350
300
250
VI
VI
0>
u
X
-------
lOOr
ro
ro
Oo
0)
&
I
X
o
-------
100 r
ro
ro
Catalyst A-037 (Co203/Pt)
100 120 140
Theoretical air, percent
160
Figure 4. NH3 conversion characteristics, natural gas doped with ammonia
Co203/Pt catalyst.
/\ 2 atm
3 atm
<
180
-------
UOP catalyst
Natural gas fuel
Nitrogen concentration = 0.5 wt percent
Fuel flowrate =25.3 MJ/hr (24,000 Btu/hr)
100
ro
CJ
o
c
0)
u
OJ
Q.
X
o
o
'E
O>
o
u
CO
80
60
40
20
60
I
I
1478K (2200°F)
A 1367K (2000°F)
I
I
I
80 100 120 140
Percent Theoretical Air
160
180
Figure 5. Effect of bed temperature on NHa conversion to NH3+HCN+NOX, UOP catalyst,
-------
120
110
ox 90
o 80
t
8
70
60
50
UOP Catalyst
Propane Fuel
Theoretical A1r = 310 percent
Fuel flowrate = 25.3 MJ/hr (24,000 Btu/hr)
Bed Temp. = 1478K (2200°F)
I
I
I
I
I
I
0.5 1.0 1.5 2.0 2.5 3.0
Chemically Bound Nitrogen Content, Ht. % of Fuel
N
0
O
3.5 4.0
Figure 6. Effect of nitrogen content on NH« conversion to
NOX, UOP catalyst. J
231
-------
-Refractory lining
ro
w
PO
Catalyst
coated cylinder
Mud
drum
Radiative heat
transfer section
Monolith bed
- Adlabatlc
Combustor
Convectlve heat exchanger
Matertube
Figure 7. Radiative water-tube boiler concept.
-------
ro
co
CO
150
140
125
120
100
100
I
O
75
50
80
1 60
>>
O)
UJ
25
Total available energy
Fuel mass flowrate =2.1 Kg/hr
Total Energy Release
0-L
OD
i
60
80 100 120 140 160 180 200 220(
Theoretical air, percent
Figure 8. Radiative catalyst/watertube system energy release vs theoretical air.
-------
TOO r
0
80
c
(U
NH3 + HCN + NOX
X
o
60
NH.
no
to
0)
c
o
u
r
40
20
0
50
60
70
80 90 TOO
Theoretical air, percent
110
130
Figure 9. Radiative catalyst/watertube system, fuel nitrogen conversion
-------
Figure 10. Model gas turbine combustor.
235
-------
ro
co
iWCTION
2
INLIT « oururr
COOC'MA
MILK KNC 0>HH )
Figure 11. Two stage catalytic arrangement.
-------
IVJ
w
IUU
80
X
o
f
£> 60
3
j conversion
-P»
0
%*
3:
20
0
»
A 505
O 605
a7n«
/u/
r»j
\ &O (D
1 Stage ±
-
A
^+ ^ ^***
^^ ^ta» '^V ^^ ^^
^ «*"
^ .202 MPa
0J2^ Q-70%
^ 60%
1 1
50
One-stage combuitlon
50% First Staye Theoretical Air
O 60% First Stage Theoretical Air
70% First Stage Theoretical A1r
i atni
2 atm
First stage
stoichlometry
50% T.A.
100
Overall theoretical air, percent
150
Figure 12. Two-stage combustor fuel nitrogen conversion.
-------
TABLE I. GRADED CELL CATALYST MODELS
S*Ml* M.
AEM-OM
/UM-027
ACM-OB
ACTO-029
AEM-030
AEM-031
AEM-OJ!
AEM-033
AOW-034
AEMMI3S
WKO-036
AEM-037
AEMO-038
AEM-041)
AEM-M1
Ho. Of
TC
6
*
-
1
6
1
i
T
7 r
7
7
4 .'
5
S
4
_ Svbitrit*
""" Typi ,
OuPont
OuPont
DuPont
Corning
Corning
OuPont
OuPont
Coming
Corning
OuPont
Corning
Corning
OuPont
DuPont
DuPont
AlurtM
AluulM
Thorl* «
! llrcenli 1*.
i Zlreonli
tplntl
IlrconU
iplml
AlartM
AluriM
Zt
11 9/4 VI
_
-
.47/.76/.80
.3./.I4/.17
.M/.Ot/.OO
WS/.31/0
"17.8/5.0
4^1/0/1 ,
_^
-
18/0.7/0 Pt
1.0/t.V
4.1 mo
3.0/0/0 Pt
7.7/9.9/
M.t C*iO]
3.0/3.2/
4.0 (HO
t.0/1.4/
0 Pt
4.S/4.8/
t.inriffl
10.3/1.9/0.8
-
tUtM
4/Z5 - 4/Z9/77
6/M - 8/W77
8/30 - 7/1/77
-
7/10 - 7/11/77
.7^14 - 7/11/77
0/7. - 1/10/77
7/M - 7/J9/77
1/14 - 9/M/77
t/tt - g/t7/77
10/1 - 10/0/77
10/10 - 10/M/7;
1Z/5 - 1Z/JW77
12/28 - 12/29/77
I/2S - 2/6/78
12/30/77 -T/3/7B
Mt. <»
Mt. tjt
Mt,0,t
-
Mt. til
Mt. 0*1
Mt. 041
Mt. Mt
Mt. «»
Mt. Oil
Mt. HI
Mt. Mt
Mt. «l
Mttnint
Mt. Sis'
Mt. CM
Mt. Gi<
Scrtn «r*c* Pt/lr c4U1yit
Scram UOP cMlytt
OotoratM offtctt of MgnMltfng point praclout MUM ind
eittlytt Hfthoiit Mtncott
Mt to bt tnUtf Mtttf at rttnltl of A-027.
InrtitlflU Mtil mid* citalytt eipibHUItt (M
Mthco>t). Ptrfor. nlgb tnp. oo*r»t1m (3100T),
Mt*1 oi1d* ctovirlton, nt*Mln tviliMtlon
Ctnwr* to A-032
fcrttn Mtth*y tlihop utilytt
ru*1 nltrogm ond praitvr* tatting 1 T«t dlfflcullUi
> procluvtu dit*
1 ratultt
CoMM-tmi t* A-033 )
ScrMnliNj ciMp*r1un
Fwl nltragm totting
Fwl nttragm ond pntiim Uttlng
tnv*tt1gtt* ctulyst-support 1nt*rtct1ont
Scrwn Johnion Pktthty cittlytt
ClUlytt toltvp, 6.06-inch dtwcttr
ro
CO
00
-------
TABLE II. SURFACE AREA AND DISPERSION MEASUREMENTS ON GRADED CELL CATALYSTS
ro
to
vo
Sample No.
AERO-025
AERO-026
AERO-027
AERO-OM
AERO-029
AERO-030
AERO-031
AERO- 032
AERO-033
AERO-034
AERO-035
AERO-036
AERO-037
AERO-OM
AERO-MO
AERO- 041
Surface Area **/,
Pretest Post-test
1.55-2.49
5.94
0.44
0.06
0.60
0.15
24.36
11.99
-
5.17
-
0
4.00
6.37
0
0
-
-
0
.01
0
0
-
0.09
-
-
0
0.50
Dispersion 1
Pretest Post-test
1.5-4.9
20.64
8.33
mm
-
36.16
f.TI
-
-
4.09
-
-
-
0
-
-
"
0
0.20
-
-
0
-
-
Catalyst
Pt/Ir
Proprietary
Pt/fr/Os
Pt/Ir/Os
NIO/Pt
CojtyPt
Stab. Pt
Stab. Pt
NIO/Pt-M
COgOj/Pt
Stab. Pt
NtO/Pt
Itjfiym
cozo3
Proprietary
Proprietary
SEH/EDAI Results and CoMtents
No Pt or Ir found on back tw
seojmts by SEH/EOAX
Not combustion tested
Invalid test data
Invalid test data
Zero surface area on each sevaant
-------
TABLE III. FUEL NITROGEN BALANCE
Run
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14-
Catalyst
Type
Pt
Pt
NiO/Pt
WO/Pt
UOP
Fuel-N
Content
(Wt. % of Fuel)
2.
3.
2.
2.
3.
3.
3.
0
0
4
0
6
0
0
Theoretical
A1r
(%)
98
77
200
75
100
90
80
75
70
100
75
90
80
70
Percent Conversion of Nitrogenous
Species in the Combustion Products
NOX
76.00
0.83
80.57
9.63
34.13
1.06
0.58
0.29
1.00
23.77
0.25
0.68
3.28
2.47
N2
15.97
79.05
13.19
58.94
71.54
58.93
79.29
98.68
74.86
68.20
74.01
61.50
54.68
45.60
NH3
8.12
18.07
1.81
41.20
4.77
44.11
25.17
4.49
4.29
5.90
5.36
42.34
37.92
41.91
HCN
1.22
1.51
T.21
3.09
1.52
2.96
2.03
2.59
2. 95
0.72
2.33
4.71
16.79
7.00
N20
0
Error
.(*)'
0.10
-0.56
-3.21
12.86
11.97
6.81
7.07
6.05
-16.91
-1.43
-18.02
9.23
12.66
-3.02
ro
4*
o
-------
SESSION III: SPECIAL TOPICS
BY
Joshua. S. Bowen
241
-------
EPRI LOW COMBUSTION
NO RESEARCH
BY
D. P. Telxeira
Electric Power Research Institute
Palo Alto, California
243
-------
EPRI Low Combustion NOX Research
by
D.P. Teixeira
Electric Power Research Institute
ABSTRACT
Recent results of EPRI's research and development efforts on the
primary combustion furnace (PCF) concept to achieve low NOX with
pulverized coal will be presented. Principal of operation and
earlier 4 million Btu/hr data will be reviewed. Recent data
from nominal 50 million Btu/hr scale tests will be covered.
Full scale application issues related to the PCF will be discussed.
A technical and economic comparison of combustion and post combus-
tion control technologies will be made.
244
-------
FLUE GAS TREATMENT TECHNOLOGY
FOR NO CONTROL
A
By:
J. David Mobley
Process Technology Branch
Utilities & Industrial Power Division
Indr°trial Environmental Research Laboratory
J.S. Environmental Protection Agency
Research Triangle Park, N.C. 27711
245
-------
ABSTRACT
The Environmental Protection Agency has maintained a program to further
the advancement of NO control by flue gas treatment technology since the
2£
early 1970's. The program consists of technology assessment and control
strategy studies in conjunction with small scale experimental projects.
These activities have shown that 90% reduction of NO emissions by selective
JL
catalytic reduction with ammonia has been commercially demonstrated on gas-
and oil-fired sources in Japan, and that such processes are ready for test
application on coal-fired sources. Based on the experience in Japan, success-
ful application of the technology in the U.S. can be expected. Further,
recent assessments indicate that 90% control of both NO and SO emissions
x x
by dry simultaneous control techniques warrants further investigation.
However, wet simultaneous NO /SO removal processes are not currently attrac-
tive. The paper also assesses the need for highly efficient NO control in
X
the U.S. and suggests that, although the nationwide need for flue gas treat-
ment technology has not been established, certain regions will probably
require application of the technology, in the future.
246
-------
INTRODUCTION
Nitrogen oxides (NO ) and sulfur oxides (SO ) in the atmosphere have
X X
been determined to have adverse effects on human health and welfare. To
aid in preventing these adverse effects, the Industrial Environmental
Research Laboratory at Research triangle Park, N. C. (IERL-RTP) is leading
the U.S. Environmental Protection Agency's (EPA's) efforts to develop and
demonstrate NO and SO control technologies for stationary combustion
X X
sources.
Flue gas desulfurization technology has progressed to commerical
application and has achieved 90% control of S09. Although NO control by
£ ' 3£
combustion modification technology has been applied commercially, NO
X
control by flue gas treatment technology has not been utilized in the U.S.
Combustion modification technology reduces NO emissions by approximately
X
50% in a relatively cost effective manner. Flue gas treatment technology
should be able to reduce NO emissions by 90% and has the potential for 90%
X
control of both NO and SO emissions. The focus of this paper will be on
2( 3t
flue gas treatment technology.
EPA is proceeding with small scale NO and NO /SO flue gas treatment
X XX
experimental projects in parallel with technology assessment and control
strategy studies. To save both development time and money, EPA is investi-
gating Japanese technology for potential application to the U.S. coal-fired
situation. Through these actions, the basic foundation will be established
if the technology is required in the U.S. and acceleration of the develop-
ment program becomes necessary.
This paper presents a summary of the past, current, and planned NO
flue gas treatment programs of EPA and an overview of the technology in
Japan.
x
247
-------
EPA'S NO FLUE GAS TREATMENT PROGRAM
A
The thrust of EPA's current NO flue gas treatment (FGT) program is
2W
technology assessment and control strategy studies in conjunction with
small scale experimental projects. The technology assessment and control
strategy studies are mainly paper studies which examine various aspects
of NO and NO /SO control technology, estimate if and when highly
efficient NO control will be needed in the U.S., and assist in determining
the appropriate scale of the experimental projects.
This paper will begin with a discussion of the following technology
assessment studies:
o Assessment of Technology in Japan
o Assessment of Technology for U.S. Application
o Assessment of Technology for an Industrial Boiler New
Source Performance Standard
Results of the control strategy assessment studies will then be
presented:
o Assessment of Point Source Impact on Ambient NO Levels
-' ...... x
o Assessment of the Need for NO Flue Gas Treatment Processes
Subsequently, experimental projects will be discussed:
o Hitachi Zosen NO Pilot Plant
o UOP-Shell NO /SO Pilot Plant
X, J±
Finally, conclusions that can be drawn from these activities will
be identified.
248
-------
TECHNOLOGY ASSESSMENT STUDIES
Technology assessment studies report the rapid advances of the
technology worldwide. These studies, which have concentrated on Japanese
technology, evaluate the various processes and process features for
interested parties in the United States. In addition, the feasibility
of application of the most promising processes to combustion sources in
the United States is addressed from technical, economic, energy, and
environmental standpoints. This effort also includes assessment of the
technology for potential regulatory actions.
Assessment of Technology in Japan
Due to stringent emission standards, Japanese technology for control
of NO and simultaneous control of NO and SO by flue gas treatment
X X X
techniques is more advanced than any other country's. EPA has sponsored
the publication of periodic reports and papers to facilitate the transfer
of information from Japan to the United States. These documents have
been mainly prepared by Jumpei Ando of Chuo University, Tokyo, Japan
(References 1,2,3). An overview of current Japanese technology, both
dry and wet processes for NO and NO /SO control, follows.
X X X
Dry NO Processes (Reference 3)
Jv
Numerous dry process types are being developed. However, selective
catalytic reduction (SCR) processes are the only ones that have achieved
notable success in treating combustion flue gas for removal of NO and
«£
have progressed to t .^ point of being commercially applied. SCR processes
are based on the preference of the reaction of ammonia (NH_) with NO
J 3t
rather than other flue gas constituents. Since oxygen (02) enhances the
reduction, the reactions can best be expressed as:
4NH3 + 4NO + 02 - catalyst , ^ + ^
4NH3 + 2N02 + 02 - CaayS » 3N2 + 6H20 (2)
Reaction 1 predominates since approximately 95% of the NO in
combustion flue gas is in the form of nitrogen oxide (NO). Therefore, a
249
-------
stoichiometric amount of ammonia can be used to reduce NO under ideal
conditions to harmless molecular nitrogen (N_) and water vapor (H~0).
An NH.:NO mole ratio of about 1:1 has typically reduced NO emissions by
3 *
90% with a leak ammonia rate of less than 20 ppm.
The SCR processes are relatively simple, requiring only a reactor,
a catalyst, and an ammonia storage and injection system. Some increase
in boiler fan capacity, or possibly an additional fan, may be necessary
to account for the increase in pressure drop which may be in the range
of 500-700 Pa.
The optimum temperature for the reaction is about 1000°C. However,
the catalyst effectively reduces the reaction temperature to the 300-
450°C range. To obtain flue gas temperatures in this range and to avoid
the requirement for large amounts of reheat, the reactor is usually
located between the boiler economizer and the air preheater. Depending
on system design, the reactor may be located either before or after the
particulate control device.
Many different types of catalyst compositions and configurations
have been developed. Initially, catalysts were developed for flue gases
without particulate and SO concentrations such as from natural gas
firing. For these applications, a catalyst of platinum (Pt) on an
alumina (Al_0_) support material was used. Alumina was poisoned by SO ,
fc J X
particularly SO., in the flue gas. Titanium dioxide (TiO.) was found to
be resistant to SO poisoning and to be an acceptable catalyst carrier
for applications with SO in the flue gas such as from oil or coal
firing. The catalyst metals also tend to react with SO , especially
SO.. Vanadium compounds were found to be resistant to attack from SO.
and promote the reduction of NO with ammonia. Therefore, many SO
3fc 2£
resistant catalysts are based on TiO. and V.O,.. Other active metals are
also used including C, Co, Cr, Cu, Fe, Mn, Mo, Ni, W, their oxides,
sulfates, or combinations thereof. However, the exact compositions and
constituents of most catalysts are proprietary.
250
-------
Reactor configurations also vary with the application, primarily to
accommodate the different particulate concentrations. For natural gas
firing, a fixed, packed bed employing a granular or pellet type catalyst
can be used. However, particulates in the gas stream will plug a fixed,
packed bed.
For applications with moderate particulate concentrations, a moving
bed arrangement can be employed and still utilize the granular or a ring
shaped catalyst. (The catalyst is charged from the top of the reactor
and moves down intermittently or continuously while the flue gas passes
through the catalyst layer in a cross flow direction. The catalyst is
discharged from the bottom of the reactor, is screened to remove particulates,
and can be regenerated to eliminate any contaminants before being returned
to the reactor.) Moving bed reactors can treat gases containing less
than about 200 mg/Nm of particulates and are applicable to oil firing.
They are also applicable to coal firing if an adequate amount of particulates
is removed from the flue gas upstream of the NO reactor. In most
Ji
cases, this would require application of a hot-side electrostatic
precipitator (ESP) which is usually more expensive than the more commonly
used cold-side ESP.
Parallel flow reactors and catalysts were developed to tolerate
relatively high particulate concentrations, such as from coal firing.
(Parallel flow indicates that the direction of flue gas is parallel,
rather than perpendicular, to the catalyst surface.) Parallel flow
designs utilize various catalyst configurations including tubular,
metallic and ceramic honeycombs or plates, and parallel passage reactors.
The selection of the catalyst and reactor configuration will probably
depend on site specific considerations since there are advantages and
disadvantages associated with each design.
Even though much progress has been made in catalysts and reactor
design, some problems still remain. The catalysts may not be resistant
to all contaminants in flue gas. In addition, fine particulates, smaller
than about 1 micron, may blind the catalyst surface. Catalyst life
251
-------
needs to be extended from the current guarantees of 1 to 2 years for
applications with SO and participates in the gas stream.
3t
One of the major concerns with SCR processes is the formation of solid
ammonium sulfate [(NH,)2SO,] and liquid ammonium bisulfate (NH^HSO^). The
problem occurs if S0_, NH-, and H20 are present in sufficient quantities,
and the flue gas cools to the formation temperature of (NH^) 2S04 and NH4HS04
For example, NH.HSO^ will form at 210°C if the flue gas contains 10 ppm of
NH, and 10 ppm of SO-. The formation conditions are difficult to avoid
J J
since one expects some leak ammonia from a SCR system and some SO- from
combustion of sulfur containing fuels. In addition, some catalysts tend to
promote the oxidation of SO- to S0«; however, this conversion is usually
less than 5% of the S02 in the flue gas stream.
The biggeat problem seems to be deposition of (NH,)2SO, and NH.HSO, on
the air preheater. These compounds are highly corrosive and interfere with
heat transfer. The problem appears to be most severe with high sulfur oil
firing. With low sulfur oils, the SO. concentration is too low to cause
problems. Early tests with coal indicate that the (NH.KSO, and NH.HSO, may
deposit on the fly ash or be removed from the heat exchanger surface by the
erosive action of the fly ash. However, more operating experience on coal-
fired sources is needed to quantify the extent of the problem.
Numerous countermeasures have been proposed to minimize the (NH,)2SO,
and NH.HSO, problem. The most desirable techniques are to avoid formation.
This can be accomplished by reducing the SO- and NH- in the reactor effluent
or increasing the exhaust temperature of the flue gas. Soot blowing and
water washing techniques are useful in removing the deposits after formation.
Heat exchanger modifications, which increase the effectiveness of the removal
techniques or minimize the likelihood of deposit, are also being considered
as well as use of corrosion resistant heat exchanger surfaces.
There are numerous developers of SCR processes and numerous pilot,
prototype, and commercial scale applications. Table I lists 13 process
3
developers and 62 applications over 10,000 Nm /hr (normal cubic meters per
hour) along with information on the type of source, fuel, plant size, type
of reactor, and start-up date.
252
-------
The cost of SCR systems varies with process developer and application.
However, the Japan Environment Agency has investigated the costs of a SCR
application to a new oil-fired, 300 MW installation in Japan. To remove 80%
of the NO , the capital investment was estimated to be about $15.5/kW and
2£
the annualized cost, about 1.6 mills /kWh.
EPA is sponsoring a pilot scale evaluation of SCR technology which will
be discussed later in the paper.
Wet NO Processes (Reference 3)
2£
The wet NO processes developed to date cannot compete economically
jt
with dry SCR processes for removal of NO from combustion flue gas. This is
3t
primarily due to the complexity, limited applicability, and water pollution
problems associated with the wet NO processes.
j£
Dry Simultaneous NO /SO Processes (Reference 3)
A A.
Although there are several dry simultaneous NO /SO removal systems,
3C X.
the only commercially demonstrated system is the Shell Flue Gas Treating
system. This process was originally designed for S02 control but was also
found to be adaptable for NO control.
2k
The process uses copper oxide supported on stabilized alumina placed in
two or more parallel passage reactors. The reactions can be expressed as
follows :
CuO + 1/2 02 + S02 - > CuS04 (3)
4NO + 4NH + 0 - ^^ - » 4N + 6H0 (4)
CuS04 + 2H2 - » Cu + S02 + 2H20 (5)
Cu + 1/2 02 - > CuO (6)
Flue gas is introduced at 400° C into one of the reactors where the S0x
reacts with copper oxide to form copper sulfate. The copper sulfate and, to
a lesser extent, the copper oxide act as catalysts in the reduction of N0x
with ammonia. When the reactor is saturated with copper sulfate, flue gas
is switched to a fresh reactor for acceptance of the flue gas, and the spent
reactor is regenerated. In the regeneration cycle, hydrogen is used to
reduce the copper sulfate to copper, yielding a S02 stream of sufficient
253
-------
concentration for conversion to sulfur or sulfuric acid. The copper in the
reactor is oxidized, preparing the reactor for acceptance of the flue gas
again. Between acceptance and regeneration, steam is injected into the
reactor to purge the remaining flue gas or hydrogen to eliminate any pos-
sibility of combustion. The process can also be operated in the NO -only
2t
mode by eliminating the regeneration cycle, or in the SO -only mode by
X
eliminating the ammonia injection.
The process has been installed at the Showa Yokkaichi Sekiyu plant in
Yokkaichi by Japan Shell Technology, Ltd. on a heavy-oil-fired boiler treating
3
120,000 Nm /hr of flue gas. In this installation, the process removes 90%
of the SO, and 40% of the NO to meet local regulations. The unit has
£* &
demonstrated 90% SO. removal and 70% NO removal.
£, Jt
UOP Process Division is the licensor of the process in the United
States. Hence, the process will be referred to as the UOP-She 11 process in
this paper. EPA is sponsoring a pilot scale evaluation of the UOP-Shell
process which will be discussed later.
Wet Simultaneous NO /SO Processes (Reference 3)
Ji A
Although the wet NO removal processes cannot compete economically with
dry NO processes, wet simultaneous NO /SO processes may be competitive
X XX
with the sequential installation of NO control by SCR followed by S0»
A te
control by flue gas desulfurization (FGD).
The first wet simultaneous NO /SO systems, called oxidation/absorption/
reduction processes, evolved from FGD systems. Since the NO in flue gas is
fairly insoluble in aqueous solutions, a gas-phase oxidant is injected
before the scrubber to convert NO to the more soluble nitrogen dioxide
(N02). The absorbent varies with the type of FGD system being modified.
The absorbed S02 forms a sulfite ion which reduces a portion of the absorbed
NO to molecular nitrogen. The remaining NO are removed from the waste
x
water as nitrate salts. The remaining sulfite ions are oxidized into
sulfate by air. and removed as gypsum.
The oxidation/absorption/reduction processes have the potential to
remove 90% of both SO and NO from combustion flue gas. However, there
X. X
several drawbacks remaining to be overcome before the processes can be
254
-------
widely applied. The process chemistry is complex, and the use of a gas-
phase oxidant, such as ozone (0_) or chlorine dioxide (C10-), is expensive.
Chlorine dioxide, although cheaper than ozone, adds to the waste water
problems created by the nitrate salts.
Absorption/reduction processes were seemingly developed to avoid the
use of a gas-phase oxidant. A chelating compound, which has an affinity for
the relatively insoluble NO, is added to the scrubbing solution. Ferrous-
EDTA (ethylene-diamine tetraacetic acid) has typically been used as the
chelating compound. The NO is absorbed into a complex with the ferrous ion
and the SO- is absorbed as the sulfite ion. The NO complex is reduced to
molecular nitrogen by reaction with the sulfite ion. A series of regeneration
steps recovers the ferrous chelating compound and oxidizes the sulfite ion
into sulfate which is removed as gypsum.
The absorption/reduction processes also have the potential to remove
90% of both the NO and SO in combustion flue gas. Although the processes
X X
seem to have advantages over the oxidation/absorption/reduction processes,
there are obstacles to be overcome before the processes can be widely applied.
Even with the addition of the chelating compounds, a large absorber is
required to absorb the NO. The sum of replacement, recovery, and regeneration
costs of the chelating compounds, although potentially less than using the
gas-phase oxidants, are still significant. The process chemistry is complex
and is sensitive to the flue gas composition of S02> NO , and oxygen. The
molar ratio of S09 to NO must remain above approximately 2.5 and the oxygen
£ Ji
concentration must remain low.
Table 11 lists the process developers evaluating wet simultaneous
NO /SO technology in Japan. Oxidation/absorption/reduction processes are
3t X
being investigated by six process developers at nine different sites. For
absorption/reduction processes, there are four pilot or bench scale plants
currently being operated by four different process developers.
As shown in Table III, based on information from the Japan Environment
Agency, it was estimated that a FGD system would have an investment cost of
about $77.8/kW with an annualized cost of 6.7 mills/kWh. As stated previously,
255
-------
the cost of a SCR system is $15.5/kW and 1.6 mllls/kWh. However, the cost
of a sequential system employing SCR + FGD would be about $102.8/kW and 8.3
mills/kWh. The total cost increased slightly over the sum of individual SCR
and FGD systems to allow for removal of NH« from the FGD scrubbing liquor.
The cost of simultaneous NO /SO , both wet and dry, was estimated to be
X X
slightly higher or about $lll.l/kW and 9.4 mills/kWh; however, these cost
estimates were more uncertain than the FGD and SCR costs. Costs for all
systems were estimated to be about 20% higher for coal-fired sources.
Assessments of Technology for U.S. Application
The Tennessee Valley Authority (TVA), through an interagency agreement
with EPA, is developing comparative economics of NO and NO /SO flue gas
* X X X
treatment processes for application in the U.S. This multi-phase study is
over 50% complete. Phase I, a technical assessment, was an evaluation and
summary of the technical feasibility of all candidate NO control processes
X
being offered in the U.S. and Japan. Phase II, a preliminary economic
assessment, concentrated on the processes recommended in Phase I for further
study. Phase III, a definitive economic assessment, will develop detailed
engineering evaluations and cost estimates for the most promising NO
**
control cases identified in Phase II. In addition, the impact of application
of SCR processes on ammonia availability and cost was evaluated. Phases I
and II of the study were cofunded by the Electric Power Research Institute
and EPA. "
Phase I Technical Assessment of NO and NO /SO Processes (Reference 4)
2t X XI
In Phase I, 42 processes for NO and simultaneous NO /SO control were
X XX
described. The discussion on each process included the process description,
status of development, reported economics, utility and raw material require-
ments, technical and environmental considerations, and advantages and disad-
vantages.
In addition, a comparison was made of the dry NO and wet simultaneous
* A
NO /SO systems. It was found that, although there are many different types
*t x^
of dry and wet processes, in most cases the dry processes have advantages
over the wet processes:
256
-------
o Low projected capital investment and annual revenue
req uirement s.
o Simple process with few equipment requirements.
o High NO removal efficiency (>90%).
2£
o Extensive tests in large units.
o No waste stream generation.
However, the dry systems also have disadvantages:
o Sensitive to inlet particulate levels.
o Require ammonia.
o Possible emission of ammonia and ammonium sulfates and
bisulfates.
o Relatively high reaction temperatures (350-400°C).
The wet NO /SO removal processes have certain general advantages and
disadvantages, compared with the dry NO systems. The major advantages
Jv
include:
o Potential economic advantage of simultaneous NO /SO
removal. x x
o Relatively insensitive to flue gas particulates.
o High S0_ removal (>95%).
On the other hand, major disadvantages of the wet systems include:
o Expensive processes due to process complexity and insolubility
of NO in aqueous solutions.
2S
o Formation of nitrates (NO- ) and other potential water
pollutants.
o Extensive equipment requirements.
o Formation of low-demand byproducts.
o Flue gas reheat required.
o Only moderate NO removal.
J x
o Limited application of some processes due to requirement
for high SO :NO ratios.
X X
In addition to being a state-of-the-art review of all NO processes
undergoing development, one of the main purposes of the study was to identify
processes for further evaluation in Phase II of the study. Three criteria
were used to screen the processes: technical considerations, developmental
257
-------
status, and representative samples of available technologies. Based on
these criteria, the nine processes in Table IV were selected for further
study.
Phase II Preliminary Economic Estimates (Reference 5)
The primary purpose of Phase II of the study was to compare the process
economics of the different types of NO and NO /SO control techniques on a
X JL X.
consistent basis. Preliminary economics, including total capital investment
and average annual revenue requirements, were determined. Rigorous compari-
sons were not made although various conclusions can be drawn from the analysis.
The processes listed in Table IV were evaluated except for the wet NO -
only process; development of that process was terminated by the developer.
However, there appears to be little, if any, future for application of wet
NO -only processes to combustion sources. The JGC Paranox process, a dry SCR
JW
process (NO only) employing a parallel passage reactor, was substituted.
A.
Process descriptions, detailed flowsheets, material balances, and a
current status of development summary were prepared for each process based
on data supplied by the process vendors. Equipment descriptions and costs
were prepared from the available data for each process.
Although the study is not complete, preliminary estimates are indicated
in Table V. (It must be emphasized that these numbers are subject to change
when the study is completed.) However, it is apparent from these estimates
that the IHI process, an ozone based system, is not economically attractive.
(This finding is consistent with an earlier EPA study (Reference 6) that
determined that ozone systems were too expensive for coal-fired applications.)
The Moretana Calcium process, a chlorine dioxide based system, and the Asahi
process, an EDTA based system, also appear economically unattractive when
compared to the least expensive SCR + FGD system, especially based on annual
revenue requirements.
Given the uncertainties of the estimates, the comparison of the UOP-
Shell dry NO /SO process and the SCR + FGD system is inconclusive. In
addition, it is not apparent which SCR system, moving bed or parallel flow,
represents the optimum NO -only configuration. The cost estimates for SCR
'
258
-------
processes include the cost of particulate control by an ESP. The ESP is
roughly 40% of the capital cost, which places NO control alone at approxi-
X
mately $36/kW.
The study also evaluated energy consumption of the processes, and
estimates are given in Table VI. The SCR systems are projected to require
less than 0.3% of the boiler capacity. The UOP-Shell, dry NO /SO process
X X
requires the least energy of any NO and SO system due to reclamation of
3C X
heat from the byproduct streams. The combination of SCR and FGD is next;
the wet NO /SO systems require significantly higher quantities of energy.
A X.
The economic sensitivity of the SCR systems to the cost of ammonia has
been the subject of some concern. It was found that the annual ammonia cost
was below 12% of the average annual revenue requirements for the dry systems.
Therefore, the average annual revenue requirement may be expected to increase
about 10% if the ammonia cost is doubled.
Phase III Definitive Economic Estimates
In Phase III of the study, to begin immediately following completion of
Phase II, a definitive-level design and economic evaluation of the leading
NO control techniques will be performed. The study will encompass both
X.
flue gas treatment and combustion modification techniques, which will be
analyzed individually as well as in combination. This will enable a compari-
son of 90% NO control by flue gas treatment alone and combustion modification
3t
followed by flue gas treatment. In addition, dry simultaneous NO /SO
X X
systems will be evaluated against the optimum highly efficient NO control
X
system used in conjunction with a conventional FGD unit.
In the course of the study, the bases for conceptual designs will be
identified, a definitive engineering package will be prepared, and capital
investment and revenue requirements will be estimated. The final report on
Phase III is scheduled to be available in early 1980.
Impact of Ammonia Utilization (Reference 7)
A major concern of the widespread utilization of ammonia-based NO
2v
control systems has been the impact on the domestic ammonia market. It has
been speculated that severe disruptions in the availability and price of
259
-------
ammonia would occur if SCR systems are employed; this could also impact on
the availability and cost of ammonia-based fertilizers. In addition, the
impact on natural gas, the primary feedstock for ammonia production, was
unclear. A study of the situation was undertaken by TVA to address these
concerns in parallel with the three-phase technical and economic evaluation
discussed above.
The annual NH_ requirements for a 500 MW coal- and oil-fired boiler
were calculated for application of a 90% NO removal system based on SCR
Ji
technology. At an assumed NH,:NO mole ratio of 1.05:1, the NH, consumption
«J 3t J
rates were 5.7 Gg/yr (6300 tons/yr) and 1.6 Gg/yr (1800 tons/yr) for the
coal- and oil-fired units, respectively.
From these calculations it is apparent that even a large coal-fired
power plant could not justify a, captive NH_ plant since the minimum plant
capacity for economic production of NH« is about 300 Gg/yr (330,000 tons/yr).
Therefore, NH« would be purchased from an off-site NEL plant and stored in
tanks at the power plant.
The projected U.S. electrical generating capacity was estimated for the
period from 1975 to 2000. Assuming application of a SCR system removing 90%
of the NO to all new large industrial arid 'utility boilers (>260 GJ/hr or
gf
250 x 10 Btu/hr) beginning in 1985, the NH, demand for NO control was
determined. The ammonia supply to meet both the conventional NH, demand and
annual growth of this demand for NO control was projected using an assumed
annual growth rate for ammonia production of 3.0%. It was found that the
NH- demand for NO control ranged from 0.6% of the total NH, supply in 1985
OX j
to 20-25% in the year 2000.
Thus, although requiring substantial amounts of NH, in the future, the
need for NH~ for NO control apparently would not have an abrupt adverse
impact on the availability and price of NH_. Under the assumptions of the
study, the primary Impact on the domestic NH, market would be to cause the
U.S. NH_ demand to increase at 4.5%/yr during the period 1985-2000 rather
than the assumed 3.0%. This increase could be met with the addition of one
260
-------
ammonia plant per year. In addition, although use of natural gas is counter
to the National Energy Plan, utilization of NH~ for NO control should not
J X
adversely affect the price ,and availability of natural gas.
Assessment of Technology for an Industrial Boiler NSPS
IERL-RTP is sponsoring a series of studies to determine the applica-
bility of various emission control technologies to industrial boilers. The
primary purpose of these studies is to support the development of a new
source performance standard (NSPS) for industrial boilers. The following
technologies are being considered:
o Oil Cleaning and Existing Clean Oil
o Coal Cleaning and Existing Clean Coal
o Synthetic Fuels
o Fluidized Bed Combustion
o NO Combustion Modification
x
o NO and NO /SO Flue Gas Treatment
X XX
o Flue Gas Desulfurization
o Particulate Control
Following completion of reports on the individual technologies, a
comprehensive assessment report will be prepared which discusses how the
technologies should be integrated for optimum multi-pollutant control.
Subsequently, EPA's Office of Air Quality Planning and Standards (OAQPS)
will study the impact of various control options. These studies will
ultimately lead to the establishment of a NSPS for industrial boilers.
Acurex Corporation is the contractor assisting OAQPS and IERL-RTP in this
overall effort.
As shown above, NO and NO /SO flue gas treatment processes are among
X XX
the technologies being considered. Radian Corporation is preparing the
technology assessment report on these processes. The report will contain
the following sections:
o Emission Control Techniques
o Candidates for Best System of Emission Reduction
o Control System Costs
261
-------
o Energy Considerations
o Environmental Considerations
o Emission Source Test Data
The cost, energy, and environmental considerations will be based on
standard boilers listed in Table VII. The coal-fired boilers will be
analyzed for high and low sulfur Eastern coal and low sulfur Western coal
cases.
The final report on the assessment of NO and NO /SO technologies
3x A. X,
should be available in the spring of 1979. The comprehensive assessment
report on optimum multi-pollutant control systems should be available in the
Fall of 1979. (Note that consideration of a technology for a standard does
not imply that the ultimate standard will be based on that technology.)
CONTROL STRATEGY ASSESSMENT STUDIES
While the primary purpose of the technology assessment studies is to
determine the performance of the technology, the primary objective of the
control strategy assessment studies is to determine if and when the tech-
nology will be needed in the U.S. Recent studies in this area have utilized
computer modeling techniques to evaluate alternative control strategies and
investigated the issues (such as prevention of significant deterioration)
which could eventually require increased NO control.
X
Assessment of Point Source Impact on Ambient N0? Levels (Reference 8)
A study was undertaken for EPA by Radian Corporation to determine the
impact of various stationary source NO control strategies on attaining and
JL
maintaining the National Ambient Air Quality Standard (NAAQS) for N0_. The
Chicago Air Quality Control Region (AQCR) was selected for use in this study
since it has historically encountered ambient NO problems.
3£
Annual Impact
The original purpose of the study was to determine the effect on
annual average ambient NO- levels of applying NO control technology to
6 ^
large point sources (> 105 GJ/hr or 100 X 10 Btu/hr) in the AQCR. A
262
-------
dispersion model was used to relate NO emissions to ambient NO- concentrations
3C £
in Chicago.
It was found that* although the major point sources account for nearly
40% of the total NO emissions in Chicago, they account for less than 10% of
Jv
the ambient N0_ levels, on the average. Considering major point source
"hotspots" (i.e., localized areas of the city where major point source
impact is the greatest), modeling results indicate that these sources account
o
for 12% of a predicted N02 level of about 60 yg/m . Taking a worst-case
approach and assuming that all NO emissions from major point sources are
3£
converted to NO., it was found that the predicted cumulative impact of all
major point sources at locations of maximum annual impact is still only 15%
of the standard.
Therefore, it was concluded that total removal of large point source
NO emissions would result in only a small improvement in annual average NO-
X fc
air quality in Chicago, and control of large point sources alone would not
be adequate to achieve and maintain the annual average NAAQS.
Short-term Impact
Based on the Clean Air Act Amendments of 1977 and the possible estab-
lishment of a short-term NAAQS, this Radian study was expanded to investi-
gate the effect of NO control techniques on short-term ambient concentra-
JH
tions in the AQCR for present and future years. Standards were assumed of
3
250, 500, 750, and 1000 ug/m of N02 based on a 1 hour average.
The short-term impact assessment was made using Gaussian-type disper-
sion models. A significant part of the effort of this study was directed
toward defining the short-term NO emission rates that should be used in the
2£
model. This was done primarily by adjusting the annual average emission
rates. Adjustments were made for season of the year, day of the week, and
time of day. The entire emissions inventory for the Chicago AQCR, including
vehicular and other area sources, was modeled using this approach.
The computer-predicted ambient NO concentrations were converted to
2t
ambient NO- concentrations by applying a ratio of NO- to NO determined from
2 fc X
measured air quality data in Chicago. This ratio is a function of season of
263
-------
the year and time of day; ratios used in the study varied from 0.25 to 0.5.
The accuracy of this approach is not known since several photochemical
reactions are involved in the conversion of NO to N0_ and other species.
2£ fc
For assessment of future year impacts, the growth in NO emissions was
it
estimated.
The study found that individual large point sources may account for 60%
3
of a predicted 1 hour N0? concentration of 1100 yg/m in industrial areas
3
and 90% of a level of 800 yg/m in non-industrial areas. This indicates
that controlling large point sources may provide significant improvements in
short-term N02 air quality. However, the degree of control required is
highly dependent on any short-term N0? NAAQS adopted by EPA. The results
summarized in Table VIII show the percentage of the 14 largest existing
point sources which would require controls for various standard levels if
those standards were currently in effect. It was assumed that the more cost
effective combustion modification techniques would be employed first, and
that the flue gas treatment techniques would only be applied if necessary to
maintain the established ambient level.
The percentages of Table VIII are based on individual large point
source impacts added to the impacts of other point sources, vehicular sources,
and non-vehicular area sources. When large point sources are located near
each other so that their Impacts interact, the degree of control required
increases significantly and more flue gas treatment is required.
Projections for NO emissions to 1985 indicate that the same level of
JH
control as shown in Table VIII would be required. There are two reasons for
this unexpected result. First, the highest predicted short-term concentra-
tions are dominated by large point sources to the extent that changes in the
impacts from other sources do not make a large difference. Second, the
change in impact of other sources by 1985 is small because increases in non-
vehicular emissions are counterbalanced by the decrease in projected vehicular
emissions.
It is concluded that control of large point source NO emissions.would
X
result in a significant improvement in short-term N02 air quality and may be
necessary to attain and maintain compliance with a short-term NAAQS.
264
-------
Assessment of the Need for NO Flue Gas Treatment Processes (Reference 9)
^""^^«^_«H^_^^«^__^«^^^^^__^__
To date, EPA's strategy for controlling impacts of NO emissions has
3t
focused on combustion modification technology. This approach has been taken
since these techniques represent the most cost effective approach to achieving
initial reductions in NO emissions. However, it is uncertain whether the
2£
NO emission reductions attainable by use of combustion modification techniques
alone can continue to provide the margin of control necessary to meet NO
3£
ambient air quality standards.
In parallel with the Chicago study discussed above, another study was
undertaken by Radian Corporation to determine if and when the application of
NO flue gas treatment technology would be necessary in the U.S. The report
X
of this study (Reference 9) addressed factors which will influence the
levels of NO emission control needed to comply with both existing and
j£
future NO standards. Topics treated include NO emission sources, nationwide
Ji 2t
trends, regional emission profiles, and atmospheric transport and reactions
of NO . Also addressed were current NO regulations and trends in NO
X X ° X
legislation: National Ambient Air Quality Standards including the possible
short term N0_ standard, New Source Performance Standards, Mobile Source
Standards, Prevention of Significant Deterioration, and Nonattainment Provi-
sions. In addition, other uncertainties were assessed such as the oxidant
problem, health effects research, and the nitrosoamines issue. Further, the
major NO emission control alternatives control of mobile sources and
control of stationary sources by combustion modification and flue gas treat-
ment were evaluated.
The study concluded that the number of AQCRs with NO compliance problems
2k
can be expected to increase significantly in the next decade. It was further
concluded that progressively larger reductions in NO emissions will be
2£
required in order to attain and maintain compliance in "problem" AQCRs. The
study does not establish conclusively whether or not flue gas treatment
technology will be required. However, current trends indicate that the
technology may be necessary in the future to achieve compliance with NO
X
standards in specific AQCRs. This conclusion follows from the regionally
specific nature of U.S. NO compliance problems as well as uncertainties
265
-------
concerning both future NO emission reduction requirements and the ultimate
effecti
cation.
effectiveness of alternative NO control methods such as combustion modif i-
x
EXPERIMENTAL PROJECTS
The technology assessment and control strategy studies and financial
constraints aid in determining the number and scale of experimental projects.
These studies, as well as previous experimental projects, supported a jump
to pilot scale evaluations of NO and NO /SO processes on a coal-fired
3t j£ 2£
application. Previous experimental projects included laboratory projects on
development of catalysts for NO reduction with ammonia (References 10, 11),
m
a pilot-scale project evaluating a selective catalytic reduction process on
a gas-fired application (Reference 12), and a laboratory evaluation of NO
2C
reduction with metal sulfidea (Reference 13).
After solicitation and evaluation of proposals from interested parties,
EPA awarded two contracts in May 1978 for the pilot-scale evaluation of flue
gas treatment technology on a coal-fired source. One contract went to
Hitachi Zosen for demonstration of NO removal and the other went to UOP
Process Division for demonstration of simultaneous NO /SO removal.
x x
Both pilot plant projects have similar objectives and scope. The basic
objective is to demonstrate the feasibility of the processes for highly
efficient control of NO or NO /SO on a coal-fired source. Pollutant
Jt 2*. Jt
removal efficiencies of 90% are expected.
The scope of the projects is divided into four phases:
I - Design
II - Procurement and erection
III - Startup, debugging, and optimization
IV - Long term operation and assessment.
The design phases were completed within approximately 3 months of
contract award on both projects. Procurement and erection, Phase II, is
expected to require about 9 months to complete. Phase III, which will
require 4-8 months to complete with the majority of time spent on optimiza-
266
-------
tion testing, will contain a parametric study of process variables including:
flow rate, temperature, NH.:NO ratio, and NO , SO, S0~, and particulate
concentrations. Long-term operation and assessment will be conducted during
a 90 day continuous run in Phase IV.
Project manuals will be available on both projects in early 1979. The
final reports on the results of the projects should be available in mid-
1980.
Hitachi Zosen NO Pilot Plant (Reference 14)
Hitachi Zosen1s NO removal process is based on SCR technology. Since
these processes were discussed previously in the paper, only key process
features will be discussed further.
Hitachi Zosen utilizes a metallic honeycomb catalyst arrangement for
applications with high particulate and SO concentrations. The reactor is
located between the boiler economizer and the air preheater and in front of
any particulate control device. The reaction temperature is in the range of
350-420°C. The pressure drop is about 200-500 Pa. For 90% NO reduction, a
NH3:NO mole ratio of 1:1 is used.
The process will be evaluated on a pilot plant scale (^2000 Nm /hr) on
a coal-fired source. The host site for the pilot plant will be Georgia
Power Company's Plant Mitchell near Albany, Georgia. Flue gas will be
obtained from Unit 3, a pulverized coal-fired Combustion Engineering boiler
with a nameplate rating of 125 MW.
Envirotech/Chemico Air Pollution Control Corporation (CAPCC), Hitachi
Zosen1s American licensee, will be the major subcontractor on the project.
CAPCC will provide detailed engineering, design, procurement, erection, and
operation of the pilot plant in cooperation with Hitachi Zosen. Fabrication
and procurement will be done within the U.S. as far as possible.
UOP-Shell NO /SO Pilot Plant (Reference 15)
^^^^^^^""^^^^~" J^^^^^^^^^^^"^^^^^^^^"*^"!^^^^
The UOP-Shell dry NO /SO removal system is based on the Shell Flue Gas
Treating process which has also been discussed previously in the paper.
Briefly, the process employs a parallel passage reactor using CuO as the
267
-------
sorbent for SO- and CuSO, as the catalyst for the reduction of NO with NH-.
fc f X J
The reactor is located between the boiler economizer and the air preheater
and in front of any particulate control device. The process operates at a
temperature of 400°C during both acceptance and regeneration cycles. The
cycles will be controlled for 90% removal of both NO and S09. For 90% NO
X £ X
reduction, a NH,:NO mole ratio of 1:1 is used.
2
The process will be evaluated on a pilot plant scale (^2000 Nm /hr) on
a coal-fired source. An existing pilot plant, previously used by UOP to
evaluate SO. removal, will be modified for simultaneous removal of NO and
t~ " X
SO-. The host site for the pilot plant will be Tampa Electric Company's Big
Bend Station in North Ruskin, Florida. Flue gas can be obtained from Unit 1
or 2, which are Riley-Stoker pulverized coal-fired boilers with a nameplate
rating of 400 MW each.
268
-------
CONCLUSIONS
This paper has summarized the status of NO and NO /SO flue gas
j£ Jt J\
treatment technology and presented the results of recent EPA studies of the
technology. The following conclusions can be drawn.
1. Dry NO processes, based on selective catalytic reduction (SCR) of NO
X X
with ammonia, can remove 90% of the NO from the flue gas of combustion
A
sources.
2. SCR processes have been extensively demonstrated on commercial-scale
gas- and oil-fired sources in Japan.
3. Based on the experience in Japan, SCR processes can be expected to be
successfully applied to gas- and oil-fired sources in the U.S.
4. Reactors and catalysts for SCR systems have been developed to tolerate
high SO and particulate concentrations and are ready for test applica-
j£
tion on coal-fired sources in the U.S. and Japan.
5. The formation of ammonium sulfate and bisulfate remains a problem for
SCR systems, but solutions are being investigated.
6. Wet processes cannot economically compete with SCR processes for
control of NO .
-------
10. The widespread application of SCR processes is not expected to have an
abrupt adverse impact on the domestic ammonia market.
11. NO and NO /SO flue gas treatment processes warrant consideration as
X X X
the basis for a New Source Performance Standard for industrial boilers.
12. Large point sources do not have a significant impact on annual average
ambient NO- levels.
13. Large point sources may have a significant impact on short term ambient
NO- levels.
14. NO flue gas treatment processes may be needed to attain and maintain
X
compliance with EPA ambient standards in certain Air Quality Control
Regions.
15. EPA's pilot plant project with Hitachi Zosen will enable evaluation of
a SCR process for 90% control of NO on a coal-fired source.
16. EPA's pilot plant project with UOP Process Division will enable evalua-
tion of a dry simultaneous NO /SO process for 90% control of both NO
A A, X
and S0_ on a coal-fired source.
17. EPA experimental projects, in conjunction with technology assessment and
control strategy studies, will enable an assessment of the feasibility of
NO and NO /SO flue gas treatment processes for application in the U.S.
X XX
270
-------
REFERENCES
1. Ando, Jumped., Katsuya Nagata, and B. A. Laseke. NO Abatement for
Stationary Sources in Japan. PEDCO Environmental, Inc., EPA-600/7-77-
103b (NTIS No. 276-948), September 1977. U.S. Environmental Protection
Agency, Research Triangle Park, N.C.
2. Ando, Jumpei. SO., Abatement for Stationary Sources in Japan. EPA-
600/7-78-210, November 1978. U.S. Environmental Protection Agency,
Research Triangle Park, N.C.
3. Ando, Jumpei, and Katsyua Nagata. NO Abatement for Stationary Sources
in_Japan. (Draft Report; to be published January 1979 by EPA.)
A. Faucett, H. L., J. D. Maxwell, and T. A. Burnett. Technical Assessment
of NO Removal Processes. Tennessee Valley Authority, EPA-600/7-77-127
(NTIS^No. 276-637, EPRI No. AF-568, TVA No. Y-120), November 1977.
U.S. Environmental Protection Agency, Research Triangle Park, N.C.
5. Burnett, T. A., J. D. Maxwell, and H. L. Faucett. The Preliminary
Economics of Alternative NO Flue Gas Treatment Processes. Tennessee
Valley Authority. (12/78 Draft Report; to be published early 1979 by
EPA and EPRI.)
6. Harrison, J. W. Technology and Economics of Flue Gas NO Oxidation
by Ozone. Research Triangle Institute, EPA-600/7-76-033 (NTIS 261
917), December 1976. U.S. Environmental Protection Agency, Research
Triangle Park, N.C.
7. Burnett, T. A., and H. L. Faucett. Impact of Ammonia Utilization
by NO Flue Gas Treatment Processes. Tennessee Valley Authority,
EPA-600/7-79-011, January 1979.uTs. Environmental Protection
Agency, Research Triangle'Park, N.C.
8. Eppright, B. R., E. P. Hamilton, III, M. A. Haecker, and Carl-Heinz
Michelis. Impact of Point Source Control Strategies on N02 Levels.
Radian Corporation, EPA-600/7-78-212, November 1978. U.S. Environmental
Protection Agency, Research Triangle Park, N.C.
9. Corbett, W. E., G. D. Jones, W. C. Micheletti, R. M. Wells, and G. E.
Wilkins. Assessment of the Need for NO Flue Gas Treatment Technology.
Radian Corporation, EPA-600/7-78-215, November 1978.U.S. Environmental
Protection Agency, Research Triangle Park, N.C.
271
-------
10. Nobe, K., G. L. Bauerle, and S. C. Wu. Parametric Studies of Catalysts
for NO Control from Stationary Power Plants. University of California,
Los Angeles, EPA-600/7-76-026 (NTIS No. PB 261 289), October 1976.
U.S. Environmental Protection Agency, Research Triangle Park, N.C.
11. Koutsoukos, E. P., J. L. Blumenthal, M. Ghassemi, and G. L. Bauerle.
Assessment of Catalysts for Control of NO from Stationary Power Plants.
TRW, Inc., EPA-650/2-75-001a (NTIS No. PB£239 745), January 1975, U.S.
Environmental Protection Agency, Research Triangle Park, N.C.
12. Kline, J. M., P. H. Owen, and Y. C. Lee. Catalytic Reduction of Nitrogen
Oxides with Ammonia: Utility Pilot Plant Operation. Environics, Inc.,
EPA-600/7-76-031 (NTIS No. PB 261 265), October 1976. U.S. Environmental
Protection Agency, Research Triangle Park, N.C.
13. McCandless, F. P., and Kent Hodgson. R^duc^ion oj^_NitricJPxide with Metal
Sulfides. Montana State University, EPA-600/7-78-213, November 1978.
U.S. Environmental Protection Agency, Research Triangle Park, N.C.
14. Wiener, R. S., and Rafat R. Morcos, Project Manual; Evaluation of Hitachi
Zosen NO Flue-Gas Treatment Process. Chemico Air Pollution Control
Company. (Draft Report; to be published January 1979 by EPA.)
15. Pohlenz, J. B., and Nooy, F. M. Project Manual; Evaluation of UOP-
Shell NO^/SO Flue Gas Treatment Process. UOP Process Division.
(Draft Report; to be published in January 1979 by EPA.)
272
-------
TABLE I. SELECTIVE CATALYTIC REDUCTION PLANTS IN JAPAN
ro
^i
CO
Process Developer
Asahi Glass
Hitachi, Ltd.
Hitachi, Ltd.
Hitachi, Ltd.
Hitachi, Ltd.
Hitachi, Ltd.
Hitachi, Ltd.
Hitachi, Ltd.
Hitachi, Ltd.
Hitachi, Ltd.
Hitachi, Ltd.
Hitachi, Ltd.
Hitachi Zosen
Hitachi Zosen
Hitachi Zosen
Hitachi Zosen
Hitachi Zosen
Ishikawaj ima H.I.
Ishikawaj ima H.I.
Ishikawaj ima H.I.
Ishikawaj ima H.I.
Ishikawaj ima H.I.
Ishikawaj ima H.I.
Ishikawaj ima H.I.
JGC
JGC
JGC
Kobe Steel
Kurabo
Mitsubishi H.I.
Mitsubishi H.I.
Gas
Source
Furnace
Coke Oven
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Furnace
Furnace
Furnace
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Furnace
Boiler
Coke Oven
Coke Oven
Boiler
Boiler
Boiler
Fuel3
BIBTT,
COG
HO(HS)
Kerosene
HO(LS)
HO(LS)
Crude Oil
LNG
HO
Kerosene
HO(LS)
HO(LS)
CO
HO(HS)
Kerosene
Coke, Oil
HO
Crude Oil
HO(LS)
HO(LS)
HO(LS)
Crude Oil
HO(LS)
HO(LS)
HO (MS)
CO
COG
COG
HO(HS)
HO(LS)
LNG
Capacity
(Nm3/hr)
70,000
500,000
15,000
16,000
20,000
19,000
300,000
2,000,000x2
20,000
30,000
490,000
550,000
350,000
440,000
71,000
762,000
10,000
20,000
180,000
960,000
480,000
1,000,000
1,900,000
1,660,000
50,000
70,000
150.000
104,000
30,000
300,000
15,000x2
Type of
Reactor^
1MB
1MB
1MB
FB
FB
FB
FB
FB
1MB
FB
PPC
PPC
FB
FB
FB
FB
FB
HC
HC
HC
HC
HC
HC
HC
PPR
PPR
PPR
1MB
CMB
1MB
FB
(Continued)
Start-up
mmm-
Nov. 1976
Oct. 1977
Oct. 1977
Aug. 1977
July 1977
July 1977
Apr. 1978
Apr. 1978
Dec. 1978
June 1978
June 1978
Oct. 1975
Nov. 1975
May 1976
Nov. 1976
Dec. 1976
Apr. 1977
Jan. 1978
Apr. 1978
June 1978
Apr. 1979
July 1979
Aug. 1981
Nov. 1975
July 1976
Mar. 1977
Aug. 1977
Aug. 1975
Sept. 1976
Dec. 1976
-------
TABLE I (Continued). SELECTIVE CATALYTIC REDUCTION PLANTS IN JAPAN
ro
Process Developer
Mitsubishi H.I.
Mitsubishi H.I.
Mitsubishi H.I.
Mitsubishi H.I.
Mitsubishi H.I.
Mitsubishi H.I.
Mitsubishi Kakoki
Mitsubishi Kakoki
Mitsubishi Kakoki
Mitsubishi Kakoki
Mitsubishi Kakoki
Mitsui Engineering
Mitsui Engineering
Mitsui Toatsu
Mitsui Toatsu
Mitsui Toatsu
Mitsui Toatsu
Mitsui Toatsu
Sumitomo Chemical Co.
Sumitomo Chemical Co.
Sumitomo Chemical Co.
Sumitomo Chemical Co.
Sumitomo Chemical Co.
Sumitomo Chemical Co.
Sumitomo Chemical Co.
Sumitomo Chemical Co.
Sumitomo Chemical Co.
Sumitomo Chemical Eng.
Sumitomo Chemical Eng.
Sumitomo Chemical Eng.
Sumitomo Chemical Eng.
Gas
Source
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Furnace
Boiler
Boiler
Boiler
Boiler
Boiler
Furnace
Furnace
Furnace
Furnace
Boiler
Boiler
Furnace
Furnace
Furnace
Furnace
Furnace
Furnace
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Fuel3
HO(LS)
HO(LS)
LNG
HO(LS)
HO(LS)
HO(LS)
HO(HS)
Naphtha
Kerosene
Kerosene
Kerosene
CO
HO(LS)
Off Gas
Off Gas
Off Gas
Off Gas
Off Gas
HO(HS)
LPG
LPG
LPG
LPG
LPG
LPG
HO (MS)
HO (MS)
Naphtha
Naphtha
Naphtha
Naphtha
Capacity
(Nm3/hr)
40,000
200,000
1,690,000x2
1,010,000
490,000
1,920,000
14,000
19,000
10,000
30,000
43,000
200,000
220,000
87,000
91,000
170,000
363,000
300,000
30,000
200,000
250,000
200,000
200,000
100,000
10,000
240,000
300,000
31,000x2
23,000
23,000
19,000
Type of
Reactor1*
PPC
HC
FB
PPC
PPC
HC
FB
FB
FB
FB
FB
FB
TC
FB
FB
FB
FB
FB
FB
FB
FB
FB
FB
FB
FB
FB
1MB
FB
FB
FB
FB
(Continued)
Start-up
Mar. 1977
Jan. 1978
Oct. 1978
Feb. 1978
July 1978
Feb. 1980
July 1978
Oct. 1978
Nov. 1978
Oct. 1977
Oct. 1978
Oct. 1975
Apr. 1978
Feb. 1976
Sept. 1976
Jan. 1977
June 1977
Oct. 1977
July 1973
May 1974
Jan. 1975
Apr. 1975
Apr. 1975
Mar. 1976
Oct. 1976
Oct. 1977
Dec. 1977
June 1978
July 1978
-------
TABLE I (Continued). SELECTIVE CATALYTIC REDUCTION PLANTS IN JAPAN
ro
Source: Reference 3
a. Fuel Codes
CO Carbon Monoxide
COG Coke Oven Gas
HO Heavy Oil
HO(HS) High Sulfur Heavy Oil
HO(LS) Low Sulfur Heavy Oil
HO(MS) Medium Sulfur Heavy Oil
LNG Liquefied Natural Gas
LPG Liquefied Propane Gas
b. Reactor Codes
CMB Continuous Moving Bed
FB Fixed Bed
HC Honeycomb Catalyst
1MB Intermittent Moving Bed
PPC Parallel Plate Catalyst
PPR Parallel Passage Reactor
TC Tubular Catalyst
-------
TABLE II. WET NOY/SO_, CONTROL PLANTS IN JAPAN
A. A
Capacity Source of
Process Developer (NmJ/hr) Gas Start-up
OXIDATION/ABSORPTION/REDUCTION PROCESSES
Chiyoda 1,000 Oil-fired Boiler Aug. 1973
Ishikawahima H.I. 5,000 Oil-fired Boiler Sep. 1975
Mitsubishi H.I. 2,000 Oil-fired Boiler Dec. 1974
Osaka Soda 60,000 Oil-fired Boiler Mar. 1976
Shirogane 48,000 Oil-fired Boiler Aug. 1974
Sumitomo Metal-Fujikasui 62,000 Oil-fired Boiler Dec. 1973
Sumitomo Metal-Fujikasui 100,000 Heating Furnace Dec. 1974
Sumitomo Metal-Fujikasui 39,000 Oil-fired Boiler Dec. 1974
Sumitomo Metal-Fuj ikasui 25,000 Sintering Machine 1976
ABSORPTION/REDUCTION PROCESSES
Asahi Chemical 600 Oil-fired Boiler Apr. 1974
Chisso Corp. 300 Oil-fired Boiler Apr. 1974
Kureha Chemical 5,000 Oil-fired Boiler Apr. 1975
Mitsui S.B. 150 Oil-fired Boiler Apr. 1974
Source: Reference 3
276
-------
TABLE III. COSTS OF NOV AND NOV/SOV CONTROL SYSTEMS IN JAPAN
A A A
Process Type
Selective Catalytic Reduction (SCR)
Flue Gas Desulfurization (FGD)
SCR + FGD
Simultaneous NO /SO (Dry and Wet)
Pollutant
Removed
NO
X
so2
N0x & S02
NO & S0_
Capital Cost
₯/kW $/kW
2800 15.5
14000 77.8
18500 102.8
20000 111.1
Operating Costs
₯/kWh
0.3
1.2
1.5
1.7
mills/kWh
1.6
6.7
8.3
9.4
ro
Basis for estimate:
Plant Size
Fuel
NO Concentration
SO Concentration
Particulate Concentration
Temperature
NO Removal Efficiency
S02 Removal Efficiency
Depreciation
Interest Per Year
300 MW, new
oil
200 ppm
1500 ppm <,
200 mg/Nm
380°C
80%
90%
7 years
10%
Maintenance
Insurance
Overhead
Catalyst Life
Annual Operation
Ammonia
Power
Steam
Kerosene
Monetary Conversion
Rate
3% of investment cost
2% of investment cost
5% of investment cost
2 years
8000 hours
₯80/kg
₯12/kWh
₯2/kg
₯32/kg
₯180/$
Source: Information from hearings of the Japan Environment Agency as reported in Reference 3.
-------
TABLE IV. PROCESSES RECOMMENDED FOR FURTHER STUDY
UNDER THE EPA/EPRI/TVA ASSESSMENT PROJECT
Process
Type of Process (classification)
UOP-Shell
UOP-Shell
Hitachi Zosen
Kurabo Knorca
JGC Paranox
Moretana Calcium
Ishikawajima H.I. (IHI)
Asahi Chemical
MON Alkali Permanganate
Dry Simultaneous S0»/N0
(Selective catalytic reduction of
NO and sorption of S0» with CuO)
3t £,
Dry NO only
(Selective catalytic reduction:
parallel passage reactor)
Dry NO only
(Selective catalytic reduction:
metallic honeycomb reactor)
Dry NO only
(Selective catalytic reduction:
moving bed reactor)
Dry NO only
(Selective catalytic reduction:
parallel passage reactor)
Wet Simultaneous S02/N0
(Oxidation/absorption/reduction:
chlorine dioxide as oxidant)
Wet Simultaneous S02/N0
(Oxidation/absorption/reduction:
ozone as oxidant)
Wet Simultaneous S02/N0
(Absorption/reduction: ferrous-
EDTA as chelating compound)
Wet NO only
x J
Source: Reference 4
^ot initally recommended; added as a substitute for the wet NO
only process.
Development discontinued by the process developer; hence, the
process was subsequently dropped from the recommended list.
278
-------
TABLE V. COSTS OF NOV AND NOY/SO SYSTEMS IN THE U.S.
X XX
Revenue
Pollutant Capital Cost Requirement
Process Type Removed ($/kW) (mills/kWh)
Selective Catalytic
Reduction (SCR) NO ,Part.
3t
Flue Gas Desulfuri-
zation (FGD) S02
SCR + FGD NOx,S02,Part.
Dry, Simultaneous
(UOP-Shell) NOx,S02,Part.
Wet, Simultaneous NO ,80-, Part.
Moretana (CIO-) X
Asahi (EDTA)
IHI (Ozone)
>60
100
>160
>155
>180
>200
>380
>2.2
4.2
>6.4
>5.7
>16.0
Basis for the Estimate:
Particulate Control System
FGD System
S09 Removal Efficiency
NO Removal Efficiency
Particulate Removal Efficiency
Boiler Size
Fuel
Heating value
Sulfur content
Ash content
Operation
Capital Investment
Annual Revenue Requirement
ESP for dry systems; wet scrubber
for wet simultaneous systems
Limestone
90%
90%
99.5%
500 MW, new
Coal
24.4 MJ/kg (10,500 Btu/lb)
3.5%
16%
7000 hr/yr
mid-1979
mid-1980
Source: Preliminary results from Reference 5.
279
-------
TABLE VI. COMPARISON OF ENERGY REQUIREMENTS FOR
VARIOUS NOV AND NOV/SOV PROCESSES
X XX
Process Type
Pollutant
Removed
of boiler capacity
Selective Catalytic Reduction (SCR)
Flue Gas Desulfurization (FGD)
SCR + FGD
Dry Simultaneous (UOP-Shell)
Wet Simultaneous
Mpretana (C10_)
Asahi (EDTA)
IHI (Ozone)
NO ,Part.
Xso2
NO ,SO-,Part.
X £f
NO ,S07,Part.
Jt £
NO ,S09,Part.
2£ £m
^0.3
3.4
^3.7
'v-ll.S
^18.6
SBasis for the estimate: See Table V
Includes heat credit for byproduct streams
Source: Reference 5.
TABLE VII. STANDARD BOILERS SELECTED FOR EVALUATION
Package,
Package,
Package,
Package,
Boiler Type
Scotch firetube
Scotch firetube
watertube
watertube, underfeed
Fuel
Natural gas
Distillate oil
Residual oil
Stoker coal
Thermal Input
GJ/hr (106 Btu/hr)
15.8
15.8
158.2
31.6
15.0
15.0
150.0
30.0
Field-erected, watertube,
chain grate
Field-erected, watertube,
spreader
Field-erected, watertube
Stoker coal 79.1 75.0
Stoker coal 158.2 150.0
Pulverized coal 210.9 200.0
280
-------
TABLE VIII. CONTROL REQUIRED OF THE 14 LARGEST
POINT SOURCES IN CHICAGO TO COMPLY
WITH A SHORT-TERM NAAQS
Short-Term NAAQS
(yg/m )
1000
750
500
250
Percentage of Plants Requiring
No Control Moderate Control High Control
79
36
14
0
21
64
57
7
0
0
29
93
3Combustion modification 0\/50% NO control)
b ^
Flue gas treatment or combustion modification plus flue gas treatment
(^90% NO control)
JH
Source: Reference 8.
281
-------
CHEMILUMINESCENT MEASUREMENT OF
NITRIC OXIDE IN
COMBUSTION PRODUCTS
By:
Blair A. Folsom
Craig W. Courtney
Energy and Environmental Research Corporation
Santa Ana, California 92705
283
-------
ABSTRACT
The response of commercially-available chemiluminescent NO analyzers to
X
simulated combustion products containing NO was investigated. An accurate flow
metering system was used to combine 0^ CO^ CO, NZ, Ar, HO, CH,, and H. into
simulated combustion products modeling a wide range of fuels and excess air
conditions. The simulated combustion products were doped with known amounts
of NO and then monitored by two commercial chemiluminescent NO analyzers.
A
The results indicate that chemiluminescent analyzers spanned with NO/N.
generally indicate less than the actual amount of NO in sampled combustion
products. The errors range from one to 20 percent depending upon fuel type,
oxidant composition and the sample conditioning system.
284
-------
ACKNOWLEDGEMENT
The work upon which this publication is based was conducted under
Contract No. 68-02^-2631 with the Environmental Protection Agency. The
authors wish to express their appreciation to W. S. Lanier of the Environ-
mental Protection Agency and to M. P. Heap, J. S. Johnsen and J. M. Keene
of Energy and Environmental Research Corporation for their assistance in
various portions of the work.
285
-------
SECTION 1
INTRODUCTION
The measurement of nitric oxide (NO) concentrations in combustion pro-
(1 2 3)
ducts by chemiluminescence has several advantages over alternative methods ' '
and as a result the chemiluminescent NO analyzer (CLA) has become a standard
instrument for most laboratory and field emission tests. The usual procedure
for CLA calibration is to set the instrument zero and span with high purity
nitrogen (N ) and a known concentration of NO in N_ respectively. The instru-
ment is then used to measure NO in combustion products. Ideally the composition
of the combustion products, which varies according to fuel type and excess air
level, will not affect the measurement of NO. However, several investigators
have found that background gas composition can measurably affect the NO concen-
tration indicated by a CLA. ' ' Extrapolating these results to typical com-
bustion product compositions indicates that the common procedure of neglecting
background gas composition variations could introduce errors as large as 28% in
indicated NO concentrations. The objectives of this study were to:
Assess the accuracy of commercial CLAs measuring NO in combustion
products following the procedures recommended by the instrument
manufacturers
Examine methods of correcting CLA indicated NO concentrations
for the effects of background gas composition variations
Examine methods to improve calibration procedures
Two commercially available CLAs were zeroed and spanned with N and NO in
N- respectively and then used to measure known concentrations of NO in a
variety of background gases simulating combustion products from a wide range
of fuel compositions and combustor operating conditions. Binary mixtures were
also tested. The NO concentrations indicated by the CLAs were then compared
with the actual NO concentrations to determine the effects of background gas
composition.
286
-------
SECTION 2
CHEMILUMINESCENT MEASUREMENT OF NO
The chemiluminescent measurement of NO is based upon the following
four reactions:
NO + 03 * N02* + 02 (1)
NO + 03 N02 + 02 (2)
N02* " N02 + hV (3)
N02* + M -N02 + M (4)
Nitric oxide and ozone (0-) react readily to form nitrogen dioxide (N0_) in
either an excited state (N02*) or a ground level state (NO ). The yield of
NO * is about 10% at ambient temperatures and increases with temperature by
(3)
approximately 0.9%/KV . The excited molecules can decay to ground state
giving off light of a characteristic frequency (chemiluminescence) or can
collide with any third body (M) and decay to ground state without chemilumi-
nescence (quenching). The relative importance of the chemiluminescent and
quenching reactions depends upon the temperature and the amounts and types of
molecules available for quenching. If these factors are constant and the
amount of 0. present is large, the intensity of chemiluminescence is directly
proportional to NO concentration.
Figure 1 is a simplified schematic diagram of a NO analyzer based upon
this reaction scheme. A sample gas containing NO is metered into a reaction
chamber at a constant rate by a suitable flow metering system. Ozone pro-
duced from 0- is also metered into the reaction chamber at a constant rate.
The 02 flowrate is selected to produce a reaction chamber 0_ concentration
many times the maximum anticipated NO concentration. Thus the reaction
between NO and 0_ does not measurably deplete the 0. concentration and the
intensity of chemiluminescence is directly proportional to NO concentration in
the sample gas. The chemiluminescent emission is filtered to remove
287
-------
extraneous light and directed upon a photomultiplier tube (PMT). The PMT
output is amplified and suitably scaled to read out directly in concentration
units (PPM). The amplifier zero offset and gain are adjustable to permit
calibration with standard gases. High purity N_ (free from NO) is used to
set the zero offset. This allows the PMT dark current, which varies with PMT
age and temperature, to be nulled out. A known concentration of NO in N^ back-
ground gas is used to adjust the instrument gain or span.
Commercial CLAs are designed and constructed to maintain the variables
affecting instrument calibration very close to constant. Calibration drift on
state-of-the-art instruments is usually less than 3% of span per day. A CLA
calibrated as described above will indicate the correct NO concentration
(within about 1%) of any mixture of NO in N« within instrument range. However
if the sample gas contains species other than NO or N?, these gases could
cause the instrument to read incorrectly.
There are four potential ways in which sample gas composition variations
could alter CLA response to NO; these are:
Depletion of 0_ by chemical reaction
Chemiluminescence of gases other than NO within the optical
filter band width
Sample flowrate variations due to sample gas property changes
Sample gas species with quenching efficiencies different from N
The first two items are not usually important in measurements of combustion
products. However the last two items can cause significant variations in CLA
calibrations.
The CLA measures the number of NO molecules entering the reaction chamber
per unit time. If the sample gas f lowrate is constant, the number of molecules
of NO entering the reaction chamber per unit time is directly proportional to
the NO concentration. The sample gas flowrate in many commercial CLAs is
maintained constant by passing the sample gas through a capillary and regulating
the capillary pressure drop. However, the volumetric flowrate through a
capillary with fixed pressure drop is inversely proportional to the sample gas
dynamic viscosity (p). Thus sample gas compositions with viscosity different
from that of N will result in sample gas flowrates different from N« and
288
-------
therefore the CLA will not respond properly to NO. Sample gases with low
viscosity will result in indicated NO concentrations greater than actual and
vice versa.
The NO * produced by reaction of NO and 0- can either decay to the ground-
state producing chemiluminescence or transfer energy to other molecules (M)
without chemiluminescence (quenching). Some quenching is unavoidable due to
the finite probability of NO-* molecules colliding with other molecules in the
reaction chamber. The rate of the quenching reaction depends upon the amount
and types of molecules in the reaction chamber. Reaction chamber pressure
affects quenching by altering the mean free path and collision frequency. Low
pressure tends to reduce quenching and several commercial CLAs operate with
reaction chamber pressures less than 10 torr. The quenching reaction rate
affects the intensity of chemiluminescence and hence variations in the quench-
ing reaction rate will alter CLA response to NO.
The quenching reaction rate is significantly influenced by the type of
molecules in the reaction chamber. Experimental evidence confirms that the
rate of quenching increases with the number of degrees of freedom of the
molecules present in the reaction chamber and varies substantially for gases
commonly found in combustion products. Mathews et al recently com-
piled quenching data from several investigators.
The variations in chemi luminescent response to NO due to quenching and
viscosity effects are generally of the same order of magnitude. However, the
magnitude and direction of the effects vary for specific gases. For some
gases such as H_ and 0_ the combined effects of quenching and viscosity result
in a greater error than for quenching alone. For other gases such as H?0, CO.
and Ar the error is reduced by inclusion of viscosity effects.
289
-------
SECTION 3
EXPERIMENTAL PROCEDURE AND APPARATUS
Two CLAs were selected for evaluation:
Beckman Model 951
Thermal Electron Corporation (TECO) Model 10A
The Beckman analyzer utilizes an atmospheric reaction chamber. Sample gas and
ozone flows are controlled by regulating the pressure drop across capillaries.
Sample pressure is provided by an internal sample pump.
The TECO analyzer utilizes a subatmospheric pressure reaction chamber
operating at nominally 9 torr. As with the Beckman analyzer, flowrates in the
TECO are controlled by capillaries. A vacuum pump exhausts the reaction
chamber and creates the pressure drop across the capillaries. The use of a
subatmospheric reaction chamber adds considerable complexity to the instrument
but reduces the importance of quenching thus increasing sensitivity to low NO
concentrations.
The manufacturers' instructions for both CLAs include detailed operating
information specifying the proper pressures, flowrates and calibration pro-
cedures. Filtering and drying the sample gas to a dewpoint less than
operating temperature is required to prevent sample capillary blockage.
The operating procedure.used by technicians in laboratory and field tests
varies somewhat with sample train design and test conditions. Most commonly
the instrument is utilized intact, operated according to the manufacturers'
instructions and zeroed and spanned with N9 and NO in N_ respectively. This
£ £
was the procedure utilized in this investigation. However, in tests of sub-
atmospheric combustors, such as flat flames, a low reaction chamber pressure
CLA (such as the TECO) is often used without the sample gas capillary. This
allows the sample gas to pass directly from the probe to the reaction chamber
without water removal. Sample flowrate is controlled by the probe tip which
290
-------
acts as a sonic flow orifice. The results of this investigation cannot be
directly applied to CIAs operated in this manner.
Table 1 shows the gas mixtures tested in this investigation. These
mixtures were prepared by blending high purity gases and mixtures of NO in N~
or Ar with a precision gas flow metering system. Nitric oxide in N« span gases
were prepared similarly so that inaccuracies in mixed gases supplied in
cylinders could not affect the CIAs' relative responses to span gases and
sample gases.
For each experiment in Table 1, the CIAs were adjusted to read correctly
on N_ with no NO and a mixture of NO in N_. The instruments were then used to
measure the concentrations of NO in the mixture compositions tested. The
actual concentrations of NO in span and test gases were identical so that the
ratios of NO measured in the test gases to actual NO concentrations were direct
measurements of background gas composition effects. The CIAs' responses to N«
zero and NO in N~ span gases were checked after each datum point. Concentra-
tions of 0_, CO and CO- in zero, span and test gases were also monitored as
a check on metering system accuracy and to verify no leaks in the sampling
train.
A schematic diagram of the experimental apparatus is shown in Figure 2.
The only materials in contact with gas mixtures containing NO were stainless
steel, glass and teflon. The total inaccuracy in the flowrate of each gas was
less than 0.5%. Water vapor was added by bubbling some of the gases (0- and N^)
through distilled water. Contacting of the NO mixtures with 02 and H20 was
controlled by maintaining the NO mixtures separate from the 0_ and HO. The
two streams were blended Immediately upstream of the instruments in a short
length of teflon tubing. The instruments included Beckman and TECO CIAs and
other instruments to measure 0^, CO, CO. and dewpoint.
All gases were high purity grade (99.97% or better) with the exception of
CO (99.0%). NO was supplied as mixtures in N. and Ar (+ 5% accuracy) to
facilitate metering. All mixed gases were calibrated against a National Bureau
of Standards (NBS) traceable mixture. To confirm the concentrations of NO in
the test gas mixtures, several gas mixtures were analyzed utilizing the
phenolydsulfonic acid (PDSA) method. The maximum deviation between calculated
and PDSA measured concentrations was 3.0%.
291
-------
SECTION 4
RESULTS
EXPERIMENT NO. 1 - BINARY BACKGROUND GASES
Figures 3 and 4 show the results of the binary background gas tests as
the ratios of indicated to actual NO concentrations as functions of the
concentrations of gases in the M/N- mixtures. As the concentration of gas M
approaches zero, the ratio of indicated to actual concentration must approach
1.0 since 100% N_ was the background gas used for spanning the instruments.
The results for all gases except H_ at high concentrations show either
negligible changes or decreases in the ratios of indicated to actual NO
concentrations as the concentrations of the gases in the background mixtures
increase. The results for CO and H9 are similar for both CLAs. This is
(5) L
expected since quenching data indicates that the relative quenching
efficiency of CO, H_ and N_ are nearly the same. Carbon monoxide has
essentially the same viscosity as N_ and thus the viscosities of CO/N
mixtures are the same as N«. This accounts for the negligible effects of CO
concentration in the background gas on the ratios of indicated to actual NO
concentrations, d. has a viscosity much smaller than N? and the viscosities
of mixtures of EL in N« decrease rapidly as the concentrations of H? increase
above 20%. This accounts for the increase in the ratio of indicated to actual
NO concentration at high H_ concentrations.
The results for the other binary background gas mixtures in Figures 3 and
4 can also be explained by examining the combined effects of viscosity related
flowrate variations and quenching efficiency variations. For example, CH, has
a viscosity less than N- and the viscosity related flowrate variations with
CH,/N_ mixtures tend to increase the ratio of indicated to actual NO concen-
trations as the concentration of CH. increases. However, since CH. has more
4 -4
degrees of freedom than N2, its quenching efficiency should be greater tending
to decrease the ratio of indicated to actual NO concentration as the concen-
292
-------
trations as the concentration of CH, increases. However, since CH, has more
degrees of freedom than N_> its quenching efficiency should be greater tending
to decrease the ratio of indicated to actual NO concentration as the concen-
tration of CH, increases. The results in Figure 3 show that the combined
effects of quenching and viscosity variations produce a nearly flat response
with the TECO. With the Beckman, the quenching efficiency effects are much
larger (due to the higher reaction chamber pressure) and over compensate the
viscosity effects to produce a decrease in ratio of indicated to actual NO
concentration as CH. concentration increases.
4
EXPERIMENT NO. 2 - LEAN COMPLETE COMBUSTION PRODUCTS (DRY)
Figure 5 shows the results of tests on the simulated complete dry com-
bustion products from burning C and CH, in air. These background gas mixtures
contained 0_, C09 and N . In all tests the ratios of indicated to actual NO
concentrations were less than 1.0. The TECO generally indicated NO concentra-
tions closer to the actual levels than the Beckman. This is consistent with
the results of experiment No. 1. For both instruments, the simulated com-
bustion products from C exhibit more deviation than those from CH,. This is a
direct result of the higher C0« concentration in the simulated combustion pro-
ducts from C.
These experiments were conducted with two concentrations of NO in the
sample and span gas mixtures. Figures 5A and B show the results for 200 ppm
NO for C and CH, fuels. Figure 5C compares the results for C as the fuel with
100 and 200 ppm NO and shows that the ratio of indicated to actual NO concen-
tration is independent of the NO doping level.
EXPERIMENT NO. 3 - LEAN COMPLETE COMBUSTION PRODUCTS (WET)
The previous experiments were conducted with dry sample gas mixtures
prepared by blending dry compressed gases from cylinders. The dewpoints of
these mixtures were about 205K. In laboratory and field emissions tests the
combustion products generally contain several percent H_0 and the dewpoint
may be as high as 350K. The usual procedure is to draw the sample gas through
a water/ice bath to remove most of the HO and reduce the dewpoint beneath
ambient temperature. This prevents condensation in the sample capillary and
the associated plugging.
293
-------
In this experiment the water vapor was added to the dry simualted com-
bustion products to produce dewpoints in the range of 275 to 294K. Figure 6
shows the results for C burning in air at 100, 250 and 500% theoretical air.
At each stoichiometry increasing the dewpoint decreased the ratio of indicated
to actual NO concentration. The effect was small for the TECO (1.0 to 2.0%)
but substantial for the Beckman (up to 6.0%)
EXPERIMENT NO. 4 - INCOMPLETE LEAN COMBUSTION PRODUCTS (DRY)
The method of simulating incomplete combustion products utilized in these
tests was to allow 20% of the C0? in the simulated complete combustion pro-
ducts to be reduced to CO and 0_. Figure 7 shows the results for simulated
combustion products burning C in air assuming complete and incomplete com-
bustion. The TECO's performance is essentially unaffected while the Beckman's
improves with incomplete combustion. This is a result of the decrease in CO2
concentration.
EXPERIMENT NO. 5 - RICH COMBUSTION PRODUCTS (DRY)
The simulated rich combustion product mixtures contained CO, CO,,, H» and
N_. The results of those tests are shown in Figure 8. As with the lean com-
bustion products, the TECO generally indicated closer to the actual NO concen-
tration than the Beckman. Largest deviations were at near stoichiometric
conditions with C as the fuel. Under very rich conditions (50% theoretical
air) with CH, as the fuel, both CLAs indicated slightly higher than the actual
concentation of NO. At 50% theoretical air the concentration of H? is over
17% and the binary background gas tests demonstrated that high concentrations
of H_ result in indicated NO concentations greater than actual.
It should be noted that the products of combustion from practical corn-
bus tors operated fuel rich will generally be different from the compositions
utilized in these tests. These simulated combustion product compositions were
selected for simplicity and to give an approximate indication of the CLA per-
formance sampling rich products.
EXPERIMENT NO. 6 AND 7 - COMBUSTION PRODUCTS IN Ar ATMOSPHERES (DRY)
In experiments involving combustion of fuels containing bound nitrogen
294
-------
it is difficult to differentiate between NO formed from atmospheric N and NO
formed from nitrogen bound in the fuel. Several investigators have recently
combusted fuels containing bound nitrogen in an oxidant containing only 0 ,
CO 2 and Ar to eliminate NO formation from N . The oxidant mixtures most
commonly used are 21% 02 balance Ar and 21% CL, X% CO balance Ar. Carbon
dioxide is sometimes added to adjust the flame temperature to match the flame
temperature produced with air as the oxidant. The concentration of CO
required is usually 20%.
Experiments 4 and 5 were repeated with 0 /Ar and 0 /CO /Ar as the oxidants
^ +- £
and the results are shown in Figures 9 and 10. The results for lean combustion
products obtained with the TECO are essentially independent of stoichiometry,
oxidant (02/Ar or 02/C02/Ar) and fuel (C or CH^) . However, while similar
results for combustion of fuels in air gave ratios of indicated to actual NO
concentrations of 0.95 to 0.99, these results are much lower (0.82 to 0.87).
The results for lean combustion products obtained with the Beckman are
also lower than comparable results obtained with air as the oxidant but the
differences are smaller than with the TECO. The Beckman results are also
affected by the CO content of the oxidant; increasing CO concentration
decreases the ratio of indicated to actual NO concentrations.
The results of the rich combustion products tests shown in Figure 10 show
the same general trends as the lean tests: the TECO response is nearly con-
stant independent of stoichiometry and oxidant under rich conditions and the
Beckman results are very sensitive to the concentration of CO^ in the oxidant.
EXPERIMENT NO. 8 - OFF DESIGN POINT CLA OPERATION
All of the previous experiments were conducted with the CLAs operated
strictly according to the manufacturers' instructions. Figure 11 shows the
results of operating the CLAs under other conditions. Simulated combustion
products from burning C in air and 0 /Ar oxidants were utilized and the CLAs
were zeroed, spanned and used to measure these sample gases with the operating
parameters set at several conditions. Figure 11 shows the ranges of instrument
response observed as the independent parameters were varied. In all cases,
variations in the operating parameters resulted in only minor changes (+ 3%)
in instrument response.
295
-------
SECTION 5
DISCUSSION
The results of the experiments discussed above show that commercial CLAs
generally Indicate NO concentrations lower than actual when spanned with
mixtures of NO in N_ and used to sample NO in combustion products. The errors
introduced by neglecting sample background gas composition variations cover
the following ranges:
COMBUSTION PRODUCTS ERROR %
FROM FUEL AND; TECO BECKMAN
21% 02/N2 +1 to -5 +2 to -11
21% 02/Ar -11 to -16 -6 to -14
21% 02/21% C02/Ar -12 to -15 -15 to -20
These results apply directly to the two CLAs tested but are probably representa-
tive of the performance of other instruments of similar design as well. Both
instruments utilize capillaries to control sample flowrates. The TECO reaction
chamber operates at approximately 9 torr while the Beckman reaction chamber
operates at atmospheric pressure.
The gases responsible for the majority of the errors are C02, HO, 0« and
Ar. Each of these gases decrease the CLA's sensitivity to NO. Carbon dioxide
is often present at high concentrations in combustion products and accounts for
the majority of the errors at near stoichiometric conditions. Water vapor has
a much higher quenching efficiency than N . At the low H?0 concentrations
typical of combustion products dried in an ice bath, the quenching effects far
outweigh the effects of viscosity. Sample dewpoint variations can change CLA
response to NO by several percent. For 0 , the effects of viscosity an
quenching variations are additive and this combined with the high concentra-
tions of 02 present in lean combustion products results in errors as high as
4%.
296
-------
The effects of viscosity and quenching variations are opposite for Ar. In
the CLAs tested here, the viscosity related variations over compensated the
quenching variations resulting in a decrease in CIA sensitivity to NO with
increasing Ar concentration. In the experiments reported in reference 5, the
sample flowrate was maintained constant and the results did not include the
viscosity related flowrate variations present in these experiments. This
accounts for the reported opposite effect of Ar on CLA response to NO. In com-
bustion experiments utilizing 0-/Ar or 0 /CO /Ar as the oxidant, the concentra-
tion of Ar in combustion products is high and results in CLA errors as high
as 20%.
The variations in CLA response due to background gas composition var-
iations were generally smaller with the TECO than with the Beckman except for
Ar. Based upon these results a low reaction chamber pressure CLA such as the
TECO would be preferred to monitor products of combustion of fuels burned in
air. However, the choice of instruments for experiments involving 0 /Ar or
0_/CO /Ar oxidants is unclear. The following example illustrates these points.
A fuel containing bound nitrogen is burned in three oxidants: air,
O./Ar and 0_/CO_/Ar. It is assumed that 500 ppm of NO are formed from fuel
bound N independent of oxidant composition. It is further assumed that an
additional 500 ppm of NO are formed from N_ when the fuel is burned in air.
Ideally, the experimental results would indicate 1000 ppm in the air experi-
ment and 500 ppm in both Ar experiments. The thermal NO could then be calcu-
lated as the difference in NO concentrations (500 ppm).
Table 2 shows the results which would be obtained in this experiment
utilizing the TECO and Beckman CLAs but without correcting the results for
background gas composition variations. The fuel was assumed to be C burning
at 100% theoretical oxidant and the instrument response was taken from the
results in section 4.
In the air experiments the TECO reads closer to the correct concentration.
However, with the 0_/Ar oxidant both CLAs indicate incorrectly by the same
amount. The thermal NO concentrations calculated by subtracting the NO con-
centrations are in error by equal amounts for both CLAs but the errors are in
opposite directions. With the 02/C02/Ar oxidant the TECO reads closer to the
297
-------
correct concentration. However, the Beckman yields better results for the
thermal NO. This example shows no clear cut advantage for either instrument
design.
The results of this study can be used to correct the results of NO
emission tests for the effects of background gas composition variations.
Accurate corrections can be made if:
The CLA utilized in the emission tests was one of those
tested here
The CLA was operated according to the manufacturer's
recommendations
The combustion product compositions are known
Figure 12 shows the application of the results of this study to an actual
NO emission test conducted at Energy and Environmental Research Corporation.
A No. 6 oil was fired in a bench scale furnace over a range of excess 0? with
air and 09/CO /Ar oxidants. Nitrogen oxides (NO ) were measured with a TECO
£* & &
CLA equipped with a stainless steel converter to reduce N0_ to NO. The indi-
cated NO concentrations were corrected for background gas composition var-
H
iations by applying the results in section 4. The C and CH, data were inter-
polated to account for the oil's H/C ratio and "the results obtained for 21%
0_/21% CO /Ar were used (neglecting the difference between 20 and 21% CO- in
the oxidant). Application of these results to the experimental data slightly
increases the NO concentration for combustion in air and substantially
Ji
increases (60 ppm) the NO concentration for combustion in the oxidant contain-
Ji
ing Ar. The conversion of fuel nitrogen to NO is higher and the thermal NO
Ji ' 2£.
formation is lower than if the effects of background gas composition varia-
tions had been neglected.
298
-------
SECTION 6
RECOMMENDATIONS
In many laboratory and field emissions tests the variability in operating
conditions and other factors introduce errors of the same order of magnitude
or larger than the errors produced by background gas composition variations.
Under these conditions it is recommended that the NO concentrations indicated
by the CLAs be used directly without corrections recognizing that the results
will tend to be low by a few percent.
If the experimental errors in the emissions tests are small compared to
the errors produced by background gas composition variations, the indicated NO
concentrations should be corrected as discussed in the previous section.
However the results of this investigation can only be applied to CLAs designed
and operated similar to those tested. In low pressure flat flame experiments
where the CLA sample gas metering system (capillary) is replaced by a different
type of metering system, the effects of background gas composition variations
on sample flowrate may be significantly different. Under these conditions the
results of this investigation cannot be applied. It is recommended that either
the effects of background composition variations on sample gas flowrate be
evaluated and quenching data such as Reference 5 be applied, or the overall
effects of background composition variations be emperically determined as in
this investigation.
As an alternative to correcting the indicated NO concentrations, the CLA
could be operated so as to minimize the effects of background gas composition
variations. Three methods for accomplishing this are to:
Calibrate the CLA with NO in a background gas similar to the
anticipated combustion products
Dilute the sample with a known amount of N^
Dry the sample to a low dewpoint
299
-------
Calibrating the CLA with NO in a background gas with a composition identical to
the combustion products to be sampled would eliminate these problems. However,
the concentration of NO in the mixture must be accurately known. Some gas
mixture suppliers determine the concentration of NO in their mixtures by using
a CLA. This introduces errors similar to calibrating the laboratory CLA on
NO in N2.
If the sample gas is diluted with a large amount of N_, variations in
sample gas compositions will have less effect on the flowrate into the reaction
chamber and quenching. For this technique to be effective the dilution factor
must be large (such as 10/1) and precisely known independent of sample gas
composition variations. This accurate dilution is difficult to achieve in
practice and has the further disadvantage of reducing the overall sensitivity
of the CLA by the dilution factor.
The sample gas must be dried to a dewpoint less than the CLA operating
temperature to avoid plugging the sample capillary with condensation. However
at 293R dewpoint the sample gas contains 2.3% H_0 vapor and as a result the
CLA will indicate less than the actual NO concentration due to background and
'dilution effects. Lowering the dewpoint to 273K reduces the H_0 vapor concen-
tration to 0.60% and reduces the error in indicated NO concentration.
In experiments involving combustion products of fuels in air, a low
pressure reaction chamber CLA such a.3 the TECO is preferred. However in
experiments involving combustion in O./Ar or O./CO Ar oxidants, high and low
reaction chamber pressure CLAs perform equally well.
Small variations in CLA operating conditions such as adjustment of the
sample gas or ozone flowrates and changes in the reaction chamber pressure
do not have any clear advantages in reducing the effects of background gas
composition variations. However if the reaction chamber pressure is main-
tained low, the sample gas flowrate is maintained constant and the ratio of
0. to sample gas flowrates in high, the effects of background gas composition
variations could be significantly reduced.
300
-------
REFERENCES
1. Clough, P.N., and Thrush, B.A. Mechanism of Chemiluminescent Reaction
Between Nitric Oxide and Ozone. Transactions of the Faraday Society,
Vol. 63, 1967. pp. 915.
2. Fontijn, A., Sabadell, A.J., Ronco, J. Homogeneous Chemiluminescent
Measurement of Nitric Oxide with Ozone. Analytical Chemistry, Vol. 42,
1970. pp. 575.
3. Niki, H., Warnick, A. and Lord, R. An Ozone - NO Chemiluminescence
Method for NO Analysis in Piston and Turbine Engines. In: Proceedings
of Automotive Engineering Congress. Detroit, Michigan, 1971. pp. 246.
4. Maahs, H.G. Interference of Oxygen, Carbon Dioxide and Water Vapor on
the Analysis for Oxides of Nitrogen by Chemiluminescence. NASA Technical
Memorandum NASA TM X-3229. August 1975.
5. Mathews, R.D., Sawyer, R.F. and Schefer, R.W. Interferences in Chemilum-
inescent Measurement of NO and NO- Emissions from Combustion Systems.
Environmental Science and Technology, Vol. 11, 1977. pp. 1092.
301
-------
o
ro
Sample
Reaction Chamber
Optical Filter
P.M. Tube
Light Tight Housing
Vent
Ozone
Generator
Power Supply,
Amplifier
and Readout
Figure 1. Chemiluminescent Measurement of NO.
-------
co
s
Compressed High Accuracy
Gas Metering
Cylinders System
o
V X
O
O
O
O
1 " CH4
»»
1 >
f »
I »
»
>
H2
CO or Np
NO/N2
NO/Ar
°2
co2
Ar or N2
Variable Mixing
Ox1d1zer Stream
t Water Addition
-*&-H
»
°'»
»
t
t
t
.*»
t
Heat
Fuel Stream
j
i
*-^
*
^
j
j
i \
\
. J *
Bypass
Psychometer
Beckman
NOX Analyzer
TECO
NOY Analyzer
A
Paramagnetic
02 Analyzer
Beckman
Infrared
CO Analyzer
Beckman
Infrared
C02 Analyzer
Vent
Figure 2. Experimental Apparatus.
-------
o
tI
o
1.021
0.98
0,94
0.90
0.86
0.82
0.7*
,
L
-
mm
-
I i 1 i Qxjpgtfi '
O"
A o
fi f*
u D
0 0"
n a.
i i i i i
j
i
L i- - L
h °
O
O
1 |
1 C4rbonlQ1oxid !
0 a
o
o
1 I p
co
o
1.0IJ
0.98|
0.94
0.90
0.86
0.82
0, 78
\f% 1 V
'(
L *" i * ^ 213
Carbon Monoxide H
-
~ o Beckman
Q TECO
^ ^
1 i i i 1
D 10 20 30 40 50
J
<
' ' ° 1 JL
_ Methane _
O
-
0 o
-
0
o-
^ ^
1 1 1 1 1
0 10 20 30 40 50
Concentration of Gas 1n Mixture (%) Balance
Figure 3. Experiment No. 1: Binary Background Gases.
-------
liO?
1
0.98
0.94
0.90
0.86
0.82
"S 0.78
ID r 1 ,0
(J IO
r- 3
T3 -t->
£ £ 1.02
0 O
OJ 2T 2? 1
O 1
01 0.98
0.94
0.90
0.86
0.82
0.7S
1 i i . ' I 1
I D S " D ° °
-
1 1 1 1
a r
i i i IP *
^ Hydrogen
M M _ ._, - - - __
~
-
-
1 1 1 1
1 Argjon
0
1
O'
~l
1
Tc i nn
fi) Ivli
O Beeknrftn
a TECO
0
10
20
30 40 50
Concentration of Gas 1n Mixture {%) Balance
Figure 4. Experiment No. 1 (continued): Binary Background Gases.
-------
1.03
1.01
.99
.97
.95
.93
.91
ii i i i i i i
I TECO - 200 ppra
1 D
! o°° a~-
. ° o o
oG
" o CH4 I
I 1 I J J 1 1 1 L^M
wo
500
Theoretical Air (X)
oo
o
V
+»
o
o
t-4
3
1.03
1.01
.99
.97
5 .95
0 .
f .93'
.91
QHi
i i i i i i i i i
-
DO
- Q o o-
o °
o
L n CH^^ I
pi i i ii ii ii [V
f>
oo
1.D3
1.01
.99
.97
.95
.93
.91
.m
| | 1 II Till
"^^ Bcctnui*- ^ ^^ TfCOc
0 AAA - i- i-,' . «Uljk/ * - J'- M
L\nj ppm tuu j>pjn
'*": ! . I B:
~ ft °
o °
o
" ° * fl»l C I
i i i i i i i i _,i_M
Figure 5. Experiment No. 2: Lean Complete Combustion
Products Burning Fuels in Air (dry).
306
-------
ItJW T.A.
.99
i
.97
.95
.93
.9*
i
.89
4P
.99
1
97
o
V
% r- -95
5 ^» 4
TJ 3 .93
c o
2~* °
* & .91
.89
i
.99
.97*
i
.95
.93
.91
.89
.87
D n
"6 °
-
-
,
8
.
- K , . ,- , -
250X T.A.
Mi
^
0 o o -
-
1
_
o
~ o
o
_ «.
p K ,
N
500% T.A.
.
D
0 DO.
-
O
O
o
N ,
Dry\ 272 283 294
Dewpo!nt
Figure 6. Experiment No. 3: Lean Complete Combustion
Products from Burning C in Air (wet).
3Q7
-------
TECO
.95
.93
.91
.89
_ '
*
O Complete
o
1/2!
I I I I I I I I I
tedcmrt
.99
.97
.95
.93
: .
__ O
.8!
1 1 1 1 1 1
1 1
.100
300 500 700
Theoretical A1r (*)
WO
Figure 7. Experiment No. 4: Incomplete Lean Combustion
Products Burning C in Air (dry).
308
-------
co o
s z
1.02
i.or
Inn
UU
.99
.98
.97
.96
to
5.95
0*
.94
o93
.92
.91
.90
.89
1 1 1 1
L
TECO
- D D 1
0 D
0 0 <
O
_ .
_
- O c
D CH4
1 1 1 1
1
'
-
] 1 1 1 1
Beckman
- ' D
- D
O
a
o
" 0 ]
_
o
mm ' MM
1 1 1 1 <
50
60 70 80
Theoretical A1r («)
90 100 50 60 70 80
Theoretical Air (*)
90
100
Figure 8. Experiment No. 5: Rich Combustion Products Burning
Fuels in Air (dry).
-------
1,021i r
0.98
0.94
0.90 -
8 8 a °
J L
0.82 -
0.78
100 200 300 400 500
1
TECO 21* 02/Ar .
TECO 21% 02/21X C02/Ar
o 6 Q g
Beckman tl% 02/Ar _
V UN-
D CH4
1 1 1 1
1 ,
1000 |/«100 200 300 400 500
Theoretical Air (%)
V
1 1 1 1 1
I Beckman 21
-
[088°
r °
r i i i i
i
[% Q2/2l% C02/Ar
B
i A.
1
^
MM
1000
Figure 9. Experiment No. 6: Lean Complete Combustion Products Burning
Fuels in 0,,/A and QjCOjf^ Oxidants (dry).
-------
NOInd1cated
Inn
.00
0.96
<
0.92
13
0.88
1
0.84
0.80
(
fill
Beckman
)
0
O
O
1
OQ
9
50 70 80 90 100
;
1 1 1 1
TECO
O 21% 02/Ar
a 21* o2/2is; co2/Ar
'> 8 . .-
o
-
1 1 1 I
60 70 80 90 100
Inn
UU
0.96
0.92
0.88
0.84
0.80
Theoretical A1r (%}
Theroetlcal Air (%)
Figure 10. Experiment No. 7: Rich Combustion Products Burning C in
02/A and 02/C02/A Oxidants (dry).
-------
TECO
1.0 < Sample Vacuum < 6.0 1o. HG
0.5 <02 Pres. < 6.0 PSIG
8.5 < Chamber Pres. < 14;0 Torr.
, Beckman
2*0 < Sample Pres. < 4.25 PSIG
5.0 < 02 Pres. < 30.0 PSIG
l.UU
.98
0.96
^ 0.94
2 « -
00 £ §
"g t5 0.92
o o
z z 0,90
0.88
0.86
0.84
0.82
0.80
TECO
- 21% 02/N2
Oxidant
_
Beckman
__ ^_ 2151 0 /N Perkman
"" TEO) Oxidant 21% 02/Ar
21% 02/Ar Oxidant
Figure 11. Experiment No. 8: Off Design Point CLA Operation Burning C
at 125% Theoretical Oxidant (dry).
-------
CO
CO
Argon Furnace, Sonlcore Ultrasonic Atomizer, No Swirl A1r,
No, 6 Alaskan/North Slope 011 - 0.51% N - 100% Conversion « 815 ppm (Dry 0% 02)
O Oxldant 21% 02/20% C02/N2 D
Raw Data Corrected Data
Oxldant Dry A1r 21% 02/N2
800
700
-*, '
o
* 1
o
^ 600
X
o
o.
D-
500
400
i i i i
; - s - *
I
-
-
i i i i
2 34 5 6
.
1
i i i i
D ° ° *.
1
o .
o
o
1 1 1 1
23456
% Excess 0,
% Excess 0,
Figure 12. Example: Correcting Data for Background Gas Composition.
-------
TABLE I. TEST MATRIX
EXPERIMENT
NUMBER
1
2
3
4
5
6
7
8
CONDITIONS SIMULATED
Binary Background Gases
Lean Complete Combustion
Product - Dry
Lean Complete Combustion
Product - Wet
Lean Incomplete Combustion
Product - Dry
Rich Combustion Product -
Dry
Lean Complete Combustion
Product - Dry
Rich Combustion Product -
T»
Dry
Lean Complete Combustion
Product - Dry
FUELS
C, CH.
4
C
C
COU
v»n *
A
C, CH,
T
C
C
OXIDANTS
21% 00/N_
2 2
21% 02/N2
21% 0 /N
21% 02/N2
21% 02/A
21% 02/21% C02/A
21% 02/A
21% 02/N2
21% 02/A
STOICHIOMETRY
% T.A.
100 - cio
100,250
500
125 - 300
50 - 100
100 - oo
60 - 100
125
COMMENTS
02,CO,C02,CH4,H2 or A
in N2
Dry to 294 K Dewpoint
20% CO.-»- CO + 1/2 0
2 2
[02] = 0, Relative C
and H Stoichiometries
Equal, No Soot or
H C
x y
[0_] = 0, Relative C
and H Stoichiometries
Equal, No Soot or
H C
x y
Instruments Operated
off Design Points
CO
-------
TABLE II. COMBUSTION EXPERIMENT EXAMPLE
CO
en
OXIDANT
Air
02/A
02/A
02/C02/A
02/C02/A
NO SOURCE
Fuel + Thermal
Fuel
Thermal (By Difference)
Fuel
Thermal (By Difference)
INDICATED NO CONCENTRATIONS (PPM)
BECKMAN
890
430
460
395
495
TECO
970
430
540
430
540
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/7-79-050b
2.
3. RECIPIENT'S ACCESS!Of*NO.
4. T.TLE AND SUBTITLE proceediiigs of the Third Stationary
Source Combustion Symposium; Volume n. Advanced
Processes and Special Topics
5. REPORT DATE
February 1979
B. PERFORMING ORGANIZATION CODE
7 AUTHOR(S) Joshua S. Bowen, Symposium Chairman, and
Robert E. Hall, Symposium Vice-chairman
B. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
See Block 12.
10. PROGRAM ELEMENT NO.
EHE624
11. CONTRACT/GRANT NO.
NA (Inhouse)
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PI
Proceedings; 3/79
PERIOD COVERED
14. SPONSORING AGENCY CODE
EPA/600/13
15. SUPPLEMENTARY NOTES JERL-RTP project officer is Robert E. Hall. MD-65, 919/541-
2477. EPA-600/7-77-073a thru-G*73e and EPA-600/2-76-152a thru -152c are pro-
ceedings of earlier symposiums on the same theme.
16. ABSTRACT Tne proceedings document the approximately 50 presentations made during
the symposium, March 5-8, 1979, in San Francisco. Sponsored by the Combustion
Research Branch of EPA's Industrial Environmental Research Laboratory-RTP,
the symposium dealt with subjects relating both to developing improved combustion
technology for the reduction of air pollutant emissions from stationary sources,
and to improving equipment efficiency. The symposium was in seven parts, and
the proceedings are in five volumes: I. Utility, Industrial, Commercial, and Resi-
dential Systems; n. Advanced Processes and Special Topics; m. Stationary Engine
and Industrial Process Combustion Systems; IV. Fundamental Combustion Research
and Environmental Assessment; and V. Addendum. The symposium provided contra-
ctor, industrial, and government representatives with the latest information on EPA
inhouse and contract combustion research projects relating to pollution control,
with emphasis on reducing NQx while controlling other emissions and improving
efficiency.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Air Pollution
Combustion.
Field Tests
Assessments
Combustion Control
Fossil Fuels
Boilers
Gas Turbines
Nitrogen Oxides
Efficiency
Utilities
Industrial Pro-
cesses
Hydrocarbons
Air Pollution Control
Stationary Sources
Environmental Assess-
ment
Combustion Modification
Trace Species
Fuel Nitrogen
I3B
21B
14B
21D
13A
13CT
07B
13H
07C
18. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
319
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (»-73)
316
------- |