950R80040
U.S. Environmental Electric Power IERL-RTP-1083
Protection Agency Research Institute October 1980
Proceedings of the Joint
Symposium on Stationary
Combustion NOx Control
Volume I
Utility Boiler NOx Control
by Combustion Modification
2 [IDA m i KTDDI
<3rtrn 1, J CKril
«eri\^Seph
-------
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. Socioeconomic 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 MISCELLANEOUS
REPORTS series. This series is reserved for reports whose
content does not fit into one of the other specific series.
Conference proceedings, annual reports, and bibliographies
are examples of miscellaneous reports.
EPA REVIEW NOTICE
This report has been reviewed by the U.S. Environmental
Protection Agency, and approved for publication. Approval
does not signify that the contents necessarily reflect the
views and policy of the Agency, nor does mention of trade
names or commercial products constitute endorsement or
recommendation for use.
-------
IERL-RTP-1083
October 1980
Proceedings of the Joint
Symposium on Stationary
Combustion NOx Control
Volume I
Utility Boiler NOx Control
by Combustion Modification
Symposium Cochairmen
Robert E. Hall, EPA
and
J. Edward Cichanowicz, EPRI
Program Element No. N130
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, D.C. 20460
and
ELECTRIC POWER RESEARCH INSTITUTE
3412 Hillview Avenue
Palo Alto, California 94303
-------
PREFACE
These proceedings document more than 50 presentations given at the
Joint Symposium on Stationary Combustion NOx Control held October 6-9,
1980 at the Stouffer's Denver Inn in Denver, Colorado. The symposium was
sponsored by the Combustion Research Branch of the Environmental
Protection Agency's (EPA) Industrial Environmental Research
Laboratory-Research Triangle Park and the Electric Power Research
Institute (EPR1)• The presentations emphasized recent developments in
N0X control technology. Cochairmen of the symposium were Robert E.
Hall) EPA, and J. Edward Cichanowicz, EPRI. Introductory remarks were
made by Kurt E. Yeager, Director, Coal Combustion Systems Division, EPRI,
and the welcoming address was given by Roger L. Williams, Regional
Administrator, EPA Region VIII* Stephen J. Gage, Assistant Administrator
for Research and Development» EPA, was the keynote speaker. The symposium
had 11 sessions:
I: N0X Emissions Issues
Michael J. Miller, EPRI, Session Chairman
II: Manufacturers Update of Commercially Available Combustion
Technology
Joshua S. Bowen, EPA, Session Chairman
III: N0X Emissions Characterization of Pull Scale Utility
Powerplants
David G. Lachapelle, EPA, Session Chairman
IV: Low N0X Combustion Development
Michael W. McElroy, EPRI, Session Chairman
Va: Postcombustion N0X Control
George P. Green, Public Service Company of Colorado,
Session Chairman
Vbs Fundamental Combustion Research
Tom W. Lester, EPA, Session Chairman
VI: Status of Flue Gas Treatment for Coal-Fired Boilers
Dan V. Giovanni, EPRI, Session Chairman
VII: Small Industrial, Conmercial, and Residential Systems
Robert E. Hall, EPA, Session Chairman
VIII: Large Industrial Boilers
J. David Mobley, EPA, Session Chairman
IX: Environmental Assessment
Robert P. Hangebrauck, EPA, Session Chairman
X: Stationary Engines and Industrial Process Combustion Systems
John H. Wasser, EPA, Session Chairman
XI: Advanced Processes
G* Blair Martin, EPA, Session Chairman
* «
it
-------
VOLUME I
TABLE OF CONTENTS
Page.
Session I: NOx Emissions Issues
"Regulatory Pressures for Increased NOx Controls,"
R. E. Wyzga *
"Development and Revision of Air Quality Standards with
Special Attention to the NO2 Standard Review,"
M. H. Jones *
"Acid Rain Issue," R. A. Luken *
"State of California Perspective on Stationary Source
N0X Controls," R. Tuvell *
Session II: Manufacturers Update of Commercially Available
Combustion Technology
"Fossil Fuel N0X Control Update," J. A. Barsin *
"Current Developments and Field Experience in Low N0X
Firing Systems," D. J. Prey and T. Kawamura *
"Development and Field Operation of the Controlled
Flow/Split-Flaoe Burner," J. Vatsky 1
"An Evaluation of N0X Emissions from Coal-Fired
Steam Generators," R. A. Lisauskas and J. J. Marshall .... *
Session III: N0X Emissions Characterisation of
Full Scale Utility Powerplants
"Fireside Corrosion and N0X Emission Tests on Coal-Fired
Utility Boilers," E. H. Manny and P. S. Na tans on 43
"Arch-Firing as a Low N0X Design Approach,"
T. W. Sonnichsen *
*See Volume V, Addendum.
iii
-------
"Combined-Cycle Powerplant Emissions," P. L Langsjoen,
R. E. Thompson, L. J. Muzio, and M. W. McElroy 93
"Relationship Between N0X and Fine Particle Emissions,"
M. W. McElroy *
Session IV: Low N0X Combustion Development
"Commercial Evaluation of a Low N0X Combustion System
as Applied to Coal-Fired Utility Boilers," S. A. Johnson
and T. M. Sonmer *
"Pilot Scale Evaluation of a Low H0X Tangential Firing
Method," J. T. Kelly, R. A. Brown, J. B. Wightman,
R. L. Pam, and E. K. Chu 131
"The Development of Distributed Mixing Pulverized
Coal Burners," D. P. Rees, J. Lee, A. R. Brienza,
and M. P. Heap *
"The Development of a Low N0X Distributed Mixing Burner
for Pulverized Coal Boilers," B. A. Folsom, L. P. Nelson,
and J. Vatsky 172
"Field Evaluation of Low Emission Coal Burner Technology
on a Utility Boiler," E. J. Campobenedetto 209
"Operating Experience and Field Data of a 700 MW Coal-Fired
Utility Boiler with Retrofit Low N0X Staged Mixing
Burners," K. Leikert and S. Mi che If elder 251
"Japanese Technical Development for Combustion N0X
Control," K. Mouri *
*See Volume V, Addendum.
iv
-------
DEVELOPMENT AND FIELD OPERATION OF THE
CONTROLLED FLOW/SPLIT-FLAME BURNER
By:
Joel Vatsky
Foster Wheeler Energy Corporation
9 Peach Tree Hill Road
Livingston, New Jersey 07039
1
-------
ABSTRACT
An advanced low NOx coal burner has been installed in a 375 MW front-wall
fired steam generator. Unstaged N0X levels below 0.4 lb/million Btu are being
consistantly obtained with burners having a maximum liberation rate of 285
million Btu/hr. Prototype tests of this burner, in a 50 million Btu/hr test
furnace, have resulted in unstaged NOx emissions of 0.25 lb/million Btu; when
staged using overfire air ports emissions were reduced below 0.20 lb/million
Btu.
This high capacity low N0X burner permits new steam generators to be equipped'
with the same number of burners and the same type of burner management system
as were used prior to the advent of emission regulations.
The Foster Wheeler low NOx system is also available for retrofit to older
steam generators.
This availability is timely in that it provides an option for utilities, which
must convert from oil to coal, to use a modern combustion system. This can be
of particular importance to those units which were designed to fire "future
coal", based on the boiler, firing system and performance coal availability of
the 1950's and *60's, but have instead been firing oil. A further advantage
may be provided by the large N0X reductions attainable since these may permit
trade-offs within the EPA's "bubble concept". However, the actual NOx levels
attainable for older units would be site-dependent.
2
-------
INTRODUCTION
The nearly simultaneous imposition of emission controls and fuel restrictions
on the design of power plants has resulted in a significant increase in the
quantity of hardware which must be included between the fuel storage area and
the stack. Coal, which is currently the only fuel permissible for new utility
boilers in the U.S.,now requires more extensive particulate and sulfur
control equipment. Although this type of equipment is supplied by boiler man-
ufacturers or their subsidiaries, the primary area of responsibility for the
steam generator designer is that of N0„ control. Significant N0„ reductions,
X l>
as compared to levels emitted prior to the advent of N0X limits, can be
attained by proper boiler and combustion system design.
A primary consideration in the design of low N0X combustion systems should be
attainment of minimum NO* levels simultaneously with minimum increase in
system complexity. Clearly, as more components are added to the overall
power plant train it is desirable to maximize the availability of each com-
ponent. Systems of reduced complexity result in lower first cost, decreased
maintenance and simpler operation.
Foster Wheeler's method of minimizing boiler and firing system complexity is
based upon the use of large capacity (up to 300 million Btu/hr) low NOx coal
burners which minimizes the number of burners required. The burner design and
its flexibility and controllability permit a simple common windbox to be
retained.
Foster Wheeler's first generation low NOx burner, the Controlled Flow design,
had been retrofitted in 1976 ^ to an older unit. NOx was reduced about 50%
(to 0.42 lb/106 Btu) without staging and to 0.3 lb/106 Btu with staging. More
recently the advanced Controlled Flow/Split-Flame burner haB been achieving un~
3
-------
staged N0X levels between 0.35 and 0.41 lb/10® Btu on a 375 MW steam generator
to which it had been retrofitted.
(2)
If the EPA's research goals can be used as a guide then even more strin-
gent limits may be instituted in the future. These research goals are 200 ppm
by 1980 and 100 ppm by 1985. Also in some states, units sold prior to 1971
are required to reduce NOx emissions to below 0.7 lb/10® Btu.
In 1979, Foster Wheeler's experimental development program achieved NOx levels
below 0.2 lb/10® Btu, with the Controlled Flow/Split-Flame burner, in a 50
million Btu/hr research furnace. This development program and the field oper-
ation of this burner will be presented here.
4
-------
NOx EMISSIONS: EFFECT OF COAL CHARACTERISTICS AND BOILER DESIGN
The N0X emissions from coal fired steam generators consist of three com-
ponents :
- Thermal N0X: formed by fixation of molecular nitrogen contained
in the combustion air. This quantity is exponentially dependent
on flame temperature; below about 2800°F the formation of thermal
NOx is negligible.
- Volatile and Char Fraction Fuel N0V; formed from the atomic
nitrogen which is chemically bound within the structure of the
coals nitrogen is contained in both the volatile and char frac-
tions. N0X formed from the volatile fraction is typically re-
sponsible for 75% of the fuel N0X emission thereby being the
largest component in the total emission.
These sources of N0X can be controlled/ to varying degrees, by proper boiler
and firing system design. The N0X level which is ultimately produced is de-
pendent on numerous variables. The following is a discussion of those
variables which are of primary consideration:
(a) Burner Zone Liberation Rate (Q/BZS); This is defined as the
net heat input to the burner zone divided by the effective
projected surface. This quantity has been discussed in detail
(3 4)
previsouly . Q/BZS affects flame temperature thereby de-
termining the amount of thermal NO generated.
X
The boiler designer is provided with a useful tool for re-
ducing N0x by increasing burner zone surface (lowering Q/BZS).
Reducing Q/BZS from 450,000 to 300,000 Btu/hr-ft2, on units
5
-------
equipped with high turbulent burners, lowers NOx by about
30%.
(b) Firing Geometry: Opposed or Single Wall; Figure 1 is a
summary of the NO^ emissions as a function of Burner Zone
Liberation Rate for both single-wall and opposed-fired units.
Both plots are composits of full load data from units all
utilizing similar high turbulent intervane burners. It can be
seen that, for any particular value of Q/BZS, opposed-firing
always has lower N0X emissions than does single-wall firing.
A third curve represents the N0X variation with Burner Zone
Liberation Rate (using the above definition and derived from
published emission data ^'®'^)for tangential-fired boilers
operated with oyerfire air ports closed. Although tangen-
tial firing has a lower uncontrolled NOx emission, when com-
pared with wall-fired boilers equipped with high-turbulent pre-
NSPS* burners* the sensitivity to heat release rate appears
to be greater. This differential is of no consequence when
wall-fired units are equipped with modern low NOx combustion
systems.
Although this is useful information, the boiler's firing con-
figuration cannot be dependent on NOx requirements. The
choice of boiler configuration is generally governed by the
specified boiler capacity, economic considerations and plant
requirements.
(c) Coal Characteristics: That part of the NO^ emission which is
formed from the fuel-bound nitrogen is primarily dependent
upon the percent fuel nitrogen content, heating value and
volatility. The total nitrogen content is a function of
both the percent nitrogen and heating value and can be ex-
pressed as lb NOx/million Btu. Varying fuel characteristics
will shift the curves of Figure 1 in the vertical direction.
*NSPS: NEW SOURCE PERFORMANCE STANDARD
6
-------
Knowledge of the fuel nitrogen effect is important so that
the designer can accurately determine the effect of various
N0X control measures. Foster Wheeler has developed such a
technique. The uncontrolled N0X level produced with the high
turbulent intervane burner is determined to provide a base-»
line against which the effectiveness of N0X reduction devices
can be compared. Figure 2 illustrates the accuracy of the
technique for four boilers equipped with intervane burners
and staging ports for NO control. It can be seen that the
baseline prediction method is quite accurate.
This prediction technique was modified in 1976 to incorporate
the first generation low NO^ burner as the primary N0X control
method. The effectiveness of the newest low NO burner, the
Controlled Flow/Split-Flame design, on a 375 MW front-wall
fired unit, is shown for comparison.
Coal particle size distribution is also of importance for NO^
control. Good fineness has always been important from a com-
bustion efficiency viewpoint. However, fineness can also have
a significant impact on NO emission in low N0V systems.
x «
(d) Burner Characteristics: Flame conditions have a major impact
on N0X emissions. High turbulent burners which provide rapid
mixing between the fuel and total combustion air and produce
short intense flames will have the highest N0x emissions. The
rapid mixing increases flame temperature while simultaneously
permitting the fuel-bound nitrogen to be liberated from the
coal particles in an excess oxygen environment, thus promoting
a relatively high fraction to be converted to NO.
The most effective means of reducing this conversion rate is to
reduce the availability of oxygen to the fuel-found nitrogen.
The two most practical means of accomplishing this are:
7
-------
- Reduce burner air by use of staged combustion
Controlled mixing of air and fuel at the burner
With a suitably designed burner, staged combustion can be
used to provide additional N0X reduction. However, coal
properties must be carefully evaluated for any adverse
effect on furnace conditions. Operation at burner
stoichiometrics below about 96-100% is, in general, not
recommended.
8
-------
N0X CONTROLS AND THEIR EFFECTIVENESS
Prior to the advent of N0x emission regulations the historical trend in burner
and boiler design had been toward hot, high turbulent systems. Burners were
designed to be as close to the premixed concept as possible in order to max-
imize flame temperature and minimize unburned carbon carryover. Figure 3
shows a typical 1965 vintage high turbulent intervane burner. A key reason
for the excellent flame stability and high carbon burnout of this design is
the annular coal nozzle and its tangential coal inlet.
The inlet/nozzle configuration produces a uniform coal distribution around
the periphery of the nozzle outlet; there is no roping of the coal. The
single register and throat geometry cause the coal/primary air stream to be
rapidly entrained by the secondary air flow thereby maximizing the oxidizing
regions of the flame. A short, hot flame results which has high thermal and
fuel N0X emissions.
The only operating control used with this type of burner is that which
operates the register. In order to minimize the complexity of low N0X
systems, which must have increased numbers of components, it is advantageous
to use the same type of control, scanning and ignition systems as were
utilized on high turbulent burners.
High turbulent burners are amenable to NO^ reduction by controlling both flame
temperature and oxygen availability to the coal.
Flame temperature can be reduced by:
- Increasing cooling surface (reducing Burner Zone Liberation Rate)
- Flue Gas Recirculation to the windbox (ineffective for coal firing)
9
-------
Figure 1 has shown that NOx emissions, from units equipped with high turbulent
intervane burners, can be reduced by about 30% when Burner Zone Liberation
2
Rate is decreased from 450,000 to 300,000 Btu/hr-ft .
The combination of reduced Burner Zone Liberation Rates and staged combustion
was used as the NOx control method for meeting the first EPA limits
(0.7 lb/MMBTu) on units sold by Foster Wheeler until 1976.
Although effective in reducing NC>x emissions, staged combustion has two
primary limitations:
- Tube wastage which occurs with high iron, high sulfur coals can
affect unit life and reliability.
- Increased slagging which can affect unit availability.
These prolems tend to occur when the burner zone is operated substoichio-
metrically so that reducing atmospheres exist along the waterwalls. In order
to avoid this situation we have limited burner stoichiometry to 96% minimum
TM
and incorporated BOUNDARY AIR to provide an oxidizing atmosphere along the
walls in the burner zone. Figure 4 shows the location of the BOUNDARY AIR
ports and slots. The locations were arrived at through the combined use of
cold flow modeling of the lower burner zone and field experience. BOUNDARY
AIR also acts as passive air flow balance technique to minimize slagging
potential during load changes, mill out of service or unequal mill load
operations.
The operational limits placed on staged combustion, so as to control wastage
and slagging, inhibit its usefulness as a NOx control measure. Consequently,
a more flexible primary N0X control technique was needed. This requirement
has resulted in the development of the low N0X controlled flow family of
burners.
TM - A trademark of Foster Wheeler Energy Corporation
10
-------
LOW NOx BURNER DEVELOPMENT
A low N0X burner development program was commenced in 1975. The primary goal
was to develop a burner which would produce N0X reductions greater than those
attainable with overfire air ports. However, since the burner is only part of
the overall system, requirements were also placed on compatability with the
steam generator's design. These requirements can be summarized as follows:
- Burner capacity should be the same as that of the pre-NSPS high
turbulent intervane burners (up to about 300 million Btu/hr).
- Excess air requirements and unburned carbon levels should be
equal, and preferably superior, to those of the intervane
burner.
- The common windbox should be retained to minimize secondary air
system complexity.
- Combustion system controls and flame scanning should require no
additional complexity.
Low NOx Concept
Within a flame there is strong competition for O2 between the carbon and
nitrogen released from the coal. Under turbulent excess air conditions,
substantially all of the carbon is oxidized to C02 whereas }about 30% of the
fuel nitrogen becomes NO, the remainder being emitted as molecular nitrogen
(the conversion varies inversely with nitrogen content).
When the early phase of combustion occurs under reducing conditions, with
sufficient residence time, the formation of NO is significantly reduced. In
particular, when fuel devolatilization takes place in a reducing environment,
11
-------
volatile-fraction fuel N0X will be minimized. If the residence time in the
reducing zone, or low excess air zone, is of sufficient duration, the char-
fraction fuel NO^ will also be decreased. Also if the local flame stoichio-
metry is below about 95% theoretical air, the flame temperature will be
depressed thereby reducing the formation of thermal N0X. The reduction in
flame temperature will also depend on the available cooling surface; a high
Burner Zone Liberation Rate will prevent minimum N0X levels from being
obtained. The sensitivity of fuel N0X to available oxygen level and mixing
rates can be used advantageously to control NOx via burner modifications.
Low N0V Burner Functional Description
The test program has resulted in the successful development and field demon-
stration of two low N0X burners: The Controlled Flow and the Controlled Flow/
Split-Flame designs. As shown by Figure 5, the two burners are similar; the
coal nozzle being the major difference. The burners operate in the following
manner:
- Secondary Air Control
A series register arrangement, common to both designs, allows simple
burner controls to be used. The inner register, which regulates
the degree of swirl around the coal nozzle, is controlled by a
manual drive since continuous adjustment is not required.
The outer register is controlled by a standard electric drive for
operation at "closed", "ignite" and "Operate" positions. The re-
gister arrangement divides the secondary air into two concentric
streams which independently vary swirl. The secondary air flows
axially into the furnace with almost no component directed radially
inward, toward the burner centerline. Two registers permit the
mixing rate between the primary and secondary air streams, and the
rate of entrainment of furnace gases, to be controlled.
Note that both registers are well shaded from direct furnace
radiation so that parts operate cooler and the tendency to Warp or
bind is minimized.
12
-------
- Secondary Air Balancing
In order to maintain low levels of excess air while minimizing CO,
unburned carbon and slag formation it is advantageous to attain a
balanced secondary air flow to all burners. Foster Wheeler
achieves this by measuring the pressure drop across the perforated
plate air hood thereby obtaining an index of the secondary air flow
to each burner. The flows are then balanced by positioning the
axially movable sleeves which optimizes the secondary air distri-
bution, both vertically and horizontally, in the windbox. The air
hood also improves secondary air distribution around the periphery
of the burner to minimize unwanted turbulence.
- Split-Flame Coal Nozzle
Coal is injected into the furnace through an annular nozzle which
has been modified to separate the coal into concentrated streams,
each of which forms an individual flame. Four streams are in-
dicated in Figure 5 but the number is a design variable.
The split-flame nozzle minimizes mixing between the coal and the
primary air. The combination of the concentrated coal streams and
the staged secondary air produces near throat flame stoichio-
metrics in the 60-70% range up to about two throat diameters into
the furnace. At that point, the swirling secondary stream from the
outer annulus,containing the remainder of the combustion air, com-
bines the flames and provides sufficient mixing to ensure adequate
carbon-burnout.
Uniform distribution of coal about the periphery of the annular
passageway is attained, as with pre-NSPS burners, by use of the
tangential coal inlet. There is no undersirable roping of the
coal.
The basic Controlled Flow burner employs a tapered annular nozzle instead of
a split-flame nozzle. The innner sleeve tip is axially movable thereby
providing a means for varying the primary aic velocity while primary air flow
13
-------
is maintained constant. The velocity adjustment is used to optimize the
primary air/secondary air velocity ratio which minimizes shear-induced
turbulence.
The primary air velocity adjustment has also been incorporated into the
Controlled Plow/Split-Flame burner.
The flows from both the Controlled Plov) and the Controlled Flow/Split-Flame
burners are axial and symmetrical about the burner axis. The only flow-in-
duced turbulence is that which is deliberately provided by register swirl
setting. This produces early staging of the flames with a minimum of unin-
tentional mixing between the fuel particles and the combustion air.
Similarity to the Pre-NSPS Burner
After the low N0X burners three manual adjustments are made(inner register
setting, coal nozzle velocity and air hood sleeve) the respective mechanisms
are locked. The burners then operate in precisely the fashion as would a
pre-NSPS intervane burner. No additional controls are required, beyond the
outer register drive. There are no feedback systems interlocked with the
pulverizers and no continuous air flow controllers. Mills are taken in and
out of service in exacly the same manner as on pre-NSPS boilers.
N0X Emission Test Data
Development testing of our low N0X combustion systems has been performed on
a four-burner, 125,000 lb/hr steam generator that has been previously des-
cribed in detail 4^. More recently, additional testing has been performed
/o g\
on a 50 million Btu/hr single burner test facility
To summarize here:
The original Controlled Flow burner design achieved 35-40% NOx reductions on
the four-burner unit. The design was then modified, scaled-up and tested on a
utility steam generator.
This first field installation was a retrofit of a 265 MW opposed-fired unit
in Japan. The work was performed by Foster Wheeler's Japanese licensee,
14
-------
Ishikawajima-Harima Heavey Industries Co. (IHI), in 1976. Figure 6 summa-
rizes the NO results obtained: 45-50% NO reduction with owerfire air ports
X X
closedr 65-70% with overfire air ports open (at a burner stoichiometry of 96%).
Similar results were obtained in the sister unit when it was retrofitted in
1977.
It should be noted that, prior to installation of the Controlled Flow burner,
these units had slagging problems which were made worse by staging for NOx
control. When the low NOx burners were installed, Boundary Air was simul-
taneously added. The combination of the cooler flames produced by the
Controlled Flow burner and the oxidizing atmosphere maintained along the lower
furnace walls by Boundary Air has significantly alleviated the slagging
problem. The units can now operate continuously with overfire air ports open.
It is interesting to note that overfire air is still effective in combination
with the low NOx burner. This is however, as expected, since the two stage
effect, with the burners operated at approximately the stoichiometric ratio,
permits most of the char to burn out in a low excess air environment, con-
i
sequently, as nitrogen is released from the char it has a greater tendency to
form ^2 than NO. The greater the separation between the overfire air ports
and the burner zone, the greater the residence time at low 0^ and the lower
the NO level will be.
All low NOx burners developed by Foster Wheeler, including the most advanced
designs, can still be used in conjunction with overfire air ports. NOx can be
reduced an additional 25-30% when burner stoichiometry, at 20% excess air, is
reduced from 120% theoretical air to about 100%. For this reason overfire air
poets may still be considered on a case-by-case basis, as a supplementary N0X
control measure as required by coal characteristics and emission limits. How-
ever, it must be reiterated that the degree of staging is generally limited be-
cause of its potentially adverse effect on slagging and tube wastage*
The Controlled Flow burner produces a flame shape which is similar to, although
slightly longer than, that of the high turbulent intervene burner. The
similarity of flame shapes and simplicity of design permits the Controlled Flow
burner to be retrofitted to many pre-:NSPS steam generators on a plug-in basis.
-------
This is confirmed by the success of the retrofit installations of this burner
in Japan; the units there being 10 years old at the time of the retrofit.
Functionally the Controlled Flow burner reduces NO^ by minimizing turbulence
and mixing between the primary air/coal stream and the secondary air. Al-
though only sufficient turbulence is generated to maintain flame stability,
the initial mixing between primary air and coal and the inner secondary air
flow results in near-throat flame stoichiometry of about 85-90%. This is
sufficient to yield 40-50% NO^ reductions and is equivalent to operating a
high turbulent burner, at the same stoichiometry, in conjunction with over-
fire air ports.
Figure 7 compares the NC>x reduction attainable from the intervane burner, when
staged, to the Controlled Flow burner unstaged and staged, as a function of
near throat burner stoichiometry. It can be seen that as the Controlled Flow
burners' stoichiometry is reduced by staging (with overfire air ports) from
120% to 96%, the attainable NO^ reduction is increased from about 50% to about
65-70%.
In order to attain greater NO^ reductions, it is therefore desirable to reduce
the near-throat flame stoichiometry. Since the Controlled Flow burner is a
dual register type which two-stages the secondary air, it does not control the
mixing or the distribution of the primary air and coal. If the coal can be
substantially separated from the primary air and concentrated (i.e., fuel-side
staging) then the Controlled Flow burner will be essentially triple-staged.
Near-throat flame stoichiometries of about 60-70% should thus be attainable and
would provide N0X reductions of at least 65%.
The split-flame coal nozzle is one method of achieving this goal and has been
successfully developed by Foster Wheeler. Development of the initial design
concept ^ on the 125,000 lb/hr industrial steam generator and initial field
(8 9)
operation ' have been discussed in detail previously. This design is
effective, functionally, in reducing N0X 55-60% as shown by Figure 8 for the
development tests and Figure 9 for the early field tests.
16
-------
Figures 10 and 11 are photographs the flames developed on the test boiler and
the utility boiler respectively.
The original split-flame design did not incorporate a mechanism for primary
air velocity control. The new design, which is now in use commercially and
whose operation was described above, now contains this control.
The new burner, which is best described as the Controlled Flow/Variable
Velocity Split-flame (CF/SV) design, has been prototype tested on a 50
million Btu/hr test facility and has been successfully scaled-up to full size
utility scale. In order to evaluate the performance of the CF/SV design on the
test furnace the two earlier burner designs, the Controlled Flow and original
Controlled Flow/Spiit-Flame design, were also tested. These burners provided
a basis against which the CF/SV design was compared.
The results of these tests are summarized in Figure 12. It can be seen that
the original split-flame design reduces NO^ 35% below the Controlled Flow
burner with annular nozzle; the new nozzle design incorporating the variable
velocity feature reduces NOx 50% below the Controlled Flow burner.
For comparison, three full load data points from the 125,000 lb/hr four-
burner boiler are also included in Figure 12. They represent the NOx levels
attained by the intervane burner, the Controlled Flow Burner and the Con-
trolled Flow/Split-Flame (original design) burner. Agreement between the
single burner test facility and the four-burner industrial boiler is excellent.
Figure 13 is a photograph of the flames produced by the Split-Flame nozzle and
Figure 14 is a photo of the flames from the Variable Velocity Split-Flame
nozzle. The new design produces more distinct flames with a greater included
angle. Furnace observations indicate the flame's length is somewhat shorter
than with the original Split-Flame design. This represents an improvement on
an already good situation since there was no flame impingement on the rear wall
of the actual utility boiler tested.
17
-------
FIELD DEMONSTRATION OF THE SPLIT-GLAME CONCEPT
The utility boiler referred to above is the San Juan #1 unit of Public Service
Co. of New Mexico. The original split-flame design was retrofitted to this
unit in November 1978; the new Controlled Flow/Split-Flame (CF/SV) burner was
retrofitted in November 1979 and has been performing satisfactorily, and
meeting all design and performance goals, since then.
The boiler is a front-wall fired unit with 16 burners, each of which has a
maximum capacity of 285 million Btu/hr, arranged in a 4 x 4 matrix. Four
Foster Wheeler medium speed MB pulverizers are used. The boiler can achieve
full load with any three pulverizers in service. The fuel used is a high has
New Mexico sub-bituminous coal, with heating value typically in the 9,000-
10,000 Btu/lb range? although coal with heating value as low as 8*000 Btu/lb
has been fired successfully.
Figure 15 compares the emissions attained on San Juan #1 with the results
achieved on the test furnace. A reduction over 65% is being achieved in the
field, which is quite similar to the reduction achieved on the prototype.
This indicates that the scale-up parameters deyeloped (scale-up ratio is 6:1)
are accurate.
The Controlled Flow/Split-Flame concept produces significantly greater NO^
reductions than does a dual register type of burner when there is no staging of
the throat. However, when the burner is staged,NOx can be reduced an addi-
tional 25% to 301.
On the test furnace, N0X is reduced to 0.185 lb/10^ Btu when overfire air
ports are open with the Controlled Flow burner and the new Variable Velocity
Split-Flame nozzle; with CO increasing to about 125 ppm. This is a 65% reduc-
tion from the level measured with the Controlled Flow burner, and 77% from the
18
-------
pre-NSPS intervane burner level. But, as shown on Figure 15, the asymptotic
decrease in the attainable NOx levels may imply that a practical floor is be-
ing approached.
Figure 16 summarizes and compares the results of the low NO^ burner conversions
on the 265 MW opposed-fired unit and the 375 MW front-wall fired San Juan unit.
Although there is a differential of about 0,25 lb/106 Btu between the uncon-
trolled NO levels, the low NO burners reduce the differential to less than
6 X
0.1 lb/10 Btu. Note also that the primary reason that the San Juan emission
is higher than the boiler in Japan is that the former is single-wall fired and
the latter is opposed fired. Other variables,such as Burner Zone Liberation
Rate and coal characteristics, are important but secondary in influence.
Figure 16 also contains the test results from the 50 million Btu/hr test
furance from which the burner in use at San Juan #1 was scaled. This data
comparison clearly shows the advantages of the Controlled Flow/Split-Flame
concept.
The effectiveness of the air hood and sleeve is demonstrated in Figure 17
which shows the necessity for horizontal adjustment of the air distribution in
the windbox. With all burners adjusted equally (all air hoods and swirl vanes
at the same settings)/ CO stratified along the sidewalls while 0 and NO
b A
peaked in the center of the furnace where CO was zero. After balancing the
secondary air, by adjusting the air hoods' sleeves to obtain equal pressure
drops across all air hoods, a nearly uniform gas distribution was obtained
across the furnace. This permitted the unit to operate at lower exess air with
reductions in both NO and CO.
x
19
-------
THE LARGE CAPACITY BURNER'S SIGNIFICANCE
New Equipment Design
The large capacity Controlled Flow/Split-Flame low N0X burner, which emits N0X
levels at least 60% below those of high turbulent burners, increases the flex-
ibility of the steam generator designer. The design constraints and added
costs imposed by a large number of small burners are alleviated.
Figure 18 summarizes the low N0x combustion system for a typical 600 MW coal-
fired steam generator. The following enumerates the advantages provided by the
large capacity Controlled Flow/Split-Flame low NO^ burner design.
1. A simple common windbox can be retained because the secondary air flow
to each burner can be measured and controlled. It is not necessary to
modulate the air hood sleeves when pulverizers are taken in and out of
service or operated unequally loaded or when load is changed.
i
2. Lower furnace geometry can be made less dependent on the requirements
of the burner. The economics of size, arrangement of heat absorbing
area and NOx requirements can be more advantageously balanced.
3. Standard flame sensing, ignition and control equipment can be used:
one electric register drive per burner with no auxiliary controls
required. The combustion control system is the same as that for a
pre-NSPS boiler.
4. Only two sets of movable vanes per burner: 600 MWe boiler would have
24 - 250 MMBtu/hr burners (48 sets of vanes) versus 48 to 60 small
burners (96-120 ets of vanes, assuming only two sets of movable vanes
per burner).
20
-------
5. Burner design is such that no secondary air control vanes receive
direct radiation from the flame: parts operate cooler and are less
likely to overheat and bind.
6. Pulverizer controls and conduit designs are greatly simplified due
to the reduced number of burners. Therefore, there are fewer areas
subject to erosion which results in reduced maintenance requirements
over the unit's life.
7. The BOUNDARY AIR system provides a passive means of maintaining an
oxidizing atmosphere along the sidewalls and in the hopper: during
unit operation it is not necessary to modulate any dampers as load
changes or as mills are taken out of service. This system is now
standard on all Foster Wheeler units, even those where slagging is
not expected to be a problem# since it also reduces CO formation
during low NO^ operation,
8. Although slightly longer than the high turbulent flame produced by
the pre-NSPS intervene burner, the relatively short flame produced by
the high capacity Controlled Flow/Split-Flame burner permits this low
NO design to be amenable to further NO reduction via the use of over-
X *
fire air ports. However, the use of overfire air must be considered on a
case-by-case basis since fuel quality and properties must be carefully
evaluated.
Retrofit Capability to Older Boilers
The Controlled Flow and Controlled Flow/Split-Flame burners can be designed
over the capacity range of pre-NSPS equipment; up to a maximum burner libera-
tion of 300 million Btu/hr. Since these burners are identical, except for the
design of the coal nozzle and their resultant NOx capabilities, and require no
special windbox» structural modifications or additional controls, they can be
installed on a plug-in basis in most units equipped with circular burners.
Units equipped with other types of burners may require pressure part
changes.
21
-------
The retrofit capability of these burners is due both to the flexibility and
controllability designed into them and to their relatively short flame
lengths. The flame envelope of the Controlled Flow design is nearly identical
to that of the pre-NSPS Intervane burner; that of the split-flame design is
slightly longer. Flame envelope and length can be modified by adjusting
register settings and primary air velocity.
Other advantages of this system, which combines a modern low NO^ burner with
BOUNDARY AIR and is essentially the same as that which is now offered for new
equipment, exclusive of low NO^ are:
1. Cooler, less turbulent, flames.
2. Improved balance of secondary flows to yield less variation in
burner-to-burner stoichiometries.
These should permit:
- Lower excess air operation for improved efficiency.
- Decreased slagging for improved availability.
The degree of improvements in operation which could be afforded to older
equipment would obviously depend on the situation of the individual unit
concerned.
22
-------
SUMMARY
A prototype advanced low NO^ burner, the Controlled Flow/Split-Flame design,
has demonstrated N0X reduction of 65% without staging (to a level of 0.25
lb/million Btu) and up to 77% with staging (to a level of 0.185 lb/million
Btu). This burner has been successfully scaled-up and retrofitted to a 375 MW
front-wall fired steam generator where N0x emissions below 0.4 lb/million Btu,
also representing a 65% reduction, have been consistently obtained without the
use of overfire air. This design is now the standard offering for new Foster
Wheeler steam generators.
The controlled Flow/Split-Flame burner, and the earlier Controlled Flow design,
are capable of being retrofitted to older steam generators. Both are in
successful field operation in retrofitted units.
23
-------
REFERENCES
1. Vatsky, J., "Attaining Low NO Emissions by Combining Low Emission
Burners and Off-Stoichiometric Firing", presented at the 70th Annual
Meeting of the American Institute of Chemical Engineers, November, 1977.
2. Mason, H. B. and Waterland, L. R., "Environmental Assessment of Stationary
Source NO Combustion Modification Technologies.: Proceedings of the
Second Stationary Source Combustion Symposium, Vol. 1, EPA-600/7-77-073a,
July, 1977.
3. Sommerlad, R. D.f R. P. Welden, and R. H. Pai, "Nitrogen Oxide Emission,
An Evaluation of Test Data for Design", presented at the 66th Annuall
Meeting of the American Institute of Chemical Engineers, Philadelphia,
Pennsylvania, November, 1973.
4. Vatsky, J., "Experience in Reducing NO Emissions on Operating Steam
Generation:, Presented at the Second nS Technology Seminar, Denver,
Colorado, November 8 and 9, 1978.
5. Selker, A. P., "Program for Reduction of NO from Tangential Coal~Fired
Boilers Phase II" Combustion Engineering Inc., Windsor, Connecticut,
Environmental Protection Technology Series Report EPA-650/2-73-005-1.
6. Bartok, W., Crawford, A. R. Manny, E. H., "Control of Utility Boiler and
Gas Turbine Pollutant Emissions by Combustion Modification - Phase I",
Exxon Research and Engineering Company, Linden, New Jersey, Interagency
Energy - Environment Research and Development Series Report EPA-600/7-78-
036a.
7. Burrington, Rw L., Cavers, J. D., Selker, A. P. "Overfire Air Technology
for Tangentially Fired Utility Boilers Burning Western U.S. Coal" Combus-
tion Engineering Inc., Windsor, Connecticut, Interagency Energy - Environ-
ment Research and Development Series Report EPA-600/7-77-117.
8. Hansen, A.M., H. Ikebe, "Significant Developments to Limit N0X Formation
in Steam Generators", presented at the Joint Power Generation Conference,
October, 1979.
9. Vatsky, J., "Modern Combustion Systems for Coal-Fired Steam - Generators"
Presented at the Pacific Coast Electric Association Conference, San
Francisco, California, March, 1980.
24
-------
1.4
1.3-
1.2
1.1-
1.0
09-
0.8-
0.7•!
0.6-
OlS-
0.4-
0.3-
0.2-
ai
NO.
LB/106 BTU
FIG.t NOx vs Q/BZS
UNCONTROLLED EMISSIONS
SINGLE WALL FIRED UNITS
OPPOSED-FIRED UNITS
6
TANGENTIAL-FIRED UNITS
NOTE; ALL DATA NORMALIZED TO THE SAME FUEL CHARACTERISTICS
BURNER ZONE LIBERATION RATE: 1000 BTU/HR-FT2
SO
I'
100
150
I '
200
250
300
i
350
400
450
-------
FIGURE 2 PREDICTED VS. MEASURED N0X
NOx
lb/10* BTU
1.0 ¦
0.9
0.8-
GUARANTEeJLZ.
LEVEL
0.6
0.5 ¦
0.4
0.3 '
0.2
0.1 ¦
M
MmIaSUrId} ST"5""5 PORTS CL0SED
O MEASURED : STAGING PORTS OPEN
M
M
M
M
LNB*
660 MW
OPPOSED
670 MW
OPPOSED
EASTERN
BITUMINOUS
125 MW
FRONT WALL
WESTERN
BITUMINOUS
650 MW
OPPOSED
375 MW
FRONT WALL
WESTERN
SUB-BITUMINOUS
LNB - LOW NO„ BURNER
-------
FLAME
DETECTOR
Oil
Igniter
Flame \
Detector
REGISTER—
INTERVANEBURNER
FIG. 3
27
-------
FIG. 4 FLAME BASKET OF TYPICAL LARGE BOILER-BURNER
ZONE SURFACE DEFINED BY H. D. AND W.
28
-------
to
vo
PERFORATED Pi ATr All; HOOD
MOVABLE SLUVE
TANGENTIAL
COAL INLET.
\^.i_SPHT HAME COAL NOZZLE
X. //// /VARIABLEVfj
/// ^velocity/
FIGURE 5
CONTROLLED FLOW / SPLIT FLAME
LOW NOv BURNER
AryuLAR COAL NOZZLE
CONTROL FLOW / BURNER
-------
FIG. 6: N0„ EMISSIONS
265 MW. UTILITY BOILER
U
L0- •
NOx
lb/IOfBTU
HIGH TURBULENT BURNER
0.6
0.6
0.4- •
0.3-
0.2-
0.1-
TYPICAL
OPERATING
REGION
LOW N0X BURNER
OPTIMUM OPERATING REGION
+
+
4-
+
+
0 20 40 60 80 100
OVERPIRE AIR PORT DAMPER OPENING { % )
30
-------
FIG. 7 NOx REDUCTION WITH STAGING
%NOx
REDUCTION
INTERVANE BURNER WITH
OVERFIRE AIRPORTS
40-
/
/
/
/
CONTROLLED FLOW
BURNER
/
60 70 00 90 100 110 120
BURNER STOICHIOMETRY (%)
31
-------
FIG. 8 : NOx vs LOAD
NOx
lb/106 BTU 125,000 Ib/hr INDUSTRIAL BOILER
a9-|
INTERVANE
BURNER
CONTROLLED FLOW
BURNER
CONTROLLED FLOW/SPLIT FLAME
(SAN JUAN DESIGN)
65 70 75 80 85 . 90 95 100
LOAD (% MCR)
32
-------
FIG. 9 : CONTROLLED FLOW/SPLIT-FLAME
INTERVANE
BURNER
/
BURNER EFFECTIVENESS
(ORIGINAL DESIGN)
0.6-
0.5-
0.4
03-
55% REDUCTION
INTERVANE
BURNER
A
i k
if
CONTROLLED
FLOW SPUT-FLAME
60% REDUCTION
M
CONTROLED FLOW
SPLIT-FLAME
SAN JUAN ' I
4- BURNER
INDUSTRIAL BOILER
33
-------
FIG.10: SPLIT FLAMES 125.000 Ib/hr TEST BOILER
-------
FIG. 11: SPLIT FLAMES SAN JUAN #1 BURNER
-------
FIG. 12: NOx COMPARISON
SPLIT-FLAME BURNER TIP
NOx „
lb/106 BTU
G'-125,000 Ib/hr INDUSTRIAL BOILER
(4 BURNERS)
ZI}50MMBTU TEST FURNACE
INTERVANE
BURNER
i
o
CONTROLLED FLOW BURNER
CONTROLLED FLOW/SPLIT- FLAME
(ORIGINAL DESIGN)
CONTROLLED FLOW/SPLIT FLAME-
(VARIABLE VELOCITY)
iS 70 75 i5 85 5o is ioo
LOAD (%MCR)
36
-------
FIG.13: TEST FURNACE SPLIT-FLAMES
(ORIGINAL DESIGN!
-------
FIG.14: TEST FURNACE SPLIT-FLAMES
(VARIABLE VELOCITY DESIGNI
-------
FIG. 15: LOW NOx BURNER NOx COMPARISON
KEY
IV «INTERVANE BURNER
CF«CONTROLLED FLOW
CF/S« CONTROLLED FLOW/SPLIT FLAME
CF/SV-CONTROLLED FLOW/SPUT FLAME (VARIABLE VELOCITY)
NOx
Ib/I06BTU
SAN JUAN »I
!.l'
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
IV
50MMBTU TEST FURNACE
IV (EST)
CF/SV
CF
CF/S
CF/SV
OFA; CLOSED OPEN
(NOT YET
TESTED)
CLOSED OPEN
39
-------
FIG. 16: NOx vs Q/BZS
NOx _
LB/106 BTU
1.3
1.2
SINGLE WALL FIRED UNITS
SAN JUAN 1
UNCONTROLLED NO.
1.0
0.9
OPPOSED FIRED UNITS
265 MW UNIT
IN JAPAN
0.6
CONTROLLED FLOW/SPLIT-FLAME
BURNER (OFA CLOSED)
CONTROLLED FLOW BURNER
(OFA CLOSED)
RANGE OF TEST DATA
50 x 10P BTU TEST
FURNACE
OFA CLOSED—
0.3
OFA OPEN
0.2
OFA OPEN
0.1.
BURNER ZONE LIBERATION RATE: 1000 BTU/HR—FT'
0
350
450
100
150
200
250
400
50
300
-------
BOILER SIOCMALL
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
SECONDARY
"AIR DUCT
16*285 MMBTU/HR BURNERS
STRATIFIED OASES
IN FURNACE
NC£0.45
COstOOppn
<^«5%
GO
pm
20C
-1—
100 ( 00
NO*
NO*
Ib/MMBTU
0.45
BALANCED CO
SECONDARY A*
N0»"0.4 200
C0*40ppm
0^*4.0% 100
I
-4——<2?
I *NOk
3/
NOx
Ib/MMBTU
0.45
Of(%)
5
¦4
3
2
I
5
4
3
2
I
FIG. 17 EFFECTIVENESS OF SECONDARY AIR BALANCING SYSTEM
(375 MWe STEAM GENERATOR)
-------
BOUNDARY AIR
SIDE WALL SLOTS
Q
BOUNDARY AIR PORTS
BOUNDARY AIR
HOPPER SLOTS
FIG. 18 :
600 M STEAM GENERATOR
24 Controlled Flow/Split Flame Burners
w
<0®
MBF MILL
SPLIT FLAME BURNERS
-------
FIRESIDE CORROSION AND NOx EMISSION TESTS
ON COAL-FIRED UtILITY BOILERS
By:
E. H. Manny and P. S. Natanson
Exxon Research and Engineering Company
Exxon Engineering Technology Division
Florham Park, New Jersey 07932
43
-------
ABSTRACT
This paper will describe the status of an EPA-sponsored field study
of NO^ emissions from coal-fired utility boilers. Previous reports
discussed the effectiveness of combustion modification techniques to
significantly reduce NO^ emissions. The simultaneous investigation of
side effects (e.g., particulate emissions, boiler slagging, boiler
performance) did not identify any significant problems. However, one
potential side effect — fireside corrosion of the boiler waterwalls —
was only partially studied. Fireside corrosion rates obtained via probes
(short-term exposure) could not be correlated conclusively with actual
furnace tube wastage experience. Therefore, a long-term corrosion test
was undertaken to obtain representative furnace tube corrosion rate data.
Results of this test, conducted on the 500 MW No. 7 pulverized-coal-fired
boiler at the Crist Station of the Gulf Power Company, are presented and
discussed. Details and a progress update are also given for ongoing
corrosion investigations sponsored by EPA on four large coal-fired utility
boilers designed to meet NSPS N0x emission standards. Information is
also included on a field test using additives to suppress slag formation
in a 330 MW pulverized-coal-fired utility boiler.
44
-------
ACKNOWLEDGMENTS
The authors wish to acknowledge the constructive participation
of Mr. R. E. Hall, EPA Project Officer, in planning the field test programs
and providing coordination with boiler operators and manufacturers. The
assistance and cooperation of the General Electric Company personnel in
helping the selection of gas turbines for testing is also gratefully
acknowledged. The helpful cooperation, participation and advice of Babcock
and Wilcox, Combustion Engineering and Foster Wheeler were essential in
selecting representative boilers for field testing and conducting the pro-
gram. The voluntary participation of electric utility boiler operators in
making their boilers available is gratefully acknowledged. These boiler
operators include the East Kentucky Power Cooperative, Inc., Public Service
Electric and Gas Company, Louisville Gas and Electric Company, Houston
Lighting and Power Company, Public Service Company of Colorado and the
Gulf Power Company. The authors also express their appreciation for the
extensive coal analyses services provided by Exxon Research's Coal Analysis
Laboratory at Baytown, Texas and to Messrs. A. A. Ubbens and E. C. Winegartner
for their contributions and advice on coal related matters. The valuable
assistance of Messrs. L. W. Blanken, R. W. Schroeder, W. Petuchovas, and
Mrs. M. V. Thompson in performing these field studies is also acknowledged.
45
-------
SECTION 1
INTRODUCTION
Exxon Research and Engineering Company (ER&E) under contract to EPA has
been conducting field studies since 1970 on combustion modification tech-
niques to control NC^ and other pollutant emissions from utility boilers.
In early studies significant reductions of NOx were achieved in gas and oil-
fired boilers under EPA Contract No. CPA 70-90 (1) without optimizing the
technology. In a follow up investigation, emphasis shifted to the more
difficult task of controlling N0X emissions in pulverized coal-fired boilers
and the assessment of potential side effects. Twelve coal-fired boilers were
tested under EPA Contract No. 68-02-0227 (2) in cooperation with boiler owner-
operators and boiler manufacturers. In this study reductions in N0X emissions
averaging 39 percent (ranging from 12 to 59%) were achieved with no apparent
adverse side-effects. In addition to the optimization of NOx emissions the
study included particulate and unburned combustible measurements, furnace
corrosion rate probing, determination of boiler efficiency and observations
on changes in boiler operability, i.e., slagging, fouling, flame impingement
or 'instability, etc.
In the current program, presently nearing completion, sponsored by EPA
(Contract No. 68-02-1415) and partially by the Electric Power Research
Institute (EPRI Project No. 200), five coal-fired and 2 coal, mixed-fuel
fired boilers were tested in the Phase I program (3) and four coal-fired
boilers, two gas turbines and one oil-fired boiler were tested in the Phase II
program now reaching its conclusion. The scope of the program was broadened
under these contracts to explore the effectiveness of equipment modifications
designed for N0X control, such as boilers constructed with overfire air ports
and use of low NOx emitting improved burner designs. N0X emissions in the
coal-fired boilers tested were reduced by 34% in the Phase I program and by
38% in Phase II. Potential combustion modification adverse side-effects such as,
46
-------
particulate mass and size distribution, boiler performance and operability,
furnace tube corrosion, etc., received increased emphasis and were studied
in more detail than previously. Since combustion modifications for N0X
control potentially may cause increased slagging problems in boilers, as a
part of this program a series of tests were conducted with promising results
using additives to suppress slag formation in a coal-fired boiler.
Furnace tube corrosion, which may be aggravated by low NOx operation,
is a potential major side-effect. Data developed in past programs with
corrosion probes, however, could not be conclusively related to actual furnace
tube corrosion. The importance of this problem dictated a major effort to
specifically address this question. An extensive long-term corrosion study
was undertaken to obtain corrosion rates on actual furnace tubes. This
program encompassed the use of corrosion probes, exposure of pre-measured
furnace tube panels in the furnace and ultrasonic thickness measurement
(mapping) of furnace tubes to determine actual corrosion rates.
Under EPA Contract No. 68-02-2696 a major effort is being expended in
an on-going program to obtain long-term corrosion data on three additional
coal-fired boilers designed to NSPS standards of 0.7 lbs of NOx/10^ Btu. Two
tests are presently in progress. The 3rd is in the active selection stage
and will be combined with Combustion Engineering's test of a new firing con-
cept (rich fireball) for tangentially fired boilers. The scope of this
program has also been expanded to include level 1 testing, continuous
monitoring of pertinent gaseous emissions and extensive pollutant assessment
of solid, liquid and gas streams entering or leaving the boilers.
47
-------
SECTION 2
TEST PROGRAM UPDATE
Details of test program designs, gaseous sampling and analysis, particu-
late, S0„, corrosion rate and boiler performance measurements and calculations
A
have been covered in prior reports (1)(2)(3)(4). This report will update
work performed under EPA Contract No. 68-02-1415 which was partially sponsored
by the Electric Power Research Institute under EPRI Project No. 200. Field
tests conducted under this program were carried out in Phases I and II. Five
pulverized coal-fired boilers and two mixed-fuel fired (coal/oil, coal/gas)
boilers were tested under Phase I, results of which were reported in EPA
Report EPA-600/7-78-036a (NTIS No. PB281078)(3). Phase II, covered by this
report, will update the program covering field tests on four coal-fired and
one oil-fired boiler with special emphasis on the long term corrosion test
conducted on Gulf Power Company's No. 7 boiler at their Crist Station.
NITROGEN OXIDE EMISSIONS
Field tests conducted in the Phase I program are summarized in Table 1
for record and comparison purposes. Included in the table are details
concerning the boiler manufacturer, the type of firing, kind of fuel burned,
numbers of burners, test variables, number of tests run, and emission data
for baseline and optimum low NOx operation on each boiler tested. Referring
to Table 1 it may be seen that uncontrolled (baseline) emissions ranged from
341 to 1383 ppm with only four out of the seven boilers tested meeting the
New Source Performance Standard (NSPS) of 0.7 lbs of N0x/10^ Btu. Three of
these, Barry No. 2, Navajo No. 2 and Comanche No. 1 were equipped with over-
fire air ports while the 4th, Gaston No. 1, had been retrofitted with B&W's
new low N0X burners. Also note that application of combustion modification
techniques successfully reduced emissions below the new NSPS standard of 0.6 lbs
N0x/10^ Btu in all cases but two (Mercer No. 1 and TVA No. 5) and even TVA No. 5
could meet the original (old) standard. N0X reductions ranged from 22 to 45%,
averaging 34 percent" commenserate with reductions achieved in prior programs.
48
-------
Results of field tests conducted during Phase II of the program are
tabulated in Table 2. It may be noted from Table 2 that baseline emissions
in the four coal-fired boilers tested, ranging from 533 to 827 ppm, did not meet
the original NSPS standard of 0.7 lbs of NO /10^ Btu. Under low NO firing
X K
conditions, however, two boilers (Cooper No. 2 and Comanche No. 1) met the
new NSPS standard 0.6 lbs of N0x/10^ Btu. There is little doubt however that
Louisville Gas and Electric Company's Mill Creek, No. 1 boiler could have met
both NSPS requirements, but low NOx firing was not applied to this unit
during the additive tests due to a lack of time. Average N0X reductions were
38% in the coal-fired boilers tested, ranging from 22 to 62 percent. This
is consistent with NO reductions achieved in Phase I and earlier programs.
Emission reductions obtained in boilers representative of the utility
boiler population and on various current design configurations complying with
recent low N0X requirements or guarantees have been discussed and published
elsewhere (1, 2, 3, 4). NO emission reduction and optimisation achieved on
X
the No. 7 horizontally opposed fired Foster Wheeler boiler at Gulf Power
Company's, Crist Station, which was selected for long-term corrosion testing,
will be presented here to illustrate slightly different applications of com-
bustion techniques.
Crist Station Boiler No. 7 is a horizontally opposed fired, dry bottom,
single furnace Foster Wheeler boiler rated at 500 MW capacity. This unit was
selected for testing because it is a large, pulverized coal burning unit of
modern design. It also appeared to have the necessary operating flexibility
and management support so that it was a good candidate for the Phase IV, long-
term corrosion test program. The furnace measures 52 feet 5 inches wide and
40 feet in depth. Six pulverizers feed 24 burners arranged in three rows of
four burners each in the front and rear walls of the furnace.
The operating variables found to have a statistically significant influence
on NOx emission levels were load, excess air level, and firing patterns. Fig-
ure 1, "ppm N0X vs. % Oxygen in Flue Gas," has been constructed to indicate the
most important relationships found in analyzing the test data. The numbers
within the symbols indicate the run number while the symbols indicate the vari-
ous firing patterns tested. The lines drawn on Figure 1 are least squares,
linear regression lines for ppm N0X vs. % oxygen calculated for each firing pattern.
49
-------
The strong influence of excess air level on NC^ emission levels for
all firing patterns is indicated by the steep slopes of the lines drawn on
Figure 1. Very close agreement was found in the calculated regression
coefficients (change in ppm NQjj for a 1% change in oxygen) for the various
firing patterns, i.e., 69, 80, 81, 59, and 76 for firing patterns S-j_, S2, S3,
Sg, and S7, respectively. Since excess air levels could be reduced by at
least 2% to as much as 5% from normal to low excess air operation without
violating the 200 ppm CO maximum emission level guideline or increasing
stack plume opacity, this represents an important operating variable for
NOjr emission control. Thus, N0X emissions were reduced by 16% in changing
from a full load (480-510 MW) baseline operation (4% O2) of 827 ppm to 696
ppm under low excess air (2.1% O2) operation.
Reducing load by 62% from the 480 to 510 MW range to 190 MW under normal
excess air firing operation resulted in lowering N0X emissions by about 37%.
Staged firing generally resulted in reduced loads as well as reduced NOx
emission levels. Separating the effect of staged firing on N0X emission
levels from the load effect indicate the following. Staged firing operation,
S2, top burners fired lean (by reduced coal flow to top row of burners) and
normal excess air (4% O2) resulted in a 12% reduction in NOx emissions (827
ppm to 728 ppm) with about 5% due to load reduction (496 to 451 MW average)
and the remaining 7% due to staged firing. Staged firing operation, S3, (1
top mill on air only) resulted in a 39% reduction in NOx emissions (to 509
ppm from 827 ppm) with about 12% due to load reduction and 27% due to staged
firing. Finally, Sg, staged firing with both top mills on air only produced
a 72% NOx emission reduction with about 32% due to reduced load (230 MW vs.
495 MW). Part of the load reduction experienced during the test period,
however, was due to abnormal operating difficulties such as partial air heater
plugging.
The combined effect of low excess air and staged firing operation re-
sulted in further NOx emission reductions as would be expected. Thus, the
ppm NOx levels (and % NOx reduction from the 827 ppm measured under baseline
operation) were 451 (-31%), 400 (-52%) and 244 (-70%) for S2, S3 and S$ staged
firing patterns, respectively. These results indicate that this boiler has an
excellent NC^ reduction capability through modified combustion operation.
50
-------
PARTICULATE EMISSIONS AND BOILER PERFORMANCE
Low NOx combustion modification techniques, especially staging the
firing pattern in combination with low excess air firing, results in less
intense combustion conditions than conventional firing methods. A tendency
toward increased burnout problems, therefore, may occur which, potentially,
could increase particulate mass loading as a consequence of increased carbon
in the fly ash. In addition, these effects could also result in changes in
particulate particle size distribution. Changes in particulate mass loading
and particle size distribution could adversely affect collection efficiency
in electrostatic precipitators or in other collection devices while an in-
crease in unburned combustibles could have a corresponding adverse effect on
boiler efficiency. A further potential adverse side effect of low N0X opera-
tion could be a change iu fly ash resistivity which might have a similar
adverse affect on precipitator collection efficiency. Measurements of
resistivity, however, were beyond the scope of this program.
Low N0X combustion modification effects on dust loading were investigated
using an EPA Method 5 type sampling train incorporating a Brink cascade
impactor for particle size determination. Measurements of total mass loading
and particle size distribution were made under baseline and optimized low
N0X operating conditions upstream of the electrostatic precipitators. In the
latter phase of the contract dust loading measurements were made with EPA's
SASS train sampling system. Results of the analyses of the latter tests,
however, are not available at this writing.
A summary of particulate emissions and particle size distribution
determination results for boilers tested in the Phase II program are tabu-
lated in Tables 3, 4, and 5, respectively. Comparing particulate mass loading
data in Table 3 for baseline against low NOx operation, it may be seen that
mass emissions under low N0X firing conditions, for the tests in the Phase II
program, are essentially the same as for baseline operation, requiring little
or no change in electrostatic precipitator collection efficiency. Referring
to Tables 4 and 5 it again may be seen that low N0X operation has very little,
if any, effect on particle size distribution. Aside from potential changes
in resistivity, therefore, it may be concluded from these data, as in the
51
-------
Phase I and prior programs, that there are no significant differences in
particulate mass loading or particle size distribution under low N0X com-
bustion conditions.
Increases in percent carbon on particulate are noted for low NOx firing
conditions in Table 3 which do not seem to have a corresponding direct effect
on mass emissions. Furthermore, the expected decrease in boiler efficiency
(Table 6) not only failed to materialize but for the low N0X conditions
efficiency, if anything, is even greater by a small margin leading to the
conclusion that low N0X firing has only insignificant effects on boiler
performance.
52
-------
SECTION 3
ANTI-SLAGGING ADDITIVE TESTS
Low NOx combustion modifications, especially staged firing in combina-
tion with low excess air operation, can result in lower net reducing atmos-
pheres in the bottom of the furnace often accompanied by higher temperatures.
Under reducing atmospheres, coal ash fusion temperatures generally are about
200°F lower than for oxidizing conditions. This fact, coupled with higher
furnace temperatures can affect the character of the slag formations making
them more fluid and sticky with potentially greater slagging difficulties.
Where boilers may be operating near incipient slagging conditions, the
application of N0X reduction techniques could result in increased slagging
problems.
A part of the Phase II program was devoted to investigation of means
to control increased slagging if ever this problem should occur when applying
N0X control modifications. The potential use of additives gave promise of
being the most cost effective solution to control or ameliorate slagging
conditions in coal-fired boilers. Accordingly, arrangements were made with
Basic Chemicals and the Louisville Gas and Electric Company to conduct
Cooperative tests on LG&E's No. 1 boiler at the Mill Creek Station. Rated
output of the No. 1 boiler is 325 MW but LG&E had arbitrarily derated the
unit to 300 MW in order to keep slagging conditions within mangeable bounds.
A series of eight tests were run during June, 1979; four without additive
injection to develop "baseline" operating information and four while injecting
Basic Chemicals UltraMag additive, an ultra fine (<2 microns) dispersion of
MgO in heating oil. Additive was injected at three different rates at each of
the four corners of the furnace at the B and C slag blower elevations immedi-
ately above the top burners. Boiler loads of 325-330 MW were maintained during
the tests, sufficient to promote slagging, and the effectiveness of the
53
-------
additive was judged by the length of time that load could be maintained at
this level before operating parameters became critical, forcing a cut-back
in load.
Results of the anti-slagging additive trials are summarized in Table 7.
Referring to Table 7, it may be seen that tests 200, 201, and 203 (baseline -
no additive) achieved 12 hours operation at full load, rated conditions
(325-330 MW) before superheat and reheat steam temperatures bordered on un-
controllability. Note test 202 (no additive). However, where the boiler
was slagged to the point of being out of control in 4 1/2 hours; a very
short period. The reasons for this drastic performance were not readily
apparent but may possibly be attributed to the fact that furnace clean-up
prior to the test may not have been as effective as before or that a change
to a higher slagging coal may have occurred for that day.
In tests 204 and 205 additive was injected continuously at the rate of
15 GPH. Referring to Table 7 it may be observed that full load capability
of the boiler could be maintained under these conditions for a period of 15
and 17 hours, respectively, or an additional 3 and 5 hours longer than with-
out additive injection. These results testify to the technical feasibility
and effectiveness of the use of additives at low injection rates with
pulverized coal firing. Other potential benefits, which were beyond the
scope of these investigations, may also acrue from additive usage, such as,
easier clean-up of the boiler during nightly reduced load periods. For
example, it may not be necessary to reduce load as much and the clean-up
period possibly may be shortened to achieve the same degree of cleanliness.
Load carrying capabilities, which are of special importance in tight load
demand situations, therefore could be improved.
Tests 206 and 207 were run in an attempt to optimize the additive
injection rate and to test the effectiveness of other injection methods.
Neither results, however, were quite as effective as injecting the additive
continuously at the rate of 15 GPH.
It is concluded from these tests that anti-slagging additives may be
effective in controlling or ameliorating slagging problems in pulverized
coal fired utility boilers especially when "low NOx" combustion modifications
54
-------
may be employed. The degree of slag reduction, the benefits of increased
load carrying capability, the optimum rate and the most effective injection
method, however, need to be defined in more extensive testing to shed more
light on the economics and technical feasibility of additive usage for this
purpose.
55
-------
SECTION 4
CRIST NO. 7
LONG-TERM FURNACE TUBE CORROSION INVESTIGATIONS
Furnace tubes in pulverized coal-fired boilers corrode under oxidizing
atmospheres due to the corrosive effect of iron alkali sulfate attack". Under
reducing conditions, which may occur as a result of combustion modifications
for NOx emission control, furnace tube corrosion may be accelerated, partic-
ularly when high sulfur, high iron content coal is fired, due to intergranular
penetration of the tube surfaces by iron sulfide.
Earlier investigations of this potential side effect employed corrosion
probes under accelerated conditions to develop corrosion rate data. Average
coupon corrosion rates obtained in these programs were approximately 50 mils/
year with considerable scatter between high and low values. Subsequent in-
vestigations conducted under decelerated conditions approximating the actual
tube environment, reduced this average to around 19 mils/year with less
scatter in the range of the data. Corrosion rates obtained via probes,
however, are still an order of magnitude higher than the 1 to 3 mil/year
corrosion normally experienced in boiler furnace tubes.
Since corrosion data developed by probes could not be handily related
to actual tube corrosion experience and this question is of major importance
to the NOx emission control program, a special long-term corrosion field study
was undertaken. These studies were conducted on Gulf Power Company's, Crist
Station, No. 7 boiler with the participation and cooperation of Foster Wheeler
Energy Corporation. The major purpose of this long-term study was to obtain
quantitative measurements of furnace tube fireside corrosion rates under both
baseline and staged combustion operation. Three corrosion measurement tech-
niques were used: (1) corrosion probes, (2) ultrasonic furnace tube wall
thickness measurements and (3) replaceable wall tube test sections.
56
-------
CORROSION PROBE INVESTIGATIONS
Corrosion probes provide a relatively simple, quick and economical means
for determining corrosion rates. Even though corrosion rate data developed
in these and previous programs could not readily be related to actual furnace
tube experience, this type of measurement was continued in the long-term
corrosion investigations with the objective of eventual correlation with rates
developed by ultrasonic measurement of actual furnace tubes and from exposure
of furnace tube test panels.
Corrosion probe testing on the No. 7 boiler at Gulf Power Company's
Crist Station was amplified extensively in order to obtain more data and
information on the effect of corrosion with time. Conditions of exposure
were maintained the same as in prior testing simulating actual furnace tube
environment but exposure, rather than at 300 hours only, was varied to include
30 and up to 1000 hours under both baseline and low NQjj conditions to deter-
mine initial, intermediate, and longer term corrosion effects. In addition,
special ports were installed in the furnace for the installation of the
probes in the most desirable areas. Two of these ports were located in the
middle of the sidewalls within the burner zone (elevation 129.8 ft.) in the
most corrosion prone area and two others were located at the middle of the
sidewalls but in the upper furnace area (elevation 157.8 ft.) outside of the
expected corrosion area, in order to provide "control" data.
A comparison of corrosion rate data developed in this program on Gulf
Power Company's No. 7 boiler at the Crist Station is best illustrated in
Figure 2 showing a plot of corrosion rate vs. exposure time for probes exposed
to both baseline and low N0X firing conditions. Referring to Figure 2 it may
be noted that coupon corrosion rates decrease with exposure time asytntotically
up to 1000-hour exposure. Initial corrosion rates developed at 24 to 30-hours
exposure are high with considerable scatter in the data. At 250 to 300
hours exposure, corrosion rates are much lower and more consistent in range.
Above 450 to 500 hours exposure, corrosion rates level out to an average rate
of 10 to 12 mils/year with very little scatter in the range of the data. These
rates, however, are still much higher than the 1 to 3 mil/year wastage expected
in actual furnace wall tubes.
57
-------
It is concluded from these corrosion probe investigations that:
• There are no major differences in corrosion rates for process
exposed to low NOx vs. baseline firing conditions, especially
for exposure exceeding 450 hours.
• Corrosion rates developed via corrosion probes decrease with
exposure time through 1000 hours approaching an asymtote above
450 hours exposure,
• Corrosion probes exposed for short terms (up to 30 hours) within
the burner areas in the furnace sidewalls experience significantly
greater corrosion rates than probes exposed outside the burner
levels under low NOx firing conditions. A similar trend is
indicated for baseline operating conditions but more data is
needed to reach firm conclusions.
• Probes exposed for periods of 300 to 1000 hours experienced no
significant differences in corrosion rates due to furnace location
(burner vs. nonburner area) or furnace operating mode (baseline
vs. low NOjj firing).
• Effective correlation of actual long-term furnace tube corrosion
rates require corrosion probe exposure of a minimum of 450 hour
exposure.
CORROSION TEST PANELS
In planning the long-term corrosion program discussions with the major
boiler manufacturers indicated that the most definitive assessment of furnace
tube corrosion would be obtained through use of test panels (premeasured and
metallurgically characterized) installed as integral sections of the furnace
water tube walls. Since boiler walls are very large and vulnerable corrosion
areas are difficult to define in advance of exposure, it was decided to con-
centrate the panels used in the test on the Crist No. 7 boiler mostly on one
furnace wall.
58
-------
Figure 3 presents a schematic side elevation of the Crist No. 7 boiler
showing the location of the furnace test panels. Note that panels 1 and 2
were located below the bottom burner centerline (at 93' elevation) while
panels 7 and 8 were installed far above the top burners (elevation 157'-8")
with panel 8 located in the right hand sidewall only. These locations were
felt to be in areas of relatively low corrosion since both temperatures and
depth of slag deposit would be lower than in the burner area where panels
3, 4, 5 and 6 were located. It was also felt that panels 2 and 5 could be
affected by curtain air at the rear of the furnace and might experience less
corrosion than panels 1 and 3 located in areas without curtain air.
The test panels were fabricated five tubes wide and about 5 feet long.
Tubes 1, 3 and 5 were made of SA 210 grade T1 ASTM specification steel while
tubes 2 and 4 were made from SA-213 grade T-2 ASTM specification steel.
Tube wall thickness measurements were made ultrasonically at three inch inter-
vals (hot side) and six inch intervals (cold side) prior to insertion into
Crist No. 7 furnace and after removal from the furnace. Two tubes were
removed for each test panel during the November, 1977 boiler overhaul period
and replaced with premeasured ASTM specification SA 210 grade T1 steel tubes
which is the same material specified for the original wall tubes on Crist
No. 7. Ultrasonic tube wall thickness measurements were also made in the
field during October, 1976 following the baseline operating period. After
removal from the furnace, tubes in the test panels were cleaned, ultrasonically
measured by Foster Wheeler, sectioned for photomicrographic examination and
corrosion determination.
Based on load demand considerations on the boiler, original test plans
provided for a 5-month sustained baseline operating test run starting in
May, 1976 followed by a 6-month "low NOx" operating period with ultrasonic
furnace tube thickness measurements at the beginning and end of the baseline
operation and middle and end of the "low N0X" operating period. The prime
reasons for including a baseline operating period was the desirability to
include "control" measurements and to allow full load operation of the boiler
during the summer peak loads. In the period of October, 1976 through January,
1977, a number of operating problems (including pulverizer overhauls furnace
tube failures, excessive cold weather, etc.) delayed the start of sustained
59
-------
"Low NOjj" operation until February, 1977 . Subsequently, it was mutually
agreed to extend the "low-NOx" operation until the normal spring outage in
1978, since the unit came off line in July 1977 due to a generator problem.
The final low NOjj operating period turned out to comprise a total of 12
months.
Table 8 summarizes the results of the statistical analysis of the
furnace test panel measurements. Columns 1 and 2 present test panel measure-
ment data for exposure during 5 months baseline operation and 12 months
"low NOx" operation. Column 3 presents data for the two replacement tubes
which were exposed during the last 4 months of the "low NOx" operating
period and columns 4 and 5 present the 13 and 17 month mixed operation data.
To facilitate comparisons of different length operating periods, the average
change in furnace tube wall thickness is shown in mils per year as well as
mils.
The results of statistical analysis of Table 8 can be summarized as
follows:
(1) The average loss in furnace tube (hot side) wall thickness during
the 5 month baseline operation was about 2 mils (or 5 mils per year)
except for panel number 5 which experienced about double that loss.
Panels in the non-burner area experienced about the same loss as
panels located In the burner area (except panel number 5).
(2) The average loss in furnace tube (hot side) will thickness was
about 5 mils (or 5 mils per year) during the 12 month "low NO*"
operating period. However, the test panels located in the burner
area experienced a significantly higher loss level (10 mils per
year) than the panels located within the non-burner area (+0.2 mils
per year).
(3) Comparing the corrosion rates (mils per year) for the 5 month
baseline and 12 month "low NOx" operating periods revealed:
60
-------
(a) panels in the non-burner area experienced significantly less
corrosion during "low NOjj" operation than during baseline
operation (+0.2 vs. -5.3 mils per year). No explanation
could be found for this result.
(b) the test panels within the burner elevation experienced
significantly greater corrosion during the "low NOx" operating
period than during the baseline operating period (10.4 vs.
6.2 mils/year) .
(4) Measurements on the replacement tubes, exposed for the last four
months of "low-NOx" operation, showed no metal loss and the small
gain shown must be due to measurement bias or miscalibration.
(5) Columns (4) and (5) of Table 8 containing measurement data
representing both baseline and "low-NOx" operation lead to similar
conclusions as in (2) above for "low-NOx" operation. Table 9,
below, has been constructed to make this comparison more meaningful.
The 17 month "mixed" operation data has been used to estimate
12 month "low-NOx" operation measurements by deducting the 5 month
baseline operation data measured in the field. The estimated
12 month "low-NOx" operation data is very similar to the actual
field measured 12 month "low NOjj" operation measurements and
statistical analysis using the estimated results lead to the same
conclusions as use of the field measurement data.
The data in Table 8 in columns 4 and 5 are also illustrated in Figure 3
in a more readily understood form. Average changes in tube thickness
(hot side) are shown for each of the panels for the 13 month and 17 month
mixed (baseline and "low NQx") operations. As seen earlier, relatively low
corrosion rates were experienced on panels 1, 2, 7 and 8 located at some
distance from the burners. Panel 5 experienced considerably less corrosion
than panels 3, 4 and 6 located within the burner elevation. The existence of
rear curtain air might explain this difference.
61
-------
In summary, the following conclusions can be stated for the analysis of
corrosion data from test panels exposed to furnace conditions as Crist No. 7
Unit.
(1) Corrosion rates of furnace tube panels are generally similar at
different sidewall furnace locations (below, within and above
the burner elevations) when exposed to baseline operation. The
average corrosion rate is about 5 mils per year during the first
5 months of exposure. (95% confidence limits are 3.8 to 6.2 mils/
year.)
(2) Corrosion rates of furnace tube panels, within the burner elevations
are about 10 mils/year while tubes at least 20 feet above or below
the burner show little or no corrosion during "low NOx" operation.
62
-------
ULTRASONIC FURNACE TUBE THICKNESS MEASUREMENTS
Another method to obtain corrosion rate data is to measure the tabes
ultrasonically before and after the desired exposure period, using the
difference in thickness measurements to calculate the corrosion rate. Ultra-
sonic thickness "mapping" of the furnace tubes was employed in the long-term
corrosion investigations on Gulf Power Company's Crist Station, No. 7 boiler.
Extensive time and effort were expended in planning the tube mapping program
to assure reliable measurements. Major considerations involved in determining
how, where and how many measurements should be made included the following:
1. Location of most likely and least likely corrosion areas within
each wall.
2. Precision and accuracy of the ultrasonic measuring instruments.
3. Changes In normal corrosion rates due to necessary tube cleaning
before measurement (possible bias).
4. Additional measurements for quality control purposes.
5. Costs of cleaning tubes, ultrasonic measurement and necessary
supervision.
Operating experience indicated that the middle area of boiler side walls,
within and just above the top row of burners, is most likely to experience
the highest rate of tube wastage. Consequently, four of the six ultrasonic
measurement levels were located at the center of the three burner elevations
and about 8 1/2 feet above the centerline of the top burner level. The "control"
levels were 13 feet below the centerline of the bottom burners and 28 feet
above the top burner level. Measurements were more highly concentrated within
the middle of the side wall burner area than near the furnace corners or below
and above the burner elevation.
Sandblasting or wire brushing necessary to clean the tubes prior to
measurement may remove a protecting coating and result in Increased wastage.
Elaborate precautions were taken to avoid this possible bias through a program
of random cleaning and measurement. Pains were also taken to assure that only
sufficient sandblasting was done to clean the tubes without metal removal. In
addition, special precautions were taken to ensure that measurements were made
63
-------
at precise elevations on the tubes by welding nuts on the web between tubes
in the comers at the given elevations and snapping chalk lines each time
prior to measurement.
Statistical Analysis of Ultrasonic Data
To assess the statistical significance of the average tube wall thick-
ness change, the most appropriate "error term" is calculated from the
differences between level averages within burner walls. Instrument calibra-
tions were not made more frequently than once per level measured. The levels
were far enough apart to expect different corrosion rates and the spots to be
measured were determined at each level independently.
This approach does not bias the analysis if calibration errors are
insignificant. Comparisons between level means provide a proper error time.
The only loss, if calibration variance is negligible, is a slight loss of
sensitivity due to a reduced number of degrees of freedom. However, experi-
ence has shown that calibration variance appears to be far from negligible.
Table 10 contains a summary of the furnace wall tube (hot side) metal
thickness changes (mils/year) for the baseline operation (5 months) and low
NOx operation <12 month duration). The 5 month baseline operation data has
been converted to a mils/year basis for ready comparison with the 12 month
"low-NOx" operation data. The body of the table presents the average thickness
change for each of the six measurement levels of each furnace wall. The
averages have been calculated from all of the paired data (before and after
measurements) available for each level after screening out obvious outlier
data (20 out of 974 differences). The weighted averages (X), number of
differences (n) and pooled standard deviation(s) are shown for half-walls
(burner and nonburner areas), walls and the whole furnace for both the baseline
and low NOx operating periods. Individual wall level averages vary from a high
loss of -8.26 mils (left wall, middle burner level under low NO* operation) to
a gain of 0.81 mils (right wall, level 1 at 27 feet above top burners under
low NOx operation). These data were also analyzed by Shewhart Control Charts
and 95% confidence limits were calculated by classical and successive
difference methods.
64
-------
The furnace average tube wall thickness loss during low NO* operation was
3.A3 mils. However, two of the half-wall averages were outside of the 97.3%
probability control limits; left wall, burner area (-6.24 mils) and right wall
nonburner area (+0.36 mils). In addition, the right side wall, burner area
average of -5.29 mils is outside of the lower limit of -5.10 mils.
Conclusions from these data and the control chart analysis are:
• The variation of measurement level averages within half-walls is
uniform within half-walls during both baseline and low NOx operation
and equals about 1.5 mils.
• With the possible exception of the right sidewall nonburner average,
all of the half-wall averages during the baseline operation are
about equal after allowing for random level to level variation.
• The low NOx operation produced significantly high metal loss within
the left wall burner area and possibly, significantly high metal
loss within the right wall burner area. The right nonburner area
showed significantly less metal loss than the furnace average.
Comparison of Baseline and Low NOx Data
Table 11 presents a comparison of five month baseline and 12 month low
NOx furnace tube wall thickness loss data. Column 1 lists the nonburner
and burner average loss for each furnace wall (mils, mils per year and mils
per \ year) for the 12 month low NOx operating period. For comparison purposes,
columns 2, 3 and 4 list the corresponding loss data for the 5 month baseline
operating period in mils, mils per year and mils per ^year, respectively.
Since the right wall, nonburner area results are so much different from the
rest of the data (see Table 11, the nonburner grand average has been
calculated with and without the right wall results. 95% confidence limits for
the low NO* operation nonburner (less right wall) and burner grand averages
are also shown in Table 11.
The baseline and low NOx loss data for both nonburner and burner furnace
area can be compared on three different bases: mils, mils per year and mils
per year. Other bases could be used out these three seem adequate considering
65
-------
the state of present knowledge of coal-fired, furnace tube corrosion. Since
the comparison using rail loss data is obviously biased in favor of the base-
line operation (because of the difference in exposure times), no importance
can be attached to its statistically significant difference. Omitting the
right wall, nonburner area averages, all 7 of the half wall low NO
averages produced larger metal losses than the corresponding baseline
averages calculated as mils.
However, both the mils per year (equal loss per unit of time) and
mils per "^year (decreasing rate of loss with time) can provide valid compari-
sons. On a mil per year basis, four of the seven half-wall comparisons
produced lower baseline loss than low NOx loss and the overall average
difference is not statistically significant. However, it should be noted that
noted that the baseline operation average of 4.0 mil/year loss for the
overall burner area is outside of the 95% confidence limits for the low
NO burner area (-6.0 to -4.1 mils/year),
x
If it is believed that the corrosion rate decreases with time (most
boiler corrosion coupon data support this concept) and a mils/^year model is
used, then the low NOx operating period corrosion loss is statistically
significantly higher than the baseline operation loss. Six of the 7 half
wall differences produce a higher loss during low NOx operation than the
corresponding loss during baseline operation. However, separating the
nonburner area from the burner area, it is found that the average difference
(-3.05 less -2.12 = -0.93) for the nonburner area is not statistically
significant while for the burner area the average difference (-5.05 less
-2.58 = -2.47) is statistically significant at the 5% probability level.
Thus, it is concluded that the low NOx operation produces significantly
higher corrosion than baseline operation within the burner area.
Figure 4 presents a visual comparison of 5 month baseline and 12 month
low NOx data for average tube wall thickness loss of nonburner area
(omitting right wall) and burner area. The projected baseline data on a
mils per year basis are shown as dashed lines on Figure 4. The 12-month
projected nonburner area value of-3.30 mil loss is very close to the actual
66
-------
-3.05 mil loss experienced during the low NOx operating period. However, the
projected baseline operation, burner area corrosion loss of 4.00 mils is
less than the actual low NC^ operation loss of 5.05 mils. This latter
difference is not statistically significant when allowance is made for the
uncertainty of both of these loss averages. Note, however, that the 12 month
projected baseline operation loss of 2.58 mils for the mil per year model _i£
statistically significantly less than the 5.05 mils per^year loss during the
low NOjj operation. Thus, the conclusion reached depends upon which model is
accepted.
Summary of Corrosion Measurements
Major results developed in the long-term corrosion investigations
conducted on Gulf Power Company's, Crist Station, No. 7 pulverized coal-fired
boiler using corrosion probes, furnace tube panels, and ultrasonic measurement
of both the panels and furnace tubes have been discussed and summarized in the
foregoing sections. Table 12, which is self explanatory, provides an
encapsulated summary of the overall results of the three methods used in these
long-term corrosion tests.
67
-------
SECTION 5
CORROSION TESTING OF UTILITY BOILER COMBUSTION MODIFICATIONS
The influence of NOx combustion modifications on fireside corrosion,
slagging, particulate emissions, boiler performance and other potential side
effects has been assessed in prior EPA sponsored studies by Exxon Research
and others. As indicated above these adverse side effects, with the possible
exception of fireside furnace tube corrosion, have either been proven to be
nonexistent or of negligible magnitude at the levels of N0X control practiced
in these investigations. Corrosion assessment studies, however, for the most
part, utilized corrosion probes which, in general, provided only a qualitative
estimate of furnace tube wastage rates. These studies, normally, were of
relatively short duration under "low NOjj" operating conditions; typically two
weeks.
EPA recognized very quickly the critical importance of this matter to
the N0X program and the necessity to obtain definitive data and a resolution
to this question. As a result, Exxon Research under contract to EPA (3)(4)
was directed to undertake the first long-term corrosion investigation, dis-
cussed above, on the pre-NSPS No. 7 boiler at Gulf Power Company's Crist
Station, in Pensacola, Florida. During the course of this investigation it
became apparent that more long-term corrosion data would be required to address
the same question on boilers of various designs which were being supplied at
that time by the boiler manufacturers to comply with the old N0X standard of
0.7 lbs N0x/106 Btu. Subsequently, a new EPA contract (Contract No. 68-02-2696)
was awarded to Exxon Research to conduct long-term corrosion tests on four
NSPS designed boilers.
The scope of the new contract was described at the "Third Stationary
Source Combustion Symposium" in San Francisco in March 1979. Investigation
of furnace tube corrosion essentially is the same as employed on the Crist,
No. 7 boiler, i.e., a three pronged approach using corrosion probes, furnace
68
-------
tube panels, and ultrasonic thickness measurement of furnace tubes and test
panels to develop corrosion rate information as described earlier. However,
the scope of the new contract is much more extensive including level 1 pollu-
tant assessment tests involving all pertinent solid, liquid and gas streams
entering or leaving the boiler and 30 day continuous monitoring of NO/NO2,
CO and O2 levels, according to EPA guidelines, with frequent reference method
checks,
PRESENT STATUS AND PLANS
Original contractual requirements calling for testing of four boilers,
one each of the major boiler manufacturers designs, has been changed to more
extensive testing of 3 NSPS designed boilers. Two tests have been in progress
for the past year and a half in cooperation with the respective utility and
boiler manufacturer, as indicated below, and test sites for the third candi-
date boiler are actively being screened. The investigative program on the
third boiler will be combined in a cooperative joint venture with Combustion
Engineering in a test of their new "rich fireball" firing concept for the
further reduction of NOx emissions in tangentially fired boilers.
Tests in progress are:
• Columbus and Southern Ohio Electric Company
Conesville, Unit No. 5 (410 MW)
Tangential firing with overfire air (5 pulverizers)
Manufacturer: Combustion Engineering, Inc.
Start of Testing: December 1978
Corrosion Panels Installed: Nov./December 1979
Ultrasonic Thickness Measurements: December 1979
Corrosion Probes: 30, 300, 1000 hours completed July 1980
• Louisville Gas and Electric Company
Mill Creek, No. 3 (450 MW)
Horizontally Opposed Firing-Low NO* Burners (4 pulverizers)
Manufacturer: Babcock & Wilcox Company
69
-------
Start of Testing: March 1979
Corrosion Panels Installed: December 1979
Ultrasonic Thickness Measurements: December 1979
Corrosion Probes
Corrosion probes provide a relatively simple, quick and economical
means for determining corrosion rates. Even though corrosion rate data
developed in these and previous programs could not readily be related to
actual furnace tube experience, this type of measurement is being con-
tinued in the long-term corrosion investigations with the objective of
eventual correlation with rates developed by ultrasonic measurement of
actual furnace tubes and from exposure of furnace tube test panels.
Corrosion probes in the Gulf Power Company, Crist No, 7 tests were
exposed for 30, 300 and up to 1000 hours to show the effect of corrosion
with time. The type of probe used in this and prior tests conducted by
Exxon Research is shown in Figure 5. The orientation of the coupons on
this probe at 90° to the furnace wall came under criticism as being un-
realistic to furnace tube orientation and results, therefore, were viewed
as being unreliable.
To overcome this problem probes designed and built by Combustion
Engineering's corrosion research section are being used in the present
contract. Design of the new probe is shown in the schematic sketch in
Figure 6. Note that the corrosion probe coupons (Figure 6) are installed
in a 4"xl0" inspection door in the plane of the furnace wall parallel to
the furnace tubes, simulating furnace tube exposure. In this position only
half of the coupon is exposed to the furnace atmosphere; the same as furnace
tubes.
Five of the new probes were used on the Conesville, No. 5 boiler to
obtain corrosion data under 30, 300, and 1000 hours exposure. Four of the
probes were located in the burner area in panels 2, 4, 5, and 6 (Figure 7)
with the fifth probe installed in panel 1, outside of the burner area, to
70
-------
obtain "control" data. Similarly, the location of the four probes which
will be used on the Louisville No. 3 boiler is shown in Figure 8. The
large rectangles on Figure 7 and 8 are test panels and the corrosion probes
are represented by the smaller rectangles (in inspection doors).
Furnace Tube Test Panels
The schematic sketches depicted in Figures 7 and 8 show the location
of the furnace tube corrosion test panels and corrosion probes installed in
the Conesville, No. 5 and Louisville, No. 3 boilers, respectively. Because
the area of greatest expected corrosion in the Conesville tangentially fired
No. 5 boiler is in the vicinity of the front and rear furnace walls adjacent
to smallest angle between the flame and the wall (Figure 7), panel locations
in these walls were chosen judiciously to provide the greatest amount of
information. Note that five panels are located in the front wall and three
in the rear wall. Panels 2, 3, 4, 5, 6 and 7 are located in the burner area
(greatest expected corrosion) with panels 1 and 8, above and below the bur-
ners for "control" purposes. Panels 2, 4 and 6 are in the most vulnerable
areas closest to the flame while panels 3 and 6 are somewhat further from the
flame but still in corrosion prone areas.
The sketch in Figure 8 shows locations of the furnace panels in the
horizontally opposed fired Louisville, No. 3 boiler. Panels 4, 5 and 6 in
the left and right walls are in the most vulnerable corrosion areas. Panels 1
and 2 are located in the hopper slopes where corrosion in this type of boiler
has also been experienced but to a lesser extent. Panel 3, in the rear wall
between the burners, covers the possibility of corrosion in the burner walls
and panel 7, far above the burners, will provide "control" information.
As indicated above, installation of the corrosion panels in the Conesville
and Louisville boilers was completed in December 1979. One half of each panel
in both boilers is scheduled for removal during the annual outage in December
1980. Interim ultrasonic tube thickness measurements are also scheduled to
be made on the furnace tubes and the panels at that time.
71
-------
REFERENCES
1. Bartok, W., Crawford, A. R. , and Piegari, G. J., "Systematic Field
Study of N0X Control Methods for Utility Boilers," Esso Research and
Engineering Company Final Report No. GRU.4G.NO.71; (EPA No. APTD
1163, NTIS No. PB 210739), December, 1971.
2. Crawford, A. R., Manny, E. H., and Bartok, W., "Field Testing:
Application of Combustion Modifications to Control NO^ Emissions
from Utility Boilers," EPA-650/2-74-066, NTIS NO. PB-237344, Exxon
Research and Engineering Company Final Report, June 1974.
3. Crawford, A. R., Manny, E. H., and Bartok, W., "Control of Utility
Boiler and Gas Turbine Pollutant Emissions by Combustion Modification -
Phase I," EPA-600/7-78-036a (NTIS No. PB 281078), Exxon Research and
Engineering Company report, EPRI Project No. 200, March 1978.
4. Bowen, J. S., and Hall, R. E., "Proceedings of the Third Stationary
Source Combustion Symposium;" Volume 1. Utility, Industrial, Com-
mercial, and Residential Systems, EPA-600/7-79-050a, February 1979,
page 157.
72
-------
S.-Normal Firing
1(480-510 MW)
900
S^-Top Burners Lean
(430-470 MW)
IB
800
1A
Top Mill on Air
(390-420 MW)
700
7A
S--2 Center Mills
on Air ONly-190 MW
600
Sg-2 Top Mills
on Air Only
(200-260 MW)
500
400
300
200
4
3
5
8
9
2
7
6
Average X Oxygen Measured in Flue Gas
Figure 1. PPM NO* vs. % Oxygen in Flue Gas
(Crist No. 7 Unit)
73
-------
140
Legend
120
Low NO,
Base Data
100
a)
c
o
60
01
O
to
u
o
CJ
40
100
200
300
400
600
0
500
700
800
900
1000
Exposure Time, Hours
Figure 2. Comparison of Corrosion Rates
Gulf Power Company, Crist Station, Boiler No. 7
Pulverized Coal Firing.
-------
LEFT SIDE WALL
(+0.3)
(-0.4)
-0.4
Note: Panel No. 8 was
located in the
right side wall
opposite Panel No. 7
in the left side wall.
160-
150-
Furnace Depth 40 Feet
(Furnace Width 52.4 Feet)
140-
-7.3
-9.5
BURNERS:
130
TOP
1—
LU
UJ
U.
-1.1
-3.2
-35.0
-10.0
-14.1
-13.0
z
o
MIDDLE
120
5
UJ
-J
UJ
15 Feet
¦15 Feet
BOTTOM
110-
Exposures;
13 Months
17 Months
100
+0.5
-2.1
90
Rear /
Curtain
Air
80
Figure 3. Test panel measurements and locations (average change in tube
wall thickness - mils).
75
-------
0
I 1 I
Non-burner Area (Omitting Right Wall)
t
Burner Area
(Baseline Operation
Actual Loss-Mils)
1
1
Baseline Projected
Mils/^/Year :
JNon-burner Area
El Burner Area
( )Non- burner Area (Low Nox)
95% Confidence
Interval
1
C 3Burner Area (Low Nov)
(Low Nox Operation
Actual Loss
1
Mils, Mils/Year, Mi I s^/Year)
4 6 8
EXPOSURE TIME - MONTHS
10
12
Figure k. Comparison of 5 month baseline and 12 month low NO* corrosion data.
-------
•S4
NJ
2-1/2" PIPE EXTENSION
1/4" S. S. GAS
SAMPLING TUBE
THERMOCOUPLE SOCKETS
END CAP
I
CORROSION
COUPONS
3/4" S. S. COOLING AIR SUPPLY TUBE
1-1/4"
¦1-1/4'' ^-6/8^
1/4" PL
L— i".-
U- Face OF FURNACE WALL TUBES
Figure 5. Corrosion probe, detail of corrosion coupon assembly (Inside of furnace),
-------
Cooling
Air
Out
Cooling
Air
Mi
Air
Out
Thermocouples
Air
In
Furnace
Figure 6. Corrosion probes
(Redrawn from C. E. Design)
78
-------
Panels
Probes
6
~
Rear
Front
Front
Right
Rear
Figure 7, Corrosion probe and test panel locations
CSOE, Conesville, No. 5 boiler.
-------
12 3 4 5
o o o o o
o o o o o
o o o o o
o o o o o
Front
ID
Panels
<
D
0
Right
Figure 8. Corrosion probe and test panel locations
L6&E, Mill Creek, No. 3 boiler.
D
5
4
3
2
1
O
o
o
o
o
o
o
3
o
o
o
o
o
O'
o
o
o
o
o
o
o
Rear
-------
TABLE 1
SUMMARY OF COAL AMD MIXED FUEL FIRED BOILERS TESTED DURING PHASE X
No. of NUX Ealabi ona
X HOx(h)
Reduction
22
29
45
27
43
37
37
36
(«) KM - Babcock cod Vilcox^ CE * Combustion Engineering, FV - Foster Wheeler
(b) Hf - rear wall, HO - horlEontally opposed, T - tangential, IV - Crone well
(c) C - coil, C-C - cotl-gu Blued, C-0 - coal-oil nixed
(d) Particulate and corrosion probe teata performed on theee boiler*
(e) Special low M0X esdaaion burners
(f) Overf ire air porta
(g) Wet bottom furnace
(h) X N0k reduction at full or near full load
(i) PPM N0X - 3TL 02f dry basis
Boiler Operator
1. Tennessee Valley
Authority
2. Southern Electric
Generating Coapany
Station and
Boiler Mo.
boiler Type of
Fuel(c) MCR No. of
Test
Test
Baseline
Low N0V
Mfr.(a) Flrlng(b) Burned (MMe) Burners Variables Runs pp«(1b/10^ BTU) ppio(lb/lQ6 BTU)
Widows Creek - 5 B6W
E. C. Gaston - I B6M
3. Alabama Power Company Barry - 2
4. Potomac Electric
Power Company
Morgantovn - 1
5. Salt River Project Nsvajo - 2
6. Public Service Compsny Comanche - I
of Colorado
7. Public Service Electric Mercer - I
and Gas Company
ce
CE
C£
CE
FW
KU
HD(e)
T< f )
T
T(f)
T(f)
FW(g)
CG
CO
C
c
125
270
130
575
800
350
270
16
18
16
40
56
20
24
31(d)
37(d)
38
27
36(d)
30(d)
597 (0.81)
189 (0.53)
341 (0.46)
552 (0.75)
492 (0.67)
417 (0.57)
Ave rage of Coal Fired Boiler*
33(d) 1383 (1.88)
33 656 (0.89)
468 (0.64)
278 (0.38)
189 (0.26)
403 (0.55)
282 (0.38)
261 (0.35)
876 (1.19)
433 (0.59)
-------
TABLE 2
SUHMAKY OF UTILITY UNITS TESTED DURING PHASE U
Boiler Operator
I. East Kentucky Power
Cooperative, Inc.
Station and
Boiler Ho.
Cooper - 2
Boiler Type Fuel
Mfr. (a) FirlngfM Burned
FW
Coal
MCR Ho. of
(«»*) Burners
Te st
220
18
No. of
Te st
N0X Emissions
Low N0X
% N0x(g)
Variables Runs ppm (lb/lO^ BTU) ppm (lb/tO^ BTU) Reduction
101
(h)
557 (0.76)
(h)
433 (0.59)
22
2. Public Service Company
of Colorado
Coaanclke - 2
HO(e) Coal
3. Public Service Electric Sewaren - 5
and Gas Conpany
oo
to
4. Houston Lighting and
Power Company
5. Houston Lightlug and
Power Company
6. Louisville Electric
and Gab Company
7« Gulf Power Company
Wharton - 43
(Gas Turbine)
Wharton - 42
(Gas Turbine)
Hill Creek - 1
Criat - 7
B 6U
GE
GE
fV
HO
(f)
-------
TABLE 3
PAKT1CULATK EMISSION TEST RESULTS
Utility
2 East Kentucky
Power Cooperative, Inc., Cooper
Station, Boiler No. 2
Public Service Electric & Gas
Company
Sewareu Station Boiler Ho. 5
Uulf Power Company, Crist
Station, Boiler No. 7
Emissions
Req. Eff.
To Meet
2
Carbon
Coal
Date
Test
No.
Firing
Condi tion
Load,
MW
mg/m3
GR/SCF
ng/J
Lb/10"
Btu
0.1 Lb/
106 Btu
On
Particulate
Ash
Wt, X
HHV. 1
Cal/g Be
3/9/77
41
Base*
178
1.06
4.65
3280
7.63
98.7
1.48
12.78
11,742
3/11/77
43
Base*
155
.72
3.12
2361
5.49
98.2
.94
12.48
12,217
3/25/77
59
Low N0X*
123
.78
3.41
2520
5.86
98.3
1.81
11.30
12,312
3/28/77
60
Low N0X *
123
.87
3.82
3130
7.28
98.7
1.87
10.47
12,291
9/17/76
4U
Base**
288
.0059
.026
17.2
0.04
-
-
-
-
9/17/76
6C
Low N0X**
280
.0063
.0274
17.2
0.04
-
-
-
-
6/20/78
150
Base*
436
.686
3.00
2301
5.35
98.1
-
_
_
6/21/78
151
Low N0X*
432
.874
3.82
1926
4.48
97.8
2.98
12.45
11,263
6/22/78
152
Base*
417
.864
3.78
2881
6.70
98.5
0.87
16.48
10,782
6/23/78
153
Low N0X*
434
.909
3.97
2468
5.74
98.3
1.71
14.32
11,033
*Pulverize«l coal firing.
**0i1 firing.
EHM:jbg
-------
TABLE 4
PARTICLE SIZE DISTRIBUTION. WT%
EAST KENTUCKY POWER COOPERATIVE
COOPER STATION - BOILER NO. 2
PULVERIZED
COAL FIRING
Baseline Firing
LOW N0X
Firing
Size
Te st No.
Test No.
Test No.
Test No
Range
41
43
Average
59
60
Average
>2.5
98.86
92.68
94.8
95. 65
91. 75
93. 7
2.5
2.04
3. 71
2.9
3. 28
5.34
4. 3
I. 5
0.50
1. 24
0.9
0. 78
1. 32
1. 1
1. 0
0.41
0.95
0. 7
0. 61
1.07
0.8
0.5
0.63
1. 13
0.9
1.04
1.65
1.4
<0.5
0.36
0. 62
0. 5
0. 67
1. 11
0.9
-------
TABLE 5
PARTICLE SIZE DISTRIBUTION, WTZ
GULF POWER COMPANY
CRIST STATION - BOILER NO. 7
PULVERIZED COAL FIRING
Baseline Firing LOW NOx Firing
Size
Test No.
test No.
Test No.
Test No.
Range
150
152
Average
151
153
Average
>2.5
93. 10
94.00
93.6
92.50
89. 80
91.2
2.5
2. 15
3.48
2.8
3. 15
3. 68
3.4
1.5
0.90
0. 72
0.8
1.12
1.94
1.5
1.0
0. 83
0.84
0.7
0.83
1.43
2.3
0.5
1.24
01.20
1.2
1. 16
1.84
1.5
<0.5
1.80
0. 72
1.3
1.24
1.33
01. 3
-------
TABLE 6
SUMMARY OF BOILER PERFORMANCE CALCULATIONS
Coal
Boiler
Firing
Test
Load,
%
N0X Emissions
(3% 02)
>
Ash, %
(Wet
Z Carbon
on
Boiler
No.
Mode
No.
MW
^2-
4.2
PPM
Lb/10b Btu
rig/J
Basis)
Particulate
Efficiency
2
Baseline
41
178
612
0.82
351
12.37
1 .48
89.82
East Kentucky
Power Cooperative
Inc, Cooper
2
2
Baseline
Low N0X
43
59
155
123
5.4
5.0
574
381
0.77
0.51
329
218
12.76
11.30
0.94
1.81
90.12
90.44
Station
2
Low N0X
60
123
6.6
490
0.65
281
10.41
1.87
90.36
Gulf Power
Company,
7
7
Low NOx
Baseline
151
152
432
417
3.1
5.5
508
848
0.68
1.13
291
486
12.45
16.48
2.98
0.87
88.92
88.86
Crist Station
7
Low N0X
153
430
1.9
456
0.61
261
14.32
1.71
89.27
-------
TABLE 7
SUMMARY OF ANTI-SLAGGIUG ADDITIVE TEST RESULTS
Test
No.
Date
200 6/11/79
201 6/13/79
202 6/14/79
203
206
207
6/18/79
204 6/20/79
205 6/21/79
Time
6/25/79
07:27
13:53
16:16
19:27
10:00
11:06
15:36
22:00
10:00
13:18
*14:30
16:15
07:00
08:16
13:45
15:58
19:00
07:10
09:26
13:35
16:00
22:20
08:15
09:38
13:00
15:35
01:15
09:00
21:00
09:30
23:10
Start
End
Start
End
Start
End
Load Cut]
Back.
Start
End
Start
End
Start
End
Start
End
Start
End
Hours (3
Full Load
(325-330 MV7)
12
S
!
12
4.5
12
15
17
12
14
°2.
%
5.0
5.3
4.7
5.2
5.0
4.8
4.8
4.9
4.5
4.6
4.6
5.3
5.3
5.7
N0X PPM Test
(3% 0?) Condition
584
599
512
566
557
448
479
517
447
473
572
514
534
581
Baseline
(No Additive)
Baseline
(No Additive)
Baseline
(No Additive)
Baseline
(No Additive)
Additive
(15 GPH)
Additive
(15 GPH)
Additive
(7.5 GPH)
Additive
(15 GPH
Slugs)
*SH/RH sprays max. @ 14:30.
87
-------
TABt.K 8
TEST PAHEL MEASUREMENTS
(Average Change In Tube Hue Side Wall Thickness)
B««eUre(l) Luv *>x<2> •%ov «0x"(3) "Mixed"14* "Wxed,,(5)
Teat Panel Nuaiber and Location Ope rat Ion-5 Ho. Operatloo-12 Mo. Operation-^ Mo. Operatlon-13 Mo. Operatlon-17 Mo
(5/76 to 10/76) (11/76 to 4/78) (12/77 to 4/78) (5/76 to 11/77) (S/™ 4/7»)
Mils
_lni_
Mlla/Yr.
Ml la
("?
Mils
<">
Ml 1 s
(">
Mlla/Yr.
Mils
(<0
Mlls/Yr.
Non-fturner Area
(20 Teet Below Burners)
).
Left Wall-Front
-1.9
(6)
-4.6
-1.2
(4)
+1.3
(20)
-0.2
(34)
-0.2
-2.5
(51)
-1.8
2.
Left Wall-Rear
-2.4
(8)
-6.0
+0.4
(5)
+1.4
(20)
+0.5
(34)
+0.5
-2.1
(5t)
-1.5
(29 feet above burners)
7.
Left Wall-Center
-2.3
(6)
-5.4
+1.3
o>
+1.1
-0.4
(30)
-0.4
-1.5
(42)
-1.1
8.
Right Wall-CeuCer
-
-
-
+1.9
(20)
+0.3
(34)
+0.3
-0.4
<51)
-0.3
Average
-2.20
(20)
-5.33
-M). 17
(14)
+ 1.42
(78)
+0.05
(132)
+0.0S
-1.62
< 195)
-1.18
Burner Area
(Middle Burner Level)
1.
Left Hall-Front
-1.7 (7)
-4.2
- 8.7
(4)
+1.6
<18)
-14.1
(30)
-12.9
-13.0
(42)
-9.2
4.
Left Wall-Center
-1.5 (7)
-3.5
-12.8
o>
+1.4
(18)
-35.0
(30)
-32.3
-10.0
<40)
-7.1
5.
Left Hall-Rear
-4.4 (5)
-10.6
-
+1.4
<18)
- 1.1
(30)
- 1.0
- 3.2
<42)
-2.3
(Top Burner Level)
6.
Left Wall-Center
-2.6 (8)
-6.4
- 9.7
<4)
+2.4
<18)
- 7.3
(30)
- 6.7
- 9.5
<45)
-6.7
Burner Area Average
-2.55 (27)
-6.2
-10.40
(11)
+1.70
<72)
-14.4
(120)
-13.3
- 8.92
<169)
-6.30
Grand Average
-2.40 (47)
-5.72
- 5.12
(25)
+1.56
<150)
-7.16
(252)
-6.61
- 5.28
(364)
-3.73
(n) Ibifcer ot paired aeasureaents used in calculatiuo iwrAge Cube wall thickness changes.
<1) S Cubes per panel. Initial aeasureaentft In laboratory; final aeaaureaeot in the field.
(2) 3 tubes per panel. Initial uasurewents in tlie field; final aeasureaenta in the laboratory.
(3) 2 replacsaeaC totei per panei. Initial teiaureaeoc la the Held; final mmurement fn the Isbmcory.
<4) 2 tubes per panel. Initial and final Measurements aarir in the laboratory.
(5) 3 tubes per panel. Initial and final rmsasureaents aide in the laboratory.
-------
TABLE 9
COMPARISON OF FIELD AND LABORATORY MEASUREMENTS
Panel No.:
Field Measurements:
5 Month
Baseline
OperatIon
12 Month
"Low NOx"
Operation
Lab Measurements:
17 Month
"Mixed"
Op era t ion
Estimated
12 Month
"Low NOx"
Operation*
Outside Burners
1
2
7
Average
-1.9
-2.4
-2.3
-2.20
+0.4
+0.17
-2.5
-2.1
-1.5
-2.03
-0.6
+0.3
+0.8
+0.17
Inside Burners
3
4
5
6
Average
-1.7
-1.5
(-4.4)
-2.6
-1.93
—8.7
-12.8
-9.7
-10.40
-13.0
-10.0
(- 3.2)
-9.5
-10.83
-11.3
-8.5
-6.9
-8.90
*17 month "mixed" operation minus 5 month baseline operation
12 month "low NOx" operation measurements.
estimated
89
-------
TABLE 10
WALL TUBE HOT SIDE METAL THICKNESS CHANGE RATE-MILS/YEAR
(Paired Data)
LEVEL MEASURED
BASELINE OPERATION - 5 MONTHS
"LOW NOx" OPERATION - 12 MONTHS
RELATIVE TO
FURNACE WALL
FURNACE
FURNACE WALL
FURNACE
BURNERS
LEFT
RIGHT
FRONT
REAR
AVERAGE
LEFT
RIGHT
FRONT
REAR
AVERAGE
NON BURNER AREA
1
-3.62
-12.96
+1.10
-7.49
-3.89
+0.81
-2.92
-1.50
-1.94
2
-3.24
-10.66
-5.74
-0.31
-5.66
-4.45
+0.07
-4.04
-4.10
-2.96
6
-0.30
-5.30
-
-9.05
-5.74
-2.48
+0.04
-
+0.21
-0.51
ic
-3.17
-10.56
-3.24
-3.62
-6.24
-3.94
+0.36
-3.37
1.84
-2.11
n
73
85
22
29
209
102
94
55
89
340
ap
5.57
6.47
4.59
4.86
5.78
4.71
3.40
3.47
5.43
4.42
BURNER AREA
3
+1.75
- 9.41
-4.27
-1.49
-3.24
-4.99
-7.14
-4.31
-3.46
-5.48
4
+0.10
- 2.11
-4.30
-3.74
-2.06
-8.26
-3.49
-2.68
-5.11
-4.86
5
-5.57
-10.18
-7.30
+0.41
-6.74
-7.17
-2.47
-3.51
-4.13
-4.41
X
-1.51
- 6.94
-5.21
-1.73
-3.98
-6.24
-5.29
-3.50
-4.17
-5.05
n
54
56
29
29
168
83
91
48
54
276
sp
4.49
5.59
4.80
4.13
4.88
6.77
4.78
3.71
6.00
5.55
X
-2.46
- 9.12
-4.36
-2.68
-5.23
-4.97
-2.42
-3.431
-2.72
-3.43
n
127
141
51
58
377
185
185
103
143
616
-------
TABLE 11
COMPARISON OF 5-MONTH BASELINE AND 12-MONTH LOW NOx
AVERAGE TUBE WALL THICKNESS LOSS DATA
Area
Low NOjj
12 Months
Mils, Mils/Year,
Mils A^ear
Baseline
5 Months
Mils Mils/Year Mil /^i ear
Nonburner:
Left Wall
Right Wall
Front Wall
Rear Wall
Total
Less Right Wall
-3.94
+0.36
-3.37
-1.84
-2.11
-3.05
-1.32
-4.40
-1.35
-1.51
-2.60
-1.37
-3.17
-10.56
-3.24
-3.62
-6.25
-3.30
-2.04
-6.82
-2.09
-2.34
-4.03
-2.12
Burner:
-4.6 to -1.5 95% simultaneous confidence limits
Left Wall
Right Wall
Front Wall
Rear Wall
Total
-6.24
-5.29
-3.50
-4.17
-5.05
-0.063
-2.89
-2.17
-0.72
-1.66
-1.51
-6.94
-5.21
-1.73
-4.00
-0.98
-4.48
-3.36
-1.12
-2.58
-6.0 to -4.1 95% simultaneous confidence limits
91
-------
TABLE 12
QUALITATIVE SUMMARY OF CORROSION MEASUREMENTS
Method Used to Obtain
Corrosion Measurements
Baseline Operation
Low NOjj Operation
Low NOjj vs. Baseline
Operation
1. Furnace Wall Tubes
(Ultrasonic)
2. Test Panel Hibes
(Ultrasonic)
3. Probe
Coupons
(Weight)
One Day
Exposure
Burner area equal
to nonburner area
No significant
differences (except
panel No. 5)
Burner area loss
greater than non-
burner area
Burner area loss
greater than non-
burner area loss
Burner area loss
greater than non-
burner area loss
Burner area loss
greater than NOx
burner area loss
Low NOx loss > base-
line within burner
area
Low NQx loss > base-
line within burner
area. Low NOx loss
< baseline outside
burner area
No significant
difference
10-42 Days
Exposure
No significant
difference
No significant
difference
No significant
difference
-------
COMBINED-CYCLE POWERPLANT EMISSIONS
By J
P. L. Langsjoen, R. E. Thompson, L. J. Muzio
KVB, Inc.
Irvine, California 92714
and
M. W. McElroy
Electric Power Research Institute
Palo Alto, California 94304
93
-------
ABSTRACT
The retrofit of existing utility steam boilers with a combustion gas
turbine to supply hot vitiated combustion air to the windbox of a fired
boiler, in place of the normal forced-draft fans and air preheaters (i.e.,
repowering), can lead to increased power output at improved heat rates. A
major consideration in converting to combined-cycle operation is the impact on
the nitrogen oxides (N0X) emissions from the system.
A field test program was conducted to determine the NO^ characteristics
of a 220-MW supplementary-fired unit. A primary objective was to determine
the fraction of the gas-turbine-generated NO^ that can potentially be reduced
upon passage through the combustion zone of the boiler. As part of this test
program, the boiler was operated in a low-NO^, staged-combustion configuration
by removing selected burners from service.
Baseline NOx emissions from the combined-cycle system were found to be
substantially lower than NOx emissions from the boiler alone when operated
with ambient air supplied by forced-draft fans: 1.4 lb NC^/MW-hr compared to
2.3 lb N02/E4W-hr at boiler loads of 190 MW and 200 MW, respectively. In a
staged-combustion configuration with four of twenty burners removed from
service, the combined-cycle NOx emissions were reduced to 0.9 lb N02/MW-hr.
The fraction of gas-turbine-generated NO^ reduced upon passage through
the combustion zone of the boiler was determined by doping the gas turbine
fuel with nitrogen (ammonia) to artificially vary the boiler inlet N0x levels
during combined-cycle operation. The results showed that during normal opera-
tion of the combined-cycle system with all burners in service, 10 to
28 percent of the NOx produced by the gas turbine was reduced (destroyed) in
the supplementary fired boiler. During operation of the boiler in a combus-
tion configuration staged by removing four burners from service, a greater
portion of the gas-turbine-generated N0x was reduced in the boiler.
94
-------
ACKNOWLEDGMENT
This project was sponsored by the Electric Power Reserach Institute
(project RD 782) and conducted by KVB, Inc., the Babcock and Wilcox Company,
and General Electric Company. KVB, Inc. appreciates the cooperation and
efforts extended on the project by the B&W staff (D. Anaki, G. Brechun and
E. Campobenedetto), and General Electric.
Also we are very appreciative of the cooperation and technical support of
the Oklahoma Gas and Electric Company which provided a combined-cycle unit for
testing during this programs Horseshoe Lake Station Unit 7. The station
engineering and operating staff at Horseshoe Lake was very helpful in
equipment installation, instrumentation, establishment of numerous modified
operating modes, net heat rate determinations, control system analysis, and
engineering support.
95
-------
SECTION 1
INTRODUCTION
The retrofit of existing utility steam boilers with a combustion gas
turbine to supply hot vitiated combustion air to the windbox of a supplemen-
tary-fired boiler in place of the normal forced-draft fans and air preheaters
(i.e., repowering), can lead to increased power output at improved heat
rates. One of the main considerations in converting to combined-cycle
operation is the impact of this modification on the N0x emissions from the
system. The purpose of the study reported in this paper was to assess the NO
characteristics of a supplementary-fired combined-cycle system and the
effectiveness of combustion modifications to the boiler in reducing N0X
emissions.
Prior EPRI-sponsored research projects have indicated that there can be
emission as well as efficiency benefits in certain combined-cycle operating
modes. Specifically, a combined-cycle system has the potential of operating
with substantially lower emissions of N0X, CO, hydrocarbons, and smoke
compared to a gas turbine and a boiler operating separately (1-2) or even
compared to a boiler operating with some degree of combustion modifications
for NO control. In the case of NO , the gas turbine exhaust products with
X X
reduced oxygen content (approximately 17 percent Oj), supplied to the windbox
of the associated boiler, provide some of the benefits that normally would be
obtained with flue gas recirculation into the combustion air of a conven-
tionally-fired boiler. This effect would generally outweigh the N0„
disadvantage of higher windbox temperatures encountered in the combined-cycle
boiler as compared to a conventional boiler (typically 1000°F compared to
650°F). Thus, the N0x produced in the boiler of a combined-cycle system will
96
-------
be less than that produced utilizing a conventional preheated combustion air
system. Furthermore, laboratory-scale tests have indicated that the
application of two-staged combustion to the boiler could lead to additional
NO^ reductions of up to 55 percent from a supplementary-fired combined-cycle
system (2). This reduction would be accomplished by operating the burner
flame zone of the associated boiler in a low-NOx staged-combustion mode where
local regions in the boiler are operated fuel-rich. The staged combustion
mode not only reduces NO^ formed in the boiler but also leads to partial
destruction of gas-turbine-generated N0X as it passes through the fuel-rich
region.
Although the repowering concept is very attractive from both a heat rate
and emissions standpoint, there has been a lack of full-scale operational data
from existing combined-cycle installations to confirm the laboratory-scale
results for N0X control. The application of two-stage combustion is the key
element in emissions reduction. Therefore, the principal objective of this
test program was to obtain full-scale data to confirm the feasibility of
reducing gas-turbine-generated N0X in the fuel-rich flame of the associated
boiler.
Oklahoma Gas & Electric's Horseshoe Lake Plant was selected since the
combined-cycle unit at this station has one of the largest supplementary-fired
boilers permitting flexibility in establishing low-NOx operating modes.
During this study, both the gas turbine and boiler were fired with
natural gas, due to the unavailability of fuel oil at the time. Although this
might be viewed as a shortcoming of this work in light of current utility fuel
trends, it is believed that the results may be semi-quantitative for liquid
fuels as well.
PROGRAM APPROACH
To accomplish the major program objectives, it was desirable to establish
not only the total combined-cycle NOx emissions with modified combustion modes
applied to the boiler, but also to determine the specific reductions in
turbine-generated NOx in the fuel-rich boiler flame zone and the reduction in
boiler-generated N0X» A characterization of the amount of turbine-generated
97
-------
NO^ reduced in the boiler under varied boiler operating modes and principal
test parameters (load, burner air/fuel ratios, etc.) could then be used to
predict emissions reductions at other existing or proposed combined-cycle
units. It should be noted that these results are for a natural-gas-fired
system. Oil fuel N0x emissions reductions with modified combustion modes more
frequently depend on boiler firing configuration than those with natural gas
fuel.
An important aspect of the field testing at Horseshoe Lake was to isolate
the contributions to the total combined-cycle stack emissions by the boiler
and gas turbine. Since gas turbine emissions can be conveniently measured at
the boiler windbox inlet, and combined-cycle emissions are readily measured at
the stack, a determination of the boiler's contribution to N0X emissions
appears to be the only obstacle to determining the final contribution of the
gas turbine to the stack emissions. That is (combined-cycle stack N0X) -
(boiler NO contribution) = (net or final turbine N0X).
The Horseshoe Lake unit was equipped with back-up forced-draft fans
capable of providing full-load boiler combustion air requirements, so it might
intially appear that the boiler contribution to total N0X emissions could
conveniently be determined by operating the boiler alone with the FD fans.
However, this is not the case for two important reasons: (1) the combustion
air temperature would be ambient with only the forced-draft fans instead of
900°F when operating with the gas turbine (boiler air preheaters were not
provided); and (2) the unpreheated combustion air would contain 21 percent
oxygen instead of the approximately 17 percent oxygen present in the gas
turbine exhaust. Both air preheat temperature and combustion air oxygen
content (or equivalent gas recirculation rate) are known to have an important
influence on NO formation, particularly on gas fuel. Accurate means do not
exist to predict the effect of these two variables; hence, calculating the
boiler-produced N0X with combined-cycle operation using the data obtained with
forced-draft fan operation was not feasible.
The only practical approach was to preserve the desired boiler inlet
preheat and combustion air oxygen content while varying the inlet gas turbine
NO, thus generating a plot of stack total NO emissions as a function of gas
98
-------
turbine inlet NO to the boiler. In this fashion, the boiler contribution to
the total combined cycle (stack) NOx was obtained by extrapolating to a hypo-
thetical condition of zero gas turbine NO . This is illustrated in Figure 1.
Because the normal range of turbine operating parameters did not give the
desired range of inlet N0X levels to the boiler, doping the gas turbine fuel
with ammonia (NH^) was selected as a practical way to artificially increase
NO for purposes of this test. NH, reacts with oxygen in the gas turbine
corabustor, producing N2, NOxe and H2° as the reacti°n products? in effect, it
acts as a fuel nitrogen compound.
The principal boiler operating mode test parameters were load, excess air
level, burner air/fuel ratio, and burner firing pattern. The test program was
designed to provide answers to the following questions:
» To what degree can combined-cycle N0X emissions be reduced by
implementation of two-stage combustion in the boiler?
. What portion of the N0X reduction is attributable to reduced
N0X formation in the boiler and what portion is due to
destruction of turbine-generated N0X?
. How dependent is the N0X reduction on boiler burner firing
pattern, excess air level, load, etc., and to what degree are
these parameters coupled?
* What potential operational problems, impacts on efficiency,
etc., may exist with the implementation of this approach?
A number of limitations in the operating configuration of the boiler were
dictated by the gas turbine, which is essentially a constant-volume-flow
device with virtually no turndown capability. Because of this the
excess-oxygen-versus-load dependence of the boiler is essentially fixed, by
the size of the turbine, to provide all the boiler combustion air at full
boiler load. At reduced boiler loads, some of the excess combustion air must
be directed into the boiler hopper or dumped to the atmosphere.
Although this study is particularly relevant to the repowering of
existing gas- and oil-fired plants, there is a wide range of concepts or
interpretations of the term "repowering." As we have said, the elementary
approach is substitution of a gas turbine for the forced-draft fans so that
99
-------
the turbine exhaust satisfies the boiler's combustion air requirements are
full load. Other considerations, in addition to emissions, are (1) modifica-
tions to the windbox to accommodate 900°F+ preheat, (2) additional boiler
pressure drop and added heat load in the convective section, (3) physical
space for the added gas turbine, and (4) added combustion controls to
integrate turbine and boiler firing with accommodations for continued boiler
operation with gas turbine loss (3).
A complex approach to repowering is to add gas turbine capacity in excess
of that required for the burners in the supplementary-fired boiler, using the
boiler convective sections to absorb the extra heat load. This can be
accomplished by bypassing the burner regions with a portion of the gas turbine
flow and discharging it into the boiler through tempering ports at the top of
the radiant or furnace section (3). Other considerations include the ability
of the boiler steam circuits to absorb the additional heat load, possible
substitution of convective sections, and steam turbine load capacity.
In the more extreme cases of repowering, the boiler may have reached the
limits of its serviceable life or it is an inefficient low-pressure,
low-reheat design and it is therefore practical to totally replace it with a
new waste heat boiler. The existing steam turbine is usually left intact.
For these applications, other options must be evaluated: whether the boiler
should be supplementary-fired and what power split is most efficient between
the gas turbine and the steam turbine output. These are the considerations
involved in new combined-cycle unit design; frequently it is more effective
not to fire the new boiler. The reader is referred to EPRI Report FP-862 (3)
for a more detailed discussion of these considerations from the efficiency and
emissions standpoints. Other considerations are existing steam turbine
capacity, unit maintenance and life cycle costs, and lost generation during
repowering construction.
This paper deals with the current combined-cycle system at Oklahoma Gas
and Electric, Horseshoe Lake Station, which was designed so that the turbine
exhaust provides all the boiler combustion air. Therefore, these results are
directly applicable to the elementary repowering approach of direct substitu-
tion for the forced-draft fans. With some caution in interpretation, these
100
-------
results could also be applied to the case of additional gas turbine bypass
into the upper furnace region as long as this flow does not influence boiler
combustion conditions or emissions.
The final configuration already discussed requires repowering with a new
unfired waste heat boiler and thus does not directly relate to the
combined-cycle configuration tested during this program. Emissions are
directly dependent on the most recent advances in gas turbine combustor
design. Nor will these results be directly applicable to repowering
situations where duct burners are used to heat the gas turbine exhaust only
100 to 300°F. In this latter case, the duct burner fuel combustion consumes
only a minimal amount of the oxygen in the gas turbine exhaust, with the
majority of the gas turbine flow effectively bypassing the burners.
101
-------
SECTION 2
COMBINED-CYCLE SYSTEM DESCRIPTION
Oklahoma Gas & Electric's Unit No. 7, placed in service in 1963, is a
220-MW combined-cycle system incorporating a General Electric frame 8 gas
turbine and a front-wall-fired Babcock & Wilcox boiler with the following
general characteristics:
Gas turbine: Maximum output 25 MW
Boiler: Maximum continuous steam flow 1,286/000 lb/hr
Maximum continuous output 205 MW gross
Combined cycle: Heat rate 9629 Btu/kW-hr
(at 219.3 MW net station output, natural
gas fired)
The gas turbine exhaust normally provides all the oxidizer to the
boiler. Two forced-draft fans furnish combustion air to the boiler when the
gas turbine is not operating. Under normal operation, these fans also supple-
ment the turbine exhaust from the most economical load of approximately
1,145,000 lb steam/hr to the top load of 1,339,000 lb steam/hr. No air pre-
heating, other than turbine exhaust, is provided for the boiler; ambient-
temperature air is used for combustion when the gas turbine is out of service.
A. large economizer instead of an air preheater lowers the stack
temperature to nominally 321®F. With the gas turbine operating, all of its
exhaust must either pass through the burner region or the furnace hopper
bottom. (Gas tempering ports to bypass turbine exhaust to the boiler convec-
tive section were not in service, and a turbine startup bypass duct to the
atmosphere was not sufficient to handle the full turbine exhaust flow.)
102
-------
GAS TURBINE SYSTEM
The gas turbine generator is a General Electric Company frame 8 two-shaft
turbine rated at 25,000 kW with an 80°F compressor inlet temperature and
14»17 psia at the compressor inlet and turbine exhaust flange. At rated
conditions, the high-pressure turbine (which drives the compressor) turns at
approximately 3285 rpm. Gas temperature at the inlet to the first-stage
nozzles is 1500°F; the exhaust temperature is 870°F. This gas turbine unit
has no provisions for NOx control.
Control of the gas turbine fuel flow is dictated by control of the
exhaust gas temperature within specified limits. The control system is such
that, if ambient conditions change, the electrical load from the gas turbine
generator will change since the gas turbine compressor, running at constant
speed, will pump a constarc volume, but not constant weight or air.
No other means is employed to vary the steam generator excess air. The
gas turbine output depends on ambient temperature only. Air tempering dampers
installed with the unit to permit turbine exhaust bypass to the boiler convec-
tive section are not in use, and a bypass duct which diverts the turbine
exhaust to a separate stack is used only for startup.
BOILER SYSTEM
The steam generator is a front-wall-fired Babcock & Wilcox radiant-type
boiler with a nominal rated steam flow of 1,286,000 lb/hr (205 MW). The
physical configuration of the combined cycle unit is shown in Figure 2. The
pressurized furnace is normally fired with natural gas fuel, with oil on
standby. Une 20 center-fire Laredo-type burners are arranged in a four-wide
by five-high pattern as shown in Figure 3. Steam temperature control is by
means of spray attemperation and flue gas recirculation into the hopper bottom
of the furnace. Outlet superheat and reheat steam temperatures are 1005°F and
1005#F, respectively.
Over most of the combined-cycle load range, the gas turbine supplies all
the combustion air for the steam generator. However, combustion air is
supplemented by forced-draft fans at boiler loads about 180 MW, depending on
103
-------
the ambient temperature. At steam turbine gross loads below about 150 MW, the
excess air level at the burners becomes excessively high. When this occurs,
dampers are opened to divert some of the gas turbine exhaust into the bottom
hopper of the boiler, reducing the 02/fuel ratio through the burners.
AMMONIA INJECTION SYSTEM
A system was devised for injecting ammonia (NH3) into the fuel (not the
exhaust) system of the gas turbine so that the turbine exhaust N0X level could
be varied from approximately 50 to 100 ppm (dry at 15 percent 02). The
reacted in the gas turbine flame to form NOx, N2, and H20, greatly increasing
the total NOx output when injected in small concentrations. The limit of
100 ppm was chosen since in a prior laboratory study (2) it had been shown
that the gas turbine reduction factor was independent of the initial NOx level
up to 100 ppm, and data were not available to confirm the independent
relationship at levels much greater than 100 ppm. The ammonia was sprayed as
a liquid upstream against the direction of the natural gas flow at a point
just after the gas line control valve and just prior to a "T" in the line
which carried the gas to either side of the circular gas header to the
combustor cans. Subsequent evaporation of the maximum NH^ flow in the fuel
supply line would cool the natural gas by 5°F at most.
GASEOUS EMISSIONS SAMPLING SYSTEM
Boiler flue gas was sampled at the stack with six three-point probes.
Multi-point water-filled gas bubblers were used to obtain a composite flue gas
sample. The flue gas was drawn into the bubbler with an air aspirator. The
gas was analyzed with a Dynascience NOx analyzer, Beckman 02 and CO analyzers/
and an Orsat analyzer.
The gas turbine exhaust was sampled at the turbine outlet just before it
divides to supply the north and south sides of the boiler. Nine probes were
installed in the turbine exhaust duct. NO/NOx measurements at the turbine
exhaust were made with a Thermo-Electron chemiluminescent analyzer. Other
gaseous measurements were made with the following analyzers: Teledyne
104
-------
Model 326A, 0^; Horiba Model A1A-21 NDIR, CO2; and Beckman Model 315A NDIR,
CO.
SYSTEM TEST CONFIGURATIONS
The test program comprised a series of tests over a wide range of system
configurations. This included both combined-cycle and single cycle operation
(forced-draft fans only, no preheat). Staged combustion was also investigated
on both basic configurations. Table I outlines the system configurations and
range of variables investigated.
A more extensive test program was initially planned than is outlined in
Table I to investigate boiler performance and emissions over a wider range of
load and firing configurations. Unfortunately, the difficulties encountered
in gas turbine firing rate control with NH^ injected into the fuel could not
be resolved, and the extent of the resulting test data, although sufficient to
accomplish major project objectives, is considerably less than what was
initially desired for a thorough study of the topic.
DATA REDUCTION
Before discussing data reduction methods it is appropriate to briefly
reiterate the basis for which emissions reductions for combined-cycle systems
are being calculated. We are concerned with NO emissions control or N0„
reduction potential, so it is important to establish the conventional baseline
operating configurations to which the reduced emissions operating mode is
being compared.
There are two basic measure of interest in combined-cycle emissions in
the current study:
1. Reduction in total stack emissions by implementing staged
combustion in the boiler using as-found combined-cycle
emissions as a baseline.
2. Reduction in gas turbine emissions by the more fuel-rich boiler
flame zone; judged by comparison to the initial gas turbine
exhaust N0x level.
105
-------
The latter gas turbine NOx reduction potential can be expressed as:
r = 1
M M
NO NO
x , - x , . ,
stack boiler
mno
Xgas turbine
where:
r = fraction of oxides of nitrogen produced in the gas
turbine which is destroyed in passing thorough
the boiler
M
NO
stack
= mass emission rate of oxides of nitrogen in the combined
cycle exhaust products (lb/hr)
M,
NO = mass emission rate of oxides of nitrogen produced by the
boiler boiler while fired with gas turbine exhaust with no inlet
nitric oxide (lb/hr)
ft
NO
turbine
mass emission rate of oxides of nitrogen from the gas
turbine (lb/hr).
The interrelationship of these terms is shown in Figure 4.
k clear distinction should be made of the basis upon which the
combined-cycle system emissions are compared. This requires that boiler
operating conditions be clearly stated:
(a) The emissions from a boiler operating normally without flue
gas recirculation but with preheated combustion air
(21 percent O2, approximately 600°F).
(b) The emissions from a boiler operating without flue gas
recirculation but with combustion air at ambient temperature
(21 percent 02» approximately 70°F).
(c) The hypothetical emissions from a boiler operating on gas
turbine exhaust that has all the normal characteristics except
zero N0X content (approximately 17 percent C>2 and 900#F).
106
-------
Case (a) above would be of interest as representative of the current
emissions of a boiler being considered as a candidate for "repowering."
Case (b) is unlikely to be encountered in normal utility practice but is
of interest in terms of plant emissions when the gas turbine is out of
service. It also was investigated during the current program as an operating
configuration that would provide more insight into the emissions reduction
potential of low N0X firing modes in the boiler.
Case (c) is the hypothetical condition necessary to calculate the gas
turbine N0X reduction factor as indicated in the previous equation. It is
desirable to isolate the effect of the boiler flame zone conditions on the
destruction of the turbine exhaust N0X from the overall unit HO emissions so
that the results of this study can be generalized to other combinations of
turbine and boiler N0X. Otherwise, the study conclusions would be applicable
only to repowering cases having similar boiler and candidate gas turbine
emissions.
This raises the subject, touched on previously, of the possible
dependence of the N0X reduction factor on the gas turbine exit or boiler inlet
NO level. The subscale laboratory study (1-2) showed that the NO reduction
X X
factor (r) was not a function of the N0X levels in the gas turbine exhaust, at
least up to concentrations of 100 ppm (at 15 percent Oj). Data were not
available to assess whether or not the reduction factor was independent of the
boiler inlet NOx levels above 100 ppm. Based on the assumption that the
reduction factor is independent of inlet NOx level, the resulting nitric oxide
emissions from the combined cycle would plot linearly versus gas turbine NO
and could be linearly extrapolated to a condition of zero NOx in the gas
turbine exhaust, thus determining boiler-produced N0X« This approach was
followed in the reduction of Horseshoe Lake field test data and indeed did
confirm the assumption that the reduction factor is independent of gas turbine
NOx concentration.
The primary step in the data reduction procedure was the calculation of
the gas turbine NOx reduction factor. To do this it was necessary to express
all oxides of nitrogen emissions on a mass flow (lb/hr) basis. Other
quantities which were calculated from the measured parameters include the gas
107
-------
turbine and boiler air/fuel ratios (or stoichiometric ratios), and the
approximate stoichiometric ratios at the burners while operating in a staged
combustion configuration. The resulting data of stack NOx versus gas turbine
NO was plotted and the data extrapolated to zero gas turbine NO .
X X
This plot of stack N0X versus gas turbine N0X for a typical test
condition is shown in Figure 5. For this particular case the tests show that
the boiler would have produced 115 lb/hr of oxides of nitrogen if it had been
fired with gas turbine exhaust with no N0X present (intercept). Also, the
test results indicate that for this configuration, 49 percent of the oxides of
nitrogen produced in the gas turbine were destroyed upon passing through the
supplementary-fired boiler (i.e., r = 49 percent).
108
-------
SECTION 3
TEST RESULTS AND INTERPRETATION
BOILER-ONLY TESTS
The boiler-only tests provide an understanding of the responsiveness of
the unit to load variations, changes in excess oxygen, and burner-out-of-
service patterns for low N0X emissions. Performing these tests with the
forced-draft fans prevents problems in data interpretation introduced by the
gas turbine and allows an assessment of the boiler's behavior. These tests
were also conducted to select an effective pattern of burners out of service
to be used during the combined-cycle tests. Finally, the data obtained when
the boiler was operated with only the forced-draft fans provided a common
basis of comparison with the combined-cycle emissions. As previously
mentioned, this does not allow a totally valid comparison with normal utility
boiler operation as the combustion air is not preheated in this unit when
operating with the forced-draft fans. Since N0X formation, particularly for
natural gas firing, is a direct function of the combustion temperatures and
air preheat, the NO^ emissions for the boiler-only tests with no air preheat
are expected to be lower than would be emitted from a unit with combustion air
preheated (nominally 600°F).
The sensitivity of the N0X emissions to excess oxygen levels and load for
single-stage (ABIS: all burners in service) and staged combustion (BOOS:
burners out of service) is shown in Figure 6. At a load of 169 MW, increasing
the stack O2 level from 2 to 4 percent resulted in an 8 percent increase in NO
levels. At the lower load of 130 MW, the unit exhibited little sensitivity to
excess oxygen changes with only a 4 ppm increase in NOx as the 02 level in the
stack was increased from 3.9 to 5.0 percent. If the unit had been operated
109
-------
with the combustion air preheated to 600°F, it is estimated that boiler
emissions could be two to three times higher than measured during this program
with ambient temperature combustion air.
Staged combustion was implemented by removing burners from service. This
entailed terminating the fuel flow to selected burners while maintaining load,
resulting in an increase in fuel flow to the remaining in-service burners.
The nominal air flow was maintained through the out-of-service burners
(registers open). Staged combustion tests were conducted with forced-draft
fan operation to determine the overall sensitivity of the unit to staging and
identify the burner patterns to be used during the combined-cycle tests.
Again, during these tests, the boiler overall excess air level was maintained
at the same level as in the normal all-burners-in-service operating mode, and
the minimum excess air limit was not determined for each burner pattern
tested. During combined-cycle operation, the stack level was determined by
the power split between the gas turbine and supplementary-fired boiler and was
not at the discretion of the operator. These test results are summarized in
Table II.
In performing these tests, burner air registers were open to their normal
open position (60 to 70 percent) whenever burners were taken out of service.
The burner pattern testing focused on removing burners from the top two rows,
either the entire row—13, 14, 15, 16; 17, 18, 19, and 20—or the two middle
burners of each row—14, 15 or 18, 19. (See Figure 3 for the burner numbering
system.) Previous studies (4) have shown that removing burners from the upper
rows is more effective in reducing N0X than removing bottom rows of burners,
but it was beyond the scope of the present program to explore all possible
burner patterns for this unit. This is consistent with the objectives of the
program in investigating the combined-cycle NOx emission reduction potential
(in particular the reduction of the gas-turbine-generated NC>x through staged
combustion in the boiler). Since time was insufficient to explore the NOx
reduction potential of all burner patterns, a burner pattern was sought that
resulted in N0X reductions that were typical (not necessarily optimum) of
previous experience with gas-fired utility boilers. Further, it is not
currently known whether a burner pattern which produces minimum NOx with
110
-------
forced-draft fan operation necessarily results in the largest reduction in
gas-turbine-generated N0X> Previously reported laboratory studies (1-2)
indicate that the gas turbine N0X reduction increases as the stoichiometric
ratio of the in-service burners decreases.
Removing the middle burners of a row from service (14, 15) resulted in a
N0X reduction of 25 percent at a load of 169 MW (Test 15). When the middle
burners of the top row (18, 19) were removed from service, the NC>x reduction
was only 2 percent (Test 3); these tests were conducted at a higher overall
boiler excess air level. Based on the N0x-versus-02 characteristic shown in
Figure 6, the differences in the excess 02 level may have accounted for 25 ppm
or 14 percent of the reduction.
When four burners were removed from service, the N0X reductions obtained
were not highly dependent on whether the top row or second row was removed.
At a load of 169 MW and nominal stack 02 level of 4 percent, a 22 percent
reduction in N0X was measured with the top row out of service and 24 percent
reduction in N0X with the second row from the top removed from service. When
reducing the stack 02 level to 2.25 percent at 168 MW (Test 14) the N0X reduc-
tions increased to 43 percent with the second row from the top out of service
(13, 14, 15, 16).
Because removing from service either two burners or four burners of the
second row from the top resulted in N0X reductions typical of those achieved
in other gas-fired utility boilers, this pattern was selected for use during
the combined-cycle tests.
COMBINED-CYCLE TEST RESULTS
The primary objectives of the combined-cycle tests were to evaluate the
potential for destroying a portion of the N0X generated in the gas turbine and
to determine the relationship between this destruction process and the boiler
operating configuration. The boiler configurations evaluated included:
Baseline operation (all burners in service).
. Staged operation (burners 13# 14, 15, and 16 out of service).
Low load, with turbine exhaust bypass through the ash hopper.
Ill
-------
The results of these tests are summarized in Table III. Each test series
represents a fixed turbine operating condition and set boiler configuration
(i.e., excess 02, load, burner pattern). For each test series a number of
tests were conducted. The different test numbers in a series vary only the
gas turbine exhaust NOx level by varying the NH^ injected into the gas turbine
fuel. In the table the gas turbine emissions and combined-cycle stack
emissions are reported for conditions with no ammonia doping of the gas
turbine fuel.
Although a number of different boiler operating configurations were
tested to characterize the N0X destruction in the boiler flame, the gas
turbine was operated at relatively constant conditions throughout the test
series.
In analyzing combined-cycle test results, the following aspects need to
be assessed:
. Stack N0X emissions rate and the effect of operating
conditions.
. Boiler contribution to the stack N0X emission rates#
. Gas turbine reduction factor.
First/ consider the stack N0X emissions from the combined cycle. For
normal combined-cycle operations, the stack N0X emission rates were substan-
tially lower than those when the boiler was operated with only the forced-
draft fans. For example, comparing Tests 17 and 47 indicates that at a boiler
load of about 200 MW the combined cycle emits 305 lb/hr of NOx (as NC^)
whereas single-cycle operation with the forced-draft fans yielded stack N0x
emissions of 467 lb/hr. Combined-cycle emissions are 35 percent lower, even
when compared to operation with cold boiler air.
N0X emissions from the combined cycle are substantially lower if they are
normalized on a unit load basis as shown in Table IV and Figure 7a. In
Table IV and Figure 7a, the emissions are normalized by the load of the
specific device. Thus, for single-cycle operation, the boiler emissions are
normalized by the boiler load, the gas turbine by the gas turbine load and.
112
-------
for combined-cycle operation, the emissions are normalized by the
combined-cycle load (boiler plus gas turbine).
N0X emissions are plotted on a mass-emission-rate basis in Figure 7b,
illustrating an interesting comparison to the N0X emissions normalized on an
output-load basis. First, it can be seen that for single-cycle operation, the
mass emission rate from the gas turbine was substantially less than the N0x
emission rate from the boiler. Second, the emissions reduction with
combined-cycle operation compared to single-cycle operation was primarily due
to a reduction in the emissions from the boiler. Finally, with
staged-combustion operation of the combined cycle (4 BOOS), the NO^ emissions
were reduced by 36 percent compared to combined-cycle operation with all
burners in service. This 36 percent N0X reduction was a result of a
40 percent reduction in the boiler emissions and a 27 percent reduction in the
NOx contribution from the gas turbine. Thus, even though a substantial
fraction of the gas-turbine-generated N0X was destroyed in the boiler, the
overall NOx reduction from combined cycle operation with the implementation of
staged combustion was primarily due to reductions in the N0X contribution from
the boiler.
It should be noted that the single-cycle emissions in Table II were
obtained at an excess oxygen level approximately 1 percent higher than the
combined cycle operating mode. Even if the results were normalized to a
common O2 level, this would reduce the single-cycle emissions by about
10 percent, indicating that the combined-cycle emissions are still substan-
tially lower.
For normal operation of a typical natural-gas-fired utility boiler, the
NOx emissions were estimated to be up to two to three times higher than the
2.3 lb N02/MW-hr tabulated in Table IV due to the use of combustion air
preheat, which was not used in this case.
Similar comparisons can be made at other loads from Tables II and III,
with similar conclusions. Two factors affected the NOx emissions from the
combined-cycle configurations, in addition to the fact that the oxidizer
stream (gas turbine exhaust) contained NOx» The primary factor was the
presence of combustion products in the oxidizer stream (CC>2 and which
113
-------
acted as inerts and reduced the oxygen concentration in the boiler oxidizer
flow to 16.6 percent 02 compared to 21 percent. This was equivalent to
approximately 26 percent flue gas recirculation. The diluent effect resulted
in reduced flame temperatures and lower NOx formation in the boiler, similar
to flue gas recirculation. Second, counteracting this effect was a gas
turbine exhaust temperature of 800°F, whereas with the forced-draft fans the
air was at ambient temperature. However, the results suggest that the effect
of preheat temperature was less compared to the reduced oxygen content. This
may not be the case for an oil-fired unit with high fuel nitrogen content as
it is generally known that flue gas recirculation with high nitrogen fuels
(e.g., residual oil) is less effective than with natural gas fuel.
Next consider the effect of a staged boiler operating configuration on
stack NOx emissions from the combined-cycle operation. At the high boiler
load condition (190 to 198 MW), removing four burners from service resulted in
a 36 percent reduction in NOx emissions. It should also be noted that the
boiler O2 dropped to 1.9 percent during the tests with four burners out of
service at a load of 198 MW compared to the baseline O2, which was in the
range of about 2.7 percent at 190 MW. This should result in a further NOx
reduction beyond just removing the four burners from service at the same
level, as indicated in the general behavior of the unit with excess air
(forced-draft fan operation) as depicted in Figure 6. Thus the 36 percent NOx
reduction was due to both removing four burners from service and reducing the
02 level from 2.7 to 1.9 percent. It was not possible to independently
control the overall boiler 02 and the load during the combined-cycle operation
and thus establish the relationship between boiler excess 02 and
combined-cycle NO^ emissions at a fixed load. This limitation exists because
the gas turbine is essentially operated at a fixed condition with all of the
combustion products fed to the boiler. The boiler load, or fuel flow,
determined the flue gas oxygen concentration.
At the lower boiler load (150 MW) the stack NO emissions were not altered
by removing four burners from service {compare Tests 27-29 and 33-34). A
major contributing factor appears to be the unexpected increase in
boiler-produced NO accompanying staged combustion. Although no clearcut
114
-------
explanation can be offered for this, as it is contrary to the behavior with
the forced-draft fan operation, the excess oxygen level was high (6.2 percent}
during these combined-cycle tests. So, even with four burners out of service,
the in-service burners were still operating with excess oxygen.
During Tests 30-32, an attempt was made during combined-cycle operation
to lower the air/fuel ratio at the burners by diverting a portion of the gas
turbine exhaust away from the burners and adding it through the bottom hopper
as the only available means to simulate low-excess-air operation at the
burners. As seen in Table III, this resulted in a 28 percent increase in NOx
emissions. In actuality, this configuration did not simulate low-excess-air
operation as visual observation showed that the bypass flow through the hopper
flowed up the furnace walls into the burner region. This appears to have
resulted in enhanced mixing with the burner flow, producing more intense
flames. This in turn resulted in higher NOx formation in the boiler; none of
the gas-turbine-produced N0X was destroyed with this configuration.
Next consider the reduction of the gas-turbine-generated NOx which
occurred in the boiler. The test results for normal and low NOx
combined-cycle operating modes showed that from 10 to 49 percent of the NOx
generated in the gas turbine was destroyed upon passage through the boiler.
The specific reduction factors are tabulated as (r) in Table III. At a boiler
load of 150 MW, the reduction factor increased from 10 to 27 percent when four
burners were removed from service. However, it can be noted in Table III that
while the reduction factor increased, the boiler-produced N0X apparently
increased, resulting in no change in the stack NOx emission rate. No defini-
tive explanation of this result is presently available. As discussed above,
with bypass through the ash hopper to simulate low-excess-air operation at the
burners, none of the gas turbine N0X was destroyed in the boiler.
The reduction factors measured at the higher load conditions (188 to
198 MW, baseline) exhibited a similar trend. During Tests 27 and 28 with all
burners in service, the test results indicated that 28 percent of the turbine
NOx was destroyed. With four burners out of service, the reduction factor was
determined to be 49 percent. This, coupled with a 40 percent reduction in the
boiler-generated NOx as a result of the staged combustion mode, produced the
115
-------
overall 36 percent reduction in stack NO^ emissions. It should be noted that
between Tests 37-38 and 23-26 there was a decrease in excess 02, tending to
lower the NO^ emissions, and a slight increase in load, tending to offset the
reduction.
The previous laboratory studies (1-2) showed that the gas turbine N0X
reduction factor, r, was a function of the equivalence ratio (i.e., the reci-
procal of the stoichiometric ratio, = 1/SR, ) of the in-service burners.
b o
Thus as the in-service burners were made to operate more fuel-rich
(higher $ ), (r) increased. For the test series tabulated in Table III, the
b
approximate in-service burner equivalence ratio $ was calculated. The
resulting relationship between the NC>x reduction factor and the in-service
burner equivalence ratio is shown in Figure 8.
The data plotted in Figure 8 include both high and low load as well as as
all-burners-out-of-service test configurations. The relationship appears to
be linear with a correlation coefficient of 0.99 for a least squares linear
fit. However, the number of data points is limited, and further data are
needed to substantiate the relationship.
GENERALIZATION OF TEST RESULTS
The test results have generally shown that (1) combined-cycle NOx
emissions are less than single-cycle emissions and (2) a substantial portion
of turbine-generated N0X can be destroyed in the boiler. The extent to which
these results can be treated as "general" warrants discussion.
The extent to which the combined-cycle emissions are lower than normal
will be dependent on the function of the power split of the combined-cycle
system. For these tests approximately 27 percent of the fuel is burned in the
gas turbine. As lower-heat-rate turbines come into use, less fuel will be
burned in the boiler; this will affect both the N0X formed in the boiler and
the reduction of turbine-generated NOx* In fact, the most thermally efficient
combined-cycle system is comprised of a gas turbine and unfired waste heat
boiler. The emission characteristics of this system will be greatly different
from the unit tested during this program.
116
-------
Bypassing a portion of the gas turbine exhaust to the ash hopper produced
interesting and unexpected results. Operational situations can exist at low
load where bypassing a portion of the gas turbine exhaust from the burners is
desirable. One might be inclined to conclude that the NOx emissions should
fall. Instead, these tests results indicated that bypassing the gas turbine
exhaust to the ash hopper yielded more intensely mixed flames, producing
higher N0X» A more successful approach might be to divert the flow and inject
it in a region above the burners, avoiding mixing of the bypass with the
burner flow.
Also worthy of discussion is the method of implementing staged combustion
for low N0X operation. In this study burners were removed from service and
the fuel diverted to in-service burners to create fuel-rich regions. With
this configuration, the portion of the gas turbine exhaust which passes
through the out-of-service burners does not directly pass through a fuel-rich
flame. One might anticipate that this portion of the gas-turbine-generated
NOx cannot be reduced, limiting the reduction factor. This turns out not to
be the case. In reported small-scale tests (2), it was shown that not only
was the N0X reduced in the portion of the gas turbine stream fed directly to a
fuel-rich burner, but also a portion of the N0X in the staged oxidizer flow
that bypassed the burner was also reduced. It was not possible to quantify
the relative portion representing the fraction of the N0X destroyed passing
through the fuel-rich region and the fraction reduced as the staged air is
mixed with the fuel-rich stream. Reduction factor (r) reported in (2) repre-
sents the total reduction of the gas turbine U0X and is thus directly
comparable to the reduction factor calculated in this study (i.e., the reduc-
tion factor is not apportioned to the fuel through the burner and that which
is bypassed as staged air).
A similar situation will probably exist with low N0X burners that may be
available in future new installations. These low N0X burners will more than
likely achieve reduced N0X by aerodynamically staging the flame at the burner
rather than staging the entire furnace, as occurs with burners-out-of-service
operation. Since the low N0X burners use a form of staging with probably
comparable boiler N0X levels, it is reasonable to anticipate a reduction in
117
-------
gas-turbine-generated N0X using low NOx burners in combined-cycle systems.
However, data are not available to quantify these anticipated reductions
relative to the results with burners out of service tested during this study.
A final note of caution is appropriate in the use of gas turbine and
boiler emissions data obtained during this program. As with all combustion
equipment N0X emissions, they are design type, fuel type, operating mode, and
site specific. Therefore, one should use considerable care in applying the
data obtained during this test series to predict emissions from future
potential repowering candidates. This is particularly true because the
General Electric frame 8 two-shaft gas turbine is not a recent design and is
not completely representative of recent gas turbine emission factors for
NO . Although the data here can be effectively used in a proportional NO
reduction manner for comparative purposes in evaluating operating modes, its
use in an absolute sense for emissions forecasting is ill advised.
Additionally it must be noted that this study was conducted with a
natural-gas-fired system (both turbine and boiler). Current interest focuses
on oil-fired systems. A note of caution is thus offered in extrapolating
these natural-gas-fired results to oil firing.
118
-------
REFERENCES
1, Hunter, S.C., "Combined Cycle N0X Study," final report, EPRI report 224-
1, Volume I, February, 1975.
2. Arand, J.X., "Reduction of NO Through Staged Combustion in Combined Cycle
Supplemental Boilers," EPRI report 224-1, Volumes II and III, February,
1975*
3. Foster-Pegg, R.W., "Combustion Turbine Repowering of Reheat steam Power
Plants," EPRI Rpoert FP-862, August, 1978 (final report of research
project 528-1, Westinghouse Electric Corporation.J
4, McGuire, W.G. et al., "Theory and Application of Nitric Oxide Emission
Reduction in Utility Boilers," presented at the First Annual Symposium on
Air Pollution Control in the Southwest, Texas A&M University, November,
1973.
119
-------
X
O
z
M
u
fl
¦u
w
(3
¦P
0
tr<
Gas Turbine NO
(Inlet to Boiler)
a * normal turbine NOx
emissions
b ¦ higher NO due to
NH3 injection in fuel
c » max NO with NH
injection in tGrbine
fuel
arrow indicates boiler
contribution to stack NO
Figure 1. Sketch of stack NOx levels versus gas turbine NO levels
120
-------
STAC*
to'-o"
BSE
+. ¦ * ¦ i I. L
L
SECONDUTf
SUP £ PLATER
A
i
RDitAT
SUP£Kh£ATE«
?^?*35E
* - --^ "5 WKMAirr &2
r
Xtm
IM®
5'"* ii i*^T
iSUMH!
312
/MV
SECONDARY
ECONOMIZER
economizer
147
J Tcas-T
TEMPERING
OAS TURBINE
WCED
0MT
FAM
eoMBusnoN
m
M-0
Stack Sampling
Location
OKLAHOMA OAS a ELECTRIC COMPANY
HORSESHOE LAKE STATION-UNIT NO. 7
HARRAH, OKLAHOMA
.Turbine Exhaust Sampling
Location
Figure 2. Physical configuration of OG&E Unit No. 7
121
-------
Sldevall Tubes
Cn
cn
B'-O
B'-O
8"-0
Sidevall Tube*
Figure 3. Arrangement of burners in the steam generator
122
-------
Air
Turbine
Fuel-
Turbine Exhaust
Fuel
to Boiler Windbox
"HJOXg,, turbine'
Supplementary-
Fired Boiler
Boiler Exhaust to Stack
*NOxk
"Wxatack " "NO*boiler
+ (1-r) Mjgox„ . . .
^"¦"'gas turbine
Figure 4. Sketch of the NOx contributions for the combined-cycle system
123
-------
400
300
200
100
/¦
-\
"Hue*
= 115 + 0.51 ft
TlOx
turbine
r = 1-0.51 ¦ 0.49
^NOxk = lb/hr
(coefficient of determination:
0.97)
100
200
300
Vx U,/BR
TURBINE
Figure 5. Typical combined-cycle stack NOx versus gas turbine NOx
124
-------
169 MW (ABIS)
200
130 MW (ABIS)
169 MW (4 BOOS)
-------
in
9
«
w
>-)
M
o
m
S W
S co
O
n cm
w
z
M
s
D
H
O
SINGLE CYCLE
w
M
9
T7T
o
o
CQ
COMBINED CYCLE
800
600
14
X
400
O
2
X
o
ss
200
*1200 lb/hr
to
OS
u
tJ
H
O
03
V)
8
a
0C
Cd
iJ
M
O
CQ
D
E-"
3
O
SINGLE CYCLE
CO
S*
u
o£
a
M
O
CQ
E-<
O
or;
a
M
O
CO
COMBINED CYCLE
(a)
Normalized by unit load
(b)
Mass emission rates
ABIS = All Burners In Service
BOOS = Burners-Out-Of-Service
Figure 7. Comparison of single-cycle and combined-cycle emissions,
load approximately 200 MW, gas turbine load 25 MW.)
(OG&E Unit No. 7 boiler
•Arrows indicate estimated increase in boiler NOx emissions if operated with normal 600°F
combustion air preheat instead of ambient air.
-------
1
1 1 1
1
1
1
In-Service
Fuel Lean
Burners
1
I
Fuel Rich
—
1
1
1
—
¦
1
1
(23-26]
1 .
•" 4 BOOS
(37-38)
ABIS 1
(3 3-34 J
4 BOOS |
|
v«/(27-29) j
...1
ABiq | ,
»
0.6 0.7 0.8 0.9 1.0
BURNER EQUIVALENCE RATIO
1.1
1.2
Figure 8. Correlation of NOx reduction factor (r) with the in-service
burner equivalence ratio
127
-------
TABLE I. BASIC TEST PLAN
Single-Cycle Operation
Variable
Range Tested
Number of
Variations
Load
130 - 200 MW
3
Boiler O2 concentration
2.25 - 5.0%
3
Burner configuration
All burners in service
1
2 burners out of service
2
A burners out of service
2
Combined-Cycle Operation
Variable
Range Tested
Number of
Variations
Gas turbine load
25 MW
1*
Gas turbine NO level
(NH^ injection into
turbine fuel)
50 - 105 ppm
3
Boiler load
150 - 190 MW
2
Boiler O2 concentration
1.9 - 6.2%
3
All burners in service
1
4 burners out of service
1
Partial turbine exhaust
bypassed to bottom hoppert
1
*Gas turbine load was essentially constant throughout the test program.
tTo simulate low excess air at the burner.
128
-------
No.
1
2
3
4
5
6
7
8
9
10
11
13
14
15
16
17
TABLE II. BOILER-ONLY TEST RESULTS
NO
Burners Out Load Flue Gas Flow 0 CO ppm @ lb/hr
of Service MW lb/hr % ary ppm 3% Oj as NO2
_0 138 l.lOxlO6 4.75 0 152 212
0 169 1.39xl06 3.85 0 199 367
2 (#18,19) 167 1.38xl06 3.80 0 196 360
4 (#17,18,19,20) 165 1.39xl06 4%05 0 155 283
4 (#13,14,15,16) 167 1.40xlQ6 4.10 0 151 276
0 130 1.16xl06 5.00 0 186 268
2 (#14,15) 130 1.15xl06 4.90 0 166 240
4 (#13,14,15,16) 130 1.16xl06 4.95 0 138 200
4 (#13,14,15,16) 129 1.09xl06 4.15 0 122 174
_0 130 1.08xl06 3.90 Q 182 260
_0 130 5.00 0 186 263
JO 169 1.37xl06 3.60 0 200 367
4 (#13,14,15,16) 168 1.26xl06 2.25 105 99 179
2 (#14,15) 167 1.27xl06 2.58 235 130 234
_0 165 1.27xl06 2.50 0 173 311
0 200 1.68xl06 3.75 0 209 467
2 (#14,15) 19? 1.68xl06 3.75 0 160 357
-------
TABLE III. SUMMARY OF COMBINED-CYCLE TEST RESULTS
Test
Series
Test
Description
Turb.
Load
KW
Boiler
Load
MM
Boiler
02
* Dry
Stack
CO
ppm
Turbine
NOx
lb/hr*
Boiler
NOx*
lb/hr*
Stack
NOx
lb/hr«
r
SRb+
4 s
b
27,28,29
ABIS
26
150
6.2
0
146
96
2 29
10
1.37
0. 73
37,38
ABIS
24
190
2.7
0
156
193
305
26
1.12
0.69
30,31,32
ABIS, hop-
25
150
5.B
0
147
145
293
0
—
—
per bypass)
23,24,25
4 BOOS (#13,
25
196
1.9
eo-
148
115
195
49
0.86
1.16
26
14,15,16)
125
33,34
4 BOOS (#13,
23
150
6.2
0
137
129
229
27
1.10
0.91
14,15,16)
• As N02
+ SR. - stoichiometric ratio of the in-service burners, stoichiometric
based on the boiler 02 measurement
SR^ >1 —> oxygen lean
SRfc <1 —> fuel rich
S $bm i/sflj,
# Determined by a least squares fit! thus nay not exactly satisfy the equation
+ (l-r) ®H0gt' u>in9 the values in the table.
TABLE IV. COMPARISON OF COMBINED- AND SINGLE-CYCLE EMISSIONS
Single Cycle Operation
Boiler (cold air operation, ABIS) : (467 lb/hr)/(200 MW) « 2.3 lb N9.2-
MW-hr
(cold air operation, 2 BOOS): (357 lb/hr)/(199 MW)
» i o ±k-$°2_
1,8 MW-hr
Gas turbine: (148 lb/hr)/<25 KM] =5.9 Np2-
MW-hr
Combined- Cycle Operation
Combined cycle (ABIS) : (305 lb/hr)/(24 MW+190 MW) = 1.4 lb N°2-
MW-hr
Boiler NOx (ABIS): (193 lb/hr)/(l90 MW) » 1.0 ~~?2-
MW-nr
Combined cycle (4 BOOS): (195 lb/hr)/(198 MW+25 MW) « 0.9 ^ N?2-
MW—nr
130
-------
PILOT SCALE EVALUATION OF A LOW NOx
TANGENTIAL FIRING METHOD
By:
J. T. Kelly, R. A. Brown, J. B. Wightman
R. L. Pam, E. K. Chu
Acurex Corporation
Energy & Environmental Division
485 Clyde Avenue
Mountain View, California 94042
131
-------
ABSTRACT
The Environmental Protection Agency (EPA)/Acurex 293 kWt pilot-scale
facility was used to develop a low-NOx pulverized coal-fired tangential
system. Low NO^ is achieved by directing the fuel and less than
20 percent of the secondary combustion air into the center of the furnace
with the remaining secondary combustion air directed parallel to the
furnace walls. The separation of secondary combustion air in this manner
creates a fuel-rich zone in the center of the furnace where NO
x
production is minimized. This combustion modification technique has
lowered NO 65 percent relative to conventional tangential firing. In
X i
addition, CO, UHC, and unburned carbon emissions are substantially
unaffected by the modification. Also, the modification places a blanket
of air on the furnace walls which is beneficial from a wall corrosion and
slagging point of view. Finally, the modification shows a decrease in
N0x emissions as firebox gas temperature is increased. This
characteristic might be beneficially applied in a large-scale system to
reduce furnace volume, and thereby capital cost, for a given combustion
heat release.
Tests are now underway to further optimize and characterize this
low-NO combustion modification technique.
x
132
-------
ACKNOWLEDGEMENTS
This work is supported by the EPA under Contract 68-02-1885.
David G. Lachapelle is the EPA Project Officer. The assistance of Acurex
staff, P. M. Goldberg and E. B. Merrick, in this study is gratefully
acknowledged.
133
-------
INTRODUCTION
Maintenance of ambient air quality in the United States requires the
restriction of N0x emissions from stationary combustion sources.
Tangentially coal-fired utility boilers account for about half of the
steam-electric capacity generated in the U.S. by coal-fired utility boilers
and produce about 40 percent of the N0^ emissions attributed to coal-fired
boilers (Reference 1). These significant N0x emissions and the projected
increase in the number of these boilers make them candidates for emission
control development both in terms of retrofit and new boiler designs.
During the combustion of pulverized coal, NO^ is generated from the
nitrogen chemically bound in the fuel as well as from the oxidation of
atmospheric nitrogen. For typical bituminous coals, NO emissions from
the fuel-bound nitrogen can be a significant fraction of the total
(Reference 2). NO emissions have been shown to respond to combustion
X
modification techniques that alter oxygen concentration, residence time, and
temperature during combustion (Reference 3). Lowering the oxygen
concentration surrounding the fuel, either locally by fuel/air stratification
or globally by limiting'the air flow in the combustion volume, shifts the
fuel and atmospheric nitrogen emission reactions from predominantly NO^
formation to a balance between N0x and molecular nitrogen formation
(Reference 3). In addition, given sufficient residence time at oxygen
deficient conditions, previously formed NO can be reduced to molecular
x
nitrogen by homogeneous (Reference 4) and heterogeneous (Reference 5)
catalyzed and noncatalyzed reactions.
Lowering peak temperature under excess air conditions decreases
atmospheric nitrogen N0x formation (Reference 6). However, under very
fuel-rich staged combustion conditions, lowering first stage temperature can
134
-------
increase NO^ (Reference 6). This is due to less fuel nitrogen being
volatilized in the first stage and carrying over and being converted into
NO^ in the oxygen-rich second stage.
Even though imperfectly understood, these basic relationships between
system parameters and N0x emissions have been employed to moderately
reduce N0x emissions from tangential as well as other types of utility
boilers (Reference 7). Some combustion modification techniques, such as
staged combustion, can lead to significant N0x reduction. However, a
major portion of the firebox is operated under reducing conditions, tdiich is
undesirable from a wall corrosion and slagging point of view. Significant
further reductions in N0x from tangentially-fired boilers under acceptable
combustion conditions requires a better understanding of the combustion
processes that control NO^ formation/reduction. Therefore, this study is
separated into two phases. The objective of the first phase is to develop
an understanding of the processes controlling NO formation/reduction in
X
pulverized coal-fired tangential boilers. Using the results of the first
phase, the objective of the second phase is to develop and demonstrate, in
pilot-scale, low-NOx combustion modification techniques that can be
retrofitted to existing, or incorporated into new, tangential boiler designs.
135
-------
DEFINITION OF COAL-FIRED TANGENTIAL SYSTEMS
Figure 1 illustrates the main features and flow patterns of a
tangentially-fired boiler. Fuel and air are introduced into the furnace
through rectangular registers located in the four corners. The bulk of the
combustion air enters above and below the fuel jet as shown. The jets are
nonswirling and fuel/air mixing is slow relative to burners used in
wall-fired boilers.
The tangential alignment of the centerlines of the corner jets to the
circumference of a circle in the center of the furnace promotes the
formation of a large-scale vortex within the furnace. Ignition of the fuel
is provided by impingement of hot burnt gases from laterally adjacent
burners and large-scale internal recirculation of combusted gases. Because
ignition occurs primarily on the vortex core side of the fuel jet,
combustion is asymmetric in the horizontal plane.
In addition to providing ignition, jet impingement and vortex
interaction help mix fuel and air to complete combustion. Partially burnt
gases from lower burner levels sweep past and interact and mix with higher
burner level fuel and air jets, helping to mix the fuel and air.
Pilot-Scale Combustion Facility
Figure 2 shows the EPA/Acurex pilot-scale facility used for the
baseline and combustion modification testing during this study. This
facility has been used to study various firing modes (tangential, front-wall-
fired) and fuels (coal, oil, gas, coal-oil mixture, refuse-derived fuel)
(References 6 and 8). The maximum firing rate is approximately 600 kWt.
The firebox is a 99-cm refractory-lined cube attached to a 61-cm
refractory-lined tower in which convective heat exchangers are placed.
Volumetric heat release, overall residence time, and furnace exit gas
136
-------
temperature are matched between pilot- and full-scale facilities. Also,
burners and their placement in the firebox are patterned after full-scale
tangential systems. During tests, NO, C>2, C02, CO, and UHC emissions
are continuously monitored with particulate samples taken intermittently.
Table 1 lists the continuous measurement instrumentation. A complete
description of the facility is given in Reference 6.
The simulation of full-scale firebox flow patterns and mixing by the
pilot-scale facility was evaluated by comparing water model flow and
pilot-scale flow and flame patterns to corresponding full-scale results. In
the comparison, similarities were found for (1) ignition standoff and
character, (2) flame spreading angle from burners, (3) apparent jet
centerline angle from corners, and (4) vortex size.
Baseline Test Results
Figure 3 compares the NO emission levels achieved by the pilot-scale
facility at various excess air levels on several coal types to full-scale
utility boiler levels. The pilot-scale results correspond well with the
full-scale results. Matching of the NO trend with excess air is encouraging
in that this, as well as the abovementioned comparison of flame patterns,
may be an indication of the matching of mixing processes between the full-
and pilot-scale systems.
During baseline testing the CO, UHC, and carbon loss emissions were
small and comparable to full-scale system levels indicating that complete
combustion is occurring in the pilot-scale facility.
137
-------
JET AND VORTEX CHARACTERIZATION TEST RESULTS
To assist in the definition of NO emission control strategies,
conventional tangential-fired tests were carried out to characterize the
important processes to NO formation/reduction in this system and establish
the effect of design variables on these processes. During the
characterization tests, the facility was operated at baseline conditions.
The fuel chosen for all combustion system definition and modification
testing was Utah bituminous coal. The fuel properties are given in Table 2.
In-Flame Sampling Test Results
In-flame gas and solid samples were taken by a water-quenched probe
at a variety of firebox locations to determine the relative importance of
near-burner, jet interaction, and vortex zones (see Figure 1 for zone
definitions) on NO processes. The probe quenches particle and gas reactions
by injecting water directly into the sample stream at the probe entrance. A
limited number of sampling locations 7.6 cm below, 26.7 cm above, and at the
burner centerline were chosen to characterize near-burner, jet interaction,
and vortex zones.
Figure 4 presents the as-measured NO, 0and CO^ concentrations,
respectively, at the fuel tube elevation superimposed on a plan view of the
firebox. These results show that, near the burner face, minimum NO occurs
in the center of the fuel jet, where the fuel is relatively unburnt, and
peaks on the vortex core side of the jet. The peak NO levels occur in hot
ignition zones observed during testing. These zones, defined as jet
interaction zones in Figure 1, are created by the interaction and mixing of
the fuel jet with hot combustion gases from the adjacent upstream corner
burner. Since the fuel is only 60 percent burnt and has been in this zone
less than 100 ms, most of the combustion and NO production in this zone can
138
-------
be associated with the volatile components of the fuel. The bulk of the
total net NO production occurs in this near-burner region.
Downstream of the burner face and just prior to the next burner
interaction, 0^ and CO 2 concentrations are more uniform and the NO peak
is not as pronounced. Oxygen concentrations are low and NO concentrations
are high in this zone, and the net NO production is small. This zone is
downstream of where the vortex was observed to significantly impact with and
cause rapid mixing of the fuel and air jets. Burning in this zone and
downstream is mostly char combustion.
Probing results 26.7 cm above the burner level show C02> 0^, and
NO levels are relatively flat and consistent with furnace exit conditions.
Very little net NO production occurs in this zone and solid samples show
that only 10 percent of the fuel is unburnt at this level.
Staged combustion probing tests were run at a firebox stoichiometric
ratio (SR) of 0.85 with an overall stack excess air level of 15 percent. As
expected (Reference 6), staged combustion significantly lowered stack NO
through reduction of prior-formed NO and lower conversion efficiency of fuel
N to NO. Of greatest interest were the probing results presented in
Figure 5, which showed that, near the burner face, the vortex core side NO
formation peak was eliminated by operating the firebox fuel rich. Operating
the active zone on the vortex side of the fuel jet under 02~deficient
conditions appeared to eliminate the peak NO formation leading to a low
firebox burner elevation NO level which is even further reduced by decay
processes occurring above the furnace burner elevation.
NO Dopant Test Results
Information on the NO reduction capability of various zones in the
firebox was obtained by injecting NO dopant into the firebox. A ceramic
injector tube 1.9 cm in diameter was used to distribute NO over roughly a
horizontal line source 15 cm long. The NO flow in the injector was set at a
level required to give 375 ppm in the stack for an inert injectant.
Injection locations were chosen above, at, and below the burner centerline
to assess the NO reduction potential of the near-burner and vortex zones
probed earlier. NO dopant was also injected into the primary air/coal
139
-------
supply line to assess the effect of the very-near-burner zone on NO
reduction. Figure 6 presents, on a plan view of the furnace, the percent of
NO dopant remaining in the stack for the indicated injectant locations. The
percent remaining is found by subtracting the exit NO concentration without
dopant from the measured doped level and dividing by the theoretical inert
injectant concentration of 375 ppm. Conventional as well as staged
combustion conditions at a first-stage SR of 0.85 were tested.
For conventional combustion conditions, injecting NO on the fuel jet
centerline near the burner gives NO reduction efficiencies of 70 to
80 percent. Away from the burner and in the vortex, reducton efficiency is
about 50 percent. In the active zone on the vortex core side of the fuel
jet, explored in the probing tests, reduction is 94 percent. Below the
burner level, NO reductions of 80 percent were measured whereas above the
burner level reductions were small, being less than 13 percent.
Under staged combustion conditions, at a first-stage SR of 0.85,
reductions at the burner elevation were better by a factor of two from the
unstaged results, except at a single point for which no explanation can be
given. Below the burner level, reductions observed were about the same as
conventional unstaged reductions. Above the burners, the reductions were
significantly better for the staged conditions.
These results show that NO is most effectively reduced if the NO is
injected into the active combustion and peak NO production zone formed by
the interaction of hot burnt gases and the fuel jet. In this zone, reaction
is probably fast and addition of NO can drive the reactions from NO^
production toward a balancing of NO production and reduction. Another
X
effective reduction zone is near the burner face at the burner elevation.
In this zone, oxygen is not abundant and NO is reduced. Below the burner
level, NO reduction is effective for both staged and unstaged conditions.
Since reduction is not observed above the burner level for lean conditions,
this indicates that NO injected below the burner centerline must get
entrained and reduced in burner elevation flame zones.
140
-------
TANGENTIAL BURNER CHARACTERIZATION TEST RESULTS
Different burner designs were tested to assess their impact on NO and
to determine the relationship between burner design parameters and NO.
Burner designs tested were limited to nonswirling slow mix designs such as
those presently used in tangential systems. Compact intense flames produced
by swirl burners are not compatible with tangential firing due to potential
corner slagging and deposition problems.
Based on the probing test results and the known importance of 0^
availability in the fuel jet to NO (Reference 6), burner designs were tested
where the exposure of the fuel jet to combustion air and the hot combustion
gases that provide ignition was varied. Also, in some of these tests the
relative position between the fuel and air was varied to operate the vortex
side of the flame more fuel-rich. For the burner design tests, three
conventional baseline burners firing on gas and one experimental burner
firing on coal were used to generate a conventional tangential system vortex.
Figure 7 presents the NO results and the burner fuel and air port
configurations tested. In the burner configurations given in Figure 7, the
open circles correspond to the end view of the burner combustion air supply
tubes and the filled circles represent the fuel supply tubes. The NO
results are given as a function of temperature since baseline testing showed
that NO increases with temperature. Also shown is the NO level when all
four corner burners are operated on gas. The results are uncorrected for
the dilution effect of the lower NO gas combustion products. Actual
differences in NO due to coal burner alterations are greater than, by up to
a factor of four, the differences appearing in Figure 7. Going from
configuration 1 to 5, exposure of the fuel jet to hot combustion gases
decreases. Configuration 1 is the fuel jet alone with all of the secondary
combustion air distributed to the other three corner burners firing on gas.
141
-------
The NO results in Figure 7 show that configurations 2 and 5 have, for
a given temperature, the lowest and highest NO levels of all the burners
tested. In configuration 5, the fuel jet is surrounded by combustion air
and entrainment of hot combustion gases into the fuel jet is limited.
Configuration 2 has the 'fuel jet located on the vortex core side of the
furnace with the secondary air jets aligned in a vertical row behind the
fuel jet. It was observed that configuration 2 had rapid ignition and
burning compared to configuration 5.
Configuration 1, the fuel jet by itself, had the greatest exposure to
hot combustion gases of all the configurations tested, and had NO results
similar to conventional tangential burner type arrangements such as
configurations 3 and 4. It was observed that the fuel jet by itself was
heavily impacted by the vortex flow, and fuel jet spreading and mixing was
extreme compared to other configurations.
142
-------
FIREBOX MIXING TESTS
Burner configurations 2, 3, and 4 were tested for the effect of
firebox mixing intensity and jet breakup on NO by varying the three gas
burner firing rates while maintaining the coal burner at a constant 73 kW.
Varying the gas firing rate changes the vortex strength and th^ turbulence
intensity in the firebox, which alters the coal burner mixing.
Figure 8 shows the NO results for configurations 2, 3, and 4 for gas
firing rates from 0 to 293 kW. At 0 kW gas firing, the vortex is absent and
the burner flames are long and narrow. As the gas firing rate is increased,
the firebox vortex develops and the coal flame begins to broaden somewhat
near the burner face and bend slightly while ignition and burning increase
on the vortex side of the fuel jet. An even more significant change-in
flame character with increasing gas firing rate is the strong impact and
dispersion of the fuel jet about one-half the firebox length from the burner
face. At this location, vortex interaction with the jet is significant
enough to destroy the original collimated character of the coal jet and
disperse the fuel in the firebox. This fuel dispersion in an overall
oxygen-rich environment is detrimental to NO emissions. Figure 8 shows that
the as-measured NO results do not change significantly for gas firing rates
from 0 to 293 kW. However, these results are uncorrected for dilution by
the considerably lower NO content gas burner combustion products. A simple
correction for dilution would show a consistent and large (an estimated
900 ppm at maximum gas firing rate) increase of NO for the coal burner as
the gas firing rate is increased. This demonstrates the importance of
firebox mixing on NO.
The results in Figure 8 also showed the importance of the vortex to
maintaining ignition and preventing lifted flames which have high NO. At
50 kW gas firing rate, or higher, ignition of the coal flame by the vortex
143
-------
is very positive and the difference in NO between configurations is small.
At 0 kW gas firing rate, where the vortex is absent and ignition is somewhat
tenuous, configuration 4 experienced positive ignition at the burner face
whereas configuration 3 did not and the flame became detached. This resulted
in an increase in NO from the 200 to the 430 ppm NO level. This abrupt
increase in NO has been observed elsewhere (Reference 11) for detached
flames.
144
-------
DISCUSSION OF TANGENTIAL SYSTEM CHARACTERIZATION RESULTS
Probing tests showed that, near the burner face, where the bulk of
total NO is formed, combustion is asymmetric with ignition, intense burning
and peak NO production occurring on the vortex core side of the fuel jet in
the jet interaction zone. At this location approximately 60 percent of the
fuel has been burned and this fraction can be associated primarily with the
fuel volatiles. Fuel/ai' mixing in this zone is enhanced by the crossflow
of hot combustion gases over the fuel and air jets. Since the initial fuel
nitrogen volatiles see an abundance of oxygen in this zone, NO formation is
very high (Reference 6). Lifted flame and dispersed fuel jet burner test
results were extreme examples of how high concentrations at the fuel
ignition point can lead to high NO. In addition, this zone has a high gas
temperature, due to reduced wall heat transfer and high entrained combustion
gas temperature. High temperature under O^-rich conditions generates
significant atmospheric nitrogen NO (Reference 6). NO dopant tests also
showed that NO can be significantly reduced in the near burner zone, with
the reduction most effective under fuel-rich combustion conditions. Adding
NO in concentrated form to this very chemically active zone reduces some of
the NO to N2 even under overall lean conditions.
As shown by the staged probing and burner configuration tests, the
high NO production rate of the jet interaction zone can be reduced by
operating this zone fuel-rich through limits on fuel/air mixing. Under
these conditions the volatilized fuel nitrogen will be in a more
oxygen-deficient environment and the fuel nitrogen NO formation reaction
will shift to a balance between NO and molecular nitrogen formation
(Reference 3). Also, atmospheric nitrogen NO formation will be reduced
under 02 deficient conditions (Reference 3).
145
-------
Downstream of the near burner zone, beyond roughly half the firebox
length, the vortex. inreracCs vrch thn burner jets and causes the fuel and
air to mil- rapidly. Combustion and NO production in this zone and beyond
are dominated by char burning and die net NO production is small relative to
the n
-------
LOW-NO SYSTEM
x
Based on the observations detailed in the last section, the
requirements for a low-NO^ tangential system were identified. These are:
(1) initiate burning sooner to minimize 0^ availability at the ignition
point, (2) operate the jet interaction zone fuel-rich, (3) protect fuel jet
from dispersion by vortex flow, (4) lower firebox mixing with the constraint
of positive ignition, and (5) operate a portion of the char burnout zone
oxygen-deficient to get NO decay. In addition to these low-NO^
requirements, constraints must be applied on the system relative to boiler
size and efficiency, wall corrosion and slagging, and heat transfer. These
constraints dictate that, to minimize corrosion and slagging problems,
oxygen-deficient combustion gases should not contact the walls. Also,
sufficient time and oxygen must be available to fully burn out the fuel and
minimize carbon loss, CO, and UHC emissions. Finally, furnace volume and
exit gas temperatures must be constrained to those typical of presently
operating units.
Figure 9 presents a top and side view schematic of the low-NO^
system with rich and lean zones identified. The major system features are:
(1) fuel directed at conventional tangential 6° yaw angle into the center
of the furance, (2) some secondary air, either displaced toward the wall
side of the firebox or surrounding the fuel jet, directed parallel to the
fuel jet, (3) the bulk of the secondary air directed along the wall at and
above the fuel jet elevation.
These major system features create oxygen-deficient conditions in the
active near burner and the char burnout zone and fuel-lean conditions on the
furnace walls at and above the fuel jet elevation. These system characteristics
address both the low NO requirements and operational constraints noted above.
147
-------
The initial low-NO system tests used a configuration that had
x
variable wall air injection angles with respect to both the furnace wall and
the horizontal plane. This configuration, denoted System 1, is shown
schematically in Figure 10. The later testing used a configuration where-
all of the wall air jets were confined to the burner blocks set in the
corners of the firebox. This configuration, denoted System 2, is also shown
schematically in Figure 10. Though less flexible than the initial
configuration, this later configuration is a closer representation of how
the system might be retrofitted into a full-scale boiler. The primary
emphasis of System 2 testing is to define the benefits of distributing the
secondary combustion air over a wider elevation in the firebox than the
initial testing. It should be noted that the true optimal low-NOx
configuration might occur when secondary combustion air is more uniformly
distributed over the furnace wall than by the few discrete jets tested.
However, the primary emphasis of this program is on retrofittable concepts
that can achieve significant NO reductions without requiring major
X
changes to tangentially-fired boiler hardware. This constrains the wall air
introduction jets to the corners of the firebox.
System 1 Test Results
For System 1 testing, the corner burners in the pilot-scale facility
were modified as shown in Figure 10 to simulate the low-NO^ system
illustrated in Figure 9. Tests were then initiated to optimize the primary
fuel, secondary and wall jet configuration, placement, direction and
velocity. As shown in Figure 11, tests where the primary fuel and secondary
jet configurations were varied showed that directing approximately 20 percent
of the secondary combustion air into the center of the furnace and 80 percent
along the walls at the fuel jet elevation gave the lowest NO for most of the
System 1 primary configurations tested. For the fixed air jet area cases
tested, increasing the percent air on the wall increases the velocity of the
wall jet as well as distributes more of the air on the wall. The minimum NO
levels may represent a balance between the benefits of the outward
displacement of the combustion air, causing a wider separation of fuel and
air, and the negative effects of increased fuel/air mixing caused by the
higher velocity wall jets. Higher percentages of wall air than the optimal
will give high wall air velocity and excessive mixing, and lower percentages
148
-------
of air flow will place too much air in the center of the furnace, yielding
nonoptimal results.
At 80 percent air on the wall, probing results showed that the center
of the furnace is oxygen-deficient with this zone typically occupying
40 percent of the firebox at 20 cm above the burner level for most
configurations tested. As the wall air mixes into this oxygen deficient
zone it typically shrinks to 20 percent at 46 cm above the burner centerline.
At this location, NO formation/reduction processes are essentially complete
and NO levels are comparable to stack values.
Vicinity of wall oxygen concentrations are 10 percent at 20 cm above
the burner elevation for the low-NO system versus 4 percent typical of
X
conventional tangential firing in the pilot-scale facility.
Figure 3 shows the effect of excess air at the optimal 80 percent air
on the wall for several primary configurations. As shown in Figure 3, the
best configuration was a coannular primary/secondary air configuration. In
this configuration the circular primary is surrounded by an annular passage
that contains 20 percent of the secondary combustion air. Flame observation
showed that this configuration had the smallest amount of fuel dispersion
prior to entering the fuel-rich core zone. Various length slot primaries,
though initially burning much sooner than the other configurations, had fuel
dispersion problems. The dispersed fuel would burn in fuel-lean zones and
yield high NO. Circular primaries having swirling or diverging flow also
yielded higher NO at 15 percent excess air.
Figure 12 shows the effect of varying the wall air jet inclination
angle on NO. The primary and secondary air jets for this case were similar
to the configuration 2 jet locations shown schematically in Figure 7. At
0°, the wall air jet is behind the fuel jet and is shielded from the direct
impact of the vortex flow. As the wall jet inclination angle is varied from
the horizontal, the wall jet is directed out from behind the fuel jet which,
due to the lack of fuel jet shielding, might be more easily entrained and
mixed into the fuel-rich core. The decrease in fuel/air separation caused
by this mixing might be the reason for the observed NO increase.
149
-------
Besides yielding minimum NO, directing the wall air jet horizontally
is desirable for raultiburner level firing. For a system with several burner
levels, directing the wall air upwards or downwards with respect to the fuel
jet might result in undesirable burner to burner air jet interactions.
In addition to varying the inclination angle of the wall jets, one
test considered the effect of directing the wall air at 10° away from the
furnace wall. As shown in Figure 11, directing the wall air 10° off of
the wall (configuration 0) increased the minimum NO approximately 25 ppm
over the 0° (configuration N) wall jet angle results. In addition,
firebox probing tests showed that 0^ concentrations near the wall 20 cm
and 46 cm above the fuel tube elevation were decreased for the 10° wall
angle case. The higher wall O2 concentrations and the lower minimum NO
levels for the case where the air is directed along the wall makes this jet
orientation the optimum.
Comparison of the best low-NOx concepts results to conventional
pilot- and full-scale tangential firing results in Figure 3 shows that the
System 1 concept reduces NO emissions by roughly 60 percent and lowers the
sensitivity of NO to excess air. This reduced sensitivity may be a result
of the more diffusive burning nature of the fuel-rich core.
Combustion characteristics for the low-NOx System 1 are not markedly
different from conventional pilot scale tangential firing. Carbon monoxide,
UHC, and percent carbon in flyash levels for this sytem are <36 ppm, <9 ppm,
and <3 percent, respectively, versus conventional tangential firing results
of <22 ppm, <1 ppm, and <7 percent. These NO reductions and good combustion
efficiency are achieved while increasing vicinity of wall oxygen
concentrations to 10 percent near the burner elevation. This oxygen
blanketing of the wall is beneficial from a wall corrosion and slagging
point orf view.
An additional feature of the low-NO^ System 1 configuration is the
improvement in NO emissions as temperature rises. Figure 13 shows that as
gas temperature is increased for two different air-on-wall system
configurations, NO decreases. As discussed earlier, under the fuel-rich
conditions existing in the center of the furnace, increases in temperature
150
-------
will lead to more volatilization of the fuel nitrogen in the rich zone and
more conversion of this nitrogen to molecular nitrogen rather than to NO.
This attractive emission behavior with temperature might be beneficially
used to reduce boiler size and capital cost for a given heat release.
Also shown in Figure 13 is the NO reduction caused by decreasing load
at a fixed gas temperature. As discussed previously, reducing load decreases
firebox mixing thereby maintaining rich zones in which NO is minimized.
System 2 Test Results
The burner configuration used in System 2 testing is shown schematically
in Figure 10. This system differs from the initial configuration in that four
wall jets, instead of two, are used and these jets are confined to the
corner burner blocks. This configuration more closely simulates the
retrofitting of the low-NOx concept into a full-scale tangential boiler.
In addition, operating the four levels of wall air in this configuration
defines the emissions and efficiency benefits of vertically distributing the
combustion air.
Figure 14 presents the variation of NO levels with percent of
secondary combustion air on the wall for the System 2 configuration. Each
curve represents a different vertical distribution of air flow between wall
air ports la, lb, 2, and 3 as defined in Figure 10. The SR's achieved at
each wall air level as a result of the vertical distribution of air is given
in Figure 14. Configurations lb and lab denote tests where air was flowing
only in port lb or air flow was equally split between ports la and lb,
respectively. Also included in Figure 14 are the optimal results from
System 1 testing.
The cases where SRj, SR2> and SR^ are 1.15 do not have wall air
flowing in ports 2 and 3. Therefore, these cases are comparable to System 1
configurations where wall air ports 2 and 3 are absent. As shown in
Figure 14, System 1 gives the lowest NO results and System 2, with only the
IB jet operational, gives the highest levels. The percent air on the
wall at minimum NO falls between 80 and 85 percent for these cases.
Differences in wall jet configuration and location between these
cases probably account for the differences in NO. As shown in Figure 10,
151
-------
the wall jet for System 1 exits downstream of the corner, nearly half way
down the furnace wall, whereas the wall jet exit for System 2 is in the
corner. In addition, the wall jets in System 1 protrude into the firebox
whereas the System 2 jets are confined to the corner burner blocks. These
configurational factors can impact aerodynamics and mixing between the wall
and fuel jets leading to the observed differences in NO. For example, the
protruding wall jets may create separated flow zones in the firebox corners
which help set up pressure fields to deflect the wall air away from the root
of the fuel jets. This may reduce fuel/air mixing and benefit NO
emissions. In addition, when only wall jet lb is operating, the wall jet
flow area is one-half that when la and lb are working or when System 1 is
tested. Therefore, the lb test conditions represent very high wall air
velocities as compared to the other cases. As discussed previously, higher
velocities can lead to enhanced fuel/air mixing and thereby higher NO.
Even though System 2 NO results are marginally higher under these
conditions, this configuration is preferred, since a retrofittable full-scale
system simulation is more accurate with the System 2 configuration.
For both configurations la and lab the Figure 14 results show that,
when wall air is distributed vertically to ports 2 and 3, NO levels decrease
with the minimum NO point shifting to higher levels of percent air on the
wall. In these cases, the vertical separation of the fuel and air is
creating a larger and more fuel-rich zone at the fuel entry elevation where
NO production is minimized.
Figure 15 shows the variation of configurations la and lab NO for a
fixed percent air on the wall for a wide range of vertical distributions of
wall air. These conditions were taken at a different temperature than the
results in Figure 14. As can be seen in Figure 15, NO decreases for SR^
less than 0.88, reaches a local minimum and then increases before finally
achieving a second low level for the case where all of the wall air is
evenly distributed between ports 2 and 3. As wall ports la, lb, 2, and 3
air flow is varied, both the distribution and velocity of wall air is
changing. For the lab case at the two extreme conditions, with no flow in
wall air ports 2 and 3 and all of the air flow in ports 2 and 3, all of the
jet velocities are equal and the observed NO reduction is due entirely to
152
-------
the vertical displacement of wall air. Therefore, the results presented in
Figure 15 under these extreme conditions show a distinct advantage in
vertically separating the fuel and wall air. At intermediate air splits,
the jet velocities are lower, which might be part of the reason for the
local NO minimum. This point will be explored in further testing.
For application to real systems, conditions at SRj=0.8 are more
attractive than at SR^=0.2. This is because at SR^=0.8 significant
amounts of wall air are present along the walls at and above the fuel tube
elevation. Probing tests at the wall showed 15, 12, and 12 percent oxygen
at, 20 cm, and 46 cm above the fuel tube elevation, respectively . Lower
levels at the fuel tube elevation are expected for the SR^=0.2 conditions.
As indicated previously, maintaining a high oxygen concentration on the wall
is beneficial from a wall corrosion and slagging point of view.
The minimum NO achieved with System 2 is lower than System 1 levels
and represents a 65 percent reduction from baseline tangential system
levels. Combustion efficiency during these tests was excellent with CO,
UHC, and percent carbon in flyash emissions being less than 40 ppm, 3 ppm,
and 2.4 percent, respectively. These levels are comparable to System 1 and
conventional tangential system results.
153
-------
CONCLUSIONS
Based on coal-fired and pilot-scale tangential system burner and
vortex characterization tests, a low-NO^ system was defined. The major
system feature is the dividing of the secondary combustion air between
injection into the center of the furnace and injection along the furnace
walls. The delayed mixing of the wall air into the vortex causes a rich
combustion zone to develop in the furnace, minimizing fuel and atmospheric
NO^ formation. Primary, secondary, and wall jet configurations and
flowrates strongly influence the effectiveness of the system to lower NO^
emissions. Testing showed that the best results are achieved with (1) wall
air directed horizontally and along the wall (2) wall air flow equal to or
greater than 80 percent of secondary combustion air (3) primary/secondary
air coannular configuration (4) wall air vertically distributed at and above
fuel entry location. At optimal parameter settings, reductions of 65 percent
in NO from conventional tangential system levels can be achieved at
comparable combustion efficiency. In addition, wall oxygen concentration at
the burner level is significantly increased over conventional tangential
firing. This increase is beneficial from wall corrosion and slagging points
of view. Also, the system shows a decrease in NO^ emissions as
temperature is increased. This characteristic could be beneficially applied
to reduce furnace volume for a given heat release.
154
-------
REFERENCES
1. Lim, K. J., L. R. Waterland, C. Castaldini, Z. Chiba, and G. B.
Higginbotham. Environmental Assessment of Utility Boiler Combustion
Modification N0X Control. Acurex Draft Report TR78-105, April 1978.
2. Habelt, W. W. The Influence of the Coal Oxygen to Nitrogen Ratio on
N0X Formation. Presented at 70th Annual AIChE Meeting,
November 13-17, 1977, New York, New York.
3. Macek, A. Seventeenth Symposium (International) on Combustion, The
Combustion Institute, 1978, p. 65,
4. Wendt, J. 0. L., C. V. Sternling, and M. A. Matovich. Fourteenth
Symposium (International) on Combustion, The Combustion Institute,
1973, p. 897.
5. Gibbs, B. M., F. J. Pereira, and J. M. Beer. Sixteenth Symposium
(International) on Combustion, The Combustion Institute, 1976, p. 461.
6. Brown, R. A., J. T. Kelly, and P. Neubauer. Pilot Scale Evaluation
of N0X Combustion Control for Pulverized Coal: Phase II Final
Report. EPA-600/7-79-132, Environmental Protection Agency, June 1978.
7. Breen, B. P. Sixteenth Symposium (International) on Combustion, The
Combustion Institute, 1976, p. 19.
8. Brown, R. A. and C. F. Busch. Pilot Scale Evaluation of Waste and
Alternate Fuels: Phase III Final Report. EPA-600/7-80-043,
Environmental Protection Agency, March 1980.
9. Selker, A. P.: Program for Reduction of N0X from Tangential
Coal-Fired Boilers, Phase II and Ila. EPA-650/2-73-005a and b,
Environmental Protection Agency, June 1975.
10. Crawford, A. R., et al. The Effect of Combustion Modification on
Pollutants and Equipment Performance of Power Generation Equipment.
EPA-600/2-76-152c, Environmental Protection Agency, September 1975.
11. Pershing, D. W. and J. 0. L. Wendt. Sixteenth Symposium
(International) on Combustion, The Combustion Institute, 1976, p. 389.
155
-------
VORTEX CORE ZONE
NEAR BURNER ZONE
JET INTERACTION ZONE
SIDE VIEW
TOP VIEW
FUEL
AIR
ANNULAR
AIR
SECONDARY
AIR .
FUEL
BURNER
Figure 1. Tangential boiler schematic.
156
-------
1,
2
3,
4,
5
6
7
8
9
10
Combustion chamber
Ignition and flame safeguard
Observation ports
Ashpit
440 kW swirl burner for front
wall firing
Corner burners for tangential
firing
2033K refractory
Tower sections
Heat exchanger assemblies
Staged air Injection ports
CO rp
Figure 2. Pilot-scale test facility.
157
-------
600,
500
400
300
200
O
~
0
A
CONVENTIONAL TANGENTIAL
PILOT-SCALE
UTAH
WESTERN KENTUCKY
MONTANA
PITTSBURG -8
FULL-SCALE
293 kW LOAD, 589K AIR PREHEAT,
15';.: PRIMARY AIR, 6° YAW ANGLE
9
REF. 10
REF. 11
AIR-ON-WALL
PRIMARY
SWIRL
*>-
rvi
SHORT SLOT
CIR. DIVERGENT
CIR (CIR. SEC.)
CIR
LONG SLOT
COAN.
100
I 1 1 I 1 [ I
12 16 20
Excess Air (percent)
24
28
Figure 3. Effect of excess air on pilot-, full-scale and air-on-wall
system NO emissions.
158
-------
NO (ppm)
600
CORNER
BURNER
FLAME
BOUNDARY
*777777777*
20
C0Z (PERCENT)
UTAH COAL, 293 kW LflAO,
589K AIR PREHEAT,
15% PRIMARY AIR.
15% EXCESS AIR
6 YAU,
£
N
<
FURNACE
WALL
Figure 4. Concentration distributions at the fuel tube elevation
superimposed on a plan view of the firebox.
-------
500i—
400
300
o
T3
CL)
S_
3
l/)
IT)
-------
Utah coal
12° yaw
293 kU load
15% excess air
1.15
.85
h (cm)
Plan View
© --
10
0
of Furnace
15% primary air
1.151.65 h (cmT
TOO
® 56
45
27
15
+21.0
-21.0
1.15
© 87
h (cm)
~21.6
Burner jet
Approximate gas flame/
vortex visible boundaries
Percent NO injectant remaining in the
stack for Injection at the indicated
locations below, at and above the fuel
entry elevation (h). Both unstaged
(SR=1.15) and staged (SR*.85) conditions
were doped.
Figure 6.
NO dopant remaining following injection at
indicated locations.
-------
BURNER CONFIGURATIONS4
1 COAL/3 GAS BURNERS
293 kW LOAD
589K AIR PREHEAT
15» EXCESS AIR
15% PRIMARY AIR
6° YAW
400
1
2
3
4
5
V
A
~
o
o
•
OOOOO
~
o o
oo# oo
ooo
•
ooo
o°o
o • o
OqO
ri
w
~Fuel is designated by solid circles and air by open
circles. Vortex flow on burner is from right to left.
300
200
O
~
A ^
100
GAS ONLY
1200
1250 1300
FURNACE EXIT TEMPERATURE (K)
1350
1400
Figure 7. Burner configuration NO levels.
162
-------
500
400
csj 300
o
o
&
a.
§ 20(
100
1 COAL/3 GAS BURNERS
73 kW COAL BURNER LOAD
589 K AIR PREHEAT
15% EXCESS AIR
15% PRIMARY AIR
6° YAW
LIFTED FLAME BURNER CONFIGURATIONS Q
— S
~
A
O
c*
I
C
^ B-—
-o
100 200 300
GAS FIRING RATE (kU)
Figure 8. Burner configuration NO level variations with firing rate.
163
-------
FUEL LEAN
TOP VIEW
FUEL RICH
FUEL LEAN
FUEL RICH
WALL AIR
SECONDARY AIR
6" YAW
FUEL/PRIMARY AIR
SIDE VIEW
WALL AIR
FUEL/PRIMARY AIR
SECONDARY AIR
CORNER BURNER DETAIL
Figure 9. Low-NOx air-on-wall system schematic.
-------
Top view
Top view
Wall air
Wall air
Annular
Primary
Front view
System 1
K
Annular
Primary
Front view
System 2
Figure 10. Low-NOx system burner types.
-------
293 kW load
589K air preheat
15% excess air
15% primary air
400 _
CM
O
o
300
200
Configurations*
Primary Sec.
~ - H
cir.
cir.
O- I
long slot
slot
A- j
short slot
slot
O - K
cir.
Slot
k- L
cir. swirl
slot
O - M
cir. diff.
slot
O- N
cir. coan.
slot
0 - o*
cir. coan.
slot
*A11 with slot wal 1 air
**With wall air directed 10c
of furnace wall
off
50 60 70 80 90 100
Secondary airflow along wall (percent)
Figure 11. Comparison of NO for a variety of primary configurations
with percent secondary air-on-wall,
166
-------
400
293 kW load
15% excess air
15% nrimary air
6° yaw
80 percent of secondary air along wall
1"
r-
N
\
O Test 291
• Test 292
) Di
I te
Different gas
temperatures
CM
o
>5
o
a*
"vi
300
200
1
1
1
1
-20 -10 0 +10 +20
Injection angle (degrees)
Figure 12. Effect of wall jet inclination angle.
-------
300
^ 250
a-c
O
a.
o
(DO °0
200
1100
589K AIR PREHEAT
15% PRIMARY AIR
15" EXCESS AIR
6° YAW
PRIMARY*
LOAC
kW
CONFIGURATION
293
220
LONG SLOT y |
o
•
CIR. COAN.(0)
~
J)
£
-------
350
\
B
ft
\
300
g.
o.
CO
o
g 250
o
200
150
60
\
/
/
/
/
¦^er
$
I
\
\
\
\
\1
\
/
293 kW load
15» excess air
15% primary air
54>9K air oreheat
t
\
ft
V
Q/ /
/
'/
/'
System 1
System 2
SR.,
sr2
sr3
IB
1AB
o
O
~
1.15
1.15
1.15
A
1.02
1.08
1.15
•
~
0.88
1.02
1.15
_l
X
_L
70 80 90
Secondary air on wall (percent)
ion
Figure 14. NO variation with percent air-on-wall for
vertical wall air configurations.
169
-------
350
293 kU load
15% excess air
15% primary air
589K air preheat
300
O IB 85% air-on-wall
~ 1AB 100% air-on-wall
CSJ
o
0
1
200
0.88
0.75
1.15
1.02
0.62
0.49
0.35
0.22
Stoichiometric ratio at fuel entry level, SR^
Figure 15. NO variation with wall air flow above fuel entry elevation.
170
-------
TABLE I. ANALYTICAL POLLUTANT MEASUREMENT EQUIPMENT
N0/N0x
Air Modeling Model 32C chemiluminescent analyzer
°2
Intertech Model Magnos 5T paramagnetic 02 analyzer
CO
Anarad Model AR500R NDIR analyzer
co2/co
Anarad Model AR600R NDIR analyzer
UHC
Intertech Model FID0008 FID H/C analyzer
so2
TEC0 Series 40 Pulsed Fluorescent analyzer
4.
TABLE II. UTAH COAL ANALYSIS
Ultimate Analysis
Carbon
Hydrogen
Nitrogen
Sulfur
Oxygen (diff)
Ash
Dry basis
(% wt)
69,91
5.15
1.5
0.98
12.46
10.0
As received
Proximate Analysis (% wt)
Moisture 5.6
Ash 9.4
Volatile 40.6
Fixed carbon 44.4
Heating value
(cal/gm) (dry basis) 6,966
Sieve analysis 77.3% wt percent through 200 mesh
171
-------
THE DEVELOPMENT OF A LOW N0X DISTRIBUTED MIXING BURNER
FOR PULVERIZED COAL BOILERS
By:
B. A. Folsom and L. P. Nelson
Energy and Environmental Research Corporation
8001 Irvine Boulevard
Irvine, California 92705
and
J. Vatsky
Foster Wheeler Energy Corporation
110 S. Orange Avenue
Livingston, New Jersey 07039
172
-------
ABSTRACT
This paper describes the development of a low N0X pulverized coal
burner for demonstration in two small pulverized coal-fired boilers by
1982. The "Distributed Mixing Burner" concept provides for controlled
mixing of the coal with the combustion air to minimize N0X emissions while
maintaining an overall oxidizing environment in the furnace to minimize
slagging and corrosion. The design of a prototype field operable burner
is discussed, and test data in a research facility suggesting that N0X
emissions less than 84 ng/J (0.2 lb/10® Btu) might be attainable in the
field are presented.
173
-------
ACKNOWLEDGMENTS
This paper is based upon work conducted under Contract Nos* 68-02-
1488, 68-02-2667 and 68-02-3127 with the U. S. Environmental Protection
Agency. The authors wish to express their appreciation to Mr. G. B.
Martin of the U. S. Environmental Protection Agency for his considerable
help and advice.
174
-------
SECTION 1
INTRODUCTION
The projected Increase in pulverized coal use for industrial and utility
steam generation will result in significant increases in pollutant emissions.
One recent report predicts that N0X emissions will increase 80 percent by the
year 2000, even if current control technology is fully employed (1). Main-
taining or improving the current air quality will clearly require the devel-
opment and application of advanced N0X control techniques. Combustion modi-
fication through the installation of low N0X burners is one of the most
promising techniques for reducing emissions without greatly increasing over-
all costs or reducing efficiency.
For the last several years Energy and Environmental Research Corporation
has been working with the EPA to develop a low N0X pulverized coal burner.
This "Distributed Mixing Burner", DMB, provides for controlled mixing of the
coal with combustion air to minimize N0X emissions while maintaining an over-
all oxidizing environment in the furnace so as to minimize the detrimental
effects of furnace corrosion. DMBs have been tested at firing rates up to
29 thermal MW (100 x 10** Btu/hr) in single and four-burner arrays in two
research furnaces. The tests covered wide ranges of burner adjustments and
operating conditions. Under optimum experimental conditions, N0X levels
less than 65 ng/J (0.15 lb/10^ Btu) were obtained (2). The minimum N0X
levels obtainable in field operating boilers will, of course, vary with
boiler design and operating conditions and may be higher than these experi-
mental results.
Application of the DMB concept to new pulverized coal-fired boilers
requires an evaluation of long-term performance in field boilers. This
175
-------
evaluation is being conducted on small pulverized coal-fired boilers (less
than 455 x 10^ kg: 1.0 X 106 lb of steam per hour) under EPA Contract 68-02-3127
and on larger utility boilers under EPA Contract 68-02-3130 (3). The small
boiler demonstration, which is the subject of this paper, has been scheduled
first so that the larger utility boiler demonstration can benefit from the
test results. However, the actual burner components being utilized in each
program are representative of those which are used commercially by the
respective boiler manufacturers. Consequently, the utility program (68-02-
3130) will benefit from the small boiler program (68-02-3127) only in that
the concept will be first demonstrated in the field in the small boiler pro-
gram. The burner components provided to the small boiler program will not
be utilized in the utility program.
The small boiler demonstration involves two Foster Wheeler front wall-
fired units which are representative of industrial size boilers. The first
unit has a capacity of 98 x 10^ kg/hr (215 x 10^ lb/hr) and has four burners
in a 2 x 2 array. The second unit being considered has a capacity of 175 x
10-3 kg/hr (385 x 10-* lb/hr) and has six burners in a two-high by three-wide
array. The DMBs to be installed in these units will be designed by incor-
porating DMB flow parameters into a commercial burner design and adding the
requisite number of tertiary air ports. Prototype burners will be fabricated
by Foster Wheeler using commercial burner components and construction methods.
The design parameters will then be optimized by testing in the research fur-
nace. The final configurations will be installed in the field boilers and
tested for 18 months duration. During this long-term evaluation all aspects
of the burner/boiler performance will be assessed, including: NOx emissions,
other environmental impact, burner and boiler efficiency, durability, opera-
fa ility, safety, boiler performance and integrity and heat transfer surface
duties.
This paper discusses the status of the DMB under development for the
initial demonstration. The prototype burner has been designed and fabricated.
The test data leading to the DMB design criteria and the prototype burner
design are discussed in the following section. The prototype burner and a
burner similar to the burners currently installed in the initial demonstra-
tion boiler have been tested in the research furnace. The results of these
176
-------
tests and an estimate of potential DMB performance in the initial
demonstration boiler are discussed in Section 3-
177
-------
SECTION 2
PROTOTYPE DMB DESIGN
The prototype DMB for the initial field demonstration has been designed
to meet the following requirements:
• DMB design criteria
• Compatibility with host burner/boiler specifications
• Commercial burner construction methods and operational
characteristics
DMB DESIGN CRITERIA
The DMB concept involves staging the combustion process to minimize NOx
emissions while maintaining an overall oxidizing atmosphere in the furnace
so as to have minimal impact on furnace slagging and corrosion. N0X pro-
duction from fuel nitrogen compounds is minimized by driving a majority of
the compounds into the gas phase under fuel-rich conditions and providing a
"stoichioraetry/temperature history which maximizes the decay of the evolved
nitrogen compounds to N£. Thermal N0X production is also minimized by
enthalpy loss from the fuel-rich zone which reduces peak temperatures.
A schematic representation illustrating how the DMB design sequentially
stages the fuel/air mixing is shown in Figure 1. The combustion process
occurs in three zones. In the first zone pulverized coal transported by the
primary air combines with the inner secondary air to form a very fuel-rich
(30 to 50 percent theoretical air) recirculation zone which provides flame
stability. The coal devolatilizes and fuel nitrogen compounds are released
to the gas phase. Outer secondary air is added in the second "burner zone"
where the stoichiometry increases up to about 70 percent theoretical air.
178
-------
This is the optimum range for reduction of bound nitrogen compounds to
N2 (2). Air to complete the combustion process is supplied through ter-
tiary ports located outside the burner throat. This allows substantial
residence time in the burner zone for decay of bound nitrogen compounds to
N2 ami radiative heat transfer to reduce peak temperatures. The tertiary
ports surrounding the burner throat provide an overall oxidizing atmos-
phere in the burner zone.
Figure 2 shows the DMB as it would be installed in the small wall-fired
boiler. This DMB design uses components illustrated below:
• Four independently controlled airstream systems divide the
total combustion air into:
1. An annular coal nozzle;
2. A central passageway containing an ignitor for start-up;
3. Two concentric secondary airstreams to provide inner and
outer secondary airflow; and
4. Four outboard tertiary air ports.
• The annular fuel nozzle and tangential coal inlet produce a
uniform annular coal distribution.
• Two adjustable air registers connected to the secondary air
passages to control both the airflow rates and degree of swirl.
• Ignition and main flame scanner systems for safety.
Six DMB configurations using other components have been evaluated in two
research furnaces to determine the effects of design parameters on flame
shape and stability, as well as efficiency (as indicated by CO emissions)
and N0X emissions. These burners ranged in capacity from 3.6 thermal MW
(12.5 x 106 Btu/hr) to 29 thermal MW (100 x 10& Btu/hr), and were tested
as single burners. In addition a 2X2 array of 3.6 thermal MW (12.5 X 106
Btu/hr) burners was also tested. Each configuration was evaluated over a
range of operating conditions and the ersults will be discussed in terms of
the following burner systems:
• Fuel System
• Secondary Air System
• Tertiary Air System 17q
-------
Fuel System
The DMB fuel system includes the following geometric variables and
fuel/air characteristics through the fuel nozzle:
• Primary stoichiometry
• Nozzle size and axial position
• Exit velocity
• Swirl
The effects of each of these variables (except nozzle size) on burner per-
formance have been evaluated. Although nozzle size has not been specifi-
cally evaluated, it is considered to have only secondary effects on burner
performance over the practical range.
The DMBs tested in the research furnaces utilized a circular coal nozzle
located on the burner axis. In two of the burners there was no provision for
primary swirl, and as a result, the rate of mixing between the primary and
secondary streams was low. This produced a longer diffusion-type flame and
low N0X emissions. Long flames can be utilized in furnaces with sufficient
depth and when low burner liberations are used. However, the goal of this
program is to develop a burner which has a short flame, requires no reduction
in burner capacity, and is more compatible with a wide range of furnace con-
figurations including short furnace depths of 6 m (20 ft) or less. The other
DMB configurations tested in the research furnaces included impellers in the
nozzle to impart swirl and increase the rate of mixing between the primary
and secondary airstreams. The resulting flames were considerably shorter,
but had slightly higher NOx emissions. The following discussion will include
the test results from the short flame burners. Test results from the long
flame burners are discussed in References 2, 3 and 4.
Figure 3 shows some typical results obtained with a 14.6 thermal MW
(50 x 106 Btu/hr) single secondary DMB firing at full load and 125 percent
theoretical air. Minimum N0X levels are in the range of 100 ppm* which
A
All NO and CO concentrations in this paper are corrected to 0 percent 02»
dry conditions.
180
-------
corresponds to 46 ng/J (0.11 lb/10^ Btu). CO emissions are in the range of
250-500 ppm which is considerably higher than typical burner performance in
field boilers (-50 ppm). This difference is believed to be a consequence of
the research furnace design, and will be discussed in a subsequent section.
With the swirl vanes removed (axial flow) it was not possible to obtain
acceptable flame stability. At 30° swirl angle an acceptable flame was
achieved with minimum emissions, and no measurable increase in N0X was
observed when the swirl angle was increased to 45°. Further increasing the
swirl angle to 60° resulted in a large increase in N0X. Due to the limited
range and scatter in the CO data, the effect of swirl angle on CO is unclear.
The indicated lines represent a best judgment as to interpretation of the
results. This interpretation would indicate that both the fuel/air mixing
characteristics and ignition flame characteristics are changing significantly.
Therefore, there should be a design trade-off between NO and CO with primary
swirl, with lower swirl angles providing lower N0X levels. However, if in
the final configuration the CO levels are extremely sensitive to burner
parameters, then this will limit its N0X capabilities. Ignition stability
represents the lower limit of operation, although it would not be prudent to
design at that low a level. Since ignition stability plays a key role in
the ability to minimize both NO and CO, the specific design parameters will
depend on the site-dependent conditions. It is not yet known if the burners'
N0X capabilities will vary with these site-dependent conditions.
Secondary Air System
The DMB secondary air system includes all components which provide air
through the burner throat except for the coal nozzle. Specific variables
include:
• Air distribution between secondaries
• Velocity
• Swirl
• Burner zone stoichiometry
The DMBs tested in the research furnaces included designs with single, as well
as dual secondary air passages. The dual secondary design is preferred
because of the additional control over fuel/air mixing rates which can be
181
-------
achieved by varying the flow rates and swirl in the two secondary passages.
However, for simplicity, most of the DMBs were designed with a single second-
ary air passage.
Figure 4 shows the effects of secondary air system variables on the
performance of the single secondary 14.6 thermal MW (50 x 10^ 8tu/hr) burner.
The operating conditions were the same as for the data presented in Figure 3,
and the primary swirl was 45°. Variations in burner zone stoichiometry are
produced by varying the amount of air through the outboard tertiary air pas-
sages. For burner zone stoichiometrics equal to the overall stoichiometry
(125 percent theoretical air) all of the air passes through the burner throat
(unstaged condition). Figure 4 shows that for unstaged conditions the DMB
produces NO and CO levels of 400 and 300 ppm, respectively. As the burner
zone stoichiometry is decreased (air diverted through the tertiary ports)
keeping the overall stoichiometry constant at 125 percent theoretical air,
NO decreases. At a burner zone stoichiometry of 60 percent theoretical air
the range of N0X is 90-180 ppm. CO remains essentially constant down to about
70 percent theoretical air where the airflow through the secondary air pas-
sages has decreased to the point where the flame stability decreases and CO
emissions rise. While the flame is still stable initially, continued reduc-
tion in burner zone stoichiometry eventually leads to instability and flame-
out. Due to the onset of instability, operation near the knee of the CO curve
is unacceptable. Consequently, actual NO levels in the field may be higher
than those shown in Figure 4. For this burner, operation at 70 percent throat
stoichiometry appears to be acceptable. Under this minimum condition NOx is
about 0.2 lb/10^ Btu.
In comparing the CO and NO characteristics, note that the impact of
burner zone stoichiometry on NO remained unchanged even as the CO markedly
increased. This suggests that the NO formation is being controlled princi-
pally by the burner zone stoichiometry and not by the primary/secondary mixing
rates. The conditions producing the increase in CO shown in Figure 4 were
found to be strongly dependent on the burner settings, as well as design
velocity. In fact, at full load it was possible to obtain minimum CO and NO
over a wide range in excess air levels by simply varying both burner zone
stoichiometry and swirl conditions. (This is not meant to imply that it is
the intent of this program to vary the burner setting in the field during the
182
-------
load or excess air changes). Although not shown in Figure 4, at zero swirl
the flame could not be operated stably, even at high burner zone stoichi-
ometry, and resulted in flameout. As secondary swirl was increased to 30°
it was possible to achieve a stable flame over a burner zone stoichiometry
range from about 70 percent theoretical air to essentially unstaged condi-
tions (120 percent of theoretical air). Lower burner zone stoichiometries
were possible as the secondary swirl angle was increased. Note that the NO
characteristics were essentially unchanged with secondary swirl angle, while
the burner performance (CO) was improved by increasing secondary swirl.
Thus, there is a trade-off in NO and CO emissions as burner zone stoichi-
ometry and swirl are varied. Optimum values with this burner for invarient
operating, conditions are: (1) secondary swirl angles of 30° to 45°, and
(2) a burner zone stoichiometry of about 60 percent theoretical air.
Tertiary Air System
The DMB tertiary air system includes all components which provide air
through the tertiary air passages. Specific variables include:
• Velocity
• Number of ports
• Location of ports
Variations in tertiary velocity have been found to have little effect
on NO and CO emissions for tertiary port configurations of practical signi-
ficance. However, the number and location of the tertiary ports have been
found to affect CO but not NO emissions. Figure 5 shows some results from
tests of a four-burner array of 3.6 thermal MW (12.5 x 10^ Btu/hr) DMBs.
The tertiary ports for adjacent burners were combined as shown to minimize
the total number of tertiary ports. The tests were conducted by maintaining
the secondary air flow rate (burner zone stoichiometry) constant and varying
the tertiary air flow rate (overall stoichiometry) and the number of ports
in service. The decrease in NO as the overall stoichiometry is reduced is
typical of performance for most burners. CO emissions for all ports in
service are constant, independent of overall stoichiometry, over the range
tested. Although not shown in Figure 5, lower values of overall stoichi-
ometry (less than 110 percent theoretical air) caused a sharp increase in
183
-------
CO and smoke emissions. This is also typical performance for most burners.
Tests were conducted with various rows and/or columns out-of-service
to define the sensitivity of NO and CO to the number and location of the
tertiary ports. The results presented in Figure 5 show that within experi-
mental accuracy the NO characteristics were completely unaffected by the
variations in tertiary air velocity or tertiary port location. The general
conclusion from this is that for these short flames the burner zone is
essentially premixed and burns at the burner zone stoichiometry and, there-
fore, NO is formed in the burner flame zone before substantial interaction
with the tertiary air. While NO was not affected by tertiary port location
in the multiburner configuration, the variations in CO were dramatic. Note
that unsymmetric changes in tertiary ports out of service were accompanied
by an increase in CO. This result reveals that the tertiary port location
relative to the burners has a strong impact on the overall mixing in the
lower furnace, and that improper port location will result in incomplete
CO burnout in the high temperature region of the furnace.
Based upon the test results presented above and other data involving
variations in exit geometry, operating conditions and coal composition,
nominal prototype burner design criteria have been identified and are listed
in Table I. These criteria are nominal in that they are the starting point
for prototype burner testing and final optimization.
HOST BURNER/BOILER SPECIFICATIONS
The boiler selected for the initial demonstration is Pearl Station
which is owned and operated by Western Illinois Power Cooperative (WIPCO).
Table II lists the characteristics of this unit which impact prototype
burner design. The coal currently supplied to this unit is an Indiana
bituminous coal with the characteristics listed in Table III.
The Pearl Station characteristics are compatible with the DMB design
criteria listed in Table I with the following exceptions:
• Windbox Depth and Burner Exit Length. The windbox depth is
designed for the currently installed pre-NSPS Intervane burners
which have a single register and a short exit. However, it is
too short to accommodate the dual registers and exit length
184
-------
specified in the DMB design criteria. The maximum exit length
which can be used corresponds to about 40 percent of the proto-
type DMB throat diameter, whereas the design criteria specify
100 percent. Burner exit geometry affects flame stability
and the effects of this short exit length were evaluated in the
prototype tests (Section 3).
« Burner Spacing and Tertiary Port Locations. The burner spacing
is too close for the tertiary ports to be located as specified
in the design criteria. In addition, windbox and furnace struc-
tural considerations limit the potential tertiary port locations.
Figure 6 shows the tertiary port locations selected for the WIPCO
unit. Although this configuration deviates from the design
criteria, the tertiary air is well-distributed around and between
the burners and the velocity matches the design criteria.
It should be noted that installation of the tertiary air ports
results in removal of a buckstay and elimination of most of the
firing wall support tubes. Consequently, an under hopper support
must be installed. This technique is questionable for larger
utility boilers and an alternate approach would have to be developed
for this equipment.
PROTOTYPE BURNER DESIGN
The prototype DMB for the initial industrial boiler demonstration has
been designed by integrating the DMB design criteria and the host firing
system/boiler characteristics into a design using commercial burner compo-
nents. The prototype DMB has been designed to allow variation of several
burner parameters during testing. The optimum configuration will then be
used as the basis for field operable burners.
Figure 7 is a cross-sectional view of the prototype burner showing its
design details. The following Foster Wheeler commercial burner components
have been incorporated into the design:
• Annular Coal Nozzle
• Telescoping Coal Nozzle for Primary Velocity Control
• Registers for Swirl Control and Flow Shutoff
185
-------
• Air Hoods for Secondary Flow Rate Control and Measurement
• Commercial Flame Scanning and Ignition System
The coal and primary air enter the annular coal nozzle through a tangential
inlet to provide swirl. The coal nozzle is a tapered annulus. The exit
end of the coal nozzle is formed from two concentric cones. The inner cone
can be moved axially to vary the exit area, and hence, the primary velocity.
The outer portion of the coal nozzle is constructed of removable sections
so that the overall length of the coal nozzle may be changed to vary the
nozzle setback within the burner throat. There are two concentric secondary
air passages supplied from a common windbox. Each is equipped with an
adjustable register for swirl control, an axially movable sleeve for flow
control and perforated plate (air hood) for flow rate measurement- Pres-
sure taps on both sides of each perforated plate allow the pressure drop
to be measured and correlated with sleeve position and flow rate.
The portion of the burner inside the coal nozzle is termed the core,
and is supplied with a small amount of cooling air from the windbox. A
retractable ignition system, including spark ignition, a 2.9 thermal MW
(10 x 10^ Btu/hr) oil gun and an ignition flame scanner is located in the
core on the burner centerline. The main flame scanner views the flame
through the inner secondary air passage.
Tertiary (staging) air is supplied from the windbox through outboard
tertiary air ports. During testing in the research facility the airflow
rate through each port will be controlled and measured by the assembly shown
in Figure 8. The hemispherical poppet pivots to vary the open area, and
thus, control the flow rate. The pressure drop across the perforated plate
air hood provides an indication of flow rate. Alternate tertiary air port
control system designs will also be evaluated during the prototype burner
tests.
In summary, the prototype burner has been designed with "flexible"
parameters to permit the flow rates, velocities and swirl in the burner
passages to be varied to optimize burner performance. The dimensions and
ranges of adjustment are listed in Table IV. All of the dimensions listed
match the prototype design point except the exit length and tertiary port
locations as previously discussed. The effects of these deviations from the
design criteria will be evaluated during prototype tests.
186
-------
SECTION 3
COMMERCIAL BURNER AND PROTOTYPE DMB TEST RESULTS
The prototype DMB design which will be recommended for installation and
demonstration in a field-operating steam generator must be capable of (1)
stable characteristics while (2) providing low NOx levels. Since the DMB
prototype burner design was based on the results of tests in the research
furnace, some assurance is required that these test results can be used to
predict burner performance in the field boilers. Though indirect, the
approach taken was to evaluate the operating characteristics of a commercial
burner in the research furnace and compare the results with field experience.
Similarity of the research furnace and field results were then taken as an
indication that the operating characteristics of the DMB in the research
furnace and in field operating boilers would also be similar. In addition,
burner manufacturers have found that when two burners are tested in a research
facility that the ratio of N0X emissions measured in the test facility will
be comparable to the ratio of N0X emissions when the identical burners are
operated in field boilers. Thus, the ratio of N0X emissions for a commer-
cial burner and the prototype burner tested in the research furnace should
be approximately the same as the ratio of N0X emissions from the same burners
operated in the host boilers.
COMMERCIAL BURNER TESTS IN THE RESEARCH FURNACE
A Foster Wheeler Intervane burner similar to those in service at the
initial demonstration site has been tested in the research combustor. In
general, the burner performance was similar to that observed in field boilers.
The flame was short, bright, intense and stable, but somewhat narrower and
longer than those observed with multiburner arrays in the field. This suggests
187
-------
chat the DMB flame shape and stability characteristics observed in the
research furnace will be similar in the field.
The CO and NO characteristics for the Intervane burner tested in the
research furnace are shown in Figure 9. The CO characteristics are more
sensitive to excess air variations than observed in the field. For example,
in field units it is expected that the sharp rise in CO would occur at a
lower value of excess air ( 110 percent T.A.) than that shown in Figure 9.
In addition, the asymptotic value of CO measured in the research furnace was
about 185 ppm (corrected to 0 percent O2) which is greater by a factor of
four than that observed in multiburner field units, less than 50 ppm- This
suggests that the high CO levels measured in the research furnace may be a
result of the furnace design and sampling point location (at the furnace
exit). The LWS is currently being modified to correct the suspected causes
of high residual CO emissions. The Intervane burner will be retested to
determine if the resulting CO emissions match those obtainable in the field.
The variations of NO with load and overall stoichiometry are within the
normally observed field ranges. However, the NO levels are 50 to 100 ppm
lower than expected in a multiburner field operating boiler. Figure 9 also
shows typical NO levels from four-burner bituminous coal-fired boilers
similar to the initial demonstration site. Extrapolation of the research
furnace results to full load (16.4 thermal MW (90 x 10^ Btu/hr)) gives an
NO level of 500 ppm while 600 ppm are typically measured at full load in
field units. The NO emissions from the initial demonstration boiler are
expected to be in this range. This will be verified during baseline tests
of the unit.
PROTOTYPE DMB TESTS IN THE RESEARCH FURNACE
The prototype DMB has been tested in the research furnace. This initial
test series was brief and primarily evaluated performance without tertiary
ports (unstaged). The flame was bright and somewhat longer and wider than the
Intervane burner flame. There was no indication of flame impingement on the
target wall which was 4.88 m (16.0 ft) from the firing face. The DMB should,
therefore, be compatible with the initial demonstration boiler where the fur-
nace depth is 6.28 m (20.6 ft).
188
-------
The NO and CO characteristics for the prototype DMB operating unstaged
and the Invervane burner are shown in Figure 10. The NO emissions for the
prototype DMB are comparable to those measured when previous DMB designs
were tested unstaged. However, they are 50 to 100 ppm lower and more
sensitive to overall stoichiometry than the Intervane burner. The lower
NO level is expected because the prototype DMB has dual registers and lower
throat velocity than the Intervane burner, and thus, lower fuel/air mixing
rates. CO emissions at a typical operating condition of 125 percent theo-
retical air are also somewhat lower for the prototype DMB. However, this is
probably not significant as discussed above.
It should, however, be noted that this prototype DMB's geometrical con-
figuration has not been optimized. Since this design is based on Foster
Wheeler's commercial hardware, which has been adapted to meet DMB velocity
and flow criteria, further geometrical changes may be necessary to obtain
optimum NO and CO performance with acceptable stability. Among the dif-
ferences between this prototype and EER's previous experimental DMBs are:
Prototype DMB Experimental DMBs
Registers for swirl control Block-type swirl generators
Annular coal nozzle with Cylindrical pipe nozzle with
tangential inlet articulated inlet
Open coal nozzle outlet with Impeller/swirler at outlet
adjustable velocity feature fixed velocity
Medium quarl depth Long quarl depth
It is reasonable and prudent to assume that these differences can result
in NO and CO performance which are also different. The prototype results
were obtained with an initial geometry which is likely to be nonoptimum.
Consequently, as part of the test program, the prototype's geometrical
and flow parameters will be varied to determine the configuration which pro-
vides optimum NO, CO and stability performance.
A limited test series was conducted with the prototype DMB operating
staged to obtain an approximate indication of its potential for low NO emis-
sions. These tests were preliminary; no attempt was made to optimize burner
adjustments and the burner was not operated at the design point. The results
189
-------
in Figure 11 show Che same decrease In NO as burner zone stoichiometry
decreases as that observed in tests of other DMB designs. Extrapolation
to the DMB design point (overall stoichiometry 118 percent theoretical air
and burner zone stoichiometry 70 percent theoretical air) results in approxi-
mately 100 ppm NO. CO emissions increased to unacceptable levels as burner
zone stoichiometry decreased. This is probably due to nonoptimum burner
configuration and adjustment. Optimization of burner adjustments and of
burner geometry should result in the characteristic knee as in Figure 4.
ESTIMATE OF NO EMISSIONS FROM THE FIELD DEMONSTRATION
A rough estimate of the potential NO emissions from the DMB operating
in a small pulverized coal-fired boiler can be made from these data. The
prototype DMB and Intervane burners have been tested in the research furnace
at 23.4 thermal MW (80 x 10^ Btu/hr) firing rate and 125 percent theoretical
air. Field data are also available for Intervane burners operating in small
pulverized coal-fired boilers. The table below summarizes these results and
presents an estimate of the NO emissions from the DMB operating in a small
pulverized coal-fired boiler based on the ratio technique described above.
Research Furnace Field Burner
NO (ppm) Basis NO (ppm) Basis
Intervane 450 Measured 600 Typical Field
Prototype DMB 100 Extrapolation 120-150 Rough Estimate
of Test Results
to Design Point
The 120 to 150 ppm NO range corresponds to approximately 59 to 76 ng/J
(0.14 to 0.18 lb/10^ Btu) which is less than the program objective of 82 ng/J
(0.2 lb/10^ Btu). This estimate is, of course, very rough and Is based on
preliminary test results. An improved estimate will be made as soon as addi-
tional prototype burner tests are conducted in the research furnace, and the
performance of the host boiler (with Intervane burners) is measured.
As discussed earlier, the prototype DMB tests involve no optimization
of either burner geometry or adjustments and resulted in unacceptably high
CO levels at the design operating point. Therefore, the "rough estimate"
prediction given above is achievable only if modification and optimization
of the burner can reduce CO levels in the field boiler typical to those
190
-------
attainable with the steam generater's original equipment Intervane burners
(typically less than 50 ppm).
191
-------
SECTION 4
SUMMARY
Preparations are underway for the demonstration of the low N0X DMB
concept on two small pulverized coal-fired boilers. Several DMB designs
have been tested in research furnaces over a wide range of conditions and
nominal design and operating parameters have been identified. These
nominal design criteria and the characteristics of the first demonstration
boiler have been integrated into a prototype burner design which incorpo-
rates commercial burner components. The prototype burner has been briefly
tested in the research furnace and additional tests will be conducted to
fully evaluate and optimize performance prior to installation in the initial
demonstration boiler.
Analysis of test data from the prototype DMB and a commercial burner
tested in the research furnace indicates that in the field the DMB flame
characteristics will be acceptable. Field N0X emissions are estimated to be
in the range of 59 to 76 ng/J (0.14 to 0.18 lb/10^ Btu).
Further developmental tests are planned to optimize burner geometry and
to ensure that satisfactory stability and combustion efficiency exist over
an acceptable range of burner adjustments and normal load and excess air
variation. Prior to field installation satisfactory operation, as well as
compatibility with commercial ignition and flame scanning equipment, will
be demonstrated in the test furnace.
192
-------
REFERENCES
1. Salvesen, K. G. et al. Emission Characteristics of Stationary N0X
Sources; Volume I Results. EPA Report No. EPA-600/7-78-120A, NTIS
No. PB284-540/AS, June 1978.
2. Zallen, D. M. et al. The Generalization of Low Emission Coal Burner
Technilogy. In: Proceedings of the Third Stationary Source Combus-
tion Symposium; Volume II Advanced Processes and Special Topics. EPA
Report No. EPA-600/7-79-050B, February 1979. pp. 73.
3. Martin, G. B. Field Application of Low N0X Coal Burners on Industrial
and Utility Boilers. In: Proceedings of the Third Stationary Source
Combustion Symposium; Volume I Utility, Industrial and Residential
Systems. EPA Report No. EPA-600/7-79-050A, February 1979. pp. 213.
4. Gershman, R. et al. Design and Scale-Up of Low Emission Burners for
Industrial and Utility Boilers. In: Proceedings of the Second
Stationary Source Combustion Symposium; Volume V Addendum. EPA Report
No. 600/7-77-073E, July 1977.
193
-------
TERTIARY AIR
OUTEfl
SECONDARY AIR
INNER
SECONOARY AIR
X
COAL AND_3
PRIMARY AIR
VEflV FUEL men PnOQnESSlVE AIR AOOMION ZONE
zone (average (ovenAit sroicinoMcinv /ox)
SI (MCINOME in Y
riNAL Am AOUIIION ZONE FOR PUHNOU1
(OVEflALL 9lOICJIIOMtin* |M*|
Figure 1. DMB Concept.
-------
Figure 2. Distributed Mixing Burner.
-------
• 14.6 Thermal MW (50 x 10s Btu/Hr) Single Secondary DMB
• Coal ¦ Utah Bituminous
• Load - 14.1 Thermal MW (48 x 10s Btu/Hr)
vO
o*
8.
a
u
N
o
N
O
o
z
sua
400
300
zoo
100
0
Qyecall StoichiometKy - 125^ T.A.
Secondary Swirl Vanes = 45°
1000
1 1 I 1
i i i i i
" /,
Prlaary
// /
Swirl Vauaa
- vy
O 30°
/j?
~ '15°
A 60°
-
o
o
400
200
mill
m
40 SO 60 70 00 90 100 110 120 130 140 40 SO 60 70 B0
Burner Zone Stofchaaetry (X T.A.)
90 100 U0 1?0 130 140
Figure 3e Effects of Fuel System Variables on DMB Performance,,
-------
• 14.6 Thermal MU (50 x 10* Btu/Hr) Single Secondary UHB
• Coal = Utah Bituminous
• Load » 14.1 Thermal MU (48 x 10® Btu/Hr)
• Overall Stolchlometry = 125Z T.A.
• Primary Swirl Vanes - 45°
1000,
SOOi
400 —
O ooo
r?
C*
M
o
600 —
O
Secondary
Swirl Vanes
o W
200 —
-------
vO
00
4 Burner Array of 3.6 Thermal MW
(12.5 x 10s Btu/Hr) Single Secondary DMBs
Coal ¦> Utah Bituminous
Burners¦
Load B 19.6 Thermal MW
(67 * 10s Btu/Hr)
Burner Zone Stoichiometry
h00
1
1
1
1
a
p.
400
—
—
17
a
300
—
n
o
Cty
N
O
200
-T
¦
—
<2J
o
U
S5
100
—
>••
—
0
1
1
1
1
622 T.A.
I (JOG
ttOO
p.
>>
U
o
C4
o
UM)
O
<2J
O
u
200
Tertiary
Air Ports
o
o#o
i i i r~
O All Ports Open
# One Row or Column Closed
¦ One Row and Column Closed
oO"
| (Ml 120 130 140 150 160 WII0 120
0VCRAIL STOICillOHEIRV (X T.A.)
J JO
140
150
160
Figure 5. Effects of Tertiary Air System Variables on DMB Performance.
-------
4 Top Ports
39.4 cm (15.5 in) Diameter
4 Side Ports
3).7 cm (12.5 in) Oiameter
II Center Port
40.6 cm (16.0 in) Oiameter
Burner
I
6 Bottom Ports
13.3 X 31.8 cm (5.25 X 12.5 In)
6. Tertiary Port Configuration for Initial Demonstration Boiler*
-------
Ignitor
Core Air Valve
Perforated Plate
Air Hoods
L
Removable Nozzle
N
Rings
Telescoping
Inner Nozzle
flrrT
Inner Register
Sliding
/ Sleeves
Outer
Register
Firing
Face
Coal Inlet
Cast Refractory
Exit
Figure 7. Prototype DMB (Tertiary Ports not Shown)
-------
Weight for
lUpId Cloture
Pivot
firing
f*ct
^ Slots for Am
OlfferentUI f
Pressure
vindicator
/...
Ilewlsplierlctl Poppet,
Closed Position
Pulley
Cable to
Actuator
Ver/or*te«l
?l»t§ 44|> Rood
Sleeve to Adjust
Call Are*
Sectio« ll»rou«jh Pivot [
Kick Hall of Mlndltox
Figure 8. Tertiary Port Control Valve.
-------
Coal « Utah Bituminous
Full Load « 26.4 Thermal MW (90 x 106 Btu/lir)
GOO
OVERALL STOICIIIOflCTRY - 1251 T.A.
LOAD - 891
Typical v
Field NQ
Data Trend '
500
NO
300
200
100
100 110 120 130 1
-------
Coal - Utah Bituminous
Load - 23.A Thermal MU (80 x 10® Btu/Hr)
600
600
500
500
P.
r
£,0°
*
<4
O
M 300
300
200
20(1
-I 100 1 1 I I I
150 160 |00 Ho m 130 HO fSO fiTo
OVERALL STOlCIIIOflCTHY (X TA)
100
120
130 140
Figure 10. Unstaged Prototype DHB and Interyane Burner Performance
in the Research Furnace.
-------
Load ** 23.4 Thermal MW
(80 x JO*" Jltu/hr)
600 -
500 -
2:
a.
a.
¦s
Csl
o
o
ca
t 1 1 1 1 1 r~i 1 r
§? 200
100 -
1
Staged at
400 LPverall Stoichioiuetn
^ 130% TA
300 -
Unataged (Buiner Zone
Stoichiometti
Overall Stoi<
600 -
500 -
6^
CD
ca
CD
CJ
Equals
biometry)
» «/i » I I I * * I
50 / 100 150
l i i 1 1 i 1 i 1 r
z:
a.
a.
^400
o
300 -
200
Staged (At Overall
Spoichiometry 130% TA)
100 s
Unst/aged (Burner Zone Stoichlometf
Equal6 Overall Stoichiometry)
I I III f I I I... I...,
50
100
150
Extrapolation to Design Point
(Overall Stoichiometry 125% TA,
Burner Zone Stoichiometry 70X TA)
BURNER ZONE STOICHIOMETRY
(% T.A.)
Figure 11. Prototype Burner Performance in Research Furnace.
-------
TABLE I. DMB NOMINAL DESIGN CRITERIA
Burner System
Parameter
Nominal Design Point
Fuel
Swirl
45° Vanes
Axial Velocity
22.9 m/sec (75 ft/sec)
Secondary Air
Stoichiometry
50-70% T.A.
Swirl
Variable
Axial Velocity
18.3 m/sec (60 ft/sec)
Tertiary Air
Number/Shape
4/Round
Radius/Throat Diameter
2.2
Velocity
15.2 m/sec (50 ft/sec)
Exit
Half Angle
25°
Length/Diameter
1.0
Setback:
0
205
-------
TABLE II. INITIAL DEMONSTRATION SITE CHARACTERISTICS
Parameter
Boiler
Conf iguration
Capacity MCR
Peak
Burners
Type
Array
Capacity MCR
Throat Diameter
Spacing Horizontal
Vertical
Furnace
Construction
Depth
Width
BZLR
Fuel
Coal Type
Pulverizers Type
Number
Air Supply
Air Heater Type
Secondary Air Temperature
Draft
Windbox Depth
Burner Pressure Drop (Nominal)
Value
Single Wall-Fired
98 x 103 Kg/hr (215 x 103 lb/hr)
111 x 103 Kg/hr (245 x 103 lb/hr)
Foster Wheeler Intervane
2x2
20 Thermal MW (69 x 10^ Btu/hr)
61 cm (24 in.)
182 cm (71.5 in.)
198 cm (78.0 in.)
Membrane Wall
6.28 m (20.6 ft)
5.01 m (16.4 ft)
492 Thermal kW/m^ (156 x 103 Btu/hr-ft^)
Indiana Bituminous
Foster Wheeler MB
2
Regenerative
265°C (510°F)
Pressurized
132 cm (52 in.)
8.9 cm H20 (3.5 in. H20)
206
-------
TABLE III. INITIAL INDUSTRIAL BOILER DEMONSTRATION SITE COAL CHARACTERISTICS
Proximate Analysis
% Moisture
% Ash
% Volatile
% Fixed Carbon
Ultimate Analysis
Moisture
Carbon
Hydrogen
Nitrogen
Chlorine
Sulfur
Ash
Oxygen (diff)
As Received
13.03
10.14
36.08
40.75
100.00
13.03
61.30
4.36
1.30
0.03
2.87
10.14
6.97
100.00
Dry Basis
11.66
41.49
46.85
100.00
70.48
5.01
1.49
0.04
3.30
11.66
8.02
100.00
207
-------
TABLE IV. INDUSTRIAL PROTOTYPE BURNER DIMENSIONS AND ADJUSTMENTS
Nominal
* Adjustment
Design Variable Value Range Method
Fuel Injector
Core Diameter (in.)
Core Area (in.2)
Nozzle Diameter (in.)
Annulus Thickness (in.)
Primary Area (in.2)
Setback (in.)
Inner Secondary Channel
9.75
74.7
14.5
2.375
90. 5
0
^68-136 } Telescoping Nozzle
-2-+12 Remove/Replace Nozzle Rings
Outer Diameter (in.)
Annulus Thickness (in.)
Area (in.2)
Setback (in.)
20.75
3.125
173.0
0
Variable
Variable
Variable
Variable
Replace Secondary Divider
Outer Secondary Channel
Outer Diameter (in.) 30.25
Annulus Thickness (in.) 4.75
Area (in.) 380.5
Setback (in.) 0
Tertiary Ducts
Distance from Burner (in.) 53
Spacing Around Burner (deg) 90
Number of Ports 4
Injection Angle (deg) 0
Axial Position (in.) 0
Diameter (in.) 13.25
Total (in.2) 553
Variable
Variable
Variable
Variable
Variable
Variable
Variable
Variable j
Recast Exit
Modify Windbox
Sleeve
Throat and Exit
Throat Diameter (in.)
Throat Area (in.)
Half Angle of Exit (deg)
Length of Exit (in.)
Length/Diameter
30.25
718.7
25
12.0
0.40
Variable
Variable
Variable
Variable
Variable
Recast Exit
Conversion Factors:
1.0 in. » 2.54 cm
1.0 in.2 = 6.45 cm2
208
-------
FIELD EVALUATION OF LOW EMISSION COAL BURNER
TECHNOLOGY ON A UTILITY BOILER
By:
E. J, Campobenedetto
Babcock & Wilcox Company
Fossil Power Generation Division
Barberton, Ohio 44203
209
-------
ABSTRACT
A program is currently in progress to demonstrate the Npx reduction
potential of the EPA distributed mixing burner applied to a utility
boiler. The demonstration program will be designed to evaluate both
emissions attributed to this burner as well as the effects of the burner
retrofit on overall boiler performance and efficiency.
Ihe boiler selection process is near completion with a single wall-
fired unit being reviewed by the EPA prior to final negotiations.
Several opposed-fired units are still under consideration, pending final
decisions by the utilities as to their interest in participating in the
retrofit demonstration program.
210
-------
ACKNOWLEDGEMENT
The vork presented in this paper was performed as part of the
Utility Boiler Field Evaluation of Lew Emission Coal Burner Technology
Program funded under EPA contract 68-02-3130. Hie contract is sponsored
by the Industrial Environmental Research Laboratory, Combustion Research
Branch. Hie author acknowledges the assistance of the EPA Project
Officer, G.B. Martin and the Energy and Environmental Research Corp. in
the planning stages of this program.
211
-------
INTRCDUCTION
The Environmental Protection Agency has regulated the emissions of
nitrogen oxides (NOjj) from utility boilers sold and erected since 1971
through enactment of the New Source Performance Standards. The Clean
Air Act Amendments of 1977 required the EPA to review and revise the
NSPS at four year intervals. The 1978 revisions lowered the NSPS for
Npx emissions from utility boilers and it is anticipated that further
reductions in the maximum allowable emission level of NO* will occur.
In addition, utilities in NOx non-attainment areas could be required to
retrofit NC^ control equipment to existing units as part of the state
implementation plans.
The EPA has developed a low emissions burner to provide for optimum
reductions in NO^ levels produced during the combustion of pulverized
coal. The development of this low emission coal burner technology began
in 1970 under a contract with the International Flame Research Foundation
in an attempt to characterize the effects of burner design parameters on
NO^ emissions from pulverized coal combustion^. The results of this
and further studies indicated that the levels of NO^ could be controlled
to low levels through combustion modifications.
212
-------
A nine task contract was awarded to the Babcock and Wilcox Co. in
1978 to demonstrate the use of this technology on a multiple burner
utility boiler. The purpose of this paper is to highlight the program
and to describe the progress to date in the areas of the boiler selection
and the burner design.
213
-------
TEST PLAN STRUCTURE
Ihe objective of the program is to evaluate the EPA lew emission coal
burner on a multiple burner utility boiler. The evaluation will attempt
to determine the lowest practical NO^ limit under normal operating conditions
without a sacrifice in boiler availability, performance, or efficiency.
The program consists of nine tasks as detailed below:
TASK
DESCRIPTION
1.0
Program Definition
2.0
Prototype Fabrication & Testing
3.0
Host Boiler Baseline Evaluation
4.0
Burner Fabrication & Installation
5.0
Unit Performance Evaluation
6.0
Industry Coordination Panel
7.0
Unit Restoration
8.0
Data Analysis
9.0
Guideline Preparation & Reports
A brief description of each task is provided below:
214
-------
l.o PROGRAM DEFINITION
This task provides for the initial planning and management plan
for the overall program effort. The task is divided into four subtasks
to provide planning in each of the major areas of the initial program
plan. Hie subtasks are: 1.1 Boiler Selection, 1.2 Burner Engineering
Design, 1.3 Analytical Measurements Plan, and 1.4 Overall Program
Plan.
1.1 Boiler Selection
The prime purpose of this subtask is to identify awl select
an existing unit to retrofit and evaluate the distributed mixing
burner. The unit selection is based on willingness of the
utility to participate as well as the state-of-the-art for
pulverized coal firing as applied to utility boilers both in the
existing population and the current new boiler market. Based on
the above, a detailed selection criteria was established to
evaluate potential candidate boilers. The utilities were then
contacted to determine interest in the program. A detailed
summary of the site selection process will be presented later in
the paper.
1.2 Burner Engineering Design
The experimental data on the lew emissions burner was
reviewed to establish the optimum design criteria for the prototype
215
-------
burner arxi subsequent retrofit burner. This information will be
incorporated with the site specific requirements of the unit
selected to finalize the prototype burner design. Also included
in the resultant design package will be the specifications for
the auxiliary equipment required for the retrofit and any constraints
placed on the burner design due to the limitations of the existing
boiler.
1.3 Analytical Measurements Plan
In this subtask, the overall approach for evaluation of the
emissions resulting frcm this burner will be defined. In addition,
the entire boiler performance evaluation program will be established
in the performance of this subtask. It is essential that both
boiler performance and emissions data be collected simultaneously
to determine the overall performance of the distributed mixing
burner system vs. the existing burner system which will be
characterized during the boiler baseline evaluation.
Energy and Environmental Research Corporation, EER, as a
subcontractor to B&w has the lead responsibility in the development
of the Analytical Measurements Plan for monitoring pollutant
emissions. This structure will provide for consistency between
the Industrial and Utility emissions monitoring programs. Also,
EER will establish the overall procedures necessary to satisfy
the requirements established by the EPA's Office of Air Quality
216
-------
Planning and Standards. Further discussions of the Analytical
Measurements Plan will be presented by EER.
1.4 Overall Program Plan
The results of the above subtasks will be incorporated into
a final program plan detailing the performance of the remainder
of the burner evaluation program. The plan will be based on the
selection of a boiler and the resultant outage schedule, burner
design, and measurements plan established for the specific unit.
This plan will be reviewed and revised periodically, as required,
to insure successful cottpletion of the demonstration program.
2.0 PROTOTYPE FABRICATION AND JESTING
A prototype burner will be fabricated based on the results of
Sub task 1.2 detailed above. The prototype burner will be a full size
burner of the design to be installed in the host boiler. This
burner will be installed in the EPA's large watertube simulator (LWS)
to verify the performance of the burner prior to finalization of the
design and fabrication of the multiple burners required for the field
demonstration. Tests will be performed to evaluate the burner performance
over a range of operating conditions and with several coal types
(including the coal to be burned during the utility demonstration) to
compare the results with those obtained during the experimental
burner test programs. A burner of the design currently utilized on
the selected utility unit will also be fabricated and tested on the
217
-------
UWS for comparison of emissions and burner performance.
3.0 HOST BOILER BASELINE EVALUATION
The baseline evaluation will consist of 4-6 weeks of boiler
testing prior to any modifications. These tests will provide a data
base for direct comparison of the distributed mixing burner. The
data to be collected will be used to evaluate overall system performance,
unit efficiency, and pollutant emissions. A continuous nonitoring
system will be utilized to collect the gaseous emissions data. Hie
data will be collected over the normal operating range of the boiler
to characterize the combustion system currently installed. Parameters
to be varied and reviewed will be boiler load, burner configuration,
excess air levels, and burner zone liberation rate.
4.0 BURNER FABRICATION AND INSTALLATION
The prototype testing will culminate in a final design package
for the burner system to be applied in the field. When the final
design package has been approved by the EPA and the utility, the
required number of burners will be fabricated and shipped to the
site. All required auxiliary equipment will also be fabricated or
purchased during this period. The above equipment will be installed
in the unit during an outage scheduled by the utility.
At the completion of the outage, the boiler system will proceed
through a normal start-up sequence. All burner settings will be
218
-------
determined and preliminary tests will be performed to insure that
overall operation of the unit over its normal load range has not been
adversely affected by the retrofit.
5.0 UNIT PERFORMANCE EVALUATION
At the completion of the burner system start-up, the performance
evaluation program will ccmence. Hie first step in the program will
be burner system optimization tests. Burner settings will be varied
to determine the optimum settings for burner stability, carbon utili-
zation, and ndjiimization of gaseous pollutant missions. Hie final
burner settings to be utilized during the long term evaluation program
will be established during this period.
She eighteen (18) month unit performance evaluation will begin
after the optimization tests are complete. Emissions data will be
collected on a continuous basis throughout the demonstration. In
addition, detailed measurements of boiler performance and emissions
will be conducted during the initial 30 day period of operation and
during the eighteenth months of operation. These detailed measurements
will be repeated periodically throughout the program to evaluate the
long term operation of the burner system.
6.0 INDUSTRY COORDINATION PANEL
Ihe purpose of this task is to secure the advice of manufacturers
and users of pulverized coal burners on the practical aspects of the
219
-------
technology and to assure the timely dissemination of the developing
technology to the potential users. Hie technical advisory panel
established includes potential users, boiler and burner manufacturers,
representatives of research and academic organizations involved in
coal combustion studies, and members of the EPA's Combustion Research
Branch. The members of this panel review all aspects of the experimental
programs as well as the industrial and utility programs to facilitate
the technology transfer through the various stages of development for
the distributed mixing burner.
7.0 UNIT RESTORATION
At the completion of the testing period, the boiler will be
restored to its original configuration during an outage scheduled by
the host utility. At the completion of the site restoration, the
system will then proceed through a normal start-up to insure that the
unit is operating in a mode acceptable to the utility.
8.0 DhTh ANALYSIS
The boiler and emissions data collected throughout the test
program (prototype, baseline, and retrofit) will be analyzed and
tabulated continuously throughout the program to provide the input to
direct the ongoing program. Ihe data will be analyzed to evaluate
overall system performance, emissions, and the effects of lew emissions
operation on boiler performance. Hie specific analyses include:
220
-------
1) Comparison of the experimental and prototype burner data to
identify the optimum design for the retrofit burner.
2) Comparison of the prototype and existing host boiler equipment
in the UWS to provide an estimate of the performance of the
distributed mixing burner in the host boiler.
3) Evaluation of the baseline data to determine the existing
performance and emissions from the unit. This will provide a
reference point to evaluate the benefits of the distributed
mixing burner.
4) Evaluation of the burner settings during the optimization stage
to determine the flexiblity of the burner performance in meeting
low N0X over the normal load range of the utility boiler while
maintaining acceptable system performance as evaluated against
the baseline tests.
5) Continuous review of the long term burner performance to document
any changes in emissions or unit efficiency due to fuel changes
or burner system deterioration. This analysis program will
provide input to the ongoing design review to identify potential
design modifications required to apply this technology to the
wide range of wall fired boilers both as a retrofit and as an
application to new units. The data will be evaluated utilizing
the baseline data as well as the burner evaluation techniques
221
-------
developed by B&W for evaluating N0X emissions and boiler performance
to generalize the technology for future applications.
GUIDELINE PREPARATION MP REPORTS
A guideline manual will be compiled at the completion of the
test program to sunmarize the results of the design study, field
tests, and the overall data evaluation. The manual will provide a
forum to present the program results and problems and solutions to
the manufacturers and utilities and will be used to assess the potential
applicability of this low NOy system on a commercial basis.
The manual will include:
1) Final design information for the prototype and field burners.
2) Design problems that must be addressed for various fuels.
3) Data sutmaries for each phase of the burner development program
(experimental, prototype, baseline, and retrofit).
4) Prediction curves developed for application of this technology
to utility boilers.
5) Functional and mechanical operation of the burner, including
problems that developed and the solutions employed to rectify
222
-------
the problems.
6) Effect of the low NC^ operating conditions on boiler performance
and other pollutants monitored throughout the program.
7) Parameters to optimize to obtain system performance equivalent
to the baseline performance.
8) Benefits and detriments of this method of lew NO operation
ccrrpared to other proven NOjr control methods.
223
-------
PROGRAM STATUS
The major effort for this program has been in the development of the
overall plan leading to a prototype test and subsequent field demonstration.
The two major items leading to conclusion of the planning stage were
selection of a suitable host site and development of the preliminary
distributed mixing burner design package for the host site selected.
The following section summarizes the work leading to the boiler
selection and preliminary burner design:
Boiler Selection
The two major factors in establishing the host site selection criteria
were: 1) The unit selected was representative of a large
segment of the existing population of units or of current designs
being sold and, 2) The economics of a retrofit of this magnitude
must be evaluated to establish a maximum boiler size. Based on the
above, the initial criteria was established to evaluate opposed fired
units less than 300 MWe capacity. The final selection criteria and prior-
itization schedule were established after a review of the trends in
fossil-fired utility boilers.
224
-------
A) History
Over the past ten years, the trend in the electric utility
fossil-fired population has been to coal fired units. This shift to
coal has been caused by the economics of coal vs. oil as well as the
uncertainty of oil supply from the Middle-East. Figures 1 and 2
provide a surrmary of the fossil-fired plant orders for the electric
utilities. Since 1975, virtually all units sold have been designed
for pulverized coal. The percentages shewn in Figure 1 are somewhat
low because the "coal only" megawatts do not include the units designed
for fuel combinations (coal/oil/ gas) as shewn in Figure 2. In
addition, the Fuel Use Act of 1978 prohibits the use of natural gas
or petrolevxn products as a primary fuel for a utility boiler.
Ihe average size of an electric utility boiler has also increased
significantly over the past twenty years. Figure 3 indicates the
trend in boiler size. In 1960, the average size unit place in
service was 133 MW with the largest unit being 500 W. This average
stayed fairly constant through the mid-1960"s. Fran 1965 on, the
average size of the steam generating units placed in service by the
electric utility industry increased steadily. ItxLay, the average
size unit is about 500 Mtf. The largest units in service are rated at
1300 fSJ.
Figure 4 provides an indication of the distribution pattern by
size of operating boilers as of 1979. As indicated, the majority of
225
-------
the units (65%) are less than 250 MW. In addition, most of these
units are one wall fired due to their size. Hcwever, with the trend
toward larger units, the majority of the units sold today are opposed-
fired units. A review of firing pattern and unit size of operating
B&W utility boilers is provided in Figure 5. As previously stated
the majority of the units are one wall-fired units (ranging in size
to about 250 MW maximum). The opposed fired units in the size range
initially established (300 MW maximum) represent a very small segment
(7.98%) of the existing population. Although the one wall-fired
units represent almost half of the total operating units, the one
wall and small opposed-fired units represent a very small portion of
the total steaming capacity of the electric utility boiler population
(14.1%).
In addition to the overall boiler population, a review of the
B&W furnace design criteria relating to N0X was also performed. The
two factors evaluated were the burner zone heat release rate (HA/SCi)
and the burner input. Figure 6 provides a summary of the changes in
HA/SC^ as NOx emission limits became a factor in boiler design.
Prior to promulgation of the New Source Performance Standards (NSPS)
of 1971, the burner zone heat release rate for the larger units
averaged between 700-800 KBtu/ HR-FT^, Field tests of these units
indicated a direct correlation between this parameter and NOjj emissions
for a given burner system. The HA/SC^ was reduced to the range of
400-500 KBtu/HR-FT2 in 1971 to meet the NQx emission levels established
by the 1971 NSPS. Currently, the average boiler is designed with a
226
-------
HA/SC^ of 350-450 KBtu/HR-FT2 to meet the 1978 USPS NO^- emission
levels.
Hie burner heat input has remained constant although the pulverized
burner design has changed significantly for NO* considerations.
Hie burner input has remained in the range of 100-135 x 106 Bt^HR as
shewn in Figure 7. The standard burner designed by B&W has a nominal
g
input of 125 x 10 Btu/HR and varies in the above range based on
matching of the number of pulverizers and burners required. For the
smaller units, the burner input can be as lew as 75 x 10^ Btu/HR
again dependent on the matching of the required nunber of pulverizers
and burners.
Selection Criteria
The preliminary selection criteria which was established based
on the above review was as follows:
Unit Size - 350 Maximim
Burner Zone Heat Release Rate - 500 KBtu/HR-ET^ Maximum
Firing Mode - Opposed Wall
Burner Type - B&W Circular or Dual Register
Fuel - Pulverized Coal
The initial list of potential units which fit the above criteria
was assembled and is shown in Figure 8. Ihese units were considered
227
-------
as the prime potential candidates for the burner retrofit program.
As the initial contacts were made, however, the responses received
frcm the utilities detailed above indicated minimum interest in a
program of this magnitude. In general, the responses from the utilities
were as follows:
The unit under consideration by B&W was a base loaded unit
and the utility could not risk an experimental program on a
unit which was the heart of their system.
With the generating capacity constraints what they are today,
the utility could not afford an extended outage to retrofit the
combustion equipment to their boiler.
The utility could not risk deterioration of unit performance or
a loss of availability or reliability for the unit as is possible
with an experimental program.
Utilities were very reluctant to cooperate with the EPA on a
program which could adversely impact future boiler designs with
additional constraints on pollutant emissions. This concern not
only included NO but carbon monoxide, hydrocarbons, fine particulate,
and trace elements which will be determined throughout the
program.
228
-------
Inadequate space was available in the burner zone to fit the
burner/tertiary air port configuration required. Burner-to-
burner side and vertical spacings were not adequate for the
installation of the required DMB hardware.
As a result of this response, the initial candidate list was
expanded and a selection prioritization schedule was developed. Uiis
unit prioritization is detailed in Figure 9. The initial seven (7)
units under review represented categories 1 and 2. Several larger
units, up to 750 MW were reviewed as potential candidates with the
same results as described above. A comparison of the design parameters
of several of these units are detailed in Figures 10 and 11. All
units to this point were opposed fired units.
The next series of units for review were the single wall-fired
units. Based on a review of several units, discussions were held
with severed utilities that were interested in the program. These
discussions led to three units which the utilities were willing to
conmit to this program. The design parameters for these three units
are shown in Figure 12. Based on this review, unit 3 was eliminated
because it was too small (75 NW). In the design review of units 1
and 2, it was determined that unit 2 would be more difficult to
retrofit because of obstructions in the windbox which could cause
improper air flew distribution to the tertiary air ports. Therefore,
it was determined that unit 1, a 175 unit operated by Northern
States Rwer (NSP) would be the prime candidate for the EPA law NOjj
burner retrofit.
229
-------
This unit, located at NSP's Highbridge Station, has a maximum
capacity of 1,200 MLB/HR main steam flow and is currently equipped
with 16 circular burners arranged in a 4 x 4 matrix on the front
wall. The burner zone is 48 feet wide by 21 feet deep. The unit was
originally designed for an Illinois high sulfur coal, but currently
burns a blend of Illinois and Montana coals for SC>2 emission consider-
ations. A sketch of this unit is shewn in Figure 13.
NSP has consented to participate and negotiations are continuing
pending approval of this unit by the EPA Project Officer .
In addition, discussions are still continuing with utilities
which are operating the larger, opposed-fired units to determine if
an opposed fired unit can be made available for this demonstration.
Burner Design
A preliminary burner design was developed during the early stages
of the contract. This was established based on a nominal 125 x 106
Btu/HR burner input. A general set of criteria was established to identify
the practical requirements for the low N0X burner retrofit. These criteria
are identified in Figure 14. These criteria are based on the practical
limitations of pulverized coal combustion and will be used to evaluate the
success of this burner retrofit along with the NC^ reduction potential of
the burner.
230
-------
Hie final design specifications for the prototype and field retrofit
burners will be established based on the requirements of the host boiler
selected. A preliminary design, based on the experimental results obtained
g
by EER, the current B&W design limits, and a nominal 125 x 10 Btu/HR
burner are detailed in Figure 15. Figure 16 is the preliminary sketch of
the prototype burner design based on the design specifications detailed
above.
One problem that has narrc*\ed the field of potential host boilers has
been the burner spacing on several units under preliminary review. To
fit the tertiary ports around the burners, there must be adequate clearances
to install the ports and the required airflow control systems while maintain-
ing an adequate number of straight tubes in the furnace walls for support.
Several units under review had inadequate burner to burner spacings to
even consider the retrofit program.
Another problem that has developed is the problem of retrofitting
this burner design to an operating unit. Besides the burner design, it
appears that major revisions to the structural supports in the windbox
will be required to install this burner. Also a stress review of any unit
under consideration must also be made to evaluate the structural integrity
of the furnace wall after bending all the additional furnace wall tubes
required to form the tertiary air ports. At present, these studies have
not been pursued in detail pending final approval of the host boiler by
EPA.
231
-------
After approval has been received, a complete evaluation of the unit
will be performed to determine the modifications to the boiler and supports
required to retrofit this low N0X burner.
232
-------
OQNCULJSIONS
A tentative host boiler has been selected at Northern States
Power's Highbridge Station.
Hie utility burner retrofit program is behind schedule due to
delays in obtaining a host site.
The current schedule, pending EPA approval, is to start the
demonstration program late in 1981.
The retrofit will be more difficult than originally anticipated
dug to the structural modifications required to install the
burner and tertiary air por 3.
Efforts are still proceeding to locate a utility operating
a snail opposed-fired B&W boiler for possible participation in
this program.
233
-------
REFERENCES
1. Martin, G.B., "Field evaluations of Lew NO Coal Burners
on Industrial and Utility Boilers". Presented at the Third
Stationary Source Combustion Symposium, San Francisco, California,
March 5-8, 1979.
2. Kidder, Peabody & Company, Fossil Boilers Status Reports.
3. EBASCo, 1979 Business and Economic Charts.
4. Heil, T.C. & Durrant, O.W., "Designing Boilers for Vfestern
Coal". Presented at the Joint Power Generation Conference,
Dallas, Texas; September 10-13, 1978.
234
-------
PLANT ORDERS (MWE) IN THE U.S.
BY ELECTRIC UTILITIES2
BOILERS
YEAR
COAL
ONLY
TOTAL
FOSSIL
COAL AS
% OF
TOTAL FOSSIL
1963
5,103
13,767
37.1
1964
4,684
15,535
30.2
1965
13,435
21j613
62.2
1966
13,242
20,797
63.7
1967
14,006
24,975
56.1
1968
13,383
22,810
58.7
1969
18,720
27,012
69.3
1970
12,234
29.635
41.3
1971
6,784
15,817
42.9
1972
9,424
16,956
55.6
1973
21,434
26,605
30.6
1974
29,438
32,522
90.5
1975
7,281
7,690
94.7
1976
5,318
6,018
38.4
1977
12,373
12,431
99.5
1978
13,457
14,424
93.3
1979
5,467
5,467
100.0
##GAS, OIL,, OR COAL (OR SOME COMBINATION THEREOF).
FIGURE 1
235
-------
Year
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
NOTE
NUMBER OF FOSSIL BOILER ORDERS BY ELECTRIC UTILITIES'
Gas
Oil
(?as/Qil
Oil/Coal
Oil & Gas
Coal
3
6
9
-
6
29
2
16
11
-
8
24
1
11
3
-
4
15
-
10
4
2
7
20
-
4
3
-
5
38
-
-
4
-
7
60
-
-
1
-
1
14
-
-
-
-
1
12
-
1
-
-
-
24
-
-
-
-
3
27
_
—
_
_
-
14
THIS TABLE INDICATES THAT MOST OF THE FUTURE ORDERS OF FOSSIL
BOILERS FOR THE ELCTRIC UTILITY INDUSTRY WILL BE FUELED BY
COAL.
FIGURE 2
236
-------
STEAM-ELECTRIC GENERATING UNITS
PLACED IN SERVICE
BY THE ELECTRIC UTILITY INDUSTRY3
M
LO
1500
1400
1300
1200
1100
<0
1000
o
is
900
"o
800
w
"O
700
c
<0
M
600
3
o
500
K
400
300
200
100
0
Size of units
SS Largest unit
Average unit
S $ S Ls_LL*
1948 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78
Figure 3
-------
DISTRIBUTION PATTERN OF COAL FIRED UNITS
IN SERVICE IN THE U.S. IN EARLY 1979 2
NUMBER OF UNITS
RANGE (MW) COAL
50-250 481
250-450 130
450-650 76
650-850 45
850-1050 2
1050-1250 3
1250-1450 5
FIGURE 4
238
-------
FIRING PATTERN & UNIT SIZE DISTRIBUTION
7o BY % STEAM
UNITS FLOW
ONE WALL FIRED UNITS 42,9 8.8
OPPOSED-LESS THAN 300 MW 8.0 5.3
opposed-300 to 600 mw 22.3 27.7
OPPOSED-GREATER tHAN 600 MW 26.8 58.2
FIGURE 5
239
-------
4
Pulverized Coal Fired Boiler Experience
N
O
Burner
Zone Heat
Release
Rate,
Thousand
Btu/sq. ft-hr
800
700
600
500
400
300
••
• •<
• •• •
M • •
• •
• •
•• • •
••
- *
• •
• •
• •
— •
• •
• • •
•• • * .
• •
• •
1962
1970
Order Year
1975
1978
Figure 6
-------
Pulverized Coal Fired Boiler Experience^
4*
Burner
Heat
Input
Million
Btu/hr
140
130
120
110
100
90
80
70
_ •• t«*
• •
» ••
• • ••
••
••
• ••
• •
• • •
• •
• •
••
• •
1962
1970
Order Year
1975
1978
Figure 7
-------
POTENTIAL HOST SITE CANDIDATES
UTILITY
A
HEAT
SIZE RELEASE BURNER
nW RATE PATTERN
NUMBER
BURNER OF COAL NOx
TYPE BURNERS TYPE PORTS
350 384 opposed circular 32 subbit. yes
B
333
508
//
DUAL REG. 24
BIT, NO
265 502
CIRCULAR 18
BIT, NO
D
250 325
//
CIRCULAR 18
DUAL REG.
BIT, NO
250 299
CIRCULAR 24
BIT, NO
F
350 390
DUAL REG. 30 SUBBIT. NO
6
216 342
//
CIRCULAR 20 LIGNITE NO
FIGURE 8
242
-------
SELECTION PRIORITIZATION
NEW UNITS - REPRESENTATIVE OF CURRENT DESIGNS
OLDER UNITS - 300 MW RANGE AND REPRESENTATIVE FURNACE
DESIGN.
LARGER OPPOSED UNITS
ONE WALL FIRED UNITS
UNITS WHERE NO BASELINE IS POSSIBLE.
FIGURE 9
243
-------
DESIGN PARAMETER COMPARISON
PRIMARY UNITS
UNIT 1
UNIT 2
STEAM FLOW
(mlb/hr)
2300
2534
S.Hi TEMP C°F)
1005
1005
S.H, PRESS.
2610
2500
FEEDWATER TEMP.
485
486
SEC, AIR (°F)
541
600
TEMP, GASLVG,
BOILER (°F)
655
690
TEMP. GAS LVG,
AH (OF)
291
266
UNIT EFF. it)
88,88
84,57
FUEL; BTU/LB
11,500
8250
ASH
16,0
5,2
h2o
5.0
29.0
n2
1,3
0,9
DRUM PRESS.(PSIG)
2770
2660
ha/sc CKbtu/hrft^)
508.4
383,5
FIGURE 10
244
UNIT 2
1450
1005
1990
460
647
750
270
85,23
8624
4,87
28,11
0,67
2115
314.5
-------
DESIGN PARAMETER COMPARISON
SECONDARY UNITS
UNIT 5.
unit 6
unit 7
STEAM FLOW (mLB/hr)
2,404
2,355
4,050
S.Hi TEMP (°F)
1,000
1,005
1,005
S.H. PRESS,(PSIG)
3,575
2,625
3,805
FEEDWATER TEMP.(°F)
500
542
550
SEC. AIR (°F)
603
565
589
TEMP. GAS LVG.
BOILER (OF)
646
635
628
TEMP GAS LVG,
AH (°F)
244
267
240
UNIT EFF. (%)
91.20
39.13
89,90
FUEL., BTU/LB
13,550
12,180
11,000
ASH I
8,5
11.6
12.48
h2o %
4.0
5.2
5.2
n2
1.3
1.2
1.0
DRUM PRESS. (PSIG)
NA
NA
NA
ha/sc (Kbtu/hr FT^)
600
460
690
FIGURE 11
245
-------
DESIGN PARAMETER COMPARISON
ONE WALL-FIRED UNITS
STEAM FLOW (MLB/HR)
S.H, TEMP. (°F)
S•Hi PRESS. (PSIG)
FEEDWATER TEMP, (°F)
SEC. AIR (°F)
TEMP. GASLVG.
BOILER (°F)
TEMPa GAS LVG.
AH (6F)
UNIT EFF. (%)
FUEL, BTU/LB
ASH %
h2o %
n2 %
DRUM PRESS. (PSIG)
o
ha/sc (Kbtu/hr ft^)
UNIT 1
UNIT 2
UNIT 3
1,200
1,000
600
1,005
1,003
1,005
1,875
1,825
1,990
470
485
463
593
608
590
705
719
695
287
292
283
87.31
88.33
87.58
10,400
11,530
11,683
12.2
13.0
9.31
13.0
8.5
6.90
0.97
1.4
1.21
2150
2,050
2,225
254
235
290
FIGURE 12
246
-------
Northern States Power
Highbridge #12
21'
- (£ BNR
Figure 13
247
-------
PRACTICAL REQUIREMENTS OF LOW NOx BURNERS
SUITABLE FOR EXISTING ENVELOPE
MINOR MODIFICATION TO HEAT FLUX PROFILE
NO EFFECT ON THERMAL EFFICIENCY
NO EFFECT ON UNIT PERFORMANCE
NO INCREASE IN RELATED POLLUTANTS
NO INCREASE IN LOWER FURNACE CORROSION POTENTIAL
FIGURE 14
-------
LOW EMISSIONS BURNER DESIGN
ASSUMPTIONS
1
BURNER SIZE - NOMINAL 125 MKB/HR.
AREA INNER ANNULUS = AREA ,OUTER ANNULUS
THEORETICAL AIR = 7.57 LB/10 KB (AS FIRED) n
SECONDARY AND TERTIARY AIR TEMPERATURE = 550UF
PRIMARY AIR TEMPERATURE = 150OF SI BURNER
WINDBOX PRESSURE = 4 INCHES W,C.
PRIMARY AIR PRESSURE = 5 INCHES W.C. 3 BURNER
BAROMETRIC PRESSURE = 29,92 INCHES HG
DESIGN CRITERIA
DUAL AIR ZONE BURNER
TOTAL EXCESS AIR = 15%
STOICHIOMETRY
PRIMARY AIR - 0,2-0.25
BURNER - 0.7 TOTAL (INCLUDING PRIMARY
air)
TERTIARY PORTS - 0.45
4. VELOCITIES
THROAT -,3600 FPM
PORTS - 3000 FPM (A PORTS)
NOZZLE - 4500 FPM (3000 FPM MINIMUM)
5. THROAT (QUARL) HALF ANGLE - 25°
FIGURE 15
249
-------
DISTRIBUTED MIXING BURNER
PRELIMINARY DESIGN
Ui
O
FIGURE 16
-------
OPERATING EXPERIENCE AND FIELD DATA OF A 700 MW COAL-FIRED
UTILITY BOILER WITH RETROFIT LOW N0X STAGED MIXING BURNERS
By:
K. Leikert and S. Michelfelder
L. & C. Steinmueller GtabH
Postbox 1949/1960
D-5270 Gummersbach 1
West Germany
251
-------
ABSTRACT
Forthcoming new federal regulations on emission control for
stationary combustion systems will clearly define tolerable NO -
emission levels in Germany and thus replace the present "best x
technical means" approach.
This fact gave reason to the initiation of a R&D program
for the development of cost effective low NC>x-combustion equip-
ment for use in pulverized coal fired boilers.
The program which was financially assisted by the Federal
Ministry for Research and Development (Bundesministerium ftir
Forschung und Technologie, BMFT) was subdivided into two parts:
As a first step a distributed mixing burner design concept
adopted for an envisaged 50 % NOx-reduction was tested and opti-
mised in a pilot plant test program with a 2.5 MW burner. The
burner design concept was based on a conventional circular bur-
ner with additional tertiary air nozzles distributed concentri-
cally about the burner mouth. Within this program a 65 % NOx
reduction was achieved with an optimised configuration of the
distributed mixing - respectively staged mixing burner (SM-bur-
ner) without disadvantageous changes in combustion and emission
characteristics.
Following the successful pilot plant tests the combustion
equipment of a 700 MW coal fired power station was changed to
SM~burners. Both, to ease the retrofit and to safely allow the
execution of a measuring program as retrofit burner a modified
version of the optimized pilot burner design was used for the
boiler. The modification resulted in a limitation of the safely
operable tertiary air mass flows and thus to a certain extent
reduced the staging capability of the large burner.
In spite of the narrower operational limits the envisaged
goal of a 50 % NOx reduction was met in the boiler demonstra-
tion tests. Furthermore a burner load dependent automatic secon-
dary/tertiary air flow control system developed to guarantee a
safe burner operation at low NOx levels over a wide turn down
ratio, was operated successfully.
252
-------
Parametric studies, in flame and flue gas measurements
were carried out to identify effects on NOx-formation and
potential problem areas at low NOx-operation. Similarly to
the pilot plant tests the staged mixing - dividing the flame
into a fuel rich primary and a secondary burnout zone - proved
to be the most effective single parameter on N0X reduction. In
the past 24 months of operation neither short term nor long
term performance problems due to low N0X operation could be
detected.
SECTION 1
INTRODUCTION
Both federal government and local regulations on NOx~
emission from power station boilers in Germany force utilities
and power station manufacturers to develop techniques which
allow the operation of power station boilers in compliance with
present and future regulations.
Unlike to flue gas desulphurization it has been shown by
various international R&D programs that the most cost effective
method to meet present and medium future N0X legislation lies in
the reduction of NOx-formation by optimizing the combustion pro-
cess , respectively the combustion equipment.
Numerous researches have clearly indicated that it has to
be differentiated between basically two NOx-formation mechanisms,
. the 'thermal N0X' route, comprising all
reactions which in their final result lead
to the oxidation of originally molecular
nitrogen, and
. the 'fuel NOj,' route, describing the oxidation
of fuel bouna nitrogen.
It is now well established that the thermal NOx-formation
predominantly depends upon the reaction temperature and the pre-
vailing residence time at high temperatures, whereas the fuel
NOx~formation is more dependent on the oxygen concentration in
the pyrolysis zone of the flame.
The relative importance of these individual reaction routes
on the overall N0X formation varies with the nitrogen content of
the fuel. With natural gas firing due to the absence of fuel
bound nitrogen only thermal NOx can be formed, and thus the emis-
sion be minimized by controlling the time temperature history of
the combustion products.
253
-------
At residual fuel oil and pulverized coal firing the fuel
NO formation mechanism often plays the dominant role. Experi-
ments by Pershing et al (1) have shown that upto 80 % of the ove-
rall NOx emission at coal firing was contributed by the fuel N0X-
formation mechanism.
As coal is or soon will be the only important fossil fuel
for power generation in Germany it is evident that controlling
NOx emission from power stations will only be possible when the
fuel NOx formation mechanism can be controlled.
As stated above, this calls for an oxygen concentration con-
trol in the pyrolysis zone of the flame, which at the simultane-
ous requirement of short intense flames can be achieved by a sta-
ged respectively distributed mixing burner concept.
Using the Steinmiiller circular burner configuration as ba-
sis, a novel distributing respectively staged mixing prototype
burner (SM-burner) was designed which was subsequently further
optimized and demonstrated in a two-phase program upon which is
reported in this paper. A pilot test and optimization program
with a 2.5 MW test burner was followed up by a largescale demon-
stration at a 700 MW power station which is equipped with 24 re-
trofit SM-burners.
Both program phases were carried out in cooperation with
and financial assistance of the German Federal Ministry of Rese-
arch and Development (Bundesministerium fiir Forschung und Techno-
logie, BMFT).
SECTION 2
THE SM-BURNER CONCEPT
As Fig. 1 shows the SM-burner is derived from the reliable
circular swirl burner. The pulverised coal is fed to the burner
throat by an annular fuel gun and injected concentrically to the
burner axis.
The former secondary air is now split up into two streams,
a reduced secondary air flow which is passed as usual through
the annular air duct surrounding the fuel gun, and a tertiary
air flow to four tertiary air nozzles which are placed concentri-
cally about the burner quarl exit.
At 100 % coal firing the core air butterfly valve is al-
most closed. It is adjusted in such a way that just sufficient
core air is passed through the central duct to avoid recircula-
tion flows induced by the swirling secondary air into the core
air duct.
254
-------
A variable swirl device allows a control of the secondary
air swirl intensity.
The tertiary air quantity, respectively the tertiary/secon-
dary air ratio , can be controlled by the butterfly valve in the
tertiary air duct.
To control the overall burner stoichiometry, the total air
flow to each burner is measured individually by a venturi type
flow meter.
Fig. 2 shows a schematic of the anticipated flow and mixing
regime of the SM-burner. As a result of the air supply chosen,
the flame, which is stabilized within the burner mouth due to
the prevailing aerodynamic conditions, is characterized by two
distinct zones of different stoichiometry, a fuel-rich primary
zone close to the burner exit, followed by an overstoichiometric
secondary or burnout zone. The design concept thus enables to
fulfil the low fuel NOx-formation requirement of low oxygen con-
centration in the fuel pyrolysis zone of the flame and additio-
nally, but to a lesser degree, reduces the thermal NOx-formation
as a consequence of lowering the maximum flame temperatures due
to interstage cooling and increase of recirculation matter en-
trainmen t.
One of the main problems to be solved was to optimize the
burner design and control in such a way to achieve optimum, low
N0X' conditions without substantial loss of combustion intensity/
turndown potential and burnout performance.
The realization of simultaneous optimum, low NOx-conditions,
high combustion intensity and good burnout performance was achie-
ved by an experimental optimization of
- shape, number and outlet area of
staging air nozzles
- cross-sectional area of secondary
air duct
- swirl intensity of secondary air flow
- fuel injection angle and velocity, and
- length of burner quarl.
As described in more detail in (2) a burner configuration
could be derived from the pilot plant results for the 2.5 MW
prototype burner which yielded the anticipated flow pattern
(Fig. 3, type II flame) for a wide range of input conditions. At
an overall burner stoichiometry of nt = 1.25 a stable type II
flame flow pattern was achieved at full load down to a 'staging
ratio' of np = 0.55, 'staging ratio' n_ being defined as the
stoichiometry of the primary zone (Fig. 2).
255
-------
A further decrease of rip caused a dramatic increase in flame
length due to a change in flow pattern (Fig. 3, type I flame).
As long as the type II flow pattern is preserved (n > 0.55)
staged mixing operation results in only moderate changes^of flame
length and combustion intensity. This fact is manifested by the
axial decay of CO concentration on Fig. 4. Also plotted on Fig. 4
is the axial concentration of free oxygen with the staging ratio
as parameter. The oxygen pattern, proves the presence of a fuel-
rich primary respectively pyrolysis zone close to the burner at
n <1 .
P
The necessity to keep the fuel transport air quantity vir-
tually constant independent of burner load in order to maintain
the transportation characteristics does at decreased burner loads
result in a reduction of the secondary air to primary air momen-
tum ratio as long as the staging ratio np is kept constant. This
momentum ratio, however, controls the flow pattern. At load re-
duction this causes the danger of a flame pattern change from
type II to type I (Fig. 3).
To overcome this problem, an automatic load-dependent secon-
dary/tertiary air distribution control procedure has been deve-
loped for the SM-burner, which was first tested during the demon-
stration of the 24 retrofit SM-burners at the 700 MW power sta-
tion.
The load-dependent air flow control procedure is graphically
illustrated on Fig. 5. Whereas the total air (n = 1.25) is redu-
ced linearly with load, the secondary air flow follows the total
air reduction only up to a certain load limit. At a further load
decrease the secondary air flow is only degressively reduced
which, of course, can be verified only at the expense of the ter-
tiary air flow. As a consequence thereof the staging ratio is
increased at reduced load in order to safely maintain the desi-
red combustion characteristics at low load and yet to enable a
maximal, obtainable NOx reduction at full load where the prevai-
ling conditions allow safe operation at lower staging ratios.
SECTION 3
DEVELOPMENT DATA AND RETROFIT EXPERIENCE
The single burner pilot plant tests (2) had clearly indi-
cated that at the conditions investigated the by far strongest
individual effect on N0X emission was caused by the staging
ratio np. Fig. 6 exhibits the NOx emission as a function of the
staging ratio with the burner load as parameter. From this in-
formation it is also obvious that with respect to NOx control
the low load burner air flow control illustrated on Fig. 5 is
tolerable, as load reductions result in considerably lower N0X
emissions aiyhow.
256
-------
The retrofit action was carried out at a 700 MW opposite
wall fired single pass power station boiler with 24 burners. The
boiler radiant section was subdivided by an evaporator-integra-
ted division wall to minimize the slagging potential.
To reduce the retrofit costs and to minimize the operatio-
nal risks it was decided to make use of the existing burners,
thus adding only the tertiary air ports, the necessary ducting,
flow control devices and instrumentation. Whereas this decision
ensured that by only closing the valves of the tertiary air duct
the original burner conditions could be restored at any time it
limited the full exploitation of the SM-burner design for maxi-
mum NOx reduction, as the optimum secondary air duct design could
not be incorporated in the retrofit burner.
Results obtained from the power station after retrofitting
the SM-burners are presented in two subsections. The data dis-
cussed and presented below refer always to flue gas measurements
and simultaneous operation of 24 burners under same conditionst
unless otherwise specified.
The demonstration tests as well as the pilot investigations
were carried out with German SAAR coal of the following average
composition:
Total water:
10
%
by weight
Inherent moisture:
1
.9
%
by weight
Ash:
8
.24
%
by weight
daf
Volatiles:
32
.2
%
by weight
daf
Sulphur:
0
.89
%
by weight
daf
Nitrogen:
1
. 1
%
by weight
daf
Net calorific value:
26
.42
MJ/kg
Analyses taken over a
longer period
and
illustrated
on Fig.
7 show a considerable variation of the nitrogen content with
time, which can of course affect N0X emissions. The data for
each individual parameter study have, however, been taken over a
comparatively short time (several hours) so that a constant N-
content can be assumed during the investigation of individual
parameters.
PARAMETRIC EFFECTS ON NO EMISSIONS (700 MW BOILER)
Although the staging ratio in the retrofit burners was li-
mited due to the above-mentioned design compromise, it proved to
be again the most effective individual parameter. Fig. 8 shows a
strong dependency of flue gas emission levels on the staging
ratio. The overall excess air level, characterized by the vari-
ous symbols on Fig. 8, has only a negligible influence.
257
-------
Furnace respectively burner loading has also a substantial
effect on NOx emission, as is illustrated on Fig. 9. This fact
is important as it more than equalizes the effect of the limited
staging ratio at low load burner operation due to the automatic
air flow control (Fig. 5).
Whereas some authors have reported an effect of combustion
air swirl intensity on N0X formation, this investigation revea-
led, neither at unstaged nor at staged operation, an influence
of secondary air swirl intensity on NO emission (Fig. 10,
n = 0.9). x
P
Heap and Pershing (3) measured a pronounced increase in NO
formation with decreasing particle size distribution in a one- x
dimensional flow reactor at unstaged condition. A change of the
coal fineness during this demonstration test from nominal 35 %
> 90 % nm to 29 %> 90 um indicated an opposite trend (Fig. 11).
The likely reason for this discrepancy lies in the different
flow and mixing regimes. An increase of fineness in the one-di-
mensional flow reactor results at unstaged conditions in the
first place in a better mixing between fuel and oxidant and thus
also increases the probability that nitrogen species are oxidized
immediately after devolatilization. Under staged conditions with
the SM-burner a decrease in particle size distribution also leads
to a better mixing resulting in a faster reaction, however with
the important difference that at the overall fuel rich equiva-
lence ratio in the primary zone this rather leads to an increase
in oxygen deficiency than to an increase in nitrogen species
oxidation.
IN-FLAME MEASUREMENTS AND OPERATIONAL EXPERIENCE
(700 MW BOILER)
Flame Measurements
Measurements within the radiant chamber of the boiler have
been carried out to identify any significant changes due to the
staged mixing operation. Visual observations did not show any
important changes.
Temperature measurements carried out with a radiation pyro-
meter (Pyropto) illustrate the relativechanges in ignition zone
(primary zone) temperature and the overall axial temperature dis-
tribution (Fig. 12). It can be seen that the primary zone tempe-
rature, respectively radiation intensity, drops when the burner
is operated in the distributed mixing mode, whereas the axial
temperature pattern exhibits a slight increase. Inspite of these
changes the heat absorption characteristics of both radiant sec-
tion and superheater remained, however, unchanged. A further re-
duction of the slagging tendency at the burner level cross-sec-
tions of the boiler, observed when the burners were operated
258
-------
in the distributed mixing mode, may be attributed to the tempe-
rature reduction in the primary zone.
Another important question which could be clarified parti-
ally by measurements within the furnace atmosphere is if corros-
ion processes at the outer tube surfaces would be stimulated or
accelerated due to the distributed mixing burner operation. As
potential danger an increased CO-content, with its detrimental
effect on CO-corrosion mechanisms, was identified. To investi-
gate these potential problems, two measurements have been per-
formed:
. CO-concentration measurements in the furnace
chamber
. tube wall thickness measurements before the
retrofit action and again after 24 months
of boiler operation with staged mixing bur-
ners.
The 2nd series of tube wall thickness measurements have not
yet been carried out, but results should be available at the
time this paper is being presented.
Gas concentration measurements in the external recircula-
tion matter, which are illustrated on Fig. 13 however, do not
indicate any increase in CO-corrosion danger. Inspite of the in-
spected CO concentration increase towards the burner axis, sta-
ged mixing operation does not lead to an increase of CO at the
furnace wall areas.
Long-Term Flue Gas Data
At no time any change in flue gas CO emissions was obser-
ved, the retrofit of SM-burners thus had no detrimental effect
on the CO emission characteristics.
Long-term measurements of uncombusted carbonaceous matter
in the fly ash did exhibit a moderate but distinct increase (Fig.
14). Before the retrofit of the SM-burners, ash analyses revea-
led a 2-3 % content of combustible matter in the fly ash. After
the retrofit of SM-burners, the carbonaceous matter in the fly
ash increased to a maximum of 4-6 %. This tendency seemed to
present basic problems in as much as it results in an efficiency
loss and, furthermore, leads to potential problems in the use
and sale of fly ash to cement works who limit the content of
carbonaceous matter in fly ash for their use generally to 5 %.
259
-------
A crosscheck with fly ash obtained from an unstaged opera-
tion, which was possible due to the retrofit burner compromise
(reuse of old burner with controllable tertiary air streams),
gave however no change in fly ash composition. This indicated
that the fly ash problem had to be attributed mainly to another
cause which was found in the mill system- The mills had run more
than 15 000 hours without any change of wear parts, resulting in
an increase in particle size distribution up to 4 2 %> 90 nm.
After an appropriate adjustment of the classifiers the un-
burnt matter in fly ash could again be reduced to an average of
3.5 %, as shown for the 2nd half of 1979 in Fig. 14.
Additionally to the problem of unburnt matter in fly ash, it
was of interest if and to what extent the collection efficiency
(respectively the fly ash properties) of the electrostatic pre-
cipitator would be effected by the distributed mixing mode opera-
tion of the SM-burners. Neither precipitator efficiency and total
solids emission measurements nor electron microscopic evaluations
of ash samples revealed any influence of staged mixing. The lat-
ter results are to some extent contradictory to preliminary re-
sults from Martin (4) who reported that retrofitting low NO coal
burners to a boiler in the U.S. had resulted in a distinct ¥n~
crease of fine particulate formation.
Finally, long-term NO emission data are plotted on Fig. 15,
which are representative f§r today's continuous boiler operation
with the adopted automatic combustion air control illustrated on
Fig. 5. Whereas the triangles on Fig. 15 are points from a mea-
suring series which has been taken to assess the dependency of
NOx emission on burner load for the adopted air flow control, the
circular points represent data which have been collected over a
10-week period in March/April 1979 and two weeks in February
1980. In spite of the considerable scatter of + 40 ppm, which may
be attributable to irregularities in the coal composition, all
values were below the 1979 EPA NSPS.
SECTION 4
SUMMARY - CONCLUSIONS
The paper reports on a development and design optimization
programme for a distributed mixing burner and on retrofit expe-
rience from a 700 MW opposed wall single pass boiler. Notable re-
sults can be summarized as follows:
1) The staged mixing, respectively distributed mixing
principle, was proven to be an effective tool for the
reduction of NO emission from coal-fired power sta-
tion boilers. T^e achieved reduction of more than 50 %
is predominantly attributable to the suppression of
fuel NO^. formation.
260
-------
2) The SM-burner concept realizes the NO reduction with-
out significant changes in combustionxintensity and flame
shape nor loss in turndown capacity.
3) It can be expected from the pilot experiments that a
full size SM-burner design based upon the optimized pi-
lot burner configuration - without compromise as accep-
ted for the retrofit burner - will result in a further
increase in N0X reduction efficiency to approximately
70 %.
4) The slagging tendency in the radiant section reduced
when the SM-burner was operated in the distributed mi-
xing mode. No disadvantageous effects on operational sa-
fety and CO corrosion potential were detectable up to
now. Tube wall thickness measurements which will be car-
ried out shortly after a 24-month operation cycle shall
allow final conclusions on this point.
5) Neither a change in fly ash properties and particle size
distribution nor an increase in CO emission have been
found at distributed mixing operation.
6) A slight increase in uncombusted carbonaceous matter in
the fly ash has been detected. The absolute values of
approximately 3.5 %, however, lie within the tolerable
limits. This increase in carbonaceous matter in the fly
ash is noted here as a potential disadvantage although a
crosscheck with non-staged mixing operation indicated
that it has to be rather attributed to a change in coal
quality than to the distributed mixing concept.
7) Until now the operational experience is limited to one
coal and one boiler only. Before a final generalization
of the reported results and experience is possible more
information is required. The effect of coal type will
be investigated by further tests in the 2nd half of 1980
as part of a joint program between EPA, BMFT and Stein-
mUller.
These tests will be carried out atEnergy & Environmental
Research Corporation, Santa Anna in California. Further
experience will be gained from two new 600 MW boilers
presently under construction in Germany.
8) Once the information to be generated from the additional
tests and the new boilers is available it can be hoped
that with the SM-burner concept a cost effective low
NO technology is available for both new and via retro-
fit also existing installations.
261
-------
SECTION 5
REFERENCES
(1) Pershing, D.W., Brown, J.W. , Martin, G.B. and Berkau, E.E.:
Influence of design variables on the production of thermal
and fuel NO from residual oil and coal combustion. Presented
at 66th Annual AIChE Meeting, Philadelphia, Nov. 1973.
(2) Michelfelder, S., Leikert, K.
Development program and operating experience with pulverised
coal staged combustion burners for steam boilers.
American Flame Research Committee,
Oct., 22-23, 1979, Houston, Texas
(3) Heap, M.P., Pershing, D., EER- SANTA ANNA
Private communication on experimental data
obtained in bench scale tests
Sept. 1978
(4) Martin, B., EPA - RTP
Private Communication
262
-------
TOTAL AIR
SUPPLY
SECONDARY AIR
SWIRL DEVICE
CORE AIR
to
o>
CO
"N-i
^1' Si J
't k\.« ^ *
IGNITION DEVICE
FUEL SUPPLY
TERTIARY AIR
VALVE
TERTIARY AIR
NOZZLE
SECONDARY AIR
Figure 1. Staged Mixing Burner (SM-Burner).
-------
EXTERNAL RECIRC
to
o»
I*
STAGING
AIR
r
INTERNAL RECIRC
BURNOUT ZONE
PRIMARY ZONE
Figure 2. SM-Burner, Schematic of Flow and Mixing Regime.
-------
j HIGH ^ LOW
tertiary air
tertiary air
seondaiy air
Secondary air
transport -^2
air* coal _
transport
air * coal
annular recirculation zone
closed recirculation zone
Figure 3. SM-Burner, Flowpattern Depending on Secondary Air/Transport Air (S/T) Ratio.
-------
14
%0-
12
0.5
%co
0,4-
0.3
0.2-
0.1 -
0
10 —
I
np=
I11
T
V
\
0,75 J
m
Op=0(5
\co
\
5
\
\
\
\
\
I
\
V
>
X..
V
===^=;
¦A
V
V/
/
/
\
\
\
\
I
*
yw
\
\
\
\
0,5
1.5
2.5
3.5
<~,5
5 (m) 5.5
4. Axial CO and O2 Concentration Profiles as Function of Primary Zone Stoichiometry
-------
oc
<
POSITION TERTIARY AIRVALVE
controled -4-
TRANSPORT AIR
BURNER LOAD
Figure 5. Air Flow Distribution as Function of Burner Load (Control
Procedure).
267
-------
1000
900
800
100% LOAD
700
60% LOAD
6Q0
500
^00
300
200
100
0,75
0,55
0
1.0
0,2
1,2
0
0,8
0,6
TRANSPORT » SECONDARY AIR
nP* STOICHIOMETRIC REQUIREMENT
Figure 6. Optimized Pilot Burner, NO-E miss ion as Function of Primary
Zone Stoichiometry (np).
268
-------
Figure 7. Variation of Coal Nitrogen Content with
269
Time.
-------
too
500-
• operating conditions
CM
O
680 MW
350
e
a
»
300
HWUf™) Qfliptj »j|m tf;
660 to 660 MW
twirl 1,5
x
O
Z
200
1,0
0,7
0,0
0,9
1.2
1,3
1,1
n _ transport- air + secondary air
primary stoichiometric air requirement
0/ content of flue gasIVOP/J
<~-4.5
4,5-5
5*5,5
5,5-6
6-7
7-9
symbols
O
®
O
•
~
A
Figure 8. Retrofit Burner, NO* Emission as Function of Primary Zone
Stoichiometry (np).
270
-------
800
700
600
500
400
300
200
100
0
1
jA
•7
/
0)
i
500 1000 1500
STEAMFLOW [t/h]
2000
Figure 9. NOx - Emission as Function of Boiler Load (Unstaged).
271
-------
400
375
E
cx
>
INJ
o
ox
NO
X
o
350
325
300
c
)
0 MW
65
0 12 3 4 5
SECONDARY AIR SWIRL —
Figure 10. Retrofit Burner, Effect of Swirl on NOx Emission.
272
-------
600
680 MW
n= 1,35
SWIRL 1,5
500
a.
ld
;• 9
300
0,9
0.8
0.7
nprimar
Figure 11. Retrofit Burner, Effect of Coal Fineness on NOx Emission.
-------
1600
1500
STEAM PRODUCED
2060 Wh
UNSTAGED
1400
STAGED
STAGED
IGNITION
ZONE
1300
UNSTAGED
V_J
AXIAL
DISTRIBUTION
1200
ce
1100
or
1000
66
27
33
21
58
RADIANT CHAMBER HEIGHT
Figure 12. Retrofit Burner, Ignition Zone and Axial Temperature Distribution.
-------
BURNER
MEASURING POSITIONS
© ©
CO-CONCENTRATION
STAGED
-a- UNSTAGED
l vpm
vpm
I vpm ]
Figure 13. Retrofit Burner, In-Flame CO - Concentration Measurements.
-------
6-
— %
^ 5
cc
i-u
h- <*¦
1—
<
2:
B-
~
LU
§ .
CD
21
o
2 '
ZD
fi'
r 1
1 1
J 1
i
I
i
i
l—.
i
• i
1
1
I
1
1
—n
i
i
i
i i
i< i
i i—
-j
i
L_r~
1 1-
1
L..J "J
1
Ln
Pn
1
l_t"
mm.m.m.ym ' 1 I ¦ » ¦ 1 t* ' » 1 A 1 1 • 1 * >
Juti Juli Oklt Nov 0«i. Jan. f*b rton April Mai Juni Julj August
1978 | 1979
Figure 14. Uncombusted Carbonaceous Matter in Fly Ash, Variation
with Time.
276
-------
standard 19
scatter^ }
area of
400 500
generator load
600 ( m w) 700
DATA-COLLECTION
1979 (MARCH/
APRIL)
1980 (FEBRUARY)
Figure 15. Retrofit Burner, Long Term Fluegas NOx Data as Function of Boiler Load.
-------
TECHNICAL REPORT DATA
(Please read Inuructions on the reverse before completing)
1. REPORT NO. 2.
3. RECIPIENT'S ACCESSION^ NO.
4. TITLE AND SUBTITLE
Proceedings of the Joint Symposium on Stationary
Combustion NOx Control. Vol. 1. Utility Boiler NOx
Control by Combustion Modification
6. REPORT DATE
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Symposium Cochairmen: Robert E. Hall (EPA) and
J.E. Cichanowicz (EPRI)
8. PERFORMING ORGANIZATION REPORT NO.
IERL-RTP-1083
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 PERIOD COVERED
Proceedings; 10/6-9/80
14. SPONSORING AGENCY CODE
EPA/600/13
is. supplementary notes EPA-600/7-79-050a through -050e describe the previous sympo-
sium. IERL-RTP project officer is R.E. Hall, Mail Drop 65, 919/541-2477.
16 abstract proceedings document the approximately 50 presentations made during
the symposium, October 6-9, 1980, in Denver, CO. The symposium was sponsored
by the Combustion Research Branch of EPA's Industrial Environmental Research
Laboratory, Research Triangle Park, NC, and the Electric Power Research Institute
(EPRI), Palo Alto, CA. Main topics included utility boiler field tests; NOx flue gas
treatment; advanced combustion processes; environmental assessments; industrial,
commercial, and residential combustion sources; and fundamental combustion re-
search. This volume relates to the use of combustion modification to control NOx
emissions from utility boilers.
17. KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
b.IDENTIFIERS/OPEN ended terms
c. COSATI Field/Group
Pollution Flue Gases
Combustion Engines
Nitrogen Oxides
Boilers
Tests
Assessments
Pollution Control
Stationary Sources
Environmental Assess-
ment
13B
2IB 21K
07B
13A
14B
18. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
282
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (»-73) 278
------- |