EPA-R2-73-292-B
June 1974
Environmental Protection Technology Seri
es
•w&iijiii^
-------
EPA-R2-73-292-B
PACKAGE BOILER FLAME MODIFICATIONS
FOR REDUCING NITRIC OXIDE EMISSIONS •
PHASE II OF III
by
L. J. Muzio, R. P. Wilson, Jr., and C. ?'cComis
Environmental and Applied Sciences Division
Ultrasystems, Inc.
2400 Michelson Drive
Irvine, California 92664
Contract No. 68-02-0222
ROAP No. 21ADG-43
Program Element No. LAB014
EPA Project Officer: G. B. Martin
Control Systems Laboratory
National Environmental Research Center
Research Triangle Park, North Carolina 27711
Prepared for
AMERICAN PETROLEUM INSTITUTE
1801 K Street, N. W.
WASHINGTON, D.C. 20006
and
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C 20460
June 1974
-------
This report has been reviewed by the Environmental Protection Agency
and approved for publication. Approval does not signify that the
contents necessarily reflect the views and policies of the Agency,
nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.
-------
TABLE OF CONTENTS
.SUCTION Page
I. SUMMARY, CONCLUSIONS AND RECOMMENDATIONS .... 1
II. INTRODUCTION 7
III. EXPERIMENTAL FACILITY 8
IV. TEST MATRIX AND BASELINE EMISSIONS 14
A. Matrix of Flame Modification Tests 14
B. Baseline Performance 17
V. EFFECTS OF BURNER MODIFICATIONS 20
A. Summary 20
B. Primary/Secondary Air Ratio 21
C. Swirl 21
D. No. 6 Oil Temperature 25
E. Atomization Air Pressure 25
VI. EFFECTS OF FLUE GAS RECIRCULATION 30
A. Summary 30
B. Effect of Fuel Type on FGR Effectiveness 31
C. Effect of Location of FGR Delivery 32
D. Effect of FGR on Heat Flux Distribution 36
E. FGR With No. 2 Fuel Oil and Natural Gas 41
VII. EFFECTS OF STAGED COMBUSTION 45
A. Summary 45
B. Interpretation for Oil Firing 45
C. Axial Boom Injection Results , 48
D. Sidewall Staging Results - Effect of Injection Point
and Angle 48
E. Sidewall Staging Results - High Excess Air 54
F. Sidewall Staging Results - Heat Flux Distribution .... 54
VIII. EFFECTS OF COMBINED FGR AND STAGED COMBUSTION ... 60
IX. EFFECT OF REFRACTORY SLEEVES ON NO AND SMOKE
CONTROL 62
REFERENCES 70
TABLE OF CONVERSION FACTORS 72
iii
-------
ACKNOWLEDGEMENTS
This project has been sponsored by the Environmental Protection Agency,
the American Petroleum Institute/ and Foster Wheeler Corporation.
Throughout this research project* we have received valuable guidance
and suggestions from the Environmental Protection Agency and American Petro-
leum Institute task force. The following individuals have been of service in
this regard and their contributions are greatly appreciated:
G. B. Martin Environmental Protection Agency
D. Lachapelle Environmental Protection Agency
D. E. Glass Shell Oil Company
C. W. Siegmund ESSO Research and Engineering Company
J. H. Weiland Texaco, Inc.
A. R. Rescorla American Petroleum Institute
In addition, we greatly appreciate the suggestions and comprehensive
review of our work provided by the Foster Wheeler Corporation through Messrs.
R. E. Sommerlad and R. Zoshak.
iv
-------
I. SUMMARY. CONCLUSIONS. AND RECOMMENDATIONS
During Phase II of this three-phase program, tests with a 3.7 x 10
Btu/hr multifuel combustion facility were conducted in order to develop NO
JC
control techniques for oil-fired package boilers. Both single burner retrofits
and suggestions for factory redesign are sought to control emissions from
c e
commercial (3 to 30 x 10 Btu/hr) and industrial (30 to 400 x 10 Btu/hr)
package boilers, which together contribute 16% of the nitric oxide produced
by stationary combustion sources . Since NO emissions are more readily
X
controlled by flame modifications than by processing stack gases, three con-
trol methods were explored: (1) flue gas recirculation, (2) staged combustion,
and (3) burner modifications. These categories each contained a wide variety
of variables as shown below:
EXPERIMENTAL VARIABLES
(All tests over range of excess air and load)
Flue Gas Recirculation Burner Modifications
Level of FGR Primary/Secondary Air Ratio
Location of FGR Addition: S^f1^6^!^ c « o ™i *T * , ^ \
„. .jo.. Fuel Type (No. 6, No. 2 Oil; Natural Gas)
Primary Air Stream ^-t •* +
. ,, „ Oil Temperature
Secondary Air Stream __ . f. .. _
_ ^ . _ . ^ ... Atomization Air Pressure
Total Combustion Air _ . ... . , _. , ,
_ Inert Atomizing Fluid
Gas Ports
Quart Injectors Staged Combustion
Combinations Level Qf gtaging
Combined FGR and SC ^cation of Staged Air Addition:
Vary Injection Location Rear Injection Boom
Vary Injection Level Sidewall Injectors
SC w/Refractory Liners for Smoke Control
-------
The experimental tests (approximately 1400) conducted on this
laboratory combustor incorporating a modified Ray oil burner allow the follow-
ing major conclusions to be drawn:
1. Simple burner modifications to atomization air pressure
or primary/secondary air ratio can result in 20% reduc-
tions in NO.
(2) (3)
2. Unlike prior data of Heap et al. and Wasser et al.,
a factor of 4 change in swirl affected NO less than 10%.
3. NO reductions of 45% without increased soot were avail-
able with FGR, provided the flue gas added to that por-
tion of air which mixes with the fuel in the near vicinity
of the burner.
4. Staged combustion was limited as an NO control tech-
nique in this combustor due to a direct tradeoff in smoke.
Approximately 25% reductions in NO could only be rea-
lized before smoke emissions became excessive. Inade-
quate mixing in the near vicinity of the burner is suggested
as the limiting factor.
5. The flame in this combustor is "naturally staged" by the
secondary air with NO formation principally controlled by
the aerodynamics of the flowfield.
6. Combined flue gas recirculation and staged combustion
resulted in substantial NO reductions (up to 50%) with
only moderate increases in smoke.
An expanded discussion of our results with each of these control techniques
follows.
Burner Modifications
Of the various burner modifications tested, the emissions were most
sensitive to the primary-to-secondary air ratio, with 20% reductions in NO
possible by lowering the P/S ratio. Increasing the atomization air pressure
was also found to reduce the NO emissions on the order of 20% with No. 6
oil firing. These reductions were attributed to a retarded rate of fuel/air
mixing (essentially "staging" the combustion internally) due to the follow-
ing aerodynamic changes in the burner flowfield:
-------
• Shifting air from primary to secondary presumably
places this air flow near the walls of the combustor
where it mixes with the fuel at a slower rate.
• Increased atomization pressure collapses the oil
spray pattern from a wide cone to a more cylindrical
form. This initial maldistribution, for a given mix-
ing intensity, presumably results in delayed mixing
of the fuel initially placed at the jet centerline.
The other burner variables such as swirl and oil temperature were found to
have a small effect (less than 10%) on the emissions from the combustor.
Flue Gas Recirculation
The effectiveness of flue gas recirculation as an emission control
technique was found to be dependent on the level of recirculation, the loca-
tion at which the flue gas is added to the burner, and also the type of fuel
being burned. The results indicated that the largest increment in NO reduc-
tion is accomplished with the first 10% recirculated. This diminishing-retums
(4 5)
effect has been observed by other investigators ' . Also, it was found that
the flue gas must be added such that it intimately mixes with the air in the
near vicinity of the burner in order to be effective. For instance, adding the
FGR to the secondary air stream or through quart injectors had no effect on
the NO emissions from the combustor, whereas reductions in NO on the order
of 30% could be obtained during No. 6 oil firing when the flue gas was added
to the total air, the primary air stream, or through the natural gas manifold.
Also, NO reductions with FGR were accomplished with little tradeoff in smoke
emissions. Flame instabilities or flame blowout were not observed with any
of the configurations tested, with levels of FGR up to 40%*. Regarding fuel
type, the greatest percent reductions in NO were obtained with natural gas,
then No. 2 oil, and No. 6 oil respectively. The differences due to fuel type
are attributed to the nitrogen content of the fuels (the fuel-bound nitrogen
conversion to NO is not temperature-sensitive and thus not very controllable
by FGR).
m
FGR(%) = -
m, , + m .
fuel air
-------
Staged Combustion
Staged combustion was not found to be as attractive as flue gas rccir-
culation in this bumer/combustor system due to increases in smoke emissions
during staging. Nitric oxide reductions of 25% during No. 6 oil firing could
not be obtained without the smoke levels exceeding a number 8 Bacharach smoke
number, regardless of the staging configuration. In addition, fora given stag-
ing configuration, the NO emissions were not very sensitive to the amount of
staged air. It is not certain whether the increase in smoke was predominantly
due to the overall fuel-rich condition near the burner during staging or due to
prolonged lifetime of local rich regions near the burner due to the reduced
velocities and less vigorous mixing.
Combined Methods
Combined staged combustion and flue gas recirculation was found to
be an attractive method of reducing NO emissions while still maintaining
acceptable smoke performance. With 25% of the air staged through the rear
boom and 24% FGR, the NO emissions could be reduced by 45% with only a
modest increase in smoke (2 Bacharach smoke numbers above baseline).
A question which still remains is the general applicability of the
results in this facility. Since the study was conducted with one particular
manufacturer's burner, it remains to be demonstrated that the results are
widely applicable to other bumer/combustor systems.
. Recommendations
It is recommended that a fire tube and a water tube boiler operating
in the field be modified and tested in order to assess the effectiveness of the
control techniques uncovered during Phase II. The fire tube or scotch marine
boiler would represent a lower size range, close in characteristics to the
.laboratory combustor, and provide the easiest transition from the laboratory
to the field. The water tube boiler in the larger size range would provide
information as to the scale over which the techniques are applicable.
-------
It is also recommended that further experiments be conducted on
the laboratory combustor in direct support of the field testing. For instance,
if the fuel delivery system of the fire tube field unit is different than the
laboratory combustor (e.g., steam or pressure atomized, different nozzle
pattern) then it would be prudent to set up a similar fuel delivery system in
the laboratory for preliminary and parallel studies. These tests would aid
in rationalizing the field data and would also help in planning modifications
to the field units. Similar studies may be warranted with the air delivery
system, quarl shape, or perhaps refractory located in portions of the combustor.
-------
-------
II. INTRODUCTION
The effectiveness of flame modifications in reducing NO emissions
x
from package boilers is presently being investigated in a three-phase program.
During Phase I, a versatile oil-fired facility was designed and constructed to
serve as an experimental arena for investigating control techniques (a sketch
of the facility is shown in Figure 1). The present report documents the work of
Phase II which utilized the experimental facility in a systematic investiga-
tion of burner modifications, flue gas recirculation, and staged combustion
as methods of emission control for oil-fired package boilers. During the test-
ing , control technique s were sought that require both minor modifications
(applicable to existing boilers) and extensive modifications (primarily applica-
ble to a redesign of new boilers). Having determined promising emission
control techniques during the activities of Phase II, we recommend that these
techniques be applied to actual field operating boilers during Phase III.
Oil Supply
Suction
Figure 1
SKETCH OF COMBUSTOR FACILITY
7
-------
III. EXPERIMENTAL FACILITY
As discussed in detail in References 1 and 6, the experimental
facility was designed to provide an authentic package boiler flame, versatile
enough for testing, and heavily instrumented. Without moving the baseline
design away from that of a typical package boiler, the facility provides the
experimenter with as many options as possible both in terms of standard para-
meters (e.g., fuel type, atomization type, load, excess air), and especially
in the modes of external flue gas recirculation and staged combustion. The
nominal specifications of the facility are given below:
Load: 3.7x10 BtuAr 5 3
Combustion Intensity: 1.7 x 10 Btu/hr-ft
Combustor L/D: 3.9
Wall Temperatures: 450 F
Fuel: No. 6 oil, No. 2 oil, natural gas
The furnace is composed of three cylindrical modules, each 30" long;
this then allows changes in both combustor volume and L/D simply by remov-
ing one of the furnace modules. In addition, an 11" throat module connects
the burner to the furnace and provides a portion of the refractory burner cone.
Downstream of the throat, the steel walls of the three furnace modules are
cooled to 250 to 450°F. Refractory may be added to the furnace as cylindri-
cal inserts.
During burner operation, the coolant distribution system provides a
continuous flow of liquid coolant to the flue gas convective section and to the
combustor wall cooling jacket. Dowtherm was selected as the cooling fluid
to avoid the high pressures associated with 400 F wall temperatures when
water is used.
A commercial 9.0 HP, dual-fired Ray burner was selected as the skele-
ton for a the research burner. Designed for application in multi-pass scotch mar
ine, water tube or firebox boilers, the burner is set up for low-pressure air atom-
ization of oil and forced-draft operation. The adaptation of this burner to a con-
figuration amenable for research included the following changes:
8
-------
(i) Physical separation of primary and secondary air
supplies for independent metering
(ii) Provision for variable swirl rate
(iii) Provision for variable fuel/air ratio independent
of load.
A sketch of the modified burner is shown in Figure 2.
Primary Air Manifold
Fuel Gas Tubes
Primary Air Tube
Secondary Air Chamber
Oil Inlet
Oil Nozzle
Natural Gas
Manifold
Gas Supply Duct
Ray Air Control Valve
(Swirl Level Control)
Secondary Air Inlet Duct
Figure 2
MODIFIED BURNER
The effect of swirl on the emissions from the combustor was investi-
gated in the course of the burner variation studies. The swirl is varied by
varying the inlet velocity of the secondary air through the secondary air control
valve (Figure 2). Since the secondary air enters the burner tangentially, this
velocity determines the angular momentum at the throat which can then be
expressed in terms of a swirl parameter for the burner.
-------
In calculating a swirl parameter for the burner, a swirl parameter is
calculated independently for the secondary and primary (the swirl parameter
for the primary is zero for this burner) and the two numbers are combined
linearly using the flow distribution as a weighting factor. (Details of the
assignment of a swirl parameter for the burner are given in Reference 6 .)
The overall air and flue gas distribution system is presented in sche-
matic form in Figure 3. Basically, eight supply points can be fed up to 1000
and 500 cfm of air and flue gas, respectively, from centrifugal blowers at
elevated pressure (1 psig).
The eight supply points include seven injection points for air, and
four for flue
1.X UC ^Uv>.
Injection Point
(i) Burner primary
(ii) Burner fuel gas ports
(ill) Burner secondary
(iv) Throat injectors
(v)-(vii) Sidewall injectors
Provision
for Flue Gas
X
X
X
X
~
(viii) Axial injection from rear
Provision
for Air
X
—
X
X
XXX
X
combustor:
Staged combustion can be accomplished in two ways in the present
**
Sidewall:
Air addition through downstream Sidewall injectors
(three locations: 20-3/8", 30-3/8", and 50-3/8"
from the oil nozzle, SI, S2, S3 respectively)
Axial Boom:
Air addition through a central boom in direction counter
to or at right angles to the main hot gas stream (rear
boom position can be continuously varied along the
axis of the combustor).
*This burner swirl parameter is based on an idealization of the inlet angular
momentum to the primary and secondary windboxes and should not be inter-
preted in the classical sense as a swirl number.
**These methods involve explicit downstream air injection. In addition, certain
burner modifications probably result in natural "internal" staging of the com-
bustion process which may be simply reduced mixing rates.
10
-------
^
Wall
'&.
i
I
I
I
From Convective
Section (6" dia)
Filter
/ Throat \
Manifold \
Boost Blower
4 Injection
Ports per
Manifold
Flexible
Hose
Water Cooled
Rear Injection
Lance
+
Main
Stack
Figure 3
AIR AND FLUE GAS DISTRIBUTION SYSTEM
11
-------
A deflector is mounted on the end of the boom to disperse the staged air in
a radial direction (see Figure 4 below). Visual observations were made
of the interaction between the main combustor air and the staged air by adding
Boron Trifluoride to the rear boom air. These observations indicated that
over the range of flows tested, the staged air was dispersed throughout
the cross sectional area of the combustor.
Woler_J
-------
Table 1
FUEL OIL CHARACTERISTICS
No. 6 Oil No. 2 Oil
Gravity, °API 16.7 35.0*
Flash Point, PMCC °F 265 170*
Pour Point, °F 80 -15*
Viscosity, SSF at 122°F, sec 97 35*
Heat of combustion, gross, Btu/lb 17,746 19,242*
Water and sediment, % 0.08 0.00*
Ash, % 0.02
Sulfur, % 0.42 0.24
Nitrogen, KJeldahl, % 0.36 0.05
Carbon, % 87.68 86.21
Hydrogen, % n-61 12-68
*Typical values, Ref. 7
13
-------
IV. TEST MATRIX AND BASELINE EMISSIONS
A. MATRIX OF FLAME-MODIFICATION TESTS
'•)-;'•..'.•.
All combustion modification tests were run on No. 6 fuel oil at two
firing rates and two excess air levels, with all other burner parameters held
constant. This 4-point operating matrix was useful for the following reason:
Since package boilers are fired under widely varying load and excess air, a
technique which controls emissions at condition A may produce excessive
smoke when applied at condition B. However, since there were so many
techniques of applying FGR and SC to the combustor, it was impractical to
attempt to test each configuration technique under all boiler operating condi-
tions. It was necessary to select a limited (4-point) set of representative
operating conditions, wide enough to establish a reasonable level of confi-
dence that each technique would be applicable over a range.
In addition, a limited number of experiments were performed using the
low-nitrogen No. 2 oil. This allowed a rough assessment of the nitric oxide
arising from nitrogen fixation and from fuel-bound nitrogen.
The testing was performed in four main blocks according to the test
matrix of Table 2: burner modifications, flue gas recirculation (FGR), staged
combustion (SC), and combined FGR and SC.
The first tests were run making simple burner changes; variations
included:
Swirl level
Primary/secondary air ratio
Oil temperature
Oil atomization air pressure
14
-------
Table 2
TEST MATRIX
Baseline Conditions:
iLoad (3.4 x 10 Btu/hr, 2 x 10 Btu/hr) I Standard 4-point
(Excess Air (17%, 35%) I matrix for all tests
Air Distribution (Primary/Secondary - 50%/50%)
Oil Temperature (200°F)
Atomization Air Pressure (18 psig)
Combustor (L/D = 4)
Coolant Inlet Temperature (250
Fuel (No. 6 Oil)
'F)
Swirl (Swirl Parameter =1.8)
(a)
Focus of Test
Test Conditions
BURNER VARIATIONS
Primary/secondary - 30/70, 40/60, 60/40, 70/3A
Secondary swirl - low, high (swirl no..76, 3.8)tD'
Oil temperature: 170, 180, 190°F
Atnmination air pressure: 10 to 36
FLUE GAS RECIRCULATION
(c)
3 recycle ratios - 10%, 20%, 30% '
5 injection positions - primary, secondary, throat,
gas ports, and total
STAGED COMBUSTION
Three Levels of Staging: 17%, 25%, 35%
Three Sidewall Injection Stations, each with
4 orientations (upstream, downstream,
co-swirl and counterswirl)
Rear Boom Injection with variable position
COMBINED STAGING
AND RECIRCULATION
External recirculation through the total air stream,
combined with 3 types of staged combustion:
2 levels of FGR(10%, 30%)
2 levels of staging (17%, 25%)
INTERSPERSED
STANDARDIZATION
Performed at random to detect any systematic
drift in combustor behavior and to test No. 2 oil
and natural gas.
(a) Secondary air control valve opened 1-3/8" (Ref. 1, also see Fig. 2)
(b) For these tests, the primary/secondary air ratio was also varied (primary = 40%, 50%,
60%) (i.e., 12-point baseline matrix instead of the usual 4-point matrix)
(c) FGR(%) = 100 (
15
-------
In general, these "front-end" modifications call for much less severe hard-
ware changes than the furnace modifications to be described in the following
sections, and therefore are logical candidates for quick field modification.
The effectiveness of these burner-oriented techniques relies on the well-
known sensitivity of emissions to near-burner mixing patterns. They were
tried first because of their simplicity. The burner results, in addition to
establishing control data for comparison with combustion modification, were
repeated periodically throughout the test program to verify standardization of
the furnace. Any systematic drift in furnace behavior such as soot accumu-
lation could be readily detected by these repeat calibration tests.
The next combustion modification series utilized flue gas recircula-
tion. Three injection schemes were used, listed in decreasing order of esti-
mated rate of flame dilution and cooling:
1. Conventional annular injection (primary, secondary passages)
2. Injection through fuel gas ports
3. Quarl throat exit injection
The percent of flue gas recirculation was varied for each injection scheme.
Tests of staged combustion followed the FGR testing. Variations in
injection schemes included the following five locations of introducing air
for delayed combustion:
1. Quarl throat injection
2-4. Downstream wall Jets Sj, S2, S_ (Figure 3)
3. Downstream axial countershower
These concepts were tested singly and in combinations.
Temperature sensors are located in the coolant stream along the length
of the combustor. The incremental temperature rise along the combustor can
then be used to monitor the distribution heat transfer from the combustion
products along the combustor walls '. This then provides a tool for deter-
mining whether changes in emissions are due to (1) significant changes in
16
-------
mixing patterns which affect the temperature of the combustion products and
thus the heat flux to the walls, or (2) local changes in mixing and combustion
patterns which, while significant to NO and smoke formation, do not affect
the overall thermal performance of the combustor.
B. BASELINE PERFORMANCE
The combustor was first operated over the baseline conditions out-
lined in Table 2. Further baseline tests were also conducted at an inter-
mediate load setting of 2.8 x 10 BtuAr. The emissions of NO, CO, and
smoke for these conditions are presented in Table 3 for firing with both No.
6 and.No. 2 oil. Throughout the test program the CO emissions were always
low and fairly insensitive to the combustor modifications. In seeking low
NO configurations, the main tradeoff was in the smoke emissions (the CO
levels never increased substantially until after an excessive smoke level
was reached). Thus only the smoke and NO emissions results are presented
hereafter. Note NO reduced to 3% O_.
X £f
Table 3
BASELINE EMISSIONS
Combustor Exit
Excess Load NO Smoke CO Gas Temperature
Fuel Air(%) CIO6 Btu/hr) ppm(3%O ) (Bacharach) ppm(3%O ) (°F)
No. 6
Oil
No. 2
Oil
17
17
17
35
35
35
17
17
17
35
35
35
3.4
2.8
2.0
3.4
2.8
2.0
2.5
3.5
2.0
2.5
2.0
3.5
273
235
214
288
286
274
87
91
114
118
4
7
7
2
3
4
.5
-
0
1.5
0
-
28
23
27
27
20
24
9
—
17
10
10
—
1560
1516
1359
1562
1587
1397
1606
—
1603
1448
1510
—
17
-------
For comparison with emissions from field-operated boilers, a few
data points are plotted in Figure 5 along with a curve of typical commercial
(8)
boiler emissions obtained during a field study by Battelle. The perfor-
mance of the experimental facility in terms of emissions of NO and smoke
can be seen to be similar to those of actual field boilers.
As seen in Figure 6, with a change in load from 3.4x10 to 2.Ox
10 Btu/hr, the heat flux from the combustion products to the coolant only
drops by 20%. Further, the heat flux distribution along the combustor changes
at the lower loads with a greater percentage of the heat flux taking place near
the combustor entrance. This is probably due to the reduced velocity and
longer residence times of the gases at the lower load setting.
18
-------
0.5
fi
I 0.3
•
M
E
g*0.2
0.1
Ultras/stems Combustor
• No. 6 Oil Fired (80% Load)
• No. 2 Oil Fired (80% Load)
• Ref. 8
(83% Load
9
8
7
o 6
i.
to
° 4
2
i
0
Ultras/stems Combustor
• No. 6 Oil Fired
• No. 2 Oil Fired
Ret 8 (83% Load)
01020304050607080
Excass Air, %
w Figure 5
TYPICAL EMISSIONS FROM COMMERCIAL BOILERS
Fuel: No. 6 Oil
Excess Air: \7%
P/SAIr: 50%/50%
3.4 x I06 Btu/hr
2.0 x 106 Btu/hr
Entrance
Combustor Axis
Figure 6
Total Coolant
Temperature Rise
Exit
EFFECT OF LOAD ON THE COOLANT
TEMPERATURE RISE ALONG THE COMBUSTOR
Low sulfur residual oil (S = 1.0%, N = 22%, API gravity = 23)
19
10 20 30 40 50 60 70
Excess Air, %
(b)
-------
V. EFFECTS OF BURNER MODIFICATIONS
A. SUMMARY
Using the baseline tests as a reference, a series of burner variation
experiments were performed. These entailed fairly minor modifications to the
system and thus are suitable for field conversion of existing units. Results
of the burner variation studies have shown two variables (exclusive of the load
to affect emissions by more than 20%: primary-to-secondary air ratio and
atomization air pressure. A summary of these tests is presented in Table 4.
Table 4
BURNER VARIATION SUMMARY
Change in Smoke Emission!
Variable Reduction in NO (Bacharach Smoke Numbers
Load 6 22% +2
(3.4 to 2x10 Btu/hr)
Excess Air 18% +4
(35% to 17%)
Primary/Secondary Air Ratio 28% +7
(70/30 to 40/60)
Swirl Level 13% +1
(5.6 to 0.7, swirl parameter)
Atomization Air Pressure
(10 to 20 psi) 4% +1/2
(20 to 36 psi) 21% +1/2
Oil Temperature 2% +1
(170°F to 200°F)
Atomization Fluid 3% -1/2
(Air -• Nitrogen)
These burner variation tests were performed primarily with No. 6 oil,
with a limited number of tests performed with No. 2 distillate oil. In all
cases, the burner and combustor variables were held at the baseline condi-
tion except for the particular variable under investigation (e.g., primary/
secondary air ratio, swirl, oil atomization air pressure. No. 6 oil temperature,
oil atomization fluid).
20
-------
B. PRIMARY/SECONDARY AIR RATIO
The distribution of air between the primary and secondary passageways
has a significant effect on the emission of nitric oxide and smoke from the
combustor. As shown in Figure 7', increasing the primary air from 40% to 70%
increases the NO emissions by 28% and decreases smoke from a number 9 +
(off scale) to a 2 Bacharach number. The increased NO and reduced smoke
emissions with increasing primary air may result from net higher mean tempera-
tures in the combustor and/or increased mixing between the fuel spray and
primary air stream due to the increased primary air velocity.
As the primary-to-secondary air ratio is increased, the overall heat
flux to the coolant is decreased 6%, resulting in higher mean temperatures
in the combustor; however one sees very similar temperature distributions for
the two different settings (see Figure 8). This decrease in heat flux was also
reflected in a 3-1/2% increase in the combustor exit temperature of the com-
bustion products (1617°F at P/S = 70%/30% vs. 1560°F at P/S = 50%/50%).
For comparative purposes, the effect of the primary-to-secondary ratio
was investigated while firing distillate No. 2 oil. As seen in Figure 9 , there
is also a marked effect on P/S ratio on both the NO and smoke emissions when
firing with No. 2 oil.
C. SWIRL
As is seen in Figure 10, changing the swirl in this burner has a
very small effect on the NO and smoke emissions from the combustor. Also,
the heat transfer profiles to the coolant for high and low swirl conditions were
virtually identical, indicating that the changes in swirl did not influence the
thermal behavior of the unit.
One explanation for the lack of sensitivity of the combustor to changes
in swirl derives from the fact that the swirl component is contained solely in
the secondary air. In the near vicinity of the nozzle the primary air dominates
mixing. Hence changing the swirl by increasing the angular momentum of the
21
-------
400
CM
o
65
CO
300 -
0)
•V*
S
200
Fuel: No. 6 Oil
Atomization Air Pressure: 20 psic
Oil Temperature: 200°F
Swirl Parameter: 1.8
3.4x 10 Btu/hr
2. Ox 10 BtuAr
A 3.4 x 10 Btu/hr
2. Ox 10 BtuAr
% Primary Air
Figure 7
EFFECT OF PRIMARY/SECONDARY AIR RATIO
ON EMISSIONS
80
*Dashed line indicates that smoke level is off scale. A smoke number of 9 is
the maximum on the Bacharach scale.
22
-------
50r-
40
a
8
u
2 30
X
S
?
g
Fuel: No. 6 Oil
Load: 3.4 x I06 Btu/hr
Excess Air: 17%
Swirl Parameter: 1.8
P/S Air: 50%/50%
P/S Air: 70%/30%
Total Coolant
Temperature Rise
„ 60°F
64°F
Entrance
Exit
Combustor Axis
Figure 8
EFFECT OF PRIMARY/SECONDARY AIR DISTRIBUTION
ON THE COOLANT TEMPERATURE RISE ALONG THE COMBUSTOR
140
120
ow
*«
ro
*•*
D
O.
ji.
i
"x
O
100
80
60
S 40
- FUEL: No. 2 Oil
4.00* 2.8 x I06 Btu/hr
FvrA« Air- 17%
Swirl Parameter:' 1.8
Atomlzation Air
PrMsura: 20 psig
80
PRIMARY AIR .
-,10
Figure 9
EFFECT OF PRIMARY/SECONDARY AIR RATIO
ON EMISSIONS
(No. 2 Oil Fired)
23
-------
350 1
§•300
£250
5
,0
«S 200
Z
ISO
0
Excess Air {.oad., |Q6 Btu/hr
• .,« I 3.4
• I7X f 2.0
A ,_ 1 3.4
„ • 35X / 2.0
+ * .
••
^
A A
1 1 1 1 1 1 III
Fuel: No. 6 Oil
P/S Air: 50%/50
Oil Temperature
Atomization Air
Pressure: 20
n
0 N Jk 0 OD C
Smoke No. (Bacharadl)
0. I 0.2 0.4 0.6 0.8 1.0 2
Swirl Parameter
Figure 10
EFFECT OF SWIRL ON EMISSIONS
e 10
24
-------
secondary air may not affect the mixing patterns between the fuel spray and
the air. One might expect different results if the same "swirl parameter" had
been obtained by varying the angular momentum in the primary air, with an
axially directed secondary air flow.
D. NO. 6 OIL TEMPERATURE
The effect of oil temperature with No. 6 oil firing was investigated
over the range of 170 F to 200 F. Over this range, the oil viscosity changed
from 260 SUS* to 180 SUS. The results of these tests are plotted in Figure 11.
During this test, variations in NO were limited to about 5%.
E. ATOMIZATION AIR PRESSURE
1. Design Range (10 to 20 psig)
The Monarch wide angle (80-70 deg) nozzle which is used in the
burner operates with an air pressure range of approximately 10 to 20 psig.
Over this range, the atomizing air flow rate varied from 9 to 15 cfm, or
approximately 1% to 2% of the total air flow to the burner. Figure 12 presents
the results of these tests, and it is seen that there was little effect on the
emissions. No. 2 distillate oil results, which are plotted in Figure 13, are
similar.
2. Extended Pressure Range (20 to 36 psig)
As can be seen in Figure 14, when the air atomization pressure was
increased above 20 psig, the NO emissions dropped essentially linearly such
that at an atomization pressure of 36 psig the NO emissions were reduced by
21%. Although the air for these tests was supplied by house air and not from
the compressor supplied with the burner, no difference was found in the emis-
sions using either house air or air from the burner compressor for a given
atomization air flow. The smoke emissions only increased by 1 Bacharach
smoke number as the pressure was raised from 20 to 36 psig. This reduction
*Saybolt Universal Seconds
25
-------
400
_ .. •
O
tn
Ejtcess Air L
•
•
«
S
17% \
35% f
17% \
35% f
,oadr 10 Btu/hr
2.8
2.0
200
170
180 190
OIL TEMPERATURE (° F)
FUEL: No. 6 Oil
P/S Air: 50%/50%
Swirl Pai-ameter: 1.8
Atomization Air
Pressure: 20 psig
8 ^
"o
z
6l
D
S
•
4 i
o
o
21
Figure 11
EFFECT OF OIL TEMPERATURE ON EMISSIONS
26
-------
Excess Air Load. 10 Btu/hr
FUEL: No. 6011
OIL Temperature: 200°F
P/SAir: 50%/50%
Swirl Parameter: 1.8
14 15 16 17 18 19 20
ATOMIZATION PRESSURE (PSI)
Figure 12
EFFECT OF ATOMIZATION PRESSURE ON EMISSIONS (No. 6 OIL)
*,.
s '*
j,«
5 120
&
1 100
z
80
0
Excess Air Load. I06 Btu/hr
• 17% 1
• 35% f 3'4
* 17% 1
> 35% f 2'°
_
B B S — -
• "— f
.
*V^
4>--___^J^--5^^_ * H
^^^^-J.^.^^
— « — • — m — u__5:^^szi
FUEL: No. 2 Oil
P/SAIr: 50%/50%
Swirl Parameter: 1.8
— .
, 1
0 0
a
£
4 5
o
2
-------
in NO with increasing air atomization may be due to smaller droplets or
collapsed spray angle:
• Increasing atomization pressure results in a spray with
smaller drop sizes. Bowman and Kesten' / have suggested
that NO ~ d2 for drops burning in an environment of con-
stant ambient conditions. Thus the smaller drop sizes may
lead to lower NO emissions.
• With increased air atomization pressure, the spray
was observed (when sprayed in a stagnant atmos-
phere) to change from a conical form to a more
cylindrical form. This would have the effect of
delaying the mixing of fuel and air, and in effect
"staging" the combustion.
Illumination of the exact mechanism requires detailed diagnostics of the
atomization and flow characteristics which were beyond the scope of the
present program.
3. Inert Atomizing Fluid
Also shown in Figure 14 are the results that were obtained using
nitrogen as the atomization fluid. The rationale for performing this experi-
ment was that although the atomization air only comprises 1% to 2% of the
total air flow, it is intimately mixed with the oil and thus may have a sub-
stantially higher potential for NO formation. One might then expect the
elimination of oxygen from the atomizing fluid to suppress the conversion of
fuel-bound nitrogen to nitric oxide since only 0.2% of the total air flow to
the burner is required to oxidize all of the fuel-N to NO. The experiments
show that this did not seem to be the case, in that using nitrogen as the
atomizing fluid resulted in only about a 10 ppm decrease in NO emissions.
This indicates that although the atomizing fluid is initially initimately mixed
with the fuel oil, the oxidation step of fuel nitrogen to NO occurs after a
delay time such that sufficient mixing of the fuel and burner air has taken
place in the combustor. This then diminishes the importance of the oil
atomizing air in the fuel-N to NO conversion process.
28
-------
30C
2
u
£ 200
z
ISO
• Air Atomized
A Nitrogen Atomized
10
20 30
ATOMIZATION PRESSURE (PSIG)
FUEL: No. 6011
EXCESS AIR: 17%
LOAD: 3.4 X 106
BTU/HR
P/SAir: 50%/50%
OIL Temperature: 200°F
Swirl Parameter: 1.8
8
u
6 2
5
u
o
to
20
&
40
Figure 14
EFFECT OF ATOMIZATION PRESSURE
AND FLUID ON EMISSIONS
29
-------
VI. EFFECTS OF FLUE GAS RECIRCULATION
A. SUMMARY
A summary of the flue gas recirculatlon (FGR) results is presented in
Table 5. A number of observations may be made from the results presented in
this section.
• FGR is most effective in reducing emissions when
added so that the FGR comes into immediate contact
with the oil spray (e.g., through the primary duct
or gas ports).
• The effectiveness of FGR can be compromised if the
FGR intensifies the mixing between fuel and air in
the near vicinity of the burner.
• The heat transfer from the combustion products to the
coolant is markedly reduced in the first part of the
combustor due to the presence of FGR.
• FGR reduces NO emissions with relatively minor
increases in smoke.
• Nitrogen content of the-fuel limits the reductions in
NO that can be obtained.
TableS
FLUE GAS RECIRCULATION SUMMARY*
(No. 6 Oil, 17% Excess Air, 3.4x 10 Btu/nr)
NO Reduction Smoke Increase
Location of FGR with 30% FGR (Bacharach Smoke No.)
From To
Gas Ports 42% 4 9
Total Air 25% •.,-•-. 4 ' 5
Primary Air 15% 4 5
Secondary Air 04 5
Quarl Injectors 0 4 3-1/2
*FGR was added to the particular stream while the primary/secondary air ratio
was held constant.
30
-------
B. EFFECT OF FUEL TYPE ON FGR EFFECTIVENESS
FGR is much more effective with natural gas firing than with oil firing,
as shown in Figure 15. In these tests, the flue gas was added to the total air
supply, with equal amounts to the primary and secondary air. Thirty percent
recirculation results in a 67% reduction in nitric oxide under gas firing, whereas
only 25% and 37% reductions are realized with No. 6 and No. 2 fuel oils, res-
pectively.
The most reasonable explanation for this strong effect of fuel type
relates to fuel nitrogen. FGR is expected to have a much greater effect on
the NO formed via the thermal fixation of nitrogen than on the formation of
NO due to the fuel-bound nitrogen, because the recirculated flue gas acts to
reduce local temperatures and fixation is more temperature sensitive. Hence
the more fuel nitrogen, the less NO reduction expected from FGR. For the
fuel oils tested, the fuel-bound nitrogen in the No. 6 oil would contribute a
maximum of 512 ppm NO whereas the No. 2 oil would contribute 65 ppm (both
at 100% conversion and 3% O ). The results presented in Figure 15 can be
b
used to speculate that approximately 40% of the fuel-bound nitrogen in the
No. 6 oil has been converted to NO, and about 91% of the No. 2 oil's fuel
nitrogen is converted. This is based on the expectation of 67% reduction of
thermal NO at 30% FGR regardless of fuel type and with no effect of FGR on
fuel-N conversion.
Another conceivable explanation is based on differences in fuel/air
mixing with each fuel. The gas/air mixing process may be more sensitive
to FGR than the oil-spray combustion. Certainly the heavy oil vaporizes at
a slower rate such that FGR is dispersed over the entire air supply by the time
ignition of the spray core occurs.
*The NO comes from two sources: (NO)totaj = (NO)fue} + (NO)air. We expect
67% reduction in (NO)air at 30% FGR from the gas results. Instead, we
observe 20% reductions for No. 6 oil. Therefore we infer that (NO)air = 20/67
(N0)fuel for No. 6 oil, or (NO)fuel = 67/87 (NO)total. At (NO)total = 271 ppm
this means that (NO)fuel * 210 ppm or about 40% conversion. For No. 2 oil,
similar reasoning leads to (NOj^gj = 2/3 (NO)totaj « 62 ppm, or about 91%
conversion. This estimate ignores obvious differences in aerodynamics due
to fuel type or FGR.
31
-------
Load: 3.4x 10° Btu/hr
Excess Air: 17%
P/SAir: 50%/50%
Swirl Parameter: 1.8
Flue Gas Added to Total AirFi
Flue Gas Temperature: 300 F
O No. 6011 (NCb =27lpj
D No. 2011 (NCU = 94 pi
A Natural Gas ( N0rt = 64
20
% RECIRCULATION
30
Figure 15
EFFECT OF LEVEL OF FGRAND
FUEL TYPE ON NO EMISSIONS
-------
C. EFFECT OF LOCATION OF FGR DELIVERY
The point at which flue gas.enters the combustor determines the
effectiveness of FGR in reducing nitric oxide. In the present case, flue
gas can be introduced into the system at five discrete locations, only three
of which proved to be effective as summarized below:
• Most Effective (40% reduction in NO): Addition of
FGR through the natural gas ports.
• Moderately Effective (15-25% reduction in NO);
Addition of FGR to the primary air, or total air stream
or combinations which include either the primary air
stream or gas ports.
• Ineffective (no reduction in NO): Addition of FGR
through the secondary, quarl throat injectors or the
secondary plus throat.
The above points are illustrated in Figure 16. A data listing for the
FGR tests is given in Table 6. For all FGR tests, 50% of the air entered the
combustor through the primary and 50% through the secondary, hence the mass
flowrates through the primary and secondary were no longer equal when the
flue gas was added.
FGR was most effective when introduced through the natural gas ports:
17.5% FGR reduced the NO emissions by 41%. However, the NO reduction
was accomplished at the expense of increased smoke emissions (e.g., a 4
Bacharach smoke number with zero FG.R, and a greater than 9 smoke number
with 17.5% FGR). Qualitatively, when the flue gas was added through the
gas ports, visual observations showed the flame to be grossly distorted. The
flame was no longer symmetrical, with the FGR jets from the gas ports tearing
the flame into isolated sections.
FGR was less effective (up to 25% reductions) when delivered either
into the total air stream (primary plus secondary) or into just the primary.
However, beyond 14% FGR, flue gas additions to the primary air stream
resulted in an upward trend in NO emissions. This counteracting effect of
33
-------
•
1.8
0.6
0.5
Recycle
Injection Point
A Gas Ports
• Total Air
A Primary Air
H All combinations
• Throat
Secondary Air
10 20
% RECIRCULATION
Figure 16
EFFECT OF FGR INJECTION LOCATION
ON EMISSIONS
-lift
Fuel: No. 6 Oil
Load: 3.4x I06 Btu/hr
Excess Air: 17%
P/S Air: 50%/50%
Swirl Parameter: 1.3
Oil Temperature: 200°F
Atomization Air
Pressure: 20 psig
FGR Temperature: 300WF
NO Baseline: 273 ppm
tJ
-------
Table 6
CO
en
Excess
Air
17%
35%
FGR Delivery
Point
Total
Primary
Secondary
Gas Port
Quarl Injectors
Primary/Gas Port
Secondary/Gas Port
Quarl/Gas Port
Primary/Quart
Secondary/Quarl
Primary/Secondary
Total
Primary
Secondary
Gas Port
Quarl Injectors
LOAD
3.4x 106 Btu/hr
% Reclrculatlon
o
NO Smoke
(ppm)
274 4
274 4
262 2
«
10
NO Smoke
MOQ
.87 4
.85 5
.99 5
.75 5.5
1.0 4
.83 6
.81 6
.83 7
.89 5
.99 4.5
.87 4.S
.89 2.5
.83 2.5
-
.87 3
-
-
20
NO Smoke
NOQ
.80 5
.84 5
1.03 5.5
.59 >9
1.0 3.5
.83 6
,73 6.S
.81 6
.83 5.5
1.07 5. 5 '
.79 5
.87 3
.75 5
-
.74 5(16%)
1.1 3(14%)
30
NO Smoke
ij°o
.'5 5
..'58 5(24%)*
1.04 5(22.5%)
-
"
.116 6
.C2 8
,:-5 7
.ii6 5
I.(i9 4.5
.V5 S.5
.{5 5
.(7 5
1.C1 3(25%)
-
-
g
2.0x10 Btu/hr
% Reclrculatlon
0
NO Smoke
ppm)
221 2
-
-
-
—
245 3.5
-
-
-
~
10
NO Smoke
NOQ
.96 6.5
-
.99 7
— "*
.9 4
.9 4
-
.88 4
"
20
NO Smoke
N00
.91 6.5
-
1.03 7
"
1 2 6.5
.86 5.5
.93 5
- ~
.85 6
1.1 5
30
NO Smoke
NOQ
.89 7.0
""
.99 8.0
_ —
.83 6.0
.98 6.5
1.03 6.5
~
* Numbers In parenthesis indicate FGR levels other than corresponding to co umn.
-------
increased mass flow rates through the primary is probably associated with
changes in the rate of fuel/air mixing. The effect disappears at higher (35%
vs. 17%) excess air, as shown in Figure 17. Here the nitric oxide emissions
decrease monotonically as the flue gas is added through the primary with a
33% reduction occurring at 27% recirculation.
Adding the flue gas to the secondary air stream or through the throat
injectors in the quart had no effect on the nitric oxide emissions from the
combustor. This insensitivity reiterates that NO formation appears to be
X
controlled by processes occurring in the core close to the axis of the furnace.
As seen in Figure 16, multiple injection configurations are basically
no more effective than the single injection schemes. At a load of 3.4 x 10
BtuAr and 17% excess air, a maximum reduction of 38% is obtained by inject-
ing the flue gas (30% recirculation) in equal quantities into the secondary and
gas ports. However, this results in an increase in smoke from a number 4 to
a number 8 Bacharach smoke number, which presumably would be an unaccept-
able penalty in an actual field application.
At reduced load (2 x 10 BtuAr), the NO reductions are not as sub-
stantial (18% as opposed to the 30% reductions obtained at full load).
D. EFFECT OF FGR ON HEAT FLUX DISTRIBUTION
Insight into the differences in relative effectiveness of the various
configurations of adding flue gas to the combustor can be obtained by com-
paring the thermal performance of the combustor. Figure 18 shows the exit
temperatures and reduction in heat loss to the walls for various FGR delivery
points. The amount of heat transfer to the coolant decreases about 35% as
the level of FGR increases up to 20%. Possible explanations for this reduction
are as follows:
• Lower mean temperature of the gas mixture after
complete combustion.
• Lower achieved temperature due to heat release
occurring out of the combustor in the flue.
36
-------
1.1
J.O
.9
o
in
.6
.5
Fuel: No. 6011
P/SAIr: 50%/50%
Oil Temperature: 200 F
Atomlzatlon Air Pressure: 20 psig
FOR Temperature: 300°F
FOR Location: Primary Air
Load
I06 Stu/hr
3.4
3.4
2.0
Excess Air NObaseline
17% 274 ppm
35% 262 ppm
35% 245 ppm
10
8
10 20
% Recirculation
Figure 17
EFFECT OF LOAD AND EXCESS AIR
ON THE EFFECTIVENESS OF FGR
37
30
-------
1.0
0.8
1
|0,
O
0)
tS 0 4
p
•
s
ffi
o
H
0.2
170Q
2
5
2
0)
o,
I 1600
(0
10
O
1500
Fuel: No. 6 Oil
Load: 3.4 x 106 Btu/hr
Excess Air: 17%
P/SAir: 50%/50%
Oil Temperature:
Atomlzation Air Pressure: 18 psig
FGR Temperature: 300°F
200°F
Recycle Injection
• Total Air
D Secondary Air
A Primary Air
O Gas Ports
O Throat
10 20 30
% Recirculatlon
Figure 18
EFFECT OF FGR INJECTION LOCATION
ON COMBUSTOR HEAT LOAD CHARACTERISTICS
3ft
-------
• Reduced residence time for heat transfer due to
the increased flow through the combustor.
• Cool insulating layer of air or flue gas formed
between the wall and hot combustion products.
• Increased blockage of radiative flux from the
flame core due to increased soot.
The first two mechanisms above give a reduction in the exit temperature,
whereas the last three would predict an increase. Exit temperature is
observed to decrease when the FGR is added through the primary stream and
the natural gas manifold, supporting the first two hypotheses.
FGR reduces the heat flux to the coolant mainly in the first third of
the combustor with fairly minor changes in the distribution occurring in the
latter two-thirds of the combustor, as shown in Figures 19 and 20.* This
axial distribution is reasonable because the effect of FGR should be greatest
where the baseline flame is hottest (the primary).
Reduced heat loss in the first third of the combustor appears to corre-
late with reduced NO emissions, as shown in Figure 21. Adding the flue gas
to the total air supply and the primary air resulted in the greatest reductions
of both.
Adding the flue gas through the gas ports had little effect on the heat
transfer distribution, but resulted in the greatest increases in smoke emis-
sions. These clues, coupled with the visual observations of a greatly dis-
torted flame with FGR addition through the gas ports, suggest that the changes
in emissions in this case are primarily due to severe changes in the aerodyn-
amic patterns in the near vicinity of the oil nozzle and not solely due to a
thermal effect of the FGR.
*In Figures 19 to 21 only the heat flux distribution from the burner to the com-
bustor midpoint is presented in that the distribution in the back half of the
combustor is virtually identical.
39
-------
Fuel:
No. 6 Oil
.6
25
0"* 20
s
15
10
O
1
.8
Load: 3.4 x 10" Btu/hr
Excess Air. 17%
P/SAln S0%/50%
Flue Gas Reclrculatlon
Injection Point: Total Air Stream
Level of FGR
O Baseline
• 10% FGR
19% FGR
30% FGR
Total Coolant
Temperature Rise
65°F
53°F '
««F
40°F
I
Entrance
Midpoint
25
No. 6 Oil
2.0 x 106 BtuAr
Access Air: 17%
?/SAln 50%/50%
Tlue Gas Recirculation
'nJecUon Point: Total Air Stream
Level of FGR
O Baseline
• 10% FGR
' A 17% FGR
• 29% FGR
T Total Coolant
'emperature Rise
S6°F
37°F
33°T
29°F
J
Nc. 6 Cil
.6
Fuel:
Load: 3.4 x 106 Btu/hr
Combustor Axis
Entrance
Com buster Axis
Midpoint
Excess Air: 17%
P/SAlr: 50%/50%
20% Flue Gas Recirculation
FGR Injection Location
O Baseline
Gas Forts
Secondary
Throat
Primary
Total
Entrance Midpoint
Combustor Axis
Figure 19
EFFECT OF FGR LEVEL
ON COMBUSTOR HEAT FLUX
DISTRIBUTION (HIGH LOAD)
Figure 20
EFFECT OF FGR LEVEL
ON COMBUSTOR HEAT FLUX
DISTRIBUTION (LOW LOAD)
Figure 21
EFFECT OF FGR INJECTION
LOCATION ON COMBUSTOR
HEAT FLUX DISTRIBUTION
-------
The series of tests reported in Figure 22 were designed to add FGR
to the burner while minimizing the flowfield changes. Flue gas was added
to the primary or secondary, keeping the mass flow rates between primary
and secondary equal. For example, as recirculated flue gas was added to
the primary, an equal amount of air flow was rerouted through the secondary.
This presumably kept the flowfield and relative mixing rates of primary,
secondary, and fuel spray essentially unchanged (although the overall gas
velocity through the burner increased due to the addition of the flue gas).
Unavoidably, these tests involved not only FGR but also P/S ratio as varia-
bles. With a 50%/50% primary/secondary ratio, 26% reduction in NO was
obtained by adding 30% flue gas to the primary air stream. This reduction in
NO is greater than that obtained when the flue gas was added to the primary
without compensating for the air (Figure 16). This is due to the two factors
mentioned above: (1) presence of the flue gas, and (2) reduced primary-to-
secondary air ratio at the burner. When the flue gas was added to the secon-
dary, the NO emissions increased above the baseline conditions, probably
due to the increased air flow through the primary air stream. An increase in
the P/S ratio was shown to increase NO (see Figure 7).
E. FGR WITH NO. 2 FUEL OIL AND NATURAL GAS
The No. 2 fuel oil responded to FGR in a similar manner to the No. 6
oil, with the most effective injection locations being the gas ports, primary
air, or total air stream. Some of the results have already been presented in
Figure 15, the remainder are shown in Figures 23 and 24. As with No. 6 oil,
during all of these FGR tests, the CO emissions were less than 100 ppm.
With natural gas, the most effective location for FGR was the total
air stream. This is due to the construction of the gas manifold. The gas
manifold is built with slots that inject natural gas into both the primary and
secondary streams. Thus one would expect the most effective configuration
to be one that injects the flue gas into both the primary and secondary.
41
-------
Primary Air
A Secondary Air
FUEL: No. 6 Oil
EXCESS AIR: 17%
LOAD: 3.4X106
BTU/HR
FOR Temperature:
NO Baseline: 273 pp
" '"sec
10 20
% RECIRCULATION
Figure 22
EFFECT OF FGR LEVEL, WITH PRIMARY
AND SECONDARY FLOWS CONSTANT,
ON EMISSIONS
42
-------
g
1.0
0.9
e.8
O.7
o.6
0.4
a.3
0.2
0.1
Recycle
Injection
Total Air
Secondary Air
Primary Air
Gas Ports
Throat
NO
O
a
O
A
0
FUEL: No. 2 Oil
Load: 3.3 x I O6 Btu/hr
P/S Air: 50%/50%
Excess Air: 17%
FGR Temperature: 300
NO Baseline: 94 ppm
10 20 30
% RECIRCULATION
Figure 23 ••
EFFECT OF FGR INJECTION LOCATION
ON EMISSIONS
(No. 2 Oil Fired)
43
-------
NO/NO.
— A Secondary Air
FUEL: Natural Gas
EXCESS AIR: 17%
LOAD: 3.4X106
BTU/HR
P/S Air: 50%/50%
FOR Temperature: 30
NO Baseline: 64 ppm
— D Primary Air
Total
20 30
% RECIRCULATION
40
Figure 24
EFFECT OF FGR INJECTION LOCATION
ON EMISSIONS
(Natural Gas Fired)
44
-------
VII. EFFECTS OF STAGED COMBUSTION
A. SUMMARY
Results of the SC tests indicate the following:
Sidewall and rear-boom give 24% and 20% NO reductions,
respectively, for the present burner configurations.
Combustion staging has been found more effective in prior
investigations,of a 50 HP Cleaver Brooks boiler'5) and of
utility boilers <10>
NO reductions are limited by increased smoke.
The large smoke emissions could be due to either (1) the
overall rich fuel/air ratio near the burner, or (2) the reduced
fuel/air mixing rate caused by reduced air flow at the
burner.
Staged combustion results in increased heat transfer from
the gases to the coolant. This is probably due to changes
in flow patterns in the combustor.
Downstre'am air injection increases NO unless injected
at least 1.5 combustor diameters downstream from the
oil nozzle. This result indicates that the near-burner
flame is effectively naturally staged without any deliberate
SC.
With rear boom staging, half of the ultimate reduction is
obtained with the first 5% of the air staged.
B. INTERPRETATION FOR OIL FIRING
"Staged combustion," in the sense of a fuel rich premixed region
near the burner and delayed air addition downstream is a misleading over-
simplification in the combustion of No. 6 oil. The mixture produced by
an oil-fired burner is a heterogeneous mixture of liquid fuel drops and air
burning in a diffusion flame. Realistically, what probably is occurring in this
diffusion flame situation is staged burning of drops such that drops are
burning in "series," in the products of previously burned drops. This is
contrasted to the baseline or non-staging case where there is sufficient air
near the burner for all (or at least a large majority) of the drops to burn in
45
-------
"parallel". For this reason SC is not expected to be as effective for oil as
for natural gas. In fact, gas turbine manufacturers have found that this
limitation can only be circumvented by prevaporizing the fuel.
This is illustrated below in Figure 25. Consider a fuel nozzle which
injects two fuel drops of such a size that under stoichiometric conditions each
requires one unit of air to burn, thus for overall stoichiometric conditions
two units of air are added to the combustor. Figure 25 compares two configu-
rations in which the fuel may be burned: (a) all air and fuel added at the
burner, and (b) insufficient air added at the burner, the remainder of the air
added downstream.
AMume
INJECTION
o
pfr
IGNITION
AFTER MIXING
DROPS BURN
SIMULTANEOUSLY
PRODUCTS
(a) CONVENTIONAL CASE
INJECTION
IGNITION
AFTER MIXING
FIRST FUEL
BURNS AS IN
CONVENTIONAL
CASE
STAGED AIR
MIXES WITH
PRODUCTS
(b) STAGED COMBUSTION CASE
LAST FUEL
BURNS IN
DILUTED AIR
Figure 25
PHENOMENOLOGICAL SKETCH OF STAGED COMBUSTION
IN OIL FLAMES
46
-------
In the conventional case (a), there is sufficient air in the local
environment of each fuel drop or element near the burner that both fuel
elements can burn simultaneously. Thus the environment that each fuel
element will see will consist initially of air and as burning proceeds, the
environment will progress towards the final level of excess air plus the fuel
elements' own products.
The situation is distinctly different with staging [case (b)], where
there is not sufficient air at the burner for all of the fuel elements to burn
simultaneously. The first element will burn as a diffusion flame in an
environment of air and its own combustion products, largely unaware of the
presence of the other fuel element (this is due to the self-regulation of a
diffusion flame with regard to its stoichiometry).
The second fuel element (drop or vapor pocket) will not have air to
burn and must wait until the air is added downstream. When this secondary
air is added, the environment which the second fuel element "sees" will no
longer consist initially of air; rather it will initially see air plus the products
of combustion of the first fuel element as it burns.
Four effects important to NO and smoke formation will occur as follows:
(1) While awaiting the addition of secondary air, the second
fuel element will be preheated and vaporized by the
combustion products of the first element. This will
"initiate" pyrolysis and soot formation.
(2) Also, the presence of the combustion products from the
first element around the second will reduce the oxygen
mole fraction.
(3) Raising the inital temperature of the reactants will raise
the flame temperature of the second fuel element. This
effect may be offset by the dilution effect which reduces
flame temperatures. Heating cannot occur without dilution.
(4) Heat will be transferred to the coolant prior to the addition of
the staged air, thus smoothing the heat release profile in
the combustor.
Wilson et al. have shown from a simplified analysis of NO forma-
tion in a diffusion flame that the second effect described above dominates,
thus lowering the NO formation rate.
47
-------
C. AXIAL BOOM INJECTION RESULTS
During these tests the rear boom was located at various distances
from the burner and the level of staged air varied from 10% to 30% (105% to
82% theoretical air at the burner). The results of these tests are shown in
Figure 26. The optimum NO reduction (20%) was attained with 30% of the
air staged and with the boom located 2.2 combustor diameters from the oil
nozzle. For this condition, smoke increases from 4 to 8 BSN. In fact, the
flame turned completely opaque when the boom was moved farther than 2 .2
combustor diameters from the nozzle. Figure 26 also shows that delivering
the axial air flow closer than 1.5 combustor diameters from the nozzle results
in increased NO emissions, presumably because of intensified mixing or
higher effective primary air. The unmodified combustor appears to be naturally
staged. Most of the NO reduction is attained with the first 10% of staged
A
air; beyond 10%, further reductions in NO are very small. Since the burner is
at 105% theoretical air at 10% staging, the NO reduction is certainly not to
*•>
be understood in terms of the premixed concept of staging. More likely the
mechanism is a shift in the flowfield pattern, or in the drop burning histories.
With 20% rear boom staging, NO and soot were found insensitive to
both primary-to-secondary air ratios and also to levels of swirl at the burner ,
as shown in Figure 27. The observed insensitivity to primary/secondary air
distribution is quite different from the no staging case, and was found to
persist regardless of boom position. As shown in Figure 28 the smoke
emissions under staging were lowered by 2 BSN when the P/S ratio was
shifted to 20-80.
D. SIDEWALL STAGING RESULTS—EFFECT OF INJECTION POINT AND
ANGLE
Side wall injectors were used in a series of experiments to investigate
not only the effect of the axial location of the staged air delivery (in a manner
similar to the rear boom), but also the effect of orientation of the staged air
flow angle. The individual injectors are designed such that the air leaves
the injector at an angle from the injector axis—see Figure 29.
48
-------
Oil Nozzle
200
Location of Main Combustor Sections
No staging- --
| durner operating
stO% excess air
• No Staging-
Combustor Exit
1
jFuel: No. 6 Oil
Load: 3.4x10° Btu/hr
Excess Air: 17%
P/S Air: 50%/50%
J % Theor. Air
% Staged at Burner
10.3
20.3
30
105
93
82
2345
Rear Boom Location (Combustor Diameters from Oil Nozzle)
10
u
°i
i
o
2
0
Figure 26
EFFECT OF REAR BOOM POSITION AND LEVEL
OF STAGING ON EMISSIONS
49
-------
Fuel: No. 6 OIL
Load: 3.4 x IO6 Btu/hr
Excess Air: 17%
Staged Air: 20% (93% Theoretical Air at Burner)
Rear boom located 1.7 combustor diameters
from oil nozzle
Swirl Parameter = 1.8
With staging
Without staging
300
2
0200
u
II
100
Swirl Parameter
.8 1.4 0.9 0.6 0.46
10 12
•I
6 ra
12345
Air Control Valve Opening (in.)
(b)
Figure 27
EFFECT OF P/S RATIO AND SWIRL ON
EMISSIONS DURING STAGING
50
-------
350
*
250
o
200
• No Staging
No Staging
No Staging
» No Staging
20
30
Fuel: No. 6 Oil
Load: 3.4 x I06 Btu/hr
Excess Air: 17%
Swirl Parameter: 1.8
Oil Temperature: 200°F
Atomization Air Pressure: 20 pslg
Staged Air: 20%
(93% Theoretical Air at Burner)
10
D
D
P/S Air
50%/50%
20%/80%
£
O
in
J_
40 50
Inches from Oil Nozzle
60
Combustor Diameters from Oil Nozzle
REAR BOOM LOCATION
Figure 28
EFFECT OF PRIMARY/SECONDARY AIR RATIO
ON EMISSIONS
51
-------
Burner Swirl
Direction
Rotate to
Change
Jet Angle
Orifice Area
Some as Pipe ID
Figure 29
CUTAWAY SKETCH OF INJECTOR TIP
Four orientations relative to the burner flows were tested:
• Co-swirl (in the direction of the burner air swirl)
• Counterswirl (against burner air swirl)
0 Upstream (towards burner)
9 Downstream (away from burner)
When air was delivered through the first set of injectors (Sj) which
are located 20-3/8" (0.9 combustor diameters) from the oil nozzle, it was
found that the NO emissions were greater than the case of no staging.
This result corroborates previous evidence that the primary mixing field of
this burner is readily intensified by air delivered within one combustor
diameter.
The second ($2) and third (83) set of injectors were then tested for
their effectiveness in adding the staged air, as shown in Figure 30.
NOX reductions were smoke limited to about 24%, with 83 slightly more
effective than 82- With the 82 and 83 injectors, the downstream orientation
could not be used due to excessive smoke emissions. In addition, tests
could not be performed with more than 17% of the air staged through the 83
injectors in the co-swirl or the counterswirl orientations due to excessive
smoke formation.
52
-------
Fuel: No. 6 Oil
Load: 3.4 x 10 Btu/hr
Excess Air: 17%
P/SAir: 50%/50%
300
.03
0 250
#•
CO
4-»
(0
I
A 200
Injector
Orientation
Upstream
Counterswirl
Co-swirl
150 •
10
20 30
% Air Staged
J.
L
40
_L
117
110 100 90 80 70
Burner Stoichiometry, % Theoretical Air
EFFECT OF S2 AND
Figure 30
S^ INJECTOR ORIENTATION
53
10
o
o
a
CD
2|
co
0
50
60
-------
Various combinations of 82 and 83 were also tested with results
similar to the tests using the individual injectors. Limited testing was also
performed at reduced load (2 x 10 BTU/hr) using the 82 and 83 injectors;
however, the results were less than promising, with reductions in NO of
only 14% before smoke emissions exceeded a 9 Bacharach smoke number.
E. SIDEWALL STAGING RESULTS—HIGH EXCESS AIR
Similar tests were performed with the burner operated at 3.4 x 10
Btu/hr but with 35% overall excess air. These results are reported in
Figure 31.
The results showed that:
9 The 83 injectors are more effective in reducing NO emission
than the S2/ presumably because they are farther downstream,
thus increasing the probability of late-burning drops burning
in the products of earlier drops.
• The most effective orientation in terms of NO reductions
was when the air was injected in the same direction
of the burner swirl. This orientation minimizes the
relative velocity between the combustion products and
staged air and thus maximizes the staging effect.
• The reductions in NO were attained only by accepting
increased smoke. Typically when the NO emissions
were reduced to 200 + 5 ppm (approximately 25% reduction),
the smoke emissions were greater than a 9+ BSN (off scale!).
F.. SIDEWALL STAGING RESULTS—HEAT FLUX DISTRIBUTION
In attempting to analyze these results, it is of value to look at the
changes in the thermal profile along the combustor with staging through the
rear boom (Figure 32), and with the sidewall injectors S2 and 83 (Figures
33 and 34). The results show that with staging the heat flux distribution
is not more uniform and the total heat transfer to the coolant was greater.
In fact the total heat flux increased as the point of staged air addition was
moved farther downstream from the burner. The only characteristic difference
between the boom configuration (Figure 32) and sidewall configuration
(Figures 33 and 34) was some sensitivity to the orientation of the sidewall
injectors. Adding the staged air counter to the direction of swirl from the
burner resulted in a greater overall heat flux to the coolant than addition of
54
-------
Fuel: No. 6 Oil
Load: 3.4 x 10 BtuAr
Excess Air: 35%
P/SAir: 50%/50%
300
a
a
200
Injector
Orientation
Upstream
Counters wirl
Co-swirl
150
10
L
20 30
% Air Staged
I
I
10
o
8 J
O
6 3
O
S3
0)
^:
o
g
50
135 130 120 110 100 90 80
Burner Stoichiometry, % Theoretical Air
70
Figure 31
EFFECT OF S2 AND S INJECTOR ORIENTATION
ON EMISSIONS (35% EXCESS AIR)
55
-------
26
24
22
18
16
14
12
g 10
O
08
u.
o
w
4)
/S Air: 50%/50%
35% Staged Air (88% Theoretic
Air at Surner)
Boom Loc. Total Coolant
Point fComb Dias) Temp. Rise
Combustor
Entrance
I — no staging
2—- 1.7
3 2.2
65F
66°F
* I Combustor
Boom 3oom Exit
Location (2) Location (3)
Distance Along Combustor Axis
Figure 32
EFFECT OF BOOM LOCATION ON
COMBUSTOR HEAT TRANSFER DISTRIBUTION
56
-------
u.
o
«
o
(A
26
24
22
20
16
O
£14
a>
13
gio
u
2 8
I 6
v
Fuel: No. 6 Oil
Load: 3.4 x I04 Btu/hr
Excess Air: 17%
P/S Air: 50%/50%
17% Staged Air (93% Theoretical
Air at Burner)
Injectors
no staging
Ct> burner swirl
counter burner swirl
i
Total Coolant
Temp. Rise
65°F
66°F
67°F
\
Injector
Location
Combustor
Entrance
Distance Along Combustor Axis
Com bus tor
Exit
Figure 33
EFFECT OF S2 INJECTORS ON
COMBUSTOR HEAT TRANSFER DISTRIBUTION
57
-------
26
24
u.22
o
0*20
*I8
12
810
o
B 8
o .
I 6
Fuel: No. 6 Oil
Load: 3.4 x I06 Btu/hr
Excess Air: 17%
P/SAir: 50%/50%
17% Staged Air (93% Theoretia
Air at Burner)
S~ Injectors
no staging
Co burner swirl
counter burner swirl
Total Coolant
Temp. Rise
6I°F
7I°F
73°F
"3
Injector
Location
Combustor
Entrance
Distance Along Combustor Axis
Figure 34
EFFECT OF S INIECTORS ON
COMBUSTOR HEAT TRANSFER DISTRIBUTION
Combustor
Exit
-------
the staged air in the direction of the burner swirl.
One might have expected the opposite result if staged air when-
injected participated immediately in combustion. Compared to the no-staging
case, the rich region upstream of the staged air addition should release less
heat with more heat subsequently released downstream of the staged air
addition. Staging could conceivably smooth'the heat distribution along the
boiler, and by "clipping" temperature peaks, reduce the total heat loss. The
observed increase in heat flux although unexpected may have at least three
plausible explanations: (1) staging may replace the cool, secondary air near
the wall by gas which is hot and more turbulent, (2) the fuel spray cone may
deliver the combustion zone closer to the walls due to the reduced burner air,
or (3) an increase in radiation in the first section of the combustor due to the
rich combustion with increased soot formation.
59
-------
VIII.
EFFECTS OF COMBINED FGRAND STAGED COMBUSTION
Tests were conducted with 24% of the total air added through the
rear boom (89% theoretical air at the burner) and 25% of the flue gas
recirculated through the burner. This arrangement essentially kept the
volumetric flowrates through the burner within 5% of the flowrates under
conditions of no staging. The results of those tests, shown in Figure 35,
show that the NO emissions could be reduced 45% to 150 ppm with smoke up
only 2 BSN above baseline. Furthermore, the rear boom could be operated at
much greater distances from the burner without producing excessive amounts
of smoke as occurred with staging alone.
-300
JO
"E
4*
5 200
— Level with No FGR or Staging
«- Level with
Fuel: No. 6 Oil
Load: 3.4 x I06 Btu/hr
Excess Air: 17%
P/SAIr: 50%/50%
Staged Air. 24%
(69% Theoretical Air at Burner)
FGR: 25% to Burner Air
O NO
Only
Smoke
Level with FGR Only
Level with No FGR or Staging-4 ~
10
8 -5
o
•1
2
4 3
o
to
2
2345
Rear Boom Location (Combustor Diameters from Oil Nozzle)
Figure 35
EFFECT OF REAR'BOOM LOCATION
ON EMISSIONS DURING COMBINED FGR AND STAGING
60
-------
Recall that FGR alone gave 30% reduction in NO and staging alone
gave 20% reduction (high smoke with staging). The combination gives 45%
reduction in NO without smoke penalty. It is possible that the rich smoke-
producing regions normally formed with staged air were dispersed by the
FGR added through the burner. Another possibility is that the suppression
of smoke may have been the result of a chemical effect with the presence
of CO0 or H0O in the gas [e.g., CO0 + C(soot) -* 2CO or C + H0O -* CO + H0].
(451213) ^ ^ *•
Other workers ' ' ' have obtained similar results and report that with the
addition of FGR, the flame has become less luminous indicating a decrease in
the soot concentration in the flame.
61
-------
DC.
EFFECT OF REFRACTORY SLEEVES ON NO AND SMOKE CONTROL
Tests were performed with the combustor lined with varying lengths
of refractory sleeve in an effort to raise the mean temperature in selected
sections of the combustor, stimulating soot burnup. Some increase in
thermal NO was also expected.
The refractory was added by casting 2-inch thick, 15"-long rings
and sliding them into the combustor (6 complete rings totally lined the
combustor), see Figure 36.
One 11" Throat Section
Three 30" Modular Furnace Sections
\&
Refractory Sleev(
X. X \ X
X X X X
1C
"JIT ' ' "
Figure 36
REFRACTORY SLEEVES
3/r.: :;!i:: :^i
TJair '••*'• u
TJ
Six Annular
Coolant Jackets
Tests were then conducted with pairs of the sleeves located in various
30-inch sections of the combustor. The tests with the refractory liners
concentrated on the effectiveness of staged combustion and FGR. During these
tests staging was done with the rear injection boom. A summary of the data
for staged combustion is presented in Figure 37 for 25% of the air staged
through the rear boom. Further data is presented in Figures 38, 39 and 40.*
As can be seen in Figure 37, the presence of the refractory raised the base-
line NO emissions and lowered the baseline smoke emissions. Surprisingly
the most effective location for reducing the baseline smoke emissions was
with the center third of the combustor lined with refractory. This may be due
to the center refractory raising the temperature of a recirculation zone in the
combustor.
*In Figures 37 through 42, the cross-hatched region indicates the section of
the combustor that was lined with refractory.
62
-------
01
400
ozzlc
Combustor Exit
O
a
A
00
!30C
o
I
Q.
(Cross-hatch indicates location of refractory inserts)
Y////////A
No Staging
It
L.
+*
z
200
V////////
Y//////7/
Fuel: No. 6 Oil
Load: 3.4x I06
3tu/hr
RXCCSE Air: 17%
P/5 Air: 50%/50%
Staged Air 35%
(08^ Theoretical
Air to the Burner)
12345
Rear ?oom Location (Combustion Diameters from Oil Nozzle)
Figure 37
SUMMARY OF THE EFFECT OF REFRACTORY
LOCATION ON EMISSIONS
63
-------
Oil Nozzle
g.400
•8 300
5
4*
z
200
Combustor Exit
Y/////////////////////////////.
No staging or FOR
O No staging or FOR
Fuel: No. 6 OH
Load:3.4xl068tu/hr
% Theor. Air
% Staged at Burner
> I 2 3 4 5
•Rear Boom Location (Combustor Diameters from Oil Nozzle)
Figure 38
EFFECT OF LEVEL AND LOCATION OF STAGED AIR ADDITION
ON EMISSIONS (COMBUSTOR COMPLETELY LINED WITH REFRACTORY)
10
64
-------
our
400
cf
»<
15
c
o.
•=300
•8
0
0
*u
Z
200
0
ozzle Combustor Exit
j
V //////// A
\/ //////// A*uv\\ No. 6 Oil
Load: 3.4xl06 Btu/hr
Excess Air: 17%
P/S Air: 50%/50%
% Theor. Air
% Staaed at Burner
• 17 97
• No staging or FGR • 25 88
*X. A 36 73
^^^
*^^_
^^C2*^~ ••
SsOnsJ
j/S^*
^K^*/
*^--t
-
• No staging or FGR
-
1 1 1 1 '
10
5
8|
•
0
4|
2"
0
2345
Rear Boom Location (Combustor Diameters from Oil Nozzle)
Figure 39
EFFECT OF LEVEL AND LOCATION OF STAGED AIR ADDITION
ON EMISSIONS (THIRD COMBUSTOR MODULE REFRACTORY LINED)
65
-------
OIIN
400
i
o
"Z
z
200
0
ozzle Combustor Exit
Y/////S///X Fuel: No. 6 OH
Excess Air: 17%
P/SAir: 50%/50%
% Theor. Air
% Staged at Burner
ft No staging • 17 97
•^^XA • 25 88
^S5^^^ A 36 73
Extrapolated ^^"^v^
+-- Burner operating ^*ft —
at 0% excess air
' J/ •
^f
§ No staging
L 1 1 1 1
o ro .ik o o> o
Smoke No. (Sacharach)
I 2 3 4 5
Rear Boom Location (Combustor Diameters from Oil Nozzle)
Figure 40
EFFECT OF LEVEL AND LOCATION OF STAGED AIR ADDITION
ON EMISSIONS (FIRST COMBUSTOR MODULE REFRACTORY LINED)
66
-------
It was found that with the partial refractory linings, NO reductions
were not very sensitive to the amount of air staged*. In addition, excessive
smoke formation limited NO reductions to only 200 ppm as indicated in Table 7.
Table 7
EFFECT OF REFRACTORY POSITION ON MINIMUM NO
NO at Limiting Smoke
Refractory Configuration Level (No. 8 Bacharach)**
No refractory 220
First third 225
Middle third 200
Back third 185
First and last third 195
Full refractory 185
Combined FGR and staged combustion tests were also conducted
with two of the partial refractory configurations. The results of these tests
are reported in Figures 41 and 42 and compared with the results obtained
with staging only. The major difference is again the suppression of smoke
emissions from the combustor when flue gas is recirculated; the differences
in NO emissions were minor.
*This is not true when the entire combustor is lined with refractory. However,
under these conditions, the unit no longer functions as a boiler.
**The limiting smoke level of a No. 8 Bacharach was arbitrarily set, and is
high from a practical standpoint. From Figures 41 and 42, it can be seen that
if the smoke limit was taken as a No. 6 Bacharach smoke number then the
NO levels could only be reduced to approximately 230 ppm, again essentially
independent of the level of staged air or the location of the partial refractory.
67
-------
Oil Nozzle
400
-300
200
Y////////S
No staging or FOR
No staging or FCR
Combustor Exit
Fuel: No. 6 Oil
Excess Air: 17%
P/SAIr: 50%/50%
Staged Air: 25%
Staged Air Only
(88% Theor. Air at Burner)
Staged Air and 25% FOR
added to burner air
12345
Rear Boom Location (Combustor Diameters from Oil Nozzle)
Figure 41
EFFECT OF COMBINED FGR AND STAGING ON EMISSIONS
(FIRST COMBUSTOR MODULE REFRACTORY LINED)
10
6 cr>
o
&
68
-------
Oil is
400
ozzle
Combustor Exit
o"
o
•o
"x
O
300
No staging or FGR
200
No staging or FGR
Fuel: No. 6 Oil
Excess Air: 17%
P/S Air: 50%/50%
Staged Air: 25%
(88% Theoretical Air
at Burner)
Staged Air only
Staged Air plus 25% FOR
added to burner air
10
\j
a
o
•6
&
o
4I
o
2"
12345
Rear Boom Location (Combustor Diameters from Oil Nozzle)
Figure 42
EFFECT OF COMBINED FGR AND STAGING ON EMISSIONS
(THIRD COMBUSTOR MODULE REFRACTORY LINED)
69
-------
REFERENCES
1. Muzio, L. J. and Wilson, R. P. Jr., "Experimental Combustor for
Development of Package Boiler Emission Control Techniques," Phase I
of III, EPA Report R2-73-292a, 1973.
2. Heap, M. P. and Lowes, T. M., "Nitric Oxide Production in Large
Scale Natural Gas Flames," PR10, EPA Contract No. 68-02-0202,
International Flame Research Foundation, December 1972.
3. Wasser, J. H., Berkau, E. E., and Pershing, D. W., "Combustion
Intensity Relationship to Air Pollution Emission from a Model Com-
bustor System," EPA, Combustion Research Section, August 1971.
4. Martin, G. B. and Berkau, E. E., "Preliminary Evaluation of Flue Gas
Recirculation as a Control Method for Thermal and Fuel Related Nitric
Oxide Emissions," presented at WSS/CI Meeting, University of
California, Irvine, October 1971.
5. Turner, D. W. and Siegmund, C. W., "Staged Combustion and Flue
Gas Recycle: Potential for Minimizing NOX from Fuel Oil Combustion,"
presented at The American Flame Research Committee Flame Days,
Chicago, Illinois, September 6-7, 1972.
6. Muzio, L. J. and Wilson, R. P. Jr., "A Versatile Combustor for
Developing Emission Control Techniques—Reference Report," avail-
able from American Petroleum Institute, 1973.
7. Burner Fuel Oils, Mineral Industries Survey, U.S. Department of
the Interior, Bureau of Mines, 1972.
8. Barrett, R. E. and Miller, S. E., "Field Investigation of Emissions
from Combustion Equipment for Space Heating," API Publication 4180,
1973.
9. Bowman, C. T. and Kesten, A. S., "Kinetic Modeling of Nitric Oxide
Formation in Combustion Processes," WSS/CI Paper 71-28, Fall
Meeting, Combustion Institute, 1971.
10. Thompson, R. E. and Teixeira, D. P., "The Effectiveness of Combus-
tion Control in Reducing NO* for Oil Fired Power Plant Boilers,"
Central States Section/Combustion Institute, Champaign, Illinois,
March 1973.
70
-------
11. Wilson, R. P. Jr., Waldman, C. H. and Muzio, L. ]., "Foundation
for Modeling NOX and Smoke Formation in Diesel Flames," Final
Report for Phase I, Contract No. EPA 68-01-0436, January 1974.
12. Andrews, R. L. et al., "Effects of Flue Gas Recirculation on Emissions
from Heating Oil Combustion," 61st Annual Meeting, Air Pollution Con-
trol Association, St. Paul, Minnesota, June 1968.
13. Siegmund, C. W. and Turner, D. W., "NOX Emissions from Industrial
BoHers: Potential Control Methods," ASME Paper 73 1PWR10, 1973.
71
-------
TABLE OF CONVERSION FACTORS
To Convert
LENGTH
AREA
VOLUME
MASS
PRESSURE
ENERGY
POWER
TEMPERATURE
inches
feet
'meters
square inches
square feet
square meters
cubic inches
cubic feet
liters
gallons
cubic centimeters
pounds
grams
pounds/in
pounds/in
Btu
BtuAr
kilowatts
horsepower
horsepowor(boiler)
horsepower (boiler]
degrees Farenheit
degrees Rankine
Into
meters
meters
centimeters
square meters
square meters
square centimeters
cubic centimeters
cubic meters
cubic meters
liters
cubic meters
grams
kilograms
dynes/cm
feet of water
gram-calories
gram-calories/second
gram-caloriesAour
kilowatts
BtuAr
kilowatts
degrees .Celsiu s (°C)
degrees Kelvin (°K)
Multiply By
2.54 x 10"2
3.048 x 10
l.OxlO2
6.452
9.29x!0"2
l.Ox 104
1.639 x 101
2.832x 10"2
l.Ox 10"3
3.785
l.Ox 10"6
4.54 x 102
l.Ox 103
6.9x 104
1.60x 10~2
2.52x 102
7. Ox 10~2
8.6x 105
7.46 x 10"1
3.35 x 101
9 -.803
(°F - 32)
x 5.56 x 10'1
5.56 x 10~*
72
-------
TECHNICAL REPORT DATA
(ricasf rcail /nuruciionx on tlic reverse before
i. MI PORT NO. 2.
EPA-R2-73-292-b
.J. TITLE AND SUMTITLE
Package Boiler Flame Modifications for Reducing
Nitric Oxide Emissions—Phase n of in
7. AUTHOH(K)
L.J.Muzio, R.P.Wilson, Jr. , and C. McComis
------- |