United States
Environmental Protection
Agency
Air and Energy Engineering
Research Laboratory
Research Triangle Park NC 27711
Research and Development
EPA/600/S7-85/017 July 1875
Project Summary
Laboratory Evaluation of NO
Reduction Techniques for
Refinery CO Boilers
H. B. Lange, J. K. Arand, M. N. Mansour, and S. C. Hunter
A laboratory test program was con-
ducted to investigate NO, emissions
from refinery CO boilers. The program
had three major objectives: (1) to simu-
late in the laboratory a full-scale refin-
ery CO boiler, (2) to investigate the
effects of operational variables on NO,
formation in the boiler furnace, and (3)
to evaluate combustion modification
techniques to reduce the NO, emissions.
The laboratory model was shown to
accurately represent combustion and
NO, formation processes occurring in
the full-scale boiler by comparison of
combustion products analyses over a
range of operating conditions. The most
significant operating variables influenc-
ing NOX formation in the CO boiler were
found to be the ammonia level in the
CO-gas and the burner stoichiometry.
Oxidation of ammonia to NO, was found
to account for most of the NOX formed
in the boiler furnace. The data also
showed that NO, contained in the CO-
gas is partially destroyed in passing
through the CO boiler.
NO, reduction methods that were
tested included combustion air staging,
staged admission of fuel, fuel injection
in the CO-gas, redesign of the fuel injec-
tors, and burner throat redesign. The
most effective NOX reduction technique
was found to be combustion air staging.
Part of the combustion air was diverted
away from the burner and added down-
stream in the furnace to complete the
combustion process. This reduced the
oxygen level in the high-temperature
burner zone of the furnace and thus
reduced oxidation of ammonia to NO,.
The effectiveness of combustion air
staging in achieving NO, reduction
without simultaneously increasing the
CO level leaving the boiler was found to
be strongly dependent on the place-
ment of the second-stage combustion
air.
NO, formation in the CO boiler was
reduced to zero (NO, emission equiva-
lent to NOXentering with CO-gas) in the
laboratory while maintaining acceptable
(<200 ppm) CO levels in the stack gas.
While zero NO» formation will probably
not be achievable in the field where
conditions are far less controllable,
substantial NO, reductions should be
achievable under practical operating
conditions.
This Project Summary was developed
by EPA's Air and Energy Engineering
Research Laboratory. Research Trian-
gle Park, NC, to announce key findings
of the research project that is fully doc-
umented in a separate report of the
same title (see Project Report ordering
information at back).
Background and Objectives
Most petroleum refineries contain one
or more carbon monoxide (CO) boilers for
incineration of off-gases from catalytic
cracking operations. These devices con-
stitute a significant fraction of total refin-
ery NOx emissions. Methods that have
been used to reduce NOx emissions from
conventional boilers either are not appli-
cable or have performed poorly when ap-
plied to refinery CO boilers burning off-
gases with high CO and ammonia con-
tent.
The work reported here had three major
objectives: (1) to simulate in the labora-
tory a full-scale refinery CO boiler, (2) to
-------
investigate the effects of operational var-
iables on NO. formation in the boiler fur-
nace, and (3) to evaluate combustion
modification techniques to reduce the
NOx emissions.
The refinery CO boiler basically oxi-
dizes CO contained in the catalytic crack-
ing off-gas (CO-gas) to CO2. In the CO
boiler, CO-gas mixes with combustion
products and excess air from a burner or
burners operating typically on refinery
gas, and CO is oxidized to C02. Prior field
measurements by KVB had established
that the CO-gas contains substantial
amounts of NO. and ammonia both of
which could contribute to the NO, emis-
sions from the boiler.
Test Program
A laboratory model CO boiler was con-
structed in KVB's combustion laboratory.
This model, which operated in the range
of 1 to 2 million Btu/hr* total heat input,
was based on a full-scale CO boiler that
had been tested extensively by KVB. The
full-scale boiler selected employed a sin-
gle multispud burner and a horizontal
furnace cavity. The burner, the CO-gas
entry throat, and the furnace were scaled
down from the full-scale boiler based on
preservation of critical velocities, resi-
dence times, and heat transfer parame-
ters.
Simulated CO-gas was generated by
operating a firetube boiler fuel-rich on
natural gas. The stoichiometry of the CO-
gas generating furnace was adjusted to
achieve the level of CO concentration
desired for testing. A flue gas bypass
around the firetube section was adjusted
to achieve the desired CO-gas tempera-
ture. Ammonia was injected both before
and after the burner to achieve desired
levels of NOX and ammonia in the CO-gas.
The overall laboratory setup is shown in
Figure 1.
Results
A comparison of operating ranges and
NO, levels produced in the laboratory ver-
sus those found with the full-scale boiler,
shown in Table 1, indicated that the
laboratory model was a good facsimile of
the full-scale boiler.
Tests were conducted to determine the
effects of operating variables—including
CO-gas composition, ratio of CO-gas to
burner gas, and burner stoichiometry—
and to evaluate various NOx reduction
techniques.
*1 Btu = 1.055kJ.
The most significant operating vari-
ables influencing NO, formation in the
CO boiler were found to be the ammonia
level in the CO-gas and the burner
stoichiometry. Burner stoichiometry, Si,
is defined as the ratio of burner air flow to
the air flow theoretically required to burn
only the supplemental natural gas fuel.
Figure 2 shows the effects of these two
variables on net NO, formation in the CO
boiler—i.e., the excess of NO, appearing
in the stack over NO, entering with the
CO-gas. NO, in the CO-gas was found to
be partially destroyed in the CO boiler
furnace. The data indicated an approxi-
mate 75 percent survival rate for CO-gas
NOX under normal operating conditions.
In view of the strong role of ammonia
oxidation in determining the CO boiler
NO, emissions, testing of NO, reduction
techniques focused mainly on combus-
tion staging. The aim was to create a fuel-
rich combustion zone in which ammonia
(and NO,) would be largely reduced to ni-
trogen. The fuel-rich stage was followed
by addition of excess air to complete the
combustion process.
Several methods of adding the second
stage air were tested. These included
staged air ports (NO, ports) located around
the burner, an air lance located on the
burner centerline, a torus (donut)-shaped
air plenum located downstream in the
furnace, and stoppage of fuel to selected
injectors on the multispud burner (injec-
tors-out-of-service). As shown in Figure
3, the abilities of the various staging
techniques to reduce NO, without cor-
respondingly increasing CO emission dif-
fered widely. The center air and injectors-
out-of-service techniques, while achiev-
ing some NO« reduction, were less ef-
fective than the unstaged burner in
achieving significant NO, reductions with-
out loss of CO control. The NO, ports pro-
vided a measurable reduction in NOX
emissions, and the achieved reduction
varied with the location of the NO, ports.
The most significant result was the good
effectiveness of combustion staging using
the torus method of injecting second
stage air. As shown in the figure, the
torus method was highly successful—
reducing net NOX to zero (NO, emission
equal to NO, contained in the CO-gas)
while maintaining good CO control.
Conclusions
In reviewing the following major con-
clusions, drawn from the results of the
laboratory testing, it should be borne in
mind that the tests were performed on a
single-burner, horizontal furnace with CO
gas admitted via a concentric annulus
surrounding the burner throat. All tests
were on natural gas fuel. The ratio of CO-
gas to total stack gas was about 0.65 (dry
basis). The CO-gas contained about 2
percent CO, 500 ppm ammonia, and 120
ppm NO,. Some of the conclusions may
be very specif ic to some of these test con-
ditions, and others may be more general.
1. All the test program objectives were
achieved: the laboratory model was
shown to accurately reproduce the
emission characteristics of a full-
scale CO boiler, the effect of operat-
ing variables was determined, and
several promising combustion mod-
ifications for NO. reduction were
identified.
2. NOx formed in the CO boiler is
mainly due to conversion of ammo-
nia entering with the CO-gas.
3. The most important variables influ-
encing the NOx level are the ammo-
nia content of the CO-gas and the
burner stoichiometry (air/fuel ra-
tio).
4. Conversion of ammonia to NOx is
normally (without staging) in the
range of 10 to 55 percent, depend-
ing on burner stoichiometry.
5. Part of the NO, entering with the
CO-gas is destroyed in the furnace:
the survival rate is normally about
75 percent, but drops to as low as
about 10 percent with reduction in
burner stoichiometry.
6. Air staging was found to be the best
NO, control method. Addition of
staged air well downstream of the
burner with good cross-mixing pro-
duces the best results.
7. Asymmetric air register setting to
bias the air is reasonably effective
but would be difficult to implement
in the field.
8. Taking fuel injectors out of service
(multispud burner) can produce
modest NOX reductions but tends to
cause high CO. CO characteristics
at various loads could be a problem
with this technique.
9. Fuel injection angle and velocity
have substantial effects on NO,,
and it appears possible to optimize
these variables for lower NO, oper-
ation.
10. Staged air ports located around the
burner are basically effective; ports
located below the burner are most
effective. Overall, however, NO,
ports are less effective than down-
-------
Combustion Air
Natural Gas •
i
tJ
4-
1
°A
Cold Flue Gas /*\~
Simulated Regenerator Gas 1 ~*T~"~"
H t Flue Gas
(Mfe.—
1
CO Generator
^^^
M)
n i j.
II*-
JL
T«
NH3 1
j
Sag House
Gaseous Sample Port — —
Continuous Oz. NO. /V0» CO, C02
— (Z)
_p^
~"^-'
for WO Addition
CO flo/'/e/-
(7
-
•r
i
-4
1 /!//•
I Preheater
P
L
•
rf>-
t
j
; Venturi
*> Manometer
-G
Ignitor
1
G/
*,-^
H/)
Pi
Supply *« M
pi Rotameter Continuous 02 , NO. NOX . CO. CO 2 1 '
r2
>
/
/v/y3 C?ra/) Sample
£
1
-4
i
Heavy Insulation
Gaseous
Sample Port
Thermocouple
Natural Gas
Figure 1. Schematic of laboratory CO boiler simulation system.
7"«/>/e /. Comparison of Operating Ranges and /VO, Levels Produced—Laboratory Boiler
Versus Full-Scale Boiler
Normal Operation Boiler Only —
with CO-Gas No CO-Gas
CO-Gas (dry):
CO, %
/VO» ppm
NHy, ppm
Temp., K
Fuel to Burner, MWt
Stack Gas (dry):
Oa%
CO, ppm
W0» ppm'
Dry Volumetric
Ratio of CO-gas
to Stack Gas
Full-Scale
1.8-3.4
45-120
470-620
-975
-20.5
1.0-3.5
50-700
105-270
0.52-0.62
Laboratory Full-Scale
1.2-4.8
48-150
490-675
-935
-0.2
1.4-3.6 3.4-3.6
30-265 120-600
117-325 40-51
0.54-0.71
Laboratory
—
-
-
--
3.5
30-40
40-60
—
'Corrected to 3% 02-
stream air staging using the method
described in conclusion 5, above.
11. NO. reduction approaches that are
ineffective include diversion of fuel
outboard of the burner or to the CO-
gas, less divergent burner throat,
and extended burner throat.
Applicability to Full-Scale
Boilers
The research program has shown that:
• Key operational variables influencing
NOX formation in refinery CO boilers
are ammonia level in the CO-gas and
burner stoichiometry.
• The most effective NO, control method
is to stage the combustion using a
second-stage air admission down-
stream in the furnace similar to the
torus air used in the laboratory.
Tothedegreethattheammonia level in
the CO-gas can be reduced—possibly by
selecting lower-nitrogen crudes or by
adjustment of catalytic cracking unit
-------
300 -
a 250 -
200 -
I
§
c
•S
5
00
i
6
Tests 1-36. 40
Si = Burner Stoichiometry
S2 = Furnace Stoichiometry
Symbols denote Si
and Si conditions as
noted on each curve
ISO -
100 -
200
400
600
800
1000
1200 1400
- Ammonia in CO-gas. ppm (Dry)
Figure 2.
Effect of ammonia content of CO-gas and burner Stoichiometry on NO*
formation in CO boiler.
operating parameters (if possible)—this
will reduce NO, formation in the CO
boiler. This approach, however, will prob-
ably not be practical in most cases. Con-
trol of burner Stoichiometry (low excess
air) is probably practiced to the extent
possible on most CO boilers but is limited
by the need to run sufficient excess air to
burn out all CO contained in the regener-
ator gas.
Staging the combustion provides the
added flexibility needed to control burner
Stoichiometry while meeting the objec-
tive of good CO oxidation. Use of a staged-
air torus downstream in the furnace, as
practiced in the laboratory, could be adapt-
ed to full-scale boilers or could be trans-
lated into staged air ports in the furnace
walls. Previous efforts to stage refinery
CO boilers have not generally produced
the desired results. The laboratory pro-
gram has demonstrated that, when prop-
erly designed, combustion air staging can
be an effective NO* reduction technique
for refinery CO boilers.
The results of this program are intended
for low temperature catalytic cracking
processes where the catalyst is regener-
ated in a deficiency of air, thus producing
high CO and ammonia. The data is not
intended to be representative of high
temperature catalyst regeneration in
which CO is burned to C02 and ammonia
is converted to mostly NOX prior to leaving
the regenerator.
4
-------
SI
Q^
S
i
150
140
130
120
110
100
90
80
70
60
50
40
30
20
10
0 -
-10 -
-20
/VOX Ports
(Bottom
Ports
Only)
Staged Air Method
O Unstaged
Q Burner NOx Ports
A Injectors Out-of-Service
<^Center Air Lance
Torus Air
Torus and Injectors
Out-of-Service
100 200 300
CO in Stack, ppm (Dry)
400
500
Figure 3. /VOx versus CO for various methods of staging the combustion.
H. Lange, J. Arand, M. Mansour, and S. Hunter are with KVB, Inc., Irvine, CA
92714.
Robert E. Hall is the EPA Project Officer (see below).
The complete report, entitled "Laboratory Evaluation of NO^R eduction Techniques
for Refinery CO Boilers. "(Order No. PB 85-200 285/AS; Cost: $11.50, subject
to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Air and Energy Engineering Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
U. S. GOVERNMENT PRINTING OFFICE: 1985/559-111/20610
-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Official Business
Penalty for Private Use $300
EPA/600/S7-85/017
0000329 PS
U S ENVIR PROTECTION AGcNCY
REGION 5 LIERftRK
230 S DEARBORN STREET
CHICAGO IL eQ£04
------- |