vvEPA
United States
Environmental Protection
Agency
Industrial Environmental Research
Laboratory
Research Triangle Park NC 27711
Research and Development
EPA-600/S7-82-029 Oct. 1982
Project Summary
Effects of Fuel Properties and
Atomization Parameters on
NOX Control for Heavy Liquid
Fuel Fired Package Boilers
G. C. England, D. W. Pershing, M. P. Heap, and J. E. Cichanowicz
Experimental studies were con-
ducted to (1) relate the formation of
NO* in liquid fuel flames to the
chemical/physical fuel properties and
to the atomizer design, and (2)
investigate the interaction between
liquid sprays and airflow pattern for
conditions typical of package boilers.
These experiments were conducted in
an 880 kW firetube boiler simulator
and in a 21 kW tunnel furnace, under
both normal and staged combustion
conditions.
The fuels studied were primarily
conventional heavy residual fuel oils
covering a wide range of properties;
however, shale- and coal-derived
liquids were also tested. Fuel nitrogen
content of the liquid fuels was found
to be the only first order parameter
affecting NOx emissions for a given
spray/flow field configuration under
excess air conditions. The tunnel
furnace results indicated, however,
that (during staged combustion) fuel
nitrogen speciation has a second-
order effect on the minimum achiev-
able NOx level.
The boiler simulator studies showed
that atomizer design, spray/flow field
configuration, and fuel composition
are significant dependent parameters
affecting NOx emissions. Cold-flow
spray characterization using laser
diffraction showed that, in general,
drop size information alone was
insufficient to predict NOX emissions
in complex flow fields under normal
combustion; however, staged com-
bustion was found to be more effective
with atomizers which produced smaller
droplets.
Staged combustion was an effective
NOx control for all liquid fuels tested.
Optimization of staging parameters
(e.g., mixing and primary zone time/
temperature history) is indicated to
achieve maximum NOX reduction
using staged combustion.
This Project Summary was developed
by EPA's Industrial Environmental
Research Laboratory. Research Tri-
angle Park, NC. to announce key
findings of the research project that is
fully documented in a separate report
of the same title (see Project Report
ordering information at back).
Introduction
Although increased coal use partially
solves the U.S. energy crisis, many
industrial users will continue to burn
liquid fuels. However, the composition
of these fuels will change as premium
fuels are reserved for transportation
and domestic use. Industrial users will
be required to burn heavy petroleum-,
coal-, and shale-derived residual oils, all
of which have relatively high nitrogen
contents and low hydrogen-to-carbon
ratios. Unless appropriate pollution
controls are applied, use of these fuels
will increase NOx or paniculate emis-
sions from stationary combustors firing
-------
these fuels. The production of N0xand
participate in turbulent diffusion flames
depends on the fuel composition and
the fuel/air contacting process. The
latter depends on the complex interaction
of the liquid fuel spray from the
combustion air flow field.
The overall goal of this study was to
provide the information necessary to
develop and generalize low NOX oil
burner technology for application to
package firetube boilers. Specific objec-
tives of the research were:
• To relate the formation of NOX in
liquid fuel flames to the chemical
and physical fuel characteristics
and to the atomizer design.
• To investigate the interaction be-
tween liquid sprays and the airflow
pattern for conditions typical of
package boilers under normal and
staged conditions.
Two combustors (Figure 1) of different
scales were used to provide information
to assess the impact of liquid spray
characteristics on NOX formation and to
relate the formation of NOX under
normal excess air and staged heat
release conditions to the liquid fuel
properties.
Experimental Systems
Tunnel Furnace
The down-fired tunnel furnace was
designed to allow utilization of commer-
cially available atomizers and to be fired
with both artificial oxidants and air. The
small-scale combustor was 2.1 m long
with an I.D. of 0.2 m. The walls were
insulating and high temperature castable
refractory, and the nominal full-load
heat release was 70,000 Btu/hr (21
kW). In certain investigations the
combustion air was enriched or replaced
with varying amounts of CO2, Ar, and
02, all of which were supplied from high
pressure cylinders. The furnace was
fired with a commercial air-assisted
ultrasonic atomizer which was used
because it provided adequate atomiza-
tion of the viscous liquid fuels at these
low flow rates.
Boiler Simulator
Experiments at larger scale were
carried out in an axisymmetric calori-
metric combustor with a nominal firing
rate of 3 x 106 Btu/hr (880 kW). This
pilot-scale combustor was divided into
calorimetric sections cooled by heat
transfer fluid and had a length of 3.2m
and an I.D. of approximately 0.58m.
The double-concentric burner could
accept a wide range of commercially
available fuel nozzles. The unheated
combustion air was supplied through an
annular duct with an axial velocity of
30m/sec. Swirl level could be varied by
interchangeable fixed-vane annular
swirl generators.
Analytical System
Identical sampling and analysis
systems, used for both combustors,
Viewing
port
allowed for continuous monitoring of
NO, NOX, CO, C02, 02, and SOz using
commercially available instruments
Flue gases were withdrawn from the
appropriate combustor exhaust through
a water-cooled stainless steel probe
Fuel Effects
The work focused primarily on 13
petroleum-derived heavy fuel oils, but
also considered three distillate oils, four
Oil heater
connection
Combustion
air
air
__O/7 Pressure
Tap
Oil inlet
Tunnel furnace
Flue
Recycled flue products
To atmosphere
Stac,1'
To A ux.
heat
exchanger
Connective heat exchanger
Fuel
Observation
window
Access
port
Modular calorimetric
cooled sections
Figure 1. Combustors used in experiments.
Boiler simulator
-------
shale oil blends, and one coal-derived
liquid (Figure 2).
Excess Air Combustion
Figure 3 is a composite plot of total
and fuel NOx emissions from the tunnel
furnace for a wide range of petroleum-,
coal-, and shale-derived liquid fuels.
Fuel NOx was defined by replacing
combustion air with a synthetic oxidant
mixture containing 21 percent Oa, 20
percent C02, and 59 percent Ar. C02
was added to the mixture to match
theoretical flame temperature for the
air and artificial oxidant cases. Both
total and fuel NOx emissions increase
with increasing fuel nitrogen content.
These data are representative of a high
mixing intensity burner with a finely
atomized spray. Fuel NOX emissions
increase by approximately 95 ppm per
0.1 percent nitrogen. Experiments were
Legend
O Alaskan diesel
A W. Texas diesel
9 California diesel
O Essex County
k Middle east
O Low sulfur #6
A Indo/Malaysian
Q Desulfurized
Venezuelan
b Pennsylvania
O Gu/f Coasf
O Venezuelan
O Alaskan
V California #1
O California #2
0 California #3
O California #4
• Cructe s/?a/e and
blends with o
<1 Shale-derived DFM
• S/?C //
• Synthoil blends
with 9
• CM + /V//3
2.0
7.S
/.6
*
gO.5
0.5
0.4
0.2
A. I
-------
Staged Combustion
Staged combustion, a cost-effective
control for the reduction of NOX,
involves operating the combustor to
ensure that the fuel originally burns
under oxygen-deficient conditions.
Second stage air is then added to
complete the oxidation process. Figure
6 shows data obtained from both the
boiler simulator and the tunnel furnace
using the ultrasonic atomizer and
operated at 3 percent overall excess
oxygen. The data is presented as a
function of the stoichiometric ratio of
the primary zone, defined as the fraction
of the theoretical air required for
complete combustion. Since distillate
oil is essentially nitrogen-free, these
results represent the influence of the
primary zone oxygen on NOx produced
from molecular nitrogen. The data for
700
600
500
I
400
300
200
100
4% Excess 62
Son/core 2507 Atomizer
[/VOx] = 174 + 572 (%Nj
(fit - 0.99>
30° Swirl
45° Swirl
[/VOx] = 158 + 305 I%N)
(fit = 0.99)
1
I
0.2
0.4 0.6
Wt%N in Fuel
0.8
1.0
Figure 4. Effect of fuel nitrogen content -- boiler simulator.
4
the three residual oils indicate that, as
the primary zone becomes progressively
more fuel-rich, the ultimate NOx emis-
sions decrease dramatically to a mini-
mum and then increase again. This
upturn in emissions at very low stoi-
chiometric ratios is probably due to an
increase in the concentration of oxidiz-
able fuel nitrogen species at the exit of
the primary zone. Minimum NOX emis-
sions achievable under staged combus-
tion conditions for a range of liquid fuels
are compared in Figure 7 Minimum
emissions from the boiler simulatorare,
in general, higher than those from the
tunnel furnace, indicating that the
small-scale furnace provides more
optimum conditions for the minimization
of NOx production than does the boiler
simulator; e.g., time and temperature in
the fuel-rich zone. However, the general
trends for both combustors are the
same. Minimum NOX emissions increase
strongly with fuel nitrogen content up to
fuel nitrogen levels of approximately 0.5
percent, and then show only a minor
increase as the nitrogen content is
increased above 2 percent by weight.
The volatility of the nitrogen compounds
in the liquid fuel appears to be the
limiting factor in NOx control by staged
combustion. Experiments involving a
residual fuel and a distillate oil doped
with pyridine and thiophene to give the
same nitrogen and sulfur content
showed minimum emission levels
under staged conditions 30 percent
lower for the doped distillate than for
the residual fuel.
s
.270
§60
S?
50
0.4 0.8 1.2 1.6
Wt%N in Fuel
2.0
Figure 5. Composite fuel nitrogen
conversion to NOx as a
function of fuel nitrogen
content -- tunnel furnace.
-------
Atomizer Effects
A goal of this program was to assess
the impact of atomizer characteristics
on pollutant production under both
excess air and staged combustion
conditions. Commercial atomizers were
selected for use in the boiler simulator
and the characteristics of six of these
atomizers were determined in a cold-
flow test rig using a laser diffraction
droplet size analyzer.
Flow Field Interaction
In a given flow field, NOX formation
depends on the atomizer type. Figure 8
shows data for several commercial
atomizers, four of which were char-
acterized by the laser diffraction analy-
zer. The Peabody and Todd nozzles pro-
duced similar drop-size distributions
(mean diameter approximately 40/urn).
The largest mean drop size (90/urn) was
produced by the Monarch nozzle; the
700
£600
Q
O
S?
500
400
S
§;300
700
0
O Residual 0.51 %N
O Residual 0. 36% N
& Residual 0.24% N
* Distillate 0.007% N
T.unnel
rurnace
I i i i i
Package boiler simulator
0.6 0.8 1.0 0.6 0.8
Primary zone stoichiometry ratio
1.0
Figure 6. The effect of fuel composition on NOxemissions under staged-combustion
conditions.
£ 500
£ 200
i
100
Boiler simulator _+ +
Liquid Fuel
Petroleum Shale Coal
Boiler simulator • *
Tunnel furnace o a O
_L
i
j
I
i
i
0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Wt % in Fuel
Figure 7. Minimum NOx achievable under staged conditions.
Sonicore produced the smallest (21 //m).
Thus, emissions do not correlate
directly with drop size. Figure 9 shows
that the impact of nozzle design also
depends'on flow field characteristics.
The effect of increasing swirl is seen to
be different for each atomizer. Varia-
tions in swirl level produce a large
change in NOX emissions with the
Monarch nozzle, and relatively mild
change with the Delavan atomizer. The
data in Figure 9b illustratethatdropsize
information alone is not sufficient to
correlate emissions and that the flow
field characteristics can have a signifi-
cant effect with some atomizers.
Staged Combustion
Increasing the mean drop size under
staged combustion conditions effectively
decreases the gas-phase residence
time in the fuel-rich primary zone
because it increases droplet lifetimes.
Table 1 shows the effect of increasing
mean drop size on staged NOX emis-
sions measured in the tunnel furnace
for a given primary zone stoichiometry
ratio and overall excess air level. As
drop size increases, the final emission
level also increases. However, this
increase is less apparent for high nitro-
gen fuels. Figure 10 compares NOX
emissions obtained in the boiler simu-
lator under staged combustion condi-
tions with the Sonicore and the Monarch
nozzles for two different fuels. In both
instances the Sonicore nozzle gives
higher NOX emissions under excess air
conditions, but lower emissions under
staged combustion conditions. Smoke
emissions increased dramatically as
NOX emissions decreased when burning
petroleum-derived residual fuel, and
nozzle design has the most significant
impact on exhaust smoke number in
this case. However, with shale-derived
fuel, the smoke emissions are accept-
able with both nozzles under all primary
zone stoichiometries, but less NO* is
produced when using the nozzle with
the smaller dropsize distribution.
Summary
The results of these small -and pilot-
scale studies of the influence of fuel
properties and spray/flow field interac-
tion indicate that:
1. With liquid fuels, fuel nitrogen
content is the primary fuel com-
position variable affecting fuel NO
formation. NOx emissions in-
crease with increasing fuel nitro-
gen content for petroleum-, shale-,
and coal-derived liquid fuels.
-------
Atomizers
£ 260
O
§: 220
i
180
2.0 3.0 4.0
Percent Excess O2
Figure 8. The effect of spray/flow field interactions.
5.0
Monarch
Peabody
Industrial combustion
Delavan
Son/core
T/#///y%g>^' '-
Todd
O
Delavan
320
300
280
260
§240
i
220
200
0.2 0.4 0.6
Swirl number (SJ
fa)
Figure 9. NOx emissions as a function of drop size.
6
0.8
O Sonicore
Q Monarch 30 gph
A Todd "CDS"
280
^.
Q
CN-
o
I 240
O
200
Peabody
Delavan "Swirl air"
40 60 80
Mean drop size, jum
(b)
100
-------
Table 1. Influence of Mean Drop Size on Staged NOt Emissions3
NO* ppm (O% Oa Dry)
Fuel Nitrogen, %
0007
0 16
0.24
0.51
20 fjm
48
85
no
165
180 pm
75
105
140
170
3Tunnel furnace, primary zone stoichiometry is 0.70.
1400
1200
WOO
Atomizer
• Son/core Monarch f
o a Petroleum N= 0.5,1%
• • Shale N , 2.08%/
Q
S
fi
•8
<0
•5
800
I
IS'
O
p 600
I
*
O
400
200
0
8
6
4
2
0.6 0.8 1.0 1.2
Primary zone stoichiometric ratio
Figure 10. Influence of atomizer type
on staged NOx and smoke
emissions for the shale
and petroleum fuels --
boiler simulation.
Staged combustion as a control
technique is effective for all three
fuels, and minimum levels achiev-
able with high nitrogen alternate
fuels approximate those of petro-
leum fuels with much lower nitro-
gen contents.
2. Decreasing droplet diameter re-
duces both NOx and smoke emis-
sions under staged combustion
conditions.
3. Fuel nitrogen speciation has an
impact on the effectiveness of
staged combustion as a control for
minimizing NOX emissions pro-
duced with high nitrogen fuels.
G. C. England, D. W. Pershing, and M. P. Heap are with Energy and Envi-
ronmental Research Corporation, Santa Ana, CA 92705; J. E. Cichanowicz is
with Electric Power Research Institute, Palo Alto. CA 94304.
W. Steven Lanier is the EPA Project Officer (see below).
The complete report, entitled "Effects of Fuel Properties and Atomization
Parameters on /VOX Control for Heavy Liquid Fuel Fired Package Boilers,"
(Order No. PBS 2-230 715; Cost: $31.50, subject to change) will be available
only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park. NC 27711
•USGPO:1982-659-095-537-
-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Postage and
Fees Paid
Environmental
Protection
Agency
EPA 335
Official Business
Penalty for Private Use $300
PS 0000329
U S ENVIR PROTECTION AGENCY
REGION 5 LIBRAKY
230 S DEARBORN STREET
CHICAGO IL 60604
------- |