United States
Environmental Protection
Agency
 Air and Energy Engineering
 Research Laboratory
 Research Triangle Park NC 27711
Research and Development
 EPA/600/S7-86/044 Apr. 1 987
 Project Summary
The  Establishment of  Design
Criteria  for Optimum  Burners for
Application to Heavy Fuel  Fired
Package  Boilers

G. C. England, D. W. Pershing, and M. P. Heap
  This report describes results of a re-
search  program to develop  low-NOx
heavy oil burners for application to in-
dustrial package boilers.  Bench scale
studies have been conducted which
define the mechanisms of NOX forma-
tion and control during staged com-
bustion. These studies showed that
>80% of total NO, emissions can be
due to conversion of fuel-bound nitrogen
to NO,. Conceptual design criteria were
established based on these bench scale
studies for an optimized two-stage
combustion system which minimizes
NOX emissions from conversion of both
fuel-bound  nitrogen  and molecular
nitrogen in the combustion air.
  Volume I  of the report documents
Phase 1 of this program, bench scale
studies which defined optimum condi-
tions for two-stage combustion. This
information  led to a conceptual  two-
stage low-NOx burner design. Volume
II gives results of pilot scale experiments
conducted  in two test facilities  with
nominal capacities of 0.9 and 2.9 MWt,
including tests of commercial burners
for both firetube and watertube boilers.
A wide  range of petroleum-, coal-, and
shale-derived fuels were investigated.
Tests were also conducted with proto-
type advanced low-NOx burners which
demonstrated that NOX emissions <
100 ppm (corrected to 0% O2) could be
achieved almost independently of the
bound nitrogen content in the fuel. The
conceptual design was successfully
scaled from 21 kWt to 0.9 MW, to 2.9
MW, with  similar  NOX emissions
performance.
  This Project Summary was developed
by EPA's Air and Energy Engineering Re-
search  Laboratory, Research Triangle
Park, NC, to announce key findings of
the research project that Is fully docu-
mented in two separate volumes (see
Project Report ordering information at
back).
Introduction
  NOX emissions from stationary combus-
tors can be reduced by combustion modi-
fication techniques which involve staging
the heat release process. However, the
application  of these techniques to re-
sidual-fuel-oil-fired combustors has been
only partially successful because  of a
tradeoff in pollutant emissions. The de-
crease in NOX emissions is often ac-
companied by an increase in paniculate
emissions. The optimization of combustion
modification techniques to control NOX
emissions from liquid-fuel-fired combus-
tors has also proved difficult because of
the limited knowledge of the controlling
phenomena.
  The conversion of the nitrogen  con-
tained in liquid fuels to NOX can contribute
significantly to the total NOX emissions
when such fuels are burned in stationary
combustors. The amount of fuel NO (NO
produced from fuel nitrogen) increases
with increasing fuel  nitrogen  content.
The other contribution to NOX emissions,
thermal NO  (produced from molecular
nitrogen), depends primarily on tempera-
ture and oxygen availability and is  rela-
tively insensitive to fuel properties. Thus,
as fuel nitrogen content increases, fuel

-------
NO constitutes an increasing fraction of
the total emissions. The formation  of
thermal NO depends strongly on tempera-
ture; therefore, combustion firing density
and heat extraction control the amount of
thermal NO produced. Fuel NO formation
is strongly affected by fuel/air mixing
conditions and depends on burner design.
  Many studies have  dealt  with the
mechanism  of fuel nitrogen  oxidation,
resulting in the following simplified ex-
planation. Initially the nitrogen in the fuel
is released during thermal decomposition.
The nitrogen-containing hydrocarbons are
pyrolized,  and the  dominant  nitrogen
compound evolved is probably HCN. The
fate  of this  HCN is then almost totally
determined by the local reactant stoichio-
metry:
  •  Fuel-Lean. The presence of oxygen
     favors the formation  of  NO. For
     homogeneous premixed  condition,
     >75% of the fuel nitrogen could be
     converted to NO. The heterogeneous
     nature  of  liquid fuel  combustion
     usually limits the conversion to <50%
     (unless the fuel nitrogen content is
     low), and this depends on atomizer
     and air delivery system design.
  •  Fuel-Rich. The absence of oxygen
     favors the  formation of N2. Under
     fuel-rich conditions,  fixed nitrogen
     species exist in the gas phase mainly
     as NO, NH3, or HCN, and their relative
     concentrations depend on stoichio-
     metry and temperature.
  For typical fuels and combustion condi-
tions,  equilibrium  total  fixed nitrogen
(TFN) concentrations (the sum of NO,
HCN, and NH3 concentrations) are  <1.0
ppm for mixtures containing between 60
and 70% of the theoretical air require-
ments.  Thus, if  a  combustor could  be
operated to  achieve these near-equilib-
rium TFN levels, fuel nitrogen conversion
would be minimized.
  Staged combustion is the modification
of the combustion process to minimize
fuel NO formation  and,  provided  heat
removal occurs in the  second stage,
thermal NO formation is  also restricted.
In staged combustion, fuel is mixed with
less than the stoichiometric air require-
ment. These conditions maximize  con-
version of fuel nitrogen to N2 prior to the
addition of the remaining  air to complete
combustion. Staged combustion can  be
achieved through burner design by re-
ducing the rate of fuel/air mixing pro-
ducing a fuel-rich zone in the core of the
flame, or by physically dividing the total
airstream,  thus giving  a  well-defined
fuel-rich zone. Under most practical con-
ditions the limit of NOX control for liquid
fuels  by staged combustion has tradi-
tionally been dictated by  the onset  of
smoke emissions.
  The objective of the beginning of this
program was to extend the studies con-
ducted under an earlier EPA contract.
The specific goals were to:
  • Document  conceptual designs  for
    low-emission heavy fuel oil  burners
    suitable  for firetube and watertube
    package  boilers.
  • Determine  the  optimum time/tem-
    perature/stoichiometry history dur-
    ing staged liquid fuel combustion in
    order to minimize NOX and particulate
    emissions in advanced low-emission
    burners and combustors.
  • Develop  optimum atomizers for ad-
    vanced low-NOx combustors.
  • Extend a prior study of fuel effects to
    include a wide range of synthetic
    liquid fuels.

Bench Scale  Studies
  Volume I of the report documents the
bench scale studies conducted  in a 21
kW, down-fired refractory tunnel furnace.
The total range of fuels included in the
fuel effects studies  included 3 distillate
petroleum oils, 13 petroleum-derived
residual fuel oils, a heavy petroleum crude
oil, and 9 coal-and  shale-derived liquid
fuels. An analysis of each of these fuels
is presented in the full report.
  Fuel  NOX  emission was determined
directly in the  tunnel furnace  by sub-
stituting (for combustion air) a nitrogen-
free oxidant consisting of argon, oxygen,
and carbon dioxide (the furnace is dubbed
the "Argon" furnace for  this  reason).
Total and fuel NOX emissions for all of the
fuels tested under excess air conditions
are given in  the full report. These data
were  taken at 5% excess O2, 132°C air
preheat, and 0.53 cmVsec firing rate. An
ultrasonic fuel atomizer was  used with
an air atomization pressure of 103 kPa
(gauge). Fuel  viscosity was maintained at
12 cSt (0.000012  m2s) by appropriate
selection of fuel temperature. Total NOX
refers to emissions  produced with air as
the oxidant. Fuel NOX is defined as the
emissions measured when burning the
fuel in the argon/oxygen/carbon dioxide
mixture. Both total and fuel NOK emissions
increase with increasing  fuel  nitrogen
content. The difference between total and
fuel NOX  is defined to be thermal NOX
which is approximately constant for all of
the heavy residual oils. In general, alter-
native fuels have higher nitrogen contents
than petroleum-derived fuels and there-
fore produce  higher fuel NOX emissions.
With few exceptions, fuel NOX emissions
from  both  alternative and  petroleum-
derived fuels appear to correlate well on
the basis of fuel-nitrogen content. The
total NOX emissions from some alternative
fuels  are higher than from pure petro-
leum-derived fuels  of similar nitrogen
content which suggests  a greater pro-
duction of thermal NOX. Distillation results
indicate that both total mass and nitrogen
are evolved for these alternative fuels at
significantly lower temperatures than for
the petroleum-derived residual fuel oils.
Thermal NO formation may be increased
for these fuels because of higher peak
flame temperatures in the early portion
of the flame since more  fuel is burned
prior to vitiation of the combustor air with
recirculated combustion  products,  or
because the relatively high vaporization
rates  produce higher  local combustion
intensities.
  TFN levels were measured at the  exit
of the fuel-rich first zone as a function of
first-stage stoichiometric ratio (SR,) for
the Alaskan diesel oil, three petroleum-
derived liquid fuels, three alternate liquid
fuels, and methane doped  with 0.79%
nitrogen  (by weight)  as  ammonia.  All
fuels investigated showed similar trends.
TFN concentration  decreased with  de-
creasing SR, and minimized at about  SR,
= 0.8. TFN increased significantly as the
primary zone stoichiometry was further
reduced. TFN for both the ammonia-doped
methane and  the fuel oil with  similar
nitrogen  content showed similar char-
acteristics  under fuel-lean  conditions.
However, below SRi = 0,7 the heavy oil
produced significantly less TFN than the
gaseous fuel. These results suggest that,
although the impact of fuel composition
is reduced under staged conditions, there
can still be a significant impact on NOX
emissions.
  The ultimate  level of exhaust NOX
produced during staged combustion re-
sults from conversion of TFN  existing the
first stage and any thermal NOX produc-
tion  which  occurs  during  burnout.
Second-stage  thermal NOX  production
was not considered to be significant in
this study because changes  in heat ex-
traction during burnout had  almost no
effect on final emissions. Minimum ex-
haust NOX and TFN  concentration at the
exit of the first stage were measured as a
function of total fuel-nitrogen content for
seven liquid fuels  and the  methane/
ammonia mixture. In general, the mini-
mum exhaust NOX concentration occurred
at SRi = 0.78 ± 0.02. NOX emissions and
TFN under  staged  conditions correlate
well with total fuel-nitrogen content, but

-------
the slope is significantly less than under
excess air conditions.  Only the SRC-II
heavy distillate exhibited unusually high
exhaust NOX emissions, and this was the
direct result of high first-stage TFN yield.
NO, NH3, and HCN  were substantially
higher at the optimum  stoichiometry for
this fuel than for fuels with comparable
nitrogen content. The increased TFN may
be the result of  basic chemical  bonding
differences between the parent coal from
which the SRC-II was derived and petro-
leum liquids. Since  this fuel  was  not
tested at pilot scale, this anomalous be-
havior could not be confirmed; however it
is well outside the limits of normal ex-
perimental errors. The results do indicate
that high  exhaust NOX  levels  can be
directly associated with high first-stage
TFN levels.
  TFN concentrations were found in
excess of equilibrium levels. It has been
suggested that increasing the tempera-
ture of the primary zone would prove
beneficial. The impact of first-stage heat
extraction was investigated in the bench
scale tunnel  furnace by inserting or re-
moving cooling coils to extract heat along
the walls of the furnace, and by inserting
or removing choke sections in the furnace
which reduce radiative heat loss. The
results presented in  the full report, ob-
tained with crude shale oil, demonstrate
the impact of first-stage heat loss on the
fate of fuel nitrogen. The hot conditions
refer to the furnace without primary zone
cooling and with  a choke section installed
at the exit of the first stage to minimize
radiative heat loss from the first stage.
Cooling coils were  added to  the first
stage, and the radiation choke  was re-
moved for the cool furnace conditions. Ex-
haust NOX and  primary zone gas tem-
perature are  reported  as  a  function of
stoichiometric ratio for these two furnace
conditions. Minimum NOX emissions were
reduced for the hot furnace case, and the
optimum stoichiometry  was  shifted
toward more fuel-rich conditions.
  Axial profiles in the report explain this
shift in the minimum  emission levels.
Heat extraction from the first stage  im-
pacts the  rate of decay of TFN  species.
Under cold conditions, both NO and HCN
essentially freeze; whereas, without heat
extraction, the initial rate of decay for all
three species is  much faster, leading to
low TFN concentrations at the exit of the
fuel-rich first  stage. Heat extraction also
affects the rate of CO oxidation.
  The bench  scale results demonstrate
that a two-stage  combustion system can
be driven toward equilibrium levels by
increasing the temperature in the fuel-
rich first stage. Since the ultimate level of
TFN at the  exit of the first stage is
kinetically limited, the minimum NOX level
is  controlled by  a tradeoff  between
temperature  in the rich stage and the
time  required  for  TFN  decay at that
temperature. This suggests that  an
optimum  staged  combustion system
should minimize heat loss from the first
stage to maintain gas temperatures as
close to adiabatic as possible. Conditions
in  the  second  stage should also  be
optimized. The oxidation of TFN to NO is
much lower  in diffusion flames than in
premixed flames; therefore, mixing should
be  controlled to provide an attached dif-
fusion flame in  the second  stage. This
suggests  that relatively slow  mixing
should take place between second-stage
air  and first-stage combustion products.
Mixing should also be controlled to mini-
mize peak second-stage flame tempera-
tures, thereby minimizing thermal NOX
formation.  Studies undertaken  in  the
tunnel furnace under an earlier EPA study
showed that two methods of secondary
air  injection (radial versus axial injection)
providing  different  mixing rates in the
second stage produced similar NOX emis-
sions,  suggesting that NOX emissions
were insensitive to second-stage mixing
in that experiment. The scale of the bench
scale tunnel furnace is so small that
mixing was probably rapid in both cases.
However,  results obtained at pilot scale
showed that mixing can have a significant
effect on NOX emissions. Since mixing is
a  scale-dependent phenomenon,  its
significance can be expected to increase
as scale increases. Second-stage mixing
effects were investigated in this study in
the pilot scale experiments.
  In practice there are  many possible
approaches to controlling mixing,  includ-
ing variation  in  air  injection velocity,
injection angle, number of injectors, and
swirl intensity. The approach taken for a
specific application must provide  mixing
which most closely achieves the  mixing
goals of the conceptual design  while
maintaining  a  flame  shape which is
commensurate with furnace geometry.

Pilot Scale Studies
  The pilot  scale  studies described  in
Volume II were aimed at identifying the
impact of fuel type and operating condi-
tions on  NOX emissions  using conven-
tional package boiler burner  equipment,
and developing advanced Iow-N0x burner
concepts based  on the optimization  of
two-stage combustion. Experiments were
conducted in two test facilities: a 0.9 MW
test furnace which simulates the thermal
environment and geometry of a firetube
package boiler furnace (Firetube Simula-
tor), and a 2.9 MW test furnace which
simulates the thermal environment of a
watertube package boiler (Small Water-
tube Simulator — SWS).
  The control of  NOX  emissions  from
practical boiler/burner systems is usually
limited by a tradeoff between decreased
NOX  emissions  and increased  smoke
emissions.  The  report  shows NOX vs.
smoke emissions produced in the  SWS
firing  a California  No.  6 oil with two
different burners. Results were obtained
with  a commercially available burner
which is  widely applied on  packaged
watertube boilers. The combustion  air is
divided into  two coaxial streams  with
variable swirl and flow through  each
passage. Air  and fuel are mixed in the
refractory throat of the burner and the
flame is stabilized by recirculation of hot
partial products of combustion. The data
show that, as the  minimum NOX  level
decreases through adjustments  to the
burner and excess 02 level, smoke emis-
sions increase. The tests indicate that
minimum NOX emissions are similar for
similar smoke levels regardless of burner
settings. Data are shown with increased
furnace insulation to raise gas tempera-
ture  at the  exit  of the furnace.  This
resulted in  slightly  lower  achievable
minimum NOX levels for equivalent smoke
emissions,  suggesting that the burnout
of soot which is formed  in the flame is
enhanced  by the  increased furnace
temperature.
  Data were  also  obtained with a dis-
tributed mixing  coal burner (designed
under a prior EPA contract ) which was
modified for oil firing. The burner is similar
to the commercial burner except for two
major differences: (1) throat diameter is
larger to decrease air velocity at the throat,
and (2) a portion of the combustion air is
added through  four ports  around the
burner throat. This permits the burner to
be operated with less than the stoichio-
metric air requirements at the burner
throat. The results is a  fuel-rich zone
within the core of the flame with  pro-
gressive air addition farther downstream.
Minimum NOX was again limited by the
onset of smoke emissions regardless of
burner settings.  Comparison  of results
obtained with the two  burners shows
that about the same minimum level  of
NOX emissions (150 ppm) was achieved
at a Bacharach smoke number of  6  in
spite  of the  substantial difference  in
fuel/air mixing. This suggests that there
is a  limit  to  the level  of NO, control

-------
possible  with  conventional  burner
systems
  The bench scale studies indicated that
staged combustion  can be  further  op-
timized by providing a physically separate
first stage. This permits the optimization
of conditions in the  first stage without
affecting the second stage, or vice versa.
The bench scale studies suggested that
an optimum two-stage combustion system
should  be designed  according  to  the
criteria shown in Tahie 1 The combustion
process is divided into three zones: a fuel
injection zone,  a fuel-rich holdup zone,
and a second stage burnout zone.
  The table shows the desired optimum
conditions for each  zone.  Fuel injection
should provide rapid mixing of small fuel
droplets with the primary air and mini-
mum heat extraction in order to drive the
liquid fuel into the gas phase as quickly
as possible. The fuel-rich  holdup zone
should provide a plug flow residence time
of >200 ms with minimum heat extraction
in order to maximize TFN  decay  Condi-
tions  in the second-stage burnout zone
should provide slow mixing of the second-
stage air  with  the primary  combustion
products in order to minimize oxidation of
TFN to NOX, and peak flame temperatures
should not exceed  1500°C in order to
minimize thermal NOX formation.
  Two prototype Iow-N0x burners, shown
in  Figure 1,  were  constructed  with
nominal design firing rates of 0.6 and 2.9
MW.  The first  stage was  constructed
from  high-temperature refractories  and
insulation to minimize heat loss through
the walls of the first stage, permitting
near-adiabatic  temperatures.  Fuel  and
primary air were injected at the entry of a
divergent quarl which forms the initial
 section  of  the fuel-rich holdup zone.
Calculated plug flow residence  time in
the  fuel-rich  zone  for  a first-stage
stoichiometric ratio of 0.7 was about 0.56
sec for the small prototype burner,  and
0.43  sec  for the intermediate prototype
burner. The primary combustion products
were directed through a convergent sec-
tion which serves to minimize radiative
heat loss, ensure complete mixing of the
primary products, and minimize back-
mixing of the secondary air into the pri-
mary zone. Secondary air was injected
parallel  to  the  burner  axis  annularly
around the primary  exit. Two  different
commercial  atomizers were  used,  an
ultrasonic air-assist atomizer and an in-
ternal mixing steam-assist atomizer.
   Also shown in Figure 1 are the Firetube
Simulator and Small Watertube Simulator
furnaces. The  combustion  chamber of
 Table 1.    Design Criteria for Scaleup of Low-NO* Concept
A. Fuel Vaporization Zone

   1) Fuel/Air Mixing


   2) Fuel Injection


   31 Heat Extraction

   4) Geometry

B. Fuel-Rich Holdup Zone

   1} Residence Time

   2) Heat Extraction


   3) Geometry

   4) Stoichiometry

C. Second-Stage Burnout

   1} Secondary Fuel/Air
     Mixing

   2) Heat Extraction
Rapid mixing of fuel and air for maximum utilization of residence
time.

Finely atomized droplets to allow rapid vaporization of fuel. 98% of
droplets < 50 nm.

Minimum heat loss to maximize droplet vaporization.

Not defined.
Mean residence time >200 msec at high temperature.

Minimize heat extraction to maximize rate of decay of XN species
to N2. Preferably exit temperature >1400K

Designed to maximize true residence time/volume ratio.

Variable from 0.4 to J.I 6. Optimum at 0.65-0.85.
Slow to minimize peak flame temperature, thus minimizing second
stage thermal NOX formation.

Optimize to limit peak flame temperature <1811K.
the Firetube Simulator is 3.2 m long with
an internal diameter of 0.6 m. The furnace
is  comprised  of  calonmetric sections
cooled by heat transfer fluid, and the wall
temperature  is nominally  230°C. The
furnace of the Small Watertube Simulator
is 5.2 m long with an internal diameter of
1 8 m. The furnace is  externally spray-
cooled and has a partial refractory lining
to control total heat extraction
  The report shows the influence of first-
stage stoichiometric ratio (SR, on exhaust
emissions from the bench scale studies
carried  out  in the tunnel  furnace and
from the prototype tests with residual oils
containing about 0.6%  nitrogen. The
dependence of NOX  emissions on SR,
was similar  in  all three experiments.
Smoke and CO emissions from the small
prototype burner were  low for SR!>0.7,
but increased sharply  as SR, was de-
creased further. CO emissions were <100
ppm,  and  smoke  number was generally
<5 at the point where the minimum NOX
level occurred. CO and  smoke emissions
performance for the  intermediate  proto-
type was similar. The minimum NOX level
ranged from about 89 ppm  in the tunnel
furnace to 84 ppm m the small prototype
burner and 74 ppm in the  intermediate
prototype  burner. The  optimum  SR,,
appears to shift downward as scale  in-
creases, going from 0.8 for  the tunnel
furnace to 0.7  in the small prototype
                burner and to 0.56 in the intermediate
                prototype burner.  This shift in the opti-
                mum SR, and minimum NOX level is most
                likely due to reduced first-stage heat loss
                associated with  the increase in  scale
                which  leads to higher  temperatures in
                the  first stage. Equilibrium calculations
                (free energy minimization) confirmed that
                the optimum SR, should move more fuel-
                rich with increased temperature; how-
                ever, the TFN concentrations which were
                measured in the tunnel furnace and small
                prototype burner  are greatly (about  2
                orders of magnitude) in excess of equilib-
                rium values  Hence, even with first-stage
                exit temperatures of 1600°C at SR, = 0.7
                and a first-stage residence time of 500
                ms,  kinetic  limitations are significant.
                Decreasing SR, reduced first-stage NO;
                but, below about  SR, = 0.8,  significant
                amounts of NH3 and HCN were formed.
                Thus, as in the bench scale studies, there
                exists a minimum in the exhaust NOX
                because of  competition  between de-
                creased first-stage NO and  increased
                oxidizable nitrogen species. Other than
                shifting the optimum SR,, concept scale-
                up did little  to alter this  competition or
                the TFN species distribution.
                  The residence  time  in the fuel-rich
                zone  to achieve  the desired TFN con-
                centration is an important practical design
                parameter since it largely determines the
                size of the precombustor. Figure 2 sum-

-------
                                             Second Stage
                                                 Air
                                                       Recycled
                                                     Flue Products
         First Stage
            Air
             (a) Small Prototype Burner (0.6 MW1425°C. Further, the rate of  NO
                                        decay decreases  significantly  when  the
                                        NH3 concentration falls to zero, indicating
                                        that TFN speciation is important in first-
                                        stage processing.  These calculations sup-
                                        port the concept that reactions in  the
                                        form: NO + NH, — N2 are controlling  the
                                        fuel-rich TFN reduction, and  that it is
                                        essential to have  both high temperatures
                                        (>1425°C) and  a  proper  TFN  species
                                        distribution in order to achieve very  low
                                        exhaust NOX levels. The calculations also
                                        indicate the  need for a  long  residence
                                        time  in the primary zone, but that  the
                                        required time to achieve a given level of
                                        TFN decreases with  increasing first-stage
                                        temperatures.
                                         Figure 3- summarizes  NOX emissions
                                        from  each  of the  pilot- and bench-scale
                                        experiments as a function of fuel nitrogen
                                        content. Figure 3a  shows NOX from  the
                                        commercial burners tested in the Firetube
                                        Simulator  and the Small Watertube
                                        Simulator,  the 3  MW distributed mixing

-------
    200
l/oo

S
<0
-C
Uj
I
I
c
5    „
O IP Burner/SWS (2.3. 2.9 &
                 3.5 MWJ
O SP Burner/SWS (0.59 MWJ
D SP Burner/ FS (0.59 & 0.88
          A Tunnel Furnace (21 kWJ
                                      0  400
                                      as
                                      o
                                      ^
                                      •6
                                        O
                                        -C
                                        ki


                                        3
                                        .5
                                        «J

                                        u.
                                  200
     0.2     0.4    0.6     0.8          0     0.2
                  Primary Zone Residence Time, sec
                                                            0.4     0.6    0.8
Figure 2.
  Effect of first-stage residence time on exhaust NO, emissions and TFN decay
  for bench-scale, small pilot-scale (SP). and intermediate pilot-scale (IP) staged
  combustors.
program to systems of different scale in
order to achieve NOX emissions <100
ppm even when firing high-nitrogen fuels.
Successful scaling criteria for advanced
low-NOx burners were  developed  and
verified over a range in scale of 750 to 1.
burner, the bench scale tunnel furnace,
and the two pilot-scale prototype burners.
Data from the tunnel  furnace  showed
that NOX emissions were > 1000 ppm for
fuels with nitrogen content X).6%. How-
ever, the application and optimization of
staged combustion dramatically  reduced
NOX emissions to 190 ppm at this same
level of fuel-bound  nitrogen. NOX emis-
sions from the Firetube Simulator with
commercial  burner A  were also high,
reaching  1060 ppm when burning the
high nitrogen (2.08 wt % nitrogen) crude
Paraho shale oil. In contrast, NO,; from
commercial burner B in the same furnace
were only 550 ppm when firing the same
fuel. This illustrates the strong  depen-
dence of NOX emissions on burner design
for conventional burner systems.  The
application of staged combustion to com-
mercial burner A further decreased  NOX
emissions to 490 ppm when burning the
shale oil, but at the expense of increased
smoke emissions
   NOX emissions from a  3 MW com-
mercial  watertube burner in  the Small
Watertube Simulator were much lower
than either of the  commercial  firetube
burners for fuels with  equivalent bound
nitrogen  content. NOX emissions were
very strongly dependent on fuel nitrogen
content with all of the conventional burner
systems  which  were  tested. The  NOX
                               emissions  from the  optimized  tunnel
                               furnace and the prototype pilot scale
                               burners were influenced by fuel nitrogen
                               to a much lesser extent and were only a
                               fraction of the emission level from the
                               commercial burners.
                                 Figure 3b shows  the fuel nitrogen
                               dependence more clearly for the tunnel
                               furnace and for the pilot-scale prototype
                               burners. Data are shown for various fuels
                               including a distillate oil (with and without
                               pyridine),  residual oils, a  coal-derived
                               liquid, and a propane/NH3  mixture.  The
                               data  indicate  that NOX is  only  weakly
                               dependent on  nitrogen content with the
                               small and intermediate prototype burners.
                               Comparison of results obtained with the
                               NH3-doped gas, pyridine-doped distillate
                               oil, and actual residual oils shows that
                               differences in NOX due to fuel volatility
                               are small in absolute terms, especially for
                               systems with short residence time in the
                               first stage where the characteristic time
                               for the vaporization of fuel approaches
                               the time available for  TFN decay. Scale-
                               up of  the concept resulted in dependence
                               of NOX emissions on fuel nitrogen level.
                               This  is probably  due to the  increased
                               temperature in the fuel-rich zone because
                               of the increase in scale.
                                  In  conclusion, these results indicate
                               that it should be possible  to  apply the
                               Iow-N0x burner concept developed in this

-------
    woo
      800
  I
  O
      600
     400
     200
                          3.0% Excess 02
                 21 kW  j
                 Tunnel /
                Furnace' (Unstaged)
                             3.2
                           Commercial
                         Watertube Burner
  MW Commercial
Firetube Burner B


  MW Commercial
Firetube Burner A
        Unstaged

        •— Staged

      21 kW tunnel
        Furnace
    (Optimum Staged) —
                        \
                              Prototype Burners
                               (0.6 & 2.9 MWJ
                                                                  200
                      1.0           2.0

                  Fuel Nitrogen (by Wt), Percent



                              (a)
                                     3.0
                                                               I
                                                  O  100 -
                                                              O
                                                                            1,0

                                                                         Wt % Fuel Nitrogen
                                                          Legend:

                                                            Intermediate
                                                              Prototype
                                                              2.3 MW,
                                                           Small
                                                          Prototype
                                                          0.59 MWt
                                                             O
                                                             O

                                                             a
                                                             O

                                                            (b)
                                                                                                            2.0
 Argon
Furnace
 21 kW,
  €
  n
  O
Propane + NH3

Distillate
(+ Pyridine)

Residual Oil

Coal Liquids

Shale Liquids
Figure 3.
Effect of fuel nitrogen content on NO* emissions from: fa) all bench- and pilot-scale burners tested, and (b) low-NO, prototype
burners.

-------
     G.  England.  D. Pershing, and M. Heap are with  Energy and Environmental
      Research Corp., Irvine, CA 92718.
     Joseph A. McSorley is the EPA Project Officer (see below).
     The complete report consists of two volumes, entitled "The Establishment of
      Design Criteria for Optimum Burners for Application  to Heavy Fuel Fired
      Package Boilers:"
       "Volume I.  Laboratory Scale Tests,"(Order No.  PB 87-145 637/AS; Cost:
      $18.95, subject to change).
       "Volume II. Pilot Scale Tests," (Order No.  PB 87-145 645/AS; Cost $18.95,
      subject to change).
     The above reports 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
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Official Business
Penalty for Private Use $300
EPA/600/S7-86/044
             0000329    PS
                                             *GENCT

-------