United States
 Environmental Protection
 Agency
 Industrial Environmental
 Research Laboratory
 Research Triangle Park NC 27711
Research and Development
 EPA-600/S7-84-076 Sept. 1984
Project  Summary
Evaluation and  Demonstration  of
Low-NOx  Burner Systems  for
TEOR Steam  Generators  -
Design  Phase  Report

G.C. England, M.P. Heap, Y. Kwan, and R. Payne
  A program to translate design criteria
developed in bench- and pilot-scale
combustors to a 16 MW commercial
grade prototype burner that can be
retrofitted to existing thermally en-
hanced oil recovery (TEOR) steam
generators is currently in progress.
  Laboratory- and pilot-scale data have
been used to develop a low-NOx burner.
capable of operating with high-nitrogen
liquid fuels such as crude oils and
synthetic fuels. Emissions of NO« can
be minimized by application of a staged
combustion process in which the first
stage is thermally isolated and provides
long residence time under high temper-
ature, optimally fuel-rich conditions.
The design criteria for this process were
developed in a previous EPA-sponsored
study, carried out in a 21 kW (70,000
Btu/hr) tunnel furnace and applied to
two pilot-scale prototype burners at 0.6
MW (2 x 106 Btu/hr) and 3 MW (10 x 106
Btu/hr).
  The report summarized here describes
results of the design development tests
and compares results of all three scales of
experiment. Firing a heavy fuel oil con-
taining above 0.6 percent bound  nitro-
gen, test results indicate that NO« emis-
sions below 85 ppm (at 3 percent O2)
are  possible while CO and smoke are
low. Translation of the design require-
ments to practical burner designs, and
additional requirements for initial
fuel/air contacting  and second-stage
air addition are  presented  and dis-
cussed. Data obtained over a  scale
range of 30:1 show that residence time
and temperature in the first stage are
the most important scaling parameters
for this low-NOx burner design and that
aerodynamic effects are second order
parameters. Detailed design of the 16
MW prototype burner is also discussed,
which was selected to meet the design
criteria as well as requirements for com-
mercial-grade oil field  equipment. Key
aspects of the hardware design are the
use of commercial primary burner
equipment a refractory-lined first stage
chamber which is regeneratrvely air-
cooled, and variable second-stage air in-
jection to optimize emissions and flame
shape. Discussion of the detailed
design includes retrofit  requirements
and specifications for  primary throat/
atomizer design, first-stage  residence
time and temperature,  refractory lining
design, and preliminary  requirements
for second-stage air injection. The bur-
ner will be briefly evaluated in a test fur-
nace prior to retrofit and performance
testing in a field-operating steam gener-
ator in Kem County, California.
  This Project Summary was developed
by EPA's Industrial  Environmental
Research Laboratory. Research Triangle
Park, NC, to announce key findings of
the research project that is fully docu-
mented in a separate report of the same
title (see Project Report ordering infor-
mation at back).

Introduction
  Enhanced oil recovery processes are
applied to oil field production in order to
extract heavy, viscous crude oil and tar
sands  which cannot otherwise be pro-
duced. A significant fraction of total U.S.

-------
oil reserves require application of enhanced
oil  recovery  in  order to be realized.
Thermally enhanced oil recovery (TEOR)
involves injection of wet steam which is
produced by combusting crude oil in oil
field steam generators, typically ranging
in size from 7 to 15 MW capacity. More
than 90  percent of all oil field  steam
generators in the  United States are
located in California, one-third (approxi-
mately 1000 units) of which are located in
Kern County.  Approximately one-third of
the produced crude oil is consumed by
the steam generator, amounting to over
200,000  barrels  of crude oil consumed
per day at full capacity. The crude oils
which are fired in these steam generators
are typically high in nitrogen (—0.8 to 1.0
percent) and sulfur content. Uncontrolled
emissions of  NOX and  SOX can therefore
reach high levels and potentially worsen
ambient air quality.

  Local legislation regulating emissions
of NO, from  oil  field  steam generators
threatens to  limit oil  production  unless
NOX control  methods are applied. A
ceiling on total NOX emissions from all
steam  generators in  Kern  County has
been established which limits total
emissions to 1979 levels; thus new
generators cannot be brought on line
without reducing emissions from existing
ones.  New generators are required to
have "best available control technology"
(BACT) which may  include Iow-N0x
burners, postflame NH3 injection, or other
postflame treatment methods. In addi-
tion, if ambient N02 level in Bakersfield
exceeds a specified level, the total NO*
ceiling is lowered  to  the equivalent of
0.14 Ib NO,/10s Btu (60 mg/MJ) or 105
ppm NO, corrected to 3 percent oxygen, if
all steamers are in operation. Enhanced
oil  recovery operations are projected to
expand through  1995. The level  of NOX
control required to meet both growth and
air quality goals  has typically been
difficult to achieve with available technol-
ogy while maintaining  acceptable CO and
particulate emissions and practical flame
conditions within the steamer.
  This program  addresses the need for
advanced NO* control  technology for oil-
field steam generators, and concerns the
development of a full-scale  burner
system, for which the concept is based on
fundamental  studies. The burner will be
retrofittable to existing steam generators
and will produce acceptable emissions of
CO and particulate while maintaining a
flame form compatible with the steamer
design. The emission goals for NOx are to
achieve  less than  85  ppm  NOx at 3
percent excess oxygen.
  The major elements  of this program
are:
 1.  Conceptual design development—to
    verify and refine  design scaling
    criteria developed  under previous
    EPA contracts in order  to develop
    hardware designs.
 2.  Detailed burner design and construc-
    tion—to  produce  a commercial-
    grade full-scale burner suitable for
    retrofit on an existing steam genera-
    tor.
 3.  Burner optimization and evaluation—
    to minimize pollutant emissions in a
    field-operating TEOR steam genera-
    tor firing high-nitrogen crude oil and
    evaluate performance over an  ex-
    tended period of time.
  The report discusses the first two
elements of the project,  through comple-
tion of the detailed engineering design of
the full-scale  (16 MW) commercial
prototype burner.


/VOX Formation and Control
  NOX emissions  from  stationary com-
bustors  can  be reduced by combustion
modification techniques which involve
staging  the heat release process.  How-
ever, optimization of these techniques to
control  NOx emissions  from  liquid-fuel-
fired combustors has  proven difficult
because of the limited knowledge of the
controlling phenomena and the problems
associated with the scaling of combustion
systems.
  Many studies have been carried out to
determine the  mechanism of NO forma-
tion  in  flames, both  from  molecular
nitrogen (thermal  NO) and from nitrogen
contained in the fuel (fuel NO). Field test
data and laboratory studies have estab-
lished the importance of fuel NO to the
total emission  of NOX from residual fuel
oil flames. Thermal NO formation is strong-
ly dependent upon temperature, and its
formation can  be  reduced by decreasing
flame temperatures by  techniques such
as the  addition of cooled combustion
products to the combustion air (flue gas
recirculation). However, since  most NO
formed  in residual fuel  oil flames is fuel
NO and its formation  is not strongly
dependent upon temperature, flue  gas
recirculation is not an effective NOX
control  technique. Fuel NO formation is
very sensitive to reactant stoichiometry.
Fuel-rich conditions promote the forma-
tion  of  molecular nitrogen  from fixed
nitrogen species (NO + NH3 + HCN) which
are formed as the liquid fuel decomposes.
Consequently,  staged  combustion  (in-
volving  the formation of a fuel-rich zone
either by delaying fuel/air mixing or by
dividing the combustion air sup_ply into
two and injecting one portion downstream
from the burner)  is the most effective
NOx control technique for residual fuels.
  In a previous study, parameters control-
ling the effectiveness of staged combus-
tion  for  a  wide  range of liquid fuels
containing bound  nitrogen were defined
in a series of bench-scale studies which
were carried out  in a  refractory tunnel
furnace. Exhaust NO levels from a staged
combustion system depend on:
  1. First-Stage Processing. Total fixed
    nitrogen  (TFN)  is  minimized if  the
    residence time of the  fuel-rich
    reactants  is maximized at high
    temperatures and at an optimum
    stoichiometry which is temperature
    dependent.
  2. Mixedness.  Since stoichiometry
    must be controlled, it is  important
    that the reactants are  well mixed.
  3. Second-Stage Processing. Exhaust
    NO levels are the sum  of thermal NO
    formation and conversion (or reten-
    tion) of TFN  species  to NO  during
    heat release in the second stage, and
    the second stage must be designed to
    minimize these effects.
  4. Fuel Properties. The rate of evolution
    of nitrogen species from  the liquid
    fuel droplets has an impact because
    it dictates the time available for N2
    formation in the first zone. This can
    be overcome  by minimizing droplet
    size.
  This information has been used to
construct bench- and pilot-scale combus-
tors having NOX  emissions below 100
ppm (corrected to  0 percent oxygen) and
acceptable  emission levels for CO and
smoke when firing fuels containing
greater  than 0.5 percent fuel-bound
nitrogen.
  The bench-scale  tests indicated that
TFN concentration  is  minimized most
effectively by increasing the temperature
of the fuel-rich zone. The combination of
high first-stage temperature, optimum
stoichiometry, and long first-stage  resi-
dence time resulted in exhaust NOX levels
below 100 ppm regardless of the nitrogen
content of the liquid fuel. To translate this
information into a full-scale (i.e., 18 MW)
burner design  requires information on
how  the controlling parameters scale.
The approach used.was to obtain experi-
mental data with 2 x 106 Btu/hr (0.6 MW)
and 10 x 106 Btu/hr (3 MW) prototype
combustors, and thereby confirm  the
appropriate criteria  for the design  and
fabrication  of the full-scale combustor. In
addition,  isothermal (cold flow) spray/flow^B
field  tests were  conducted  to  assessM
mixing and  atomization characteristics

-------
 produced by full-scale hardware in the
 region near the burner throat.

 Pilot-Scale Development

 Prototype Combustors
  Two prototype low-NOx burners, shown
 in  Figure 1, were  constructed with
 nominal design firing  rates of 2  x  106
 Btu/hr (0.6 MW) (small pilot-scale or SP)
 and 10x 106Btu/hr(3 MW)(intermediate
 pilot-scale or IP).  Each  test burner
 contains three major elements
  1  An air/fuel injection system
  2  A fuel-rich holdup zone.
  3  A secondary air injectiqn section
  The primary zone is constructed from
 high temperature refractories and insula-
 tion to minimize heat  loss through the
 walls  and permit near-adiabatic condi-
 tions  in  the  primary  zone  Fuel and
 primary air are injected at the entry of a
 divergent which forms the initial section
 of the fuel-rich holdup zone. The residence
 time in the fuel-rich zone, calculated at a
                             first-stage stoichiometric ratio of 0.7, is
                             approximately 0.50 sec for the SP burner
                             and 0.42 sec  for the  IP burner. The
                             primary products are directed through a
                             convergent  section  which  serves to
                             minimize radiative heat loss, ensure
                             complete mixing of the primary products,
                             and minimize backmixmg of  secondary
                             air into the primary zone  Secondary air is
                             injected parallel  to  the burner axis
                             annularly around the primary  exit

                             Fuel Injection
                               A number of different commercial
                             atomizers were used in the two combus-
                             tors. Most of the  data  were, however,
                             obtained with  an  ultrasonic, air-assist
                             atomizer and an internal-mixing steam-
                             assist atomizer. At the 0.6 MW scale, the
                             ultrasonic atomizer produced droplets
                             with a narrow  size distribution about a
                             mean  diameter  of 20 um,  and  the
                             internal-mixing atomizer produced drop-
                             lets with a much broader size distribution
                             about a  mean  diameter of 40 //m  At
                                      larger scales the ultrasonic atomizer has
                                      been found to produce droplets with a
                                      mean size greater than 120/um.
                                        Several air/fuel  injection hardware
                                      changes were investigated during the 3
                                      MW burner testing. Table 1 summarizes
                                      the first-stage hardware variations which
                                      were evaluated. Most of the configura-
                                      tions were variations on the fuel/air pre-
                                      mixer concept with different exit geome-
                                      tries. The premixer design is scaled up on
                                      the basis of constant velocity and  swirl
                                      numbers from the 0.6 MW design utilized
                                      in the small pilot-scale prototype.

                                      Second-Stage Design
                                        Second-stage mixing effects were
                                      extensively evaluated  in  the 3-MW
                                      studies, to determine limits of additional
                                      NO reduction possible through variations
                                      in the  second-stage mixing and to
                                      develop  practical information on the
                                      tradeoffs between NOx/smoke emissions
                                      and flame shape. In addition to the first-
                                      stage exit  sleeve  shown in .Table 1,
                                                                        Recycled
                                                                      Flue Products
     Atomization
        Air
                                     Second
                                    Stage Air
                                                                     Burner
       First Stage
          Air
                                                                    (bl Firetube Simulator
                (a) Small Prototype Burner (0 6 MW)
       Fuel & Air
       Injection
           Fuel-Rich
            Holdup
           Sections
 Oil and
Atomizing
  Fluid
. First-stage
 Secondary
Air Injection
                        Second-stage Air
      Fuel/Air
      Premixer
          (c) Intermediate Prototype Burner (3 MW)
                                                               Id) Small Watertube Simulator
Figure 1.    Prototype burners and experimental test furnace

-------
several air injection port configurations
were tested to vary second-stage mixing.
The number of ports, diameter (yielding
two velocity levels), and distance from the
centerline were varied. Figure 2 illustrates
the different hardware configurations
used.
Test Facilities
  The pilot-scale burner  tests were
conducted at two test facilities:
  • Firetube Simulator (FS) (Previous
    Study). The combustion chamber is
    3.2 m long with an internal diameter
    of 0.6 m. The chamber  is composed
    of calorimetric sections  cooled by
    heat transfer fluid and has a nominal
    wall temperature of 230°C. Only the
    SP burner was tested in this facility.
  • SmaH Watertube Simulator ISWS)
    (This Study). The furnace is 5.2 m in
    length and 1.8 m inside diameter, is
                 externally spray-cooled, and has  a
                 partial refractory lining. The total
                 heat  extraction  is  controlled by
                 varying the percentage of internal
                 surface which is covered by refrac-
                 tory. Both the SP and IP burners were
                 tested in this facility.
             Pilot-Scale Results

             First-Stage Parameters
               Figure 3 shows the influence of the
             first-stage stoichiometric ratio (SRi) on
             exhaust emissions from the bench-scale
             studies carried out in the tunnel furnace
             and the  prototype tests with different
             residual oils containing approximately
             0.6 percent  nitrogen. The  curves are
             generally lower but similar in shape to
             those obtained in other staged combus-
             tion studies.  As SRi  decreased, NOx
             emissions decreased rapidly until a
                                      minimum was reached and then increased.
                                      Smoke and CO emissions from the SP
                                      burner were low for SRi > 0.7, but under
                                      more fuel-rich conditions emissions
                                      increased sharply. CO was less than 100
                                      ppm  and smoke number was generally
                                      less  than 5  at the point where the
                                      minimum  NOx level  occurred.  CO and
                                      smoke emissions from the IP burner
                                      were also low for SRi greater than the
                                      point where minimum NOx occurred. CO
                                      and smoke emissions from the IP burner
                                      were also low for SRi greater than the
                                      point  where  minimum  NOx occurred;
                                      however, for many of the configurations
                                      tested, the sharp increase in emissions
                                      below this point was less severe. Differ-
                                      ences in second-stage thermal  environ-
                                      ment  between the  2 x 106 Btu/hr (0.6
                                      MW) firetube simulator and the 10 x 106
                                      Btu/hr (3 MW) small watertube simulator
                                      may contribute to the observed difference
                                      in burnout of CO and smoke. Comparison
Table 1.   First-Stage Hardware Configurations in 3 MW Tests
Configuration

. 	 ...-LJ
jJ / sws
	 T~\

	 v 1 1
^ // x sws
I S 1 1
• • - n

... V1_J
i| /// ^"^ sws
——*\ . s . 	 1
n
First-Stage

— Premixer
— Swirl(S=O.4)
— Parallel Exit


— Premixer
— Swirl (S=0.4)
— Set Back Parallel Exit


tt
— No Swirl
— Parallel Exit

Second-Stage
— Axial Air Injection
— Annulus Injector
— 51 CM First-Stage Exit Diameter
— Axial Air Injection
— Annulus Injector
— 5 1 CM First-Stage Exit Diameter

— Annulus Injector
— 51 CM First-Stage Exit Diameter
                           sws
                           sws
                — Premixer
                — Swirl(S=O.4)
                — Divergent Exit (L/D=2.6)
                — 2 Channel (Low Velocity)
                — Swirl
                — Divergent Exit (L/D=t.1)
                                 Axial Air Injection
                                 Annulus Injector
                                 51 CM First-Stage Exit Diameter
                                 Axial Air Injection
                                 Annulus Injector
                                 36 CM First-Stage Exit Diameter
         v,
sws
— Premixer
— Swirl(S=0.4)
— Divergent Exit (L/D=2.6)
Axial Air Injection
Annulus Injector
36 CM First-Stage Exit Diameter
                L-T
                          sws
                — Premixer
                — Swirl(S=O.4)
                — Parallel Exit
                                 Axial Air Injection
                                 Annulus Injector
                                 36 CM First-Stage Exit Diameter

-------
(al
Secondary Air

   Primary
   Spriman
    Zone
                            SWS
                                         Annular
                          Furnace
                                         Injector
      Secondary Air
         Primery
                            SWS
                                       Inboard Ports
          Zone
                    Furnace
(c)
Secondary Air <—*^JT

   Primary
\ Primary
 j  Zone
                            SWS
                                       Outboard Ports
                           Furnace
Figure 2.   Secondary air injection configurations investigated with 3 MW pilot scale burner
           (ports shown in axial orientation).
of the three sets of results indicates that
concept scaleup shifted the optimum SRi
from 0.8 to 0.7 in going from 70,000
Btu/hr (0.02 MW) to 2 x 106 Btu/hr (0.6
MW), and to 0.56 in  going to 8 x 106
Btu/hr  (2.5  MW). This shift  can be
attributed to  reduced heat losses and
therefore  higher temperatures, and to
changes in fuel/air mixing  characteris-
tics, associated with the  increase in
scale.
  The influence of burner load with the
IP burner is shown in Figure 4 with and
without air preheat. Fuel flowrate was
varied above and below the baseline load
from 2.3 MW to 3.5 MW  throughput.
Minimum NOx emissions increased with
load, ranging from 95 ppm to 77 ppm at
3.5 to 2.3 MW load, respectively, without
first-stage air preheat. With first-stage air
preheat, minimum NOx ranged from 88 to
63 ppm at the same loads,  respectively.
The optimum stoichiometry in the first
stage was not shifted  significantly.
Optimum SRi varied from 0.54 to 0.56
with preheat and from 0.55 to 0.57 with-
out preheat.  Smoke  emissions  were
insensitive to  load.
  The effect of atomizer and air distribu-
tion variations can be seen  in Figure 5.
NOx emissions increased substantially
when the swirl vanes were removed from
the premixer assembly (Configuration III).
Optimum stoichiometry was also shifted
                                 to the fuel-lean side. The steep depend-
                                 ence  of  NO,  on SRi  was  somewhat
                                 reduced at high  primary air preheat for
                                 Configuration  III; however, only limited
                                 data was obtained. Observation of the
                                 flame from the rear of the SWS indicated
                                 that a rather  bright,  narrow jet flame
                                 was formed in the first stage chamber.
                                 Utilization of available volume for  resi-
                                 dence time is probably very poor in this
                                 configuration, accounting for the increased
                                 NOX emission.
                                   The experimental narrow-angle  Y-jet
                                 atomizer and ultrasonic atomizer produced
                                 similar NOx emissions  under equivalent
                                 conditions, although slightly higher than
                                 baseline  in both  cases. Smoke numbers
                                 from  the  Y-jet  and  internal-mixing
                                 atomizer  were 2 for all data shown in
                                 Figure 5.

                                 Second-Stage Effects
                                   The method of secondary air injection
                                 can potentially  impact  exhaust  NOX
                                 levels, and different modes of injection
                                 have been examined in both prototype
                                 burner systems.  The standard annular
                                 (parallel)  air injection system provided a
                                 relatively slow mixing rate and resulted in
                                 long diffusion-type flames. In the small
                                 prototype  burner, rapid mixing was ob-
                                 tained by  injecting the  second-stage air
                                 radially through   a water-cooled boom
                                 inserted along the firetube axis from the
rear. NO, emissions remained essentiaHy
constant when the ratio of air between
the axial and radial injectors was varied:
however,  smoke  emissions were  sub-
stantially decreased as secondary air was
transferred from the parallel injector to
the radial injector. Similar results were
obtained in both prototype burners when
swirl vanes were added to the annular
second-stage air injection system,  sug-
gesting a relative insensitivity of NOx
emissions to changes in  mixing while
providing some measure of control over
flame  shape.  In larger practical-scale
burner systems, however, mechanical
considerations may well preclude the use
of narrow annular passages for secon-
dary air injection. To overcome this
limitation the  potential use of discrete
secondary air  injection ports has  been
investigated in the IP burner. Typical data
are presented in Figure 6a, and show that
with axial injection there is no difference
in minimum NOX between the annulus
and discrete ports; with radial injection to
induce rapid mixing, an increase of some
25 ppm NOx was observed. This configu-
ration produced short intense flames and,
although the level of smoke emission was
generally low, the range of operation over
which  low smoke could be maintained
was extended. The reason  for this
apparent increased sensitivity to mixing
at the larger scale is not immediately
apparent from the data, but is believed to
be due to increased formation of thermal
NOx in local high-intensity combustion
zones close to the  air jets.
  Figure 6 shows the results of doubling
the  velocity of first-stage combustion
products entering the second stage. This
was accomplished by adding a sleeve to
the first-stage exit. The flame  produced
by this configuration was very short with-
out  a  visible smoky trail.  The increase
in velocity by a factor of 2 decreased
flame length by a  factor of 3 to 4, and
increased NOX emissions to 95 ppm from
72  ppm.  Further  optimization of exit
velocity and swirl level may be possible to
improve NO, and flame shape character-
istics with the annular air  injector.
  All the TFN  which  exists in the first
stage  does not  result in NO in the
combustion products of the second stage.
Thus, it is important to design the second-
stage air injection system to minimize NO
formation from TFN species. In addition,
thermal NO formation must be minimized
which  requires that second-stage flame
temperatures must be maintained below
some critical threshold. Heat extraction
between the  fuel-rich zone  and the
secondary  air   injection system allows
some control  over temperature in the

-------
   200
   150
   100
 5
o'
    so
   -  
                i
                »«

                r 100
                                       Q.
                                      O'
      0.4
          0.6
                       0.8
                            1.0
                     0.4
        0.6
                    S/f,
        o.a
      s/?,
                                                                  i.o
Figure 3.    Influence of scale on combustor performance.
         second  stage. Tests in the SP burner
         showed that interstage cooling,  in the
         form of coils around the periphery of the
         second-stage throat (approximately 6
         percent of the heat input was removed),
         decreased the minimum exhaust NOx by
         approximately 20 ppm. Exactly analogous
         results were  achieved when 40 percent
         flue gas recycle was introduced with the
         second-stage air (ANOX = 18 ppm). In both
         tests there was a minor impact on smoke
         emissions; thus, it appears that  at the
         optimum operating  conditions, second-
         stage thermal NOx formation accounts for
         approximately 20 percent  of the total
         exhaust NOx emissions. Use of interstage
         cooling  may become more important in
         subsequent  scaleup  if air  preheat is
         added  and/or rich-zone  heat loss is
         reduced further. Tests in the IP burner at
         3.8 MW in which excess air was varied
         from 0.3 to 4.5 percent produced a 16
         ppm variation in NO,, suggesting about
         the same order of magnitude for thermal
         NOx formation as in the SP tests.

        Without Preheat
              o
                   o
              o    o
               o o
                                            With Preheat
                                                  o
                                                     O
                                                 •O
                                                  o    o
     0.4
           0.5
0.6
        0.7
0.4
0.5
                                                             0.6
0.7
   200
    ISO
 I
 I
 i
roo
    sa
        Primary Air Temp = 52°C
         • 3.5 MW
         O 2.9 MW
         O 2.3MW
        	t
                                          200
                                      150
                                      too
                                           50
                                          _ 93.5 MW (Primary Air Temp •-
                                                     238 ± 5°C)
                                            O2.9 MW (Primary Air Temp -
                                                    239 ± 5°C)
                                            O2.3 MW (Primary Air Temp =
                                                  | 222±20°CJ	
     0.4
           0.5
0.6
          0.7
                                            0.4
                                                     0.5
                                                              0.6
                                                                   0.7
                    Sfl,
                                                       Sfl,
Figure 4.   Influence of load on emissions from IP burner.
Scaling and Full-Scale  Burner
Design

Design Criteria
  The experimental work carried out on
the bench and prototype combustors has
provided an understanding of the control-
ling mechanisms  of NOx formation and
control in staged liquid fuel combustion.
The data suggest that application of these
concepts can be  achieved in a staged
burner which consists  of three major
components with the  following charac-
teristics:
  • Fuel Injection System
   — High quality atomization to  pro-
      duce a small mean droplet size.
   — Rapidly mixed fuel and air.
   — Minimal  heat extraction  and air
      preheat to maximize fuel vaporiza-
      tion rate.
  • First-Stage Fuel-Rich Zone
   — Stoichiometric ratio approximately
      0.7 or  lower  to  minimize  TFN
      concentration.
   — Residence time in excess of 400
      msec  to  allow sufficient  TFN
      decay.
    — Control of heat loss to produce a
      first-stage exit temperature in
      excess of 2600°F (1425°C).
  • Second-Stage Air Injection
    — Partial quenching to minimize
      thermal NO formation in  the
      second stage.
    — Refractory choke to prevent back-
      mixing and reduce heat loss.
    — Intermediate second-stage mixing

-------
      to  reduce smoke formation with-
      out affecting minimum N0«.
The residence time in the fuel-rich zone
required to achieve the  desired  TFN
concentration i% an important practical
design parameter. Figure 7 summarizes
minimum exhaust NO, and TFN data for
the tunnel furnace  and the prototype
burners  firing  in  both the  firetube
simulator and the small watertube simu-
lator  as  a  function of the  first-stage
residence time. At both bench and small
pilot  scales,  increasing the first-stage
residence time significantly decreased
exhaust  NO, because of an associated
decrease in TFN. (Increasing first-stage
residence time allows HCN, NH3, and NO
to decay further toward equilibrium
val ues.)Both the exhaust NOX and the TFN
from the  SP burner were lower than from
the  tunnel furnace because  of the
increased rich-zone temperature asso-
ciated with the increase in scale.

  The first-stage residence time  was
varied in the SP burner/firetube simula-
tor tests by altering the firing rate, and in
the SP burner/watertube simulator tests
by changing the reactor length at a fixed
firing rate. A strong correlation between
the two sets of experiments suggests that
the controlling processes are dominated
by kinetic effects (time and temperature)
and  not  by aerodynamics  and that the
principal  influence  of firing  rate is on
residence time. Clea/ly, increasing firing
rate  also increases  first-stage tempera-
ture, but this effect is small because even
at 2 x 106 Btu/hr (0.6 MW) the prototype
heat  losses were  low.  First-stage exit
temperatures measured at SRi  of 0.7
were 2950°F (1620°C) compared  to
calculated adiabatic flame temperature of
3180°F(1750°C).
  The key  elements in these  design
characteristics are the requirements for a
high-temperature, optimally fuel-rich,
primary  combustion zone with  a  long
residence time. A number of the remaining
parameters are scale-dependent and must
be optimized at the burner scale under
consideration. The  mean droplet  size
produced by a given atomizer design, for
example,  does not remain constant  with
scale, but  tends to increase as  size
increases.  Drop-size distributions  pro-
duced by full-scale  (18  MW) atomizers
under isothermal conditions showed that
the mean drop size was similar for an
ultrasonic atomizer and a Y-jet atomizer
(Figure 8); however,  radial distribution of
fuel mass within the swirling flow  field
was  very different.  While small droplets
are more favorable,  combustion condi-
tions are such that droplet size alone does
    200
    150
     roo
     .
          Primary Air Temp = 52°C
          P Baseline - Swirl Number =i 0.4
          • Zero Swirl
                                   200
                                          150
                                   100
                                           50
                                         Primary Air Temp = 230°C
0.4       0.5      0.6      0.7
              SR,
                                            0.4      0.5
                   0.6      0.7
                                         (a)
    200
     150
     roo
  i   so
         Primary Air Temp = 52°C
  - O Baseline - Ultrasonic; Air;
         69kPa
    • 75 Degree Y-Jet; Steam;
        483 kPa; Smoke No. = 2
                                          2OO -
                                          150
                                          100
50
                                        Primary Air Temp = 230°C
^Baseline - Ultrasonic; Air;
    69kPa
^Internal Mixing; Steam;
    483 kPa; Smoke No. = 2
       0.4
         0.5      0.6
             SR,
                                 0.7
                                             0.4
          0.5
               •0.6
                                                                        7
                                        Ib)
  Figure 5.    Effect of (a) swirl and (b) atomizer type on NOn emissions.
not have  a large  impact  on minimum
NOX. Rather, increased droplet size leads
to greater  sensitivity to those parameters
which control fuel/air mixing and spray
penetration. For practical applications,
optimum results are obtained by selecting
and matching spray  angles and air
register designs to provide for uniform
distribution and rapid mixing.
  Similarly, the overall burner perform-
ance  has been  shown  to be  more
sensitive  to the  mode of secondary
combustion air injection at larger burner
scales.  Rapid  mixing is required to
suppress  smoke formation, but  may
impact minimum  NOX  performance. In
                                  practical applications a major additional
                                  concern is the shape of the second-stage
                                  flame and its interaction with its heat
                                  transfer surroundings,  which can be
                                  strongly  influenced by mixing. Final-
                                  designs of secondary air injection  will
                                  therefore  represent  a compromise be-
                                  tween NOx emissions, flame shape, and
                                  other pollutants such as particulates and
                                  CO.
                                   The impact of fuel type  on the ability of
                                  such burners  to achieve  low-NOx emis-
                                  sions is also important. The influence of
                                  fuel-nitrogen content on the performance
                                  of  the small-scale  Iow-N0x prototype
                                  burner was established  by conducting

-------
    200
    ISO
 ^ roo
 I
 i  so
          2.9 MW
        .  No Preheat
              o      Vv»
        .  O Baseline (annular injector)
          • 8 Axial Ports
          + 8 Radial Ports  '
                                                200
                                              O 1BO
                                                roo
                                                 50
                                                       2.9 MW
                                                       Primary Air Temperature
                                                       = 240°C
                                                                     34m/s
              17m/s   £xit
                       Velocity
      O Baseline (Configuration I)
      • Configuration VII
                                                   0.4
      0.4
                                0.7
               0.5      0.6
                    SR,

                    (a)

Figure 6.    Second-stage mixing effects.
          i).5     0.6
                SR,

                (b)
                                                                         0.7
                 O  IP Burner/SWS (2.4, 3. 3.6 MW)

                 O   SP Burner/SWS (0.6 MW)

                 D   SP Burner/FS (0.6. 0.9 MW)

                 A   Tunnel Furnace (21 kW)
~ 200
^
5
i
1
«
Uj

I
c
   100
                                                 4OO
                                            i
                                            I,
| 200
       0    0.2    0.4    0.6    0.8                    0    0.2   0.4    0.6   0.8

                            Primary Zone Residence Time, sec

Figure  7.    Second-stage NO, production and first-stage TFN decay for bench scale and pilot
            scale.
tests with a distillate oil, residual oils, a
coal liquid, and  propane +  NH3. Data
obtained in these tests are shown  in
Figure 9 which indicates that with the SP
burner minimum NO* emissions are only
weakly dependent upon total fuel nitrogen
content. Figure 9 also summarizes data
obtained  in  the small-scale  tunnel
furnace, and in the larger 10 x 108 Btu/hr
(3 MW) prototype. Concept  scaleup
reduced the dependence on fuel-nitrogen;
this is probably because of the increased
fuel-rich temperature. Thus this low NOX
concept should be capable of producing
NOx emissions less than 100 ppm at 0%
Oz (85 ppm at 3% Oz) even when firing the
high-nitrogen crude  oils typical, for
example, of the Kern County oil fields.


Final Engineering Design
  Two full-scale design concepts evolved
from  the  detailed engineering design
study. The first  utilized a  multilayer,
castable refractory lining to insulate the
first stage, similar to the  pilot-scale
prototypes. Analysis of this design proved
that, although the criteria for temperature
and residence time could be achieved, the
refractory  lining  would  be  excessively
prone to failure due to thermal stresses,
especially during start-up and shutdown.
This is primarily due to the large thermal
inertia of the  thick monolithic refractory
lining. An alternate design was developed
which  utilizes a thin refractory  lining
contained by  a  hallow cylindrical shell
which is cooled by the primary combus-
tion air. This "regenerative"  design,
based in part on a commercial direct-fired
air heater design,  was selected  for
construction after consideration of sever-
al factors, including structural integrity of
the refractory lining, heat loss to sur-
roundings, temperature profile within the
first stage, space-efficiency (compact-
ness), and transient performance. Both
concepts employ  commercially available
primary burner components,  and secon-
dary air is injected  through two sets of
discrete ports, oriented axially and radially,
to allow control over second-stage flame
shape..
  Figure 10 illustrates the basic construc-
tion of  the  regenerative burner design.
The refractory lining  is made up of 15 cm
(6-in.) thick,  90  percent-alumina brick
installed   directly  against  the  inner
stainless steel shell.  Primary air enters
the hollow shell near the burner exit and
flows toward  the primary burner throat,
cooling the shell and gaining  preheat
(hence  the term "regenerative"). Materi-
als and dimensions were selected com-
mensurate with the design operating

-------
conditions outlined in Table 2. A substan-
tial degree of conservatism was incorpor-
ated in the final design. The outer shell is
externally insulated to reduce heat losses
and  personnel safety hazard. Second-
stage air is  injected through a separate
module and can be  mixed radially or
axially at the first-stage exit. After optimi-
zation of the secondary air injection, the
variable components can be fixed to
reduce complexity. Second-stage air
injection will be optimized for velocity and
radial/axial  split.
  Commercial-grade control equipment
has  been  used  to accommodate the
additional control requirements  of the
burner  and provide simple interfacing
with the existing steam generator con-
trols. The control system is  designed to
maintain first-stage stoichiometric ratio
and  flue gas oxygenations through
closed-loop  control. The  system uses a
pneumatic  primary control loop with a
feedback trim signal generated by a pro-
grammable controller and flow measure-
ment instrumentation.
  Construction of the burner was com-
pleted in May 1983, with commissioning
in the field in October 1983. The burner
will be briefly evaluated in an experiment-
al test furnace prior to the field installa-
tion.
to
§
I
   200
   ISO
   700
    50
       Hyvis 04
       Y-Jet
    O Centerline
    • 16° Half Angle
                1OO
                            200
30O
                                                    400
500
                                            200
                                         .8
                                         co
                                         §  100
                                                            I
                             I
                                                                                         Hyvis 04
                                                                                         Ultrasonic
                                                                                      O Centerline
                                                                                      • 8° Half Angle
                                         Figure 8.
                100         200         300

                            Atomizing Pressure. kPa

         Drop size from ultrasonic and Y-jet atomizers.
                                                                                                400
                                                                                                            500

-------
   200
T»
O
   100   -
Legend:
                                    i.o
                                  Fuel Nitrogen, wt. %
Intermediate
Prototype
2.3 MW

•
•

Small
Prototype
0.6 MW
0
0
a
O
Argon
Furnace
21 kW

c
c
A
                                                       Propane + NH,
                                                       Distillate (+ Pyridine)
                                                       Residual Oil
                                                       Shale Liquids

Figure 9.    Influence of fuel nitrogen on second-stage NO, - bench- and pilot-scale data.
                                                                                          10

-------
               Igniter
            Oil Gun

Flame Detector.


Control Panel
Steamer
                          \  Air
         Skid     \        Compressor

             Feedwater Pump
                                       Expansion
                            Combustion    Joint
                            Air Blower
                       Ignitor
                Oil Gun
     Flame Detector
             Skid
                        Feedwater
                          Pump
                                                                     Steamer
Figure 10.    Conceptual burner designs for TEOR steam generators.


 Tattle 2.    First-Stage Design Specifications and Performance

                                                       Horizontal Cylinder
                                                       Flat Wall Entrance
                Configuration                     30-Degree Convergent Exit Cone
Volume
Inside Diameter
Outside Diameter
Overall Length
Refractory Lining Thickness
Primary Burner Throat Diameter
First-Stage Exit Diameter
» at Design
g Point '
IPerformanc
Operatin
Load
First-Stage Stoichiometric Ratio
Gas Temperature
Hot-Face Refractory Temperature
Inner Shell Temperature
Primary Air Velocity at Primary Throat
Exit Gas Velocity
Mean Residence Time
18.3 m3 (650 ft3)
21 3 cm (7.0 ft)
262 cm (8.6 ft)
554 cm (18.0 ft)
15 cm (0.5 ft)
38 cm (1.25 ftj
130 cm (4.25 ft)
18 MW (60 x JO1 Btu/hr)
0.6O
165O°C(3000°F)
156O°C(2850°F)
335°C(650°F)
40m/s(130ft/s)
16m/s(50ft/s)
0.8s
                                                                                         11
                                                                                                 "USGPO:  1984-759-102-10677

-------
      G. C.  England,  M. P. Heap, Y. Kwan, and R. Payne are with Energy and
        Environmental Research Corporation. Irvine, CA 92714-4190.
      W. S. Laniar is the EPA Project Officer (see below).
      The complete report, entitled "Evaluation and Demonstration of Low-NO* Burner
        Systems for TEOR Steam Generators—Design Phase Report," (Order No. PB
        84-224 393; Cost: $28.00, 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:
             Industrial Environmental 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

-------