oEPA
         United States     Industrial Environmental Research  EPA-600/7-79-050b
         Environmental Protection  Laboratory          February 1979
         Agency       Research Triangle Park NC 27711
Proceedings of the Third
Stationary Source
Combustion Symposium;
Volume II.
Advanced Processes
and Special Topics

Interagency
Energy/Environment
R&D Program Report

-------
                  RESEARCH  REPORTING SERIES


 Research reports of the Office of Research and Development, U.S. Environmental
 Protection  Agency, have been grouped into nine series. These nine broad cate-
 gories were established to facilitate further development and application of en-
 vironmental technology. Elimination  of  traditional grouping was consciously
 planned to foster technology transfer  and a maximum interface in related fields.
 The nine series are:

     1. Environmental Health Effects Research

     2. Environmental Protection Technology

   .  3. Ecological Research

     4. Environmental Monitoring

     5. Socioecondmic Environmental Studies
                   •»

     6. Scientific and Technical Assessment Reports (STAR)

     7. Interagency Energy-Environment Research and Development

     8. "Special" Reports

     9. Miscellaneous Reports

 This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
 RESEARCH AND DEVELOPMENT series. Reports in this series result from the
 effort funded under the 17-agency Federal Energy/Environment Research and
 Development Program. These studies relate to EPA's mission to protect the public
 health and welfare  from adverse effects of pollutants associated with energy sys-
 tems. The goal of  the Program is to assure the rapid development of domestic
 energy supplies in  an environmentally-compatible manner by providing the nec-
 essary environmental data and control technology. Investigations include analy-
 ses of the transport of energy-related  pollutants and their health and ecological
 effects; assessments  of, and development of, control technologies  for energy
 systems; and integrated assessments of a wide'range of  energy-related environ-
 mental issues.
                        EPA REVIEW NOTICE
This report has been reviewed by the participating Federal Agencies, and approved
for publication. Approval does not signify that the contents necessarily reflect
the views and policies of the Government, nor does mention of trade names or
commercial products constitute endorsement or recommendation for use.

This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.

-------
                              EPA-600/7-79-050b

                                   February 1979
     Proceedings of the  Third
Stationary Source Combustion
             Symposium;
 Volume II. Advanced Processes
         and Special  Topics
            Joshua S. Bowen, Symposium Chairman,
                     and
            Robert E. Hall, Symposium Vice-chairman

              Environmental Protection Agency
              Office of Research and Development
            Industrial Environmental Research Laboratory
            Research Triangle Park, North Carolina 27711
               Program Element No. EHE624
                   Prepared for

           U.S. ENVIRONMENTAL PROTECTION AGENCY
             Office of Research and Development
                Washington, DC 20460

-------
                                 PREFACE
       These proceedings document more than 50 presentations and discussions
presented at the Third Symposium on Stationary Source Combustion held March
5-8, 1979 at the Sheraton Palace Hotel, San Francisco, California.   Sponsored
by the Combustion Research Branch of the EPA's Industrial  Environmental
Research Laboratory - Research Triangle Park, the symposium papers  emphasized
recent results in the area of combustion modification for NOX control.  In
addition, selected papers were also solicited on alternative methods for
NOX control, on environmental assessment, and on the impact of NOX  control
on other pollutants.

       Dr. Joshua S. Bowen, Chief, Combustion Research Branch, was  Symposium
Chairman; Robert E. Hall, Combustion Research Branch, was  Symposium Vice-
Chairman and Project Officer.  The welcoming address was delivered  by Clyde
B. Eller, Director, Enforcement Division, U.S. EPA, Region IX and the opening
Address was delivered by Dr. Norbert A. Jaworski, Deputy Director of IERL-RTP.

       The symposium consisted of seven sessions:

       Session I:
       Session II:


       Session III:


       Session IV:


       Session V:



       Session VI:


       Session VII:
Small Industrial, Commercial and Residential  Systems
Robert E. Hall, Session Chairman

Utilities and Large Industrial Boilers
David G. Lachapelle, Session Chairman

Advanced Processes
G. Blair Martin, Session Chairman

Special Topics
Joshua S. Bowen, Session Chairman

Stationary Engines and Industrial Process Combustion
Systems
John H. Wasser, Session Chairman

Fundamental Combustion Research
W. Steven Lanier, Session Chairman

Environmental Assessment
Wade H. Ponder, Session Chairman
                                    ii

-------
                                VOLUME II

                             Table of Contents

                     Session III:  Advanced Processes

                                                                         Page

"The Influence of Fuel Characteristics on Nitrogen Oxide
Formation - Bench-Scale Studies,"  M. P. Heap, D. PerShing, G. C.
England, J. W. Lee and S. L. Chen ............. ......    3
"The Control of Pollutant Formation in Fuel Oil Flames -
The Influence of Oil Properties and Spray Characteristics,"
G. C. England, M. P. Heap, R. T. Horton, D. W. Pershing and
G. Flament  ..... ....................... ..  .    41

"The Generalization of Low Emission Coal Burner Technology,"
D. M. Fallen, R. Gershman, M. P. Heap and W. H. Nurick  ........    73

"Alternate Fuels and Low NO  Tangential Burner Development
                           jL
Program," R. A. Brown .........................   Ill

"Pollutant Formation During Fixed-Bed and Suspension Coal
Combustion," D. W. Pershing, B. D. Beckstrom, P. L. Case
and G. P. Starley ............. ..............   147

"Advanced Combustion Concepts for Low BTU Gas Combustion,"
B. A. Folsom, C. W. Courtney, T. L. Corley and W. D. Clark  ......   163

"Catalytic Combustion System Development for Stationary
Source Application," J. P. Kesselring, W. V. Krill, E. K.
Chu, and R. M. Kendall  ........................   207


                       Session IV:  Special Topics


"ERPI Low Combustion NO  Research," D. P. Teixeira (Abstract)*  ....   243
                       2w

"Flue Gas Treatment Technology for NO  Control," J. D. Mobley .....   245
                                     j&

"Chemiluminescent Measurement of Nitric Oxide In Combustion
Products," B. A. Folsom and C. W. Courtney  ....... . ......   283
(*) See Volume V.
                                         ill

-------
    SESSION III:

ADVANCED PROCESSES
  G.  BLAIR MARTIN
     CHAIRMAN

-------
                THE INFLUENCE OF FUEL CHARACTERISTICS
                    ON NITROGEN OXIDE FORMATION
                        - BENCH-SCALE STUDIES
                                By:

M. P. Heap, D. W. Pershing, G. C.  England, J. H. Lee, and S. L. Chen
            Energy  and Environmental  Research Corporation
                         Irvine, California

-------
                                  ABSTRACT

     Results of two bench-scale experimental studies are presented showing
the effect of the physical and chemical characteristics of both liquid and
solid fuels on fuel nitrogen conversion.  The formation of thermal NO was
prevented by the use of artificial oxidant mixtures (Ar/C02/02> free from
molecular nitrogen.  Results are presented showing the effect of fuel nitro-
gen content, initial fuel/air mixing process, and droplet/particle size for
both staged and unstaged conditions.  The effectiveness of staged combustion
as an NOX control technique appears to be dependent upon fjuel nitrogen
volatility.

-------
                               ACKNOWLEDGEMENT
      The authors are pleased to acknowledge the assistance of several of
their colleagues, R. C. Horton, P. C. Lackey and J. Small in the conduct of
the experiments and to G. B. Martin and W. S. Lanier of the Environmental
Protection Agency for their support and encouragement.

-------
                                  SECTION 1
                                INTRODUCTION

     Fossil fuels contain trace specie which may form pollutants during the
combustion process, e.g., sulfur oxides, nitrogen oxides and fine particu-
late matter.  The oxidation of chemically-bound nitrogen during the combus-
tion of fossil fuels provides a significant source of nitrogen oxides, and
sulfur emissions are directly related to the fuel sulfur content.  Although
this paper is concerned with the fate of fuel nitrogen, the studies from
which the results are drawn include an assessment of fuel characteristics on
the formation of other pollutants, particularly the size distribution and
trace metal content of fine particulate emissions.
     The combustion of coal and residual fuel oils accounts for the major
fraction of nitrogen oxide emissions from stationary combustion sources, and
since it is generally recognized that the oxidation of chemically-bound nitro-
gen contained in these fuels provides a significant fraction of the total
emissions, the factors affecting the fate of fuel nitrogen compounds are of
major importance in the development of low emission combustion systems.  The
nitrogen content of coals and petroleum varies considerably.  Typical values
range from 0.2 to 0.5 percent for residual fuel oils, and 1.1 to 1.7 percent
by weight for coals although for both fuels examples can be found outside
these ranges.  Alternate liquid fuels derived from shale or coal have
higher nitrogen contents than conventional petroleum-derived liquid fuels
and thus, have the potential to produce significant nitrogen oxide emissions
if burned without the application of control techniques.
     An effective method of controlling NO  emissions from combustors fired
                                          x
with nitrogen-containing fuels is to modify the combustion process in such
a way as to 'ensure that heat release occurs in two stages.  Initially, the
fuel reacts under oxygen-deficient conditions which serves to maximize the

-------
production of N  from the bound fuel nitrogen specie.  Then, provided heat is
lost from the first stage, NO production during burnout in the second stage
is minimized.  The optimum design of a staged combustor requires a knowledge
of the fate of the fuel-bound nitrogen as a function of the characteristics
of the fuel and the combustion process.
     The effect of combustion on the fate of nitrogen contained in solid and
liquid fuels is illustrated by the simplified flowchart shown in Figure 1.
Before combustion can take place it is necessary that the fuel undergo some
physical transformation to enhance combustion rates by providing for effi-
cient fuel/air mixing.  Coal will be pulverized or crushed external to the
combustor, and liquid fuels are injected through a nozzle which atomizes the
liquid, providing a spray of small droplets.  The combustion of liquid or
solid fuels in turbulent diffusion flames can be conveniently divided into
three processes.
     •    Devolatilization - As the fuel is heated, matter is driven from
          the particle or droplet in the form of a gas or tar which may
          undergo secondary pyrolysis reactions producing carbonaceous
          solids.
     •    Gaseous Combustion - The gaseous fuel components are burned when
          they contact oxygen, provided conditions are suitable for ignition.
     •    Solid Burnout - The char remaining after devolatilization and any
          solid produced by pyrolysis reactions in the gas phase are
          oxidized.  The fraction of nitrogen contained in the fuel which is
          converted to nitric oxide will depend upon the availability of
          oxygen during the combustion of the gaseous and solid components.
     It is generally recognized that fuel nitrogen can be considered in two
fractions:  that which can be released from the fuel in the devolatiliztion
process (volatile fuel nitrogen), and that which remains with the solid
(refractory or char nitrogen).   The relative amounts of volatile and char
nitrogen will depend upon the droplet/particle heating rate and their final
temperature.  The division of fuel nitrogen into two components is compli-
cated by the fact that volatile nitrogen fractions may undergo pyrolysis
reactions where some of that nitrogen remains with the carbonaceous residue,
and is not converted to a simple nitrogenous gaseous molecule.

-------
     The major factors governing the conversion of gaseous nitrogenous
specie to NO are well-known, whereas char nitrogen reactions are less well-
understood.  However, it appears that in turbulent diffusion flames char
nitrogen conversion to NO is low.  Staged combustion processes attempt to
provide a fuel-rich zone with optimum conditions to maximize the conver-
sion of fuel nitrogen compounds to molecular nitrogen because these specie
may form NO during the second stage burnout.  An optimized rich first stage
for minimum NOX emissions will prevent char nitrogen from entering the burn-
out zone since conversion to NO in that zone will be small but finite, and
provide the appropriate gas phase stoichiometry and residence time to mini-
mize the content of gaseous nitrogenous specie.  If the first stage is too
rich, then the formation of HCN and NH3 is favored; whereas if the first
stage is too lean NO itself will enter the burnout zone.  Thus, the two
factors which will control the design of the rich first stage .of a low NOX
combustor are those which pertain to the partition of fuel nitrogen between
the volatile and char nitrogen fractions, and the kinetics (either homoge-
neous or heterogeneous) involved in the production of N2 from nitrogneous
species.
     This paper describes a series of bench-scale experiments designed to
identify the influence of both the chemical and physical characteristics
of fuels on fuel nitrogen conversion to NO with the objective of building
a data base which will allow combustor designers to assess the effect of
fuel characteristics on pollutant emissions.  The data base will be extended
to include development studies and field tests using the same fuels.  Con-
sequently, a body of data will be assembled which will relate the influence
of scale,  combustion condition and fuel  characteristics  to pollutant
emissions.
                                      8

-------
                                  SECTION  2
                                  APPROACH

     The combustion of fossil fuels allows the production of nitric oxide
from two sources.  In order to assess the conversion of fuel nitrogen to
NO, it is necessary to eliminate or estimate the production of NO from
molecular nitrogen (thermal NO).   Various investigators have used different
techniques to achieve this objective.  The addition of cooled combustion
products (FGR) to the combustion air has been used as a method of eliminat-
ing thermal NO production with nitrogen-containing fuels (1), or emissions
from a non-nitrogen-containing fuel have been used to assess thermal NO
production (2).  In this study thermal NO production was prevented by the
use of artificial oxidants which did not contain molecular nitrogen, thus
any NO appearing in the exhaust was formed either directly or indirectly
from the nitrogen contained in the fuel.  The artificial oxidant consisted
of mixtures of argon, carbon dioxide, and oxygen to give the appropriate
oxygen concentration and adiabatic flame temperature.  This method is not
free from criticism and has two major drawbacks:
     1.   The formation of fuel NO may be affected by the large excess
          of nitrogen molecules present when air is used as the oxidant.
     2.   The addition of C02 which is used to balance flame temperatures
          because of the different heat capacities of argon and nitrogen
          may affect H, OH and 0 concentrations and this affect is most
          likely to have a strong influence under fuel-rich conditions.
Although the use of artificial oxidants may not be ideal, it probably repre-
sents the most suitable method of assessing fuel nitrogen conversion during
the combustion of real fuels.

-------
EXPERIMENTAL SYSTEMS
     Liquid and solid fuels were burned in two separate, but similar,
refractory down-fired tunnel furnaces whose design has been described  in
detail elsewhere (3).  The modular construction of both furnaces allowed
ready access for flame observation and staged air injection through side-
wall ports.  Combustion air and artificial oxidant mixtures were supplied
from a high pressure source and metered by 600 mm length rotameters, and
they were preheated by electric circulation heaters.  The furnace wall tem-
perature was monitored continuously and when not in use, both furnaces were
fired with propane to maintain their thermal equilibrium.  Thermal NO pro-
duction was found to be very strongly dependent upon furnace wall tempera-
ture, and it was necessary to learn by experience to match propane heat
release patterns to those obtained with the test fuels to minimize experi-
mental scatter.
     The burner used in the liquid fuel furnace is shown in Figure 2,  and
several design iterations were necessary in order to provide a stable flame,
and yet prevent coke buildup at the nozzle for the wide range of fuels
investigated.  Combustion air was introduced axially and its velocity could
be varied by the use of interchangeable sleeves inserted in the burner
throat.  The interchangeable fuel injection system consisted of a 19 mm
diameter stainless steel tube which contained the atomizing air supply tube,
the fuel oil supply tube, a cartridge heater for final oil temperature con-
trol, and a chromal/alumel thermocouple positioned at the injector tip prior
to the nozzle for accurate oil temperature measurement.  The major portion
of the experimental investigations carried out to date utilized a commercial
ultrasonic oil atomizer (capacity 0.55 cc/sec) because it provided adequate
atomization of the heavy fuel oils at relatively low flow rates.  The tip of
the fuel nozzle was positioned at the beginning of the burner divergent
section, and in general, the visible flame front was displaced approximately
one nozzle diameter from the nozzle tip.  The liquid fuels were preheated
electrically in a small supply tank drawn through a 60 micron filter into
a variable-speed Zenith metering pump.  The pump outlet pressure was main-
tained constant during calibration, and operation by means of a micro-needle
valve.  The oil flow rate was varied by direct mechanical adjustment of the
pump speed.
                                     10

-------
     Pulverized coal was supplied pneumatically from a hopper-fed screw
feeder to one of two burners which allowed the coal to be burned entirely
premixed or as a diffusion flame.  Premixing was achieved by direct impinge-
ment of the coal jets with the main combustion air supply in a premixing
chamber which was separated from the combustion zone via a series of water-
cooled tubes which prevented flashback (see Figure 3).  The coal/air mixture
burned in a plug flow mode with the ignition zone situated in the refractory
divergent.  The coal, plus transport air could also be supplied to a variable
swirl burner and injected either axially or radially; thus, allowing experi-
ments to be conducted with very different fuel/air mixing rates.  The variable
swirl was achieved by dividing the total oxidant flow into two streams, one
of which flowed axially around the fuel injector, and the other through fixed
swirl vanes providing a tangential velocity cpmponent to the mixed flow.
Details of the pulverized coal burners are presented in Figure 3.
     The results presented in this paper all refer to exhaust measurements
made under excess air conditions, and the same sampling and analysis system
was used for both furnaces.  It allowed for continuous monitoring of NO, NO ,
                                                                           3C
CO, CO-, 0~, and S0_.  Flue gas was withdrawn from the appropriate exhaust
duct through a water-cooled stainless steel probe.  Sample conditioning
prior to the instrumentation consisted of an ice bath water condenser, quartz
wool filters, and a  stainless/Teflon sampling pump.  All sample lines were
Teflon and stainless steel.
FUEL SELECTION
     The  characteristics of the  liquid and solid  fuels tested to date are
presented in Tables  I and  II, respectively.  The more significant fuel
selection criteria were:
     •    Coals.  Although the full range of coals selected has yet to
          be investigated, -examples of each coal rank are to be tested,
          as well as variations in nitrogen and sulfur content within in
          each rank.  Secondary criteria used in the selection process
          were oxygen-to-nitrogen ratio and the form of the sulfur  (either
          organic or pyritic).
                                     11

-------
Liquid fuels.  The major selection criterion was the need to
include residual fuels from the major crude sources available
to U.S. consumers.  Alternate fuels were used to extend the
nitrogen content range.  Future experiments will use fuels of
widely different nitrogen mass evolution rates as a function
of temperature.
                           12

-------
                                  SECTION 3
                                   RESULTS

     This section of the paper presents the results obtained to date for
both liquid and solid fuels under normal and staged combustion conditions.
The studies are far from complete, and therefore, only preliminary conclu-
sions can be drawn from the results presented below.  Whenever possible,
liquid and solid fuels are presented in parallel in order to contrast the
fate of fuel nitrogen in liquid and solid fuels.
FUEL NITROGEN CONTENT, UNSTAGED
     Figures 4 and 5 show the emission of fuel NO and the conversion of
fuel nitrogen to NO as a function of fuel nitrogen content for liquid and
solid fuels, respectively.  Both sets of results refer to 5 percent overall
excess oxygen, and an air preheat level of 700°F.  The results with coal
were obtained with the premixed burner and a particle size distribution of
70 percent through 200 mesh.  Combustor heat release rates were similar for
all fuels.  For liquid fuels, fuel NO emissions increased almost linearly
with fuel nitrogen content.  However, for the coals tested to date there is
almost no effect of weight percent nitrogen on fuel NO emissions.
EFFECT OF FLAME ZONE TEMPERATURE, UNSTAGED
     The flame zone temperature was varied by changing the carbon dioxide
content of the artificial oxidant, while maintaining the oxygen concentra-
tion constant.  The results presented in Figure 6 indicate that flame zone
temperature has a much stronger influence on certain fuels than on others.
In general, fuel nitrogen conversion in solid fuels is more strongly depen-
dent upon the thermal environment than for liquid fuels, and certain coals
show a very strong dependence upon temperature.  These results can be attri-
buted to the influence of heating rate and final temperature on the partition
of fuel nitrogen between the volatile and the char fractions.
                                    13

-------
EFFECT OF INITIAL FUEL/AIR MIXING
     Pulverized coal was burned in three modes in order to significantly
modify the fuel/air contacting process, and Figure 7 shows the effect of
the initial fuel/air mixing upon fuel NO formation for one coal as a function
of excess oxygen.  It can be seen that premixed conditions lead to the maxi-
mum conversion of fuel nitrogen to fuel NO, and that minimum conversions are
obtained with an axial diffusion flame.  The rate of mixing achieved in the
radial swirling flame is intermediate between these two limits, and the fuel
NO emissions are also between the premixed and the axial diffusion flame.
This result is not unexpected and is in agreement with several other investi-
gations (4,5), and can be attributed to the fact that as the coal/air mixing
varies from premixed to an axial diffusion flame there is less oxygen in
contact with the volatile fuel nitrogen fractions, and therefore, the forma-
tion of N£ from the volatile nitrogenous specie is maximized.  Figure 7 also
shows that the effect of the initial fuel/air mixing conditions is also
dependent upon the coal composition, and not all coals are as sensitive to
fuel/air mixing rates as others which can be attributed to a variation in
the split between volatile and char nitrogen fractions with coal type.  Fig-
ure 8 shows the influence of the initial fuel/air mixing upon emissions from
staged combustion systems with pulverized coal.  These results are difficult
to interpret because the actual rich first stage residence time will depend
upon the fuel/air mixing type, but it appears that a backmixed first stage
gives lower final emissions than either the premixed or the axial diffusion
flame.  These results were obtained using air as the oxidant because of the
potential influence of C0? under; fuel-rich, conditions/,
THE EFFECT OF PARTICLE/DROPLET SIZE
     The influence of droplet size on the fate of fuel nitrogen in liquid
fuels has been investigated by varying the atomizing air pressure to the
ultrasonic nozzle, and by sieving the pulverized coal to burn narrow particle
size distributions.  Figures 9 and 10 show the effect of mean drop size on
both fuel nitrogen conversion and thermal NO production for four liquid fuels.
It appears that the fuel nitrogen conversion for alternate fuels is much less
sensitive to liquid droplet size than for petroleum-derived fuels, where as
                                     14

-------
drop size increases fuel nitrogen conversion decreases.  Figure 10 shows
there is a very strong influence of drop size on thermal NO production
with an apparent maximum around a mean drop size of 30 microns.  The ultra-
sonic atomizer was characterized under cold flow conditions using a laser
diffraction system to determine both the mean and the drop size distribution
of the spray as a function of liquid fuel flow rate, atomizing pressure, and
liquid viscosity.  These tests were carried out in a purpose built test rig
and not with the actual oils used in the combustion tests (6).
     Figure lla shows that under premixed conditions fuel NO production from
pulverized coal also decreases as particle size increases.  However, as shown
in Figure lib, emissions under staged combustion conditions were lowest for
the smallest particles.
                                      15

-------
                                  SECTION 4
                                 DISCUSSION

     Heap (4) suggested that under rapid mixing unstaged combustion
conditions volatile fuel nitrogen compounds were mainly responsible for
the NO produced in pulverized coal flames,  and that in order to minimize
emissions it would be necessary to maximize the evolution of these volatile
nitrogen compounds from the coal under fuel-rich conditions.  It was sug-
gested that fuel nitrogen conversion was highest when the volatile fuel
nitrogen fractions were rapidly mixed with the available combustion air.
Thus, nitrogen volatility would be a major factor in fuel nitrogen conver-
sion to NO under normal combustion conditions.  Figure 12 presents a com-
posite plot of all fuels tested to date showing the fractional conversion
of fuel nitrogen to NO as a function of weight percent nitrogen in the fuel.
If the nitrogen is expressed on a dry-ash-free basis, then the conversions
are much lower with coal than with any of the liquid fuels.  This is entirely
consistent with the hypothesis presented above, since it is known that some
nitrogen remains with the char and there may not be an equivalence between
char fractions for liquid and solid fuels.   Figure 12 also contains one data
point for a gaseous system (NH3 in methane) which represents the extreme in
nitrogen volatility.   It can be seen that fuel nitrogen conversions are high.
     Many of the results presented in the earlier section for coal can be
explained in terms of an effect upon fuel nitrogen volatility.  If heating
rate and final temperature have a strong influence on the division of nitro-
gen between the refractory and volatile fractions, then either increase in
temperature or decreasing particle size would tend to maximize the volatile
nitrogen fraction, thus increasing fuel NO production under normal condi-
tions.  Conversely, if in a practical system char nitrogen could not be pre-
vented from entering the second stage of a staged combustion system, then
                                     16

-------
the use of fine particles would minimize NO emissions since the volatile
nitrogen fraction would be maximized.
     The effect of fuel nitrogen volatility with liquid fuels is much less
pronounced under normal conditions.  The Gulf Coast and the Alaskan oils
appear to have the largest mass of refractory nitrogen assessed by com-
paring nitrogen mass evolution rates obtained under vacuum distillation,
and yet their fractional fuel nitrogen conversion are somewhat higher than
other oils.  Also, pyridine is an extremely volatile nitrogen compound, and
yet residual oils doped with pyridine to give the same nitrogen content as
undoped residual fuels give very similar NO emissions to the undoped fuel
(see Figure 13).
     The normal combustion conditions used'for the liquid fuels may be such
that fuel volatility has little effect upon fuel NO conversion because the
spray and air distribution give a well-dispersed rapidly mixed heteroge-
neous system which is relatively insensitive to the rate at which the fuel
nitrogen compounds are released from the liquid drops.  Under staged com-
bustion conditions, however, if all of the nitrogen has not been vaporized
in the rich stage then the effectiveness of staged combustion operation will
be minimized.  Experiments have been carried out with a range of liquid fuels
which have different amounts of refractory nitrogen, i.e., that nitrogen
which remains  in a residual fraction after vacuum distillation.  Figure 14
compares the effect of staged combustion on total NO emissions from several
liquid fuels.  Figure  15 compares  the fractional reduction in NOX emissions
achieved under staged combustion conditions as a function of the fraction of
the original nitrogen remaining in the residue for two residence times, and
it can be seen that this parameter is remarkably effective in correlating
the fractional reduction measured in the small-scale furnace.
                                     17

-------
                                 SECTION 5
                                CONCLUSIONS

     The results of the bench-scale studies on the fate of fuel nitrogen
during the combustion of solid and liquid fuels indicate that:
     •    With liquid fuels, fuel nitrogen content is the primary
          composition variable affecting fuel NO formation.  With
          coal, fuel nitrogen content may not be the primary variable
          controlling fuel NO production.
     •    Fuel NOX formation appears to be relatively insensitive to
          thermal environment for both liquid and solid fuels burning
          in a diffusion flame, whereas there is a strong dependence
          of fuel NO production on flame temperature in premixed pulverized
          coal flames.
     •    Initial fuel/air  contacting has a major impact upon fuel NO
          production under  staged and unstaged conditions.
     •    With liquid fuels, minimum emissions under staged combustion
          conditions depend upon the amount of refractory nitrogen in
          the original fuel.
A major factor affecting the fate of fuel-bound nitrogen, and therefore,
fuel NO production during fossil fuels combustion appears to be the rela-
tive volatility of the fuel nitrogen compounds.  Early evolution of nitrogen
compounds may maximize the benefits to be achieved from staged combustion,
and although alternate liquid fuels contain significant quantities of fuel-
bound nitrogen, NOX control may not be difficult to achieve because the
nitrogen compounds are volatile.
                                     18

-------
                                REFERENCES


1.   Turner,  D.  W.  and C. W* Siegmund.  Staged Combustion and Flue Gas Recir-
     culation:   Potential for Minimizing NO  Emissions from Fuel Oil Combus-
     tion.   Presented at the American FlameXResearch Committee Flame Days,
     Chicago, Illinois,  September 6-7, 1972.

2.   Appleton,  J. P. and T.  B.  Heywood.  The Effects of Imperfect Fuel-Air
     Mixing in a Burner on NO Formation from Nitrogen in the Air and the Fuel.
     Fourteenth Symposium on Combustion, published in The Combustion Institute,
     Pittsburgh, Pennsylvania,  1973.

3.   Pershing,  D. W., J. E.  Cichanowicz, G. C. England, M. P. Heap, and G. B.
     Martin.   The Influence of Fuel Composition and Flame Temperature on the
     Formation of Thermal and Fuel NO  in Residual Oil Flames.  Presented at
                                     "5C"
     the Seventeenth Symposium (International) on Combustion, Leeds, England,
     1978.

4.   Heap,  M. P., T. L. Lowes, R. Walmsley, H. Bartelds and P. LeVaguerese.
     Burner Criteria for NO  Control Volume I Influence of Burner Variables
     on NO  in Pulverized Coal Flames.  EPA-600/2-76-061a, March, 1976.
          J\.

5.   Wendt, J. 0. L., J. W. Lee and D. W. Pershing.  Pollutant Control Through
     Staged Combustion of Pulverized Coal.  U.S. DOE Report No. 1817-4, U.S.
     DOE Technical  Information Center, Oak Ridge, Tennessee, 1978.

6.   England, G. C., M. P. Heap, R. C. Horton, D. W. Pershing and G. Flament.
     The Effect of  Fuel Properties and Atomizer Design on Emission Control
     from Heavy Fuel Oil Fired Combustors.  Paper presented at the Third
     Stationary Source NO  Symposium, San Francisco, California, March, 1979.
                         X
                                     19

-------
            Physical Transformation
                Droplet/Par tide
           Heatlng/Devolati1i zati on
     Secondary
     Reaction
Tar
Secondary
Reaction
                                                NO
                                 'N2
Figure 1.   The Fate of Nitrogen in Solid  or  Liquid
           Fuels During Combustion.
                       20

-------
                Ultrasonic Twin-fluid Atomizer
               Burner
Viewing Port   Section
      Thermocouple
      Connection

      Oil  Heater
      Connection

    Combustion
    Air         \
    JT      1L
    Tunnel Furnace
Atomizing Air

•*— Oil  Pressure
    Tap

 Oil Inlet

Viewing. Port
                                                       Ultrasonic
                                                       Nozzle
                                        Burner Detail
                               Insulating
                               Block
                                                Insulating
                                                Refractory
                             High Temperature
                             Refractory
Flue
                                     Furnace Cross-Section
 Figure 2.  Details of Liquid Fuel-Fired Furnace  System.
                            21

-------
             Insulating
Insulblock    Refractory
                       High
                          Temperature
                          Refractory
Coal + Transport Air








a
/
i
i
i


i

i
k
^
o'l
0
o
0
0
o
0
o


_^B«^B-
•^•^—•"•M
J




L

r







.^•WH



3



r~
                                                                      Main
                                                                      Combustion
                                                                      Air
                                               Coal + Transport Air
                                                                        Cooling
                                                                        Water
        Figure 3.  Details of Pulverized Coal Furnace System and Burners.

                                       22

-------

o
o
c
01
O)
o
•M
•r-
0)
3
Lt,
«*-
0
O
»r-
00
o>
c
o
o
>e
60

50

40

30

20

10

0
1 1 T 1 1 II II ||

-
Scranton A
- N' Dak- Utah ^
A Q Black Creek
A Western AL. O
~ Montana KY. ft ~
Savage Rga
- AL'
6
_ W. Va.

	 1 	 1 	 L 1 I 1 I I i i i
        0.8 0.9  1.0  1.1  1.2 1.3  1.4  1.5  1.6  1.7  1.8  1.9  2.0
                    Weight % Nitrogen in Fuel (DAF)
1400
1300
^1200
° 1100
£ 1000
a
ll 90°
a.
o 800
^ 700
600
500
400
1 1 1 1 1 II II II
U?ah Black Cre^k .
- Scrfnton c ^ Western O _
N Dak Savage KY- Rosa AL.
- ' ' Mont. _

- O
W. Va.
—
~ —
I I 1 I I 1 i 1 1 I 1
        0.8 0.9  1.0  1.1  1.2 1.3  1.4  1.5  1.6  1.7  1,8  1.9 2.0
                        % Nitrogen in Fuel (DAF)
Figure 4.  Fuel NO and Percent Conversion of Fuel Nitrogen to
           Fuel NO - Pulverized Coal Premixed.
                              23

-------
 CVJ
o
  X
o
 Q.
1600
1400
1200
1000
800
600
400
200
90^
- 5% Excess 02
-
-
-
*
mm ^p
^* ^N^j
••• J^^1^
r 1 1 1 1
1 1 II
        80
o
in
o
o
       50
       L£Q£NJ2
»  ALASKAN DIESEL
,  W.  TEXAS DIESEL
  CALIFORNIA NO. 2
  ESSEX COUNTY
  MIDDLE EAST
  LOW SULFUR NO. 6
  INDO/MALAYSIAN
 DESULFURIZED VENEZUELAN
  PENNSYLVANIA/AHARADA HESS
  GULF COAST
  VENEZUELAN
  ALASKAN
  CALIFORNIA #1
  CALIFORNIA #2
  CALIFORNIA #3
  CALIFORNIA W
  CRUDE  SHALE AND BLEf4i>s
 WITH O
 SHALE-DERIVED  DFM
 SRC II
 SYNTHOIL BLENDS
 WITH 9
 CHj,  * NHj
                                cf
                    .4    .6    .8  1.0   1.2   1.4 1.6   1.8 2.0
    Figure 5.   Fuel  NO and Percent Conversion of Fuel Nitrogen to
                 Fuel  NO - Liquid  Fuels,  Ultrasonic Atomizer.
                                       24

-------
   700
 CM

5 60°"
O
% 500 -


S 400
  X
O
     300
 0)
if 200-


^ 100
               I       I

                Wilmington
                         (1607
                                 (1655)    (17U°K)
               (1514)
                 E>--
              Alaskan  Gulf^Coast

                Venezuelan
              East Coast
               I
                    I
I
I
I
I
             2100   2200   2300    2400    2500   2600°K
                   Theroetical "Flame Temperature
l_ O Utah
     Savage, Montana
 ~O Western Kentucky
     Scranton, N.  Dak.
r-a W. Va.
   O Black Creek,  Al.
   O Rosa, Al.
       I        I       I
                   Volume % of C02 in Oxidizer
     Figure  6.  Effect of Flame Zone Temperature on
                Fuel NO Formation.
                            25

-------
   CM
  O
1300

1200

1100

1000

 900
        U Ta.r =  6500F
   . 800

  £. 70°

  §. 600
  Q.
  o 500

  *5 400

  "" 300

     200

     100
           Utah  Coal
           70 x  103 Btu/hr
         0.0
    cT*
1400

1200
i

1000
       8°°

       600


       400

       200
                   2.0            4.0
                        02% in Stack
                                    \<>
                                    l
                                        t
  6.0
     0

     0
Total NO
5% 02
   I
                  1.0     1.2      1.4     1.6      1.8

                        % Nitrogen in Fuel (DAF)

Figure 7.  The Effect of Initial Fuel/Air Mixing -  Coal.
                             26

-------
   700 -
   600  -
   500 -
 CVJ
o
o.
Q.
   400
   300
   200
   100
     0
          Utah Coal
             0.4
0.6
                   Axial  Diffusion

                        O
                 Radial  Swirling
0.8
1.0
                      First  Stage  Stoichiometric Ratio
  Figure  8.  The  Effect  of  Initial Fuel/Air Mixing - Staged, Coal.
                                   27

-------
0>
o
o
tt>
   70 -
    65
    60
    55
    50
    45
A Indo/Malaysian

Q Alaskan

• SRC II

• Crude Shale
          20          40     60   80 100         200


                 Rosin-Rammler Mean  Dropsize  (ym)
 Figure 9.  Effect of Droplet Size on Fuel Nitrogen Conversion.
                                 28

-------
   300
•» 200
O)
Q.
a.
    100
                                         A  Indo/Malaysian
                                         D  Alaskan
                                         + SRC  II
                                         • Crude  Shale
           20          40     60   80 100         200
                   Rosin-Ranmler Mean Dropsize (urn)
    Figure 10.  Effect of Droplet Size in Thermal NO  Formation.
                                                    A
                                 29

-------
  CM
 o
 c.
o

"ttJ
  evj
 o
       130C -
       1100-
        900-
 700-
        500-
        300
                                             Normal  Distribution
       130C -
      1100 -
900-
       700

       500

       300
                                    4.0
                               %  in Stack
                                                             6.0
                      a)  Premixed  Unstaged
               .4
                    .6           .8           1.0
                 First Stage Stoichiometric  Ratio
1.2
                      b)  Staged
Figure 11.  The Effect of Coal Particle. Size on  NO  Formation.
                                 30

-------
         90 -
         80 -
       .X
      o
      2=   70 -
      o
      «r-
      10
      V
      O)
         60 -
          50
          40
          30
          20
O

 O
                 I	I
     ^   K k
               1
                         I          I         I
                           OCH4 + NH3
                           O Distillate Oils
                           b Residual Oils
                           DAlternate Liquid Fuels
                           • Coals
                                              a
L
I
I
                     .2        .6        1.0       1.4      1.8
                         Weight % Nitrogen In  Fuel  (DAF)
Figure 12.  Overall Composite of Fuel Nitrogen Conversion - Unstaged.
                                   31

-------
       700
       600
      CM
     O

     o 500
     £•400
     0)
     2  300
       200
       100
                   0.77% N
                   1.63% S
                     0.76% N
                   0.51% N
                   1.63% S
                     0.40% N
                 2.22% S
• Pure Residual Oil
O Distillate Oil + Pyridine + Thiophene
A Residual Oil + Pyridine + Thiophene

                      I        I        I
                  _L
I
                  1
234
   % Overall Excess 0,
Figure 13.  Relative Volatility Effects - Liquid Fuels - tlnstaged.
                                  32

-------
 2000
 1800
  1600
  1400
  1200
§1000
    T—i—r—i—i—i
•  Unrefined Shale Oil
"  Alaskan Bunker C
O  Gulf Coast No. 6
A  Indo/Malayslan No.  6
    Alaskan Diesel Oil
              0.6      0.7     0.8      0.9      1.0      1.1
                      Primary Zone Stoichiometric Ratio
       Figure 14.   Effect  of  Liquid Fuel Composition on Emissions -  Staged.
                                    33

-------
     0.24 —
     0.22 —
     0.20 —
     0.18
  •a
  4>
  o>
  S2 0.16
 . « ^
o o
     0.14
     0.12
     0.10
     0.08
     0.0$
I
I
                                           I
                                   I
                  I
0.3      0.4     0.5
       %N  in  Residue \
        %  N in Fuel  /
                      (
0.6      0.7      0.8
   '% Residue Mass
         100
                                                                   0.9     1.0
                 (Fraction of Original Nitrogen Left in  the  Residue)
    Figure 15.  Importance of Refractory Nitrogen - Liquid Fuels - Staged.
                                     34

-------
u>
en

Ultimate Analysis:
Carbon, %
Hydrogen, %
Nitrogen, %
Sulfur, %
Ash, %
Oxygen, % (by difference)
Conradson Carbon Residue, %
Asphalt ene, %
Flash Point, °F
Pour Point, °F
API Gravity at 60°F
Viscosity, SSU, at 140°F
at 210°F
Heat of Combustion:
Gross Btu/lb
Net Btu/lb
Calcium, ppm
Iron, ppm
Manganese, ppm
Magnesium, ppm
Nickel, ppm
Sodium, ppm
Vanadium,, ppm
Alaskan W. Texas California
Diesel Diesel Distillate

86.99 88.09 86.8
12.07 9.76 12.52
0.007 0.026 0.053
0.31 1.88 0.27
<.ooi <.ooi <.ooi
0.62 0.24 0.36




33.1 18.3 32.6
33.0 32.0 30.8
29.5 28/8 29.5

19,330








Essex
County

86.54
12.31
0.16
0.36
0.023
0.61
2.1
0.34
205
50
24.9
131.2
45

19,260
18,140
7.1
16
0.09
3.7
6.7
37
14
Middle E.
(Exxon)

86.78
11.95
0.18
0.67
0.012
0.41
6.0
3.24
350
48
19.8
490
131.8

19,070
17,980
1.2
2.6
0.02
0.08
13
0.98
25
Low S.*
No. 6 Oil

86.57
12.52
0.22
0.21
0.02
0.46
4.4
0.94
325
105
25.1
222.4
69.6

19,110
17,970
9.52
123.6
0.46
2.23
14.10
3.74
3.11

-------
CO
en

Ultimate Analysis:
Carbon, %
Hydrogen, %
Nitrogen, %
Sulfur, %
Ash, %
Oxygen, % (by difference)
Conradson Carbon Residue, %
Asphaltene, %
Flash Point, °F
Pour Point, °F
API Gravity at 60°F
Viscosity, SSU, at 140°F
at 210°F
Heat of Combustion:
Gross Btu/lb
Net Btu/lb
Calcium, ppm
Iron, ppm
Manganese, ppm
Magnesium, ppm
Nickel, ppm
Sodium, ppm
Vanadium, ppm
West
Texas

86.01
9.72
0.46
3.53
0.048
0.23
20.08
9.24
205
94
6.2
29,720
1,824

17,800
16,910
5,9
43
0.13
0.9
29
3.2
60
Alaskan

86.04
11.18
0.51
1.63
0.034
0.61
12.9
5.6
215
38
15.6
1,071
194

18,470
17,580
6.9
24
0.06
1.4
50
37
67
California California
#1 #2

85.75 85.75
11.83 11.44
0.62 0.77
1.05 1.63
0.038 0.043
0.71 0.71
8.72
5.18
—
38
19.5 15.4
246.1 854
70.00 129

18,470
17,430
21
73
0.8
5.1
65
21
44
California
#3

85.41
11.23
0.79
1.60
0.032
1.02
9.22
5.18
150
30
15.1
748.0
131.6

18,460
17,440
14
53
0.1
3.8
82
2.6
53

-------
TABLE I.  LIQUID FUELS TESTED (CONTINUED)

Ultimate Analysis:
Carbon, %
Hydrogen, %
Nitrogen, %
Sulfur, %
Ash, %
Oxygen, % (by difference)
Conradson Carbon, Residue %
Asphaltene, %
Flash Point, °F
Pour Point, °F
API Gravity at 60°F
Viscosity, SSU, at 140°F
at 210°F
Heat of Combustion:
Gross Btu/lb
Net Btu/lb
Calcium, ppm
Iron, ppm
Manganese, ppm
Magnesium, ppm
Nickel , ppm
Sodium, ppm
Vanadium, ppm
Indo/
Malaysian

86.53
11.93
0.24
0.22
0.036
1.04
3.98
0.74
210
61
21.8
199
65

19,070
17,980
14
16
0.13
3.6
19
15
101
Venezuelan
Desulphurized

85.92
12.05
0.24
0.93
0.033
0.83
5.1
2.59
176
48
23.3
113.2
50.5

18,400
17,300
8.7
6.5
0.09
3.6
19
15
101
DFM Pennsylvania Gulf
(Shale) (Amarada Hess) Coast

86.18
13.00
0.24
0.51
0.003
1.07
4.1
0.036
182
40
33.1
36.1
30.7

19,430
18,240
0.13
6.3
0.06

0.43
0.09
<.l

84.82
11.21
0.34
2.26
0.067
1.3
12.4
4.04
275
66
15.4
1049
240

18,520
17,500
9.2
13.2
0.10
3.3
32.7
64.5
81.5

84.62
10.77
0.36
2.44
0.027
1.78
14.8
7.02
155
40
13.2
835
181

18,240
17,260
4.4
19
0.13
0.4
29
3.6
45

-------
CO
CO

Ultimate Analysis:
Carbon, %
Hydrogen, %
Nitrogen, %
Sulfur, %
Ash, %
Oxygen, % (by difference)
Conradson Carbon, Residue %
Asphaltene, %
Flash Point, °F
Pour Point, °F
API Gravity at 60°F
Viscosity, SSU, at 140°F
at 210°F
Heat of Combustion:
Gross Btu/lb
Net Btu/lb
Calcium, ppm
Iron, ppm
Manganese, ppm
Magnesium, ppm
Nickel, ppm
Sodium, ppm
Vana,diuro» PPm . , . , , . . „
California
#4

86.66
10.44
0.86
0.99
0.20
0.85
15.2
8.62
180
42
12.6
720
200

18,230
17,280
90.6
77.2
0.87
31.4
88.0
22.3
66.2
Synthoil

86.30
7.44
1.36
0.80
1.56
2.54
23.9
16.55
210
80
S-1.14
10,880
575

16,480
15,800
1,670
109
6.2
170
2.6
148
6.5
Shale
(Crude)

84.6
11.3
2.08
0.63
.026
1.36
2.9
1.33
250
80
20.3
97
44.1

18,290
17,260
1.5
47.9
0.17
5.40
5.00
11.71
<.3

-------
                                        TABLE II.   SOLID FUELS INVESTIGATED
oo
vo
EER COAL DESIGNATION
Coal Source
Ultimate Analysis;
Carbon, %
Hydrogen, %
Nitrogen, %
Sulfur, %
Ash, %
Oxygen, %
(by difference)
Proximate Analysis:
Moisture, %
Volatiles, %
Ash, %
Fixed Carbon, %
Calorific Value
(Btu/lb)
Forms of Sulfur
Pyritic, %
Organic, %
Sulfate, %
B
West
Virginia

71.57
4.94
1.33
1.09
11.66
7.54
1.87
32.53
11.66
53.94




D
Utah

68.49
5.20
1.24
0.64
7.40
10.64
6.39
38.89
7.40
47.32
12,340



4
Scranton
N.D.

42.02
2.71
0.54
0.99
7.50
11.28
34.96
28.85
7.50
28.69
6,446
0.32
0.65
0.02
11
Montana
Savage #11

41.36
2.57
0.72
0.27
4.61
14.11
36.36
—
4.61
—
6,995
0.01
0.26
0.00
18
Rosa
Alabama

74.72
4.36
1.60
0.96
6.79
3.55
8.02
21.81
6.79
63.38
13,394
0.43
0.47
0.06
19
Black Creek
Alabama

80.27
5.09
1.75
0.74
4.45
5.45
2.25
28.28
4.45
65.02
14,284
0.07
0.67
0.00
20
Western
KY

69.50
4.76
1.33
2.95
7.81
8.85
4.80
36.10
7.80
51.1
12,450




-------
           THE CONTROL OF POLLUTANT FORMATION IN FUEL OIL FLAMES
                   - THE INFLUENCE OF OIL PROPERTIES AND
                           SPRAY CHARACTERISTICS
                                    By:

    G. C. England, M. P. Heap,  R. T. Horton, D. W. Pershing and G. Flament*
               Energy  and  Environmental Research Corporation
                         Irvine, California 92715
International Flame Research Foundation,  IJmuiden,  Holland
                                     41

-------
                                   ABSTRACT

     Pollutant production in liquid fuel flames is controlled  by  the
complex interaction of the liquid fuel spray and the combustion air flow
field.   This paper presents results of a study to investigate  the influ-
ence of atomizer design and fuel properties under both normal  and staged
combustion conditions on nitrogen oxide formation.  The sprays produced
from the various atomizers were characterized under isothermal conditions
by a method involving the diffraction of laser light.
                                      42

-------
                               ACKNOWLEDGEMENTS

     The work presented in this paper was carried out under EPA Contracts
68-02-2624  and 68-02-3125, and it is a pleasure to acknowledge the support
and assistance of  Mr.  G. B.  Martin and Mr. W.  S. Lanier,  the Project Officers,
                                       43

-------
                                  SECTION 1
                                INTRODUCTION

     Although the increased utilization of coal provides a partial solution
to the U. S. energy crisis, a balanced fuel economy necessitates that in the
future many industrial users will burn petroleum, coal or shale-derived
residual fuels.  The relatively high nitrogen content and low hydrogen-to-
carbon ratios associated with these liquid fuels suggests that their use may
be accompanied by an increase in two pollutants unless preventive measures
are taken.  The production of nitrogen oxides (NO ) and particulate matter
                                                 X
from a given fuel in turbulent diffusion flames is controlled by the fuel/air
contacting process, which is dependent upon the complex interaction of the
liquid fuel spray and the combustion air flow field.  A major fraction of the
nitrogen  oxides produced in residual fuel oil flames is formed by the oxidation
of nitrogen contained in the fuel.  The fraction of fuel nitrogen which is
converted to fuel NO is almost completely dependent upon the oxygen avail-
ability during the thermal decomposition of the liquid droplets.  Particulate
matter produced in residual fuel oil flames consists of soot, which is pro-
duced from the gas phase specie, and two other constituents, cenospheres and
coke, production of which is associated with the multi-component nature of the
fuel and  the characteristics of droplet combustion.  The formation of these
latter two components can be attributed to the occurrence of liquid phase
thermal cracking because of high droplet temperatures.  Soot formation, which
occurs in three distinct phases:  nucleation, surface growth, and coagulation,
is associated with the existence of fuel-rich zones within the flame.  Thus,
local stoichiometry which is controlled by the turbulent mixing process and
is spray/flow field-dependent plays a dominant role in the formation of NO
and particulate matter.
     Bench-scale studies (1) indicate that the total nitrogen content of
liquid fuels is the major parameter controlling fuel NO formation under
                                      44

-------
overall fuel-lean conditions.  The fuel nitrogen speciation appears to
be a secondary parameter unless the fuel is being burned under staged com-
bustion conditions, when fuel nitrogen volatility appears to limit NO
                                                                     X
reduction achievable under staged combustion.  Field studies (2) have investi-
gated the influence of fuel properties and atomizer design on both the pol-
lutant emissions from, and thermal efficiency of operating boilers.  Although
these studies indicated that both of these parameters were important, the
nature of the tests precluded the identification of the controlling param-
eters.  The results presented in this paper were obtained under conditions
which were typical of industrial practice, and yet allowed the influence of
atomizer type and fuel properties to be identified under comparable conditions,
The specific objectives of the research program were:
     •     To relate the formation of nitrogen oxides in liquid fuel flames
           to atomizer design and to fuel oil properties.
     •     To investigate the interaction between liquid sprays and the
           airflow pattern for conditions typical of package boilers under
           normal and staged conditions.
The information generated by this program will provide for the generalization
of low-NO  oil burner technology for application to package boilers.
                                       45

-------
                                  SECTION 2
                            EXPERIMENTAL SYSTEMS

     The combustion experiments were conducted in an axisymmetric combustor
at a nominal firing rate of 3 x 10  Btu/hr. (.88 MW).  This combustor, which
has been described in detail elsewhere (3), allowed the addition of cooled
combustion products to the combustion air supply and had the capability of
dividing the combustion air into two separate streams to allow operation
under staged combustion conditions.  In these investigations the staged air
was injected through an axial boom inserted from the rear of the combustion
chamber.  The distance of the air injector from the fuel atomizer could be
varied.  Standard techniques and instrumentation were employed for the
measurement of flue gas composition.
BURNER AND ATOMIZERS
     The commercial burners used in this  investigation were fitted to an axial
oil gun in a simple burner system which consisted of:
           a refractory divergent exit, 45  half angle
           an annular air duct with interchangeable fixed-vane swirlers
           generating an axial velocity of 100 ft/sec (31 m/sec) and swirl
           numbers of 0.21, .45 and .79.
The oil gun was precisely positioned to locate all the atomizers at the throat
of the divergent and the combustion air supply was designed to give a uniform
air distribution across the throat.  A steam/auxiliary electric heating system
was used to maintain the oil temperature  (measured by a thermocouple) at the
oil nozzle constant.   Table 1 lists the atomizers used in the study which were
of the twin fluid (air or steam assist) type.
SPRAY CHARACTERIZATION
     The characteristics of the sprays produced by the commercial atomizers
used in the study were not available at the time this paper was written.
                                     46

-------
Droplet size distribution will be determined under quiescent isothermal
conditions using a laser droplet size analyzer based on the Fraunhofer
Diffraction pattern of a laser beam by the particles moving through a cross-
section of the beam (4).  The particle cloud of a spray causes a deflection
of a portion of the parallel beam giving a conical pattern of diffracted
light.  There is a direct relationship between the diameter of the particle
(assumed spherical) and the angle of the light diffracted by this particle.
Therefore, it is possible to calculate the particle size distribution of a
cloud of particles by an evaluation of the diffracted light energy.  The
conical diffracted light pattern is measured in the focal plane of a lens by
a photo detector element composed of 30 annular rings whose radii increase
from 0.148 mm to 14.56 mm.   The instrument evaluates the drop-size distribution
and expresses it in a Rosin-Rammler form together with the weight percentage
and number of drops in each of 15 groups.
     The advantages of this laser diffraction technique are that it is fast
in comparison to photographic techniques,  requires no calibration for parti-
cles with diameters ranging from 10 to 500 ym and it does not interfere with
the spray.  However,  there may well be major limitations associated with the
use of this technique and a comparison of  various optical techniques for spray
characterization has been initiated under the FCR program.  The primary limi-
tation of the diffraction technique is that imposed by spray density which
arises due to multiple scattering.  Other limitations include an undefined
influence of droplet velocity, a minimum drop size detectability of 10 ym and
the use of the Rosin-Rammler distribution.
     Figure 1 provides an.illustration of the use of this instrument to
compare drop-size distribution from three atomizers:
     -     a pressure jet  (mechanical)
           a single lobe of a Y-jet (twin fluid)
     -     an ultrasonic atomizer  (twin fluid)
It can be seen that the pressure jet and the Y-jet give very similar drop-
size distributions while the ultrasonic atomizer gives small droplets with a
narrow size distribution around 20 ym.
                                       47

-------
FUELS
     The fuels used during the investigation are listed in Table II, together
with selected properties.  Figure 2 presents the results of a Differential
Thermal Analysis (DTA) and Figure 3 shows the mass evolution of nitrogen as
a function of distillation temperature.  The DTA clearly shows the multi-
component nature of the petroleum-derived residual fuels.  It can also be
seen that the nitrogen in the shale-derived fuels (Nos. 4 and 8) is evolved
at a much lower temperature than that from petroleum-based residual fuels.
Apparently the nitrogen species in the Gulf Coast and North Slope oils
(Nos. 5 and 6) appear to be refractory in nature since almost all the nitrogen
remains in the residuum.
                                       48

-------
                                  SECTION 3
                                   RESULTS

     The series of investigations planned to define the influence of atomizer
design and fuel characteristics on the emission of NO  from residual fuel oil-
                                                     X
fired flames can most conveniently be described under two separate headings:
those concerned with normal excess air conditions, and those which pertain to
two-stage combustion in which the first stage was operated fuel-rich and air
was added to the rich combustion products downstream from the fuel nozzle to
complete combustion.
NORMAL EXCESS AIR CONDITIONS
     The results presented in Figure 4 compare NO  emissions as a function
                                                 X
of overall excess air level for a single fuel (No. 3) with a burner swirl
number of 0.45.  It can be seen that in general the NO  emissions increase
                                                      X
with increasing excess air level for each atomizer.  However, there is a wide
range in emission level for different atomizers.  The results presented in
Figure 5 show the influence of fuel type on NO  emissions at two swirl levels,
                                              X,
S = 0.45 and 0.79, with a single nozzle (B).  These results clearly show the
impact of the flow field on NO  formation.
                              x
     It is generally recognized that NO  emissions increase with increasing
fuel nitrogen content.  Figure 6 presents a plot of NO  emissions as a function
                                                      JV
of fuel nitrogen content for four different atomizers and two swirl levels.
For each atomizer there appears to be an almost linear relationship between
NO  emission and fuel nitrogen content.  However, this relationship is not
  X
general and depends both upon atomizer design and the flow field conditions.
For instance, atomizer B gave very high NO  emissions which were similar to
                                          X
those produced by atomizer A at the low swirl level.  However, at the higher
swirl condition emissions from atomizer A were almost unaffected by the swirl
level, whereas those from atomizer B were considerably reduced.  At the inter-
mediate swirl level the four atomizers appear to fall into two categories,
whereas at the high swirl level three of the four atomizers behaved similarly.

                                       49

-------
The data presented in Figure 6 illustrates some of the difficulties associated
with any general correlation relating fuel nitrogen content to emissions since
the nature of the combustion system has a significant influence.
STAGED COMBUSTION CONDITIONS
     Staged combustion, i.e., the operation of a combustion system in which
the fuel originally burns under oxygen-deficient conditions, provides the
most cost-effective control technique for reducing fuel NO.  The results pre-
sented in Figure 7 show the influence of atomizer design on the emission of
NO  at an overall excess air level corresponding to three percent excess oxygen
as a function of the percentage of theoretical air in the fuel-rich primary
zone.  The burner throat conditions were unmodified, and therefore, as the
percentage of theoretical air in the primary zone decreases, the burner throat
velocity also decreases.  The second stage air was added at the same  location
for each atomizer.  The atomizer design has a significant influence upon the
minimum NO  achieved under staged combustion conditions.  All of the nozzles
show  the same characteristic performance.  As the primary zone becomes more
fuel-rich NO  emissions decrease to a minimum and then increase.  This increase
            ii
is probably associated with  the escape of fuel nitrogen specie from the primary
zone  where  they are oxidized to NO when  the second staged air is added.  Com-
paring the  minimum emission  levels with  those reported for the bench-scale
furnace with the same  fuels,  it can be seen that  the results shown in Figure 7
do not show the same degree  of control.  This can be attributed to two factors.
The bench-scale combustor has refractory walls, and therefore the initial fuel
rich  stage will be at a higher temperature.  Residence times in the bench-scale
combustor rich  zone were also much longer than those achieved in the cold-wall
combustor.
      The trade-off between the decrease in NO  emissions and the increase in
                                             X
smoke emissions is illustrated by the data presented in Figure 8 which shows NO
                                                                               3
and Bacharach smoke number as a function of percent theoretical air in the
primary zone for three swirl conditions with a single nozzle.  The NO  emis-
sions  appear to be relatively insensitive to primary zone swirl number for
this  particular nozzle.  However, it can be seen  that the smoke emissions
differ considerably for the same range of swirl levels.
                                       50

-------
     A cylindrical refractory extension 36 inches (.91 meter) long was
added to the divergent burner exit to ascertain whether the reduced heat
loss in the first stage could account for the different emission levels
achieved in the combustor and the bench-scale experiments.  The addition of
this refractory extension had almost no effect upon exhaust NO  emissions.
     Figures 9 and 10 present data comparing the performance of two atomizers
under staged combustion conditions for two different fuels and zero swirl in
the primary zone.  One fuel was essentially free from nitrogen, and it can
be seen that the emission levels from the two nozzles are dependent upon fuel
nitrogen content.  Also, at low theoretical air ratios in the primary zone
nozzle A gives higher smoke emissions than nozzle B for both fuels.
                                       51

-------
                                   SECTION 4
                                  DISCUSSION

     Data available from both field and development studies indicates that
the emission of nitrogen oxides from cold-wall combustors is a complex
function of fluid dynamics and fuel chemistry.  The results presented in this
paper were part of an investigation planned to provide data under well-
controlled conditions to define the effect of fuel properties and atomizer
parameters on pollutant emissions from combustion systems typical of those
encountered in practice.  At the time of writing the paper one section of
the investigation, that concerned with spray characterization, had not been
completed, and therefore the interpretation of the results can only be con-
sidered as preliminary.
     The primary fuel parameter controlling NO  formation in liquid fuel
                                              X
flames is the fuel nitrogen content of the oil.   NO  emissions increase with
                                                   3C
increasing nitrogen content.  However,  the conversion of fuel nitrogen to NO
                                                                            3C
is strongly dependent upon the mixing of fuel and air within the flame.
This interaction is illustrated by the results presented in Figure 11 which
shows the NO  emissions as a function of fuel nitrogen content obtained with
atomizer B.  The open data symbols refer to measurements made with pure com-
bustion air.  The partially shaded data symbols refer to NO  emissions mea-
sured when the combustion air was vitiated with approximately 20 percent
flue gas  recirculation.  It can be seen that with flue gas recirculation
emissions are almost independent of swirl level and very similar to those
obtained  with unvitiated air at a swirl level of S = 0.79.  This could be inter-
preted as meaning that the difference in emission levels between the two
swirl levels is due to thermal NO production.  However, this is probably an
erroneous conclusion since the addition of flue gas recirculation will not
only lower flame temperatures due to the addition of an inert thermal ballast
but will also modify the mixing process due to the increased throat velocities.
Extrapolation of the two unvitiated lines to zero percent nitrogen indicates
                                       52

-------
 a thermal N0x production  of approximately  170 ppm which is higher than that
 achieved with low nitrogen diesel  fuels.   Figure 11 also allows a comparison
 between the bench-scale results  (1) and  those obtained in this investigation.
 The thermal characteristics of the bench-scale  furnace are different from the
 3,000,000 Btu/hr combustor, and  therefore, only the fuel NO  emissions have
                                                           x
 been plotted.  The same atomizer design was used for both investigations,
 although the capacity was different.  Apparently fuel nitrogen conversions
 were higher in the bench-scale investigations than those obtained in the
 combustor.  This might well be associated  with  a change in drop size as the
 atomizer capacity is increased,  i.e., for  a given atomizer design drop size
 is not independent of capacity.
     The influence of atomizer design on NO  emissions is probably due to the
                                           X.
 combined effect of drop-size distribution, spray pattern, and the momentum of
 the droplets.  Since these parameters will control the heating rate, penetra-
 tion of the droplets through recirculation zones and the local stoichiometry.
 Figure 12 compares the performance of five nozzles with a single fuel as a
 function of swirl level.  As the swirl level increases the size and intensity
 of the internal axial recirculation zone will increase, and the NO  emissions
                                                                  x
 are for the most part, associated with the interaction of the spray and the
 internal recirculation zone.  It appears that different nozzle designs will
 influence this interaction.   For example, nozzle A shows a continual decrease
of emissions with increasing swirl level, whereas many of the other nozzles
show a minimum emission at the intermediate swirl level.  Table III presents
the ratio of NO  at swirl level  equal to 0.5 and 0.79 for four atomizers and
               2£
four fuels.   This table indicates that there is also an influence of fuel
properties on the interaction of the spray and  the flow field since for a
given atomizer increasing the swirl level may increase or decrease the nitrogen
oxide emission.
     It was  noted earlier that the NO  reductions achieved in the bench-scale
                                     x
experiments  were higher than those obtained in  the combustor.  This can be
seen in Figure 13,  which compares the emissions from a staged crude shale oil
containing over 2 percent nitrogen.  In the bench-scale studies at the shortest
primary zone residence time of 400 msec minimum emissions were approximately
                                       53

-------
200 ppm.  In the combust or primary zone residence time was much shorter,
about 200 msec, and emissions were always greater than 300 ppm.
     It can be seen that there is a very significant effect of atomizer
design upon NO  emissions.  The atomizer giving the lowest emissions under
              X
staged conditions gave the highest value unstaged.   Comparison with informa-
tion presented by Heap et al (1) would suggest that nozzle B has a much
smaller drop-size distribution than nozzle A,  thus  allowing for more rapid
vaporization of the fuel nitrogen compounds in the  primary zone giving more
time for the formation of N.,  and thus,  lower  overall emissions.
                                       54

-------
                                REFERENCES

1.    Heap, M. P., D. W. Pershing, G. C. England, J. W. Lee, and S. L.
     Chen.  The Influence of Fuel Characteristics  on Nitrogen Oxide
     Formation - Bench-Scale Studies.  Third  Symposium on Stationary
     Source Combustion, March  1979.

2.    Cato, G. A., L. J. Muzio,  and P.  E.  Shore.  Field Testing:  Application
     of Combustion Modification to Control Pollutant Emissions from Industrial
     Boilers, Phase II.  EPA Report No. 600/l-76-086a, April 1976.

3.    Muzio, L. J., and R. P. Wilson.   Experimental Combustor for Development
     of Package Boiler Emission Control Techniques.  Phase  I of III, EPA-R2-
     73-292a  (NTIS PB No. 224274/AS),  July  1973.

4.    Swithenbank, J., J. M. Beer, D. S. Taylor, D.  Abbott,  and G. C. McCreath.
     A Laser Diagnostic for the Measurement of Droplet and  Particulate Size
     Distribution.  University of Sheffield,  Department  of  Chemical Engineering
     and Fuel Technology.  Report 1976.
                                        55

-------
      70
      60
      50
                                                                 n

                                                                 o
Pressure Oet


Ultrasonic





Y-Jet Single Hole
en
    0.40
    GO
      30
     20
      10
                                                              Diameters (pm)


                                       Figure 1.  Dropsize Distribution for Three Atomizers.

-------
                                                                                    Indo/M»laysian
                                                                                    Middle Eastern
                                                                                    Gulf Coast
                                                                                    California
                                                                                    North Slope
                                                                                    North Slope Diesel
                                                                                    DFM
0.0
                 100
                                                                                            600
                                           Temperature,  C
                             Figure 2.  DTA Analysis of Eight Fuel Oils.

-------
  40
            k 2 - Middle Eastern
            A 3 - Indo/Malaysian  	
               4 - DFM (Shale Derived
               5 - Gulf Coast
            D 6 * North Slope
            O 7 - California      	
            • 8 - Shale (Unrefined)
   30
QJ
£  20
   10
                500
800
900
                                                                              1000
1100
                                     Distillation Temperature (°F)
                      Figure 3.  Vacuum Distallation Results - Nitrogen  Evolution.
                                              58

-------
   260
   220
 CM
o
o
•4->
10
    180
    140
    100
D Atomizer A
O Atomizer B
   Atomizer C
   Atomizer D
         I
_L
 k Atomizer  E
 Q Atomizer  F
 Q Atomizer  G
 O Atomizer  H
	I
                  2.0
                        3.0             4.0
                   % Excess Oxygen (Stack)
                               5.0
       Figure  4.  Comparison of NO  Emissions for Eight Commercial Atomizers
                 in a Common Flow rield.
                                      59

-------
       600
                                             Swirl No = 0.45
       500
       400
01
o
    Q.
    Q.
300
       200
                                                          _L
                                                             Swirl No = 0.79
                                                             O  California
                                                             Q  North Slope
                                                             ^  Gulf Coast
                                                             £  Indo/Malays1an
                                                             k  Middle Eastern
                                                           52
                                                          % Excess Air
               Figure 5.  Influence of Fuel Type on N0x Emission for a Single Atomizer  in  Two  Swirl  Levels.

-------
             Atomizer B
             Atomizer A
             Atomizer F
             Atomizer D
    600
 CM
o
o
4->
•a
     400
a.
a.
     200
       0
                                              O  Atomizer B
                                              D  Atomizer A
                                              O  Atomizer F
                                                 Atomizer D
        0
0.2     0.4     0.6     0.8
  Wt. % Nitrogen in Fu$l
1.0
0     0.2     0.4     0.6     0.8
         Wt. % Nitrogen in Fuel
1.0
                        Figure 6.   Influence of Atomizer Type at Two Swirl Levels.

-------
          400
                         Swirl  = .45
                         3% Excess Oxygen
                   Atomlzer
                   DA

                   A  D

                   OB
          300
01
no
       CM
      O
o

4->
IB
          200
      i
          100
                                                   I
                                                                      I
                         60
                                            80                      100
                                          % Theoretical Air - Primary Zone
120
                        Figure 7.
                             Comparison of W  Emission From Four  Atomizers Under Staged Combustion.

-------
    400
    300
 t
 •a
S  200
Q.

Q.
    100
            D 15° Swirl Vane     Open Symbols - NO

            O 30° Swirl Vane     Solid Symbols - Smoke

            A 45° Swirl Vane
                  0.6
0.7          0.8          0.9          1.0

         Primary  Zone  Stoichiometric Ratio
1.1
1.2
                                                                               8
                                                                                  s-
                                                                               _  O)
                                                                               6  -
                                                                                                                 
-------
      200
      150
CTl
    CM

   O
      100
   a.
   a.
       50
              —i	r

               Alaskan Diesel  011

               Zero Swirl

               3%  Excess 00
                  Atomi zer
                Atomizer

                   A

                  A   NOX

                  A   Smoke
                     0.6
0.7
0.8         0.9          1.0

Primary Zone Stoichiometric Ratio
1.1
                       Figure 9.   Influence of Atomizer Type on NO  and Smoke Emissions Under

                                  Staged Combustion - North Slope ftiesel Oil.
                                                                             9





                                                                             8
                                                                6





                                                                5





                                                                4





                                                                3





                                                                2





                                                                1
                                                                                o

                                                                                £
                                                                                to
                                                                                JC
                                                                                u
                                                                                to
                                                                                CO

-------
   400
   300
 CM

O
a.
a.
   200
   100
  Alaskan  RFO, 3%

•  Zero Swirl



 O  Atomizer B


 A  Atomizer A
                  I
                           Open Symbols - NO


                           Solid Symbols - Smoke
                      I
I
I
I
                 0,6
                     0.7         0.8          0.9          1.0

                             Primary Zone Stoichiometric Ratio.
                                   1.1
                                   1.2
                                                                                               9


                                                                                               8


                                                                                               7




                                                                                               6i
                                                                                                  50
                                                                                                  j*
                                                                                               4  <•>

                                                                                                  2
                                                                                                  1C



                                                                                               3  tJ
                                                                                               J  OQ
                                                             2





                                                             I
                   Figure 10.   Influence  of Atomizer Type on NO  and Smoke Emissions Under

                               Staged Combustion - North Slope Bunker C.

-------
    800
    600
  CM
 o
 o
 4-»
 •0
    400
    200
      0
                           Fuel NO
                                  n
                           Bench Scale (1),
                   I
I
                                                     Swirl No.
                                                  O  0.45
                                                  D  0.79
0.0     0.2
                         0.4      0.6
                         Wt. % Nitrogen
        0.8
1.0
1.2
Figure 11.  Comparison of NO  Emission From Bench—Scale and Cold-Wall
            Combustor  Experiments.
                             66

-------
                                                                                          T
   320
   300 —
   280
    260
 CM
O
ox 240
o.
Q_
    220
    200
    180
TIndo/MalayslanT

     4% Excess O  (Stack)
Atomizers
D  A

O  B

O  c
                  .2
.4     .                 .6
   Swirl  Number'at Throat
          .8
                Figure 12.  NO  Emissions as a Function of Swirl Number for Five Atomizers.
                              JL

-------
                                                          Firetube Combustor
                        	-a	
1500

1400

1300

1200

1100
oo
          0.6
                Crude Shale  |
                3% Excess Op
                No Swirl/Refractory Insert
                O Atomizer B
                D Atomizer A
                £3 Bench  Scale Experiments
                                                              I
                                          0.8                0.9
                                            Primary Zone Stoichiometric Ratio
1.1
                   Figure  13.   NO   Reduction  -  Comparison of  Bench  Scale  Experiments With  Cold  Wall  Combustor.
                                X

-------
                    TABLE I.  LIST OF COMMERCIAL ATOMIZERS


ATOMIZER                                         DESCRIPTION


   A                            Internal mixing, air/steam assist,  swirl
                                chamber and pintle

   B                            Ultrasonic air assist

   C                            Internal mix, prefilming, air/steam assist,
                                pintle

   D                            Y-jet internal mixing, air/steam assist

   E                            Y-jet internal mixing, air/steam assist

   F                            Y-jet internal mixing, air/steam assist

   G                            Y-jet internal mixing, air/steam assist
                                swirl cap

   H                            Low pressure Y-jet internal mixing, swirl
                                cap air assist
                                       69

-------
                                 TABLE II.  SELECTED FUEL OIL PROPERTIES
Property
Ultimate Analysis
Nitrogen %
Sulfur %
Asphalt ene %
API Gravity @ 60°F
Gross Heat of Combustion
N. Slope
Diesel

0.007
0.31
-
33.1
19,400
Middle
Eastern

0.18
0.67
3.24
19.8
19,070
Indo/
Malaysian

0.24
0.22
0.74
21.8
19,070
Diesel
Fuel
Marine

0.24
0.51
0.036
33.1
19,430
Gulf
Coast

0.36
2.44
7.02
13.2
18,240
N. Slope
Bunker C

0.51
1.63
5.6
15.6
18,470
California

0.77
1.63
5.18
15.4
18,470
Shale
Oil

2.08
0.63
1.33
20.3
18,290
  Btu/lb
Vanadium ppm
25
101
45
67
44
<0.3

-------
TABLE III.  RATIO OF NOV EMISSIONS AT INTERMEDIATE AND HIGH
                       A

            SWIRL LEVELS FOR FOUR ATOMIZERS AND FOUR FUELS
                                        ATOMIZER

  FUEL                         A       B       C       F





    2                         1.07    1.26    0.98    0.91


    3                         1.46    0.96    0.94    0.96


    6                         0.96    1.52    0.84    0.93


    7                         1.02    1.53    0.84    1.15
                             71

-------
          THE GENERALIZATION OF LOW EMISSION
                COAL BURNER TECHNOLOGY
                          By

D. M. Zallen, R. Gershman, M. P. Heap and W. H. Nurick
     Energy and Environmental Research Corporation
             Santa Ana, California  92705
                            73

-------
                                  ABSTRACT



     This paper describes the development of a low NO  pulverized coal
                                                     jt

burner.  Results are presented from a series of tests designed to define


optimum burner design parameters at three scales 10, 50 and 100 Btu/hr


heat input.  NO  emissions were found to be most sensitive to burner zone
               x

stoichiometry and fuel injection parameters.  Results are presented for


three bituminous coals.  The use of dry sorbents for S0_ control is


discussed and results are presented which suggest that low NO  coal burners
                                                             A

may provide optimum conditions for combined NO  - SO  control using flame
                                              Jv     Jv

zone sorbents.
                                         74

-------
                               NOMENCLATURE
SR






T



SA



S




ppm






Primary Air



Secondary Air



Tertiary Air
                        (air/fuel) actual/

                        (air/fuel) stoichiometric
Stoichiometric Ratio






Temperature



Swirl angle


_ .  ,    ,      G_    Angular momentum
Swirl number = -r^-r  = . ? ,	
               G R    Axial momentum x characteristic  radius
                Xt-


Parts per million by volume on a dry basis and  reduced to


zero percent excess air



Air in fuel injector



Burner air in coannular passage around the fuel injector



Air used for staging
                                SUBSCRIPTS



               Air with coal in fuel injector



               Air in the coannular flow passage around  fuel  injector



               Staging air injected from separate  ports  located


               at some radial distance from the burner centerline
                                       75

-------
                               ACKNOWLEDGMENTS

     The work described in this paper was carried out during the conduct
of EPA Contract 68-02-2667.   The authors are pleased to acknowledge the
considerable help and encouragement provided by G. B. Martin of the EPA
and of several of their EER colleagues C. McComis, M. Deming, R. Thomas,
R. Thalken and J. Lee during the course of this investigation.
                                        76

-------
                                 SECTION I

                               INTRODUCTION

     An increase in the direct combustion of coal, either pulverized and
burned in suspension or crushed and burned in fixed or fluidized beds, repre-
sents the only near term solution to decrease our dependence upon the import
of foreign petroleum products because of increased utilization of coal.  How-
ever, the increased use of coal enhances environmental problems associated
with its extraction, transportation and utilization.  Coal contains several
trace compounds which have the potential to produce atmospheric pollutants
during the combustion process.  This paper is concerned with the generaliza-
tion of technology which is controlling the emission of two pollutants,
nitrogen oxides (NO ) and oxides of sulfur (SO ) from pulverized coal-fired
                   2t                          j£
combustors.  Direct coal combustion represents the major source of nitrogen
oxides emitted by stationary sources.  Consequently, unless appropriate
emission control technologies are applied an increase in the utilization of
coal in direct fired combustors has the potential to significantly increase
NO  emissions.
  x
     Historically, application of NO  control to utility boilers represents
the most cost-effective approach to controlling total emissions from stationary
sources.  Emissions from coal-fired utility boilers have been estimated to
account for 32 percent of all emissions from stationary sources.  The EPA has
established an intensive program to demonstrate control technology for pul-
verized coal-fired boilers; industrial as well as utility, which will prevent
adverse impact on the environment in the event of increased coal firing in
water wall boilers.  Technical goals have been established for 1985 which are
equivalent to 20 percent of the existing New Source Performance Standards
(NSPS) for NO  emissions.  These goals are to be met by combustion modifica-
             2m
tion techniques (i.e., changing the process by which fuel and oxidant are
brought together then how reaction occurs and heat is released).  These
techniques have been successful in allowing boilers to comply with NSPS.
                                       77

-------
However, the 1985 technical goal might well represent the limit of their
application since the near term requirement dictates that the control
techniques be compatible with existing state-of-the-art boilers.  Major modi-
fications to the boiler would involve an extensive development program and
necessarily the control technology would take time to be accepted by the
utility industry.
     Two processes are involved in the emission of nitrogen oxide from coal-
fired boilers:  1) the oxidation of molecular nitrogen (thermal NO), and
2) the oxidation of nitrogen contained in the coal (fuel NO).  Insofar as
bulk gas temperatures are low, and thermal NO formation is mainly restricted
to the heat release zone, combustion modifications which minimize peak flame
temperatures and maximize combustion product quench rates will control its
production.  Fuel NO formation is determined by the conditions in the heat
release zone; fuel residence time under oxygen-deficient conditions being
the major factor in minimizing fuel NO production.  Peak flame temperatures,
product quench rates and residence time in fuel-rich zones can be controlled
by burner design parameters (e.g., method of fuel and air injection, distri-
bution of axial and tangential velocity, burner geometry, etc.).  Effective-
ness of these parameters as a means of controlling NO  emissions from pul-
verized coal flames has been demonstrated in both pilot- and large-scale
systems.  These investigations have led to the development of a burner which
can be applied to existing boiler designs to satisfy the technical NO  emis-
                                                                     JC
sion goals described above.  This burner is based upon the concept of dis-
tributed mixing* which minimizes NO  formation while satisfying other pro-
cess requirements (e.g., heat flux distribution, burnout, etc.).
     The basic burner design features, the watertube simulators, and the
ancillary equipment used in this study have been described in detail else-
where (1).  The results of earlier pilot-scale studies were used to develop
a prototypte burner whose performance has been demonstrated at a scale and
under conditions which are similar to actual watertube boilers.  This paper
summarizes the results obtained to date with single burners at three scales
*  Distributed Mixing Burner (DMB)
                                       78

-------
with different coals, and discusses  the  implications  associated with multi-
ple burner firing.  In addition  to NO  control,  the DMB has been used in a
series of investigations to assess the possibility of in-situ sulfur capture
using dry sorbents.  The sorbents were added  to  the coal prior to pulveriza-
tion and it appears that the DMB has the potential to satisfy the process
requirements for both NO  and  SO control.
                       X        X
                                       79

-------
                                  SECTION 2
                         LOW NO  BURNER DEVELOPMENT
                               x

     Pulverized coal combustion in burner stabilized, turbulent diffusion
flames is a complex phenomenon which cannot be described fully by a mathe-
matical model.  Consequently, the development of the low NO  DMB based
                                                           A
upon the original pilot-scale concept has progressed emperically.  Two
different burner systems have been tested to date.  These are:
     •    A burner with a divided secondary throat which allows control
          of  axial and tangential velocity in two annular air streams.
          Fuel is injected axially from an annulus surrounding a central
          oil gun used for ignition.
     •    A simple double concentric burner with fixed swirl vanes in both
          the annular fuel and secondary air channels.  Swirl is used in
          the fuel injector to enhance flame stability.   This burner design
          was selected because of ease of construction in providing for
          exact geometric and velocity similarity at three scales.
The velocity characteristics of the two burners are illustrated in Figure 1
for 50 x 10  Btu/hr burners operating at 4 percent overall excess air
as a function of the burner zone stoichiometric ratio.  The influence of
several burner and operational parameters have been investigated with both
burners, and the composition of the three different coals which have been
used to date are given in Table 1.
     The burner with the divided throat allows a wide variation in the
distribution of both axial and tangential velocity of the secondary air
stream and provides for considerable control over flame shape with the wide
angle burner exit (35  half angle).  Three basic flow patterns were observed
in both cold and hot flow (see Figure 2).  These were:
          Type A with all axial flow except for an annular toroidal
          recirculation zone which provided for flame stability.

                                       80

-------
          Type B with an axial toroidal recirculation zone.
          Type C in which the flow attached to the combustion chamber
          wall producing a wall jet.
The transition from A through C occurred as the swirl level  in the outer
channel increased.   Figure 2 also presents sketches of two general flame
types which were obtained with a narrower burner exit (25° half angle).   The
shorter flame is much more compatible with existing boiler designs.
     The low sufur  Utah coal listed in Table 1 was used for  the major
portion of the tests performed using the divided throat burner.  The general
effects of overall  excess air and burner zone stoichiometry  were as expected.
NO  decreased as the overall excess air and the burner zone  stoichiometry was
  Jv
decreased.  Unlike  more conventional burner designs, emissions were relatively
insensitive to variations in load for a given burner size.  Various burner
exit geometries were tested and it was found that flame stability was improved
with a narrower burner exit whose length was equal to the throat diameter.  A
schematic of the divided secondary throat burner is presented in Figure 3.
Figure 4 compares NO  emissions for the divided secondary throat burner at
                    A
two scales.  The burner was designed for constant air velocity and geometric
similarity.  The larger burner yielded lower NO  values.  For the scaling
                                               2v
principle used, the size of the fuel-rich burner zone would  be proportional
to a linear burner  dimension implying that the residence time in this region
is also proportional to burner scale for a given velocity.  This provides one
simple explanation  for the lower values obtained at larger scale.  However,
other.parameters may well be important.
     Further development studies have been carried out with the less com-
plex double concentric burner shown in Figure 5 at three design firing rates
of 12.5, 50 and 100 x 10  Btu/hr.  The use of a single channel for the burner
throat provided for a more straightforward interpretation of the results, and
also allowed "exactly similar" burners to be produced at three scales.  All
three burner scales employed similar distribution of tangential velocity and
similar locations for the outboard staged air injectors, and had a burner
exit half angle of  25°.  The NO  emission characteristics of this low N0x
DMB design have been evaluated and the influence of the following operational
                                       81

-------
and design parameters assessed:
     •    Burner zone stolchiometry.  Both the overall burner zone stoi-
          chiometry and the stoichiometry of the primary airflow.
     •    Burner parameters including the degree of swirl in both the
          fuel and secondary air channels.  The location and velocity of
          the outboard staged air injectors and the location of the fuel
          injector.
     •    Firing rate.
     •    Fuel type.
Typical results obtained to date are presented below.
BURNER ZONE STOICHIOMETRY

     The effect of burner zone stoichiometry for a range of excess air
levels is shown in Figure 6 for three different scales at comparable con-
ditions.  The two large burners were tested in the large watertube simulator.
Burner zone stoichiometry is calculated based upon the total fuel flow and
the air being delivered with the primary and secondary streams (total air
minus outboard staged air).   Comparison of the large and intermediate scale
burner test results (see Figure 6) show a moderate increase in NO  with
                                                                 2v
scale.   However the results obtained with the smallest burner were highest
of all.  It is possible that variations in combustor configuration may con-
tribute to this difference since the smallest burner was fired in a different
combustor than the two large burners but at a similar burner zone heat  release
rate.
     For a given fuel injector geometry variations in the amount of primary
air will not only change the local stoichiometry in the zone close to the
fuel injector it will also affect the primary velocity and the results pre-
sented in Figure 7 may be attributed to changes in either or both of these
properties.  The data indicates that the primary stoichiometry for optimum
burner performance is dependent upon burner scale.  This is probably
associated with the fact that the residence time in the fuel rich zone is
different for the three burner scales.
                                      82

-------
BURNER PARAMETERS

    The operation of the simple burner  shown  in Figure 5  involves a reduc-
tion of secondary airflow as the burner  zone stoichiometry is reduced.   This
reduces the tangential momentum of  the secondary stream; thus, any balance
between primary and secondary momenta will be  affected  as  the burner zone
stoichiometry is reduced at a given primary flow.   It was  found necessary
to provide the burner with fixed annular swirl vanes in the primary pipe to
ensure flame stability under staged conditions.   This simple vane system
not only provides some tangential momentum, but  also acts  to divide the
coal stream into separate jets.  Figure  8 compares  NO   emissions for various
primary swirl vane angles and it appears that  a  vane angle of 45  provides
the optimum performance.  NO  emissions  were also found to be sensitive to
the level of swirl in the secondary stream.  In  general, NO  emissions  were
                                                           X
reduced as the swirl level increased, as can be  seen in Figure 9.
    The effect of both the location and velocity of the staged air jets
was investigated.  Typical results  are presented in Figure 10.  Optimum per-
formance appears to be obtained with low velocity,  and  staged air injection
with the injectors located two burner-throat-diameters  from the burner  axis.
For the smallest burner tested the  emissions and flame  stability were found
to be sensitive to the location of  the fuel injector.   However, good per-
formance was obtained with the two  larger burners when  the fuel injector, was
located at the throat of the burner.
FUEL TYPE
    It can be seen from the results presented in Figure 11 that the perfor-
mance of the low NO  DMB is dependent upon fuel  type for the 12 x 10 Btu/hr
                  J&
burner.  There are considerable differences in the  emission characteristics
for the same burner with different  fuels, which  is  not  necessarily associ-
ated with overall nitrogen content  since under highly staged conditions
emissions from the three coals are  comparable.  The results obtained at
100 x 10  Btu/hr compared to that shown  for 12 x 10  Btu/hr for the West
Virginia and Utah coals suggest that the effect  of  fuel type is also
associated with scale.  For example at the smaller  scale the Utah coal  gives
a lower value of NO  than the West  Virginia coal.  This characteristic  is
                  x
reversed at the larger scale.
                                      83

-------
                                  SECTION 3
                               SULFUR CAPTURE

     Control of oxides of sulfur can be achieved by one of three general methods:
1) the sulfur can be removed prior to combustion, either in a gasifier or
by physical or chemical cleaning; 2) sulfur may be absorbed by a solid during
the combustion process itself as in a fluidized bed combustor; and 3) the
sulfur oxides can be removed from the products of combustion.  It is known
that certain compounds have the potential to produce sulfates, and certain
coals with high calcium and sodium ashes give high sulfate ash contents.
The use of sorbents such as limestone to remove sulfur oxides has been applied
unsuccessfully in several instances (2,3).  Both small-scale studies and field
test have suggested that the retention efficiency of such sorbents can be
quite low and under these conditions the mass of sorbent required to achieve
the necessary level of S0_ control is excessive.  However, sorbent utiliza-
tion efficiency has been shown to be sensitive to such parameters as tempera-
ture, stoichiometry, particle size,  etc.  A preliminary series of experiments
have been carried out with the low NO  DMB with the intention of assessing
whether or not low NO  operation enhances the possibility for in-situ sulfur
                     ji
capture , since the DMB provides a temperature/time stoichiometry history that
should maximize the sorbent utilization.
     Preliminary tests were carried out using the Utah coal in a small
refractory tunnel furnace to assess the effect of sorbent dispersion, flame
temperature, the method of fuel/air contacting, and combustion air staging
upon sulfur capture.  The effectiveness of the sorbent was assessed by
measuring the S0_ content of the combustion products.  The two most signifi-
cant results of these small-scale studies were that the sulfur capture was
enhanced considerably when the sorbent was mixed with the coal prior to pul-
verization,  and that increasing flame temperature enhanced rather than
inhibited the effectiveness of the sorbent.  The influence of sorbent dis-
                                       84

-------
persion can be readily understood since to achieve high sorbent efficiencies
it is necessary that the sorbent be uniformly mixed throughout the combustion
gases.  The effect cf increasing temperature was somewhat surprising since
with limestone as the additive, it could be concluded that maximum flame
temperatures would produce dead-burning, and therefore, limit the activity
of the sorbent.
     A series of preliminary exploratory investigations were also carried out
in the small watertube simulator with a 12 x 10  Btu/hr low NO  DMB.  The
                                                              x
objective of these preliminary studies was to assess the effect of sorbent
type, (calcium and sulfur compounds were used), sorbent ratio and coal type
upon sulfur capture, as well as to determine (1) whether there were signifi-
cant advantages of operating under low NO  conditions, and (2) whether or not
                                         X
sulfur capture influenced NO  emissions.  It is important to note that the DMB
design is such that under non-staging conditions the coal/air mixing profile
still provides for a slow mixing control core.  Consequently even though
unstaged the DMB flame characteristics will be quite different from those
produced with typical commercial turbulent burners.  The results presented in
Figures 12 and 13 for a low and high sulfur coal, respectively, are typical of
the results obtained in these preliminary experiments.  Since exhaust S0_
concentration was used as the criterion for sulfur capture, care was taken in
the sampling procedure to prevent condensation in the sample lines.  Also,
measurements were made of SO,, concentrations before and after tests with
additive to ensure that the sulfur dioxide levels were reproducible.  With a
calcium-to-sulfur molar ratio of one, 50 percent of the sulfur was captured
by the additive for both the high and low sulfur coals, and with this parti-
cular calcium-to-sulfur ratio there appeared to be a significant effect of
operation under staged combustion conditions.  In these preliminary investi-
gations no attempt was made to optimize the burner configuration for both NO
                                                                            X
reduction and sulfur capture.
     The results presented in Figure 14 show that there appears to be no
effect of sulfur capture upon NO  emissions.  From gas phase kinetic con-
siderations it might be expected that SO /NO  would influence the overall
                                        X   X
NO  emissions when the burner was operated staged.  The results obtained to
  X
date do not indicate that this is a significant factor.
                                       85

-------
                                  SECTION 4
                                 DISCUSSION

     The primary design objectives of  the low NO  DMB are:
                                                x
     -    To maximize both the rate and  total volatile evolution from
          the coal particle.  This does  not only refer to nitrogen specie,
          but also to fuel fragments since gas phase stoichiometry will
          be dependent upon the volatile fuel fraction.
          To provide an initial oxygen-deficient zone which minimizes NO
          production, but then to add  sufficient oxygen to the rich com-
          bustion products to maximize the rate of decay of nitrogenous
          specie (such as NH_, NO,  HCN)  to molecular nitrogen.
          To provide optimum residence time and sufficient temperature
          to maximize N_ production.
          To maximize char residence time in the fuel-rich zone since
          some nitrogen will remain in the solid with the potenital for
          forming NO during burnout.
     -    To provide second stage air ensuring complete burnout.
     -    To produce an overall oxidizing envelope around the fuel-rich
          core thereby minimizing the possibility for corrosion on the
          combustor walls.
These objectives are obtained by providing for the optimum interaction
betwen the primary fuel jet and the swirl-stabilized recirculation zone,
together with delayed air addition from the outboard staged air injectors.
     Figure 15 compares the NO  levels achieved at three different scales for
a low NO  DMB.  Figure 16 presents the NO characteristics for the optimum
configuration at the same three scales.  The differences between the various
scales may be associated with the variation in fuel residence time for
                                      86

-------
different scales, or burner/chamber interaction affecting bulk gas tempera-
tures, and therefore, reaction zone temperatures.
    In addition to the optimization of burner design parameters and the
development -of empirical scale criteria, the program to generalize low
emission coal burner technology has three other important objectives which
are:
    1.   Construct a fuel data base which will establish whether it is
         necessary to vary the burner design as a function of coal
         properties.
    2.   Determine the effects of burner/chamber interaction and
         develop information for multiple burner configurations.
    3.   Provide direct comparison between prototype low NO  burners
                                                           X
         and commercially available burners.
    4.   To assess the effect of low NO  operation on the emission of
                                       X
         other pollutants such as fine particulate matter and sulfur
         oxides.
A comparison between the prototype and commercial burners will provide
information enabling the performance of the low NO  DMB to be assessed when
                                                 X
operating under more practical conditions.
    Three investigations are planned to assess the affect of burner/
chamber interactions on the performance of the low NO  DMB.  These include:
                                                    3C
    •    Single-wall arrays to assess the effect of burner/burner
         spacing, burner/sidewall spacing and the use of shared
         staged air injectors.
    •    Opposed-wall firing.
    •    Off-axis firing.  The minimum NO   emissions achievable by the
                                         X
         DMB concept may provide excessive  flame lengths.  In order to
         overcome flame impingement upon combustion chamber walls, four
         burners will be tested firing off-axis in order that the inter-
         action between the various long flames will prevent wall
         impingement.
                                      87

-------
     The preliminary results involving the use of dry sorbents  for  SCL
capture are encouraging,  and currently investigations are underway  to pro-
vide confirmatory results by closing the sulfur balance after capturing the
solid.  Figure 17 summarizes the results obtained to date with  three additives,
two coals and various additive/sulfur ratios.  It can be speculated that
the effectiveness of the  DMB as a means of enhancing sulfur capture is
associated with the following:
     •    Sorbent dispersion.   Effective sorbent utilization requires that
          its concentration must be high and that it is evenly  distributed
          in these regions where the sulfur specie are evolved  from the
          coal.  In the tests described above the sorbent was mixed with
          the coal before the pulverizer and therefore was evenly dispersed
          in the form of  small particles.
     •    Sorbent activity.  The sorbent is in contact with high sulfur
          concentrations  initially after calcination at the time it is
          most reactive.   Also the low NO  DMB probably provides the
                                         Ji
          optimum time-temperature history for maximum sorbent  activity
          and minimal deadburning.
     •    Peak flame temperatures occur under fuel rich conditions. The
          sulfur and sorbent initially make contact under oxygen deficient
          conditions and  probably form a sulfide which is more  stable at
          higher temperatures than the sulfate.  Consequently decomposition
          of the "sulfur  containing sorbent" will be minimized.
     •    Reduced peak temperatures and low bulk gas temperatures.   Low
          NO  operation requires that peak temperatures are minimized thus
          ensuring maximum residual sorbent activity to capture sulfur  in
          the bulk gases  after the heat release zone.  The fact that the
          sorbent is evenly dispersed will also tend to maximize sulfur
          capture in this region.
These preliminary results suggest that the use of a flame zone  sorbent  plus
baghouse might well be a viable SO  control technology for low sulfur Western
                                  3C
coals or in addition to physical or chemical cleaning for high  sulfur  coals.

-------
                                  SECTION 5
                                   SUMMARY

     The increased utilization of coal for power generation has the potential
for adverse environmental impact due to increased emissions of atmospheric
pollutants.  One of the pollutants, NO  can be controlled by modifying the
                                      Ji
combustion process to maximize residence time under fuel rich conditions
in order to prevent the formation of NO from nitrogen contained in the coal.
The combustion of pulverized coal involves injection of the fuel and air
through a burner whose function is to ensure ignition stability and complete
combustion in the necessary volume.  A burner has been developed which
controls the rate of pulverized coal air mixing that not only satisfies
normal process requirements but also minimizes the formation of nitrogen
oxide.  This concept was originally evaluated at pilot scale but it has now
been demonstrated at a scale and in an environment which is typical of
waterwall boilers.  The low NO  DMB has been tested at three scales using
                              X
three bituminous coals and the results obtained to date can be summarized by:
     •    Under optimum design conditions NO  emissions are typically below
                                            2£
          200 ppm (dry at 0% 0-) for three burner scales (12, 50 and 100 x
          106 Btu/hr).
     •'    NO  emissions are not independent of scale when a simple geometric
            3C
          scaling law is applied.
     •    The emission characteristics of the low NO  DMB appears to be
          mildly dependent upon coal characteristics.
     A preliminary series of investigations have been carried out to assess
the use of the DMB to control both NO  and SO  by adding dry sorbents to
                                     3C       X
the coal before the pulverizer.  It was felt that the DMB would provide
ideal conditions to maximize sorbent efficiency.  Limestone mixed with the
coal in calcium to sulfur molar ratios of 1, 2 and 3 gave 53, 73 and 88%
reduction in S0_ emissions respectively for a low sulfur coal.  At low
calcium/sulfur ratios SO- reduction is  enhanced by low NO  operation.
                        ^                      .          3C
                                       89

-------
     It should be noted that the impact associated with the use of this
technology for SO  control in field operating boilers requires a complete
                 X
assessment.  The most immediate areas requiring process and economic analyses
are increased solids handling (both prior to and after combustion), mill
capacity, fly ash characteristics  (size distribution and resistivity), slagging
and fouling in the boiler and the  environmental problems associated with solids
disposal.  Should the development  and small scale studies continue to provide
encouraging results it is intended that the technology will be demonstrated
in an industrial boiler fitted with an EPA low NO  DMB.
                                                 x
                                       90

-------
                                REFERENCES

1.   Gershman,  R.  E.,  M. P. Heap, and T. J. Tyson.  Design and Scale Up
    of Low Emission Burners for Industrial and Utility Boilers.
    Proceedings of the Second Stationary Source Combustion Symposium,
    Volume 5,  Addendum EPA 600/7-77-073e, July, 1977.

2.   Attig, R.  C.  and P. Sedor.  Additive Injection for Sulfur Dioxide
    Control -  A Pilot Plant Study, National Air Pollution Control Adminis-
    tration, Report No. 5460, March, 1970.

3.   Gartrell,  F.  E.  Full Scale Desulfurization of Stack Gas by Dry
    Limestone  Injection, Volume I, Tennessee Valley Authority, Chattanooga,
    EPA 650/2-73-019A, (PB-228447), August, 1973.
                                      91

-------
£    A»)  Divided Secondary Throat Design.
u
o
      B.)  Simple Double Concentric Burner Design
                                                               1.2   1.3
Figure 1.  Air Velocity Characteristics for two 50 x 10  Btu/hr Coal Fired

           Distributed Mixing Burners.
                                 92

-------
      Type A
     Low Swirl
  Type  B
Medium Swirl
  Type C
High Swirl
 Flame Types
                                                      Low Swirl
                                                      S * 0.30
           NO  »  125  ppm
             /\          "~—
                                                   Medium Swirl
                                                   S * 0.44
B.)  Low NO  Flame Shapes
            Figure 2.  Flame Characteristics for DMB.
                                93

-------
                Secondary
                   Air
Terti ary
  Air
Furnace
 Wall
       Swirl Vanes
 Coal &
Primary
  Air
            Secondary
              Air
                                                               Retractable
                                                               Oil Nozzle
                                                                 Carwrlc
                                                                 Quarl
                                           Tertiary
                                             Air
                Figure  3.  Divided  Secondary Throat Burner Design.
                                      94

-------
                            Utah Coal   6.3% Stack
  400
fr
a
 CM
o
 .300
I
a.
  200
  100
                   I              I
        Scaling:
        Geometric/Constant Velocity
        Primary Zone <* Burner Dimension
.°.
           tres «  Burner Scale
                                        I
                                        w   9 x 10° Btu/hr
                                        O  50 x 106 Btu/hr
     1.25
          1.00
0.83
0.71
0.62
0.56
                                        SR,
       Figure A.  Comparison of NO  Emissions at Two Burner Scales
                  Divided  Secondary Throat Design.
                                     95

-------
                                                       c
                                                       60
                                                      •H
                                                       CO
                                                       0)
                                                      Q

                                                       S-i
                                                       
-------
  [Utah Coal, SR1 = 0.25, SA1 = 45°, SA2 = 60°, Tert-Mid]
  600
            I
      I
      I
  500

  400

  300

  200

   100

     0
1*400
 cvj
        12 x 10° Btu/hr
    SR2 = 1.0
   200
   100
T
I
I
        50 x 10° Btu/hr
                              J	L
400


300

200

100
n
- - I I 1 1 1
6
100 x 10° Btu/hr
^ 	 *S*2
y>0.6
~~ O 	 ^^^S*bV5
0.45
—
1 1 1 1 1
123456
1


, = 0.7




1
7



^~"



~

8
                  Excess  02  (%  Dry)
    Figure 6.  Effect of Burner Zone Stoichiometry.
                            97

-------
500

400

300

200

100

  0
500
1*400
       [Utah  Coal,  SR2 = 0.65,  SAj_ = 45°,  Ter-M1d]
            1  ,   !
        12 x 105
              0.350-
 CM
   300
   200
   100
   500
   400
   300
   200
   100
     0
            I
          50 x 10
    1
100 x  10
                   1
                 1
                              SA2 = 45
1
                                       = 0.15
                             I
             1
       SA2 = 60°
                                     =  0.30
      I
                                         1
                                   SA  = 60
                                  1
                   3456
                      Excess Q2 (% Dry)
                                         8
   Figure 7.  Effect of Primary Stoichiometry.
                        98

-------
10
           500
         £400
         o
          *
          CM
          ^200
           100
                 12 x 10°  Btu/hr
                                               [Utah Coal, SRj = 0.25,  SAg  =  60°]
Tert-Far
                                            	O0.6
                                             SA, -
                                       .... SA:
    60°  _
    45°
    J	
                                                   7     8
                                                      Excess 0,
50 x 10b Btu/hr
                                      I
                   1     2
                    Dry)
Tert-Mid
SR2 =0.6
                                                                                                             8
                                           Figure 8.   Effect  of  Primary Swirl.

-------
     CM
    o
8   e
         500
         400
         300
         200
         100
   I      I

12 x 106 Btu/hr
                  I      I
    [Utah Coal, SRj = 0.25,  SAj • 45°, Tert-M1d]


                          500
                                              SR2 = 0.6
SA2=45<
I	 I
                                                        8
                          400
                                                300
                          200
                         100
    I      I


50 x 106 Btu/hr
                                                                                           I      I
                                                             SR2 =0.7
                                                                                                60"
                                                                      I      I
                             I	I
                                                                         8
                                                     Excess  02  (%  Dry)
                                    Figure 9,  Effect of Secondary .Swirl.

-------
       [Utah Coal, SRj = 0.25, SR2 = 0.6, SAj = 45°, SA2 = 60°]

s;
o
A
CM
O
O
ft
Q.
Q.
X
o

500
400

300

200
100
0
II 1 1 1 1
12 x 105 Btu/hr
_ —

_ • Mid/Low Vel —
9 ^*?^ Tr Q
_ ^ Far/Low Vel _
_ —
1 1 1 I 1 1
L 2 3 4 5 6 7 £
                                           I      I
                                        50 x  106 Btu/hr
                         Excess 02 (% Dry)
                                                                                8
Figure 10.  Effect of Tertiary Location and Velocity.

-------
     700
r\>   O
  [SR, * 0.25, SA. = 45°, SA2 * 60°. Tert-Mid]
1	H—I	L
    I      I      I
4.2% Excess 02
                                                                                               12 x lOpBtu/hr
         J
                                        I
                                                                                     I - L
   Utah Coal
   W.V. Coal  .
—W. KY. Coal
_l	I	
                                                          9  0.4  0.5   0.6   0.7   0.8   0.9
                                                                                        SRo
                                                        1.0   1.1   1.2
                                             Figure 11.  Fuels Performance.

-------
  900
  800
  700
  600
fc
Q
o

o


1400
a
 CM
o
V)
  300




  200




  100



    0
 • Unstaged  Utah
 H Staged  Utah
'O Unstaged  Utah
                        +  Sorbent
          Staged Utah + Sorbent
                                                til  /S^  /Zl
                                                               8
                                                                9     10
                                    % Excess 0
    Figure  12.   The Effect of the Addition  of Limestone to the Utah Coal Prior
                to Pulverization on SO- Emissions  (Ca/S - 1.0).
                                      103

-------
   3600
  .3000
  2500
o
«
o
   1500
 CVJ
o
in
   1000
    500
Urtstagad High Strifur Coat
Staged High Sulfur Coal
ISnitaged High Sulfur Coal * Sorbent
Staged High Sulfur Coal + Sorbent
Uasttged High Sulfur Coal After Sorber
                                                                8
                                                                  10
   Figure 13.  The Effect of the Addition of Limestone to the Western Kentucky
               Coal Prior to Pulverization on S0_ Emissions  (Ca/S - 1.0).
                                        104

-------
900


800



700


600
o
**
o
   400


   300


   200



   100


     0
                    T
T
             Unstaged  Utah
             Staged  Utah
             Unstaged  Utah  + Sorbent
             Staged  Utah +  Sorbent
             Staged  Utah After Sorbent
                                   4      5
                                  % Excess 0,
                                    8
                                                                         10
     Figure 14.  The Effect of Sorbent on NO  Emissions (Utah Coal Ca/S = 1)
                                       105

-------
     [utah Coal, SR,  * 0.25,  SAj  =  45°,  SA2  =  60°, M1d, 3 % Excess OJ
   500 i	1	1	1   J
   400
0^300
Q.
Q.
   200
   100
     0
      0.4
                                              42 x 106 Btu/hr
                                      100 x 10° Btu/hr
                                    •50 x 10° Btu/hr
0.6
                                 SR,
0.8
                     Figure  15.  Effect of Scale.
                                106

-------
   700
                        [Utah Coal,  3% Excess 02]
  600
   500
Q 400
 CM
O
   300
ex
Q.
   200
   100
                           I         I         I
12 x 10° Btu/hr
                 I          I          I         I
      0.3       0.4       0.5       0.6       0.7
                               SR2
        Figure 16.   Performance at Optimum Configuration.
                                 107

-------
     100
    CM
   o
   f
   ce.
Mineral
Limestone +
Utah Low S.
Coal






^^•i
53

^^^m
73





88







NaC03



Limestone +
High S. Coal


50



mmm^m
41
* NaHC03
80]








Pure
NaHCO,



52


o


123 1 24 2
Ca/S Ca/S N*/S Na/S
Figure 17.  Summary of  Sulfur  Capture  Results Using Various Sorbents

            and Molar Ratios.
                                  108

-------
TABLE I.  COAL COMPOSITION

Proximate Analysis, %
(as received)
Moisture
Volatile
Ash
Fixed Carbon
Heating Value, Btu/lb
Ultimate Analysis, %
(DAP)
Carbon
Hydrogen
Nitrogen
Sulfur
Oxygen (by difference)
UTAH


6.39
38.89
7.4
47.32
12,340


79.45
6.03
1.44
0.74
12.34
W. VIRGINIA


1.29
31.01
13.76
53.94
12,500


83.92
5.66
1.55
2.06
6.81
W. KENTUCKY


3.68
38.19
17.90
40.23
10,640


77.57
5.7
1.63
3.8
11.3

-------
    ALTERNATE FUELS AND LOW

     NOX TANGENTIAL BURNER

      DEVELOPMENT PROGRAM
              BY:
       Richard A. Brown
      Acurex Corporation
Energy & Environmental Division
Mountain View, California 94042
                111

-------
                                   ABSTRACT

      The EPA is continuing to explore control technology in the areas of alter-
nate fuels, waste fuels and control of NO  emissions from tangentially fired
boilers in the government owned test facility located at Acurex Corporation.

      Results from baseline and combustion control technology tests on coal oil
mixtures show that in general NO  emissions of coal oil mixtures fall between
                                A
the NOY emissions of the parent oil and coal.  Tests of NO  control techniques
      A                                                   A
including staging and a low NO  burner design showed varying degrees of control
                              A.
depending on the particular fuel combination.

      Baseline tests on refuse derived fuels (RDF) fired either with pulverized
coal or natural gas were determined.  Emissions assessments made include NO ,
                                                                           A
CO, particulate loading and size distribution, twelve trace metals and a cursory
search.for FCB and POMs.  NO  emissions decreased as the percent RDF increased
                            A
even though the available fuel nitrogen increased.  Particulate loadings from
the RDF were concentrated in the less than 1 ]i size fraction.

      An extensive research program has been initiated to develop a low NOY
                                                                          A
burner design for coal fired tangential boilers.  The initial tasks consist
of firing  the pilot scale facility on natural gas, natural gas doped with
Nitrogen species, char, char plus natural gas, and coal to ascertain the
relative importance of the various flame regions, coal volatiles and nitrogen
evolution.  In addition, a water model of the pilot scale unit is being con-
structed to ascertain the mixing patterns of the pilot scale unit as compared
to the full scale unit.
                                        112

-------
                                   SECTION 1
                                 INTRODUCTION

      The EPA multiburner multifuel test facility is continuing to explore
control technology for alternate fuels and waste fuels.  In addition, an
extensive research and development program has recently been initiated to
control NOX emissions from tangentially fired utility boilers.  This paper
will:   (1)  review the data from tests on coal-oil mixtures and cofiring of
refuse-derived fuels, and (2) present the test plans and objectives for the
NO  tangential burner development program.

      The EPA experimental multiburner, furnace facility (Figure 1) was
developed to study NO  control technology problems associated with large - scale
                     A
utility and industrial boilers.  Details on this facility design have been dis-
cussed in other papers (References 1 and 2).

      The furnace fires from 293kw-thermal to 880kw-thermal (1 to 3 x 10  Btu/hr)
depending on the fuel and heat release per unit volume being simulated.  The
facility may either be front-wall fired using one to five variable swirl block
burners,, or it may be corner-fired using four to eight tangentially fired burners
patterned after Combustion Engineering's design.  A horizontal extension confi-
guration (Figure 2) with a single large variable swirl block burner was used to
conduct the coal-oil mixture tests.  Radiant section cooling was simulated by
placing water cooling coils on the inside circumference of the refractory tunnel.
A convective section quenched the flue gases at the appropriate residence time.

      The standard gaseous emissions (NO, CO, C0?, 02» S0_) were continuously
monitored throughout the test program.'  Particulates were obtained using a high
volume EPA method 5 stack sampler; trace metals and organics sampling was

                                        113

-------
obtained using the Source Assessment Sampling System  (SASS)  (Reference 3).

Coal-Oil Mixture Emissions

      In many industrial and utility boilers, conversion  to  coal or partial coal
could be accelerated if only minor modification to  the boiler were required.
Therefore, there is a growing interest in  firing a  mixture of coal and fuel oil
in existing units.

      It is now necessary to determine if  current control technology for  coal
combustion is applicable to coal-oil systems or if  further work is needed to
ensure that pollution standards can be met.

      The coal-oil mixtures examined were  prepared  from parent fuels which
represent a broad range of fuel compositions.  The  compositions of the parent
oils, coals, and coal-oil mixtures are listed in Table I.  The coals, pulverized
to 70 percent 200 mesh, were blended with  the fuel  oils and  a suspension  additive
supplied by Carbonoyl Company that constituted 3.75 percent  by weight of  the
mixtures.  The mixtures were heated, thoroughly agitated  by  a variable speed
rotary pump, and then delivered to the burner nozzle  through heat traced  lines.

      The two nozzle types used were an air atomized  combustion swirl Delavan
Nozzle and an air atomized Sonicore Nozzle.  Both of  these nozzles underwent
erosion during the tests.  The Delavan 440 hardened stainless steel nozzle
lasted approximately 3 to 4 hours on a coal-oil mixture of 30 percent coal
until significant erosion produced an unstable flame.  The Sonicore nozzle
with a stellite tip lasted about 8 to 12 hours until performance deteriorated.
Other operational problems included the rapid deterioration  of pump seals and
the gradual buildup and plugging of lines  in the delivery system.

      Figure 3 illustrates the results of  the baseline emission tests.   The NO
levels for the Chevron-based fuels fall in an intermediate range between the
parent fuels, while the data for the Amerada base mixtures are closer to the
parent oil.   This difference may be caused by sulfur in the fuel,  atomization
characteristics,  and how the particular oil and  coal volatilize.
                                       114

-------
     At  this point  there  is  insufficient data to determine which mechanism is
causing this effect; however,  not all fuel combinations behave in the same
manner.

     In  this study, established combustion control technology for pulverized
fuel was  applied  to  the  coal-oil mixture, including staging the combustion air,
burner air distribution  or "low NO  burner" configurations and combined staging
,                                  A
and low NO  burner design.  Only the combined low NO,, and staging tests will be
reported  here.  The  low  NOX burner involves axial injection of tertiary air at
the periphery of  the burner diameter.

     The curves  in Figure 4  compare the  results  of applying burner air distri-
bution plus staging to straight burner air distribution.   The purpose of this
comparision was to evaluate the mixture responses with an enriched flame zone.
For the western Kentucky/Amerada mixture, further enriching the flame zone re-
sulted in higher  NO levels.   This result  may further illustrate the response of
the Amerada high  sulfur  parent oil to fuel-rich conditions.   The Montana/Chevron
mixture results validate that each mixture responded favorably to the air distri-
bution, but that  the composition of each  mixture  leads to a unique emission
curve.

     The following  conclusions can be drawn:

     •   NO emissions from coal-oil combustion are affected by the composition
         of the parent fuels  which make up the mixture
     •   Present  conventional control technology  used for pulverized fuel com-
         bustion  reduces NO emissions produced by coal-oil mixture combustion
     •   Coal-oil mixture  flames do not react consistently to some forms of
         control  technology,  implying that the chemical properties of the fuel
         and the  physical  processes which take place during combustion are both
         important
     •   Coal-oil mixture  combustion is different from either pulverized coal
         or residual oil combustion.  Therefore,  predictions of NO emissions
         based on emission levels of the  parent fuels often will not be valid
                                       115

-------
      In order to analyze and understand coal-oil mixture combustion,  better
understanding of the combustion processes of each parent fuel must be  obtained.
Also, the combustion of coal-oil mixture must be examined to determine if it is
merely a combination of the two individual processes or instead is a completely
different physical and chemical phenomena.

      Future work should examine the role of fuel-bound nitrogen utilizing flue
gas recirculation, nitrogen evolution studies of the parent oils, and  the effect
each fuel exerts on the other, such as shielding or physical separation in the
droplets.
                                       116

-------
                                   SECTION 2
                          REFUSE-DERIVED FUEL TESTING
      There is condiderable heating value (4000 to 7000 Btu/lbm) associated with
municipal solid waste.  If this resource could be used in steam boilers rather
than lost by incineration, a significant energy resource would be tapped.

      Research studies have included using heat recovery incinerators, spreader
type stokers and suspension firing in large electric utility boilers.  The EPA
has supported experiments in cofiring the refuse-derived fuel (RDF) in full
scale boilers in St. Louis, Missouri, Ames, Iowa, and Columbus, Ohio (References
4 through 7).

      Although these full scale experiments are providing useful data, problems
associated with the many varieties of RDF need to be studied.  Because the refuse
comes from local municipalities, there can be significant variations in the com-
bustion and environmental characteristics of the fuel from season to season or
from locale to locale.

      There is little published data on the emissions from RDF when cofired with
other fuels.  Kilgroe (Reference 8) reported that the St. Louis demonstration
site produced a moderate increase in chloride emissions but that the RDF did not
significantly affect the SO  or NO  emissions.  Little information is available
                           i      X
on the amount of trace metals and organics or the nature of the particulates.
The work performed here provides the initial data base to answer environmental
questions on four RDF's.
                                       117

-------
       The refuse-derived fuel tests were conducted in the main firebox in  the
 tangentially fired mode.   First,  a feed system (Figure 5) was designed to  control
 and measure from 10 to  60 Ibs/hr  of varied refuse-derived materials.   RDF  is
 delivered to the upper  part of two diagonally opposed corner-fired burners.
 Natural gas or coal is  also delivered  to these burners and  to the other two
 burners.  The RDF feed  system consists of a rotating drum hopper which deposits
 the material on a conveyor belt.   The  conveyor delivers the material  to a  vertical
 downcomer, where it is  pushed through  by a blast  of air.  Additional  air at the
 junction of the vertical downcomer and horizontal feed tube conveys the RDF into
 the furnace through a horizontal  water-cooled feed tube.  The RDF feedrate is
 controlled by a variable-speed drive on the feed  belt;  the  drum is maintained at
 constant optimum speed  to keep the feed belt full.

       The delivery tube is sized  to prevent blockage while  minimizing the  trans-
 port air.   This sizing  is critical for a small-scale facility where the minimum
 pipe size is governed by the maximum particle size and  the  minimum conveyance
 air needed to keep the  material suspended.   The received  RDF material was
 shredded from a nominal 2 to 4 inches  down to 1 to 2 inches using a conventional
 garden shredder to reduce the feed tube diameter  and transport air flow to an
 acceptable level.

      The  test program  determined  the gaseious, particulate trace metal and organic
 emissions of refuse-derived fuel  from  San Diego,  California;  Richmond,  California;
 the Americology Facility in Milwaukee,  Wisconsin;  and Ames,  Iowa.

       All of these materials had gone through metals and glass separation  and
a primary shredding.  Table II shows the composition and heating value for  each
fuel type.

       Figure 6 shows the effect of NO  versus excess air for the four  fuels at
 20  percent RDF and 80 percent natural  gas.  Although the  NO levels are not par-
 ticularly high,  there was a definite difference between the fuels.  The NO also
 increases with both excess air and increases in RDF. Also, when the  percent
 RDF is increased and cofired with coal, the NO levels decrease (Figure 7)  while
 total fuel nitrogen increases.  This is possibly  the result of enriched fuel
                                       118

-------
jets at the coal-refuse injection  guns.   Except at very low excess air (5 percent)
CO levels were always less than  100  ppm.

     Results from the particulate,  trace metal,  and organics sampling also
provide some interesting preliminary information of RDF emissions.  Table III
shows the results of particulate concentrations in the various size cuts for
four RDF materials cofired with  natural  gas.   Although the total particulate
quantity was quite low in all  cases, the majority of particles were smaller
than 1 U and collected only on the filter.  Although the particulate may be
rather friable and break up in the sampling equipment during collection, it may
eventually end up in the respirable  size range.

     Table IV shows particulate loading in the same size cuts for two levels of
Richmond RDF cofired with coal and for coal alone.   With the substitution of RDF,
the total grain loadings decreased with  increasing RDF.   However, in both cases,
adding RDF increased the percent of  material  in the. 1 u size cut over coal alone.
Thus, it appears that adding RDF may increase the grain loading in size cuts
less than 1 U.  This result could  produce problems for flyash collection equip-
nent.  Percent combustibles in the particulate were generally less than 2 percent
except when the excess air levels  were 10 percent or less.  This result also
corresponds to generally low CO  (< 100 ppm) and unburned hydrocarbon levels.

     Table V lists the total  trace  metals  in micrograms/Btu found in the parti-
culates and the condensable vapor  for:   1)  coal only, 2)  coal plus 10 percent
BDF, and 3) gas plus 10 percent  RDF. Increases in trace metal concentrations
varied among the three tests.  In  the coal  only test, the lead concentration was
exceptionally high.  In addition,  no correlation was found as the percent of RDF
was increased.

     It is difficult to draw  any  conclusions from this trace metal data.
Several factors may be contributing  to  the  data variability:

     •  The RDF material is nonhomogeneous and will vary from minute to minute,
        hour to hour, and season  to season
     •  Metals from the furnace and sampling system could contribute to the
        trace metal loadings
                                       119

-------
      •  Hold up in the convective section
      •  Analytical error

      These factors indicate the need for a broader data base to draw meaningful
conclusions on trace metal concentration when cofiring RDF with other fuels.
This data will contribute to that base, but a larger sample of data is needed
to statistically determine real trace metal effects.

      In addition to trace metals, a limited search for organics in terms of
(PNA or PCBfs) was undertaken.   A portion of the participate and the XAD-2
organic section resin of the SASS train were analyzed by liquid chromatography
according to EPA Level 1 procedures (Reference 9).  Of the material divided into
the standard seven cuts, only cuts two and three were expected to contain PNA
and PCB.   Therefore, these two fractions were combined for a single GC/MS
analysis.  Five out of nine tests where organic samples were taken contained no
detectable compounds.  Test conditions, and the PNA found in the remaining
samples,  are listed in Table VII.   No PCB was found in any of the test samples.

      Little organic material,  combustible CO, and unburned hydrocarbons were
found in the particulate and gaseous streams.  It has been reported (Reference
8) that significant quantities  of unburned material have been found in full-scale
tests.  These pilot scale tests have higher combustion efficiency over full-scale
tests possibly because of the additional shredding and/or hot refractory walls
providing an improved ignition source.

      In summary,  up to 30 percent RDF may be cofired in a subscale test facility
without experiencing a reduction in combustion efficiency.  Additional studies
are necessary to determine how this technology can be implemented in a full-scale
facility using the same degree of efficiency but a higher RDF percentage.
Furthermore, the flame's heat transfer characteristics must be clearly defined
to determine the effect on the boiler stearnside or to design a boiler specifically
for cofiring RDF.   Finally, more data is needed to statistically determine the
trace metal and organic makeup of RDF cofired boilers emissions.
                                       120

-------
                                  SECTION 3
                   LOW NO  TANGENTIAL BURNER DEVELOPMENT
                         A




INTRODUCTION



      The RDF work is being followed by development of a low NO  tangential
                                                               X

burner for utility boilers.  The following sections describe the background and


program plans for this work.
      Tangential coal-fired boilers produce approximately 10 percent of all


stationary source NO  emissions and consume, in Btu's, approximately 9 percent
                    A.

of all fuel used in stationary sources.  Reducing tangentially coal-fired boiler


NO  emissions is necessary to maintain ambient air quality in the United States.
  A.

The need to address the long-term capabilities of the corner-fired boiler is


becoming more apparent as recent developments in low NO  burners for front wall-
                                                       A.

fired boilers have reached the 125 ppm range (Reference 10).
      Therefore, the EPA has contracted with Acurex Corporation to conduct a


     ith program to develop a low NOX burner concept for 1


boilers.  The two principle goals of this program are to:
31-month program to develop a low NOX burner concept for tangentially fired
      •  develop a better understanding of processes controlling NO  formation
                                                                   A

         during combustion of pulverized coal in tangentially fired furnaces


      •  develop low (50 to 100 ppm) NO  combustion concepts for retrofitting
                                       A.

         or new designs of tangentially fired boilers





       Figure 8 shows the flow patterns typically present in the tangentially


fired system.  Air and fuel are distributed in a nonswirling slow mix character


through registers located in vertical orientation at each of the four corners.


Generally there are three to five fuel register levels.  Fuel ignition is pro-



                                       121

-------
vided by impingement of hot laterally adjacent streams and large-scale internal
recirculation of combusted gases.  Thus, the ignition and combustion of each
corner jet serves to ignite and promote burning in the downstream adjacent jet.
A large-scale vortex is formed by the tangential nature of the air and fuel
injection.  This vortex motion promotes mixing of fuel-rich regions, internal
recirculation flow patterns, and the slow-mix nature of the jets should inherently
make the tangentially fired system low in NO  production.   The pilot-scale
                                            A.
facility models most of the significant parameters of the tangential system.  For
example, the fuel and air are distributed in a manner similar to the full-scale
system.  The bulk residence time and heat transfer rates are also maintained.
However, the residence time between adjacent jets is condiderably different
than in the full-scale system,  and fuel is introduced only at one or two.
levels.

      Although there is a three-dimensional aspect to the tangential boiler flow
patterns, our primary concern will be the flows encountered at a single elevation.
Figure 9 and the following paragraphs define three regions in a tangentially
fired boiler:

      •  Near burner-early mixing
      •  Intermediate jet flame
      •  Recirculating fireball

Near Burner-Early Mixing Region
      In this region, flow is predominantly an axial jet that entrains air and
small amounts of flue gas.  The region consists of approximately the first five
flow path jet diameters where significant coal devolatilization occurs, including
the flame's ignition region.  It is believed that a high peak NO is formed early
in this region.

Intermediate Zone
      In this region, consisting of five to perhaps 30 fuel jet diameters, the
remaining air is mixed with the fuel, the majority of the fuel is burned, and
considerable flue products are entrained.  In this region, one side of the flame
interacts with the fireball, and the opposite side interacts with the boiler

                                        122

-------
wall.  Locally rich combustion zones which have the potential  to  reduce entrained
burnt gas NO  by flame processing can exist here.
           A

Fireball
     The fireball consists of the recirculating  zone  in  the firebox center.   It
is believed that this zone consists mostly of  flue gas products and regions where
char is burned out.  However, it is not known  at  this  time  if  any additional
HO is formed here, or if NO formed early  is reduced.

     The proposed research program focuses on the physical and chemical pro-
cesses associated with NO  formation in the near, intermediate and fireball
''                        A
zones.

APPROACH
The proposed approach is deiplayed in Table VII.

Task A — Tangentially Fired System Definition and Evaluation
     The primary goal of Task A is to define  the phenomena which control the
formation and reduction of both fuel and  thermal  NO in a  classical tangentially
fired boiler.  This task will isolate the critical chemical and physical processes
(e.g., early mixing, flame shielding, devolatization,  etc.) and establish the
design parameters effect on these processes.   Task A will also investigate the
uniqueness of tangential firing and attempt to definitively show why such a
system generally has lower NO  emissions  than  wall-fired units.   The majority of
                            A
this work will be done in the existing pilot-scale  facility.   In each region the
importance of the volatiles and char burnout will be  investigated by doping char,
natural gas flames, and natural gas plus  char  with  nitrogen and sulfur compounds.

     We will also  focus on hot sampling  for NO  and flame-flame interaction
studies.  It is possible that NO  formed in one flame  can be reduced by passing
through an adjacent flame.  Figure  10  shows parameters that will be investigated,
including the intersection point  of  the  two flames, X, and the intersection angle
over a range of fuel richness  in  each  flame.

     Therefore,  these subtasks will help focus  subsequent tests on the those
chemical/physical processes offering  the  greatest potential for NOX reduction.

                                       123

-------
      At the completion of Task A, sufficient understanding and information on
tangential system NO., formation processes will be available to permit the design,
construction, and test of specialized low NO  burner/firebox subcomponent
                                            X.
hardware.

Task B — Optimization of Near Burner and Intermediate Zone NOV Levels
                                                              A
      The primary objective of this task is to follow through on the guidance
provided by Task A; and to design, construct, and test (in subscale), low NO
                                                                            A
subcomponent concepts of near and intermediate zones.  Sufficient flexibility
will be incorporated into the hardware to define optimal low NO  conditions for
                                                               A
both the near burner and intermediate field zones.  Parameters optimized during
these tests will be air and flue gas distribution, and mixing and temperature
history, or  local  cooling rates.  The B tasks will utilize a new versatile
burner which fires into the horizontal furnace extension.  In these initial
 tasks,  the near burner early mixing zone oxygen,  flue gas availability and
 temperature/cooling  rate history will be optimized for low NO .  Optimal and
                                                             A
 off-optimal  near  burner configurations will then be tested with intermediate
 zone air and flue gas distribution, and wall cooling.  For these tests, air
and flue gas introduction ports, and wall cooling panels will be incorporated
into the horizontal  firebox appendage.

      Optimal near and intermediate zone low NOV  concepts will emerge from the
                                               A
Task B tests.  These concepts will be applied in the pilot-scale tangentially
fired tests carried out in Task C.
Task C — Optimization of Coal-Fired Tangential System Low NO  Concepts
        	                  	    	         	              _   A    	
      The primary objective of this task is to define and demonstrate in pilot
scale, low NOX tangentially fired systems for several coal types.  This task
will adapt the low NO  subcomponent concepts developed in Task B into the
tangential firebox and optimize the complete system for low NOY.  The system
                                                              A
will be optimized for each coal type tested.

      The optimized low NOX tangential system concept will be demonstrated on a
single characteristic coal.  Sufficient parameter variation will be built into

                                       124

-------
the hardware to optimize the system for the characteristic coal used in this
task or any other commonly used coal.  Then the low NOV concept will be opti-
                                                      x
mlzed for several additional coal types.

Fireball Simulation Study
      Throughout the 31-month program, a background study will determine how
to best simulate a tangentially fired system in subscale.  The study will
identify and quantify the following parameters important to modeling:

      •  Mixing rates
      •  Temperature/time history
      •  Fuel type and size
      •  Burner firebox geometry

      Previous hot and cold flow studies on tangential systems, such as those
by Juniper (Reference 11) will be studied.

      To assess the overall flow simulation of the pilot scale versus the full
scale, a water model of each system is being constructed.  Flow patterns will
be observed using a dye injection technique and recorded by still and motion
pictures.  To assess the mixing patterns for the new concept in the tangential
mode, models will also be made of the low NOY burner design concepts.
                                            A.

CONCLUSION
      Completion of all proposed program tasks will result in a pilot-scale
demonstration of a low NO  tangentially fired system for several coal types.
                         A

      Following the program's completion, research will be conducted to either:
1) demonstrate the low N0_ concept in larger scale with possibly a greater
variety of fuels, or 2) carry out additional subscale tests.  This will eventually
lead to an optimal low NO  full-scale tangentially fired system capable of firing
                         A
several coal types.
                                        125

-------
                                  SECTION 4
                                   SUMMARY

      The paper presented three facets of EPA's program on  emission character-
 ization and control of alternate waste and  conventional fuels.  First  the
 facility simulated the combustion characteristics  of  a  package boiler  to test
 emission characteristics of  coal oil mixtures.  Different fuel combinations of
 a  coal-oil mixture responded in a unique manner to conventional control techno-
 logy.  In addition, many operational problems  (pumping  characteristics, pump
 seal wear, and nozzle erosion) were discovered.

      Tests were also conducted to determine the feasibility of subscale
 evaluation of cofiring refuse-derived fuels.  A unique  feed system was developed
 to deliver a variety of EDFs from 10 to 60 Ibs/hr.  However, it was found that
 the test furnace was even more efficient in achieving complete combustion than
 full-scale units.

      Tests showed slight increases in NO and SO  emissions over natural gas but
 demonstrated a reduction in these levels when cofired with pulverized coal.
When cofired with natural gas, the particualte was found concentrated in less
 than a 1 y range.  The trace metal analysis showed no conclusive trends.  Few
 PNA organics emissions and no PCB materials were found.

      The program plan to develop a low NOX tangential burner was presented.
This plan includes tests to: 1) determine formation and destruction of NO in
current designs;  2)  optimize early and intermediate zone environmental conditions,
and 3)  optimize low NOX concepts in the pilot-scale facility.   A water modeling
                                        126

-------
study will also be employed  to  study the mixing patterns in the pilot-scale and
full-scale systems.
                                        127

-------
                             SECTION 5
                            REFERENCES

1.  Brown, R. A. et al.  Pilot Scale Investigation of Combustion Modification
    Techniques for N<>x Control in Industrial and Utility Boilers.  EPA-600/2-
    76-152b.  Proceedings of the Stationary Source Combustion Symposium,
    Volume II, June 1976.
2.  Brown, R. A. et al, Investigation of Staging Parameters for NO  Control
                                                                  A
    in  Both Wall and Tangentially Coal-Fired Boilers.  EPA-600/7-77-073C.
    Proceedings of the Second Stationary Source Combustion Symposium,
    Volume III, Stationary Engine, Industrial Process Combustion Systems,
    and Advanced Processes, July 1977.
3.  Blake, D.  Source Assessment System Design and Development, EPA 600/7-
    78-018, August 1977.
4.  Vaughan, D. A.  and Associates.  Report of First Year Research on
    Environmental Effects of Utilizing Solid Waste as a Supplementary
    Powerplant Fuel, Battelle Columbus Ohio Laboratories, EPA Research
    Grant R-804008, June 1975.
5.  Nydick S. E. and Hurley, J. R.  Study Program to Investigate Use of
    Solid Waste as a Supplementary Fuel in Industrial Boilers, Thermo-
    Electron Corporation.  EPA Contract No. 68-03-3005, January 1976.
6.  Riley, B. T.  Preliminary Assessment of the Feasibility of Utilizing
    Densified Refuse Derived fuel (DRDF) as a Supplementary Fuel for
    Stoker Fired Boilers, published report to EPA, 1975.
7.  Vaughan, D. A., Krause, H. H., Hunt, J. F., Cover, P. W., Dickson,
    J. D., and Boyd, W. K.  Environmental Effects of Utilizing Solid
    Waste as a Supplementary Powerplant Fuel.  Seventh Quarterly Progress
    Report, EPA Research Grant 804008-02-1, 1975.
                                    128

-------
 8.   Kilgroe, J.D., Shannon, L.J., Gorinan, P.  Environmental Studies on
     the St. Louis Union Electirc Refuse Firing Demonstration.
 9.   IERL/RTP Procedures Manual:  Level 1, Environmental Assessment,
     Second Edition.  June 1976.
10.   New Coal Burner May Reduce Nitrogen Dioxide Emission 80 to 85 percent.
     EPA/Environmental News Release, Office of Public Awareness (A-107),
     Washington, D.C., September 22, 1978
11.   Juniper, L. A. Flame Measurements in a Brown Coal Fired Furnace,
     Combustion Institute European Symposium,  Academic Press, 1973.
                                    129

-------
Figure 1.   Photograph of experimental  multiburner furnace.

-------
IFRF Burner
                          ,
                         i"i/i;J''*!
                         i'i i" i/A'/1!
                         "" 1r.il'5!
                          '!
                 Radiant

                 Section
                                 I  /v y\r
                                             I
                         I
i]
Corrective  Main Rrebox

  Section
                Figure 2. Package Boiler Configuration


                                131

-------
1400
           Chevron Oil
Q.
1200

1000

 800

 600

 400

 200
                                     1
        10   20   30   40
          % Excess Air
     O Chevron No. 6
     Q Virginia/Chevron 30%
     & Montana/Chevron 30%
     0 Montana Coal
     o Virginia Coal
1400

1200

1000

 800

 600

 400

 200
                                                   Amerada Oil
                                                10    20   30   40
                                                   % Excess Air
                                                O Amerada No. 6
                                                & Montana/Amerada
                                                o W. Kty/Amerada
                                                O W. Kentucky Coal
                                                a Montana Coal
               Figure 3.  COM Baseline Emissions
                              132

-------
      30% Virginia Coal/
         Chevron Oil
    30% W. Kty Coal/
       Amerada Oil
                                                      30% Montana Coal/
                                                         Chevron Oil
,

I
800

600-
 1200
0
       .55  .65  .75 .85
                      D
                      O
  800

^600

SW

&200
                          ^8001

                          260*
                          o
                          Z
         .55 .65 .75 .85

           SRiA
                                                            .55 .65 .75 .85
SRi = 0.95 — Staging + Low NOx Burner
SRt ^ 1.20 — Low NOx Burner Only
                  Figure 4.  Combustion Modification Results
                                      133

-------
                                 Tangential Burner
Main Firebox
   Figure 5.  RDF Feed  System and Firebox




                     134

-------
        20% Refuse/Natural Gas
   200
CM
o
Q.
Q.
   100
% NDMMF
      1.26
        o Richmond
        A Americology
       i° San Diego
                               o
                                 Ames
            10    20

                 EA%
30
    Figure 6.  Baseline NOX - Emissions - RDF & Gas
                    135

-------
        Richmond Refuse/Pittsburg Coal
 CM
O
Q.
   600
    500
    400
    300
   200
    100
       0      10      20      30
             % RDF (Heat Input)
40
    Figure 7.  Effect of Percent RDF Coal & RDF

                       136

-------
CO
                          Top View  /air
                                                                                             bulk  of  combustion
                                                                                             air
                                                                       Side View
                                      air and fuel  registers

                            Figure 8.   Tangentially Fired  System Flow Patterns

-------
Figure 9.  Tangential Burner Flame Regions




                       138

-------
Short flame
               Intersection distance
                           Figure 10.   Intersecting Flame
                                            139

-------
TABLE IA.  COAL ANALYSIS
Proximate (Z WO^-— 	
^ Coal
Moisture
Volatiles
Fixed Carbon
Ash
Rank
Ultimate (Z Wt)
Carbon
Hydrogen
Nitrogen
Oxygen
Sulfur
Ash
Heat of Combustion Btu/lb

Montana
21.23
35.16
34.27
9.34
Sub-bit. C.

53.26
3.35
0.87
11.16
0.78
9.34
8,972

Virginia
0.31
31.9
51.4
16.5
High-Vol. A

71.11
4.46
1.68
. 4.24
2.02
16.5
14,079

W. Kentucky
5.0
36.55
50.98
7.47
High-Vol. B

69.79
4.79
1.34
8.65
2.95
7.47
12,349
TABLE IB.  OIL AND  COM ANALYSIS
-— -Jlixture 30Z (Wt)
Analysis """— — ^JJoal
Carbon
Hydrogen
Nitrogen
Oxygen
Sulfur
Ash
Moisture
Heat of Combustion Btu/lb
API Gravity
Flashpoint COC°F
Viscosity, SSO at 100°F
V. Kentucky/
Amerada
80.23
9.00
0.63
3.92
2.44
2.26
1.52
17,600



Montana/
Amerada
75.27
8.57
0.49
4.67
1.79
2.82
6.39
16,600



Montana/
Chevron
75.88
8.37
. 0.83
4.77
0.89
2.87
6.37
15,500



Virginia/
Chevron
81.23
8.70
1.07
2.73
1.26
4.95
0.09
17,000



Amerada
Hess 16
84.71
10.75
0.36
1.93
2.22
0.93

19,867
15.3
. 204.0
2,500
Chevron
#6
85.57
10.52
0.81
2.08
0.93


18,292
12.3
182.0
4,900.0
	 •*
             140

-------
                        TABLE II.  RDF FUEL ANALYSES
Ultimate Analysis*
Carbon %
Hydrogen %
Oxygen %
Nitrogen %
Sulfur %
Ash %
Moisture %
(as received)
Chlorine %
Heating Value
Btu/lb
Fuel Type
Pittsburg
No. 8 Coal
75.23
5.15
8.12
1.49
2.51
7.50

0.93
0.14

13,545
Richmond
Refuse
42.60
6.26
37.90
0.83
0.16
12.25

23.8
.46

7696
Ames
Refuse
40.49
6.01
30.04
0.73
0.35
22.38

15.2
.43

7831
Americology
Refuse
40.29
5.88
25.20
0.91
0.17
27.55

24.4
.72

7164
San Diego
Refuse
38.01
5.64
17.40
0.69
0.21
38.05

26.3
.79

7146
Dry Basis
                                    141

-------
TABLE III.  EFFECT OF RDF TYPE ON PARTICULATE  SIZE DISTRIBUTION
Fuel
20% Ames
+ Nat Gas
20% Richmond
t- Nat Gas
20% Americo-
logy
f- Nat Gas
20% San Diego
f- Nat Gas
Filter
Qty (gr/ft3)
%
0.039
(69)
0.032
(91)
0.041
(86)
0.062
(80)
>10y
Qty (gr/ft3)
%
0.011
(19)
0.0004
(1.)
0.002
(5)
0.007
(9)
>3y
Qty (gr/ft3)
%
0.003
(6)
0.0004
(1)
0.002
(5.)
0.002
(3)
>iy
Qty (gr/ft3)
%
0.004
(6)
0.0024
(7)
0.002
(5)
0.006
(8)
                            142

-------
TABLE IV.  EFFECT OF COAL  & RDF CONCENTRATION ON PARTICIPATE SIZE DISTRIBUTION
„
m Richmond
Li RDF
ICoal
m Richmond
[ RDF
iCoal
bal Only
Filter
Qty (gr/ft3)
%
.026
(9.1)
.044
(7.6)
.021
(2.1)
>10y
Qty (gr/ft3)
%
.134
(47.3)
.269
(46.1)
.539
(56.3)
>3y
Qty (gr/ft3)
%
.107
(38.0)
.226
(38.8)
.344
(35.9)
>iy
3
Qty (gr/ft )
%
.016
(5.5)
.044
(7.6)
.055
(5.7)
                                        143

-------
 TABLE  V.   TRACE METAL CONCENTRATION FOR COAL VS  10%  RDF + COAL VS  10% RDF + GAS
                                   (yg/Btu)
ELEMENT
Cu
Zn
Mn
Pb
Cd
Be
Ti
Sb
Sn
Hg
As
Coal Only
Test #40*
1.9581
0.5294
0.1526
17.5319
0.0091
<0.0176
<1.7540
<0.0020
0.1300
<0.0015
<0.0881
10% RDF + Coal
Test #38a
<0.3319
0.8227
<0.2693
0.3091
0.0048
0.0013
<0.0587
<0.0090
<0.0913
<0.0173
<0.0323
10% RDF + Gas
Test #llba
0.3402
0.4468
0.0209
1.4996
0.0062
<0.0034
<0.0277
0.0333
<3.5033
<0.0009
<0.0184
20% excess air
                                       144

-------
TABLE VI.  ORGANICS FOUND
Test Condition
Gas Cofire
10% RDF
20% EA
Ames Fuel
Gas Cofire
10% RDF
20% EA
Richmond Fuel
Gas Cofire
10% RD?
20% EA
Americology Fi*el
Coal Cofire
10% RDF
20% EA
Organic
Fluoranthene
Pyrene
Phenanthrene
Fluoroanthene
Pyrene
Diphenyl Ether
Blphenyl Phenylether
Phenanthrene
Pyrene
Phenanthrene
Amount (ug/108 Btu)
10
332
64
160
576
3395
1697
59
104
98
             145

-------
  TABLE VII.   PHASE IV — LOW NO  TANGENTIALLY FIRED SYSTEM
                          TASK A


 Tangentially fired system definition and  evaluation


 —  chemical/physical processes defined

 —  early mixing evaluated

 —  flame processing of NO in intermediate  and  far zone evaluated




                          TASK B


 Optimization of tangentially fired near and intermediate zone NO


 —  early mixing,  cooling,  and entrainment  of flue gas optimized

 —  intermediate zone mixing,  cooling,  and  entrainment of flue
    gas optimized




                         TASK  C


Optimization of low NOX tangentially  fired  systems for several coals

—  optimal burners in firebox with optimal  fuel/air feeding of the
    fireball
                              146

-------
       POLLUTANT FORMATION DURING
FIXED-BED AND SUSPENSION COAL COMBUSTION
                   Bv:
             D. W. Pershing
             B. D. Beckstrom
             P. L. Case
             G. P. Starley
   Department of Chemical Enaineerina
           University of Utah
      Salt Lake City, Utah   84112
                     147

-------
                                  ABSTRACT

      This paper summarizes the overall scope and the progress made during
 the  first five months of a grant to study the formation of pollutant species,
 particularly nitrogen and sulfur oxides, during the fixed-bed and suspension
 combustion of coal.  The work will consider conditions typical of both
 pulverized and stoker-fired coal systems with the major emphasis on spreader/
 stoker fired boilers and furnaces.  Specifically the program will consider:
 1) the evolution and oxidation of fuel nitrogen and sulfur; 2) the retention
 of sulfur oxides by the ash and/or solid chemical sorbents in suspension-
 and  fixed-bed burning; and, 3) the effectiveness of distributed air addition
 for  nitrogen oxide control.  In addition, the study will attempt to quantify
 the  combustion process in a stoker environment and consider possible detri-
 mental effects of control technology on boiler operation.
      The approach is primarily experimental, utilizing a controlled mixing
 history furnace and fixed-bed reactor to investigate suspension and fixed-
 bed  combustion, respectively.  In addition, a model spreader/stoker system
will be fabricated, characterized, and used for evaluating advanced combustion
control concepts.  The suspension furnace has now been designed and most
of the refractory casting is complete.  The flow-control, system for this
furnace has beer, specified and essentially all of the required instrumentation
has been procured.   Design of the fixed-bed furnace has just been initiated.
                                      148

-------
                               ACKNOWLEDGMENTS

      This research was supported by the U. S. Environmental  Protection
Agency under Grant No. R-805899-01.  The considerable help and advice of
G. Blair Martin, the EPA Project Officer, Robert D. Giammar,  Battelle,
and Michael P. Heap, Vice President, Energy and Enviromental  Systems is
gratefully acknowledged.  In addition, thanks are due to Mrs. Colleen
Anderson for her help in completing the secretarial duties associated with
this project.
                                      149

-------
                              PURPOSE AND SCOPE

      This paper describes the progress of an investigation  to  study the
formation of air pollutant species, particularly NO,  N02>  S02,  and S03> during
the combustion of coal particles with a view toward providing technology for
the design of high-efficiency, low-emission boilers and furnaces.  The work
will consider conditions typical of both pulverized-coal and stoker-fired
systems, but the emphasis will be on the latter because they represent a
major source which has been largely overlooked previously.
      At present stoker-fired boilers are significant in terms  of both coal
consumption and environmental impact.  In 1974 almost 20 percent of the
coal consumed in the United States was burned in stoker systems.  Stoker-
fired,  water-tube boilers are the single largest source of  parti oil ate
emissions and the fourth largest source of SOX emissions, because neither
is  in general controlled.  They account for 13 percent of the  NOX emissions
from all coal systems and their environmental impact  may be  even more
significant than indicated by the mass emissions, because stoker systems
are often located in congested metropolitan areas.  Further, the energy
and environmental importance of stoker firing seems certain  to grow with
the increase in industrial coal utilization that is projected.   It  is
important,  therefore, that further knowledge be obtained on  the formation
and control  of nitrogen and sulfur oxides under conditions typical of  stoker
firing.
      In particular,  this program will consider the following  major  research
areas:
                   1.   The  evolution and oxidation of fuel
                       nitrogen  and sulfur;
                   2.   The  retention of sulfur oxides by ash
                       and/or  solid-chemical sorbents in both
                       suspension-  and fixed-bed burning;
                   3.   The effectiveness  of distributed air
                       addition  for NOx control  in both
                       pulverized-  and stoker-fired coal
                       systems.

-------
In addition,  the study will  attempt to quantify the combustion process  in
a stoker environment and consider possible detrimental  effects of control
technology on boiler operation.
      The approach is primarily  experimental,  utilizing a controlled mixing
history furnace and a fixed-bed  reactor to investigate  suspension and thick-
bed combustion, respectively.   In addition, a  model spreader-stoker system
will  ultimately be fabricated, characterized,  and used  for evaluating advanced
combustion control  concepts.   The next section of the paper describes the
details of the approach.  The  third and fourth sections then discuss the
progress made during the first five months of  the program.
                                       151

-------
                                  APPROACH
 OVERVIEW
       This  study is  scheduled for three years and is divided into five
 separate tasks:
                   Task  1:  Preparation of a detailed program plan.
                   Task  2:  An  experimental investigation of
                            suspension burning of bituminous coal.
                   Task  3:  An  experimental investigation of
                            fixed-bed burning of bituminous coal.
                   Task  4:  An  experimental and analytical optimi-
                            zation of combustion modification
                            technology using a model stoker system.
                   Task  5:  Experimental studies to extend the
                            control-technology results to other
                            fuels.
 Each task is  discussed in detail in the following paragraphs.  Tasks 1, 2, and
 3 have been initiated.   Task 4  will start in the second year of the program
 and Task  5 will  begin in the third year.
 EXPERIMENTAL PLAN PREPARATION
      Work is currently  underway on a detailed program plan for the first
 eighteen months of the study.   It will include a description of the initial
 experiments  to be conducted in  each task and detailed designs for the fixed-
 bed and suspension experimental furnaces.  To ensure the practicality
 of the proposed experimentation, conversations will be held with stoker
manufacturers  and users  to define in detail the problems associated with
industrial stoker usage,  the current trends, and future utilization of stoker
technology.  This will  include  site visits to at least one fabricator, one
industrial user,  and  other researchers in the field.
                                     152

-------
SUSPENSION BURNING
      The purpose of the second task is to investigate the formation
of pollutant emissions during the suspension-burning phase of the
combustion process.  These studies are intended to be relevant to both
pulverized-coal combustion and the suspension-burning phase of a spreader/
stoker system.  Initial studies will focus on the relationship between
the combustion parameters and the evolution and subsequent oxidation
of nitrogen and sulfur species.  The role of fuel-bound nitrogen in the
formation of NOX will  be established, both for small particles that are
essentially completely burned in the suspension phase and for large
particles that only undergo partial  oxidation.  The possibility of gas-
phase sulfur capture will also be explored in this task.
      Both the furnace and the flow-control system for the suspension-phase
studies have been designed as described in the next section of this paper.
Furnace fabrication is now more than 50 percent complete and essentially all
of the components for the flow-control system have been procured.
FIXED-BED COMBUSTION
      The purpose of the third task is to utilize a simple, relatively inexpen-
sive experimental test facility to study the combustion of coal in a fixed-
bed environment.  The experimental system will be designed to provide a
Lagrangian simulation of the time/temperature/environmental history seen by
a small section of i  ^ stoker bed as it passes through an actual furnace.
Both thick- and thin-jed systems will be considered.  The use of a simple,
fixed-bed coal furnace will allow evaluation of a large number of different
conditions, fuels, and additive materials without the complications associated
with moving stoker  grates.
      The design of the experimental system has not yet been resolved; however,
it is likely that the fixed-bed furnace will be a refractory chamber with a
removable bed section and auxiliary gas heating capability.  Thermocouples and
sampling probes will be positioned both within the bed and above the bed.
Provisions will be made for independent control of both underfire and overfire
air.  Present plans are to design the bed for firing rates up to 400,000
Btu/ft2/hr at a coal-feed rate of 20 Ibs/hr.
                                        153

-------
      The first set of experimental studies will focus on the relation
between fixed-bed combustion parameters and overall pollutant emissions.
The parameters of interest include:
                  . Bed composition
                  . Bed thickness
                  . Bed and gas-phase  temperature
                  . Free-stream  composition  (stoichiometry)
                  . Particle  size
                  . Bed-heating  rate and amount
                  . Bed-residence  time
                  . Amount and location of overfire air
Later testing will focus  on  the evolution and subsequent oxidation of fuel-
nitrogen and sulfur  species  both  within and above the bed.  Measurements of
nitrogen and sulfur  intermediates (e.g. NHg.HCNjHgS, etc.) will be attempted.
Solid samples will be removed from  the bed after various residence times and
analyzed for ultimate composition.  Other studies will investigate the sulfur-
retention  capabilities of the coal  ash itself and the potential of solid
sorbent  materials.   Of  particular  interest will be the distribution of the
sorbent  materials throughout the bed.
MODEL STOKER STUDIES
      The  purpose of the  fourth task  is to apply the  pollutant reduction
 concepts  developed in Tasks II  and III to an actual  stoker
environment.  An experimental furnace will be fabricated to directly simulate
both  spreaker/stoker and  fixed-bed  systems.  Initial experiments will be
conducted to define the interaction between the suspension burning and the
fixed-bed combustion since each of  these will have been considered separately
in the previous tasks.  Later testing will be directed toward defining and
optimizing new combustion control  technology.
ALTERNATE FUELS
      The  purpose of the final  four-month task is to study the impact of coal
composition on  combustion control  technology.   It is  anticipated that the fuels
to be studied here will  include  at least one desulfurized coal, a lignite, and
perhaps  a markedly different  alternate fuel  such as wood chips.
                                      154

-------
                            SUSPENSION BURNING

      The suspension furnace was designed to  meet the following  criteria:
           1.   Capable of simulating the environment seen by a  particle
               in the suspension burning phase of spreader/stoker
               combustion or in the burner zone of a PF  furnace.
           2.   Independent wall-temperature control  with high-
               temperature capabi1i ty.
           3.   Adequate probing and visual  access.
           4.   Variable residence time.
           5.   Well  defined fluid dynamics.
      The final furnace design utilizes  three types  of interchangeable  furnace
sections.  In  each case the innercombustion chamber Js 6 inches  in  diameter  and
the outer steel shell is 28-inches square.  A cylindrical  innerchamber  was
chosen (in preference to a square one) because it is more amenable  to analytical
modelling at some later date.  A square outside surface  was  selected because it
reduced the cost of fabrication and provided  a more  suitable base for mounting
probes, thermocouples, etc.
      The walls of the main furnace sections  consist of  an outer steel  shell,
5 inches of 1900°F insulating block, 4 inches of 2500°F  insulating  refractory,
and 2 inches of 3400°F high-temperature  castable refractory. Each  of the mid-
sections contain 5 2-inch diameter ports for  second-stage air injection and/or
insertion of species and temperature probes.
       As presently  conceived,  the  furnace  can be operated in two modes:
The self-sustaining flame mode and the tubular-reactor,  control-mixing  history
mode.  Figure  1 illustrates the first mode in which the  pulverized  coal and
at least part  of the combustion air (or artificial atmosphere)  enter the com-
bustion chamber at the top via a premixed burner.  Current plans are to use  the
premixed burner developed by Howard and Essenhigh (1).  A modification  of this
burner system  has proved suitable for studying NOY formation in  premixed, one-
                                                 /%
dimensional flows (2).  This mode of operation will  be used  for characterizing the
nitrogen and sulfur evolution for particle sizes and size distributions smaller
than 50 percent through 100 mesh (i.e., pulverized coal  and  the fines  in stoker-
coal system).   In this mode, the flue gases will exit at the bottom of  the
furnace through the horizontal section.
                                      155

-------
      In order to simulate the temperature and environment of the suspension-
 phase burning in a spreader/stoker system, the second mode of operation will
 be  utilized  (Fig. 2).  In this case, a high-intensity gas burner will  be
 attached to  the horizontal extension at the furnace bottom.  The large, stoker-
 sized coal particles will be fed from the top of the furnace and allowed to
 fall downward into the stream of hot combustion products flowing vertically
 upward.  In  this case, the premixed coal burner will be replaced with  a
 large, multihole coal injector and the majority of the combustion air  will be
 supplied from the bottom with the gas flame.  In an actual spreader/stoker
 system, the  suspension phase combustion occurs with reduced oxygen concentra-
 tions (less  than 10 percent) and with both nitrogen and sulfur species already
 present from the fixed-bed combustion.  To simulate this, S0« and/or ammonia
 will be added to the gas flame to produce the concentrations of SOg. S03, NO,
 and NOp that would normally be present in the combustion gases leaving the
 stoker bed.  Both acetylene and methane will be used as fuels for the  gas
 burner to  investigate potential XN/hydrocarbon interactions.  In this  mode,
 the main flue gases will exit at the top of the furnace through the horizontal
 section as shown (Fig. 2).
      As shown in Fig. 3, the side walls have provision for auxiliary  heating
 (or cooling) by firing natural gas or passing cooling air through outer
channels.   These channels are 2 inches in width and cover 50 percent of the
circumference.   This will allow partial independent control of the.radient
 heat transfer to the particles and more complete control of wall temperature.
The auxiliary firing sections are completely interchangeable so any portion
 of  the furnace can be heated (or cooled).
      The flow-control system for the suspension furnace has been designed
and the required components procured.   The flow system was selected to
provide precise metering of the streams entering the main combustion chamber
and adequate control  of the other auxiliary gas and air streams.  Flow
metering  and control  of the following  subsystems are included:
                                     156

-------
           1.  Main-chamber air.
           2.  Side-burner air.
           3.  Auxiliary-burner air.
           4.  Artificial oxidants
           5.  Side-burner gas.
           6.  Auxiliary-burner gas.

The combustion air is supplied by a 100 SCFM, 100 psi  rotary  vane  compressor.
Primary, secondary, and tertiary air are metered with  600 mm high acuracy
Brooks rotameters.  The second-stage air is metered with a large, 10-inch
Dwyer rotameter.  All other air and gas flows are metered with 10-inch
Dwyer rotameters.  The flow system also includes an artificial oxidant  supply,
so that the combustion air can be completely replaced  with an atmosphere
composed of argon, carbon dioxide,  and oxygen in varying proportions.  In
each case, the pure gases are supplied from high-pressure cylinders and
metered with rotameters.
                                      157

-------
                           ANALYTICAL MEASUREMENTS

       The furnace exhaust gases will be continuously monitored for NO, N02,
 CO, C02, and 02.   Using the instrumentation listed below:

               Gases                             Instrument
            NO, N02                      Thermoelectron, Model 10AR,
                                         chemiluminescent analyzer
            CO, C02                      Anarad,  Model AR-600 NDIR
            02                           Beckman, Model 755, paramagnetic
                                         oxygen analyzer
       The other stable, gas-phase pollutants and intermediate species of
 interest will be measured on a batch bases.  These include hydrogen sulfide
 (H2S), sulfur dioxide,  and trioxide (S02,  S03), carbonyl sulfide (COS),
 carbon disulfide  (CS2), ammonia (NH3),  and hydrogen cyanide (HCN).  S03 will
 be  measured using the standard ASTM condensation method.  Measurement of
 NH3 will  be attempted using the Axworthy chemiluminescent technique.  The
 other  species (S02,  H2S,  COS,  CS2>  and  HCN)  will be analyzed with a Hewlett-
 Packard, Model 5830A, dual  column,  temperature programmable gas chromatograph.
The sulfur compounds will be sensed  with a flame photometric detector.  The
nitrogen containing species (e.g., HCN)  will be measured with  the rubidium-bead
nitrogen/phosphorus flame detector.  Current plans are to use an acetone
washed 1/8-inch teflon tube, 4 feet  long,  packed with acetone washed Porapak
QS for separation of the sulfur compounds fcfter deSouza (3)).
                                      158

-------
                                REFERENCES


1.   Howard,  J.  B.  and R.  H.  Essenhigh.   Pyrolysis of Coal  Particles  in
    Pulverized  Fuel  Flames.   Ind. 'Eng.  Chem.,  Process Design  and  Development,
    6 (1),  1967.

2.   Wendt,  J. 0. L.,  J. W. Lee,  and D.  W.  Pershing.   Pollutant Control  Through
    Staged  Combustion of  Pulverized Coal.   FE-1817-4, U. S. Department  of
    Energy,  Washington, D.C.,  1978.  158pp.

3.   de Souza, T. L.  C., D. C.  Lane, and S.  P.  Bhatia. Analysis of Sulfur-
    Containing  Gases  by Gas-Solid Chromatography  on  a Specially Treated
    Porapak  QS  Column Packing.   Analytical  Chemistry, 47 (3), 1975.
                                      159

-------
                              JL-L
en
o
                               TT
                                     O O O O O
O  O O O  O
O  O O O  O
                                                                                  TT
                                                                                                           c < x

                                                                                                           (B fO fl)
                                                                                                           _I-J Q.


                                                                                                           OO N*
                                                                                                           C fl>

                                                                                                           -J Q.
                                Figure 1.    Suspension Furnace  -  Self-sustaining  Flame Mode

-------
\\\\\

SNNNX
  \\\\\\NS\\N\N\\\\\\
    \v\\\\\. \.\\vs.\\\\\
N\\\\S\\\\\\\\\\\\s
  W W\v vv\ vv vvWVX
\\\\\\\\\\
                       \\\\\\\\\\\\K\\N\\'
                       \\\\\\\\S\\\K\\\\\'
Figure 2.    Suspension Furnace - Tubular Reactor Mode

-------
IN3
         \VAN\N\\\\\\\\\\
         \\N\\\\\\\\\\\\\\
                              Figure 3.   Detail of Auxiliary Heating (Cooling)  System

-------
        ADVANCED COMBUSTION CONCEPTS
                     FOR
           LOW BTU GAS COMBUSTION
                     By:

                B. A. Folsom
               C. W. Courtney
                T. L. Corley
                 W. D. Clark
Energy and Environmental Research Corporation
          Santa Ana, California 92705
                       163

-------
                                  ABSTRACT
     The low Btu gas combined cycle power plant is an alternative to the
direct coal-fired steam cycle with the potential for low sulfur emissions and
high overall efficiency.  However, low Btu gas contains ammonia which could
lead to high nitrogen oxide emissions.
     This paper discusses the development of low NO  combustor concepts for
                                                   X
this application.  The thermodynamic performance of several alternative low
Btu gas-fired combined cycles is investigated and the combustor operating
conditions necessary to optimize thermodynamics performance are identified.
     A kinetic mechanism describing fuel nitrogen conversion to NO is used
to analyze idealized combustors with these operating conditions and several
potential low NO  combustor concepts are investigated.
                X,
     Synthetic low Btu gases with varying compositions were fired in three
atmospheric pressure flame reactors:  diffusion flame reactor, premixed flat
flame reactor, and premixed catalytic reactor.  NO  emissions were found to
                                                  2t
be sensitive to the concentrations of ammonia and hydrocarbon gases in the
low Btu gas.  Lowest NO  emisssions were produced by the diffusion flame
                       Ji
reactor operating fuel-lean and the catalytic reactor operating fuel-rich.
                                       164

-------
                             ACKNOWLEDGEMENTS

      This paper is based upon work conducted under Contract No. 68-02-2196
with the-Environmental Protection Agency.  The authors wish to express their
appreciation to Mr. G. B. Martin of the Environmental Protection Agency and
Messrs. J. Johnsen and J. Miltko of Energy and Environmental Research
Corporation for their assistance in various portions of the work.
                                      165

-------
                                  SECTION 1
                                INTRODUCTION

     This paper discusses the development of low nitric oxide (NO) emission
combustors for low Btu gas (LEG)-fired combined cycle power systems.  The
rising demand for electrical power coupled with the limited availability of
petroleum makes the construction of many alternate fuel-based power plants
by 1985-1990 very desirable.  Since these plants will commence operation in
the future when environmental restrictions will probably be much more
restrictive than at present, the environmental design goals should be to
minimize emissions rather than to meet current New Source Performance Standards.
     Several advanced coal-based power cycles are currently being developed.,for
this application  (1,2,3,4).  The integrated gasifier LBG-fired combined gas
turbine-steam turbine cycle is one of these alternatives.  The primary advan-
tage of this cycle is the potential for low emission of sulfur products without
the economic and efficiency penalties associated with stack gas sulfur
removal (5,6).  Since the coal gasification process operates fuel-rich, the
sulfur in the coal is converted mainly to hydrogen sulfide (H_S) in the low
Btu offgas which can be removed (potentially) more easily than sulfur dioxide
(S0_).   Energy losses associated with the gasification and cleanup process can
amount to as much as 30 percent of the coal's heating value.  However, if the
clean LBG is fired in a gas turbine combustor as part of a combined cycle power
plant the gains in overall power plant performance, when compared to conventional
steam cycles, may more than offset the losses in the gasification and cleanup
processes.
     The minimization of NO emissions from LBG-fired combined cycles is an
important and at present unsolved problem.  In the gasification process, some
of the nitrogen in the coal is converted into ammonia (NH~) in the LBG.  Under
typical gas turbine combustor operating conditions a large fraction of the NH~
may be converted to NO resulting in significant emission.
                                       166

-------
     The concentration of NKL in the LEG depends upon gasifier design and
operating parameters and may be as high as 0.38 percent (6).   The gas cleanup
system,  which functions primarily to remove sulfur products and particulates
from the LEG, may remove a portion of NH.,, but the remainder will enter the
gas turbine combustor and may be oxidized to nitrogen oxides (NO ).   A concen-
                                                                Jv
tration of 0.38 percent NH, in the LEG would produce approximately 1370 ng/J
              6
(3.2 Ib N02/10  Btu) if all NH- is converted to N02.  At present there are no
New Source Performance Standards (NSPS) for LEG fired combined cycle power
systems.  However, it is reasonable to expect that when NSPS are promulgated
they will be at least as stringent as current NSPS for other power systems
such as gas turbines or gas fired steam generators.  The current NSPS for
gaseous fossil fuel-fired steam generators with greater than 73.3 MW (250 x 10
Btu/hr)  heat input is 86 ng/J (0.2 Ib N02/10  Btu) (7).  For a combined cycle
firing an LEG with 0.38 percent NH- to meet this emission level, the overall
conversion of NH  to NO would have to be less than 6.3 percent.  It is well-
known that the conversion of fuel nitrogen compounds (such as NH.) to NO  is
sensitive to combustor design and operating parameters.  A recent analytical
study of combustion modification techniques applied to LBG-fired combined
cycle combustors has demonstrated the potential for significant reductions in
N0x emissions (8).
     The development of low NO emission combustor concepts for LBG-fIred
combined cycle systems is the objective of a current Environmental Protection
Agency program (Contract No. 68-02-2196, G. Blair Martin, Project Officer).
This paper presents some preliminary analytical and experimental results from
this program.  The following section discusses combustor design requirements
based on maximizing the thermodynamic performance of LBG-fired combined cycles.
The subsequent section outlines a kinetic mechanism for NO formation from NH_
and presents several low NO  combustor concepts based upon kinetic analysis.
The next section presents the results of combustion experiments burning
synthetic low Btu gases with varying compositions in several types of com-
bustors.  Discussion and conclusions follow.
                                       167

-------
                                  SECTION 2
                     DEFINITION OF COMBUSTOR REQUIREMENTS

     Figure 1 is a simplified schematic diagram of an LBG-fired gas
turbine-steam turbine combined cycle power plant with integrated gasifier.
Only the major heat, work and mass flow paths have been shown for clarity.
The system consists of a gas turbine topping cycle with exhaust heat trans-
ferred to a steam turbine bottoming cycle through a waste heat recovery
boiler.  Low Btu gas is produced in a gasifier supplied with coal, compressed
air from the gas turbine air compressor* and steam from the bottoming cycle.
This "integration" with other cycle components significantly reduces the
energy losses attributable to the gasification process.
     "Raw" LB6 exits the gasifier at temperatures as high as 1370 K depend-
ing upon gasifier design and operation containing three pollutant precursors:
     •    Sulfur compounds which can be oxidized to SO  in the combustor,
                                                      a
     9    Farticulates including carbon, tars and ash which may damage
          turbine blades and may be emitted to the atmosphere, and
     •    Ammonia which may be oxidized to NO  in the combustion process.
                                             4m
     Raw LBG is processed through a gas cleanup system to reduce the concen-
trations of these materials prior to combustion.  Several gas cleanup sys-
tems have been developed or are under development for this application (9) .
Hot gas cleanup systems process the LBG without significantly reducing its
temperature and remove  a majority of sulfur products and particulates.  Cold
gas cleanup systems  require cooling the LBG to near ambient temperature (400 K)
and remove a majority of the sulfur products and particulates.  They may also
remove a substantial portion of the ammonia.  The sensible heat removed from
the LBG may be used  for a variety of purposes, but it effectively bypasses
the gas turbine topping cycle and contributes to reduced efficiency.  The
thermodynamic trade-offs in hot and cold gas cleanup processes have recently
                                       168

-------
been investigated (8,10).  Depending upon gasifier design and operating
parameters, the sensible heat in LEG offgas can amount to as much as 20 per-
cent of the total heat release and the losses due to intercooling the LEG
with a cold gas cleanup system can be substantial.
     From a combustor design point of view, the choice of a hot gas cleanup
system as opposed to a cold system could well have two important effects:
     •    Higher adiabatic flame temperature
     •    Higher NH_ concentration
The sensible heat in the LEG contributes directly to the adiabatic flame
temperature.  For typical LEG compositions, cooling the gas by 400 K will
decrease the adiabatic flame temperature at stoichiometric conditions by about
200 K, thus reducing the potential for thermal NO  formation.  If a low
                                                 A
temperature cleanup system is employed, a substantial portion of the NH_ may
be removed.  Residual NH_ concentrations of 100 to 400 ppm have been predicted
for full-scale systems (3,4).  These concentrations correspond to approximately
34 to 145 ng/J (0.08 to 0.34 Ib N02/106 Btu) for full conversion.  While the
lower prediction meets current NSPS (neglecting the contribution from thermal
NO ), further NO  control may be necessary to meet future standards.
  x             x
     The term low Btu gas does not refer to a specific gas composition, but
rather to a family of fuels, produced by reforming coal with air and steam.  The
                                                        3                  3
heating value of LEG may range from 16,000 to 41,000 J/m  (80 to 200 Btu/ft )
and the primary combustible specie are CO and H-.  The ratio of CO to H  con-
centrations ranges from 0.5 to 2.  Hydrocarbon fuel gases (normally CH.) may
also be present in qua».-ities up to 10 percent, and nitrogen is the primary
diluent comprising 35 to 55 percent of the fuel gas.  Other diluents include
CO., and HO.  Trace amounts of H_S, COS and NH  may also be present, since
  £      £.                      f,             j
these specie are not entirely removed by the gas cleanup system.  The relation-
ship between LEG composition and combustor performance is currently under
investigation as part of this program and Section 4 discusses the performance
of combustor firing LBGs with a wide range of compositions.
     Combustor operating conditions are dictated by thennodynamic and materials
considerations.  High turbine inlet temperature (TIT) is desirable for improved
thermodynamic performance.  Current materials and blade cooling technology
                                        169

-------
limit TIT for stationary gas turbines to 1273 K to 1473 K.  The combustor
pressure producing an optimum balance between cycle efficiency and specific
work output is usually about 12 atmospheres.  The corresponding compressed
air temperature for high efficiency  compressors is about 638 K.  Improvements
in high temperature materials and  turbine blade cooling are expected to
increase allowable TIT in  the future.   One  proposed LBG-fired combined cycle
was designed around a 1644 K TIT  (4).
     If the exhaust heat from an optimized  gas turbine cycle precisely matched
the heat requirements of an optimized steam turbine cycle  there would be no
need for cycle variations.  However, the gas turbine  exhaust temperature is
usually too low for an optimized  steam  turbine cycle.  The overall efficiency
of the combined cycle can  be  Improved by modifying the cycle arrangement to
reduce the mismatch between the energy  content of the gas  turbine exhaust and
energy requirements of  the steam turbine cycle.  This can  be accomplished by
incorporating any of  the cycle  modifications listed below.
     •    Direct  fired  turbine  exhaust  heater
     •    High  excess air  supercharged  boiler gas turbine  combustor
     •    Reheat  gas  turbine  cycle
     The thermodynamic performance of LBG-fired  combined cycles incorporating
these cycle modifications  was analyzed  as part of this program and the analy-
tical details are discussed in  references 8,  10, 11,  12 and  13.  The results
are summarized in Table 1  which lists the overall efficiency and combustor
operating conditions for these cycles with  two TITs:  1366 K typical of state-
of-the-art stationary gas  turbines and  1700 R which may be achievable in the
future utilizing advanced  turbine blade cooling  techniques.  These results
were calculated subject to the component efficiencies and  other assumptions
listed in Reference 10.   Since these assumptions may  not necessarily be ful-
filled in future generation combined cycle  systems, the efficiencies and
operating conditions listed should be considered as approximate.
     The combustors listed in Table  1 can be grouped  into  two general types:
primary combustor where the oxidant  is  nonvitiated compressed air, and
secondary combustors where the oxidant  is the partially vitiated combustion
products from the primary  combustor.  The primary combustors will operate at
                                       170

-------
several atmospheres pressure with warm air (from the heat of compression)
and will generate a hot gas stream for the turbine.  They will operate either
adiabatically or with only a small amount of heat (less than 20 percent) trans-
ferred to the bottoming cycle.  The overall stoichiometry will be very fuel-
lean, greater than 250 percent theoretical air.  The NO  control goals for
                                                       X
primary combustors are to minimize conversion of NH_ in the LEG to NO  (fuel
                                                   •J                 Vt
NO) and minimize formation of NO  from N  (thermal NO ).
  *                              x       L            x
     The secondary combustors will operate at pressures less than the primary
combustors and may produce either a hot gas stream for a low pressure turbine
or a warm gas stream for a waste heat recovery boiler.  The overall stoichi-
ometry of these combustors will be fuel-lean but richer than the primary com-
bustors.  The amount of heat released in the secondary combustors will be
small compared to the primary combustors.  The NO  control goals for secondary
                                                 ji
combustors are the same as for primary combustors except that additional NO
control might be achieved by reducing some of the NO  formed in the primary
                                                    X
combustor to N_.
                                       171

-------
                                 SECTION 3
                NITRIC OXIDE FORMATION IN LBG-FIRED COMBUSTORS

     Tyson et al. (14) have assembled a kinetic reaction set based upon the
work of many investigators which describes the formation of NO and N£ in com-
bustible mixtures of CO, H_ and CH,.  Those reactions involving nitrogenous
species are presented in Table 2.  Ammonia breakdown occurs via hydrogen
abstraction allowing nitric oxide formation by the oxidation of any of the
nitrogenous fragments.  Under fuel-lean conditions, Reaction 4 predominates
and the major portion of the NH« is converted to NO.  Under fuel-rich condi-
tions NO might  still be formed via Reaction 4 if all the oxygen has not been
consumed, or by Reaction 6.  Hydrogen cyanide has been observed by several
investigators,  and for high temperature flames Morley (15) has shown that
regardless of their nature, nitrogen compounds are quantatively converted to
HCN in the reaction zone of premixed rich (less than 80% theoretical air)
hydrocarbon flames.  Hydrogen cyanide formation from NO could be possible by
reactions such  as 7 and 8.  Hydrogen cyanide can also be produced from mole-
cular .nitrogen  via reaction with hydrocarbon radicals (16).  The subsequent
oxidation of HCN allows the interchange of XN specie since NH is one of the
specie formed as HCN is oxidized via the intermediary NCO.  Once NH is pre-
sent then the reverse of Reactions 1, 2 and 3 allows the formation of ammonia.
Nitrogen formation can occur via the reverse of Reaction 5, or by reaction
with NO and other nitrogenous species.  Although Reaction 13 is included in
Table 3, there  is some question as to whether a reaction of this type can
take place.  The reactions shown in Table 3 illustrate how ammonia in LEG
could be converted to either N_, NO, HCN or back to NH, in rich combustion
products.  Since N. is thermodynamically favored at certain reactant stoichio-
metries and temperatures, the objective of the combustor designer is to
provide the most favorable temperature and stoichiometry histories to maxi-
mize N_ production and thus minimize total bound nitrogen species (EXN).
                                       172

-------
     Decoupling the .practical considerations of the fuel/air contacting pro-
cess, it would appear that several combustor configurations might minimize
NO  emissions from LEG containing NH .   Four such concepts were analyzed by
  X                                 j
applying the kinetic mechanism described above to idealized combustor models
consisting of various combinations of well-stirred and plug flow modules.
     The simplest case considered was the rich/lean series staged combustor
shown in Figure 2.  The ignition reactor provides the feedback of heat and
radical species for flame stabilization and the plug flow reactor, which
operates at an optimum stoichiometry, allows sufficient time for the reactions
listed in Table 2 to proceed towards equilibrium.  Secondary air is added
uniformly over a short time period to burn out the remaining combustibles and
to dilute the combustion products to the desired TIT.
     This combustor configuration was analyzed subject to the fuel composition
and combustor operating conditions listed in Reference 3.  This is an adiabatic
gas generator operating at 12.0 atmospheres and 1589 K TIT fired with a LEG
containing 600 ppm NH3.  Figure 3 shows the adiabatic equilibrium EXN concentra-
tion as a function of reactant stoichiometry.  ZXN is strongly dependent upon
reactant stoichiometry, and the minimum occurs at 65 percent theoretical air.
On the rich side of this minimum the XN species are primarily NH_ and HCN,
whereas on the lean side the XN species are almost totally NO.
     Figure 4 shows the results of kinetic calculations for a 1.0 msec ignition
zone followed by a plug flow section operating at 65 percent theoretical air.
Although £(XN) decays toward the equilibrium value of 2.9 ppm, the rate of
decay decreases with increasing residence time.  For a finite primary zone
residence time,. both the decay rate and the equilibrium value are important.
Consequently, the stoichiometry which produces minimum Z(XN) at equilibrium
may not necessarily be the optimum stoichiometry for a combustor with limited
residence time.  However calculations similar to those discussed above were
conducted for a range of stoichiometries and the minimum ZXN for all residence
times was found to occur with 65% theoretical air.
     A secondary stage following a 65 percent theoretical air 50 msec primary
stage was analyzed for two air injection patterns.
     •    Uniform air addition over 4.0 msec
     •    Uniform air addition over 60.0 msec
                                      173

-------
The results shown, in Figure 5 indicate that rapid air addition gives lower
emissions.  Slow air addition allows the mixture to spend considerable time
near stoichiometric conditions where temperatures are high.  Thus, a large
amount of thermal NO is formed.  The decay after the peak is dilution.  Full
conversion of the initial  18 ppm of NH  would result in 8 ppm when fully
diluted.  The rapid air addition case produces  14 ppm which can be attributed
to  full conversion of  the  NH. emerging the secondary stage plus 10 ppm of
thermal NO  .  The total NO emissions are equivalent to 14 ng/J (0.033 Ib
       f   X                X
N0_/10 Btu).  Thus this combustor concept has  the potential for very low NO
  *~                                                                        X
emissions.
     Three additional  low  NO  combustor  concepts are shown in Figure 6.  These
                            2C
have been analyzed similar to the  rich/lean  series staged concept discussed
above  and the results  demonstrate  the potential for low NO  emissions and
                                                          X
certain improvements over  the rich/lean  series  staged concept.
      If the reduction  of fuel XN species to N.  is kinetically limited, the
rate of reduction can  be accelerated by  increasing the reactant temperature
 (17).  Concept A in Figure 6 is one way  to achieve higher reactant tempera-
tures  in  an overall adiabatic system.  Heat  is  removed from the hot rich
combustion products and transferred  to the cooler reactants prior to ignition.
The results  of kinetic calculations applied  to  this concept show that the
optimum stoichiometry  for  maximum XN decay shifts to richer conditions as the
amount of feedback heat transfer increases.  For comparable overall residence
times, a combustor with 10 percent of the heat  release fed back to the
reactants produces NO  emissions 16 percent lower than the simple rich/lean
                     3E
series staged concept.
     Kinetic calculations  indicate that the decrease in XN decay rate as
residence time increases in a rich plug flow reactor is partially due to
depletion of the radical pool from the high superequilibrium concentrations
present immediately after  ignition (14).  By staging the air addition as in
Figure 6-B, the radical pool is periodically replenished leading to a more
rapid overall XN decay rate.  The stoichiometry, air addition rate, and
residence times can be optimized for each rich  stage to achieve a minimum
final XN concentration.
                                      174

-------
    Ammonia  conversion to N. might well be maximized by contacting certain
species  that  would not normally occur in a sequential process by dividing the
total  combustor into two parallel streams with different stoichiometries, and
then mixing the products of these two combustors prior to secondary burnout.
This concept  referred to as parallel staging is shown in Figure 6-C.
    Although in the limit, kinetic considerations dictate the ultimate con-
version  of fuel nitrogen to N  or NO, conversions in practical systems are
dictated by the realities of the fuel/air mixing process.  The reactants can
be premixed to ensure that reaction takes place at a single stoichiometry or
the fuel and  air may be supplied to the combustor separately, thus allowing
reaction to take place over a range of stoichiometries which will depend upon
the intensity of the mixing process.  The following section describes the
experimental  results obtained from firing LEG in several types of combustion
reactors.
                                       175

-------
                                   SECTION 4
                      LOW BTU GAS COMBUSTION EXPERIMENTS

      The idealized fuel/air contacting assumed in the preceeding kinetic
 analysis cannot be achieved in practice.  The combustor designer must  take
 account of practical constraints such as combustor cost, maintainability,  and
 operability.   Pressure drops must be minimized since they  represent  energy
 losses, and combustor volumes (and therefore, residence times) must  be kept
 within bounds to reduce combustor cost and the potential for heat  loss.  The
 approach of this study is to use the results of kinetic analysis to  guide  the
 development of low NO  combustion concepts primarily through experiments.   This
                      JL
 section discusses the results of a series of flame reactor experiments where
 LBG containing various XN compounds was combusted over a wide  range  of fuel/air
 contacting conditions.  The objectives of these tests were to:
      •    Evaluate the relationship between L'BG composition and XN
           processing
      *    Investigate XN processing in LBG-fired reactors  as a function
           of  fuel/air contacting method.
 Three  simple  combustors were tested:
     •   Laminar and turbulent diffusion flame
     •   Premixed pseudo one-dimensional flame (flat flame)
     •   Premised catalytically supported reactor
 The nitrogen  species  included in these investigations are  NH3, HCN and NO.
Hydrogen cyanide 'and NO were included because they are representative of the
XN species which might exit a rich combustion stage, thus providing informa-
tion on the optimum second stage reactor.
                                       176

-------
EXPERIMENTAL SYSTEMS
     The experimental equipment used in this investigation includes:
     •    Reactant flow system
     •    Interchangeable flame reactors
     •    Analytical train
These components are shown schematically in Figure 7.
     The LEG was synthesized by blending together high purity gases from
cylinders.  All gases were high purity grade (99.97 percent or better) with
the exception of CO (99.0 percent).  NH^, NO and HCN were supplied as custom
grade mixtures in nitrogen (+ 2 percent accuracy) to facilitate metering.  The
oxidant was dry air.  All gases were metered with sapphire jewel orifices
operated in the critical flow (sonic) regime.  The pressures upstream of the
orifices were measured with high accuracy variable capacitance pressure trans-
                                              f*
ducers which were calibrated periodically against a laboratory reference.  The
pressures downstream of the orifices were maintained constant and the flow rate
through each orifice was calibrated by filling an evacuated tank.  The esti-
mated total inaccuracy in each gas flow rate was 0.5 percent.
     Variations in water vapor content of the LBG were achieved by metering
distilled water with a calibrated rotameter and prevaporizing before mixing
with the other reactant gases.  The mixture was maintained well above the
dewpoint temperature to prevent subsequent condensation.  Other design details
of the reactant flow system are discussed in Reference 18.
     The diffusion flame reactor utilized in these experiments is shown in
Figure 8.  Fuel gases entered through a stainless steel tube centered in a
sintered stainless steel disc.  Air passed through the sintered disc and
mixed with the fuel gases by laminar or turbulent diffusion depending upon
flow conditions.  A fully developed velocity profile was assured by a tube
length greater than 100 diameters.  The surrounding tube was enclosed by a
water jacket to control heat losses.  Combustion products exiting the reactor
were mixed by several rows of stainless steel water-cooled tubes normal to
the flow.  Samples were withdrawn  for analysis downstream of this mixing
section through a water-cooled stainless steel probe.  All water cooled
                                        177

-------
surfaces were maintained at 343 to 363 K.
     The premised pseudo one-dimensional flame reactor is shown in Figure 9.
In this "flat flame" reactor the premixed reactants passed through a water
cooled stainless steel sintered disc.  Once the reactants were ignited a thin
planar flame stabilized approximately 1.0 mm above the disc.  The flat flame
reactor was enclosed by a quartz tube and combustion products were sampled
with stationary probes.  The flat flame reactor was moved axially to obtain
combustion product samples at any desired distance downstream from the flame
front.
     The premixed catalytically supported reactor is shown in Figure 10.  The
catalyst used in these tests was a platinum coated graded cell monolith
supplied by the Acurex Corporation.  The desirable features of the graded cell
catalyst have been discussed previously (18).  The maximum recommended tempera-
ture for this catalyst is. 1588 K (2400 F).  Type K thermocouples were cemented
in two of the cells in the downstream segment to monitor maximum monolith
temperature.  To minimize heat losses and approximate adiabatic conditions, the
catalyst was mounted in a refractory tube and electrically backheated.  The
reactants passed through a sintered stainless steel disc immediately upstream
of the catalyst and combustion products were sampled with a stainless steel
water cooled probe.  Additional details of this reactor's construction and
operation are documented in Reference 13.
     The same sampling and analysis systems were used for all flame reactor
studies.   Combustion products were analyzed for 09, CO, C09, NO, NO , NH  and
                                                 £•        &        Jt    .3
HCN.  The sample train components were constructed entirely of stainless steel,
glass and Teflon.  Ammonia and HCN were trapped by bubbling known volumes of
combustion products through three water baths in series.  The absorbed NH-
and HCN concentrations were then measured by ion specific electrodes.  The
other product species were monitored continuously.  The sample was dried to
a dewpoint of 273 K in a cyclone water trap located close to the sample
probe to minimize residence time between probe and trap.  Nitric oxide and
NO  were measured with a Thermal Electron Corporation Model 10A chemilumines-
  Jt
cent analyzer operated as discussed in Reference 19.  NO  concentrations
                                                        JL
were measured after the sample had passed through a stainless steel converter
operated at 1073 K.  This converter could not be operated under oxygen-
deficient conditions because of the well-known reduction of NO.
                                        178

-------
EFFECT  OF LEG COMPOSITION ON FUEL NITROGEN CONVERSION
     The experiments involving the effect of LEG composition on the forma-
tion of fuel NO  were carried out in the diffusion flame and flat flame
               3t
reactors.   The experimental procedure was to fire the reactors with and with-
out an  XN dopant and to calculate the percentage conversion of the dopant to
IN as follows:

                                   (XN measured\            / XN measured  \
                                   (with dopant I     -      I without dopant)
                                    /    j\/            \   /•    j  \  /
                                    (ppm dry) /	\   (ppm dry)  /
              nopant co AIM  i  =    	"	Trrr—'—-	:	—	—rrr	*	'
                K ,„.        I             /XN  calculated for full\
                           /             I   conversion of dopant  j
                                         \      (ppm dry)        /
For experiments at fuel-lean stoichiometries, only NO and NO  were measured.
                                                            Ji
For experiments at fuel-rich stoichiometries, NO, NH_ and HCN were measured
and EXN was calculated as the sum.  For rich experiments the XN measured with-
out the dopant was negligible and for lean experiments, thermal NO  was
                                                                  Ji
typically 50 ppm while full conversion of the dopants was usually a factor
of 20 higher.
     Figures 11, 12 and 13 show the results for a laminar diffusion flame.  All
data refer to a constant fuel volumetric flow rate, 366 K fuel and air preheat
and 150 percent theoretical air.  The presence of hydrocarbon fuel in the LEG
was found to have the most significant effect upon the conversion of NH»  to
fuel NO   (see Figure  11).  Without methane  (CH,), percentage  conversions  were
       X.                                      "
typically 5 percent of the dopant.  This rapidly increased  to greater than 20
percent when the fuel gas contained 5 percent CH..  Above  this concentra-
tion further increases in CH, content  did not have  a  significant  effect
upon fuel fuel nitrogen conversion.  The influence  of  ethylene (C-H.) as  the
hydrocarbon fuel is similar to that of  CH,;  however,wmaximum  conversions
are lower typically of the order of 20 percent for  an LEG  containing 10 per-
cent C_H,.   It  can  also  be  seen in Figure  11 that  the influence of CO/H2
ratio was minimal compared to that of  the presence  of  a hydrocarbon.  Similar
experiments varying the diluent composition  to include N«,  CO^ and H_0  showed
diluent composition effects to be on the same order as CO/H_  ratio effects  (12),
     The partial products of combustion of  the first  stage in a staged  heat
                                        179

-------
release combustor can contain several nitrogenous species (NO, NH_ and
HCN)  as well as CO, H  and a hydrocarbon fuel.   The results in Figure 12
show that the conversion of a given nitrogenous  specie in an LBG fuel is not
only dependent upon the hydrocarbon fuel content, but also on the nature of
the fuel nitrogen specie.  High conversions of small quantities of HCN and
NO were measured for small dopant concentrations, and these conversions were
almost independent of hydrocarbon content.  However, at the same dopant level
the conversion of NR- is strongly affected by hydrocarbon content.
     Figure 13 shows the effects of hydrocarbon fuel concentration and dopant
leval on NH_ and NO conversion.  With the exception of NO as the dopant and
zero percent CH,, percentage conversion of dopants decreased as the dopant
levels increased.  However, this decrease in conversion was insufficient to
counter the increase in XN available for conversion and total NO  emissions
                                                                x
increased with dopant level.  At NO dopant levels greater than 500 ppm there
was a significant influence of hydrocarbon content upon fuel nitrogen conver-
sion (NO to NO ).  This suggests that if the effluents of a rich primary con-
tained a hydrocarbon fuel gas and NO, as the only bound nitrogen species,
significant reductions in emissions might be achieved in a lean secondary burn-
out stage.
     Figure 14 shows the results of similar experiments varying CH, concen-
tration in LBG-.fired flat flames.  The conversion of NH3 to NO is essentially
the same with and without CH, except at very rich conditions where some
is formed in the case with CH,.
                                                                        HCN
     The effect of a hydrocarbon fuel on the conversion of NH_ to NO can
be explained by consideration of the reactions shown in Table 2.  In a lami-
nar diffusion flame NH_ breakdown occurs on the fuel-rich side of the heat
release zone, and N_ formation can occur by reaction of the nitrogenous
fragments.  Any NO that is produced can react with these nitrogen fragments
producing N_.  When a hydrocarbon fuel is present hydrogen cyanide can be
produced from NO via reactions such as 7 and 8, thus effectively removing
nitrogen specie from the zone in which they can be more easily converted to
N_.  Once HCN is produced, then the XN specie exchange reactions allow the
regeneration of NO as 0_ diffuses into the reaction zone.
                                      180

-------
     In the flat flame the reactants are premixed and NH  breakdown occurs
in the presence of 0 .  Thus NH_ and the. resulting fragments compete with
the fuel species for 0-.  Reduction reactions are not favored.  At very rich
stoichiometries much less 0_ is 'available and the effects of hydrocarbon
content become more pronounced.
     It is well-known that the fuel/air contacting conditions in a diffusion
flame are dependent upon fuel and oxidant flow rates.  For a simple coaxial
system such as the one tested here, a laminar flame is produced at low flow
rates.  As the flow rate is increased there is a transition to a turbulent
flame with noticeable audible noise and a brush-liek appearance.  As the
flow rate is increased further, the flame lifts off the port allowing partial
premixing of fuel and air beneath the reaction zone.  Figure 15 shows the
conversion of NH, to NO  in LBG-fired diffusion flames operated in all three
fuel/air contacting conditions.
     Under laminar conditions, the hydrocarbon content is an important variable
as discussed previously.  Under attached turbulent conditions the conversion
increases somewhat but the strong effect of CH, concentration is similar to
the laminar case.  However, in the lifted condition the effect of CH,
disappears and conversion is nearly complete.  Thus from a fuel nitrogen per-
spective the attached laminar and turbulent flame produces fuel nitrogen
conversion similar to a premixed system.

EFFECT OF COMBUSTOR CHARACTERISTICS ON FUEL NO  FORMATION
     The effect of combustor characteristics on fuel NO  formation was
                                                       2C
measured by firing an NH_ doped LBG in the three reactors described previously
over a range of reactant stoichiometries.  The LBG composition synthesized for
these tests is listed below:
CO
H2
CH4
N2
NH3
20 percent
20 percent
5 percent
55 percent
471 ppm
                                        33            3
The heating value of this gas is 37 x  10  J/m   (182 Btu/ft ).
                                        181

-------
Measured concentrations of NO, NH. and HCN are presented in Figure 16 for
the diffusion flame, the premixed catalytic reactor and the premixed flat
flame reactor operating fuel-rich.  For the diffusion flame and catalytic
reactors, combustion products were sampled downstream of the mixing device
described previously.  For the flat flame reactor combustion products were
sampled on  the axis approximately 45 msec from the primary reaction zone.
Axial probing traverses showed no significant variation in these concentra-
tions over  a product residence time range of 2 to 50 msec.
     The results in Figure 16 are similar for all reactors:  as the percentage
of theoretical air in the reactants decreased, the concentration of NO in the
reactants decreased, the concentration of NO in the products decreased while
the NH_ and HCN concentrations increased.  The diffusion flame utilized for
these tests incorporated a water-cooled housing.  As the percent theoretical
air was reduced from overall lean conditions the flame expanded toward the
cooled walls.   Under rich conditions the base of the diffusion flame operated
in the normal fashion with a near stoichiometrie flame front, but the upper
portion was quenched along the cold walls.  This accounts for the high NH.
throughput under rich conditions.
     For the premixed catalytic reactor the HCN concentration was virtually
zero down to 70 percent theoretical air.  As the percent theoretical air
was further decreased, HCN increased rapidly until at 45 percent theoretical
air the measured NH. and HCN concentrations were the same.  In the premixed
flat flame NH. concentrations were very low, and the HCN concentration
increased as the percentage theoretical air decreased.  The NO concentrations
in the premixed flat flame were generally higher than those in either of the
other two reactors.
                                       182

-------
                                  SECTION 5
                         DISCUSSION AND CONCLUSIONS

     The results presented in this paper represent the initial portion of a
study to develop combustor concepts to minimize NO  emissions from LEG fired
                                                  2£
combined cycle power systems.  Low Btu fuel gases may contain up to 3800 ppm
NIL and this could result in high NO  emissions under typical gas turbine
  J                                 X
combustor operating conditions.  However it is known that the conversion of
NIL. to NO can be restricted by modifying the combustion process in such a way
as to ensure that the NH  reacts initially under fuel-rich conditions.
Kinetic analysis of idealized combustors has shown the potential for very low
NO  emissions even with high NH_ concentration in the LEG.
  *                            J
     Before the development of bench-scale combustors can commence it is
necessary to ascertain the parameters controlling fuel NO  formation in the
LEG—fired systems.  LEG fuel gases will vary in composition depending upon the
gasifier design and operating parameters.  Thus, if the combustor concepts are
to be universally applicable then it is necessary to establish whether the
combustor must be tailored to the particular fuel gas being produced.  The
results indicate that by far the most significant effect of LEG composition
on fuel NO  formation is the presence of a hydrocarbon gas.
          2C
     Kinetic considerations suggest that optimum temperature/stoichiometry
histories exist to maximize N- production.  However, practical considerations
require that the fuel and oxidant be mixed either before or after injection
into the reaction chamber, thus providing the potential for different types
of combustors based upon the fuel/air contacting process in the primary zone.
A series of simple reactor experiments have been carried out with synthesized
LBG to determine the influence of combustion characteristics on the processing
of NH_ in LEG.  Figure 17 compares the EXN  (measured rich) and NO  (measured
     j                                                           X
lean) for the three reactors firing the baseline LEG doped with NH  and similar
                                       183

-------
data for the diffusion flame reactor firing an LEG with no hydrocarbons.  For
the baseline LEG containing CH,, the diffusion flame and catalytic reactors
produce comparable minimum conversions and the flat flame reactor produces
much higher conversion over the range of stoichiometries tested.  (It could
not be operated richer than 60 percent theoretical air due to flame
instability.)  It is difficult to compare the three reactor types since their
operational characteristics are different.  The maximum operating temperature
of the catalyst was 1588 K, and the undiluted reactants could not be burned
between 40 and 230 percent theoretical air.  Therefore, to obtain the results
shown  in Figures 16 and 17 it was necessary to dilute the reactants with
nitrogen to ensure that this maximum temperature was never exceeded.  In a
real system this would require recirculation of cooled combustion products or
heat loss from the primary section.  Fenlmore et al. (20) have carried out
similar .diffusion flame experiments and have shown that as the.flame becomes
turbulent there is a pronounced increase in fuel nitrogen conversion to NO .
                                                                          3t
In the experiments reported here, the transition from laminar to turbulent
flames did not produce comparable results.  This can be explained only if it
is accepted that at these turbulence levels the turbulent flame front pro-
vides a continuous reaction zone around the fuel stream.  If turbulent levels
were increased to such levels that the reaction zone were stretched and
extinguished then fuel oxidant mixing could take place without reaction and
the subsequent formation of fuel NO  would increase.
                                        184

-------
                               REFERENCES
 1.    Evaluation of Phase 2 Conceptual Designs and Implementation Assessment
      Resulting from the Energy Conversion Alternatives Study (EGAS),  pre-
      pared under Interagency Agreement E(49-18)-1751,  NASA Report No.
      TM X-73515.

 2.    Shaw, H., A. E. Cerkanowicz,  and S.  E. Tung.  Environmental Assessment
      of Advanced Energy Conversion Technologies - Interim Report, Vol.  1,
      State-of-the-Art.   Contract No.  68-02-2146, U. S. Environmental Pro-
      tection Agency, Cincinnati, Ohio, 1977.

 3.    Harris, L. P.,  and R. P.  Shah.   Energy Conversion Alternatives Study
      (EGAS), General Electric Phase II Final Report:   Vol.  II,  Advanced
      Energy Conversion Systems - Conceptual Designs:   Part 3,  Open Cycle
      Gas Turbines and Open Cycle MHD, General Electric Report  No. SRD-76-
      064-2, NASA Report No. NASA CR-134949, 1976.

 4.    Beecher, D. T., et al.  Energy Conversion Alternative Study (EGAS),
      Westinghouse Phase II Final Report:   Vol.  II - Combined Gas-Steam
      Turbine Plant Using Coal-Derived Fuel, Westinghouse Report No. 76-
      9E9-ECAS-R2v.2, NASA Report No.  NASA CR-134942,  1976.

 5.    Fluor Engineers and Constructors, Inc.  Gasification-Combined-Cycle
      Power Plants.  EPRI Journal,  July/August,  1978.   p. 43.

 6.    Robson, F. L.,  W.  A. Blecher, and A. J.  Giramonti.  Combined-Cycle
      Power Sytems.  EPA-600/2-76-149, U.  S. Environmental Protection
      Agency, Washington, DC, 1976.  p. 359

 7.    Environmental Protection Agency Title 40,  Chapter 1, Subcharter C,
      Part 60-Standards  of Performance for New Stationary Sources, Federal
      Register, Vol.  36, No. 247, December 23, 1971.

 8.    Tyson, T. J., M. P. Heap, C.  J.  Kau, B. A. Folsom, and N.  D. Brown.
      Low NOX Combustion Concepts for Advanced Power Generation Systems
      Firing Low Btu Gas.  EPA-600/2-77-235, U. S. Environmental Protection
      Agency, Washington, DC, 1977.

 9.    Dravo Corporation, Handbook of Gasifiers and Gas Treatment Systems,
      NTIS Report No. FE-1772-11, Feburary 1976.

10.    Folsom, B.A., T. L. Corley, M. H. Lobell, C. J. Kau, M. P. Heap,
      and T. J. Tyson.  Evaluation of Combustor Design Concepts for Advanced
      Energy Conversion Systems.  In:   Proceedings of the Second Stationary
      Source Combustion Symposium,  Vol. 5, Addendum, EPA-600/7-77-073,
      July 1977.

                                       185

-------
11.   Heap, M. P., T. J. Tyson, J. E. Cichanovicz, R.  Gershman, and C.  J.  Kau.
      Environmental Aspects of Low Btu Gas Combustion.  In:  Sixteenth Sym-
      posium on Combustion, The Combustion Institute,  1977.  p. 535.

12.   Folsom, B. A., C. W. Courtney, M.  F. Heap,  and G.  B. Martin.  The Effect
      of LBG Composition and Combustor Characteristics on Fuel NO  Formation.
      In:  Fourteenth Annual International Gas Turbine Conference, A.S.M.E.,
      Gas Turbine Division, March 1979.

13.   Folsom, B. A., C. W. Courtney, and M. P. Heap.  Environmental Aspects
      of Low Btu Gas-Fired Catalytic Combustion.   Proceedings of the Third
      Workshop on Catalytic Combustion,  October 3-4, 1978.

14.   Tyson, T.  J.,  et al.  Fundamental Combustion Research Applied to Pollu-
      tant Control.   First Annual Report, EPA Contract 68-02-2631.  In prepa-
      ration.

15.   Morley,  C.   The Formation and Destruction of Hydrogen Cyanide from
      Atmospheric and Fuel Nitrogen in Rich Atmospheric-Pressure Flames.
      Combustion and Flame, Vol. 27, 1976.  ppi 189-204.

16.   Myerson, A.  L.  The Reduction of Nitric Oxide in Simulated Combustion
      Effluents  by Hydrocarbon-Oxygen Mixtures.  Fifteenth Symposium (Inter-
      national)  on Combustion, The Combustion Institute, Pittsburgh,
      Pennsylvania,  1975.

17.   Sarofim, A.  F., J. H. Pohl, and B. R. Taylor.  Mechanisms and Kinetics
      of NO Formation Recent Developments.  Paper presented at 69th Annual
      Meeting  AIChE, Chicago, November 1976.

18.   Kesselring,  J. P., W. Y. Krill, and R. M. Kendall.  Design Criteria
      for Stationary Source Catalytic Combustors.  EPA-600/7-77-073c,
      U. S.  Environmental Protection Agency, Washington, DC, July 1977.
      p. 193.

19.   Folsom,  B.  A., and C. W. Courtney.  Chemiluminescent Measurement of
      Nitric Oxide in Combustion Products.  In Proceedings of the Third
      Stationary Source Combustion Symposium, March 1977.

20.   Fenimore,  C. P.  Effects of Diluents and Mixing on Nitric Oxide from
      Fuel Nitrogen  Species in Diffusion Flames.  Seventeenth Combustion
      Symposium,  published by the Combustion Institute, 1978.  p. 1065.
                                       186

-------
oo
               Coal  Input
                    Steam
     Pressurized
     Gaslfler And
    Cleanup  System
                Air
                Inlet
                           Compressor
LBG
              Adiabatlc
             Gas Generator
                                     Turbine
                     Steam Turbine
        Generator
                     Condenser
                       AWi
                       i
                   Cooling Water
Feed-
Water
Pump
                        -3-
Generator
                                                                            Waste  Heat
                                                                            Boiler
                    Figure 1.   Simplified Schematic of Basic LBG-Fired  Combined Cycle.

-------
       Air
       620K
CO
00
        LBG
        Fuel
        422K
        A1r
        620K
  Well
 Stirred
Ignition
                                                  Plug Flow
                                                                                  Secondary A1r
Plug Flow
 To  Turbine
-  1589K
 218% TA
                                       Rich Primary
                                                  Secondary Burnout
                                                        Stage
                                 Figure 2.  Limit Case Schematic Diagram of a
                                            Rich/Lean Series Staged Combustor.

-------
 1000.0 -
  100.0 -
to

•r-
CO



O)




t
Q.
   10.0 -
            65 100
                               300
400    500
             Stoichiometry (% Theoretical Air)
 Figure 3.   Adiabatic Equilibrium £XN for LEG.
                       189

-------
0
 80     100
Residence Time (MSEC)
                  Figure 4.  Kinetic Analysis of 65% Theoretical
                             Air Rich Primary.

-------
10
 0
I
                                  Secondary Air Added Over 60 MSEC
                      Secondary Air Added Over 4  MSEC
   •<	   100%  Conversion  of  XN  Entering  Secondary to NO  (Zero Thermal  NO )
                                                           A                 A
I
I
I
I
J_
I
I
  0   10      30      50       70      90     110     130     150.

                           Secondary Stage  Residence  Time,  MSEC
                                             170
                                             190
                                             210    230
                        Figure  5.   Second  Stage Air Addition.

-------
Fuel
Air
         Heat
         Addition
Rich Com-
bustion
  Zone
Heat
Removal
         To  Second
         Stage
         Burnout
                          A.  Heat Feedback
ieL /^ ^\
In taMMMM^^M
i *R- / ^
r_ \ 1

r


L u j

I
*,
*1


^^^







i .
I *"
I
i h

^^





j ,

i
1 ,
1 '3
i
M t


To Second
Staae

Bumout
                    B.  Distributed Air Addition
                                            Mixing Zone
                     C.   Parallel  Staging  .  f
                                                           To Second
                                                           Stage
                                                           Burnout
            Figure 6.  Idealized Low NO  Combustor Concepts.
                                       X
                                    192

-------
ID
CO
                                                     Exhaust
                                                     Stack
                                        High Purity Gas
                                           Cylinders
                                                                                               \/ent
                              Gas Bubbler
                              Train for Wet
                              Sample
8 Channel
Critical Orifice
Flow Metering
System
                                                                 In Situ
                                                                 Calibration
                                                                 System
                Water Trap
                                                              Sample
                                                              Probe
                                                      Vaporizer
                                                                                   Calibrated
                                                                                   Rotameter
Zero and Span Gases
     O  O
                                                        Reactant
                                                        Preheater
                                                        500°C
                                                        Max.
                                                                               Rotameters
NO/NO
Chemi luminescent
Analyzer

CO
NDIR
co2
NDIR

°2
Paramagnetic
                                                                                               Bypass and
                                                                                               Wet Sample
                                                                                               Flowrate
                                                                                               Measurement
                                     Figure 7.  Experimental Apparatus.

-------
Sample Probe:
Stainless
Baffle Plate
         Cooling Water
         Circulation
             Pump
                                                          Water Cooled
         Exhaust Stack

         Stainless Steel
         Water Cooled
         Mixing Section
        Stainless Steel
        Water Cooled
        Flame Tube

        Sintered
        Stainless Steel
        Disc lOOy
                                          30xidizer Plenum (Heated)
                                                  0.63 cm
                                                  0.55 cm
                0.250 in
                0.218 in
O.D.
I.D.
                  Oxidizer
                             Fuel
              Figure 8.   Diffusion Flame Reactor.


                               194

-------
Movable
Burner
Assembly

Stack
1 '

y
91 c
(36
Water C(
Stainle;
Probe


:m
in




1

^V^B
M^BB


^•a
^^m
1
                          4.8 cm
                          (1.9 in,
     OOOOQ
        \

                             OO
                            I

                      ^
                  Reactants u n
                  Inlet     H2°
                           Ih
Thermo
coupl
.1
                                 Sample
                                                        =3
                                       Ul K
                                       tl
                                       In  Out
                                  Flat  Flame
                                  Sintered
                                  Stainless
                                  Steel disc (20 y)
                                  Cooling Coil
                                  7.6 cm (3.0 in)
                                  Quartz Tube

                                  Stainless
                                  Steel Weldment

                                  Perforated
                                  Plate
                    Out
Figure 9.  Flat Flame Burner.
             195

-------
Exhaust
Stack
                                        	Water-Cooled
                                            S.S.  Sample Probe
                                                    Graded Cell
                                                    Catalyst
                                   Ceramic Wool
                                     Insulation
                                                        Electrical
                                                        Heating
                                                        Element
                                                        1500 w.
  WO y
  Sintered
  S.S.  Disc
Castable
Refractory Cylinder
                        0)

                        Q.
    Figure 10.   Catalytic Reactor and Housing.


                          196

-------
    30
    20
O
-P
c
O
•r*
to

0)


O
O
_  NH3 =  3059  ppm

   Diluent = 55% N,

   150% T.A.
             CH4 Hydrocarbon
                         0
O

O  CO/Hg =1.0

&  CO/H2 = 2.0
                             Hydrocarbon
                     O CO/H2 =  1.0
     0
       0
                                                10
                  Fuel  Hydrocarbon Content
   Figure  11.   Laminar Diffusion Flame Processing
               of NH
                          197

-------
              100 ppm NO
              100 ppm HCN
              100 ppm NH3
              CO/H2  = 1.0
              Diluent =  55%  N?
              150% Theoretical  Air
                  Fuel  CH4  Content (%)
                                                 10
Figure 12.  Diffusion Flame Processing of XN
            at Low Concentrations.
                     198

-------
10
                              1000
                                                                   Dopant NO
                                                                   O 0 % CH
                                                                -  D 5 %
                                                                   A 10% CH
                                                                   A 10%
                                                                 	I
2000
3000         0       1000
   Dopant in LBG (ppm)
2000
3000
                               Figure 13.  Laminar Diffusion Flame Process of NH~ as Function
                                           of Dopant and Hydrocarbon Concentrations.

-------
                               0% CH,
5% CH,
                                                       Full  Conversion
o
o
        X
        o
        Qi
        1
       o
100
 90

 80

 70

 60

 50

 40

 30

 20
                                                       I
                                             120      140
                                                       % Theoretical Air
O EXN (rich) or NO   -
                  (lean)
                                                                                        O
                                                                                        D
       NO
       NH3
       HCN
             120
                                                                                                 140
                            Figure 14.   Flat Flame:  Effects of Fuel Hydrocarbon Content
                                        on NH3 Conversion to XN.

-------
                         LBG  Diffusion Flame
                         (NH3)  =  500  ppm
                         150% Theroetical  Air
                         CO/H2  -1.0
                         (N2) = 55%
                                     Full Conversion
                                                                O Q% CH^
                                                                D 5% CH,
ro
O
                  100
    80
  X
o
o

I  60
10
S-

O
^.co 40
•z.
                   20
                    0
                                    I
                    II
                               I
                      Attached
                      Laminar
                       Flame
                  Attached Turbulent Flame
                                   1
                      Lifted Turbulent Flame
                     1
                    1
                      0
               2000
4000
6000    8000
10000    12000
14000
                          Figure  15.  Diffusion Flame:   Effects of Flow Characteristics
                                      on NH.  Conversion to XN at 150% Theoretical Air.

-------
            \   Diffusion Flame


               N471 ppm NH, 1n LBG
                N          J
   300
   20°
o
£
+j

0)
u

o
o


X
   100
T—I—I—i—r
 Flat Flame

 471 ppm NH- 1n LBG

 Thermal NOJ
—I—I—I—I—I—I—

Catalyst


            1n LBG


                 • ZXN

                 D NH3


                 A HCN

                 O NO


            Diluted with N,

            to 1477 K     *

          \ Equilibrium

           • AcMabatic Flame

           •Temperature
                                       I
      0  20  40  60  80  100  120   0  20  40  60
                                        Theoretical Air (%)



            Figure  16.  Comparison of XN Emissions from Three Reactors.

-------
                  Low Btu Gas
                  CO/H2 =1.0
                       55%
                  NH
         2 = 471 ppm
                   O  Diffusion Flame  (5% CH4)
                   •  Diffusion Flame  (0% CH4)
                   D  Catalyst (5% CH4)
                   A  Flat Flame (5% CH4)
          100
ro
o
co
        s-
        o
        o
        4J
        c
        o
        •I—
        I/)
        0)
        c
        o
        o
50
                                                   1
                                                  i
                             1
                      20
                     40
60
80       100       120
  % Theoretical  Air
140
                                                  Diffusion (5%)  ~
                                                                ""^V J
160
180
200
                                Figure 17.  NH« Conversion to ZXN for Three Reactors,

-------
                      TABLE  I.  OVERALL  EFFICIENCY AND COMBUSTOR OPERATING CONDITIONS
                                FROM THERMODYNAMIC ANALYSIS
CYCLE/TIT (K)
Basic/ 1366
Basic/ 1700
Duct Heater /I 366

Duct Heater/1700

Supercharged n „,,.
Boiler /1JD
Supercharged/1700
Boiler
Reheat/1366

OVERALL
EFFICIENCY
39.1
46.3
40.1

46.8

41.0
47.3
44.6

COMBUSTOR OPERATING CONDITIONS
TYPE*
AGG
AGG
AGG.
ADH
AGG
ADH
SB
SB
AGG
ARC
PRESSURE
(ATM)
16.0
30.0
16.0
1.1
30.0
1.1
16.0
30.0
20.0
6.9
OUTLET
TEMP,
(K)
1366
1700
1366
914
1700
914
1366
1700
1366
1366
OXIDANT AIR
STOICH.
(% TA)
NV**
NV
NV
377
NV
286
NV '
NV
NV
425
TEMP.
(K)
697
844
697
762
844
840
697
844
536
1078
STOICH.
(% TA)
377
286
377
314
286
268
309
265
425
283
HEAT***
0
0
79.3
20.7
91.0
9.0
16.3
6.9
62.3
37.7
ro
2
          *  AGG - Adiabatic Gas Generator;  ADH - Adiabatic Duct Heater;  SB - Supercharged Boiler;
                   ARC = Adiabatic Reheat Combustor
         **  NV = Nonvitiated Air
        ***  For duct heater and reheat combustor, Heat % is % of total heat liberated in combustor.
             For supercharged boiler it is % of total heat transferred to bottom cycle.

-------
          TABLE II.  MAJOR FEATURES OF KINETIC SCHEME DESCRIBING
                     THE FATE OF AMMONIA DURING COMBUSTION
Ammonia Breakdown

   NH3 + H -»• NH2 + H2                                        1

   NH_ + H -> NH +  H                                          2

   NH + H + N + H                                            3
Nitric Oxide Formation

   N + 0  -*• NO + 0                                           4

   N  + 0 -»• NO + N                                           5

   N + OH -»• NO + H                                           6
Hydrogen Cyanide Formation

             via CH NO
   CH, + NO 	, no «mn '   HCN + .  .  .                       7
     3       and CH NOH

   CH + NO ^ HCN +  0                                         8

   CH -t- N  + HCN +  N                                         9


XN Specie Exchange

   HCN + OH -»• CN +  H20                                       10

   CN + OH •*• NCO +  H                                         11

   NCO + H -*- NH + CO                                         12

   NH + H2-*NH2-fH                                         2R

   NH2 + H2 •*• NH3 + H                                       1 R


Nitrogen Formation

   N + NO + N+0                                           5R

   NH  + NO -> N +  H0                                       13
                                      205

-------
          CATALYTIC COMBUSTION SYSTEM DEVELOPMENT
             FOR STATIONARY SOURCE APPLICATIONS
                            By:
J. P. Kesselring, W. V. Krill, E. K. Chu, and R. M. Kendall
                    Acurex Corporation
             Mountain View, California 94042
                             207

-------
                                   ABSTRACT

     An experimental program has been conducted for the Environmental Pro-
tection Agency to develop design criteria for catalytic combustors as applied
to stationary systems.  The program included catalyst screening tests from
which the graded cell concept was developed.  The graded cell catalyst exhibits
greatly enhanced combustion characteristics in terms of Increased maximum
throughput.  Advanced testing of the graded cell catalysts showed high heat
release rate capabilities with low emissions of CO, HC, and thermal NO .
                                                                      j£,
Operation of the catalysts under fuel-rich conditions showed capability to
control fuel NO  emissions.  Additional criteria for system scaleup and opera-
               X.
tion under varying preheat and pressure conditions were also generated.
     Catalysts developed during the program were incorporated into three small-
scale systems with heat extraction.  A radiative catalyst/watertube system,
utilizing direct heat removal from the catalyst, was devised and tested.  The
concept has potential application to watertube boilers.  A model gas turbine
combustor was tested at pressures between 0.101 and 1.01 MPa (1 to 10 atmos-
pheres) to investigate operating characteristics and fuel nitrogen conver-
sion to NO .  The final system, a two-stage combustor, was constructed to
          2k
utilize fuel-rich first stage combustion for fuel NO  control.  Measured
                                                    A
emission results make the concept attractive for a variety of future system
applications.
                                      208

-------
                               INTRODUCTION

     Catalytic combustors have shown promise in reducing the levels of CO,
HC, and NO  emissions over those of conventional burners in laboratory tests.
          X.
In order to develop and later demonstrate catalytic combustors for commer-
cially viable systems, two related development activities were performed.
The first activity involved the development and characterization of catalyst
systems, while the second activity focused gn the analytical and experi-
mental evaluation of three system concepts employing catalytic combustors.
     The following section of this paper presents data on a variety of
catalyst models with varying substrate, washcoat, and catalyst materials.
This data indicates the catalyst system heat release capabilities, preheat
requirements, temperature limitations, and operational life.  The subse-
quent section presents three subscale designs and operational results for
watertube boiler, gas turbine combustor, and two-stage combustion systems
using catalytic combustors.  The concepts, designs, analyses, and data
presented here were generated under EPA Contracts 68-02-2116 and 68-02-2611,
Task 11.
                                       209

-------
                           CATALYST DEVELOPMENT

     In order to develop catalyst models for subsequent system application,
a number of desirable catalyst characteristics were identified.  These
characteristics include:
     •   Low ignition temperature
     •   Low preheat requirements for sustained combustion
     •   Combustion uniformity throughout the bed
     •   High heat release capability
     •   High combustion efficiency
     •   Low pollutant emissions
     •   High operating temperature
     •   Fuel flexibility
     •   Long life
Catalyst models tested included variations in substrate, washcoat, and catalyst
materials — as well as substrate geometry and preparation techniques.
     As reported at the Second Stationary Source Combustion Symposium (Ref-
erence 1), an initial series of combustion tests were performed at the Jet
Propulsion Laboratory and at Acurex.  Results from those tests indicated
that catalyst performance was improved by:
     1.   Increased catalyst loading, resulting in lower initial lightoff
         temperatures, higher mass throughputs, and increased lifetime at
         1367K (2000°F).
     2.   Increased cell size, allowing higher possible mass throughputs at
         the expense of increased hydrocarbon emissions.
     3.  Heavier hydrocarbon fuels which promote lightoff at lower ignition
        temperatures.
     4.  Hydrogen  sulfide  (H^S)  fixation of platinum catalysts to promote
        retention of platinum surface area.
                                      210

-------
     5.  Presintering of catalyst washcoats which reduces burying of active
         catalyst below the surface during combustion.
     6.  Stabilization of y-AlJO,. washcoats with cesium oxide (Cs20) up to
         5 weight percent increased surface area.  Stabilization of alumina
         washcoats with ceria up to 5 weight percent had a negative effect
         on surface area.
     7.  Decreased cell size to significantly reduce unburned hydrocarbon
         and carbon monoxide emissions.
     8.  Catalyst beds of combined large cell and small cell monoliths also
         significantly increase throughput (at a given preheat temperature)
         and overall catalyst life with low emissions.
     9.  Bed temperature uniformity was increased by operation at higher
         temperatures.
     Catalysts with combined large and small cell segments represent the best
concept developed by the test series.  This graded cell concept (three seg-
ments shown in Figure 1) was further developed through predictions of the
PROF-HET catalytic combustion computer code (Reference 2).  Finally, prelimi-
nary data with ammonia-doped natural gas indicated a potential for control of
fuel nitrogen compounds to NO  under lean conditions.
                             Ji

GRADED CELL CATALYST TESTS
     Additional catalyst screening tests were conducted at Acurex on the
graded cell configuration.  The primary objective of these tests was to
identify the best catalysts for system application.  Catalysts were also
tested for high temperature capability, system scaleup design criteria, and
conversion data of fuel nitrogen to nitrogen oxides under varying operating
conditions.
     Screening catalysts were obtained from six sources, including W. R.
Grace and Company, Universal Oil Products, Inc., William Pfefferle (a
private consultant), Matthey Bishop, Inc., Johnson Matthey, and Acurex.
Substrate materials were either DuPont alumina or Corning zirconia spinel.
Washcoats varied from proprietary preparations with high pretest surface
                                     211

-------
 area to no washcoat with low pretest surface area.   Catalysts were  either
 precious  metal, metal oxide, or mixtures.   The catalyst  loadings  and fuels
 used are  listed in Table I.
      To support combustion test results, pre- and post-test catalyst physical
 measurements were made.   These measurements included catalyst surface area
 and dispersion performed in the EPA/Acurex catalyst  characterization labora-
 tory.  Additional scanning electron microscopy (SEM)  and energy dispersive
 analysis  by X-ray (EDAX) tests were performed at the Jet Propulsion Labora-
 tory as required.  Table II is presented for reference,  and summarizes all
 surface area and dispersion measurements.
      Catalysts .were aged for 10 hours and  compared in maximum throughput,
 varying atoichiometry, and minimum preheat tests.  Some  catalysts,  primarily
 precious .metals with low loadings, exhibited lifetimes of less.than 10 hours,
 precluding extensive data evaluation.  Others were tested in a complete  20
 to 30 hours test sequence.  In addition, some catalysts  showed preferential
 activity  for either fuel-rich or fuel-lean combustion or for operation only
 at higher preheats of 700K to 811K (800°F  to 1000°F).
      The  maximum throughput (volumetric heat release rate) capabilities  were
 compared  for various graded cell catalysts to indicate relative activity.
 Catalyst  data selected from Table I are compared below.
                      MAXIMUM THROUGHPUT COMPARISON DATA**
 Catalyst Type
UOP Proprietary
Pt-Ir/Al203

Stab. Pt/Al203

JM Proprietary
 Spinel
Noble Metal/
Manufacturer
UOP 2
W. R. Grace
Matthey
Bishop
Johnson
Matthey
Pfefferle 3
Acurex 3
Bed
K
1522
1256-
1589
1611
1589
1617
1644
SV, 1/hr
343,900
327,400
348,000
603,300
443,100
661,400
in J
q »hr-Pa-m3
3.0 x 106
2.8 x 106
5.2 x 106
7.2 x 106*
5.1 x 10s*
7.8 x 106*
Bed
Uniformity
Excellent
Ragged
Ragged
Excellent
Excellent
Excellent
                                 1756  2,070,000
                  Acurex 4
 Proprietary
**A11 results with natural gas/methane fuel
 *Bed not blown out
 tMultiply entry by 2.7 for Btu/hr-atm-ft3
38.9 x 10
                                                              6*
Excellent

-------
The first three catalysts, either precious metal or proprietary materials,
exhibited volumetric heat release rates  (q111) of approximately 3.7 x 106
J/hr-Pa-m3  (10 x 106 Btu/hr-atm-ft3) at blowout conditions.  The next three,
either proprietary or cobalt oxide catalysts, reached approximately 7.4 x 106
J/hr-Pa-m3  (20 x 106 Btu/hr-atm-ft3) without blowing out.  This heat release
rate represents the maximum facility capability in the screening test con-
figuration.  It is apparent that the cobalt oxide catalysts are generally
capable of higher heat release rates than the precious metals in the graded
cell configuration.  The oxides required operation at higher temperature,
however, to produce combustion efficiencies comparable to the precious metals
at lower temperatures.  The last catalyst, noble metal on a proprietary
substrate, was tested to the limits of the test facility without blowing out.
     During graded cell catalyst testing, operating temperatures were varied
from 1367K to over 1978K (2000 to 3100°F).  Thermal NO  data were obtained
                                                      2k
with natural gas as the primary test fuel.  A number of the test models are
compared in Figure 2.  From 1367 to 1644K (2000 to 2500°F), little variation
among the NO  emissions with catalyst type is apparent with NO levels below
            Jv
20 ppm.  At 1756 to 1867K (2700 to 2900°F), the NO  production rate begins to
                                                  3C
increase, as with conventional flame combustion.
     In Figure 2,  two different catalyst geometries are shown for fuel-lean
operation in the 1644 to 1978K (2500 to 3100°F) range.  The upper curve is
for a two-segment graded cell model (A-029) that acted primarily as a flame
holder for downstream gas phase reactions.  The lower curve represents data
for a three-segment catalyst (A-030) where significantly increased combustion
occurs on the surface of the catalyst bed.  It is apparent that maximizing
the amount of surface reaction occurring in a catalytic combustor minimizes
the production of thermal NO .  These results indicate the importance of the
graded cell configuration, particularly for thermal NO  control in high excess
                                                      X
air gas turbine configurations.
     In addition to baseline catalyst screening tests, selected catalysts
were tested to characterize conversion of fuel-bound nitrogen to nitrogen
oxides with catalytic combustion.  Ammonia was added to natural gas to simu-
late fuels of varying nitrogen content.  Exhaust gas analyses for NO  by
                                       213

-------
chemiluminescent analyzer and for ammonia  (NH«) and cyanide (HCN) by specific
ion electrode were performed.
     A nickel oxide/platinum catalyst was  prepared at Acurex and tested over
a range of stoichiometries from 55 to 200  percent theoretical air at a nominal
bed temperature of 2400°F.  Fuel dopant concentration ranged from 2500 to
10,000 ppmv NH- in the  fuel.  The results  are shown in Figure 3 as the percent-
age of the incoming NH, converted to NH,,  HCN, and NO.  The NH3 conversion to
NO increased from zero  levels under very fuel-rich conditions to better than
90 percent on the fuel-lean side.  NH. conversion to HCN showed the opposite
trend — high under rich conditions and decreasing to zero on the lean side.
Unconverted ammonia was highest below 70 percent theoretical air and remained
at low levels under lean combustion.
     The total of these three curves (dashed line and cross symbols) repre-
sents all assumed NO  precursor species for the NiO/Pt catalytic combustor.
A distinct minimum occurs between 70 and 80 percent theoretical air, where
only 20 percent of the  fuel nitrogen is converted to NO  precursors.
     A second graded  cell catalyst  (cobalt oxide/platinum) was tested with
simulated fuel nitrogen.  The fuel nitrogen conversion is shown in Figure 4.
The ammonia conversion  to nitric oxide provided the samp, characteristic curve
as the nickel oxide/platinum catalyst.  Differences in the HCN and NH- species
measured, however, resulted in lower total NO  precursor (NO + NH^ + HCN)
levels under fuel-rich  conditions.  The m-tn-tmim occurred at a lower value of
theoretical air (60 percent) than that of  the previous nickel oxide catalyst
(75 percent).  The cobalt oxide catalyst could thus be operated fuel rich
without dilution to achieve low conversion of fuel-bound nitrogen to nitrogen
oxides.
     The low fuel nitrogen conversion of these two catalysts at 60 to 80
percent theoretical air has important system implications.  Combustors which
could operate fuel-rich, possibly with staging, have potential for fuel NO
control.   This characteristic of the catalytic combustor under fuel-rich
conditions is the basis for the two-stage  catalytic combustor developed
under this program.
                                      214

-------
     Because of the significance of controlling fuel NO  with catalytic
                                                       A.
combustors, additional testing of NH~-doped propane and methane was con-
ducted.  It should be noted that fuel N conversion in rich catalytic com-
bustion is achieved for much shorter residence times than those associated
with conventional rich-lean combustion systems.  Tests covering a range of
stoichiometries from rich to lean were run to determine the effect of theo-
retical air, bed temperature, mass throughput, fuel type, and fuel-bound
nitrogen type and concentration on the conversion of fuel-bound nitrogen to
NO  or NO  precursors.  In addition, an oxygen/argon mixture was used as the
  X      3t
oxidant in a series of tests to perform a nitrogen balance across the cata-
lytic combustor.  Test results indicate that operating the combustor at
lower bed temperatures (1367K compared to 1478K) minimizes the formation of
fuel NO  under lean conditions but does not affect the formation of NO
       X                                                              X
precursors under rich conditions.  Figure 5 shows the results for a proprie-
tary UOP catalyst tested with ammonia-doped natural gas fuel.  The lower
bed temperature is postulated to minimize gas phase reactions, thereby mini-
mizing NO  formation under lean conditions.  Under fuel-rich conditions,
however, the stoichiometry has a much greater effect than bed operating
temperature.
     The effect of fuel nitrogen concentration on NO  formation under lean
                                                    2v
conditions is shown in Figure 6.  Using the proprietary UOP catalyst at 1478K
bed temperature, the conversion of NH- to NO  was seen to decrease as nitrogen
                                     J      3t
weight percent in the fuel increased.
     The results of nitrogen mass balance investigations using oxygen/argon as
oxidizer instead of air are shown in Table III.  As the table indicates, a
maximum deviation of ±12 percent between the input fuel nitrogen content and
the measured nitrogen species in the combustion products was obtained, except
for two data points.  N_0 was not detected for any test condition.  Under
fuel-rich conditions, as expected, Nj was the dominant product of the chemi-
cally bound nitrogen conversion process.
     Design criteria for the graded cell catalyst concept also included scaleup,
pressure, and preheat characteristics.  Catalyst blowout was again selected to
compare activity.  Based on the results of small-scale testing, a Universal
Oil Products catalyst was selected for scaleup.  Analytical modeling predicted
                                      215

-------
 that combustion throughput capability would scale proportionately  to  bed
 frontal area.  Therefore,  the bed diameter was increased  to provide a 2.7
 increase in frontal area over the small-scale model.   Bed length remained
 fixed at 7.6 cm (3.0 inches).
      Test data showed that maximum throughput (at blowout) did scale  approxi-
 mately with frontal area.   Maximum throughput reported for the small-scale
 catalyst was 258.5 MJ/hr (245,000 Btu/hr)  and 3.42 x  106  J/hr-Pa-m3 (9.3 x
 106 Btu/hr-atm-ft3) volumetric heat release rate.  This compares to a volumet-
 ric heat release of 4.38 x 106 J/hr-Pa-m3  (11.9 x 10s Btu/hr-atm-ft3)  at
 926.3 MJ/hr (878,000 Btu/hr) for the scaleup catalyst at  672K preheat.
      A series of blowout tests were then conducted to determine the operational
 mass throughput limit of the catalyst for  varying preheat and pressure condi-
 tions.  The blowout points used are shown  below.

                         BLOWOUT DATA — CATALYST A-041
Data
Point
1
2
3
4
Bed Temp
K (°F)
1588
1588
1588
1588
(2400)
(2400)
(2400)
(2400)
Preheat Temp
K (°F)
608
478
389
603
(635)
(400)
(240)
(625)
Max. Fuel Flowrate
Kg/hr (Ibm/hr)
15
10
8
22
.0
.5
.4
.2
(33
(23
(18
(49
.0)
.1)
.5)
.0)
Pressure
MPa (atm)
.195
.140
.134
.301
(1.
(1.
(1.
(2.
93)
39)
33)
98)
 Two  things  are evident from the data:
     o   Blowout  scales linearly with  pressure
          •                   •
            fuel max    atm   mfuel max, 1 atm)
     o   Blowout  for catalyst A-041 was exponential in preheat temperature,
         although a relatively  weak exponential factor was determined.
     Operation of the  catalyst  was, of course, possible at any combination of
preheat and fuel  flowrate below the blowout limits.
     The exponential behavior of preheat temperature and its effect on blowout
was shown in PROF-HET code predictions (Reference 2).  The blowout data ob-
tained support this prediction and can be employed as design criteria for
operation of catalyst  A-041 under varying conditions.
                                      216

-------
     The graded cell catalyst screening tests identified the design criteria
required for incorporating the graded cell configuration into system appli-
cations.  Specifically, these criteria included mass throughput and heat
release capabilities, emissions under varying operating conditions and with
nitrogen-containing fuels, lightoff requirements, and current lifetime capa-
bilities.  In addition, the state of the art in catalyst development has been
evaluated, including an understanding of preparation techniques, material
capabilities, and material interactions.

                            SYSTEM CONCEPT TESTING

     The design criteria generated for graded cell catalyst configurations
were used in the specification of three small-scale systems incorporating heat
extraction techniques.  A radiative catalyst/watertube system exhibited a
stoichiometric, water-cooled combustor for boiler application.  A model gas
turbine combustor utilized high excess air for catalyst and exhaust gas tem-
perature control.  Finally, a two-stage combustor was constructed to evalu-
ate fuel nitrogen control with application to either boiler, furnace, or
turbine equipment.

RADIATIVE CATALYST/WATERTUBE SYSTEM
     The radiative catalyst/watertube concept is shown in Figure 7.  A
stoichiometric fuel/air mixture is fed to the radiative section which contains
a close-packed array of catalyst elements and watertubes.  The mixture is par-
tially combusted by the catalyst which is kept at a low surface temperature by
radiation heat loss to the watertubes.  The combustion products and remaining
unburned fuel and air are then passed to a downstream catalytic adiabatic com-
bustor to complete combustion reactions.  A final convective section extracts
energy from the fully combusted gases.  Both catalyst sections operate well
below the maximum use temperature of the catalyst supports — the radiative
section by radiative cooling and the adiabatic section by dilution of the fuel/
air mixture with exhaust products from the radiative section.  The radiative
section was constructed and tested independently of the downstream adiabatic
combustor and convective sections.
                                      217

-------
     An initial test series was conducted using a platinum catalyst on Coors
alumina cylinders.  Stoichiometry was varied from 50 to 219 percent theoreti-
cal air and fuel mass flowrate from 2.1 to 6.7 kg/hr (4.7 to 14.8 Ibm/hr) of
natural gas.  Preheat conditions were also varied.  Figure 8 shows the energy
extracted by the cooling tubes out of the total available energy at the bed
inlet as a function of stoichiometrie ratio.;  The total available energy in-
cludes the fuel heating value and the sensible preheat energy.  Thermal
input to the catalyst cylinders is primarily controlled by the adiabatic
flame temperature of the fuel/air mixture.  This temperature peaks near unit
Stoichiometry, and as a consequence, the tube temperatures have a correspond-
ing maximum.  As theoretical air percentage increases above 100 percent,
catalyst surface temperature begins to decrease, decreasing the radiant
exchange to the watertubes.  The higher total mass throughput, however,
also increases convective  heating of the watertubes such that at fixed fuel
flowrate the energy exchange does not fall off rapidly.  Overall combustion
efficiency at 100 percent  theoretical air was calculated as approximately
37 percent from  the data.   Significant emissions of CO and HC were measured
for  the radiative section  due to the incomplete combustion.  NO  levels
                                                               Zb
were consistently below 2  ppm as measured.
     The radiative section was also tested to evaluate fuel nitrogen conver-
sion characteristics of the system.  For those tests, natural gas was doped
with ammonia and Stoichiometry was varied from 52 to 120 percent theoretical
air.
     Figure 9 shows the fuel nitrogen conversion characteristics of the
radiative system for natural gas doped with 2000 ppm of ammonia.  Low NO  and
                                                                        3E
high NH. values above 100  percent theoretical air are consistent with the
incomplete combustion characteristics of the radiative section.  The low point
in the total NO  precursor curve (NH, + HCN + NO ) at 60 percent theoretical
               A                    J           X
air is similar to those obtained for metal oxide graded cell catalysts.  It
should be noted that this  low level of conversion is attained even though the
fuel is not fully combusted.
     The radiant  section  as tested is not fully suited for complete system
development.  The addition  of  the downstream adiabatic catalytic combustor
                                     218

-------
would result in too high a temperature in that region at stoichiometric con-
ditions.  This occurs since combustion efficiency in the first stage was not
quite as high as expected at the nominal 4.3 kg/hr (9.5 Ibm/hr) design con-
dition.  The"radiative catalyst/watertube section did exhibit excellent per-
formance at stoichiometric conditions with very low levels of NO .  The
                                                                3v
potential for control of fuel nitrogen conversion and the extremely stable
operation experienced under all test conditions make it attractive for future
boiler applications.

MODEL GAS TURBINE COMBUSTOR
     Since the graded cell catalyst was demonstrated to have the low preheat,
high heat release, and pressure capabilities required for gas turbine applica-
tions, a 1055.1 HJ/hr (106 Btu/hr) model turbine can and fuel injection system
were constructed.  The system and catalyst are shown in Figure 10.  Testing was
performed at Acurex and Pratt and Whitney Aircraft (West Palm Beach, Florida)
facilities.  Acurex and UOP catalysts were prepared on both duPont alumina
and Corning zirconia spinel support materials of varying configurations.
     The model gas turbine combustor was first tested at Acurex with propane
at pressures between 0.101 and 0.354 MPa (1 and 3.5 atmospheres).  A heat re-
lease rate of 263.8 MJ/hr (250,000 Btu/hr) at 1478K (2200°F) bed temperature
was run as the nominal test condition.  No significant emissions of NO  or CO
                                                                      JL
were obtained.
     Pratt and Whitney test data were obtained with propane, No. 2 oil, and
No. 2 oil with 0.5 weight percent nitrogen as fuels.  Heat release rates to
844 MJ/hr (800,000 Btu/hr) were achieved with low NO  emissions for both pro-
                                                    Jv
pane and No. 2 oil.  Some difficulty was encountered with flashback and flame-
holding on the fuel nozzles when running No. 2 oil.  High CO and unburned
hydrocarbon emissions resulted from operating at the low bed temperatures
(near the breakthrough limit) required to avoid flashback.  Variations in
pressure under lean conditions were not found to affect emission levels.
Tests run with pyridine-doped No. 2 fuel oil, however, increased the NOX emis-
sions levels, representing percentage conversions of fuel nitrogen to NO  of
                                                                        2£
100, 61, and 55 percent for test pressures of 0.303, 0.505, and 0.707 MPa,
                                       219

-------
respectively.  Subsequent inspection of the test hardware showed that low
liquid fuel inlet velocities due to a fabrication error causing all liquid
fuel to be introduced through the large diameter gaseous fuel injection ports
were responsible for the flashback and flameholding.
     A final test series was conducted at Acurex (after the test hardware
had been reworked) with natural gas, natural gas doped with ammonia, and No. 2
oil.  Pressures from 0.101 to 0.808 MPa (1 to 8 atm) with natural gas and 0.101
to 0.505 MPa  (1 to 5 atm) with fuel oil were included.  A decrease in ammonia
converted to NO  with pressure was observed.  These results are consistent
with those obtained at Pratt and Whitney with pyridine-doped No. 2 fuel oil.
Maximum throughput for the catalyst at 0.303 MPa, 1244K (3 atm, 2100°F) pres-
sure and bed temperature, and 561K (550°F) preheat temperature was also in-
vestigated.  At space velocities near 200,000 per hour, the catalyst began to
break through with increasing CO and unburned hydrocarbon emissions.  Nitrogen
oxide emissions remained at near zero levels throughout the test.  Full blow-
out was not  achieved as control of catalyst temperature during breakthrough
produced difficulties in system control.  The maximum heat release obtained
was 615 MJ/hr (583,000 Btu/hr).
     Final tests were conducted with diesel fuel to compare emissions with
those from natural gas.  An increased bed temperature was maintained for the
oil tests to maintain uniform bed conditions and suppress soot formation.
The NO,  levels were higher than for natural gas (15 ppm compared to 3 ppm) due
primarily to the small amount of fuel nitrogen in diesel fuel.
     The results of the model gas turbine testing demonstrated the application
of the graded cell concept in a system similar to current turbine combustor
designs.  High mass flowrates were achieved in a relatively small volume com-
bustor.  Overall pressure drop for the combustor and fuel injector were meas-
ured at less than 1 percent at 0.303 MPa (3 atmospheres) test pressure.

TWO-STAGE COMBUSTOR
     The two-stage catalytic combustor is attractive for two reasons.  First,
it allows control of bed temperatures to those compatible with the support
material without large excess air requirements.  Second, the first stage can
be operated fuel-rich, which has been shown to be advantageous for reduced
                                       220

-------
 conversion  of  fuel nitrogen to  nitrogen oxides.  A  two-stage combustor was
 designed and constructed to exhibit  these  concepts.
     The two-stage combustor is shown  schematically in Figure 11.  A fuel-
 rich mixture is  introduced  into the  primary  stage which  contains a graded
 cr"1.! catalyst  bed.   The  fuel is partially  combusted, and the released energy
 is removed  by  an interstage heat exchanger.   Sufficient  secondary air is
 then injected  into the combustion products to complete combustion of the
 remaining fuel in the second stage.  The full system combustor would also
 include a second heat exchanger to remove  the combustion energy released in
 the second  stage.
     The two-stage combustor containing  two  cobalt  oxide catalysts was tested
with natural gas  at  0.101 MPa and 0.202  MPa  pressures (1 and 2 atmospheres).
Lightoff and steady-state operation presented no unusual control problems.
The combustor was tested at  an  overall stoichiometry varying from 70 to 150
percent theoretical  air at a nominal fuel  flowrate equivalent to 211 MJ/hr
 (200,000 Btu/hr)  heat release rate.  The first stage stoichiometry was varied
from 40 to  70 percent theoretical air.   Ammonia was added to the natural gas
fuel at a rate of 0.2 to 0.4 weight percent.
     Bed temperatures ranged from 1256K  to 1660K depending on theoretical air,
for a relatively  constant preheat of 617K  (650°F).  The  energy extracted in
the interstage heat exchanger represents 50  to 60 percent of the combustion
energy generated  in  the first stage.
     The results  of  vu2 fuel nitrogen conversion data are shown in Figure 12
as a function of  overall combustor stoichiometry.   The data show that when
operating above 100 percent  theoretical  air,  only nitrogen oxides are normally
present.  Under overall fuel-rich conditions,  fractions  of ammonia and cyanide
 are also present.  These results are consistent with fuel nitrogen data ob-
tained on the single cobalt  oxide catalyst (model A-037, Figure 4).  The data
in Figure 12 show a  nominal  30  percent conversion rate of fuel nitrogen to NO
                                                                             X
precursors with  a value of approximately 27  percent near overall stoichiometric
conditions.  A slight decrease  in conversion was noted at 0.202 MPa (2 atm)
pressure.
     The data shown  in Figure 12 at approximately 10 percent conversion levels
varied in test conditions from  the other data in two respects:
                                       221

-------
     1.  Most significantly, the first stage was operated at higher values
         of theoretical air  (60 and  70 percent) compared with 50 percent for
         the initial  data, and
     2.  The first  stage  catalyst had experienced some sooting by later
         test  times when  the data were taken, causing the catalyst to
         operate at lower temperatures with incomplete combustion.
 The first  stage sooting of the cobalt catalyst proved to be a limiting factor
 in the test life of the system.  The incomplete combustion occurring at later
 test times was evident by increasing measured carbon monoxide levels.
     The demonstration of the two-stage  combustor showed a number of important
 results:
     1.  The two-stage combustor is  effective in controlling conversion of
         fuel  nitrogen to nitrogen oxides under stoichiometric and fuel-
         lean  conditions.
     2. A slight decrease in nitrogen conversion was found at 0.202 MPa
          (2 atmospheres)  pressure.
     3. The variation of first stage stoichiometry impacts overall fuel
         nitrogen conversion.
     4.  First stage  sooting of the  cobalt oxide catalyst was a limiting
         factor in combustor operating life.
 Application of  the  concept of both boiler and turbine systems is possible.
 Convective heat exchangers downstream of each catalyst stage would provide
 for  steam raising in  boiler  applications.  Interstage cooling would not be
 required for gas turbine  systems where high excess air in the second stage
 could be used to control  the catalyst and exhaust gas temperatures.  Further
 work is required for  optimization of the system and catalyst elements for
 specific applications.

                                  CONCLUSIONS

     As a result of this  research and development program, significant progress
has been made toward developing a practical catalytic combustion system.  Be-
fore the step to demonstration  can be taken,  however,  additional work relating
                                      222

-------
to the integration of the catalytic combustor into the total combustion system
must be performed.
     Based upon the analysis and test results of this program, the design,
fabrication, and operation of catalytic combustors with high volumetric heat
release rates and low emissions have been demonstrated.  Both precious metal
and oxide catalysts have been tested over a wide operating temperature range.
The precious metal catalysts should be limited to temperatures below 1589K
(2400°F) for catalyst life considerations, while oxide catalysts can be oper-
ated for long periods at temperatures above 1644K (2500°F).  Catalyst perfor-
mance has been greatly enhanced through the use of graded cell monoliths and
higher catalyst loadings.
     Catalytic combustors have been shown to be effective in controlling
both thermal and fue} NO  emissions.  The thermal NO  control appears to re-
                        X                           X
suit from maximizing surface reactions in the combustor, while fuel NO  can
                                                                      3t
be minimized by operating at a rich fuel/air ratio which minimizes the forma-
tion of HE-, HCN, and NO,  with complete combustion of CO and HC at a later
time.
     The maximum throughput of a catalytic combustor is a linear function of
pressure and an exponential function of preheat.  Thus, for a given preheat,
the catalyst is face velocity limited in throughput ability.
     Small-scale catalytic combustion system configurations have been tested
and show the feasibility of direct radiative removal of bed heat for
temperature control, two-stage catalytic combustion for temperature and fuel
NO  control, simulated exhaust gas recirculation through the use of nitrogen
  X
diluent for temperature control, and high excess air operation.  The combus-
tion system concepts that have been operated show that  it is possible to
operate near stoichiometric conditions with less than 10 ppm NO  and CO in a
                                                               A
natural gas-fired catalytic combustor.
     A number of areas in catalytic combustion need to  be addressed to capital-
ize on the progress to date.  Additional testing of simple and mixed oxide
catalysts for combustion and fuel nitrogen conversion abilities is needed,
along with life testing of selected catalysts to 1000 hours at various
pressures.
                                       223

-------
     Exploratory work with heavy fuel oils (Nos.  4,  5,  and  6)  and pulverized
coal should be conducted to determine system feasibility and fuel preparation
problems.  The potential of catalytic combustion in controlling NO  emissions
from the combustion of these fuels is great and needs early experimental
verification.
     Development of auxiliary systems required to interface with the catalytic
combustor is also needed.  This includes lightoff systems,  temperature control
systems, and fuel and air introduction systems.  In addition,  further testing
of the radiative catalyst/watertube, two-stage combustor, and gas turbine com-
bustor systems is needed to more thoroughly define operating ranges with a
variety of fuels.
     Finally, the design, fabrication, and operation of a demonstration unit
should be undertaken when the above work is completed.   The demonstration unit
would be operated as a laboratory device for several months prior to the initi-
ation of field demonstration tests.
                                      224

-------
                                 REFERENCES
1.  Kesselring, J. P., ^t &L., "Design CritgEia_for Stationary Source Cata-
    lytic Combustors," published in Proceedings of the Second Stationary
    Source Combustion Symposium, Volume III, EPA-600/7-77-073c, July 1977,
    pp. 193-228.

2.  Kelly, J. T., et al.., "Development and Application of the PROF-HET
    Catalytic Combustor Code," Paper No. 77-33 presented at the Western
    States Section of the Combustion Institute, October 1977.
                                       225

-------
ro
ro
   *******
  *.******
  *******
*******
********
*******
mmmmm*l
mmmmmmm
              Figure 1. Corning square-celled extruded monolith structures.

-------
   350
   300
   250
VI
VI
0>
u
X

-------
            lOOr
ro
ro
Oo
        0)
        &
         •I

        X

        o

-------
          100 r
ro
ro
                                                                         Catalyst A-037  (Co203/Pt)
                                                       100         120         140


                                                    Theoretical air, percent
160
                        Figure 4.  NH3 conversion characteristics, natural gas doped with ammonia

                                  Co203/Pt catalyst.
                                                                                                      /\ 2 atm
                                                                                                        3 atm
            <
                                                                                                       180

-------
                                         UOP catalyst

                                         Natural gas fuel

                                         Nitrogen concentration  = 0.5 wt percent

                                         Fuel flowrate =25.3 MJ/hr  (24,000 Btu/hr)
                            100
ro
CJ
o
                         c
                         0)
                         u

                         OJ
                         Q.
X

o
                         o

                         'E
                         O>

                         o
                         u
                         CO
                             80
                             60
     40
                             20
                                    60
                                               I
                                I
                                                1478K (2200°F)


                                            A  1367K (2000°F)
I
                              I
I
80       100       120      140


    Percent Theoretical Air
                                                           160
                           180
                  Figure 5.   Effect of bed temperature on NHa conversion to NH3+HCN+NOX, UOP catalyst,

-------
    120
    110

ox   90
o    80
t
8
70
     60
     50
                    UOP Catalyst
                    Propane Fuel
                    Theoretical  A1r =  310 percent
                    Fuel  flowrate = 25.3 MJ/hr  (24,000 Btu/hr)
                    Bed Temp.  =  1478K  (2200°F)
            I
I
I
I
I
I
                0.5      1.0       1.5      2.0      2.5      3.0
                 Chemically Bound Nitrogen Content, Ht. % of Fuel
                                                                          N
                                                                          0
                                                                          O
                                                                  3.5   4.0
             Figure 6.  Effect of nitrogen content on NH«  conversion to
                        NOX, UOP catalyst.              J
                                       231

-------
             -Refractory lining
ro
w
PO
                                             Catalyst

                                             coated cylinder
                 Mud

                drum
                                                                      Radiative heat
                                                                      transfer section
                                                                                                 Monolith bed

                                                                                                - Adlabatlc
                                                                                                 Combustor
                                                                                                                    Convectlve heat exchanger
                                             Matertube
                                              Figure 7.   Radiative water-tube  boiler concept.

-------
ro
co
CO
           150
                    140
125
                    120
                    100
           100
                  I
                  O
            75
 50
                     80
1  60
>>
                  O)


                  UJ
            25

                                                                            Total  available  energy
                                                                 Fuel mass flowrate =2.1 Kg/hr
                                                                Total  Energy Release
            0-L
                                                                                                            OD
                                                                                                  i
                                60
                               80       100      120      140       160      180      200      220(


                                             Theoretical  air,  percent
                     Figure 8.  Radiative catalyst/watertube system energy release vs theoretical air.

-------
            TOO r
                                                    0
             80
         c
         (U
                                                                                                 NH3 + HCN + NOX
         X
         o
             60
                                                                                                       NH.
no
to
0)
c
o
u

r
             40
             20
              0
               50
                  60
70
80          90          TOO

 Theoretical air, percent
110
130
                         Figure 9.  Radiative catalyst/watertube  system,  fuel  nitrogen  conversion

-------
Figure 10.  Model gas turbine combustor.
                    235

-------
ro
co
                                                                      iWCTION
                                                                        2
                                                                  INLIT « oururr
                                                                  COOC'MA
                                                                  MILK KNC 0>HH )
                                        Figure 11.   Two stage catalytic arrangement.

-------
IVJ
w
IUU
80
X
o
•f
£> 60
3
j conversion
-P»
0
%•*
3:
20

0
•»
A 505
O 605
a7n«
/u/
r»j
\ &O (D
1 Stage ±
-

A
^•+ ^ ^***
^^ ^ta» '^V ^^ ^^
^ «•*"
^ .202 MPa
0J2^ Q-70%
^ 60%
1 1
                      50
                                                                                  One-stage combuitlon

                                                                                  50% First Staye Theoretical  Air
                                                                              O  60% First Stage Theoretical  Air
                                                                                  70% First Stage Theoretical  A1r

                                                                                     i  atni

                                                                                     2  atm
                                                    First stage
                                                    stoichlometry
                                                    50% T.A.
               100

Overall theoretical air, percent
150
                               Figure 12.  Two-stage combustor fuel nitrogen conversion.

-------
                                        TABLE I.  GRADED  CELL CATALYST MODELS
S*Ml* M.



AEM-OM
/UM-027
ACM-OB
ACTO-029
AEM-030
AEM-031
AEM-OJ!
AEM-033

AOW-034

AEMMI3S

WKO-036


AEM-037
AEMO-038
AEM-041)

AEM-M1
Ho. Of
TC



6
*
-
1
6
1
i
T

7 r

7

7


4 .'•
5
S

4
_ Svbitrit*
"""• Typi ,



OuPont
OuPont
DuPont
Corning
Corning
OuPont
OuPont
Coming

Corning

OuPont

Corning


Corning
OuPont
DuPont

DuPont
AlurtM


AluulM
Thorl* «•
! llrcenli 1*.
i Zlreonli
tplntl
IlrconU
iplml
AlartM
AluriM
Zt
11 9/4 VI
_

-

.47/.76/.80



.3./.I4/.17
.M/.Ot/.OO
WS/.31/0
"17.8/5.0
4^1/0/1 ,
_^
-
18/0.7/0 Pt
1.0/t.V
4.1 mo
3.0/0/0 Pt
7.7/9.9/
M.t C*iO]


3.0/3.2/
4.0 (HO
t.0/1.4/
0 Pt
4.S/4.8/
t.inriffl
10.3/1.9/0.8


-
tUtM
4/Z5 - 4/Z9/77


6/M - 8/W77
8/30 - 7/1/77
-
7/10 - 7/11/77
.7^14 - 7/11/77
0/7. - 1/10/77
7/M - 7/J9/77
1/14 - 9/M/77

t/tt - g/t7/77

10/1 - 10/0/77

10/10 - 10/M/7;


1Z/5 - 1Z/JW77
12/28 - 12/29/77
I/2S - 2/6/78

12/30/77 -T/3/7B

Mt. <»


Mt. tjt
Mt,0,t
-
Mt. til
Mt. 0*1
Mt. 041
Mt. Mt
Mt. «»

Mt. Oil

Mt. HI

Mt. Mt


Mt. «l
Mttnint
Mt. Sis'
Mt. CM

Mt. Gi<

Scrtn «r*c* Pt/lr c4U1yit


Scram UOP cMlytt
OotoratM offtctt of MgnMltfng point praclout MUM ind
eittlytt Hfthoiit Mtncott
Mt to bt tnUtf Mtttf at rttnltl of A-027.
InrtitlflU Mtil mid* citalytt eipibHUItt (M
Mthco>t). Ptrfor. nlgb tnp. oo*r»t1m (3100T),
Mt*1 oi1d* ctovirlton, nt*Mln tviliMtlon
Ctnwr* to A-032
fcrttn Mtth*y tlihop utilytt
ru*1 nltrogm ond praitvr* tatting 1 T«t dlfflcullUi
> procluvtu dit*
1 ratultt
CoMM-tmi t* A-033 )

ScrMnliNj ciMp*r1un

Fwl nltragm totting


Fwl nttragm ond pntiim Uttlng
tnv*tt1gtt* ctulyst-support 1nt*rtct1ont
Scrwn Johnion Pktthty cittlytt

ClUlytt toltvp, 6.06-inch dtwcttr
ro
CO
00

-------
                   TABLE  II.   SURFACE AREA AND DISPERSION MEASUREMENTS ON GRADED CELL CATALYSTS
ro
to
vo
Sample No.
AERO-025
AERO-026
AERO-027
AERO-OM
AERO-029
AERO-030
AERO-031
AERO- 032
AERO-033
AERO-034
AERO-035
AERO-036
AERO-037
AERO-OM
AERO-MO
AERO- 041
Surface Area **/•,
Pretest Post-test
1.55-2.49
5.94
0.44
0.06
0.60
0.15
24.36
11.99
—
-
5.17
—
-
0
4.00
6.37
0
0
-
-
0
.01
0
0
—
-
0.09
—
-
-
0
0.50
Dispersion 1
Pretest Post-test
1.5-4.9
20.64
8.33
mm
-
36.16
f.TI
-
-
4.09
-
—
-
-

0
-
-
•"•
—
0
0.20
-
-
0
-
—
—
-

Catalyst
Pt/Ir
Proprietary
Pt/fr/Os
Pt/Ir/Os
NIO/Pt
CojtyPt
Stab. Pt
Stab. Pt
NIO/Pt-M
COgOj/Pt
Stab. Pt
NtO/Pt
Itjfiym
cozo3
Proprietary
Proprietary
SEH/EDAI Results and CoMtents
No Pt or Ir found on back tw
seojmts by SEH/EOAX

Not combustion tested




Invalid test data
Invalid test data


Zero surface area on each sevaant




-------
                                          TABLE III.  FUEL NITROGEN BALANCE
Run
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14-
Catalyst
Type
Pt
Pt
NiO/Pt
WO/Pt
UOP


















Fuel-N
Content
(Wt. % of Fuel)
2.



3.




2.
2.
3.
3.
3.
0



0




4
0
6
0
0
Theoretical
A1r
(%)
98
77
200
75
100
90
80
75
70
100
75
90
80
70
Percent Conversion of Nitrogenous
Species in the Combustion Products
NOX
76.00
0.83
80.57
9.63
34.13
1.06
0.58
0.29
1.00
23.77
0.25
0.68
3.28
2.47
N2
15.97
79.05
13.19
58.94
71.54
58.93
79.29
98.68
74.86
68.20
74.01
61.50
54.68
45.60
NH3
8.12
18.07
1.81
41.20
4.77
44.11
25.17
4.49
4.29
5.90
5.36
42.34
37.92
41.91
HCN
1.22
1.51
T.21
3.09
1.52
2.96
2.03
2.59
2. 95
0.72
2.33
4.71
16.79
7.00
N20
0


























Error
.(*)'
0.10
-0.56
-3.21
12.86
11.97
6.81
7.07
6.05
-16.91
-1.43
-18.02
9.23
12.66
-3.02
ro
4*
o

-------
SESSION III:  SPECIAL TOPICS
             BY




     Joshua. S. Bowen
            241

-------
       EPRI LOW COMBUSTION
          NO  RESEARCH
               BY
          D. P. Telxeira
Electric Power Research Institute
      Palo Alto, California
                243

-------
              EPRI Low Combustion NOX Research
                             by
                        D.P. Teixeira
               Electric Power Research Institute
                          ABSTRACT

Recent results of EPRI's research and development efforts on the
primary combustion furnace  (PCF) concept to achieve low NOX with
pulverized coal will be presented.  Principal of operation and
earlier 4 million Btu/hr data will be reviewed.  Recent data
from nominal 50 million Btu/hr scale tests will be covered.

Full scale application issues related to the PCF will be discussed.
A technical and economic comparison of combustion and post combus-
tion control technologies will be made.
                                244

-------
        FLUE GAS TREATMENT TECHNOLOGY
               FOR NO  CONTROL
                     A
                     By:

               J. David Mobley
          Process Technology Branch
    Utilities & Industrial Power Division
Indr°trial Environmental Research Laboratory
    J.S. Environmental Protection Agency
     Research Triangle Park, N.C. 27711
                    245

-------
                                  ABSTRACT
     The Environmental Protection Agency has maintained a program to further
the advancement of NO  control by flue gas treatment technology since the
                     2£
early 1970's.  The program consists of technology assessment and control
strategy studies in conjunction with small scale experimental projects.
These activities have shown that 90% reduction of NO  emissions by selective
                                                    JL
catalytic reduction with ammonia has been commercially demonstrated on gas-
and oil-fired sources in Japan, and that such processes are ready for test
application on coal-fired sources.  Based on the experience in Japan, success-
ful application of the technology in the U.S. can be expected.  Further,
recent assessments indicate that 90% control of both NO  and SO  emissions
                                                       x       x
by dry simultaneous control techniques warrants further investigation.
However, wet simultaneous NO /SO  removal processes are not currently attrac-
tive.  The paper also assesses the need for highly efficient NO  control in
                                                               X
the U.S. and suggests that, although the nationwide need for flue gas treat-
ment technology has not been established, certain regions will probably
require application of the technology, in the future.
                                    246

-------
                                INTRODUCTION
     Nitrogen oxides (NO ) and sulfur oxides (SO ) in the atmosphere have
                        X                       X
been determined to have adverse effects on human health and welfare.  To
aid in preventing these adverse effects, the Industrial Environmental
Research Laboratory at Research triangle Park, N. C. (IERL-RTP) is leading
the U.S. Environmental Protection Agency's (EPA's) efforts to develop and
demonstrate NO  and SO  control technologies for stationary combustion
              X       X
sources.

     Flue gas desulfurization technology has progressed to commerical
application and has achieved 90% control of S09.  Although NO  control by
                                              £• •'             3£
combustion modification technology has been applied commercially, NO
                                                                    X
control by flue gas treatment technology has not been utilized in the U.S.
Combustion modification technology reduces NO  emissions by approximately
                                             X
50% in a relatively cost effective manner.  Flue gas treatment technology
should be able to reduce NO  emissions by 90% and has the potential for 90%
                           X
control of both NO  and SO  emissions.  The focus of this paper will be on
     • •            2(       3t
flue gas treatment technology.

     EPA is proceeding with small scale NO  and NO /SO  flue gas treatment
                                          X       XX
experimental projects in parallel with technology assessment and control
strategy studies.  To save both development time and money, EPA is investi-
gating Japanese technology for potential application to the U.S. coal-fired
situation.  Through these actions, the basic foundation will be established
if the technology is required in the U.S. and acceleration of the develop-
ment program becomes necessary.
     This paper presents a summary of the past, current, and planned NO
flue gas treatment programs of EPA and an overview of the technology in
Japan.
x
                                     247

-------
                      EPA'S NO  FLUE GAS TREATMENT PROGRAM
                               A

      The thrust of  EPA's  current NO  flue  gas treatment  (FGT) program is
                                    2W
 technology assessment and control strategy studies in conjunction with
 small scale experimental  projects.   The technology assessment and control
 strategy studies are mainly paper studies  which examine various aspects
 of NO and NO  /SO  control technology,  estimate if and when highly
 efficient NO  control will be needed in the U.S., and assist in determining
 the  appropriate scale of  the  experimental  projects.

      This paper will begin with a discussion of the following technology
 assessment studies:
          o     Assessment of  Technology in Japan
          o     Assessment of  Technology for U.S. Application
          o     Assessment of  Technology for an Industrial Boiler New
                Source Performance Standard
      Results of the control strategy assessment studies will then be
 presented:
          o     Assessment of  Point Source  Impact on Ambient NO  Levels
                        -'        ......                  x
          o     Assessment of  the Need for  NO  Flue Gas Treatment Processes

      Subsequently,  experimental projects will be discussed:
          o     Hitachi  Zosen  NO  Pilot  Plant
          o     UOP-Shell  NO /SO  Pilot  Plant
                           X,    J±
     Finally,  conclusions that can be drawn from these activities will
be identified.
                                    248

-------
TECHNOLOGY ASSESSMENT STUDIES
     Technology assessment studies report the rapid advances of the
technology worldwide.  These studies, which have concentrated on Japanese
technology, evaluate the various processes and process features for
interested parties in the United States.  In addition, the feasibility
of application of the most promising processes to combustion sources in
the United States is addressed from technical, economic, energy, and
environmental standpoints.  This effort also includes assessment of the
technology for potential regulatory actions.
Assessment of Technology in Japan
     Due to stringent emission standards, Japanese technology for control
of NO  and simultaneous control of NO  and SO  by flue gas treatment
     X                               X       X
techniques is more advanced than any other country's.  EPA has sponsored
the publication of periodic reports and papers to facilitate the transfer
of information from Japan to the United States.  These documents have
been mainly prepared by Jumpei Ando of Chuo University, Tokyo, Japan
(References 1,2,3).  An overview of current Japanese technology, both
dry and wet processes for NO  and NO /SO  control, follows.
                            X       X   X
Dry NO  Processes  (Reference 3) —
      Jv
     Numerous dry process types are being developed.  However, selective
catalytic reduction  (SCR) processes are the only ones that have achieved
notable success in treating combustion flue gas for removal of NO  and
                                                                 «£
have progressed to t .^ point of being commercially applied.  SCR processes
are based on the preference of the reaction of ammonia  (NH_) with NO
                                                          •J         3t
rather than other flue gas constituents.  Since oxygen  (02) enhances the
reduction, the reactions can best be expressed as:

     4NH3 + 4NO + 02  - catalyst  ,    ^ + ^
     4NH3 + 2N02 + 02  - CaayS   »    3N2 + 6H20             (2)
     Reaction 1 predominates since approximately 95% of the NO  in
combustion flue gas is in the form of nitrogen oxide (NO).  Therefore, a

                                    249

-------
stoichiometric amount of ammonia can be used to reduce NO  under ideal
conditions to harmless molecular nitrogen  (N_) and water vapor (H~0).
An NH.:NO mole ratio of about 1:1 has typically reduced NO  emissions by
     3                                                    *
90% with a leak ammonia rate of less than  20 ppm.

     The SCR processes are relatively simple, requiring only a reactor,
a catalyst,  and an ammonia storage  and injection system.  Some increase
in boiler fan capacity, or possibly an additional fan, may be necessary
to account for the increase in pressure drop which may be in the range
of 500-700 Pa.

     The optimum temperature for the reaction is about 1000°C.  However,
the catalyst effectively reduces the reaction temperature to the 300-
450°C range.  To obtain flue gas temperatures in this range and to avoid
the requirement for  large amounts of reheat, the reactor is usually
located between the  boiler economizer and  the air preheater.  Depending
on system design,  the reactor may be located either before or after  the
particulate  control  device.

     Many different  types of catalyst compositions and configurations
have been developed.   Initially, catalysts were developed for flue gases
without particulate  and SO  concentrations such as from natural gas
firing.  For these applications, a  catalyst of platinum (Pt) on an
alumina (Al_0_) support material was used. Alumina was poisoned by  SO ,
           fc J                                                        X
particularly SO.,  in the flue gas.  Titanium dioxide (TiO.) was found to
be resistant to SO   poisoning and to be an acceptable catalyst carrier
for applications with  SO  in the flue gas  such as from oil or coal
firing.  The catalyst  metals also tend to  react with SO , especially
SO..  Vanadium compounds were found to be  resistant to attack from SO.
and promote  the reduction of NO  with ammonia.  Therefore, many SO
                               3fc                                  2£
resistant catalysts are based on TiO. and  V.O,..  Other active metals are
also used including C, Co, Cr, Cu,  Fe, Mn, Mo, Ni, W, their oxides,
sulfates,  or combinations thereof.  However, the exact compositions  and
constituents of most catalysts are  proprietary.
                                   250

-------
     Reactor configurations also vary with the application, primarily to
accommodate the different particulate concentrations.  For natural gas
firing, a fixed, packed bed employing a granular or pellet type catalyst
can be used.  However, particulates in the gas stream will plug a fixed,
packed bed.

     For applications with moderate particulate concentrations, a moving
bed arrangement can be employed and still utilize the granular or a ring
shaped catalyst.  (The catalyst is charged from the top of the reactor
and moves down intermittently or continuously while the flue gas passes
through the catalyst layer in a cross flow direction.  The catalyst is
discharged from the bottom of the reactor, is screened to remove particulates,
and can be regenerated to eliminate any contaminants before being returned
to the reactor.)  Moving bed reactors can treat gases containing less
than about 200 mg/Nm  of particulates and are applicable to oil firing.
They are also applicable to coal firing if an adequate amount of particulates
is removed from the flue gas upstream of the NO  reactor.  In most
                                               Ji
cases, this would require application of a hot-side electrostatic
precipitator (ESP) which is usually more expensive than the more commonly
used cold-side ESP.
     Parallel flow reactors and catalysts were developed to tolerate
relatively high particulate concentrations, such as  from coal firing.
(Parallel flow indicates that the direction of flue  gas is parallel,
rather than perpendicular, to the catalyst surface.)  Parallel flow
designs utilize various catalyst configurations including tubular,
metallic and ceramic honeycombs or plates, and parallel passage reactors.
The selection of the catalyst and reactor configuration will probably
depend on site specific considerations since there are advantages and
disadvantages associated with each design.

     Even though much progress has been made in catalysts and reactor
design, some problems still remain.  The catalysts may not be resistant
to all contaminants in flue gas.  In addition, fine  particulates, smaller
than about 1 micron, may blind the catalyst surface.  Catalyst life
                                    251

-------
needs to be extended from the current guarantees of 1 to 2 years for
applications with SO  and participates in the gas stream.
                    3t
     One of the major concerns with SCR processes is the formation of solid
ammonium sulfate  [(NH,)2SO,] and liquid ammonium bisulfate (NH^HSO^).  The
problem occurs if S0_, NH-, and H20 are present in sufficient quantities,
and  the flue gas  cools to the formation temperature of (NH^) 2S04 and NH4HS04 •
For  example, NH.HSO^ will form at  210°C if the flue gas contains 10 ppm of
NH,  and 10 ppm of SO-.  The formation conditions are difficult to avoid
   J                 J
since one  expects some leak ammonia from a SCR system and some SO- from
combustion of sulfur containing fuels.  In addition, some catalysts tend to
promote the oxidation of SO- to S0«; however, this conversion is usually
less than  5% of the S02 in the flue gas stream.
     The biggeat  problem seems to  be deposition of (NH,)2SO, and NH.HSO, on
the  air preheater.  These compounds are highly corrosive and interfere with
heat transfer.   The problem appears to be most severe with high sulfur oil
 firing. With low sulfur oils, the SO. concentration is too low to cause
problems.   Early tests with coal indicate that the (NH.KSO, and NH.HSO, may
deposit on the fly ash or be removed from the heat exchanger surface by the
erosive action of the fly ash.  However, more operating experience on coal-
fired sources is needed to quantify the extent of the problem.
      Numerous countermeasures have been proposed to minimize the (NH,)2SO,
and  NH.HSO, problem.  The most desirable techniques are to avoid formation.
This can be accomplished by reducing the SO- and NH- in the reactor effluent
or increasing the exhaust temperature of the flue gas.  Soot blowing and
water washing techniques are useful in removing the deposits after formation.
Heat exchanger modifications, which increase the effectiveness of the removal
techniques or minimize the likelihood of deposit, are also being considered
as well as use of corrosion resistant heat exchanger surfaces.
     There are numerous developers of SCR processes and numerous pilot,
prototype, and commercial scale applications.  Table I lists 13 process
                                             3
developers and 62 applications over 10,000 Nm /hr  (normal cubic meters per
hour) along with information on the type of source, fuel, plant size, type
of reactor, and start-up date.
                                    252

-------
     The cost of SCR systems varies with process developer and application.
However, the Japan Environment Agency has investigated the costs of a SCR
application to a new oil-fired, 300 MW installation in Japan.  To remove 80%
of the NO , the capital investment was estimated to be about $15.5/kW and
         2£
the annualized cost, about 1.6 mills /kWh.
     EPA is sponsoring a pilot scale evaluation of SCR technology which will
be discussed later in the paper.

Wet NO  Processes (Reference 3) —
      2£
     The wet NO  processes developed to date cannot compete economically
               jt
with dry SCR processes for removal of NO  from combustion flue gas.  This is
                                        3t
primarily due to the complexity, limited applicability, and water pollution
problems associated with the wet NO  processes.
                                   j£
Dry Simultaneous NO /SO  Processes (Reference 3) —
                   A   A.
     Although there are several dry simultaneous NO /SO  removal systems,
                                                   3C   X.
the only commercially demonstrated system is the Shell Flue Gas Treating
system.  This process was originally designed for S02 control but was also
found to be adaptable for NO  control.
                            2k
     The process uses copper oxide supported on stabilized alumina placed in
two or more parallel passage reactors.  The reactions can be expressed as
follows :
        CuO + 1/2 02 + S02  - >   CuS04                (3)

        4NO + 4NH  + 0      - ^^ - »•   4N  + 6H0           (4)
        CuS04 + 2H2         - »•   Cu + S02 + 2H20      (5)
        Cu + 1/2 02         - >•   CuO                  (6)
     Flue gas is introduced at  400° C  into  one of the  reactors where  the  S0x
 reacts with  copper oxide to form copper sulfate.   The copper sulfate and,  to
 a  lesser extent, the copper oxide act as catalysts in the reduction  of N0x
 with ammonia.   When  the reactor is saturated with  copper sulfate,  flue gas
 is switched  to  a fresh reactor  for acceptance of the  flue gas,  and the spent
 reactor  is regenerated.  In the regeneration cycle, hydrogen is used to
 reduce the copper sulfate to  copper,  yielding a S02 stream of  sufficient

                                     253

-------
 concentration  for conversion to sulfur or sulfuric acid.   The  copper  in  the
 reactor is  oxidized,  preparing the reactor for acceptance  of the  flue gas
 again.   Between acceptance and regeneration,  steam is  injected into the
 reactor to  purge the remaining flue gas or hydrogen to eliminate  any  pos-
 sibility of combustion.   The process can also be operated  in the  NO -only
                                                                   2t
 mode by eliminating the  regeneration cycle, or in  the  SO -only mode by
                                                        X
 eliminating the ammonia  injection.
      The process has been installed at the Showa Yokkaichi Sekiyu plant  in
 Yokkaichi by Japan Shell Technology, Ltd. on  a heavy-oil-fired boiler treating
          3
 120,000 Nm  /hr of flue gas.  In this installation,  the process removes 90%
 of the  SO,  and 40% of the NO  to meet local regulations.   The  unit has
          £*                   &
 demonstrated 90% SO. removal and 70% NO  removal.
                    £,                   Jt
      UOP Process Division is the licensor of  the process in the United
 States.  Hence, the process will be referred  to as  the  UOP-She 11  process in
 this paper. EPA is sponsoring a pilot scale  evaluation of the UOP-Shell
 process which  will be discussed later.
 Wet Simultaneous NO /SO   Processes (Reference 3)—
                    Ji  A
      Although  the wet NO  removal processes cannot  compete economically  with
 dry NO   processes, wet simultaneous  NO /SO  processes may  be competitive
      X                               XX
 with the sequential installation of  NO control by  SCR  followed by S0»
                                       A                              te
 control by  flue gas desulfurization  (FGD).
      The first wet simultaneous  NO /SO  systems, called oxidation/absorption/
 reduction processes,  evolved from FGD  systems.  Since the  NO in flue  gas is
 fairly  insoluble in aqueous solutions, a gas-phase oxidant is  injected
 before  the  scrubber to convert NO to the more soluble nitrogen dioxide
 (N02).   The absorbent varies with the type of FGD system being modified.
 The absorbed S02 forms a sulfite  ion which reduces a portion of the absorbed
 NO  to molecular nitrogen.   The remaining NO  are removed  from the waste
                                            x
water as nitrate  salts.  The remaining sulfite ions are oxidized  into
sulfate by air. and removed as gypsum.
     The oxidation/absorption/reduction processes have the potential to
remove 90% of both SO  and NO  from combustion flue gas.  However, there
                     X.       X
several drawbacks remaining to be overcome before the processes can be
                                    254

-------
widely applied.  The process chemistry is complex, and the use of a gas-
phase oxidant, such as ozone (0_) or chlorine dioxide (C10-), is expensive.
Chlorine dioxide, although cheaper than ozone, adds to the waste water
problems created by the nitrate salts.
     Absorption/reduction processes were seemingly developed to avoid the
use of a gas-phase oxidant.  A chelating compound, which has an affinity for
the relatively insoluble NO, is added to the scrubbing solution.  Ferrous-
EDTA (ethylene-diamine tetraacetic acid) has typically been used as the
chelating compound.  The NO is absorbed into a complex with the ferrous ion
and the SO- is absorbed as the sulfite ion.  The NO complex is reduced to
molecular nitrogen by reaction with the sulfite ion.  A series of regeneration
steps recovers the ferrous chelating compound and oxidizes the sulfite ion
into sulfate which is removed as gypsum.
     The absorption/reduction processes also have the potential to remove
90% of both the NO  and SO  in combustion flue gas.  Although the processes
                  X       X
seem to have advantages over the oxidation/absorption/reduction processes,
there are obstacles to be overcome before the processes can be widely applied.
Even with the addition of the chelating compounds, a large absorber is
required to absorb the NO.  The sum of replacement, recovery, and regeneration
costs of the chelating compounds, although potentially less than using the
gas-phase oxidants, are still significant.  The process chemistry is complex
and is sensitive to the flue gas composition of S02> NO , and oxygen.  The
molar ratio of S09 to NO  must remain above approximately 2.5 and the oxygen
                 £      Ji
concentration must remain low.
     Table 11 lists the process developers evaluating wet simultaneous
NO /SO  technology in Japan.  Oxidation/absorption/reduction processes are
  3t   X
being investigated by six process developers at nine different sites.  For
absorption/reduction processes, there are four pilot or bench scale plants
currently being operated by four different process developers.
     As shown in Table III, based on information from the Japan Environment
Agency, it was estimated that a FGD system would have an investment cost of
about $77.8/kW with an annualized cost of 6.7 mills/kWh.  As stated previously,
                                    255

-------
 the  cost of a  SCR system is  $15.5/kW and 1.6 mllls/kWh.  However, the  cost
 of a sequential system employing  SCR + FGD would be about $102.8/kW and  8.3
 mills/kWh.  The total  cost increased slightly over the sum of individual SCR
 and  FGD systems to  allow for removal of NH« from the FGD scrubbing liquor.
      The cost  of simultaneous NO  /SO , both wet and dry, was estimated to be
                                X  X
 slightly higher or  about $lll.l/kW and 9.4 mills/kWh; however, these cost
 estimates were more uncertain than the FGD and SCR costs.  Costs for all
 systems were  estimated to be about 20% higher for coal-fired sources.
 Assessments of Technology for U.S. Application
      The Tennessee  Valley Authority (TVA), through an interagency agreement
 with EPA, is developing comparative economics of NO  and NO /SO  flue  gas
       *                                            X       X   X
 treatment processes for application in the U.S.   This multi-phase study is
 over 50% complete.   Phase I, a technical assessment, was an evaluation and
 summary of  the technical feasibility of all candidate NO  control processes
                                                        X
 being offered in the U.S. and Japan.  Phase II, a preliminary economic
 assessment, concentrated on  the processes recommended in Phase I for further
 study.  Phase III,  a definitive economic assessment, will develop detailed
 engineering evaluations and  cost  estimates for the most promising NO
                                                                    **
 control cases identified in  Phase II.  In addition, the impact of application
 of  SCR processes on ammonia  availability and cost was evaluated.  Phases I
 and  II of the  study were cofunded by the Electric Power Research Institute
 and  EPA.                                                "
 Phase I —  Technical Assessment of NO  and NO /SO  Processes (Reference  4) —
                                     2t       X   XI
      In Phase  I,  42 processes for NO  and simultaneous NO /SO  control were
                                    X                    XX
 described.  The discussion on each process included the process description,
 status of development,  reported economics, utility and raw material require-
 ments,  technical and environmental considerations, and advantages and  disad-
 vantages.
      In addition, a  comparison was made of the dry NO  and wet simultaneous
       *                                             A
NO /SO  systems.  It was found that, although there are many different types
  *t   x^
of dry and wet processes, in most  cases the dry processes have advantages
over the wet processes:
                                    256

-------
           o     Low projected capital  investment  and  annual  revenue
                req uirement s.
           o     Simple  process with  few equipment requirements.
           o     High NO  removal  efficiency  (>90%).
                       2£
           o     Extensive tests in large units.
           o     No  waste  stream generation.
     However,  the  dry  systems also  have disadvantages:
           o     Sensitive to  inlet particulate levels.
           o     Require ammonia.
           o     Possible  emission of ammonia and  ammonium  sulfates and
                bisulfates.
           o     Relatively high reaction temperatures  (350-400°C).
     The wet NO /SO removal  processes  have certain general advantages and
disadvantages,  compared  with  the dry NO systems.  The major advantages
                                        Jv
include:
           o     Potential economic advantage of simultaneous NO /SO
                removal.                                       x   x
           o     Relatively insensitive to flue gas particulates.
           o     High S0_  removal  (>95%).
      On the other  hand,  major disadvantages of  the wet systems include:
           o     Expensive processes  due  to process complexity and insolubility
                of NO   in aqueous solutions.
                    2S
           o     Formation of nitrates  (NO- ) and  other potential water
                pollutants.
           o     Extensive equipment  requirements.
           o     Formation of low-demand  byproducts.
           o     Flue gas  reheat required.
           o     Only moderate NO  removal.
                  J            x
           o     Limited application  of some processes due  to requirement
                for high  SO :NO  ratios.
                          X   X
     In addition to being a state-of-the-art review of all NO  processes
undergoing development,  one of the main purposes of the study was to identify
processes  for further  evaluation in Phase II of  the study.  Three criteria
were used  to screen the  processes:  technical considerations, developmental
                                     257

-------
status, and representative samples of available technologies.  Based on
these criteria, the nine processes in Table IV were selected for further
study.
Phase II — Preliminary Economic Estimates (Reference 5)—
     The primary purpose of Phase II of the study was to compare the process
economics of the different types of NO  and NO /SO  control techniques on a
                                      X       JL   X.
consistent basis.  Preliminary economics, including total capital investment
and average annual revenue requirements, were determined.  Rigorous compari-
sons were not made although various conclusions can be drawn from the analysis.
     The processes listed in Table IV were evaluated except for the wet NO -
only process; development of that process was terminated by the developer.
However, there appears to be little, if any, future for application of wet
NO -only processes to combustion sources. The JGC Paranox process, a dry SCR
  JW
process  (NO  only) employing a parallel passage reactor, was substituted.
           A.
     Process descriptions, detailed flowsheets, material balances, and a
current  status of development summary were prepared for each process based
on data  supplied by  the process vendors.  Equipment descriptions and costs
were prepared from the available data for each process.
     Although the study is not complete, preliminary estimates are indicated
in Table V.  (It must be emphasized that these numbers are subject to change
when the study is completed.)  However, it is apparent from these estimates
that the IHI process, an ozone based system,  is not economically attractive.
(This finding is consistent with an earlier EPA study (Reference 6) that
determined that ozone systems were too expensive for coal-fired applications.)
The Moretana Calcium  process, a chlorine dioxide based system, and the Asahi
process, an EDTA based system, also appear economically unattractive when
compared to the least expensive SCR + FGD system,  especially based on annual
revenue requirements.
     Given the uncertainties of the estimates,  the comparison of the UOP-
Shell dry NO /SO  process and the SCR + FGD system is inconclusive.  In
addition, it is  not  apparent which SCR system,  moving bed or parallel flow,
represents the optimum NO -only configuration.   The cost estimates for SCR
        '
                                    258

-------
processes include the cost of particulate control by an ESP.  The ESP is
roughly 40% of the capital cost, which places NO  control alone at approxi-
                                                X
mately $36/kW.
     The study also evaluated energy consumption of the processes, and
estimates are given in Table VI.  The SCR systems are projected to require
less than 0.3% of the boiler capacity.  The UOP-Shell, dry NO /SO  process
                                                             X   X
requires the least energy of any NO  and SO  system due to reclamation of
                                   3C       X
heat from the byproduct streams.  The combination of SCR and FGD is next;
the wet NO /SO  systems require significantly higher quantities of energy.
          A   X.
     The economic sensitivity of the SCR systems to the cost of ammonia has
been the subject of some concern.  It was found that the annual ammonia cost
was below 12% of the average annual revenue requirements for the dry systems.
Therefore, the average annual revenue requirement may be expected to increase
about 10% if the ammonia cost is doubled.
Phase III — Definitive Economic Estimates—
     In Phase III of the study, to begin immediately following completion of
Phase II, a definitive-level design and economic evaluation of the leading
NO  control techniques will be performed.  The study will encompass both
  X.
flue gas treatment and combustion modification techniques, which will be
analyzed individually as well as in combination.  This will enable a compari-
son of 90% NO  control by flue gas treatment alone and combustion modification
             3t
followed by flue gas treatment.  In addition, dry simultaneous NO /SO
                                                                 X   X
systems will be evaluated against the optimum highly efficient NO  control
                                                                 X
system used in conjunction with a conventional FGD unit.
     In the course of the study, the bases for conceptual designs will be
identified, a definitive engineering package will be prepared, and capital
investment and revenue requirements will be estimated.  The final report on
Phase III is scheduled to be available in early 1980.
Impact of Ammonia Utilization (Reference 7)—
     A major concern of the widespread utilization of ammonia-based NO
                                                                      2v
control systems has been the impact on the domestic ammonia market.  It has
been speculated that severe disruptions in the availability and price of
                                    259

-------
ammonia would occur if SCR systems are employed; this could also impact on
the availability and cost of ammonia-based fertilizers.  In addition, the
impact on natural gas, the primary feedstock for ammonia production, was
unclear.  A  study of the situation was undertaken by TVA to address these
concerns in  parallel with the three-phase technical and economic evaluation
discussed above.
     The annual NH_ requirements  for a 500 MW coal- and oil-fired boiler
were calculated for application of a 90% NO  removal system based on SCR
                                           Ji
technology.  At an assumed NH,:NO mole ratio of 1.05:1, the NH, consumption
                             «J    3t                             J
rates were 5.7  Gg/yr (6300 tons/yr) and 1.6 Gg/yr (1800 tons/yr) for the
coal- and oil-fired units, respectively.
     From these calculations it is apparent that even a large coal-fired
power plant  could not justify a, captive NH_ plant since the minimum plant
capacity for economic production  of NH« is about 300 Gg/yr (330,000 tons/yr).
Therefore, NH«  would be purchased from an off-site NEL plant and stored in
tanks at the power plant.
     • The projected U.S. electrical generating capacity was estimated for the
period  from  1975 to 2000.  Assuming application of a SCR system removing 90%
of the NO  to all new large industrial arid 'utility boilers (>260 GJ/hr or
        gf
250 x 10  Btu/hr) beginning in  1985, the NH, demand for NO  control was
determined.  The ammonia supply to meet both the conventional NH, demand and
annual growth of this demand for  NO  control was projected using an assumed
annual growth rate for ammonia  production of 3.0%.  It was found that the
NH- demand for  NO  control ranged from 0.6% of the total NH, supply in 1985
   OX                                         j
to 20-25% in the year 2000.
     Thus, although requiring substantial amounts of NH, in the future, the
need for NH~ for NO  control apparently would not have an abrupt adverse
impact on the availability and  price of NH_.  Under the assumptions of the
study, the primary Impact on the  domestic NH, market would be to cause the
U.S. NH_ demand to increase at  4.5%/yr during the period 1985-2000 rather
than the assumed 3.0%.   This increase could be met with the addition of one
                                     260

-------
ammonia plant per year.  In addition, although use of natural gas is counter
to the National Energy Plan, utilization of NH~ for NO  control should not
                                              •J       X
adversely affect the price ,and availability of natural gas.
Assessment of Technology for an Industrial Boiler NSPS
     IERL-RTP is sponsoring a series of studies to determine the applica-
bility of various emission control technologies to industrial boilers.  The
primary purpose of these studies is to support the development of a new
source performance standard (NSPS) for industrial boilers.  The following
technologies are being considered:
          o    Oil Cleaning and Existing Clean Oil
          o    Coal Cleaning and Existing Clean Coal
          o    Synthetic Fuels
          o    Fluidized Bed Combustion
          o    NO  Combustion Modification
                 x
          o    NO  and NO /SO  Flue Gas Treatment
                 X       XX
          o    Flue Gas Desulfurization
          o    Particulate Control
     Following completion of reports on the individual technologies, a
comprehensive assessment report will be prepared which discusses how the
technologies should be integrated for optimum multi-pollutant control.
Subsequently, EPA's Office  of Air Quality Planning and Standards  (OAQPS)
will study the impact  of various  control options.  These  studies will
ultimately lead to the establishment of a NSPS for industrial boilers.
Acurex Corporation is  the  contractor assisting OAQPS  and  IERL-RTP  in  this
overall effort.
     As shown above, NO  and NO /SO  flue gas treatment processes  are among
                       X       XX
the technologies being considered.  Radian Corporation is preparing the
technology assessment  report on these processes.  The report will  contain
the following sections:
          o    Emission Control Techniques
          o    Candidates  for Best  System of  Emission Reduction
          o    Control System Costs
                                     261

-------
          o    Energy Considerations
          o    Environmental Considerations
          o    Emission Source Test Data
     The cost, energy, and environmental considerations will be based on
standard boilers listed in Table VII.  The coal-fired boilers will be
analyzed for high and low sulfur Eastern coal and low sulfur Western coal
cases.
     The final report on the assessment of NO  and NO /SO  technologies
                                             3x       A.   X,
should be available  in the spring of  1979.  The comprehensive assessment
report on optimum multi-pollutant control systems should be available in the
Fall of 1979.  (Note that consideration of a technology for a standard does
not imply that the ultimate standard  will be based on that technology.)

CONTROL STRATEGY ASSESSMENT STUDIES
     While  the primary purpose of the technology assessment studies is to
determine the performance of the technology, the primary objective of the
control strategy assessment studies is to determine if and when the tech-
nology will be needed in the U.S.  Recent studies in this area have utilized
computer modeling  techniques to evaluate alternative control strategies and
investigated the issues  (such as prevention of significant deterioration)
which could eventually require increased NO  control.
                                           X
Assessment  of Point  Source Impact on  Ambient N0? Levels (Reference 8)
     A study was undertaken for EPA by Radian Corporation to determine the
impact of various stationary source NO control strategies on attaining and
                                       JL
maintaining the National Ambient Air  Quality Standard (NAAQS) for N0_.  The
Chicago Air Quality  Control Region (AQCR) was selected for use in this study
since it has historically encountered ambient NO  problems.
                                                3£
Annual Impact—
     The original purpose of the study was to determine the effect on
annual average ambient NO- levels of  applying NO  control technology to
                                            6   ^
large point  sources  (> 105 GJ/hr or 100 X 10  Btu/hr) in the AQCR.  A
                                    262

-------
dispersion model was used to relate NO  emissions to ambient NO- concentrations
                                      3C                        £•
in Chicago.
     It was found that* although the major point sources account for nearly
40% of the total NO  emissions in Chicago, they account for less than 10% of
                   Jv
the ambient N0_ levels, on the average.  Considering major point source
"hotspots" (i.e., localized areas of the city where major point source
impact is the greatest), modeling results indicate that these sources account
                                                 o
for 12% of a predicted N02 level of about 60 yg/m .  Taking a worst-case
approach and assuming that all NO  emissions from major point sources are
                                 3£
converted to NO., it was found that the predicted cumulative impact of all
major point sources at locations of maximum annual impact is still only 15%
of the standard.
     Therefore, it was concluded that total removal of large point source
NO  emissions would result in only a small improvement in annual average NO-
  X                                                                        fc
air quality in Chicago, and control of large point sources alone would not
be adequate to achieve and maintain the annual average NAAQS.
Short-term Impact—
     Based on the Clean Air Act Amendments of 1977 and the possible estab-
lishment of a short-term NAAQS, this Radian study was expanded to investi-
gate the effect of NO  control techniques on short-term ambient concentra-
                     JH
tions in the AQCR for present and future years.  Standards were assumed of
                            3
250, 500, 750, and 1000 ug/m  of N02 based on a 1 hour average.
     The short-term impact assessment was made using Gaussian-type disper-
sion models.  A significant part of the effort of this study was directed
toward defining the short-term NO  emission rates that should be used in the
                                 2£
model.  This was done primarily by adjusting the annual average emission
rates.  Adjustments were made for season of the year, day of the week, and
time of day.  The entire emissions inventory for the Chicago AQCR, including
vehicular and other area sources, was modeled using this approach.
     The computer-predicted ambient NO  concentrations were converted to
                                      2t
ambient NO- concentrations by applying a ratio of NO- to NO  determined from
          2                                         fc      X
measured air quality data in Chicago.  This ratio is a function of season of
                                    263

-------
the year and time of day; ratios used in the study varied from 0.25 to 0.5.
The accuracy of this approach is not known since several photochemical
reactions are  involved in the conversion of NO  to N0_ and other species.
                                              2£      fc
For assessment of future year impacts, the growth in NO  emissions was
                                                       it
estimated.
      The study found that individual large point sources may account for 60%
                                                    3
of a  predicted 1 hour N0? concentration of 1100 yg/m  in industrial areas
                              3
and 90% of  a level of 800 yg/m  in non-industrial areas.  This indicates
that  controlling large point sources may provide significant improvements in
short-term  N02 air quality.  However, the degree of control required is
highly dependent on any short-term N0? NAAQS adopted by EPA.  The results
summarized  in  Table VIII show the percentage of the 14 largest existing
point sources  which would require controls for various standard levels if
those standards were currently in effect.  It was assumed that the more cost
effective  combustion modification techniques would be employed first, and
that  the flue  gas  treatment techniques would only be applied if necessary to
maintain the established ambient level.
      The percentages of Table VIII are based on individual large point
source  impacts added to the impacts of other point sources, vehicular sources,
and non-vehicular  area sources.  When large point sources are located near
each  other  so  that their Impacts interact, the degree of control required
increases significantly and more flue gas treatment is required.
      Projections for NO  emissions to 1985 indicate that the same level of
                       JH
control as  shown in Table VIII would be required.  There are two reasons for
this  unexpected result.  First, the highest predicted short-term concentra-
tions are dominated by large point sources to the extent that changes in the
impacts from other sources do not make a large difference.  Second, the
change in impact of other sources by 1985 is small because increases in non-
vehicular emissions are counterbalanced by the decrease in projected vehicular
emissions.
     It  is  concluded that control of large point source NO  emissions.would
                                                          X
result in a significant improvement in short-term N02 air quality and may be
necessary to attain and maintain compliance with a short-term NAAQS.

                                   264

-------
Assessment of the Need for NO  Flue Gas Treatment Processes (Reference 9)
^""^™—^«^_«H^_^^«^__^«^^^^^__^__
     To date, EPA's strategy for controlling impacts of NO  emissions has
                                                          3t
focused on combustion modification technology.  This approach has been taken
since these techniques represent the most cost effective approach to achieving
initial reductions in NO  emissions.  However, it is uncertain whether the
                        2£
NO  emission reductions attainable by use of combustion modification techniques
alone can continue to provide the margin of control necessary to meet NO
                                                                        3£
ambient air quality standards.
     In parallel with the Chicago study discussed above, another study was
undertaken by Radian Corporation to determine if and when the application of
NO  flue gas treatment technology would be necessary in the U.S.  The report
  X
of this study (Reference 9) addressed factors which will influence the
levels of NO  emission control needed to comply with both existing and
            j£
future NO  standards.  Topics treated include NO  emission sources, nationwide
         Ji                                      2t
trends, regional emission profiles, and atmospheric transport and reactions
of NO .  Also addressed were current NO  regulations and trends in NO
     X                                 X   °                         X
legislation:  National Ambient Air Quality Standards including the possible
short term N0_ standard, New Source Performance Standards, Mobile Source
Standards, Prevention of Significant Deterioration, and Nonattainment Provi-
sions.  In addition, other uncertainties were assessed such as the oxidant
problem, health effects research, and the nitrosoamines issue.  Further, the
major NO  emission control alternatives — control of mobile sources and
control of stationary sources by combustion modification and flue gas treat-
ment — were evaluated.
     The study concluded that the number of AQCRs with NO  compliance problems
                                                         2k
can be expected to increase significantly in the next decade.  It was further
concluded that progressively larger reductions in NO  emissions will be
                                                    2£
required in order to attain and maintain compliance in "problem" AQCRs.  The
study does not establish conclusively whether or not flue gas treatment
technology will be required.  However, current trends indicate that the
technology may be necessary in the future to achieve compliance with NO
                                                                       X
standards in specific AQCRs.  This conclusion follows from the regionally
specific nature of U.S. NO  compliance problems as well as uncertainties
                                     265

-------
concerning both future NO  emission reduction requirements and the ultimate
effecti
cation.
effectiveness of alternative NO  control methods such as combustion modif i-
                               x
EXPERIMENTAL PROJECTS
      The  technology assessment and  control strategy studies and financial
constraints aid  in determining the  number and scale of experimental projects.
These studies, as well as previous  experimental projects, supported a jump
to pilot  scale evaluations of NO  and NO /SO  processes on a coal-fired
                                3t       j£   2£
application.  Previous experimental projects included laboratory projects on
development of catalysts for NO  reduction with ammonia (References 10, 11),
                               m
a pilot-scale project evaluating a  selective catalytic reduction process on
a gas-fired application  (Reference  12), and a laboratory evaluation of NO
                                                                         2C
reduction with metal  sulfidea  (Reference 13).
      After solicitation  and evaluation of proposals from interested parties,
EPA  awarded two  contracts  in May 1978 for the pilot-scale evaluation of flue
gas  treatment technology on a  coal-fired source.  One contract went to
Hitachi Zosen for demonstration of  NO removal and the other went to UOP
Process Division for  demonstration  of simultaneous NO /SO  removal.
                                                     x   x
      Both pilot  plant projects have similar objectives and scope.  The basic
objective is to  demonstrate the feasibility of the processes for highly
efficient control of NO  or NO /SO   on a coal-fired source.  Pollutant
                       Jt     2*.   Jt
removal efficiencies of  90% are expected.
     The  scope of the projects is divided into four phases:
          I    -    Design
          II    -    Procurement and erection
        III    -    Startup, debugging, and optimization
          IV    -    Long term operation and assessment.
     The design phases were completed within approximately 3 months of
contract award on both projects.  Procurement and erection, Phase II, is
expected to  require about 9 months  to  complete.  Phase III, which will
require 4-8 months  to complete with the majority of time spent on optimiza-
                                    266

-------
tion testing, will contain a parametric study of process variables including:
flow rate, temperature, NH.:NO ratio, and NO , SO™, S0~, and particulate
concentrations.  Long-term operation and assessment will be conducted during
a 90 day continuous run in Phase IV.
     Project manuals will be available on both projects in early 1979.  The
final reports on the results of the projects should be available in mid-
1980.
Hitachi Zosen NO  Pilot Plant (Reference 14)
     Hitachi Zosen1s NO  removal process is based on SCR technology.  Since
these processes were discussed previously in the paper, only key process
features will be discussed further.
     Hitachi Zosen utilizes a metallic honeycomb catalyst arrangement for
applications with high particulate and SO  concentrations.  The reactor is
located between the boiler economizer and the air preheater and in front of
any particulate control device.  The reaction temperature is in the range of
350-420°C.  The pressure drop is about 200-500 Pa.  For 90% NO  reduction, a
NH3:NO mole ratio of 1:1 is used.
     The process will be evaluated on a pilot plant scale (^2000 Nm /hr) on
a coal-fired source.  The host site for the pilot plant will be Georgia
Power Company's Plant Mitchell near Albany, Georgia.  Flue gas will be
obtained from Unit 3, a pulverized coal-fired Combustion Engineering boiler
with a nameplate rating of 125 MW.
     Envirotech/Chemico Air Pollution Control Corporation (CAPCC), Hitachi
Zosen1s American licensee, will be the major subcontractor on the project.
CAPCC will provide detailed engineering, design, procurement, erection, and
operation of the pilot plant in cooperation with Hitachi Zosen.  Fabrication
and procurement will be done within the U.S. as far as possible.
UOP-Shell NO /SO  Pilot Plant  (Reference 15)
  ^^^^^^^"™"^™^^^™~"  J^^^^^^^^^^^"^^^^^™^^^"*^"!^™^^^—
     The UOP-Shell dry NO /SO  removal system is based on the Shell Flue Gas
Treating process which has also been discussed previously in the paper.
Briefly, the process employs a parallel passage reactor using CuO as the
                                    267

-------
sorbent for SO- and CuSO, as the catalyst for the reduction of NO  with NH-.
              fc         f                                        X        J
The reactor is located between the boiler economizer and the air preheater
and in front of any particulate control device.  The process operates at a
temperature of 400°C during both acceptance and regeneration cycles.  The
cycles will be controlled for 90% removal of both NO  and S09.  For 90% NO
                                                    X       £•             X
reduction, a NH,:NO mole ratio of 1:1 is used.
                                                                   2
     The process will be evaluated on a pilot plant scale (^2000 Nm /hr) on
a coal-fired source.  An existing pilot plant, previously used by UOP to
evaluate SO. removal, will be modified for simultaneous removal of NO  and
           t~                                                  "      X
SO-.  The host site for the pilot plant will be Tampa Electric Company's Big
Bend Station in North Ruskin, Florida.  Flue gas can be obtained from Unit 1
or 2, which are Riley-Stoker pulverized coal-fired boilers with a nameplate
rating of 400 MW each.
                                    268

-------
                                CONCLUSIONS
     This paper has summarized the status of NO  and NO /SO  flue gas
                                               j£       Jt   J\
treatment technology and presented the results of recent EPA studies of the
technology.  The following conclusions can be drawn.
1.   Dry NO  processes, based on selective catalytic reduction (SCR) of NO
           X                                                              X
     with ammonia, can remove 90% of the NO  from the flue gas of combustion
                                           A
     sources.
2.   SCR processes have been extensively demonstrated on commercial-scale
     gas- and oil-fired sources in Japan.
3.   Based on the experience in Japan, SCR processes can be expected to be
     successfully applied to gas- and oil-fired sources in the U.S.
4.   Reactors and catalysts for SCR systems have been developed to tolerate
     high SO  and particulate concentrations and are ready for test applica-
            j£
     tion on coal-fired sources in the U.S. and Japan.
5.   The formation of ammonium sulfate and bisulfate remains a problem for
     SCR systems, but solutions are being investigated.
6.   Wet processes cannot economically compete with SCR processes for
     control of NO .
                  
-------
10.  The widespread application of  SCR processes  is not expected to have an
     abrupt adverse impact  on the domestic  ammonia market.
11.  NO  and NO  /SO   flue gas treatment processes warrant consideration as
       X       X  X
     the basis for a  New Source Performance Standard for industrial boilers.
12.  Large point sources do not have  a significant impact on annual average
     ambient NO- levels.
13.  Large point sources may have a significant impact on short term ambient
     NO- levels.
14.  NO  flue gas treatment processes may be needed to attain and maintain
       X
     compliance  with  EPA ambient standards  in certain Air Quality Control
     Regions.
15.  EPA's pilot plant project with Hitachi Zosen will enable evaluation of
     a SCR process for 90%  control  of NO  on a coal-fired source.
16.  EPA's pilot plant project with UOP Process Division will enable evalua-
     tion  of a dry simultaneous NO  /SO  process for 90% control of both NO
                                  A  A,                                   X
     and S0_ on  a coal-fired source.
17.  EPA experimental projects, in  conjunction with technology assessment and
     control strategy studies, will enable an assessment of the feasibility of
     NO  and NO  /SO   flue gas treatment processes for application in the U.S.
       X       XX
                                    270

-------
                                 REFERENCES
1.   Ando, Jumped., Katsuya Nagata, and B. A. Laseke.  NO  Abatement for
     Stationary Sources in Japan.  PEDCO Environmental, Inc., EPA-600/7-77-
     103b (NTIS No. 276-948), September 1977.  U.S. Environmental Protection
     Agency, Research Triangle Park, N.C.

2.   Ando, Jumpei.  SO., Abatement for Stationary Sources in Japan.  EPA-
     600/7-78-210, November 1978.  U.S. Environmental Protection Agency,
     Research Triangle Park, N.C.

3.   Ando, Jumpei, and Katsyua Nagata.  NO  Abatement for Stationary Sources
     in_Japan.  (Draft Report; to be published January 1979 by EPA.)

A.   Faucett, H. L., J. D. Maxwell, and T. A. Burnett.  Technical Assessment
     of NO  Removal Processes. Tennessee Valley Authority, EPA-600/7-77-127
     (NTIS^No. 276-637, EPRI No. AF-568, TVA No. Y-120), November 1977.
     U.S. Environmental Protection Agency, Research Triangle Park, N.C.

5.   Burnett, T. A., J. D. Maxwell, and H. L. Faucett.  The Preliminary
     Economics of Alternative NO  Flue Gas Treatment Processes.  Tennessee
     Valley Authority. (12/78 Draft Report; to be published early 1979 by
     EPA and EPRI.)

6.   Harrison, J. W.  Technology and Economics of Flue Gas NO  Oxidation
     by Ozone.  Research Triangle Institute, EPA-600/7-76-033  (NTIS 261
     917), December 1976.  U.S. Environmental Protection Agency, Research
     Triangle Park, N.C.

7.   Burnett, T. A., and H. L. Faucett.  Impact of Ammonia Utilization
     by NO  Flue Gas Treatment Processes.  Tennessee Valley Authority,
     EPA-600/7-79-011, January 1979.uTs. Environmental Protection
     Agency, Research Triangle'Park, N.C.

8.   Eppright, B. R., E. P. Hamilton, III, M. A. Haecker, and Carl-Heinz
     Michelis.  Impact of Point Source Control Strategies on N02 Levels.
     Radian Corporation, EPA-600/7-78-212, November 1978.  U.S. Environmental
     Protection Agency, Research Triangle Park, N.C.

9.   Corbett, W. E., G. D. Jones, W. C. Micheletti, R. M. Wells, and G. E.
     Wilkins.  Assessment of  the Need for NO  Flue  Gas Treatment Technology.
     Radian Corporation, EPA-600/7-78-215, November 1978.U.S. Environmental
     Protection Agency, Research Triangle Park, N.C.


                                    271

-------
10.  Nobe, K., G. L. Bauerle, and S. C. Wu.  Parametric Studies of Catalysts
     for NO  Control from Stationary Power Plants.  University of California,
     Los Angeles, EPA-600/7-76-026  (NTIS No. PB 261 289), October 1976.
     U.S. Environmental Protection Agency, Research Triangle Park, N.C.

11.  Koutsoukos, E. P., J. L. Blumenthal, M. Ghassemi, and G. L. Bauerle.
     Assessment of Catalysts for Control of NO  from Stationary Power Plants.
     TRW, Inc., EPA-650/2-75-001a (NTIS No. PB£239 745), January 1975, U.S.
     Environmental Protection Agency, Research Triangle Park, N.C.

12.  Kline, J. M., P. H. Owen, and Y. C. Lee.  Catalytic Reduction of Nitrogen
     Oxides with Ammonia:  Utility Pilot Plant Operation.  Environics, Inc.,
     EPA-600/7-76-031 (NTIS No. PB 261 265), October 1976.  U.S. Environmental
     Protection Agency, Research Triangle Park, N.C.

13.  McCandless, F. P., and Kent Hodgson.  R^duc^ion oj^_NitricJPxide with Metal
     Sulfides.  Montana State University, EPA-600/7-78-213, November 1978.
     U.S. Environmental Protection Agency, Research Triangle Park,  N.C.

14.  Wiener, R. S., and Rafat R. Morcos, Project Manual;  Evaluation of Hitachi
     Zosen NO  Flue-Gas Treatment Process.  Chemico Air Pollution Control
     Company.  (Draft Report; to be published January 1979 by EPA.)

15.  Pohlenz, J. B., and Nooy, F. M.  Project Manual;  Evaluation of UOP-
     Shell NO^/SO  Flue Gas Treatment Process.  UOP Process Division.
     (Draft Report; to be published in January 1979 by EPA.)
                                    272

-------
                                 TABLE I.   SELECTIVE CATALYTIC REDUCTION PLANTS IN JAPAN
ro
^i
CO
Process Developer
Asahi Glass
Hitachi, Ltd.
Hitachi, Ltd.
Hitachi, Ltd.
Hitachi, Ltd.
Hitachi, Ltd.
Hitachi, Ltd.
Hitachi, Ltd.
Hitachi, Ltd.
Hitachi, Ltd.
Hitachi, Ltd.
Hitachi, Ltd.
Hitachi Zosen
Hitachi Zosen
Hitachi Zosen
Hitachi Zosen
Hitachi Zosen
Ishikawaj ima H.I.
Ishikawaj ima H.I.
Ishikawaj ima H.I.
Ishikawaj ima H.I.
Ishikawaj ima H.I.
Ishikawaj ima H.I.
Ishikawaj ima H.I.
JGC
JGC
JGC
Kobe Steel
Kurabo
Mitsubishi H.I.
Mitsubishi H.I.

Gas
Source
Furnace
Coke Oven
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Furnace
Furnace
Furnace
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Furnace
Boiler
Coke Oven
Coke Oven
Boiler
Boiler
Boiler

Fuel3
BIBTT,
COG
HO(HS)
Kerosene
HO(LS)
HO(LS)
Crude Oil
LNG
HO
Kerosene
HO(LS)
HO(LS)
CO
HO(HS)
Kerosene
Coke, Oil
HO
Crude Oil
HO(LS)
HO(LS)
HO(LS)
Crude Oil
HO(LS)
HO(LS)
HO (MS)
CO
COG
COG
HO(HS)
HO(LS)
LNG

Capacity
(Nm3/hr)
70,000
500,000
15,000
16,000
20,000
19,000
300,000
2,000,000x2
20,000
30,000
490,000
550,000
350,000
440,000
71,000
762,000
10,000
20,000
180,000
960,000
480,000
1,000,000
1,900,000
1,660,000
50,000
70,000
150.000
104,000
30,000
300,000
15,000x2

Type of
Reactor^
1MB
1MB
1MB
FB
FB
FB
FB
FB
1MB
FB
PPC
PPC
FB
FB
FB
FB
FB
HC
HC
HC
HC
HC
HC
HC
PPR
PPR
PPR
1MB
CMB
1MB
FB
(Continued)
Start-up
mmm-
Nov. 1976
Oct. 1977
Oct. 1977
Aug. 1977
July 1977
July 1977
Apr. 1978
Apr. 1978
Dec. 1978
June 1978
June 1978
Oct. 1975
Nov. 1975
May 1976
Nov. 1976
Dec. 1976
Apr. 1977
Jan. 1978
Apr. 1978
June 1978
Apr. 1979
July 1979
Aug. 1981
Nov. 1975
July 1976
Mar. 1977
Aug. 1977
Aug. 1975
Sept. 1976
Dec. 1976


-------
                          TABLE I (Continued).   SELECTIVE CATALYTIC REDUCTION PLANTS IN JAPAN
ro
Process Developer
Mitsubishi H.I.
Mitsubishi H.I.
Mitsubishi H.I.
Mitsubishi H.I.
Mitsubishi H.I.
Mitsubishi H.I.
Mitsubishi Kakoki
Mitsubishi Kakoki
Mitsubishi Kakoki
Mitsubishi Kakoki
Mitsubishi Kakoki
Mitsui Engineering
Mitsui Engineering
Mitsui Toatsu
Mitsui Toatsu
Mitsui Toatsu
Mitsui Toatsu
Mitsui Toatsu
Sumitomo Chemical Co.
Sumitomo Chemical Co.
Sumitomo Chemical Co.
Sumitomo Chemical Co.
Sumitomo Chemical Co.
Sumitomo Chemical Co.
Sumitomo Chemical Co.
Sumitomo Chemical Co.
Sumitomo Chemical Co.
Sumitomo Chemical Eng.
Sumitomo Chemical Eng.
Sumitomo Chemical Eng.
Sumitomo Chemical Eng.

Gas
Source
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler
Furnace
Boiler
Boiler
Boiler
Boiler
Boiler
Furnace
Furnace
Furnace
Furnace
Boiler
Boiler
Furnace
Furnace
Furnace
Furnace
Furnace
Furnace
Boiler
Boiler
Boiler
Boiler
Boiler
Boiler

Fuel3
HO(LS)
HO(LS)
LNG
HO(LS)
HO(LS)
HO(LS)
HO(HS)
Naphtha
Kerosene
Kerosene
Kerosene
CO
HO(LS)
Off Gas
Off Gas
Off Gas
Off Gas
Off Gas
HO(HS)
LPG
LPG
LPG
LPG
LPG
LPG
HO (MS)
HO (MS)
Naphtha
Naphtha
Naphtha
Naphtha

Capacity
(Nm3/hr)
40,000
200,000
1,690,000x2
1,010,000
490,000
1,920,000
14,000
19,000
10,000
30,000
43,000
200,000
220,000
87,000
91,000
170,000
363,000
300,000
30,000
200,000
250,000
200,000
200,000
100,000
10,000
240,000
300,000
31,000x2
23,000
23,000
19,000

Type of
Reactor1*
PPC
HC
FB
PPC
PPC
HC
FB
FB
FB
FB
FB
FB
TC
FB
FB
FB
FB
FB
FB
FB
FB
FB
FB
FB
FB
FB
1MB
FB
FB
FB
FB
(Continued)
Start-up
Mar. 1977
Jan. 1978
Oct. 1978
Feb. 1978
July 1978
Feb. 1980
July 1978
Oct. 1978
Nov. 1978
Oct. 1977
Oct. 1978
Oct. 1975
Apr. 1978
Feb. 1976
Sept. 1976
Jan. 1977
June 1977
Oct. 1977
July 1973
May 1974
Jan. 1975
Apr. 1975
Apr. 1975
—
—
Mar. 1976
Oct. 1976
Oct. 1977
Dec. 1977
June 1978
July 1978


-------
                          TABLE I (Continued).  SELECTIVE CATALYTIC REDUCTION PLANTS IN JAPAN
ro
Source:  Reference 3

a.  Fuel Codes
     CO        Carbon Monoxide
     COG       Coke Oven Gas
     HO        Heavy Oil
     HO(HS)    High Sulfur Heavy Oil
     HO(LS)    Low Sulfur Heavy Oil
     HO(MS)    Medium Sulfur Heavy Oil
     LNG       Liquefied Natural Gas
     LPG       Liquefied Propane Gas
b.  Reactor Codes
     CMB       Continuous Moving Bed
     FB        Fixed Bed
     HC        Honeycomb Catalyst
     1MB       Intermittent Moving Bed
     PPC       Parallel Plate Catalyst
     PPR       Parallel Passage Reactor
     TC        Tubular Catalyst

-------
              TABLE II.  WET NOY/SO_, CONTROL PLANTS IN JAPAN
                               A.   A
                              Capacity         Source of
  Process Developer           (NmJ/hr)            Gas              Start-up


                 OXIDATION/ABSORPTION/REDUCTION PROCESSES

Chiyoda                         1,000        Oil-fired Boiler      Aug. 1973
Ishikawahima H.I.               5,000        Oil-fired Boiler      Sep. 1975
Mitsubishi H.I.                 2,000        Oil-fired Boiler      Dec. 1974
Osaka Soda                     60,000        Oil-fired Boiler      Mar. 1976
Shirogane                      48,000        Oil-fired Boiler      Aug. 1974
Sumitomo Metal-Fujikasui       62,000        Oil-fired Boiler      Dec. 1973
Sumitomo Metal-Fujikasui      100,000        Heating Furnace       Dec. 1974
Sumitomo Metal-Fujikasui       39,000        Oil-fired Boiler      Dec. 1974
Sumitomo Metal-Fuj ikasui       25,000        Sintering Machine      —  1976

                      ABSORPTION/REDUCTION PROCESSES

Asahi Chemical                    600        Oil-fired Boiler      Apr. 1974
Chisso Corp.                      300        Oil-fired Boiler      Apr. 1974
Kureha Chemical                 5,000        Oil-fired Boiler      Apr. 1975
Mitsui S.B.                       150        Oil-fired Boiler      Apr. 1974
Source:  Reference 3
                                    276

-------
                         TABLE III.   COSTS OF NOV AND NOV/SOV CONTROL SYSTEMS IN JAPAN
                                                A       A   A
Process Type
Selective Catalytic Reduction (SCR)
Flue Gas Desulfurization (FGD)
SCR + FGD
Simultaneous NO /SO (Dry and Wet)
Pollutant
Removed
NO
X
so2
N0x & S02
NO & S0_
Capital Cost
₯/kW $/kW
2800 15.5
14000 77.8
18500 102.8
20000 111.1
Operating Costs
₯/kWh
0.3
1.2
1.5
1.7
mills/kWh
1.6
6.7
8.3
9.4
ro
         Basis for estimate:
              Plant Size
              Fuel
              NO  Concentration
              SO  Concentration
              Particulate Concentration
              Temperature
              NO  Removal Efficiency
              S02 Removal Efficiency
              Depreciation
              Interest Per Year
300 MW, new
oil
200 ppm
1500 ppm <,
200 mg/Nm
380°C
80%
90%
7 years
10%
Maintenance
Insurance
Overhead
Catalyst Life
Annual Operation
Ammonia
Power
Steam
Kerosene
Monetary Conversion
  Rate
3% of investment cost
2% of investment cost
5% of investment cost
2 years
8000 hours
₯80/kg
₯12/kWh
₯2/kg
₯32/kg

₯180/$
         Source:   Information from hearings of the Japan Environment Agency as reported in Reference 3.

-------
        TABLE IV.  PROCESSES RECOMMENDED FOR FURTHER STUDY
             UNDER THE EPA/EPRI/TVA ASSESSMENT PROJECT
          Process
Type of Process  (classification)
UOP-Shell
UOP-Shell
Hitachi Zosen
Kurabo Knorca
JGC Paranox
Moretana Calcium
Ishikawajima H.I. (IHI)
Asahi Chemical
MON Alkali Permanganate
Dry Simultaneous S0»/N0
(Selective catalytic reduction  of
 NO  and sorption of S0» with CuO)
   3t                   £,

Dry NO  only
(Selective catalytic reduction:
 parallel passage reactor)

Dry NO  only
(Selective catalytic reduction:
 metallic honeycomb reactor)

Dry NO  only
(Selective catalytic reduction:
 moving bed reactor)

Dry NO  only
(Selective catalytic reduction:
 parallel passage reactor)

Wet Simultaneous S02/N0
(Oxidation/absorption/reduction:
 chlorine dioxide as oxidant)

Wet Simultaneous S02/N0
(Oxidation/absorption/reduction:
 ozone as oxidant)

Wet Simultaneous S02/N0
(Absorption/reduction:  ferrous-
 EDTA as chelating compound)
Wet NO  only
      x    J
Source:  Reference 4

^ot initally recommended;  added as  a substitute for the wet NO
 only process.

 Development discontinued by  the process developer;  hence,  the
 process was subsequently dropped from the recommended list.
                               278

-------
       TABLE V.  COSTS OF NOV AND NOY/SO  SYSTEMS IN THE U.S.
                            X       XX
                                                         Revenue
                          Pollutant     Capital Cost   Requirement
     Process Type          Removed         ($/kW)       (mills/kWh)
Selective Catalytic
Reduction (SCR) NO ,Part.
3t
Flue Gas Desulfuri-
zation (FGD) S02
SCR + FGD NOx,S02,Part.
Dry, Simultaneous
(UOP-Shell) NOx,S02,Part.
Wet, Simultaneous NO ,80-, Part.
Moretana (CIO-) X
Asahi (EDTA)
IHI (Ozone)
>60

100
>160
>155
>180
>200
>380
>2.2

4.2
>6.4
>5.7
>16.0
Basis for the Estimate:
  Particulate Control System

  FGD System
  S09 Removal Efficiency
  NO  Removal Efficiency
  Particulate Removal Efficiency
  Boiler Size
  Fuel
    Heating value
    Sulfur content
    Ash content
  Operation
  Capital Investment
  Annual Revenue Requirement
ESP for dry systems; wet scrubber
for wet simultaneous systems
Limestone
90%
90%
99.5%
500 MW, new
Coal
24.4 MJ/kg (10,500 Btu/lb)
3.5%
16%
7000 hr/yr
mid-1979
mid-1980
Source:  Preliminary results from Reference 5.
                                  279

-------
            TABLE VI.   COMPARISON OF ENERGY REQUIREMENTS FOR
                   VARIOUS NOV AND NOV/SOV PROCESSES
                             X       XX
           Process Type
    Pollutant
     Removed
of boiler capacity
 Selective Catalytic Reduction  (SCR)
 Flue Gas Desulfurization  (FGD)
 SCR + FGD
 Dry Simultaneous (UOP-Shell)
 Wet Simultaneous
  Mpretana  (C10_)
  Asahi (EDTA)
  IHI (Ozone)
    NO ,Part.
      Xso2
  NO ,SO-,Part.
    X   £f
  NO ,S07,Part.
    Jt   £
  NO ,S09,Part.
    2£   £m
       ^0.3
        3.4
       ^3.7
                          'v-ll.S
                          ^18.6
SBasis for  the estimate:  See Table V
  Includes heat credit for byproduct streams
Source:  Reference 5.
          TABLE VII.   STANDARD  BOILERS SELECTED FOR EVALUATION

Package,
Package,
Package,
Package,
Boiler Type
Scotch firetube
Scotch firetube
watertube
watertube, underfeed
Fuel
Natural gas
Distillate oil
Residual oil
Stoker coal
Thermal Input
GJ/hr (106 Btu/hr)
15.8
15.8
158.2
31.6
15.0
15.0
150.0
30.0
Field-erected, watertube,
  chain grate
Field-erected, watertube,
  spreader
Field-erected, watertube
Stoker coal          79.1        75.0

Stoker coal         158.2       150.0
Pulverized coal     210.9       200.0
                                    280

-------
              TABLE VIII.  CONTROL REQUIRED OF THE 14 LARGEST
                     POINT SOURCES IN CHICAGO TO COMPLY
                         WITH A SHORT-TERM NAAQS
Short-Term NAAQS
(yg/m )
1000
750
500
250
Percentage of Plants Requiring
No Control Moderate Control High Control
79
36
14
0
21
64
57
7
0
0
29
93
3Combustion modification  0\/50% NO  control)
b                                ^
 Flue gas treatment or combustion modification plus flue gas  treatment
 (^90% NO  control)
         JH

Source:  Reference 8.
                                    281

-------
       CHEMILUMINESCENT MEASUREMENT OF
               NITRIC OXIDE IN
             COMBUSTION PRODUCTS
                     By:

               Blair A. Folsom
              Craig W. Courtney
Energy and Environmental Research Corporation
         Santa Ana, California 92705
                      283

-------
                                   ABSTRACT
     The response of commercially-available chemiluminescent NO  analyzers to
                                                               X
simulated combustion products containing NO was investigated.  An accurate flow
metering system was used to combine 0^  CO^ CO, NZ, Ar, HO, CH,, and H. into
simulated combustion products modeling a wide range of fuels and excess air
conditions.  The simulated combustion products were doped with known amounts
of NO and then monitored by two commercial chemiluminescent NO  analyzers.
                                                              A
     The results indicate that chemiluminescent analyzers spanned with NO/N.
generally indicate less than the actual amount of NO in sampled combustion
products.   The errors range from one to 20 percent depending upon fuel type,
oxidant composition and the sample conditioning system.
                                       284

-------
                               ACKNOWLEDGEMENT
     The work upon which this publication is based was conducted under
Contract No. 68-02^-2631 with the Environmental Protection Agency.  The
authors wish to express their appreciation to W. S. Lanier of the Environ-
mental Protection Agency and to M. P. Heap, J. S. Johnsen and J. M. Keene
of Energy and Environmental Research Corporation for their assistance in
various portions of the work.
                                        285

-------
                                  SECTION 1
                                INTRODUCTION
     The measurement of nitric oxide  (NO) concentrations in combustion pro-
                                                                          (1 2 3)
ducts by chemiluminescence has several advantages over alternative methods  ' '
and as a result the chemiluminescent NO analyzer (CLA) has become a standard
instrument for most laboratory and field emission tests.  The usual procedure
for CLA calibration is to set the instrument zero and span with high purity
nitrogen (N ) and a known concentration of NO in N_ respectively.  The instru-
ment is then used to measure NO in combustion products.  Ideally the composition
of the combustion products, which varies according to fuel type and excess air
level, will not affect the measurement of NO.  However, several investigators
have found that background gas composition can measurably affect the NO concen-
tration indicated by a CLA.  ' '    Extrapolating these results to typical com-
bustion product compositions indicates that the common procedure of neglecting
background gas composition variations could introduce errors as large as 28% in
indicated NO concentrations.     The objectives of this study were to:
     •    Assess the accuracy of commercial CLAs measuring NO in combustion
          products following the procedures recommended by the instrument
          manufacturers
     •    Examine methods of correcting CLA indicated NO concentrations
          for the effects of background gas composition variations
     •    Examine methods to improve calibration procedures
     Two commercially available CLAs were zeroed and spanned with N  and NO in
N- respectively and then used to measure known concentrations of NO in a
variety of background gases simulating combustion products from a wide range
of fuel compositions and combustor operating conditions.  Binary mixtures were
also tested.   The NO concentrations indicated by the CLAs were then compared
with the actual NO concentrations to determine the effects of background gas
composition.
                                       286

-------
                                 SECTION 2
                     CHEMILUMINESCENT MEASUREMENT OF NO
     The chemiluminescent measurement of NO is based upon the following
four reactions:
                          NO + 03 	* N02* + 02                      (1)
                          NO + 03	 N02  + 02                      (2)
                             N02* 	" N02  + hV                      (3)
                          N02* + M	-N02  + M                       (4)
Nitric oxide and ozone (0-) react readily to form nitrogen dioxide (N0_) in
either an excited state (N02*) or a ground level state (NO ).  The yield of
NO * is about 10% at ambient temperatures and increases with temperature by
                    (3)
approximately 0.9%/KV  .  The excited molecules can decay to ground state
giving off light of a characteristic frequency (chemiluminescence) or can
collide with any third body (M) and decay to ground state without chemilumi-
nescence (quenching).  The relative importance of the chemiluminescent and
quenching reactions depends upon the temperature and the amounts and types of
molecules available for quenching.  If these factors are constant and the
amount of 0. present is large, the intensity of chemiluminescence is directly
proportional to NO concentration.
     Figure 1 is a simplified schematic diagram of a NO analyzer based upon
this reaction scheme.  A sample gas containing NO is metered into a reaction
chamber at a constant rate by a suitable flow metering system.  Ozone pro-
duced from 0- is also metered into the reaction chamber at a constant rate.
The 02 flowrate is selected to produce a reaction chamber 0_ concentration
many times the maximum anticipated NO concentration.  Thus the reaction
between NO and 0_ does not measurably deplete the 0. concentration and the
intensity of chemiluminescence is directly proportional to NO concentration in
the sample gas.  The chemiluminescent emission is filtered to remove
                                       287

-------
extraneous light and directed upon a photomultiplier tube (PMT).  The PMT
output is amplified and suitably scaled to read out directly in concentration
units (PPM).  The amplifier zero offset and gain are adjustable to permit
calibration with standard gases.  High purity N_ (free from NO) is used to
set  the  zero offset.  This allows the PMT dark current, which varies with PMT
age  and  temperature, to be nulled out.  A known concentration of NO in N^ back-
ground gas is used to adjust the instrument gain or span.
     Commercial CLAs are designed and constructed to maintain the variables
affecting instrument calibration very close to constant.  Calibration drift on
state-of-the-art instruments is usually less than 3% of span per day.  A CLA
calibrated as described above will indicate the correct NO concentration
(within about 1%) of any mixture of NO in N« within instrument range.  However
if the sample gas contains species other than NO or N?, these gases could
cause the instrument to read incorrectly.
     There are four potential ways in which sample gas composition variations
could alter CLA response to NO; these are:
     •    Depletion of 0_ by chemical reaction
     •    Chemiluminescence of gases other than NO within the optical
          filter band width
     •    Sample flowrate variations due to sample gas property changes
     •    Sample gas  species with quenching efficiencies  different from N
The first two items are not  usually important in measurements of combustion
products.  However the last  two items can cause significant variations in CLA
calibrations.
     The CLA measures the number of NO molecules entering the reaction chamber
per unit time.  If the sample gas f lowrate is constant, the number of molecules
of NO entering the reaction  chamber per unit time is directly proportional to
the NO concentration.  The sample gas flowrate in many commercial CLAs is
maintained constant by passing the sample gas through a capillary and regulating
the capillary pressure drop.  However, the volumetric flowrate through a
capillary with fixed pressure drop is inversely proportional to the sample gas
dynamic viscosity (p).   Thus sample gas compositions with viscosity different
from that of N  will result in sample gas flowrates different from N« and
                                        288

-------
therefore the CLA will not respond properly to NO.  Sample gases with low
viscosity will result in indicated NO concentrations greater than actual and
vice versa.
     The NO * produced by reaction of NO and 0- can either decay to the ground-
state producing chemiluminescence or transfer energy to other molecules (M)
without chemiluminescence (quenching).  Some quenching is unavoidable due to
the finite probability of NO-* molecules colliding with other molecules in the
reaction chamber.  The rate of the quenching reaction depends upon the amount
and types of molecules in the reaction chamber.  Reaction chamber pressure
affects quenching by altering the mean free path and collision frequency.  Low
pressure tends to reduce quenching and several commercial CLAs operate with
reaction chamber pressures less than 10 torr.  The quenching reaction rate
affects the intensity of chemiluminescence and hence variations in the quench-
ing reaction rate will alter CLA response to NO.
     The quenching reaction rate is significantly influenced by the type of
molecules in the reaction chamber.  Experimental evidence confirms that the
rate of quenching increases with the number of degrees of freedom of the
molecules present in the reaction chamber and varies substantially for gases
commonly found in combustion products.     Mathews et al    recently com-
piled quenching data from several investigators.
     The variations in chemi luminescent response to NO due to quenching and
viscosity effects are generally of the same order of magnitude.  However, the
magnitude and direction of the effects vary for specific gases.  For some
gases such as H_ and 0_ the combined effects of quenching and viscosity result
in a greater error than for quenching alone.  For other gases such as H?0, CO.
and Ar the error is reduced by inclusion of viscosity effects.
                                        289

-------
                                 SECTION 3
                    EXPERIMENTAL PROCEDURE AND APPARATUS
     Two CLAs were selected for evaluation:
     •    Beckman Model 951
     •    Thermal Electron Corporation  (TECO) Model 10A
The Beckman analyzer utilizes an atmospheric reaction chamber.  Sample gas and
ozone flows are controlled by regulating the pressure drop across capillaries.
Sample pressure is provided by an internal sample pump.
     The TECO analyzer utilizes a subatmospheric pressure reaction chamber
operating at nominally 9 torr.  As with the Beckman analyzer, flowrates in the
TECO are controlled by capillaries.  A vacuum pump exhausts the reaction
chamber and creates the pressure drop across the capillaries.  The use of a
subatmospheric reaction chamber adds considerable complexity to the instrument
but reduces the importance of quenching thus increasing sensitivity to low NO
concentrations.
     The manufacturers' instructions for both CLAs include detailed operating
information specifying the proper pressures, flowrates and calibration pro-
cedures.  Filtering and drying the sample gas to a dewpoint less than
operating temperature is required to prevent sample capillary blockage.
     The operating procedure.used by technicians in laboratory and field tests
varies somewhat with sample train design and test conditions.  Most commonly
the instrument is utilized intact, operated according to the manufacturers'
instructions and zeroed and spanned with N9 and NO in N_ respectively.  This
                                          £            £
was the procedure utilized in this investigation.  However, in tests of sub-
atmospheric combustors, such as flat flames, a low reaction chamber pressure
CLA (such as the TECO) is often used without the sample gas capillary.  This
allows the sample gas to pass directly from the probe to the reaction chamber
without water removal.  Sample flowrate is controlled by the probe tip which
                                       290

-------
acts as a sonic flow orifice.  The results of this investigation cannot be
directly applied to CIAs operated in this manner.
     Table 1 shows the gas mixtures tested in this investigation.  These
mixtures were prepared by blending high purity gases and mixtures of NO in N~
or Ar with a precision gas flow metering system.   Nitric oxide in N« span gases
were prepared similarly so that inaccuracies in mixed gases supplied in
cylinders could not affect the CIAs' relative responses to span gases and
sample gases.
     For each experiment in Table 1, the CIAs were adjusted to read correctly
on N_ with no NO and a mixture of NO in N_.  The instruments were then used to
measure the concentrations of NO in the mixture compositions tested.  The
actual concentrations of NO in span and test gases were identical so that the
ratios of NO measured in the test gases to actual NO concentrations were direct
measurements of background gas composition effects.  The CIAs' responses to N«
zero and NO in N~ span gases were checked after each datum point.  Concentra-
tions of 0_, CO and CO- in zero, span and test gases were also monitored as
a check on metering system accuracy and to verify no leaks in the sampling
train.
     A schematic diagram of the experimental apparatus is shown in Figure 2.
The only materials in contact with gas mixtures containing NO were stainless
steel, glass and teflon.  The total inaccuracy in the flowrate of each gas was
less than 0.5%.  Water vapor was added by bubbling some of the gases (0- and N^)
through distilled water.  Contacting of the NO mixtures with 02 and H20 was
controlled by maintaining the NO mixtures separate from the 0_ and HO.  The
two streams were blended Immediately upstream of the instruments in a short
length of teflon tubing.  The instruments included Beckman and TECO CIAs and
other instruments to measure 0^, CO, CO. and dewpoint.
     All gases were high purity grade  (99.97% or better) with the exception of
CO  (99.0%).  NO was supplied as mixtures in N. and Ar  (+ 5% accuracy) to
facilitate metering.  All mixed gases were calibrated against a National Bureau
of Standards  (NBS) traceable mixture.  To confirm the concentrations of NO in
the test gas mixtures, several gas mixtures were analyzed utilizing the
phenolydsulfonic acid  (PDSA) method.  The maximum deviation between calculated
and PDSA measured concentrations was 3.0%.
                                       291

-------
                                  SECTION 4
                                   RESULTS

EXPERIMENT NO. 1 - BINARY BACKGROUND GASES
     Figures 3 and 4 show the results of the binary background gas tests as
the ratios of indicated to actual NO concentrations as functions of the
concentrations of gases in the M/N- mixtures.  As the concentration of gas M
approaches zero, the ratio of indicated to actual concentration must approach
1.0 since 100% N_ was  the background gas used for spanning the instruments.
     The results for all gases except H_ at high concentrations show either
negligible changes or  decreases in the ratios of indicated to actual NO
concentrations as the  concentrations of the gases in the background mixtures
increase.  The results for CO and H9 are similar for both CLAs.  This is
                             (5)   L
expected since quenching data    indicates that the relative quenching
efficiency of CO, H_ and N_ are nearly the same.  Carbon monoxide has
essentially the same viscosity as N_ and thus the viscosities of CO/N
mixtures are the same  as N«.  This accounts for the negligible effects of CO
concentration in the background gas on the ratios of indicated to actual NO
concentrations,  d. has a viscosity much smaller than N? and the viscosities
of mixtures of EL in N« decrease rapidly as the concentrations of H? increase
above 20%.  This accounts for the increase in the ratio of indicated to actual
NO concentration at high H_ concentrations.
     The results for the other binary background gas mixtures in Figures 3 and
4 can also be explained by examining the combined effects of viscosity related
flowrate variations and quenching efficiency variations.  For example, CH, has
a viscosity less than N- and the viscosity related flowrate variations with
CH,/N_ mixtures tend to increase the ratio of indicated to actual NO concen-
trations as the concentration of CH.  increases.  However, since CH. has more
                                   4                              -4
degrees of freedom than N2,  its quenching efficiency should be greater tending
to decrease the ratio of indicated to actual NO concentration as the concen-
                                       292

-------
trations as the concentration of CH, increases.  However, since CH, has more
degrees of freedom than N_> its quenching efficiency should be greater tending
to decrease the ratio of indicated to actual NO concentration as the concen-
tration of CH, increases.  The results in Figure 3 show that the combined
effects of quenching and viscosity variations produce a nearly flat response
with the TECO.  With the Beckman, the quenching efficiency effects are much
larger (due to the higher reaction chamber pressure) and over compensate the
viscosity effects to produce a decrease in ratio of indicated to actual NO
concentration as CH. concentration increases.
                   4

EXPERIMENT NO. 2 - LEAN COMPLETE COMBUSTION PRODUCTS (DRY)
     Figure 5 shows the results of tests on the simulated complete dry com-
bustion products from burning C and CH, in air.  These background gas mixtures
contained 0_, C09 and N .  In all tests the ratios of indicated to actual NO
concentrations were less than 1.0.  The TECO generally indicated NO concentra-
tions closer to the actual levels than the Beckman.  This is consistent with
the results of experiment No. 1.  For both instruments, the simulated com-
bustion products from C exhibit more deviation than those from CH,.  This is a
direct result of the higher C0« concentration in the simulated combustion pro-
ducts from C.
     These experiments were conducted with two concentrations of NO in the
sample and span gas mixtures.  Figures 5A and B show the results for 200 ppm
NO for C and CH, fuels.  Figure 5C compares the results for C as the fuel with
100 and 200 ppm NO and shows that the ratio of indicated to actual NO concen-
tration is independent of the NO doping level.

EXPERIMENT NO. 3 - LEAN COMPLETE COMBUSTION PRODUCTS (WET)
     The previous experiments were conducted with dry sample gas mixtures
prepared by blending dry compressed gases from cylinders.  The dewpoints of
these mixtures were about 205K.  In laboratory and field emissions tests the
combustion products generally contain several percent H_0 and the dewpoint
may be as high as 350K.  The usual procedure is to draw the sample gas through
a water/ice bath to remove most of the HO and reduce the dewpoint beneath
ambient temperature.  This prevents condensation in the sample capillary and
the associated plugging.

                                      293

-------
     In this experiment the water vapor was added to the dry simualted com-
bustion products to produce dewpoints in the range of 275 to 294K.  Figure 6
shows the results for C burning in air at 100, 250 and 500% theoretical air.
At each stoichiometry increasing the dewpoint decreased the ratio of indicated
to actual NO concentration.  The effect was small for the TECO (1.0 to 2.0%)
but  substantial for the Beckman (up to 6.0%)

EXPERIMENT NO. 4 - INCOMPLETE LEAN COMBUSTION PRODUCTS (DRY)
     The method of simulating incomplete combustion products utilized in these
tests was to allow 20% of the C0? in the simulated complete combustion pro-
ducts to be reduced to CO and 0_.  Figure 7 shows the results for simulated
combustion products burning C in air assuming complete and incomplete com-
bustion.  The TECO's performance is essentially unaffected while the Beckman's
improves with incomplete combustion.  This is a result of the decrease in CO2
concentration.

EXPERIMENT NO. 5 - RICH COMBUSTION PRODUCTS (DRY)
     The simulated rich combustion product mixtures contained CO, CO,,, H» and
N_.  The results of those tests are shown in Figure 8.  As with the lean com-
bustion products, the TECO generally indicated closer to the actual NO concen-
tration than the Beckman.  Largest deviations were at near stoichiometric
conditions with C as the fuel.  Under very rich conditions (50% theoretical
air) with CH, as the fuel, both CLAs indicated slightly higher than the actual
concentation of NO.  At 50% theoretical air the concentration of H? is over
17% and the binary background gas tests demonstrated that high concentrations
of H_ result in indicated NO concentations greater than actual.
     It should be noted that the products of combustion from practical corn-
bus tors operated fuel rich will generally be different from the compositions
utilized in these tests.  These simulated combustion product compositions were
selected for simplicity and to give an approximate indication of the CLA per-
formance sampling rich products.

EXPERIMENT  NO.  6 AND 7 - COMBUSTION PRODUCTS IN Ar  ATMOSPHERES (DRY)
     In experiments involving combustion of fuels containing bound nitrogen
                                       294

-------
it is difficult to differentiate between NO formed from atmospheric N  and NO
formed from nitrogen bound in the fuel.  Several investigators have recently
combusted fuels containing bound nitrogen in an oxidant containing only 0 ,
CO 2 and Ar to eliminate NO formation from N .  The oxidant mixtures most
commonly used are 21% 02 balance Ar and 21% CL, X% CO  balance Ar.  Carbon
dioxide is sometimes added to adjust the flame temperature to match the flame
temperature produced with air as the oxidant.  The concentration of CO
required is usually 20%.
     Experiments 4 and 5 were repeated with 0 /Ar and 0 /CO /Ar as the oxidants
                                             ^         +-   £•
and the results are shown in Figures 9 and 10.  The results for lean combustion
products obtained with the TECO are essentially independent of stoichiometry,
oxidant (02/Ar or 02/C02/Ar) and fuel (C or CH^) .  However, while similar
results for combustion of fuels in air gave ratios of indicated to actual NO
concentrations of 0.95 to 0.99, these results are much lower (0.82 to 0.87).
     The results for lean combustion products obtained with the Beckman are
also lower than comparable results obtained with air as the oxidant but the
differences are smaller than with the TECO.  The Beckman results are also
affected by the CO  content of the oxidant; increasing CO  concentration
decreases the ratio of indicated to actual NO concentrations.
     The results of the rich combustion products tests shown in Figure 10 show
the same general trends as the lean tests:  the TECO response is nearly con-
stant independent of stoichiometry and oxidant under rich conditions and the
Beckman results are very sensitive to the concentration of CO^ in the oxidant.

EXPERIMENT NO. 8 - OFF DESIGN POINT CLA OPERATION
     All of the previous experiments were conducted with the CLAs operated
strictly according to the manufacturers' instructions.  Figure 11 shows the
results of operating the CLAs under other conditions.  Simulated combustion
products from burning C in air and 0 /Ar oxidants were utilized and the CLAs
were zeroed, spanned and used to measure these sample gases with the operating
parameters set at several conditions.  Figure 11 shows the ranges of instrument
response observed as the independent parameters were varied.  In all cases,
variations in the operating parameters resulted in only minor changes (+ 3%)
in instrument response.
                                       295

-------
                                 SECTION 5
                                 DISCUSSION
     The results of the experiments discussed above show that commercial CLAs
 generally Indicate NO concentrations lower than actual when spanned with
 mixtures of NO in N_ and used to sample NO in combustion products.  The errors
 introduced by neglecting sample background gas composition variations cover
 the following ranges:
     COMBUSTION PRODUCTS                                 ERROR %
        FROM FUEL AND;                            TECO	     BECKMAN
      21% 02/N2                                 +1 to -5          +2 to -11
      21% 02/Ar                                -11 to -16         -6 to -14
      21% 02/21% C02/Ar                        -12 to -15        -15 to -20
These results apply directly to the two CLAs tested but are probably representa-
tive of the performance of other instruments of similar design as well.  Both
instruments utilize capillaries to control sample flowrates.  The TECO reaction
chamber operates at approximately 9 torr while the Beckman reaction chamber
operates at atmospheric pressure.
     The gases responsible for the majority of the errors are C02, HO, 0« and
Ar.  Each of these gases decrease the CLA's sensitivity to NO.  Carbon dioxide
is often present at high concentrations in combustion products and accounts for
the majority of the errors at near stoichiometric conditions.  Water vapor has
a much higher quenching efficiency than N .  At the low H?0 concentrations
typical of combustion products dried in an ice bath, the quenching effects far
outweigh the effects of viscosity.  Sample dewpoint variations can change CLA
response to NO by several percent.  For 0 , the effects of viscosity an
quenching variations are additive and this combined with the high concentra-
tions of 02 present in lean combustion products results in errors as high as
 4%.
                                       296

-------
     The effects of viscosity and quenching variations are opposite for Ar.   In
the CLAs tested here, the viscosity related variations over compensated the
quenching variations resulting in a decrease in CIA sensitivity to NO with
increasing Ar concentration.  In the experiments reported in reference 5, the
sample flowrate was maintained constant and the results did not include the
viscosity related flowrate variations present in these experiments.  This
accounts for the reported opposite effect of Ar on CLA response to NO.  In com-
bustion experiments utilizing 0-/Ar or 0 /CO /Ar as the oxidant, the concentra-
tion of Ar in combustion products is high and results in CLA errors as high
as 20%.
     The variations in CLA response due to background gas composition var-
iations were generally smaller with the TECO than with the Beckman except for
Ar.  Based upon these results a low reaction chamber pressure CLA such as the
TECO would be preferred to monitor products of combustion of fuels burned in
air.  However, the choice of instruments for experiments involving 0 /Ar or
0_/CO /Ar oxidants is unclear.  The following example illustrates these points.
     A fuel containing bound nitrogen is burned in three oxidants:  air,
O./Ar and 0_/CO_/Ar.  It is assumed that 500 ppm of NO are formed from fuel
bound N  independent of oxidant composition.  It is further assumed that an
additional 500 ppm of NO are formed from N_ when the fuel is burned in air.
Ideally, the experimental results would indicate 1000 ppm in the air experi-
ment and 500 ppm in both Ar experiments.  The thermal NO could  then be calcu-
lated as the difference in NO concentrations  (500 ppm).
     Table 2 shows the results which would be obtained in this  experiment
utilizing the TECO and Beckman CLAs but without correcting the  results for
background gas composition variations.  The fuel was assumed to be C burning
at  100%  theoretical  oxidant and  the instrument response was taken from the
results  in section 4.
     In  the air experiments the  TECO reads closer to the correct concentration.
However, with the 0_/Ar oxidant  both CLAs indicate incorrectly  by  the same
amount.  The thermal NO concentrations calculated by subtracting the NO  con-
centrations are in error by equal amounts for both CLAs but the errors are  in
opposite directions.  With  the 02/C02/Ar oxidant the TECO reads closer to the
                                        297

-------
correct concentration.  However, the Beckman yields better results for the
thermal NO.  This example shows no clear cut advantage for either instrument
design.
     The results of this study can be used to correct the results of NO
emission tests for the effects of background gas composition variations.
Accurate corrections can be made if:
     •    The CLA utilized in the emission tests was one of those
          tested here
     •    The CLA was operated according to the manufacturer's
          recommendations
     •    The combustion product compositions are known
     Figure 12 shows the application of the results of this study to an actual
NO emission test conducted at Energy and Environmental Research Corporation.
A No. 6 oil was fired in a bench scale furnace over a range of excess 0? with
air and 09/CO /Ar oxidants.  Nitrogen oxides (NO ) were measured with a TECO
         £*   &                                  &
CLA equipped with a stainless steel converter to reduce N0_ to NO.  The indi-
cated NO  concentrations were corrected for background gas composition var-
        H
iations by applying the results in section 4.  The C and CH, data were inter-
polated to account for the oil's H/C ratio and "the results obtained for 21%
0_/21% CO /Ar were used (neglecting the difference between 20 and 21% CO- in
the oxidant).  Application of these results to the experimental data slightly
increases the NO  concentration for combustion in air and substantially
                Ji
increases (60 ppm)  the NO  concentration for combustion in the oxidant contain-
                         Ji
ing Ar.   The conversion of fuel nitrogen to NO  is higher and the thermal NO
                                              Ji              '               2£.
formation is lower  than if the effects of background gas composition varia-
tions had been neglected.
                                       298

-------
                                 SECTION 6
                               RECOMMENDATIONS
     In many laboratory and field emissions tests the variability in operating
conditions and other factors introduce errors of the same order of magnitude
or larger than the errors produced by background gas composition variations.
Under these conditions it is recommended that the NO concentrations indicated
by the CLAs be used directly without corrections recognizing that the results
will tend to be low by a few percent.
     If the experimental errors in the emissions tests are small compared to
the errors produced by background gas composition variations, the indicated NO
concentrations should be corrected as discussed in the previous section.
However the results of this investigation can only be applied to CLAs designed
and operated similar to those tested.  In low pressure flat flame experiments
where the CLA sample gas metering system (capillary) is replaced by a different
type of metering system, the effects of background gas composition variations
on sample flowrate may be significantly different.  Under these conditions the
results of this investigation cannot be applied.  It is recommended that either
the effects of background composition variations on sample gas flowrate be
evaluated and quenching data such as Reference 5 be applied, or the overall
effects of background composition variations be emperically determined as in
this investigation.
     As an alternative to correcting the indicated NO concentrations, the CLA
could be operated so as to minimize the effects of background gas composition
variations.  Three methods for accomplishing this are to:
     •    Calibrate the CLA with NO in a background gas similar to the
          anticipated combustion products
     •    Dilute the sample with a known amount of N^
     •    Dry the sample to a low dewpoint
                                        299

-------
Calibrating the CLA with NO in a background gas with a composition identical to
the combustion products to be sampled would eliminate these problems.   However,
the concentration of NO in the mixture must be accurately known.   Some gas
mixture suppliers determine the concentration of NO in their mixtures  by using
a CLA.  This introduces errors similar to calibrating the laboratory CLA on
NO in N2.
     If the sample gas is diluted with a large amount of N_, variations in
sample gas compositions will have less effect on the flowrate into the reaction
chamber and quenching.  For this technique to be effective the dilution factor
must be large (such as 10/1) and precisely known independent of sample gas
composition variations.  This accurate dilution is difficult to achieve in
practice and has the further disadvantage of reducing the overall sensitivity
of the CLA by the dilution factor.
     The sample gas must be dried to a dewpoint less than the CLA operating
temperature to avoid plugging the sample capillary with condensation.   However
at 293R dewpoint the sample gas contains 2.3% H_0 vapor and as a result the
CLA will indicate less than the actual NO concentration due to background and
'dilution effects.  Lowering the dewpoint to 273K reduces the H_0 vapor concen-
tration to 0.60% and reduces the error in indicated NO concentration.
     In experiments involving combustion products of fuels in air, a low
pressure reaction chamber CLA such a.3 the TECO is preferred.  However in
experiments involving  combustion  in O./Ar or O./CO  Ar oxidants,  high  and low
reaction chamber pressure CLAs perform equally well.
     Small variations  in CLA operating conditions such as adjustment of the
sample gas or ozone flowrates and changes in the reaction chamber pressure
do not have any clear advantages  in reducing the effects of background gas
composition variations.  However  if the reaction chamber pressure is main-
tained low, the sample gas  flowrate is maintained constant and the ratio of
0. to sample gas flowrates  in high, the effects of background gas composition
variations could be significantly reduced.
                                       300

-------
                                 REFERENCES
1.    Clough, P.N., and Thrush, B.A.   Mechanism of Chemiluminescent Reaction
     Between Nitric Oxide and Ozone.   Transactions of the Faraday Society,
     Vol. 63, 1967.  pp. 915.

2.    Fontijn, A., Sabadell, A.J., Ronco, J.  Homogeneous Chemiluminescent
     Measurement of Nitric Oxide with Ozone.  Analytical Chemistry, Vol.  42,
     1970.  pp. 575.

3.    Niki, H., Warnick, A. and Lord,  R.  An Ozone - NO Chemiluminescence
     Method for NO Analysis in Piston and Turbine Engines.  In:  Proceedings
     of Automotive Engineering Congress.  Detroit, Michigan, 1971.  pp.  246.

4.    Maahs, H.G.  Interference of Oxygen, Carbon Dioxide and Water Vapor on
     the Analysis for Oxides of Nitrogen by Chemiluminescence.  NASA Technical
     Memorandum NASA TM X-3229.  August 1975.

5.    Mathews, R.D., Sawyer, R.F. and Schefer, R.W.  Interferences in Chemilum-
     inescent Measurement of NO and NO- Emissions from Combustion Systems.
     Environmental Science and Technology, Vol. 11, 1977.  pp. 1092.
                                       301

-------
o
ro
           Sample
Reaction Chamber
  Optical  Filter
                    P.M.  Tube
           Light Tight Housing
                                               Vent
                                                                    Ozone
                                                                  Generator
                                                                 Power Supply,
                                                                   Amplifier
                                                                  and Readout
                                  Figure 1.  Chemiluminescent Measurement of NO.

-------
co
s
Compressed High Accuracy
Gas Metering
Cylinders System
o
V— X
O
O




O
O



1 	 " CH4
»»




1 >

f 	 »
I 	 »
»

>
H2
CO or Np

NO/N2

NO/Ar
°2
co2

Ar or N2
Variable Mixing
Ox1d1zer Stream
t Water Addition
-*&-H



»









°'»
• »
t

t















•
t













.*»
•
t











Heat




































Fuel Stream



























j



i
*-^
*










^

j




j
i \
\



. J *
                                                                                           Bypass
                                                                                        Psychometer
                                                                                          Beckman
                                                                                        NOX Analyzer
                                                                                           TECO
                                                                                       NOY Analyzer
                                                                                         A
                                                                                        Paramagnetic
                                                                                        02 Analyzer
                                                                                          Beckman
                                                                                         Infrared
                                                                                        CO Analyzer
                                                                                          Beckman
                                                                                         Infrared
                                                                                       C02 Analyzer
                                                                                                            Vent
                                       Figure 2.   Experimental Apparatus.

-------
    •o

    t—I
   o
1.021
0.98
0,94
0.90
0.86
0.82
0.7*
,
L
-
mm
-
I i 1 i Qxjpgtfi '
O"
A o
fi f*
u D
0 0"
n a.
i i i i i
j
i



L i- - L
h °
O
O
1 |
1 C4rbonlQ1oxid€ !
—
0 a
o
o
1 I p
co
o
1.0IJ
0.98|

0.94

0.90

0.86

0.82
0, 78
\f% 1 V
'(
L *" i * ^ — 213
Carbon Monoxide H

-

—

~ o Beckman
Q TECO
^ ^
1 i i i 1
D 10 20 30 40 50
J










<
' ' ° 1 JL
_ Methane _
O
— -
0 o
— -
0
o-

^ ^
1 1 1 1 1
0 10 20 30 40 50
                                             Concentration  of Gas 1n Mixture (%) Balance
                                    Figure 3.   Experiment No.  1:   Binary Background Gases.

-------
liO?
1
0.98
0.94
0.90
0.86
0.82
"S 0.78

ID r— 1 ,0€
(J IO
•r- 3
T3 -t->
£ £ 1.02
0 O
OJ 2T 2? 1
O 1
01 0.98
0.94
0.90
0.86
0.82
0.7S
1 i i . ' I 1
I D S " D ° °

-
1 1 1 1
a r
i i i IP *
^ Hydrogen
• 	 M M _ ._, - - - __

— ~
-
-
1 1 1 1
1 Argjon
0
1
O'
•
~l
1
Tc i nn
fi) Ivli
O Beeknrftn
a TECO



0
10
20
30       40        50
     Concentration  of Gas  1n  Mixture  {%)  Balance
          Figure  4.  Experiment No.  1  (continued):  Binary  Background Gases.

-------
1.03
1.01
.99
.97
.95
.93
.91
ii i i i i i i
I TECO - 200 ppra
1 D
! o°° a~-
. ° o o
oG
" o CH4 I
I 1 I J 	 J 	 1 	 1 	 1 	 L^M
                wo
   500
Theoretical Air (X)
                                                       oo


•o
V
+»
o
•o
t-4
3


1.03
1.01
.99
.97
5 .95
0 .
f .93'
.91
• QHi
i i i i i i i i i
-
DO
- Q o o-
o °
o

L n CH^^ I
pi i i ii ii ii [V
                                                          f>
                                                       oo
1.D3
1.01
.99
.97
.95
.93
.91
.m
| | 1 II Till
"^^ Bcctnui*- ^ ^^ TfCOc
• 0 AAA - i- i-,' . • «Uljk/ * - J'- M
L\nj ppm tuu j>pjn
•
'*": ! . I B:
~ ft °
o °
o
" ° * fl»l • C I
i i i i i i i i _,i_M
Figure 5.  Experiment No.  2:   Lean Complete Combustion
           Products Burning Fuels in Air (dry).
                            306

-------
                              ItJW T.A.
.99
i
.97
.95
.93
.9*
i
.89
4P

.99
1
• 97
•o
V
% r- -95
5 ^» 4
TJ 3 .93
c o
2~* °
* & .91

.89

i

.99
.97*
i
.95
.93

.91

.89
.87

•
• •
D n
"6 °
-
-
, —
8 „
.
- K , • . ,- , -
250X T.A.
• Mi
^
0 o o -

-
1
• _
o
~ o

o
_ «.
p K ,
N
500% T.A.
.
D
0 DO.
-
O
O

o
—
• N ,
Dry\ 272 283 294
                            Dewpo!nt
Figure 6.  Experiment No. 3:  Lean Complete Combustion
           Products from Burning C in Air (wet).
                          3Q7

-------
                              TECO
 .95
 .93
 .91
 .89
_ •'
           *
       O  Complete
       o
                             1/2!
     I      I     I     I      I     I     I      I     I
                             tedcmrt
 .99
 .97
 .95
.93
:     ••    .
__ O
.8!
         1     1     1      1     1     1
                                       1      1
   .100
300        500        700
         Theoretical A1r  (*)
                                           WO
  Figure 7.  Experiment No. 4:  Incomplete Lean Combustion
            Products Burning C in Air (dry).
                        308

-------
co   o
s   z
1.02
i.or
Inn
• UU
.99
.98
.97
.96
to
5.95
0*
.94
o93
.92
.91
.90
.89
1 1 1 1
L
TECO
- D D 1
0 D
0 0 <
O
— —
_ — .
_ —
- O c
D CH4
1 1 1 1
1

'
-






] 1 1 1 1
Beckman
- ' D
- D
O
a
o
" 0 ]
— _
o
mm ' MM
1 1 1 1 <
             50
60        70        80
     Theoretical A1r («)
90      100 50       60         70        80
                            Theoretical  Air (*)
90
100
                              Figure 8.  Experiment No. 5:  Rich Combustion Products Burning
                                         Fuels in Air  (dry).

-------
1,021—i     r
0.98
0.94
0.90 -
         8   8    a   °
         J	L
0.82 -

0.78
    100 200  300  400   500
                                                1
                                     TECO  21* 02/Ar   .
TECO 21% 02/21X C02/Ar  „
                                                            o    6   Q   g
                                    Beckman  tl% 02/Ar  _
V UN-
D CH4
1 1 1 1
1 ,
                                              1000  |/«100  200  300 400  500

                                             Theoretical Air (%)
                                                                                                      V
1 1 1 1 1
I Beckman 21
-
[088°
r °
r i i i i
i
[% Q2/2l% C02/Ar
B

i A.

—
1
^
•MM
                1000
                  Figure 9.  Experiment No. 6:  Lean Complete Combustion Products Burning
                             Fuels in 0,,/A and QjCOjf^ Oxidants  (dry).

-------


NOInd1cated




Inn
.00
0.96
<
0.92
13
0.88
1
0.84
0.80
(
fill
Beckman
)
0
O
O
1
OQ
9
50 70 80 90 100


;




1 1 1 1
TECO
O 21% 02/Ar
a 21* o2/2is; co2/Ar
'> 8 • . .-
o
-
1 1 1 I
60 70 80 90 100
Inn
• UU
0.96
0.92
0.88

0.84
0.80

    Theoretical  A1r (%}
Theroetlcal Air (%)
Figure 10.  Experiment No. 7:  Rich Combustion Products Burning C in
            02/A and 02/C02/A Oxidants (dry).

-------
	TECO	
1.0 < Sample Vacuum < 6.0 1o. HG
0.5 <02 Pres. < 6.0 PSIG
8.5 < Chamber Pres. < 14;0 Torr.
,	Beckman	
2*0 < Sample Pres. < 4.25 PSIG
5.0 < 02 Pres. < 30.0 PSIG
l.UU
.98
0.96
^ 0.94
2 « -
00 £ §
"g t5 0.92
o o
z z 0,90
0.88
0.86
0.84
0.82
0.80

TECO
- 21% 02/N2
Oxidant
_

— Beckman
__ 	 ^_ 2151 0 /N Perkman
"" TEO) Oxidant 21% 02/Ar
21% 02/Ar Oxidant

—
 Figure 11.  Experiment No. 8:  Off Design Point CLA Operation Burning C
             at 125% Theoretical Oxidant (dry).

-------
CO
CO
Argon Furnace, Sonlcore Ultrasonic Atomizer, No Swirl A1r,
No, 6 Alaskan/North Slope 011 - 0.51% N - 100% Conversion « 815 ppm (Dry 0% 02)
                          O Oxldant 21% 02/20% C02/N2       D
                Raw Data                                       Corrected Data
                                                                                     Oxldant Dry A1r  21%  02/N2
800
700
-*, '
o
* 1
o
^ 600
X
o
o.
D-
500
400
i i i i
; - s - *
I
-

-
i i i i
2 34 5 6
.
1




i i i i
• D ° ° *.
1
o .
o
o
•
1 1 1 1
23456
                                    % Excess 0,
                                                              %  Excess  0,
                                 Figure 12.  Example:  Correcting Data for Background Gas Composition.

-------
                                            TABLE I.   TEST MATRIX
EXPERIMENT
NUMBER
1

2

3

4

5




6

7





8

CONDITIONS SIMULATED
Binary Background Gases

Lean Complete Combustion
Product - Dry
Lean Complete Combustion
Product - Wet
Lean Incomplete Combustion
Product - Dry
Rich Combustion Product -
Dry



Lean Complete Combustion
Product - Dry
Rich Combustion Product -
T»
Dry




Lean Complete Combustion
Product - Dry
FUELS


C, CH.
4
C

C

COU
• v»n *
A




C, CH,
•T
C





C

OXIDANTS


21% 00/N_
2 2
21% 02/N2

21% 0 /N

21% 02/N2




21% 02/A
21% 02/21% C02/A
21% 02/A





21% 02/N2
21% 02/A
STOICHIOMETRY
% T.A.


100 - cio

100,250
500
125 - 300

50 - 100




100 - oo

60 - 100





125

COMMENTS
02,CO,C02,CH4,H2 or A
in N2


Dry to 294 K Dewpoint

20% CO.-»- CO + 1/2 0
2 2
[02] = 0, Relative C
and H Stoichiometries
Equal, No Soot or
H C
x y


[0_] = 0, Relative C

and H Stoichiometries
Equal, No Soot or
H C
x y
Instruments Operated
off Design Points
CO

-------
                                     TABLE II.  COMBUSTION EXPERIMENT  EXAMPLE
CO
en
OXIDANT
Air
02/A
02/A
02/C02/A
02/C02/A
NO SOURCE
Fuel + Thermal
Fuel
Thermal (By Difference)
Fuel
Thermal (By Difference)
INDICATED NO CONCENTRATIONS (PPM)
BECKMAN
890
430
460
395
495
TECO
970
430
540
430
540

-------
                               TECHNICAL REPORT DATA
                         (Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/7-79-050b
                          2.
                                                    3. RECIPIENT'S ACCESS!Of*NO.
4. T.TLE AND SUBTITLE proceediiigs of the Third Stationary
Source Combustion Symposium; Volume n. Advanced
Processes and Special Topics
                                                    5. REPORT DATE
                                                    February 1979
                                                    B. PERFORMING ORGANIZATION CODE
7 AUTHOR(S) Joshua S. Bowen, Symposium Chairman, and
Robert E.  Hall, Symposium Vice-chairman
                                                    B. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
See Block 12.
                                                    10. PROGRAM ELEMENT NO.
                                                    EHE624
                                                    11. CONTRACT/GRANT NO.

                                                    NA (Inhouse)
12. SPONSORING AGENCY NAME AND ADDRESS
 EPA, Office of Research and Development
 Industrial Environmental Research Laboratory
 Research Triangle Park, NC 27711
                                                    13. TYPE OF REPORT AND PI
                                                    Proceedings; 3/79
                                                                    PERIOD COVERED
                                                    14. SPONSORING AGENCY CODE
                                                      EPA/600/13
15. SUPPLEMENTARY NOTES JERL-RTP project officer is Robert E. Hall. MD-65, 919/541-
2477. EPA-600/7-77-073a thru-G*73e and EPA-600/2-76-152a thru -152c are pro-
ceedings of earlier symposiums on the same theme.
16. ABSTRACT Tne proceedings document the approximately 50 presentations made during
the symposium, March 5-8, 1979, in San Francisco.  Sponsored by the Combustion
Research Branch of EPA's  Industrial Environmental Research Laboratory-RTP,
the symposium dealt with subjects relating both to developing improved combustion
technology for the reduction of air pollutant emissions from stationary sources,
and to improving equipment efficiency. The symposium was in seven parts, and
the proceedings are in five  volumes: I. Utility,  Industrial, Commercial, and Resi-
dential Systems; n. Advanced Processes and Special Topics; m. Stationary Engine
and Industrial  Process Combustion Systems; IV. Fundamental Combustion Research
and Environmental Assessment; and V. Addendum. The symposium  provided contra-
ctor, industrial, and government representatives with the latest information on EPA
inhouse and contract combustion research projects relating to pollution control,
with emphasis  on reducing  NQx while controlling other emissions and improving
efficiency.
17.
                            KEY WORDS AND DOCUMENT ANALYSIS
                DESCRIPTORS
                                        b.IDENTIFIERS/OPEN ENDED TERMS
                        c. COSATI Field/Group
 Air Pollution
 Combustion.
 Field Tests
 Assessments
 Combustion Control
 Fossil Fuels
 Boilers
                     Gas Turbines
                     Nitrogen Oxides
                     Efficiency
                     Utilities
                     Industrial Pro-
                       cesses
                     Hydrocarbons
Air Pollution Control
Stationary Sources
Environmental Assess-
 ment
Combustion Modification
Trace Species
Fuel Nitrogen
I3B
21B
14B
21D
13A
13CT
07B
13H
07C
18. DISTRIBUTION STATEMENT
 Unlimited
                                         19. SECURITY CLASS (This Report)
                                         Unclassified
                                                                 21. NO. OF PAGES
                                                                    319
                                        20. SECURITY CLASS (Thispage)
                                         Unclassified
                                                                 22. PRICE
EPA Form 2220-1 (»-73)
                                       316

-------