&EPA
United States     Industrial Environmental Research  EPA 600 "/S'9*
Environmental Protection  Laboratory          0-* 
-------
                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.  Socioeconomic 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.
                           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-600/7-78-191

                                             October 1978
Pollutant  Emissions from  "Dirty
   Low-  and  Medium-Btu  Gases
                           by

                R. T. Waibel, E. S. Fleming, and D. H. Larson

                     Institute of Gas Technology
                    Applied Combustion Research
                   III Center, 3424 South State Street
                      Chicago, Illinois 60616
                     Contract No. 68-02-2643
                    Program Element No. EHE624A
                 EPA Project Officer: David G. Lachapelle

                Industrial Environmental Research Laboratory
                 Office of Energy, Minerals, and Industry
                   Research Triangle Park, NC 27711
                         Prepared for

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

-------
                             CONTENTS
Abstract                                                        iv
Executive Summary                                                v
Figures                                                         vi
Tables                                                        viii
   1.  Introduction                                              1
   2.  Test Facilities                                          10
   3.  Experimental Plan                                        20
   4.  Results                                                  25
            Natural gas:  base-line tests                       25
            Wellman-Galusha oxygen:  emission studies           35
            Wellman-Galusha air:  emission studies              49
            Particulate studies                                 53
            Furnace efficiency                                  58
   5.  Discussion                                               68
            Fuel-nitrogen effects on NOX                        68
            Fuel-sulfur effects on NOX                          71
References                                                      74
Conversion Table                                                76
                                iii

-------
                                 ABSTRACT



     When low-Btu fuels, resulting from coal gasification,are not subjected to


post-gasifier cleanup, their combustion can result in NO  levels greatly above
                                                       X

the clean fuel values.  In general, for turbulent diffusion combustion the


level of NO  increases with an increase in a) the amount of fuel-nitrogen,
           X

b) the amount of fuel-sulfur, c) the level of excess air, and d) the degree


of initial fuel/air mixing.  Besides enhancing NO , fuel-sulfur is converted
                                                 X

to SO .  Results can be interpreted to support the contention that, in the
     X

presence of NO , large amounts of SO- may be formed in turbulent diffusion


flames characterized by poor fuel/air mixing.  Further research is needed,


however, to definitely determine the fate of fuel sulfur in this type of


flame.
                                     iv

-------
                             EXECUTIVE SUMMARY

     Data were collected to determine the emissions from "dirty" low- and

medium-Btu gases when combusted on industrial process burners.  The fuels

utilized were blended to have the composition typically found for Wellman-

Galusha oxygen (WGO) and air (WGA) fuel gases.  Base-line data were collected

for natural gas, ambient WGO and WGA, and hot WGO (700K) and WGA (616K).

Then ammonia, hydrogen sulfide, coal tar and char were added singly at vari-

ous levels and in combinations to the hot fuels in a parametric study to de-

termine the effects of these contaminants on pollutant emissions.  The burn-

ers used in this study were a forward-flow baffle burner and a gas momentum

controlled kiln burner.  These burners were mounted in turn on a pilot-scale

test furnace equipped with water tubes as a load and were each fired at

1.03 + 0.07 MW  (3.50 + 0.25 X 106 Btu/hr) with 10% excess air, and 477 K

(400°F) air preheat.

     Based on a detailed analysis of the" experimental data the following con-

clusions were made:

•    Low-Btu fuels not subjected to post-gasifier cleanup can yield
     NOjj levels greatly above the thermal levels for the clean fuels
     and for natural gas.

•    In turbulent diffusion 'flames, fuel-NOx  increases with an increase
     in a) the amount of fuel-nitrogen, b) the amount of fuel-sulfur,
     c) the level of excess air, and d) the degree of initial fuel/air
     mixing.

•    Attempts to close the fuel-sulfur balance were unsuccessful.
     Whether this shortfall is due to sampling/instrument effects or
     large concentrations of some unmeasured sulfur-containing species
     is not clear.  Further work should be done in this area.

•    Compared to natural gas, heat transfer to the load is reduced for
     the low-Btu fuels tested.  This heat transfer is not greatly af-
     fected by the presence of contaminants (tar and char) at levels
     characteristic of raw gasifier effluents.
                                      v

-------
                                FIGURES


Number                                                               Page

  1       Comparison of fuel conversion processes                      3

  2       Sulfur removal from coal during gasification                 6

  3       Nitrogen removal from coal during gasification               8

  4       Pilot-scale test furnace                                    11

  5       Overall furnace system                                      13

  6       Doping system to synthesize "dirty" low-btu gases           15

  7       Assembly drawing of the suction pyrometer                   16

  8       Assembly drawing of gas sampling probe                      18

  9       RAC "Staksamplr" schematic of EPA particulate sampling
          train (Method 5)                                            19

 10       Bloom baffle burner                                         28

 11       Assembly drawing of baffle burner                           29

 12       Profile of five-zone reheating furnace with overhead
          metal recuperator                                           29

 13       Scaled drawing of test furnace with cooling-tube posi-      31
          tions for PBB simulation

 14       Heat-absorption profile of preheat section of a five-
          zone reheat furnace as a function of furnace length         32

 15       Schematic drawing of a rotary cement kiln                   33

 16       Temperature and composition in a typical cement kiln        33

 17       Heat absorption profile of reaction and calcining zones
          of a rotary cement kiln                                     34

 18       Scaled drawing of test furnace with cooling tube posi-
          tions for FMCB simulation                                   36

 19       Schematic diagram of the kiln burner fuel injector          37

 20       Ammonia conversion for Wellman-Galusha oxygen fuel gas
          on the baffle burner with 10% and 20% excess air            44


                                  vi

-------
                            FIGURES, Cent.


Number

 21       Ammonia conversion for Wellman-Galusha oxygen fuel
          gas on the kiln burner with 10% and 20% excess air          45

 22       Effect of H2S on ammonia (1.0% by volume) conversion
          to NOX with Wellman-Galusha oxygen (10% excess air)         47

 23       Effects of radial vs. axial flow on ammonia (1.0%)
          conversion to NOX with Wellman-Galusha oxygen fuel
          gas on the kiln burner                                      48

 24       Particulate emissions for char doping of Wellman-
          Galusha fuel gases on the baffle and kiln burners
          (10% excess air)                                            55

 25       Heat absorption profiles for ambient, "clean" Wellman-
          Galusha oxygen and natural gas on the baffle.burner         59

 26       Heat absorption profiles for hot "clean" Wellman-
          Galusha oxygen and natural gas on the baffle burner         60

 27       Heat absorption profiles for ambient, clean Wellman-
          Galusha air and natural gas on the baffle burner            61

 28       Heat absorption profiles for hot, clean Wellman-Galusha
          air and natural gas on the baffle•burner                    62

 29       Heat absorption profile for ambient,  clean Wellman-
          Galusha oxygen and natural gas on the kiln burner           63

 30       Heat absorption profile for hot, clean Wellman-Galusha
          oxygen and natural gas on the kiln burner                   64

 31       Heat absorption profile for Wellman-Galusha. air fuel
          gas compared to natural gas on the kiln burner              66

 32       Heat absorption profile for hot, clean Wellman-Galusha
          air and natural gas on the kiln burner                      67
                                 vii

-------
                                TABLES


Number                                                               Page

  1       Primary Converters — Distribution of Energy into
          Products and By-Products                                     4

  2       Tar Analysis                                                 9

  3       Char Analysis                                                9

  4       Fuel Composition for Low- and Medium-Btu Gases Tested       14

  5       Summary of Test Data for Baffle Burner                      26

  6       Summary of Test Data for Kiln Burner                        27

  7       Base-Line Data for Clean Fuels:  Natural .Gas and
          Wellman-Galusha Oxygen (WGO), 3.5 Million Btu/hr
          With 10% Excess Air                                         38

  8       Char Doping:  Wellman-Galusha Oxygen (700 K), 3.5
          Million Btu/hr With 10% Excess Air                          40

  9       Tar Doping:   Wellman-Galusha Oxygen (700 K), 3.5
          Million Btu/hr With 10% Excess Air                          41

 10       Ammonia Conversion to NOX on the Baffle Burner              42

 11       Ammonia Conversion to NOX on the Kiln Burner                43

 12       Parametric Doping:  Wellman-Galusha Oxygen (700 K),
          3.5 Million Btu/hr With 10% Excess Air                      50

 13       Parametric and "Dirty" Doping:  Wellman-Galusha Oxygen
          (700 K), 3.5 Million Btu/hr With 10% Excess Air             51

 14      . Base-Line Data for Clean Fuels:  Natural Gas and
          Wellman-Galusha Air (WGA), .3.5 Million Btu/hr With
          10% Excess Air                                              52

 15       "Dirty" Doping:  Wellman-Galusha Air (616 K), 3.5
          Million Btu/hr With 10% Excess Air                          54

 16       Baffle Burner Cascade Impactor Results for Char-Doped
          Wellman-Galusha Oxygen (700 K) — Doping Rate:  1.14  g/s     56

 17       Kiln Burner Cascade Impactor Results for Char-Doped
          Wellman-rGalusha Oxygen (700 K) — Doping Rate:  0^76  g/s     57
                                 viii

-------
                               INTRODUCTION
     The objective of the research program was to provide and evaluate quan-
titative data on the differences in the environmental quality of effluent
combustion products and furnace efficiency when retrofitting a natural gas/
oil industrial burner with "dirty" intermediate- and low-Btu gases.  These data
were collected from the IGT pilot-scale industrial test furnace.
     This program was intended to complement a recently completed EPA pro-
gram,   which evaluated the emissions resulting from burning "clean" low- and
medium-Btu gases in boilers.  This previous work provided quantitative infor-
mation on emissions using gases processed by ambient temperature sulfur
cleanup systems and enabled correlating emissions to gas composition and
operating conditions.
     Work is currently being done by the government and industry to develop
high temperature-post gasification sulfur cleanup systems.  This is an at-
tractive development because it would enable utilizing the sensible heat con-
tained in the product gas from the gasifier.  The high-temperature cleanup
processes leave varying amounts of tars, oils, and ammonia in the product gas
stream.  Since these contaminants can contribute to emissions resulting from
combustion of fuel gas, it is important that the magnitude of the potential
problem be evaluated and the results used to determine if high-temperature
sulfur removal systems are feasible from an environmental viewpoint.  A
cursory look at the combustion process emissions using a "dirty" intermediate-
Btu gas reveals a potential increase in the NO  emission level from 50 ppm
                                              X
to 500 ppm.  Thus, before a decision on the suitability of retrofitting
gas/oil-fired industrial burners with "dirty" intermediate- and low-Btu gases
can be made, experimental values on NO  emissions must be available and eval-
uated relative to emission standards.
     A second area of concern when retrofitting an industrial furnace with
intermediate- or low-Btu gases is loss in furnace efficiency.  This decrease
in efficiency could arise due to a decrease in flame emissivity, lower adia-
batic flame temperature, and increased volume of combustion products.
"Dirty" low- and intermediate-Btu gases could have higher flame emissivities
than "clean" gases due to tar-oils and char, which could result in increased

-------
furnace efficiency.  However, the ambient "clean" fuel often has a higher
adiabatic  flame  temperature due to a lower water content in the fuel.  This
program was designed to quantify such changes in efficiency and provide data
on which a decision can be made on retrofitting an industrial process furnace
with low-  or intermediate-Btu gases.
     This  program was designed to provide data for two (2) gases and two (2)
burners as a means to broadly scope the potential environmental problem.
ENVIRONMENTAL ASSESSMENT OF COMMERCIAL GASIFICATION PROCESSES
Tars and Oils
     A wide variety of coal conversion systems can be used to produce a low-
or intermediate-Btu gas.  However, the operating conditions of the gasifier
and cleanup system will have a significant effect upon the types of potential
pollutants present in the off-gas.  In most of the gasification processes,
the initial treatment of the coal determines the major characteristics of the
raw gas.  These  initial treatment conditions are illustrated by the tempera-
ture-pressure coordinates of Figure 1.  Initial dimensions of contact time,
gas-to-coal ratio, contacter types (entrained bed, fluidized-bed, fixed-bed,
stirred liquid), and contacter mode (cocurrent, countercurrent, back-mixed
reactor), and perhaps other variables should be included for a complete pro-
file, but Figure 1 can be used to gain insight as to the types of pollutants,
particularly higher organic compounds that might be present in the raw gasi-
fier product.
     Comparison  of Figure 1 with Table 1 indicates how the initial operating
regimes of the various processes affect the products of the individual con-
verters, particularly the quantities of tars and oils that are present in the
raw off-gas.  These tars and oils may contain a variety of high-molecular-
weight materials.
     These liquids would be condensed and removed from the gas stream using
ambient temperature sulfur cleanup systems, but would not be removed when
using the high-temperature processes.  If not condensed they would be burned
during the combustion process with the off-gas.  The combustion of these tar-
and oil-containing gases in an environmentally acceptable manner was part of
the overall experimental evaluation.

-------
CYCLONE
BURNER •



LU
a:
r>
h-
a:
UJ
a.
s
LU
1-
Z
^ •

-------
     The basic character of the complex coal organic structure is aromatic.
Therefore, the tars that are expelled from coal during devolatilization in
lower temperature reactors may be expected to contain naphthalenes, indenes,
anthracenes, and similar compounds.  Oxygenated compounds such as phenols and
cresylic acids may be expected, in addition to nitrogen- and sulfur-containing
ring structures.  In moderate  temperature reactors, these complex aroma-
tics are hydrocracked and possibly hydrodealkylated to simpler BTX (benzene-
toluene-xylene) streams.
              TABLE 1.  PRIMARY CONVERTERS - DISTRIBUTION OF
                   ENERGY INTO PRODUCTS AND BY-PRODUCTS
                        	Percent of Output Energy
Process
Koppers-Totzek
Texaco
U-Gas
BI-GAS
C0_-Acceptor
HYGAS-Oxygen
Synthane
Westinghouse
Union Carbide
Winkler
Wellman-Galusha
Lurgi
COED
TOSCOAL
Garrett
H-COAL
CSF
SCR
     Of the conversion processes illustrated in Figure 1 and listed in
Table 1, only four are commercially available:  Koppers-Totzek, Winkler,
Wellman-Galusha, and Lurgi.  Of these, only Wellman-Galusha and Lurgi produce
significant amounts of tars and oils.
Gas
95
95
96
95
94
95
71
96
96
80
72
69
25
6
35
10
10
20
Tars and Oils
—
—
—
—
—
3
4
—
—
—
18
16
20
18
15
75
75
55
Chars , Ash ,
and Residue
5
5
4
5
6
2
25
4
4
20
10
15
55
76
60
15
5
25

-------
Sulfur
     The sulfur that is present in coal is one of the primary reasons that
low- and>intermediate-Btu gasification processes are being developed.  Much
of the coal in this country contains significant quantities of sulfur, and
present methods of sulfur oxides reduction, such as stack-gas scrubbing, are
not viewed as sufficiently reliable, effective, or operable by many potential
coal users.  The concept of low- and intermediate-Btu gasification permits
the removal of the sulfur from the gas before combustion, and overall sulfur
oxides emissions may be significantly reduced utilizing this approach.  The
other major impetus for the development of low- and intermediate-Btu gasifi-
cation is the higher overall efficiency achieved when the gas is utilized to
produce electricity in combined-cycle operations.
     The sulfur that occurs naturally in the coal will be largely driven into
the raw product gas, as illustrated in Figure 2.  These data are based upon
analyses of the initial coal and gasification residues; hard data do not
exist on the form of the sulfur in the off-gas.
     Thermodynamically, the great majority of the sulfur should exist as
hydrogen sulfide, with approximately 4% of the sulfur existing as carbonyl
sulfide (COS) and only trace amounts of carbon disulfide.  However, the
thermodynamics are based upon gas-phase equilibria and may not be applicable
in the gasification system.  In most calculations of moderate-temperature
gasifier operations, the gas-solids reactions are assumed to occur, but only
a limited number of gas-gas reactions will occur at temperatures below the
range of 1144  to 1366 K (1600° .to .2000°F).  Data available from Texaco and
Koppers-Totzek coverters indicate that the COS-H S equilibrium is attained;
comparable data from coke ovens, however, indicate that the concentrations of
COS and CS  are much higher than would be expected thermodynamically.  The
disposition of the sulfur in the gas in moderate-temperature gasifiers cannot
be accurately predicted at this time.
     Purification systems are available for removing a large quantity of the
hydrogen sulfide that is present in the raw product gas.  Currently, much
work is being done to develop and demonstrate the use of high-temperature
sulfur clean-up systems.  However, even the most optimistic projections show
an H S concentration of 100 ppm in the off-gas.

-------
    1.0
   0.9
O
UJ


fe 0.8
O
CD
tr
<
O

z
o:
u.
   °6
   0.5
    0.4

      0.5
           O  PITTSBURGH  COAL


           A  ILLINOIS NO.6 COAL
0.6       0.7       0.8       Q9

 FRACTION SULFUR GASIFIED
1.0
Figure  2.  Sulfur removal from coal during gasification

-------
Nitrogen
     Nitrogen that is present in coal may be considered as fixed nitrogen.
Fixed nitrogen may be defined as nitrogen that is chemically bound to other
species in contrast to molecular nitrogen (N£> that is present in the air.
                                                                    Q
The nitrogen in coal tends to gasify simultaneously with the carbon,  as il-
lustrated in Figure 3.  Generally, the nitrogen is expected to react with the
hydrogen during gasification to form ammonia.  The existence of ammonia in
raw gasifier effluents has been confirmed by many investigators.
     Hydrogen cyanide (HCN) is also present in the raw gas effluents.  Pub-
lished data report  concentrations of less than 10% of the ammonia concentra-
tion.  Thus, the major contribution in NOX formation will be ammonia.
     Although a number of potential pollutants'might be present in the raw
gasifier products, the data now available on their occurrence and/or concen-
tration are poor.  Consequently, the trials concentrated on the fate of species
for which some data exists (tars, oils, ammonia, particulate, and hydrogen
sulfide).  These contaminants were "doped" into, the hot experimental fuel gas.
Although coal-tar does not have an identical chemical analysis to that of tars
.and oils found in raw off-gas, it does contain all the aromatic and oxygenated
compounds that are found.  Thus, realistic characterization of the combustion
process and pollutant emissions can be anticipated. Analyses of the tar and
char are presented in Tables  2 and 3.

-------
100


 90
 70
I
O
.2
LU
 60
 50
8  40
t
z
    30


    20


    10
                             O IRELAND MINE COAL
                               MONTANA SUBBITUMINOUS COAL
              10
                            20          30
                         CARBON CONVERSION, %
40
50
     Figure 3.  Nitrogen removal from coal during gasification
                                                       8

-------
             TABLE  2.   TAR ANALYSIS
Ultimate Analysis

Ash
Carbon
Hydrogen
Sulfur
Nitrogen
Oxygen  (by  difference)
  Wt % (Dry Basis)

        0.0
       84.39
        5.65
        0.47
        0.55
        8.94
General Analysis

Solids
Heavy Fraction
Light Fraction  (Toluene,
  Benzene, Xylene)
 Wt  %  (As. Received)

       27.1
       55.3
       17.6
            TABLE 3.  CHAR ANALYSIS
Ultimate Analysis

Ash
Carbon
Hydrogen
Sulfur
Nitrogen
Oxygen (by difference)
Wt %  (Dry Basis)
     22.88
     66.30
        ,75
        ,78
      0.72
      6.57
1.
1,
Proximate Analysis

Moisture
Volatile Matter
Ash
Fixed Carbon
Wt %  (As Received)

      7.8
      13.6
      21.1
      57.5
Sieve Analysis

Screen
200
230
270
325
Pan
Wt % Retained
      67.7
       3.3
       4.9
       3.6
      20.5

-------
                               TEST FACILITIES
 PILOT-SCALE  FURNACE
     The experimental work was carried out in the pilot-scale furnace shown
 in Figure 4.  This furnace is 14 feet  long and has a cross-sectional area of
 21.3 sq ft.  The facility can be used for firing burners rated up to 6 million
 Btu/hr  (6M Btu/hr).  Combustion air temperatures up to 1000°F can be generated
 with a separately fired air preheater.
     The furnace is also equipped with 58 water cooling tubes, each of which
 can be independently inserted through the roof, along the sidewalls.  Varying
 the number of tubes, their location, and the depth of insertion allows control
 over the magnitude and character of the load that can be placed on the fur-
 nace.  The amount of heat absorbed by each tube can be determined by measuring
 the water flow through each tube and the temperature difference between the
 inlet and outlet.  The water temperature measurements are made with a
Vertronix digital thermometer.  The stated accuracy is  0.25°F.  These measure-
ments were checked with a mercury-in-glass thermometer, accurate to 0.1°F, and
were found to agree within 0.2°F.  A temperature difference of 25° to 60°F
was maintained between the water inlet and outlet.  This helped minimize the
effect of temperature measurement error on the heat balance.  The water flow
rate from each tube was determined by measuring the time to fill a bucket of
known capacity.  The time was measured using a quartz digital stopwatch.
     In addition to the combustion air preheater,  a separately fired fuel
preheater is available that can heat 12,000 SCFH of low-Btu gas to any de-
sired temperature up to 800°F.  Temperatures up to 1200°F are attainable with
lower flow rates.
   It is EPA policy to use metric units.  However, in this report,
   English units are occasionally used for convenience.  See
   conversion table at back of this volume.
                                     10

-------
Figure 4.   Pilot-scale test furnace

-------
     The overall system, shown in Figure 5, has the flexibility to indepen-
dently vary —
•    Fuel firing rate
©    Air input
•    Furnace load
•    Air preheat temperature
•    Fuel preheat temperature.
There are 33 panels or "sampling doors," shown in Figure 4, along one side-
wall that allow insertion of probes at any axial position-from the burner wall
to the rear wall.
LOW-Btu GAS GENERATING SYSTEM
     The low- and medium-Btu gases are generated using a special gas-generat-
ing and fuel-preparation facility.  The critical items are the special gas
generators or reformers that can produce varying ratios of hydrogen and car-
bon monoxide.  Natural gas, carbon dioxide, and steam are passed through re-
action retorts contained in a vertical cylindrical furnace.  The catalyst-
filled retorts are heated by the furnace and the input gases undergo e'ndo-
thermic chemical reactions at a temperature of 1422 K (2100°F).  The gases are
then quenched and compressed (maximum 80 psig).  Facilities are available to
remove excess carbon dioxide, if necessary.  After compression, the product
gas is blended with nitrogen, methane, carbon dioxide, and/or steam, as
required, to obtain the specified composition of the fuel gas to be tested.
     Up to 1.69 X 10  J/s (5.75 million Btu/hr) of simulated low- or medium-
Btu fuel gas can be generated.  This corresponds to 0.39 m /s (50,000 SCFH)
            /:    o            o
of 4.29 X 10  J/m  (115 Btu/ft ) low-Btu gas.  Table 4 gives the composition
of the Wellman-Galusha oxygen (WGO) and Wellman-Galusha air (WGA) fuel gases,
which were chosen to be simulated as test gases for the program.
DOPANT SYSTEM
     Figure 6 gives a schematic diagram of the doping system.  Raw coal-tar
from a coke oven was used as the tar introduced into the hot gas stream.
Ammonia and hydrogen sulfide were blended into the fuel.gas stream from cyl-
inders.  The flow of these dopants was adjusted using rotameters.  The tar,

                                     12

-------
                                                                                COMBUSTION AIR I
                                                                                METERING PITOT '
                                                                                            LEGEND
                                                                                        .X  PRESSURE . TEMPERATURE GAUGE
                                                                                     ;   T   THERMOCOUPLE
                                                                                           FLOW CONTROL VALVE
                                                                                        6 BALL \ALVE
                                                                                          . BUTTERFLY VALVE
                                                                                        B PNEUMATIC VALVE
                                                                                        6 GATE VALVE
                                                                                        (§( SOLENOID VALVE
L .
                                                 Figure  5.    Overall furnace system

-------
 which was  a liquid  at  room temperature, was  forced  from  a  container under
 nitrogen pressure through  a nozzle  in the hot  fuel  feed  line where it was
 steam atomized.  The tar nozzle was  about 2  m  from  the nozzle exit.  Feed
 rates were continually monitored  by  measuring  the weight loss of  tar from  the
 pressurized feed tank  over time.  Char was screw-fed  into  the hot fuel,  the
 feeder  being about  2.5 m from the nozzle exit.  Doping rates were controlled
 by  varying the screw speed.
      TABLE 4.  FUEL COMPOSITION FOR LOW- AND MEDIUM-Btu GASES TESTED
Fuel
Wellman-Galusha
   Oxygen
Wellman-Galusha
   Air
Temp,
°F
90
800
90
650
Heating
Value ,
CO
39
29
26
25

.2
.7
.9
.2
H2
40.4
30.6
14.3
13.4
C02
16.2
12.4
7.4
6.9
CH^'
'0.9
0.7
2.6
2.4
N2
1.4
1.1
46.9
44.1
H?0 Btu/SCF.
1.
25.
1.
8.
9
5
9
0
267
202
159
149
Adiabatic
Flame
Temp , *
K (
2248
2190
2045
2044
;°F)
(3587)
(3483)
(3222)
(3220)
*  10% excess air at 477 K (400°F).   The adiabatic flame temperature for
   ambient (298 K, 77°F) natural gas is 2231 K (3556°F).
INSTRUMENTATION
     A major task of this program is to measure temperature profiles at the
flue and flue-gas composition.  Modified designs of the International Flame
Reserach Foundation were used to construct probes that enabled this type of
data collection.
     Temperature data were collected using a suction pyrometer; the design
is illustrated in Figure 7.  A Pt/Pt-13% Rh thermocouple was used.  The
efficiency of the pyrometer was monitored and was better than 95% with a
15-second response time.
     The flow direction was measured using a water-cooled Hubbard probe, with
the upstream and downstream pressure taps connected to a Datametrics Barocel
transducer and Datametrics CGS electric manometer.
                                     14

-------
  CLEAN SYNTHETIC
    LOW-BtuGAS
ROTAMETERS
                           FUEL
                        PREHEATER
HYDROGEN
  SULFIDE
                        AMMONIA
                                                     SCREWFEEDER
                                                         N2
                                                   PRESSURE
                                                                                  •"DIRTY" LOW-Btu GAS
                                                                                   TO BURNER
                                                                    TAR
                     Figure 6.  Doping system to  synthesize "dirty"  low-btu gases

-------
         XL
Suction
                         Water
                                                                 Cool ing .Jacket
                                                                                      Radiation
                                                                                    /   Shield
                                                                                 Pt-Pt/13% RHI
                                                                                 Thermocouple
                                                                                                    Gas In
End View
                                                               SUCTION TIP FOR MEASUREMENTS IN
                                                               NATURAL GAS AND OIL FLAMES
                                                               SUCTION TIP FOR MEASUREMENTS IN
                                                               PULVERIZED-COAL FLAMES
                                              Alternate Probe Tips
                          Figure  7.   Assembly drawing of the suction pyrometer

-------
     Figure 8 shows the assembly drawing of the gas-sampling probe used in
the flame and the flue.  To minimize N0~ losses, the probe is water-cooled
                                  ®
stainless steel joined to a Teflon • sample line.  At the end of the probe is
a section of Teflon tube heated to 190°F, followed by a Millipore filter and
           ®
a Permapure  gas dryer.  This dryer reduces the dew point to less than 32 F .
In the dryer, water in the sample gas diffuses through a thin membrane into a
stream of dry nitrogen.  Tests have shown that only water is lost from the
sample stream.
     The analytic instrumentation equipment consists of the following items:
•    Beckman 742 Polarographic Oxygen (02)
•    Beckman Paramagnetic Oxygen (02)
•    Beckman NDIR Methane (CH^)
•    Beckman NDIR Carbon Monoxide (CO)
•    Beckman NDIR Carbon Dioxide (CO 2)
•    Varian 1200 Flame lonization Chromatograph (Total CH and C2 to Cg)
•    Beckman NDIR Nitric Oxide (NO)
•    Beckman UV Nitrogen Dioxide (N02)
•    Thermo Electron Pulsed Flourescent Sulfur Dioxide (S02)
•    Hewlett-Packard Thermoconductivity Chromatography, Hydrogen (H),
     Nitrogen (N2) , Argone (A2), CO, C02, C± to C5, Oxygen (02)
•    Beckman Chemiluminescent NO-N02
•    Data Integration System.
     To evaluate radiation intensity, which is needed for determination of
flame emissivity, a PR 200 Pyroelectric radiometer, manufactured by Molectron
Corp. in Sunnyvale, California, was used.  This radiometer uses a permanently
poled lithium tantalate detector that is capable of resolving radiant power
in the nanowatt range while maintaining a continuous spectral response from
the vacuum UV to 500 ^m.   A built-in optical calibration system, in the form
of a highly stable LED (light-emitting diode) that is calibrated against an
NBS traceable-standard of total irradiance, permits a direct correlation of
experimental data from different trials.
     To measure stack particulates, a Research Appliance Corp. (RAC) Model
2414 "Staksamplr" sampling train (Figure 9) was employed.  To provide greater
                                     17

-------
oo
         l-l/2-ln. 304 SS TUBING


l/2-ln. 3O4 SS TUBING
                                                  -l-3/4-ln. 304 SS TUBING



                                                         2-ln 3O4 SS TUBING



                                                                2-1/2-h. 3O4 SS TUBING
                 3/4-in. 3O4 SS

                        TUBING
                in.     \    	
                 	 T In.  \ • •	7 In.
       l/8-ln. 3O4 SS TUBING
                                                                                  *- 3/4-to. 3O4 S3 TUBING


                                                                  -l/4-l»i. TEFLON TUBING
                               Figure 8.  Assembly  drawing of gas  sampling probe

-------
  accuracy  in measuring stack-gas velocity, the differential pressure of the
  stack-gas velocity  probe  was  measured by a Barocel differential pressure
  transducer rather than the inclined manometer originally provided with the
  RAC instrument.  The  output of the pressure transducer was electronically
  amplified and  recorded continuously during a trial.
                                                             .Thermometer
          Reverse-type
           Pitot Tube
                                                                    Check
                                                                    Valve
                                                                    Vacuum
                                                                      Line
                                                      Main  Vacuum
                                                      Valve  Gauge
                                     Dry Gas
                                     Meter
Air-tight
 Pump
Figure 9.  RAC "Staksamplr"  schematic  of EPA particulate sampling train (Method 5)
       Particulate  size  distribution was determined using an Anderson Mark III
  high-temperature, multistage,  cascade impactor inserted into the RAC sampling
  train.  Data were analyzed by  the D^Q Method.  Because of sampling time re-
  strictions, size  distribution  data could only be taken at the higher char
  feed rates.
       For all particulate  trials,  the sampling probe assembly was inserted
  into the stack about 10-stack-diameters downstream of the flue and 3-diameters
  upstream of the top of the stack.  Isokinetic sampling conditions were main-
  tained for all trials  to within +8%, at worst.  Sampling times were normally
  from 1/2 to 1-1/2 hours, depending on the particulate loading rate.
                                        19

-------
                             EXPERIMENTAL PLAN

PHASE I.  EMISSION LEVELS OF PORTED BAFFLE BURNER
Task 1.0  Base-line Evaluation of NOX Emissions With Natural Gas
     To insure a level of NOX emissions characteristic of an industrial pro-
cess furnace and allow an accurate evaluation of changes in process furnace
efficiency when retrofitting to a "dirty" off-gas, the pilot-scale industrial
test furnace was modified to model an industrial process furnace.  The ported
baffle burner (PBB) was fired with natural gas, and water-cooling tubes were
inserted vertically along the furnace sidewalls until the pilot-scale furnace
reached an efficiency typical of industrial practice.  This efficiency was
determined by a mass energy balance around the furnace; the input data
collected for this balance included —
9    Temperature profile at the furnace flue using a suction pyrometer
•    Detailed flue-gas analysis with a thermal conductivity chromatograph
«    Flue-gas velocities with a modified Hubbard probe using an electric
     manometer
e    Inside wall temperatures using an optical pyrometer
•    Outside wall temperature with thermocouples
0    Water volume through the cooling tubes with a flow meter
a    Water inlet/outlet temperatures with thermocouples.
Once the furnace load had been fixed (number and position of cooling tubes)
the overall furnace efficiency, flame emissivity, flue-gas concentration
measurements, and particulate emissions were measured.  The furnace effi-
ciency and NOX flue concentration measurements were taken throughout the
program anytime the fuel or furnace operating conditions were altered.
Task 2.0.  NO,, Emissions and Furnace Performance for "Clean" Wellman-Galusha
0"xygen Gas
     To evaluate the changes in NOX emission levels and furnace efficiency
when retrofitting to a "clean" producer gas, the PBB was fired with "clean"
Wellman-Galusha oxygen gas.  The two fuel compositions tested during this
task are listed in Table 4.  To quantify these changes and provide means
to correlate .the natural gas test results with the "dirty" Wellman-Galusha
data, the following subtasks were conducted.
                                     20

-------
Task 2.1..  "Clean" Ambient Temperature Wellman-Galusha Oxygen—
     During this subtask, the PBB was fired with "clean" Wellman-Galusha
oxygen gas delivered to the burner at ambient temperature.  The fuel composi-
tion being tested at its delivery temperature was typical of a gasification
system that includes a complete post-gasifier cleanup system.  The data
gathered during this trial (NOX emissions and furnace efficiency) provided a
first line of correlation with the natural gas data.
Task 2.2.  "Clean" Hot Wellman-Galusha Oxygen Gas —
     The hot "dirty" producer gas that was studied during the program con-
tained ammonia, tar-oils, and char.  A base-line test was conducted to
establish the changes in NOX emissions and furnace efficiency that were due
to increases in the adiabatic flame temperature and the contributions that
were made by the ammonia, tar-oils, and char.  This test employed "clean"
Wellman-Galusha oxygen gas delivered hot.  Data were collected on NOX
emmission levels and overall furnace efficiency.
Task 3.0.  NOY Emissions and Furnace Performance for "Dirty" Wellman-Galusha
Oxygen Gas
                                     «
     Hot "dirty" producer gas contains a number of components (ammonia, tar-
oils, and char) that were expected to affect NOX emissions and flame emis-
sivity (resulting in a change in furnace efficiency) as compared with natural
gas and/or "clean" producer gas.  The following subtasks were designed to
evaluate the relative contributions of each of these sources to the total
NOX emission level and quantify any changes in furnace efficiency that re-
sulted from their presence in the gas.
Task 3.1.  Addition of Ammonia to "Clean" Hot Wellman-Galusha Oxygen Gas—
     To investigate the conversion efficiency dependence on fuel gas ammonia
concentration, NOX emission levels were measured for various concentrations
of ammonia in the fuel.  The effect of excess air was determined by making
the above measurements with 10% and 20% levels of excess air.
     As an initial experiment, with 1% ammonia addition, the oxygen in the
flue was set to the level required for 10% excess air with the undoped fuel.
The ammonia doping level was then reduced in five steps to 0.2% and the NOX

                                     21

-------
and C>2 levels were monitored.  This experiment showed the effect of the
ammonia doping on the oxygen levels in the flue.  At high doping levels, it
was necessary to include the ammonia as fuel when setting the excess air.
Task 3.2.  H2S Effect on NOX Emissions From Hot Wellman-Galusha Oxygen Gas
Containing Ammonia—
     To provide a quantitative answer to the possible interaction of fuel-
sulfur and fuel-nitrogen, a hot Wellman-Galusha oxygen gas stream with 1.0%
ammonia was doped with different amounts of HLS.
Task 3.3.  Addition of Tar-Oils to "Clean" Hot Wellman-Galusha Oxygen Gas—
     Since tar-oils contain chemically bound nitrogen, the purpose of this
task was to determine the contribution that they made to NO  emissions and
                                               3           x
what effect they had on the overall furnace efficiency.
     To permit an evaluation of tar-oils conversion to NO , several doping
levels were used.  Overall furnace efficiencies were measured.  The heating  .
value of the tar was included in the fuel enthalpy input.
Task 3.4.  Addition of Char to "Clean" Hot Wellman-Galusha Oxygen Gas—
     Like tar-oils, char contains fuel-bound nitrogen that could convert to
NO .  To evaluate the percentage conversion of the char-bound nitrogen to NO
  X                                                                         X
and the dependence of furnace efficiency on fuel char concentration, several
levels of char were doped into the gas stream.  At each doping level, NC^
emissions,  particulate emissions, and furnace efficiency were measured.
Task 3.5.  "Dirty" Wellman-Galusha Oxygen Gas—
     The purpose of this task was to quantify NO  and particulate emissions
                                                X
and to evaluate furnace efficiency for the PBB burning "dirty" Wellman-
Galusha oxygen gas.  "Dirty" Wellman-Galusha oxygen gas was defined for this
program as "clean" hot Wellman-Galusha oxygen gas doped with ammonia, char,
and tar.  This gas was similar in composition to that received from a high-
temperature sulfur-removal system.
Task 3.6.  Parametric Study of "Dirty" Wellman-Galusha Oxygen Gas—
     It was not anticipated that the separate contributions of ammonia, tar-
oils, and char to the level of NO  emissions will equal the concentration
    '                            x                  •

                                     22

-------
measured for "dirty" Wellman-Galusha oxygen gas.  To evaluate the relative
contribution of each of these potential sources of NO  emissions during the
combustion of Wellman-Galusha oxygen gas, a number of possible combinations
of ammonia, tar-oils, and char were tested.  Doping combinations tested in-
cluded —
e    Ammonia/Tar-Oils
•    Ammonia/Char
•    Char/Tar-Oils.
For each test condition, the NO  emission level was measured and overall fur-
                               x
nace efficiency was evaluated.
Task 3.7.  Evaluation of Reduced Emissions Through Combustion Aerodynamics
Using Wellman-Galusha Oxygen Gas—
     "Dirty" hot Wellman oxygen gas was tested with 15% external flue-gas
recirculation to determine the effect on NO  emissions.
                                           x
Task 4.0.  N0y Emissions and Furnace Performance for Wellman-Galusha Air Gas
     The two controlling factors in NO  formation and furnace performance,
adiabatic flame temperature and combustion aerodynamics, are significantly
different between Wellman-Galusha oxygen and air gases.  To quantify these
differences in NO  emission levels and furnace efficiency for a "clean" and
"dirty" Wellman-Galusha air gas for an industrial PBB, the following subtasks
were performed.
Task 4.1.  "Clean" Ambient and Hot Wellman-Galusha Air—
     This test established a furnace efficiency base line and NO  emission
level for "clean" ambient Wellman-Galusha air gas with the PBB.  To evaulate
the effect of fuel preheat on emissions and furnace performance, hot Wellman-
Galusha air gas was also tested.
Task 4.2.  "Dirty" Wellman-Galusha Air—
     "Dirty" Wellman-Galusha air gas was fired on the PBB to quantify the
level of NO  emissions and evaluate furnace efficiency.
           X
                                     23

-------
PHASE II..  EMISSION LEVELS OF FUEL MOMENTUM CONTROL BURNER



     Phase II of the proposed program was designed to evaluate changes in NO
                                                                            X

emissions and furnace efficiency caused by combustion aerodynamics.  During


this program phase, tasks identical to those in Phase I were repeated with


the fuel momentum control burner (FMCB).
                                     24

-------
                                  RESULTS
     Summaries of all the test data for the ported baffle burner and the fuel
momentum controlled burner are presented in Tables 5 and 6 respectively.  The
following sections present the detailed findings by fuel type.

NATURAL GAS:  BASE-LINE TESTS.
Baffle Burner Tests
     A baffle burner, representative of the forward-flow type,  was the first
burner to be utilized.  The burner tested, illustrated in Figures 10 and 11,
is full scale and is available as an off-the-shelf item from the manufacturer
(Bloom Engineering).  The burner consists of a centrally located gas nozzle,
surrounded by a high-temperature refractory baffle that has ports for the
injection of combustion air into the furnace.  The flame patterns produced
by this burner can be altered by changing the angles of these air ports or
their diameters.
     This type of burner is found on many large-scale industrial process
heating furnaces such as steel reheating, batch glass melting,  aluminum
holding, and tunnel kilns.  The baffle design selected for testing produces
a flame-to-furnace length ratio equal to the flame-to-preheat section length
ratio found in a five-zone steel slab reheat furnace.  Figure 12 illustrates
a five-zone steel slab reheating furnace.
     More than 60% of the fuel used in a reheat furnace is consumed in the
preheat section.
     The baffle burner selected for testing was fired at 3.5 million Btu/hr.
This fuel input produces a pilot-scale volumetric heat release rate of 12,000
Btu/ft -hr, which is within the range (10,000 to 25,000 Btu/ft -hr) of in-
dustrial firing densities.  The baffle used to produce the desired flame-to-
furnace length ratio has eight holes, 2.2 inches in diameter, rotated at a
15-degree angle.
     The rate of heat absorption and efficiency of the furnace preheat section
was simulated on the pilot-scale furnace by inserting water cooling tubes
along the furnace sidewalls.  A scaled drawing of the test furnace is shown in

                                      25

-------
                 TABLE 5.    SUMMARY  OF  TEST  DATA  FOR BAFFLE  BURNER
Inputs


Fuel Type
Natural Gas
WGO* (322K, 10%)
Excess Air
WGO* (700K, 10%)
Excess Air




WGO* (700K, 20%)
Excess Air




WGO* (700K, 10%)
Excess Air
Tar Doping
WGO* (700K, 10%)
Char Doping

WGO* (700K, 10%)
Ammonia /Tar
WGO* (700K, 10%)
Ammonia/Char
WGO* (700K, 10%)
Char/Tar
WGO* (700K, 10%)
Ammonia/Char/Tar
WGO* + 15% FGR
(700K, 10%)
Ammonia/Char/Tar
WGO* (700K, 10%)
Hydrogen Sulfide

WGA+ (322K, 10%
Excess Air)
WGAt (616K, 10%
Excess Air)
Rate
m3/s

0.027
0.111

0.140
0.134
0.134
0.134
0.134
0.134
0.134
0.134
0.134
0.134
0.134
0.134
0.141
0.141
0.141
0.138
0.138
0.138
0.139

0.138

0.138

0.138
0.136
0.138


0.134
0.134
0.136
0.175

0.175

NH3
	 %

0
0

0
0.19
0.38
0.60
0.81
1.03
0
0.19
0.38
0.60
0.81
1.03
0
0
0
0
0
0
0.99

1.00

0

1.00
1.02
1.00


1.04
1.04
1.02
0

0

H2S
	

0
0

0
.0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0

0

0

0
0
0


0.02
0.52
2.53
0

0

Tar
— g/i

0
0

0
•0
0
0
0
0
0
0
0
0
0
0
0.36
0.50
0.71
0
0
0
0.39

0

0.46

0.47
0.38
1.22


0
0
0
0

0

Char
3 	

0
0

0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0.13
0.24
0.86
0

0.13

0.13

0.13
0.07
0.13


0
0
0
0

0

Flue-Gas Composition
NO
	

112
38

30
310
360
400
420
430
38
290
355
410
435
465
52
55
58
30
33
44
520

470

40

500
510
510


480
655
950
16

22

H02
ppm

10
10

5
10
10
15
20
20
5
10
10
20
20
20
7
5
4
3
2
4
10

43

8

42
15
20


29
33
50
1

2

CO
	

10
200

115
55
55
53
55
60
45
• 45
45
40
40
40
108
93
93
59
100
90
108

65

63

75
24
45


40
50
37
260

35

COj
	 %

10.9
24.0

24.0
24.0
24.0
24.0
24.0
24.0
22.0
22.0
22.0
22.0
22.0
22.0
24.0
24.0
24.0
24.0
24.0
24.0
24.0

24.0

24.0

24.0
24.0
24.0


24.0
24.0
24.0
18.4

18.4

(as measured)
22.
—

1.9
1.8

1.8
1.9
1.8
1.9
1.9
1.9
3.3
3.3
3.2
3.2
3.2
3.3
1.8
1.9
1.9
1.8
1.7
1.9
1.7

1.8

1.9

1.8
1.9
1.9


1.8
1.8
1.9
1.3

1.4

S02 HC
— ppm —

__ —
— —

	 	
— —
— —
— —
— —
— —
— —
— —
— —
— —
— —
— —
1 <1
1 < 1
2 < 1
4 < 1
6 < 1
29 < 1
2 < 1

4 < 1

7 < 1

8 < 1
7 < 1
4 < 1


83 —
2546 —
11146 —
	 	

	 —

Particulate,
g/m3 g/s




















0.037 0.009
0.070 0.017
0.263 0.064







0.026 0.006










WGA* (616K, 10%)
   Ammonia/Char/Tar
0.186  1.04  0
                 0.58  0.35
                              610  20
                                        40
                                              18.4  1.4
                                                            8 < 1
                                                                   0.044   0.016
*   Wellman-Galusha oxygen fuel  gas.
+   Wellman-Galusha air  fuel gas.
t   Wellman-Galusha air  fuel gas.  Uses 3-inch nozzle.  (All others used 2-1/2 inch nozzle.)
                                                26

-------
                      TABLE 6.    SUMMARY  OF TEST DATA FOR  KILN  BURNER
                   Rate
Natural Gas a
WGOa'b (316 K, 10%
Excess Air)
WGO (705 K, 10%
Excess Air)




WGO (705 K, 20%
Excess Air)




WGO (705 K, 10%
Excess Air)
Hydrogen Sulfide
WGO (705 K, 10%)
Char Doping

WGO (705 K, 10%)
Tar Doping

WGO (705 K, 10%)
Ammonia/Tar
WGO (705 K, 10%)
Ammonia/Char
WGO (705 K, 10%)
Char/Tar
WGO (705 K, 10%)
Ammonia /Char /Tar
WGO + 15% FGR
(705 K, 10%)
Ammonia/Char/Tar
WGAC (322 K, 10%)
Clean
WGA (620 K, 10%)
Hot/Clean
WGA (620 K, 10%)
Ammonia /Char /Tar
0.
0.

0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.

0.

0.

0.

0.


0.

0.

0.

027
104

127
127
127
127
127
127
127
127
127
127
127
127
127
127
127
124
127
134
136
136
136
129

144

136

136

136


164

177

172

Inputs
NH3

0
0
0
0.20
0.40
0.63
0.85
1.09
0
0.20
0.40
0.63
0.85
1.09
1.09
1.09
1.09
0
0
0
0
0
0
1.07
0.98
0
1.02
1.02
H^S Tar Char

0
0
0
0
0
0
0
0
0
0
0
0
0
0
0.02
0.54
2.89
0
0
0
0
0
0
0
0
0
0
0,
B/s
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0.38
0.60
0.73
0.63
0
0.48
0.25
0.57

0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0.13
0.29
0.67
0
0
0
0
0.13
0.35
0.35
0.13
NO

70
35
24
270
360
420
460
500
40
270
370
420
480
550
470
700
920
30
35
55
50
52
58
540
480
65
540
525
Flue-Gas
NO 2
- ppm
5
5
4
20
30
30
40
40
5
20
30
30
40
50
30
40
80
5
2
5
5
3
4
40
30
5
40
30
CO

150
405
34
30
34
32
25
34
25
25
25
25
25
20
34
34
34
30
40
75
50
57
70
75
33
45
40
50
Composition
CO 2

10.2
24.0
24.0
24.0
24.0
24.0
24.0
24.0
22.0
22.0
22.0
22.0
22.0
22.0
24.0
24.0
24.0
24.0
24.0
24.0
24.0
24.0
24.0
24.0
24.0
24.0
24.0
24.0
°2

2,
1.
1.
1.
1.
1.
1.
1.
3.
3.
3.
3.
3.
3.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.

.1
,9
.8
.8
.8
.8
.8
.8
,3
.3
.3
.3
,3
.3
,7
,7
,7
.8
.9
,7
,8
,8
,7
,9
,8
,9
,8
,8
S02 HC
ppm
—
— —
	 	
—
—
—
—
—
—
—
—
—
—
—
90
2579
12800 —
3 <1
6 <1
10 <1
1 <1
1 <1
2 <1
1 <1
3 <1
6 <1
7 <1
4 <1
Particulates ,
g/m3 g/3
—
—
	 	
—
—
—
—
—
—
—
—
—
—
—
—
—
--
0.037 O.OO1
0.050 0.01
0.129 0.03
—
—
—
—
—
—
0.065 O.Oli
—
                              0    0

                              0    0

                              0.99  0
0    0

0    0
22

19
                                           0.45  0.13  460
 4    200    18.0   1.3

 2     30    18.0   1.3

30     35    18.0   1.3     7   <1
                                                                                                  0.025   0.010
a   Used 1-1/2  inch Schedule 10 (axial) fuel injector; all others used  4-inch Schedule 10 injector.
b   Wellman-Galusha oxygen  fuel gas.
c   Wellman-Galusha air fuel gas.
d   Fuel flow times heating  value adjusted to give  from 1.00 to 1.05 MJ/s input.
                                                     27

-------
10
oo
                                        Figure 10.  Bloom baffle burner

-------
 OBSERVATION PORT
  FUEL IN
                NOZZLE ASSEMBLY     BAFFLE
                                               PORT BLOCK
           Figure  11.   Assembly drawing of baffle burner

X
1
JJLJH
I

I
J
T] _
A
I/
I
Figure  12.   Profile of  five-zone reheating furnace with  overhead
                         metal recuperator
                                  29

-------
Figure 13.  The locations of the 58 water cooling tubes, which can be placed
within the furnace, are indicated by circles.  The heat-absorption profile
typical of a five-zone reheat furnace preheat section is given in Figure 14.
The curve has been plotted in dimensionless units to make scaling easier.  In
addition to reproducing this heat absorption profile, the furnace should be
operated with a thermal load efficiency (heat absorbed by water cooling tubes
divided by the fuel input enthalpy)of between 30% and 40%.  To achieve an
efficiency and heat absorption rate in this range, nine cooling tubes were
inserted into the furnace.  These tubes are denoted in Figure 13 by the
filled-in circles.  The heat absorption curve of Figure 14 was reproduced in
our furnace by adjusting the cooling surface area of each heat sink.  This is
accomplished by altering the length of each cooling tube exposed within the
furnace.  The entire thermal load was placed on the right furnace sidewall,
which would simulate the hearth of the reheat furnace.  Thus all temperature,
radiation, gas compsition, and flow direction profile data represent varia-
tions in the vertical plane of the reheat furnace.  If the flame were symmet-
rical about the burner axis, the relative orientation of the test furnace to
reheat furnace would not matter.  Photographs show that the flame is asymmet-
rical.  Thus, it is important to interpret the results as viewing the combus-
tion process from the roof of a reheat furnace.
Fuel Momentum Controlled Burner Tests
     A fuel momentum controlled kiln burner (FMCB) was installed on the pilot-
scale test furnace and set up to specifically simulate a cement kiln.  The
critical operating parameters were 1) a length sufficient to simulate the
calcining and reaction zones, 2) the firing density, and 3) the heat absorp-
tion profile.
     A typical kiln, shown schematically in Figure 15, consists of preheat,
calcining, reaction, and cooling zones.  The cross-sectional area (20 sq ft)
and length (14 ft) of the IGT pilot furnace allow for a near-ideal simulation
of the two zones of primary importance: the calcining and reaction zones.  It
is the heat transfer in these zones that is sensitive to fuel type.  In the
calcining zone, calcium and magnesium carbonate are decomposed to calcium and
magnesium oxide.  In the sintering or reaction zone, several complex reac-
tions occur, and among these the oxides combine with SiO. to form silicates.
                                     30

-------


400—





300 —

E
o
l"
K
0
2
LU
_J
0 200—
<
cr
u.





100-






0
O
O
O
0
O
O
O
O
O
O
O



O
0
O
O
O
O


O
O
O
O
o
o
o
o
o
o
o
o

1 1 1 1 1 1
-60 -40 -20 0 20 40
FURNACE WIDTH, cm
o
0
©
o
0
o
•
o
®
0
0



o
®
o
«
o



o
•
o
0
0
o
o
o
o
o
0
9

\
60

Figure 13.  Scaled drawing of test furnace with cooling-tube positions for
                              PBB simulation
                                    31

-------
10
NJ
           0.35
o
§   0.30
_l
CD
O
UJ   0.25
O
UJ
I

I
u_
o
g
>-
o
<
cr
           0.20
            0.15
           0.10
           0.05
                    TOTAL HEAT REMOVED BY LOAD
                    IN PILOT-SCALE TEST FURNACE
                    0.547 MWt (1.87 M Btu/hr)
                0
O.I
                        0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
                                          DISTANCE FROM BURNER WALL
                                                FURNACE LENGTH
               Figure 14.  Heat-absorption profile of preheat section of a five-zone reheat
                                 furnace as a function of furnace length

-------
 FEED
         DRYING
HEATING
CALCINING   REACTION COOLING
                                                                  DISCHARGE
            Figure 15.  Schematic drawing of a rotary cement kiln

These two zones occupy  about one-third of the overall kiln length, with  the
calcining zone being  twice  the length of the sintering zone.   The  flame  usu-
ally extends three-quarters of the length of these zones.
                                                                         3
     Firing densities in rotary kilns range from 10,000 to 20,000  Btu/hr-ft .
                                         3
A firing density value  of 12,500 Btu/hr-ft  was used for these tests  and is
typical of many kilns.  This requires a firing rate of 3.5 million Btu/hr for
our test furnace volume.  Cement kilns require about 5 to 7 million Btu/ton
of clinker.  Assuming 6 million Btu/hr, our furnace would simulate a  produc-
tion rate of 1150 Ib/hr.
     In order to determine  the heat absorption profile, we must consider the
temperature variation of the solid material (clinker) in the kiln.  Figure 16
                  DRYING     HEATING      CALCINING REACTION
           cr
           UJ
           o
           V)
           m
           in
           CD
                                                          4000°R
                                  (CoO)3Si02
                               -2000°R
                                  (CoO)2Si02
                          FRACTION OF KILN LENGHT
     Figure 16.  Temperature and composition in a typical cement kiln
                                     33

-------
shows a typical plot for this temperature.  This plot shows that the clinker
temperature is relatively constant through the calcining zone and then rises
rapidly as it enters the reaction zone.  In the calcining zone, calcium and
magnesium carbonate decompose by an endothennic reaction to calcium and mag-
nesium oxide;
                CaCO  = CaO + CO   -1350 Btu/lb CaO
                MgCO  = MgO + CO   -1264 Btu/lb MgO
With a composition of about 68% CaO and 1.5% MgO, the heat requirement in the
calcining zone is 950 Btu/lb of clinker produced.  Therefore, the thermal
load in the IGT furnace for the calcining section must be 950 Btu/lb times
the 1150 Ib/hr simulated production rate or 1.1 million Btu/hr, absorbed
relatively uniformly throughout the section.
     In the reaction zone, the temperature of the solid (clinker) materials
increases rapidly by about 900°F because of the completion of the calcining
reactions and the onset of the exothermic reactions.  The overall effect is
exothermic, releasing approximately 200 Btu/lb of clinker produced, which is
equivalent to 230,000 Btu/hr on the IGT test furnace.  The increase in the
temperature of the materials requires 233,000 Btu/hr (1150 Ib/hr X 0.225
Btu/lb-°F X 900°F), so virtually no heat is transferred by the flame in this
zone.
     Figure 17 shows the required heat-absorption profile.  The flame length
should be about 10.5 ft, and the total thermal loading of the furnace should
be 1.1 million Btu/hr.
HEAT REMOVED PER COOLING TUBE
TOTAL HEAT REMOVED BY LOAD
_ P 0
O — r\>

7 TUBES
REACTION
1
CALCINING
1 1
D 0.2 0.4 0.6 0.8 1.
DISTANCE FROM BURNER WALL
                                    FURNACE LENGTH
            Figure  17.  Heat  absorption profile  of  reaction  and
                 calcining zones of a rotary cement kiln
                                     34

-------
     This rate of heat absorption and this efficiency were simulated on the
pilot-scale furnace by inserting water cooling tubes along the furnace side-
walls, while firing natural gas at 3.5 million Btu/hr.  A scaled drawing of
the test furnace is shown in Figure 18.  The locations of the 58 water-cooling
tubes, which can be placed within the furnace, are indicated by circles.  The
seven filled circles indicate those tubes inserted into the furnace to simu-
late the desired efficiency and heat absorption rate for the kiln furnace.
Figure 19 is a schematic of the kiln burner fuel injector.
WELLMAN-GALUSHA OXYGEN:  EMISSION STUDIES
     Prior to the doping studies, base-line emissions were obtained for natu-
ral gas and Wellman-Galusha oxygen (WGO) fuel gas on the baffle burner and
kiln burner and are shown in Table 7.  Both fuels were fired at 1.03 + 0.07
MW  (3.50 + 0.25 X 10  Btu/hr) with 10% excess air.  The above variation in
fuel heat input represents the range of firing rates from run to run and not
the firing rate fluctuation during a given run, which was minimal.  For all
tests the furnace was at positive pressure.  All concentrations presented are
dry analyses at the flue entrance.  For NO  levels under 100 ppm the repro-
                                          X
ducibility of the reported values was + 5 ppm while at the higher levels it
was + 10 ppm.  Recent studies   have shown that quenching may cause interfer-
ence in NO  measurements.  Our results, however, have not been compensated
          X
for such effects.
     With natural gas and WGO on the baffle burner a 0.063-m inside diameter
(ID) (nominal 2-1/2 inch Schedule 40) fuel nozzle was used.  Fuel velocities
at the nozzle exit were 40 m/s (130 ft/s, ambient WGO) and 110 m/s (360 ft/s,
hot WGO).  Natural gas velocity was 9.5 m/s (28 ft/s).  Combustion air en-
tered the furnace at velocities of 19 m/s (63 ft/s) for the low-Btu fuels
and 23 m/s (76 ft/s) for natural gas.
     On the kiln burner a 0.043-m ID (nominal 1-1/2 inch Schedule 10) axial
fuel nozzle with a 0.019-m ID (7/8 inch, 16 gauge) tube as the radial injec-
tor was employed for natural gas and ambient (322 K) WGO.  For hot (700 K)
WGO a 0.108-m ID (nominal 4 inch Schedule 10) axial fuel nozzle with a 0.064-
m (nominal 2-1/2 inch Schedule 10) radial injector was required to achieve
comparable fuel velocities as well as to maintain flame stability.  The amount
of radial flow for natural gas on the kiln burner was 95% of the total.  This
                                     35

-------
                   400-
                   300-
                E
                o
                x"
                o
                z
                UJ

                O  200—
                <
                o:
                u_
                    100—
                      0
— — — — — l-Ll.
o
o
o
o
o
o
o
0
o
o
o
0
o
o
o
0
o
o
o
o
0
o
o
0
o
o
o
o
o
III!
-60-40-20 0
Jt 1
@
0
o
o
©
o
®
o
o
o
®
o
©
o
o
o
•
®
o
o
o
o
o
o
o
o
o
0
0
1 1 1
20 40 60
                               FURNACE WIDTH, cm
Figure 18.  Scaled drawing of test furnace with cooling tube positions for
                           FMCB simulation
                                  36

-------
I-1/2 in.
                XX\\X\\\\\\
                \//s ssssssssss
56-11/16 in.
I




T
3 in.

                                    \\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\X\\1
SSSSSSSSSSSSSSS/SSSfSSSSSSSSSSS/SS
                                                                   5/16 in.
                                                                  GAS INLET
                                                                                  GAS INLET
                       Figure  19.   Schematic diagram of the kiln burner  fuel injector

-------
                                  TABLE 7.  BASE-LINE DATA FOR CLEAN FUELS:
                                NATURAL GAS AND WELLMAN-GALUSHA OXYGEN (WGO),
                                   3.5 MILLION Btu/hr WITH 10% EXCESS AIR
                                                          Fuel           NOX*             CO
                                  Fuel Type          Temperature, K      	 ppm 	

w               Baffle            Natural Gas             298  .          135              10
00               Burner            WGO                     322             53             200
                                  WGO                     700             38             115

                Kiln              Natural Gas             298             83             150
                Burner            WGO                     322             44             405
                                  WGO                     700'            31              34
                   NO plus NO  (Dry, corrected to 0% excess air).

-------
value was selected to give a stable flame of a size compatible with the  fur-
nace dimensions.  For both ambient and hot WGO the flow was 22% radial,
chosen to give a flame length comparable to that of natural gas.  With the
fuels studied, gas injected radially was at sonic velocity while the axial
component entered at 85 m/s (280 ft/s) £or ambient WGO.  The axial velocity
was 52 m/s  (170 ft/s) for hot WGO on the larger nozzle.  With natural gas
the axial flow was 1.2 m/s (4 ft/s).  Air velocity with the kiln burner was
3.1 m/s  (10 ft/s).
     In order to determine the effects of the various potential sources of
fuel-NO  on the measured NO  levels, char, tar, and ammonia were individually
       X                   X
doped into hot WGO.  The doping system and the char (0.66 weight percent
nitrogen) and tar (0.55 weight percent nitrogen) analyses were presented in
the preceeding section.  The results of the char and tar doping trials are
presented in Tables 8 and 9.  For char, the doping rate varied from 0.13 to
0.86 g/s or 0.4 to 2.7 grains/SCF, while the tar levels of 0.36 to 0.73 g/s
correspond to 1.1 to 2.3 grains/SCF.  The total NO  levels measured for char
                                                  X
or tar are comparable and show a NO  increase from 0 to 30 ppm.  The fact
                                   X
that bound sulfur in the tar (0.47 weight percent sulfur) and char (1.64
weight percent sulfur) is converted to sulfur oxides is evidenced by the
SO. levels.   Increases in NO  levels above undoped thermal NO  levels cannot
  2                         x                  v             x
simply be ascribed to fuel-nitrogen conversion when fuel-sulfur is also pre-
                                          1 R'
sent.   Work of IGT(unpublished) and others   has shown that in turbulent
diffusion flames thermal NO , as well as fuel NO , is enhanced by fuel-
                           x'                   x'              3
sulfur.  In the tar doping studies the problem of char-like residue collect-
ing inside the fuel nozzle was encountered.  On the baffle burner this resi-
due amounted to about 3% by weight of the doped tar, whereas on the kiln
burner the residue was around 7% of the tar input.
     The effects of various levels of fuel-nitrogen, in the form of ammonia,
on the conversion of fuel-nitrogen to NO  are shown in Tables 10 and 11 and
Figures 20 and 21 for the baffle and kiln burners with 10% excess air.
Ammonia was metered into the hot WGO at levels of 0.02 to 1.00 volume percent
of fuel input.  The oxygen level at the flue was maintained at 1.8 +_ 0.1%
(dry analysis), corresponding to 10% excess air.  The fraction of ammonia
converted to NO  decreases as the fraction of ammonia in the fuel increases.
               x
The results are comparable for both burners with the kiln burner giving a
somewhat higher conversion.
                                     39

-------
                           TABLE 8,   CHAR DOPING:
                      WELLMAN-GALUSHA OXYGEN (700 K),
                  3.5 MILLION Btu/hr WITH 10% EXCESS AIR
                    Char Input,
                       g/s
Baffle                 0
Burner                 0.13
                       0.24
                       0.86

Kiln                   0
Burner                 0.13
                       0.29
                       0.67
NOX*
38
36
38
53
31
38
41
66
CO
115
59
100
90
34
30
40
75
so2
0
4
6
29
0
3
6
10
   NO plus NO- (Dry, corrected to 0% excess air).

-------
                           TABLE 9.  TAR DOPING :
                      WELLMAN-GALUSHA OXYGEN (700 K),
                  3.5 MILLION Btu/hr WITH 10% EXCESS AIR
Baffle
Burner
Kiln
Burner
Tar Input,
g/s
0
0.36
0.50
0.71
0
0.38
0.60
0.73
*
N0x
38
65
66
68
31
60
60
68
CO
115
108
93
93
34
50
57
70
S02
0
1
1
2
0
1
1
2
   NO plus NO- (Dry, corrected to 0% excess air).

-------
                   TABLE 10.   AMMONIA CONVERSION TO NOX
                           ON THE BAFFLE BURNER
NH3,
% in fuel
0.02
0.11
0.19
0.38
0.60
0.81
1.03
1.42
0.19
0.38
0.60
0.81
1.03
Excess Air,
%
10
10
10
10
10
10
10
10
20
20
20
20
20
Fuel NO ,*
X
ppm
77
213
312
367
416
443
454
509
305
382
459
489
524
NH Conversion,
J %
52
29
26
15
11
9
7
6
26
16
13
10
8
*  Dry, corrected to 0% excess air.
                                    42

-------
                    TABLE 11.  AMMONIA CONVERSION TO NOX
                             ON THE KILN BURNER
NH3,
% in fuel
0.03
0.10
0.20
0.40
0.63
0.85
1.09
0.20
0.40
0.63
0.85
1.09
Excess Air,
%
10
10
10
10
10
10
10
20
20
20
20
20
Fuel N0xt
ppm
123
243
287
396
462
516
560
291
421
480
563
658
NH Conversion,
J %
60
38
23
16
12
10
8
23
17
12
11
10
*  Dry, corrected to 0% excess air.
                                     43

-------
    15.0
O
UJ
O
<->  10.0
 ro
X
cr
ii-
it.
o
o
o
cr
9:
cj
UJ
Ct
    5.0
     1.0
          O  10% EXCESS AIR

          A  20% EXCESS AIR
      0.0
                  0.2
0.4
0.6
0.8
1.0
                           PERCENT NH3 IN FUEL
 Figure 20.  Ammonia conversion for Wellman-Galusha oxygen fuel gas on the
                 baffle burner with  10%  and 20% excess air

                                  44

-------
   12.0
                                     O  10% EXCESS AIR
                                     A  20% EXCESS AIR
                                                              1.0   I.I
                         PERCENT NH3 IN FUEL
Figure  21.  Ammonia conversion for  Wellman-Galusha oxygen fuel gas on the
               kiln burner with 10% and  20% excess  air
                                 45

-------
     The effects of changing the amount of excess combustion air are also
shown in Tables 10 and 11 and in Figures 20 and 21 for the baffle and kiln
burners with 20% excess air.  Here, ammonia constituted from 0.20 to 1.09
volume percent of the fuel input.  The oxygen level at the flue was held at,
3.3 + 0.1%  (dry analysis) to keep the excess air level at 20%.  A comparison
of Figures  20 and 21 shows that the fraction of ammonia converted to NO.  in-
                                                                       x
creases with an increase in the availability of oxygen.  Again the kiln
burner gives a slightly higher ammonia-to-NO  conversion than the baffle
•burner.
     The effects of ammonia doping on flue oxygen levels were determined by
setting the flue oxygen to the level required for 10% excess air with 1.0%
ammonia addition and then reducing the ammonia level in five steps to 0.2%.
For a given reduction of ammonia the oxygen level rose by an amount consis-
tent with the reaction —
                 NH3 + (0.75 + |)02 = f NO + (1 ~ f)N2 + 1.5 H 0
where f is the fraction of ammonia converted to NO .
                                                  x
     Since raw hot gasifier effluents contain both fuel-nitrogen and fuel-
sulfur the effects of the latter on fuel-nitrogen.conversion to NO  were
                                                                  X
determined by adding various levels of hydrogen sulfide to 1.0% ammonia-doped
hot WGO.  The results for both burners are shown on Figure 22.  Hydrogen
sulfide was metered in from 0.02 to 2.89 volume percent of fuel input.  Oxygen
in the flue was kept at 1.8%, corresponding to the 10% excess air level.  As
can be seen, small amounts of fuel-sulfur significantly enhanced fuel-nitrogen
conversion to NO .  Above about 1.5%, and up to 2.9%, hydrogen sulfide did
                X
not add greatly to further total NO  enhancement.  These results were essen-
                                   X                            n
tially the same for both burners.
     A review of the literature  indicated that the degree of initial fuel/air
mixedness had an appreciable effect on fuel-nitrogen conversion to NO .  With
                                                                     X
1.0% ammonia-doped hot WGO at an excess air level held at 10% on the kiln
burner, the amount of radial flow was varied from 0 to 36% of the total.  The
results, shown in Figure 23, confirmed the effect of initial mixing on fuel-
NO  emissions.  The ammonia conversion to NO  rose sharply from 10% radial
  X                                         X
flow to 36%.  The thermal NOx was not appreciably affected over the same range.

                                    46

-------
                                                       O  BAFFLE BURNER
                                                       A  KILN BURNER
   0.4
Figure  22.
  0.8
.2        1.6        2.0
   H2S IN FUEL,%
2.4
2.8
Effect of H2S on ammonia (1.0% by volume) conversion  to NOX
 with Wellman-Galusha oxygen (10%  excess air)
3.2

-------
          800
  a
   i  O
  -
   J
          300
                                20       30
                             RADIAL FLOW, %
Figure 23.  Effects of radial vs. axial flow on ammonia (1.0%) conversion
     to  NOX with Wellman-Galusha oxygen fuel gas on the kiln burner
                                48

-------
     Following the single-dopant tests, so-called parametric studies were
performed, wherein the various contaminants were added in combinations of two.
The results of the ammonia-plus-tar and ammonia-plus-char parametric trials
are presented in Table 12.  The doping rates were representative of those
used in the single-dopant tests.  The total NO  levels were found to be about
                                              X
5% to 20% higher than expected on a simple additive basis (derived from the
single-dopant studies), indicating some kind of synergistic effect.
     The char-plus-tar parametric and the ammonia-plus-char-plus-tar ("dirty")
doping results are shown in Table 13.   Doping levels were again consistent
with previous tests, and excess air was held at 10%.  Total NO  levels in the
case of char-plus-tar appear to be purely additive while the results of the
"dirty" doping trials show the same kind of NO  enhancement as seen in the
                                              X
ammonia-plus-char and ammonia-plus-tar tests.  Total fuel-nitrogen conversion
to NO  is slightly higher on the kiln burner than on the baffle burner (VLO%
     X
versus ^9%) if one assumes that thermal NO  levels are not greatly affected
                                          X
by the sulfur in the dopants.
WELLMAN-GALUSHA AIR:  EMISSION STUDIES
     As with Wellman-Galusha oxygen (WGO) fuel gas, base-line data were ob-
tained for clean Wellman-Galusha air (WGA) fuel gas at ambient and elevated
temperatures on both the baffle and kiln burners.  These data are presented
in Table 14 along with the base-line natural gas data previously shown in
Table 7.  The fuels were fired at 1.03+0.07 MWfc (3.50 + 0.25 X 10  Btu/hr)
with 10% excess air.  For the baffle burner, natural gas and ambient (322 K)
WGA were burned on the same 0.063-m ID (nominal 2-1/2 inch) fuel nozzle as
WGO.  Hot (616 K) WGA required a 0.078 m ID (nominal 3 inch Schedule 40) fuel
nozzle to overcome the stability problems encountered on the 0.063-m ID noz-
zel.  WGA fuel velocities were 63 m/s (206 ft/s) for ambient WGA and 83 m/s
(273 ft/s) for hot WGA with combustion air at 19 and 30 m/s (62 and 99 ft/s) ,
respectively.
     On the kiln burner, the 0.108-m ID  (nominal 4 inch) axial fuel nozzle
was used for both ambient and hot WGA.  The amount of radial flow was selected
to give flame lengths comparable to that of natural gas; ambient WGA was fired
with 10% radial flow, while hot WGA required 16% radial flow.  Radial gas

                                     49

-------
                       TABLE 12,   PARAMETRIC DOPING :
                      WELLMAN-GALUSHA OXYGEN (700 K),
                  3.5 MILLION Btu/hr WITH 10% EXCESS AIR
Baffle
Burner
Kiln
Burner
Inputs
NH3,,
vol %
0
1.03
0.99
1.00
0
1.09
1.07
0.98
Tar
p/g
0
0
0.39
0
0
0
0.63
0
Char
0
0
0
0.13
0
0
0
0.13
NO
                                                  X
                                                            CO
S0
38
492
580
561
31
591
635
558
	 J^Lll 	
115
60
108
65
34
34
75
33
0
0
2
7
0
0
1
3
   NO plus N09 (Dry, corrected to 0% excess air).

-------
                 TABLE 13.  PARAMETRIC AND "DIRTY" DOPING
                      WELLMAN-GALUSHA OXYGEN (700 K),.
                  3.5 MILLION Btu/hr WITH 10% EXCESS AIR
Baffle
Burner
Kiln
Burner
Inputs
NH3,
vol %
0
1.03
0
1.00
0
1.09
0
1.02
Tar
~ /
0
0
0.
0.
0
0
0.
o.
B/

46
47


48
25
Char
0
0
0.13
0.13
0
0
0.35
0.35
NOV
A
38
492
53
593
31
591
77
635
CO
ppm
115
60
63
75
34
34
45
40
S02
0
0
7
8
0
0
6
7
   NO plus N0_ (Dry, corrected to 0% excess air).

-------
                  TABLE 14.  BASE-LINE DATA FOR CLEAN FUELS:
                  NATURAL GAS AND WELLMAN-GALUSHA AIR (WGA),
                   3.5 MILLION Btu/hr WITH 10% EXCESS AIR
                  Fuel Type          Temperature, K      _ ppm
                                          Fuel           NOX              CO
Baffle            Natural Gas             298            135              10
Burner            WGA                     322             18             260
                  WGA                     616             26              35

Kiln              Natural Gas             298             83             150
Burner            WGA                     322             28             200
                  WGA                     616             22              30
   NO plus NO- (Dry, corrected to 0% excess air).

-------
velocities were sonic while axial fuel velocities varied from 33 m/s (107
ft/s) for ambient WGA to 74 m/s (244 ft/s) for hot WGA.  Air velocity was
3.1 m/s (10 ft/s).
     Only "dirty" doped, hot WGA was studied.  The results are shown in Table
15.  If the additives did not greatly affect thermal NO  levels then the
                              0     J                  x
baffle burner is somewhat more efficient in converting fuel-nitrogen to NO
                                                                          X
than the kiln burner, namely, VL1% versus ^9%, respectively.  This is opposite
to the WGO results, where the kiln burner appeared to be slightly more effi-
cient,  in converting fuel nitrogen.
PARTICULATE STUDIES
     Besides measuring the effects of the various dopants on gas-phase pol-
lutants, total particulates in the stack were measured as well.  The instru-
mentation and sampling technique used were described in a preceding section.
Results of these total stack-particulate measurements are presented in Figure
24 for both WGO and WGA on the baffle and kiln burners.  The fraction of char-
derived particulates surviving in the stack increased from ^5% at a char in-
put of 0.02 g/s to ^11% at a char input rate of 1.1 g/s.  At all char doping
rates,  it was visually noted that a qualitatively large number of glowing
particles were deposited on the furnace hearth.  The presence of ammonia
and/or tar did not affect the measured total stack-particulates from char.
     In addition to total particulates, particle size distributions were
measured for char-doped WGO on the baffle and kiln burners using a cascade
impactor and methods described previously.  The results are shown in Tables
16 and 17.  On the baffle burner, with a char doping rate of 1.14 g/s, 63%
of the particles are smaller than roughly 0.7y, while on the kiln burner,
for a char rate of 0.76 g/s, only about 5% of the particles are below 0.7y.
This shift in particle size depends more on char input level than on burner
type for the cases here studied.  This conclusion is supported by the data
obtained from the total particulate measurements wherein an increase in
"coarse" particles trapped in the cyclone preseparator relative to "fine"
particles staining the filter was observed with decreasing char doping rate
for both burners.
                                     53

-------
Ln

                   Baffle Burner


                   Kiln Burner
                                            TABLE 15.  "DIRTY" DOPING:
                                           WELLMAN-GALUSHA AIR (616 K),
                                     3.5 MILLION Btu/hr WITH 10% EXCESS AIR

NH3,
7
0
1.04
0
0.99
Inputs
Tar
P/s
0
0.58
0
0.45

Char

0
0.35
0
0.13

NOX

26
672
22
522

CO
ppm
35
40
30
35

S02

0
8
0
7
                   *  NO plus NO  (Dry, corrected to 0% excess air).

-------
o»
(/>"
UJ
§
D
O
      O
           0.14
           0.12
           0.10
          0.08
      <   0.06
      GL
          0.04
          0.02
           0.0
            WGOWGA
              O   O  BAFFLE BURNER
              D   A  KILN BURNER
              0.0      0.2      0.4      0.6      0.8
                                 CHAR INPUT, g/s
                                                  1.0
1.2
Figure 24.   Particulate emissions for char doping of Wellman-Galusha fuel
            gases on the baffle and kiln burners (10% excess air)
                                  55

-------
Oi
                                     TABLE 16.  BAFFLE BURNER CASCADE IMPACTOR
                               RESULTS FOR CHAR-DOPED WELLMAN-GALUSHA OXYGEN (700 K) -
                                               DOPING RATE: 1.14 g/s
Stage
0
1
2
3
4
5
6
7
Filter
Size

10.
7.
5.
3.
1.
1.
0.
0

7
3
0
2
8
1
7

(P)
±.17
- 17
- 10
- 7
- 5
- 3
- 1
- 1
- 0

.1
.1
.1
.3
.0
.2
.8
.1
.7
Wt
11.
1.
0.
2.
2.
8.
4.
5.
62.
*
% % Cumulative
1
4
4
9
5
9
7
6
5
88
87
87
84
81
72
68
62
0
.9
.5
.1
.2
.7
.8
.1
.5

                       *  Wt % of particulates passing through stage.

-------
Ul
                                     TABLE 17.  KILN BURNER. CASCADE  IMPACTOR

                              RESULTS FOR CHAR-DOPED WELLMAN-GALUSHA OXYGEN  (700  K)-
                                              DOPING RATE:  0.76  g/s
                                                                                       *
                                        Size  (y)        Wt  %               %  Cumulative
0
1
2
3
4
5
6
7
Filter
>0.2.6
7.9 -
5.3 -
3.6 -
2.4 -
1.2 -
0.7 -
0.5 -
0
12.6
7.9
5.3
3.6
2.4
1.2
0.7
0.5
48.5
4.9
9.2
7.2
5.8
12.1
7.8
0.2
4.3
51.5
46.6
37.4
30.2
24.4
12.3
4.5
4.3
0
                      *  Wt % of particulates passing  through  stage.

-------
FURNACE EFFICIENCY
     Aside from the pollution aspects of burning low-Btu fuel gases as sub-
stitutes for natural gas, potential users of low-Btu gas are also very in-
terested in retrofitting problems that might be encountered.  One phase of
any retrofit evaluation is the effect of firing low-Btu gas on furnace
efficiency and heat transfer to the load.  As noted in an earlier section,
water-cooling tubes were positioned along the furnace wall to simulate the
load of a steel reheat furnace for use with the baffle burner and then to
simulate the calcining and reaction zones of a cement kiln for use with the
kiln burner.
     For the baffle burners, fired at 1.03 MW  (3.5 X 10  Btu/hr), furnace
thermal efficiencies, defined as the heat absorbed by the load divided by the
fuel heat input, for clean ambient and clean hot WGO were found to be 30% and
31% as compared with the natural gas base-line value of 35%.  Average flue
temperatures were measured to be 1541 K (2314°F) for ambient WGO, 1564 K
(2356°F) for.hot WGO, and 1436 K (2126°F) for natural gas.  The fraction of
heat transferred to the load as a function of axial distance down the furnace
for ambient and hot WGO relative to natural gas are shown in Figures 25 and
                                                            i.
26.  Ambient WGO and natural gas exhibit similar heat-transfer curves, while
the hot WGO curve is shifted toward the flue.
     For clean ambient and clean hot WGA on the baffle burner, furnace ther-
mal efficiencies were 24% and 29% with corresponding average flue temperatures
of 1453 K (2156°F) and 1503 K (2246°F).  The heat-transfer curves, presented
in Figures 27 and 28, manifest a shift downstream relative to natural gas for
both fuel temperatures.  These shifts are due to the longer flame lengths
obtained with the low-Btu fuels on the baffle burner.
     With the kiln burner fired at 1.03 MWfc (3.5 X 10  Btu/hr) the base-line
natural gas efficiency was 31% with an average flue temperature of 1532 K
(2298°F).  Clean ambient and clean hot WGO yielded efficiencies of 29% and 30%
with average flue temperatures of 1524 K (2284°F) and 1545 K (2322°F).  As
shown in Figures 29 and 30, the heat transfer curves for ambient and hot WGO
differ somewhat from the natural gas base line and from each other.  The heat —
                                     58

-------
Ul
VO
               0.4
            Q
            LU
            CD

            O  0.3
            LU
               0.2
                o.
                                     322 K WGO

                            	r— NATURAL GAS
WGO FIRING RATE-. 1.06 MWt(3.62 M Btu/hr)

WGO LOAD RATE: 0.32 MWf (109,260 Btu/hr)

WGO EFFICIENCY: 30%
                                0.2
         0.4
0.6
0.8
1.0
                                         DISTANCE FROM BURNER WALL

                                               FURNACE LENGTH
                 Figure 25.  Heat absorption profiles for ambient,  "clean" Wellman-Galusha
                               oxygen and natural gas on the baffle burner

-------
    0.4
o
UJ
m
ac
o

-------
    0.4
O
UJ
CD
CE
O

-------
NJ
               0.4
o
UJ
GO
or

en  0.3
oo
            UJ

-------
CT>
CO
              0.20
s
CD
O 0.15
CD
£ 0.10
p
*
z
o
•h 0.05
                       WGO FIRING RATE: 1.03MWt (3.49M Btu/hr)
                       WGO LOAD' 0.30 MWt (1,005,000 Btu/hr)
                       WGO EFFICIENCY-. 29%
                                                            322 K  WGO
                                                    	NATURAL GAS
                               0.2           0.4         .  0.6
                                      DISTANCE FROM BURNER WALL
                                            FURNACE LENGTH
                                                              0.8
1.0
                 Figure  29.  Heat absorption profile for ambient, clean Wellman-Galusha
                              oxygen and natural gas on the kiln burner

-------
   0.20
Q
o
It?   0.15
UJ
cr
£   0.10
o
   0.05
              WGO FIRING RATE:  1.00  MWf (3.39 M Btu/hr)

              WGO LOAD RATE'  0.30  MWf (1,032,000 Btu/hr

              WGO EFFICIENCY:  30%
                                                        700 K WGO
                                                 	 NATURAL GAS
                     0.2
                                    0.4
0.6
0.8
                                     DISTANCE FROM BURNER WALL

                                           FURNACE LENGTH
1.0
1.2
             Figure 30.  Heat absorption profile for hot, clean Wellman-Galusha oxygen

                                and natural gas on the kiln burner

-------
transfer curves for ambient WGO essentially rises continuously toward the
flue, while for hot WGO the general trend of the curve approximates natural
gas.
     On the kiln burner, clean ambient and clean hot WGA gave furnace ther-
mal efficiencies of 24% and 28% with average flue-gas temperatures of 1479 K
(2203°F) and 1553 K (2336°F).  The heat-transfer curves for ambient and hot
WGA, shown in Figures 31 and 32, are comparable to that of natural gas.  This
similarity in heat-transfer curves is not unexpected because on the kiln
burner all flame lengths were adjusted to approximate that of natural gas.
     In simulating the hot raw gasifier off-gas, the addition of dopants to
the clean fuel might be expected to affect overall furnace thermal efficiency
in two ways.  First, the ammonia, char, and tar are fuels themselves and will
therefore affect the fuel heat input.  Second, char and tar could affect
flame emissivity.
     In the studies conducted, the level of contaminant doping was such that
the maximum contribution to the total fuel heat input was less than 6%.  Flue
temperature measurements were essentially constant for a given fuel/burner
with and without doping, indicating that the doping levels employed did not
significantly affect fuel heat input.  In any case, calculations of furnace
efficiencies included the contributions of doped material to the total fuel
enthalpy.
     For ammonia additions of 1.0 volume percent, no effects on furnace ther-
mal efficiency were observed for WGO or WGA on either burner.  Char addition
of 0.13 g/s (0.4 grains/SCF of WGO, 0.3 grains/SCF of WGA) also had no effect
on efficiency for both fuels on the burners studied.
     At a tar feed rate of 0.58 g/s (1.8 grains/SCF of WGO, 1.4 grains/SCF of
WGA), furnace thermal efficiencies for WGO and WGA were increased by about
1.0% to 2.0% on both burners.  This enhancement is probably due to the In-
crease in flame luminosity that was visually observed.  The heat-transfer
curve obtained is basically the same as for the undoped fuels.
                                     65

-------
   0.20
UJ
I
o
u
CD
a:
WGA FIRING RATE'1.07 MWt (3.63 M Btu/hr)

WGA LOAD RATE'- 0.26 MWt (871,200 Btu/hr)

WGA EFFICIENCY^ 24%
    0.15
5   0.10
UJ
X

u.
o
g

§
IT
   0.05
                                      322 K WGA

                                      NATURAL GAS
     0.0
       0.0
         0.2           0.4           0.6

                 DISTANCE FROM BURNER WALL

                       FURNACE LENGTH
0.8
1.0
      Figure 31.  Heat absorption profile  for Wellman-Galusha air fuel gas com-

                     pared to natural gas on the kiln burner

-------
o
W
S  0.15

ui
5  0.10
&

o

g  0.05
o
           WGA FIRING RATE« 0.98 MWt (3.32MBtu/hr)
           WGA LOAD=0.27MWt (9l7,400Btu/hr)
           WGA EFFICIENCY.28%
616  K WGA
NATURAL
                    0.2
    0.4           0.6
DISTANCE FROM BURNER WALL
      FURNACE LENGTH
            0.8
1.0
      Figure 32.   Heat absorption profile for hot, clean Wellman-Galusha  air and
                          natural gas on the kiln burner

-------
                                DISCUSSION
FUEL-NITROGEN EFFECTS ON NOY
                           A.



     In premixed flames, NO  levels associated with thermal  fixation of'-.atmp-
                           X

spheric nitrogen are dependent primarily on flame temperature and, second-.,

                                       4                •
arily, 'on the amount of combustion air.   In turbulent diffusion  flames,  ther-


mal NO  has also been found to vary with the initial degree  of  fuel/air mix-

 -•,  'i *  ' ..     '                             •
ing;.   For the fuels and burners studied, base-line thermal  NO  levels, pre-
                                                              X

sented in Tables 7 and 14, are mainly ordered by adiabatic flame  temperatures.


Mixing effects on thermal NO  associated with the different  aerodynamic char-
                            X

acteristics of the two burner types appear to be similar except for 'natural


gas where a higher thermal NO  level is found with the baffle burner- than with
                             X

the kiln burner, suggesting (after Reference 1) that natural gas/air mixing


is somewhat better on the baffle burner.



    - Another source of NO  in combustion is chemically bound..nitrogen  in  the


fuel.  .Since.fuel-nitrogen bonds are much weaker than the bond  in molecular
ndtrqgen, fuel-nitrogen can give rise to higher amounts of NO^ than from  ther-


                                      nt, fi

                                      4,14.
             14                                  ,           x
mal fixation.' '   In a flame environment, fuel-nitrogen is generally believed
to react through the competitive paths



                           NH. + Ox = NO                                 (1)



                           NH. + NO = N2                                 (2)



where NH^ is some fuel-nitrogen intermediate, usually considered to be atomic


nitrogen-  or a cyano or amine derivative,   and Ox is an oxygen-containing


species such as 0, OH, or 0 ,  In fuel-lean combustion, fuel-nitrogen appears
                           ^                                •

in the exhaust gases mostly as NO and N_; under fuel-rich conditions, signifi-

                                   17
cant HCN and NH,, can also be found.



     Factors that affect fuel-nitrogen conversion to NO  are availability of
                                                       X

oxygen, initial fuel-nitrogen concentration, temperature, and general fuel

     4 17
type. '    As the amount of oxygen available for combustion increases,. the

                                             18 4
conversion of fuel-nitrogen to NO  increases.  '   Where f-uel/air mixing is
                                 X

incomplete, this conversion is strongly affected by local oxygen concentra-

                                          13
tions as well as local flame temperatures.   Combustion of  fuel-nitrogen under
                                      68

-------
locally fuel-rich conditions can lower the amount of NO  formed   with the

                                        4              X
fuel-nitrogen preferentially forming N .    As initial fuel/air mixing is im-

                                      L                               1
proved, the conversion efficiency of fuel-nitrogen to NO  is enhanced.   For

                  10                                    X
example, one group   reported a doubling in fuel-NO  when going from diffu-
                                                   X

sional to premixed combustion.



     With increasing levels of fuel-nitrogen, the fraction converted to NO


decreases  '   even though the absolute amount increases.  Unlike thermal NO ,

                                                                 4          x
temperature  does not greatly affect fuel NO  in premixed flames >  probably

                                            X         1
because overall fuel-nitrogen reactions are exothermic  and are therefore less


temperature dependent.



     Recent results   imply that fuel-nitrogen conversion to NO  depends on
                                                               X

fuel type; i.e., ammonia conversion to NO  in hydrocarbon combustion was


found to be much greater than with a hydrogen/carbon monoxide fuel.  The au-


thors attributed this to a difference in the intermediate fuel-nitrogen spe-


cies (NH^ in Equation 1), depending on whether the main fuel was a hydrocarbon


or not.17



     In order to gauge the magnitude of several of the above parameters on


fuel-nitrogen conversion to NO  for raw gasifier effluents, char, tar, and
                              X

ammonia were added to hot clean Wellman-Galusha fuel gases.  To ascertain the


contribution of each dopant to total NO  emissions, single dopant and combi-
                                       X

nation tests were also performed.  The tests were performed using two indus-


trial burners: a forward flow and a kiln burner.



     With char-doped hot WGO on both burners, we have seen (Table 8) that


total NO  levels were not greatly increased above thermal at the char feed
        X

rates employed (maximum increase 30 ppm).  The contribution of char-nitrogen


(0.66 weight percent) to these increments is difficult to interpret, owing to


the small change (if any) over thermal NO  relative to measurement reproduc-
                                         X

ibility (+ 5 ppm), and considering the known, but in this case unmeasurable,


enhancement of thermal NO  by fuel-sulfur (1.64 weight percent of char) in

                         x 18
turbulent diffusion flames.    Measured SO  at the flue entrance accounted for


about 50% of the char-sulfur in all cases.  Attempts to close this sulfur


balance were unsuccessful.
                                     69

-------
     As with char, hot WGO tar doping results, shown in Table 9, cannot be

unambiguously analyzed because of the presence of sulfur in the tar (0.47

weight percent).  For the tar feed rates employed, measured NO  exceeded the
                                                              X
clean thermal values by about 30 ppm on both burners.  This increase in NO

cannot be solely accounted for by tar-nitrogen (0.55 weight percent) even

assuming a 100% conversion to NO .  The implication is that tar-sulfur is
                                X
enhancing thermal NO .  The contributions to total NO  of sulfur-enhanced
                    X                                X
thermal and fuel NO  cannot be apportioned.  If all tar-nitrogen were con-
                   X
verted to NO , then sulfur enhancement of thermal NO  could be up to 10 ppm
            X                                       X
over the clean, undoped value.  Conversely, enhanced thermal NO  could be much
                                                               X        '  '
higher, with fuel NO  contributing relatively little to the observed increase.
                    X
Measured SO  accounted for only 20% of tar-sulfur in all cases, while the tar

residue trapped in the fuel nozzles accounted for another 10% of the fuel-
sulfur.

     When fuel sulfur is absent, as in the case of ammonia doped hot WGO, the

effects of fuel-nitrogen on NO  are more clearly evident.  Varying the amount
                              X
of ammonia in the fuel shows that the fraction of fuel-nitrogen converted to

NO  decreases with increasing fuel-nitrogen content, as can be seen from
  X
Figures 20 and 21, even though absolute NO  levels increased.  This observa-
                                          X
tion is in agreement with the references cited at the beginning of this sec-

tion.  For example, on the baffle burner at 10% excess air, ammonia doped at

1.0 volume percent of fuel input yielded a  7% conversion to NO , while 0.4%
                                                               X
ammonia gave a 14% conversion.  With the kiln burner, the corresponding con-

versions were 8% and 16%.

     Examination of Figures 20 and 21 shows that an increase in excess air,

from 10% to 20%, enhanced fuel-nitrogen conversion to NO  as expected.  On
                                                        X
the baffle burner with 1.0% ammonia in hot WGO, the conversion efficiency to
NO  increased from 7% to 8% with the increase in excess air.  At the 0.4%
  x
ammonia level, the conversion went from 14% to 16%.  Similar results were ob-

tained with the kiln burner.  For 1.0% and 0.4% ammonia, the respective in-
creases in conversion with increased excess air were 8% to 10% and 16% to 17%.

     At the radial flow chosen to give the proper flame length for the kiln
burner (22% of the total hot WGO flow), fuel NO  is only slightly higher than


                                     70

-------
for the baffle burner at comparable ammonia doping rates and excess air.   In
other words, the anticipated mixing/aerodynamic effects of the different
burner types were not evident at the operating conditions employed.  However,
changing the amount of radial flow drastically affected ammonia conversion to
NO ,  as can be seen from Figure 23.  Increasing radial flow from 22% to 36%
  X
resulted in about a 50% increase in ammonia conversion, indicating that improved
fuel/air mixing raises fuel-nitrogen conversion, as expected from the brief
literature survey presented earlier.  (This effect was also confirmed when
fuel-nitrogen conversion was found to be 50% higher on a highly mixed high-
forward-momentum burner than on the kiln and baffle burners in other tests
done at IGT.)  Raising the radial flow from 0 to 15% lowered fuel NO .  In
                                                                    X
this region, the apparent loss and then recovery of fuel/air mixedness is
probably due to a trade-off between increasing radial mixing and decreasing
axial fuel momentum, the net effect of which is to decrease overall mixing up
to about 10% radial, where radial flow becomes the dominant mixing parameter
due to the radial flow penetration of the axial flow.
     Since the ammonia-to-NO  conversion is nearly the same for the baffle
                            x                    J
burner and the kiln burner (22% radial) with the same doping rate, it may be
inferred that the baffle burner gave about the same degree of initial hot WGO/
air mixing as the kiln burner at 22% radial flow.
FUEL-SULFUR EFFECTS ON NO
                         x
     As noted in the discussion of the char and tar results, fuel-sulfur is
known to affect thermal NO .  In turbulent diffusion flames, characterized by
                          x                                                 J
relatively poor initial fuel/air mixing, fuel-sulfur enhances thermal NO
                                              18                        "
while in premixed flames an inhibition occurs.    Besides thermal NO  fuel-
sulfur also affects fuel NO .  In premixed flames it may enhance, inhibit, or
                           X
have no effect on fuel NO  depending on the point of sampling and/or the burn-
                         X
ing mixture's residence time in the combustion apparatus, while in turbulent
                                                                18
diffusion flames fuel-sulfur has been found to enhance fuel NO .
                                                              x
     In order to determine the effects of various levels of fuel-sulfur
(hydrogen sulfide) on fuel-nitrogen conversion to NO , hot WGO, doped with
                                                    X
1.0 volume percent ammonia, was fired on both burners with the results shown
in Figure 22'.  Neglecting fuel-sulfur/thermal NO  interactions, the antici-
                                                X
pated enhancement of fuel-nitrogen conversion during turbulent diffusion
                                     71

-------
combustion is evident at hydrogen sulfide levels of 0.5 to 2.9 volume percent
fuel input.  Fuel-sulfur effects on fuel NO  are essentially the same for both
                                           X
burners.  This is not surprising, since the kiln burner, operated at 22% radi-
al flow, appears to give the same degree of initial hot WGO/air mixing as the
baffle burner.  This was also implied by the ammonia doping tests as pre-
viously noted.  Further, as fuel-sulfur levels are increased the enhancement
of NO  app.ears to reach a maximum, as suggested by Figure 22.
     X
     On both burners, measured S09 corresponds to about 80% of the sulfur in-
put (as hydrogen sulfide) at all doping rates.  The fate of the remaining
sulfur  is uncertain.  If it were present as some other species, a possible
candidate is 803 .  Although equilibrium considerations predict negligible
amounts of SO    in hot (T >_ 1300 K) combustion gases, relatively high con-
centrations of S0~ are possible under combustion conditions, where rapid
                                            1 f\ ?
cooling of combustion gases can "freeze" SO.,  '  at superequilibrium values.
                                                            o
Even so, reported SO  levels are usually only a few percent,  though levels
                                  7
as high as 10% have been recorded.   Such high levels are possible where
quenching of SO  takes place by rapid cooling or by short residence time in
the combustion chamber.
     With hot WGO fired on the baffle and kiln burners, combustion takes place
by turbulent diffusion, resulting in wide variations in local species concen-
trations and temperatures.   The S02, readily formed from the added t^S ,  "
might form SO  in two ways: 1) by reaction with 0 atoms in high temperature,
fuel-lean regions; and 2) in lower temperature regions (T _> 1000 K) where the
right-hand-side of the equilibrium process, SO  + 1/2 09 = SO.,, is not
                 1 fi
negligibly small.    Although the S0,/S09 approach to equilibrium is slow at
                  3
lower temperatures , the presence of NO  can catalyze the formation of SO,
,3,9                                  X
by
                           NO + 1/2 0  + NO                               (3)
                           NO  + SO  + SO  + NO                           (4)
     For the tests performed with ammonia and hydrogen sulfide-doped hot WGO,
high concentrations of NO  were present, making more plausible the possibility
                         X
of high SO .  That S09 levels measured were not severely depressed by some
          -3          £
artifact of the sampling system/instrumentation is supported by subsequent IGT

                                     72

-------
tests performed on a high-forward-momentum burner where measured SO. accounted
for 95% of the hydrogen sulfide added to a 1.0 volume percent ammonia-doped
low-Btu fuel (ambient WGO plus 25% N ).   Since the mixing characteristics of
this kind of burner are superior to the baffle or kiln (22% radial) burners,
one would expect less low-temperature formation of SCL if Reaction 4 is of any
importance.  The results are in good qualitative agreement with this tentative
mechanism, though more research on this possibility is required before any
definite conclusions can be made.
                                     73

-------
                                  REFERENCES

 1.  Appleton, J. P. and Heywood, J. B. Fourteenth Symposium (International)
     on Combustion, The Combustion Institute, 1973.  pp. 777.

 2.  Chigier, N. A. Prog. Energy Combust. Sci: 1, 3, 1975.

 3.  Cullis, C. F. and Mulcahy, M. F. R. Combustion and Flame,  18: 225, 1972.

 4.  DeSoete, G. G. La Rivista dei Combustibli, 29: 35, 1975.

 5.  Feldkirchner, H. L. and Schora, F. C., Jr. Coal Desulfurization Aspects
     of the HYGAS™ Process.  Paper presented at the Second International
     Conference on Fluidized Bed Combustion, Hueston Woods State Park, Ohio,
     Oct. 4-7, 1970.

 6.  Haynes, B. S. Combustion and Flame, 28: 81, 1977.

 7.  Hedley, A. B.  J. Institute of Fuel, 40: 142, 1967.

 8.  IGT Process Research Division. HYGAS®: 1964 to 1972, Pipeline Gas From
     Coal Hydrogenation (IGT Hydrogasification Process).  Final Report,
     FE-381-T9-P3, Washington, D.C., July 1975.

 9.  Levy, A., Merryman, E. L., and Reid, W. T. ES&T, 4: 653, 1970.

10.  Lisauskas, R. A. and Johnson, S. A. NOX Formation During Gas Combustion.
     CEP, Aug. 1976.  p. 76.

11.  Matthews, R. D., Sawyer, R. F., and Schefer, R. W. ES&T, 11 (12): 1092,
     1977.

12.  Merryman, E. L. and Levy, A. Fifteenth Symposium (International) on
     Combustion, The Combustion Institute, 1975.  pp. 1073.

13.  Sarofim, A. G., Pohl,  J. H., and Taylor, B. R. Mechanisms and Kinetics
     of NOX Formation: Recent Developments.  69th Annual Meeting, AIChE,
     Nov. 30, 1976.

14.  Seery, D. J. and Zabielski, M. F. Combustion and Flame, 28: 93, 1977.

15.  Shoffstall, D. R. and  Waibel, R. T. Burner Design Criteria for NOX Control
     From Low-Btu Gas Combustion.  EPA Final Report, EPA-600-7-77-094b,
     Dec. 1977.

16.  Sternling, C. V. and Wendt, J. 0. L. Kinetic Mechanisms Governing the
     Fate of Chemically Bound Sulfur and Nitrogen in Combustion.  EPA Final
     Report, PB-230895, Aug, 1972.

17.  Takagi, T., Ogasawara, M., Daizo, M., and Tatsumi, T. Sixteenth Symposium
     (International) on Combustion, The Combustion Institute, 1977.  pp. 181.


                                      74

-------
18.  Wendt, J. 0. L., Corley, T. L., and Morcomb, J. T.  Interaction Between
     Sulfur Oxides and Nitrogen Oxides in Combustion Processes.   Second
     Symposium on Stationary Sources Combustion, New Orleans, Aug. 29-Sept. 1,
     1977.

19.  Wendt, J. 0. L. and Edmann, J. M. Effect of Sulfur Dioxide and Fuel Sulfur
     on Nitrogen Oxide Emissions.  EPA Progress Report,  Grant R-802204,
     Sept. 1974.
                                       75

-------
            CONVERSION TABLE



ENGLISH TO SI METRIC CONVERSION FACTORS
To Convert From
106
Btu
SCFH
ft/s
Inch
Feet
Feet
Inch
°F
J
g
s
MWt
m
M
F
K
SCFH
Btu/hr





2
2

• Joule
° gram
= second
= megawatts
= metre
= mega (106
= Fahrenhei
= Kelvin
= standard
To
MWt
J
m3/s
m/s
m
m
m2
m2
K



thermal

)
t

cubic feet per hour
Multiply
2.928751 E
By
-01
1.055 E E + 03
7.865790 E
3.048000 E
2.540000 E
3.048000 E
9.290304 E
6.451600 E
t°C = (t°F









-06
-01
-02
-01
-02
-04
- 32J/1.8









                   76

-------
                                TECHNICAL REPORT DATA
                         (Please read Instructions on the reverse before completing)
1. REPORT NO.
 EPA-600/7-78-191
                           2.
                                                      3. RECIPIENT'S ACCESSION NO.
4. TITLE ANDSUBTITLE
 Pollutant Emissions from "Dirty" Low- and Medium-
 Btu Gases
            5. REPORT DATE
             October 1978
            6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)

R.T. Waibel, E.S. Fleming, andD.H. Larson
                                                      8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
 Institute of Gas Technology
 Applied Combustion Research
 IIT Center, 3424 South State Street
 Chicago, Illinois  60616	
                                                       10. PROGRAM ELEMENT NO.
            EHE624A
            11. CONTRACT/GRANT NO.
             68-02-2643
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 PERIOD COVERED
            Final; 5/77-4/78
            14. SPONSORING AGENCY CODE
              EPA/600/13
15. SUPPLEMENTARY NOTES ffiRL-RTP project officer is David G. Lachapelle, Mail Drop 65,
919/541-2236.  EPA-600/7-77-094a and -094b are earlier reports in this series.
is. ABSTRACTrj.^ report gives results of a. study to determine the emissions from 'dirty'
low- and medium-Btu gases when combusted on industrial process burners. The fuels
tested were blends with compositions typical of Wellman^Galusha oxygen (WGO) and
air (WGA) fuel gases. Baseline data were collected for natural gas, ambient WGO and
WGA, and hot WGO (700 K) and WGA (616 K). Ammonia, H2S, coal tar, and char were
added to the hot fuels  to determine their effects  on pollutant emissions. Study conclu-
sions include: (1) low-Btu fuels not subjected to post-gasifier  cleanup can yield NOx
levels greatly above the thermal levels for the clean fuels and for natural gas; (2)  in
turbulent diffusion flames, fuel-NOx increases with an increase in a) the amount of
fuel-nitrogen, b) the amount of fuel-sulfur, c) the level of excess air,  and d) the de-
gree of initial fuel/air mixing; (3) attempts  to close the fuel-sulfur balance were
unsuccessful (further  work should be done in this area); and (4) compared to natural
gas, heat transfer to the load  is reduced for the low-Btu fuels tested (this heat trans-
fer is not greatly affected by the presence of contaminants—tar and char--at levels
characteristic of raw gasifier effluents).
17.
                             KEY WORDS AND DOCUMENT ANALYSIS
                DESCRIPTORS
                                          b.lDENTIFIERS/OPEN ENDED TERMS
                         c. COSATl l-'icld/Group
Air Pollution
Nitrogen Oxides
Fuels
Burners
Industrial Processes
Combustion
Additives	
Air Pollution Control
Stationary Sources
Low-Btu Gas
Medium-Btu Gas
Dopants
13 B
07B
2 ID
ISA
13H
21B
11G
18. DISTRIBUTION STATEMENT
 Unlimited
                                           19. SECURITY CLASS (ThisReport)
                                           Unclassified
                         21. NO. OF PAGES

                           84
20. SECURITY CLASS (This page I
Unclassified
                         22. PRICE
EPA Form 2220-1 (9-73)
                                         77

-------