78-1521!
Eiwlrewuwtal Pratwtiw Teclmtoiy Serfes
EPA
6002/2
76
152B
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into jfive series. These five broad
categories were established to facilitate further development and application of
environmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The five series are: i
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research j ;
4. Environmental Monitoring !
5. Socioeconomic Environmental Studies i
This report has been assigned to the ENVIRONMENTAL PROTECTION
TECHNOLOGY series. This series describes research performed to develop and
demonstrate instrumentation, equipment and rpethddology to repair or prevent
environmental degradation from point and non-point sources of pollution. This
work provides the new or improved technology rejquired for the control and
treatment of pollution sources to meet environmental quality standards.
EPA REVIEW NOTICE
j • ; - . - •:
This report has been reviewed by the JU. 8. Environmental
Protection Agency, and approved for publication. Approval
does not signify that the contents necessarily reflect the
views and policy of the Agency, nor does mention of trade
names or commercial products constitute endorsement or
recommendation for use. '.
This document is available to the public throjjgh the National Technical Informa-
tion Service, Springfield. Virginia 22161.
-------
EPA-600/2-76-152b
June 1976
PROCEEDINGS OF THE
STATIONARY SOURCE COMBUSTION SYMPOSIUM
VOLUME H--FUELS AND PROCESS RESEARCH AND DEVELOPMENT
JoshuaS. Bowen, Chairman
Robert E. Hall, Vice-Chairman
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, NC 27711
ROAPNo. 21BCC
Program Element No. 1AB014
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
. ...-lESTirN AGENCY
-------
I
L
-------
PREFACE
The Stationary Source Combustion Symposium was held on September
24-26, 1975, at the Fairmont Colony Square Hotel in Atlanta, Georgia.
The symposium was sponsored by the Combustion Research Branch of E.P.A.'s
Industrial Environmental Research Laboratory (IERL). The Combustion
Research Branch has been involved in developing improved combustion
technology for the reduction of air pollutant emissions from stationary
sources, and improving equipment efficiency.
Dr. Joshua S. Bowen, Chief, Combustion Research Branch, was Symposium
Chairman; Robert E. Hall, Research Mechanical Engineer, Combustion Research
Branch, was Symposium Vice Chairman and Project Officer. The Welcome
Address was delivered by Dr. John K. Burchard, Director of the Industrial
Environmental Research Laboratory. Frank Princiotta, Acting Director of
the Energy Processes Division of E.P.A.'s Office of Energy, Minerals,
and Industry, was the Keynote Speaker.
The Symposium consisted of four Sessions:
Session I; Fundamental Research
Co-chairmen: Dr. Joshua A. Bowen
W. Steven Lanier, Research Mechanical Engineer, E.P.A.,
IERL, Combustion Research Branch
iii
-------
Session II; Fuels Research andDevelopment
Chairman: G. Blair Martin, Chemical Engineer, E.P.A., IERL,
Combustion Research Branch
Session III; Process Research and Development
Chairman: David G. Lachapelle, Research Chemical Engineer,
E.P.A., IERL, Combustion Research Branch
Session IV: Field Testing and Surveys
Co-chairmen: Robert E. Hall
John H. Wasser, Research Chemical Engineer, E.P.A.,
Combustion Research Branch
These Session Chairmen have reviewed the transcriptions of the
question and answer sessions, and, in addition, have worked with authors
to clarify and revise presentations, where appropriate, and to make them
clear and meaningful for these printed proceedings.
We are grateful for the cooperation of Marjorie Maws, Project Leader;
Anita Lord, Symposium Administrator; and Margaret Kilburn, Program Director,
of Arthur D. Little, Inc., who coordinated the symposium for E.P.A.
IV
-------
CONTENTS
Preface ....
Welcome Address
Keynote Address
ill
J.K. Burchard 1-1
F.T. Princiotta 1-3
SESSION I - FUNDAMENTAL RESEARCH
Formation of Soot and Polycyclic Aromatic Hydrocarbons (PCAH)
in Combustion Systems 1-18
J.D. Bittner, G.P. Prado, J.B. Howard, R.A. Kites
Questions and Answers 1-32
Effects of Fuel Sulfur on Nitrogen Oxide Emissions 1-35
J.O.L. Wendt, J.M. Ekmann
Questions and Answers 1-88
Two-Dimensional Combustor Modeling 1-91
R.C. Buggeln, H. McDonald
Questions and Answers (n/a)
Effects of Interaction Between Fluid Dynamics and Chemistry on
Pollutant Formation in Combustion 1-109
C.T. Bowman, L.S. Cohen, L.J. Spadaccini, F.K. Owen
Questions and Answers 1-119
Fate of Coal Nitrogen During Pyrolysis and Oxidation 1-125
J.H. Pohl, A.F. Sarofim
Questions and Answers 1-147
-------
A Detailed Approach to the Chemistry of Methane/Air Combustion:
Critical Survey of Rates and Applications
V.S. Engleman
Questions and Answers
1-153
1-182
Chemical Reactions in the Conversion of Fuel Nitrogen to NOx 1-185
A.E. Axworthy, G.R. Schneider, V.H. Dayan
Questions and Answers 1-211
Prediction of Premixed Laminar Flat Flame Kinetics, Including
the Effects of Diffusion 1-217
R.M. Kendall, J.T. Kelly, W.S. Lanier
Questions and Answers 1-266
Estimation of Rate Constants 1-267
S.W. Benson, R. Shaw, R.W. WooIfoik
Questions and Answers (n/a)
Production of Oxides of Nitrogen in Interacting Flames 1-291
C. England
Questions and Answers 1-316
Concurrent Panel Discussions:
1. Combustion Chemistry and Modeling: An Overview
A.F. Sarofim - Combustion Chemistry and Modeling 1-325
Questions and Answers 1-345
T.J. Tyson - The Mathematical Modeling of Combustion Devices .... 1-347
Questions and Answers 1-410
vi
-------
2. Federal, Regional, State, and Local Air Pollution
Regulations: An Overview
S. Cuffe - Federal 1-413
G.T. Helms - Regional 1-429
R.H. Collum - State 1-443
H.W. Poston - Local I- 447
SESSION II - FUELS RESEARCH AND DEVELOPMENT
Assessment of Combustion and Emission Characteristics of
Methanol and Other Alternate Fuels II-3
G.B. Martin
Questions and Answers 11-30
Burner Design Criteria for Control of Pollutant Emissions
from Natural Gas Flames 11-31
D.F. Shoffstall
Questions and Answers (n/a)
Integrated Low Emission Residential Furnace 11-81
L.P. Combs, W.H. Nurick, A.S. Okuda
Questions and Answers . 11-101
The Control of Pollutant Emissions from Oil Fired
Package Boilers 11-109
M.P. Heap, T.J. Tyson, E. Cichanowicz, R.E. McMillan, F.D. Zoldak
Questions and Answers 11-160
Pilot Scale Investigation of Catalytic Combustion Concepts for
Industrial and Residential Applications 11-163
J.P. Kesselring, R.M. Kendall, C.B. Moyer, G.B. Martin
Questions and Answers . 11-196
vii
-------
The Optimization of Burner Design Parameters to Reduce NOx
Formation in Pulverized Coal and Heavy Oil Flames ...... 11-197
M.P. Heap, T.J. Tyson, G.P. Carver, G.B. Martin, T.M. Lowes
Questions and Answers 11-239
Pilot Scale Investigation of Combustion Modification Techniques
for NOx Control in Industrial and Utility Boilers 11-241
R.E. Brown, C.B. Moyer, H.B. Mason, D.G. Lachapelle
Questions and Answers 11-267
SESSION III - PROCESS RESEARCH AND DEVELOPMENT
Overfire Air as an NOx Control Technique for Tangential
Coal-Fired Boilers Ill-3
A.P. Selker
Questions and Answers Ill-26
Control of NOx Formation in Wall Coal-Fired Boilers 111-31
G.A. Hollinden, J.R. Crooks, N.D. Moore, R.L. Zielke,
C. Gottschalk
Questions and Answers Ill-77
The Effect of Additives in Reducing Particulate Emissions
from Residual Oil Combustion III-S3
R.D. Giammar, H.H. Krause, A.E. Weller, D.W. Locklin
Questions and Answers Ill-115
System Design for Power Generation from Low Btu Gas Boilers Ill-119
M.P. Heap, T.J. Tyson, N.D. Brown
Questions and Answers (n/a)
viii
-------
SESSION IV - FIELD TESTING AND SURVEYS
The Effect of Combustion Modification on Pollutants and
Equipment Performance of Power Generation Equipment IV-3
A.R. Crawford, E.H. Manny, M.W. Gregory, W. Bartok
Questions and Answers . IV-109
Analysis of Gas-, Oil-, and Coal-Fired Utility Boiler
Test Data IV-115
O.W. Dykema, R.E. Hall
Questions and Answers IV-161
Influence of Combustion Modifications on Pollutant Emissions
from Industrial Boilers IV-163
G.A. Cato, L.J. Muzio, R.E. Hall
Questions and Answers IV-219
Emission Characteristics of Small Gas Turbine Engines IV-227
J.H. Wasser
Questions and Answers IV-252
Systems Evaluation of the Use of Low-Sulfur Western Coal in
Existing Small- and Intermediate-Sized Boilers
K.L. Haloney
Questions and Answers
IV-255
IV-316
A Survey of Emissions Control and Combustion Equipment Data in
Industrial Process Heating IV-321
P.A. Ketels, J.D. Nesbitt, D.R. Shoffstall, M.E. Fejer
Questions and Answers (n/a)
ix
-------
POM and Particulate Emissions from Small Commercial
Stoker-Fired Boilers IV-411
R.D. Glammar, R.B. Engdahl, R.E. Barrett
Questions and Answers IV-439
Concluding Remarks - J.S. Bowen IV-441
Appendix
List of Speakers A-l
List of Participants
Alphabetically by Name A-2
Alphabetically by Organization A-6
-------
SESSION II
FUELS RESEARCH AND DEVELOPMENT
G. B. Martin, Chairman
II-l
-------
-------
ASSESSMENT OF COMBUSTION AND EMISSION
CHARACTERISTICS OF METHANOL AND
OTHER ALTERNATE FUELS
by
G. Blair Martin
Combustion Research Branch
Energy Assessment and Control Division
Industrial Environmental Research Laboratory
U. S. Environmental Protection Agency
Research Triangle Park, North Carolina
II- 3
-------
INTRODUCTION
The Combustion Research Branch (CRB) of the EPA's Industrial Envir-
onmental Research Laboratory at Research Triangle Park, N. C., has the
responsibility for carrying out a combustion modification R&D program
directed toward control of nitrogen oxides and other pollutants from
stationary combustion sources. In addition to pollutant emission con-
trol it appears that the technology can also lead to equal or improved
efficiency of energy utilization. The majority of the R&D is carried
out under EPA contracts initiated and directed by CRB and performed by
private organizations. However, CRB also maintains an in-house research
program in a number of areas. This in-house activity provides direct
project officer expertise in combustion research and provides the cap-
ability for initial evaluation of potential control techniques or
promising novel concepts applicable to combustion systems. The results
of many of these studies are well documented.
For several years, one in-house project has been devoted to estab-
lishing the combustion and emission characteristics of alternate fuels
(e.g., alcohol fuels, low Btu gas), an area that is now receiving
increased emphasis in the United States. The goals of this project
have been threefold: first, to determine the potential advantages or
problems related to the combustion of these fuels; second, to obtain
some understanding of the important factors governing the combustion
II-4
-------
and emission characteristics and of potential control techniques; and,
finally, to define necessary R&D for application of these fuels in
practical combustion systems. The supply of alternate fuels available
for evaluation has been quite limited to date; however, one class,
the alcohol fuels, is quite readily simulated with chemical grade pure
alcohols. Therefore the combustion and emission characteristics of
methanol and methanol-isopropanol mixtures have received the primary
emphasis to date.
The purpose of this paper is to summarize the results of the
alcohol fuel testing and to describe briefly the future direction of
the project.
ALCOHOL FUELS
This section provides a short background on sources of alcohol
fuels, a summary of the available information on combustion and
emission characteristics of alcohol fuels as compared to conventional
fuels, and information on a recent series of tests on emulsions of
methanol and distillate oil.
Background
The bulk of the methanol produced in the United States today is
II-5
-------
synthesized from natural gas for end use as a chemical. There is obvi-
ously little benefit to be derived from using this type of alcohol as a
fuel; however, other approaches may be attractive. It has been shown
that it is feasible to produce methanol from a synthesis gas generated
by coal gasification. In addition, Duhl has proposed a Methyl
Fuel, consisting of about 90% methanol and the balance higher alcohols,
as a more economical and efficient synthesis process for a variety of
starting materials. Finally, biological synthesis from waste materials
may be possible. The ultimate place of alcohols in the alternate fuel
picture will depend on a number of factors, including: fuel cost, over-
all energy efficiency relative to other fuels, and environmental con-
siderations of production and use. It would appear that alcohols might
be in competition with other clean high Btu fuels (e.g., Synthetic
Natural Gas) to supplement or replace dwindling supplies of natural gas
and light fuel oils in combustion equipment requiring clean fuels.
Although there is a large body of information on the use of alcohol
fuels in automobiles, the combustion characteristics of these fuels in
continuous combustion devices of the type used in stationary sources
(2)
have received very little attention. Duhl published very limited
data for a small boiler which showed some promising results; however,
little detail was given. This paucity of detailed data led to this
study.
II-6
-------
Summary of Combustion Results
The studies were directed toward defining the combustion and emis-
sion characteristics of alcohol fuels and understanding the critical
phenomena involved. Most of the results have been reported in detail
(3)
elsewhere and are summarized in this section. In addition, the ex-
periments were recently extended to examine the use of alcohols as a
supplement to conventional fuels, similar to the methanol-gasoline
blends being considered for automotive use. The approach taken was
to study emulsions of distillate oil and methanol.
Exp^er imenta 1 _E_quiipmen t
The experimental equipment was also described in detail in refer-
ence 3, and the key features are summarized here. The basic combustion
chamber is a refractory lined cylindrical chamber 0.267 meters in diam-
eter and 1.22 meters long. The nominal heat input was maintained con-
stant at 88,000 watts for all fuels and the combustion air was 115% of
theoretical for all cases. The burner is a movable block swirl burner
(4)
of well known design and is capable of controlling the ratio of axial
to tangential momentum of the combustion air over a wide range. Liquid
fuels were atomized through high pressure nozzles and propane was intro-
duced through a radial hole injector. Flue gas recirculation was simu-
lated by withdrawing cooled product gas from the stack and r-ixing it with
II- 7
-------
the combustion air. Emulsions were prepared with a Total device on loan
to our laboratory.
Experimental Results
The initial activity in evaluation of an alternate fuel is to
compare its combustion and emission performance to conventional fuels.
The baseline fuels were a distillate fuel oil and propane, represent-
ing both liquid and gaseous conventional fuels. The alcohols were
tnethanol, isopropanol and a mixture containing 50% of each, thereby
allowing an assessment of the effect of the presence of higher alcohols
on the methanol combustion. Figure 1 shows a comparision of NO emis-
sions as a function of the swirl block position (SBP), where the tangen-
tial momentum of the combustion air increases as the swirl block posi-
tion increases. The salient features of this figure are as follows:
1) The NO emission for distillate oil shows a distinct
peak at SBP of 4.
2) The NO emissions for propane and the alcohol fuels
decrease continuously as the SBP is increased from 2 to 8.
3) The alcohol fuels have lower NO emissions than either
of the conventional fuels.
II-8
-------
300
o
UJ
oc
CO
GO
QC
a
a
x
o
o
oc
200
100
ODISTILLATE OIL A ISOPROPANOL •METHANOL
DPROPANE O50% ISOPROPANOL - 50% METHANOL
8
SWIRL BLOCK POSITION
FIGURE 1. COMPARISON OF BASELINE NITRIC OXIDE EMISSIONS FOR
VARIOUS FUELS AS A FUNCTION OF SWIRL PARAMETER.
II-9
-------
4) The NO emissions of the alcohol fuels decrease as
the percentage of isopropanol decreases.
It is interesting that, although the alcohol fuels are liquids, their
emission characteristics are more like those of gaseous propane
than of the liquid distillate oil. This is probably related to vapor-
ization of the low boiling alcohols early in the combustion zone as
indicated by observation of the flame. The spray is only visible
for about 2 to 5 cm and the flame has a blue to nonluminous appear-
ance which is quite similar to propane flames. The low NO for the
alcohols and the trend as a function of alcohol type may be postulated
to result from the partially oxidized nature of the alcohol having
the effect of diluent addition to the flame zone. Viewed in this way
methanol can be viewed as a CEL'H-O fuel and compared to a CH. „
/ Z 1. o
representation of the distillate oil. The methanol therefore requires
nearly twice the fuel mass to supply the same heat input to the
furnace. For convenience, the term "bound water" will be used for
the preoxidized part of the alcohols. The bound water is 54, 43 and
30 mass percent for methanol, the 50/50 mixture and isopropanol,
respectively. Increasing the diluent mass would reduce flame temper-
ature and therefore thermal NO.
To examine the validity of the postulated effect of bound water,
a series of tests on water-in-oil emulsions was run and the results
11-10
-------
are plotted in Figure 2 for an SBP of 4. The line was drawn for the
water-in-oil emulsions and the alcohol results plotted for comparison.
For any given alcohol fuel the NO emissions compare quite well to
the emulsion results of the same water content.
To allow comparision of the alcohol results with another dilution
technique, the effect of the flue gas recirculation was examined and
the results are presented in Figure 3. This shows that a further
reduction of NO emissions is possible for all of the fuels with about
85% reduction possible at 25% FGR.
The effect of any diluent on thermal NO is based on a reduction
of peak flame temperatures and, thereby, a reduction in the rates of
formation. To allow a comparison of all the techniques on a common
basis, the NO emissions are plotted versus theoretical flame tempera-
ture as shown in Figure 4. Note that the temperature is calculated
based on fully equilibrated products at the overall stoichiometry of
the system for adiabatic conditions. Since the combustion system
is hot wall (about 1400 C) and designed for minimum heat loss, the
assumption of adiabatic conditions may be reasonably good. The
assumption that the reaction occurs at the overall stoichiometry is
more open to question for the diffusion flame system and may aid in
explaining the differences seen in the figure. The most important
factor may be the differences in fuel vaporization and atomization
11-11
-------
v»
O
X
O
O
ac
I I
O DISTIL LATE OIL-WATER EMULSIONS
O 50% ISOPROPANOL • 50% METHANOL
AISOPROPANOL
OMETHANOL
• ISOPROPANOL AND WATER
20 30 40
FUEL WATER CONTENT, MASS %
FIGURE 2. COMPARISON OF NITRIC OXIDE EMISSIONS FOR VARIOUS
FUELS AS A FUNCTION OF WATER CONTENT.
II-12
-------
O DISTILLATE OIL
D PROPANE
A ISOPROPANOL
O METHANOL
0.05
0.10 0.15
FRACTION RECIRCULATED.f
0.20
0.25
FIGURE 3. EFFECT OF FLUE GAS RECIRCULATION ON NITRIC OXIDE
EMISSIONS FOR VARIOUS FUELS.
11-13
-------
300
Q
ut
g 200
<
LU
5
CO
<
*
DC
Q
a.
a.
Q
X
o 100
o
DC
• DISTILLATE OIL WITH FGR
• DISTILLATE OIL EMULSIONS
A ISOPROPANOL
O 50% ISOPROPANOL - 50% METHANOL
OMETHANOL
A ISOPROPANOL - WATER
1.900
1.800
1.700
1.600
1.500
1.400
THEORETICAL FLAME TEMPERATURE, 10 °C
FIGURE 4. COMPARISON OF NITRIC OXIDE EMISSION REDUCTION
AS A FUNCTION OF THEORETICAL FLAME TEMPERATURE FOR VARIOUS
DILUENT ADDITION TECHNIQUES.
II-14
-------
characteristics coupled with the fuel-air mixing characteristics of
the system. An explanation may be summarized as follows:
1) For the water-in-oil emulsion, the water droplets are
believed to shatter causing a secondary atoraization of the oil. It
is also possible that at the higher water levels the emulsions may
tend to separate. In either case the oil may burn in an environment
where the diluent effect of the water is less than that calculated
and the actual flame temperature is higher.
2) For the flue gas recirculation the diluent is uniformly
mixed with the combustion air and should have the same effect on the
flame temperature no matter where the fuel burns; however, the fuel
reactions still may occur at conditions closer to stoichiometric than
the overall mixture which would have the effect of increasing peak
temperatures locally.
3) For the alcohols, not only is the "diluent" intimately
bound in the fuel molecule, but the early vaporization may allow more
rapid mixing with the combustion air. This would most closely approxi-
mate the basis on which the flame temperature calculations were made.
4) Finally the alcohol fuels have high latent heats of
vaporization which may result in a significant local heat extraction
II- IS
-------
and an associated NO reduction.
The final series of experiments examined the effects of alcohol
and distillate oil emulsions on NO emissions. The total heat input to
the furnace was kept at a constant value and the percent of heat input
in the form of methanol was increased from zero to 100%. The average
results of three series of tests are shown in Figure 5. The first 25%
methanol reduced the NO emissions by about 50% relative to distillate
oil alone. To allow a comparison on a common basis the "bound water"
in the methanol was calculated as a water content for the total mix-
ture and plotted along with the water in oil emulsion in Figure 6.
For a given water content in the fuel mass the methanol-oil emulsions
give consistently lower NO emission than do the water-oil emulsions.
This may be attributable to a number of factors including:
1) The methanol has a lower boiling point than does water
and may provide a more efficient secondary atomization effect.
This would promote better mixing and a greater effect of the
diluent on the combustion temperature.
2) Not only does the methanol have a higher heat of
vaporization per mass than the water, but also there is a
larger mass to be vaporized for a given "water" content, re-
sulting in a more significant effect on peak flame temperature.
II- 16
-------
I
I I
O DISTILLATE OIL
O DISTILLATE OIL • METHANOL EMULSIONS
A METHANOL
25 50 75
METHANOL CONTENT, % INPUT BTU
FIGURE 5. NITRIC OXIDE EMISSIONS FOR METHANOL
DISTILLATE OIL EMULSIONS AS A FUNCTION OF
METHANOL HEAT INPUT.
11-17
-------
O DISTILLATE OIL-WATER EMULSIONS
D DISTILLATE OIL -METHANOL EMULSIONS
A METHANOL
20 30 40
FUEL WATER CONTENT, MASS %
FIGURE 6 COMPARISON OF NITRIC OXIDE EMISSIONS FOR WATER AND METHANOL
WITH DISTILLATE OIL AS A FUNCTION OF WATER CONTENT.
II-18
-------
It should be stated that, although the explanations offered here
(based on temperature effects on thermal NO) appear to explain most of
the observed phenomena, some anomalies still exist in the methanol
data. At this time it is not possible to assess the effect of the
methanol on flame chemistry. It is possible that the methanol reaction
scheme may reduce concentrations of certain radical species and thereby
reduce non-Zeldovitch NO. This possibility will be explored in the
laboratory by combustion of a mixture of 33% CO and 76% H2 to simulate
a decomposed methanol.
H-19
-------
Conclusions
Based on the results presented here and other Information contained
in reference 3, the following conclusions can be drawn:
1) Alcohol fuels produce lower emissions of nitrogen oxides
than distillate oil or propane.
2) The NO emissions of alcohol fuels increase as the per-
centage of higher alcohols increases.
3) The low NO emissions for alcohol fuels appear to function
from the presence and level of oxygen in the fuel molecule, which
can be viewed as a diluent carried in the fuel. The operative
mechanism appears to be related to thermal effects of the fuel,
latent heat of vaporization, and/or decreased flame temperature;
however, chemical effects cannot be totally ruled out.
4) In the hot wall experimental system, NO levels similar
to those for methanol (e.g., 65 ppm) could be attained for distil-
late oil and propane with flue gas recirculation and could be ap-
proached at 50% water for distillate oil emulsions.
5) Based on the calculations, the use of methanol and a
.11-20
-------
water-oil emulsion appears to impose a 6 to 7% increase in stack
heat loss compared to distillate oil with flue gas recirculation;
however, it is anticipated that there will be some compensating
factors that can be designed into practical systems to minimize the
losses, based on lower excess air operation as seen in the package
boiler simulator.
6) The emissions of CO, UHC, and particulate from alcohol
fuels were generally the same as, or less than, those for the con-
ventional clean fuels tested (propane and distillate oil). Alde-
hydes do not appear to be a significant problem based on limited
data generated under EPA contract 68-02-1498 with Ultrasystems,
lnc.<3>
7) From a technical standpoint, methanol appears to be a
satisfactory fuel for stationary combustion systems; however, the
final commercial feasibility will probably depend more on cost
and availability than on the ability to burn the fuel.
FUTURE PROGRAM
The future work in this area will be on other alternate fuels
that appear to have some potential for application to stationary com-
bustion sources. The fuels of interest and current plans of research
11-21
-------
direction are summarized below.
Background
In addition to the alcohol fuels there are a number of other fuels
that may be derived from coal, oil shale or refuse at some time in the
future. These possibilities cover the range of gaseous, liquid and
solid fuels, with a number of candidates for each of these types.
Gaseous Fuels
The gaseous fuels of interest can be divided into the broad cate-
gories of Synthetic Natural Gas (SNG) and producer gas. The SNG is a
high Btu fuel to replace or supplement natural gas in pipelines and
the properties of the two types of SNG are compared to natural gas in
Table 1. Unless some information develops to indicate a large diver-
gence of combustion and/or emission characteristics of SNG from those
of natural gas, the current technology for methane combustion should
be adequate for emission control. The properties of low Btu gas
vary widely depending on both the feed material and on the process
3
used; the energy content can range from 4403 to 8806 KJ/M as shown
in Table 1. The combustion characteristics of these fuels are
just beginning to be studied; however, several aspects already have
come to light. First, due to the nature of the fuel, the adiabatic
11-22
-------
CO
1-1
e
&
8
%
o
co
s
M
1
CM
0
2
O
CO
1
o
•
.H
iJ
H
•O
ki
0)
C
I-l
iN!
H
^
0
C >i
O X
•H X
*J CJ
•H
01
0
6*
O
O CM
35
<3
o
U
ocNiio ^o vO^ooin
ro ^ co co
^•^ '^^
v2 s5
co to to co 2^
fl tO W cb ^s ^-^ /^\ x*s
OOO O,D Hui.O
cd cd ntc cij bo j>^ cd ^]
M l-< W-' t-i h C 00 CJ
333^3 3 -H >, 1
U 4J 4-1 (1) 4J (-J&3CC3
2KK4JZ W W 0) CO
g"gOHO^ 0(Sc3o
« «o 4-1 n ye 3333
?d ^ 4J QJ 4J tiO
x ai c a c u 3 ^ 3 3
H U C0%4 (AhJ >J i-] hJ »4
•
C
o
J 1
•H
S
• ^
en
-
0
t t
•o
1
14
S-i
<1>
CM
.
S3
X
0
00
(3
•H
M
0)
C
•H
a
iH
cd
a
°s
S
A
o
J^
0
o
s
jg
0}
ON
•
^
I
O
iH
4J
a.
(U
CO
»
>.
CO
o
Q
c
CJ
(U
CA
0
M-l
(U
3
4-1
•H
4J
CO
C
M
^
M
V
ex
CM
1
•H
•0
a.
P^
CO
rt
0
u
B
H
iw
at
rT
C
s
p
^Q
•a
vO cd
S3
CM O
a «M
T3 *
ed CM
O
•j- o
CJ <4.J
fft
*J 4-1
41 -H
CO CO
-------
flame temperature is lower than conventional fuels and thermal NO prom-
ises to be low. However, if the high temperature processes for H_S
removal are developed and the fuel gas is presented to the burner at 550
to 850°C, significant thermal NO may result. Second, the presence of
significant amounts of ammonia appears to exist in the product gas at
the gasifier exit for many processes. Unless this bound nitrogen is
removed in the gas clean up sequence prior to combustion, significant
fuel NO can result. Robson states that 600 ppra of NH^ can be present
in the fuel gas following some low temperature l^S removal processes,
and projects 3800 pptn for some of the hot H2S removal processes under
development. This then identifies high temperature fuel gas and chem-
ically bound nitrogen as two potential sources of NO that must be con-
sidered in low Btu gas combustion technology development.
Liquid Fuels
The liquid fuels of interest can be derived from either coal liq-
uefaction or shale processing and some potential products are shown in
Table 2. It can be noted that, while the sulfur content is relatively
low (0.2 to 0.76%), the fuel nitrogen is quite high (0.68 to 1.77%) as
compared to a residual oil produced from petroleum stock. While the
hydrotreated COED oil shows that the heteroatom content can be reduced,
this represents another processing step and extra cost for the finished
fuel. If these fuels can be burned in the crude form and the emissions
II- 24
-------
CO
w
5
Q
rH
:=>
M
J
S3
0
M
P£
£>
fc
O
8««
4J
00
•H
Ol tO
13
CO
•H
CO
£>s,
1— 1
cC
C
<£
CU 2
4-1
CD
g
•H
4-1
rH
E>
ar
CJ
m co
O o rH en
o O r^ d
VO
ON ON in rH
O CO O") O
• • • •
O O O O
ro
CM CM rH
O CN rH O
• • » «
0 0 rH O
t— I U"> sf ^O
» » * *
f*"} CM 00 rH
rH I— I rH
CF» 00 O O
» » * *
\jD \^ ro 00
CO 00 00 00
^D -^- O
1 t • 1
rH 1 1 ON 1 O
in
rH ON CO ^Q
CO rH ^
ao
0
rH
O
C
,G
O
0!
H
CO
0
Q-t
o
CU
4-1
3
4-J
•H
4J
CQ
C
M
;
2
O
en
I
O
ffl
(LI
3
C
•H
q
p«^
II O 1
ii - i
O^i
s~^
.a
J2 T3 H
O X> 4-1 O rH
^ o jr IT)
Q Q VJ 4-1 O
w w T: c u
o o x >^ i
C_J C_) S3 C/3 ffi
|
|
1
1
>J
3
H-l
i— 1
3
V3
U
O
j_J
rH
TO
O
CJ
1
r-»
•
O
00
H
CO
M
r-
en
^— ^
o
x_^
rH
b
rH
CO
O
O
CO
5
vD
*
»^|-
00
ON
^3-
•«
CO
-------
of NO and smoke effectively controlled by combustion technology, sign-
A
ificant cost and/or overall energy efficiency savings may be realized.
The thrust of the in-house research will be exploring the possible
approaches to minimizing the emissions.
Planned Program
The near term emphasis in the in-house research project will be low
Btu gases; liquid fuels will be included as suitable candidates
become available.
GasGenerator
A catalytic partial oxidation reactor developed by the Jet Pro-
pulsion Laboratory for automotive use is being adapted as a gas genera-
tor to supply a variety of simulated low Btu gases. The reactor will
accept a variety of hydrocarbons and reform these to yield a product
fuel gas closely approximating the predicted equilibrium composition
based on the input fuel to air stoichiometry. The reactor can be operated
with about 40 to 70% of theoretical air. By changing the fuel type
from methane to benzene and the input stoichiometry, the CO to H? ratio
and the energy content of the product gas can be controlled over wide
ranges. At design condition the reactor will produce approximately
88,000 watts of fuel gas at a temperature of 650 to 720°C. The product
II- 26
-------
gas temperature will be controlled down to 120 C by additional heat
exchange.
Research Areas
The program will address the two areas previously identified as
requiring attention, fuel gas temperature and bound nitrogen, for
several gas compositions. The gas compositions will strongly depend
on the capability of the gas generator as delivered; however, the varia-
bles of interest are: 1) CO to H,, volume ratio in the range of 0.5 to 2;
3
and 2) energy content between 2980 and 7460 KJ/M . The composition of the
gas will also be altered by the addition of methane to simulate some
of the processes yielding significant quantities of hydrocarbons. In
addition the effect of chemically bound nitrogen will be studied by
ammonia addition in the range of 100 to 4000 ppm. The gas temperature
will be controlled by a variable heat exchanger and will range from
120 to 720 C in three or four increments.
In addition to establishing the emission characteristics of the
various gases as a function of composition and temperature, potential
control techniques will be explored for gases that yield high levels
of NO. The approaches to be considered are the classical techniques
of burner design and/or staged combustion. The burner design variables
will allow changes in fuel and air mixing rates and should be applicable
i r- 27
-------
to either thermal or fuel NO. The staged combustion techniques will be
primarily directed toward control of fuel NO from ammonia bearing fuels,
Based on fundamental and small scale data, there is reason to believe
that operating with a burner input of 60 to 70% of theoretical air and
second stage air injection to complete fuel burn out will yield fuel
NO levels below 100 ppm even with high ammonia levels in the input
fuel.
The results of this study will complement and reinforce related
low Btu gas combustion experiments being carried out under contract
to EPA by IGT and Ultrasystems.
SUMMARY
This in-house research project has examined the combustion of
alcohol fuels and identified potential advantages for NO control.
X
The next stage is to explore emission and combustion characteristics
of low Btu gases and to study applicable NO control techniques.
A
The alternate liquid fuels have received only limited considerations
because of limited availability; however, as these fuels are secured
they will be factored into the effort.
11-28
-------
REFERENCES
1. Hoogendoorn, Jan C., "Experience with Fischer-Tropsch Synthesis
at Sasol." Symposium Papers - Clean Fuels from Coal, pp 353-336,
Institute of Gas Technology, Chicago, Illinois (September 1973).
2. Duhl, R. W. and T. 0. Wentworth, "Methyl Fuel from Remote Gas
Sources." Presented at llth Annual Meeting of the Southern
California Section AIChE, Los Angeles, California, April 1974.
3. Martin, G. B. and M. P. Heap, "Evaluation of NOV Emission
X
Characteristics of Alcohol Fuels for Use in Stationary Combus-
tion Systems." Presented at 80th National AIChE Meeting, Boston,
Mass., September 9, 1975.
4. Heap, M. P. , T. M. Lowes, R. Walmsley, and H. Bartelds, "Burner
Design Principles for Minimum NO Emission." Proceedings,
X
Coal Combustion Seminar, EPA-650/2-73-021, pp 141-172; NTIS No.
PB 224-210/AS (September 1973).
5. Robson, F. L. and A. J. Giramonti, "The Environmental Impact of
Coal-based Advanced Power Systems." Symposium Proceedings
Environmental Aspectsof Fuel Conversion Technology, EPA-650/2-
74-118, pp 237-257; NTIS No. PB 238-304/AS (October 1974).
11-29
-------
8:30 a.m.
Assessment of Combustion and
Emission Characteristics of
Methanol and Other Alternate
Fuels
G. Blair Martin, U.S. E.P.A.
Qs Can you comment on the quality of the raethanol
distillate oil emulsions as characterized by dis-
persed phase particle size, and also, do you have
any idea of the residence time of the emulsions
between the point of preparation and the point of
injection into your combustion?
A: We did not look at the particle size. The one
thing that we did for both the oil and methanol
emulsions was to spray them at a container and to
watch the separation characteristics. At high water
content, the emulsions tended to separate relatively
quickly. The residence time between the emulsifier
and the burner is fairly short in that the emulsion
is formed in a device immediately before the fuel
tube to the burner. The fuel tube is approximately
18 inches long from that point to the nozzle and is
constructed with a 1/8 inch I.D., so the amount of
fuel that's actually in the fuel tube at any one
time is relatively small.
11-30
-------
BURNER DESIGN CRITERIA FOR CONTROL OF
POLLUTANT EMISSIONS FROM NATURAL GAS FLAMES
by
D. R. Shoffstall
Paper Presented at
SYMPOSIUM ON STATIONARY SOURCE COMBUSTION
Sponsored by
COMBUSTION RESEARCH SECTION OF
U.S. ENVIRONMENTAL PROTECTION AGENCY
September 24-26, 1975
Atlanta, Georgia
II-31
-------
BURNER DESIGN CRITERIA FOR CONTROL
OF POLLUTANT EMISSIONS
FROM NATURAL GAS FLAMES
by
D, R. Shoffstall
Institute of Gas Technology
Chicago, Illinois 60616
ABSTRACT
This program has developed a number of simple burner modifications which
resulted in significant reductions in NO emission levels. These modifications
can be utilized on existing equipment or can be incorporated in the design of
new combustion systems. Because of the influence of furnace geometry and
multiple flame interactions on NO emission levels, these modifications may
not result in the size reduction in emission levels observed during the test
program. However, applying the principles of aerodynamic NO control
developed by this program, coupled with the use of suitable monitoring
equipment, reductions in NO emission levels up to 60% are possible.
11-32
-------
BURNER DESIGN CRITERIA FOR CONTROL,
OF POLLUTANT EMISSIONS
FROM NATURAL GAS FLAMES
by
D. R. Shoffstall
Institute of Gas Technology
Chicago, Illinois 60616
For 3 years, IGT has conducted a program for the Environmental Pro-
tection Agency on NO (nitrogen oxides: NO, nitric oxide; and NO2, nitrogen
X
dioxide) emissions control from natural gas combustion on a pilot-scale test
furnace. The objective of the program is twofold: 1} to determine whether
combustion aerodynamics can be used to control NO emissions from a
Ji
natural gas flame and 2) if aerodynamic control is successful, to determine
the lower emission limits that can be achieved by using this technique. The
resulting data could be used by industry and/or burner manufacturers to aid in
reducing NO emissions from existing combustion processes and to help in
X-
the development of new low-NO -emission burners.
X
The effect of combustion aerodynamics on NO emission levels was tested
on a number of industrial burner types including a kiln, a flat flame, a baffle,
a movable-block swirl, and a utility boiler. This paper deals primarily
with the baffle burner because not only has the investigation resulted in
recommended low-NO -emission operating conditions but also the resulting
data have been used by a group of gas utility companies in the development
of a burner with a 60% reduction in NO emissions.
Jv
RESEARCH INSTALLATION
Most of the experimental work described in this paper was conducted on
the IGT rectangular test furnace, which has a 25-sq-ft cross-sectional area
and a length of 15 feet. This furnace can be end- or sidewall-fired at rates
up to 4 million Btu/hr. The furnace is equipped for in-the-flame sampling
(Figure 1). The overall furnace system is flexible enough that the following
parameters can be varied independently:
• Heat input, up to 4 million Btu/hr
• Air input, up to 40^ excess
II-33
-------
10/75
o
03
a
f-i
n>
&0
9
-4J
O
0)
IH
0
60
•H
1=4
1 1. - 34
-------
• Heat losses to the furnace wall, by changing flow in water-cooling tubes
cast into the refractory furnace sidewalls
• Combustion air temperature, up to 1000 F
• External flue-gas recirculation (FGR), up to 35% of combustion air
• Furnace pressure, up to 4-0.05 in H2O.
The following analytical instrumentation was used for concentration
measurements of chemical species during the program (Figure 2):
• Beckman 742 Polarographic Oxygen (O2)
• Beckman Paramagnetic Oxygen (O2)
• Beckman NDIR Methane (CH4)
• Beckman NDJR Carbon Monoxide (CO)
* Beckman NDIR Carbon Dioxide (COZ)
• Varion 1200 Flame lonization Chromatograph (Hydrocarbon Cl-C9)
• Beckman NDIR Nitric Oxide (NO)
• Beckman NDUV Nitrogen Dioxide (NOE)
* Hewlett-Packard Thermoconductivity Chromatograph Hydrogen (H2),
Nitrogen (N2), Argon (Ar), O2, CO, COZ, and hydrocarbons C] to Cs
• Beckman Cherniluminescent NO-NO2.
GENERAL. FLAME CHARACTERISTICS
A general aerodynamic characterization of a flame can be made by deter-
mining the different types of flow patterns that exist within a combustion
chamber. A detailed flow analysis of a confined flame reveals that the front
section of a combustion chamber can be divided into four zones; primary
jet, primary recirculation, secondary jet, and secondary recirculation. The
primary and secondary jets contain only particles with a forward flow direc-
tion (away from the burner), whereas the recirculation zones contain gas
particles moving in the reverse flow direction (back toward the burner). The
size, shape, and particle density of the recirculation zones are determined
by the velocity, the ratio of gas to air, the spin intensity of the secondary jet,
the burner-block angle, and, for the secondary recirculation zone, the size
and shape of the combustion chamber. Figure 3 shows the three types of flow
patterns that we observed during our pollution control trials-
II- 35
-------
10/75
Figure 2. Control room facility
II- 36
-------
TYPE I
NO SWIRL
MO PRIMARY
RE CIRCULATION
TYPE. IT
LOW SWIRL INTENSITY
PRIMARY JET VELOCITY >
SECONDARY JET VELOCITY
HIGH &WIRL INTENSITY
SKONWRY jet VELOCITY
PRIMARY JET VELOCITY
A -1.M-21 66
Figure 3. Flames types tested
11-37
-------
A type I flow pattern is observed when the secondary jet has no tangential
velocity component (no spin) and the primary jet velocity is much larger than
the secondary jet velocity. Depending on the initial jet exit velocity from the
fuel injector, the flame can be either attached to or detached from the injec-
tor. A type I flow pattern also can be generated with a secondary jet with a
tangential velocity component (spin) if the burner-block divergent angle is
equal to or less than the angle of the secondary jet relative to the burner axis.
As a result, the burner block restricts the expansion of the secondary jet and
inhibits the formation of a primary re circulation zone.
A type II flow pattern is generated when the secondary jet has a tangential
velocity component large enough to cause the particles to adhere to and pack
tightly against the burner block. This packing creates a low or negative
pressure region in the center of the burner block. The pressure differential
between the furnace and the central region of the burner block causes gas
molecules to be pulled into this region and back toward the burner, thus creat-
ing the primary recirculation zone. When the velocity of the primary jet is
greater than the velocity of the recirculating gases in the primary recircula-
tion zone, the primary jet penetrates this reverse flow region and a recircu-
lation lobe occurs on each side of the burner axis. This flow pattern is
labeled type II.
If the central region of the burner block has a large enough negative
pressure differential relative to the furnace, such that the velocity of the
recirculating gases is greater than the velocity of the primary jet, pene-
tration of the primary recirculation zone is not possible and a type III flow
pattern results. A type III flow pattern also can be generated mechanically
by introducing the fuel with only a radial velocity relative to the axis of the
burner.
DATA ANALYSIS AND DISCUSSION
Baffle Burner
Figure 4 illustrates the design of the baffle burner. The gas nozzle lies
parallel to and along the axis of the burner. It is inserted into the ceramic
baffle, thus ensuring that the gas enters parallel to the axis of the baffle
burner. Combustion air enters perpendicular to the axis and passes through
the six ports in the baffle, which impart a swirl to the air in some designs.
H-38
-------
10/75
r
o
4 O. D.
\ Baffle
I 8"
A t\r
V 'V
U—J?—
-
o a.
4r
, O
: CD
Long Rome
Short Home
infer. Flame
'A'
•fc"
2"
2"
1B'
nV'
14^*
i«i"
Baffle
Straight Thru Holes
25° Spin
15 Spin
Figure 4. Assembly drawing of baffle burner
H-39
-------
Two of these baffle ports are shown in Figure 4 with their axes parallel to the
axis of the burner. Air exiting from the ports (as illustrated) would have an
axial velocity component only, resulting in a "long" flame. To shorten the
flame length, a tangential-flow component must be added to the combustion air
velocity. This is accomplished by using a baffle where the combustion-air
ports are rotated relative to the axis of the burner. For the "intermediate"
flame length baffle (IFLB), the rotation orientation of the ports is 15 deg,
and for the "short" flame length (SFLB), 25 deg.
Standard Conditions
Figure 5 presents normalized nitrogen oxide (NO) versus excess oxygen
(O2) test data for burner conditions normal to an industrial situation (stand-
ard gas nozzle, baffle nozzle position, and 4 deg burner-block angle). Photo-
graphs illustrating flame geometry and flame luminosity with these industrial
conditions are shown in Figures 6 and 7.
The standard gas nozzle is a 2-inch pipe that rests within a centering tube
mounted between the baffle and the burner housing. Gas nozzle positions are
illustrated in Figure 8. The gas nozzle position is denoted as the baffle posi-
tion when the nozzle is even with the burner-block side of the baffle. Other
nozzle positions investigated were the throat position (gas nozzle midway
between the baffle and the front wall of the furnace) and the exit position (gas
nozzle even with the front wall of the furnace). For the IFLB, the exit gas
nozzle position produced unstable combustion and therefore is not included
in the test results.
At a gas input of 3000 SCF/hr, the relationship between NO and excess O2
became increasingly nonlinear as a function of temperature; see Figure 5.
This indicates that in the range of excess O2 we investigated, the higher the
preheat, the more closely the NO-versus-O2 relationship characterizes that
of a premix flame.
For the second set of curves in Figure 5, burner conditions were the
same, but the gas input was reduced to 2000 SCF/hr. In an attempt to main-
tain the same level of bulk NO formation, the wall temperature was held at
1420 C, compared with 1435°C for a gas input of 3000 SCF/hr. These curves
show the same linear relationship between NO and excess O2 that was observed
for the kiln burner. This linear relationship is characteristic of a diffusion
flame.
11-40
-------
600-1
Cos input 3070 8 2O05 SCFH
Oos Nozzle Bottle Position
Won Temperature 1435* c
4* Burner Slock Angle ,_
500-
400-
E
a.
a.
O
Z
300-
IM
"5
E
b
z
200-
100-
o
1 6,000
1
t
Legend •
A
O
1 I
2 3
Standard
Standard
1
4
Nozzle,
Nozzle.
I
5
Gas
Gas
Input
Input
1
6
3000
1000
CFH
CFH
IN FLUE, 96
Figure 5. Normalized NO concentration as a
function of flue O2 for the IF LB burner
with a standard gas nozzle at gas
inputs of 3070 and 2005 SCF/hr
11-41
-------
10/75
1. Divergent fuel nozzle
2. Standard fuel nozzle
Figure 6. Flame photographs of intermediate
flame length baffle burner
11-42
-------
10/75
3. Combination fuel nozzle
4. Radial fuel nozzle
Figure 7. Flame photographs of intermediate
flame length baffle burner
T T- 43
-------
10/75
Figure 8, Nozzle positions tested for the baffle burner
11-44
-------
External FGR
As a base-line control case, the effect of external FGR was tested. The
percentage of FGR is determined by using the relationship —
FGR (CF/hr)
% FGR =
Secondary Air (CF/hr) -f-
X 100
(CF/hr)
By using the data presented in Figure 9, the NO concentration measured
for standard burner operating conditions with 46 2 C secondary air preheat
can be compared with the concentration at similar burner conditions but with
the addition of 15% and 30% FGR to the secondary air. During preheat, both
the flue gas and secondary air were blended and reached the same final pre-
heat temperature, which in this particular test was 460 C. At 3% excess O2,
a 15% FGR level reducjed the normalized NO from 575 to 160 ppm, whereas
30% FGR reduced the measured NO concentration to 50 ppm.
Wall Temperature
The data for Figure 10 were collected with the burner operating in the
normal industrial mode; however, the wall temperature was decreased from
1435 to 965 C by using water-cooling tubes within the furnace sidewalls.
Although the NO was depressed by approximately a factor of 2, it is difficult
to say what part, if any, was due to a reduction in the bulk NO formation.
Because the secondary-recirculation-zone flue products were at a much lower
temperature when they became entrained by the secondary combustion air
jet, a lower flame temperature resulted and, consequently, a lower concen-
tration of NO.
Nozzle Type and Position
Figure II presents NO concentrations measured under standard burner
conditions but with a radial gas injector. This injector forces the gas to
enter the burner block radially to the axis of the burner, which causes a
flame flow pattern categorized as type III. The combination gas nozzle
shown in Figure 12, operating with radial injection only, was used.
The radial nozzle in the baffle position produced more NO tliar. the stand-
ard nozzle because of the increased rate of heat release, coupled with the
inability of the burner block to be used as a heat sink. However, when the
radial nozzle was moved into the throat position, there w-rs a dramatic drop
11-45
-------
600
50OH
£
d
a.
3OO
4)
O
20CT1
O
I.7OO
Gas input 3070 SCFH
Gas Nozzle Baffle Position
Wall Temperoturt I39O C
4 Burner BlocK
T2,OOO
30
FOR
I4,OOO
1
1
1
2
1 I
3 4
1
5
1
6
N FLUE. 96
Figure 9. Normalized NO concentration as
function of flue O2 for the IF LB burner
with a standard gaa nozzle and
15% and 30% FGR
11-46
-------
300-
25O-
E
ex
O
Z 200H
4>
N
fj 150-
o
z
100-
50-
Gos input 3070 SCFH
Cos Nozzle B of fie Position
Wall Temperature 965* C
4* Burner Block Angle
Air Preheat AS Labeled
232 C
O Standard Nozzle, Baffle Position
Wall Temperature 965° C
1
\
( —
2
1
3
4
i
e1
IN FLUE,%
Figure 10. Normalized NO concentration as a
function of flue O2 for the IF LB burner
with a standard gas nozzle at a
wall temperature of 965°C
H-47
-------
7OO-1
Gas inpgt 2970 SCFM, Radial
Gas Nozzle Baffle & Throat Position
Won Temperature 1370° C
4 Burner Block Angle
legeiv;
100-
Stand-ire! Nozzle Rottle Position
O R,j Jill N- 7.-!C.
O «odiol N--zzJe
Boffl" r<.-siti">n
Throat Position
1
t
1
2
0, IN
1
3
FLUE, 'X
1
4
1
5
1
6
Figure II. Normalized NO concentration as a
function of flue O2 for the IF LB burner
with a combination gas nozzle and
radial injection
II-43
-------
10/75
O
c
-------
in the measured NO concentration (approximately by a factor of 2-1/2).
This decrease could be due to the combustion outside the burner block,
which allows a larger mass exchange to occur among the fuel, the internal
recirculation zone, the secondary air, and the external recirculation zone,
thus lowering the flame temperature.
In an attempt to further slow the rate of heat release, half the gas was
introduced radially while the other half was introduced axially (at a high
velocity of approximately 300 ft/s). Again, data were collected for the half
axial-half radial injection in both the baffle and throat positions. These
results are presented in Figure 13. The axial gas injection had the desired
effect: It decreased the NO concentration relative to all-radial injection by
projecting the flame further into the furnace and slowing the rate of heat
release. Again, moving the nozzle position into the throat of the burner
block further reduced the NO concentrations.
Figure 14 is a composite plot showing all significant reductions in NO as
a function of nozzle type and/or nozzle position. All these data wf.'re col-
lected with a secondary air preheat at 4^0 C. The base-line conditions were
those of the standard nozzle in the baffle position; this burner configuration
produced the highest NO concentrations. Moving the standard nozzle into the
throat position did not change the shape of the NO-versus-O2 relationship,
indicating that the basic mixing patterns were not dramatically altered. How-
ever, it did decrease the NO concentration to approximately 125 ppm. With
a divergent gas nozzle (Figure 15} similar to the one used in the kiln burner,
an additional decrease to 100 ppm NO resulted. The differences in NO con-
centration produced with the divergent nozzle and the standard nozzle in the
throat position decreased as a function of excess O2, until at 6% excess O2,
the NO concentrations were equal. The divergent nozzle produced a type III
flame. The decrease in NO occurred because of the additional entrainment
of combustion products from the primary ai^i secondary recirculation zones.
Although the data are not plotted, test results with the divergent nozzle in
the throat position paralleled those with the kiln burner; i.e., the NO was
increased by moving the gas injector toward the combustion chamber. The
difference in NO concentration as a function of aozzle position, however, is
very small.
11-50
-------
50O-
40O-
•0-3OO-
N
200-
100-
,700
Gas
Gas
Wall
input 3101 SCFH
1511 Axial. I59O Radial
Nozzle Baffle ft Throat Position
Temperature I39O° C
4° Burner Block Angle
Legend
Standard Nozzle, Baffle Position
O Hatr Radial Axial Nozzle, Baffle Poeitfon
D Halt Radial Axial Nozzle, Throat Position
I
4
I
S
IN FLUE, 96
Figure 13. Normalized NO concentration as
function of flue O2 for the IF LB burner
with a combination gas nozzle and
axial and radial injection
II- 51
-------
600-,
Cos Input 3070 SCFH
Secondory Air Preheat 460° C
4° Burner Block Angle
500-
400-
E
a.
a.
*
Z 300-|
N
Z 800-
100-
OStandard Nozzle. Battle ft«ltIon
V Standard Nozzte. Throat Position
A Divergent Nozzle.Baffle Position
D Axial Nozzle, Baffle Position
OStandord Nozzle, Baffle Position, 15% FOR
I I I
234
00 IN FLUE. %
I
5
6
Figure 14. Normalized NO concentration as a
function of flue O2 for the IF LB burner
with the various gas nozzles in
different positions
II-52
-------
10/75
O
C
CO
a
bt>
u
C
V
CD
f-,
4)
3
00
H-53
-------
The most dramatic decrease in NO with burner alterations occurred when
the gas velocity was increased by a factor of 16. This was accomplished
by using all-axial gas injection but decreasing the injector opening from the
standard E inch to 0. 5 inch. As shown in Figure 14, at less than 2% excess
O2, the divergent nozzle produced the lowest NO emissions; however, above
that level of excess O2, the high-velocity axial nozzle produced the lowest
NO emissions. At 3% excess O2 (11. 4% excess air, a. typical industrial level),
comparing the normal burner output of 515 ppm with the high-velocity gas-
injection burner output of 250 ppm, we decreased the NO pollutant level by
325 ppm through a low-cost alteration of the gas injector. Noted that exter-
nal FGR remains the best depressant of NO formation that we tested. At 15%
FGR, the NO level is 150 ppm with the standard baffle-and-nozzle configura-
tion, which represents a total decrease of 425 ppm.
Burner-Block Angle
To test the influence of the primary recirculation zones on the level of
NO emissions, the burner-block angle was increased from 4 to 8 deg. The
primary recirculation zones contain not only combustion products but also
fuel and air. The entrainment of these gases by the primary and secondary
jets should help to depress the flame temperature; however, they are not as
desirable as the gases in the secondary recirculation zone because they con-
tain both fuel and air. The angular difference between the 1FLB (the air ports
are angled at 15 deg relative to the centerline of the baffle) and the burner
block (air entry being an 8-deg divergent cone) is 7 deg. Because, the angle
of the ports is greater than that of the burner block, the flow will not separ-
ate from the contour of the block. The IFLB was tested with the standard,
divergent, and high-velocity axial gas nozzles with an air preheat tempera-
ture of approximately 460 C. The results are graphed in Figure 16.
Table 1 compares the NO levels for integer values of the percentage of
excess O2 in the flue versus nozzle type between the 4 and 8-deg burner
blocks. For the standard and divergent gas nozzles at low levels of excess
air (below 12%), the emission levels from the two burner blocks are compar-
able. However, as the excess air is increased, the emission levels from
the 8-deg block become larger than those from the 4-deg block. Use of the
II-54
-------
700
Cos input 2298 SCFH
Gas Nozzle Baffle Position
Wall Temperature 1360 C
Burner BlocK Angle
8'
600
500-
E
N
O
E
300-
200-
IOO'
Legend
O standard Nozzle
A Divergent Nozzle
O Axial Nozzle
1
3
IN FLUE, %
T
6
Figure 16. Normalized NO concentration as a
function of flue O2 for the IF LB burner
with standard, divergent, and axial gas nozzles
II-55
-------
Table 1. NORMALIZED NO AS A FUNCTION OF NOZZLE TYPE,
BLOCK ANGLE, AND EXCESS O2 FOR THE IFLB BURNER
WITH AN AIR PREHEAT TEMPERATURE OF 46o°C
Nozzle Type Block Angle, deg
Standard
Divergent
Axial
4
8
4
8
4
8
% 02
1
390
408
197
204
227
174
2
487
492
273
272
255
196
3
.„. nnm
577
564
313
360
280
238
4
567
612
340
428
300
278
__5
542
636-
360
480
313
313
high-velocity (611 ft/s) axial nozzle with the 8-deg block at normal operating
levels of excess air (below 20% ) resulted in a 20% reduction in NO emissions.
Thus the desired mass exchange among the primary jet, secondary jet, and
the internal recirculation zones occurs only for the high-velocity axial noz-
zle. In the case of the standard nozzle (38 ft/s) gas velocity) and the diver-
gent nozzle (654 ft/s gas velocity), the increased burner volume resulted in
high emissions. This increase occurred because a) the rate of heat release
resulting from a high mass exchange between the primary and secondary jets
improved more than the mass exchange among the primary jet, the secondary
jet, and the internal recirculation zones or b) the volume of the secondary re-
circulation zone was decreased.
SFLB Burner
Air Velocity
Initial tests with the SFLB burner were conducted by using a high-velocity
(126 ft/s) baffle and a low-velocity (80 ft/s) baffle. (Air injection ports are
angled at 25 deg relative to the centerline of the baffle. ) Because of the
extreme pressure drop associated with the high-velocity baffle (approximately
22 in. H2O with an air flow rate of 38,000 SCF/hr at 450 °C), the tests were
conducted with a gas load of 2000 CF/hr. The 8-deg burner-block angle used
in these tests is the manufacturer' s recommended angle for industrial appli-
cations.
Figure 17 presents the test results as normalized NO emissions plotted
against the measured percentage of O2 in the flue. The high-velocity baffle
produced consistently lower emission levels than the low-velocity baffle as
11-56
-------
600-1
500-
a. 400-
o.
Gas input 2O20 6CFH
Gas Nozzle Baffle Position
Wall Temperature 1295° C
Air Preheat A« Labeled
6° Burner Block Angle
i: 300
o
w.
O
z
200"
Legend
I3OJXX) o LOW Velocity (80 fps) Short Flame Bafflt
A High Velocity (125 fps) Short Flame Baffle
8
34
IN FLUE,%
i
5
6
Figure 17. Normalized NO concentration as a
function of flue O2 for the SF LB burner
with a standard gas nozzle
11-57
-------
a result of the larger mass exchange between the secondary recirculation
zone and the secondary jet because of the increased velocity gradient. Thus,
the higher the velocity of the secondary jet, the lower the NO emission levels.
The magnitude of the reduction in NO depends not only on the velocity gradient
across the shear layer separating the secondary jet from the secondary
recirculation zone but also on the volume and temperature of the secondary
recirculation zone. The cost involved in this method of NO reduction is due
to the additional fan horsepower needed to overcome the increased pressure
drop in the burner.
Wall Temperature
Figure 18 illustrates the effect of wall temperature on the NO emission
levels of the SFLB as a function of secondary air preheat and excess air.
Conclusions similar to those for the IFLB can be drawn for the SFLB: A
wall temperature as low as possible must be maintained through increased
wall cooling, lowered gas input, or furnace design.
External FGR
To determine whether combustion aerodynamics can be used to recircu-
late combustion products to the base of the flame as an emission control tech-
nique, we simulated the idealized case (in which the combustion air and pro-
ducts are thoroughly mixed before ignition) by mixing flue gases with the
combustion air outside the burner. Figure 19 compares the NO levels as a
function of the percentage of Oz in the flue for 0%, 15%, and 25% FGR. Again,
as demonstrated for the kiln burner and the IFLB, external FGR is an ex-
tremely effective method of controlling NO emissions from the SFLB.
Nozzle Type
To determine the type of injector that would reduce NO emissions, a
series of trials were conducted by using different methods of gas injection;
the results are shown in Figure 20. The NO-versus-excess O2 relationship
for the standard nozzle (injection method recommended by manufacturer;
2-inch stainless-steel pipe with an inlet velocity of 38 ft/s and a gas input
rate of 3000 CF/hr) is shown as a reference, so that the effa:': xf !;Ae gas
injector type on the NO emission levels can be evaluated. Visual observa-
tions wer,e made to determine the flame length for each gas nozzle tested.
This allowed a first-order evaluation of alterations that can be made to obtain
the flame shape desired by the manufacturer.
11-58
\
-------
800-t
700-
6001
CL
o
z
N
£ 400H
o
z
30O-
2OO-
100-
Gas input , 3079 SCFH. Axial
Gas Nozzle Baffle Position
Wall Temperature 1050* a I45O* C
Air Preheat As Labeled
8° Burner Block An die
Leaend
O 14S06 C Wall Temperatur«
A I05O C Wall Temperature
'30,000
11,000
9
2
°2
\
4
i
6
IN FLUE,%
Figure 18. Normalized NO concentration as a
function of flue O2 for the SF LB burner
with a standard gas nozzle at
wall temperatures of 1450° and 1050°C
11-59
-------
800H
700 H
600H
O 500H
ISI
-400H
o
300 H
8001
Gas input 3070 SCFH, AKiol
Gas Nozzle Baffle Position
Walt Temperature I36O C
81* Burner Block Anflle
0% FOR
6,000
25% FOR
o-o-
i
1
o —
I
2
o—
i
3
— o
i
4
i
5
O
6
10
IN FLUE,
Figure 19. Normalized NO concentration as a
function of flue O2 for the SF LB burner
with a standard gas nozzle and 15% and 25% FGR
11-60
-------
O
z
X3
0)
N
b
IOOO
900-
8OO-
700-
600-
500-
400-
3OO-
2OO-
100-
Gas input 3O62 8CFH
Gas Nozzle Baffle Position
Wall Temperature 1420° C
8" Burner Block Angle
— Legend —
V Radial Nozzle (IO43 fps Gas Velocity)
^ Radial Nozzle (532 fps G«s Velocity)
O standard Nozzle
O Divergent Nozzle
O Axial Nozzle
O Standard Nozzl*» 15%
O
^o-— — v
6,000
I 1 i
1 2 3
4 i ^
12
Q, IN FLUE
Figure 20. Normalized NO concentration as a
function of flue O2 for the SF LB burner
with the various gas nozzles
11-61
-------
The visual flame length observed with the standard gas nozzle with the
SFL.B was 103 cm. The high-velocity (1043 ft/s) radial injector increased
the NO emissions substantially above those measured for the standard nozzle.
The shape of the emission curve closely resembled that of a premixed flame;
peak emissions occurred at 1.5% excess OE. The flame was invisible. Re-
ducing the radial velocity to 532 ft/s resulted in a shift of the maximum NO
emission to an excess O2 level of 2. 3%. Because of the reduction in the
radial injection velocity, the magnitude of the peak NO concentration was
reduced 20%. The visual^flame length was 72 cm, still 31 cm shorter than
that desired. The divergent nozzle, which was the first altered gas injector
tested with the SFL.B, resulted in lower emission levels than the standard
nozzle. The gas velocity from the divergent nozzle was 654 ft/s (at a 3000
CF,/hr gas load); however, a type III flame resulted because of the wake gen-
erated by the divergent cone in the nozzle. Despite the high inlet velocity,
the flame length was 113 cm, which compares favorably with the desired
length of 103 cm.
The injection method that resulted in the lowest emissions was the high-
velocity axial nozzle (611 ft/s and a 3000 CF/hr gas load). This nozzle pro-
duces a type II flame, even with the high-swirl SFLB, because the velocity of
the gas is large enough to split the primary recirculation zone. Although
the rate of entrainment per unit area between the primary jet and the sur-
rounding flow zones increased because of the larger velocity gradient when
compared with that from the standard nozzle, the flame length also increased,
to 286 cm. The increased flame length was a result of a lower mass exchange
rate per unit time between the primary and secondary jets. This reduced
mass exchange resulted from a decrease in the area of the shear layer caused
by the smaller primary jet. Thus, by delaying the mixing between primary
and secondary jets, we not only slowed the heat release rate of the flame, but
also approached the ideal of the EFGR by allowing more time for the sec-
ondary jet and the combustion products in the secondary recirculation zone
to mix.
Burner-Block Angle
»
Figure 21 represents NO data from the SFL.B trial with standard, diver-
gent, and high-velocity axial nozzles and a l6-deg burner block. Obviously,
the combustion aerodynamics were changed drastically as a function of excess
11-62
-------
6OO-,
Goa input 2953 SCFH
Cos Nozzle Baffle Position
Wall Temperature 1460° C
16 Burner Block Angle
Legend
o standard Nozzl*
Divergent Nozzle
Q Axial Nozzle
IN FLUE.
Figure 21. Normalized NO concentration as a
function of flue O2 for the SF LB burner
with standard, divergent, and axial gas nozzles
H-63
-------
air. There is no one general interaction theory that explains these curves
in their entirety. Therefore, we selected a 1% level of excess O2for analysis.
This level was chosen because it consistently produced the lowest levels of
NO, with concentrations below 50 ppm.
Table 2 lists the nozzle type as a function of burner-block angle for both
the SFLB and the IFLB. The inlet gas velocity is listed in parenthesis beside
Table 2. NORMALIZED NO CONCENTRATION AT 1% EXCESS O2 WITH AN
AIR PREHEAT TEMPERATURE OF 4&0 C AS A FUNCTION OF BAFFLE
TYPE, GAS NOZZLE TYPE, AND BURNER-BLOCK ANGLE
Nozzle Type
NO Concentrations
4
590
375 -»
175 ->
215 «-
120
IFLB
Block
8
--
__
400 -»
200 ->
175 -»
--
Angle, deg
16 8
nnm ,
950
700
470 490
235 440
200 230
110
SFLB
16
- -
--
*- 240
«- 210
-* 310
—
Radial (1043 ft/s)
Radial (532 ft/s)
Standard (38 ft/s)
Divergent (654 ft/s)
Axial (611 ft/s)
Standard, 15% FGR
the nozzle type. For 14% excess air, the velocity is 120 ft/s. Note that
the air velocity increases with increasing excess air. The arrows indicate
the direction of increasing NO concentration. We know from flame-length
observations that the standard and divergent nozzles have similar rates of
mass exchange between flow zones, whereas the axial nozzle has a slower
exchange rate. It is not surprising that the standard and divergent nozzle
show similar trends in NO formation as a function of burner-block angle.
For the IFLB, the NO concentration increases with burner-block angle; for
the SFLB, it decreases with an increasing angle.
For the IFLB, we increased the mixing rate between the primary and
secondary jets more than the size, shape, and mixing rate of the primary
recirculation zone with the primary and secondary jets. For the SFLB, we
were successful in improving the mass exchange between the primary recircu-
lation zone and the primary and secondary jets, which resulted in approxi-
mately a 55% reduction in NO emissions. The high-velocity axial nozzle
showed almost the opposite effect. For the SFLB, there was a 30% increase
II-64
-------
in the emission level as the block angle was increased. Thus, for high
swirl intensity and high axial velocity, increasing the burner-block angle
resulted in a larger primary-secondary jet mass exchange. We investigated
the IFLB with the 16-deg block for flow separation between the secondary jet
and the block and obtained negative results. Therefore, we assumed that it
also was not present for the SFL.B and the l6-deg block. The results for the
IFLB and the high-velocity axial nozzle were somewhat mixed. There was
a decrease in the NO level with a block angle increase from 4 to 8 deg fol-
lowed by an increase in the NO level with an angle increase from 8 to 16 deg.
The optimum burner-block angle for producing minimum NO emissions with
the axial nozzle was approximately 10 deg.
Wall Temperature
Because of the observed sensitivity of wall temperature to the flue con-
centrations of NO during these tests, we decided to gather data relating NO
to wall temperature during a warm-up cycle. The standard nozzle was fired
with 3000 SCF/hr of gas at 3% excess O2, 460°C air preheat, and the SFLB.
The results are graphed in Figure 22. The dramatic increase in the level of
NO emissions above 1300°C indicates that an accurate control of wall temper-
ature is required to produce a self-consistent set of experimental data. It
also confirms that the furnace should be operated with as low a wall tempera-
ture as possible.
Furnace Geometry
The IFL.B burner was mounted on our cylindrical (tunnel) test furnace,
which has a cross-sectional area of 1. 17 sq m (12. 6 sq ft) and a volume of
6. 4 cu m (226 cu ft). The following operating parameters were investigated:
secondary combustion air preheat, method of gas injection, and position of
gas injector.
Figure 23 shows the emissions data collected from the IFLB with the
standard gas injector (2-inch stainless-steel tube), a 2000 SCF/hr gas input,
and nonpreheated and preheated secondary combustion air. To determine
the influence of furnace geometry on NO emissions. Figure 23 compares
these data with similar data collected during a test on the larger-diameter
furnace. The cross-sectional area of this rectangular furnace is 2. 3 sq m
(25 sq ft) and the volume is 10.6 cu m (375 sq ft). With the exception of
11-65
-------
700^
6OO-
E
a.
500-
E
t->
o
40O-
Gos Input 3O08 SCFH.AXlfll
Gas Nozzle Baffle Position
3% Excess Oxygen
460* C Air Preheat
Burner flipck
IOOO
1100
1200
1300 1400 1500
Woll Temperature, C
Figure 22. Normalized NO concentration as a
function of wall temperature for the
SF LB burner with a standard gas nozzle
11-66
-------
8001
700-
600-
Nl
O
I
300H
200-
100-
Oas input 1958 SCFH
Gas Nozzle Baffle Position
Wall Temperature 1360* C
Secondary Air Preheat As Labtittf
4° Burner ..fljbcfc Angle
— Legend—
O Cylindrical
O Cylindrical
A Rectangular
v Rectangular
454° C
0-220 c
14
1 7
IN FLUE , %
Figure 23, Normalized NO concentration as a
function of flue O2 for the IF LB burner
with a low-velocity gas nozzle
11-67
-------
secondary air preheat temperature (350 C for the cylindrical furnace, com-
pared with 454 C for the rectangular furnace), all other burner and furnace
operating conditions (gas input, gas nozzle position, burner-block angle, and
wall temperature) were identical. For both conditions of secondary air pre-
heat, the cylindrical furnace had the higher levels of emissions. The area
ratio between the burner-block opening and the burner wall is 35.9 for the
rectangular furnace and 18, Z for the cylindrical furnace. These ratios
reflect the relative sizes of the secondary recirculation zones. The primary
reason for the difference in measured NO concentrations between the two
combustion chambers is the unequal secondary recirculation zone volume.
Nozzle Type
Additional tests were conducted to investigate the influence of gas injec-
tion on total NO emissions. In these trials, we used a high-velocity (127 m/s,
407 ft/s) gas nozzle and a divergent gas nozzle (injection velocity 135 m/s,
436 ft/s). Both injectors have only an axial velocity component. Figure 24
compares the normalized NO emissions as a function of excess O2 from
these gas nozzles. The low-velocity nozzle produces less NO, by a factor of
2, than the high-velocity nozzle. This result contradicts the data collected
from the rectangular furnace, which are presented in Figure 25. Although
the rectangular furnace has a higher gas input (2998 SCF/hr) and a higher
secondary air temperature (450 C), the cylindrical furnace had the higher
emission levels for a given gas nozzle. This occurs because NO production
in gas flames is entirely thermal in origin. Thus, any change in the com-
bustion chamber or fuel-air mixing that results in lower peak and average
flame temperatures reduces the rate of NO production. Conditions that
should reduce NO emissions include decreasing the rate of heat release,
increasing the mass exchange between the flame zone and secondary recircu-
lation zone, and reducing the wall temperature.
During previous trials with the rectangular furnace, the high-velocity gas
nozzle always produced less NO emissions than nozzles with lower injection
velocities. This result was explained by the different fuel-air mixing rates.
An increase in injection velocity was achieved by decreasing the cross-sec-
tional area of the injector, which caused a decrease in the fuel-air mixing
rate and permitted an increase in recirculation and combustion zone mass
exchange. The combustion zone is blended with combustion products, which
H-68
-------
BOOT
I20O-
1(00"
\ooo-
E
ex
fk
O
Z
o
o
BOO'
TOO'
Gas input 1958 SCFH
Gas Nozzle Baffle Position
WON Temperature I36O° C
Secondary Air Preheat 350°
4* Burner Block Angle
60O
5001
Legend
O LOW Momentum Nozzle
A High Momentum Nozzle
^ Divergent Nozzle
3 4
IN FLUE, %
6
Figure Z4. Normalized NO concentration as a
function of flue O2 for the IF LB burner
with low-velocity, high-velocity,
and divergent gas nozzles
II- 69
-------
6OO
Gas Input 3O70 SCFH
Secondary Air Preheat
4 Burner Block Angle
450° C
100
.Legend
OStandard Nozzle, Baffle Position
A Diver gent Nozzle. Bottle Position
Q Axial Nozzle, Baffle Position
16
I T
3 4
IN FLUE, 96
Figure 25. Normalized NO concentrations as a
function of flue O2 for the IF LB burner
•with standard, divergent, and axial gas nozzles
II- 70
-------
result in lower ilame temperatures. However, for the cylindrical fur-
nace, the high-velocity gas nozzle causes higher NO emissions. This means
that even though the same method of gas injection was used in each furnace,
the thermal history of the two flames was different. The major reason for
a higher flame temperature in the cylindrical furnace would be a smaller
mass exchange between the secondary recirculation and the combustion zones.
The smaller mass exchange occurred because of a decrease in the size of the
secondary recirculation zones, which, in turn, resulted from a smaller com-
bustor cross-sectional area.
The divergent nozzle injects the gas radially at a 45-deg angle to the
burner axis. Comparison of Figures 24 and 25 shows that the cylindrical
furnace has the higher levels of emissions for the divergent nozzle. Argu-
ments similar to those presented for high-velocity nozzle could also be made
for the divergent nozzle. However, for both combustion chambers, the
divergent nozzle has lower emissions than the high-velocity nozzle. The
diverging gas injection causes a primary recirculation zone, which allows a
larger dilution of the combustion zone with combustion products than an axial
injector. This increased dilution leads to lower flame temperatures and
reduced emissions.
Natural-Gas, Low-NO Burner I
Jt
IGT has been commissioned by three gas utilities (Consumer's Gas Co.,
Consolidated Natural Gas Co. , and Southern California Gas Co. ) to develop
a burner with low NO emission levels relative to the industrial burner it
x
was intended to replace and with a similar flame geometry and luminosity.
The initial design guidelines for this burner development work were extracted
from the data presented in this paper. After fabrication, the low-NO burner
Jt
I (LNO-1) was mounted on the cylindrical furnace for NO emission level
testing.
Standard Operating Conditions
The initial set of trials was conducted by using 25% primary t r. This
percentage is determined by ratioing the volume of primary air tc total air
needed for stoichiometric combustion and multiplying it by 100. !' igure 26
shows the results of these tests as normalized NO plotted against excess O^.
These data were collected by premixing the primary air and fuel, which gives
II- 71
-------
900-
800-
7001
200-j
ioo-
Gos Input 1980 SCFH
Secondory Air Preheat Temperature AS Labeled
!5° Spin on Secondary Air
25% Primary Air
ciear_ Flame
O Baffle Burner-Stand. Nozzle
A LOW NOX Burn«r-I
V LOW NOX Burner-X
17
3 4
0 IN FLUE, %
Figure 26. Normalized NO concentrations as a
function of flue O2 with 25% primary air
11-72
-------
rise to a clear flame. For comparison, Figure 26 also shows the emissions
curve for the IFLB burner with a standard nozzle. The LNO-I burner was
designed to replace burners in the IFLB burner category. Comparison of the
emission levels of these burners at 350 C secondary air preheat shows that
the LNO-I burner was extremely effective in reducing NO emissions. At a 2%
level of excess O2 (approximately 10% excess air), the LNO-I burner emitted
55% less normalized NO than the IFLB burner. This decrease in NO emis-
sions dropped to 18% for NO secondary air preheat.
Flame Luminosity Adjustment
Adjusting the primary air and fuel so that no mechanically induced pre-
mixing occurred prior to ignition (luminous flame) did little to alter the
LNO-I burner emissions without secondary air preheat. Figure 27 illus-
trates the normalized NO emissions versus excess O2 for luminous flame
operating conditions. With a 350 C secondary air temperature, the lumin-
ous flame had an NO level, at 2% excess O2, 13% lower than that of the clear
flame. Thus, the effect of the mixing rate of primary air and fuel on NO
emissions is directly related to secondary air temperature. At ambient
combustion air temperatures, the luminous and clear flames had nearly the
same emission levels, but at the elevated (350 C) combustion air tempera-
ture, the clear flame had the higher emissions.
Variation in Primary Air
To determine the influence of primary air volume on NO formation, the
amount of primary air was reduced to 15%. Figure 28 shows the normalized
NO-versus-excess O2 relationships determined for this primary air volume.
The clear flame with 350°C secondary air preheat displayed little change in
NO emissions between 15% and 25% primary air. The luminous flame was
somewhat more sensitive to primary air volume, yielding a 10% increase in
the NO emission levels at the lower primary air level. Thus, if a luminous
flame is needed for the desired industrial application, the higher the percent-
age of primary air, the lower the NO emissions.
With an ambient secondary air temperature, both flame conditions pro-
duced the lowest levels of NO measured during the burner trial series, indi-
cating that the percentage of primary air should be reduced as the secondary
air preheat temperature is decreased. To completely generalize the
H-73
-------
400-j
Cos Inpu 1993 6CFH
Secondary Air Temperature AS Labeled
15* Spin on Secondary Air
(5 96 Primary Air
Clear and Luminous Flames
300-
A
O.
to
O
200-1
o
6
too-j
19
365° C
Legend
O Clear Flame
O Ctsqr Float
A Luminous F|0me
V Luminous Flome
I 2 9 4 &
02 IN FLUE.K
Figure 27. Normalized NO concentration as a
function of flue O2 for the LNO-1 burner
under luminous-flame operating conditions
II- 74
-------
Gas Input 1985 SCFH
Secondary Air Temperature AS
15° Spin on Secondary Air
25 % Primary Air
Luminous Flame .
Labeled
30O-
E
a.
a.
«J2OO-
E
o
IOO
i e
12345
02 IN FLUE, %
Figure 28. Normalized NO concentration as a
function of flue O2 for the LNO-1 burner
with 15% primary air
II-75
-------
dependence of NO emissions on primary air volume would require testing
above 25% primary air; however, because of primary fan capacity, these
trials were not possible.
Although this investigation was very successful in demonstrating that the
LiNO-I burner has low NO emission characteristics, particularly at elevated
secondary air temperatures, additional testing is under way to simplify the
burner design and to determine the burner1 s limitation, not only in NO
reduction but also in the types of flames it is capable of producing.
FUTURE WORK
The final step of the program has been initiated by the American Gas
Association in its funding of a field test demonstration for the L.NO-I burners.
IGT is currently contacting industries that might be interested in having
L.NO-I burners mounted on their furnaces with IGT furnishing the burners
and monitoring the emissions during process operation.
SUMMARY
This experimental research program has demonstrated that the NO
emission levels from gas-fired industrial burners can be significantly
reduced by alterations in the combustion aerodynamics. A synopsis of the
operational variables studied and their test results is presented in Table 3.
The technique resulting in the most dramatic reductions of NO emission levels
was external flue-gas recirculation. This verified the intimate relationship
between flame temperature and NO emissions; thus, this program1 s efforts
concentrated on reducing flame temperature either by reducing the rate of
combustion or by diluting the flame with combustion products recirculated
within the furnace using combustion aerodynamics.
Although large variations in NO emissions occur because of changes in the
amount of excess air, reduction levels can be established in addition to their
relative effectiveness by comparing the emission levels at several fixed levels
of excess air. The conclusions reached below are based on an excess air
level equivalent to 3% oxygen in the flue and a combustion air preheat temper-
ature of 450°C.
II-76
-------
Table 3.
Excess Oxygen,
SYNOPSIS OF DATA COLLECTED FOR
THE BAFFLE BURNER
1
Normalized NO, ppm
IFLB Burner (Standard operating conditions: gas input, 3000 SOF/hr;
450 C, secondary air preheat; 2-inch-diameter axial fuel
injector: baffle fuel injector: baffle position; and a 1400 C
wall temperature, 4-degree burner block. )
Standard Operation
Gas Input, 2000 SCF/hr
Secondary Air Temperature:
1100°C
225°C
22°C
Wall Temperature,
EFGR 15T
EFGR 30f,
Radial Nozzle-Throat Position
Axial Nozzle-Throat Position
Throat Position
Divergent Nozzle
Axial Nozzle
8-Degree Burner Block (B. B. )
8-Degree B. B. -Axial Nozzle
8-Degree B. B. -Divergent Nozzle
390
250
195
110
225
130
40
200
110
280
190
210
410
200
220
580
310
235
125
260
180
60
260
190
390
300
250
560
230
350
SFLB Burner (Standard operating conditions: gas input, 3000 SCF/hr;
450 C, secondary air temperature; 2-inch-diameter axial
fuel nozzle; baffle position; 8-degree burner block angle;
and 1400 C average wall temperature. )
Standard Operation
Gas Input, 2000 SCF/hr
Secondary Air Velocity, 1Z5
Secondary Air Temperature:
1000 C
225 C
22°C
Wall Temperature,
EFGR 15"',
EFGR 30^
Radial Nozzle
Divergent Nozzle
Axial Nozzle
16-Degree Burner Block (B. B. )
16- Degree B. B. -Divergent Nozzle
16-Degree B. B. -Axial Nozzle
450
270
210
165
95
270
100
60
660
440
210
210
180
300
660
430
250
270
160
320
150
80
800
580
390
380
400
390
— Cylindrical Furnace —
IFLB Burner (Standard operating conditions:
350°C,
gas input, 2000 SCF/hr,
secondary air temperature; 2-inch diameter axial
fuel nozzle, baffle position, 4-degree burner block angle
and 1 100 C average wall temperature. )
Standard Operation
Axial Nozzle
Divergent Nozzle
460
880
825
610
1170
1090
LNOX-I
180
280
Gas input for this test was 2000 SCF/hr
11-77
-------
For the intermediate flame length baffle burner (IFLB), which has a
tangential-to-axial velocity component ratio of 0. 27, we found the following:
a. Reducing the firing rate from 3000 SCF/hr to 2000 SCF/hr resulted in a
47% reduction in NO emissions.
b. Reducing the combustion air preheat from 450 C to 225 °C leads to a 59%
reduction in NO emissions, and no preheat gives a 78% reduction.
c. By cooling the walls from 1400°C to 1100°C, the NO emissions diminished
by 55%.
d. External flue-gas recirculation reduced NO emissions by 69% for 15%
recirculation and 90% for 30% recirculation.
e. Several types of fuel injectors and positions were tested with a maximum
reduction of 67% measured for the high-momentum axial nozzle in the
throat position. The minimum changes in flame geometry and luminosity
were observed for the divergent nozzle, which showed a 48% decrease
in NO emissions.
f. The burner block angle was increased from 4-degrees to 8-degrees,
leading to only a 4% reduction under standard burner operating conditions.
However, by also changing the method of fuel injection a maximum reduc-
tion of 60% was measured with the high-momentum axial gas nozzle.
For the short flame length baffle burner (SFLB), which has a tangential-to-
axial velocity component ratio of 0.47, we observed the following:
a. Reducing the firing rate from 3000 SCF/hr to 2000 SCF/hr resulted in a
35% decrease in NO emissions.
b. With a 2000 SCF/hr gas input, the combustion air velocity was increased
from 80 ft/s to 125 ft/s, which diminished the NO emissions by 42%.
c. Reducing the combustion air temperature from 450°C to 225°C resulted
in a 59% NO reduction and, with no preheat, a 76% reduction.
d. Decreasing the wall temperature from 1400°C to 1000°C reduced the
NO emissions by 52%.
e. Externally recirculating flue gas and blending it with the combustion air
produced a 77% decrease in emissions for 15% recirculation and 89%
reduction for 30% recirculation.
f. Of the several types of fuel injectors tested, the high-momentum axial
nozzle produced the minimum emissions with a 41% reduction.
g. The burner block angle was increase-:! from 8-degrees to l6-degrees.
This produced a 41% decrease in NO emissions for all gas nozzles tested.
II-78
-------
The IFLB burner was tested on our cylindrical test furnace, which has a
volume of 226 cu ft and an area ratio between the burner block opening and
the burner wall of 18. 2 compared with a 35. 9 ratio for the rectangular furnace.
During these trials we have the following observations:
a. At a 2000 SCF/hr firing ratio, the cylindrical furnace emissions were
97% greater than the emissions from the rectangular furnaces.
b. The high-momentum axial and the divergent nozzles produced increases
in the NO emissions of 92% and 79%, respectively, although both produced
reductions in NO emissions on the rectangular furnace.
This program has led to the design, development, and testing of a low-NO-
emissions burner (LNOX-I), which, hopefully because of its versatile design,
could be used in a large number of industrial applications. The following
conclusion was drawn from the test data accumulated from LNOX-I:
• With 2000 SCF/hr firing ratio on the cylindrical furnace, the LNOX-I
burner resulted in a 54% reduction in NO emissions when compared
with the IFLB burner under similar operating conditions.
Thus, this program has not only demonstrated techniques and variations
in furnace and burner operating conditions, which lead to reduced NO emissions,
but these test results have produced a burner design that has characteristic
low emissions. The final success of this program will be in the near future
when the LNOX-I burner is field-tested.
11-79
-------
-------
INTEGRATED LOW-EMISSION
RESIDENTIAL FURNACE
By:
L. P. Combs, W. H. Nurick, and A. S. Okuda
Rockwell International, Rocketdyne Division
Canoga Park, California
11-81
-------
ABSTRACT
Analytical and experimental studies are establishing practical
technology for simultaneously minimizing pollutant emissions
from and improving fuel economy of oil furnaces for residential
space heating. Objectives are to reduce the emission of ni-
trogen oxides from current levels of 2-1/2 to 3 g/kg of fuel
burned to less than 0.5 g/kg, and to raise season-averaged
thermal efficiency by at least 10 percent above that achieved
by typical current conventional furnaces, while maintaining
minimum emissions of smoke, unburned hydrocarbons, and carbon
monoxide.
The approach has involved research on designing conventional
pressure-atomizing oil burners for low emissions, matching
furnace fireboxes to such low-emission burners, and operat-
ing these optimized components under fuel-conservation con-
ditions. This paper summarizes the research results and
shows how they are being applied in the design, assembly,
and test of a prototype warm-air furnace to verify that the
technology can be used in practical, cost-competitive pro-
duction furnaces.
ACKNOWLEDGEMENT
Work upon which this presentation is based, was performed
pursuant to Contract Nos. 68-02-0017 and 68-02-1819 with
the U.S. Environmental Protection Agency. The EPA Contract
Officer was Mr. G. Blair Martin, Combustion Research Section,
Industrial Environmental Research Laboratory, Research Tri-
angle Park, North Carolina.
11-82
-------
INTRODUCTION
The Environmental Protection Agency has sponsored studies over the
past few years to document the emission of air pollutants from exist-
ing residential and commercial oil-fired space heating equipment (Ref.
1 and 2). Concurrently, the EPA has also supported applied research
programs to determine the effects of design and operating parameters
on exhaust gas emission levels and to devise strategies for minimizing
pollutant emissions (e.g., Ref. 3, 4, and 5). These and related stud-
ies have shown that substantial reductions in total emissions from fur-
naces can be realized by applying combustion control technology, such
as flue gas recirculation, staged combustion, and advanced burner
designs.
As part of those EPA activities, Rocketdyne performed an intensive in-
vestigation of residential and commercial oil burners (Ref. 5) which
led to criteria for optimizing conventional burner designs with re-
spect to emissions. The optimized burners were shown to be capable of
reducing NOX emissions to below 1 g NO/kg fuel, in contrast to typical
manufactured residential burners that operate at about 1 1/2 to 3 g
NO/kg fuel or higher. During the investigation of burner optimization,
several variations in the combustion chamber construction (adiabatic
refractory walls vs metal heat-sink or water-cooled walls) and design
(relative orientation of the burner and chamber axes and relative
burner and chamber diameters) were also studied. The results clearly
demonstrated that the total emissions are sensitive to (1) the design
of various components comprising a residential heating system and (2)
design interactions among the components. It became obvious, there-
fore, that minimum emissions could be achieved only by systematically
optimizing the burner in conjunction with the furnace combustion cham-
ber and heat exchanger as well as operating mode. The research de-
scribed in this paper was undertaken to investigate such system opti-
mization and to delineate furnace design requirements for commercial-
izing the optimum furnace technology.
In view of uncertainties in fossil fuel supplies and prices, and of
the national emphasis on achieving energy independence, an optimized
furnace should, in addition to providing low levels of pollutant emis-
sions, also provide some substantial increase in overall thermal ef-
ficiency. Further, for the developed technology to be adopted commer-
cially, the optimized design must be cost-competitive. In this regard,
significantly increased fuel economy will be important in allowing con-
sumer recovery, within a reasonable number of heating seasons, of any
increased capital costs that might result from modifications required
for system optimization. The need to establish technological require-
ments for optimizing residential heating systems is important, there-
fore, because such effort addresses both the improvement of air quality
and a better utilization of the nation's dwindling energy resources.
11-83
-------
The principal objective of this research program was to establish the
technology required for engineering optimization of residential heat-
ing furnaces combining minimum pollutant emissions with maximum ther-
mal efficiency. Primary emphasis was given to systems fueled with No.
2 distillate fuel oil. General overall goals are: (1) to reduce nitro-
gen oxide emissions to less than 0.5 g NO/kg fuel burned, while main-
taining minimum emissions of CO, UHC, and smoke, and (2) to increase
overall season-averaged furnace energy efficiencies by 10 percent or
more above those achieved by current conventional systems.
BACKGROUND
Typical residential oil furnace emission and efficiency performance data
are summarized in Table 1 for the existing furnace population and for
new furnaces representative of current technology.
Cycle-averaged emissions data, abstracted from Ref. 1 and 4, show that
the products of incomplete combustion (smoke, CO, unburned hydrocarbons,
and particulates) are generally quite low for both the existing popu-
lation and current new furnaces, although on the average those from the
former set tend to be appreciably higher. The low levels result from
achieving the consistent design goal of accomplishing complete combustion
while the higher levels for the existing population are not surprising,
considering the greater component ages and broad diversity of burner
types, firebox design, conversion units, etc., which make up that
population.
Average emissions of the oxides of nitrogen from current new furnaces are
approximately the same as the average from the existing population. This
results from the preponderant use of side-fired, refractory-lined fire-
boxes, irrespective of the burner designs.
Wide ranges of thermal efficiency estimates for the existing residential
furnace population may be found -in the technical literature. Those
listed in Table 1 are representative of the mid-range, and are generally
consistent with the flue gas C02 concentrations (or excess air levels) re-
ported in Ref. 1 and 4. Recognizing that gross efficiencies, based on the
higher heating value of the fuel, are about 5 to 7 percent lower than net
efficiencies, it may be estimated that fuel utilization efficiencies
potentially might be increased by as much as 35 to 40 percent. Achieving
such large gains, however, would entail very significant departures from
current design concepts, and manufacturing, marketing, and utilization
practices. To maximize efficiency, considerably larger heat exchangers
would be needed to cool flue gases to near-room temperature and condense
their combustion-generated water vapor. A number of problems would im-
mediately arise, e.g., inadequate firebox draft, need for corrosion-
resistant furnace and flue construction materials, condensate disposal
11-84
-------
and noncompetitlve Initial costs. For these reasons, it was deemed
appropriate for the current research to address efficiency gains
achievable by less drastic modifications of existing commercial tech-
nology, i.e., improvements that can be achieved relatively near term
and remain cost-competitive in the marketplace.
IMPROVING THERMAL EFFICIENCY
The greatest source of thermal inefficiency in residential space heat-
ing equipment is the convection of heat up the flue. When the burner
is being fired, exhausted product gases carry off substantial sensible
heat and the heat of vaporization of combustion-generated water vapor.
Flue gas sensible heat losses may be reduced by lowering either excess
air or flue gas temperatures. Steady-state net thermal efficiencies
might thus be increased by 5 to 10 percent before furnace operation is
constrained by pollutant formation, low draft, and condensation problems.
During this same time, most installations use heated, humidified living-
space air for burner combustion air and for barometric draft control air.
Although the resultant heat losses are not charged against furnace
thermal efficiency, fuel utilization efficiency can be raised by 10 per-
cent or more by using a sealed air system to bring in outdoor ambient
air for these uses.
When the burner is not being fired, a natural draft flow of air through
the burner, firebox, etc., cools furnace components and continues to
convect heat up the flue. This loss can reduce net thermal efficiency
by as much as 5 percent and is a major fraction of the transient heat
losses which cause cycle-averaged efficiencies to be lower than steady
state. The lower the draft air loss is, the less sensitive is cycle-
averaged efficiency to variations of cycle duration and of fractional
burner-on time. Draft air loss can be eliminated by providing a positive
shutoff device in the combustion air supply. It can also be reduced or
eliminated by having the burner fire more nearly continuous, e.g., by
using modulated flow or high/low/off burner control. No real thermal ef-
ficiency advantage can be realized with such control schemes, however,
because the flue gas temperature varies enough to offset the reduction
or elimination of draft air heat losses.
Heat conducted to the exterior cabinet of a furnace is radiated and
convected to the surroundings. Although treated as a furnace loss,
this heat, in some installations, may contribute directly to heating
the residence and not be a true heat loss. These furnace setting
losses average about 1-1/2 to 2 percent for warm-air furnaces. They
are higher (3 to 3-1/2 percent) for hydronic boilers, because compo-
nent temperatures are more nearly constant during standby than are
11-35
-------
those in warm-air units. Reduction of these losses nominally depends
upon a straightforward economic tradeoff of fuel economy vs more ex-
pensive insulation, although manufacturers must also be concerned
about first-cost competition.
The electrical energy expended in operating furnaces amounts to no
more than 2 to 3 percent of the net heat of combustion of the fuel.
Nevertheless, the minimizing of electric power consumption is a worth-
while goal because of its greater cost, to the homeowner than fuel oil
and because of the factor of 2 1/2 to 3 on total energy savings when
the inefficiency of the electric generating plant is considered.
To maximize furnace thermal efficiency, without departing substantially
from current manufacturing and installation practices, it was concluded
that the following measures should be effected:
1. Design the burner/combustion chamber combination to operate
pollution-free at low excess air levels, e.g., 10 to 15 per-
cent excess air (13 1/2 to 13 percent C02)
2. Lower flue gas temperatures as much as possible. The minimum
practical temperature is probably around 200 C (392 F)
3. Employ sealed air systems for combustion air and barometric
control. Filter the combustion air to prevent long-term
degradation of performance
4. Close the combustion air supply during standby
5. Effect savings in electrical power consumption. Areas to
consider are proper matching of drive motors to air fans
and fuel pumps, interrupted-spark rather than continuous-
spark ignition systems, and solid-state ignition and con-
trol circuits.
REDUCING POLLUTANT EMISSIONS
When clean fuels, such as natural gas and No. 2 fuel oil, are burned
in residential space heating equipment, two classes of combustion-
generated air pollutants are emitted, viz., incomplete combustion pro-
ducts (CO, UHC, and smoke) and oxides of nitrogen. The two classes
are similar in some respects; their concentrations in the flue gases
are usually very low, so that their production does not affect signi-
ficantly combustor performance or thermal efficiency, and their con-
centrations usually are not particularly close to those representative
of thermal equilibrium, but are more likely to be determined by poorly
controlled features of combustion, e.g., regions of poor mixing, spray/
combustor wall interactions, size and strength of recirculation eddies,
11-86
-------
and flame-zone cooling phenomena• These characteristics tend to ob-
viate quantitative analytical characterization of the pollutant emis-
sions; as a result, such characterizations are usually studied experi-
mentally. Conversely, the two classes are dissimilar in some respects,
principally because their usual concentrations lie on opposite sides
of equilibrium. Thus, for example, NO production is enhanced by higher
temperatures and longer flame-zone residence times, both of which tend
to drive the hydrocarbon combustion processes to completion and reduce
carbonaceous emissions.
It was reasonably well established in the earlier oil burner studies
(Ref. 5) that an optimized conventional burner can operate reliably at
low excess air levels and produce acceptably low CO, UHC, and smoke.
Its NO emission levels, however, were a factor of 2 or more higher
than the target reduction below 0.5 g NO/kg fuel, and it was uncertain
whether that goal could be reached by the optimizing of the furnace
combustion chamber for this burner.
Therefore, the experimental investigation, which used research combus-
tion chambers rather than typical furnaces, consisted of three separ-
ate, but related parts. One part dealt exclusively with the 1 ml/s
(gph)* conventional burner, optimized to produce low emissions, as de-
scribed in Ref. 5. This effort sought to optimize the matching of the
fire box (combustion chamber) to that fixed-burner design. The other
two parts branched out to burners having, in one case, forced flue gas
recirculation (FGR) to the burner air intake and, in the other case,
forced combustion gas recirculation (CGR) from the combustion chamber
to the burner air intake. These efforts were also focused upon the
optimizing of the burner/firebox combination for low emissions.
The first of these approaches—the optimizing of the combustion cham-
ber matched to the optimum low-emission burner—was found to be the
best approach. The experimental investigation with the 1 ml/s combus-
tion gas recirculation (CGR) burner was frustrated by an inability to
simultaneously satisfy the NO, CO, and UHC emission goals. Because of
these problems and the inherent potential problems of moderately high
temperature mixed gases being passed through the combustion air fan,
it was concluded that further work with this burner concept was not
warranted.
*1.00 ml/s = 0.951 gph. The 5-percent difference is neglected in this
paper when reference is made to burner firing rates.
II- 87
-------
Tests of the 1 ml/s flue gas recirculation (FGR) burner were more suc-
cessful and showed good potential for satisfying the emission goals
under efficient operating conditions. However, the actual achieving
of all the emission goals would probably be contingent upon the use of
more-complicated burner start sequence to eliminate start-spike emis-
sions of CO, smoke, and UHC which were experienced under conditions
having acceptably low NO.
Thus, it was concluded that the optimized conventional burner in opti-
mized combustion chambers is a preferred choice over both the C6R and
the FGR burners. With this choice, the emission goals can be met, low-
excess air levels can be employed, and fewer new or peripheral problems
are likely to be encountered in the commercializing of this burner than
with the more complicated recirculation types. Additionally, burner
simplicity will ensure lower furnace costs with this choice. The re-
mainder of the paper describes the testing, results, and implementation
of that burner.
DISCUSSION
OPTIMUM BURNER/COMBUSTION CHAMBER MATCHING
EXPERIMENTS
Experimental Apparatus
The 1 ml/s (gph) optimum low-emission residential oil burner, whose
development is detailed in Ref. 5, was used during this study (Fig. 1).
The optimum burner was cyclically fired in a set of three cylindrical
research combustors. In conformity with the testing reported in Ref.
2, the burner was fired for 10 minutes of 30-minute cycles. The three
combustors constituted a matched set, allowing considerable dimensional,
structural, and coolant variations. The basic approach is illustrated
in Fig. 2 for one of the chambers. Each chamber was a 1.27 m (5.00 ft)
long, flanged section of steel pipe with a stubby, flanged, side-arm
section of the same size of pipe attached near one end. The tunnel-
fired orientation is illustrated in Fig. 2, in which the oil burner
fits into the annular flange depicted at the left end of the chamber.
In this orientation, the side-arm was redundant and so was positioned
at the opposite end from the burner and simply blanked off. The steel
chamber could be either lined with a refractory fiber insert to form
an adiabatic combustion zone, or left unlined and cooled by some means.
A movable, spiral-wound, finned-tube heat exchanger (shown inserted in
the opposite end of the combustion chamber from the burner) provided
effective variation of the combustion zone length. To achieve the
side-fired configuration, the combustion chamber was simply turned end-
for-end, with the blank and burner port flanges relocated as appropriate.
11-88
-------
Three chamber diameters were selected such that addition of refractory
linings to the larger two would produce lined chambers having inside
diameters comparable with the smaller two unlined chambers, viz.:
Steel Pipe
ID, m (in.)
0.162 (6.36)
0.222 (8.75)
0.222 (lined)
0.279 (11.0)
0.279 (lined)
Nominal Combustion
Chamber ID, m (in.)
0.162 (6.36)
0.222 (8.75)
0.175 (6.89)
0.279 (11.0)
0.22 (8.7)
Nominal Thickness
of Pyroflex, m (in.)
0.024 (0.93)
0.030 (1.18)
In use, the major axes of the chamber were vertical, with the burner
firing vertically upward when tunnel-fired and horizontally when side-
fired.
A water-cooled heat exchanger was used to accomplish rapid cooling of
the combustion gases as they flowed out of the primary "firebox" por-
tion of any of the combustion chambers. It was intended that the gas
temperature be quenched rather rapidly so that changes in heat ex-
chariger position (i.e., firebox length) could be readily correlated
with variations in pollutant emissions. The heat exchanger was a
spiral-wound, finned-tube assembly designed to fit inside all three
combustors and to be positioned anywhere along the length of a chamber.
The exchanger was designed to cool the exhaust gases down to 200 to
350 C (392 to 662 F) at a maximum water coolant flowrate of 3.8 x 10~4
ta3/s (6 gal/min). A 0.0127 m OD by 0.00124 m wall (1/2-in. x 0.049-in.)
stainless steel tubing was helically wound with 0.006 m (0.25 in.) high
by 0.00051 m (0.020-in.) thick carbon steel fins spaced at 394 fins/ra
(10 fins/in.). The carbon steel fin material was selected over stain-
less steel for its higher thermal conductivity and nickel-chrome clad
plating provided resistance to corrosive exhaust products. Approxi-
mately 4.3 m (14 feet) of finned tubing was required in the construc-
tion of the heat exchanger, resulting in a total heat transfer surface
area of about 1.4 m2 (15 ft2).
Several semicircular baffle plates were cut from 21-gage stainless
steel sheet, slipped between coils of the heat exchanger, and wired
in place. Cut so their outside diameters would just fit comfortably
in the inside diameter of a chamber, the baffles ensured that the
gases passed repetitively over the heat exchanger coils and prevented
them from bypassing around the outside of the coils.
11-39
-------
The optimum burner was fired in the various research combustion cham-
bers at an outdoor test facility whose principal components were at-
tached to a waist-high steel table. A Unistrut superstructure at one
end of that table supported the vertically mounted combustion chamber
and allowed suspension of the spiral-wound heat exchanger within it.
The facility was organized for rapid and easy changing of combustion
chambers, burner orientation, and heat exchanger position. Minimum
protection from inclement weather was provided by a simple sheet metal
roof over the test apparatus.
Experimental data requirements were primarily concerned with flue gas
pollutant concentrations. Most pollutant species concentrations were
measured by conducting a continuous flue gas sample to a train of anal-
ysis instruments located within a nearby laboratory. Flue gas smoke
content was measured intermittently at the flue with a manual smoke
meter. The instruments used, analyses performed, and types of data
obtained are described and discussed in Ref. 5.
Experimental Results
Data were recorded during 160 tests of the apparatus previously de-
scribed; in addition to chamber diameter, construction, and burner
orientation variations, the heat exchanger position and burner ex-
cess air level, or stoichiometric ratio, were varied systematically.
Typical measured emission data are illustrated for one subset of these
tests in Fig. 3. In this example, flue gas pollutant concentrations
varied fairly regularly with both stoichiometric ratio and heat ex-
changer position, although this was not always consistently so.
The results of all the parametric variations tested were synthesized
into a tabular summary form amenable to extraction of trends and de-
sign criteria. That summary is given in Table 2, where the first
three columns describe the combustion chamber, and the next three col-
umns list the chamber lengths and stoichiometric ratios for which the
pollutant concentrations in the flue gas were acceptable. Few of the
NO emissions of the combustor were acceptably low so conditions that
minimized NO production are listed under that pollutant. Further, the
right-hand column of Table 2 lists approximate flue gas NO concentra-
tions at 25-percent excess air (12-percent C02) for 0.50 m (20 in.)
long chambers. It was seen that, with a few exceptions, the combus-
tion chamber designs tested generally failed to achieve the goal of
NO < 0.5 g/kg fuel. However, the observed trends in the data support
confidence that the goal can indeed be satisfied.
I 1-90
-------
Co^bustor Design Criteria
Combustor design criteria that can be derived from Table 2 (and the
larger body of data upon which it is based), are subdivided into cate-
gories of burner orientation, chamber size, and cooling medium in the
following paragraphs.
Burner Orientation. Combustion chamber design requirements for ensur-
ing acceptable levels of CO, UHC, and smoke emissions were quite simi-
lar for the side- and tunnel-fired burner orientations. That observa-
tion held for NO emissions from water-cooled combustors, as well. How-
ever, NO production in refractory-lined and air-cooled chambers was
about 1 1/2 to 2 times as high in the side-fired orientation as in the
tunnel-fired. Ostensibly, this latter phenomenon resulted from sub-
stantially longer gas residence times at high temperatures in the
stronger and more complicated eddies of a side-fired chamber.
Combustion Chamber Size. Chambers smaller than about 0.20 m (8 in.) ID
must be refractory-lined to avoid operating with unacceptable combus-
tion roughness. Operation of all designs was acceptable in this regard
when their inside chamber diameter was 0.22 m (8 in.) or larger.
A tendency existed for larger-diameter chambers, with or without in-
sulation, to require longer chamber lengths to achieve comparable
levels of carbonaceous pollutants and, concurrently, to produce sub-
stantially lower levels of NO. Both phenomena were undoubtedly linked
to the ingestion of recirculating combustion gases into the flame zone.
For a given burner, larger and stronger recirculation eddies can be
established in bigger chambers, reducing the combustion intensity some-
what and lowering the rates of burnout of carbonaceous species. Also,
larger-diameter chambers have greater wall areas for convective and
radiant transmission of heat from the flame zone, which reduces peak
flame temperatures somewhat. Increased average gas residence times
should produce opposing trends for both NO production and carbon burn-
out, which could help account for some anomalies in the data. Also,
the steady-state recirculation and radiation effects may be partially
masked by starting transient effects, particularly with respect to
smoke and UHC data.
Short combustion chambers generally were more favorable for low NO but,
if they are too short, the fuel may not be completely burned before
combustion reactions are quenched in the heat exchanger. A comb ustor
length of 0.5 m (20 in.) is a suitable compromise for most of the cham-
ber designs tested. Slightly shorter chambers might be appropriate if
refractory linings were to be used.
H-91
-------
Chamber Cooling Medium. Refractory-lined combustion chambers generally
exhibited less combustion roughness and less sensitivity to starting
conditions but also produced higher NO concentrations than did cooled-
wall combustors. The air-cooled, side-fired configuration had better
starting characteristics (but higher NO) than did the air-cooled,
tunnel-fired chamber. Water-cooling appeared to be preferable to air-
cooling, partially because of lower side-fired NO levels, but also be-
cause of smoother starting and more consistent CO and NO emission re-
sults. The water-cooled chambers in shorter lengths, however, were
prone to produce excessive CO.
It was inferred that the key effect of both air- and water-cooling is
the reduction of flame zone temperatures that results from removing
heat from the combustor walls. The 0.279 m (11 in.) diameter water-
cooled combustor, which met all emission goals, had approximately 20
percent of the heat of combustion of the fuel extracted from the fire-
box. The single "forced draft" air-cooled chamber was designed to
duplicate that heat removal rate, whereas considerably less-effective
natural convection prevailed in the other air-cooled configurations.
That forced-draft air-cooled chamber, however, had an intermediate
0.25 m (10 in.) ID. Its flue gas NO concentrations were approximately
midway between those from the 0.222 m (8 in.) and 0.279 (11 in.) ID
water-cooled combustors, indicating that forced draft air-cooling can
be just as effective as water-cooling in controlling NO emission levels.
Optimized Combustors for the 1 ml/s (gph)
Optimum Burner
To reduce NO emission levels below 0.5 g/kg No. 2 fuel oil burned while
maintaining acceptably low carbonaceous pollutant emissions during
operation at low excess air levels (10 to 15 percent), the 1 ml/s (gph)
optimum oil burner should be fired into a combustor having the follow-
ing design attributes:
1. Cooled walls, with either water or forced-draft air coolant,
which extract approximately 20 percent of the heat of com-
bustion of the fuel from the flame zone. It is also impor-
tant to design for nearly uniform inside wall temperatures
and for retention of elevated wall temperatures during stand-
by periods.
2. The ID of the combustor should be 0.28 m (11 in.) or greater.
(It is inferred that NO emission levels will continue to go
down as diameter is increased above 0.28 m (11 in.), but this
hypothesis has not been tested.)
11-92
-------
3. The effective combustion chamber length, from the burner end
to the location where the furnace heat exchanger begins to
quench the combustion gas temperature rapidly, should be
greater than 0.5 m (20 in.) and perhaps as long as 0.75 m
(30 in.).
4. The combustor may be either side-fired or tunnel-fired, which-
ever is convenient for a particular furnace or boiler design.
CONCEPTUAL PROTOTYPE FURNACE DESIGNS
Preliminary conceptual designs were laid out for a prototype warm-air
furnace and a prototype hydronic boiler, each embodying the preferred
approaches previously discussed for improving efficiency and reducing
emissions. Each design concept layout was discussed with one or more
representatives of a manufacturing company that makes and markets res-
idential space heating equipment. Thereafter, the designs, the manu-
facturer's comments, criticisms and suggestions, and appropriate de-
sign modifications were reviewed with the EPA Project Officer prepa-
ratory to selecting one of the designs for fabrication and experimental
evaluation.
The preliminary conceptual designs are illustrated and briefly described,
Greater detail is given in Ref. 7.
Conceptual Design: Warm-Air Furnace
A conceptual design layout for a prototype low-emission, improved-
efficiency, warm-air furnace is shown in Fig. 4. It is based upon
modifying an existing, commercially purchased, warm-air furnace* to
accommodate an optimized, low-emission oil burner with sealed air sup-
ply and a large, air-cooled, steel combustion chamber. This approach
allows key concepts to be tested without the expense and delay of de-
signing and fabricating most furnace components. Further, it ensures
that commercially practical heat exchangers, air filters, blowers,
cabinets, and controls are imposed upon the concepts to be evaluated.
Referring to Fig. 4, the burner vestibule is a closed volume supplied
with combustion air, through a small filter, from an external, sealed
air system. Not shown is a positive shutoff device for this air when
the burner is inactive; this device might be a solenoid-actuated but-
terfly valve in the combustion air inlet line.
*A Lennox Model 011-140 furnace was selected.
11-93
-------
The combustion chamber is offset slightly toward the back of the fur-
nace to accommodate its large diameter. The chamber is comprised of a
standard pipe cap welded to an eccentric reducer whose small end is
mated to the furnace heat exchanger. Welded to its outside are many
longitudinal fins to enhance transfer of heat to the vertically up-
flowing, coolant air. The fins also act as a heat sink to help keep
the combustion space warm during standby periods.
Conceptual Design; Hydronic Boiler
The conceptual design layout for a prototype, low-emission, improved-
efficiency hydronic boiler is shown in Fig. 5. This design is based
on substantially all-new construction, principally because no commer-
cially available unit was identified as being readily adaptable to the
larger-diameter, water-cooled combustion chamber for the optimum
burner.
A horizontal, tunnel-fired, water-cooled combustion chamber was selec-
ted because it offers boiler structural and package volume advantages.
The horizontal firebox is followed by multiple sequential passes
through a water-cooled, fire-tube heat exchanger. Commercially avail-
able Turboflue internally finned tubing is indicated for this service.
The combustion chamber and heat exchanger tubes are all attached to
the same boiler tube sheets. Between the combustion chamber and the
first heat exchanger pass, combustion gases are reversed by passage
through an insulated ISO-degree return manifold, whose volume is con-
sidered to be part of the combustion chamber. Similarly, there are
180-degree return manifolds between successive heat exchanger passes.
The conceptual hydronic boiler design also has the usual plumbing and
control features needed for installation and operation in a residence,
including a tankless coil for heating domestic hot water. Needed, but
not shown in Fig. 5, is an exterior cabinet that should be well insu-
lated to reduce external heat losses.
FUTURE WORK PLANS
Following comparative assessments of the two preliminary prototype de-
signs, including their potential overall emissions and fuel economy
impact, the warm-air system was selected as the more promising; there-
fore, a prototype warm-air unit will be built for further study in the
continuing work. The data and experiences gained will be applied to
the final design of a cost-competitive, commercially producible, warm-
air furnace that embodies the derived low-emission, improved-efficiency
technology.
I I- 94
-------
REFERENCES
1. Hall, R. E., J. H. Wasser, and E. E. Berkau, "NAPCA Combustion
Research Programs to Control Pollutant Emissions From Domestic
and Commercial Heating Systems," New and Improved Oil Burner
Equipment Workshop, NOFI Tech. Publ. 108 ED, National Oil Fuel
Institute, Inc., New York, New York, September 1970, pp. 83-93.
2. Hall, R. E., J. H. Wasser, and E. E. Berkau, "A Study of Air
Pollutant Emissions From Residential Heating Systems," EPA-650/
2-74-003, Environmental Protection Agency, Research Triangle
Park, N.C., January 1974.
3. Martin, G. B., and E. E. Berkau, "Evaluation of Various Combus-
tion Modification Techniques for Control of Thermal and Fuel-
Related Nitrogen Oxide Emissions," presented at the Fourteenth
Symposium (International) on Combustion, Pennsylvania State
University, August 1972.
4. Barrett, R. E., S. E. Miller, and D. W. Locklin, "Field Inves-
tigation of Emissions from Combustion Equipment for Space Heat-
ing," EPA R2-73-084a (API Publ. 4180), Environmental Protection
Agency, Research Triangle Park, N.C., June 1973.
5. Dickerson, R. A., and A. S. Okuda, "Design of an Optimum Distil-
late Oil Burner for Control of Pollutant Emissions," EPA-650/
2-74-047, Environmental Protection Agency, Research Triangle
Park, N.C., June 1974.
6. Peoples, G., "Sealed Oil Furnace Combustion System Reduces Fuel
Consumption," Addendum to the Proceedings, Conference on Improv-
ing Efficiency in HVAC Equipment and Components for Residential
and Small Commercial Buildings, Purdue University, Lafayette,
Ind., October 1974.
7. Combs, L. P., and A. S. Okuda, "Residential Oil Furnace Optimi-
zation," Interim Report for Phase I of Contract 68-02-1819,
R-9815, Rocketdyne Division, Rockwell International, Canoga Park,
California, draft submitted to EPA, August 1975.
11-95
-------
EXTERNAL VIEW
OPTIMUM HEAD
Figure 1. 1 ml/s (gph) Optimum Low-Emission Residential
Oil Burner
11-96
-------
REFRACTORY INSERT
WPYHOF LEX LINER
IOOU BLE THICKNESS)
0.20 m ID /
q.JSm s -^
HEAT EXCHANGER
I WATER
\ -
,/BLANK
. // FLAWGE/
SIDE-FIRED /
iURNER PORT /
MOVABLE HEAT
, -^ EXCHANGER (SPIRAL'
WOUND. FINNED-TUBE)
- 0.2S nvOIAMETER
COMBUSTION CHAMBER
TUNNEL-FIRED BURNER PORT
END FLANGE
Figure 2. Experimental Combustion Chamber Arrangement
X X
o
S
o
s20
X
i 10
Q
2
4
o 0
7
LU
i s
rt
5
1 3
o
CD 1
— • -| OT 15
0.40 |
0.7S -BACHAHACH SMOKE • 1 •
' ' ' <
'=• fl-SD « : 0 ™
', , '' : |2
osc ow 6
\ '• \ z
• ••'' \ £
x ^\ m
V"^- ^ - r^' ' r ^ n
1.5
1
o > o
0 >S m IMT EXCHANGER POSITIONI .
1 S
1 0 50 X
1 o 40 O
\ 'ODD u 0.5
- xOx 1 ~
X^X. \^ i
-- ~"\
- - - - aACHARACH SMOKt - 1
-
0.30 m (HT EXCHANGER
040 , POSITION)
075 O.JO
0.90
o.s6""*-»^ — — ^ °75 0*
_
O.n " IHT. EXCHANGER POSITION!
.•-^-.^^.-.-^^s^*"-
^ - ' 030
1.00 1.10 1.20 1.30 140 1.60 1.00 1.10 I.JO 1.30 1.40 1.50
STOICHIOMETRIC RATIO STOICHIOMETRIC RATIO
Figure 3. Cycle-Averaged Flue Gas Emissions, 10 ml/s
Optimum Oil Burner in a 0.279 m Diameter,
Air-Cooled, Side-Fired Combustor at Various
Heat Exchanger Positions
I L-97
-------
1
U-r-^"^~^...Jl
! • ' v •
pin ....- .ill
°"cl>i
. — f - -----
! ^
E
f?
O
i
v
-------
u
•rl
G
O
M
*C»
fs
53
C
o
CO
•H
S
O
0)
ex
o
M
-------
TABLE 1. TYPICAL MISSION AND EFFICIENCY PERFORMANCE OF
RESIDENTIAL OIL FURNACES
CYCLE-AVERAGED EMISSIONS
SMOKE IBACHAflACK SCALE)
CO
UHC
PARTICUt-ATES
NOX
THERMAL EFFICIENCIES (NET)
STEADY STATE
CYCLE-AVERAGED
AVERAGE OF EXISTING
FURNACE POPULATION
~ NO 2
~ 06G/KG FUEL
~ OOBG/KG FUEL
~ 0.3G/KG FUEL
~ 2.50/KG FUEL
& TO 75%
50 TO 60%
AVERAGE OF CURRENT
NEW UNITS
~NO. t
~ 03G/KGFUEL
--0.06G/KG FUEL
— 0.2G/KG FUEL
~2.SC/KG FUEL
75 TO 80% i
66 TO 70%
TABLE 2. SUMMARY OF RESULTS FROM 1 ML/S (GPH) OPTIMUM BURNER/COMBUSTION
CHAMBER MATCHING EXPERIMENTS
CHAMBER DESIGN ATTRIBUTES
CONFIGURATION
TUNNEL-FIRED
(COAXIAL)
SIDE-FIRED
(PERPENDICULAR
PORT)
COOLING METHOD
AIR-COOLED
(STARTABLE OVER A
VERY NARROW SR
RANGE)
WATER-COOLED
INSULATED
AIR-COOLED
(FORCED DRAFT!
WATER-COOLED
INSULATED
1 D .fli
0.162
0.222
0.279
0.222
O.I 75
0.22
0.162
0.222
0.279
0.250
0.222
0.279
0.175
0.22
REQUIREMENTS FOR ACCEPTABLE (OR
MINIMUM) EMISSIONS
CO & UHC
ROUGH-BURN
Lc > 0.5 in
LC = 0.75m
Lc > 0.5 m
Lc > O.S m
Lc 2 0.5 m
ROUGH-BURN
Lc 2 0.6 m
Lc >0.4 m &
SR > 1.15
Le > 0.5 m
Lc ;> 0.5 m
Lc J 0.5 m
Lc 2 0.3 m
Lc > O.S wi
SMOKE 1 NO. 1)
NG THROUGHOUT OPE
SH > 1.05
SR > 1.1 Oft
LB >0.5m
SR >1.15
Lc >0.7S m
OR SR >1.3
Lc > 0.5 mil
SR > 1.15
NG THROUGHOUT OPE
Lc > 0.5 ni
SR > 1.15
Lc > O.S m ft
SR > 1.08
LC 2 0-5 m, SR > 1.05
Lc >O.S m &
SR > 1.2
Lc >0.5&SR J1.1
Lc > O.S m &
SR > 1.25
Le > 0.5 en &
SR > 1.1S
NO
RATING RANGE
Lc «• 0.4 m
Lc =• 0.4-5.0 m
SHORT LC'S BEST
SHORT LC'S BEST
SHORT LC'S BEST
BATING RANGE
Le - 0.75 m SLIGHT
FAVOR
SHORT LC'S BEST
SHORT Lc'S BEST
SHORT LC'S BEST
SHORT LC'S BEST
SHORT LC'S BEST
DATA MIXED
NO.g/VgFUEL
9 L • 0 .5 m &
1.25 SR
I0.6)
(0.41
0.7
1.3
0.9
0.9
0.6
0.6
0.7
0.4
2.0
1.9
II-100
-------
10:05 a.m.
Integrated Low Emission
L. Paul Combs, Rocketdyne
Am I right that you still need about 20-25% excess
air for a normal operation to obtain zero smoke number?
That chart was not typical of our current testing.
With the optimized firebox, we have designed the proto-
type furnace to operate at 15% excess air.
Have you experienced blue flame phenomenon in the
course of your work?
With the recirculating gas burners, yes, we did have
blue flame condition.
Did you think it worthwhile to consider that as an
extension of your research?
It has promise for further reducing the NO to levels
perhaps half of our target here, on the order of a
quarter of a gram per kilogram. However, it has
difficulties. The blue flame burners are hard to
start because the combustion chamber is filled with
colder gases than during steady state.
Do you have to design a combustion chamber for your
particular burner or vice versa? Does there have to
be a mating of burner design and combustion chamber
design or can this be a conversion burner?
The optimization wor.-c. that we have done has all been
done with one gallon per hour firing levels. We have
done a small amount of experimentation at different
firing rates and the indication is that the combustion
chamber can take some latitude in firing rate. I think
it is best optimized for a particular firing rate.
We are working on another contract in which we are
assessing the direct commercial application of the
£1-101
-------
optimum burner head for use as a retrofit device
for existing furnaces.
In the beginning of your talk, you mentioned that
one of your goals was to make something that would
fit codes, and I question the shutoff damper in the
exhaust flue line. I rather suspect that that has
got to be a code problem. Did you look into that?
We discussed it with the enginnering people at Lennox.
The shutoff device that we had originally sketched
in our design they objected to pretty strenuously,
from that standpoint. The use of a solenoid actuated
butterfly valve they thought would have no problem in
passing code. It may introduce other problems that
impact the codes. The draft air flow through the
burner and firebox helps to keep components such as
the electrodes, the flame sensing cadmium sulfide cell,
the electrical wiring, the control wiring and the
vestibule cooled. With that air shut off, we may have
difficulty with too high temperatures and need to
put in additional insulation and so on.
Your design burner is going to operate at 15% excess
air. Now does that represent the excess air level
as measured in the flue also? Most heat exchangers
like Lennox (I have a Lennox in my house) are not de-
signed to take the flame temperatures that result with
15% excess air. And that is why on a normal household burner,
there is a large amount of excess air used just because
heat exchangers are typically not made out of a high alloy
material, and you are talking about gas to gas heat exchangers.
Now, you say 15% excess air gives a high thermal efficiency
if you maintain it through the whole thermodynamic boundary
11-102
-------
of your system. Do you do that?
It is our intention to do that. We are helped in that
regard by the fact that we are going to remove 20-25%
of the heat from the combustion chamber so that the
combustion gases will be at least partially cooled
by the time they hit the heat exchanger.
When you say you hope to, this means that you haven't
been able to do it in your prototype furnace?
We haven't built a prototype furnace. We have de-
signed it and it is being built now.
[Moderator]
Since we shut off the questions on Paul's presentation,
he has agreed to endure double jeopardy; and if Rix Seals
and the other gentlemen are still here and would like to
ask their questions, we would welcome them at this time.
Paul, what was the final temperature, not after dilution
but after leaving the heat exchanger on the integrated
furnace setup? I understood it was operating at 15%
excess air but I couldn't see the chart from the back
and I wasn't sure what the final temperature was.
The chart didn't tell what the final temperature is.
We anticipate that we will probably have flue gas
temperatures on the order of 400° F. We have run
some computer analyses of the behavior of the existing
Lennox compact heat exchanger that we will use down-
stream with the combuster but our analyses are a little
uncertain as to what is going to happen. That's part
of the testing that will be performed, to characterize
the behavior of the firebox within the combustion
chamber and the heat exchanger combination.
11-103
-------
The second point that I wanted to ask is not intended
controversially or that sort of thing; but I quite
often see and hear comments regarding the advantages of
using an outside air source and I won't attempt to
dispute that there are energy savings. However, have
you in your consideration or percentage assignments
or evaluations of this taken into account the fact
that the load on the heat exchange or combustor system
is increased when it is looking at colder combustion
air than when it is looking at inside combustion air
(let's say the 70 degree differential that might exist
which in turn might distort the apparent advantages
of using outside air)? I might re-express that: you
are raising the same mass or volume of combustion air,
but you are raising it, say, in case (1) if it were
outside design at zero you might be raising it from zero
to 400 or 500 final exit. Whereas in case (2) using
inside air one might be raising it from 70 to 400 or 500
outside. Has that been allowed for in your evaluation?
We ran a substantial number of computer analyses
simulating the furnace system that I didn't talk about
before. The analytical portion did include examination
of the effect of varying the combustion air temperature and
it does have a very modest impact on the furnace thermal
efficiency. I might say that there is a several hundred
page interim report that has been submitted to EPA
in draft form, and if things go according to plan
that will be available toward the beginning of
II- 104
-------
next year, for those who want more detail on that material.
I didn't point out at the beginning of my presentation that
I do have co-authors. One sent me a note saying I should
have mentioned that we did run a preliminary test on the
finned firebox and found that as opposed to the cylindrical
combustion chamber I illustrated, it was capable of removing
the heat required and that the emissions were surpressed as
predicted.
Q: I would like to offer one comment to clarify the question
Rix just asked you. The comment regards the relationship
between the air that the furnace requires and the contribution
of that air to the building infiltration. Now in your analysis,
if I heard it correctly, you assume that all the air that
the furnace needs for combustion and for draft control comes
sooner or later from outside. Rix wondered what happens if
you compare that with the other situation where you have to
increase the heating for the same mass of air through a larger
differential. Now, to add to the confusion, we did some
measurements last heating season, and it turns out that not
all the air that the furnace needs for combustion and for
draft control comes directly from outside, or, in other words,
contributes infiltration. Typically between 30% and 100%
of that air contributes infiltration. Meaning that some of
the air that goes to the furnace is really coming from a
reduction in the exfiltrating air through the cracks near
the attic where you have a higher inside pressure compared to
the outdoor pressure. In some houses, depending on how they
are built, whether one or two stories, and depending on how
11-105
-------
leaky they are, all the air that Is required by the
furnace comes from increased infiltration. If I didn't
know anything about a house, I would guess that about
70% of the furnace air requirements contribute to in-
filtration. A year ago when I didn't know anything
about that, I assumed like you did, that 100% is a
contribution. The question regards the effect of
combustion chamber size on smoke and NO increase or
decrease, and I think your conclusion was as you
increase combustion chamber size your NO generally
goes down. Is the same true for smoke or was the
smoke going the other way?
It is true, there is a tendency that as the combustion
chamber diameter is increased to surpress the NO
formation, a somewhat longer chamber is required to
burn out the products of incomplete combustion. CO
and the unburned hydrocarbons did go up and so we
would use a longer combustion chamber to keep them
down.
I see; so really you have to maintain the volume,
so to speak, to maintain the soot burnout the same;
but by reducing the diameter you can reduce the NO.
Is that correct?
Yes, that is correct. Concerning your comment, I
might say that we have not tried analyzing the
residence itself. Our analysis has been restricted
to the furnace unit and the inputs and outputs that
it experiences.
Q: When you take the combustion air from outside the
building, and, of course, assuming that you possibly
II- 106
-------
could have an installation of this type in quite cold
climates, would not the inlet air temperature have
some effect on your combustion process, and possibly
your emission levels? Also, since the device has a centri-
fugal fan, the actual air temperature and the consequent
density is going to determine somewhat the fan output pres-
sure, which will somewhat affect the combustion air flow
throughout the system. Would you have any comment on that?
A: You are right, there will be effects. The thermal
effect is really pretty marginal. I don't recall, nor do
I have here, the graphs of the decrease in thermal efficiency
as a result of the colder air supply. I don't know about
the emissions; it is conceivable that there might be a
slightly higher starting spike on the CO and unburned hydro-
carbons. The colder air temperatures, as the airflow is
established before ignition is established, will tend to
chill the combustion chamber more, and it could conceivably
have a starting spike effect. Yes, I'm sure it will.
11-107
-------
-------
THE CONTROL OF POLLUTANT EMISSIONS
FROM
OIL FIRED PACKAGE BOILERS
By:
M. P. Heap, T. J. Tyson, E. Cichanowicz
Ultrasysterns, Inc.,
and
R. E. McMillan, F. D. Zoldak
The Foster Wheeler Energy Corporation
11-109
-------
ACKNOWLEDGEMENTS
The authors wish to acknowledge the help and guidance given by
Mr. G. B. Martin, EPA Project Monitor for the current projects and
all the members of the EPA-API Steering Committee. We are grateful
to Mr. W. Axtman of the ABMA for his assistance on various matters
concerning package boiler practice and boiler sales data. Mr. McComis
was responsible for all the laboratory investigations and the authors
gratefully acknowledge the help of Mr. R. E. Sommerlad of the Foster
Wheeler Energy Corporation. Dr. L. J. Muzio was responsible for
Phase I and II of the EPA-API program.
11-110
-------
1.0 INTRODUCTION
A package boiler is described in the ABMA Lexicon as:
"a boiler equipped and shipped complete with fuel
burning equipment, mechanical draft equipment,
automatic controls and accessories. Usually
shipped in one or more major sections."
Although boilers with a capacity up to 250,000 Ibs of steam per hour
can be shipped as a single unit by rail or truck, larger units
(250,000 to 350,000 Ibs of steam per hour) must be modularized. Thus
the term packaged encompasses a wide range of equipment, size range,
design type and fuel capability. This paper is somewhat more limited
in scope since it is mainly concerned with firetube boilers and
practical experiments with equipment in the lower size range (up to
25,000 Ibs of steam per hour). However, it is contended that the
general principles established from these experiments with regard to
pollutant formation are applicable to the complete range of package
boilers firing oil.
The genesis for the results presented in this paper was a
recently concluded program jointly supported by the EPA and API. The
objective of this program was to provide information on the control of
nitrogen oxide emissions from package boilers by the use of flue gas
recirculation or staged combustion techniques, and this work is currently
being continued under EPA support (EPA Contracts 68-02-1498 and 68-02-
1500). Several workers have established the influence of the fuel/air/
combustion product mixing processes on NO formation in turbulent
X
II- 111
-------
diffusion flames £l»2 and 3^j. Results of this investigation emphasize
that due regard must also be paid to those design features which control
these mixing processes in order to maximize the effectiveness of both
staged combustion and flue gas recirculation as NO control techniques.
Thus, in the ongoing projects emphasis is placed upon optimizing the
total combustion system to achieve NOX control.
The EPA-API program consisted of three separate phases:
Phase T- - Construction of a versatile combustor
Phase II - Experimental investigations in that versatile
combustor to determine the optimum method of
applying both flue gas recirculation and staged
combustion to control NO emissions
x
Phase III - Demonstration of the applications of these
techniques to operating boilers in the field
The results of the initial phases of this program have been reported
elsewhere [4,5] and the third phase has recently been completed [V] •
A decision to proceed with the field tests was made even though the
results of the experimental investigations were not encouraging. The
results of the laboratory investigation aimed at optimizing control
techniques have been summarized by Muzio, Wilson and McComis \5~\.
- a 20 percent reduction in NO emission could be achieved
X
either by reducing the amount of primary air or by
increasing the atomizing air pressure;
11-112
-------
maximum reductions of 25 percent could be achieved by
deliberately staging the heat release before smoke
emissions became excessive;
a 30 percent reduction in emission level could be
obtained by adding cooled flue gases to either the total
or the primary air streams;
- it was possible to achieve a 45 percent reduction in
NO emissions by combined application of staged com-
X
bustion and flue gas recirculation.
This paper briefly describes the results of the field test
and recent laboratory investigations which have been conducted both
to explain earlier results and to provide information on the optimum
design of combustion systems for package boilers. The optimum design
must satisfy the dual criteria of minimum pollutant emissions and
maximum efficiency where efficiency relates to both thermal and
operating efficiency.
2.0 EMISSION CONTROL FROM PACKAGE BOILERS - FIELD TESTS
The influence of flue gas recirculation on NO emissions has
X
been determined in two package boilers of different design types
(watertube and firetube). The influence of delayed air addition,
in order to stage combustion, has also been established in the same
firetube boiler. This work was supported in part by the original
EPA-API program and the field tests were carried out by the Foster
Wheeler Energy Corporation.
11-113
-------
2.1 Design of the Equipment
The boilers tested during the field demonstration were both
located in the same boiler house and represent a compromise between
that which was considered ideal and that which was practically
attainable. In selecting the field units the following criteria
were considered to be of particular importance:
the units to be tested should be typical of modern
practice. The value of the investigations would be
negated if the data were to be obtained on equipment
of outmoded design
- the units tested should reflect the bulk of the popula-
tion of industrial boilers both with respect to type
and size
- the units should be capable of burning both natural
gas and residual oil
the same oil supply should be burned in both units
- it should be possible to apply both flue gas recircula-
tion and staged combustion to the units
- the owners of the units must be cooperative since the
tests could not be carried out without some interruption
of the normal routine
The aid of the American Boiler Manufacturers Association was
solicited in order to determine the type, size and characteristics
of the "typical" package boiler. The ABMA were most helpful and
II-ll/j
-------
provided data on the total sales of boilers for a ten year period.
This data indicated that the bulk of the population was in the range
3,450 to 6,900 Ibs of steam per hour for the firetube design and
21,000 to 50,000 Ibs of steam per hour for the watertube design.
A decision was made to investigate the influence of boiler
design on pollutant emissions rather than the effect of boiler size
for the same design. Both the units chosen for testing were located
in the same boiler house and therefore one important criteria, that
of common fuels, was satisfied. A complete description of the boilers
and the modifications made to allow the application of flue gas
recirculation has been given elsewhere [&"]• The forced draft fire-
tube boiler burned both natural gas and No. 5 fuel oil and was rated
3
at 12 x 10 Ibs per hour of steam at 15 psig. The watertube boiler
was also designed to fire the same fuels, however, a steam rather
than an air atomized oil nozzle was used in this unit. Although
both boilers were built by the same manufacturer, they were fired by
different burner designs.
The flue gas recirculation systems designed and installed in
each boiler were capable of recirculating approximately 30 percent
of the combustion products at full load. The recirculation was
drawn from the stack through a metering orifice of standard ASME
design and the recirculation was controlled by a damper at the dis-
charge of the recirculation fan. The wind box of both boilers was
breached to provide an entry for the cooled recirculation. Front
II- 115
-------
views of both boilers are presented in Figures 1 and 2 which show the
location of the flue gas recirculation systems.
Fon
Flue Gas.
Ducting "
Slack
Windbox
Figure 1. Front View of Firetube Boiler
Showing the Location of the Flue Gas
Recirculation System
•Stock
Orifice Plots
Combustion
Air Fon
Flue Gas
Recirculation
Fon
Figure 2. Front View of Watertube Boiler Showing
the Location of the Flue Gas Recirculation System
11-116
-------
Staged combustion investigations were only carried out in the
firetube boiler. Figure 3 shows the general arrangement of the
staged combustion system. Air supplied by a separate fan was
injected radially into the firetube through eight fishtail injectors.
The axial location of the injectors (X see Figure 3) could be varied
to determine the optimum position for staged air injection. The
staging system was uncooled, and therefore, the staging air was
heated prior to injection. However, the temperature of the second
stage air upon injection was not measured.
Combustion
Air
Air
Supply
Manifold
Staged Air Supply
Figure 3. Details of Staged Combustion Equipment
in the Firetube Boiler
11-117
-------
Flue gas samples were withdrawn from the stack through a
stainless steel probe and cooled in an ice bath. The composition of
the flue gases were monitored continuously and the NO results
2t
reported below were obtained by an analyzer based upon the chemi-
luminscent principle.
2.2 Results
Flue Gas Recirculation
Figure 4 shows the influence of flue gas recirculation on
NO emissions from both boilers at fixed excess air and three loads.
x
Reductions of the order of 30 percent were possible in the firetube
boiler and the limit was imposed by the capacity of the recircula-
tion system. Only modest reductions could be attained by recycling
cooled combustion products through the windbox of the watertube
boiler. However, in this case the limit was imposed by flame sta-
bility and not the capacity of the recirculation system. Figure 4
also shows that for all watertube boiler loads recirculation of
combustion products initially caused an increased emission. The
addition of flue gases did not significantly affect either smoke or
carbon monoxide emissions in either boiler.
11-118
-------
FIRETUBE
86%
a
a
i
WATERTUBE
LOAD 65%
10 20
% FLUE GAS RE CIRCULATION
30
Figure 4. The Influence of Flue Gas Recirculation and Load
on NOX Emissions from a Watertube and a Firetube Boiler
at 20 Percent Excess Air
11-119
-------
Staged Combustion
The insensitivity of NO emissions to the overall excess air
during staging is shown in Figure 5. However, as expected, NO
JV
emissions are strongly dependent upon burner stoichiometry (i.e., the
fraction of combustion air passing through the windbox). The influ-
ence of the location of the staged air injectors on NO emissions
Jt
can also be seen in Figure 5. The fractional reduction increases as
the injection point moves downstream. In the system tested, there
appears to be an optimum burner stoichiometry for minimum NO emis-
X
sions. However, this minimum optimum level is probably peculiar to
this system and is probably associated due to the presence of
unreacted fuel nitrogen compounds at the point of staged air
injection. In the laboratory experiments increased smoke emissions
were observed as a direct tradeoff for decreased NO emissions,
X
whereas in the field tests excessive smoke emissions were only
observed at low excess air levels and carbon monoxide emissions give
a better indication of deteriorating combustion conditions within
the firetube. Figure 6 compares the influence of the location of
the staging injectors on both NO and CO emissions for a fixed
X
burner stoichiometry. It can be seen that the initial NO reductions
were not accompanied by a large increase in the CO emission. How-
ever, with the injectors located at the furthest distance from the
burner the CO emission increases significantly, suggesting that the
11-120
-------
180
140
120
5?
O
E
CL
o.
X
O
180
140
120
l.25
= l.50
= 1.75
0%
120
100
80
% BURNER STOlCHlOMETRtC AIR
Figure 5. The Influence of Overall Excess Air and
Location of Staged Air Injection During Staged
Combustion Tests on the Firetube Boiler
(60 Percent Load)
It-121
-------
140
M
O
300
« 200
o
E
100
ID
2D 3D
LOCATION OF AIR INJECTORS
Figure 6. The Influence of the Location of Staged Air
Addition on NOX and CO Formation with 87 Percent of
the Stoichiometric Air at the Burner and 4 Percent
Overall Excess Oxygen
II- 122
-------
optimum location for the addition of staged air in this system is
between 1.75 and 2.5 furnace diameters from the oil nozzle.
2.3 Implications 'of the giej-A Test. Results
Only a selection of the results obtained during the investi-
gations in the two package boilers have been presented in this paper.
Baseline tests were carried out to establish the operating character-
istics of the boilers and the emission levels appear to be typical of
other units of similar size. The influence of the various NO con-
x
trol techniques has been established for both natural gas and fuel
oil. Of considerably more importance than the specific results are
the implications of the tests with respect to emission control from
this class of equipment.
Pollutant emission control techniques for package boilers
must be considered from two points of view. Control of emissions
from those boilers already installed, i.e., retrofit and the control
of emissions from new equipment which either requires some minor
design modification or a radical change in boiler /burner design.
Retrofit would be expected to have the most Immediate effect upon
total pollutant emissions from package boilers. One of the major
conclusions that can be drawn from the field tests carried out in
this investigation is that many problems can occur if retrofit becomes
mandatory, which could render NO control expensive and/or ineffective
.X
in certain cases unless further work is carried out to optimize the
total combustion system for low NO and high efficiency.
X
11-12:
-------
The boilers chosen for testing were done so with a knowledge
of the requirements of the control system to be installed and yet
problems were uncovered that could not have been anticipated.
Changes in the geometry of the firetube boiler windbox to accommodate
the flue gas recirculation inlet resulted in the initiation of
severe pulsations at several boiler loads. Successive modifications
succeeded in alleviating the problem but not in eliminating it.
Flame instability occurred in the watertube boiler with the addition
of greater than 30 percent recirculation at all boiler loads. With-
out prior direct experience, neither of these practical problems
could be predicted. There is no reason to believe that these two
boilers represent special cases and it must be expected that similar
problems would occur with other units. The age of equipment and
the lack of available space in the vicinity of most package boilers
will tend to extend the problems of retrofit, particularly with
respect to flue gas recirculation even if it were to be shown to be
an effective control technique.
Based upon the results of the laboratory investigations, flue
gas recirculation was as effective as could be expected In reducing
NO emissions from the firetube boiler. However, the recirculation
x
of flue gases cannot be considered as a cost-effective control
technique for watertube boilers if the results of these tests are
typical. Comparison with the data of Cato el al Q?^ suggested that
baseline emissions from this boiler are perhaps a little lower than
II-124
-------
could perhaps be expected, but that the emission characteristics do
not differ from other boilers of this size. We believe that the minor
influence of flue gas recirculation on emissions is attributable to
the fact that the total NOX emission at intermediate boiler loads is
produced by the oxidation of fuel nitrogen compounds. The generous
furnace volume and low combustion intensity minimize thermal NO
formation. Consequently, reductions in flame temperature will not
significantly influence emissions. The slight increase in emission at
low recirculation is probably caused by improved air/fuel mixing due
to the increased burner pressure drop. It may well be that minor
changes in burner design would result in an improvement in the effi-
ciency of flue gas recirculation.
3.0 LABORATORY INVESTIGATIONS
The overriding objective of the continuing laboratory investi-
gations is to establish the optimum combustion conditions necessary to
ensure maximum thermal efficiency with minimum pollutant emissions
(EPA Contracts 68-02-1498 and 68-02-1500). Having established these
conditions, a combustion system will be designed, constructed and opera-
ted to demonstrate that they are compatible with the successful opera-
tion of a commercial system. It is not the intent to write a "text" on
burner design or to manufacture a commercial burner, but to provide
proof of the practical viability of the concepts of low emission, high
efficiency oil burners for package boilers.
A burner can be described simply as a piece of equipment which
allows safe conversion of chemical energy contained in a fuel into
1 I-125
-------
usable thermal energy. In a package boiler this requires as a
minimum the creation of conditions which ensure flame stability and
complete combustion without accelerating the mechanical deteriora-
tion of any of the boiler/burner components. These necessary con-
ditions can be created by suitable hardware design to produce the
optimum fuel-air-combustion product mixing patterns within the
combustor. The practical choices available to the burner designer
are the geometry of the burner, the method of fuel delivery and the
condition of the combustion air at the burner throat. Prior to the
necessity for NO control the major problem of ensuring complete
J\
combustion was solved by efficient fuel air mixing based upon the
combustion engineers dictum that mixed is burned. However, for fuels
containing bound nitrogen, rapid fuel air mixing maximizes nitric
oxide production. The objective of the laboratory investigations is
to determine those design options which will provide a controlled
rate of fuel air mixing which will minimize NO production and yet
ensure complete combustion at low overall excess air levels.
In the investigations in the field, fuel rich combustion was
achieved by dividing the total air supply and injecting some part
of that air at a physical location remote from the burner. In the
field this exercise was more successful than the original laboratory
investigations. The major problem in the laboratory was that smoke
emissions increased in almost direct correspondence to the reduction
of NO. Smoke emission from practical flames is the net result of
excessive carbon formation and poor carbon burnout. Staging the
11-126
-------
combustion process will undoubtedly cause an Increase in the mass
of soot produced, and therefore, it is essential that optimum con-
ditions for carbon burnout are provided after injection of the second
stage air. Lee, Thring and Beer
showed that in the temperature
range 1300-1700 K the rate of carbon combustion in gm/cm of soot
aurface can be represented by
q = 1.085 x
4 pn I"1/2 exp (-39300/RT)
0,
(1)
where
0^ is the partial pressure of oxygen (atm)
T is the temperature (°K)
R is the universal gas constant (cal/mole K)
Equation (1) illustrates the importance of both oxygen concentra-
tion and temperature on the rate of soot burnout. Consequently,
even without increased soot formation, soot emissions will increase
with staged combustion if the second stage air is mixed with the
carbon when temperatures are too low or the addition of the staged
air chills the flame products.
In the laboratory investigations it was found that NO emis-
sions were relatively insensitive to the level of staging when the
second stage air was injected from the combustor axis [_5J. If the
second stage air was added either from the axis or from the com-
bustor walls too close to the burner NO emissions increased. Species
concentration maps have been obtained for certain burner conditions
H-127
-------
in an attempt to explain the performance of the modified commercial
burner used in the Phase II laboratory investigations. The samples
were withdrawn from the flame with a cooled stainless steel probe.
Figure 7 shows a schematic layout of the combustor illustrating the
relative position of the fuel injector, the air streams and the
sampled field, and Table 1 lists the burner conditions for the tests
which will be discussed in this paper and CL, CO and NO concentra-
tion contours are shown in Figures 8, 9 and 10. The burner used in
these investigations has been described in detail by Muzio ^5^. It
was a commercial burner which was modified slightly to provide
independent control of the air distribution in the primary and
secondary streams. The original combustion air damper was retained
to vary the swirl intensity of the secondary stream.
The "baseline" conditions for the laboratory investigations
correspond to Test 3, the combustion air is split equally between
the primary and secondary flows. Figure 11 compares 0 , CO and NO
concentration profiles at the exit of the burner divergent (x/D =
0.45) for the unstaged baseline case and a staged flame, Test 6. In
Test 6, 20 percent of the total air flow was added downstream of the
burner through sidewall injectors while maintaining the 50:50 split
between the primary and secondary air flows at the burner. The most
remarkable feature of the unstaged profiles is the high axial oxygen
concentration (almost pure air) so remote from the burner throat.
This suggests that the ignition front is situated well downstream of
H-128
-------
S 2
S u
<75 'Jf
o -
O -o
"Z:
o
-O "3
«w
O
c
o
rt
o
o
(J
00 O
o n
en G
M
O PQ
4J
tfl U
3 X
i"
o *o
Hi
O M
4J O
rt *J
•I-)
c
M-l M
O -rt
a *a
O 0)
>s 00
tG rt
§
0 -H
C/3 CO
2
II- 129
-------
3
pa
o
u
•a
0)
-H
£
o
•H
O
u
(1)
.-1
•a
H
>.
i^
•H
<
03 >^
01 *^
(1)
O
3
4-1 M
3 J3
O '•••1^
&^
M 4J
«
4J
B)v£)
0) O
P3 ^H
So
cu ^
H ^
X
a
00
o
00
CS CM
. ,
^H *~^
n
r- tx r>.
rH i— 1 f- ^
-*-*-*
• . •
rn ro c^
CO -C vO
J3
a
00
o
o
1— 1
m
00
OO
CN
i
CN
CN
o
m
o
in
r»
>— i
-a-
•
^V"S
* )
00
11-130
-------
IU
O)
CC Ul
3 I-
I
8
z
UJ
X
o
UJ
/^V"
00
CO
CO
o:
^
UJ
Q
2
3
9
V)
z
UJ
S
Q
03
e
o
o
c
ID
S^'
&*'
O
3
Efl
•H
4-1
in
(U
•H
O
4-1
CO
rt
00
O)
h
a
•H
nvixv
II- 131
-------
fO
H
W
UJ
VO.J
oo
*
I-
«/>
111
to
cc
I
o
u
UJ
00
Ul
O CC
UJ
u
Q
<
00
CO
UJ
00
Ul
O
00
to
(VJ
3DNV1SIQ 1VIXV
o
J2
l-l
cd
u
t)
-------
ro
CO
5f
h-
UJ
a:
z
o
u
o
z
8 —
o
CO
UJ
vt
UJ
CO
en
to
UJ
01
2
UJ
2
0)
•o
•H
a
•H
-H
2
•a
cu
M
3
to
a
a>
o
at
$-1
00
(M
3DNV1SIO 1VIXV
TI-133
L
-------
O
It
O
"*
+J
a)
i
c
ct)
O
co
CS
o
in
O
g
•H
4-1
43
•H
M
4J
CO
•H
Q
iH
.2
T3
j
oo
•H
• NOIiDVMJ
H-134
-------
the fuel nozzle and that the primary air hardly participates in the
combustion process to that point. Muzio et al [_5~^ reported that
increasing the swirl intensity by changing the position of the
secondary air damper had almost no influence on pollutant emissions.
Species measurements indicated that the mixing process was almost
unaffected by the position of this damper, e.g., the axial oxygen
concentration was independent of "swirl level". The primary effect
of reducing the combustion air flow through the burner throat in
Test 6 appears to be to slightly increase the carbon monoxide con-
centrations and to reduce the axial oxygen concentration. Reduced
throat velocities may well allow the ignition front to move upstream
closer to the axis, thus consuming more of the primary air.
Figures 11 and 12 emphasize the fact, that in the initial
stages of heat release, the major difference between the staged and
unstaged flames is the participation of the primary air stream in
the combustion process. Figure 12 plots the radial distribution of
three derived properties: N, the degree of oxidation, M the
stoichiometric mixing factor, and K the mass of NO produced per mass
of carbon. These properties are defined as:
mass of reacted oxygen
N
M =
mass of reacted oxygen plus mass of oxygen
necessary to complete reaction
/mass of oxygen\
\ mass of fuel / flame point
/mass of oxygen\
\ mass of fuel
stoichiometric
II-135
-------
K> (0
o
n
o
0)
0)
-------
K =
mass of NO
mass of fuel reacted (CO + CO )
A more detailed account of the calculation of these quantities from
flame data has been given by Hemsath [^9j|. It should be noted that
these derived quantities relate to time average values. When N = 1
combustion is complete and values less than 1 give an indication of
the completeness of reaction at that point. In a perfectly mixed
situation (assuming mixing controlled heat release) N would be equal
to 1 whenever M = 1. M = 1 indicates that the fuel and oxygen are
present in stoichiometric proportions.
Over a large proprotion of the radial field at x/D - 0.71 the
degree of oxidation is similar for the staged and unstaged flames;
however, the bulk of the unreacted fuel shown by the carbon monoxide
concentration contours presented in Figure 9 lies off the axis and in
this region the degree of fuel air mixing, as shown by M, is almost
the same for the staged and unstaged flame. This is only true if
the unmixedness of the two conditions is similar, M = 1 could occur
and the unmixedness be such that at some instant in time there could
either be all fuel or no fuel, and the two would never be mixed on
a molecular scale.
Test 6 with 20 percent staging showed a reduction in flue gas
NO emission of only 18 percent compared to the baseline level. At
the exit of the divergent and at x/D = 0.71, actual NO concentrations
are higher for the staged case than the unstaged case (see Figures 11
11-137
-------
and 13) but the mass of NO per mass of fuel reacted is very similar
(see Figure 12).
It is contended that the method of second stage air injection
should satisfy the following criteria:
— reaction of fuel nitrogen intermediates should be
complete, otherwise the addition of air will allow
the formation of fuel NO;
— rapid mixing of the staged air and the primary
products should take place to give the maximum
opportunity for carbon burnout.
Second stage air was injected at x/D = 2.2 in Test 6 by which time
reaction was almost 80 percent complete and the unreacted fuel was
concentrated toward the axis of the combustor (see Figure 14). The
sidewall injectors were constructed to promote swirling second stage
air flow £5^. This construction directed the air jet away from the
axis of the combustor. Thus the air jets never penetrated to the
combustor axis where the bulk of the unburned fuel lay. Mixing
between the second stage air and the partially combusted products
of the primary zone was poor. Axial oxygen concentrations shown
in Figure 15 for two radial locations graphically illustrates this
point. Because combustion was almost 80 percent complete by
x/D = 2.2, it was concluded that improved mixing of the staged air
would have no influence upon NO emissions but that smoke emissions
might well be reduced. Consequently, minor modifications were made
to the sidewall injectors to test this hypothesis.
H-138
-------
X
# I
O D
CO
GO
QC
* 5
o
H
p
K
4J
«
Q
TJ
CB
s
t 9
in
O
0 ^
§
K
^
-------
I-
U)
Ui
CO
tr
8
I
in
o
<
o
<
cc
in
in
o
tn
z
UJ
5
CNJ
O
•H
•H 00
X n)
0) T3
o
N NOIJLVQIXO dO 33U93Q
11-140
-------
20
15
fM
O
ui
s
O
10
Burner Air
% Stoich.
t.17
0.93
Rodiol Position
r/R=0
O
•
r/R=0.7l
D
•
I 2
DIMENSIONLESS AXIAL DISTANCE
Figure 15. Axial Distribution of 0 Concentration at r/R = 0
and 0.71 for Test 3 and Test 6
H-141
-------
New sidewall injectors were constructed to inject the staged
air directly toward the combustor axis and Table 2 compares emissions
with the old and new injector designs for various staging levels and
injection locations. With only 20 percent of the total air added
through the staging injectors, the design of the injector had no
effect upon the NO emission. However, for the two cases where the
injection took place close to the burner, smoke emissions were
reduced. When 40 percent of the air was staged, then both the smoke
and NO emissions were influenced by the injector design. With the
new injector design NO emissions were higher when 20 of the 40 per-
cent staging air was added at x/D - 2.2.
After consideration of the results of these laboratory investi-
gations the staging system for the field tests was designed with
radial staged air injection. The number of injection points was
also increased from four to eight in order to promote mixing. An
attempt was also made to prevent flame products from impinging upon
the firetube walls, which would cause local chilling.
Several measures were taken to try to influence the mixing
conditions in the new burner field in order to improve staging per-
formance using the sidewall injectors. These measures included
varying the atomizing fluid from air to steam, varying the pressure
of the atomizing fluid , air distribution across the throat, nozzle
size and swirl distribution at the burner throat. The objective
at this time was merely to demonstrate that staging could be more
II- 142
-------
tn
a
o
•rl
(0
CO
•H
I
-u
G
n)
4J
O
PH
C
O
CJ
01
M
3
60
cfl
4-1
O
4-1
HI
S
14-1
o
OJ
^=
H
CM
cu
rH
•s
H
H
CO
CJ
c/1
A
O
CO
)_l
(0
o
CO
rt
•
0
_£!
o
$
T3
0)
V4
3
CO
CO
QJ
a
CO
B
CX
p.
O
!S
(D
T3 03
•O >~i
•* 2
M U
•rl Q)
T-t M
CO
4-* i — 1
O iH
H etJ
IH (U
O *O
•H
QJ C/3
60
4J 00
C 3
QJ O
U »-l
41 H
00
•rl
U)
(U
O
I
"Z.
c
00
•H
41
O
CM CM
^•*- ^"*.
f-l r-l
C1") CO 00 ^4D U">
s^* l/^ OQ QQ f"^
"O ;
00
0)
Q
!n
S5
00
•rH
M
CJ
O
r-t
0
CM
ITI
II
^1 o
VD
CM
II
*/ 1 /-i
?S | f— 1
CM
CM
11
*|Q
CM O O CM OS
i— I O vO 'A 00
CM CM iH i-l i-l
CM i-l i-» iH i-H
1 1001
CM CM
1 O 1 0 0
CM CM CM
0 1 1 1 O
CM CM
3
4-1
rt
3
Cu
C
M
cfl
0)
V-l
•H
-------
effective than had been demonstrated earlier and to try to approach
the results reported by Siegmund and Turner £lO^. Figures 16 and 17
illustrate some of the efforts made to minimize both NO emissions
and smoke. For a given atomizing air pressure the minimum NO emis-
sions appeared to occur between 20 and 40 percent primary air.
Staging was most effective at low primary air flow rates. The best
results (i.e, low NO, low smoke) were achieved by fitting a 15 vane
swirler to the oil nozzle to rotate the primary air in the opposite
sense to that imposed on the secondary air by the damper. 45 percent
reductions in NO emissions from the "baseline" conditions with a
smoke increase of 1.5 on the Bacharach scale to 5.5. Even so, this
cannot be considered as satisfactory performance.
The emission performance of the modified commercial burner
was found to be very sensitive to atomization parameters:
• Nozzle size - Figures 18 and 19 illustrate that
although the trends were similar, smoke and NO
X
emissions as a function of primary air percentage
are dependent upon the capacity of the oil nozzle
at a given firing rate.
• Atomization fluid - Figure 20 illustrates the effect
of changing from air to steam as the atomizing fluid
on NO and smoke.
x
11-144
-------
300r
260
•7, 220
o
s«
10
E
o.
Q.
X
O
2 180
140
*
i
Range high to low
secondary swirl
Burner
Modified
Modified
plus
-15°
vanes
% Staged
0
20
0
20
35
•
a
•
o
A
20
40
% PRIMARY AIR
60
80
Figure 16. The Influence of Primary Air Percentage and
Swirl Level on NO Emissions (17 Percent
Excess Air 3.4 x 104 Btu/hr Heat Input)
II-145
-------
I0r
8
o
LU
*C
§ 6
10
CD
Burner
Modified
Modified
plus
-15°
vanes
% Staged
0
20
0
20
35
•
D
•
O
A
20
40
% PRIMARY AIR
60
80
Figure 17. The Influence of Primary Air Percentage and
Swirl Level on Smoke Emissions (17 Percent
Excess Air 3.4 x 10^ Btu/hr Heat Input)
11-146
-------
400
360
32O
o1
10
ex
0.
X
o
280
240
200
Nozzle Size
gph
60 A
80 D
too o
20
40 60
7o PRIMARY AIR
80
Figure 18. The Influence of Primary Air Flow and Nozzle Size
on NOX Emissions from Fuel Oil Flames
(3.4 x 106 Btu/hr 17 Percent Excess Air)
11-H7
-------
Nozzle Size
gph
60 A
80 D
100 O
40
% PRIMARY AIR
60
60
Figure 19. The Influence of Primary Air Flow and Nozzle Size
on Smoke Emissions from Fuel Oil Flames
(3.4 x 106 Btu/hr 17 Percent Excess Air)
II- 148
-------
400r
o
1=
a
x
O
360
320
280
240
200
10
8
d
UJ
to
"I
CO
10 20 30 40
% EXCESS AIR
50
Figure 20. NO and Smoke Emissions as a Function of Excess
Air Level and Atomization Medium
II-149
-------
• Pressure of the atomizing fluid - it had previously been
reported £§] that increasing the air pressure and there-
for the mass flow of atomizing air decreased emissions.
The results presented in Figure 21 show that this trend
holds for larger nozzle sizes and for steam atomization.
The reasons for the observed emission characteristics for steam are
not known. However, species concentration maps give an indication
why increased atoraization pressure and nozzle size affect the
pollutant emissions.
Comparing the iso-concentration contours shown in Figures 8, 9,
and 10 for Tests 3, 4, and 8, it is apparent from the oxygen contours
that the fuel air mixing process is strongly dependent upon the oil
nozzle parameters. Increasing the pressure from 20 to 28 psig produces
more fuel rich conditions in the initial stages (see the CO contours).
Changing the size of the oil tip has a much more radical influence on
the concentration fields close to the burner nozzle (see Figures 21
and 22).
The investigations with the modified commercial burner served
to emphasize the strong dependence of pollutant formation in practical
flames on those parameters which control fuel air mixing processes.
The burners used in the laboratory investigation and the field test
were of different design and therefore it should not be expected
that optimum staging conditions defined in the laboratory would also
apply to the field boilers. Consequently, there is an urgent need
11-150
-------
fvi
o
5?
JO
ex
Q.
X
O
320r
300
280
260
240
220
O
D
Air
Steom
o
z.
L'J
•X.
40
to
CO
O
8 12 16 20 24 23 32
ATOMIZATION PRESSURE PS!G
Figure 21. The Influence of Atomization Fluid and Pressure
Pollutant Emissions from Oil Flames
(17 Percent Excess Air, 3.4 x 106 Btu/hr)
H-151
-------
20 -
UJ
O
O Test 3
D Test 4
A Test 8
Figure 22. The Influence of Atomization Parameters on
Radial Oxygen Distribution at x/D = 0.45
11-152
-------
for experiments to define optimum mixing patterns before devising
the most practical approach to produce them in staged combustion
situations.
As part of the initial stage in the production of this type
of information a versatile test burner has been constructed which
allows the distribution of both axial and tangential velocity across
the throat of the burner to be varied continuously. The air delivery
system consists of four separate plenum chambers which deliver air to
four annualar channels. Details of this test burner are shown in
Figure 23. Rotation is imparted to each of the air streams by vary-
ing the inclination of the air nozzles and rotation in a clockwise
or counterclockwise sense is possible.
It is known that the characteristics of swirling flames are
dependent upon the size and intensity of internal recirculation
zones £l]]. The static pressure field generated by the swirling flow
strongly influences both of these properties. The radial distribu-
tion of static pressure P is given by
s
dr
= P
W
(2)
where
p is density
W is tangential velocity
r is radius
II- 153
-------
Operating Handle
Counterclockwise
(CCW)Mode
- Clockwise (CW! Mode
- Elbow with turning and guide vanes
Exit Channel
(Annulus)
Test
Comtxnlor
Air Inlet (Typ.)
Channel 4
Gland for Fuel
Injector
Channel t
Channel 2
Figure 23. Multichannel Continuously-Variable-Swirl Test Burner
II- 154
-------
Consequently, the distribution of static pressure can be varied by
controlling the radial variation of tangential velocity at the
burner throat. The multichannel test burner allows the distribution
of tangential velocity to be varied. With rotation on the inner of
the four channels the tangential velocity distribution approaches
that of a free vortex, i.e., W = C/r. Conversely, distribution
given by W = c'r, solid body rotation can be produced by applying
rotation to the outer channels.
Figures 24 and 25 present initial results from this investi-
gation. NO emissions are plotted against the dimensionless swirl
X
parameter S for two different size steam atomized nozzles. S is
* P P
given by
S = 2 S
p n
where S is the swirl number associated with the flow in each channel
n
G r.
xl 1
where
G~ is the axial flux of angular momentum
G is the axial flux of linear momentum
r is the outer radius of that channel
and subscription 1 refers to channel 1.
The results presented in Figure 24 and 25 were obtained with
a refractory parallel burner exit. S is independent of air
11-155
-------
r
700 r
%A1R 01 ST.
1
50
40
2
3,4
30
50
30
ROTATION
1
20
X
X
Z
3
X
40
4
X
X
X
40
X
o
A
O
O
V
a
200
0 05 1.0 1.5 ZO 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0
SWIRL PARAMETER
Figure 24. The Influence of Tangential Velocity Distribution on
NO Emissions from Oil Flames (3 x 106 Btu/hr 60 gph Nozzle)
II-156
-------
ROTATION
I-'
in
o
K
<
i«
^r
IO
PJ
-
0*
X X
X X
o
•H M
0 43
OJ 4-1
> «
angential
(3 x 106
H tfi
0)
O tw
r- 1
OJ PM
o
C •-!
0) -H
3 O
i-H
^ €
C 0
M (-1
M-l
0)
j: w
H a
o
•H
• (0
IT) Ifl
CM -H
3 O
00 Z
U-157
-------
distribution between the channels, however, the results suggest that
axial velocity distribution also influences emissions. In general,
both nozzles exhibit similar trends. Emissions are maximized by
free vortex type distributions and minimized by solid body type
tangential velocity distributions.
4.0 CONCLUSIONS
Studies to optimize burner/combustor conditions to allow the
control of nitric oxide formation in fuel oil flames are continuing.
It has been shown to date that
1. Oil atomization parameters not only have a strong
influence upon uncontrolled emissions, they also
influence the effectiveness of staged combustion as
a control technique.
2. Pollutant formation in confined oil flames is strongly
dependent upon the distribution of axial and tangential
velocity at the burner throat.
3. Field tests have demonstrated that 50 percent reduc-
tions in NO emission from firetube boilers are
x
possible without incurring serious penalties due to
emissions of other pollutants.
4. Thirty percent reduction in NO emissions from firetube
j£
boilers firing oil containing nitrogen can be obtained by
flue gas recirculation. Reductions in watertube boilers
will be strongly dependent upon combustion intensity.
H-158
-------
REFERENCES
1. Heap, M. P., Lowes, T. M. and Martin, G. B, "The Optimization of
Aerodynamic Design Variables to Control the Formation of Nitric
Oxide in Fossil Fuel Flames." Fluid Mechanics of Combustion.
p 75, The American Society of Mechanical Engineers (1974)
2. Pershing, D. W., Brown, J. W., Martin, G. B. and Berkau, E. C.
"Influence of Design Variables on the Production of Thermal and
Fuel NOX from Residual Oil and Coal Combustion", paper presented
at the Annual A.I.Ch.E., Meeting Philadelphia, 1973.
3. Pershing, D. W., Brown, J. W., and Berkau, E.E., "Relationship
of Burner Design to the Control of NOX Emissions Through
Combustion Modification." Proceedings, Coal Combustion Seminar,
EPA-650/2-73-021, NTIS PB224-210/AS, pp 87-140 (September 1973)
4. Muzio, L. J. and Wilson, R. P. Jr., "Experimental Combustor for
Development of Package Boiler Emission Control Techniques,"
Phase I of III, EPA Report R2-73-292a, 1973.
5. Muzio, L. J., Wilson, R. P. Jr., and McComis, C., "Package
Boiler Flame Modifications for Reducing Nitric Oxide Emissions,"
Phase II of III, EPA-R2-292b.
6. Heap, M. P., Tyson, T. J., McComis C., McMillan, R. E.,
Zoldak, F. D. and Sommerlad, R. E., "Control of Package Boiler
Emissions by Staged Combustion and Flue Gas Recirculation,"
Phase III of III, Final Report EPA API Contract 68-02-0222.
7. Cato, G. A. and Robinson, J. M., "Field Testing Application of
Combustion Modifications to Control Pollutant Emissions from
Industrial Boilers," Phase I, EPA Report 650/2-74-078a,
October 1974,
8. Lee, K. B., Thring, M. W. and Beer, J. M., 1962 Combust. Flame,
6, 137-145.
9. Hemsath, K., "Mixing factors and Degree of Oxydation:
Definitions and Formulae for Computation," I.F.R.F., Doc.nr.
G 00/a/l1, 1965
10. Siegmund, C. W. and Turner, D. W., "NOX Emissions from Industrial
Boilers: Potential Control Methods," ASME Paper 73 1PWR10, 1973.
H-159
-------
10:45 a.m.
The Control of Pollutant Emissions from
Oil Fired Package Boilers
Dr. Michael P. Heap, Ultrasysterns
Q: Could you tell us what the fuel nitrogen concentration
was in the oil for the water tube test?
A: This is another sort of problem at the moment. Some
of the analyses were done with a Coleman analyzer
using the Dumas method, and there is some difficulty
in reproducibility of results. In the past I have
always had problems trying to analyze fuel for nitrogen
using the Dumas method. Using the Kjeldahl method I
always sort of got reproducible answers. I believe
that the 0.1% figure is more or less correct.
Q:
A;
It seems that exhaust gas recirculation would have
some significant thermal efficiency effects. I was
wondering if you had extended your tests at all to
consider this in relationship to the reductions or
changes in pollutants.
This is being done at the moment but I don't have
any information on it. It is part of the extension
of the field test that is just being carried out this
week.
It may even pay to be a low polluter in such cases.
Certainly in the case of the fire tube boiler
firing natural gas. One of the advantages is not only
do you get the NO down to very low levels (about
«V
15 ppmX but it also allows you to fire the boiler
at less than 2% oxygen, which you couldn't do without
the flue gas recirculation. So you could say it would
II- 160
-------
improve the thermal efficiency that way,
Q: What kind of furnace heat release rates were you
talking about in your package burner boiler tests?
Volumetric heat release?
A: The laboratory one was designed that way so that
it would be typical ot a normal fire tube package
boiler. The water tube combustion intensities-^off-
hand I can't say, but we did have difficulty getting
up to 100% load anyway. So the combustion intensi-
ties would be low. Most of the measurements were only
done at less than 70% full load.
11-161
-------
-------
PILOT SCALE INVESTIGATION
OF CATALYTIC COMBUSTION CONCEPTS
FOR INDUSTRIAL AND RESIDENTIAL APPLICATIONS
by
J. P. Kesselring
C. B. Moyer
R. M. Kendall
Aerotherm/Acurex Corporation
Mountain View, California
J. A. Cusumano
Catalytica Associates, Inc.
Palo Alto, California
II- 163
-------
PILOT SCALE INVESTIGATION
OF CATALYTIC COMBUSTION CONCEPTS
FOR RESIDENTIAL AND INDUSTRIAL APPLICATIONS
J. P. Kesselring, C. B. Moyer, and R. M. Kendall
Aerotherm Division/Acurex Corporation
Mountain View, California
J. A. Cusumano
Catalytica Associates, Inc.
Palo Alto, California
The use of catalysts in place of conventional burners for promoting
hydrocarbon oxidation reactions appears to have advantages in the con-
trol of emissions. The operating conditions of these catalytic combus-
tors are limited by the catalyst bed temperature capability, and the
temperature obtained through normal one-stage, low excess air operation
that is necessary for efficiency reasons is outside the current tempera-
ture capability of catalyst systems. It is therefore necessary to con-
sider other system techniques such as bed cooling, exhaust gas recircu-
lation, or staged combustion to hold the bed temperature down. Based
on a review of existing application and research programs involving
catalytic combustion, a research and development program to establish
design criteria for the application of catalytic combustion to low emis-
sion, high efficiency stationary combustion systems has been undertaken.
This program consists of experimental catalyst and combustion concept
11-164
-------
screening, experimental and theoretical scale-up criteria, and develop-
ment of prototype systems.
INTRODUCTION
More than 150 years ago, Sir Humphrey Davy discovered that platinum
wires could promote combustion reactions in flammable mixtures.
The resulting reactions appeared to take place on the surface of the
wires, "without flame" and with the high radiative fluxes associated
with the solid surface emittance rather than the low emittance of the
combustion products (Reference 1). Subsequent study and investigation
followed a number of paths reviewed by Spalding in Reference 2. Of this
work, a substantial fraction which might be termed "fundamental studies
of heterogeneous catalysis" has attempted to identify and quantify speci-
fic heterogeneous reactions and define the mechanisms by which the solid
surface promotes or accelerates oxidation reactions. A second fraction,
termed "computational", has dealt largely with mathematical theories
and analyses attempting to describe the diffusion of reactants to the
promoting surface, the diffusion of products away from it, and the com-
bined heat and mass transfer. A final third fraction can be termed
"reduction to practice", which includes attempts to construct and employ
practical devices exploiting heterogeneous combustion. As is frequently
the case in combustion, work in this area has proceeded with very little
reference to achievements in fundamentals and computation. Experiments
throughout the 1800*s demonstrated that various materials promoted
11-165
-------
heterogeneous combustion reactions to various degrees, and that the num-
ber of materials with the ability to promote reactions increased as sur-
face temperature increased. The earliest successful practial devices
exploited heterogeneous combustion to provide more efficient lamps.
The best example is the mantle lantern or Auer light, still sold in great
numbers as a "camping lantern", in which combustion takes place on the
ash residue of a fiber bag mantle surrounding the fuel/air orifice.
This mantle ash contains oxides, including cerium oxide and thorium
oxide, which in addition to constraining radiation emission to desirable
lines in the visible spectrum, contribute to the catalytic activity of
the mantle (Reference 3).
Concepts close to the original experiments of Davy have been used
as passive ignition devices and as combustible detectors, particularly
in the form of hydrogen leak detectors. Related applications along this
line include catalytic fume abatement devices and automotive catalytic
converters, both depending on precious metals to oxidize low concentra-
tions of combustibles in gas streams.
After extensive practical experimentation, Bone publicized various
high temperature applications of surface combustion, using for the most
part conventional ceramic materials (Reference 4, 5). He built and
tested surface combustion furnaces, boilers, and permeable panels.
This line of investigation has not, however, proven very fruitful due
to the complexity and expense of the required equipment and the limited
II-166
-------
industrial need for high radiative fluxes, although some commercial
equipment of this type is available (Reference 6).
The range of commercial devices has recently led to some speciali-
zation of terminology in a field where formerly several terms were used
interchangeably. The term "radiant burner" refers to burner types em-
ploying a refractory surface situated within or near the flame so as to
provide for radiant heat transfer from the flame area. Heterogeneous
combustion occurs in these devices to some degree, but generally does
not dominate the process and homogeneous combustion constitutes most
of the reaction mechanism. "Surface combustors" employ refractory in
a similar manner but use special provisions of air/fuel distribution to
maximize the amount of heterogeneous combustion occurring. "Catalytic
combustors" employ special additives (usually precious metals) to en-
hance the heterogeneous activity of the surface.
There are several reasons for the recent upsurge of interest in
catalytic combustion. Use of the diffusion flame for combustion appears
to place a limit on the lower level of control that can be achieved
economically for clean fuels (approximately 25 to 30 ppm of NO for small
A
sources), and catalytic combustion appears to be able to reach lower
levels of NO control (perhaps <10 ppm). In addition, catalytic combus-
X
tion is soot-free, and the potential of high combustion efficiency for
the overall system is present if the excess air can be limited in some
way, such as two-stage combustion, flue gas recirculation, or cooling of
the catalyst bed.
II- 167
-------
AVAILABLE CATALYST MATERIALS
The materials needed for a catalytic combustion system are those
associated with the support, the washcoat, and the catalyst itself.
The importance of each of these elements in the system is described
below.
Support Materials
The catalytic support serves two important functions in a catalyst
system:
• It increases surface area of the active metal or metal oxide
by providing a matrix that stabilizes the formation of very
small particles.
• It increases thermal stability of these very small particles,
thus preventing agglomeration and sintering with consequent
loss of active surface.
Most common support materials are A£ 0, and Si02. The application of
these supports in catalytic combustion presents several significant
problems that must be considered:
• The large space velocities used in a catalytic combustion unit
will result in severe pressure drops if the catalyst beds are
packed with conventional pellets like those used in the petro-
leum industry.
11-168
-------
• The high reaction rates of complete combustion usually present
severe pore diffusion limitations for the effective operation
of the system.
• The high temperatures and heat fluxes place special constraints
on the thermal properties of the system.
Over the last decade, a number of new materials and new preparative
techniques have been developed to overcome these problems. These mate-
rials have emphasized the use of monolithic support structures and sur-
face impregnated pellets.
Monolithic supports are composed of small parallel channels of a
variety of shapes and diameters. These structures may be in the form
of honeycombed ceramics extruded in one piece, oxidized aluminum alloys
in rigid cellular configurations, or multilayered ceramic corrugations.
The channels in honeycomb-like structures have tubular diameters of 1
to 3 mm. The overall diameter of the monolithic support may vary from
2 cm to 60 cm, and is limited, in the case of extrusion processes, by
availability and operation of the metal die. Materials of fabrication
are usually low surface area ceramics such as mullite (3A£,jO, • 2Si02)
or cordierite (2MgO • 5SiO • 2A&-0,). The refractory monolith is pro-
duced with macropores (ly - 10u) and may be coated with thin layers of
catalytic materials, 5-20 wt% coatings being common.
The two major advantages of monolith supports for catalytic opera-
tions are the high superficial or geometrical surface area and the low
H-169
-------
pressure drop during operation. Table I lists the currently available
monolith materials, and Figure 1 shows typical extruded monolith struc-
tures.
TABLE I. AVAILABLE MONOLITH SUPPORTS
Manufacturer
American Lava
Corporation
Corning
Glass Works
E. I. DuPont
de Nemours
& Company
General
Refractories
Company
Norton
Company
Product
Thermacomb
AASiMag 614
Thermacomb
AASiMag 776
Thermacomb
AZSiMag 795
Celcor 9475(EX-20)
TORVEX
Versagrid
SPECTRAMIC
honeycomb
Description
Dense 96%
Alpha-alumina
Porous 96%
alpha-alumina
Cordierite
Cordierite
Alumina
Mullite
Cordierite
Mullite
Silicon carbide
RX387
Silicon nitride
RX384
Silicon
oxynitride
RX385
Temperature
Limit (°F)
2800
2192
2192
2200
2732
2462
2550
3000
3000
2800
2800
In addition to the monoliths, highly porous pelleted materials can
also be used as the catalyst support. Pellets come in a wide variety
11-170
-------
fit ••**«•• •«»*•• •*»•••••••••••••••••••*•
••••••••••••••••••••••••••••••••••••••••
•••••••••••••••••••••••••••••••••••••••a
•••••••••••••••••••••••••••••••••••••••a
•••»••«»•• ••••••*•»••••••••*••••••••••••
•••••••• •••••••••• ••••••••••••••••••••
*•••••• »••••••»••••••• •••••••••••••»•••
!••••»••• •!•••••••••••••••••••••••••
)•••••*•••••••••••••••••••••••••••••
!•••••••••••••••**•••••••••••••••
!**•• •*••«•• •!•*»•• «••••*• •*!*••
::"'9g3$&S
I •»•»»•••••••••»•»*•••»•••*•••
••••»•»***••»•»»«««**•• •••••f
••••••••••••••••••••••••••I*
•*•*••»•«»• *•*»•••«**•» ••••»/
••*•*»•»•»•»•••»»»*• •••••
••»•*•••!• •••••»»••»*• •••
•••••*»•»•••*»•• •••••••••pr
Figure 1. Typical extruded monolith structures (courtesy
of General Refractories Company).
II- 171
-------
of materials, shapes, and sizes, but have the disadvantage of a higher
pressure drop through the bed.
Wash Coat Materials
The low surface area of the monolith structure requires the appli-
cation of a thin coat of oxide such as AA-O,. This wash coat, which
strongly adheres to the ceramic or refractory support, provides a uni-
form high surface area while still insuring that the catalytic material
which is subsequently impregnated on the wash coat is close to the main
flow of reactants. The thickness of the wash coat is usually between
10 x 10"6 and 20 x 20~6 m (10-20 microns).
Many different materials can be used as wash coats. ZrC^, for
example, has been used in test runs for the catalytic combustion of
hydrocarbons in automotive exhaust. Variation in the type of wash coat
can have important consequences in the stability and life of the cata-
lyst system.
Catalyst Coatings
Although the field of catalysis has progressed substantially over
the past decade, its theoretical aspects are not yet at the degree of
sophistication which would enable one, a priori, to choose or design
an active oxidation catalyst with a given set of catalytic properties.
However, a vast amount of data exists which permits the development of
correlations useful for choosing promising candidates based on excellent
11-172
-------
catalytic activity and catalyst stability. Over the last decade there
has been a rapid development of the art, science and technology neces-
sary to synthesize, test and manufacture ultrastable oxidation catalysts
for the abatement of automotive emissions. These catalysts operate
under the most stringent conditions, and sometimes at temperatures close
to 2000°F. The expertise developed in this area will be useful in the
development of a stable catalytic combustion system.
Two broad classes of catalytic coating materials are available:
metals and oxides. The metals of catalytic interest are listed in
Table 2. Of these metals, the only ones which have a possibility of
TABLE 2. METALS OF INTEREST FOR CATALYTIC COMBUSTION
Fe
Ru
Os
GROUP
Co
Rh
Ir
VIII* GROUP IB
Ni
Pd
Pt
Cu
Ag
Au
*Enclosed metals are considered noble.
remaining in the metallic state in a high-temperature, oxidizing en-
vironment are the noble metals; the others readily form oxides. Of
the noble metals, a large volume of data and correlations are avail-
able for platinum and palladium because of their use as automotive
oxidation catalysts. They are among the most active catalysts for the
oxidation of a number of fuels, including methane, methanol, and hydrogen.
11-173
-------
The high activity of these metals is related to their ability to acti-
vate H_, 0?, and carbon-hydrogen and oxygen-hydrogen bonds. Palladium
and platinum are readily prepared in a highly dispersed form on a num-
ber of support materials. Because of the high activity of these metals
per unit area of metal surface (specific activity) and the ability to
attain high dispersions, only small amounts are necessary for catalytic
combustion (0.1 - 0.5 wt%). Ruthenium metal is another possible candi-
date for catalytic combustion; however, under oxidizing conditions it
forms a volatile oxide (RuO,), which is rapidly removed from conventional
catalyst supports. One approach to solve this problem has been to anchor
Ru to a support by forming a relatively stable perovskite structure with
certain oxides such as La-O,. Osmium is even more volatile and poison-
ous than ruthenium in an oxidizing environment. Also, as with iridium
and rhodium, it is very costly, and available in limited supply. The
use of the latter two metals, if at all, would be restricted to small
quantities in multimetallic systems. Silver, while quite active at low
temperatures for the activation of 0., melts at low temperatures and
therefore sinters to a significant degree at the high temperatures of
catalytic combustion. However, it could find use as an additive in a
multimetallic catalyst. Gold is very inactive for oxidation, and would
therefore only be considered as a structural or electronic modifier for
multimetallic catalysts.
H-174
-------
Based on the above considerations, those metals which show the
greatest promise for use in a single-active-element catalyst system are
platinum, palladium, and ruthenium (stabilized). Highly active and
stable catalyst systems can be produced through the use of multi-
metallic systems (such as palladium/platinum), as well.
The catalytic properties of metal oxides have been investigated
by a number of researchers, including research in catalytic oxidation.
As for metals, it has been found that the activity for hydrocarbon oxi-
dation parallels the ability of the catalyst to catalyze the homomolecu-
lar exchange of oxygen. It has also been noted that the oxides of
transition containing ions with partially-filled d-orbitals have the
greatest activity. Based on these considerations, the most active sim-
ple oxides are Co^^, MnO,,, NiO, CuO, Co203> ^^2°3' and V2°5* Multi-
component metal oxides of interest include stable spinel structures
such as Co-O,/CuO, Cr 0 /CuO, and La20/Co_0,.
Temperature Capability of Catalyst/Support System
The choice of an optimum catalyst/support combination requires a
consideration of the temperature capability of a given combination.
The low temperature limit is dictated by the so-called "light-off"
temperature. Solutions to this problem are not complex for most cata-
lytic combustion systems, since noble metal catalysts exhibit good
light-off characteristics for most fuels of interest. The major ther-
mal limitation for catalytic combustion is the high temperature of
11-175
-------
operation. The phenomena which lead to catalyst deactivation at high
temperatures are:
• Sintering of the catalytic species
» Changes in catalyst stoichiometry
• Sintering of the wash coat
• Thermal degradation of the ceramic support material
* Mechanical failures
Sintering of the catalytic species will lead to a concomitant loss
in surface area and therefore a loss in catalyst activity per unit mass
of catalyst. The temperature at which sintering occurs is a function
of the catalytic material. In an oversimplified manner this sintering
or crystallite growth is a function of the melting point for metals.
A number of techniques have been developed to minimize or prevent sin-
tering. For example, multimetallic systems can be prepared which are
more stable than the constituent metals. A number of structural promo-
ters are also known.
Stoichiometry changes for oxides can lead to a catalyst with dra-
matically different catalytic properties. Loss of lattice oxygen can
cause an Increase in the energy required to activate 0» and thereby
decrease combustion activity. One method for overcoming this problem
is to dope the lattice with a material which causes a decrease in the
energy necessary to activate oxygen and also stabilizes the stoichiom-
etry of the lattice.
II- 176
-------
A third mechanism of thermal deactivation is wash coat sintering.
At temperatures above 1650°F high surface area r)~ or y-A£_0_ undergoes
a phase change to a-AJLO,, with concomitant sintering. The change may
be from a surface area of 300 m2/g to about 5 m2/g. This sintering
results in pore closure and a "burying" of active catalytic sites in
the AA^Oq- To minimize this problem, it is possible to work with cata-
lysts for which the catalytic phase has been deposited on a presintered
AJLO,. Alternatively, other materials which are thermally more resis-
tant may be investigated as wash coats. These might include ZrO? and
At very high temperatures, the ceramic support, be it spheres or
monolithic honeycomb structures, begins to degrade. Typical monolith
fail temperatures have been given in Table 1. Although catalytic com-
bustion units are not expected to operate continuously at the tempera-
tures listed, the possibility of local hot spots occurring due to uneven
air/fuel mixing and local catalyst bed nonuniformities must be con-
sidered.
Thermal expansion and contraction may also occur at high tempera-
tures. Both pellet and monolithic supports have thermal expansion coef-
ficients different than their mounting hardware. Considerable catalyst
mechanical attrition can occur if the catalyst becomes loose, thus caus-
ing a decrease in catalytic performance.
II- 177
-------
EXISTING APPLICATIONS OF CATALYTIC OXIDATION CONCEPTS
The existing applications of catalytic oxidation concepts in indus-
try are largely in the low- temperature area at this time. The primary
applications of catalytic combustion are in nitric acid plant tail gas
cleanup, industrial odor control, small catalytic heaters, and automotive
oxidation catalyst systems.
Nitric Acid Plant Tail Gas Cleanup
Nitric acid is produced commercially by reacting ammonia with air
to produce nitrogen oxides, which are then absorbed in water to yield
nitric acid. Large amounts of unabsorbed nitrogen oxides in the tail
gas from the absorption tower pose a pollution problem, which has been
solved in some cases by the catalytic processing of this tail gas over
platinum-group metals to yield useful energy in the form of steam and/or
power. The basic reactions for the catalytic treatment of nitric acid
tail gas over platinum, palladium, or rhodium for methane reducing fuel
are:
CH, + 4N02 -* 4NO + C02 +
CH, +
20 (decolorization)
+ 2H20 (combustion)
• CH, + 4NO •* C02 + 2H20 + 2N2 (abatement)
The tail gas reactor outlet temperature is usually between 1250°F and
1380°F. Support materials for the catalyst have included nichrome wire,
alumina pellets, and alumina honeycomb materials.
11-178
-------
Industrial Odor Control
The elimination of organic fumes is desirable from both air pollu-
tion and fire safety points of view. The platinum metals have been used
for several years as catalysts for the oxidation of carbon monoxide and
a wide range of organic molecules in the presence of air or oxygen.
Most odors are caused by the emission of low concentrations of organic
molecules to the air, with these organic molecules always containing
carbon and hydrogen and generally sulphur, nitrogen, and oxygen as well.
Since most industrial odor problems are caused by organic compound con-
centrations well below the level required for spontaneous combustion in
air, it is necessary to raise the contaminated air stream temperature to
a level at which combustion can occur. The use of catalysts lowers the
temperature needed to achieve odor removal, and also lowers the neces-
sary residence time at the combustion temperature.
Low-Temperature Catalytic Heaters
Low-temperature catalytic heaters are characterized by the small
portable radiant heaters that operate at temperatures below 800°F.
Operating characteristics include heat release rates of approximately
75 Btu/in2 and lifetimes of several thousand hours. Gaseous fuel is
distributed in a uniform fashion to all parts of a catalytic pad, and
combustion takes place on the surface of the fibers of the pad.
The Control Systems Laboratory of the EPA has conducted an extensive
emissions testing program on a number of catalytic heaters, as reported
11-179
-------
in Reference 7. No correlation between emissions and specific heat
release rates was found, and CO and UHC levels were generally quite
high. However, NO levels were quite low for nearly all heaters tested.
The high CO and UHC levels were attributed primarily to poor fuel-air
mixing.
Automotive Exhaust Catalysts
Since platinum metal catalysts were selected for emission control
on conventional automotive internal combustion engines for the 1975
model year, a great deal of development work went into these emission
control systems. In general, NO emissions are controlled by a reduc-
X
tion mechanism, and CO and UHC emissions are removed by catalytic oxida-
tion with secondary air added to the exhaust stream to ensure complete
oxidation.
A reactor for catalytic exhaust emission control has three essen-
tial design features:
• A ceramic support on which the catalyst is deposited in a manner
that prevents attrition
* Rapid warm-up of the catalyst unit, and hence rapid light-off
• Nearly complete conversion efficiency
The essential features of a satisfactory CO/UHC oxidation catalyst
for automotive emission control are (Reference 8):
• A light-off temperature in the region of 250°C
11-180
-------
* Conversion efficiencies in excess of 90 percent for both CO and
UHC at space velocities up to 150,000/hr
• High temperature stability of the support and platinic crystal-
lite system at temperatures to 950°C
• Poison resistance to compounds containing lead, phosphorous,
and sulfur.
Platinum-palladium catalysts have been extensively tested in full-scale
vehicle tests (Reference 9). The differences between platinum and pal-
ladium catalyst performances were quite small.
Reference 10 discusses possible substitute catalysts for platinum
in automobile emission control. In this study, it was concluded that
base metal catalysts generally are not as effective as platinum/palladium
catalysts for automobile emission control, especially for hydrocarbon
and carbon monoxide oxidation. For NO reduction, certain base metal
A
catalysts initially are more active and selective than are noble metal
catalysts; however, most suffer from rapid deactivation.
RESEARCH PROGRAM
Under EPA funding, an 18-month research program has been undertaken
which emphasizes the understanding of the mechanisms of catalyst per-
formance in the catalytic combustion process. The program includes cata-
lyst screening tests on five different fuels, where such performance
factors as light-off, conversion efficiency, extinction, poisoning, and
11-181
-------
temperature limits will be examined for both single- and multi-active
component catalysts; the integration of promising catalyst systems into
a catalytic combustion system for the purpose of evaluating the concepts
of bed heat removal, exhaust gas recirculation, and two-stage combustion;
and the development of an analytical model based on data obtained in the
screening tests which will be capable of evaluating the performance of
the catalyst system. All of these tasks will be accomplished at both
small-scale (nominally 100,000 Btu/hr heat release rate) and large-
scale (nominally 1,000,000 Btu/hr heat release rate) conditions. Each
of these program elements will now be described in greater detail.
Screening Experiments
It is the intent of the screening experiments to provide an effi-
cient means of evaluating catalyst bed performance, in order to permit
the design of a variety of systems incorporating each catalyst bed.
The evaluation of a catalyst for a combustion system differs in at least
one major respect from more conventional catalyst applications (e.g.,
reforming, exhaust gas clean up, partial oxidation of organic compounds,
etc.). This is the major heat release associated with the process.
Because of the exponential relation between kinetic rates and tempera-
ture, these systems tend to "bootstrap" themselves rather than simply
"cook" at an imposed temperature.
Any system concept will provide for heat removal in some way to
limit maximum surface temperatures. To do this uniformly is not
11-182
-------
generally possible for packed bed or monolith systems. To achieve uni-
formity, varying amounts of inert dilution will be fed with the test
streams into an effectively adiabatic reactor. The adiabatic flame tem-
perature of this stream will approximate the maximum catalyst surface
temperature within the system. The relative uniformity of temperature
achieved during the screening experiments will greatly enhance the
ability to obtain quantitative catalyst activity data from these tests.
The initial screening experiments, all to be run at atmospheric
pressure, will be conducted in the existing Jet Propulsion Laboratory's
catalytic combustion test facility. A photograph of this facility is
shown in Figure 2. This facility has previously been used for research
involving the catalytic partial oxidation of kerosene to produce a
hydrogen-rich fuel for automotive applications. These screening ex-
periments will initially focus on the performance variations of plat-
inum catalysts on several substrates, with varying wash coats, in
order to determine such things as the effect of wash coat deposition
procedure, the effect of thermal deactivation of wash coat and catalyst,
and the effect of chemical stabilization of the wash coat. Wash-coated
supports will be supplied to Aerothenn by substrate manufacturers, and
Aerotherm will prepare and apply the catalyst to the wash-coated support.
The planned catalyst test matrix for the first 16 tests is shown in
Table 3. All these tests will be run with natural gas as the fuel.
II-183
-------
o
«
M-l
m
fl)
a
o
a
,0
u
u
ft
U
cs
01
60
-H
II- 184
-------
X
C/)
CO
§
Cu
H
CO
i-l
PL,
ro
W
r4
OS
I
0)
•H
4J
U
CU
1-J
XI
o
u
CO
o
CJ
r*
CO
CO
4J
CO
rH
n)
ij
CO
CJ
o
a
4-1
CO
cu
H
00
e
•H
4-1
CO
OJ
0)
N
•H
e
•H
4J
a,
o
'O
a
CO
4-1
•H
C
,J
>-) CO
01
4-1 V)
3 3
o -o
CU
Q) rj
C 0
•H S-i
hJ a.
01
4-1
, j
CU V4
4J CU
•H -H
rH TJ
rH M
IS
/«v.
CJ
o
CO
O
CM
o?
^
j
^
"^
CJ
O
4-1
PM
in
I
tH
^t
4J
•H
H
•H
,Q
CO
a.
cfl
O
C
0
•H
4-1
CO
S-i
CO
a,
CU
VH
a.
4-1
co
^
rH
CO
4-1
cO
CJ
^
M-l
•H
^
cu
CU
4J
•H
rH
rH
S
/— \
CJ
o
(f)
o
CM
c^
5 co
cu
i.\-j ^|
O
•
4-1 14-1
CJ CJ
cu •*-<
M-l
M-i CU
01 VH
3
01 T3
4-1 01
cd cj
M O
•H (-1
4-1 O.
CO
01 C
> 0
C -H
M 4J
01
4-1
•H
rH
rH
1
fO
XX
g
Vfc*^
co
o
csl
o?
^3
1
o"
OS
S
4-1
PH
CT*
CU
M
CO
a
g
o
u
\-^
C
o
•rt
4-1
Cfl /~*
> TJ
•H l-l
4J CO
O 13
CO C
01 CO
•0 4J
CO
rH
cd cO
£3
H CO
cu f.
H | ] \
M-l i-l
U 3
x
CJ <^v
O T3
s*-^ CD
l-l
CO CO
0 4J
CM C
<=< iH
<; co
I ^s
*~
o"
M
ci
4_l
PH
O
rH
CU
4J
k|_|
rH
rH
3
y*-^
o ^a
VH/ Qj
^H
CO CU
O 4J
CM C
o? -H
<(J W
1 s>v/
?~
I
4-1
CM
rH
rH
CO
CO
3
*4H
o
a
0
•H
4J
CO
N
•H
rH
•H
XI
CO
4-1
CO
rH
CO
e»
£^
* CJ
•
§
s
•W
PH
CO
rH
4-1
CO
O
U
f.
ca
cO
Q)
rH
^tQ
CO
4-1
CO
CO
M
4-1
tH
3
*4-t
0
4-1
CJ
0)
M-l
MH
H
(V
4J
•H
rH
rH
1
CO
O
CN
o^
^
rH
CO
4-1
Cfl
CJ
M-J
O
4J
U
CM Cfl
O 0)
Vn M
CJ H
o"
Pi
4-1
CM
so
rH
a.
J
to
-------
Additional screening tests will be conducted with four other fuels:
#2 distillate oil, methanol, hydrogen, and low Btu gas. The remainder of
the screening experiments at JPL will focus on establishing light-off tem-
perature, extinction limits, and poisoning effects for a variety of cata-
lyst. Multi-active-component catalyst systems will be developed based on
the most promising single-active-component catalysts tested and will be
tested with the five different fuels mentioned previously. The evalua-
tion criterion for catalyst performance will be the effect of a test vari-
able on pollutant emissions, e.g., CO, UHC, and NO .
While the initial screening experiments are being conducted at JPL,
Aerotherm will be constructing a catalytic combustion test facility for
the remainder of the experimental work on this contract. A schematic
diagram of this facility is shown in Figure 3. This facility will be
capable of testing at pressures up to 10 atmospheres, and will be cap-
able of doing both small- and large-scale testing. It will initially
be used in the conduct of the high-pressure screening tests and the
catalyst life tests. An overview of all testing to be conducted during
this study is given in Table 4.
System Concept Evaluation Experiments
The focus of the system concept evaluation will be on effectiveness
of various heat extraction techniques and their effect on catalyst per-
formance (kinetic and thermal stability). All of these experiments
will be conducted in the Aerotherm test facility in the versatile
II- 186
-------
O
m
VI
ai
a_-t-> o o. X
E » E j| a. Q. o
>J OOI3E**— to
ail ^ u •
^ x *J >
\. ^ ^- o — tx. •—
1-
(
O
u
a
a
n
c
o
u
u
.
a
u
o
tn
©
X'Lddns
UO
1ujniaa
-®—
II- 187
-------
TABLE 4. OVERVIEW OF CATALYST SCREENING AND SYSTEM
EVALUATION TESTS
Element
Small-scale
catalyst
systems
Small-scale
system
studies
Large-scale
catalyst
systems
Large-scale
system
studies
Test Series
Single-catalyst
screening (1A)
Multi-component
catalyst screen-
ing (IB)
Selected catalyst
system evaluation
(1C)
Concept evaluation
(ID)
Large-scale
catalyst screen-
ing (2 A)
Concept develop-
ment (2B)
Primary Purpose
Preliminary evaluation
of single-catalyst per-
formance
Preliminary evaluation
of multiple-component
catalyst performance
Perform extensive eval-
uation of selected cat-
alyst systems
Integrate promising
catalyst systems into
a practical combustion
system
Extensive evaluation of
selected catalyst per-
formance
Development of optimum
combustion system con-
figurations
Approximate
Number of
Tests
70
13
28
34
35
34
II- 183
-------
catalytic combustion test apparatus shown in Figure 4. In this figure
many possible concepts are shown connected in series, but only for illu-
strative purposes. In actuality, only a few of the segments would be
t
assembled to test a particular concept. The versatile test apparatus
is capable of testing flue gas recirculation, two-stage combustion, and
bed heat removal concepts. Table 4 gives an overview of the system con-
cept tests to be conducted.
Data. _Ana ly s is
In order to establish the data reduction procedures associated with
obtaining catalyst activity, Figure 5 must be considered. In preparing
this idealized representation of wall events it is assumed that heat and
mass transfer coefficients are equal, that diffusion coefficients are
also equal, and that the system is adiabatic, so that enthalpy is con-
stant. It should be emphasized that these assumptions are for the ideal-
ized representation only, and do not necessarily represent what will act-
ually be used in the data analysis. The independent variable selected,
K , represents the wall concentration of the lean reactant. The depen-
w .
dent variables are n^, mR£AC, and TWAU/ mTRAN is the mass flux of
*
reactant to the wall, and n^pAp is the mass flux of reactant consumed by
the catalyst, while T T is the wall temperature and is proportional to
WALLi
the fraction of reactants consumed at the wall. This proportionality is
assumed to be linear in the sketch. Thus, when all reactants are con-
sumed away from the wall the adiabatic flame temperature is reached. A
steady-state solution is reached when the mass flux of reactants to the
II- 189
-------
I
s.
o
VI
c
O)
Cfl
S-
41
L-
41
•o
I
•a
V
at
o
•!-
*->
O — 00
X <->
u •«->
•M 4> V*
I/I (/) 01
J. S_ >,4-> C 10 O •
O "— C O +> ••-
. 0_ 10 01 ••- « +J
> 4-» £ 4-> <_> c
• O> 10 O O 0)
z> c o 01 ai -o **
-^ -r- (/> tfl 0) V
U
•M ai
1. (/)
i- O
Hi a: 4->
oo fi
CLl. 3
111
X'^ (J
UJ -M
U
4-> 01 C
CO I/) -r-
£
aiccu
3 O OI-P-
u- <_> i—<•»
I QC O
L<-> Ol O> -^ _ „
: .p- C C -O O>>— t- C
I r— •*-'*- 41 C to 9> •— ••-
) O i— G*.^ '»- '«- U 4D C
COHSUX-D10C
co-«-)*^(dP'''~
I
'•O r— oo O> o •
-------
o
*J
•o
Ul
rO
n
O
»— i— C
fd IQ *O
C C-U
o ou
f-
2 £§
Q. CXU
O ^«
u
o
M
O.
X
o
u
•H
«8
4-1
-------
wall is equal to the mass-flux of reactants consumed at the wall, and is
• *
represented by the intersection of an J^BAJJ curve with the "^pAp curve
v
shown. Both "cold" and "hot" stable solutions can occur, but the "hot"
stable solution is the one of interest.
Two items of interest in Figure 5 are events concerning self-ignition
and extinction. Self-ignition of the bed can occur at very low space
velocities if effectively adiabatic conditions occur. At high values of
space velocity the value of the mass transfer coefficient p u (L. will
also increase, and eventually extinction will occur. The extinction
space velocity values for a few adiabatic flame temperatures can provide
a characterization of a given catalyst bed/fuel combination.
In addition to the preceding "thermal" analysis, exit composition
measurements will further assist in the formulation of the surface kine-
tic reaction processes. In addition, the effectiveness of the catalyst
to yield equilibrium NO compositions at its surface will also be eval-
X
uated.
SUMMARY
The use of catalysts in place of conventional burners for promoting
hydrocarbon reactions appears to have advantages in the control of emis-
sions. A research program has been undertaken that addresses the basic
lack of knowledge concerning catalyst characteristics, wash coat, sup-
port, and bed depth required for an oxidation reaction to go to comple-
tion. In addition, the program will study system techniques such as
II- 192
-------
bed cooling, exhaust gas recirculation, and staged combustion to hold
the bed temperature down. The results of this study will be directly
applicable to the design of compact catalytic combustion systems for
residential and industrial applications.
11-193
-------
REFERENCES
1. Davy, H., "Some New Experiments and Observations on the Combustion
of Gaseous Mixtures," J. Davy, ed., The Collected Works ofSir
Humphrey Davy, Vol. VI, Miscellaneous Papers and Researches, Part I,
On the Safety Lamp for Preventing Explosions in Mines, Houses
Lighted by Gas, Spirit Warehouses, or Magazines in Ships, etc.,
With Some Researches on Flame, Section II, Papers Published in the
Philosophical Transactions and in the Journal of Science and the
Arts, on the Fire Damp, the Safety Lamp, and on Flame, Smith, Elder,
and Co., Cornhill, London, 1840, pp. 81-88.
2. Spalding, D. B., "Heat Transfer from Chemically Reacting Gases,"
Ibele, W., ed., Modern Developments in Heat Transfer, Academic
Press, N.Y., 1963, pp. 19-64.
3. Liitken, A. and Hoist, H., Opfindelsernes Bog. Vol. IV, Nordisk
For lag, Kgibenhavn, 1914, pp. 131-139.
4. Bone, W. A., "Surface Combustion and Its Industrial Applications,"
Engineering. Vol. 91, April 14, 1911, pp. 487-489.
5. Bone, W. A., "Surface Combustion," Engineering, Vol. 93, May 10,
1912, pp. 632-634.
6. Walker, W. H., Lewis, W. K., McAdams, W. H., and Gilliland, E. R.,
Principles of Chemical Engineering. 3rd Edition, McGraw-Hill Book
Company, Inc., New York, 1937, pp. 202-204.
7. Thompson, R. E., Pershing, D. W., and Berkau, E. E., "Catalytic
Combustion — A Pollution-Free Means of Energy Conversion," Environ-
mental Protection Technology Series EPA-650/2-73-018, Aug. 1973.
8. Acres, G. J. K., Bird, A. J., and Davidson, P. J., "Recent Develop-
ments in Platinum Metal Catalyst Systems," The Chemical Engineer,
March 1974.
9. Barnes, G. J. and Klimisch, R. L., "Initial Oxidation Activity of
Noble Metal Automotive Exhaust Catalysts," SAE Paper 730570, pre-
sented at the Automobile Engineering Meeting, Detroit, Michigan,
May 14-18, 1973.
10. National Materials Advisory Board, "A Semi-Delphi Exercise on Sub-
stitute Catalysts for Platinum in Automobile Emission Control De-
vices and Petroleum Refining," NMAB-314, March 1974.
II- 194
-------
TABLE OF CONVERSION FACTORS FOR METRIC UNITS
English Units
atmosphere
Btu
Btu/hour
degrees Fahrenheit
foot
inch
Metric Units
1.0133 x 105 newtons/meter2
2.52 x 102 (g) cal
2.52 x 102 (g) cal/hour
(tp - 32)/1.8 degrees Celcius
3.048 x 10"1 meters
2.540 x 10~2 meters
II-195
-------
11:40 a.m.
Pilot Scale Investigation of
Catalytic Combustion Concepts
for Industrial and Residential
Applications
Dr. John P. Kesselring, Aerotherm Acurex Corporation
Will you be considering the possibility of using a
catalyst monolithic screen in a location between a
furnace and a heat exchanger as a burnout medium for
soot that is passing through?
We will not be considering that in this program.
In fact, the avoidance of sooting conditions in cata-
lytic combustion appears important, since active sites
can be buried by soot deposition inside the monolith.
11-196
-------
THE OPTIMIZATION OF BURNER DESIGN PARAMETERS
TO CONTROL NOV FORMATION IN
J\
PULVERIZED COAL AND HEAVY OIL FLAMES
M. P. Heap, T. J. Tyson and G. P. Carver
Ultrasysteins, Inc.
G. B. Martin
Environmental Protection Agency
and
T. M. Lowes
Associated Portland Cement Manufacturers
11-197
-------
ACKNOWLEDGEMENTS
The work reported in this paper was carried out under EPA
Contracts 68-02-0202 (International Flame Research Foundation) and
68-02-1488 (Ultrasysterns, Inc.). Two of the authors (M.P. Heap and
T. M. Lowes) wich to express their appreciation to all their former
colleagues at the IFRF Research Station. Dr. E. E. Berkau and
Mr. D. W. Pershing made significant contributions to this work by
virtue of many stimulating discussions.
11-198
-------
1.0 INTRODUCTION
In recent years considerable effort has been expended in an
attempt to reduce the emission of nitrogen oxides from stationary com-
bustion sources. Several parallel paths have been followed in an
attempt to achieve this objective, including the modification of
operational procedures and system redesign. This paper describes a
series of small scale investigations which were carried out to identify
burner parameters which control nitric oxide formation in pulverized
coal and heavy fuel oil flames. The amount of nitric oxide produced
in turbulent diffusion flames appears to be strongly dependent upon
the rate of fuel/air mixing. Two burner design options have been
identified which control mixing and, therefore, influence nitric oxide
formation. These options are:
• the method of fuel injection, and
• the method of combustion air delivery.
Both options also affect flame length, flame volume and heat absorp-
tion at the boundary of the combustion chamber. The paramount objec-
tive of the investigations discussed in this paper is the satisfaction
of combustor requirements as well as the minimization of pollutant
emissions. Plans will also be discussed to investigate the influence
of scale and multiple burner systems on low emission, high efficiency
combustion systems suitable for both industrial and utility boilers.
II- 199
-------
2.0 NITRIC OXIDE FORMATION IN PULVERIZED COAL AND HEAVY OIL FLAMES
Nitrogen oxides are formed from two sources of nitrogen during
the combustion of fossil fuels, molecular nitrogen, and nitrogen com-
pounds which occur naturally in both liquid and solid fuels. The
reactions controlling the rate of oxidation of molecular nitrogen,
producing thermal NO, are strongly temperature dependent and proceed
at significant rates only above 1600°C. It was originally thought
that the reaction between nitrogen molecules and oxygen atoms was
mainly responsible for NO production in flames. However, it is now
believed that reaction of hydrocarbon radicals with molecular nitrogen
in a flame zone provides an additional path for thermal NO formation.
Nitrogeneous material in fossilized animals and vegetation provided the
source for nitrogen compounds which are present in coal and oil. During
the combustion of these fuels some of this nitrogen is oxidized giving
fuel NO. The amount of fuel NO formed in simple premixed flames is not
strongly dependent upon temperature, but it is most sensitive to oxygen
availability.
The processes involved in heat release and nitric oxide forma-
tion in pulverized coal and heavy oil flames are extremely complex.
The fuel and air are supplied separately to the combustion chamber
through a device which ensures that they are mixed and that the fuel
is in a condition to ignite and burn completely. In order to discuss
those features of practical flames which influence NO formation the
total process will be considered in three simple stages.
11-200
-------
Fuel Preparation . Liquid fuels are atomized, mixing takes
place between the fuel, the combustion air and recirculated
hot combustion products. Both liquid and solid fuels evolve
gaseous components which will ignite under suitable conditions.
Heat Release. The nature of the fuels and the turbulent
mixing process allow heat release to take place in several
modes. The volatile fuel fractions can burn: 1) in homo-
geneous premixed zones, or 2) in diffusion flames around single
droplets or clouds of droplets. Soot formed from the volatile
fuel fractions must be oxidized as must the char which remains
after the devolatilization of the coal particles.
Heat Transfer. The products of combustion leave the heat
release zone and mix with the bulk gases and lose heat before
leaving the combustion chamber.
In cold wall combustion chambers nitric oxide production can
be considered to occur only during the heat release stage. Extensive
reviews of the kinetic mechanisms associated with thermal NO formation
(1 2)
in hydrocarbon flames are available; ' but for the purposes of this
discussion it is convenient to refer to the two dominant mechanisms as
Zeldovich, i.e.,
N + 0 — > NO + N
or Fenimore, ' i.e.,
N2 + R — *• RN + N
where R is some hydrocarbon fragment such as CH, and oxidation of the
nitrogen fragments produces NO, The Zeldovich path will be dominant
in fuel lean zones, and the amount of thermal NO produced will depend
upon:
— the stoichiometry of the reaction zone;
— the amount of diluent combustion products in the
reaction zone;
11-201
-------
the temperature of those diluents; and
the rate at which the freshly formed combustion
products mix with bulk gases and thus their rate
of temperature decay.
The Fenimore path will be dominant in fuel rich zones, and the amount
of NO produced will probably depend upon the carbon/hydrogen ratio of
the fuel fragments.
(2)
Sternling and Wendt discussed the fate of chemically bound
nitrogen during the combustion of solid and liquid fuels. Lacking
definitive experimental results these workers assembled fragmentary
data and concluded that 80 percent of the nitrogen in coal would appear
in the coal char and that a smaller, but still major, portion in the
solid formed after devolatilization of oil droplets. Subsequent oxida-
tion of the char would then produce the major portion of the fuel NO.
(4 5)
Heap et al, ' postulated that fuel NO accounted for the major portion
of the emission from coal flames and that the fuel NO was mainly formed
from "volatile" fuel nitrogen compounds. Pershing et al, burned coal
with both air and oxygen/argon mixtures and showed that approximately
80 percent of the total emission could be attributed to fuel NO. The
possible fate of chemically bound nitrogen in an oil droplet or coal
particle is traced in Figure 1.
A simple working hypothesis of fuel NO formation can be con-
structed based upon available information from studies involving doped
fuels and the addition of nitrogen compounds to laboratory premixed
II- 202
-------
OIL DROPLET
OK
COAL PARTICLE
VOLATILE PRACTIONS
(HYDROCARBONS, RN etc.)
UN
Figure 1. The Possible Fate of Fuel Nitrogen Contained
in Coal Particles or Oil Droplets During Combustion
11-203
-------
and gaseous diffusion flames. The primary fuel nitrogen compounds
are broken down yielding secondary intermediate compounds RN which are
subject to two competitive reaction paths:
A — reaction with an oxygen containing species producing
NO, or
8 — reaction with a nitrogen containing species producing
Nr
Thus the formation of fuel NO can be considered simply as
Fuel
NO + ...
*»
A
B
The production of nitrogen (path B) will be favored under fuel rich
conditions and the conversion of fuel nitrogen to fuel NO in flames
can be limited by ensuring that the majority of fuel nitrogen com-
pounds react under oxygen deficient conditions.
Once formed, nitric oxide can be reduced in heat release zones,
a process which could allow the net production of NO in flames to be
reduced. An experimental study by Wendt, Sternllng, and Matovich^ '
demonstrated that "reburning" was effective and that up to 50 percent
to\
reduction in flue gas NO level was obtained. Engleman et al, added
NO to a well-stirred reactor burning methane/air mixture and found that
almost all the NO was retained in lean mixtures, but that the 65 per-
cent of the input NO was reduced when it was added to reactants con-
taining only 65 percent of the stoichiometric air requirements.
11-204
-------
(9)
Sarofim et al, demonstrated that a "diffusion flame environment"
facilitated NO destruction. Myerson has shown that NO contained
in simulated combustion products can be reduced by isobutane when
flowed through a heated ceramic tube at temperatures in excess of
1000°C. More recent information on NO reduction by hydrogen has been
presented by Flower, Hanson and Kruger . Thus, the possibility
exists that NO could be produced in the early stages of heat release
and reduced before leaving the heat release zone.
None of the previous discussion has alluded to the possibility
of N02 production in flames. Nitrogen dioxide has been observed in
both flat flames and turbulent diffusion flames. In flat flames the
N02 is produced in the visible flame region and, in most cases, is
rapidly converted to NO downstream of the visible zone. High N02
concentrations have been observed in regions of high shear or on the
(12 13)
periphery of the base of natural gas air flames. ' The effects
of N02 formation in coal and heavy oil flames are unknown. In fuel
oil or pulverized coal flames it is unlikely that the rapid formation
and destruction of N02 within the heat release zone would influence
the net production of NO unless this process competes with those
reactions leading to the formation of N2 from NO. Consequently, NO
would be converted to N02 only to be reformed at a later stage, rather
than being reduced to N2-
Two generalized schemes are presented in Figure 2 to illustrate
possible approaches to the control of the net production of NO in
H-205
-------
HEAT LOSS
BUT LOSS
SECOND STACE AIR
TOTAL AIX
// i /
——4-4-
PRIMARY
RICH
ZONE
-x^-,_
PRODUCTS
PARTIAL £ OX
— x-J
OF
-x »
IDAIIOH
SECONDARY'
BURNOUT
ZONE
TO FLUE
J
STAGED HEAT RELEASE
niMAKY FUEL
1—
i
L_
SECONDARY FUEL
HIGH INTENSIlT
LOU EXCESS AIR
PRIMARY ZONE
PRODUCTS OF
COMPLETE"! COMBUSTION*1
1
SECONDARY AIR
REBURNINC
ZONE
. _^J BURNOUT
t\ ZONE
REBURNINC
Figure 2. Generalized Schemes for the Reduction of
HO Emission from Turbulent Diffusion in Flames
11-206
-------
turbulent diffusion flames. The more familiar approach is through
staged heat release by reducing the rate at which air is mixed with the
fuel. Thus a rich primary zone is produced which creates conditions
favorable for path B. The crucial practical implications of this sys-
tem are that the second stage air must be added in such a way as to
ensure complete combustion within the constraints imposed by practical
systems; i.e., the need to produce steam and all that this implies in
terms of both thermal and operating efficiency. Although in Figure 2
the primary and secondary zones are shown as being physically separated,
this must not necessarily be the case — some compromise being necessary
to accommodate the heat release zone in the available physical space.
The alternate scheme for the NOX control during combustion
of nitrogen containing fossil fuels is based upon the concept of
reburning. High intensity rapid mixing conditions favor both NO forma-
tion and efficient combustion. In the primary region the major portion
of the fuel is burnt without regard for NO control. Secondary fuel is
added to the products of combustion and the nitric oxide is reduced
in the fuel rich mixtures. Subsequent addition of air completes com-
bustion. This scheme would require rapid mixing of the secondary fuel
with the primary products which should contain the minimum amount of
excess air, probably requiring that the secondary fuel be gaseous.
This paper is concerned with investigations to determine
the optimum mixing pattern to control nitric oxide formation by the
11-207
-------
former scheme, which aims to prevent excessive formation of nitric
oxide in the primary zone and promote burnout of the partially oxidized
fuel in the secondary zone.
3.0 SUBSCALE EXPERIMENTS TO ESTABLISH OPTIMUM CONDITIONS
FOR MINIMUM POLLUTANT EMISSIONS
EPA Contract 68-02-0202 carried out by the Staff of the Inter-
national Flame Research Foundation, IJmuiden, Holland, was concerned
with defining those burner parameters which control the rate of NO
formation in turbulent pulverized coal and heavy fuel oil flames. The
results of these investigations will be described to illustrate various
approaches which can be used to provide the conditions illustrated in
Figure 2 as staged heat release. Two methods have been identified
which allow the control of nitrogen oxide formation while maintaining
satisfactory combustion conditions. These are:
• division of the total combustion air supply into separate
streams, thus delaying mixing the total air supply with
all the fuel,
• optimization of the method of fuel injection.
The subscale experiments were carried out in a refractory tunnel
furnace of 2 meters cross section and 6 meters long; complete details
/ 1 0\
of the experimental system have been given elsewhere. Figure 3
presents a schematic showing the arrangement of the versatile test
burner used in the investigations. This burner utilized the moving
(4)
block swirl generator designed by Leuckel and readily allowed the
influence of several parameters on pollutant emissions to be investi-
11-208
-------
N
PM
iH
•H T3
° §
M
O eg
0)
M
C Vi
o FM
IN ••-(
o w
-------
gated. The most significant being 1) the method of fuel injection,
2) division of the total air supply between primary, secondary and
tertiary air streams, 3) the secondary swirl imparted by the moving
blocks, 4) the primary swirl imparted by fixed vanes, 5) the angle of
the burner divergent, 6) the number of injection points for the tertiary
air supply, and 7) the position of the fuel injector relative to the
burner exit.
3.1 Optimization of the Method of Fuel Injection
Parametric studies very quickly indicated that the method of
fuel injection had a profound effect upon nitric oxide emission from
pulverized coal flames. Measured emissions ranged from 800 to 200 ppm
for a 1.1 percent nitrogen coal with 300°C preheat and 5 percent excess
air. Maximum emissions were achieved when the fuel was injected
radially into the air stream. Heap et al, ' ' ^ postulated that this
method of fuel injection provided the most oxygen-rich conditions for
the combustion of the volatile coal fractions. Those injection con-
ditions which tended to minimize the rate of mixing between the fuel
and the total combustion air gave minimum emissions of nitric oxide,
and were provided by a high velocity axial fuel injector utilizing the
minimum amount of primary air. Figure 4 gives an impression of the
range of emissions obtained for a given primary air percentage, excess
air level and burner geometry. Although pulverized coal flames could
be produced with low NO emission, the length of these low NO flames is
unacceptable for use in existing boiler configurations. Practical
requirements in utility boilers dictate stable, high intensity
11-210
-------
800 r
A
D MEDIUM VELOCITY AXIAL
O HIGH VELOCITY AXIAL
1 2
SWIRL NUMBER
Figure 4. Range of NO Emissions Produced by Change
of Fuel Injector (1.1% N2 Coal, 300°C Preheat,
5% Excess Air)
II- 211
-------
combustion with rapid burnout. These combustion characteristics are
attained in practice by using a fuel injection system which is similar
to the radial fuel injector used in the subscale experiments. Thus,
rapid fuel/air mixing provides satisfactory combustion conditions, but
also promotes NO formation.
Subsequent experiments with a heavy fuel oil produced results
which provide similar conclusions to those obtained with pulverized
coal. Minimum NO emissions were obtained by maintaining the fuel oil
as a single coherent jet on the flame axis, thus producing a long
thin cylindrical flame similar in appearance to the low NO coal flame.
Further investigations concentrated upon changing the design of the
burner to maintain the required flame pattern, while attempting to
reduce NO emissions to the minimum level possible. Japanese workers
suggested an attractive technique for achieving these conditions which
involved splitting the flame into several discrete sections which
gives the flame a flower petal arrangement. In these investigations
this was achieved by placing a distributor tip over a standard Y-jet.
Details of two types of distributor tips which have been investigated
are presented in Figure 5; their emission characteristics, plotted
as a function of secondary swirl intensity, are shown in Figures 6 and
7. It can be seen in Figure 6 that at low secondary swirl levels emis-
sions from standard Y-jets are considerably higher than those obtained
with distributor tips. However, as the swirl level is increased, the
emissions from flames produced with distributor tips increase and, in
11-212
-------
m
O
•
O
55
4
to
0)
o
•H
^W
•H
H
O
-^s
|
QQ
a
o
SB
00 00 CM
i— 1
o m o
CM CM CM
O O O.
o o c4
o o o
o o o
^H CM f)
(fl
I
•H
O
O
SS
^H CM M
E
pq'
1
*~r
OQ
a
*
o
^H CNJ
CM •*
0 0
in m
O O
m m
CM CM
-» m
Figure 5. Details of Distributor Tips Used
NOTE: All dimensions are in millemeters
11-213
-------
g
0.
i
150 H
O
O
D
A
$0° Y-JET
70° Y-JET
DISTRIBUTOR TIP 1
DISTRIBUTOR TIP 2
DISTRIBUTOR TIP
0.4 0.6 0.8 1.0
SECONDARY SWIRL INDEX
Figure 6. Emission Characteristics of Fuel
Oil Flames (5% Excess Air, 300°C Preheat,
Primary Swirl Vanes 30°)
11-214
-------
260.
DISTRIBUTOR TIP
NO CO
4 O •
too
.4 .'6 .8
SECONDARY SWIRL INDEX
Figure 7. Emission Characteristics of Fuel Oil
Flames (5% Excess Air, 300°C Preheat,
Primary Swirl Vanes 30°)
11-215
-------
one instance, approach the levels obtained with standard Y-jets.
Distributor tips 4 and 5, which included an axial port, tended to
produce considerable amounts of carbon monoxide at low swirl levels.
As the swirl was increased, nitric oxide emissions increased and
carbon monoxide emissions decreased. Further experiments indicated
that the emission characteristics of fuel oil flames, produced with
the distributor tip design shown in Figure 5, were not only dependent
upon the configuration of the tip, but also upon other parameters such
as the angle of the burner exit and the furnace wall temperature.
From these investigations it can be concluded that this type of fuel
injection system could be readily adapted for single flame applica-
tions. However, it may well be that the interaction between flames
in multiple burner installations will negate any value with respect
to reducing nitric oxide emissions. Consequently, further work is
necessary to establish the influence of flame interactions on emis-
sions before this type of tip can be recommended for use in multiple
burner installations.
The characteristics of flames produced with the distributor tips
were found to be different from those produced from the standard Y-jet.
The extent of this difference can be judged by comparing the radial
profiles of species concentration and temperature presented in Fig-
ure 8. Measured temperature, carbon monoxide, hydrogen, oxygen and
methane concentrations are shown at an axial distance equal to
1.6 burner diameters from the face of the firing wall. It is evident
from these radial profiles that there is considerable difference
H-216
-------
STANDARD Y-JET
Flame
Axis
* TEMPERATURE
D CO
0 H2
• 1400
- 1000
Ul
u
on
\a
a 6
TIP I
2
|
<
TIP 4 /
IEOO
800
ta
U.
"
2
u
1200
800
Radial Distance
Figure 8. Radial Profiles of Temperature
and Species Concentration
11-217
-------
between the flame produced with the Y-jet and those produced with
the distributor tips. The low NO flames are considerably longer, as
shown by the carbon monoxide and hydrogen concentrations, and mea-
sured temperatures are much lower. The distributor tips also produce
wider flames at this axial location. Consequently, utilization of
this method of fuel oil injection would require that considerable
attention be paid to the problem of flame impingement.
3.2 Division of the Total Air Supply into Separate Streams
The efficiency of staged combustion as a technique for reducing
nitric oxide emissions from utility boilers is well established. The
methods used in actual installation normally involve increasing the
fuel flow through several of the burners, and using the remainder of
the burners only to supply air. Experiments have been conducted at
subscale to establish whether a combustion process can be effectively
staged using the burner as a staging device by delaying mixing of all
the fuel and air. Once more, the dual criteria of minimum emissions
and acceptable flame characteristics were imposed upon these investi-
gations since this type of control technique must be acceptable in
existing equipment.
Investigations were carried out with both fuel oil and pulverized
coal and it was found that the burner can be used as an effective staging
device for both these fuels. Emission characteristics are shown in
Figure 9 for two oil flames burning under identical furnace and burner
-------
Q.
O.
180
EXIT NO SOLIDS.
PPM 0% 02 UG/NM*
a
.4 .6
SWIRL INDEX
Figure 9. Reductions in NO Emissions From a
Fuel Oil Flame by Diverting 35 Percent of
the Total Air Through the Tertiary
Injectors (5% Overall Excess
Air, 300°C Preheat)
11-219
-------
conditions except that the staged flame is 35 percent of the air delivered
through eight parallel tertiary injectors. Included on the same figure
are values of solid emissions measured with a 0-Ray absorption device;
it can be seen that the reduced NO emissions achieved by the staged burner
are not accompanied by significant increases in the level of solid emissions.
However, the reduced rate of fuel/air mixing which is believed to be the
cause of the reduced production of NO also causes an increase in flame
length. This can be appreciated by comparing the axial temperature and
species concentrations distribution for staged and unstaged flames
(see Figure 10). Temperatures in the unstaged flame are considerably
higher than those measured under staged conditions. Flame length can
be UMMed either from visual observation, or it can be denoted by a
particular carbon monoxide concentration. It is not possible from
these measurements to ascertain whether the characteristics of the low
NO produced by the staged oil flame are acceptable for normal boiler
practice. However, these encouraging results indicate that the con-
cept shows considerable promise, and that further investigations should
be carried out under conditions more representative of large water
walled boilers.
Similar experiments were carried out with pulverized coal to
establish whether these techniques were applicable to solid fuels. Fig-
ures 11 and 12 give an indication of the success of this method of
staging for reducing NO emissions from pulverized coal flames. As the
11-220
-------
CD
H
•H
O
'O
V
00
•o
-------
FUEL
INJECTOR
SWIRL
INDEX
600-
g 400
a
a,
i
200
C.5.
RAD.
0.6
0.9
0
0.6
O
a
v
A
"3
Mtotal
.4
.6
Figure 11. The Influence of Mass Flow of Tertiary Air (143) on
NO Emissions from Pulverized Coal Flames (5% Excess Air,
300°C Preheat, 950°C Wall)
11-222
-------
1000
FUEL
INJECTOR
C.S.
RAD.
SWIRL
INDEX
0.6
0.9
0.6
0.9
0.4
0.6
o
a
Figure 12. The Influence of Mass Plow of Tertiary Air (Mj) on
NO Emissions from Pulverized Coal Falmes (5% Excess Air,
300°C Preheat, 1250°C Wall)
11-223
-------
percentage of air injected through the tertiary ports increases the
NO emission decreases markedly . It was found that the efficiency of
staging depends both upon furnace wall temperature and upon the method
of fuel injection. Figure 11 presents results obtained with an average
wall temperature of approximately 950°C, whereas the furnace cooling
load was reduced before the results presented in Figure 12 were
obtained causing the wall temperature to increase to approximately
1250°C in the region of the burner. The influence of furnace wall
temperature on emissions from pulverized coal flames had been observed
before} temperature had a strong influence on NO emission from both
natural gas and pulverized coal flames, but only a minor influence on
emissions from oil flames. The increased wall temperature and subse-
quent increase in the temperature of recirculated gases produce am
increase in NO emission of about 200 ppm for the staged condition.
It is contended that the influence of the method of fuel injection on
the efficiency of staging is associated with the residence time of the
coal particles within the rich primary zone. The intensity of swirl
in the secondary air appears to have a negligible influence upon
nitric oxide emissions at high staging levels. Under all circumstances
carbon burnout was virtually complete at the exit of the furnace for
both staged and unstaged flames. Figure 13 compares the influence of
The results refer to two methods of fuel injection, radial (rad) and
a scaled down version of a fuel injector typical of those used in
practice (c.s).
11-224
-------
tertiary air injection on carbon monoxide emissions for the two methods
of fuel injection with the lower furnace wall temperatures. It can be
seen that staging has no influence upon CO emissions from the radial
fuel injector; however, the six-fold decrease in NO emission with the
coal spreader type injector is accompanied by a three-fold increase in
carbon monoxide emissions at high levels of burner staging. The experi-
ments made no attempt to optimize conditions for low CO emissions which
could probably be achieved by further minor modifications to the fuel
injector.
3.3 Investigations Involving Petroleum Coke
Petroleum coke is a high nitrogen, low volatile, low ash solid
which was burned under identical conditions to those for burning
pulverized coal. NO emissions from petroleum coke were found to be
generally lower than those from coal. This can be appreciated by
comparing the influence of firing rate on nitric oxide emissions for
the two fuels (see Figure 14). At the nominal firing rate of approxi-
mately 5 x 10^ BTUs per hour with 300°C preheat and 15 percent excess
air, emissions from pulverized coal flames are approximately twice
those of petroleum coke when both fuels are injected to ensure rapid
mixing between the fuel and combustion air. Increasing the firing
rate increases the bulk gas temperature within the furnace which
causes an increased NO emission from both fuels. However, the
influence of firing rate is much stronger with pulverized coal,
particularly at high firing rates, once more indicating that the
temperature environment of pulverized coal flames has a very strong
influence upon nitric oxide production. The profound influence of
II-225
-------
160
s
M
Q.
Q.
Ul
Q
I
O
i
U
80-
40-
FUEL
INJECTOR
C.S.
RAD.
SWIRL
INDEX
0.6
0.9
0
0.6
O
D
V
0.2
0.6
total
Figure 13. The Influence of Mass Flow of Tertiary Air (M )
on Carbon Monoxide Emissions (5% Excess Air,
300°C Preheat, 950°C Wall)
H-226
-------
100W
g
I «oo
HOQ-I
1.0
COAL
PETROLEUM
COKE
1.6
FRACTIONAL LOAD
Figure 14. Nitric Oxide Emissions as a Function of
Load for Coal and Pulverized Coke
II-227
-------
the method of fuel injection upon the emission of nitric oxide from
pulverized coal flames has been discussed earlier. It was found that
emissions from petroleum coke were almost independent of fuel injector
type. Emissions from petroleum coke could also be decreased by using
the burner as a staging device. Figure 15 indicates that emissions
could be reduced by almost half if 50 percent of the air was supplied
through the tertiary injectors. This reduction was found to be inde-
pendent of the secondary swirl level.
Since petroleum coke is a low volatile, high nitrogen solid,
comparison of the results presented in Figure 14 could be used to pro-
vide support for the hypothesis that volatile nitrogen compounds are
responsible for fuel NO formation in pulverized coal flames. This
support would be based upon the assumption that the fuel nitrogen com-
pounds in coal and oil are similar and that their evolution from the
solid particle would be similar under similar time/temperature histories.
At this time these are rather sweeping assumptions and, although the
results tend to support the hypothesis, they cannot be considered as
definitive.
4.0 SCALE-UP CRITERIA FOR LOW EMISSION, HIGH EFFICIENCY BURNERS
The results of the subscale experiments discussed in the pre-
vious section demonstrated that low emission, high efficiency burners
could be operated with both pulverized coal and heavy fuel oil.
Burner designs which reduce the rate of coal/air mixing (and thus NO
emissions) are already being tested in large utility boilers. The
11-223
-------
400 -
0~
S
CL
a
ZOO-
PRIMARY S,
0.4 O
0.6 O
0.9 A
.6
M.3_
Mtotal
Figure 15. The Influence of Burner Staging on
NO Emissions from Petroleum Coke Flames
H-229
-------
relevance of the subscale investigations to practical situations can
be criticized on three points. Firstly, all the subscale tests were
conducted in a horizontal tunnel furnace where the flames fired
horizontally, and the exit of the furnace was on the flame axis. The
recirculation patterns were very different from those likely to be
encountered in large industrial and utility boilers. Secondly, the
results refer to single burner conditions, and the influence of flame
interactions on NO formation is ill understood. Thirdly, it has not
been established whether the principles discussed in Section 2.0 of this
paper are suitable for scaling to ensure their applicability in practi-
cal boilers. Investigations to be carried out under EPA Contract 68-02-
1488 with Ultrasystems have been planned to answer all of these questions,
Emission control techniques which reduce pollutant emissions at
the expense of fuel economy or operating efficiency are not only
financially unacceptable, but they are also socially irresponsible.
The overall objective of this new EPA program is to design, construct,
and test low emission, high efficiency burners in a size suitable for
use in industrial and utility boilers. This research program will:
• Design and construct a test facility which will allow
large scale burners to be tested under conditions repre-
sentative of those to be encountered in industrial
practice.
• Establish the influence of burner/burner and burner/
furnace interactions on pollutant formation in large
scale fossil fuel flames.
H-230
-------
• Establish scaling laws to ensure that design methods are
available to enable low emission, high efficiency burners
to be constructed for a wide range of firing rates.
• Ensure that reduced pollutant emission is not accomplished
at the expense of reduced thermal efficiency and increased
operating costs.
• Establish the optimum firing method for dual firing of low
BTU gas and pulverized coal in both and single and multi-
ple burner arrays.
• Specify burner designs which are suitable for extensive
testing in industrial and utility boilers.
This program will be carried out with three fuels: pulverized coal,
low BTU gas, and heavy fuel oil, burned singly and in combination.
The investigations will include both input/output measurements, and
detailed flame probing to establish the characteristics of the low
emissions flames.
The goals of the program can only be satisfied if a flexible
experimental facility is available. The focal point of this facility
must be a combustor which simulates wall temperature and recirculation
patterns typical of modern industrial and utility water wall boilers.
It must also allow rapid interchange of multiple burner arrays to
enable an assessment of the influence of such parameters as burner/
burner spacing and the distance of burners from adjacent walls on
pollutant emissions. Figure 16 shows a cutaway sketch of a modular
spray-cooled combustor which has been designed to satisfy the require-
ments of this project by providing the representative environment and
the required degree of flexibility. The major features of this combustor
II-231
-------
SKIN PANELS
EX05KEUTON
(SPRAY MANIFOLD
AND STRUCTURAL
FRAME)
ASH HOPPER
WATER SPRAY
MOVEABLE WALL
Figure 16. Cutaway Sketch of a Modular Moveable
Wall Spray Cooled Combustor
11-232
-------
are its low thermal inertia; up to five burners can be placed anywhere
on the firing wall; one wall of the combustor is tnoveable, thus allow-
ing the influence of burner/furnace interactions to be established;
the modular construction allows probes to be inserted in almost any
location; and Figure 17 presents a schematic showing the arrangement
of a flame probe and the corabustor wall. The entire facility will
stand alone, will be dedicated to this project, and will have a firing
rate capacity for single burners of 120 x 10 BTU/hr. Multiple burner
investigations will be able to be carried out with a maximum of five
burners firing at a total rate of 60 x 10 BTU/hr. Three single burner
sizes of 12, 60 and 120 x 10 BTU/hr will be tested and their charac-
teristics established, thus providing information on scale up criteria
of low emission burners. Some idea of the extent of this test facility
can be gained from the general layout which is presented in Figure 18,
which gives the approximate location of the heat exchanger, pulverizer,
steam supply and propane reformers which will be used to provide low
BTU gas. This program is presently in the design phase; it is expected
that construction of the test facility will commence in the near future.
5.0 CONCLUSION
This paper has presented experimental information which clearly
demonstrates that low emission, high efficiency burners can be opera-
ted in subscale. At this time it is pertinent to consider the EPA's
11-233
-------
COOLING
WATER
HOSES
HYDRAULIC
ACTUATOR
HOSES
WALKWAY
SPRAY
MANIFOLD
WATER
COOLED PROBE
IN COMBUSTION
CHAMBER
SPECIAL
APERTURE
MODULE
ROLLERS IN
OUTER SKIN
SUPPORT
CHANNELS
Figure 17. Schematic Representation Showing the Location and
Actuating Mechanism of Flame Probe
11-234
-------
•?j~Jkizgbi
* «°25 5
:u«rtO«w;i
osoESi-s
_li*iOlu-l<«c
momctuSo
i t i i i i i
m
o
id
PN
4-1
CO
0)
H
4-1
3
O
0)
g
O
0)
M
00
O
s
DC
H-235
-------
projected long and short term goals for utility boilers as stated by
Lachappelle, Bowen and Stern.
Table 1. Project Short and Long Term Goals
for NOX Control from Utility Boilers
ppm at 3 percent
Gas
Residual Oil
Coal
1980 Goal
100
150
200
1985 Goal
50
90
100
The short term goals can be achieved at present under controlled
experimental conditions. It remains to be seen whether they are
achievable in practical operating plants. Section 4.0 of this
paper discussed plans to construct a test facility which will pro-
vide much needed information on the characteristics of low emission
combustion systems for coal and heavy fuel oil. This information
will give confidence in the design concepts and give a clear indica-
tion of whether the goals presented in Table 1 are achievable.
II- 236
-------
REFERENCES
1. Bowman, C.T., The Fourteenth Symposium (International) on
Combustion, p. 729, The Combustion Institute, 1973.
2. Sternling, C.V. and Wendt, J.O.L., "Kinetic Mechanisms Governing
the Fate of Chemically Bound Sulfur and Nitrogen in Combustion."
Shell Development Company, Emeryville, California, EPA-650/2-
74-017 (NTIS No. PB 230-895/AS), August 1972.
3. Fenimore, C.P., Thirteenth Symposium (International) on Com-
bustion, p. 373, The Combustion Institute, 1971.
4. Heap, M.P., Lowes, T.M. and Walmsley, R., "Nitric Oxide
Formation in Pulverized Coal Flames," Combustion Institute
European Symposium, p. 493, Academic Press, (1973).
5. Heap, M.P., Lowes, T.M., "Burner Design Optimization to Control
NOX Emissions," Presented at the "Coal Combustion Seminar,"
Environmental Protection Agency, Research Triangle Park, N.C.,
June 19-20, 1973, EPA-650/2-73-021, (NTIS No. PB 224-210/AS)
September 1973.
6. Pershing, D.W., Martin, G.B. and Berkau, E.E., "Influence of
Design Variables on the Production of Thermal and Fuel NOX
from Residual Oil and Coal Combustion." Paper presented at
the AIChE 66th Annual Meeting, Philadelphia, 1973.
7. Wendt, J.O.L., Sternling, C.V., and Matovich, M.A., Fourteenth
Symposium (International) on Combustion, p. 697, The Combustion
Institute, 1973.
8. Bartok, W., Engleman, V.S. and del Valle, E.G., "Laboratory
Studies and Mathematical Modeling of NOX Formation in Com-
bustion Processes," Esso Research and Development Co., EPA
report APTD 1168 (NTIS No. PB 211-480), 1972.
9. Sarofim, A.F., Williams, G.C., Modell, M. and Slater, S.M.,
"Conversion of Fuel Nitrogen to Nitric Oxide in Premixed and
Diffusion Flames." Paper presented at the AIChE 66th Annual
Meeting, Philadelphia, 1973.
10. Myerson, A.L., Fifteenth Symposium (International) on Combustion,
p. 1085, The Combustion Institute, 1975.
H-237
-------
11. Flower, W. L., Hanson, R. K. and Kruger, C. H., Fifteenth
Symposium (International) on Combustion, p. 823, The
Combustion Institute, 1975.
12. Cernansky, N. P. and Sawyer, R. F., Fifteenth Symposium
(International) on Combustion, p. 1039, The Combustion
Institute, 1975.
13. Heap, H. P., et al. "Burner Criteria for NO Control, Volume I-
Influence of Burner Variables on NO in Pulverized Coal Flames."
International Flame Research Foundation, IJmuiden, Holland,
EPA-600/2-76-061a (NTIS No.-Later), March 1976.
14. Leuckel, W., Swirl Intensitites, Swirl Types and Energy Losses
of Different Swirl Generating Devices, Doc. G02/a/16, IFRF,
IJmuiden, Holland (1968).
15. Heap, M. P., Lowes, T. M., and Martin, G. B., "Fluid Mechanics
of Combustion," p. 75, The American Society of Mechanical
Engineers, April 1974.
16. Tsuji, S., Tsukada, M. and Asai, M., Ishikawima-Harima
Engineering Review 13, p. 227 (1973).
17. Lachapelle, D. G., Bowen, J. S., and Stern, R. D., "Overview/
Environmental Protection Agency NO Control Technology for
Stationary Combustion Sources." Presented at 67th Annual
Meeting of the AIChE, December 1974.
11-238
-------
1:30 p.m.
The Optimization of Burner Design Parameters
To Control NO Formation in Pulverized
Coal and Heavy Oil Flames
Dr. Michael P. Heap, Ultrasystems
Mike, I didn't understand when you said that you
added methane. When you added it the NO went up?
If you were to take an average value of what you
would expect, if you were just to prorate it, you
would find out that you didn't get any advantage
by having the gas and the coal fired together.
And you could see that instead of methane if you
put in blast furnace gas, which did also contain a
small amount of methane, that- the emissions were
lower than what you would have expected from just
the algebraic sum of the two. I believe it has
something to do with the heating rate of the
particles.
That was what was bothering me. Because 1 had
just an intuitive feeling that if you nave a fast
heating rate and get the stuff hot in the absence
of oxygen you can get the nitrogen to combine to
molecular nitrogen N?.
In fact, the burner situation was such that you
mixed the coal very efficiently with the air
anyway. You are just driving off more fuel nitrogen
where the oxygen was available. The same kind of effect
appears to happen if you increase the temperatures
considerably.
H-2'39
-------
-------
PILOT SCALE INVESTIGATION OF COMBUSTION
MODIFICATION TECHNIQUES FOR NO CONTROL
IN INDUSTRIAL AND UTILITY BOILERS
by
R. A. Brown
C. B. Moyer
H. B. Mason
Aerotherm/Acurex Corporation
Mountain View, California
D. G. Lachapelle
Environmental Protection Agency
Research Triangle Park
North Carolina
11-241
-------
PILOT SCALE INVESTIGATION OF COMBUSTION
MODIFICATION TECHNIQUES FOR NO CONTROL
IN INDUSTRIAL AND UTILITY BOILERS*
R. A. Brown, C. B. Moyer, H. B. Mason
Aerotherm Divlsion/Acurex Corporation
Mountain View, California
D. G. Lachapelle
Environmental Protection Agency
Research Triangle Park
North Carolina
INTRODUCTION
As part of a continuing effort to provide better NO control for
stationary sources, the EPA has conducted supporting laboratory experi-
ments to define NO formation mechanisms, to identify control tech-
X
niques, and to guide development work on commercial equipment. The
EPA facility described here is an important component of this support
effort. Unlike most laboratory units, it provides a fairly realistic
modeling of the geometry and aerodynamics of large multi-burner boilers.
It can fire a wide variety of fuels, including pulverized coal, and
can simulate almost all combustion modifications for NO control:
x
This project was funded at least in part with Federal funds from the
Environmental Protection Agency under contract number 68-02-1503,
68-02-1318, and 68-02-1861. The content of this publication does not
necessarily reflect the views or policies of the U.S. Environmental
Protection Agency, nor does mention of trade names, commercial pro-
ducts, or organizations imply endorsement by the U.S. Government.
11-242
-------
load changes, excess air control, staging, biasing, flue gas recircu-
lation, reduced air preheat, and burner modifications.
The important goals for this facility will include:
• Further definitions of staging and biasing in wall firing of
high nitrogen fuels (coal and oil) to provide a more complete
mapping than presently available of the effects of first and
second stage stoichiometries, residence times, and mixing
rates, so as to
(a) Guide experiments of full scale equipment toward the
most effective retrofit technology
(b) Point the way toward second generation or optimized KOX
control including advanced combined techniques, for new
designs
• A duplicate set of goals for tangentially fired units, but with
the additional topic of burner modifications, an aspect not
being fully investigated elsewhere
• Exploration of NOX control techniques for mixed fuel firing
(such as high and low nitrogen fuels) and for waste fuels
• A quick look at the emission levels of other pollutants such
as sulfates and nitrates, as influenced by NOX controls
This paper describes the physical features and capabilities of
the EPA research test furnace facility and the tentative test plans to
achieve these goals.
II-243
-------
FACILITY DESCRIPTION
Figure 1 shows the furnace facility. The multifuel/multiburner
furnace is one of the most versatile pieces of equipment of this size
and type in the United States. Table 1 lists the components and capa-
bilities relevant to the research program, and the ranges over which
parameters may be varied.
The furnace may have both front-wall or opposed-wall firing, with
from one to five burners in each wall. Alternatively, the furnace
can accept tangential burners to simulate the design practice of Com-
bustion Engineering. The combustion chamber can run with or without
internal cooling tubes.
The front wall fired burners are in two sizes, 300,000 Btu/hr
(five each) and 1.5 x 106 Btu/hr (two each). These burners, shown in
Figure 2, are patterned after the IFRF design, with variable swirl and
provision to change many other parameters. Eight CE type burners,
shown in Figure 3, are also provided. These burners are capable of
variable tilt (±30°), variable yaw (±10°), have interchangeable air
sleeves and fuel nozzles and are capable of being fired on gas, oils
or pulverized coal. The relative position of the two banks of corner
fired burners is the same as the Combustion Engineering burners.
Four identical heat exchangers designed to cool a heat input of
three million Btu/hr to 400°F are stacked atop the main combustion
II-244
-------
1. Combustion chamber (39" cube)
2. Ignition and flame safeguard
3. Observation ports
4. Ashpit
5. 1.5 x 106 Btu/hr I FRF burner
6. C.E.-type corner fired burners
7. 3200°F refractory
8. Heat exchange sections
9. Drawer assemblies
10. Staged injection ports
0
Figure 1. Experimental multiburner furnace.
11-245
-------
TABLE K PRINCIPAL COMPONENTS AND CAPABILITIES
Component
Description
Main Furnace Combustion Chamber
Max. refractory temp: 3200°F
Volume: 47 ft5
View ports: 5, 3" dia., 1 10" x 10"
bottom with periscope
Ignition ports: 6, 1" dia.
1,5 hole, 28" dia.
1, single hole, 28" dia.
8, corner, "x"
1-5 horizontal opposed
1-5 wall fired
4, 8 tangentially fired
Burner blocks
and plugs:
Burner mounting:
Ash Pit
Volume: 8 ft3
Max. temp: 2600°F
Heat Exchangers
Sections: 4 refractory lined
Max. temp: 3000°F
Inside dimensions: 25x25x32" high
Drawers: 24 with 20-5/8" b tubes/DWR -
removable
Length of Dwr.: 32"
Coolant: Dowtherm
Access ports: 4/section, 1" dia.
Mixing section: 6" above and below
drawers (access ports
are located in those
sections)
Max. heat abs.: 2.2 x 106 Btu/hr
Wt.: 1500#/section w/o drawers
Burners
2 - l.S x 10s Btu/hr Aerotherm/IFRF
5 - 300,000 Btu/hr Aerotherm/IFRF
• Interchangeable fuel tips
• Interchangeable quarls
t Variable swirl
• Air sleeves to change velocity
8 - 375,000 Btu/hr Aerotherm corner fired
• Three identical circular air &
fuel inlets/burner
• ±30° tilt - all inlets ganged
• ±10° yaw - all outlets ganged
t Interchangeable air sleeves for
each port
t Interchangeable fuel nozzles
Air Supply
Primary
• 800 SCFM 6 8 psig
• Aftercooler: 70°F dew point @ 8 psig
• Cold control valve & orifice; separate
heater
Secondary
Hot control valves & orifices
11-246
-------
TABLE 1. (Continued)
Component
Description
Individual control 4 measurement to
16 lines, 8 on each side of the fur-
nace. Allows flow control of second-
ary air to each I FRF burner and con-
trol of annular and secondary air
flows to the corner fired burners.
Staged Air
Staged air manifold parallel to heat
exchanger stack. Mixes hot secondary
and cold secondary air to achieve any
temperature up to the secondary air
temperature. Presently only total
staged air is controlled.
Heaters
Secondary air 200 kw max.
Temperature at the burner: 800CF
Primary air heater 12 kw max.
Temperature at the burner: 250°F
Continuous control from 70°F to the
maximum temperature for 10:1 flow
range
Flue Gas Recirculation
Take off point downstream of baghouse
Max. temperature: 400°F
Max. flow: 120 SCFM
Max. pressure: 2 psig
Max. firing rate permissable at these
conditions: 1.5 x 106 Btu/hr @ 10'*
excess air
Present introduction point is in the
secondary air line downstream of cne
secondary air heater. (Simple modi-
fication could be made to introduce
the flue gases in the stage air, pri-
mary air or individual burners.
No FGR heater at this time
Oil Delivery System
• Up to 25 gal/hr on #2 or #6 oil
• Single pumping & supply system for
both oils
t Max. temperature 46: 220°F at the
nozzles
• Two oil manifolds with 8 taps each on
either side of the furnace
• Quick disconnect fittings at the
manifold and burners
• Flow control valves to each burner
• Max. pressure: 250 psig
II-247
-------
TABLE 1 (Concluded)
Component
Description
Gas System
• Up to 3000 ft'/hr @ 25 psig
• Manifold with quick disconnects, 8
outlets on each side of the furnace
t Shutoff ball valves for each tap on
the manifold and needle control
valves for each burner inlet
Coal System
l up to 250 Ibs/hr of pulverized coal
• Ten delivery lines to two manifolds.
one on each side of the furnace.
Five flexible lines on each manifolc
deliver pulverized coal to one to
five burners. The small lines must
be recombined when firing the large-
burners .
• The coal and primary air flow rates
are controlled and measured to each
of the delivery lines. Uniform dis-
tribution is obtained from a fluid-
ized bed distributor.
• Bagged pulverized coal is fed into ;
bagdump from the seconc floor level
and Into a 50 ft3 hopper. This rep-
resents about a 1 day supply of coa".
• The fuel flow may be stopped in the
event of a flame-out or unstable co--
ditlon through a solenoid operated
air purge system. This purge syster
is controlled manually or by the
flame safeguard systeir.
Dowtherm System
Two Dowtherm-to-air heat exchangers
can remove up to 2.5 x 10s Btu/hr
from the Dowtherm
A bypass arrangement around these
coolers allows control of the heat
removal rate.
Induced Draft Fan
An induced draft fan with bypass
allows control of the hack pressure
in the combustion charter to ±2"
H20 over the full range of firing
rates.
11-243
-------
o>
c
a;
u.
-------
Vj
~— -1
^
y.\v
h^
J
s-
3
J3
i.
0)
c
o
o
11-250
-------
chamber. Each exchanger consists of six drawer assemblies with twenty
stainless steel u-tubes for easy removal of heat exchange surface.
Refractory plugs have also been provided to fill the drawer windows
when the heat exchange surface is removed.
Figure 4 shows the relationship between heat release rate per
unit volume and firing rate, and indicates typical industrial design
practice for heat release per unit volume. The upper boundary is
determined with the full heat exchanger surface installed (or using
the main firebox as the combustion volume). The lower boundary is de-
termined by the required heat exchange surface to lower the gas tem-
perature to 400°F. A wider performance map could be obtained if
additional heat exchange surface was added in the crossover duct or if
the baghouse was bypassed. Water cooling could also be used in the
main heat exchangers, requiring less heat transfer surface and thus
allowing longer residence times. Table 2 further illustrates the cur-
rent performance of this facility by showing the bulk residence time,
and heat release rates for various firing rates.
Figure 5 shows the general layout of the facility in Aerotherm's
laboratory.
RESEARCH PROGRAM
The 28 month operating contract for the EPA furnace is subdivided
into the following phases:
• Phase I: Installation and Shakedown (four months)
11-251
-------
0-
o
«d-
-------
TABLE 2. FURNACE PERFORMANCE PARAMETERS FOR
VARIOUS CONFIGURATIONS
Configurator
( Furnace
plus n empty
sections)
Furnace
Furnace + 1
Furnace + 2
Furnace + 3
Vol
ft3
46.60
57.27
67.93
78.60
Total Heat Release (Btu/hr
3 x 10'
R.T.
0.67/0.83
X
><
K.R.
64,378
X
kX
x
1.5 x 106
R.T.
1.34/1.68
1.65/2.06
1.96/2.45
H.R.
32,189
26,191
22,068
1.0 x 106
R.T.
2.01/2.51
2.47/3.09
2.93/3.66
3.39/4.24
H.R.
21,459
17,461
14,721
12,723*
KEY:
R.T. = Residence time (sec) at 25 percent excess air/stoichiometric
H.R. = Volumetric heat release rate Btu/hr-ft3
£>
-------
j .
_Q»L_-_.U_f
V)
i
•M
v>
5?
8.
Q.
O
8
o
I
s-
3
2
0)
s_
11-254
-------
• Phase II: Conventional Fossil Fuels Test Program (12 months)
• Phase III: Mixed and Nonconventional Fuels Test Program (12
months)
The Phase I shakedown tests served to certify the system design and
also yielded the operational and control characteristics needed to
execute the subsequent test program.
The objective of the Phase II test program is to develop advanced
techniques for the control of combustion generated air pollution for
application to large and intermediate sized steam generators firing
conventional fossil fuels. For this series the furnace will be run in
the single wall fired mode using five 300,000 Btu/hr burners and in
the corner fired mode using four or eight burner arrays. Emphasis
will be on fuels with high fuel NO forming potential — coal and re-
sidual oil — with secondary emphasis on distillate oil and natural
gas. Western Kentucky #9 coal and Chevron bay area residual oil (S =
1%, N « 0.8%) have been selected as the primary fuels. Two other coals,
Pittsburgh #8 and Big Sky Montana, and two other residual oils will be
fired periodically to generalize the results obtained with the primary
fuels.
The Phase II test plan is shown schematically on Figure 6. The
initial furnace characterization studies will define the system oper-
ating conditions and baseline emissions to be used to interpret the
11-255
-------
Phase II
Fossil Fuel-
Program
—Furnace ____
Characterization
^_ Pr«>1 iminary . . ,___.
Studies
Baseline
Emissions
Control
Development
1 — Wall Cooling
Wall fired
Mall Cooling
Corner Fired
Burner Array
Corner Fired
| — Wall Fired
Corner
Fired
Fuel NOX
Wall Fired
Fuel NOX
Corner Fired
1st Stage
Variables -
— Coal
2nd Stage var-
iables -on
2nd Stage Var-
iables - Coal
2nd Stage Var-
iables - Oil
Combined Con-
trols - Coal
, Combined Con-
trols -011
r" •"»!•»
Burner
— Parameters
Flue Gas
Recirculation
Combined
Controls
II-A
1I-B
II-C
II-D
II-E
Il-F,
II-F2
II-H,
II-H2
n-J1
II-J2
II-Kj
II-K2
II-L
II-M
Il-N
II-P
II-Q
20
Points
(20)
(4)
(87)
(97)
(15)
(15)
(42)
(42)
(21)
(21)
(65)
(65)
(23)
(50)
(46)
(20)
(25)
Figure 6, Summary of Phase II test sequences.
H-256
-------
results of the control development studies. The preliminary wall
cooling studies, II-A and II-B, will indicate the level of wall
cooling needed to give the best simulation of NOX formation in large
steam generators. The determining factors will be the level of NO
as well as the trend with excess air, preheat and load for at least
two levels of wall cooling. Gas and coal will be fired to encompass
the range of thermal/fuel NO and the range of luminous versus gas
radiation encountered in practical equipment. The wall cooling scheme
established in II-A and II-B will be carried through all remaining •:
test series. The third preliminary study, II-C, will determine if
four or eight corner fired burners yield the best simulation of NO
formation in full scale equipment.
The uncontrolled baseline emission series II-D and II-E will
provide reference emission levels over the range of preheat, excess
air, load and fuel type to be considered in the control development
series. The baseline series will also give further characterization
data to indicate the correspondence of this facility to full scale
equipment. An additional baseline series, II-F, will be run to elu-
cidate the thermal NO /fuel NO split as affected by fuel type, pre-
A X
heat and excess air. Concurrent tests at the University of Arizona
In which argon is substituted for nitrogen in the combustion air will
indicate the effect of flue gas recirculation on thermal NO and
X
thereby show the fractional conversion of fuel nitrogen to NO. These
11-257
-------
results will be translated to the EPA/Aerotherm furnace run with a
duplicate matrix and the same fuel to give semi-quantitative guide-
lines on fuel nitrogen conversion.
The control development test series will be largely devoted to
determining the best procedures for NO control through staged com-
bustion. The wall-fired control development series will involve a
systematic screening of primary first stage variables and second
stage variables followed by combined testing of the best procedures
identified in the screening study. The first stage screening study
will investigate the following variables:
• Primary flame zone mixing due to burner swirl
• Primary flame zone stoichiometry
* First stage flame heat removal
• Bulk residence time to the point of second stage air injection
• Burner air preheat
The NO control results from these tests will be interpreted through
X
the time-temperature history of the combustion gases up to the point
of second stage air injection as affected by the above variables.
Since the time-temperature histories and combustion intensities in
this facility are similar to full scale equipment, these results will
serve to establish semi-quantitative guidelines for staging in field
application.
II-258
-------
The second stage screening study, II-J , will investigate the
following variables:
• Residence time from second stage air injection to quench
• Stage air preheat
• Rate of mixing of stage air with combustion gases
These results will compliment the II-H results to indicate the best
overall approach to staging which will be tested in series II-K.
Throughout the staging studies, measurement will be made of combustion
gas temperature, local NO concentration and residence time distribu-
X
tion by means of the suction pyrometer, hot sampling probe and helium
tracer measurements respectively. These measurements will provide
insight into how staging variables affect N0x formation and control
and will also allow some generalization of the results toward full-
scale application guidelines.
Flue gas recirculation for NO control with dirty fuels, pri-
marily residual oil, will be studied on a limited basis in series
II-L. The objectives are (1) determine the effectiveness of FGR as a
function of fuel type, excess air and preheat, and (2) illustrate
qualitatively how fuel NO conversion is affected by primary system
A
variables. These FGR tests will actually be run in advance of the
staging studies. If the FGR technique shows promise, it will be In-
cluded in the combined control optimization series, II-K.
11-259
-------
The corner fired control development series will extend the
staged combustion guidelines established in the wall fired series to
the corner fired configuration. Additionally, the corner fired series
will study the effect of burner parameters on NO formation and con-
•X
trol. Series 11-M provides for study of the following corner fired
burner configurations:
• Burner firing orientation
• Eccentric fuel/air injection
• Biasing of burner stoichiometry between the upper and lower
nozzles
These tests will identify low NO,, configurations and will indicate
Jt
the effects of hardware settings on fuel/air mixing in the jets and
in the central fireball. The observations on fuel/air mixing will
guide the selection of burner configuration to be used in the sub-
sequent studies on staging, FGR and combined controls.
The Phase III mixed and nonconventional fuels test program has
the objective of defining practical design guidelines for emission
control with the firing of nonconventional fuels. Mixed fuel firing
shows promise as a new and powerful NO control option and so will
X
tentatively occupy eight months of the 12 month test program with
11-260
-------
nonconventional fuels covering the remaining four months. Mixed fuels
approaches considered for Phase III include:
• Blends of fuels to achieve low net fuel N
• Simultaneous firing of two fuels through two separate sets
of burners to provide a shielding of the high N fuel and
hence a delayed exposure of fuel N to combustion air, sus-
pected to be an effective control technique for fuel NO
X
• Simultaneous firing of two fuels through separate elements of
the same burner to provide a shielding effect on the scale of
an individual burner*
• Staged fuel addition, or "reburning," to reduce NO in a sec-
ondary combustion zone
Table 3 lists the candidate fuels for both individual firing and
dual fuel firing, and briefly indicates the selection rationale in
each case. The individual tests are limited to important potential
boiler fuels: oil shale product, solvent refined coal, coal char,
and one high nitrogen waste fuel. The particular selections are in-
fluenced by the prototype characters of these fuels as being repre-
sentative of general classes of new fuels and as having interesting
NO control difficulties. The only conspicuous absence from this
X
single-fuel-firing list are synthetic fuel gases produced from coal.
*For wall firing, other EPA programs currently are developing this
technology. Consequently, the stress here will be on tangential firing.
II-261
-------
o
1-^
<
CQ
O
CD
U. UJ
Q i— i
LU
GO
a: a.
o
a. c£.
SO
(U
Li_
5?
»^-
O
S-
ns
.c
<_>
m
o
o
O
Ol
(O
.c
<^
-o
(/>
s
p**
-------
ACKNOWLEDGEMENTS
The contributions of D. W. Pershing and Professor J. 0. L. Wendt,
of the University of Arizona, to the formulation of the test plan is
gratefully acknowledged.
11-265
-------
TABLE OF CONVERSION FACTORS FOR SI UNITS
Equivilant SI Units
0.0254m
0.3048m
0.028317m3
0.45359Kg
6895 N/m2
0.003785 m3
1055.06
37,259 J/m3
(Tf-32) /1.8°C
English Units
inch
Ft
Ft3
1b
Psi
Gallons
Btu
Btu/Ft3
°F
H-266
-------
2:25 p.m.
Pilot Scale Investigation of
Combustion Modification Techniques
for NO Control in Industrial and
Utility Boilers
Richard E. Brown, Aerotherm Acurex
In noticing most of the presentations of yesterday
and today, the effect of wall temperatures on NO
X
emissions just seems to be one of the most pronounced
effects. Now I noticed how with your U-tubes you can
affect a cooling rate of the flue gases, but they are
not down in the furnace. By the time the gases are
going up through those cooling tubes there is going
to be no back mixing into the flame zone. Maybe I
am wrong, but I see that this furnace is compared to
what everyone else seems to see as the most important
parameter, that's wall temperature -- and I don't
see how we can work that into...
We will be putting cooling walls in this furnace,
and vary the spacings to look at various cooling
wall patterns. That is part of our base line test
series.
Will you be testing two-stage combustion on a single
burner setup?
Perhaps. Burner configuration will be determined during
our baseline series and we think it will probably be the
five burner array. Most likely we will not look at
stagedcombustion for the single burner until we go
to some horizontal extensions. This furnace is also
designed such that we can go to horizontal extensions,
either end. And in that case we may look at two-staged
combustion with the single burner.
II-267
-------
Q: The word "shielding," could you interpret that please?
A: The effect of introducing fuel and air in either a
fuel-rich or fuel-lean mode, and providing another
fuel to shield the main fuel, say, from the oxygen.
Q: What method are you going to use to preheat your air?
A: We are preheating the air with electrical heaters up
to 800° F.
Q: And on the flue gas recirculation, are you taking it
directly from the stack?
A: We are taking it downstream of the baghouse; we are
taking the particulate and the fly-ash out.
Q: And where are you reinjecting it?
A: We will reinject it at this time into the secondary
air streams, although we have the ability to inject
it at any location.
II-268
-------
SESSION III
PROCESS RESEARCH AND DEVELOPMENT
D. G. Lachapelle, Chairman
III-l
-------
-------
OVERFIRE AIR AS A NO CONTROL
x
TECHNIQUE FOR TANGENTIAL COAL-FIRED BOILERS
By:
A. P. Selker
Combustion Engineering, Inc.
Windsor, Connecticut
III-3
-------
OVERFIRE AIR AS A NO CONTROL
x
TECHNIQUE FOR TANGENTIAL COAL-FIRED BOILERS
A. P. Selker
Combustion Engineering, Inc.
Windsor, Connecticut
INTRODUCTION
The purpose of this EPA sponsored program was to determine the
effectiveness of overfire air (OFA) and biased firing* as methods for
NOX control on tangentially coal fired utility boilers. In
addition, the effects of these emission control techniques on overall
unit performance were evaluated. The program was divided into two
contract phases:
Phase I (EPA Contract 68-02-0264) consisted of selecting a suitable
utility boiler to be modified for the incorporation of an overfire
air system and conducting studies in preparation for the actual unit
modifications and test program. This phase was completed in 1973.
Phase II (EPA Contract 68-02-1367) consisted of modifying and testing
the utility boiler selected in Phase I to evaluate overfire air and
biased firing as methods for NOX control. This phase also included
the completion of detailed fabrication and erection drawings, instal-
lation of analytical test equipment, updating of the preliminary test
program, and analysis and reporting of test results.
This program was conducted with the cooperation and active participa-
tion of Alabama Power Company at their Barry Steam Station Unit 2.
This unit is a natural circulation, balanced draft utility plant
boiler firing coal through four elevations of tilting tangential fuel
nozzles. Unit capacity at maximum continuous rating (MCR) is 408,000
kg/hr main steam flow with a superheater outlet temperature and pres-
sure of 538 C and 131.8 kg/cm^. Superheat and reheat temperatures
are controlled by fuel nozzle tilt and spray desuperheating. A side
elevation of the unit prior to modification is shown in Fig. 1.
OVERFIRE AIR SYSTEM DESIGN
The overfire air system (Fig. 2) incorporated in the unit provided
for introducing a maximum of 20% of the total combustion air above
the fuel admission nozzles at full unit loading. The overfire air
was introduced into the furnace tangentially through two separate
*Biased firing consists of removing individual fuel elevations from
service while maintaining full airflow through these same elevations.
III-4
I
-------
'' ^''
Fig. 1: Test unit side elevation, Alabama Power Company, Barry 2
III-5
-------
F- FUEL AND AIR
A-AIR
0-OVERFIRE AIR
Fig. 2: Test unit overfire air system schematic
compartments near each furnace corner located approximately 2.4 meters
above the fuel admission zone. While on new unit designs the over-
fire air ports (Fig. 3) are designed as vertical extensions of the
corner windboxes, on the field modified unit, separate compartments
were required due to structural considertions. The system provided
for air nozzle tilting (-30 deg. from vertical plane) and separate
compartment flow control dampers to permit studying the effects of vari-
ous airflow rates, introduction angles, and compartment airflow dis-
tributions.
III-6
-------
WINDBOX
SECONDARY AIR
DAMPERS
SECONDARY AIR
DAMPER DRIVE
UNIT
OVERFIRE AIR
NOZZLES
SIDE IGNITOR NOZZLE
SECONDARY AIR NOZZLES
COAL NOZZLES
HIGH SWIRL OIL GUN
Fig. 3: Corner windbox showing overfire air system
TEST INSTRUMENTATION
In order to evaluate the effect of using overfire air as a combustion
process modification technique for emissions control, a grid of stain-
less steel flue gas sample probes was installed on the unit's econo-
III-7
-------
mizer and air preheater outlets as shown in Fig. 4,
Fig. 4:
A. ECONOMIZER OUTLET
B. AIR HEATER OUTLET
(Oj ONLY)
STAINLESS STEEL FLUE GAS
SAMPLE PROBE LOCATIONS
Schematic showing location of sample probes
The gas sample from the economizer outlet was blended, transported,
and preconditioned prior to analysis by the following instrumentation
(Fig. 5).
1. A Scott chemiluminescence NO-NOX analyzer (0-2000 PPM)
2. An L&N paramagnetic 02 analyzer (0-25% 02)
3. Beckman non-dispersive infrared CO analyzer (0-1000 PPM)
4. A Scott flame ionization Total Hydrocarbon (THC) Analyzer (0-100
PPM)
A schematic of the gas analysis system is shown in Fig. 6.
Unit performance was monitored using the instruments and analytical
procedures shown in Table I.
TEST PROGRAM
The Phase II test program called for the evaluation of overfire air
as an emissions control system. Also evaluated were existing process
variables, including excess air, unit loading, and air/fuel distri-
bution.
To properly assess the effect of the overfire air modifications on
unit performance and emission levels, the test program was divided
into two distinct phases.
1. The first phase, run prior to modification, consisted of baseline
III-8
-------
TABLE I
INSTRUMENTS AND ANALYTICAL
PROCEDURES FOR MONITORING UNIT PERFORMANCE
PARAMETER
Flow rates
Temperatures
Pressures
Laboratory analysis
Steam and Water
Feedwater flow
Reheat and superheat
desuperheat spray
Reheat flow
Air and Gas
Total Air and Gas Weight
Overfire air
Air heater leakage
Steam and Water Deg. F
Unit absorption rates
Waterwall absorption
Air and Gas Deg. F
Steam and Water PSIG
Unit absorption rates
Unit draft loss
Temperature and Pressure
Fuel and Ash
INSTRUMENT/ANALYTICAL PROCEOURJ
Flow orifice
Heat balance (deg. F & PSIG
around desuperheater
Heat balance around reheat
extraction and estimated
turbine gland seal losses
Calculated
Pitot traverse
Paramagnetic 0. analyzer
Calibrated stainless steel
sheathed CR-C well & butti
TC's
Calibrated stainless steel
sheathed CR-C chordal WW '
CR-C TC's
Water cooled probes
Pt/Pt-IOZ Rh TC's
Pressure gauges and/or
transducers
Water manometers
C-E data logger
capacity: 400 tentperatur
50 pressures
ASTM procedures
-, 9
-------
Fig. 5: Gaseous emissions test system
111-10
-------
JUL
SAMPLE PROBES
GAS DUCTS
-TTTTn-fTrTfHii m iTi
.1.1. JL
SAMPLE PROBES
SAMPLE BLENDER
PRECONDITIONING
CONSOLE
FILTER p=
T~
PUMPQ
^^
1
1
3 r-0-i
COND
*H2O
H — h n
QD
i
3RYER 1
(~U J 1
'
H20_T__
r
1 1 FILTER
THERMAL
coNvl—l _f4
uumv. ij 0_ | t F|LTER
I
*H2° O PUMP
1 n
1 °2 D —
1 NO T
|HUx |_ 1
|L
(3 VACUUM PUMP
CONSOLE 1
CONSOLE 2
Fig. 6: Gas analysis system
and biased firing programs as well as 30-day waterwall corrosion
coupon studies at baseline and optimum biased firing conditions.
2. The second phase, conducted after unit modification, consisted of
repeating the baseline evaluation (to determine if the modifica-
tion had influenced unit performance) and testing the overfire
air system. A 30-day waterwall corrosion coupon evaluation was
also conducted at the optimum overfire air condition.
The originally planned evaluation of optimum operations with alternate
coal types could not be performed due to difficulties in obtaining
the necessary coal supplies. However, this information has been re-
Hi- 11
-------
ported by others as a result of tests on another unit at the same
station.'-1)
RESULTS
The first phase of the test program characterized the baseline emis-
sion levels and unit thermal performance including the effects of
various operating variables including unit load, excess air level,
and furnace wall deposit variation. In addition, tests were conducted
to evaluate the effect of biased firing on NO emissions.
Load and Excess Air Variation
Figure 7 shows that the NOX emission levels increased with increased
excess air but did not change significantly with changes in unit loading.
BASELINE TESTS BEFORE MODIFICATION
1,5
; 4
1 3
1SPS
1 2
1 Q
0 9
n fi
X»
/
4
/
P^
Tl
,
^
/
r
^
yS
^,
*~ —
4
• •
^
LEGEND
Unit LMjl Furiact Sin
Q HC» OL19«
Q 3/4 «CR O Noderat*
O vz MCR i"«»y
no i?o no no iso
THEORETICAL AH TO FUEL FIRING 2OIE, PtRCENT
160
Fig. 7: K02 vs theoretical air to fuel furing zone — baseline study
An average increase of 0.017 g/10 cal NCL was noted for each 1% change
in unit excess air. Acceptable CO limits of 0.05 g/10^ cal establishes
minimum excess air levels of 20 to 25% (Fig. 8). (Throughout this re-
port, NO emission levels are expressed as g/10" cal NO-.)
X £*
On these plots, excess air (EA) is expressed as percent theoretical
air (TA) to the fuel firing zone which, for unit operation without
111-12
-------
BASELINE TESTS BEFORE flDDIFICATION
0.6
0.5
•» 0.3
s
110 IJO 110- 140 150 1*0
THEORETICAL All! TO FUtL FIRING ZONE. PERCENT
Fig. 8: CO vs theoretical air to fuel firing zone — baseline study
overfire air, can be considered to be directly comparable (20% Total
EA - 120% TA).
The excess air and load variation tests were conducted at what were
considered to be clean furnace conditions to establish a baseline for
wall deposit variation tests. Furnace wall deposit variations were
determined by visual observations by C-E test personnel.
FurnaceWall Deposit Variation
Changes in furnace slagging did not exhibit a discernable effect on
NOX or CO emission levels (Figs. 7 and 8) with the exception of two
half-load tests when higher CO levels were obtained at slagged as com-
pared to clean furnace conditions. The reasons for this condition are
not completely clear and the higher CO levels were not observed under
full load heavy slag operation.
Maximum slagging conditions were obtained by operating the test unit in
excess of 24-hour periods without operating any wall blowers. In this
manner, slag deposits up to 4-inches thick could be obtained in and
above the firing zone. Wall deposit variations were again determined
III-13
-------
by visual observation.
The slagging conditions indicated should not be considered representa-
tive of normal operation at the various excess air and unit load
levels studied, but were specifically controlled for test purposes by
varying soot blower operation.
Biased Firing Variation
Figure 9 shows the results of removing individual fuel elevations from
service and operating them on an air-only basis. Maximum NOX control
was obtained with top elevation of fuel nozzles out of service at max-
B1ASED FIRING TESTS BEFORE MODIFICATION
1 3
ISPS
1 2
1 ^
IfS
06
07
Of
OC
»
>
— r
0
*/
;
i
/
/
00
f
/
1
A
/
10
s^
O
i
O
^
20
•""
13
,-*•
0
---•
1
I
-"
•
4D
li
Unit iMd
O N» Pots.
(B 3/4 NCR
0 1/2 NCR
Fuel Nozzles
Out af Sent.
Fig. 9:
THIORETICAl AIR TO FUEL FIRMS ZONE, PERCENT
NO, vs theoretical air to fuel firing zone — biased firing
study
imum and 75% maximum loading.
At 50% percent maximum loading, the high excess air levels required to
maintain unit steam temperatures appeared to negate any NOX reductions
obtained by biasing the top fuel nozzle elevation; however, the emis-
sions level obtained was below the current EPA limit (NSPS*) for coal
fired units of 1.26 g/106 cal.
It should be noted that when any of the lower three fuel elevations
were biased (out of service), the air introduced through the inactive
fuel compartment was included in the percent TA to the fuel firing
*New Source Performance Standard
III-14
-------
zone.
As can be seen from Fig. 9 biasing any of the bottom three fuel
elevations did not have a discernable effect on NOX emission levels,
although the emission level tended to be higher at reduced unit load-
ings for given percent TA levels.
Figure 10 indicates that with biased firing, low percent TA levels to
the fuel firing zone were obtained without increasing CO emission
levels. This condition is due to the ability to maintain acceptable
BIASED FIRING TESTS BEFORE MODIFICATION
0.3
f-i
0.1
0
90
d
A
I
*
9
•
•
100 110 \?0 130 140 15
LEgHP
Unit Load
Fuel Nozzles
Out of Str».
Mii. Pass.
3/« NCR
^n nc«
Top Ctr.
llol. Ctr.
Sot to*
THEORETICAL AIR 10 fUCL FIRING 2011. PERCEKT
Fig. 10:
CO vs theoretical air to fuel firing zone — biased firing
study
total unit excess air levels as measured at the economizer outlet
during biased firing operation.
The second phase of the test program was conducted to determine base-
line operating characteristics of the modified unit and to compare
these to the unmodified unit test results. The overfire air system
optimization program then investigated the effects of varying over-
fire air locations, tilt angles, and introduction rates with respect
to emission levels and unit thermal performance.
Load and Excess Air, and Furnace Wall Deposit Variation (After Modifi-
cation)
The baseline test program conducted prior to unit modification was re-
peated after installation of the overfire air system with similar re-
Ill- 15
-------
suits with respect to NO^- and CO emission levels being obtained as
shown in Figs. 11 and 12. The NO levels after modification
BASELINE TESTS AFTER MODIFICATION
1.3
NSPS
1.2
1.1
8 '-0
%
o. 0.9
0.8
0 ?
0.6
0.5
X
X
9
X
B°
**
4
^
•^
<
X
V
X
•^
X
x
.
/
*s
U6EHO
Unit Lead Furntcc SIM
O NCR O Light
O 1/4 NCR 0) NMtertU
O UZ NCR « HN«y
100 110 1ZO 130 140 ISO
THEORETICAL MB TO FIRING ZONE. PERCENT
Fig. 11: N02 vs theoretical air to firing zone, overfire air study,
load and excess air variation — modified boiler
BASELINE TESTS AFTER MODIFICATION
0.4
,- 0-3
g 0.2
ai
0
10
k
(1^
O
p
x
& t
^
t> 4
r~~-a
~Q-~
J
^
0 110 120 130 140 ISO 16
LEGtND
Unit LotJ Funiic* Sin
O NCR
a 3/4 NCR
O 1/2 NCR
0 Light
9 NMtrite
THtORCTtCU. AIR TO FUEL FIRING ZONE. PERCENT
Fig. 12: CO vs theoretical air to firing zone, overfire air study,
load and excess air variation — modified boiler
III-16
-------
tended to be slightly lower than pre-modification results; however,
this could have resulted from updating of the firing system (replace-
ment of fuel/air nozzles, etc.) during modification and a slight re-
duction in fuel N2 levels and higher heating values. Again, no dis-
cernable effect on NOX emission levels are noted due to furnace wall
cleanliness.
Overfire Air (OFA) Variation
As shown in Fig. 13, the effect on NOX levels of varying the OFA ad-
mission points was investigated with the result that NOX levels de-
creased with decreasing percent TA to the fuel firing zone,
0 V £ P. f I R E AIR LOCATION VARIATION
1.3
NSPS
1.2
1.1
3
io.a
0.7
0.6
0.5
0.4
0-3
|
^
,
^
- -
— — -
_ — -—
-
•"tS
w
•4 —
, -. 4>
3^-
r"
^
^-*
rip
—
Of*
1 ^**
t — •
- IOMIR LIMIT OF
11
H- *l
^— —
4
-""
,**
-~~~
ACCEPTABLE
vt^
A LEVE
•- •
^
^
LS
-
c
^-*
—
- •>
rt
80 85 90 95 JOO 105 110 115
LECtjID
UK. Pis.
A 0-1
& 0-2
0-3
tote
& No OF*
A 1/2 Man. OF*
A Man. OFA
ABC (§}
BCD (H)
THEORETICAL AtR TO FIRING ZOHE, PERCENT
N0? v^ Theoretical Air to Firing Zone, Overfire Air Location, Rate 4 Velocity Variation
Fig. 13: N02 vs theoretical air to firing zone, overfire air location,
rate and velocity variation
The importance of using firing zone percent TA levels now becomes
apparent as this procedure provides a mechanism for correlating test
results while total excess air levels as measured at the economizer
outlet provide no meaningful base for data correlation.
NOX levels were found to be generally higher with the top 3 elevations
of fuel nozzles in service than with the bottom three elevations. We
think this is due to the increased distance between the fuel firing
zone and the admission point for the OFA.
111-17
-------
Operation at TA levels below 95% did not result in significant reduc-
tion in NOX emission levels. As shown in Fig. 14, CO levels remained
acceptable for the duration of the test series. Total excess air as
measured at the economizer outlet was maintained at approximately 27
percent.
OVERFIRE AIR LOCATION VARIATION
0.4
0.2
LOHE8 LIMIT OF ACCEPTABLE TA LEVELS
LEGtM)
Mm. Pts
£ 0-1
t. 0-2
b o-i. 0-2
O 0-3
tote
& No OF*
A 1/2 MM. OFA
A N>x. OFA
mm in _5tr».
ABC ®
KO 151
M
95 100 105 110
THEORETICAL AIR TO FIRING ZONE, PERCENT
115
120
Fig. 14: CO vs theoretical air to firing zone, overfire air location,
rate and velocity variation
Overfire Air Tilt Variation
A test series was conducted at full unit loading to determine the
effect of varying OFA and fuel nozzle tilts on NOX and CO emission
levels. These tests were conducted with total excess air and TA
levels to the firing zone of 24 and 92 percent, respectively, and mod-
erate waterwall slagging conditions.
Essentially, varying fuel/air and OFA nozzle tilts moved the firing
zone both in the furnace and relative to the overfire air. Fuel
nozzle tilts that are maximum minus combined with OFA tilts of maxi-
mum plus increased the distance between the overfire air and the fir-
ing zone. As with previous methods of increasing this distance, the
NOX emission levels were decreased. Figure 15 shows that as the tilts
were moved toward one another (fuel nozzle tilts up, OFA tilts down),
the OFA/firing zone separation was decreased and the NOX levels were
increased.
When the OFA tilts were maximum minus and the fuel nozzle tilts maxi-
mum plus, the term overfire air became ambiguous and the actual over-
fire air was less than the calculated value, as the air was being
III-18
-------
OVERFIRE AIR TILT VARIATION
0-9
0.8
0.6
60
SO 40 10 20 10 0 10 20 30 40 SO SO 70
TOWARD EACH OTHER AMAY fdOM EACH OTHER
OFA TILT AND FUEL NOZZLE TUT A . DEGREES
Fig. 15: N02 vs OFA tilt and fuel nozzle tilt differential, OFA tilt
variation
forced down into the raised firing zone. At this point, where the
combined fuel nozzle and OFA tilt differential was 52 degrees toward
each other, the NOX emission level reached a maximum of 0.846 g/106
cal.
CO emission levels exhibited an increase as the tilt differential in-
creased; however, enough total excess air was available to maintain
an acceptable emission level as shown in Fig. 16.
Flame stability became a limiting factor as the tilts move substantial-
ly away from each other (40 - 50 degrees). NOX emission levels de-
creased to 0.596 g/106 cal; however, the CO emission levels increased
and the fire appeared less stable. Maintaining the fuel nozzle tilts
and OFA tilts at approximately equal tilt angles resulted in accept-
able flame stability as well as reduced NO emission levels.
x
For all OFA tilt variation tests, the NO emission level obtained was
well below the EPA limit of 1.26 g/106 cal.
Load Variation at Optimum Conditions
Tests were conducted to evaluate unit performance and emission levels
at optimum operating conditions as determined during the overfire air
variation tests. Tests were conducted over the unit load range at vary-
Ill- 19
-------
OVERFIRt AIR TILT VARIATION
_ 0.05
SO. 04
0.03
102
70 60 50 40 30 20 10 0 10 20 30 40 50 SO 70
TONARC UCH OTHCfl AHA* FROH EACH OTHES
OF* TILT AND FUEL IWZ2LE TUT A. DEGREES
Fig. 16: CO vs OFA tilt and fuel nozzle tilt differential, OFA tilt
variation
ing furnace waterwall slagging conditions. The NO emission level re-
sults of this series of test versus unit loading, expressed as main
steam flow, are shown in Fig. 17.
The wide range of N02 levels obtained, particularly during the base-
line tests are due to variations in unit operating parameters such as
excess air level. During the optimization tests, total excess air at
the unit economizer outlet was maintained between 20 and 28% at full
and 3/4 load and 45 to 47% at 1/2 load with fuel nozzle tilts varied,
as required, to maintain acceptable reheat and superheat outlet temper-
atures. Also, minimum excess air levels were established on the basis
of maintaining acceptable CO emission levels and flame stability.
Furnace Performance
During the test program, furnace performance was monitored by use of
chordal thermocouples installed in the furnace waterwalls. Tempera-
tures and corresponding absorption rates were found to vary signifi-
cantly with wall slag conditions making data interpretation difficult.
The method finally arrived at as representing an accurate indication
of furnace performance was as follows:
111-20
-------
BASELINE AND 0 P T I fl U !1 UNIT OPERATION
LEGEND
G Baseline Tests
A Optimization Tn
STI A" FLOW - 10J».r.,'HR
PERCENT Of fi'U IOA0 fiflTINO
Fig. 17: N0? vs main steam flow, ranges for normal and optimum opera-
tion
The front and right side wall centertube profiles were determined and
the average of these plotted as shown in Fig. 18. It should be noted
that the maximum and minimum profiles shown do not represent individual
walls in every case, i.e., at given furnace elevations the maximum
rate shown may switch from wall to wall.
For comparison of optimum and normal unit operation with respect to
furnace performance, three full load tests with similar furnace slag-
ging conditions, etc., were selected. The average centerline profiles
for these tests were determined and plotted together as shown in Fig.
19. Furnace performance remained essentially unchanged when furnace
slagging effects were taken into account.
Waterwall Corrosion Coupon Evaluation
Following completion of the steady state phases of the baseline,
biased firing, and overfire air test programs, 30-day waterwall cor-
rosion coupon evaluations were performed to determine whether any
measurable changes in coupon weight losses could be obtained for the
various firing modes studied. The individual probes were exposed at
111-21
-------
&
UJ
i
*
§
s
Date: 6/28/74
Load: 125 MM
Furn. Absorp.: 153.17 1
Total Absorp.: 276.4 1
TA to Fuel Firing Zone:
Total Excess Air: 25.9
KG-CAL/HR
KG-CAl/HR
119.9 J
O - Front HW Center Tube Profile
- Right HW Center Tube Profile
- Avg. Center Tube Profile - Both
. FUEL EL£V. A
<-'-; t-"-^
" FUEL ELEV. B
FUEL ELEV. D •
.,_ . i I
24 6 6 10 12 14 16 18
TUBE CROWN ABSORPTION RATE, KG-CAL/HR-CK2
20 22
24
26 28
Fig. 18: Average centerline absorption profile — full load operation
111-22
-------
g
—
8*
£ 18.29
d (60)
"2 4 6 8 10 12 14 16 18 20 22 24
TUBE CROWN ABSORPTION RATE, KG-CAL/HR-CM2
26 28
Fig. 19: Average centerline absorption profile for all tests — full
load operation with and without overfire air
111-23
-------
five locations on the furnace wall with coupon temperatures being main-
tained at the same levels for each 30-day run.
The individual coupon weights were determined before and after each
30-day test and the individual coupon and average probe weight losses
determined. The weight losses calculated for the biased and overfire
air portion of the test program were found to be greater than those for
the baseline test. The average weight losses for all five probes were:
Baseline
2.6381 mg/ctn
Biased Firing
2
4.6429 mg/cm
Overfire Air
4.4419 mg/cm
These values are within the range of losses which would be expected for
oxidation of carbon steel for a 30-day period. To verify this premise,
control studies were conducted in C-E's Kreisinger Development Labora-
tory using probes exposed during the biased firing study. These probes
were cleaned and prepared in an identical manner to those used for fur-
nace exposure and placed in a muffled furnace for 30 and 60-day expo-
sures at 400 C with a fresh air exchange. The test results were as
follows:
Probe
M (30 Day)
Q (30 Day)
R (60 Day)
B (60 Day)
Wt. Loss mg/cm - 30 days
4.7999
4.7741
5.1571/2 = 2.5785
8.3493/2 - 4.1746
These results indicate that the test coupons oxidized more rapidly
during the first 30 days exposure with average weight losses decreas-
ing in the second 30 days. Based on these results, it appears that
the differences in weight losses observed during the test program are
within the ranges to be expected from oxidation alone.
SUMMARY
Under normal unit operating conditions without the use of overfire
air, excess air variation is the most effective means of controlling
NOX emission levels. The use of overfire air provided a method of
further reducing fuel firing zone stoichiometries while maintaining
acceptable overall unit excess air and CO emission levels. The use
of overfire air did not affect unit performance and results of the 30-
day corrosion coupon runs indicate that longer term operation could be
111-24
-------
achieved without significantly affecting unit operating procedures.
Longer term corrosion studies of a year or longer would be required
before any definite conclusions concerning corrosion effects could be
established.
The results of this study indicate that variables normally used to con-
trol boiler operation should not be considered as NOX controls with
coal firing. These variables include unit load, nozzle tilts, damper
positions, and total excess air when overfire air is employed.
REFERENCE
1. CRAWFORD, A. P., MANNY, E. H. and BARTOK, W., "Field Testing:
Application of Combustion Modifications to Control NOX Emissions
from Utility Boilers," EPA-650/2-74-066, June 1974.
111-25
-------
A:
3:40 p.m.
Overfire Air as an NCbc Control
Technique for Tangential Coal-
Fired Boilers
Ambrose P. Selker, Combustion Engineering
When was this last experimental work done?
The last test work, you mean? It was completed right
around the end of last year. We do have
a follow-on study on Western coals. We got the first
baseline corrosion coupon results back, and they fell
in between the baseline and OFA numbers I got on this
unit.
Q: Did you notice in the off-stochiometric combustion
mode if there was any change, any observed change
in slagging characteristics between normal firing and
low NO firing?
X
A: I am going to answer that in two ways. First, we
did not see an adverse change in furnace slagging.
In other words, it didn't change unit operation where
we had to change retractable soot blower or wallblower
blowing cycles or something like that. Secondly, when
we did try to run the heavily slagged test with the
overfire system in operation, we would let the unit
go for 24-48 hours without wall blowing and come in
and look at it, and it would be dirty everywhere but
the firing zone. So we did have trouble getting the
firing zone to slag with the overfire air system in
operation.
Q: Would you care to elaborate on what you consider the
slagging characteristics of the fuel or the ash of
the coal that was tested?
IJT-26
-------
A: It wasn't a bad slagging coal; also, we have a
unit with a relatively low heat release rate here.
A pretty good quality coal. Now, this coal run in
other units in the same station where you are talking
about higher heat release units and so forth is still
not a bad slagging coal.
Q: Has Riley had any similar or totally different ex-
perience in their own work?
A: In some cases we have had increased slagging condi-
tions and in other cases decreased slagging. In some
cases with combined or overfire air conditions, we
improved the carbon loss conditions, acually decreased
NO by a factor of 2 and decreased carbon loss by a
A
factor of 2, which was not anticipated. So we have
had it go both ways.
Q: Do you feel that the corrosion rates that you obtained
on, let's say, a low NO firing mode, the differences
over base firing, are really significant?
A: I don't believe they are significant; I'll tell you why.
What I showed you were average rates; they were the
average of 5 probes, 4 coupons apiece. That is, 20
coupons. In the study where we had the low average, we
had some coupons at very high rates. Where we had the
high average, we had some with very low rates.
There was a lot of overlapping that didn't actually show
up in this average presentation I have here. It is in
111-27
-------
the final report though...the actual numbers for each
coupon.
Are you aware of the work we did on a prior EPA program
where we showed no significant difference for corrosion
rates? [Reference Exxon field tests.]
Q: This is on Barry No. 4?
A: On more than just Barry.
Q: Right.
A: Tomorrow I will present some more data.
I don't know why we saw the difference. I think it is a fluke,
I really do, right now, but I can't prove that. We are going
to find out in the Western coal program, hopefully, whether
or not we would get the difference.
I don't think anywhere in your paper you give us the
levels of the fuel nitrogen. You made some reference
that they increased at one point; but could you give
us an idea of how much of the NO that you were
A
measuring could be attributed to fuel nitrogen?
To answer the first part of your question, the fuel
nitrogen levels ran from approximately 0.9 and 1%
up to 1.5 or 1.6%, something like this. How much
is attributable to it, we haven't really established
that yet. I don't know, maybe it is 50%, maybe it is
100%.
Some work done by a very distinguished visitor, Mr.
David Pershing, with argon-oxygen experiments in a
pilot scale unit» has indicated that for coal firing
fuel nitrogen conversion can be quite high,
III-28
-------
perhaps 70% or more. Is that about right, Dave?
A: Of the total, yes. Now what the split is, I am not
sure that we know that yet.
Q: I'd like to make some comments regarding the Western
Coal Program that Ambrose has mentioned. We have a
follow-on program with Combustion Engineering, that is
looking at overfire air firing on two units of new
design that were built with the overfire air as an
extended wind box design. One is the Huntington
Canyon Unit of Utah Power and Light, and the other
is the Columbia Unit of Wisconsin Power and Light.
One will fire a Western coal and one will fire a bi-
tuminous. So, we hope to relate the effect of designs
built as extended windboxes versus retrofit; we
hope to be able to relate the effect of the unit size —
the Barry unit was 125 megawatts, the two new units
are about 400 megawatts, and we hope to be able to
get an insight as to the effect of the fuel itself.
And then, responding to Ambrose's conclusion or
recommendation regarding further work in the corrosion
area, we are planning, if we get the money from our
friends in Washington, to start a more detailed corrosion
program which would be sort of a monitoring type activity
where we would select one or two units that were equipped
with overfire air and try to keep sort of a continuous
monitoring of the corrosion rates on a couple of boilers
to establish some longer term data than we've been able
to get thus far. Technically, our corrosion runs have
been of 30-day duration. We'd like to extend that to
6 months to a year if we can.
Ill-29
-------
-------
CONTROL OF NO FORMATION IN
X
WALL COAL-FIRED BOILERS
By
Gerald A, Hollinden, James R. Crooks
N, D. Moore and R. L. Zlelke
Power Research Staff
Tennessee Valley Authority
And
Chris Gottschalk
Division Power Production
Tennessee Valley Authority
111-31
-------
This paper has been reviewed by the Environmental Protection
Agency and approved for publication. The paper summarizes data
taken from a joint EPA-TVA Study (Interagency agreement EPA-IAG-137D),
entitled "Control of NO Formation in Wall Coal-Fired Utility
Boilers." Approval by EPA does not signify that the contents neces-
sarily reflect the views and policies of the agency, nor does mention
of trade names or commercial products constitute endorsement or recom-
mendation for use. Dr. David G. Lachapelle served as EPA project
officer.
111-32
-------
ABSTRACT
It has been shown in previous work that biased firing techniques
effectively reduce NOX emissions; however, its long-term effects
on corrosion, slagging, and general thermal and operational performance
were not evaluated. The objective of this study was to provide
quantitative information from which biased firing techniques would
be evaluated as a method for control of NOX formation with respect
to its long-term effects on boiler corrosion, slagging, thermal effi-
ciency, and operational performance.
In order to accomplish the objective, particulate and gaseous
data were taken for a number of boiler load and firing patterns. In
addition, two separate furnace corrosion measurement techniques were
employed during long term runs,
Analysis of the gaseous emission data showed that certain combustion
modifications can reduce NOX emissions from 30 to 50 percent depending
on the load and the burner configuration utilized.
The collected and dried particulate was analyzed for unburnt
carbon. It was observed that the difference between baseline and the
"low NOX" condition was significant. In certain cases the carbon
losses were more than double for the "low NOX" case when compared to
normal firing. This is an important factor to investigate further.
Ill-33
-------
Of primary importance was the preliminary finding that for the
"low NO " operating mode, both corrosion measurement techniques
A
indicated a loss of tube metal significantly higher than that for
normal boiler operation. The most apparent tube metal loss occurred
on the side wall. More data will be obtained to further quantify
this finding.
111 - 34
-------
Introduct_ion_
In the last few years, TVA has participated in several research
programs aimed at both identifying and reducing NOX emissions from
coal-fired boilers and, to a lesser extent, from oil-and gas-fired
combustion turbines. 1 The overall results provided information on
control of NOX formation. However, there remained lack of information
concerning boiler performance and efficiencies.
The major combustion operating parameters investigated in the
past studies were excess air level and staged or biased firing patterns.
These parameters were investigated mainly at full or slightly reduced
load conditions.
The effectiveness of combustion modifications will vary with the
individual boiler. For example, during the earlier studies it was
shown that at full load conditions NOX emissions could be reduced
by approximately 30% over uncontrolled operation in certain boilers.
The NOX emissions were reduced by firing the operating burners in
the lower burner rows with substoichiometric quantities of air and
supplying the remaining air required through the burners in the upper row.
The purpose of the present TVA-EPA study has been to obtain more
detailed information primarily on possible corrosion and efficiency
losses that might occur while utilizing biased firing conditions as a
control measure for NOX-
III- 35
-------
The approach used in this field testing was first to characterize
baseline and "low NOX" firing condition. Using these results the
operating conditions for NOX emission control were defined. Second,
the "boiler was operated under the conditions for emission control while
particulate tests were run to evaluate ignition losses and overall
"boiler efficiency. Lastly, two long-term (300 hrs) corrosion tests
were run to determine waste rates. Corrosion rates were determined
"by two separate methods. In one method, actual boiler tube wall
thickness measurements were made before and after the testing program.
The second method involved the installation of air-cooled carbon steel
coupons exposed in the vicinity of the furnace wall tubes. The coupons
were used during the long-term corrosion tests. A weight difference
method was used on the coupons to determine the mils/yr loss rate,
Test Program
The major objective of this study was to determine whether any
significant operating problems occur during long range HOX control
testing — such as reduction in boiler performance efficiency, corrosion,
increased particulate loadings, or higher slagging rates.
To accomplish the objective, a boiler was chosen with enough
operational flexibility to permit a thorough evaluation of many biased
firing combustion modes. Widows Creek Unit 5 was selected. This Babcock
and Wilcox unit has the following principal features: single pass,
111-36
-------
divided furnace, water wall, dry bottom, radiant, reheat, 850,000 Ib.
steam per hr., 1825 psig (at superheater outlet), 1003°F (at superheater
outlet), 125 MW, pulverized bituminous coal, wall fired. The unit went
into service in September, 1955. Unit 5 has a sister (Unit 6) that was
used as a control for corrosion rate determinations for the long term tests.
The test program was designed to characterize the pollutant emis-
sions and process variables of both baseline and biased-firing conditions.
The evaluation consisted of 47 separate tests. Table 1 depicts the
conditions for each run. Figure 1 shows the numbering of the 16 burners
inside Unit #5. These individual tests vary the amount of coal and
secondary air flow through the burners. All necessary information
concerning the boiler operation (steam flow, air flow, etc) was recorded
along with the different emission levels of NO, NOX, CO, C02, 02, and
S02 for all test series.
During both the baseline and biased firing characterization test
program, the levels of gaseous emissions were continuously monitored and
recorded. The baseline emission levels of N0_, were used for comparison
A
with both low air and biased firing techniques. From an evaluation of
the characterization runs (47 tests), both baseline conditions and selected
operating conditions that significantly reduced NOX levels were run again at
each load condition (50,75,100, 125 MW). The levels of NOX, unburnt
particulate carbon, total particulate and boiler efficiencies were compared
for each of these operating conditions.
Ill-37
-------
Data Collection and Measurement
A mobile analytical van capable of monitoring the combustion
products was fabricated by an outside firm for use in the program.
The system is by necessity sophisticated, utilizing the accepted
methods in analytically measuring the components of the gas sample.
The gas sample was handled carefully to avoid contamination or loss.
The van is equipped with a nondispersive infrared analyzer to measure
CO and CC>2, a nondispersive ultraviolet analyzer to measure SC^, a
polargraphic 02 analyzer, a chemiluminescent analyzer for measuring
NO and NOX, and a. flame ionization detector for hydrocarbon analysis.
Table 2 depicts the equipment used and the ranges of detection for
each. All necessary calibration gases of appropriate concentrations
were stored in cylinders employing dry nitrogen as the carrier gas.
The instruments were calibrated daily before testing and periodically
throughout each shift to assure accurate analysis.
In running field tests, the samples were drawn before the air
preheater section from six zones in each duct providing 12 sample
regions per boiler. This sampling procedure reasonably assures a
representative sample. The sample is removed from the boiler under
111-38
-------
vacuum through stainless steel probes fitted with metal particulate
filters. The sample passes through teflon lines and its temperature
is maintained above 250°F until it reaches the gas cooler. The sample
is refrigerated to a 35°F dew point to remove water vapor and then
is sent to the analytical van for analysis. Figure 2 illustrates the
flue gas sampling and conditioning system.
Four Research Appliance Company particulate sampling trains
(EPA-method 5) were used. This setup included four sampling "boxes
equipped with U-inch filters and probes, and two sets of isokinetic
pumping systems. The measurements were taken after the air preheater
section of the boiler. Each duct was divided into 12 sample regions
to assure a fairly representive sample. The data was used to calculate
the amount of particulate emitted (gr/scf) and the uriburnt combustibles
retained in the flyash. The latter was obtained from an ignition
loss analysis.
Two different methods were used to obtain corrosion rate data.
The first method relied on actual tube wall measurements prior to
beginning the study and after its completion of both Units 5 and 6.
The tubes were cleaned to bare metal and then measured with a
Krautkamer USK5 Ultrasonic Minature Flaw Detector.
Ill-39
-------
The instrument was calibrated frequently with a standard cali-
bration "block. More than fifty measurements at different locations
were made on each boiler wall, totally at least UOO measurements per
boiler. Similar measurements were taken after completion of the
testing portion of the program.
The second approach (for corrosion detection) utilized corrosion
probes exposed in both the control boiler (normal firing—Unit 6) and
the test boiler (biased firing—Unit 5). The temperature of the SA192
carbon steel corrosion coupon was set at approximately 650°F for each
boiler. Two separate corrosion coupon tests were run. The coupons
were exposed to the test atmosphere with range of times from 315 to
755 hrs. After exposure, the slag was removed and saved, and the
coupons were then cleaned with a mild acid pickling solution. The
coupons were re-weighed to determine any weight loss. Figures 3 and
h illustrate the details of the corrosion probes. The probe design
is similar to Combustion Engineering's which basically consists of
two concentric pipes. Cooling air from the plant air supply enters
the tube-shaped coupons through a 1/2-inch stainless steel tube
roughly centered inside the coupons. The coupons (outside surface)
are exposed to the furnace atmosphere. The amount of cooling air is
controlled automatically to maintain the set-point temperature. Two
thermocouples are used for each probe—one for controlling the temper-
ature and the other for recording the temperature. The cooling air
111-40
-------
travels back through a 2-inch, extention pipe for discharge outside of
the boiler. Therefore, the cooling air does not interfere with the
test. Figure 5 gives the location of the probes and figure 6 gives
the operating procedures utilized throughout this portion of the test.
Data Evaluation
N°x: The analysis of the short term NOx test program shows that
the(lbs/MM Btu's) of NOX can be reduced by staged firing and lowering
the excess air from the normal levels at each level of load generation.
It is important to note that the excess air levels increase as the
load of the unit is reduced and that lowering load without either
staged firing or a reduction in excess air from the normal level for
that load will not in general reduce the NOX (ibs/MM Btu's).
Table 3 depicts the emission levels of Q^* C°2> CO, SOo, and
for each short term test. Figure 7 is a graph of NCL (ibs/MM Btu's) vs
Jt
02$ for normal firing modes at 125, 100, 75, and 50 MW. At 50 MW,
there were two distinct coals (in terms of heat content) used. The
higher Btu coal content coal gave the lower levels of NO and this
higher Btu coal was more representative of the majority of coal burned
during these tests. Utilizing the data at 50 MW obtained with the
higher Btu content coal and the data at 125, 100, and 75 MW, the following
relationships have been developed for normal firing conditions:
111-41
-------
loge 63 = 2.8358 - .0126 MW (1)
loge NO = -1.9068 + .02923 MW - 0.000127 MW2 (2)
d(HO) = d(HO) ^ d(MW) _ (NO) (.8) (-1.15 + 5.08 log^ NO)1/2 /-x
y* /i. A \ J *
d(02) d(MW) d(02) 02
where 52 = normal % oxygen content
NO = normal oxides of nitrogen (Ibs/MM Btu's)
MW = megawatt load
Therefore, given the megawatt load, the expected 02 level for
normal firing can be determined by equation (l), the NO emission can
be determined by equation (2), and the rate of change of NO with respect
to Op can be determined by equation (3) and thus the required level
of Op reduction necessary to meet a specified emission level.
As stated earlier, KOX emissions can be reduced by biased firing
at each level of load generation. Figures 8-11 are graphs of NOX (ibs/MM Btu's)
vs 02$ for different biased firing patterns and load conditions. The
approximate range of testing for each biased firing condition is repre-
sented by the length of the line on each figure. Extrapolation is not
recommended.
II I-42
-------
Partieulate: The particulate emission data are summarized in
Table k. Total quantities of particulate increased for the biased
firing tests but were not considered significant. However, carbon
losses increased significantly, particularly the relative differences
at high loads. Table 5 illustrates the effect of ignition losses for
both types of firing as a function of increased fuel operating costs.
Boiler Efficiency; Boiler efficiencies were calculated using the
ASME STEAM Generating Units, Power TEST Codes, using the Abbreviated
Efficiency Test Heat Loss Method. The assumptions made were that no
combustibles remained in the bottom ash and that the unmeasured losses
were 0.5 percent. Table 6 summarizes the calculated boiler efficiencies
and other pertinent information for normal and biased-firing conditions.
Figure 12 is a graph of mean boiler efficiency versus load (MW). A
consistent reduction of efficiencies throughout the load range of the
boiler is apparent under biased firing conditions.
Corrosjign: A summary of the boiler tube corrosion measurements
taken on Units 5 {biased firing) and 6 (normal firing) is contained
in Table 7. The following conclusions have been drawn based upon
a statistical analysis of the data:
(l) Unit 5, B furnace, side wall, lost an average of
7 mils thickness from the time the wall was first .
measured ("before" in Table 7) until the wall was
again measured ("after" in Table 7). The loss was
111-43
-------
statistically significant at « = .05 or only 5 percent
of time will this difference be due to random variation,
and not a true loss of wall thickness.
(2) Unit 5, B furnace, division wall, lost an average of
3 mils thickness. The loss was statistically significant
at °= = .08.
(3) No other walls tested "both before and after showed a
statistically significant (a = .10) loss.
(k) All walls in Unit 6, B furnace, showed a significant
gain in wall thickness.
(5) Comparison of Unit 6, B furnace, after measurements,
Unit 5, B furnace, before measurements, and Unit 5,
A furnace, before measurements, shows no difference,
both on an overall average, and a wall to wall average.
(6) The consistent differences on Unit 6, B furnace, on a
wall to wall comparison, in conjunction with the
conclusion above, point to a calibration error in the
measurements on Unit 6, before measurements, of
approximat ely 5 mils.
Based on the conclusion in #6 above, the tube wall measurements
for the B furnace on Unit 6 reflect this correction factor in Table T-
It is apparent that more data is needed to further quantify the
magnitude of corrosion. The unit is presently on line due to the load
III-44
-------
requirements on the power system, but as soon as it can be taken off
line a thorough tube wall measurement will be performed.
The corrosion probes were used as an indicator of possible corrosive
atmospheres in the boiler and were not intended to supply any signifi-
cant quantitative date. However, the results shown in Tables 8 and 9
are quite interesting. For meaningful comparisons, probe 1 can be
compared with probe 3 and probe 2 with probes k and 5- The comparison
cases are for those probes in similar locations in the boiler. In
the first series of tests (Table 8), it is apparent that between 2 and
5 times the corrosion rate (mils/yr) occurred on those elements in the
"low NOX" mode boiler. In the second series of tests (Table 9)» the
data shows significant corrosion increases when comparing probe U (biased
firing) to probe 2 (normal firing); however, little difference is
shown for the other location.
Conclusions
Even though much data has recently been published in the areas
of NO control, there has been little if any definitive published
A.
data on coal-fired units where such control has been correlated to
particulate emissions, boiler efficiency and corrosion. However, based
on the results of this test program the following tentative conclusions
can be made:
111-45
-------
Certain combustion modifications can reduce NOX emissions
ffom 30 to 50 percent depending on the load and the "burner
configuration utilized for pulverized coal-fired "boilers.
Carbon losses in the particulate will increase for biased
firing conditions. This carbon increase will reduce the
conductivity of the ash somewhat and could affect ESP efficiency.
Total particulate increases, however, were not considered
significant.
A consistent reduction of boiler efficiencies throughout
the load range of the boiler is apparent under biased firing
conditions.
Corrosive atmospheres will prevail in the boiler under biased
firing conditions. Such atmospheres .appear from preliminary
information to cause significant water tube losses in certain
areas of the boiler.
111-46
-------
REFERENCES
1. Hollinden, G. A. et.al., "HO Control at TVA Coal-Fired Steam
A
Plants," presented at ASME Air Pollution Control Division
National Symposium, Philadelphia, Pennsylvania, April 2k - 25, 1973-
III-47
-------
TABLE 1
Test Program for Baseline and Biased Firing Emissions Studies
At Widows Creek Unit 5
Test Program
Burners
Run;
1
2
3
U
5
6
7
8
9
10
11
12
13
lit
15
16
17
18
19
20
21
22
23
2U
25
26
27
28
29
30
31
32
33
Boiler Load
(MW)
125
125
125
125
125
125
125
125
125
125
100
100
100
100
100
100
100
100
100
100
100
100
100
100
75*
75
75
75
75
75
75
75
75
Excess Air
Normal
Low
Normal
Normal
Normal
Normal
Low
Low
Low
Low
Normal
Low
Normal
Normal
Normal
Normal
Normal
Normal
Low
Low
Low
Low
Low
Low
Normal
Low
Normal
Normal
Normal
Low
Low
Low
Normal
Firing Coal
16
16
lit
lit
lit
lit
lit
lit
11+
lit
16
16
12
12
12
12
12
12
12
12
12
12
12
12
12
12
10
10
10
10
10
10
8
On Air Only
0
0
AI jAlj.
E-i»DU
B2,Bo
C2,C3
Ai,Alj
DI ,DU
B2,B3
c2,c3
0
0
AI ,Al| ,D]_
B2,B3,C2
Aj,A2 ,A3
G! ,c2 ,c3
D-r ,D2 ,D3
A2 ,A3 ,DI
A^jAj. .Dj
B2,B3,C2
An ,A2 ,A3
C-, ,C2 ,C3
D-, ,D2 ,Do
A2,A3,D1
0
0
Bl )Bit
c2,c3
D-, ,Dj|
Bi>Bii
c2,c3
Di»Dit
B-. ,Bp »B-3
'DU
»C3
,Alj
»Cij
,Di|
>DU
,DI^
>C3
,AI^
,c^
,D^
,DU
>Bk
111-48
-------
Table 1 (cont)
Burners
Run
3k
35
36
37
38
39
ko
1*1
1*2
1*3
hk
1*5
1*6
1*7**
Boiler Load
(MW)
75
75
75
75
75
50A
50A
50B
5QB
50A
50A
50A
50A
50A
Excess Air
Normal
Normal
Low
Low
Low
Normal
Low
Normal
Low
Normal
Normal
Low
Low
Low
Firing Coal On Air Only
8 Ci.ca.cg.c^
8 B^jB^jDj ,D^
8 Bj_ , Bg , B^ , Bjj
8 C-, , Co jCo,^
O Bn , Bj, , D] , Dh
8 0
8 0
8 0
8 0
6 BO ,B^
6 cltck
6 c Ci
6 B-L^Bj^
In all runs at 75 MW, the A mill is off.
In these runs, the A and D mills are off.
In these runs, the A and C mills are off.
Used in particulate series only
111-49
-------
TABLE 2
Instruments
1. NO, NO Analyzer
Thermo Electron
Model 10A
2. CO Analyzer
Becfcman #86H
3. C02 Analyzer
Beckman #86U
k. SOg Analyzer
DuPont
5. 02 Analyzer
Beckman (7^2 Model)
Technique
Chemiluminscent,
negative pressure
nondispersive
infrared absorption
nondispersive
infrared absorption
nondispersive
ultraviolet absorption
Polargraphic sensor
6. Hydrocarbon Analyzer flame ionization
Beckman #UOO
Measuring Range
0-10 ppm
0-25 ppm
0-100 ppm
0-250 ppm
0-1000 ppm
0-2500 ppm
0-1000 ppm
0-5000 ppm
0-555
0-20$
0-1000 ppm
0-5000 ppm
0-5?
0-10J8
0-2551
0-1000 ppm
5-100,000 ppm
111-50
-------
•o oi-ooo
o^
H ift I—
fO f»"» CO
8?
cvi W Ji
cviStoo t-opw (rico wo cj
irvmmjf Jo3 >i)
51
4J
^
H
g™ ?
4»
s
«
•p
-.
1*1
os o\
111-51
-------
tfl
\D 00 ITS O\ ON CVI CM CO jgf CO JJ3 b--- ON, rH U"\ OS H Ijr
OOOOOQOQOOOOOQOaQO
o ®
ra
IA _=f
d 81
UN CM
IT* VP
OX 01
t**-
Q H Q Q fO
C^ V£ ON C\l rH
CD Q lf\ tA ON
§
ir\ b\
H CO CO ^
ft
! "A
i 6-1
+>
1
IfN ON Od
\^ !_| OD
O -5
4 H
4>
i®
8
m J- - OH CO VJ3 ^ CO
^t iTi^fiAro^r mcy
CO odfnjtcocdojrnvO^vO
on co co UN UN
vocoj*cMt^pnvoocM oon
o ON co on ON ON on UN CM CM ~t
t~-HpnvOtDCOOOON
-------
1
i-3 IA
M
SEH
H
5 o
i|
§B
SS
0
!| ^
*i
*= *
H B
K C
glk.
M Q
§
g«
wo!
cvw
KEH
o|
i^ W
£
u
g*
*,
1°
<\J
O @)
**** o
Q,
Pi-Mi
• — rn
0®
_
S CV
1°
"""en
g*®
1
"S1
i*S
cvia
»i
S3
g
T3
«
43
•"
1
ffl
_p
!
I
Pi
1
<
1
7
ffl
HP
1
J
a
1
»
^
g
T
|
g
*.
of
R
IA CO VD .^ O
-rt vO -3- IA b-
0 0 0 0 O
OO OJ OS Pn r^
C- ^JD J- J1 H
-^ O r- ( CTi O\
rn i-i H
-* ct f: 3 ^
-3- H H t7\ oo
on H r-l
CT\ t1- ON CM rn
O cVJ OO VJD j^j*
IA
t^- CO J" J~ O
\X) OJ CO IA J-
t\j cn
8 f » 8^ rt
m 3^ ro rn tA
IA PO kA CO OS
rn IA m m ,-f
m \o CM on H
^t J>- LA IA CD
vj3 cO vO O^ O
m vO ro on \x)
-if ,=f H OO V£>
^t CO O CO SjO
_j- ro en IA -nf
o\ t*- t*~ t~ ^-
1 » iSi « S
1A !A
LA LA IA ^- t-
f- t-. t«- 1 1
I I t O O
^O fp L^ H H
jH rH ri rH rH
o t*--moo-4'co ro^fcovo IAO
\jQ CO t*- -3" \O C~- tA j IA ^" ^J" -5T
oooooooooooo
St*- OJ VD W^ OJ CO CO \O VO -^f" (*^
r-t H fcr^C lAj'COOO-S ON \p
^H O O ON -j O\ ON t~- -^ r^ '^O W
rO ^f r-4 Cy OJ rH rH m PO rH F^
f1^ t^ rn ON ON CM p o ON t- ^H w
COPO OJCMHHCOPO«HH
rH CO CO t~- OO ^O r4 ON -^ ON t*- ON
co m^ w-,* iAt-iAm-^^oco
-^ N- CO y? ON \O IA J" IA OO CO PO
CM rO PO ,^ PO ^r CO \O Pn CO O ON
S!$ K^l fc^SI^^?
H^t C*-vD IA t-- H CC t*- <*^ -^ ON
UNPOH LAfO.31 H LA-it *f\ O ON
-^tvO iAcn^ ir\ j* POJ- cnmoj
CMCO ONlAlAjJCOH H-4-ONCn
^t ITX t^ -* CD O QNCOCO\OlfNCO
H
VOVO PO t-- i»t CO rOH f-CD ON^tf-
PO tA \o on t-- co o co t— IA ^f co
ONvj3 tA_4" C*--d'^t' IA ON LA 1A *r>
-* OOCVIIAONOOOCU POO^r^l
rHHH HH*HHHr-IH
o r--ONt\j HCO lAj^- cnofoon
^^^^jjd00 Pn^-^t
-------
fi*i
^
* s
d d
i 1
-as
o
o
s
a
p p
wo
cT
I
ft I
-------
M CO
I X
I 1
O
a m*
•vo
r3 0
i-nl H
oS
•H •— -
-P
•H (0
as
M 1-5
-P
* 0
M* 3
<-( .. ^ f\
G •• N f— i
•ri PL, |
T) u m
0} CO
J en
a
fl -H -p
•H n) o
S3 Q
£5J?
-p WJ
r-J gj
d r,
0 O
h Tl
0) 0>
Pn -P
O
rH 0)
^?
55
rt
^
K
-P
iri
£^
VO
CO
0
OO
vo
ON
O
OJ
•fc
-f
H
O
H
OJ
8
CO
£
H
<
H
1
IA
LA
^—
1
O
H
1
OJ
ON
t-
O
0
ON
t—
vo
(O
VO
00
OJ
LA
O
IA
CO
co
OJ
OJ
oo
H
PQ
f-^
|
VO
LA
^—
1
H
H
1
OJ
ON
LA
O
0
ON
rH
VO
00
_3"
0
CO
|S
t—
ON
t-
^
l/"N
C\J
|
t—
LT\
r-
i
OJ
^
i
OJ
OJ
LA
0
-3"
oo
H
VO
vo
oo
OJ
H
00
OJ
o
OJ
O
LA
<
ON
CO
1
OO
LA
t—
1
OJ
H
1
OJ
H
IA
0
ON
ON
*
_^
t—
H
00
OO
VO
_-,.
H
t-
ffl
OO
1
ON
LA
f-
1
OO
rH
1
OJ
oo
Jt
o
OJ
ON
H
00
OO
o
00
oo
UN
rH
t—
H
t-
o
Cf\
1
o
l-l
LA
t-
1
00
H
1
OJ
00
CO
0
OJ
vo
o
OJ
LA
ON
c-
LA
LA
rH
LA
CO
OJ
H
O
H
1
P-f
H
LA
t—
1
| 1
|
OJ
CO
t—
O
ON
tr\
i — i
vo
LA
^3-
t—
-*
t—
IA
O
CO
oo
LA
t~
pq
IA
OJ
|
OJ
H
LA
t—
1
-~f
rH
1
OJ
IA
t—
O
-=r
t-
-{•
t-
_-,.
00
ON
LA
00
CO
ON
00
t—
0
LA
OJ
1
oo
rH
LA
t—
1
CO
H
1
OJ
OO
CO
0
H
j.
ON
oo
t—
oo
H
t—
00
CO
oo
ro
oo
LA
LA
o
H
H
H
1
H
LA
t—
1
ON
H
OJ
VO H
ON -^t
0 0
ON vo
CO O
rH
-3" VO
H t-
C— t-
rH 0
t- -a-
CO LA
H LA
LA CO
00 ON
VO O
LA OO
OJ -*
O -*
H
PQ <
H I-
| I
LA VO
H H
LA LA
t— t—
1 1
rH H
OJ OJ
1 1
OJ OJ
111-55
-------
jgg pQ ^
ir\
sin °
IS
•4^ *
Tt 01 CO
as
M K-5
-p
* o -3-
W* P O\
•H EH 1 t—
0} 03
1-3 CO
C
R -H .p
•H aJ o H
£ ^ (S CO
O — - 1 t—
^
1
-P IA
CM
CM
IA
O
t—
•
CO
\o
OJ
MD
vo
CVi
V^)
|v —
ir\
-^
^o
*
ITS
j-
_jf-
CVI
H
?
CO
rH
LA
T
LA
^
H H H 00 |
rH P
H
g
W
OOOOr— LACOrHLA oj
LA LA ON -^ C— LA VO t«0
CO t— t^ VD -3" OO J" "d
§
M
-p 0)
* to w
t— VO CM O co t— ON »
C! H
•H -d
rt
3 C
T^ O
r< C '
0) -H 60
i — 1 cd W
co t""- ON oo rH ON oo p4 JH
O CM VD -* t— VD LA B-p ON ON t— t— IA LA <;aj+>
PH -p 03
M o
0) En
C HO
•H HO
PQ < pq <• pq << pq Ot—
l^HrHOOVOVD -do
J'CVICVIOOOOOOOO t) O
1 1 1 1 1 1 1 ft (U 43
OvOHCMf^-^lA O-P
HCMCMCVICMCVICVI HOSTS
0> HO
> ^ 4^
11 o o
LALAIALALA "d'HO)
f— t— t— t— t—LALA 4Jf-l
1 1 1 1 1 t- r~- ^ h rH
vot^b— cocol 1 to an o
CVICMCMCVJCM-^--^ (UPUO
1 I i i i 1 I (-5 *
CVICMCMCMCM^fO #*
HI-56
-------
!H
O
K
M
•H
<<
rH
fn
•H
<
IH
O
K
4>
D)
&&
h
•H
<
h
•H
<
P.
•H
•a;
TJ T3
s!25^^>
Q Q o Q
«*>H3l-l ««t-ljH3
D 4) 4) t/>C/)O v
tfl (A
pq pq
CO
O
CO
8
M
oooooooooooooo
oooooooooooooo
O OcOCOONi-lrH O^i O H LfN O\ t— t—
o o o o o o
irsJ-OJrHOHO\H
H
r-\
tr\<*">
w:
OO
ON OJ \O
O OJ
VO
O\VO
H
O O-*
OJOOCO PO|>-o\o\VO-3'-3-
_. l-UNOJCTkCOVO
MOJOJCVJOOONONt-
rH H r-t rH H rH
oo ono i— t—-3- co
t-t-t—IAJ-J-J- m
13
i~it— co
I/NO
CVICM
pn op J* ^± rn tn
o\ovo o\-=«- ir>
r-lr-l(HCVJCVI
«
w.
>
^
o
ON
£• 3s
^g^i^
w o
+3 VO
rH -p «J O
O 4) -P H
fVS «J ->*
CJ --^« M
« 05 43
IH +» a
H -H Oj 4)
05 S 4) O
O W
CO Oi CO
•H CO • t—
PL, O\ 0) .
• !> O
CCO < t~
o
T^ • * •
4) H OJ CO
-------
VO
Ok U
ri S
S 5
e
w w
J M
LJ 1-J
O M
CO
CO
Is
CO
w
s
-_
IGNITIO*
%
< **
>-
a c
Xw
o m 6-
^^ p
K^* "
vO
.0 O rj r*, &} ~3
0) 0) 01
C! C X X C
•H -H O 0 -H
i-H rH S5 S! iH
CO 0) 01
M CO S > CO
m vo oo co co o
m r- vo
ON i— ! in O ON "^
co in *n ON in r^
iH rH
r-l ON vO r^ O OO
CM oo oo co r^ cr*
vo -* m ON oo m
^\ in in in r*^ r*»
o o o o o o
r-. co r» m m o>
vo m m r-. oo oo
CM ON oo so m «*
^^ ^j^ c^ ^* r^ ^^
pg <• pq •< pg U
rH rH r-l in I/I in
in ci 1-1 i-*. CM co
01 01 <1> 01
C X X C C C
•H O O iH -H iH
01 0) 01 tt)
SU u co w tn
o o at cd co
in
m
00
o
oo
sr
00
CJN
0
m
o
vo
vO
CO
r-.
£
CM
t
3
a
VO
rH
H3
CO
CO
*v'
oo
00
00
vO
0
m
o
0
r*
CO
r-.
•
8
s
CSI
i
J5
3
III- 58
-------
3
H U
Pn
W
en
CO 00
st
oo
00
m
00
CM
CM
00
st
00
o
st 00 H
m m en
oo
00 sD
I— OO OO 00 OO vO OO
tn
a
O
CJ
H
O
o o o o o o o
t^j ,_( cn "*£> rH tn m
00 C^ C7^ O^ C^ f*^ 00
co CM c^ r*^ r"*- ^f oo
in m m *^~ *^ *** ro
^C PQ *^ pQ i^^ ^j pQ
^D ^o ^^ ^^ ^^ r*^ r™^*
CO OO CO CO CO *^ >3"
1 1 1 1 1 1 1
^" in oo o\ o \D CT\
CN] CM iH rH r-|
Q>
cd
o>
>
ed
0)
,c
4J
M
•H
0)
3
•H
CN|
O
QJ
•E
E-t
*
III- 59
-------
s
pgs
pa o. H
ijKX
j s o
o o a
PH H M
t-
%
S3
EH
d,
w
CQ
M
PH
K
0
CQ
g
e
1
M
O
F 1 .n
B s
B EH
a
0
o
1
i
Ix,
«
g
O
W
H
H
CO
pq
ij
^
>
fn
0)
a
h
3
PQ
H
•tH
Q
K
W
§
W
u
<
fe
S
s
H -H
a a> d -H
>H TJ O >
pi -H S -H
W CQ fc P
•P
at
a;
^
o
1
M 4J
1) C
H U
a) 'H
•H
*o 12
a) Si
•P -H
0 CQ
h 0
II
II
CQ
H a
III- 60
-------
P g CO
EH
Q ^ -H JH -H
M pq CO £ P
IS
PM
H
K
•3
,
w
^ S
a o
S fc
\£> O VD t—
oo on on on
<^J
h-1
, "1
1
§
>H -rl
CU -p CO
fl C -H
fi -d 0 l>
d -H (H .H
ft CO fc P
-^^>-
SH
0>
CO CO
CO CO S-,
CD
jjj
cd cd o
•H
-P -P Vl
C S -H
CO CCl JM
O O t£
•H -H -H-
IH SH CO
•H -H
•H -H is;
CQ CO
II
II II
CO
CM on a
III- 61
-------
8
M
CQ
8
B
CO
CM
*
CO
OO
VO
-4-
•
CXI
IA
IA
8
EH
rQ
a
CO O
H
W CO
m §
* 8
p3
13
"
o
s
-------
O
H
O
8
lr\ J-
• *
jzt LA
LT\ CO
o\
CO CO
ir\ u\
fO no
vo
LP>
m
o
a
9
EH
&H
CQ
a
o
H
K
8
i
W
EH
1
O
a
o
i-i
3 CQ <~-
1-3 VO CQ CQ
Pk OS
S PH
-------
FIGURE 1
TVA's 125MW Widows Creek Unit 6 - Burner Configuration
PULVERIZER
ROW
2 3
B
o o o o
o o o o
o o o o
o o o o
111-64
-------
111-65
-------
TII-66
-------
111-67
-------
FIGURE 5
PROBE LOCATIONS
Unit 6
I Burner Wall
Unit 5
1 Burner Wall I /} 3
V
'C" Burner Elevation
"C" Burner Elevation
V | Burner Wall L
n
n
T 8
"B" Burner Elevation
III-6S
-------
cu
W)
to
fl
O
o
o
-p
O * !>> O CU O
POH .p 0 ^!
^i <( C X to
0) O H CU
g" CO CO
OJ S-i S d h
CVJ H g Jj 0) 3 C
, — 1 «r rt O rr; o ,~i
a -H
O cfl
.H h a
40 -rt CO n) Qj
• ooo
O'OOO
oooo
• ooo
oooo
oooo
oooo
oooo
l6-Burners in service ill-Burner configura
Reduction of excess air A-^ , Aj^ burners on a
until burners show a only. Reduce exces
rapid change in CO. NOX until burners show
lowest as possible. rapid change in CO.
NOX less than 500 p
111-69
-------
!H
•H
05
•H
c8
0
-a
a
t) >» aJ
a H
i S3 H
o a
rH
03 ^l
O -H
o ai
c a
^a to c8
fn ' «VO
fn
cd
"* 6 ""'«•*-• S
^S HO I
-° 2
£H
a)
efl
3
OO
oo
®ooo
oo
c
o
-H
Id
C
O
CJ
1
a)
O
§0
J-
w «
0; rod 3 to
C < ni f> to
h -
-------
-------
o\
M O
Cu B
IfN
(M
LTk
(M
OJ
oo
•
o
U"\
O
III-72
-------
CO
HJ
•H
-------
CO
si
o
0)
O
H
B
E
o
Os
\D
if\
I
ON
*
o
CO
o
o o o
XOH
111-74
-------
!-i h H
O -H -H
H
cr%
-oo
H fn
H "es. CD
CO
QJ CO
-H O O
P^l S UA
cr\
CO
d
o
ro
O
OH
III- 75
-------
Figure 12
Boiler Efficiency versus MW
87
86
30
75
100
125
A - Normal Firing
B - Biased Firing
HI-76
-------
4:20 p.m.
Control of NO Formation in
Wall Coal-fired Boilers
Dr. Gerald A. Hollinden, Tennessee Valley Authority
Q: I tried to glance through your paper quickly I couldn't
see it, it might be in there...where did you measure the
particulate loading? Was it before the ESP, after the
ESP, or where was it?
A: It was after the air pre-heater section.
Q: What was your particular collection device for that
system?
A: It was a Research Appliance system mechanical collector,
Q: You made the statement that the carbon losses were in-
creasing but there was no overall, there was no increase
leaving the furnace. It had to be accounted for
someplace.
A: No, I said no significant increase in particulates.
The difference that we did notice was in the chemical
content of the fly-ash as opposed to the quantity of
the fly-ash.
You are aware that Exxon tested boiler number six
(Widow's Creek, No. 6) a couple of years ago and we
obtained similar results on increased carbon on
particulates. We evaluated that out on the effect
on performance and there was no significant effect
on performance or efficiency. You are also aware
that we tested your same boiler.
I am not aware that you did any ignition loss tests
on the particulate; I may be wrong.
We did check the carbon and calculate the efficiency by the
same method you did. And we found even under those
111-77
-------
A:
conditions, where we had an increase in carbon, that there
was no appreciable difference in efficiency. Remember, we
reduced the excess air and we had a net gain in efficiency
from that standpoint. So the two balance each other out.
That was in the prior work.
You know we shouldn't mislead the audience. We
would have to compare the excess air levels. I have
already mentioned that we dropped pretty low (excess
air), especially at 125 megawatts, and I don't think
we would do that again. 1 don't really remember
2 years ago what we ran for excess air levels. I
don't think it was as low as what we did in this
program.
Tomorrow I will present some other data on your
same boiler and in that case we had a decrease in
carbon loss on low NO firing on the same boiler.
A
Now, I gather from your data you are saying, in
every case, low NO firing increased the carbon
X
loss. Is that true?
That is true. I am saying that with the four low NO
X
conditions that we selected for 50, 75, 100 and 125
megawatts, compared to normal firing at those four
loads, we did have a reduction in boiler efficiency.
We did not obtain particulate data for all 47 tests.
We are also going to present some data on corrosion
based on corrosion probes taken on your boiler #5
and boiler #6. Our data show that there is no
significant difference in corrosion, by corrosion
probes.
It is very important, as I stated earlier, where
you put the corrosion probes, since the probes are
III-73
J
-------
just an indicator of corrosion. In one case under
normal firing conditions we got an equivalent 8 mils
loss per year on a particular probe and under the
same conditions several weeks later got 50 mils loss
per year. Obviously there was not a seven-fold in-
crease in corrosion for the normal firing case over
that period of time. So I don't hold much credence
from a quantitative standpoint in our corrosion probe
information except that it was strange that we always
got more corrosion or more loss of the coupons in the
stage-fired unit as compared to the normal firing.
And that is really my whole point. The important
issue is that more information is needed on the
actual tube-wall measurements. That will give you
your final answer.
Q: What I would like to know is whether or not the 0«
levels or excess air levels at the corrosion probes
were determined when you saw the increased losses.
A: No, they were not. We didn't have any sample withdrawn
in the furnace area itself.
Q: Can you give us some idea what your overall excess
air 'was, and what your burner stoichiometry was at
the conditions where you saw the dramatic reduction?
A: At 50 megawatts and 75 megawatts the 0,, level was
somewhere between 6 and 9%. We weren't even near
reducing condtions in the burner region. However,
at 125 megawatts, where we were operating under biased
firing conditions at 3% 02 we were substoichiometric in
the burner region. I don't know what all of them
were, but at 125 megawatts we got down to about
.85 to .9.
111-79
L
-------
Q: And this is where your corrosion data was taken that
you showed?
A: Yes. Again, the corrosion data for the coupons were
only done for the four selected tests. They weren't
done for all of them. However, the boiler corrosion
tube data was done over the entire program.
Q: Gerry, did you note any difference in furnace per-
formance under bias-firing conditions?
A: No, sir.
Q: No increases in spray-flows and things like that?
A: No, every once in a while we saw some soot fly
across from one region to the other.
[Comment by Symposium Vice-Chairman:]
I think it is important to point out, as you have
alrady said, some of the test data is not in yet.
There are still some questions remaining on the results
of this program. Even assuming that it all pans out the
way it looks, I think it is important to point out that
this data was taken on one unit; it is not necessarily
typical of all boilers, especially different designs,
and a lot of work in this area still needs to be done
before conclusions can be drawn.
Q: On the calculations of efficiency, was it done on a
heat rate basis, Btu's per kilowatt hour, or did you
calculate it on another basis?
A: No. It was merely a boiler efficiency calculation.
ni-30
-------
[Comment by Moderator:]
Sometimes on a heat rate basis, it may not change. In
other words, it takes the same number of Btu's to generate
the same KWH regardless of your conditions. And it is my
understanding that some boiler manufacturers (I may get
jumped on for this by some of the people who are manufacturing
wall-fired units), are aware of the potential for corrosion,
especially on the wall, on the side walls, and note that on
this unit we had apparent corrosion on the division wall
and on a side wall. Some manufacturers are building in a
curtain air concept wherein there is some bleed air that is
brought down the side wall to sort of compensate for this.
111-81
-------
-------
THE EFFECT OF ADDITIVES IN
REDUCING PARTICULATE EMISSIONS
FROM RESIDUAL OIL COMBUSTION
By
R. D. Giammar, H. H. Krause, A. E. Waller, and D. W. Locklin
BATTELLE
Columbus Laboratories
Columbus, Ohio
111-83
-------
THE EFFECT OF ADDITIVES IN
REDUCING PARTICULATE EMISSIONS
FROM RESIDUAL OIL COMBUSTION
by
R. D. Giammar, H. H. Krause, and A. E. Weller
ABSTRACT
In this study, the effectiveness of proprietary additives
and pure compounds in reducing particulate emissions was evaluated in
a 50-hp (500 kW) oil-fired packaged boiler. It was observed that
additives containing certain alkaline-earth and transition metals in
concentrations in the range of 20 ppm to 50 ppm in the oil were
effective in reducing carbon particulate by as much as 90 percent.
These additives were also observed to be effective in reducing smoke
and polycyclic organic matter. Explanations are offered suggesting
possible additive effects on the combustion mechanism that could
account for differences in performances of the various additives in
achieving particulate emission reduction.
Ill-84
-------
THE EFFECT OF ADDITIVES IN
REDUCING PARTICULATE EMISSIONS
FROM RESIDUAL OIL COMUBSTION
by
R. D. Grammar, H. H. Krause, and A. E. Waller
INTRODUCTION
The use of small amounts of fuel additives as a means of
improving the performance of fossil-fuel boilers has been of recurring
interest for many years. Trials with various fuel additives have had
different objectives, depending on the type of fuel used and the era
in which the trials were conducted. Earlier additive trials and investi-
gations were aimed at reducing the deposition of soot and ash on heat-
transfer surfaces and minimizing the corrosion that resulted from the
accumulation of these deposits and from their interactions with the
flue gases. Recently, the use of additives has been emphasized as a
possible method of reducing air pollutant emissions—mainly visible
smoke and particulate--although some claims have been made of additives
reducing emissions of sulfur oxides and nitrogen oxides.
111-85
-------
Although a substantial level of effort has been devoted by
various organizations to the development of additives to reduce air
pollution emissions, the effectiveness of some of these additives has
been in question because of the lack of consistent results in both
experimental laboratory investigations and in actual commercial usage.
These inconsistencies, in part, could be attributed to differences in
combustion equipment, operation, and maintenance; fuel type and com-
position; and perhaps, most importantly, evaluation methodology.
Thus, while qualitative observations and commercial claims
for the effectiveness of additives are numerous, relatively few quanti-
tative data are available from experimental investigations in which
boiler operating conditions were well-defined and in which emissions
were accurately measured. Consequently, there are few scientific data
to support an assessment of how, and under what conditions, additives
work.
This paper discusses a portion of an ongoing research program
for the U.S. Environmental Protection Agency* to systematically evalu-
ate the effectiveness of additives for reducing combustion-generated
air pollutant emissions, especially particulate, from an oil-fired con-
tinuous combustion system. This program is unique in that an extensive
and well-documented program was conducted to evaluate a large number
of proprietary additives and pure compounds under controlled conditions
* EPA Contract No. 68-02-0262, "The Effectiveness of Fuel Additives
in Reducing Emissions from Oil-Fired Continuous Combustion".
Ill- 86
-------
in a single unit—a commercial 50-hp (500 kW) Scotch firetube boiler
firing residual oil. This work is an extension of the work of G. B.
Martin, et al., who systematically evaluated about 300 additives for
reducing emissions from a distillate oil-fired residential burner *.
In general, most proprietary additives have not been found
to be effective in reducing air pollution emissions . However, some
additives have been found to be effective in substantially reducing
smoke and particulate .
REVIEW OF EXPERIMENTAL ADDITIVE INVESTIGATIONS
To guide the selection of additives for this program, the
literature pertaining to experimental studies of combustion additives
was reviewed**. Previous investigators have used a wide variety of
organic, organo-metallie, and inorganic compounds in efforts to reduce
the amounts of pollutants formed during combustion, or to convert the
pollutants into easily removable solids.
The greatest amount of effort in additive development has
been devoted to particulate emissions. This emphasis has been the
result of the almost universal recognition of smoke as an undesirable
emission over a period of many years. For obtaining reduction in
particulate formation in fuel-oil combustion processes, experimental
evidence indicated that compounds of the transition metals, manganese,
* Numbers in brackets designate References at the end of the paper
** This state-of-the-art review was a part of the program on EPA Con-
tract No. 28-02-0262 and will be submitted as part of the final report,
III-87
-------
(2)
iron, nickel, and cobalt, could be effective . The alkaline earth
(3 4)
metals, barium and calcium, also gave evidence of useful activity ' .
The best results were obtained with organometallic derivatives of these
metals. The organic portion of the molecule provides the needed solu-
bility in oil and also influenced the stability and volatility of the
additive molecule.
Only a few investigations have examined the effect of additives
on potential carcinogenic materials in the particulate emissions, but
this aspect is presently receiving a significant amount of attention.
Such potential carcinogenic agents are found primarily in the polycyclic
organic material (POM) that forms during the combustion of the fuel * .
Host of the early research was done with gas flames or light fuels such
as distillate oil. The additives which were reported as effective in
reducing the POM emissions included organic compounds such as nitro-
paraffins and peroxides , as well as organo-metallic compounds of
/g\
manganese and iron . Although the physical state of POM and the
carbonaceous portion of the particulate is different, their probable
common origin and chemical reactivity suggest that an additive effective
for one would have some effect on the other. This result was generally
observed in the earlier diesel engine studies, and where both types of
emissions were measured, the POM usually was reduced to a greater
extent than was the particulate .
The U.S. Environmental Protection Agency has conducted two
comprehensive experimental programs that have contributed quantitative
III- 88
-------
information on the effects of fuel-oil additives in reducing pollutant
emissions. One program involved tests of numerous commercial additives
used with distillate heating oil in a residential oil burner . In
general, the additives were found to have a minor effect compared to
the opportunity for emission control by proper adjustment and operation
of the burner. However, the additives containing manganese, iron, or
cobalt did reduce the particulate emissions. A few organic formulations
containing no metals gave moderate reductions in particulates, but the
concentrations of the additives required were too high to be practical.
No changes were observed in the amounts of nitrogen oxides or sulfur
oxides in the emissions with any of the additives.
The second EPA program was an investigation of the effective-
ness of several proprietary fuel-oil additives in reducing sulfur oxide
(9)
emissions when firing residual oil in a commercial boiler . No
effects of the additives were found for sulfur oxide emissions or for
nitrogen oxides.
The experience of the EPA investigators with the inefficiency
of additives for reducing the nitrogen oxides or sulfur dioxide in the
emissions verified the generally negative results reported by the
earlier studies. Only in the case of sulfur trioxide, which is present
to the extent of 25 to 30 ppm in the flue gases, has an additive effect-
iveness been demonstrated. Chemicals which form basic oxides such as
magnesium, calcium, and zinc have reduced sulfur trioxide levels sub-
stantially (10).
Ill-89
-------
EXPERIMENTAL PROGRAM
Commercially available additives and other materials selected
as a result of the previous considerations were experimentally evaluated
with fuel oil in a 50-hp (500 kW) firetube, packaged Scotch boiler
capable of firing natural gas, distillate oil, and residual oil at
rates up to 2 million Btu/hr and generating up to 1500 Ib/hr (680 kg/hr)
2
of steam at 15 psig (103 kN/m ). Combustion air was metered with a
thin-plate orifice, distillate oil flow with a rotameter, and residual
fuel oil flow was determined from beam-scale readings. Fuel was sup-
plied to the burner at a fixed rate by a variable-speed, positive-dis-
placement gear pump. For residual oil requiring preheating, controllers
were used to provide a constant fuel temperature at the pump inlet and
burner nozzle.
The combustion gases were vented from the boiler through a
12-in (0.305 m) diameter insulated stack. A special sampling platform
permitted sampling in a straight section of stack, 15 ft (4.6 m) above
111-90
-------
the boiler outlet. Figure 1 shows the overall experimental facility.
Analy t ica 1 Procedure s
Particulate and POM sampling and analytical procedures (modi-
fied EPA Method 5 train with an adsorbent column) have been described
by Jones and Giammar . Gaseous emissions were determined by: para-
magnetic analysis for oxygen; flame ionization detection for unburned
hydrocarbons; nondispersive infrared for carbon monoxide, carbon dioxide,
and nitrogen oxide; and a dry electrochemical analyzer for sulfur dioxide,
Smoke emissions were determined with a Bacharach smoke tester according
(12)
to the ASTM filter-paper method for smoke measurements .
Fuel
3
To provide a uniform reference fuel, 6000 gallons (22.7 m )
of a residual fuel oil were obtained that consisted of: 87.5 percent C,
11.1 percent H, 0.31 percent N, 0.95 percent S, and less than 150 ppm
trace metals. API gravity was approximately 15 . The distillate oil
used was a typical No. 2 heating oil, with 86.4 percent C, 13.6 percent
3
H, and less than 0.1 percent S. An additional 6000 gallons (22.7 m )
of a 2 percent sulfur residual oil was also procured and additives are
currently being evaluated with this oil.
Selection of Additives
Fuel additives have been used as a means of accomplishing a
number of functions that can be distinguished as occurring before,
during, or after the combustion process. For the purposes of this
III-91
-------
investigation, only additives that were intended to improve combustion
and thus reduce emissions, were considered. However, it is recognized
that fuel handling and postflame treatment additives can in some circum-
stances improve the overall performance of a combustion system, especially
in marginal cases, by keeping nozzle or heat transfer surfaces clean.
To categorize these "combustion improver" additives, the follow-
ing three classes were defined:
• Organo-metallic
• Organic
• Inorganic.
Table 1 categorizes the additives used in this study according to defined
classes and subclasses. Inorganic additives were not evaluated because
of their insolubility in oil. Selection of these additives was based
on a state-of-the-art review of additive technology and on theoretical
considerations of the mechanism of pollutant formation and possible
modes of additive action.
Concentrat ions
Concentrations for the proprietary additives were based upon
the manufacturer's recommendations or the results of an earlier EPA
study . For pure compounds, concentrations were based on approximately
27 ppm of metal in the oil (0.1 g metal/gal of fuel oil).
111-92
-------
Procedure
To ensure a valid and controlled evaluation of each additive,
procedures were developed to minimize operational variations over the
entire period of experimentation (approximately 1.5 hr) and to provide
several measurements for evaluation purposes.
For each additive evaluation run, the reference fuel oil was
fired at a constant, continuous rate in the boiler facility to establish
a baseline operating performance over a range of air/fuel ratios. Gaseous
and smoke emission data were obtained over this entire range, while a
filterable particulate emission was determined at one selected air/fuel
ratio. The additive containing fuel was fired at the same firing rate
over the same range of air/fuel ratios to generate a boiler performance
curve with the additive. A particulate sample was also obtained at the
same air/fuel ratio as for the reference oil run. Thus, the effectiveness
of each additive could be determined by comparing the smoke and gaseous
characteristic curves and the particulate loadings with and without addi-
tive for runs made on the same day. (Except as noted, all data reported
in this paper were obtained at steady state.)
Particulate consists of two basic components: (1) an ash
component originating with the fuel, and (2) a carbon component re-
sulting from incomplete combustion. For combustion at a constant
Probe and filter catch only.
111-93
-------
air/fuel ratio, the ash loading in the products of combustion should
be constant assuming no accumulation on the boiler surfaces; the carbon
loading could vary depending on the completeness of combustion. To
minimize the variation in ash loading, all particulate measurements
were made at the same air/fuel ratio corresponding to 12.8 percent CO.
(about 20 percent excess air). A baseline firing rate of 80 Ib/hr
(36 kg/hr) was selected to correspond to a commonly used boiler load of
80 percent.
SUMMARY OF RESULTS
Although additives were evaluated with both distillate and
residual fuel oils, this discussion will be limited to the residual
oil runs, because steady-state firing of distillate oil in the 50-hp
(500 kW) boiler did not generate appreciable quantities of carbon
particulate at normal air/fuel ratios (over 8 percent excess air).
Thus, a valid measure of the effectiveness of additives firing with
distillate oil could not be made. In fact, for those additives con-
taining metals, particulate loadings were observed to be increased by
the use of additives with distillate oil.
Table 2 summarizes the results of the additive evaluation runs
firing the 1 percent sulfur residual fuel oil at 12.8 percent CO . As
indicated in portions (d) and (e) of Table 2, several runs were con-
ducted at different firing rates and one run was conducted with cyclic
boiler operation so that other modes of boiler operation could be
111-94
-------
considered. The additive concentrations or dosages are either designated
as ppm metal in the fuel oil for those additives containing metal even
in trace quantities, or by volume percent for the pure organic compounds.
As for the naphthenic and 2-ethyl hexoic acids, the volume percentages
were selected to give the same volumes as used for the barium naphthenate
and barium 2-ethyl hexoic runs, while the volume percentage for toluene
was selected to give the same volume percentage of the proprietary hydro-
carbon additives. The results of the evaluation runs with the 2 percent
sulfur residual oil are incomplete at this time. However, cursory
analysis of the data indicate approximately the same level of additive
effectiveness with the 2 percent sulfur residual as with the one percent
residual oil.
Stat1stica1 Cons iderat ions
Table 2 includes the particulate loadings for both the daily
baseline run and the additive run made that same day. These daily base-
line runs were used as a measure of the variation in particulate emissions.
As discussed earlier, the ash loading for every run should be nearly the
same and, accordingly, and deviation in the baseline particulate loadings
from day-to-day would have to be attributed to the deviations in the
carbon portion of the particulate loading . The standard deviation of
III-95
-------
the particulate loadings for the baseline runs of Runs 6 through 31 .
is 10 mg/sm , less than 10 percent of the arithmetic mean of 104
2
mg/sm of these runs. Consequently, each additive was evaluated by
comparing the particulate loading of the specific additive run with
the arithmetic mean of the particulate loadings of all daily baseline
runs. This method minimizes the effect of random variations in the
particulate emissions but may bias the results somewhat in that addi-
tives with low (or high) daily baseline particulate loadings (assuming
these observed loadings are "real" rather than instrument or measurement
errors) appear more effective (less effective) than if the particulate
loading of the additive run had been compared with the specific daily
baseline loading. However, regardless of which method of comparison
is used, the conclusions would be similar.
Figure 2 graphically shows the effectiveness of the proprietary,
organo-metallic, and organic additives in reducing filterable particulate.
The arithmetic means and the estimated ash component of the particulate
loading are shown. The data indicate, in general, those additives con-
taining certain alkaline-earth or transition metals in concentrations
of about 20 to 50 ppm in the oil significantly reduced particulate
emissions. The use and effectiveness of these additives at higher con-
centrations is somewhat questionable. It has been observed that even
though carbon burnout can be further promoted at the higher concentra-
tions, the net reduction in particulate is not substantial because of
the additional particulate from the metallic inert contained in the
additive.
* The level of 100 mg/m was determined from the standard error of the
mean of the baseline runs.
ill-96
-------
Also shown in Figure 2 is a line located at two standard
deviations below the baseline average. This standard deviation is
based on the individual baseline measurements and is regarded as
equally valid for the particulate measurements with additives.
Although the use of two standard deviations strictly requires the
assumption of normality to correspond to a 95 percent confidence
level, it is used here in a broad sense to indicate a reasonably high,
but unknown, confidence level. Those particulate measurements for
additives falling below the line are concluded to be statistically
significant (not due to chance) at this reasonably high confidence
3
level. Those additive runs falling above this level (84 mg/m ) but
3*
below 100 mg/m suggest that these additives are effective, but more
experimental data are required to establish the level of statistical
significance.
In addition, the measurements indicate that some additives
are more effective in reducing particulate than others. However, by
3
assuming that a standard deviation of 10 mg/m also applies to the
additives runs, a statistical treatment of the data would indicate that
the effective additives are not statistically distinguishable.
Effect of Carbon Particulate
For the baseline boiler operating condition of 80 Ib/hr (36 kg/hr)
* 3
The level of 100 mg/m was determined from the standard error of the
mean of the baseline runs.
111-97
-------
at 12.8 percent C0_, the average carbon content of the filterable
participate was observed to be about 50 percent for a corresponding
filterable particulate loading of 104 mg/sm . Accordingly, additives
can only reduce particulate loadings by about 50 percent to about
3
50 mg/sm which represents the ash component associated with the
inerts of the fuel oil, unless the use of additives caused a larger
portion of the particulate to remain within the combustion system
(this would probably be undesirable as it would foul heat transfer
surfaces).
This observation that particulate loading can be reduced to
about 50 mg/sm is verified by the results of Runs 13, 14, 32, 33, 34,
and 35 in which the boiler was fired at 5 boiler loads. Figure 3 shows
the particulate loading generated by the boiler operated at each of 5
loads with and without one additive. For the baseline runs, particu-
late loadings were higher at the 100 percent load (Run 32) and the 35
percent load (Run 35) than at the 80 percent load (Runs 13 and 14),
65 percent load (Run 33), and 55 percent load (Run 34), but with the
use of barium naphthenate, particulate loadings were reduced to about
3
the same level of approximately 55 mg/sm , regardless of load. This
particulate loading is about 10 percent higher than the anticipated
ash level when firing the referenced fuel alone and can be attributed
to the increase in metallic inert of the oil by addition of the barium
naphthenate. Analysis of the particulate loadings of these runs (except
Run 32) indicated that over 85 percent of the particulate was ash.
111-98
-------
Additive Type and Concentration
The data obtained here indicate that some additives containing
metals substantially reduced particulate emission. These additives were
used in concentrations corresponding to the addition of between 20 ppm
and 5fl ppm of metal to the fuel oil. Proprietary additives considered
organic with some trace metals also were shown to be effective at con-
centrations that corresponded to the addition of at least 15 ppm of
metal to the fuel oil. As in the case of the additive "Formula LSD",
these concentrations were significantly higher than those recommended
by the manufacturer.
In general, the organo-metallic compounds listed in Table 2
were effective in reducing particulate except for two additives in the
subclass cyclopentadienes (CI-2 and PD 1654, a ferrocene derivative)
and zinc naphthenate. The proprietary additives (CI-2, Runs 4 and 20)
and PD 1654 (Run 5), although observed to be effective in reducing
particulate in other types of combustion systems firing distillate
oil , were found to be ineffective in the 50-hp (500 kW) boiler
firing residual oil; their ineffectiveness could be attributed to the
type of bonding structure of these compounds (for PD 1654, the additive
concentration used was that recommended by the manufacturer rather than
one based upon a 27 ppm metal in the oil). The data indicate that the
most effective agents are compounds of the alkaline-earth metals (barium
and calcium) and the transition metals (manganese, iron, and cobalt).
111-99
-------
Zinc, an element just beyond the above transition metals, was found to
be an effective additive.
Some of the pure organic additives considered appeared effec-
tive in reducing particulate emissions. However, these additives had
to be used at impractically high concentrations to be marginally effec-
tive and, therefore, could not be considered as a viable pollution con-
trol method.
Effect on POM
POM (polycyclic organic matter) measurements were made on a
limited number of additive runs. These measurements included both total
POM and species quantification for over 15 compounds (the majority of
which are carcinogenic) as listed by the National Academy of Science .
In general, additives that were effective in reducing particulate appear
to be effective in reducing POM. Even if more data were available, it
is improbable that any specific relation between additives and POM
reduction could be made. It is believed that subtle changes in the
combustion process that cannot be detected by the gross measurements
of stack gas analysis have a significant effect on the formation of POM
that would tend to mask an additive effect, especially considering that
3
POM loadings were observed to be less than 10 p.g/sm for continuous
steady-state combustion. For these relatively small POM loadings, over
99 percent of the POM found was considered innocuous.
111-100
-------
POM measurements were also made during the cyclic operation
of the boiler in Run 36. Although this size range of boiler would not
be operated in an on-off cycle, but rather modulated to meet load, it
was felt that some useful data could be obtained by operating the
boiler for a 15-minute on/5-minute off cycle. As noted in Table 2,
particulate loadings increased by a factor of two for both the baseline
run and the barium naphthenate additive. However, POM levels of the
2
baseline run increased by over two orders of magnitude (920 jjg/sm )
above the steady-state runs. For this run the additive was signi-
ficantly effective in reducing POM (to 120 p,g/sm ) although the POM
emissions still were much higher than for steady-state runs. This
observation agrees with that of an early investigation in which it
appeared that POM was reduced to greater extent than particulate by
(4)
use of an additivev . At the higher POM loadings, several compounds
considered carcinogenic were found in appreciable quantities.
Effect on Smoke
As a measure of the effectiveness of additives in reducing
particulate over a range of air/fuel ratios, Bacharach smoke measurements
were made at about six excess air levels for every daily baseline run and
every additive run. Figure 4 shows the effect of cobalt naphthenate
(Run 7) and barium naphthenate (Run 13) on smoke reduction. In general,
the relation between Bacharach smoke and (XL for the daily baseline runs
as shown by the curve in Figure 4 remained constant from day to day as
would be expected from the small variations in particulate loadings of
these runs. Accordingly, the effectiveness of the additive could
III-101
-------
semiquantitatively be evaluated by comparing the smoke characteristic
curves with and without the additive. The relative differences of
these curves could be related to the relative differences in particulate
loadings at 12.8 percent CO . From Table 2, for Run 7, the ratio of the
particulate loading with and without additive was 0.71 and was 0.46 for
Run 13, at a CO level of 12.8 percent.
Effect on Efficiency
Figure 4 is a plot of stack-gas temperature as a function of
time for 3 long-term (60 hours) runs with the 2 percent sulfur residual
oil. Barium naphthenate was used at two boiler operating conditions;
one condition matched the excess air level of the baseline run while
*
the other condition matched the smoke density of the baseline run.
The results of these runs are somewhat inconclusive because
for the relatively short duration of these runs, it is difficult to
predict the stack-gas temperature for an extended period of time (such
as several months). It appears that the slope of the baseline curve
is leveling off after 70 hours, and it is possible that by projecting
the curves of all three runs, the curves would intersect. In all likeli-
hood, however, the slopes of the additive runs should also level off and
not intersect with the slope of the baseline run.
However, by assuming that the time-rate-of-change of the stack-
gas temperature (the slope of the curves) of all three runs is the same
after 60 hours, the effectiveness of barium naphthenate to increase
III-102
-------
boiler efficiency can be estimated. A reduction in either or both the
mass flow-rate and temperature of the stack gas will increase boiler
efficiency. For the barium naphthenate run at 14.2 percent CO , an
increase of boiler efficiency of about 2 percent is realized. About
half of this increase can be attributed to the lower stack-gas temperature
while the other half can be attributed to the lower excess air level.
POSSIBLE EFFECTS OF ADDITIVES
ON COMBUSTION MECHANISMS
To explain the differences in performance by the various metals
which achieved some reduction in the particulate emissions, it is necessary
to hypothesize as to the mechanism by which they act. The transition
metals--iron, manganese, and cobalt are known to be oxidation catalysts
in a variety of chemical systems. Consequently, their activity could be
attributed to catalytic enhancement of the carbon burnout in the fly ash.
Differences among these metals could result from the degree of catalysis
displayed by the particular metal. However, in these experiments no
statistically significant difference was found in the performance of these
three metals.
The alkaline earth metals--barium and calcium—are not noted
for catalytic activity, so the mechanism by which they reduce particulates
is very likely different from that of the transition metals. It has been
(13)
suggested by Friswell that alkaline earth metals may function by
promoting the dissociation of water to H atoms and OH radicals, a process
III-103
-------
occurring only at high temperature. The OH radicals are involved in
the combustion of carbon:
C + OH - CO + H
CO + OH -* C02 + H.
The H atoms in turn react with oxygen in the combustion air to form
additional OH radicals and 0 atoms, namely, 0 + OH -* H + 0.
For differences between various compounds of the same metal,
the bond strength between the metal and the remainder of the molecule
may be the controlling factor. This aspect was explored most completely
for barium. The number of available materials was limited by the necessity
that they be soluble in the oil. In this investigation, the naphthenate,
hexoate and the sulfonate were used, all of which are barium salts of
organic acids. Although some differences in performances of these barium
derivatives can be noted in Table 2, more experiments are needed to make
a statistically valid evaluation. In the cases of iron and manganese,
the situation is similar, although the compounds used cannot be as
readily compared. These mechanistic considerations can be used to
explain differences, but the possibility should not be eliminated that
other properties such as volatility of the additive and the nature of
the oil may influence the effectiveness of the additive in reducing
particulate emissions.
III-104
-------
PERSPECTIVE ON FUEL ADDITIVES
In this study, additives containing certain metals used at
concentrations of approximately 20 ppra to 50 ppro of metal in the fuel
oil were found effective in reducing particulate emissions. The results
of this study, for the most part, have confirmed the observations of
other researchers and have extended the range of investigation of pro-
prietary additives and pure compounds to commercial size combustion
equipment firing residual fuel oil. It should be pointed out that the
mechanisms involved are not fully understood and there is evidence to
indicate that additives effective in one system firing one type of fuel
may not be effective in another system or when firing a different fuel.
Before any general or widespread use of fuel additive can be
recommended, a perspective should be developed for each type of boiler
installation to consider:
• Possible toxic effects of pollutant emissions
resulting from materials introduced with the
additives
• Practical considerations of additives use (such as
economics, corrosion resulting from use and increase
in other pollutants) other than simply the reduction
of one or more pollutants, and
• Comparison of the effectiveness of additives with
alternative control means.
It is anticipated that there should be no serious economic penalty with
the use of additives as some of the pure compounds are relatively in-
expensive. In addition, it should be noted that the level of iron
III-105
-------
introduced as iron naphthenate was less than the iron present in the
oil as ash so that use of iron containing additives for this oil may
not introduce any new pollutant. However, the long-term effects of
additive use are not well-established nor are the nature and potential
toxicity of these metallic compounds adequately defined.
ACKNOWLEDGMENTS
The research upon which this publication is based was performed
pursuant to EPA Contract No. 68-02-0262 with the United States Environ-
mental Protection Agency, Division of Control Systems. The authors wish
to thank EPA for granting permission to present this material, and to
EPA staff members, G. B. Martin and W. S. Lanier who participated in
planning this program and have provided helpful comments. The authors
also thank other members of the Battelle staff who have contributed to
this study—A. E. Weller, A. Levy, P. W. Jones, L. J. Hillenbrand,
R. E. Barrett, R. Coleman, T. C. Lyons, and H. G. Leonard--for their
advice and assistance in carrying out the research.
Ill- 106
-------
REFERENCES
1. Martin, G. B., Pershing, D. W. , and Berkau, E. E., "Effects of Fuel
Additives on Air Pollutant Emissions From Distillate-Oil-Fired
Furnaces", U.S. Environmental Protection Agency, Office of Air
Programs, Publication No. AP-87, June, 1971, pp 86.
2. Weeks, R. L., Clinkenbeard, W. L. , and Soltis, J. D., "Clean,
Efficient Combustion of Heating Oil", Proceedings of the 5th World
Petroleum Congress, New York, Section VI, Paper 25, June, 1959,
pp 381-396.
3. Riggs, R. J. Wilkinson, C. J., and Wolfe, H. R., "Combustion Im-
provers for Fuel Oils", Erdoel and Kohle, 18 (4) 282-286 (1965).
4. Perilstein, W. L., "Overbased Barium Sulfonates for Use in Distillate
Fuel Oils", U.S. Patent 3,580,707, May 25, 1971, 8 pp.
5. Hangebrauck, R. P., von Lehmden, D. J., and Meeker, J. E., "Sources
of Polynuclear Hydrocarbons in the Atmosphere", HEW, A.P. No.
999-AP-33, Durham, N.C., 1967.
6. Committee on Biologic Effects of Atmospheric Pollutants, "Biologic
Effects of Atmospheric Pollutants; Particulate Polycyclic Organic
Matter", National Academy of Sciences, National Research Council,
Washington, D.C., 1972, pp 375.
7. Ray, S. K., and Long, R., "Polycyclic Aromatic Hydrocarbons from
Diffusion Flames and Diesel Engine Combustion", Combustion and
Flame, 8, 139-151 (1964).
8. Spengler, G. and Haupt, G., "Formation of Soot and Polycyclic Aromatic
Hydrocarbons in Simple Hydrocarbon Flames and Its Inhibition by Fuel
Additives", Erdoel un Kohle, 22 (11) 679-684 (1969).
9. Pershing, D. W., Martin, G. B., Berkau, E. E., and Hail, R. E.,
"Effectiveness of Selected Fuel Additives in Controlling Pollution
Emissions from Residual Oil-Fired Boilers", U.S. Environmental
Protection Agency Report No. EPA-650/2-73, October, 1973.
10. Rendle, L. K. and Wilson, R. D., "The Prevention of Acid Condensa-
tion in Oil-Fired Boilers", J. Institute of Fuel, 29, 372 (1956).
11. Jones, P. W., Giammar, R. D., Strup, P. E., and Stanford, T. B.,
"Efficient Collection of Polycyclic Organic Compounds from Combustion
Effluents", presented at the 68th Annual Meeting of the Air Pollution
Control Association, Paper No. 75-33.3, Boston, June 15-20, 1975.
Ill- 107
-------
12. Standard Method of Test for Smoke Density in the Flue Gases
from Distillate Fuels, ASTM Designation: D 2156-65 (Reapproved
1970).
13. Friswell, N. J., Emissions from Continuous Combustion Systems.
edited by W. Cornelius and W. G. Agnew, Plenum Press, New York,
1972, p 161.
111-108
-------
TABLE 1. ADDITIVES CLASSES
ORGANO-MEX
Naphthenates
Cobalt
Iron
Manganese
Barium
Iron + barium
Zinc
Calcium
Cvclopentadienes
Manganese - CI-2 (18% Mn)
Iron - PD 1654 (1% Fe)
Sulfonates
Barium
Ethyl Hexoate
Barium
Carbony|s
Iron
ORGANIC
Hydrocarbon
LSD (C,H0, trace metals)
/ o
Improsoot (C_H0, trace metals)
/ o
Triple E
Toluene (C?Ha, pure hydrocarbons)
Watcon 130 (trace metals)
Ether
Diglyme
Alcohol
Ethyl
Hexyl
Acid
Napthenic
Ethyl hexoic
MC-7 soluble (57=, Ba, 2% Mn)
Rolf it e 404 (27, Mn)
III- 109
-------
xuiu a. suwm or EVALUATION KIMS
torn
Additive
t>OM»«
fattlculete toadIng
Maltlve Basel in*
• mrttm*
rartleulate
loading
»atlo(e)
9
10
11
12
13
14
IS
16
1?
1ft
19
20
90
31
•• Ittltlil Evaluation Kuns Vith Proprietary Additives
HJ-7 soluble (Apollo)
ronula tSD (Coisacrelal Chcnleal)
iBprotoot (Comercial Chcaical)
CI-2 (Ethyl Corporation)
10 1654 (Arapahoe)
35 pp. to. U pp. tt>
3 ppi Co. * Ppm Mn
9 pp. Co, 2 ppmCa, 10 ip.lta
21 pp. Ha
5pp*r«
b. Evaluation Runs Vtth Oreano-Metallic Additives
Zinc naphthenate
Cobalt naphthetut*
Iron B*phth«Mte
Iron a*phthcnac«
Hug*ne*c Mphchenate
Calelua Mphthcnatc
Iron and barium n»phth«n»t«
Bariua n*phthctue*
lariua naphtheiuc*
laxim naphthcMt*
••elm n«phth«i*te
Bcrlua 2-ethyl
Barium tulfocxtt
Iron pent*ttrbonyl
Cl-2 (Ethyl)
21
22
23
2*
25
2*
27
28
2*
Ethyl hexolc *cld
Itopthcnle *cid
Ethyl alcohol
•cxyl alcohol
Digly»e
Tolucn*
t-4000 (Triple E Product!)
. t-«000 (Triple E Products)
ronul* LSD (CcxmwrcUl Ch
cvical)
Vatcon 130 (Induttrial Chemicals)
kolfit* 404 (Andrew Rolfe Chemical Co.)
27 pp. ta
27 pp. Co
27 pp. F.
iipp-r.
27pp.Mn
27pp.Ca
27 pp. Fa, 27 pp. ft
31 pp. to. 0.051 by
51 pp. to
26 pp. to
10 pp. to
40 pp. to, .031 by voi
20 pp. to
27pp.ro
21 pp. to
Evaluation Runs With Organic Additives
0.03X by volume
0.05X by volume
1.5X by voluw
l.Sl by volume
1.5X by voluo*
0.2X by volume
0.031 by volume
0.21 by voluw
0.2X by voluae, 16 pp. Co,
• pp. MB
0.21 by volume, 16 ppn Fe
0.21 by voluae, 6 pp. Mn
51
77
4*
100
M
106
74
W
71
77
62
35
4$
50
75
•9
55
•3
7i
97
102
•3
1«1
79
90
114
118
101
69
66
86
•1
80
90
86
77
115
109
127
108
106
115
108
117
92
102
104
93
98
103
103
102
91
120
86
104
109
115
110
100
87
102
0.61
0.9)
0.38
1.20
1.08
1.02
0.71
0.77
0.68
0.74
0.60
0.53
0.46
0.48
0.72
0.86
0.53
0.80
0;76
0.93
0.98
0.82
0.97
0.74
0.87
1.03
1.13
0.97
0.66
0.63
0.84
4. Evaluation Run*With Variable Boiler Lead at Constant Air/fuel Ratio
Particular Loading
1HH
32
33
34
35
la
36 •
Additive
Barlm naphthenate
ditto
•
AJa-tttve
toxlua naphthenate
Dosage
51 pp.
51 pp.
51pp.
51.PP«
*. EvsltMt
Dosage
51 PP-
tiring Hate
Ib/hr
nc/srnJ
98 47
66 53
54 31
36 55
ion Run Vlth Cvclie-Boller
Partlrul.ite
Additive
134
Loading
Baseline
206
i^.Kaseline
148
115
. 105
147
Operation
(•) The ratio of the part leu late loading of the additive run and the arithmetic .eaa of th«
paniculate loadings of the appropriate baseline runs.
(»> Ib/hr - 0.454 ';c/hr.
IH-110
-------
FIGURE 1. EXPERIMENTAL 50-HP BOILER FACILITY
III-111
-------
[XXXXXXXX 1X\\\X\X\
10
a
mean
hmet
t
is
¥^ft
\\X\\'
5
\ \\X\\\\
k\\\\\
xx
\
LXXXXXX
\
xxxxxx
^v
\\\
fc
XXXXXXXX
\\x\\\\\\\\\\\
\xx\x\\\
§s
\\
•v^Y^
\\x\\\\\\\\\
0
if)
Q
K)
O 04
O O
IMENTAL DATA FOR
LOAD AND 12.8 PERC
EXPER
SUMMARY OF THE
RUNS AT 80 PERCE
E 2
FIG
£uus/6uj '
III- 112
-------
160
[40
120
o»
•o
o
o
O
O
0_
100
80
60
40
20
1 (Baseline
-
—
-
-
V///\ Barium Nopthenate
T5
O
51?
O
g
7]
/
/
X3
~~
ft
b
CO
\
f
^^^
^
o
5
. o
0s-
O
CO
\
•o
0
0
^o
in
ID
\\\X\\\\\\\\\\\\VI
^^H
T^J
o
—
,0
o
m
7
I
I
x
0
Q
0^
m
rO
!
32
14 32
Run Number
34
35
FIGURE 3. THE EFFECTIVENESS OF BARIUM NAPHTIIKNATE
IN REDUCING PARTICIPATE AT VARIABLE
BOILER LOAD FOR CONSTANT AIR/FUEL RATIO
III- 113
-------
490 (254)
480 (249)
470 (243)
~ 460 (238)
5 450 (232)
440(227)
o
o
430 (221)
420 (216)
410 (210)
400 (204)
Baseline, 12.8%
C02
iorium napththenate
(51 ppm Bo)
12.8% C02 .
Barium napththenate
(SIppmBa) 14.2 %C02
10 20 30 40 50
Time, hours
60 70 80
FIGURE 4. STACK-GAS TEMPERATURE AS A FUNCTION OF TIME
FIRING 2-PERCENT SULFUR OIL AT 80 LB/HR
IH-114
-------
5:15 p.m.
The Effect of Additives in
Reducing Particulate Emissions
from Residual Oil Combustion
Robert D. Giammar, Battelle Columbus Labs.
Could you elaborate some more by what you meant by
your cyclic operating conditions and why the POM
levels were so high?
The boiler was operated in a 15 minute on, 5 minute
off fashion for 10 cycles. The particulate measure-
ments were made during this period. We suspect that
the POM levels were higher during the cyclic mode
of operation than the continuous mode of operation
because of the characteristic of burner shutdown
and startup. During these periods there is poor
fuel atomization and some of the oil will get through
the vaporized field partially unburned. We found
out that our concentrations of POM in raw fuel oil
were orders of magnitude higher than the concentra-
tions that we have seen in the stack gases from con-
tinuous combustion, and so any unburned fuel collected
will give a high POM level.
You had a slide up there which showed stack gas
temperature on the vertical axis vs. time for no
additives and two different additives. I noticed
for one of the additives, at time zero there was
significantly lower temperature than the other two,
Should all the stack gas temperatures start at the
same levels?
The air/fuel ratios of these curves were different,
In comparison to the two top curves that were for
20% excess air, the lower curve corresponded to
10% excess air. A lower percentage of excess air
111-115
-------
will give lower stack gas temperature.
Q: That was at a common smoke number.
A: Yes,it is a common smoke number. The lower curve
with the additive was a number 4 smoke which was
the same as the baseline.
You had mentioned that you did a great deal of
testing not only with pure compounds but also with
a large number of proprietary additives. If you
were asked to make some suggestion, if one of these
manufacturers or users were to use a fuel additive,
would you care to comment on what your suggestions
would be in terms of a particular additive? Should
they go to a pure compound? Should they buy a
proprietary additive? And what are the economics
involved ?
In our study we found that the alkaline earth metals
and the transition metals were effective in reducing
particulates or carbon particulates. These were
procured in pure compounds... the naphanates, the
hexoates...which were significantly less expensive
than some of the proprietary additives.
Is any work being planned on the toxicology of the
particulates?
It's not within the scope of this program.
Q: Would you care to hazard a guess on average residual
oil and average set of conditions of operation, the
cost of treatment per thousand gallons or a range
of possible costs?
III-116
-------
A: I don't have those numbers with me. I don't have
the specific costs but I made a calculation based
upon the following: if you could save a dollar in fuel
costs by the use of additives with an increase of
efficiency of a couple percent, then you would have
to take 20% or 20 cents of that dollar to purchase
your additive. I do want to mention, however , that
we had to detune this boiler. This boiler could be
tuned so that it generated very little carbon particu-
late and there would have been no point in doing
our studies with no carbon particulate. We tried
to establish a smoke characteristic curve that was
typical of that found in the field, and we relied
upon some EPA field study data for this purpose.
Sawyer and co-workers out at Berkeley have shown
that several of these additives can significantly
affect the particle size distribution by shifting
them to lower sizes. Did you try at all to measure
•
the size distribution and,based upon your experimental
techniques,would you expect such a shift to alter
your conclusions?
We did not make any particle size measurements at
this time.
Did you run a complete metal analysis on your
fuel to find out if you had any substantial trace
metals present?
Yes. Trace metals were low, less than 150 ppm.
The question of how these metallic inerts present
in fuel affect the formation of carbon particulate
was of concern to us. Some of these trace metals
can be present in the oil at higher concentrations
than are introduced into the oil with the additive,
III-117
-------
Apparently, these trace metals are not nearly as
effective as the metallic additives. We suspect,
however, that vanadium in some oils might have
some effect.
Q: Would you care to rank your alkaline earth metals
and your transition metals in terms of their
effectiveness as seen by your test?
Q: Statistically, they are equivalent. We just don't
have enough data.
Q: Would that be on a molar basis or a weight basis?
A: On a weight basis. It appears that the alkaline
earth metals were more effective than the transition
metals. Calcium was our best performer, but on a
molar basis, calcium and barium gave about the
same results.
Several investigators in premixed laboratory flames
and diffusion flames have noticed the effect of
additive concentration on either promotion or
inhibition of soot formation. Did you notice
any effect on concentration of your additives?
Yes. In our studies we noticed that increasing
the metal concentration in the oil almost always
decreased the carbon found in the particulate
catch. For example, the metallic cyclopentadiennes
reduced carbon particulate by only 30% at concentrations
of .1 g of metal per gallon of oil but reduced carbon
particulate by over 90% at concentrations of 1 g of
metal per gallon of oil. However, the resulting total
particulate loading would increase.
Ill- 118
1
-------
SYSTEM DESIGN FOR POWER GENERATION
FROM LOW BTU GAS
By:
M. P. Heap and T. J. Tyson
Ultrasysteins, Inc.
and
N. D. Brown
Combustion Engineering
The work reported in this paper was carried
out under EPA Contract 68-02-1361
III-119
-------
1.0 INTRODUCTION
The conversion of coal to a clean fuel gas is one attractive
method of utilizing the nation's coal reserves in an environmentally
acceptable manner. Alternate approaches include the use of fluid bed
coal combustors or conventional coal-fired boilers with flue gas
desulfurization. This paper addresses the question of how existing
technology might best be used to generate electric power from gasifier
off-gas. The total study has yet to be completed, therefore this
paper represents a partial progress report. The primary objective of
the study is to identify those areas which are likely to give rise to
emission problems in the design of a 500 MW power plant fired by low
Btu coal-derived fuel gas. Having identified problem areas and assessed
their magnitude, it is the intent to suggest design modifications to
allow these problems to be overcome.
The design study is based upon the provision of a 500 MW power
plant with the following characteristics:
- low plant heat rate;
low N0x emissions (0.1 Ibs N02/106 Btu fired);
- high turndown ratio (5:1);
- low capital cost;
low operating cost.
The restrictions placed on the study prevented any consideration of
integrated power plant/gasifier systems and it was required that the
economic analysis be conducted on the basis of purchased coal gas, neces-
sitating an assessment of an "over the fence" cost.
III-120
-------
2.0 SYSTEMS STUDY
2.1 Fuel Characteristics
The systems study was carried out on the basis that two general
classes of fuel gas can be considered as representative of existing
technology:
- a low Btu gas typical of that produced by a Lurgi
air blown gasifier;
a medium Btu gas typical of that produced by an oxygen
blown Koppers Totzek (K-T) system.
The compositions of these two fuels are presented in Table 1.
Table 1. Fuel Gas Compositions for the Systems Study
(After 1)
CO
H2
CH4
co2
TT f\
EL*\v
N2
LHV (Btu/scf @
60°F 1 atmos)
Low Btu
(Lurgi)
14.1 %
20.9 %
5.8 %
12.5 %
6.6 %
40.1 %
160
Medium Btu
(K-T)
53
36.4
-
9.25
0.3
1.05
272
In addition to variations in heating value it was also considered
that the fuel gas may be available at high pressure (in excess of
10 atmospheres) and high temperature (approximately 1500°F). It has been
III-121
-------
shown that coal gas combined cycle power plant systems only have an
overwhelming economic advantage over conventional plants if advanced
turbine concepts allowing turbine inlet temperatures of 2600 F can be
(2)
used' . High temperature/pressure fuel gas was included in this study
for purposes of comparison, although the assessment of fuel cost pre-
sented a problem.
2.2 Systems Considered
Over forty conceptual power plant systems were considered which
were capable of using the fuel gases discussed above. These concepts
can be classified into four groups:
1. Conventional boiler systems with low temperature fuel gas or
systems which include either a fuel gas precooler or high
temperature burners to utilize high temperature fuel gas.
2. Combined cycle systems of the waste heat and auxiliary
fired steam boiler type.
3. Supercharged boiler systems.
4. Oxygen-fired boilers - if oxygen is available in bulk
for the gasification process then it could also be
utilized in the power plant.
Each of these four general classes includes a wide spectrum of system
concepts. After all the concepts had been studied, five were chosen
for more detailed consideration. These systems are listed in Table 2.
System Al was chosen because it represents a baseline case suitable for
retrofit. Waste heat combined cycle plants were not considered because
(2)
they were being studied in detail in another EPA-sponsored project .
Ill- 122
-------
i
3
§ 2
o— —
O •
U
§9
u
3
<
oc
at
4J
OJ
a
•a
-
zz ~
23 £
|Si§
1 . sis
CM
a)
o
m
g 2
« I
? &
£
Ul
y
111
o
u
111
III- 123
-------
Ill-124
-------
CROSS SECTION OF
PROPOSED BURNER
IGNITER
Figure 1. Sketch of High Pressure Boiler
Showing Detail of the Burner
III-125
-------
i—
i
o:
—VV
^
I
_ J
£U
V)
<5'
.8.
A
— '
~A-
-J
-A
tn
M
O
*a
01
-------
has a maximum continuous rating of 612,000 Ibs/hr main steam flow with
superheat and reheat steam conditions of 1011°F/2685 psig and 1001°F/
501 psig respectively. Superheat and reheat steam temperature control
is achieved via desuperheat spray. Flue gas leaves the boilers at 2000°F
and is delivered to three gas turbines; heat losses from the system are
minimized by routing the air from the compressors to the boilers in an
annulus surrounding the flue gas pipe. Upon leaving the gas turbines,
the flue gas passes through a waste heat recovery chain consisting of an
evaporator-superheaters, economizer and a feedwater heater. The flue
gas at 300 F then passes to the stack. The three gas turbine generator
sets deliver a total net output of 191 MW; a single steam turbine genera-
tor delivers 308 MW net.
Calculated plant performance for each of the five systems is
summarized in Table 3.
2.3 System Economics
Capital cost estimates have been prepared for each system
assuming all plants were outdoor-type located at an average U.S. site
using the standard Federal Power Commission account numbering system.
The performance and operating costs for 1974 operation are summarized
in Table 4, and the terms used in this table are defined in Table 5.
The gross electrical output of each plant is approximately 520 MW and
has been corrected for typical auxiliary power requirements for
III- 127
-------
§
o
5 m ui
8
8
-> tt.
z
§
in
O
O
.
.^ 8
8
in
8
m
• SJ 2
m
in
pi
CO
H
~
<3
oo
fc JOS
o
t-
u. zui
«O
III- 128
-------
c
o
5*:
Ou
OtVMXJ i O
B- :s
... in * *
0 — 0 «o o
in 01 o> w
r«- * f\j ru W
M •- <«• —
ni cu
oh- tntv
mm ~f-
Oin r»is>
v ^
O
CJ
S
Q o com
O ^«*
CM N
8
ota
— o
CO
VI
V
P.
O
vr
r>-
a\
OQ
- 22
O* ^ ftj OO
v ea inSS
llil
toe
to.
>O * f- O
eo • in nj
so "" « 2
•-
U
U. 5
K m 0.00
< CK
a: D
K H u
2< t
~ O ^~
X
ce
SK
K v> -J ?<
!»! an: > 5 X £
z_>.>. t ^ ri i~
? y^gg
UJ =1 H
z z «
< l-l-
z
x^ go 5a
Q. '^1 Zw Z =
< D Z»- Z<
U it. « <2
K
§x
*
* §
?SS I9S8
OQ.
•-O
OHH
»"*
til
III-129
-------
•O
V
n
-------
condenser cooling pumps, boiler circulating pumps, FD, ID and GR fans.
Transformer losses are 0.5 percent of the net output. The boiler feed
pumps are steam turbine driven and are included in the steam cycle.
The net plant heat rate is the gross (higher heating value plus sensible)
fuel heat input divided by the net output.
Annual net generations are the net full load output times 5000
and 7500 hours per year. These are equivalent to load factors of
57 and 86 percent. A fixed charge rate of 20 percent of capital cost
per year is assumed as typical for an average U.S. site. Annual fuel
costs for 1974 are based on $1.25/10 Btu as the delivered price for
10 psig Lurgi LBG and $1.40/106 Btu for 147 psig Lurgi LBG. The
delivered price for Koppers Totzek MBG is $1.50/10 Btu; plant 15 has
an additional cost of $0.55/10 Btu for oxygen for the boiler, resul-
ting in a total of $2.05/10 Btu. In order to arrive at a realistic
cost of the high pressure Lurgi gas, an evaluation of the relative
gasifier capital costs for low versus high pressure fuel gas yielded an
increased cost of $0.15/10 Btu for the high pressure gas. This is
associated with the increased auxiliary power requirements for the
gasifier because the product gas expanded to 10 psig.
2.4 Discussion
Table 6 summarizes the plant costs and performance of the five
power plant systems considered suitable for the application of existing
III-131
-------
01
4-1
in
o
o
00
c
cd
s
a
o
T3
C
rt
0)
o
d
o
U-4
M
(U
PLI
C
nJ
vD
(U
t— 1
H
^^
<:
,—1
33
i— 1
O
1—4
OQ
r~f
•<
3
pQ
§
Z
H
£5
A
O
o
lA
^
s
H
(^
g
O
H
3
CH
S
^
O
£-4
55
O
CM
ON
00
O
— i
CM
^«4
^
*^
CO
CM
^-1
vD
00
f-^
^O
F-H
^
1-^
o
«-4
CM
CM
rH
rH
f^
fVfc
LA
CM
*— i
f^»
st
ON
!A
CM
t— 1
CO
1
O
i— i
X
H
C/l
O
O
9
H
h^
to
w
^J
^J
f— 1
g
CO
CM
CO
^^
CM
^
CO
CM
N^
«*
CM
CM
lA
CM
«M
lA
CM
J3
S
H
W
^
•co-
co
CO
lA
CM
00
ON
r^b
O
O
i-H
c-
O
z
M
t—l
Z
O
M
to
C/1
M
§
a
H
^2
3
H
W
^^
CO
•
0
4-1
si
III-132
-------
technology to the utilization of gasifier off-gas for power generation.
The supercharged boiler combined cycle plant (HI) has the lowest capital
costs per KW and the lowest net heat rate which results in the lowest
fuel costs. The three other Lurgi LBG plants (Al, Bl and Cl) have
nearly identical unit energy costs approximately 6 percent higher than
Plant HI. Plant Al for the Koppers Totzek fuel has unit energy costs
approximately 16 percent higher than Plant HI; the cost for this plant
exceeds that of the other Lurgi plants due to the higher fuel cost.
Koppers Totzek plant 15 energy cost is approximately 36 percent higher
than Plant HI; the high energy cost of this plant is a result of the
use of oxygen in the boilers.
Although the supercharged boiler combined cycle system has both
the lowest capital and energy costs, the advantage is hardly over-
whelming. Supercharged boilers have been used throughout Europe,
but their use in the States has been restricted to marine boilers.
(3)
Robson had earlier considered the use of supercharged boilers and
rejected them for the following reasons:
- the development costs associated with their construction
was likely to be too high;
as permissible turbine inlet temperatures increase, the
efficiency advantage of the supercharged boiler system
disappears.
(4)
In a study conducted for the Office of Coal Research Com-
bustion Engineering compared the cost of conventional steam plants
111-133
-------
with an atmospheric gasifier and supercharged boiler combined cycle
plant with pressurized gasifiers. In this study which considered an
integrated gasification and power generation plant, the supercharged
boiler concept had a 3.0 Mills/KW energy cost advantage.
Even if a supercharged boiler can be considered as existing
technology, there does not appear to be a clear-out financial incen-
tive to recommend its use. This is compounded by the environmental
considerations presented in the following section clearly indicating
that further detailed study is required before the supercharged boiler
combined cycle system can be considered to be an acceptable method of
power generation from coal gasifier off-gas.
3.0 ENVIRONMENTAL CONSIDERATIONS IN THE COMBUSTION OF
GASIFIER OFF-GAS
The systems study discussed in the previous section was based
upon the delivery of a fuel gas almost free from sulfur (i.e., less
than that which would exceed S0_ emission standards) and particulates.
Since the gasification process is not included in the study, the major
area giving rise to environmental concern is the formation of NO
during the combustion of the low Btu gas. The kinetic mechanisms con-
trolling the rate of NO formation in combustion processes are outside
JV
the scope of this paper. However, it is known that the amount of
NO formed from molecular nitrogen is strongly temperature dependent,
A.
Thus, thermal NO formation is unlikely to be a problem in any of the
systems discussed earlier (perhaps with the exception of system HI)
III- 134
-------
provided the fuel gas is burned with air or oxygen and sufficient
quantities of recycled cooled combustion products. The emission
levels shown in Table 6 were calculated by the NO Emission Subroutine
J X
of the Lower Furnace Program . For tangentially fired units there
is sufficient experience to believe that these emission levels will be
within acceptable predictive limits. The Lurgi fired systems are so
below the target emission level that NO control techniques need not
be considered. This does not apply to the K-T fired boiler (Al), the
higher heating value and higher flame temperatures necessitate some
type of control to meet the 0.1 Ibs NO /10 Btu target.
Predictions for the supercharged boiler must be considered as
less acceptable. It should be recognized that these predictions were
made with a model tuned to a particular furnace design. The prediction
can be considered to be suspect because
- the upfired boiler may give a very different heat flux
distribution from that expected based upon tangential
firing experience;
it may no longer be permissible to treat the furnace as
a series of horizontal strips of uniform properties;
the lower furnace model does not include the capability
of accepting pressures above atmospheric.
One other real concern is the possibility of trace quantities of
ammonia in the fuel gas. It is known that ammonia is one of the pro-
ducts of most gasification schemes and that it will be oxidized to NO
HI- 135
-------
during the combustion process. The emission levels indicated above
did not include the possible conversion of ammonia contained in the
(2)
fuel gases. The best available estimates indicate that after low
temperature H_S clean-up the fuel gas may contain 500 ppm of ammonia.
This could produce as much as 200 ppm of NO in the combustion pro-
ducts, depending upon the fuel-air mixing process within the combustor.
Existing processes for high temperature H S removal do not scrub
ammonia and the high temperature gas may contain as much as 4000 ppm
of ammonia which would give NO levels in excess of the N.S.P.S. (New
Source Performance Standards) unless appropriate precautions were
taken.
4.0 CONCLUSIONS
A systems study has been conducted to assess the most appropriate
use of existing technology to produce electric power from low Btu gas in
an environmentally acceptable manner. Fired gas turbines could ulti-
mately lead to the most economically advantageous use of this fuel pro-
vided turbines can be built to operate satisfactorily for prolonged
periods with inlet temperatures of 2600°F. The present study indicates
that a supercharged boiler unfired gas turbine system provides the best
method of utilizing existing technology to convert coal-derived fuel gas
to electric power. The analysis discussed earlier gives slight economic
advantage for the supercharged boiler. However, it should be noted that
III-136
-------
there may be considerable economics associated with the design and
construction of an integrated plant. The net power plant heat rate for
this system is significantly lower than those of the other systems
studied, suggesting that a more detailed study of an integrated plant
might uncover a more significant advantage. There should be little need
for concern regarding NO production from low Btu gas provided it is
used cold and free from significant quantities of ammonia. The study
has highlighted the need for an analytical tool capable of predicting
combustor performance. Work is in progress to assess the NO emissions
A.
from a supercharged boiler fired by coal-derived fuel gas which could
contain up to 500 ppm of ammonia, and to identify those design param-
eters which will allow emissions to be controlled to a minimum without
impairing the thermal efficiency of the system.
III-137
-------
REFERENCES
1. Martin, G.B., "Environmental Considerations in the Use of
Alternate Fuels in Stationary Combustion Processes." (Symposium
Proceedings: Environmental Aspects of Fuel Conversion
Technology, EPA-650/2-74-118, pp 259-275; NTIS No. PB 238-
304/AS. (October, 1974).)
2. Robson, F.L., Blecher, W.A. and Giramonti, A.J., "Combined
Cycle Power Systems", (Presented at the Symposium on Environ-
mental Aspects of Fuel Conversion Technology II, Hollywood,
Florida, December, 1975.
3. Robson, F.L., and Giramonti, A.J., "The Environmental Impact
of Coal-Based Advanced Power Systems." (Symposium Proceedings:
Environmental Aspects of Fuel Conversion Technology, EPA-650/
2-74-118, pp 237-257; NTIS No. PB 238-304/AS. (October, 1974).)
4. Review and Evaluation of New Plant Gasifier/Power System
Alternatives, Task 4, Office of Coal Research, Project
No. 900129, Combustion Engineering.
5. Beuters, K.A., Cogoli, J.G. and Habelt, W.W., Fifteenth
Sumposium (International) on Combustion, p 1245, The Com-
bustion Institute (1075).
Ill-133
-------
SPEAKERS LIST
(NOTE: To facilitate their identification, speakers are listed alphabetically
together with the name of the organization they represent. The complete address
of each organization represented at the conference appears at the end of the list
of attendees.)
LIST OF SPEAKERS
Name
Axworthy, Dr. Arthur E.
Bittner, James D.
Bowen, Dr. Joshua A.
Bowman, Dr. Craig T.
Brown, Richard A.
Burchard, Dr. John K.
Cato, Glenn A.
Collom, Jr., Robert H.
Combs, L. Paul
Crawford, Allen R.
Cuffe, Stanley T.
Dykema, Owen W.
England, Dr. Christopher
Engleman, Dr. Victor S.
Siammar, Robert P.
Sail, Robert E.
Heap, Dr. Michael P.
'lelms, G. Thomas
:tollinden, Dr. Gerald A.
Kason, T.T.
Kendall, Dr. Robert M
Kesselring, Dr. John P.
Ketels, Peter
Lachapelle, David G.
Lanier, W. Steven
McDonald, Henry
Maloney, Dr. Kenneth L.
Manny, Srwin H.
Martin, G. Blair
Muzio, Dr. Lawrence J.
Pohl, John H.
I'oston, H. Wallace
Frinciotta, Frank
J.osenberg, Dr. Robert B.
Sarofim, Dr. Adel F.
Selker, Ambrose P.
Shaw, Dr. Robert
Shoffstall, Dr. Donald R.
Tyson, Dr. Thomas J.
Nasser, John H.
Wendt, Dr. Jost O.L.
Representing
Rockwell International, Rocketdyne Division
Massachusetts Institute of Technology
EPA, IERL, Combustion Research Branch
United Technology Research Center
Acurex, Aerotherm Division
EPA, IERL
KVB, Inc.
State of Georgia, Department of Natural Resources
Rockwell International, Rocketdyne Division
Exxon Research and Engineering
EPA, Office of Air Quality Planning and Standards
The Aerospace Corporation
Jet Propulsion Laboratory
Exxon Research and Engineering
Battelle-Columbus Laboratories
EPA, IERL, Combustion Research Branch
Ultrasysterns
EPA, Region IV, Air and Hazardous Materials Division
Tennessee Valley Authority
City of Chicago, Department of Environmental Control
Acurex, Aerotherm Division
Acurex, Aerotherm Division
Institute of Gas Technology
EPA, IERL, Combustion Research Branch
EPA, IERL, Combustion Research Branch
United Technology Research Center
KVB, Inc.
Exxon Research and Engineering
EPA, IERL, Combustion Research Branch
KVB, Inc.
Massachusetts Institute of Technology
City of Chicago, Department of Environmental Control
EPA, Energy Processes Division
Institute of Gas Technology
Massachusetts Institute of Technology
Combustion Engineering
Stanford Research Institute
Institute of Gas Technology
Ultrasystems
EPA, IERL, Combustion Research Branch
University of Arizona
A-l
-------
PARTICIPANTS LIST
(NOTE: To facilitate their identification, participants are listed alphabetically
together with the name of the organization they represent. The complete address
of each organization represented at the conference appears at the end of the list
of attendees.)
LIST OF PARTICIPANTS
Name
Alvey, Courtney D.
Anderson, Dr. Larry W.
Axtman, William H.
Bagwell, Fred A.
Baker, Burke
Ban, Stephen D.
Barrett, Richard E.
Barsin, Joseph
Bartok, William
Batra, Sushil K.
Bauman, Robert D.
Beals, Rixford A.
Beatty, James D.
Bennett, Dr. Robert
Blandford, Jr., J.B.
Blythe, R. Allen
Buechler, Lester
Bonne, Ulrich
Booth, Michael R.
Bowman, Barry R.
Bueters, K.A.
Carpenter, Ronald C.
Cernansky, Dr. Nicholas P.
Christiano, John P.
Chu, Richard R.
Clark, Norman D.
Cleverdon, R.F.
Cotton, Ernest
Creekraore, Andrew T.
Daughtridge, Jimmy T.
Degler, Gerald H.
Demetri, E.P.
DeWerth, D.W.
Dingo, T.T.
Donaldson, Thomas M.
Dowling, Daniel J.
Downey, Thomas A.
Dyer, T. Michael
Dygert, J.C.
Representing
Baltimore Gas & Electric
Acurex, Aerotherm Division
American Boiler Manufacturers Association
South California Edison
Shell Development Company
Battelle-Columbus Laboratories
Battelle-Columbus Laboratories
Babcock & Wilcox
Exxon Research & Engineering
New England Electric Systems
EPA, Office of Air Quality, Planning & Standards
NOFI
Procter & Gamble
Apollo Chemical
Englehard Industries
International Boiler Works
Systems Research Labs
Honeywell
Ontario Hydro
Lawrence Livermore Laboratories
Combustion Engineering
Armstrong Cork
Drexel University
EPA, Office of Air Quality, Planning & Standards
EBASCO
C-E Air Preheater
Chevron Research Company
American Petroleum Institute
EPA, Control Programs Development Division
Pratt & Whitney Aircraft
Systems Research Labs
Northern Research and Engineering Corporation
American Gas Association Labs
General Motors
EPA, Office of Air Quality, Planning & Standards
Union Carbide
Gamlen Chemical Company
Sandia Laboratories
Shell Development Company
A-2
-------
LIST OF PARTICIPANTS (CONT'D)
Name
Dzuna, Eugene R.
Erskine, George
Feng, C.L.
Fennelly, Paul F.
Fletcher, James
Fletcher, Roy J.
Freelain, Kenneth
Frisch, Dr. N.W.
Fuhrman, Jr., Theodore C.
Gibbs, Thomas
Goetz, Gary
Goodley, Allan R.
Graham, David J.
Greene, Jack H.
Grimshaw, Vincent C.
Grossman, Ralph
Hangebrauck, Robert P.
Heck, Ronald
Hensel, Thomas E.
Holden, Edward A.
Honea, Dr. Franklin I.
Howard, Jack B.
Hudson, Jr., James L.
Jackson, Dr. A.W.
Jepson, Dr. A.F.
Karas, Dennis T.
Kemmerer, Jeffrey
Khan, M. Ali
Khoo, Dr. S.W.
Kloecker, J.F.
Kykendal, William
Lahre, Thomas
Lange, Dr. Howard
Lavoie, Raymond C.
Lee, James E.
Lenney, Ronald J.
Levy, Arthur
Lewis, F. Michael
Lii, Dr. C.C.
Lin, Donald J.L.
Locklin, David W.
Lord, Harry C.
Loweth, Carl
Marshall, David
Marshall, John H.
Representing
Gulf Research & Development Company
Mitre Corporation
Selas Corporation of America
GCA/Technology Division
Industrial Combustion
Peabody Engineering Corporation
Federal Energy Administration
Research-Cottrell
Erie City Energy Division
EPA, Region IV
Combustion Engineering
California Air Resources Board
EPA, Office of Research and Development
EPA, Administrative Office
Process Combustion
Ralph Grossman, Ltd.
EPA, Energy Assessment and Control Division
Englehard Industries
Turbo Power and Marine Systems
General Foods Corporation
Midwest Research Institute
Massachusetts Institute of Technology
Tampa Electric Company
Ontario Hydro
Environmental Measurements, Inc.
East Chicago Air Quality Control
Fuller Company
East Chicago Air Quality Control
Canadian Gas Research Institute
Erie City Energy Division
EPA, Process Measurement Branch
EPA, Office of Air Quality, Planning & Standards
Babcock & Wilcox
Rohm & Haas Company
Facilities Engineering Command (U.S. Navy)
Ronald J. Lenney Associates
Battelle-Columbus Laboratories
Stanford Research Institute
EPA, Combustion Research Branch
Forney Engineering Company
Battelle-Columbus Laboratories
Environmental Data Corporation
The Trane Company
Babcock & Wilcox
Combustion Engineering
A-3
-------
Name
Marton, Miklos B,
Mayfield, D. Randell
Meier, John G.
Moore, Douglas S.
Moore, Edward E.
Morton, William J.
Moscowitz, Charles
Hosier, Stanley A.
Newton, Charles L.
Nurick, W.H.
Pantzer, Karl
Pershing, David W.
Pertel, Dr. Richard
Renner, Ted
Riley, Joseph
Robert, J.
Roberts, Dr. George
Robertson, J.F.
Roffe, Gerald
Rosen, Meyer
Ross, Marvin
Rulseh, Roy
Sadowski, R.S.
Samples, J.R.
Scott, Donald R.
Sheffield, E.W.
Slack, A.V.
Smith, Lowell L.
Spadaccini, L.J.
Sterman, Sam
Sullivan, Robert E.
Swearingen, W.E.
Takacs, Dr. L.
Taylor, Barry R.
Utterback, Paul M.
Van Grouw, Sam J.
Vatsky, Joel
Vershaw, Jim
Watson, Raymond A.
Webb, R.
Weiland, J.H.
Weinberger, Dr. Lawrence
White, David J.
White, James H.
White, Phil
LIST OF PARTICIPANTS (CONT'D)
Representing
IBM
EPA, Region IV
International Harvester, Solar Division
Chevron Research Company
Eclipse, Inc.
E. Keeler Company
Monsanto Research Corporation
Pratt & Whitney
Colt Industries
Rockwell International, Rocketdyne Division
Babcock & Wilcox
University of Arizona
Institute of Gas Technology
Fuel Merchants Association of New Jersey
EPA, Region IV
Canadian Department of Environment
Englehard Industries
Crystal Petroleum Company
General Applied Science Laboratories
Union Carbide
Lawrence Livermore Laboratory
Cleaver-Brooks
Riley Stoker Corporation
Union Carbide
Columbia Gas System Service Corporation
TRW
SAS Corporation
KVB, Inc.
United Technology Research Center
Union Carbide
General Motors
Koppers Company
General Motors
Massachusetts Institute of Technology
Babcock & Wilcox
KVB, Inc.
Foster-Wheeler Energy Corporation
The Trane Company
Florida Power & Light Company
The Trane Company
Texaco, Inc.
Mitre Corporation
International Harvester, Solar Division
Coen Company
Ventura Company
L
A-4
-------
Name
Wiedersum, George C.
Wilhelm, Ronald
Wilson, Jr., R.P.
Winters, Harry K.
Wittig, Dr. Sigmar L.K.
Woolfolk, Dr. Robert
Wright, Richard
Young, Dexter E.
Ziarkowski, Stanley
Zielke, Robert L.
Zirkel, Eric C.
LIST OF PARTICIPANTS (CONT'D)
Representing
Philadelphia Electric Company
Aqua-Chera, Inc.
Arthur D. Little, Inc.
Ray Burner Company
Purdue University
Stanford Research Institute
Industrial Combustion
EPA, Control Programs Development Division
Gamlen Chemical Company
Tennessee Valley Authority
Armstrong Cork
A-5
-------
LIST OF ORGANIZATIONS REPRESENTED
Kame (Represented By)
Aeurex Corporation, Aerotherm Division
(Mr. Anderson)
American Boiler Manufacturers Association
(Mr. Axtman)
American Gas Association Labs
(Mr. De Werth)
American Petroleum Institute
(Mr. Cotton)
Apollo Chemical
(Dr. Bennett)
Aqua-Chem, Inc.
(Mr. Wilhelm)
Armstrong Cork Company
(Mr. Carpenter, Mr. Zirkel)
Babcock & Wilcox
(Mr. Barsin, Mr. Lange, Mr. Moore)
(Mr. Utterback, Mr. Pantzer,
Mr. Marshall)
Baltimore Gas and Electric Company
(Mr. Alvey)
Battelle-Columbus Labs
(Mr. Ban, Mr. Barrett,
Mr. Levy, Mr. Locklin)
C-E Air Preheater
(Mr. Clark)
California Air Resources Board
(Mr. Goodley)
Canadian Department of Environment
(Mr. Robert)
Canadian Gas Research Institute
(Dr. Khoo)
Chevron Research Company
(Mr. Cleverdon, Mr. D. Moore)
Address^
485 Clyde Avenue
Mountain View, California 94042
1500 Wilson Boulevard, Suite 317
Arlington, Virginia 22209
8501 East Pleasant Valley Road
Cleveland, Ohio 44131
1801 K Street, N.W.
Washington, D.C. 20006
35 South Jefferson Road
Whippany, New Jersey 07981
P.O. Box 421
Milwaukee, Wisconsin 53201
Liberty & Charlotte Streets
Lancaster, Pennsylvania 17604
20 South Van Buren Avenue
Barberton, Ohio 44203
P.O. Box 2423
North Canton, Ohio 44720
2012 Gas and Electric Building
Baltimore, Maryland 21203
505 King Avenue
Columbus, Ohio 43201
Andover Road
Wellsville, New York 14895
1709 llth Street
Sacramento, California 95814
351 St. Joseph Boulevard
Houll, Quebec, Canada
55 Scarsdale Road, Den Mills
Ontario, M3B2R3, Canada
P.O. Box 1627
Richmond, California 94802
A-6
-------
LIST OF ORGANIZATIONS (CONT'D)
Name (Represented By)
Cleaver-Brooks
(Mr. Rulseh)
Coen Company
(Mr. J.H. White)
Colt Industries
(Mr. Newton)
Columbia Gas System Service Corporation
(Mr. Scott)
Combustion Engineering
(Mr. Bueters, Mr. Goetz, Mr. Marshall)
Crystal Petroleum Company
(Mr. Robertson)
Drexel University
(Dr. Cemansky)
EBASCO
(Mr. Chu)
East Chicago Air Quality Control
(Mr. Karas, Mr. Khan)
Eclipse, Inc.
(Mr. E. Moore)
Englehard Industries
(Mr. Blandford, Dr. Heck, Dr. Roberts)
Environmental Data Corporation
(Mr. Lord)
Environmental Measurements, Inc.
(Dr. Jepsen)
Environmental Protection Agency
EPA - Administrative Office
(Mr. Greene)
Address
3707 North Richards Street
Milwaukee, Wisconsin 53201
1510 Rollins Road
Burlingame, California 94010
701 Lawton Avenue
Beloit, Wisconsin 53511
1600 Dublin Road
Columbus, Ohio 43215
1000 Prospect Hill Road
Windsor, Connecticut 06095
P.O. Box 4180
Corpus Christi, Texas 78408
Philadelphia
Pennsylvania 19104
145 Technology Park
Norcross, Georgia 30071
900 East Chicago Avenue
East Chicago, Indiana 45312
1100 Buchanan
Rockford, Illinois 61101
Middlesex Turnpike, Wood Avenue
Edison, New Jersey 08876
608 Fig Avenue
Monrovia, California 91016
2 Lincoln Court
Annapolis, Maryland 21401
Research Triangle Park
North Carolina 27711
A-7
-------
LIST OF ORGANIZATIONS (CONT'D)
Same (Represented By)
EPA - Combustion Research Branch
(Dr. Bowen, Mr. Hall, Mr. Lachapelle,
Mr. Lanier, Dr. Lii, Mr. Martin, Mr. Wasser)
EPA - Control Programs Development Division
(Mr. Creekmore, Mr. Young)
EPA - Energy Assessment & Control Division
(Mr. Hangebrauck)
EPA - Office of Air Quality, Planning,
and Standards
(Mr. Bauman, Mr. Christiano,
Mr. Donaldson, Mr. Lahre)
EPA - Office of Research and Development
(Mr. Graham)
EPA - Process Measurement Branch
(Mr. Kuykendal)
EPA - Region IV
(Mr. Biggs, Mr. Mayfield, Mr. Riley)
Erie City Energy Division
(Mr. Fuhrman, Mr. Kloecher)
Exxon Research & Engineering Company
(Mr. Bartok)
Federal Energy Administration
(Mr. Freelain)
Florida Power & Light Company
(Mr. Watson)
Forney Engineering Company
(Mr. Lin)
Foster Wheeler Energy Corporation
(Mr. Vatsky)
Fuel Merchants Association of New Jersey
(Mr. Renner)
Address
Research Triangle Park
North Carolina 27711
Research Triangle Park
North Carolina 27711
Research Triangle Park
North Carolina 27711
Research Triangle Park
North Carolina 27711
Washington, D.C. 20460
Research Triangle Park
North Carolina 27711
1421 Peachtree Street, N.E.
Atlanta, Georgia 30309
1422 East Avenue
Erie, Pennsylvania 16502
P.O. Box 8
Linden, New Jersey 07036
1200 Pennsylvania Avenue, N.W.
Washington, D.C. 20461
P.O. Box 013100
Miami, Florida 33101
P.O. Box 189
Addison, Texas 75001
10 South Orange Avenue
Livingston, New Jersey 07039
66 Morris Avenue
Springfield, New Jersey 07081
A-8
-------
LIST OF ORGANIZATIONS (CONT'P)
Name (Represented By)
Fuller Company
(Mr. Kemmerer)
GCA/Technology Division
(Dr. Fennelly)
Gamlen Chemical Company
(Mr. Downey, Mr. Ziarkowski)
General Applied Science Laboratories
(Mr. Roffe)
General Foods - Technical Center
(Mr. Holden)
General Motors Corporation
(Mr. Sullivan)
(Mr. Dingo, Mr. Takacs)
Gulf Research and Development Company
(Mr. Dzuna)
Honeywell, Inc.
(Mr. Bonne)
IBM
(Mr. Marton)
Industrial Combustion
(Mr. Wright, Mr. Fletcher)
Institute of Gas Technology
(Dr. Pertel)
International Boiler Works
(Mr. Blythe)
International Harvester, Solar Division
(Mr. Meier, Mr. D.G. White)
Address
124 Bridge Street
Catasauqua, Pennsylvania 18032
Burlington Road
Bedford, Massachusetts 01730
299 Market Street
Saddle Brook, New Jersey 07662
Merrick & Stewart Avenues
Westbury, New York 11790
250 North Street
White Plains, New York 10625
5735 West 25th Street
Indianapolis, Indiana 46224
Technical Center
Warren, Michigan 48090
P.O. Box 2038
Pittsburgh, Pennsylvania 15230
Bloomington
Minnesota 55420
1000 Westchester Avenue
White Plains, New York 10604
4465 North Oakland
Milwaukee, Wisconsin 53211
3424 South State Street
Chicago, Illinois 60616
P.O. Box 498
East Stroudsburg, Pennsylvania 18301
2200 Pacific Highway
San Diego, California 92119
A-9
-------
LIST OF ORGANIZATIONS (CONT'D)
Name (Represented By)
KVB Engineering, Inc.
(Mr. L. Smith, Mr. Van Grouw)
E. Keeler Company
(Mr. Morton)
Koppers Company, Inc.
(Mr. Swearingen)
Ronald J. Lenney Associates
(Mr. Lenney)
Arthur D. Little, Inc.
(Mr. R.D. Wilson)
Lawrence Livermore Laboratories
(Mr. Bowman, Mr. Ross)
Massachusetts Institute of Technology
(Mr. Howard, Mr. Taylor)
Midwest Research Institute
(Dr. Honea)
Mitre Corporation
(Dr. Weinberger, Mr. Erskine)
Monsanto Research Corporation
(Mr. Moscowitz)
NOFI
(Mr. Seals)
Naval Facilities Engineering Command,
Southern Division
(Mr. J. Lee)
New England Electric Systems
(Mr. Batra)
Northern Research & Engineering Corporation
(Mr. Demetri)
Address
6624 Hornwood Drive
Houston, Texas 77036
Williamsport
Pennsylvania 17701
Koppers Building
Pittsburgh, Pennsylvania 15219
2001 Palmer Avenue
Larchmont, New York 10538
Acorn Park
Cambridge, Massachusetts 02140
P.O. Box 808
Livermore, California 94550
Massachusetts Avenue
Cambridge, Massachusetts 02139
425 Volker Boulevard
Kansas City, Missouri 64110
Westgate Research Park
1820 Dolly Madison Boulevard
McLean, Virginia 22101
Station B. Box 8
Dayton, Ohio 45407
New York, New York
2144 Melbourne Street
P.O. Box 10068
Charleston, South Carolina 29411
20 Turnpike Road
Weston, Massachusetts 01581
219 Vassar Street
Cambridge, Massachusetts 02139
A-10
-------
LIST OF ORGANIZATIONS (COKT'D)
Naroe (Represented By)
Ontario Hydro Corporation
(Mr. Booth, Dr. Jackson)
Peabody Engineering Corporation
(Mr. R. Fletcher)
Philadelphia Electric Company
(Mr. Wiedersom)
Pratt & Whitney Aircraft
(Mr. Daughtridge, Mr. Mosier)
Process Combustion Corporation
(Mr. Grimshaw)
Procter & Gamble Company
(Mr. Beatty)
Purdue University
(Mr. Wittig)
Ralph Grossman, Ltd.
(Mr. Grossman)
Ray Burner Company
(Mr. Winters)
Research-Cottrell, Inc.
(Dr. Frisch)
Riley Stoker Corporation
(Mr. Sadowski)
Rockwell International Corporation,
Rocketdyne Division
(Mr. Nurick)
Rohm & Haas Company
(Mr. Lavoie)
SAS Corporation
(Mr. Slack)
Address
620 Union Avenue
Toronto, Ontario, Canada M561X6
835 Hope Street
Stamford, Connecticut 06907
2301 Market Street, S10-1
Philadelphia, Pennsylvania 19101
P.O. Box 2691
West Palm Beach, Florida 33402
1675 Washington Road
Pittsburgh, Pennsylvania 15228
610 South Center Hill Road
Cincinnati, Ohio 45224
West Lafayette
Indiana 47907
P.O. Box 70, Town of Mt. Royal
Montreal, Canada H3P 3B8
1301 San Jose Avenue
San Francisco, California 94112
P.O. Box 750
Boundbrook, New Jersey 08805
9 Neponset Street
Worcester, Massachusetts 01613
6633 Canoga Avenue
Canoga Park, California 91304
P.O. Box 584
Bristol, Pennsylvania 19007
RFD #1
Sheffield, Alabama 35660
A-ll
-------
LIST OF ORGANIZATIONS (CONT'D)
Name (Represented By)
Sandia Laboratories
(Mr. Dyer)
Selas Corporation of America
(Mr. Feng)
Shell Development Company
(Mr. Dygert, Mr. Baker)
South California Edison
(Mr. Bagwell) '
Stanford Research Institute
(Dr. Woolfolk, Mr. Lewis)
Systems Research Labs
(Mr. Buechler, Mr. Degler)
TRW, Inc.
(Mr. Sheffield)
Tampa Electric Company
(Mr. Hudson)
Tennessee Valley Authority
(Mr. Zielke)
Texaco, Inc.
(Mr. Weiland)
The Trane Company
(Mr. Loweth, Mr. Vershaw, Mr. Webb)
Turbo Power and Marine Systems
(Mr. Hensel)
Union Carbide Corporation
(Mr. Rosen, Mr. Sterman)
(Mr. Dowling)
Address
Livermore
California 94550
Dresher
Pennsylvania 19025
P.O. Box 481
Houston, Texas 77001
P.O. Box 800
Rosemead, California 91770
1611 North Kent Street
Arlington, Virginia 22209
2800 Indian Ripple Road
Dayton, Ohio 45440
1 Space Park - R4/2020
Redondo Beach, California 90278
P.O. Box 111
Tampa, Florida 33601
524 Power Building
Chattanooga, Tennessee 37401
P.O. Box 509
Beacon, New York 12508
3600 Pammel Creek Road
La Crosse, Wisconsin 54601
1690 New Britain Avenue
Farmington, Connecticut 06032
Tarrytown Technical Center
Tarrytown, New York 10591
Box 180
Sistersville, West Virginia 26175
A-12
-------
LIST OF ORGANIZATIONS (CONT'D)
Name (Represented By)
Union Carbide Corporation
(Mr. Samples)
United Technologies Research Center
(Mr. Spadaccini)
University of Arizona
(Mr. Pershing)
Ventura County A.P.C.D.
(Mr. P. White)
Address
Box 4361
South Charleston, West Virginia 25353
400 Main Street
East Hartford, Connecticut 06040
Tucson
Arizona 85721
740 East Main Street
Ventura, California 93001
*u.-i. coi'rn,NNE?fr HUNTING OFFICE:
Region ».
A-13
-------
TECHNICAL REPORT DATA
(flcast trad liiitructiuns on the rcicrse before completing)
1. REPORT NO.
EPA-600/2-76-152b
2.
I. RECIPIENT'S ACCESSIOr+NO.
Combustion Symposium; Volume n--Fuels and Process
Research and Development
5. REPORT DATE
June 1976
6. PERFORMING ORGANtZATtON CODE
7. AUTHOR(S)
Miscellaneous
8. PERFORMING ORGANIZATION REPORT NO
9. PERFORMING OR9ANIZATION NAME AND ADDRESS
NA
10. PROGRAM ELEMENT NO.
1AB014; ROAP 21BCC
11. CONTRACT/GRANT NO.
NA (In-house)
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
Proceedings; 9/24-26/75
14. SPONSORING AGENCY CODE
EPA-ORD
5 SUPPLEMENTARY
Mail Drop 65, Ext. 2470/2477.
Chairman J.S. Bowen, Vice-Chairman R.E. Hall,
6. ABSTRACT
The proceedings document the 37 presentations made during the Stationary
Source Combustion Symposium held in Atlanta, Ga. , September 24-26, 1975. Spon-
sored by the Combustion Research Branch of EPA's Industrial Environmental Resea-
rch Labor atory--RTP, the symposium dealt with subjects related both to developing
improved combustion technology for the reduction of air pollutant emissions from
stationary sources, and to improving equipment efficiency. The symposium was
divided into four parts and the proceedings were issued in three volumes: Volume I--
Fundamental Research, Volume n — Fuels and Process Research and Development,
and Volume IE—Field Testing and Surveys. The symposium was intended to provide
ontractor, industrial, and Government representatives with the latest information
on EPA in-house and contract combustion research projects related to pollution
control, with emphasis on reducing nitrogen oxides.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COSATi Field/Group
Air Pollution, Combustion, Field Tests
Combustion Control, Coal, Oils
Natural Gas, Nitrogen Oxides, Carbon
arbon Monoxide, Hydrocarbons, Boilers
Pulverized Fuels, Fossil Fuels, Utilities
Gas Turbines, Efficiency
Air Pollution Control
Stationary Sources
Combustion Modification
Unburned Hydrocarbons
Fundamental Research
Fuel Nitrogen
Burner Tests
13B
21B 14B
2 ID 11H
07B
07C 13A
13G 14A
s, DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
430
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Fotm 2220-1 (9-73)
A-14
-------
2P 600/2 ^?A
5-152b In^ust rial
'nv .
S Lnb.
AUTHOR
Proceeding^ oF the stationary
TlTLEsource combustion pvmnosium.
V.2: Fuel"5 :\ uroct-ss research
DATE
OAYLOUD 41
-------
-------
DATE DUE
i
-------
-------
|