EPA-650/2-75-012
JANUARY 1975 Environmental Protection Technology Series
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environ-
mental Protection Agency, have been grouped into series. These broad
categories were established to facilitate further development and applica-
tion of environmental technology. Elimination of traditional grouping was
consciously planned to foster technology transfer and maximum interface
in related fields. These series are:
1. ENVIRONMENTAL HEALTH EFFECTS RESEARCH
2. ENVIRONMENTAL PROTECTION TECHNOLOGY
3. ECOLOGICAL RESEARCH
4. ENVIRONMENTAL MONITORING
5. SOCIOECONOMIC ENVIRONMENTAL STUDIES
6. SCIENTIFIC AND TECHNICAL ASSESSMENT REPORTS
9. MISCELLANEOUS
This report has been as&iyned to the ENVIRONMENTAL PROTECTION
TECHNOLOGY series. This series describes research pi-rlormed to
develop and demonstrate instrumentation, equipment and methodology
to repair or prevent environmental degradation Irom point and non-
point sources of pollution. This work provides the new or improved
technology required for the control and treatment of pollution sources
to meet environmental quality standards.
-------
EPA-650/2-75-012
ANALYSIS OF TEST DATA
FOR NOx CONTROL
IN GAS- AND OIL-FIRED
UTILITY BOILERS
by
Owen W. Dykcma
The Aerospace Corporation
Urban Programs Division
El Segundo, Cahiornia 90245
Grant No. R-802366
ROAP No. 21ADG-089
Program Element No. 1ABQ14
EPA Project Officer: Robert E. Hall
Control Systems Laboratory
National Environmental Research.Center
Research Triangle Park. North Carolina 27711
Prepared for
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
January 1975
-------
EPA REVIEW NOTICE
This report has been reviewed by the National Environmental Research
Center - Research Triangle Park, Office of Research and Development,
EPA, and approved for publication. Approval does not signify that the
contents necessarily reflect the views and policies of the Environmental
Protection Agency, nor does mention of trade names or commercial
products constitute endorsement or recommendation for use.
This document it available to the public for sale through the National
Technical Information Service, Springfield, Virginia 22161.
-------
ACKNOWLEDGMENTS
Sincere appreciation is acknowledged for the guidance
and assistance provided by Mr. Robert E. Hall of the Control Systems
Laboratory, who was the EPA Project Officer during the conduct of this
study.
A special acknowledgment is due the Los Angeles Depart-
ment of Water and Power (LADWP) for its cooperation in making avail-
able the data upon which this study is based and to LADWP personnel
Messrs. Hans Sonderling, Roy Toda, Wesley Pepper, and Robert
Centner for their extensive assistance in data acquisition and
interpretation.
Acknowledgment is also due The Aerospace Corporation
personnel Mrs. Sandra Barnes and Mr. Guy Kuncir for their assistance
in computer programming and operation and in statistical interpretation
of the data.
Approved by:
Owen W. Dykem%£/Manager
Combustion Effects
Office of Pollution and Monitoring
Director
Office of Pollution and Monitoring
Toru lura, Associate Group
Director, Environmental
Programs Group Directorate
e itze r ,
fvironmental Pro
ire ctorate
111
-------
FOREWORD
This study was initially conceived in coordination with
the Control Systems Laboratory of the Environmental Protection
Agency (EPA), The Aerospace Corporation, and the Los Angeles
Department of Water and Power. A great deal of testing on full-scale
multiple-burner utility boilers has been accomplished in recent years
by electric utility companies in the Los Angeles area, which has
resulted in significant reductions in nitrogen oxide (NO ) emissions.
JC
Data from this testing could be very valuable to utilities in other areas
of the country where NO emission regulations are only beginning to
require the large NO emission reductions already accomplished under
the more stringent regulations in the Los Angeles area. An in-depth
analysis of the data could also be valuable to supply feedback to
further research in NO reduction techniques as well as to assist
Jt
directly in pointing out combusion modification techniques likely to
yield minimum NO emissions. It appeared that the increasing inci-
dence of combustion and flame stability problems under operating
conditions modified for the purpose of NO reduction represented a
jf.
real and perhaps severe limitation on NO reduction by combustion
JL
modification methods. The Aerospace Corporation's long history of
dealing with fundamental combustion processes and combustion stability
problems in rocket engines was considered to be particularly applicable
in the areas of concern.
-------
The study reported herein was conducted for the EPA,
Control Systems Laboratory, Clean Fuels and Energy Branch,
Combustion Research Section, Research Triangle Park, North Carolina,
by The Aerospace Corporation. The study mainly concerns the analysis
of data on combustion modifications for NO reduction with natural gas
and low-sulfur oil fuels. The analysis of combustion and flame insta-
bility data is limited in this initial study to an effort to identify general
mechanisms and to verify these with the available data. A followon
study has already been initiated to conduct the same type of NO emis-
X
sion reduction study with coal fuels and with other furnace designs.
Although the general mechanisms of the observed combustion and flame
instability problems are believed identified and adequately verified, the
more extensive stability analyses required to derive acceptable solutions
have been deferred to a later study.
A brief introduction and a summary of the results of
this study are contained in Section 1. Section 2 describes the analysis
of NO reduction techniques, and Section 3 discusses the combustion
J\.
and flame stability analyses. The appendixes include related details
of these studies as well as a listing of the detailed test data used in the
study.
VI
-------
ABSTRACT
This report describes The Aerospace Corporation
analyses of a large quantity of emissions, operating conditions, and
boiler configuration data from the full-scale multiple-burner electric
utility boilers of the Los Angeles Department of Water and Power,
using natural gas and low-sulfur oil fuels. Objectives of the study
include (a) evaluation of the effects of combustion modifications on NO
H
emissions in fundamental combustion terms, (b) evaluation of techniques
for further reductions in NO emissions, and (c) determination and
.X
substantiation of general mechanisms for observed combustion and
flame stability problems. The report includes the following results:
(a) discussion of the major combustion process modifications resulting
in NO emission reductions due to two-stage combustion, burners-out-
Ji.
of-service, combustion air temperature reduction, load reduction, and
excess air variations; (b) estimates of NO minima achievable in the
X.
boilers studied with current hardware; (c) estimates of most probable
longer-term hardware and operating condition modifications likely to
yield ultimate NO reductions with these fuels; (4) identification and
verification of general mechanisms for the combustion and flame
instabilities observed; and (e) a listing of all of the hardware configura-
tion, operating conditions, and NO , CO, O , and CO emissions data
"X. £* £
for 428 tests in 8 full-scale, multiple-burner, face-fired electric utility
boilers using natural gas and low-sulfur oil fuels.
Vll
-------
CONTENTS
ACKNOWLEDGMENTS ........................... iii
FOREWORD .................................. v
ABSTRACT ................................... vii
1. INTRODUCTION AND SUMMARY ................ 1-1
1. 1 Introduction .......................... 1-1
1.2 Summary
1. 2. 1 Effects of Combustion Modifications
on NO Emissions ................. 1-7
x
1.2.2 Combustion and Flame Stability ........ 1-23
2. EFFECTS OF COMBUSTION MODIFICATIONS ON
NO EMISSIONS ............................ 2- *
x
2. 1 Data Analysis Approach .................. 2-1
Z.I.I The NO Model ................... 2-2
2. 1. 2 The General Correlation Equation ....... 2-7
2. 1. 3 Application of the Correlation
Equation ................ . ....... 2-11
2. 2 Results ............................. 2-22
2. 2. 1 Observations Directly from the
Correlations ..................... 2-25
2.2.2 Parametric Studies of the Correlations. . . . 2-37
2. 2. 3 Effects of Some Combustion
Modifications .................... 2-46
2.2.4 Burners -Out -Of -Service (BOOS) ........ 2-60
2. 2. 5 Other Boilers .................... 2-74
2.3 Further Extrapolations and Conclusions ........ 2-84
2.3.1 Further NO Reduction Techniques ...... 2-87
2.3.2 NO Minima with Existing Hardware ..... 2-90
References ............................... 2-95
-------
CONTENTS (Continued)
3. COMBUSTION AND FLAME INSTABILITIES 3-1
3. 1 Feed System Coupled Combustion Instability 3-2
3.1.1 Gain Stabilization 3-3
3.1.2 Gain Stabilization Applied to Utility
Boilers 3-6
3. 2 Correlation of Observed Instabilities with
Air Feed System Dynamic Response 3-9
3. 2. 1 Air Flow Through Individual Burners 3-11
3.2.2 Available Combustion Vibration Data 3-24
3.3 Combustion Stability Conclusions 3-31
Nomenclature 3-34
References 3-36
APPENDIXES
A. DEVELOPMENT OF THE CORRELATING
EQUATION A-l
B. FUEL ANALYSES AND COMBUSTION
CALCULATION RESULTS B-l
C. INTERMEDIATE DATA CONVERSION AND
CORRELATIONS C-l
D. INPUT DATA FOR CORRELATIONS D-l
E. EFFECT OF PARTIAL REACTION WITHIN A
BURNER ON THE RESISTANCE TO AIR FLOW
THROUGH THE BURNER E-l
x
-------
FIGURES
1-1. Frequency Distribution of Measured NO Data 1-3
JC
2-1. Schematic of Mixing Zone Model--Horizontal
Section at Lowest Burner Level 2-3
2-2. Schematic of Mixing Zone Model--
Vertical Section 2-4
2-3. Mixing Zone Model--Definition of Burner
Configurations 2-6
2-4. Schematic of Possible Variations of the Average
Molecular Scale A/F Ratio of the Combustion
Air as Gas or Oil-Vapor Mixes With It 2-24
2-5. Effects of a Single (Air) Burner on the Bulk
Gas A/F Ratio Through the H5 Boiler 2-38
2-6. Effects of NO Port Admittance 2-42
x
2-7. Effects of Increased Cooling Rate 2-43
2-8. Effects of Reduced Combustion Air
Temperatures 2-44
2-9. Effects of Load Variations: (a) With Natural
Gas Fuel; (b) With Low-Sulfur Oil Fuel 2-48
2-10. Separate Effects of Total Flow and Combustion
Air Temperature Variations with Load: (a) With
Natural Gas Fuel; (b) With Low-Sulfur Oil Fuel 2-51
2-11. Effects of Combustion Air Diverted Through NOX
Ports (Two-Stage Combustion): (a) With Natural
Gas Fuel; (b) With Low-Sulfur Oil Fuel 2-54
2-12. Effects of Excess Air: (a) With Natural Gas Fuel;
(b) With Low-Sulfur Oil Fuel 2-58
2-13. Effects of the Vertical Location of Four (Air)
Burners (Calculation Only) 2-63
XI
-------
FIGURES (Continued)
2-14. Effects of the Vertical Location of Eight (Air)
Burners on the NOX Represented in the Bulk Gas
Term in the Correlation Equations . 2-65
2-15. Effects of the Vertical Location of Eight (Air)
Burners: (a) With Natural Gas Fuel; (b) With Low-
Sulfur Oil Fuel 2-66
2-16. Effects of the Number of (Air) Burners
Located at Midlevels with Natural Gas Fuel 2-71
2-17. Effects of the Vertical Level of (Air) Burners
in the Smaller Boilers: (a) With Natural Gas Fuel;
{b) With Low Sulfur Oil Fuel 2-78
2-18. Maximum Estimate of NOX from Fuel Nitrogen
Versus Burner Equivalence Ratio 2-85
3-1. Block Diagram of a Feed System Coupled Mode
of Instability in a Rocket with Gaseous Reactants .... 3-5
3-2. Effect of Air Velocity in an (F+A) Burner on the
Admittance to Air Flow Through the Burner 3-17
3-3. Effect of Air Velocity on the Fraction of
Combustion Completed Within an (F+A) Burner 3-20
3-4. Effect of Partial Combustion in a Burner on the
A/F Ratio in the Burner with Gas Fuel 3-23
3-5. Response of Air Flow Through an (F+A) Burner to
Furnace Pressure Perturbations with Gas Fuel 3-25
3-6. SI Boiler Vibration Test Data with Gas Fuel 3-27
3-7. H5 Boiler Vibration Data with Gas Fuel 3-29
XII
-------
TABLES
2-1. Description of Boiler-Specific Input Variables 2-13
2-2. Values of Boiler-Specific Input Variables 2-15
2-3. Summary of the Total Data Sample 2-17
2-4. Description of Terms in the Correlating Equation .... 2-21
2-5. Coefficients of the Terms in the Correlation
Equations for Various Data Samples 2-26
2-6. Average Input Values for the Parameters in Each
Variable Term in the Correlation Equations for
the Various Data Samples 2-30
2-7. Average Values of the Terms in the Correlation
Equations for the Various Data Samples 2-32
B-l. Fuel Analysis B-2
B-2. Equilibrium Combustion Inputs for the
NO Generation Equation B-4
C-l. Correlations of Natural Gas Flow with Load C-3
C-2. Correlations of Low Sulfur Oil Flow with Load C-4
C-3. Correlations of Combustion Air Flow with Air-Foil
Percent Flow Indicator C-7
C-4. Correlations of Combusion Air Flow with Forced
Draft Fan Current C-8
C-5. Correlations of Combustion Air Temperature
with Load C-9
Xlll
-------
-------
SECTION 1
INTRODUCTION AND SUMMARY
1. 1 INTRODUCTION
Requirements for the reduction of nitrogen oxide (NO )
emissions from large utility boilers were established at a time when
only general guidelines concerning the desired combustion conditions
for minimum NO generation were available from laboratory research.
Methods of operating a boiler or necessary hardware modifications to
provide those combustion conditions in a full-scale multiple-burner
boiler were not clearly established. Analytical and experimental re-
search in this area continue today. As is often the case in rapid tech-
nology development, the hardware and operating side of the industry
was required to achieve certain goals using limited guidelines supple-
mented by the powerful method of "cut-and-try." A vital part of the
iterative research and development process is the feedback to research
of the results of this full-scale testing, both to provide evaluation of
the initial guidelines developed in research as well as to provide a new
source of information to guide further research.
The problem of describing and controlling the combined
aerodynamics, reaction, and heat transfer within the reaction section
of a full-scale combustor is highly complex and involves a large number
of independent variables. A reasonable analysis of full-scale test
results, then, requires a fairly large number of tests in which all of
the significant variables are varied, even though the significance of the
1-1
-------
variables cannot be easily assessed before the results are analyzed.
The stringent NO regulations imposed on electric utilities in the Los
Angeles area have caused these utilities to try more NO reduction
A
methods and to generate more data than elsewhere in the country.
Certain common techniques have been established which significantly
and reliably reduce NO emissions to 30 to 50 percent of the un-
Jk
controlled levels. In many areas of the country, these reductions are
sufficient to meet any near-term regulations. Los Angeles area
regulations, however, are sufficiently stringent that further reductions
appear necessary. These further reductions require deeper under-
standing of control techniques.
Among the leaders in the analytical-empirical study of
combustion modifications for NO reduction in utility boilers is the
Los Angeles Department of Water and Power (LADWP), a municipally
owned electric utility. Over the period 1969 to 1973, the LADWP con-
ducted more than 500 full-scale multiple-burner boiler tests for the
express purpose of developing optimum hardware configurations and
operating conditions for minimum NO emissions compatible with low
emissions of hydrocarbons (HC), carbon monoxide (CO), and smoke
and with high plant efficiency and safe, stable boiler operation.
Initial NO reductions to 30 to 50 percent of uncontrolled
levels appear to have been easily accomplished, with further reduc-
tions becoming progressively more difficult. This is indicated in
Figure (1-1), which is a frequency plot of all of the LADWP NO data
.X
accumulated in this study. The figure shows that comparatively little
testing was required to achieve a reduction in NO to 50 percent of
initial levels, but that required to achieve another 50 percent reduction
increased enormously. In fact, the increase in testing required for
decreasing NO emissions appears to be exponential. This character-
istic could result from (a) the exponential character of the NO genera-
tion rate with temperature, (b) the exponential decrease in understanding
1-2
-------
LU
«: <
s>
O O
>-
O
UJ
o
50
40
30
10
ALL BOILERS, ALL OPERATING CONDITIONS
O NATURAL GAS
LOW-SULFUR OIL
NOTE: THE FREQUENCY OF OCCURRENCE IS
A MEASURE OF THE NUMBER OF
CONFIGURATIONS TESTED WHICH
FAILED TO FURTHER REDUCE THE
NO EMISSIONS
r\ X
200
400
600
800
1000
MEASURED VALUES OF NO ppm
A
Figure 1-1. Frequency distribution of measured NO data
1-3
-------
of the remaining, minor sources of NO production at low NO emission
Ji A
levels, or (c) an asymptotic approach to real limits in minimum NO
emissions with existing hardware and techniques.
Among the possible real limits to NO reduction appear-
X
ing in this data are (a) excessive HC, smoke, or CO production and (b)
combustion stability and flame stability problems, either of liftoff or
of flames deep back in the burner, causing overheating. In the cor-
relations of the effects of combustion modifications on NO emissions
x
discussed in this report, some note is made of conditions which would
appear to yield lower NO emissions but are prohibited from doing so
because of high HC, smoke, or CO emissions. Approximately 20 per-
cent of the effort in this study was devoted to the delineation and verifi-
cation of the mechanism of observed combustion and flame instabilities
(Section 3). Detailed stability analysis and solutions to these problems
were deferred to a later study.
Complete data on hardware and operating conditions and
the resulting measured emissions of NO , CO, oxygen (O2), and carbon
dioxide (CO_) were obtained in nine face-fired boilers using natural gas
and low-sulfur oil fuels. Test conditions yielding high emissions of HC
or smoke were generally avoided; hence, all of the 485 tests analyzed
in this study can be considered to be free of problems with these two
emissions. The data resulting from these tests were sufficiently com-
plete to be used in the analyses and are included in this report.
1.2 SUMMARY
The large number of independent variables which could
have significant effects on NO emissions dictate, at least for an initial
A
study such as this, a data analysis technique which can identify not only
all of these potentially significant variables but, roughly, the proper
form of these variables in relation to NO emissions. The analysis
A\
technique should also be capable of correlating large numbers of test
1-4
-------
conditions in which each of these variables are significantly varied.
The study approach was to establish, first, a rough model of the pro-
cesses of aerodynamic mixing, reaction, and cooling of the gases in
the boiler and from that to develop a single equation which, if all of the
input information were exact, would be capable of predicting the NO
j£:
emissions. Since the information required to develop such a prediction
equation is not well known, the equation was cast in the form of a series
of linear terms, and the method of linear regression analysis was used
to correlate the necessarily large samples of data from full-scale
multiple-burner boilers. A total of 428 test conditions were used in
these correlations, each involving more than 40 independent input
variables specifying the hardware configurations and operating condi-
tions. Obviously, a computer was necessarily employed to handle the
input data, make necessary data conversions, calculate the values of
the terms in the correlating equation, effect the linear regression
analysis, and statistically evaluate the result.
The useful output of this study is not the resulting cor-
relations themselves. Unlike many linear regression analyses, the
correlation equation does not consist of a series of simple powers of
the input variables. The calculation of the value of the parameter in
each term of the correlation equation is much too complex to describe
as a simple function of the independent variables, even including some
tabular data read into the program. Use of the correlating equations
to predict NO emissions requires the exact calculation procedure used
Ji
in the correlations of this study.
The more useful output of this study lies in the insight
gained concerning the effects of certain input variables on NO emis-
Ji
sions, by study of the resulting correlations themselves and by the
parametric exercise of these equations. Such insight can be generalized
to indicate hardware and operating conditions for minimum NO emis-
sions in many boilers and, in some cases, the levels of NO emissions
1-5
-------
under these minimum conditions. Such generalizations represent the
important conclusions reached from detailed analyses of the data from
428 test conditions on 4 face-fired boiler types (8 boilers) with natural
gas and low-sulfur oil fuels.
In general, all of the effects of combustion modifications
on NO emissions appear roughly explainable by the equivalent air-fuel
J\:
(A/F) ratio, temperature, and time history of the reactants as they flow
through the aerodynamic mixing and reaction zones in the boiler and
become subject to cooling by the boiler water walls. Further insight is
gained from the effects of this history on a simple Zeldovich mechanism
for NO generation. Since combustion reactions occur at the molecular
level, the appropriate A/F ratio is the average molecular-scale
mixed air-gaseous fuel (or vapor fuel in the case of liquid fuels) ratio.
The effects of combustion modifications on NO emissions are discussed
JC
with respect to that history.
Similarly, the observed combusion instabilities appear
to result from coupling of the air feed system with combustion in the
boiler. One of the phenomena controlling this feed system coupled
instability is related to the effect of partial combustion within the
burners on the admittance (inverse of resistance) to air flow through
these burners. This effect also strongly controls the distribution of
combustion air between burners flowing both fuel and air (F+A) and
those burners flowing air-only (Air), and any NO ports. NO ports
Ji Jt
are separate ports to introduce combustion air well downstream of
(above) the burner flows. This combustion air maldistribution contri-
butes to some of the observed flame instabilities. More specific con-
clusions drawn from this study form the bases for the following
discussion.
1-6
-------
1.2.1 Effects of Combustion Modifications on
NO Emissions
In an uncontrolled boiler, the A/F ratios and
temperatures in all parts of the boiler, for significant lengths of time,
are at values which represent rates of NO generation close to the
maximum possible for that boiler. Any combination of modifications
which result in significant (a) changes in A/F ratio (either increase
or decrease), (b) reduction in reaction product temperatures, or (c)
reduction in the time spent by the reaction products under high NO
Jt
generation conditions will almost inevitably result in a significant
reduction in NO emissions. As a result, reductions to 30 to 50 per-
cent of the uncontrolled NO emissions are relatively easy to accom-
plish. Reductions below those levels, however, require much more
detailed understanding and control.
Combustion modification techniques for NO reduction
X
which could be studied with the available data sample included (a)
reduction in combustion air temperatures, (b) increased boiler cooling
rates, (c) two-stage combustion using NO ports, and (d) burners-out-
of-service (BOOS) (shutting off the fuel to some of the burners). Flue
gas recirculation (FGR) into the primary flame zone could not be
studied because FGR in the boilers in the data sample was introduced
only from the bottom of the boiler. No significant effect of FGR intro-
duced in this manner could be observed.
Reduction in the combustion air temperature reduces
NO emissions but becomes decreasingly effective at the lower tern-
J\.
peratures. Even a reduction to ambient temperatures represents a
relatively small reduction in the peak reaction temperatures (about
10 percent). Depending on how such a reduction in combustion air
temperature was accomplished, significant plant efficiency losses
could result. In any case, NO emissions with oil fuels could not be
1-7
-------
reduced below that resulting from conversion of fuel bound nitrogen by
this technique alone. In conjunction with .two-stage combustion or BOOS
techniques to reduce the conversion of bound nitrogen, combustion air
temperature reduction might be effective in reaching lower than current
NO levels, but plant efficiency would be decreased.
Increasing the cooling rate in the radiant section of the
boiler has essentially the same effects of combustion air temperature
reduction, except that the heat is always rejected to the working medium
(water-walls), and the plant efficiency should not be compromised.
Boiler modifications to increase this cooling rate, however, are exten-
sive, and care would have to be taken to maintain sufficient bulk gas
temperatures to the superheater tubes. In many practical cases, it
appears that when the cooling rate (per unit length of the radiant section)
becomes too great, FGR is introduced from the base of the boiler by
the operator to counteract this excessive cooling and drive the hot gases
more quickly up to the superheater tubes.
Two -stage combustion simply subtracts some of the
combustion air from the burners, leaving the majority of the boiler
combustion volume fuel-rich. The remaining air is added downstream
after maximum cooling has occurred. The NO reductions by this
Jv
method alone are generally limited (to levels which are still significant)
by the NO generated when the remaining excess air is added.
The major effect of BOOS is the same as that of staged
combustion. This technique is also limited in NO reduction to signi-
ficant levels, unless great care is taken as to where and how much
excess air is added to the reacting gases. Both the staged-combustion
and BOOS techniques, however, can significantly reduce the conversion
of bound nitrogen in oil fuels, leaving only the thermal NO as the
X
problem. Neither the staged combustion nor the BOOS techniques
necessarily result in significant plant efficiency losses, nor do they
require major hardware modifications to implement.
1-8
-------
Further insight into the effects of the above combustion
modification techniques on NO emissions in full-scale multiple-burner
boilers was obtained by direct study of the correlation equations and by
parametric studies using these equations, with confirmation from data
where available.
1.2.1.1 Correlation Equations
Insight into the effects of combinations of the above
combustion modifications on NO emissions was obtained, first, by
^x
direct study of the average empirical values of the terms in the corre-
lations and, secondly, by parametric exercise of these equations. The
correlations contain both positive and negative terms (Section 2,
Table 2-7). Therefore, insight into the potential for minimum NO
production can be obtained by considering those hardware and operating
conditions (Section 2, Table 2-4) which would maximize the dominant
negative terms and minimize the dominant positive terms in the cor-
relations. Two major observations from direct study of the correlations
involved differences between natural gas and oil fuels and between large
and small boilers. These differences were pursued further in the
parametic studies.
1.2.1.1.1 Differences Between Gas and Oil Fuels
In the large boiler type of this study (H5/6) (Section 2,
Table 2-1), which incorporates a large vertical array of burners (six
rows high), the dominant terms in the correlations were those related
to the configurations of (Air) and (F+A) burners in the burner array.
Those related to bulk gas and NO port mixing zones were of lesser or
indeterminant significance. In this large boiler type, the signs of the
dominant burner configuration terms were exactly opposite for natural
gas and oil fuels. From consideration of the signs of these dominant
terms, minimum NO emissions with natural gas fuels in this large
1-9
-------
boiler type appeared to result when the total number of (Air) burners
was small and the burners were located near the middle rows of the
burner array. These same considerations indicated that with oil fuels
in this large boiler type, minimum NO emissions appeared to result
JK
when the total number of (Air) burners was large and the burners were
located in or near the bottom or top rows of the burner array. This
somewhat surprising observation could not be generalized from direct
study of the correlations alone. Neither could it be ascertained, at
that point in the study, whether the conditions indicated for mini-
mum NO emissions had actually been tested or whether those conditions
x '
represented possible new burner configurations capable of further
significant NO reduction. Extensive parametric studies in conjunction
X *
with and guided by the above observations were necessary to achieve
the final observations.
I.Z.I.1.2 Differences Between Large and Small Boilers
In the study of the large boiler type incorporating a
large vertical array of burners, the dominant terms in the correlating
equations were those related to the configuration of (Air) and (F+A)
burners in.the burner array. In the smaller boilers incorporating a
more limited vertical array of burners, the bulk gas and NO port
terms were of approximately equal importance to the burner configura-
tion terms. In all sizes of boilers, the signs of the burner configura-
tion terms for gas fuels remained the same. With oil fuels, the signs
of these terms tended to reverse as the size of the boiler decreased
and to become similar to those for the gas fuels. The somewhat obvious
generalization from this result is that the arrangement of (Air) and
(F+A) burners in the array becomes less important to NO emissions
. t x
as the size of the array approaches one (single burner). This result
was significant in the progress of this study.
1-10
-------
Laboratory testing of boilers involving the mixing,
reaction, and heat transfer phenomena occurring in full-scale boilers
with large numbers of burners would be prohibitively expensive. The
most significant new insight possible from the study of data from full-
scale boilers, then, concerns the effects on NO emissions of the inter-
Jt
actions between burners in large burner arrays. Since the correlations
of data from the small boilers in this data sample indicated a markedly
reduced influence of burner configuration over that in the larger boiler
type, the parametric studies of burner configuration effects were
limited to the large boiler type (H5/6) alone. Conclusions reached
from study of these effects in the large boiler type were evaluated for
application to small boilers, using applicable small boiler data.
1.2. 1.2 Parametric Studies
For reasons discussed in 1. 2. 1. 1. 2, the parametric
studies were conducted using the correlation equations for natural gas
and oil fuels in the large boiler type (H5/6). The effects of load
(megawatt plant electrical output) variation on NO emissions are well
ji
known. Load variation, however, is not considered a desirable NO
Ji
control technique not only because it limits the available plant capacity
but it usually results in a significant loss in plant efficiency. Neverthe-
less, parameteric runs were made with load as the variable: (a) to
check the accuracy of the multiple-variable correlations when used to
predict the effects of single variables and (b) to evaluate the relative
contributions to NO emissions of the combustion air temperature and
j£
total throughput flow variations which result from the load variations
and more directly control NO emissions. Similarly, although com-
2C
bustion air temperature and boiler cooling rate control are not particu-
larly desirable NO emission control techniques because of potentially
2£
significant plant efficiency losses and/or the major hardware modifica-
tions required, their effects on NO emissions were also evaluated in
1-11
-------
single variable parametric runs. No further effort was made in this
study to evaluate the effects of these variables. Efforts were, instead,
concentrated in the potentially more productive areas concerning the
effects of excess air, NO ports, and burner configurations.
3C
1.2.1.2.1 Load, Combustion, Air Temperature, and
Boiler Cooling Rate Effects
Reduction in plant electrical load is accomplished by
reducing the total fuel and air flow through the boiler, essentially at a
constant A/F ratio. This reduction results in greater total cooling
of the combustion gases and, through the air preheater, a reduction in
combustion air temperature. Empirical correlations of both throughput
flow and combustion air temperature with the overall plant load were
derived from the data. Parameteric variation of load, then, resulted
in the simultaneous parametric variation of both of these variables.
Comparison of the calculated and measured NO emissions with load
Ji
variation showed reasonable agreement with natural gas fuel and ex-
cellent agreement with oil fuels. In both cases, reducing the load
reduced NO emissions. Calculated NO emissions at half-load were
x x
about 73 and 70 percent of those at full load with gas and oil fuels,
respectively. Separate parametric runs to evaluate the relative con-
tributions of throughput flow and combustion air temperature over the
ranges dictated by load variations, as single variables, indicated some
surprisingly large differences with gas fuels and relatively small dif-
ferences with oil fuels. Since no data were available to evaluate these
separate calculated effects, these differences were not pursued further.
Combustion air temperatures were further varied over
a wide range, involving extrapolations by more than a factor of six
beyond the temperature data in the correlations (to ambient tempera-
tures), with full-load throughput flow. Although calculations involving
such large extrapolations may or may not be valid, results indicate that
1-12
-------
NO emissions with combustion air temperatures reduced to ambient
are about 36 and 53 percent with gas and oil fuels, respectively, of
those at the full-load temperature. With the same reservations re-
sulting from large extrapolations, it would appear that the NO emis-
sions at ambient temperatures with gas fuels are still decreasing with
decreasing temperature, while those with oil fuels appear to be
asymptotically approaching a value corresponding to the conversion
of the fuel-bound nitrogen (no NO emissions from thermal fixation
Jt
mechanisms).
Extrapolations by a factor of about 2. 6 in the boiler
cooling rate {reduction in combustion product temperatures with time)
showed that the NO emissions with gas fuels rapidly approached zero,
while those with oil fuels again approached a value resulting from con-
version of fuel-bound nitrogen alone. Both of these results and those
for combustion air temperature variations are in agreement with
theoretical thermal and fuel-bound nitrogen conversion mechanisms
for the formation of NO emissions. Extrapolations much beyond 2. 6
times the data range in the boiler cooling rates resulted in unrealistic
NO emission variations.
1.2.1.2.2 Excess Air, NO Port, and Burner
Configuration Effects
Boiler operation with reduced excess air is often recom-
mended as a significant NO reduction technique, particularly since it
Ji.
appears to increase plant efficiency. There was no evidence of signifi-
cant effects of excess air variation on NO emissions in the data of this
x
study. Parametric variations of excess air, in fact, showed that NO
emissions with gas fuels increased slightly with decreased excess air.
Those with oil fuels did decrease with decreasing excess air, but
again only slightly. Although the available applicable data were widely
scattered, it appeared to confirm these trends. The magnitudes of the
1-13
-------
calculated and measured NO values were in good agreement in both
A
cases. These results are interpreted to indicate that, at normal values
of excess air, thermally generated NO increases as excess air is
.X
reduced toward stoichiometric but conversion of fuel-bound nitrogen
always decreases with decreasing excess air. With natural gas fuels,
there is no fuel-bound nitrogen; therefore, NO emissions are thermally
generated and will increase with decreasing excess air as long as that
excess air is above zero. With the oil fuels of this data sample, con-
taining very little bound nitrogen, the tradeoff between increasing
thermally generated NO and decreasing conversion of fuel-bound
JC
nitrogen with decreasing excess air results in a slight net decrease in
NO emissions. Fuels containing larger concentrations of bound
nitrogen (such as coal) should always exhibit the expected strong trend
toward decreasing NO emissions with decreasing excess air. Because
X
the variation of NO emissions with excess air with the fuels of this
x
study is so small (±7 to 10 percent over the practical range of excess
air), no further effort was made to study this variable.
Parametric variations of the fraction of combustion air
flowing through NO ports were conducted with extrapolations to about
Ji:
2.4 times the data range. Again, confirmation of the calculated values
with appropriate data was reasonable with gas fuels and excellent with
oil fuels. With both fuels, the NO emission levels asymptotically
JC
approached a limiting value as the fraction of combustion air through
the NO ports increased. With all burners operating on (F+A) and 3
percent O_, these limiting values were 32 and 52 percent, for gas and
oil, respectively, of the values with NO ports closed. Since the air-
fuel ratio in the burner region of the boiler would be uniformly very
low, the NO emissions with large fractions of combustion air flowing
through the NO ports are interpreted to be generated in, and down-
J*.
stream of, the NO port mixing zones. Thus, although the effect of
NO ports in reducing overall NO emissions with all burners operating
1-14
-------
(F+A) is dramatic, this technique alone is limited to NO levels which
X
are still quite significant. Further, when used in conjunction with BOOS
(a practical substitute for NO ports), further combustion air diverted
ji
through NO ports may have very little effect on further reduction in
JC
NO emissions and, in some cases, could even increase these emissions.
JC
The bulk of the parametric studies of NO emissions in
the large boiler type were concentrated on the effects of various numbers
and configurations of (Air) and (F+A) burners in the burner array.
Various vertical and horizontal combinations of four and eight (Air)
burners were evaluated, generally, with other input variables held
constant. In some cases, fixed configurations were evaluated with NO
X.
port air flow and combustion air temperatures as the variables. Not all
combinations with up to 8 of the 24 (Air) burners were evaluated since
more than a million combinations are possible.
Although a number of effects were observed, the two
most significant variations involved the number of (Air) burners and
their vertical location (as a group) in the burner array. In general, the
parametric calculations showed that configurations involving only one
to four (Air) burners yielded minimum (in fact, negative) values of NO
.X
emissions with gas fuels, if these (Air) burners were located in the
third to fifth row from the bottom of this six-row array. Unfortunately,
little or no test data involving one through seven (Air) burners were in
the data sample, so this possible minimum could not be confirmed.
Configurations involving no (Air) burners or eight (Air) burners, no
matter how they were arranged, yielded larger values of NO emis-
sions with gas fuels. The minimum NO emission configuration indi-
X.
cated by the available gas-fired test data with eight (Air) burners was
with the eight (Air) burners located as high as possible in the burner
array with NO ports open.
As indicated by direct observation of the dominant terms
in the correlation equations, the results of parametric studies of burner
1-15
-------
configurations with oil fuels in the large boiler type indicated almost
directly opposite configurations for minimum NO emissions from
those with gas fuels. Instead of a small number of (Air) burners
located in the middle rows of the burner array as indicated with gas
fuels, the parametric studies with the oil fuel correlation indicated
that minimum NO emissions should be achieved with a larger number
X.
of (Air) burners (eight in this study) located at either the top or bottom
of the burner array, with the bottom location yielding the lower NO
emissions. Attempted confirmation with applicable data indicated that
NO emissions monotonically decreased with (Air) burners located
lower in the burner array. Configurations with (Air) burners located
in the bottom row were not in the data sample. Minimum NO emis-
sions obtainable with existing hardware with these oil fuels, however,
appear to be most closely approached in the data sample with four of
the eight (Air) burners located in the second row from the bottom of the
array.
Evaluation of the results of the parametric studies and
data confirmation as they might apply to the smaller boilers indicated
the same general locations for (Air) burners to achieve minimum NO
emissions with both fuels (i. e., at the top for gas fuels and at the
bottom for oil fuels). The effects of (Air) burner locations, however,
were much less significant than in the large boiler type.
1.2.1.3 General Considerations for Minimizing
NO Emissions
The results discussed in Sections 1. 2. 1. 1 and 1.2.1.2
were interpreted in the light of the general mechanisms for the forma-
tion of NO from thermal fixation and fuel-bound nitrogen conversion
x
and the probable A/F ratio temperature and time history of the reacting
gases. In general, thermally generated NOx can be minimized by
maintaining the molecular-scale, mixed A/F ratio either fuel-rich
1-16
-------
or fuel-lean. Conversion of fuel-bound nitrogen can be minimized only
by maintaining the A/F ratio in the burner area of the boiler fuel-
rich. In each case, the amount of time spent by the reacting gas at A/F
ratios near stoichiometric must be minimized to minimize thermally-
generated NO . The major mixing of the reacting gases from any given
X.
burner with those from the burners below occurs when the burner gases
encounter the bulk gases. The molecular-scale, mixed A/F (gas or
vapor) ratio in the burner gases and the bulk gases as this mixing begins
determines whether the A/F ratio of the burner gases must pass
through stoichiometric during the subsequent mixing process and, there-
fore, what the relative A/F ratios of the burner and bulk gases
should be at the vertical level of that burner to minimize net NO
ji
emissions.
With the current hardware configurations of the boilers
of this study, certain combinations of operating conditions or combus-
tion modifications yield the minimum NO emissions observed in the
Ji
data. In one case, the general considerations derived from this study
indicate that further reduction may be possible with a combination of
operating conditions which were not in the data sample. These same
general considerations, however, indicate that over the longer range
view, when new power plants are built or when significant hardware
modifications might be implemented on existing plants, totally different
combinations of operating conditions would yield minimum NO emis-
j£
sions. These levels, with the new or modified hardware, are expected
to be significantly lower than those obtainable with current hardware.
Conclusions applicable to current hardware are discussed in 1.2. 1.4
and those applicable to the longer range in 1. 2. 1. 5.
1.2.1.4 Minimum NO Emission in Current Hardware
X:
Two major differences in the boilers of this study signi-
ficantly affect the operating conditions for minimum NO emissions.
X,
1-17
-------
These are the so-called large boilers, in which a large number of
burners (24) are distributed vertically in six rows, and the small
boilers, in which the number of rows of burners is less than five.
Minimum NO emissions in the hardware of this study are discussed
x '
separately with respect to the large and small boilers, with both gas
and oil fuels in each.
1.2.1.4.1 Large Boilers
Final results indicate that combinations of combustion
modifications which yielded the minimum NO emissions observed in
X.
the gas-fired test data of this study are predominantly determined by
the fact that air-gas fuel mixing is relatively rapid. Average molecular-
scale, mixed A/F ratios equal to the burner input A/F ratio are
reached before the burner flows begin to mix with the bulk gas flow.
This indicates that the bulk gas A/F ratio should be near that of
the burner.
Such configurations result in current hardware when
(Air) burners are located high in the burner array or when NO ports
are open. Average A/F ratios must eventually cross stoichiometric
to reach the overall boiler excess air condition, when the excess air is
finally added to the burner flows high in the boiler, but this is accom-
plished only after maximum cooling of the gases has occurred. All of
the reacting gases (burner and bulk) in the burner region of the boiler
are at A/F ratios well below stoichiometric.
The above is supported by the observation that, with
eight (Air) burners in the burner array, minimum NO emissions of
197 ppm were achieved (in the data sample) in the large boiler type,
fired at full load with gas guel and with the (Air) burners located as
high as possible in the boiler and NO ports open.
The large boiler parametric studies also indicated that
a possible NO minimum even lower than this might be achieved by
1-18
-------
locating just a few (Air) burners in the fourth or fifth rows from the
bottom of this six-row burner array. This observation would also be
compatible with the above observations if it were assumed that the air/
gas-fuel mixing to molecular-scale, although very fast, required a
finite time and distance into the boiler to occur. As the bulk gases
grow in volume and spread across the boiler cross section, as they flow
up the radiant section, at some level the burner gases would not have
time to complete the necessary internal mixing before they encounter
forced mixing with the spreading bulk gases. This would be the appro-
priate level, then, to introduce some (Air) burners to convert the fuel-
rich bulk gases from below to air-rich gases. Such a configuration
was not represented in the data.
Combinations of combustion modifications which yield
the minimum NO emission observed in the data of this study when
x
firing low-sulfur oil fuel, with 0. 24 percent by weight bound nitrogen,
are predominatly determined by the fact that the rate of mixing of the
air-vapor fuel is very slow. This mixing rate is controlled by the
vaporization rate of the liquid fuel rather than by the air turbulence.
As a result, the average molecular-scale mixed air-vapor fuel ratio
cannot reach that of the burner overall air-liquid fuel ratio before
significant mixing with the bulk gases begins. This indicates that the
bulk gases should be maintained air-rich to keep this average A/F
ratio from ever crossing stoichiometric and to minimize thermally-
generated NO . Unfortunately, with oil fuels, this type of A/F
3t
ratio tailoring results in maximum conversion of the fuel bound nitrogen.
Thus, with oil fuels, two possible minimum NO conditions result: one
3C
with minimum thermal NO but maximum fuel NO (the latter case) and
x x
one with the reverse resulting from fuel-rich burners and fuel-rich
bulk gases. In boilers burning oil fuels with very low bound-nitrogen
content and high thermal NO formation rates, the lower minimum may
be the air-rich bulk gases case, simply accepting the maximum
1-19
-------
conversion of the bound nitrogen for the boiler excess air level. This
is the case with all of the boilers in the sample of this study using the
low-sulfur fuel containing 0. 24 percent by weight bound nitrogen. This
case is implemented in the existing hardware by locating a few (Air)
burners low or at the bottom of the burner array (and no NO ports).
JL
The NO minima achieved in the data are approximately 220 ppm, at 3
percent O-, in good agreement with conversion efficiencies of bound
nitrogen from the literature. No significant problems of excessive
CO, HC, or smoke, combustion or flame stability, or losses in plant
efficiency were observed under conditions sufficient to reach this mini-
mum. In other boilers, with lower combustion air temperatures or
higher cooling rates, burning oil fuels containing more nitrogen, the
lower of the two NO minima could be the opposite of that observed in
this study.
1.2.1.4.2 Small Boilers
Final results indicate that the effects of (Air) burner
locations in the burner array are much less significant in small boilers
involving limited numbers of burners and limited vertical distribution
of the burner array. In a boiler with 12 burners arranged in a config-
uration of 4 burner columns wide and 3 burner rows high, filling 1 row
with (Air) burners leaves only 2 rows of (F+A) burners. Thus, there is
little room for definition of top, middle, or bottom rows. The whole
boiler begins to take on the characteristic of a. single-burner boiler,
and the burner gases begin to become indistinguishable from the bulk
gases. Under these conditions, it might be expected that only the
general case of two-stage combustion would have any significant effect
on NO emissions. Two-stage combustion could be accomplished either
by locating (Air) burners in the top row of the burner array and/or
with open NO ports) (fuel-rich to fuel-lean staging) or by locating the
1-20
-------
(Air) burners in the bottom row with NO ports closed (fuel-lean to
fuel-rich staging).
These conclusions are supported by the observations
that minimum NO emissions in small boilers, when fired with natural
gas fuel, were achieved with two-stage combustion by the use of NO
ports or by locating a few (Air) burners in the top row of the burner
array. NO emission levels, with 3 percent O (equivalent excess air)
X £
of 110 to 140 ppm were achieved in the small boilers in the sample
studied with no undesirable emissions of CO, HC, or smoke, with no
combustion or flame instabilities, and with no significant losses in
plant efficiency. The small boiler data with oil fuels also supports
these observations. Minimum NO emissions were achieved when the
x
(Air) burners were located lower in the array, as in the large boiler
type. Minimum NO emissions observed in this data were 178 to 208
7r x
ppm, again only slightly lower than the value calculated for full con-
version of the fuel-bound nitrogen at the boiler excess-air condition.
No significant problems of excessive CO, HC, or smoke, combustion
or flame stability, or significant losses in plant efficiency were
observed under conditions sufficient to reach this minimum.
1.2.1.5 Longer Range NO Emission Reduction
Since natural gas fuels contain no fuel-bound nitrogen,
minimum NO emissions can be attained by maintaining the A/F
ratio in the radiant section of the boiler either fuel-rich or fuel-lean
to minimize thermally generated NO . The overall boiler, however,
X.
is always maintained slightly fuel-lean to minimize CO, HC, and smoke
emissions. Thus, if the burner section is maintained fuel-rich, the
average molecular-scale mixed A/F ratio must cross stoichiometric
at lease twice in the reacting section of the boiler.
It appears that low NO emissions can be achieved in
gas-fired boilers by a combustion modification technique opposite to
1-21
-------
that which results in the minimum with existing hardware. This
technique would involve operating the burners and the resulting bulk
gases slightly more lean (air-rich) than the overall boiler A/F
ratio by reducing the fuel flow to the burners and adding the remaining
fuel well downstream of the burners in fuel NO ports. This scheme
would avoid the necessity for the average molecular-scale mixed
A/F ratio to ever cross stoichiometric. This combustion modification
can yield very low NO emissions because of the absence of fuel-bound
nitrogen in the natural gas fuel. No inherent problems of excessive
CO, HC, or smoke emissions, of combustion or flame instability, or
significant losses in plant efficiency are apparent in this configuration.
Further significant reduction of NO emissions with
Jf.
bound nitrogen-containing oil fuels requires that the conversion of the
bound nitrogen to NO be minimized. At the moment, the only way to
accomplish this is by initially reacting the fuel in a fuel-rich air-vapor
fuel environment. This is generally impossible, and thermally gener-
ated NO can be large, if the fuel is slowly bled into the surrounding air
Ji
stream as it vaporizes. This suggests that very fine atomization or
prevaporization of the oil is a necessary prerequisite. A general
approach to the ultimate minimum NO emissions with oil fuels, then,
X
appears to involve very fine oil atomization, strong fine-scale turbu-
lence, and use of NO ports or (Air) burners located high in the burner
array. No inherent problems of excessive CO, HC, or smoke emis-
sions or losses of plant efficiency are apparent in this configuration.
Since the finely atomized, rapidly mixed oil fuel combustion would be-
have somewhat like natural gas combustion, care would have to be
taken to avoid combustion or flame stability problems.
1-22
-------
1.2.2 Combustion and Flame Stability
Application of a general stability criterion for feed
system coupled modes derived in the liquid rocket industry indicates
that this type of combustion instability is probable in boilers coupled
with the air feed system. The resistance to air flow through the burners
from the windbox to the furnace is relatively small. As a result, the
air flow rates through the burners are only weakly controlled. Small
perturbations in the local pressure in the furnace can easily represent
large perturbations in the pressure drop across the burners and, there-
fore, in the air flow through the burners. This is the definition of a
high dynamic response feed system.
The small pressure drop across the burners also allows
any partial reaction in the diffusion flame within the burner to dominate
and greatly increase the resistance to air flow through the burner.
When other nonburning flow paths are available for air flow, such as
NO ports (in two-stage combustion) or (Air) burners (in the BOOS
X
technique), the increased resistance to air flow in the (F+A) burners
causes an unexpectedly large diversion of air flow to these nonburning
paths. This effect is nonlinear and is compounded by this diversion,
causing even further reduction in the (F+A) burners. The resulting
large imbalance in air flow distribution between the burning and non-
burning paths has strong effects on both steady-state NO emissions
J\.
as well as on the dynamic response of the air feed system to furnace
pressure perturbations and air feed system coupled combustion
stability.
The available data indicate that when about 75 percent
of the burners in the burner array or equivalent NO port flow area are
j£
configured for air flow only, the A/F ratio in the remaining (F+A)
gas-fired burners drops below or near the fuel-rich premixed flammable
limit for air-natural gas flames. Since the flames in the (F+A) burners
1-23
-------
are not well-mixed (a diffusion flame), the flame does not extinguish
in the pilot zone in the burner but the reaction may not go to completion
as moliecular-scale mixing is approached. Under these same condi-
tions, the nonlinear effect of the flame in the burner begins to strongly
amplify the air feed system response, leading to air feed system
coupled combustion instability.
Records of visual observations of gas-fired burner
flames under these operating conditions indicate all of the poorly
anchored, ragged characteristics typical of operation near a premixed
flammable limit. The available data on observed gas-fired combustion
instabilities also correlate well with this operating condition. Flame
stability problems of the type where the flame retreats deep into the
burner, causing burner and register overheating, are a further result
of these operating conditions.
Flame liftoff appears to be simply the opposite case. If
a burner gas spud is designed to limit mixing within the burner to mini-
mize the flame deep in the burner, then, under conditions of higher air
velocity through the burner, the flame can rather easily be blown off
the pilot zone and lift off of the burner.
All of the combustion and flame stability problems dis-
cussed occur in gas firing. Since oil fuels do not need to be anchored
by a pilot zone in the burner and since they have not begun significant
vaporization within the burner, the problems associated with the effects
of partial reaction within the burner are usually not significant. The
approach to ultimate reduction of NO emissions with oil fuels, i.e. ,
the finely atomized rapidly mixed oil combustion discussed in Section
1.2. 1, suggests that these oil flames could provide all of the conditions
discussed for gas flames which yield combustion and flame instabilities.
Solutions to such problems are possible, however, and involve proper
design of the oil gun spray with respect to the air flow profile through
1-24
-------
the burners. More detailed steady-state and dynamic analysis than that
reported here is necessary to assure trouble-free operation. Similar
problems could also occur if and when significant quantities of low-Btu
gaseous fuels are fired in -utility boilers.
1-25
-------
-------
SECTION 2
EFFECTS OF COMBUSTION MODIFICATIONS ON NO EMISSIONS
x
The purpose of conducting this study on large quantities
of NO emissions data from operational multiple-burner utility boilers
A
was to gain some insight into the fundamental combustion processes
occurring in boilers as they affect NO emissions. Although a multiple
regression analysis technique was used in the analysis, there was no
intention to develop a universal correlation equation which could be used
to predict NO emissions for any boiler simply by inserting the hard-
j£
ware and operating conditions of that boiler into these equations.
A large amount of raw data was accumulated during the
course of this study from the records of the LADWP on NO emissions
X.
and related hardware and operating conditions. Nearly all of the fun-
damental combustion processes pertaining to NO emission reduction in
X.
full-scale multiple-burner utility boilers are represented in that data.
This section discusses the approach taken in the analysis
of the LADWP data, observations from the correlations, parametric
studies of the correlations, applications of combustion modifications
for NO control, and estimations of NO minima for existing and future
X 5£
hardware.
2. 1 DATA ANALYSIS APPROACH
The approach taken in analyzing the data was to (a) as-
semble a rough model of NO generation in a large multiple-burner
2-1
-------
utility boiler, using generally accepted principles; (b) use that model
to generate a single equation which, if all of our input knowledge were
correct and accurate, would directly predict NO emissions; (c) use
Ji
that equation to correlate the data and correct for inadequacies in the
input assumptions; and finally (d) analyze the resulting correlation
equations to determine what these correlations indicate about the areas
of poor understanding. This procedure is little different from that
generally followed by engineers to obtain an understanding of the vari-
ations in some parameter of interest. The principal difference here is
that a fairly large number of independent variables are known to affect
NO emissions. This dictates that a fairly complicated model be estab-
Ji
lished in order that each of these variables be included in their approx-
imately correct form and with their approximately correct relationship
to each other and to the NO emissions. Each of the steps involved in
Jv
this procedure are discussed.
2.1.1 The NOX Model
Several important variables relating to NO emissions
Ji
from a large multiple-burner boiler describe the air flow through NO
A.
ports and take into account the geometric relations between burners
flowing (F+A) and those flowing (Air). Experimental data involving
these variables can come only from the large multiple-burner boilers
themselves. Therefore, it was considered of primary importance to
establish a model which included the effects of these variables as im-
proved understanding would most likely come from this area.
The model, shown schematically in Figures 2-1 and 2-2,
was established for the largest and most complex of the boilers studied.
Figure 2-1 shows a series of fixed mixing zones in the horizontal plane
at the lowest furnace level. The effects of horizontally adjacent and
opposite burner flows on each other are accounted for by mixing in the
adjacent and opposite mixing zones, and the effects of all flows in a
2-2
-------
FRONT
WALL
BURNER
RECIRCULATION /'
^\ /
( 1 /SECONDARY
/ '
£_ /
n /
PRIMARY 1 /
_J /
\ /
Xv/ ADJACENT
A MIXING
/ \
1 \
' \
_I V
\ \
f ^
o\
\
-n
z.
>
t
^
X
I
^^
CD
oa
__
\
\o
\ '
\ s_
\ i
\ i
V !. . . -.1. - -
N /
\ /
\ /
/ r
/ L 7
x /o
REAR
WALL
OPPOSITE MIXING
Figure 2-1. Schematic of mixing zone model--horizontal
section at lowest burner level
2-3
-------
PORTS
BURNERS
^»-
^
^
-^
_
FINAL-FINAL
N0x PORT MIXING
A
\ FINAL No. 6
\
\
1 \ FINAL No. 5
__ ^_ _J \_^ __ _ __ _
\FINALNo. 4
\FINAL No. 3
\
III \
j \FINAL
^FINAL
_ _ _ ^_ ^ i^
ZONE i
!
1 ^^
I
m
M^^m «
/ L
7 C
v c:
NO. 2 / p
_ _ _ _/ L_ .
^/ c:
I7
FGR
Figure 2-2. Schematic of mixing zone model-
vertical section
2-4
-------
given furnace level are accounted for by the bulk gas mixing zones.
Figure 2-2 shows the assumed spreading of the bulk gases as more
burner flows are fed in and the vertical level increases. Two effects
of the orientation of burners in the vertical level are accounted for
(a) by the mixing of burner flows at a given level in the bulk gases with
all of the flows from the burners below that level and (b) by the preven-
tion or reduction of adjacent, opposite, and secondary mixing at a given
level by the spreading bulk gases. Figures 2-1 and 2-3 show the as-
sumed recirculation of part of the primary flow from one burner back
toward the furnace firing face and upward lo join the primary flow from
the next higher level burner. This recirculation flow accounts for the
effect of the flow from one burner on the burner flow immediately above
it at all levels. Figure 2-3 shows the definition of burners adjacent,
opposite, and below a given burner.
Figures 2-1 and 2-2 also show the approximate geometries
of these various mixing zones. These zones are assumed fixed in space
in a given furnace. The individual flows from each burner are assumed
to flow through a larger or smaller number of these zones, in series,
on the way to the NO port mixing zone. Through the NO port zone
X X
and beyond, all flows are assumed to be in the furnace, and the total
flow and the A/F ratio are that of the overall furnace. As the flows
pass from one zone to the other, the average flow velocity, A/F ratio,
and equilibrium reaction products of the zone are assumed to be in-
stantly established. The resulting mixture then flows uniformly through
the zone.
Such a picture of the flow and mixing from multiple burn-
ers in a large boiler is not unreasonable. In any case, it does provide
a rough framework on which a rough NO prediction equation can be
A.
established. This equation will include, among others, variables that
account for the orientation of (F+A) and (Air) burners, horizontally
adjacent and opposite and vertically adjacent, and for the vertical level
2-5
-------
RECIRCULATION
FLOWS
FRONT*
i
:>
C3
o
00
TOP
OPPOSITE
0
©
"Viewed from the outside
REAR*
ADJACENT
BELOW
O 0
© 0
Figure 2-3. Mixing zone model--definition of burner
configurations
-------
of the individual burners. Greater sophistication in modeling than this
is not considered necessary for the purposes of this study. Further
details on the geometry of these zones are described in Appendix A.
2.1.Z The General Correlation Equation
The correlation equation used in this study was devel-
oped, essentially, by tracking the flows from the individual burners
through the various mixing zones to the top of the furnace (the end of
the radiant section). In each mixing zone, the increment of NO formed
in that zone was calculated, and the total NO emissions from the boiler
was the sum of those increments. Fundamental to this approach is the
assumption that NO levels are always well below equilibrium, and
3£
therefore the NO destruction rate is negligible.
jf.
The NO formation rate in each zone was calculated
x
from
d[NO]/dt = (2.4 x 1012)[N] [0]1/2exP(--) (2-1)
where T is equal to degrees Kelvin. Equation (2-1) was taken from
Ref. 2-1 and was modified by assuming a negligible rate of NO decom-
position and by conversion of the concentration terms [NO], [No]' and
[O,l to mole fractions. Equation (2-1) is based on a Zeldovich mecha-
£
nism of NO formation and assumes that oxygen atoms are in thermal
equilibrium with oxygen molecules. This assumption is truly valid
only for fuel-lean mixtures. Oxygen atom concentration is appar-
ently somewhat higher than equilibrium under fuel-rich conditions.
Equation (2-1) describes the formation only of thermal nitrogen and
contributes nothing to the understanding of the conversion of fuel-
bound nitrogen to NO. In view of the current ill-defined knowledge
of the kinetics of fuel-rich HC combustion processes and of
2-7
-------
bound-nitrogen conversion, Eq. (2-1) was accepted as a first-order
approximation of the formation rate of thermal nitrogen over the
full range of A/F ratios. A separate term was included in the
correlation equation to account for bound-nitrogen conversion,
consisting only of a calculation of that NO which would be formed from
the fuel-bound nitrogen if there were 100 percent conversion. It was
left to the correlation to determine a coefficient for this term which
might, at least, represent an average bound-nitrogen conversion ef-
ficiency for the data sample.
Two major problems in any attempt to model a complex
aerodynamic mixing, reaction, and heat transfer process such as oc-
curs in the radiant section of a large multiple-burner boiler are the
determination of the chemical species and the temperature at every
point in the volume. These inputs are important for each mixing zone
in the furnace and establish the inputs to Eq. (2-1). These, however,
are exactly the unknown inputs which might be at least partially ex-
plained by correlation of a large amount of data such as in this study.
Major assumptions made in this area in developing the
correlation equation were:
a. Air and fuel, either gas or oil fuels, leave the burners
in a completely mixed gaseous state.
b. Equilibrium combustion products and temperatures are
maintained as subsequent mixing occurs in downstream
mixing zones (shifting equilibrium).
c. The cooling rate of the gases is a constant function of
time only.
The first assumption, of course, is not true. Even in
the case of gas fuels, the flame issuing from the burners is a diffusion
flame, with mixing to finer scale levels and further reaction continuing
well into the primary mixing zone and beyond. Detailed analysis of the
completeness of the gas flame reaction within the burner, discussed at
length in Section 3 indicates that 20 to 30 percent of the gas flame
2-8
-------
combustion has occurred before the flame leaves the burner. This
assumption is clearly not true for oil flames. The majority of the oil
fuel leaving the burner is still in the liquid state, and the A/F vapor
ratio is very large. The oil fuel must first vaporize, and subsequent
reaction depends on the local A/F ratio surrounding the liquid droplets
and the rate of further mixing with more distant air. Thus, the actual
average A/F ratio existing in any mixing zone, on a molecular scale,
depends not only on the gross mixing, which the rough model attempts
to describe, but also on the breakdown of the gross-scale mixing into
finer scale mixing, eventually approaching molecular scale. The NO
reaction rate of Eq. (2-1) is based on the reaction product specie
concentrations and temperatures on a molecular scale. An attempt to
include all of these turbulent mixing phenomena in the rough model is
clearly beyond the scope of this study.
Instead, assumption (a) was made, with the clear recog-
nition of its inadequacy. Figures 2-1 and 2-2, however, show that
certain of the mixing zones, which are represented in the correlation
equation by individual terms, are close to the burner exit (primary
zones) and some, such as the bulk gas zones, are farther away in both
space and time. The data correlation can shed some light on both the
error in assumption (a) and on the mixing processes actually occurring,
by defining those zones which are and are not significant to the overall
NO formation. If, for example, the data correlation shows that the
n
primary and recirculation zones dominate the variations in NO for-
mation, then it can be assumed that the assumption is approximately
correct and that the bulk of the gases in those zones, on a molecular
scale, are at or are passing through A/F ratios near stoichiometric.
Critical to the arrangement of (F+A) burners and (Air) burners in the
array is the relationship betwe.cn gross-scale and molecular-scale
mixing in the full-size boiler.
2-9
-------
Assumption (b) is a reasonable one and will be discussed
no further here.
Assumption (c), on the cooling rate of the gases in the
various zones, is not necessarily a good one. It seems reasonable that
the cooling rate in zones well removed from the water walls, where
heat loss is by radiation to the walls, might asymptotically approach a
constant. The cooling rate in the recirculation zones, however, where
heat loss could be dominated by convective heat transfer, may be con-
siderably higher than in the other zones. Several attempts were made
to establish a spatial distribution of cooling rates throughout the fur-
nace, using water wall heat flux profiles supplied by the utility. Indi-
cations were, at least in that section of the furnace up through the NO
H
port mixing zone, that a constant cooling rate with time alone was a
reasonable approximation. The data were too limited, however, to
confirm this indication. Because of the limited scope of this study, a
constant rate was incorporated. The danger in this simple assumption
is that the temperature enters into Eq. (2-1) in the argument of the
exponential term. A linear regression analysis might not be able to
compensate for a large error in this nonlinear term. Assumption (a)
also strongly affects the calculated temperature in a given zone, at
least in the early mixing zones where it is known to be invalid. Both
of these potential errors had to be accepted, however. It was expected
that, at worst, significant errors due to these two assumptions would
be reflected in an inability to obtain good correlation of the data.
Further details on the development of the correlation
equation are described in Appendix A; Appendix B discusses the fuel
analyses and the equilibrium combustion calculations made with these
fuels.
2-10
-------
2.1.3 Application of the Correlation Equation
The approach to analysis of the data, involving the
development of a rough NO prediction equation from a rough model of
the combustion processes occurring throughout the furnace, was in-
tended to maximize the probability that all of the significant independent
variables affecting NO emissions were included in the correlation
JC
equation, in the approximately correct form and with the approximately
correct relationship to the NO emissions. For example, the rough
JC
analysis indicates (a) that the initial combustion air and fuel temper-
atures, the equilibrium temperature rise due to reaction at the A/F
ratio of a given zone, and the temperature decrease due to cooling en-
route to the zone are all significant independent variables, (b) that they
should be related in the form of a sum to establish an estimate of the
average temperature in the zone, and (c) that they should be related to
NO emissions in the exponential form indicated in Eq. (2-1). This is
considered a significantly different and more realistic approach to the
analysis of data involving a large number of independent variables. The
usual approach involves lumping all of the anticipated independent vari-
ables into an extended polynomial correlation equation involving each of
the independent variables directly. The approach in this study, how-
ever, does introduce the problem that the terms in the resulting correla-
tion equation are not entirely independent of each other. For example,
the combustion air temperature enters into all but the bound-nitrogen
term in the correlation equation. This problem must be kept firmly in
mind when interpreting the results of the correlations. The objective
of this study, as stated in Section 1, is to learn as much as possible
from the data about the fundamental combustion processes which in-
fluence NO emissions and not necessarily to establish influence coef-
ficients for each of the independent variables.
2-11
-------
2.1.3.1 Data Samples
Table 2-1 provides a description of the furnace geometry
input variables and a number of furnace-related derived variables used
in the correlation equation. The furnace designation (PQ) is not an in-
put variable. The flow area of the burner (AB) is calculated from the
burner diameter (HWB) and, therefore, is not an independent variable.
The mixing zone dimensions are all derived from the furnace geometry.
The cooling factor essentially accounts for variations in furnace cooling
surface-to-volume and flow rate ratios. It is derived from furnace
geometry and operating variables. The FOR flow rate constants are
empirically derived as a means of converting measured flue gas fan
amperage to flue gas flow rates. As such, they represent one indepen-
dent input variable concerning operating conditions. (These derivations
are discussed in Appendix A.) The air flow admittances are derived
empirically from air flow rate and windbox-furnace pressure drop data
(discussed extensively in Section 3). These three admittances represent
three additional independent input variables which might be considered
related to operating conditions. Thus, Table 2-1 includes nine indepen-
dent input variables related to furnace geometry and four independent
variables related to operating conditions. Values of these variables
are listed in Table 2-2.
In addition, each of the burners in the total burner array
can be operated independently in one of three configurations: (a) (F+A),
(b) (Air), or (c) shut off completely (fuel valves shut off and air regis-
ters closed). NO ports can be open or closed by means of a damper.
Ji
Variations in the operating configurations of the individual burners and
the NO ports are a large part of techniques to reduce NO ; the con-
x x
figuration used represents a number of additional independent input
variables equal to the number of burners plus the NO ports. This
number of additional input variables ranges from 12 in the smallest
boiler to 25 in the largest.
2-12
-------
Table 2-1. DESCRIPTION OF BOILER-SPECIFIC INPUT
VARIABLES
PQ
NT
NFW
IWD
NNP
HWF
HDF
HBB
HBNP
AB
HWB
ZLPa
ZLSa
ZLAO3
ZLNP3
Boiler designation code
P = Plant name
H = Haynes
S = Scattergood
L = Harbor (Long Beach)
Q = Specific boiler number within the plant
Total number of burners
Number of firing walls
Number of burners on a given vertical level
Number of NO ports
x r
Horizontal width of the furnace, between nonfinng walls
Horizontal depth of the furnace, between firing walls
(opposed) or between firing wall and back wall (single
wall)
Vertical height between burner levels
Vertical height between the top level of burners and
the NO ports
x r
Cross-sectional flow area of the burners
Horizontal width of the burners (diameter of the burners)
Length of the primary mixing zones
Length of the secondary mixing zones
Length of the adjacent and opposite mixing zones
Length of the NO port mixing zones
2-13
-------
Table Z-l. DESCRIPTION OF BOILER-SPECIFIC INPUT
VARIABLES (Continued)
ZLF = Length of the final mixing zone
FCF = Furnace cooling factor (Appendix A)
FRA, = Constants in the FGR flow calculation from flue gas
FRB fan amperage
ADMNP = Admittance of one NO port
x r
ADMA = Admittance of an (AIR) burner
ADMFG = Admittance of an (F+A) burner with gas fuel
ADMFO = Admittance of an (F+A) burner with oil fuel
All lengths in the flow direction.
All admittances are for air flow through the burners
2.14
-------
Table 2-2. VALUES OF BOILER-SPECIFIC INPUT
VARIABLES
Variable
code
Unit
Boiler type
Hi/2
H3/4
H5/6
Sl/2
L3
Furnace-burner geometry
NT
NFW
IWD
NNP
HWF
HDF
HBB
HBNP
AB
HWB
-
-
.
-
ft
ft
ft
_
ft2
ft
12
1
4
0
49.06
24.0
8.0
.
7.069
3.0
12
2
4
4
36.0
28.0
8.0
7.0
7.069
3.0
24
2
4
2
33.04
30.0
3.67
9.43
5.241
2.583
16
1
4
0
39.41
24.25
6.5
-
4.273
2.333
12
1
6
0
36.58
21.0
5.63
-
3.681
2. 165
Mixing zone
ZLP
ZLS
ZJLAO
ZLNP
ZLF
ft
ft
ft
ft
ft
6.0
6.0
12.0
-
40.0
6.0
3.0
6.0
10.0
40.0
5. 17
3.28
6.56
10.0
52.0
4.667
6.545
13.05
-
34.0
4.33
5.56
11.11
-
28.0
Cooling factor
FCF
-
1.742
1.343 | 1.0
2. 177
2. 147
FGR flow
FRA
FRB
Ib/sec-amp
amp
3.947
9.2
1.579
22.7
1.825
15.27
0.538
35.5
0
0
Burner air flow admittance
ADMNP
A DM A
ADMFG
ADMFO
ft-lb1/2/sec
ft-lb1/2/sec
ft-lb1/2/sec
ft-lb'/2/sec
0
10.0
4.635
5.277
2.5
10.0
5.056
7.114
10.0
5.4
4.036
4.692
0
6.0
3. 100
4.252
To convert Multiply
from to by
ft m 0.3048
Ib kg 0.4536
ft-lb1/2/sec m-kg'/Z/sec 0.2053
0
2-15
-------
Only three other variables related to operating conditions
are considered primary independent input variables in this study: (a)
the fuel flow rate, (b) the total combustion air flow rate, and (c) the
combustion air temperature. The fuel flow rates are directly mea-
sured by flow-meters. In nearly all cases, the air flow rates were
calculated from the A/F ratios, determined from the ratio of O9 and
£
CO2 in the flue gas analyses and the measured fuel flow rates. Com-
bustion air temperatures were measured only at the exit to the air
preheater and, lacking any other data, were taken as the combustion
air temperatures in the windbox.
Thus, the total number of independent variables which
can enter into the correlation equation is potentially as high as 41: (a)
9 related to fixed furnace geometry, (b) 4 related to fixed operating
conditions in a given furnace, (c) 12 to 25 related to variable burner
configurations, and (d) 3 related to variable operating conditions. This
strongly influences the size of the data sample necessary for meaning-
ful correlation. No specific attempt was made in this study to evaluate
the relative significance of these variables.
In some cases, not all of the data necessary to calculate
values of the primary independent operating condition variables were
recorded. This was particularly true of the combustion air tempera-
tures. In anticipation of this problem, a number of intermediate em-
pirical correlations were established relating the desired primary
variables to other more commonly recorded secondary data. These
were used to fill in missing data. In all cases, the measured fuel flow
rate and combustion air temperatures, as well as the air flow rate cal-
culation, were used where all necessary data were available. Signifi-
cant data conversion calculations and the intermediate correlations are
discussed in Appendix C. A summary of the data sample is contained
in Table 2-3. A listing of values of boiler load, air and fuel flow rates,
A/F ratios, FGR flow rates, and combustion air temperatures are
2-16
-------
Table 2-3. SUMMARY OF THE TOTAL DATA SAMPLE
Boiler
type
Hl/2
H3/4
H5/6
Sl/2
-
Rated
load, MW
240
240
350
180
82
Firing
type
Single wall
Opposed
Opposed
Single wall
Single wall
NOX
ports
No
Yes
Yes
No
No
No. of
burners
12
12
24
16
12
Boiler
code
HI
H2
H3
H4
H5
H6
SI
S2
L3
No. of tests
Gas
11
20
Total 31
39
12
Total 51
115
30
Total 145
7
33
Total 40
8
Total
Oil
12
33
45
6
8
14
39
22
61
18
23
41
16
-
I
-J
Total samples
Total gas and
oil samples
275 177
452
-------
listed in Appendix D, along with the measured values of NO, Q^* anc*
CO?, for each boiler and test condition. A brief discussion of the flue
gas analysis technique and apparatus is also contained in Appendix D.
2.1.3.2 Preliminary Correlations
The initial correlation equation consisted of 24 terms:
22 describing the thermal NO generation in the various types of mixing
ji:
zones, 1 for bound-nitrogen conversion, and 1 constant. As discussed
in Appendix A, this was reduced to eight terms for thermal NO , plus
.X
the bound-nitrogen term and the constant. The regression analysis
program analyzed the significance of each of the terms used in the cor-
relation equation. Early correlations indicated that the "final-final"
mixing zone, downstream of the NO port mixing zone, plus some of
ji
the individual bulk gas mixing zones, were of little significance to the
correlations. The "final-final" zone term therefore was eliminated,
and all six of the bulk gas zones were combined into one term. Other
combinations of terms resulted from consideration of the number of
independent variables in the correlation equation and the size of the
data samples to be correlated.
The remaining 18 terms in the correlation equation
represented 18 unknown linear coefficients to be determined by cor-
relation of the data and as many as 41 independent variables. Theo-
retically, the 18 unknown coefficients could be determined with only
18 sets of data, in which all of the terms in the correlation equation
varied. In fact, correlation of such a data sample could be exact, with
a correlation coefficient of 1.0. If there is significant data scatter or
the variations of some of the terms are small, an extremely strange
correlation could result, indicating erroneous influences of the terms.
It might be possible, however, for different combinations of the 41
possible independent variables represented in the 18-term equation
to yield the same values for the 18 terms but different values of NO .
2-18
-------
The 18 coefficients determined by correlation, then, would also be
different. Thus, a data sample significantly larger than the number of
terms in the correlation equation, perhaps larger than 41, was neces-
sary to assure that a best estimate of the coefficient of each term was
derived. For evaluation of the best combination of terms in the cor-
relating equation for the data sub samples of interest, a number of cor-
relations were made with different numbers and combinations of terms.
A decision on the appropriate terms for the desired data samples was
based partly on the correlation coefficients resulting from these
correlations.
The correlation coefficient R is defined as
R= Vl- VNCT
where VRES is the variance of the differences between the measured
values of NO and the values of NO calculated from the correlation equa
tion and VNO is the variance of the differences between the measured
values of NO and the average value of NO of the data sample.
In 108 test conditions including all boilers and both gas
and oil fuels, NO emissions were simultaneously sampled and inde-
pendently analyzed by two of three different organizations. The stan-
dard deviation of the differences between these analyses was 29.9 ppm,
and standard deviation of the NO sample itself was 117.6 ppm. Thus,
even if the values of NO calculated from the correlation equation were
exactly correct, the correlation coefficient, Eq. (2-2), could not be
higher than 0.967. Obviously, similar degrees of random scatter in
the measured data are input to the correlation equation. Because of
the complexity of the equation, however, it is difficult to estimate the
(nonlinear) effect of such scatter on the ultimate correlation coefficient.
2-19
-------
On the basis of the effect of NO scatter alone, it was estimated that
data scatter in all of the measured data could easily limit the correla-
tion coefficient, for a realistic correlation, to less than about 0.92.
If the data sample or the range of data variations be-
comes too small for the number of terms in the correlating equation,
correlation coefficients greater than 0.92 could be obtained. This
would indicate that the correlation was being dictated by random scat-
ter in the measured data rather than by true physical relations between
the variables. Therefore, any data correlations yielding correlation
coefficients greater than about 0. 9 were viewed with suspicion and
avoided. On the lower end, it would be desirable to have correlation
coefficients for all data samples near 0. 9, to explain as much of the
data variation as possible. Thus, the appropriate data samples and
terms in the correlating equation were selected, at least partly, to
achieve correlation coefficients as high as possible but not greater than
0.9. In some of the correlations obtained during the trial period, cor-
relation coefficients as high as 0.921 were obtained. In this particular
case, using 10 terms in the correlation equation to correlate 45 test
conditions on 1 boiler type, the coefficients of some of the terms be-
came large positive and negative values. This made the values of NO
calculated from the equation extremely sensitive to small changes in
the input parameters affecting those terms. Such a correlation equation
would be useless in attempting to interpolate or extrapolate the effects
on NO of those input variables. These considerations also entered into
the selection of the data sample sizes to be correlated and the terms to
be used in the correlation equation.
The final groupings into the 10 terms used in all corre-
lations discussed are listed and described in Table 2-4. Final corre-
lations involved more than 61 test conditions.
2-20
-------
Table 2-4. DESCRIPTION OF TERMS IN CORRELATING
EQUATION
Term
no.
Zone code
Type of term
1
8
9
10
X1P16
X2R16, X2S14,
and X2S05
X4R16, X4SJ4,
and X4S05
X5R16, X5S14,
and X5S05
X3O01 and X3O02
X6O01 and X6O02
DNF 1 through
DNF 6
DNST
Primary zones at all levels associated
with (F+A)
Recirculation zones at all levels and
secondary zones at levels 1-5 associated
with burner configurations involving (F+A)
vertically above (F+A) (or above a zero
flow burner); same A/F ratio as 1
Same as term 2 except associated with
burner configurations involving (F+A)
vertically above (Air)
Same as term 2 except associated with
burner configurations involving (Air)
vertically above (F+A)
Horizontally adjacent or opposite zones
at levels 1 and 2 associated with burner
configurations involving (F+A) adjacent
or opposite (F+A) (or adjacent or opposite
a zero flow burner); same A/F ratio
as 1
Same as 5 except associated with burner
configurations involving (F+A) adjacent
or opposite (Air)
All bulk gas zones
NO port mixing zone
Ji:
Bound nitrogen term
Constant
2-21
-------
2.2 RESULTS
The general objectives of this study were to explain the
effects of certain hardware and operating condition variations on NOx
emissions in fundamental combustion terms and to use this increased
understanding to suggest practical hardware and operating conditions to
achieve minimum NO emissions. As with all data correlations, small
interpolations within the conditions tested are reasonably reliable but
interpolations over large data gaps and large extrapolations outside of
the data sample must be treated with care. Such correlations, however,
make the interpretation of the effects of hardware and operating condi-
tions on NO emissions easier and more reliable. The extrapolations
to new untested configurations to minimize NO are more difficult and
j£
must depend partly on the understanding developed within the range of
conditions tested and backed up by available data. This understanding
can be achieved by (a) examining the correlation equations themselves,
(b) conducting parametric studies with the correlation equations by
varying individual input parameters over the range tested, and (c) using
the understanding developed to dictate further parametric studies over
variable ranges not tested. Each of these processes are discussed in
the following paragraphs.
One of the most important areas of understanding to be
gained from analysis of this large quantity of data from full-scale
multiple-burner boilers is that involving the smaller-scale mixing of
the air and fuel from the multiple-burners within the boiler. It would
be desirable to track the actual mean molecular A/F (gas or vapor)
ratio of each of the burner flows through the various gross mixing zones
at the burner level and after mixing with the bulk gases and NOx port
flow.
In gas firings, air and fuel are being mixed in weight
proportions larger than 9:1, beginning with gross mixing within the
2-22
-------
burner. In oil firings, air and fuel vapor mixing proportions start at
infinity at the burner and decrease to proportions as low as 9:1 in the
furnace as the liquid fuel slowly vaporizes. It seems reasonable to
visualize the small-scale mixing process as the addition of fuel to the
air. Figure 2-4 is a schematic showing how the mean molecular-scale
A/F ratio of the combustion air might change with time as the fuel is
mixed in. Since oil fuels must first vaporize before air-fuel vapor
mixing can occur, the A/F ratio of the air must change more slowly
with oil fuels than with gas fuels. The case shown is one where the
burner is being operated off-stoichiometric. Eventually, the mean
molecular-scale A/F ratio reaches the same low level for both the gas
and oil fuels. Also shown in Figure 2-4 is a band of A/F ratios near
stoichiometric where the rate of NO generation is high. This band is
shown decreasing in width, with time, to simulate the reduction in gas
temperature resulting from heat rejection to the water walls.
The schematic shown in this figure is convenient for
visualizing several important aspects of the NO control problem:
Ji
a. According to the A/F ratio, temperature, time-history
approach to minimum NOX emissions, it is necessary
to control the mean molecular-scale A/F ratio so that
it skirts the edges of the high NOX generation zone or
spends as little time in it as possible.
b. Those regions in the furnace where this mean A/F ratio
crosses the high NOX rate zone should appear in the cor-
relation equation as regions of proportionally large con-
tributions of NO to the total.
c. The region where this mean A/F ratio first crosses the
high NOX zone, in the initial mixing with and within a
given burner flow, compared with the region where this
same burner flow enters the bulk gases (enters a region
of significant mixing with the already mixed flow from
burners below) is critical to the arrangement of operat-
ing (F+A) burners and (Air) burners in the multiple
burner array.
2-23
-------
GAS OIL
REGION OF HIGH
NOX GENERATION
o
<
OVERjALL_BOILER_
STOICHIOMETRIC
BURNER
TIME AND DISTANCE FROM BURNER EXIT
Figure 2-4.
Schematic of possible variations of the average
molecular scale A/F ratio of the combustion
air as gas or oil-vapor mixes with it
2-24
-------
In the initial development of the correlation equation,
the assumption was made, although it was known to be incorrect, that
all air-fuel mixing was complete, on a molecular scale, in the flow
leaving the burners. This assumption results in calculated values of
the NO contributions from the early mixing zones, which are too high.
If the correlations of the data are good, however, small coefficients
will be assigned to those terms representing these early mixing zones
such that the assumption error is corrected. The resulting indications
of the effects of each of the zones on the NO variations should still
indicate where in the furnace the mean molecular-scale mixing is in, or
is slowly passing through, the high NO generation zone of Figure 2-4.
Ji
2. 2. 1 Observations Directly from the Correlations
Table 2-5 shows the coefficients obtained from the cor-
relation of various data samples, using the 10-term correlation equation
described in Table 2-4. The correlations are grouped separately for
natural gas and low-sulfur oil fuels. The correlation coefficients shown
in the table represent correlations of data from a single boiler (H5) or
a boiler type consisting of two identical boilers (H5/6), opposed-fired
boilers using NO ports (H3/4 + H5/6), single-wall-fired boilers with-
Ji
out NO ports (Hl/2 + Sl/2) and all data for a single fuel. The number
of test conditions used in each of the correlations and the resulting cor-
relation coefficients are also shown in Table 2-5.
The correlation coefficients of the gas-fired data shown
in Table 2-5 are all in the desired range, indicating good but not mis-
leading correlations. The correlation of the single-wall-fired boilers
without NO ports (Hl/2 + Sl/2), involving only 71 test conditions,
3t
shows a correlation coefficient slightly above 0.9. The abnormally
large coefficients for terms 7 and 8 show the tendency toward a high
correlation coefficient but, perhaps, a misleading correlation when the
amount of data correlated becomes relatively small. The gas
2-25
-------
Table 2-5. COEFFICIENTS OF THE TERMS IN THE CORRELATION
EQUATIONS FOR VARIOUS DATA SAMPLES
Term no.
1
2
3
4
5
6
7
8
Constant
No. of
tests
Correlation
coefficient
Natural gas fuel data samples
H5
-12.50
+9.922
-0.2062
-6.126
-10. 11
+0.5255
+4.383
+ 17.82
+201.1
115
0.841
H5/6
-12.08
+9.576
-0.3184
+154.0
-9.720
+0.1464
+4.971
+8.725
+213.4
145
0.859
H3/4 - 5/6
-0.9526
+0.6911
-0.0646
-192.4
-0. 1465
+0.7776
+3.043
+15.20
+220.6
196
0.844
Hl/2 - Sl/2
-0.6629
+0.6744
+0.2729
-418.8
-2.276
+6.744
-154. 1
+ 1150
+101.8
71
0.902
All gas
-0.4271
+0.3927
-0.018Z
-167. 1
-0.2393
+0.8473
+3.473
+20.07
+184.7
267a
0.837
M
I
8 tests on L3 not used in correlation
-------
Table 2-5. COEFFICIENTS OF THE TERMS IN THE CORRELATION
EQUATIONS FOR VARIOUS DATA SAMPLES (Continued)
Term no.
1
2
3
4
5
6
7
8
9
Constant
No. of
tests
Correlation
coefficients
Low -sulfur oil fuel data samples
H5/6
+0.6693
-0.5180
-0.4389
+ 1401
+0.5043
+0.7910
+0.5915
+2.589
-5.028
+ 1012
61
0.860
H3/4 - 5/6
+0. 1876
-0. 1315
-0.3846
+ 1123
+0.0643
+0.6324
+0.7151
+2.306
-1.551
+457.2
75
0.849
Hl/2 - Sl/2
-0.00623
+0.01304
+0. 1234
-433. 1
+0.3417
+2.388
-19.45
+148.0
+1.260
-13.72
86
0.680
All oil
-0.0221
+0.0261
-0.0280
-139.5
-0.0233
+1.480
+0.4563
+3.659
+2.573
-205.9
I6la
0.734
N
I
ro
16 tests on L3 not used in correlation.
-------
correlations also show that the correlation coefficients are relatively
constant regardless of the variety of types and sizes of boilers intro-
duced into a single correlated data sample. The correlation of all of
the gas-fired data yields a correlation coefficient nearly identical to
that for the single boiler (H5) sample alone.
The correlations of the oil-fired data show good corre-
lation coefficients for all of the data groupings except those involving
the single-wall-fired boilers (Hl/2 + Sl/2). The poor correlation is
related to the Sl/2 data. The data sample on Sl/2, however, is too
small (41 tests) to properly evaluate that correlation alone. Corre-
lation coefficients for the Sl/2 data only are 0. 960 and 0. 503 for gas
and oil, respectively. An error in the Sl/2 oil-fired input data is
suspected. At this writing, that error has not been found.
The coefficients listed in Table 2-5 are not sufficiently
revealing with respect to the NO generation processes without the
jt
calculated values of the corresponding parameters. Average values of
these parameters, therefore, were calculated for each data sample
and are shown in Table 2-6. The products of the coefficients and the
parameters for each term yielded values in units of NO, ppm, which
can be thought of as influence coefficients of the zones and burner com-
binations represented by each term of the total NO emissions. The
values of these terms are listed in Table 2-7.
The gas correlations involve constants with values over
the positive range from 101 to 221 ppm. In the oil correlations, the
term representing the bound nitrogen (term 9) for any one correlation
is also essentially constant. The effective constants for the oil corre-
lations, shown in Table 2-7 as "9 + CONSTANT," also have values
over the positive range, between 187 and 239 ppm. The variations in
total NO emissions, therefore, are controlled by terms 1 through 8.
From the standpoint of distance within the furnace from
the burner exits, term 1 is closest, representing the primary zones
2-28
-------
starting at the burner exits. Terms 2 through 4 represent the
recirculation and secondary zones. Terms 5 and 6 represent the
more distant adjacent and opposite mixing zones, while term 7 repre-
sents the bulk gas zones and term 8 the NO port zone. Considering
Jt
the correlations of the largest, most complex boiler type (H5/6) and
the correlations of all gas and oil data, an initial estimate of the in-
fluence of distance from the firing walls on the variations in NO can
Jt
be generated. Summing those terms in Table 2-7 describing the same
mixing zones yields the following-
Zone
Primary
Recirculation and secondary
Adjacent and opposite
Bulk gases
NO ports
x r
Value of terms
H5/6
Gas
-3376
+4251
- 849
+ 75
+ 15
Oil
+ 975
-1204
+ 225
+ 49
+ 6
All data
Gas
-120
+146
- 14
+ 36
+ 9
Oil
-26
+38
- 3
+ 15
+ 3
Several preliminary observations can be made from these data: (a)
The NO variations with gas fuels are strongly dominated by variations
2C
in the early mixing zones, while those with oil fuels are more distrib-
uted; (b) the variations in NO with both fuels are more distributed and
5C
are proportionately about the same for gas and oil when the smaller
single-wall-fired boilers are added into the sample; and (c) the effects
of mixing zones at a given furnace level'(exclusive of the bulk gas and
NO port zones) are opposite between gas and oil fuels in the large
J*.
boiler.
One of the most important observations obtained from
this data analysis concerns the location within the furnace where the
mean small-scale A/F ratio is in, or is passing through, the region
2-29
-------
Table 2-6. AVERAGE INPUT VALUES FOR THE PARAMETERS IN EACH
VARIABLE TERM IN THE CORRELATION EQUATIONS FOR
THE VARIOUS DATA SAMPLES
Term no.
1
2
3
4
5
6
7
8
Natural gas fuel data samples
H5
269.7
427.1
81.60
0.0551
84.20
2.904
16.95
0.8117
H5/6
279.5
445.8
87.42
0.0622
90.50
2.304
18. 14
0.7646
H3/4 - 5/6
258.7
408.7
70.72
0.0817
89.11
3. 148
13.88
0. 5716
Hi/2 - Sl/2
342.3
408.3
65.32
0.0707
7.650
1. 568
0.1319
0.0272
All gas
281.0
408.6
69.28
0.0788
67.45
2.728
10.22
0.4268
N
I
1*1
O
-------
Table 2-6. AVERAGE INPUT VALUES FOR THE PARAMETERS IN EACH
VARIABLE TERM IN THE CORRELATION EQUATIONS FOR
THE VARIOUS DATA SAMPLES (Continued)
Term no.
1
2
3
4
5
6
7
8
9
Low-sulfur oil fuel data samples
H5
(not
corre-
lated)
H5/6
1456
2298
125.3
0.0296
445.3
0.8339
82.65
2.303
159.5
H3/4 - 5/6
1485
2392
101.9
0.0240
546.8
0.6783
69.84
1.931
158.4
Hl/2 - Sl/2
918.9
1150
92.39
0.0414
30.09
3.888
0. 1429
0.0185
159.0
All gas
1183
1729
96.81
0.0333
270.8
2.393
32.61
0.9095
158.7
-------
Table 2-7. AVERAGE VALUES OF THE TERMS IN THE CORRELATION
EQUATIONS FOR THE VARIOUS DATA SAMPLES
Term no.
1
2
3
4
5
6
7
8
Constant
Totals
Natural gas fuel data samples
H5
-3371
+4238
-17
-0
-851
+2
+74
+ 15
+201
+296
H5/6
-3376
+4269
-28
+ 10
-880
+0
+90
+7
+213
+305
H3/4 - 5/6
-246
+283
-5
-16
-13
+2
+42
+9
+221
+277
Hl/2 - Sl/2
-227
+275
+ 18
-30
-17
+ 11
-20
+31
+ 102
+ 143
All gas
-120
+ 161
-1
-13
-16
+2
+36
+9
+ 185
+243
N
I
UJ
N
Units of each term in ppm
-------
Table 2-7. AVERAGE VALUES OF THE TERMS IN THE CORRELATION
EQUATIONS FOR THE VARIOUS DATA SAMPLES (Continued)
Term no.
1
2
3
4
5
6
7
8
9
Constant
Totals
No. 9 and
constant
Low-sulfur oil fuel data samples3-
H5/6
+975
-1190
-55
+41
+225
+ 1
+49
+6
-802
+ 1012
+262
+210
H3/4 - 5/6
+279
-315
-39
+27
+35
+0
+50
+ 5
-246
+457
+253
+211
Hl/2 - Sl/2
-6
+ 15
+ 11
-18
+ 10
+9
-3
+3
+200
-14
+207
+ 186
All oil
-26
+45
-3
-5
-6
+4
+ 15
+3
+408
-206
+229
+202
t\>
I
Units of each term in ppm.
-------
of high NO generation rates. The data show that the NO variations
X X
possible in the primary, recirculation, and secondary zones are larger
than those in the bulk gas zones by about a factor of 50 for gas fuels but
by only a factor of about 20 for oil fuels in the large boiler. Even the
total data samples for each fuel show that these zones are in the ratio
of about 3. 5 for gas to 2. 1 for oil. The adjacent and opposite mixing
zones in the large boiler are also about three times as significant with
gas fuels as with oil fuels. In the total data sample, however, the
adjacent and opposite zones are small compared with the bulk gas zones
for both fuels. A general preliminary observation from this data might
be that the small-scale mixed A/F ratio for gas fuels passes rather
quickly through the region of high NO generation rate in the early mix-
JL
ing zones (but not completely in the primary zone), while that for oil
passes through more slowly and over a larger distance into the furnace.
These observations are not surprising except, perhaps, that the gas
mixing is apparently not complete in the earliest (primary) mixing
zones.
The more distributed effects of the various zones on the
NO with the smaller boilers added into the sample are also not very
surprising when the definitions of these zones are considered. The
larger boiler has six vertical levels of burners with no divider wall,
while all other boilers have only 3 to 4 vertical levels and most have
a divider wall of some type. Thus, the early mixing between burners,
both vertically and horizontally, tend to disappear, and the boiler
begins to approach the behavior of a single-burner boiler. Hence, the
data from the smaller boilers provide less information on the effects
of multiple-burner arrays on NO , and the multiple-burner configura-
JL
tion effects tend to become obscured. The correlations tend to be about
the same for gas and oil fuels when the smaller boilers are included
in the data sample. Single-burner laboratory-type information is
much more applicable in these cases.
2-34
-------
The data for the large boiler type show that the algebraic
signs of the first three mixing zone types are exactly opposite for the
cases of gas and oil fuels. This is a significant observation, which is
discussed throughout this report. Table 2-7 shows that recirculation
and secondary zone effects (terms 2 through 4) on NO are dominated
X-
in the large boiler correlations by term 2. This term represents those
burner configurations where an (F+A) burner is directly below any
given (F+A) burner. Term 1 represents those configurations where
that same given burner is (F+A). Obviously, these two terms are not
independent of each other with respect to burner configurations. A
configuration where a given burner is in the bottom row is considered,
in this model, to be the same as one which has an (F+A) burner below
it, because the same A/F ratio results in the recirculation and second-
ary mixing zones. An (F+A) burner in the top two rows, in this model,
yields a smaller value for term 2 because the model assumes that the
bulk gases have spread sufficiently to reduce, and eliminate in the top
row, the secondary mixing zones. In the top row, flows from the
primary and recirculation zones issue directly into and mix with the
bulk gases. Starting with a burner array where all burners are (F+A),
term 1 will always be reduced by the effect of one (F+A) burner if the
fuel to one burner is shut off. Term 2, however, will be reduced dif-
ferently, depending on where the resulting (Air) burner is located
vertically in the burner array: (a) If the (Air) burner is in levels 1 to
4, the number of configurations involving two vertically adjacent (F+A)
burners will be reduced by two (both vertical combinations involving the
(Air) burner); (b) if the (Air) burner is in the top level, the number of
vertically adjacent (F+A) burner configurations will be reduced by only
one since there is no (F+A) burner above it; and (c) if the (Air) burner
is in either of the top two levels, the reduction in the value of term 2
will be less because the secondary mixing zones are reduced or absent
at these levels.
2-35
-------
Similarly, the effects of the adjacent and opposite
mixing zones in the large boiler correlations are dominated by term 5,
representing adjacent and opposite mixing between flows from two
(F+A) burners on the same vertical level. Because of the bulk gas
spreading, however, this kind of mixing was assumed to occur only
in the bottom level and partially in the second lowest level.
The foregoing observations may seem to be strongly
dependent on the specifics of the model used here. These are the
specifics, however, which were input to the correlation equation.
The subsequent correlations roughly compensate for the inaccuracy
of the specific inputs by determining the appropriate coefficients to
best fit the real data.
The large boiler correlations (Table 2-7) show that to
reduce NO emissions with gas fuels it is desirable to keep the value
of terms 1 and 5 as large as possible while reducing the value of term
2. The case of oil, however, is exactly the opposite, and the conclu-
sions, therefore, are opposite. For gas fuels, these conclusions indi-
cate that the total number of (F+A) burners should be kept as large as
possible (to maximize term 1) and no (Air) burners should be located
in either the bottom two levels or the top two levels of the burner array
(to maximize term 5 and minimize term 2). This implies just a few
(Air) burners located in the vertical midlevels of the array.
For oil fuels, the data indicate that a larger number of (Air) burners
should be used (to minimize term 1) and these (Air) burners should be
located either in the top level (to maximize term 2) or in the bottom
row (to minimize term 5). For minimum NO , the better of these two
3C
locations is dependent on the relative variations in terms 2 and 5 and,
perhaps, on the additional effects of (Air) burners in these levels on
bulk gas emissions (term 7). In smaller boilers, with less freedom in
vertical location of (Air) burners, both gas and oil fuel data tend to
look like the gas case, except that the bulk gas and NO port zones may
2-36
-------
be more important. The effect of even a single (Air) burner located
at various vertical levels in a furnace is shown schematically in
Figure 2-5. (Air) burners located low in the furnace keep the bulk
gas A/F ratio above that of the overall furnace throughout the burner
region. A single (Air) burner located in the top row may cause the
bulk gas A/F ratio to remain close to stoichiometric throughout most
of the burner region. Larger numbers of (Air) burners at these loca-
tions, of course, would drive the A/F ratios shown in Figure 2-5 to
much lower and higher values.
These general observations on the correlation equations
tend to verify the slower mixing of the air and fuel vapor in the case of
oil fuels, as postulated in Figure 2-4. There is some indication, how-
ever, that the air-gas fuel mixing may be slower than expected and may
not be complete before the burner flows at the highest levels are forced
to mix with the bulk gas flow coming from lower levels. A first indi-
cation of appropriate (F+A) and (Air) burner configurations for minimum
NO can also be obtained. These best configurations may be exactly
Ji
opposite for gas and for oil fuels in large multiple-burner boilers but
may be quite similar in smaller boilers with a limited number of ver-
tical levels. Further understanding of the correlation equations can
also be obtained by parametric exercise of the correlation equations.
2. 2. 2 Parametric Studies of the Correlations
In this section, discussion concerns some of the para-
metric computer runs intended to further explain the correlation equa-
tions. Subsequent sections discuss parametric studies conducted to
explain the effects of the combustion modifications tested and those
conducted to extrapolate, beyond the existing data, conditions leading
to minimum NO emissions. The primary question addressed in this
Jt
section, however, is the meaning of the constants in the correlations.
2-37
-------
00
14
o
5
to
GAS FUEL, NO PORTS CLOSED
A
LOCATION OF SINGLE (Air) BURNER:
FIRST LEVEL
FOURTH LEVEL
SIXTH LEVEL
l
16 18 20 22
BULK GAS A/F RATIO
24
Figure 2-5. Effects of a single (Air) burner on the bulk
gas A/F ratio through the H5 boiler
2-38
-------
Obviously, the NO contributions from each mixing zone
and term in the correlation equation, in reality, are positive. The
correlations, however, are positive and negative variations around a
constant. If the constant were properly distributed among the various
terms, each term would be positive. The difficulty in this case, how-
ever, is that those conditions which would yield a zero value of NO
when all of the variables go to zero (a zero intercept) are not in the
data sample since no zero NO measurement was ever obtained. To
force a zero intercept in this case could represent an excessive and
perhaps erroneous extrapolation. In these empirical correlations,
then, it is unlikely that the intercept would be zero. Care must be
taken in parametric exercising of the correlation equations to avoid
drawing fundamental conclusions from extrapolations far beyond the
existing data.
Only one term in the gas correlations is very nearly
constant, i.e. , term 8, the NO port mixing zone. Since all of the
X.
air and fuel flows are involved in this zone, regardless of whether the
NO ports are open or closed, the A/F ratio is the overall furnace
X
ratio and varies only slightly with overall excess air. The tempera-
ture in the zone depends only on the total cooling of the burner flows
enroute to this zone and the initial combustion air temperature. Since
this zone is always within the A/F ratio region of high NO generation
X.
(Figure 2-4), it might be expected that the NO contribution from this
X.
zone is relatively high compared with other zones, unless the cooling
rate is high. In the correlations of gas-fired data, therefore, the
average value of the NO port zone contribution to the total NO emis-
X X.
sions might be better represented by the sum of the constant and the
smaller variable term 8. The same is true for oil fuels if the constant
is taken as term 9 plus the true constant.
The possibility that the constant in the gas correlations
is associated with the NO port zone can be investigated on the basis
2-39
-------
of operating conditions in a real boiler: (a) If sufficient combustion air
flow were diverted to the NO ports, the NO contributions of all zones
J± A.
except the NO port zone should go to zero; (b) if the furnace cooling
?t
rate were increased sufficiently, the NO contributions from all zones,
particularly those further downstream (including the NO port zone),
JL.
should go to zero; and (c) if the combustion air temperature were re-
duced sufficiently, the NO contributions from all zones should go to
Jt
zero. However, because the combustion air temperature above am-
bient is small compared with the temperature rise due to reaction,
case {c) may not be reached with any realistic reductions in combustion
air temperatures. Direct observation of the H5 correlation equation
indicates that the limit of NO with large NO port flows should be
A\ 3t
slightly larger than the value of the constant. The realistic limits in
cases (b) and (c) cannot be determined by direct observation.
The same extreme limits in these three operating con-
dition variables with oil fuels should yield the same results except that,
instead of total NO values going to zero with high cooling rates and
li
reduced combustion air temperatures, these values should approach
that of bound-nitrogen conversion. Again, direct observation of the
H5/6 correlation for oil fuels indicates that the limit of NO with large
3t
NO port flows should be slightly larger than the value of term 9 plus
3t
the constant (210 ppm). If cases (b) and (c) are realistic, NO values
Ji
should also approach 210 ppm but would then represent the bound-
nitrogen conversion. In the investigation of the NOx variations result-
ing from these three extremes of operating conditions, the H5 correla-
tion for gas fuel and the H5/6 correlation for oil fuel were used in
parametric calculations in which the three variables involved were
varied (singly) over wide but realistic ranges.
For case (a), the maximum NOX port admittance in the
data for H5/6 is 4. 1 m-kg 1 /2/sec (20 ft-lb 1 /2/sec). It was decided
to parametric ally increase this admittance to three times the maximum
2-40
-------
data value, recognizing the degree of extrapolation. For case (b),
cooling rates in the smaller boilers, with dividing walls, are as much
as 2. 2 times higher than that for the H5/6 boilers. In this case, it
was decided to also increase the H5/6 cooling rate by a factor of three,
again recognizing the degree of extrapolation. For case (c), the data
indicate that combustion air temperatures were never reduced more
than about 45 degrees K (80 degrees F). With the same factor of three
for extrapolation, temperatures no lower than about 440 degrees K
330 degrees F) are obtained. This reduction represents such a small
fraction of the temperature of the reactants in any zone that it was un-
likely that total NO values would be reduced anywhere near zero.
X.
Nevertheless, it was decided to make the parametric calculation by re-
ducing the combustion air temperatures to ambient. This represents
an extrapolation by more than a factor of six beyond the data.
2.2.2.1 Natural Gas Fuel
Figures 2-6 through 2-8 show the results of the para-
metric variations of these three operating condition variables for both
gas and oil fuels. In the gas fuel case, Figure 2-6 shows that for NO
j£
port admittance values greater than about 8.4(40) (about twice the data
value for H5), the NO values asymptotically approach a value of 200
ppm, which is the value of the constant for the H5 data correlation.
Figure 2-7 shows that increasing the cooling rate for the H5 data to
values greater than about 2.9 times the data value decreases the NO
to zero and even to negative values. The negative values may indi-
cate the limits of reasonable extrapolation for the cooling rate in this
data correlation. Both of these calculations indicate that the constants
in the gas fuel correlations are very likely associated with the NO
Ji
port mixing zone. Figure 2-6 also indicates that the single approach
of increasing NO port flow (two-stage combustion) can be effective
initially but cannot reduce NO emissions below a still significantly
2-41
-------
700
600
500
400
300
200
H5/6 BOILfRS, 24 BURNERS
OPERATING, 3% 02, FULL LOAD
10
20
10 12 m-kg1/2/sec
I I I
30
40
50
TOTAL NO PORT ADMITTANCE
A
60 ft-lb1/2/sec
Figure 2-6. Effects of NO port admittance
2-42
-------
600
500
400
300
200
100
H5/6 BOILERS, 24 BURNERS
OPERATING, 3% 02, FULL LOAD
NO PORTS CLOSED
A
0 1.4 1.8 2.2 2.6 3.0
FURNACE COOLING RATE RATIO TO NOMINAL
Figure 2-7. Effects of increased cooling rate
2-43
-------
600
500
400
300
200
H5/6 BOILERS, 24 BURNERS OPERATING,
3% 02, FULL LOAD, NOX PORTS CLOSED
100
300
400
I
500
I
I
I
I
100 200 300 400
COMBUSTION AIR TEMPERATURE
500 °F
Figure 2-8. Effects of reduced combustion air
temperatures
2-44
-------
high level. Figure 2-8 appears to indicate that NO values for the gas
fuels are asymptotically approaching the value of the constant as com-
bustion air temperatures are lowered. Considering the large range
of extrapolation represented by ambient temperatures, however, it is
not known whether the calculated NO values are simply not represen-
tative with temperatures near ambient or whether a zero NO value
might be approached with further reductions. The values of terms 1
through 8, all temperature sensitive terms, have not reached zero
with ambient combustion air temperatures.
2.2.2.2 Low-Sulfur Oil Fuel
In the oil fuel calculations, Figure 2-6 shows that for
admittance values greater than about 10(50), the NO values approach
225 ppm. This is just 15 ppm greater than the constant (210 ppm).
It is entirely reasonable that this 15 ppm represents that part of the
NO formed from the fuel-bound nitrogen under these low A/F ratio
conditions in the burner region of the furnace. The remaining 210
ppm could reasonably represent that thermal NO formed in the NO
ji.
port zone. Figure 2-7 shows that the calculated values of NO reach
a minimum of 270 ppm for cooling rates about 2. 3 times the nominal
for H5/6. Beyond that cooling rate, the NO values begin to increase
again. Since there is no apparent physical reason why this should
occur, 2.3 times the nominal cooling rate is considered the limit of
extrapolation of this parameter for the H5/6 oil correlation. At this
minimum, essentially all sources of thermal NO should be reduced to
zero, and the remaining NO should all originate from fuel-bound nitro-
gen conversion. Figure 2-8 shows that, despite the large extrapolation
represented by combustion air temperatures near ambient, the cal-
culated NO values appear to be asymptotically approaching a level which
could represent the bound-nitrogen conversion. The low-sulfur fuel oil
used contained 0. 24 percent bound nitrogen, by weight. If all of this
2-45
-------
were converted to NO, 353 ppm (dry at 3% O^) would result. Single
burner laboratory tests (Ref. 2-2) indicate that the conversion effi-
ciency for 0. 24 weight percent bound nitrogen fired with 3 percent
excess O2 should be about 63 percent. This would yield NO values in
this case of 222 ppm. This is very close to the apparent asymptotic
value in Figure 2-8 and not too far from the 270 ppm minimum shown
in Figure 2-7. For the moment, the agreement appears sufficiently
close to indicate that under these low temperature conditions the con-
stant represents the NO from conversion of the fuel-bound nitrogen.
Thus, the constant in the fuel oil data correlations must be considered
associated with both the NO port mixing zone and the conversion of
bound nitrogen. Figure 2-6 further indicates that, as in the gas-fuel
case, the two-stage combustion approach to NO reduction can be
JC
initially very effective but cannot reduce NO emissions below some
jt
still significantly high level.
In general, the parametric studies of the gas correla-
tion equation indicate that the constant should more appropriately be
associated with the NO port mixing zone, in a form:
x
term 8' = term 8 + constant (2-3)
For the oil correlation equation, the constant is apparently associated
with both the NO port mixing zone and the bound nitrogen in a form:
x
term 8' = term 8 + term 9 + constant (2-4)
2.2.3 Effects of Some Combustion Modifications
A number of combustion modification techniques to
reduce NO are represented in the data. These include (a) load and
combustion air temperature reduction, (b) NO port flow (two-stage
2-46
-------
combustion), (c) excess air reduction, and (d) various BOOS
configurations. Each of these will be discussed separately, sup-
ported by such data as is available. Again, the H5 natural gas data
sample and the combined H5/6 oil data sample will be used for these
discussions.
2.2.3.1 Load and Combustion Air Temperature
In the parametric evaluation of the effects of load re-
duction on NO emissions, the only truly independent parameter which
was varied was the fuel flow rate. The empirical correlations be-
tween fuel flow and load (Appendix C) were used to relate the fuel flow
rates to the appropriate plant load. Air flow was calculated from the
fuel flow rates to maintain the excess O^ constant at 3 percent. Com-
bustion air temperatures, however, are known to be dependent on the
flue gas flow rates through the boiler and the air preheater. The
empirical correlation shown in Appendix C was used to calculate this
temperature for each load. In all cases, all burners were operating
(F+A), and the NO ports were closed. Figures 2-9a and b show the
Ji
resulting calculation of the variation of NO emissions with load for
Ji
the H5/6-type boilers. Available data are also plotted in these figures.
Figure 2-9a shows that the calculated values agree reasonably well with
the gas-fired data except at the midload levels. The large scatter in
measured NO at full load is considered real and will be discussed
Jt
further. Figure 2-9b shows good agreement with the oil-fired data at
all load levels.
Reduced total flow in the calculation of the inputs to the
correlating equations results only in longer cooling times for the gases
enroute to each of the mixing zones. Since a constant cooling rate with
time was assumed, this results in cooler gases in any zone in the fur-
nace. Actually, the cooling rate itself may be a complex function of
flow rates, particularly, if the mechanism for the overall heat rejection
2-47
-------
1000 r-
600
400
200
0
H5/6 BOILERS, 24 BURNERS OPERATING,
3% 0?l NO PORTS CLOSED
t A
O ACTUAL DATA
140
180
CALCULATED
220
260
LOAD, MW
300
O
O
340
380
(a) With natural gas fuel
Figure 2-9. Effects of load variations
2-48
-------
1000
800
H5/6 BOILERS, 24 BURNERS OPERATING,
3% 0?, NO PORTS CLOSED
£ A
O ACTUAL DATA
E
a.
a.
600
400
200
-CALCULATED
140 180 220 260 300
LOAD, MW
340
380
(b) With low-sulfur oil fuel
Figure 2.9. Effects of load variations
(Continued)
2-49
-------
to the water walls is convective to any large extent. Thus, it is
possible that the true average cooling rate could decrease faster with
flow rate than the cooling times increase. This case could result in
an increase in temperature in some or all of the zones in the radiant
section of the boiler and an increase in NO emissions with a decrease
in total flow. The temperature of the flue gases at the boiler exit, how-
ever, after passing through the entire convective section of the boiler,
is clearly reduced. Through the air preheater heat exchanger, the
combustion air temperatures, then, are also always reduced.
To evaluate these separate effects, the total flow and
the combustion air temperatures were varied independently over the
ranges corresponding to the load variations of Figures 2-9a and b.
The results are shown in Figures 2-10a and b. In the gas fuel case,
NO emissions are seen to increase, at first, with decreasing load and
then drop sharply. The decrease in combustion air temperature, how-
ever, more than compensates for the initial effects of flow reduction
such that when both of these parameters vary simultaneously, the NO
values always decrease with decreasing load. The fact that the cal-
culated overall NO curve for gas firings does not fit the data well in
the middle of the load range (Figure 2-9a) may be due to error in the
assumption of a cooling rate independent of total flow. For the oil
firings, NO values decrease with either or with both flow and temper-
ature. In this case, the flow reduction appears to have more of an
effect than does the temperature.
Figures 2-9 and 2-10 concern the effects of load on a
boiler with all burners operating and no NO ports. In such a con-
Ji
figuration, gas temperatures in the furnace are about as high as occur
in any configuration. The effects of those variables associated with
load reduction in other configurations, where temperatures may be
reduced by air-fuel variations, will be much less. In no case in the
test data (applicable to Figures 2-9 and 2-10) did the combustion air
2-50
-------
10001-
800
600
400
200
0
FLOW
I
140 180
TEMPERATURE
ONLY
I
H5 BOILER, 24 BURNERS
OPERATING, 3% 02,
NO PORTS CLOSED
A
I
220 260 300 340 380
LOAD, MW
(a) With natural gas fuel
Figure 2-10. Separate effects of total flow and combustion
air temperature variations with load
2-51
-------
1000
800
E
o.
600
400
200
H5/6 BOILER, 24 BURNERS OPERATING,
3% 0,, N0₯ PORTS CLOSED
£ A
TEMPERATURE
ONLY
FLOW
ONLY
BOTH
1
1
140 180 220 260 300
LOAD, MW
340 380
(b) With low-sulfur oil fuel
Figure 2-10. Separate effects of total flow and combustion
air temperature variations with load (Continued)
2-52
-------
temperature vary independently of load; therefore, these independent
trends cannot be verified.
2.2.3.2 NOX Port Flow
Figure 2-6 shows the independent effects of NO port
.X.
admittance over a very wide range of extrapolation for both gas and
oil fuels. To relate this calculation to the fraction of the total com-
bustion air flow diverted through the NO ports, values'of 19.9(96.9)
ji
and 23.0(112.6) were used for the total admittance of the 24 operating
(F+A) burners (24 times ADMFG and ADMFO, in Table 2-2). These
results are shown in Figures 2-1 la and b. Since only one NO port
JC
configuration was used in the H5/6 data sample and only fully open or
closed NO port data could be used, only two groups of data were
X. <
available to check the calculation from the correlation equation; NO
ji
ports either open or closed. The gas fuel calculation (Figure 2-1 la)
appears somewhat low compared to the NO ports open data, but the
Ji
fit is still reasonable. Again, the wide data scatter with NO ports
2£
closed is shown. The oil fired calculation (Figure 2-lib) fits the
available data very well. In both cases, the slight rise in NO values
as the fraction of flow through the NO ports begins to increase is due
3t
to the burner A/F ratio passing through the high NO generation region
2£
near stoichiometric. The fraction of flow diverted to the NO ports in
the current configuration with both gas and oil is such that the burner
A/F ratios, with all burners operating, are greater than 11. Further
significant reductions in NO could be achieved by increasing the NO
jf,
port admittances alone, without encountering combustion or flame
instability or any other known problems. This approach, however,
would be limited to NO values which are still fairly high (200 and 225
ppm for gas and oil fuels, respectively, in the H5/6 boilers) because
of the relatively constant NO generation in the NO port mixing zone.
Ji
As a single NO reduction technique, however, this would represent
2-53
-------
1000
800
E
o.
600
400
200
O
O
H5 BOILER, 24 BURNERS OPERATING,
3% 02, FULL LOAD
O ACTUAL DATA
I
I
I
0 8 16 24 32
PERCENT OF TOTAL COMBUSTION AIR PASSING
THROUGH N0 PORTS
40
(a) With natural gas fuel
Figure 2-11. Effects of combustion air diverted through
NO ports (two-stage combustion)
2-54
-------
1000
800
E
a.
o.
600
400
200
0
H5/6 BOILERS, 24 BURNERS OPERATING,
3% 02, FULL LOAD
O ACTUAL DATA
9
O
1
CALCULATED
1
I
0 8 16 24 32
PERCENT OF TOTAL COMBUSTION AIR PASSING
THROUGH NO PORTS
n
40
(b) With low. sulfur oil fuel
Figure 2-11. Effects of combustion air diverted through
NO ports (two-stage combustion) (Continued)
2-55
-------
reductions of 69 and 47 percent for gas and oil fuels, respectively,
from the uncontrolled levels, with no decrease in plant efficiency.
2.2.3.3 Excess Air
The interest in excess air variation as a NO reduction
Jv
technique sterns partly from the fact that low excess air improves plant
efficiency. Since the minimum temperature of the flue gas entering the
stack is limited by operational considerations, the less total flue gas
flow or excess air leaving the boiler heat exchange surfaces at this
temperature, the lower are the heat losses.
When the excess air is varied, part of the variation of
measured NO concentration values is due to simple dilution of the NO
by the excess air. Since it is actually the total NO being discharged
to the atmosphere that is significant to atmospheric pollution, direct
values of NO at the excess air level can be somewhat misleading. As
a result, both the calculated and measured values of NO were corrected
to those concentrations which would result if the same NO were diluted
by the products of stoichiometric combustion.
Figures 2-12a and b show the results of these calcula-
tions for gas and oil fuels. The oil fuel calculation, as expected, shows
a decrease in NO values with decreasing excess air but the magnitude
of the reduction is only about 70 ppm over the full 5 percent O-, range.
The data available at the test conditions is minimal but tends to confirm
both the magnitude of the NO calculation and the relatively flat slope
(but not the positive slope). No evidence of excessive CO or smoke was
indicated at any of the test conditions shown.
The gas fuel calculation, however, shows an increase in
NO with decreasing excess air, of as much as 150 ppm over the 5 per-
cent O- range. The little data available at these test conditions (all 24
burners operating, NO ports closed, and full load) is minimal and
Ji
widely scattered. This wide scatter in measured NO data at these test
2-56
-------
conditions has been noted previously. The data shown generally
support the magnitude of the calculated NO curve but cannot confirm
or deny the calculated slope of the curve. Again, there was no evi-
dence of CO or smoke problems over the range of O^ data shown.
In general, Figure 2-12 tends to indicate that, at least
under the operating conditions of those data and with little or no bound
nitrogen in the fuels, reduction of excess air is not a particularly ef-
fective method of NO reduction nor does it significantly increase NO .
Ji A-
It probably always improves plant efficiency. The general effect of
reduction in excess air is to reduce the A/F ratio in all of the mixing
zones and to decrease the total flue gas flow. Depending on many com-
plex effects of these two simultaneous variations and on the amount of
bound nitrogen in the fuel, it could be expected that NO could increase
or decrease with excess air. Clearly, if a large part of the flow in the
furnace spends considerable time at A/F ratios just above the range for
high thermal NO generation rates, then a reduction in excess air could
reduce the A/F ratio of these zones into this undesirable range, and an
increase in thermally-generated NO emissions could result. In this
3C
respect, reduction in excess air would tend to increase NO emissions.
J\.
Reduction in excess air, however, apparently always decreases NO
arising from the conversion of fuel-bound nitrogen. A reduction of NO
from this source could offset the increase in thermal NO and result in
a net decrease in NO emissions with reduction in excess air. This
3£
explanation is consistent with the slopes of the calculated curves shown
in Figure 2-12. A fuel containing large concentrations of bound nitro-
gen would be expected to exhibit a strong trend toward lower NO emis-
Ji
sions with reduced excess air. For the fuels of this study, containing
little or no bound nitrogen, the net effect of excess air is in all cases
small.
2-57
-------
1UUU
<-«
Q.
2 800
i
0
o
£ 600
LU
0
o
5 400
en
O
CORRECTED
PO
8
i 0
0 0
' ^^ CALCULATED
O
O *
o
0
H5 BOILER, 24 BURNERS OPERATING,
FULL LOAD, NOY PORTS CLOSED
A
0 ACTUAL DATA
1 1 1 1 1
01234
EXCESS AIR EXPRESSED AS PERCENT 0,
(a) With natural gas fuel
Figure 2-12. Effects of excess air
2-58
-------
E 1000
a.
a.
0
§ 800
o
o
nr
OICHIOMETF
o*
S
"» 400
o
o
o
£ 200
QL
O
O
O
0
[- H5/6 BOILER, 24 BURNERS OPERATING,
FULL LOAD, NO PORTS CLOSED
A
0 ACTUAL DATA
CALCULATED
_ ' 0^
1 1 1 1 1
1234
EXCESS AIR EXPRESSED AS PERCENT 0,
(b) With low-sulfur oil fuel
Figure 2-12. Effects of excess air (Continued)
2-59
-------
2.2.4 Burners-Out-of-Service (BOOS)
The final result of the parametric analyses of the
correlation equations is the most important because, at present, it
is only from data from full-scale multiple-burner boilers that insight
into the effects of BOOS can be obtained. No configurations were
tested in the H5/6 boilers with more than eight burners operated as
air-only. With this many (Air) burners, and particularly with NO
li
ports open, the burner A/F ratio with gas fuels becomes so low that
combustion stability can become a problem, and the flame can move
back deep into the burners, causing register overheating. Even
limiting the number of (Air) burners to less than 9 in the total array
of 24 burners, however, more than a million different arrangements of
(F+A) and (Air) burners are possible. Several preliminary observa-
tions were necessary, therefore, to limit the configurations evaluated.
One observation is based on the earlier discussion of
the time and space rates of mixing of the air and fuel, down to the
molecular level, as shown schematically in Figure 2-4. This mixing
can be considered to occur in two major phases: (a) within the flows
from burners at a given level, with some mixing with the flow recir-
culated from the burners immediately below that level, and (b) between
the burner flows and the bulk gases resulting from all of the burner
flows below that level. If the average mixing at a given level is com-
plete on a molecular scale before the flows enter the bulk gases, then
the mixed A/F ratio would have already passed through the region of
high NO generation rates. It would be desirable, then, to mix the flow
from this level with bulk gases which are also at A/F ratios below the
critical region to avoid again passing through this region until as much
cooling as possible has occurred.. This case suggests that all of the
(Air) burners should be located at the top of the burner array or that
only NO ports be used.
2-60
-------
If, however, this molecular-scale mixing is far from
complete before the flows from a given level begin strong mixing with
the bulk gases, then it would be desirable to introduce (Air) burners
at the lower levels so that the bulk gases are at A/F ratios above the
critical range. The average molecular-scale mixed A/F ratio need
never pass through stoichiometric and would only just enter the critical
A/F ratio region after the maximum cooling has occurred. This case
implies one or more (Air) burners at the lowest level in the furnace and
no NO ports.
Even one (Air) burner at the lowest level (and no NO
ports) will cause the bulk gas A/F ratio to start high and monotonically
decrease to the overall boiler A/F ratio. Figure 2-5 shows a schematic
of the variation of the bulk gas A/F ratio with furnace level, with just
one (Air) burner located at various vertical levels. A single (Air)
burner located at some midlevel causes the bulk gas A/F ratio to shift
from below to above the overall furnace A/F ratio at that level. This
latter location for an (Air) burner might be desired if the bulk gas
spreading is such that mixing is essentially complete in the burner
flows (before entering the bulk gases) below that level but is incom-
plete above that level. Thus, one aspect of (F+A) and (Air) burner
configurations that appeared worth evaluating was the vertical level of
the (Air) burners.
A second observation is based on the evaluation of the
correlation equations themselves. In the H5/6 correlations, it was
shown earlier that the first two terms in the correlation equations tend
to dominate the NO variations, particularly with gas fuels. It was
also shown in the parametric evaluations of the correlation equations
that bypassing large fractions of combustion air through NO ports and,
similarly, through (Air) burners is very effective in reducing NO
Ji.
emissions irom uncontrolled levels but cannot reduce these emissions
below still significant levels. Further, in the case of gas fuels, term 1,
2-61
-------
the primary region of (F+A) burners, has a negative coefficient. Thus,
to get very low NO emissions, it seems desirable to maximize the
?c
number of (F+A) burners or minimize the number of (Air) burners and
use no NO ports. Term 1 in the H5/6 correlation for oil fuels appears
a
to indicate an opposite trend would be desirable, but the early mixing
zones do not dominate this correlation as much as in the gas fuel case.
In any case, the above considerations suggest that the effect of the
number of (Air) burners in the array should also be investigated.
2.2.4.1 Effects of the Vertical Location of (Air) Burners
The effects of the vertical location of (Air) burners was
first investigated with four and eight (Air) burners in the array. Other
test conditions were (a) NO ports closed, (b) 3 percent O7, and (c) full
X £t
load. Figure 2-13 shows the values of NO calculated for the cases of
four (Air) burners located at each of the six vertical levels for both gas
and oil fuels. No data is available from tests of these boilers with four
(Air) burners, so these calculations represent an interpolation between
data involving either no (Air) burners or eight (Air) burners. The fig-
ure shows that the effects of the vertical location of four (Air) burners
is opposite for gas and oil fuels. This was expected as a result of the
opposite signs of terms 1 and 2 in the H5/6 gas and oil correlations.
The calculated values of NO for gas fuels with the four
(Air) burners at all levels except the top are low and even negative when
the (Air) burners are in levels 3 through 5. Obviously, negative values
of NO are meaningless and may indicate that the range of interpolation
between zero and eight (Air) burners is too large or that the true NO
values would simply be zero. The general observations which appear
reasonable from the gas fuel calculation, however, are (a) with NO
Ji.
ports closed, four (Air) burners appear to yield lower values of NO
than either zero or eight (Air) burners, and (b) the (Air) burners should
be located at the midlevels in the H5/6 boilers.
2-62
-------
1000
800
600
400
200
20 (F+A) BURNERS OPERATING, 3% 02, FULL LOAD,
N0x PORTS CLOSED
NATURAL GAS - H5 BOILER
LOW-SULFUR OIL - H5/6 BOILERS
-cf
-I.
2345
LEVEL OF (AIR) BURNERS IN BO IIIR
Figure 2-13. Effects of the vertical location of four (Air)
burners (calculation only)
2-63
-------
The oil fuel calculation in Figure 2-13 shows that all of
the NO values except those calculated with the (Air) burners at the top
level are higher than the calculated or measured data with either zero
or eight (Air) burners in any configuration. Even the NO value calcu-
lated with the (Air) burners at the top level is not particularly low.
The general observation from Figure 2-13 for oil fuels, therefore, is
that four (Air) burners at any furnace level yield little, if any, advan-
tage over other techniques to reduce NO .
j£
The calculations for the case of eight (Air) burners indi-
cated that the contributions of all terms upstream of the bulk gases were
reduced essentially to zero. Of the remaining terms, only the NO from
the bulk gas mixing zones is affected by variations in the vertical loca-
tion of the (Air) burners. Figure 2-14 shows the calculated values of
the bulk gas term (term 7) only as a function of the average vertical
level of eight (Air) burners for both gas and oil fuels. As a direct
result of the variation of the bulk gas A/F ratio with vertical location
of (Air) burners, shown schematically in Figure 2-5, the value of the
bulk gas term in Figure 2-14 is minimum when the (Air) burners are
at the bottom or top levels for both gas and oil fuels. Both of these
levels represent cases where the bulk gas A/F ratio does not enter the
critical A/F ratio until maximum cooling has occurred.
Figure 2-15a shows the total calculated values of NO
versus the average level of eight (Air) burners with the NO ports
A.
closed for the gas fuel data. The available data for these test con-
ditions are also plotted for both open and closed NO ports. Accord-
ing to the previous discussion, when (Air) burners are located in the
top levels, the bulk gas A/F ratio is below the critical A/F ratio range
up to the top levels. Thus, open NO ports would only decrease the
X.
bulk gas ratio further below the critical range, with little further effect
on NO emissions. Conversely, when the (Air) burners are located in
X.
the bottom levels, the bulk gas A/F ratio would remain above the
2-64
-------
E
o.
UJ
Of
Q_
UJ
a:
o
120
100
80
60
40
20
0
16 (F+A) BURNERS, 3% 02, FULL LOAD,
N0x PORTS CLOSED
O NATURAL GAS - H5 BOILER
O LOW-SULFUR OIL - H5/6 BOILfRS
O O o
O
I
I
123456
AVERAGE IfVEL OF (AIR) BURNERS IN BOIlfR
Figure 2-14. Effects of the vertical location of eight (Air)
burners on the NOjj represented in the bulk
gas term in the correlation equations
2-65
-------
H5 BOILER, 16 (F+A) BURNERS, 6 BURNER LEVELS,
3% 02, FULL LOAD
1.0
o
0.8
LU
Q_
O
to
a:
Of.
13
CO
o
l__l
O
0.6
0.4
0.2
NO PORTS
n
O CLOSED
OPEN
CALCULATION,
NO PORTS CLOSED
A
I I
I
I
0.2 0.4 0.6 0.8 1.0
AVERAGE LEVEL OF (AIR) BURNERS
NUMBER OF BURNER LEVELS
(a) With natural gas fuel
Figure 2-15. Effects of the vertical location of eight
(Air) burners
Z-66
-------
H5/6 BOILERS, 16 (F+A)BURNERS, 6 BURNER
LfVELS, 3% 02, FULL LOAD
1.0
- 0.8
o
o 0.6
CO
o
i__i
o
0.2
0
NO PORTS
A
O CLOSED
OPEN
-CALCULATION
NO PORTS CLOSED
A
I
I
I
I
0.2 0.4 0.6 0.8 1.0
AVERAGE LEVEL OF (AIR) BURNERS
NUMBER OF BURNER LEVELS
(b) With low-sulfur oil fuel
Figure 2-15. Effects of the vertical location of eight
(Air) burners (Continued)
2-67
-------
critical range. Open NO ports in this case would reduce the bulk
Jt*
gas A/F ratios further, at least at the upper levels, into the critical
range, causing the NO to increase. Figure 2-15a shows that when
the eight (Air) burners are located in the top two rows: (a) NO emis-
?t
»
sions are minimum; (b) the data for NO ports closed agree well with
3t
the calculated values; and (c) the effect of NO ports open or closed
J£
becomes negligible. These data are in agreement with the above
reasoning. Unfortunately, there are no data, at least in this data
sample, in which the eight (Air) burners were located in the two bot-
tom rows. However, the NO port open or closed data should reverse
A.
in this case, with the NO port closed emissions less than or at least
j£
equal to the NO port open data. The calculated values of NO (NO
X X
ports closed) indicate this trend in that they tend to decrease to the
NO port open data or slightly below. Two NO port closed data points,
X X
however, disagree with the calculated curve. The data to confirm or
deny the calculated NO values with the eight (Air) burners located in the
lowest furnace levels are insufficient. It is clear, however, that mini-
mum NO emissions with gas fuels at these operating conditions, using
3t
eight (Air) burners, are obtained with the (Air) burners in the highest
burner levels, whether NO ports are open or closed.
.X.
Figure 2-15b shows the same kind of information as
Figure 2-15a but for oil fuels. Here the agreement between the data
and the calculated trends is poor. The calculated NO values indicate
minimum NO with eight (Air) burners at either the lowest or highest
levels but with only 4Z ppm separating the minimum from the maximum.
The data indicate an apparently strong trend toward a minimum with the
eight (Air) burners located at the lowest level. With the (Air) burners
located at the highest levels, however, the data trend is somewhat
questionable. In the data available for (Air) burners located at average
levels higher than the fourth, three data points show a continuing up-
ward trend in NO , while a single data point, shown in the Figure 2-15b
2-68
-------
at the 0.75 level, indicates a decreasing trend. If this one measurement
is ignored, however, the data trend indicating minimum NO emissions
J\.
with eight (Air) burners located in the lower levels is clear and con-
sistent. The NO data shown have a range of 94 ppm from minimum to
.X.
maximum. In the case of oil fuels, the eight (Air) burners were never
located lower in the furnace than an average level of 3. 5 (equivalent to
four (Air) burners in the third level and four in the fourth level). The
calculated curve indicates that the trend toward decreasing NO with
lower (Air) burner locations should reach a minimum at some level but
there are no data available to indicate the value of that minimum level.
The measured values of NO with the eight (Air) burners at the lowest
ji.
average level tested were 239, 262, and 265 ppm. As discussed
earlier, Ref. 2-2 indicates that at 3 percent O-,, conversion of the
bound nitrogen in the fuel used in these boilers should yield about 222
ppm. This might be expected to represent the minimum NO value
with the eight (Air) burners located low in the furnace, but the actual
minimum level in the H5/6 boilers cannot be accurately predicted.
Certainly, the decreasing trend indicated in Figure 2-15b shows no ten-
dency to level off.
This result with oil fuels is entirely compatible with the
earlier reasoning involving the time and space rate of molecular-scale
mixing of the fuel with the air. It'seems likely that the large oil fuel
droplets generated by the oil guns in these boilers would not have com-
pletely evaporated by the time they enter and mix with the bulk gases.
As a result, the A/F vapor ratio is probably above the critical A/F
ratio range when this mixing begins to occur. The data indicate that
in such cases the minimum NO emissions would be achieved by locat-
x *
ing the (Air) burners in the bottom level of the furnace to keep the bulk
gas A/F ratio also above the critical range. The minimum NO
li
achieved in this manner, however, would be limited to that resulting
from conversion of the bound nitrogen in the fuel at these relatively
2-69
-------
high A/F ratios. Attempting to keep the A/F ratios always low with
existing hardware appears always to increase the thermal NO by more
than the fuel NO is decreased, resulting in a net increase in NO emis-
ji
sions. Ultimate minimum NO emissions with oil fuels appear to re-
3t
quire more radical hardware modifications.
2.2.4.2 Effects of the Number of (Air) Burners
The results of the studies on the effects of the vertical
level of four and eight (Air) burners appear to establish the conditions
for minimum NO with oil fuels within the constraint of existing hardware.
Jt
The minimum condition for gas fuels has not yet been indicated. The
results with four (Air) burners indicated that some number of (Air)
burners between zero and eight, with the (Air) burners located at mid-
levels in the burner array and with no NO port flow, yields a minimum
A
NO level. A further parametric run was made, therefore, to evaluate
the variation in the number of (Air) burners located at midlevels and
with no NO port flow. Figure 2-16 shows these results, along with the
available data. In the calculation, (Air) burners were added, and (F+A)
burners subtracted, one at a time to the fourth level of boiler H5 until
that level was all (Air) burners. In the next step, all of the burners on
level three were designated as (Air) burners, those on level four as
(F+A), and the burners in level five as (Air) burners, one at a time,
until a total of eight (Air) burners were included. The NO values cal-
culated in this manner are represented in the curve in Figure 2-16
labeled "minimum calculated NO."
The curve in the figure shows a sharp drop to negative
values in the calculated NO with the addition of only one (Air) burner
in level four. The minimum NO appears to be reached at between two
and three (Air) burners. NO values as low as -1550 ppm were calcu-
lated. The fact that negative NO values were calculated could indicate
that interpolation between the data points at zero and eight (Air) burners
2-70
-------
H5 BOILER, 24 BURNERS OPERATING,
3% 0?l FULL LOAD, NO PORTS CLOSED
Lf A
10001-
800
600
a.
o"
400
200
O
O
O
MINIMUM CALCULATED NO
O
O
6
HIGH CO
(212 ppm)
_J
4 J 1
14 16 18 20 22
NUMBER OF (F+A) BURNERS OPERATING
24
Figure 2-16. Effects of the number of (Air) burners located
at midlevels with natural gas fuel
2-71
-------
is too great. Considering the large values for terms 1 and 2 in Table
2-7 and the extreme sensitivity of NO formation to temperature at the
high temperature levels of this boiler at full load (570-degrees K or
F combustion air temperature), it is not surprising that parametric
calculations around these full load conditions are sensitive. A series
of calculations similar to the series shown in Figure 2-13 for the gas
fuel with the combustion air temperature reduced by only 30 degrees K
(50 degrees F) completely eliminated all negative values of NO, but
the minimum at midlevels remained. It seems clear that the actual
magnitude of NO which might occur at the minimum in Figure 2-16 can-
not be accurately predicted, but both the available data and the related
reasoning indicate that such a minimum does exist.
As noted, the measured NO emissions in the H5 boiler
X
under full load conditions with all burners (F+A) and the NO ports
Ji
closed are extremely scattered. Figure 2-16 again shows such data.
The data cover a range of 627 ppm, from as low as 292 to as high as
919 ppm, under supposedly identical test conditions in the same boiler.
The data shown in Figure 2-16 for the condition with eight (Air) burners
show that this variation cannot be ascribed to NO measurement error
jC
alone. The data points shown in Figure 2-16 for eight (Air) burners
represent those obtainable within ±1 percent of the 3-percent O2 excess
air. Although these data include wide variations in the arrangement of
the eight (Air) burners in the burner array, the total data spread is still
only 280 ppm. A large part of the measured data variation for the con-
dition with all burners (F+A) is considered real and due to the extreme
sensitivity to temperature variations under these high-temperature
conditions. Figure 2-5 shows that even one (Air) burner located in the
top level of the H5/6 burner array can shift the A/F ratios of the re-
maining burners and the bulk gases very close to stoichiometric. In
such a case, all of the burner flows and the bulk gas flows would re-
main at this very high NO generation rate A/F ratio throughout most
X,
2-72
-------
of the radiant section of the boiler. Conversely, that same single (Air)
burner located in the bottom level of the burner array would again shift
the A/F ratios of the remaining burners near stoichiometric, but would
also shift the bulk gas A/F ratios far above stoichiometric, at least
until all burner flows are mixed into the bulk gases and considerable
cooling has occurred. Inadvertent maldistribution in air flow to sev-
eral burners in the same general vertical locations can accomplish
these same effects. Considering the exponential effect of A/F ratio
(through temperature) on the NO generation rate, it is easy to see
JC
how such maldistribution, under high-temperature conditions of the
H5/6 boilers with all burners ooerating (F+A) and NO ports closed,
H
can cause large variations in measured NO emissions under appar-
Jt
ently identical operating conditions. Under any other off-stoichiometric
or lower-temperature operating conditions, the effects of such maldis-
tribution would be much less marked. The deliberate introduction of a
single (Air) burner to the array, under the conditions where all other
burners are (F+A) and the NO ports are closed, might be expected
A.
to result in dramatic changes in the NO _ emissions.
On the other side of the minimum shown in Figure 2- 16,
only one test condition could be found in which less than eight (but more
than zero) (Air) burners were in operation. This was a test in the
midst of a series evaluating the effects on NO of various arrangements
j£
of eight (Air) burners. Two (Air) burners were apparently restored to
(F+A) for operational test purposes. The average vertical level of the
six (Air) burners was that of the fifth level. Although the Q£ level was
2.53, CO was measured at 212 ppm, an excessively high level, and
further testing of this configuration was terminated. The measured
NO level of 261 ppm was the third lowest measured in any test with
Jt
NO ports closed and under full load. Thus, this one test appears to
X,
lend support to the existence of a minimum NO level at some operat-
3C
ing condition involving a number of (Air) burners between zero and six.
2-73
-------
Also, from the earlier analyses of the correlation
equation and the parametric studies of the effects of cooling rates,
NO port flows, and combustion air temperatures, it appears likely
j£
that the molecular-scale mixing of the gas fuel into the combustion
air in the burner flows does not approach completion until well into
the secondary mixing zone. At about the fourth burner level in the
furnace, the bulk gases have probably spread sufficiently that the
burner flows issuing from the primary mixing zones at that level begin
to mix directly with bulk gases before the molecular-scale mixing of
the gas fuel and air can achieve the overall burner A/F ratio. The
model (Figure 2-2), in fact, assumes that mixing in the secondary
zones does not exist at all in the top level. It is precisely at the fourth
level that this parametric study indicates that it is desirable to shift
the bulk gas A/F ratio from a fuel-rich mixture, resulting from levels
1 through 3 flows, to an air-rich mixture for levels 4 through 6. It
appears, then, that (Air) burners located in level 4 add further air
to unmixed flows from the burners in levels 4 through 6 and avoid the
critical A/F ratio range by keeping the average molecular-scale mixed
A/F ratios above it.
As a result, it appears that a real NO minimum exists
ji
with gas fuels and should be reached by a burner configuration consist-
ing of one to four (Air) burners located in the fourth level of the H5/6
boilers with NO ports closed. The magnitude of this minimum cannot
li
be determined. The one data point available and nearest to this mini-
mum condition exhibited high CO emissions. The reason for this is not
known, nor is it known whether excessive CO will be a problem at the
minimum NO condition. No other limiting problems were apparent.
Ji
2.2.5 Other Boilers
Most of the results discussed thus far were based on the
correlations of gas-fired data from one of the two largest boilers in the
2-74
-------
data sample and oil-fired data from these two largest boilers combined.
The reasons for this were (a) the unique opportunity to develop new under-
standings in this study concerning the effects of BOOS techniques on
NO , (b) the need for a fairly large data sample on the exact hardware
X
and operating conditions analyzed to evaluate the accuracy of the corre-
lation equations in predicting the effects of single independent variables,
and (c) the need for large numbers of test conditions covering wide
ranges of the independent variables to obtain a good correlation equation.
The first reason suggests that large boilers with large numbers of burn-
ers should be of primary interest. The second reason suggests limit-
ing the detailed analyses to a single boiler if possible. The third reason,
however, dictates that only the H5/6 data, both gas and oil firings re-
present samples large enough to obtain good correlations for a single
boiler or boiler type. Table 2-3 shows that no more than 51 test con-
ditions were available on any other single boiler type.
The approach taken to include the other boiler types in
the correlations was to separately correlate the smallest useful data
samples and then combine these until all gas data were included in one
sample and all oil data in another. This resulted in the correlation of
the data samples shown in Tables 2-5 through 2-7. Parametric cal-
culations using correlation equations resulting from data from more
than one boiler type could be made, which might indicate the effects of
some of the independent furnace variables on NO . However, nearly
Ji:
a. dozen independent variables are associated with a given furnace and
only five different furnace types. Thus, unless many of these vari-
ables could be combined or eliminated, the interpretation of such para-
metric calculations could be very difficult. It was decided, therefore,
to limit evaluation of the effects of furnace variables to observations
on the various correlation equations themselves and on such data as
were available, particularly in view of the conditions for minimum
NO determined for the large boilers.
2-75
-------
2.2.5.1 Natural Gas Fuel
For gas fuel, the general operating conditions for
minimum NO in the large boilers involved a small number of (Air)
burners at the midlevel in the furnace with no NO ports. This was
Ji
generally conjectured from the correlation equation from the condi-
tions necessary to maximize terms 1 and 5 and minimize term 2
(Table 2-7). The small number of (Air) burners was necessary to
keep term 1 as large as possible, and (Air) burners located in the
midlevels reduced term 2 as much as possible and left term 5 at its
maximum. In all of the other correlations involving the small boilers
with only three or four vertical levels, the signs of these three terms
are the same but they are less predominant in determining total NO
values compared to the bulk gas zone term 7. From considerations
such as indicated in Figure 2-14, bulk gas zone NO contributions can
be minimized by locating the (Air) burners at either the top or bottom
levels. Also, term 4, associated with burner configurations involving
an (Air) burner directly above an (F+A) burner, becomes relatively
more important in the smaller boilers. This term is negative and,
therefore, to maximize it the (Air) burners should not be located in
the bottom level.
Thus, locating the small number of (Air) burners near
or at the top burner levels would most likely lead to minimum NO in
boilers with small numbers of vertical levels. This seems reason-
able since, as the number of vertical levels of burners approaches
zero, the furnace configuration approaches that of a single burner in
which the distinction between burner flows and bulk gases disappears;
i.e. , all flows are essentially bulk gas flows. With single burners,
the BOOS technique is meaningless, and only staged combustion is
left as a technique to reduce NO .
3£
Figure 2-17a shows the available gas-fuel data, which
justify these conclusions. With only 2 burners out of 12 (1/6)
2-76
-------
operating air-only, the effect on NO is small, but the reduction is
jt
greatest with the (Air) burners located at the top (third) level. Nearly
all of the NO reduction is accomplished with between one sixth and
JC
one fourth of the burners operating air-only and with these (Air) burn-
ers located in the top level. Increasing the number of (Air) burners
to one third of the total affects no further reduction with the (Air) burn-
ers located in the top level, but a tendency toward minimum NO with
ji
the (Air) burners located somewhat below the top level is indicated.
This may simply be an indication that with four (Air) burners the whole
top level of burners are (Air) and the second level is now top level.
The NO minimum with gas fuel in the larger boiler was indicated with
j£
1/24 to 1/6 of the burners air-only at the midlevels. With one third of
the burners air-only, the location for minimum NO in the larger boil-
A.
ers shifted to the top row. Thus, the effect of reduced numbers of
burners or reduced burner levels appears to be the elimination of the
minimum, corresponding to a small number of (Air) burners at mid-
level. Perhaps this is simply geometric, resulting from the fact that
a clearcut midlevel does not exist in a boiler with a small number of
burner levels and some (Air) burners.
In small boilers with small numbers of burner levels,
the minimum NO appears to be achieved with a significant number of
JC
(Air) burners located at the top burner levels. Since all burner flows
and bulk gases below the top level should contribute little to the total
NO emissions in such a case, most must be generated when mixing
.X
of the bulk gases with the air flow from the (Air) burners causes the
average molecular-scale mixed A/F ratio to cross the critical A/F
ratio range. It would appear that if this mixing were delayed until
further cooling of the bulk gases had taken place the NO emissions
jC
could be reduced. This implies, of course, a configuration involving
no (Air) burners and instead NO ports of total admittance equivalent
3t
to the four (Air) burners placed well downstream from the top burner
2-77
-------
H1/2+H3/4+S1/2 TOTAL
BOILERS, FULL FRACTION OF No. OF
LOAD, NO NO (AIR) BURNERS BURNERS
PORTS x
1.0,-
^
o
UJ
£0.6
QL
mO.4
0.2
o
i_j
o
O 1/6
D 1/4
A 1/3
1
I
12
16
12
I
TOTAL
BURNER
LEVELS
3
4
3
I
0
0.5 0.6 0.7 0.8 0.9
AVERAGE LEVEL OF (AIR) BURNERS
NUMBER OF BURNER LEVELS
(a) With natural gas fuel
1.0
Figure 2-17. Effects of the vertical level of (Air)
burners in the smaller boilers
2-78
-------
NO/[NO WITH ALL BURNERS OPERATING (F+A)]
OOP 0 H-
,jO IN9 &» O» OO O
H1/2+ H3/4+ Sl/2 FRACTION TOTAL TOTAL
JOILERS FULL OF (AIR) No- OF BURNER
OAD NO NO BURNERS BURNERS LEVELS
5ORTS o 1/6 12 3
D 1/4 16 4
_ A 1/3 12 3
^" i»__< ^1K
0 A
D
1 1 i 1 1
.5 0.6 0.7 0.8 0.9 1.0
AVERAGE LEVEL OF (AIR) BURNERS
NUMBER OF BURNER LEVELS
(b) With low-sulfur oil fuel
Figure 2-17. Effects of the vertical level of (Air) burners
in the smaller boiler (Continued)
2-79
-------
rows (at higher furnace levels). In the data shown in Figure 2-17a,
only one of the boiler types (H3/4) had NO ports installed. The
total admittance of these NO ports was estimated to be equivalent
to only one (Air) burner. Thus, no data are available to test the
effect of air flow equivalent to the total (Air) burner flows introduced
further downstream.
From these considerations, it appears that the mini-
mum NO conditions with gas fuels in the existing small boiler hard-
ware (without introducing enlarged NO ports) have already been
tested. With one third of the burners (Air), the possibilities of com-
bustion, flame stability and excessive CO problems can become sig-
nificant. Thus, the best condition for minimum NO appears to be
with one sixth to one fourth of the burners (Air) and with these (Air)
burners located in the top row. The two tests shown at roughly this
condition in Figure 2-17a yielded 108 ppm in the four-level boiler
(S2) and 140 ppm in the three-level boiler (H3). Opening the NO
ports under the same operating conditions in the three-level boiler
only reduced the NO to 135 ppm. In both of the cases in the three-
level boiler with one third of the burners (Air), CO emissions were
slightly high (58 to 87 ppm) even though the O? level was 2.78 to 3.65
percent. The single test in the four-level boiler with one fourth of
the burners (Air) (4 of 16) also showed CO emissions of 60 ppm, but
the O2 level was significantly lower (2. 15 percent). Thus, without
major hardware changes, it appears that the NO emission minima
for the three small boiler types in this data sample (Sl/2 and HI
through H4) with gas fuels is in the range of 110 to 140 ppm. This
should be sufficiently low to meet currently established regulations
for the future.
2-80
-------
2.2.5.2 Low-Sulfur Oil Fuel
In the analysis of the oil-fired data from the large
boilers, it was concluded that the minimum NO condition in the
ji
existing hardware involved locating the (Air) burners in the lower
levels of the furnace. This was interpreted as resulting from the
fact that the molecular-scale mixing of the air and fuel vapor flow
from the burners at all levels was not complete before forced mixing
with the bulk gases occurred. Hence, it was desirable to maintain
the bulk gas A/F ratio above the critical A/F ratio range to avoid
passing through the critical range. The minimum NO level, then,
Jt
should be that resulting only from conversion of the bound nitrogen
at these fairly high A/F ratios. Attempting to keep the bulk gas A/F
ratio below the critical range can reduce the bound-nitrogen conver-
sion, but the data indicate (Figure 2-15b) that, at least in that boiler
type, the increase in thermal NO generated in crossing the critical
A/F ratio range more than offsets this reduction, resulting in a net
NO increase.
x
In small boilers, with a small number of vertical
burner levels, the A/F vapor mixing is still slow, but the distinc-
tion between burner flows and bulk gases becomes hazy. Using the
same reasoning as before but treating the flows entirely as bulk
gases, the minimum could occur under conditions maintaining the
bulk gas A/F ratio either above or below the critical range, depend-
ing on the tradeoff between thermal NO generation rates and bound-
nitrogen conversion efficiency. Lower combustion air temperatures
or high cooling rates should reduce the thermal NO generation rates
and should always favor the low bulk gas A/F ratio approach.
In the correlation equations in Table 2-7, a major
difference between the large boiler correlation (H5/6) and those
involving the smaller boilers is apparent. In the large boiler cor-
relation, the dominant terms, as in the gas-fuel case, are again
2-81
-------
terms 1, 2, and 5. The signs of these terms are exactly opposite
to those of the gas-fuel correlations. It is this difference in signs
which leads to the nearly opposite conclusions regarding the best
vertical locations for (Air) burners between gas and oil fuels in this
large boiler type. In the small boiler correlations, Table 2-7 shows
that with oil fuels, when the smaller boilers are correlated alone or
included in significant numbers in correlations with the larger boiler
type, the signs of these three predominant terms reverse and are
the same as in the gas-fired case. As in the gas fuel case, these
terms again become less dominant in the small boilers compared to
the bulk gas term 7. There are no other significant changes in the
oil-fuel correlations for smaller boilers compared to the larger (the
sign reversal in the bound nitrogen and constant terms are considered
meaningless) which might indicate the best vertical location for (Air)
burners. On the basis of the similarity between the gas and oil data
correlations for the small boilers, then, it would appear at first
glance that a small number of (Air) burners located near or in the
top burner row should yield minimum NO with these oil fuels.
2k
Figure 2-17b shows the available NO data from the
7£
smaller boilers (Hl/2 and Sl/2) with oil fuels (full load and with no
NO ports) plotted versus the average vertical level of the (Air)
ji
burners. Contrary to the initial guess, the trend, if any, appears
to be opposite that of the gas-fired case and the same as in the large
boiler type with oil fuels. The minimum NO condition appears to
be attained with a small number of (Air) burners if they are located
low in the furnace. The minimum which could be achieved with such
a configuration should be the 222 ppm calculated earlier, from the
data of Ref. 2-2, for the conversion of the bound nitrogen in the fuel
to NO . The data in which 4 of the 16 burners were (Air) and the (Air)
Ji
burners were located lowest in the furnace indicate an average of
173 ppm NO . The standard deviation of this small amount of data
2-82
-------
would easily include 222 ppm as a possible true value. The data
where one third of the burners were (Air) indicate an average of
208 ppm, also very near the 222 ppm value. Thus, Figure 2-17b
may indicate that a minimum of about 222 ppm, due only to con-
version of the bound nitrogen in the oil fuel used, can be reached
easily in these small boilers with a small number of (Air) burners
located anywhere but at the top burner levels. With (Air) burners
located in the top levels, the data indicate that the increase in ther-
mal NO under these conditions begins to outweigh the reduction in
Ji
bound-nitrogen conversion resulting from combustion at lower A/F
ratios in more of the lower levels. Table 2-2 shows that the cooling
factors for these smaller boilers are 1.3 to 2.2 times that of the
larger boiler type, indicating that thermal NO should be less of a
problem in these boilers.
Thus, it appears that with existing hardware, a mini-
mum of about 220 ppm NO can and has been achieved in the smaller
boilers, even with a small number of (Air) burners if they are located
in the middle to bottom levels of the boiler. The tradeoff between
thermal NO and that generated from the bound nitrogen in both the
X.
small and large boilers appears to be such that, with existing hard-
ware, the minimum NO achievable is the limit set by the conver-
x J
sion of the bound nitrogen at about 3 percent O^.
The frequent reference to the bound-nitrogen conver-
sion efficiency implies that an estimate might be made of the effect
of A/F ratio in the burners on this conversion efficiency. Such an
estimate can be made from observations of the correlating equations.
In the H5/6 correlation equation with oil fuel, as noted previously,
the limits of NO for very low combustion air temperatures and for
very high values of NO port admittance imply that (a) bound-nitrogen
X.
conversion with 3 percent O~ in burner contributes about 210 ppm
NO and (b) the NO port mixing zone contributes about 216 ppm
X X
2-83
-------
thermal NO with large fractions of combustion air flowing through
3t
NO ports. Figures 2-8 and 2-6 show close approaches to these
Ji
values at the limits of those calculations. With no air flow through
the NO ports (Figure 2-6), the total calculated NO (427 ppm) should
JL
consist of 210 ppm from the bound nitrogen and 217 ppm thermal NO.
Thus, the thermal NO at the limits of the calculation shown in Figure
2-6 in both cases is about 217 ppm. In between those limits it seems
reasonable that the thermal NO could be larger than 217 ppm but un-
likely that it would be smaller than this value. Thus, a conservative
upper limit on the NO from the bound nitrogen can be estimated by
assuming that the thermal NO over the entire range of Figure 2-6 is
constant at 217 ppm. It is also reasonable to assume that the fuel
NO thus calculated could never be larger than the 210 ppm at the
burner A/F ratio corresponding to 3 percent O^-
The results of this estimate are shown in Figure 2-18.
The fuel NO drops rapidly with burner equivalence ratios (A/F) less
than stoichiometric and begins to approach zero at equivalence ratios
near 0.7. This equivalence ratio corresponds to an A/F of 9.7,
which is approximately the rich limit for flame propagation with pre-
mixed natural gas fuels. If the fuel-rich products of liquid-oil com-
bustion exhibit a similar fuel-rich flammable limit, then it is rea-
sonable to assume that both the HC and the NO reactions might have
difficulty propagating throughout the air-vapor fuel mixtures. This
reasoning implies, therefore, that a reasonable estimate of fuel-NO
in fuel-rich mixtures might be a linear reduction from stoichiometric
to zero at that burner equivalence ratio corresponding to the fuel-
rich flammable limit of the air-vapor fuel mixtures.
2.3 FURTHER EXTRAPOLATIONS AND CONCLUSIONS
Throughout the analyses of the effects of various com-
bustion modifications on NO from the available data sample, it
2-84
-------
H5 BOILER, 24 BURNERS OPERATING,
FULL LOAD, 3% Q OVERALL
CM
o
£ 200
E
§ 160
120
o
a:
LL.
80
40
x
<
0.7 0.8 0.9 1.0
BURNER EQUIVAL£NCE RATIO, A/F
1.1
Figure 2-18. Maximum estimate of NOX from fuel nitrogen
versus burner equivalence ratio
2-85
-------
appeared that all of the observed effects could be understood in terms
of the probable time history of the average molecular-scale mixed
A/F ratio. Particularly in the case of oil fuels, where the air-vapor
fuel is the significant mixture, 90 to 95 percent of the total weight
flow of reactants is represented by the combustion air. Buried in
this large fluid flow is O?, which is the ingredient necessary for the
HC reactions as well as for the formation of NO . Large fractions
J\:
of these reactions cannot occur until the 5 to 10 percent weight flow
of fuel is mixed into this gross air flow on a molecular scale (i. e. ,
fuel molecules near each and every O^ molecule). Such mixing can-
not occur immediately even with initially gaseous fuel. Some time
and space is required for gross recirculation and eddies to dissipate
into the small-scale isotropic turbulence approaching the molecular
scale. This turbulence mechanism, or behavior, is discussed in
Ref. 2-3. Consequently, in all cases, 90 to 95 percent of the total
reactant weight flow enter the furnace from the burners at an infinite
A/F ratio (pure air), and 5 to 10 percent enter at zero air-fuel ratio.
While there are always interfaces between these two fluids where the
molecular-scale A/F ratio is near stoichiometric, the fraction of the
total weight flow involved in these interfaces is small until the scale
of the turbulent mixing begins to approach zero. It appears, then,
that the problem in controlling NO emissions can be thought of as
2£
one of controlling the fraction of the total weight flow of reactants
involved in molecular-scale A/F ratios near stoichiometric and the
time spent at these conditions.
Gross overall A/F ratios near stoichiometric can be
tolerated for appreciable lengths of time if more air or fuel is mixed
in before the major fraction of the flow begins to approach molecular-
scale mixing. This case is approximated when the flow from a given
burner enters (at right angles to) the relatively high-velocity fine-
scale turbulent flow in the bulk gases in a large multiple-burner boiler.
2-86
-------
Similarly, large fractions of the flow may be at small-scale A/F
ratios near stoichiometric as long as continuing finer-scale mixing
rapidly shifts the A/F ratio out of this region. This latter case is
approximated by the relatively rapid mixing of gas fuels into the air
flow from a given burner to arrive at a final, molecular-scale mixed
A/F ratio well below stoichiometric. For short periods of time,
large fractions of the air and gas fuel flow from a burner may be at
small-scale air-fuel ratios near stoichiometric, but the continuing
finer-scale mixing immediately reduces this ratio to well below
stoichiometric. In the case of oil or coal fuels, this mixing through
stoichiometric to lower A/F ratios is difficult to accomplish rapidly
because the fuel vapor, or generated gases, is slowly bled into the
surrounding air as the liquid fuel vaporizes or the coal gasifies, at
least when the rates of these state-change processes are slow com-
pared to the rate of turbulence decay in the gas stream.
2. 3. 1 Further NOX Reduction Techniques
On the basis of these considerations, it is most desir-
able, from the standpoint of minimum thermal NO , to maintain the
3t
burner A/F ratios well above stoichiometric. Then, no matter how
slow the turbulence decay or the rate of introduction of fuel vapor or
gases, the average molecular-scale mixed A/F ratio need never pass
through stoichiometric. In other words, the fraction of the total re-
actants near stoichiometric A/F ratios in the interfaces between gross
unmixed regions of air and fuel will continually decrease with time as
the turbulence decays and the mixing approaches molecular scale.
Since the overall boiler A/F ratio is always maintained somewhat
above stoichiometric (10 to 20 percent excess air) to assure complete
oxidation of CO and other carbon and HC species, stoichiometric A/F
ratios need not be crossed even at the boiler exit.
2-87
-------
Such an approach appears to provide the best opportunity
to reduce thermal NO to a minimum near zero. For fuels containing
2£
little or no bound nitrogen, this appears to be the simplest technique
to apply to existing boilers for ultimate minimum NO emissions.
Ji
Seemingly, the most direct hardware approach would be to incorporate
fuel NO ports. All of the existing burners could be made air-rich by
Ji
reducing the fuel flow to the burners, with the excess fuel added well
downstream in the fuel NO ports. The same goal might also be ac-
2v
complished by biasing the top row of burners to fuel-rich, leaving all
of the lower burners air-rich. A gross approach to this goal, which
has been tested in some boilers, is to simply shut the fuel off in some
of the low-level burners in the boiler. The risk in this approach is
that some or all of the flows from the resulting fuel-rich burners at
higher levels may achieve average molecular-scale mixing to A/F
ratios below stoichiometric before beginning mixing with the excess
air from the (Air) burners below. This A/F ratio would then have to
first decrease through stoichiometric in the burner flows and then
increase through it again as the excess air is mixed in. Apparently,
this is exactly what happens in current boilers firing natural gas fuels.
This approach to ultimate NO reduction is not partic-
Ji
ularly applicable with fuels containing bound nitrogen. It is well known
that the efficiency of conversion of bound nitrogen is a function of the
molecular-scale A/F ratio. If that A/F ratio is deliberately main-
tained above stoichiometric to minimize thermal NO , the conversion
of bound nitrogen to NO would be very efficient. In the case of a fuel
Ji
containing significant concentrations of bound nitrogen, then, the ulti-
mate combustion modification technique to minimize NO emissions
X.
appears necessarily to involve maintaining a molecular-scale A/F
ratio well below stoichiometric, at least until the nitrogen in the fuel
has been converted to free molecular nitrogen. No other method of
preventing the conversion of bound nitrogen to NO is known at this
2-88
-------
time. Some preliminary parametric studies of a correlation equation
for the oil fuels used in this study (0. 24 weight percent bound nitrogen)
indicate that with sufficiently low burner A/F ratios resulting from
large air flows diverted through NO ports, the NO generated from
A. 3t
the bound nitrogen could be reduced to less than a'bout 10 ppm.
A major problem with the fuel-rich burner technique
for minimizing NO from fuel-bound nitrogen is to avoid the formation
Ji
of thermal NO . As discussed, when burners are operated fuel-rich
the molecular-scale mixed air-vapor fuel ratio of a large fraction of
the reactants must decrease through stoichiometric to the low value
during the decay of the large-scale induced turbulence (i.e. , the frac-
tion of the total reactants near stoichiometric in the interfaces between
unmixed air and fuel must first increase to a maximum and then de-
crease to zero as the mixing approaches uniformity on a molecular
scale). Unfortunately, those fuels containing high levels of bound
nitrogen, such as oil and coal, are normally in the liquid or solid
state. Otherwise, the nitrogen could be easily removed before burn-
ing. Thus the rate of mixing of the fuel vapor, or fuel gases, with the
combustion air is usually controlled by the rate of vaporization or
gasification of the fuel and not by the decay of the gross turbulence in
the air. The mixing of the last fuel to be vaporized or gasified could,
in fact, be reduced further because the turbulence in the surrounding
air has by then decayed to fine scale.
One simple way to accelerate the initial air-fuel vapor
mixing, of course, might be to atomize oil fuels to very fine droplet
sizes. At some small droplet size, the vaporization rate would cease
to be limiting on the air-fuel mixing rate, and oil combustion would
start to behave like that with gas fuels. For further reduction of ther-
mal NO , it might also be desirable to promote finer-scale more in-
tense turbulence in the burner air flow. Oil fuel handling systems in
current boilers provide oil pressures up to 1000 psi in the oil guns.
2-89
-------
This energy should be adequate to finely atomize the oil with little or
no increase in energy requirements on the plant. Pulverizing coal to
very fine particle sizes may be more difficult and expensive in terms
of energy demands on the plant; this area will be investigated in a later
study. In the final mixing region, the molecular-scale mixed A/F ratio
of the total boiler flow must once again be driven (rapidly) through
stoichiometric when the final excess air is added (through NO ports,
for example), but care should be taken to assure adequate cooling of
the flue gases before this mixing is attempted. Many other methods
of reduction of thermal NO , such as FGR through the burners, re-
J±
duced combustion air temperatures, or water sprays in the combus-
tion air, could be employed to assure minimum thermal NO formation
Ji
in the discussed configuration, but these appear to require greater
structural modifications than the approach described or could involve
significant plant efficiency losses. Such techniques might be required
in burning coal fuels, however. Existing hardware in current gas-- and
oil-fired boilers appear to be such that neither of the combustion modi-
fication techniques discussed provide minimum NO emissions. Mini-
Ji
mum NO , in fact, is apparently achieved in the boilers studied here
by almost exactly opposite techniques.
2. 3. 2 NOX Minima with Existing Hardware
In the case of gas fuels, the existing NO ports are not
Ji.
readily adaptable to conversion to fuel NO ports. The secondary
Ji
approach of limiting the air flow to the top row of burners to bias this
row to fuel-rich and all the others to air-rich could be accomplished
by reducing the flow area in the air registers in the top row. This
reduction in air flow, however, could lead to combustion and flame
stability problems in those burners, without a simultaneous and some-
what empirical redesign of the gas spuds. In any case, no test data
are available under conditions even close to these two approaches to
2-90
-------
minimum NO . The air-rich primary, fuel-rich secondary type of
JC
reversed two-stage combustion with gas fuels, perhaps, should be
evaluated first in single-burner laboratory furnaces for other undesir-
able side effects.
The data in this study, with gas fuels, indicate that
complete, molecular-scale mixing of the air and gas is achieved
rapidly in the burner flows but still requires significant time and
distance in the furnace. In the large boiler type with six burner
levels, it appears that this burner flow mixing is complete before
mixing with the bulk gases begins for the lower three burner levels
but is not complete in burner levels five and six. It would appear
desirable to maintain the bulk gas A/F ratio well below stoichio-
metric for the first three burner levels but well above stoichiometric
in levels five and six. This can be accomplished by locating a few
(Air) burners in the fourth burner level with no NO port flow. Since
Ji
such a configuration has not been specifically tested in these boilers,
the parametric calculations vising the correlation equation represent
a rather large interpolation between existing data. The calculation,
however, indicates the existence of a minimum in NO emissions with
3£
this burner configuration which could be very low, even approaching
zero NO . The small amount of appropriate data available appears to
X
confirm the existence of a NO minimum at these conditions but cannot
suggest the value of that minimum. Although the reason is not apparent,
some data indicate a potential high CO emissions problem in this con-
figuration. There should be no problem of combustion or flame stability
in this configuration.
In gas-firing in the smaller boilers with smaller numbers
of burner levels, the separation of burner flow mixing from bulk gas
mixing becomes somewhat academic, since the flow configuration begins
to approach that of a single burner (i.e. , the burner flows and bulk gas
2-91
-------
flows become in series or are one and the same). In this case,
minimum NO appears to result with the simple two-stage combus-
JC
tion technique, and the BOOS technique becomes meaningless. The
data from these smaller boilers indicate that with existing hardware
minimum NO with gas fuels is achieved either by locating a rela-
Ji
tively larger number (more than in the larger boiler type) of (Air)
burners in the highest levels or by using few or no (Air) burners and
large NO port air flow. These conditions have already been tested
li
in the three smaller boiler types and yield NO minima in the range
Ji
of 110 to 140 ppm. Some indications point to possible CO and com-
bustion and flame stability problems if large fractions of air flow are
diverted through (Air) burners or NO ports, but such large flows do
ji
not appear necessary to reach the NO minima. The data correlations
ji
tend to indicate that (Air) burners located in the bottom levels even in
these smaller boilers cannot provide the reversed two-stage conditions
discussed and that NO emissions in such a case would be higher than
ji
the 110 to 140 ppm range. No data is available to confirm this latter
observation.
With gas fuels, short-term minima involving no signifi-
cant hardware modifications required combustion modifications which
are nearly exactly opposite to those which can achieve further reduc-
tions over the longer term. These short-term minima, however,
represent large reductions in NO from uncontrolled levels and could
satisfy even the stringent Los Angeles Air Pollution Control District
(APCD) Rule 68 regulations for the immediate future. It is unlikely
that the longer-term modifications, or even the possibly dramatic
but simple tests in the larger type boiler, will be accomplished soon
because of the phasing out of natural gas fuels in utility boilers. Such
verification may become increasingly important in the longer term,
however, if low-Btu gas from coal becomes a significant fuel in utility
boilers.
2-92
-------
In the case of oil fuels, the data appear to indicate that
the vaporization of the liquid droplets in the furnace is sufficiently
slow that the molecular-scale mixed air-vapor fuel ratio does not,
at any burner level in any boiler, begin to approach levels below
stoichiometric before mixing with the bulk gases (the flows from
lower burner levels) begins to occur. Minimum thermal NO must
X,
then be attained by maintaining the bulk gas A/F ratio above stoichio-
metric. This, of course, leads to high efficiency of conversion of
the bound nitrogen in the fuel to NO . Two NO minima are possible:
Xi X,
(a) In the case where the thermal NO is minimized but the bound nitro-
x
gen conversion is high, and (b) in the case where the bound nitrogen
conversion is minimized but the thermal NO is high. With existing
Xi
hardware, the possibility of minimizing both sources of NO simul-
X.
taneously appears to be unlikely. Thus, the only concern is which of
these two minima is the lower. In the boilers studied here, the lower
minima occur for the case where the thermal NO is minimized. The
.X.
minimum NO achieved with oil fuels in the existing hardware studied
here is specified by that value resulting from conversion of the bound
nitrogen in the fuel at about 3-percent O2- If 100 percent of the nitro-
gen in the fuels used in this study (0. 24 percent by weight) were con-
verted to NO at 3-percent O?, a concentration of 353 ppm would result.
X. fc»
The available literature indicates that for a fuel with 0. 24-percent
nitrogen, the conversion efficiency should be about 63 percent at 3-
percent O,. Thus, the minimum NO level in the hardware of this
£ X.
study should be about 220 ppm. All of the oil-fired data in the sample
of this study indicated minimum NO values very near to 220 ppm.
X.
These minima were achieved with a small number of
(Air) burners located low in the furnace and with no NO port flow.
No significant problems of CO emissions or combustion or flame
stability were apparent at these minima, and no sacrifice of plant
efficiency was apparent. As with the gas fuels, these minima still
2-93
-------
represent significant reductions from uncontrolled levels and appear
to satisfy the stringent Los Angeles APCD, Rule 68, NO emission
Ji
requirements scheduled for the near future. Where testing at the
conditions for this minimum have not already been accomplished, it
appears that such testing will be conducted in the near future.
To generalize these conclusions to apply to other exist-
ing boilers with oil fuels containing different concentrations of bound
nitrogen is somewhat risky. The question of the appropriate firing
configuration to achieve the lower of the two NO minima depends on
the relative values of thermal and fuel nitrogen NO under these firing
conditions. In plants firing high nitrogen oil fuels in boilers with large
cooling rates, low combustion air temperatures or FGR, the fuel nitro-
gen NO may outweigh the thermal NO . In such cases, the opposite
Ji ji
configuration from that indicated by the data of this study may yield
the lower minimum (i.e. , (Air) burners high in the furnace or NO
ports). The smaller boilers in this sample with cooling rates 1.3 to
2.2 times higher than that of the larger boiler type indicate a trend in
this direction.
The amount of fuel nitrogen NO generated is not a
JC
linear function of the weight percent of nitrogen in the oil at the same
excess air. The literature indicates that the conversion efficiency of
this nitrogen to NO decreases with increasing concentrations of nitro-
j£
gen in the fuel. Thus, increasing the nitrogen in the fuels of this study
from 0. 24 weight percent, for example, by a factor of two (to 0.48
percent) only increases the NO resulting from this conversion from
Ji
222 ppm to 297 ppm, a factor of only 1.34. This level, however, may
not meet emission requirements, and the minimum NO conditions
JL
indicated by APCD, for the oil fuels of this study, may not be accept-
able. Thus, when oil fuels contain significantly more than the 0.24-
weight-percent nitrogen the other minimum condition may have to be
accepted, and more drastic measures may be needed to reduce the
thermal NO .
2-94
-------
REFERENCES
2-1. W. Bartok, et al. , System Study of Nitrogen Oxide Control
Methods for Stationary Sources, Final Report, GR-2-NOS-69,
VII, NAPCA Contract No. PH-22-68-55, ESSO R&E Co. (20
November 1969).
2-2. D. W. Turner and C. W. Siegmund, "Staged Combustion and
Flue Gas Recycle: Potential for Minimizing NOX from Fuel Oil
Combustion," paper presented at American Flame Research
Committee Fiame Days, Chicago (6-7 September 1972).
2-3. H. B. Palmer and J. M. Beer, Combustion Technology, Some
Modern Developments, Academic Press, New York (1974).
2-95
-------
-------
SECTION 3
COMBUSTION AND FLAME INSTABILITIES
One of the undesirable side effects of certain combustion
modifications for the purpose of reducing NO emissions from utility
X.
boilers is the occurrence of various types of flame and combustion in-
stabilities. These instabilities appear to be among those problems
which limit the application of certain combustion modification tech-
niques for NO reduction. If a general type of solution could be ob-
tained for these problems, it may be possible that these combustion
modification techniques might be extended to the point where future
regulations could be met. The literature (Ref. 2-3) indicates that a
wide variety of mechanisms could possibly explain the numerous types
of flame and combustion instabilities observed in boilers, including
both the feedback-coupled and the uncoupled varieties. If this is true,
solutions to stability problems may be unique to each boiler.
In the current study, the effort in the area of flame and
combustion instabilities was limited to a study of instability data in the
sample obtained for the main (NO reduction) analysis. The purpose
3t
was to determine if a mechanism or mechanisms for the observed in-
stabilities could be discovered and verified by the data. If this could
be accomplished and the resulting mechanism was sufficiently general
to be widely applicable, then actual analysis and/or modeling of that
mechanism would be conducted at a later date.
3-1
-------
This limitation, of course, implies that no new
mechanisms could be generated and that the data would have to be
examined in the light of known mechanisms. Few well developed
analyses of instability mechanisms in the utility boiler literature
were found, as compared with an abundance of data and analyses in
the air-breathing engine and rocket engine fields, particularly, in the
latter. Initially, then, the data were examined with respect to appro-
priate mechanisms developed for rocket engine applications. It is con-
cluded here that the observed combustion instability data do verify a
common rocket engine mechanism and that the flame instability data
correspond to a nonlinear part of that mechanism.
3.1 FEED SYSTEM COUPLED COMBUSTION
INSTABILITY
The combustion instability data obtained from the boilers
studied occurred exclusively when natural gas fuel was in use. There-
fore, the most destructive and widely studied mechanism of combustion
instability in liquid rocket engines, the liquid-droplet combustion cou-
pled mechanism, would not apply. Occurring more frequently but less
destructively in rocket engines, including those engines where one or
both of the propellants are injected in the gaseous state, is the feed
system coupled mode of instability. This basic mechanism exists and
provides a source for potential combustion instability in continuous
combustion devices of all types. Instabilities of this general type have
been involved in combustion instability problems in the development
stages of nearly all liquid rocket engines and have occurred frequently
in air-breathing turbine engines and even in solid rocket motors, al-
though the feed system must be defined rather loosely in the latter
case.
All continuous combustion systems require continuous
feed of fresh reactants into a flame zone, as well as continuous dis-
charge of exhaust products. All that is required to develop a classical
3-2
-------
feed system coupled instability is (a) a finite time delay from the time
that reactants enter the combustor until a significant portion of those
reactants are reacted (combustion time delay) and (b) a feedback cou-
pling of the dynamic flow and combustion in which the gain (the ampli-
tude of one cycle compared to the previous cycle) around the whole
closed loop is greater than one at the frequency of the loop. The gain
of the loop is frequency-dependent. The frequency of the loop is domi-
nated by the combustion time delay. Both the variation of the gain with
frequency and the combustion time delay are usually poorly known and,
in any case, are specific to the combustor. For these reasons, exact
solutions to feed system coupled instability problems in one combustor
cannot normally be generalized to other combustors. One combustor
might be stabilized by changing the combustion time delay and the loop
frequency so that the loop gain at the new frequency is less than one
(phase stabilization). Since the gain in another combustor may be a
different function of frequency from the first, such a loop frequency
change may not have the same effect on stability in the second com-
bustor. In the latter case, the system might have to be stabilized by
increasing the oscillatory energy dissipation in the components in-
volved in the loop, thereby decreasing the loop gain to less than one
at the same loop frequency (gain stabilization).
3.1.1 Gain Stabilization
Although the general mechanism of feed system coupled
instability is well known and has been analyzed extensively in rocket
engine applications (Ref. 3-1), exact solutions for stability are usually
difficult and specific to hardware. An asymptotic case can be estab-
lished, however, which can lead to general design criteria for feed
system stability. The hardware configuration for this asymptotic case
in a liquid rocket engine is one in which the reactants are fed into the
combustor from a large reservoir, through a number of short ports of
negligible inertia, into a choked combustor of dimensions that are
3-3
-------
small compared to the wavelength of the oscillation. The latter
criterion is satisfied at very low frequencies in any combustor. Such
a system is shown, modeled in block diagram form, in Figure 3-1.
The general linearized equation for the loop gain of such a system
(Ref. 3-2) shows a number of frequency-independent and frequency-
dependent terms. As the loop frequency approaches zero, however,
all of the frequency-dependent terms approach one, and the loop gain
approaches a maximum. Thus, if this maximum gain for this most
unstable feed system coupled loop can be made to be less than one,
then the loop gain at any other frequency or in any other combustor
configuration must also be less than one. Such systems would, there-
fore, be stable in feed system coupled modes of instability at all fre-
quencies and with all other combustor configurations at the same oper-
ating conditions. Satisfying this criterion is called gain stabilization.
Since the gain inherent in the combustion itself and in
the combustor pressure feedback path are difficult or undesirable to
change, the gain stabilization criterion is applied to the pressure drop
which causes the reactant flow into the combustor. This pressure
drop is a resistance term in the analysis and tends to dissipate oscil-
latory energy and to reduce the loop gain. Higher pressure drops are
stabilizing. For a rocket engine injecting both reactants in the gase-
ous state, the gain stabilization criteria for the two reactants are
Fuel:
Gain stabilization criteria such as shown in Eqs. (3-1)
and (3-2) have been derived and recommended for liquid rocket engines
of various types since 1951 {Ref. 3-3) and were recently modified for
3-4
-------
m
UJ
i
INFINITE MANIFOLD COMPRESSIBILITY
NEGLIGIBLE INJECTOR INERTIA
COMBUSTION CHAMBER DIMENSIONS« INSTABILITY WAVELENGTH
AP
2Rw
w
INJECTOR
(1 - K)e
-TS
K
~
Vp
COMBUSTION
CHAMBER
(COMBUSTION
TIME DELAY
Pc RESPONSE
DUE TO OTHER
REACTANT
Figure 3-1.
Block diagram of a feed system coupled mode
of instability in a rocket with gaseous reactants
-------
the case of rocket engines with gaseous reactants (Ref. 3-2). These
criteria are recommended in Ref. 3-4 for designing stable combustion
into rocket engines. Reference 3-4 also indicates that, since the gain
stabilization criteria may represent a "worst-worst" case, the pres-
sure drops indicated by Eqs. (3-1) and (3-2) might be overly conserva-
tive. In liquid rocket engine history, feed system stability problems
seem to have been absent if the feed system pressure drops were
greater than about one half of those values. It can generally be ex-
pected, however, that if pressure drops are allowed to decrease to
much less than half of the theoretical gain stabilization values, then
feed system coupled instabilities might be expected. Thus, these cri-
teria provide both an initial safe-design starting point and can be used
as an indicator of when an observed instability is likely to be of the
feed system coupled variety. This latter application was used in this
study to evaluate the potential for feed system coupled instability in
the utility boilers studied.
3.1.2 Gain Stabilization Applied to Utility Boilers
Because of the limited scope of investigation of com-
bustion stability in this study, no attempt was made to derive gain
stabilization criteria specifically for utility boilers. The only differ-
ence would be in the pressure response of the combustor. The exhaust
flow in a rocket is determined by choked flow conditions in the nozzle,
while the boiler flue gas flow is not choked. It is expected that the
rocket case would yield higher response than that of the boiler, so
direct application of Eqs. (3-1) and (3-2) to boilers would at least be
conservative.
Most boilers operate at mean pressures in the boiler
(P, ) near 1 atm. The mean, overall weight A/F ratio (F) is normally
D
about 16. Because of the dominance of the nitrogen in the air and in
the flue gases, the molecular weights of the combustion products are
nearly that of air. The molecular weight of the natural gas fuel used
3-6
-------
in the boilers studied was 18.4. The density ratios in Eqs. (3-1) and
(3-2) are given by
The temperature of the combustion products for a range
of A/F ratios around the overall boiler ratio is about 2300 degrees K
(3700 degrees F). For the data analyzed in this study, the combustion
air temperatures were about 560 degrees K (550 degrees F) and that of
the fuel was estimated to be near ambient (300 degrees K). Therefore,
the approximate density ratios for the boilers studied are
p
Air: -& = 0.244 (3-5)
pa
p
Fuel: -^s 0.206 (3-6)
pf
With these data inserted into Eqs. (3-1) and (3-2), the
pressure drops across the burners and gas spuds required to establish
theoretical gain stabilized feed system coupled modes are
Air: AP a 248 gm/cm2 (98 in. of water) (3-7)
cL
Fuel: APf s 17. 8 gm/cm (7 in. of water) (3-8)
In large utility boiler common design practice, pressure
drops across the gas fuel spuds or rings are normally quite large, in
3-7
-------
order to promote mixing of the small weight flow of fuel with the large
weight flow of air. Thus gas-fuel pressure drops tend to be of the
2
order of 1000 to 1800 gm/cm (400 to 700 in. of water or 15 to 25 psi),
well above even the most conservative estimate of that necessary for
gain stabilization of fuel feed system coupled instability. In oil-fired
units, the theoretical gain stabilization criterion is higher, about 30
gm/cm (12 in. of water), but the fuel pressure drop necessary to
atomize and distribute the heavy liquid fuel is very large, approach-
4 2
ing 7 X 10 gm/cm (1000 psi). Thus, fuel feed system coupled com-
bustion instability should not be a problem in gas- or oil-fired utility
boilers.
Such is not the case, however, with air feed system
coupled instability. Because such large quantities of air must be
pumped through the boiler (16 times as much as the fuel), the pump-
ing energy requirements become significant in the overall plant effi-
ciency. As a result, pressure losses in the air feed system tend to
be minimized. Typical air pressure drops across the burners in the
data studied are less than 25 gm/cm (10 in. of water) at full load.
This is less than 10 percent of the theoretical gain stabilization re-
quirement indicated in Eq. (3-1). Even reducing the gain stabilization
requirement to eliminate all of the conservatism discussed above, the
actual air pressure drops appear to be well down into the range where
air feed system coupled instabilities could be expected. In fact, the
question appears to be one of why there are not more combustion in-
stability problems with all boilers firing air, oil, or coal.
Probably, air feed system coupled instabilities were
encountered in the initial development of the burners for these boilers,
which were perhaps empirically stabilized by the manufacturers. The
resulting system, however, should be classified as phase stabilized.
As discussed, a phase stabilized system is stable only under certain
operating conditions and can become unstable as a result of perhaps
3-8
-------
minor changes in these conditions, even though the hardware is the
same. As long as the boiler is operated as designed, no stability prob-
lems should occur.
In recent years, however, existing boilers have been
subjected to a wide range of user onsite modifications for the purpose
of reducing NO emissions. NO ports have been added which remove
X X
air flow from the burners to implement the two-stage combustion NO
reduction technique. Fuel flow has been turned off in some burners
and diverted to the remaining burners to implement the NO reduction
JL
technique variously called "off-stoichiometric," "biased-firing," or
"burners-out-of-service (BOOS).11 Combustion air temperatures have
been lowered, water has been sprayed into the combustion air,and
other diluents have been added to it. Many of these changes affect the
combustion time delay and thereby the loop frequency and phase stabi-
lization, as well as the loop gain at the same frequency. Since more
detailed analysis is required to evaluate the effects of such combustion
modifications on the combustion time delay and the frequency-dependent
loop gain function, this study examined the effects on the frequency-
independent loop gain term represented by the resistance of the burners
to air flow through the burners. This was done by considering only the
air feed system response to boiler pressure perturbations. Coupling
the combustion and combustor geometry response to complete the loop
was left to more detailed studies. If air feed system coupling is re-
sponsible for the observed instabilities, then the instabilities should
occur predominantly under operating conditions where the air feed
system response is high.
3.2 CORRELATION OF OBSERVED INSTABILITIES
WITH AIR FEED SYSTEM DYNAMIC RESPONSE
The air feed system response in its simplest linearized
form is represented by the first diagram block of Figure 3-1. The
3-9
-------
term 2Rw is the linearized form of the flow resistance in the square
law flow relation:
2 AP ,-, Qv
w = -£- (3-9)
This flow relation is often expressed as
(3-10)
where ADM is the admittance to flow. The admittance form of the
flow equation is usually used when flows are in parallel, as in a paral-
lel multiple -burner array fed from a single windbox, because the total
admittance of the array is the simple sum of the individual burner
admittances. Similarly, the resistance form is used for the same
reason when components are in series. Clearly, flow resistance is
related to flow admittance by
R= - j- (3-11)
(ADM)
Thus, the linearized air feed system response can be represented in
terms of admittances by
1 (ADM)2
Flow response = - =- = - - =r-*- (3-
V ZRw Zw
This feed system response describes the perturbations
in flow (in this case, air flow) from each individual burner. Assuming
that the windbox volume is sufficiently large that small perturbations in
flow through one burner do not affect local windbox pressures suffi-
ciently to in turn affect adjacent burner flow (a reasonable assumption),
then the response of each burner is independent of the others. Since
3-10
-------
the burner flow response concerns only the feed system and not the
combustion or combustor response, it is independent of the type of
combustor downstream of the feed system (e.g., rocket thrust cham-
ber, turbojet combustor, or boiler).
The first observation from Eq. (3-12) is that if the
mean air flow w through a burner is reduced, the flow response is
increased, and the likelihood of instability is increased. This is
what occurs when NO ports are introduced into an existing boiler
with no change in burner resistance. The NO ports introduce an
Ji
additional air flow path, and less air is divided between the burners.
One observation is that if the observed instabilities are of the air
feed system coupled variety, then they should occur most frequently
in older boilers that have been retrofitted with NO ports and more
X
frequently when the NO ports are open rather than closed.
Ji
3.2.1 Air Flow Through Individual Burners
If the fuel that is being injected into the air stream in
operating (F+A) burners or the initial flame anchored in or near the
burner has no effect on the admittance to air flow through the burner,
then the admittance of (F+A) burners would be the same as (Air)
burners. The total air flow, therefore, would be equally divided
between all burners whether they were (F+A) or (Air), and the BOOS
technique for NO reduction would not have any effect on air feed sys-
Ji
tern coupled stability. This assumption of equal air flow distribution
between burners, not only between (F+A) and (Air) burners but despite
windbox aerodynamic variations and nonuniform burner configurations,
appears to be common in the industry and in boiler literature. Partic-
ularly with gas fuels and in the BOOS configuration, this assumption
appears to be seriously in error.
Unfortunately, burner pressure drop instrumentation
2
in most cases is inadequate. Small differences (13 to 25 gm/cm )
(5 to 10 in. of water) between large numbers (over 1000 gm/cm )
3-11
-------
(400 in. of water) must be measured. The measurement is usually
made with single-point, standard, water manometers. At low boiler
loads, the measured air pressure drop across a burner array may
actually reach negative values although flow is still clearly in the
positive direction. This measurement is not significant for any other
operational or research purpose and, therefore, has not received
great attention.
No measurements have been made of the air flow through
(F+A) and (Air) burners in full-scale, operating, multiple-burner
boilers. Not only are the combustion air temperatures high, greater
than 530 degrees K (500 degrees F), but the flow into and through the
burners is extremely complex. Some swirl is introduced in the air
registers. At full load and with full open registers, however, the
flames entering the boiler normally show little evidence of a positive
swirl. Often the air register vanes are not at uniform angles to the
circumference of the air register. In some cases, in the earlier data
of this study, some of the register vanes may not have been open at
all (tangent to the circumference), while others in the same register
might have been hanging radially. These register problems have
largely been corrected, but the data sample obtained may be clouded
by such anomalies.
Flow measurements with a pitot-static rake or hot wire
anemometers appear to be useless because of the flow complexity.
Measurements might be made in an (F+A) burner by rake sampling
the gases at the burner exit and measuring the fuel flow to the burner.
This would be a complex and expensive task and has not been done to
date.
With a large amount of data, reasonably accurate values
of average windbox and furnace pressures and, therefore, average
pressure drops across the entire burner array served by the windbox
can be obtained. Reasonably accurate values of total air flow can also
3-12
-------
be established. From Eq. (3-10), average values of total burner
array admittance (ADMT), plus NOx ports where applicable, can be
obtained for all configurations of (F+A) and (Air) burners and NOx
ports tested. It seems quite reasonable, considering the basic data
accuracy and the averages taken, to assume (a) equal flow through all
(F+A) burners, (b) equal flow through all (Air) burners, and (c) that
the admittance to air flow through (Air) burners (ADMA) and the NOx
ports (ADMNP) are constant (i.e., independent of flow through the
burners or NO ports). It does not appear reasonable to assume that
the admittance to air flow through an (F+A) burner (ADMFA) is inde-
pendent of burner operating conditions. Thus, for each operating con-
figuration (for example, 16 burners (F+A), 8 burners (Air),and 2
NO ports open), an admittance equation can be written:
ADMT = 16 (ADMFA) + 9 (ADMA) -f 2 (ADMNP) (3-13)
where ADMT is averaged over a number of tests with these values of
NFA, NA, and NNP, but not necessarily with the same burners (F+A)
or (Air). If the ADMFA were independent of flow conditions in the
(F+A) burner, only three such equations would be necessary to deter-
mine all of the admittances. Since this cannot be assumed, at least,
with gas fuels, ADMFA is different for each test condition, and there
is always one more unknown than available equations. Thus, an addi-
tional equation is needed.
There are two possible effects of the introduction of fuel
in a burner on the air flow through the burner: (a) one is the axial mo-
mentum interchange between the air flow and the fuel flow in cases
where there is appreciable mixing between the two fluids within the
burner; (b) the other is the effect of heating of the A/F mixture by
3-13
-------
partially complete combustion within the burner. In the case of gas
fuels, at least some of the gas is injected perpendicular to the air flow.
A simple analysis indicated, however, even if all of the gas fuel were
injected in this manner, with no axial velocity, the momentum exchange
would have negligible effect on the admittance to air flow through the
burner. In the case of oil fuels, the liquid spray is injected primarily
in the direction of air flow and could act to aspirate the airflow, thereby
increasing the admittance. The oil guns, however, are close to the
exit of the burner. In this case, also, a similar analysis indicated a
negligible effect on burner air flow admittance. The effect of a flame
in the burner, however, can have a significant effect on the air flow
admittance.
With gas fuels, it is necessary to provide a pilot flame
in the burner to continuously ignite the fresh mixture and to control
the point of flame initiation. Some of this pilot flame is provided in
the low axial-velocity region in the center of the burner, due to the
vortex-like flow resulting from the air swirl. In some cases, this
pilot flame can be seen deep in the burner, even overheating the air
registers. Some of the pilot flame could be provided in the flow re-
circulation zones just downstream of bluff bodies, such as the gas
spuds or rings, or the oil gun and/or diffuser. It is clear that partial
combustion of the air-gas fuel mixtures must and does occur within
the burner.
With oil fuels, however, no pilot flame need be pro-
vided in the burner. The flame is instead stabilized around the liquid
fuel droplets themselves, and the whole cloud of droplets can be con-
tinuously ignited by gross recirculation within the furnace. This kind
of gross recirculation within the furnace could also provide air-gas
fuel ignition except for the great danger of sudden flashback to the
burner. Such flashback cannot occur with oil flames because the
liquid fuel is not vaporized and mixed with the air close to the burner.
3-14
-------
As mentioned, in operation the oil gun tip is near the exit of the
burner and injects axially into the furnace. Thus, except for early
fine atomization on the periphery of the spray and the possibility of
some small reverse flow in the center of the burner due to the swirl,
flames are unlikely to be back inside the burners with oil fuels, nor
were they observed there.
Clearly, with almost any kind of pilot flame, the frac-
tion of combustion completed near the flame holder, within the burner
in this case, is some inverse function of the main stream velocity. If
the velocity is increased sufficiently the flame can be blown completely
out of the burner or liftoff of the burner. At a sufficiently slow velo-
city, all of the combustion could be completed within the burner. There-
fore, at the highest velocities, the ADMFA for gas-firing burners should
begin to approach the ADMA (no flame in the burner). Since flame lift-
off represents a dangerous situation because of the possibility of flash-
back, however, such conditions are usually not allowed to exist. As a
result, calculated values of ADMFA from gas-fired tests can approach
but never be exactly equal to ADMA. At the higher burner velocities
with oil fuels, however, the ADMFA should be very close to the proper
values of ADMA. These observations provide the additional equation
needed, with which all of the admittances for all burner and NO port
x r
configurations can be estimated. Analysis of the effect of heating of
the A/F mixture within the burner on the air flow admittance (discussed
later in this section) indicates that the presence of the flame in the
burner will reduce this admittance. Therefore, the ADMA for all of the
burners for each of the furnaces in this study were estimated according
to the following procedure: (a) it was assumed that at high total air flow,
using oil fuels, the measured ADMFA is equal to the proper value of
the ADMA; (b) configurations were selected, where possible, where all
burners were operating (F+A), a number of burners were completely
shut off (no air or fuel), and the total air flow was large; and (c) tests
were selected which yielded the highest calculated values of ADMFA.
3-15
-------
The need to select only certain test conditions from
which to calculate ADMA eliminated the advantage of averaging large
amounts of data. Thus, the experimental errors in the small sample
of data are more significant. No way to improve this method of esti-
mating ADMA was found, and the derived values of ADMA remain esti-
mates only. No pressure drop data were available for the smallest
(Riley) furnace studied, and this furnace was subsequently eliminated
from both the stability and the NO analyses. In the case of boilers
Si/2, the desired oil-firing data were not available, and the value of
ADMA was estimated by proportioning the ADMA for Hl/2, with simi-
lar burners, according to burner flow area. Despite the obvious dif-
ficulties in estimating values of ADMFA and ADMA, the resulting air
flow distribution is considered more correct than the common, but
clearly erroneous, assumption of a uniform air flow distribution (equal
values of ADMFA and ADMA) with gas fuels.
Figure 3-2 shows an example of ADMFA data calculated
from cases where all burners were operating (F+A) but different num-
bers of burners were shut off completely (recognizing that there may
be considerable but unknown air leakage through a fully closed regis-
ter). In this case, only six test cases were available from gas-firings
and eight from oil firings. Even to include this much data required
that half of the cases used include open NO ports. It was necessary,
JC
therefore, to estimate an average value for ADMNP and correct both
the ADMT and the air flow velocity through the burners for the NO
Jt
port open cases. Since there were no (Air) burners in any of this data,
ADMFA was simply the measured or corrected value of ADMT divided
by the total number of operating burners.
The H3 boiler is of the face-fired horizontally opposed
configuration with a partial divider wall and a total of 1Z burners in
3 levels. Rated load is 240 MW. For the gas-fired data, all 12burners
were operating. For the oil-fired data, the number of operating
3-16
-------
o
at
in
CM
H3 BOIIf R DATA
O NATURAL GAS
D LOW-SULFUR OIL
NO(AIR) BURNERS
«4_
in
1U
a:
UJ
as
co 9
5
i
0
=> 8
0
x
»
^
§ 7
0£
O
t
LU f.
o o
1 1 1 1 1
10 20 30 40 50m/sec
1 1 1 1 1 1 1 1
20 40 60 80 100 120 140
AIR VELOCITY IN (F+A) BURNER
160 ft /sec
Figure 3-2. Effect of air velocity in an (F+A) burner on the
admittance to air flow through the burner
3-17
-------
burners varied from 12 to 6, as shown in Figure 3-2. The gas-fired
data show a well-defined linear increase in ADMFA with air flow
velocity through the burners. This is interpreted as a direct indica-
tion of the effect of the air velocity on the degree of combustion oc-
curring in the burner. The absence of gas-fired data at velocities
higher than about 37 m/sec (120 ft/sec) may be an indication that the
flame will lift off the burner at these high velocities. The ADMFA
would then rapidly rise to the ADMA value.
The oil-fired data, with the same number of burners
operating as for the gas-fired data (12), show ADMFA values higher
than those for the gas-fired case at the same burner velocities. The
oil-fired ADMFA values, however, increase rapidly as the number
of operating burners decrease. This may be due to the presence of a
significant flame within the burners, even with oil fuels, or to sig-
nificant air leakage through fully closed air registers. Just why an
oil flame within a burner should be blown out of the burner more with
fewer number of burners operating at the same flow velocity is not
clear. It seems likely that an oil flame would not be well-anchored
in the burner, because there is no need to anchor it there. As a re-
sult the initial flame position may "wander" significantly under the in-
fluence of minor flow configuration variations. An admittance value for
a fully closed air register of as much as one-third that of a fully open
register is necessary to explain the full variation of the ADMFA data
for oil firings on the basis of air leakage alone. Introducing a fourth
unknown admittance value to be determined from the data (for fully
closed air registers) would further complicate an already nearly im-
possible task. Therefore, the commonly used assumption of zero ad-
mittance for a fully closed air register was maintained throughout this
study. This effectively lumps any closed-register air leakage into that
of the fully open (Air) burners (into ADMA). Since the ADMA should be
higher than ADMFA with any possible flame present in the burner, ADMA
3-18
-------
was estimated slightly higher than the highest ADMFA calculated
from the oil-fired data.
That ADMFA for the gas-fired cases is nearly a factor
of two lower in admittance than the estimated value of ADMA is shown
in Figure 3-2. This has significant implications to both stability and
NO control. A remaining significant question concerns whether this
value of ADMA could be even larger than estimated. Since the ADMFA
values calculated from the oil data are increasing as the number of
operating burners decrease and no limit is indicated by the data, it
might be possible that ADMFA for less than six burners operating
could be even greater than the estimated value for ADMA. The ques-
tion concerns the value of ADMFA which represents the complete ab-
sence of oil flame in the burner (assuming no air leakage through
closed registers). To answer this question, an analysis of the frac-
tion of combustion (C, } which must be completed within the burner to
yield the calculated values of ADMFA for both gas and oil fuels was
undertaken. This analysis is discussed in Appendix £.
The results of the C, calculation for the data from gas-
and oil-fired tests on boiler H3 are shown in Figure 3-3. The calcu-
lated values of C, fall between the limits of zero (flame liftoff) and one
{all combustion completed in the burners). A similar calculation was
made to evaluate the effect of possible large air leakage through closed
air registers. Again, the calculated values of C, were between zero
and one, although the maximum value was only about 0.5 for this case.
Significant error in the estimated values of ADMA could easily yield
values of C, greater than one or less than zero. The highest values
of ADMFA for the oil-fired data in Figure 3-2 indicate almost no com-
bustion in the burner (C, =0), justifying the estimate of ADMA made
earlier on the assumption that ADMFA for these oil tests were very
nearly equal to ADMA. These conclusions, however, depend rather
strongly on the single test which yielded a calculated value of C, near
3-19
-------
1.0
0.8
BOILER DATA
o- _
2 0.6
OO
I
N)
O
o~ 04
u_ Z
O
Z I
g*
a 0.2
Of
2.205
FUEL FLOW RATE
IN BURNER. Ib/sec
O
D
0
A
O
OIL
2.1 - 2.7
GAS
3.4-3.6
2.7-2.95
1.9 -2.1
1.6-1.75
0.8-1.05
10
20
30
40
m/sec
i
1
i
1
20 40 60 80 100
AIR VELOCITY IN (F+A) BURNER
120
140 ft/sec
Figure 3-3. Effect of air velocity on the fraction of com-
bustion completed within an (F+A) burner
-------
1.0. The next largest calculated value of C, is less than 0.5. It is
unlikely that more data could be obtained which would yield high values
of C, because that is an undesirable operating condition, which would
probably result in overheating of the air registers and perhaps damage
to the burners. Lacking such further data, the results had to be
accepted as they are, with reservations.
The data from gas firings could be empirically fit
rather well with a function:
= 2.205 (3-14)
where velocity v& is given in ft/ sec. The range of fuel flows at any
given air flow velocity indicates that the fuel flow rate has little effect
on C, and, therefore, that the air flow velocity is the major control-
ling parameter, as conjectured earlier. The magnitude, negative
slope, and shape of the C, versus flow velocity curve in Figure 3-3 have
several significant implications with regard to both steady-state A/F
ratios in (F+A) burners and the dynamic response of these burners to
furnace pressure oscillations. They stem from the following
observations:
a. The presence of partial combustion within a burner
has a strong effect on the resistance to air flow through
the burner. Thus, the steady-state A/F ratio in (F+A)
burners can be significantly lower than would be cal-
culated from an assumption of equal flow in (F+A) and
(Air) burners .
b. The increased resistance to air flow in (F+A) burners
tends to be a stabilizing influence on the air feed sys-
tem response, but both the reduced air flow and the
negative slope of C^ with velocity are destabilizing
influences. When a small rise in furnace pressure
causes a small reduction in air flow in an (F+A) burner,
the air flow velocity through the burner will be re-
duced, the flame will retreat further into the burner,
resistance will be increased, flow will be decreased
3-21
-------
further, until a stable point is reached. This
nonlinear response will be stronger in configurations
where the steady-state air flow through (F+A) burners
is reduced and the slope of Ch versus flow velocity is
more negative. The use of NOx ports and BOOS tech-
niques for NOX control both create this reduced air
flow.
c. Below some reduced air flow rate through (F+A)
burners, and below some critical air flow velocity,
the air flow can become what is called statically un-
stable. That is, the reduction in air flow due to the
nonlinear effect can become self-accelerating, stop-
ping only when the flame is fully back in the burner
and C, is equal to one.
The results of calculations of the steady-state A/F
ratio in (F+A) burners are shown in Figure 3-4. Test data were ob-
tained with as many as four (Air) burners in the array (NFA = 8).
Under these conditions, the assumption of uniform air flow distribu-
tion to all burners indicates that the A/F ratio in (F+A) burners
should be in the range of 10 to 11. Burner A/F ratios calculated
from admittance data indicate that these ratios were actually in the
range of 6 to 8. Note that the premixed flammable limit for a typical
natural gas-air flame is about 9.7. Because the flame is actually a
diffusion flame, local A/F ratios above this limit will exist in the
burner, and a pilot flame can be sustained, but further flame propa-
gation could be quenched as further mixing reduces the A/F ratio
below the limit. Engineers' notes on the character of gas flames
under such conditions tend to indicate lazy, sparking, diffuse, and
opaque flames that appear ragged (detaching from the burners in
chunks and quenching in the furnace) and not attached to ring burners
around the full periphery. These indications are descriptive of lower
than expected air flow rates and velocities, lower than expected A/F
ratios, and A/F ratios below the premixed flammable limits. These
indications are considered to support the conclusions of this study
with regard to the effect of flames within the burners on air flow
through the burners.
3-22
-------
18
16
14
o
i
ec.
5 12
ex.
^5
CO
10
H3 BOILER, 12 BURNERS OPERATING,
3% 02, FULL LOAD
ASSUMED UNIFORM AIR.'
DISTRIBUTION
FUEL-RICH
FLAMMABLE LIMIT
ACTUAL AIR
DISTRIBUTION
NO PORTS
ClOSED
OPEN
7 8 9 10 11 12
NUMBER OF (F+A) BURNERS OPERATING
Figure 3-4. Effect of partial combustion in a burner on
the A/F ratio in the burner with gas fuel
3-23
-------
Again a check was made on the possible effects of large
air leakage through closed air registers, in this case on the actual air
distribution shown in Figure 3-4. The slope of the actual air distribu-
tion is somewhat shallower in this case, but still significantly steeper
than that resulting from a uniform distribution assumption. The uni-
form distribution case, of course, would also include the assumption
of zero leakage through closed air registers.
An attempt was made to include, also, the dynamic
effect of the flame within the burner on the response of the air feed
system to furnace pressure perturbations. This was approximated by
including the slope of C, with air flow velocity at each flow rate in the
linearized resistance term, Eq. (3-12). Figure 3-5 shows the air feed
system response calculated in this manner, again using H3/4 data, as
a function of the fraction of the total number of operating burners
[(F-fA) plus (Air) burners] which are (F+A). The calculated response
with all burners operating either as (F+A) or (Air), begins to rise
rapidly when less than about two thirds (less than eight) of the burners
are (F+A). In this case, Figure 3-4 shows that the steady-state A/F
ratio in the (F+A) burners would be well below the premixed flammable
limit. Attempts to reduce the fraction of operating burners further,
which might have resulted in unstable operation, were frustrated by
large increases in CO emissions, even with the NO ports closed. As
A.
a result, some of the burners were shut off completely. This increased
the fraction of (F+A) burners and not only increased the A/F ratio in
these burners and cleared up the CO problem but also increased the air
flow velocity through the burners and strongly reduced the air feed
system dynamic response.
3.2.2 Available Combustion Vibration Data
In general, a review of the available data from the
standpoints of CO emissions and combustion instability problems
verifies that (a) increases in CO emissions are generally related
3-24
-------
1.2
1.1
1.0
O-
q
o
23
ID
0.9
0.8
0.7
0.6
BECOMES VERY
LARGE AT 0.44
MUCH CO
ACTIVITY
O NO PORTS CLOSED
A
N0x PORTS OPEN
NUMBER OF BURNERS
OPERATING
O 12
Q 10
8
H3 BOILER, 3%
I
I
0.4 0.5 0.6 0.7 0.8 0.9
FRACTION OF BURNERS OPERATING (F+A)
1.0
Figure 3-5. Response of air flow through an (F+A) burner to
furnace pressure perturbations with gas fuel
3-25
-------
to actual A/F ratios in (F+A) burners near or below the premixed
flammable limit and (b) the CO emissions and combustion stability
problems occur when about 75 percent or less of the operating burners
are (F+A) and 25 percent are (Air) burners. Both of these conclusions
are verified by data from a series of 12 tests conducted on boiler SI
for the purpose of evaluating combustion vibrations. These data are
summarized in Figure 3-6. Although only four data points are shown,
they represent repeated attempts to come back to, reproduce, and go
beyond these configurations as well as all of the transition configura-
tions between each condition shown. High vibration occurred with both
configurations where 75 percent of the burners were (F+A) even though
the total number of operating burners in the one case was half of that
in the other case. Under those conditions creating large vibrations,
the observed flames included all of the characteristics discussed,
tending to indicate lower air flows in the (F+A) burners and A/F ratios
near or below the premixed flammable limit. The high vibrations
which occurred with only half of the burners operating occurred at
part load where the air feed system response would again be high even
though the full flow was diverted through only half of the burners.
Figure 3-6 also indicates that the observed vibrations
are not likely to be due to typical induced draft fan instabilities. Fan
instabilities tend to occur in operating regions where air volume flow
is low and the fan discharge pressure is high. For any given number
of burners operating (F+A), this unstable fan region should correspond
to smaller numbers of burners operating (Air). That the observed
trend toward vibration is opposite to this is shown in Figure 3-6.
One other observed case of high vibrations occurred in
the largest boiler in the sample, H5. These tests involved the first
firings with all new gas spuds installed. These spuds were designed
to promote rapid mixing of the gas with the air within the burner.
From the analyses and discussion to this point, it could be expected
3-26
-------
?5
o
or 4
LLJ H
O.
O
I/)
ct:
UJ
I 3
CO
u. 2
o
o:
LU
CO
1
TOTAL OF 12 TESTS
si BOILER VIBRATION^
O LIGHT
LARGE
8
10
12
14
NUMBER OF (F+A) BURNERS OPERATING (nfa>
0.75
FLAIV\ES BREAKING
UP-HIGH CO
16
Figure 3-6. Si boiler vibration test data with gas fuel
3-27
-------
that the improved mixing would hold more of the flame within the
burner, at any given air flow velocity, increase the (F+A) burner re-
sistance to air flow, and reduce the (F+A) burner A/F ratio for any
given burner configuration. All of this should have increased the
probability of CO and combustion instability problems. As it turned
out, the system was so severely unstable that these spuds and the
design had to be discarded.
All of the instability occurred under part load condi-
tions. The boiler could not be brought up to full load because of the
violence of the instabilities. Neither could the operating conditions
be held long enough to take data on windbox-furnace pressure drops
or flue gas samples. All of the data was obtained from the environ-
mental engineer's notes. To make the data useful at all, a subjective
rating scale was used, rating the magnitude of the instabilities on a
scale from 1 to 10 from comments in these notes. Clearly, such data
are not accurate but do reveal some trends which are of value.
These rough average ratings are shown in Figure 3-7
on a plot similar to Figure 3-6. The same general trend appears as
in Figure 3-6, in that the magnitude of the vibration increases mark-
edly when 75 percent or less of the burners are (F+A). The point
rated 10.0 was so severe that the steaming rate began to rise mark-
edly and was almost out of control.
A microphone external to the boiler was used to tape
record the vibration to allow frequency analysis. The most violent
vibrations appeared to be in the ranges of 10 to 12. 5 and 40 to 50 Hz.
With the speed of sound of 900 m/sec (2950 fps), the lower frequency
was found to correspond to a standing acoustic resonance (12.5 Hz)
between the bottom and top of the furnace, a vertical distance of about
36 m (118 ft). This vertical height is unimpeded by any water tubes
transverse to the flow. No apparent source of eddy-shedding exists
in this volume. The higher frequency corresponds most closely to
3-28
-------
12
c
o
O.
O
(/I
CO
£
o
on
UJ
CO
GAS SPUD 3, LOADS GREATER THAN 100 MW
H5 BOILER VIBRATION:
= AVERAGE
1.0 = LIGHT (minimum)
10.0 = SEVERE (maximum)
12
14
16
18
20
22
NUMBER OF (F+A) BURNERS OPERATING (nfa>
Figure 3-7. H5 boiler vibration data with gas fuel
24
3-29
-------
the 9. 14-m (30-ft) horizontal resonance (49.2 Hz) between the burner
walls (firing faces in this opposed-fired boiler). The horizontal reso-
nance is exactly what would be expected from coupling with burner
feed system oscillations since all burners would be at a pressure anti-
node in that resonance. The vertical resonance is somewhat less
likely since the burners are spread vertically, approximately from
the pressure antinode at the bottom of the furnace nearly to the pres-
sure node at the middle of the furnace. Some data involving this mode
indicate that vertical location of the (F+A) burners had an effect on the
instability.
Figure 3-7 also shows that at the extreme case of 50
percent of the burners (F+A) and 50 percent (Air) (NFA = NA = 12)
that the vibrations appear to be minimum. It can only be conjectured,
without further data or analysis, that this condition represents the
case of flames fully back in the burners, where the nonlinear effect of
flame position on air flow resistance is no longer a factor in the
feed system response. This is the condition represented in Figure
3-3 by the constant value of C, at 1.0. The more rapid mixing pro-
moted by this gas spud makes such a case more probable.
An intriguing possibility of damping some of these
modes with a Helmholtz resonator, analogous to the corner absorbers
and acoustic liners of liquid rocket engines and afterburners of turbo-
jet engines, is presented by the ash pit of this boiler. This pit, of
the same horizontal dimensions as the furnace (9. 14 X 10. 1 m) (30 X
33 ft) and 0.914 m (3 ft) deep, is separated from the main furnace
volume by a neck running the full width of the furnace (10. 1 m) (33 ft)
and approximately 0.76 m (2.5 ft) in the front-to-back direction. The
FGR flow is introduced into the furnace through this pit and neck.
This general configuration is that of a classical Helmholtz resonator.
The current dimensions yield a resonant frequency of about 27 Hz.
By appropriate modification of the pit volume and neck geometry and
3-30
-------
by the addition of two partitions, this volume could be converted to
two resonators near the side walls tuned to the 40 Hz of the furnace
width resonance and a single central resonator tuned to the 12.5 Hz
of the vertical resonance. The neck is in the wrong location (at a
pressure node) for effective damping of the horizontal furnace depth
resonance. Although there does not appear to be any need for the gas
spud design associated with the high vibrations observed in this fur-
nace, such an application of damping devices to stabilize a large boiler
has never been attempted. The inadvertent presence of this resonant
damping device could explain the absence of vibrations in the frequency
range around 27 Hz.
3.3 COMBUSTION STABILITY CONCLUSIONS
It appears that the available data on combustion insta-
bility and flame stability problems can all be explained by air feed
system problems. Both appear to result largely from the nonlinear
effect of the presence of the flame within the burners on the resis-
tance to air flow through the (F+A) burners. In general, this resis-
tance appears to be strongly controlled by the heat addition of partial
combustion within the burner. As long as all combustion air flow
paths are equal [no NO ports, no (Air) burners], the air flow dis-
Ji:
tribution to all (F+A) burners will be the same. Any changes in air
flow resistance will simply result in changes in the overall windbox-
furnace pressure drop required to maintain the same flow. The re-
sistances to air flow through NO ports 'and (Air) burners, however,
are less than that of the (F+A) burners and appear to be independent
of air flow rates. In cases where NO ports are introduced (two-
3t
stage combustion) or fuel is shut off to some of the burners (BOOS)
for purposes of reducing NO , not only is the proportion of the com-
Ji.
bustion air flowing through the remaining (F+A) burners substantially
decreased, but th'e resistance to air flow through (F+A) burners be-
comes a nonlinear function of the flow itself.
3-31
-------
A reduction in air flow through an (F+A) burner allows
more of the combustion to be completed within the burner. This in-
creases the resistance to air flow through these burners and diverts
a larger proportion of the air flow to the NO ports and the (Air)
X
burners. This diversion of air flow away from the (F+A) burners
allows even more complete combustion within the burners, further
increasing resistance, diverting more flow to the nonburning paths
and compounding the effect. When a certain fraction of the air flow is
being diverted to nonburning paths, this nonlinear effect causes prob-
lems in both the combustion and flame stability areas.
The air feed system response is high with any burner
configuration because of the low windbox-to-furnace pressure drop.
In fact, this low pressure drop allows the flame in the burner to domi-
nate the air flow resistance. As more air flow is diverted away from
the (F+A) burners, the air feed system response becomes larger and,
at some critical fraction, air feed system coupled modes can become
unstable. The data appear to indicate, at least with the hardware
studied, that this critical fraction occurs when less than 75 percent
of the operating burners are (F+A).
This same nonlinear effect causes the A/F ratio in the
(F+A) burners to decrease to much lower values than would be cal-
culated under the assumption of uniform distribution of air flow be-
tween all burners, both (F+A) and (Air). As the air flow through and
the A/F ratio in the (F+A) burners decreases, the flame advances
further into the burner, and register overheating problems can result.
Again, when about 75 percent or less of the burners are (F+A), the
A/F ratio appears to decrease below the rich premixed flammable
limit for natural gas fuels. Ragged flames, difficulty in flame-
holding, and pulsating flames can occur. On the other extreme, gas
spuds designed empirically for less rapid gas-air mixing to promote
stable operation under two-stage combustion or BOOS conditions
3-32
-------
may not adequately anchor gas flames at high flows, and flame
liftoff problems can occur.
It is not as important to anchor oil flames in the burners
as it is gas flames. As a result, little burning is apparent within the
burners with oil fuels, and none of the nonlinear effects related to gas
flames are apparent. Combustion instability, register overheating,
and flame liftoff problems are not apparent in oil-fired data. Future
oil-firing configurations designed for minimum NO , however, may
require much finer oil atomization and may introduce all of the gas-
fired combustion and flame stability problems discussed herein.
Thus, the general mechanism of combustion and flame
stability problems observed in the data appears to have been identified
and verified. Although the analytical techniques are available, this
mechanism was not modeled or examined further. Complete analysis,
including design recommendations for avoiding these problems, has
been deferred to later studies.
3-33
-------
NOMENCLATURE
ADMA
ADMFA
ADMNP
ADMT
c.
c*
K
M
NA
NFA
NNP
P
AP
s
T
V
admittances to air flow through an air-only (Air)
burner, a fuel plus air (F+A) burner, a NOx port.
and the total of all of these, respectively, m-kg*' 2/
sec (ft-lb1/2/sec)
= area of the choked throat of a rocket motor nozzle,
m2 (ft2)
= fraction of fuel reacted within the burner
= characteristic velocity of the combustion products in
the throat of a rocket motor nozzle, m/sec (ft/sec)
= ratio of densities of the combustion products to the
reactant being injected
= molecular weight
= number of air-only (Air) burners
= number of fuel-and-air (F+A) burners
= number of NO ports open (flowing combustion air)
Ji
2 2
= pressure, kg/m (abs) (Ib/ft )
= pressure drop across an injector orifice or a burner,
kg/m2 (Ib/ft2)
= flow resistance, sec /kg-m (sec /lb-ft ), defined by
Eq. (3-9)
= universal gas constant
= the Laplace operator
= temperature, degrees K (degrees F)
= volume of tiie rocket motor combustor (to the choked
throat), m3 (ft3)
= velocity, m/sec (ft/sec)
3-34
-------
w = weight flow rate, kg/sec (Ib/sec)
24 24
p = mass density, kg-sec /m (Ib-sec /ft )
T = combustion time delay, sec
Subscripts
a = air
b = burner
c = combustor
d = delayed flow (in time)
f = fuel
m = manifold (or windbox)
p = reaction products
3-35
-------
REFERENCES
3-1. D. T. Harrje, and F. H. Reardon, Liquid Propellant Rocket
Combustion Instability, NASA SP-194 (1972).
3-2. O. W. Dykema, "Feed System Coupled Instability in Gas/Gas
Combustors," paper presented at the llth JANNAF Combus -
tion Meeting, Pasadena, California (9-13 September 1974).
3-3. M. Summer field, "A Theory of Unstable Combustion in Liquid
Propellant Rocket Systems," J. Am. Rocket Soc.
(September 1951).
3-4. Design and Development Procedures for Combustion Stability
in Liquid Rocket Engines, CPIA Publication 256
(September 1974).
3-36
-------
-------
APPENDIX A
DEVELOPMENT OF THE CORRELATING EQUATION
A. 1 INTRODUCTION
The approach taken in this study in an attempt to
elucidate the effects of hardware and operating condition variations on
the NO emissions from large multiple-burner utility boilers was
Ji
(a) set up a crude model of the NO generation process in these boilers,
(b) develop an equation from that model consisting of the sum of a
number of linear terms which if all processes were accurately de-
scribed would predict the NO emissions, (c) use that equation to
Ji,
correlate a large amount of NO emissions data from such boilers,
and finally (4) gain understanding of some of the poorly known processes
and the overall effects of these modifications on NO emissions by
x '
exercising and evaluating the resulting correlation equations. Major
assumptions involved in the development of the correlation equation
are discussed in the text of this report, while some of the details of
this development are described in this Appendix.
Development of the correlation equation was initiated
by defining the simple mixing zones shown in Section 2,Figures 2-1 and
2-2. Further definitions of relative burner combinations and of the
assumed recirculation flow are shown in Figure 2-3. A brief descrip-
tion of the mixing zones follows.
A-l
-------
A. 2 PRIMARY MIXING ZONE
The primary mixing zone is the core flow zone where the
flow from an individual burner issues into the furnace. It is taken to be
two burner diameters in length. The A/F ratio is that of the burner
alone. Stay time in the zone is calculated from the constant zone
length divided by the burner flow velocity. Cooling time is taken as one
half of the stay time.
A. 3 RE CIRCULATION MIXING ZONE
The recirculation zone represents a complex flow and
mixing region. Clearly, entrainment by the flow from a burner causes
some kind of flow toward the burner but where it comes from, where it
goes after some entrainment, and what its prior mixing and cooling
history has been in all cases is not clear. Since it represents a flow
path in parallel to the main flow of gases from a burner through the
furnace, it complicates an attempt to sum the series contributions of
each mixing zone through which the flow passes. It seems clear that
there should be a general upward flow of these gases along the wall.
For purposes of this approximate model, then) the following simplifying
assumptions were made:
a. The recirculation flow travels in a circle of mean
diameter equal to the length of the primary zone
(Section 2, Figures 2-1 and 2-2) at the velocity of the
primary zone. As a result, the stay time in the zone
is IT times that of the primary zone.
b. For the sake of simplicity, the total amount of flow
recirculating is taken as 32 percent (1/ir) of the primary
zone flow.
c. Since the amount of NO generated in a given zone de-
pends on the product oithe amount of flow involved and
the stay time, the later two assumptions make the pro-
duct for the recirculation region equal to the product
for the primary zone flow.
A-2
-------
d. Since the recirculation flow is assumed to leave the
main flow at one level and reenter the main flow in the
level above (Section 2, Figure 2-3), the recirculation
flow at any one level is treated as in series with and
between the primary and secondary flows at that level.
e. The recirculation flow entering and mixing with primary
zone flow at a given level is taken at the A/F ratio of
the burner below.
f. The flow departing the main flow at a given level to
recirculate (the flow treated as in series with the
primary and secondary flows at that level) is taken at
the A/F ratio resulting from mixing of 68 percent of
the primary flow at that level and 32 percent of the
primary flow from the burner below.
A. 4 SECONDARY MIXING ZONE
The secondary mixing zone is considered physically to
be that zone where the primary and recirculation flow mixing is taking
place. The uniform A/F ratio in the zone is taken to be the same as
that in the recirculation region. The length of the zone at the lowest
level is taken to be one third of the distance from the end of the primary
zone to the furnace centerline (opposed firing) or to the backwall (single
wall firing). This length is reduced by the bulk gas flow, linearly, to
zero from the fourth to the sixth level. The flow is assumed to fill
one half of the furnace width, and therfore the average velocity in the
zone is that of the primary zone reduced by one half of the ratio of the
burner-to-furnace horizontal width. Cooling time to the zone is the
sum of the stay times in the primary and recirculation regions plus
one half of the stay time in tiie secondary zone.
A. 5 ADJACENT AND OPPOSITE MIXING ZONES
In the adjacent and opposite mixing zones, one half of
the secondary zone flow from a given burner is taken to mix with that
from the horizontally adjacent burner and the other half with that from
the horizontally opposed burner (if opposed firing). Adjacent mixing
A-3
-------
is prohibited where a divider wall may exist. The lengths of both
adjacent and opposite zones at the lowest level are taken as two thirds
of the distance from the end of the primary zone to the furnace center-
line (opposed firing) or the back wall (single wall firing). The length of
these zones is linearly reduced to zero, by the bulk gases, from the
first to the third level. The flow velocities in these regions are taken
to be the same as that of the average in the secondary mixing zones.
Therefore, the stay times in the adjacent and opposite mixing zones
is twice that of the secondary zones. Cooling time to the adjacent and
opposite mixing zones is the sum of the stay times in the primary,
re circulation, and secondary zones plus one half of that in the adjacent
and opposite zones.
A. 6 BULK GAS MIXING ZONES
The gases entering the bulk gas mixing zones at each
level are the sum of those from the burners at that level plus the
bulk gases from below. All of the recirculated flue gas is assumed
to enter the bulk gases, only, at the first level. Since the NO introduced
Ji
into the boiler with the FGR is again removed from the flue gases at
the boiler exit, forming a steady closed loop, flue gases are treated
throughout as inert gases containing no NO . From the standpoint of
Jt
NO concentrations, the main fuel and air reaction products are
treated as undiluted by the (parallel) stream of recirculated flue gases.
The FGR in this model, however, does affect the bulk gas temperature
and velocity. The length of each bulk gas mixing zone is taken as the
distance between burner centerlines (levels). The bulk gas velocity is
calculated from the total fuel, air, and FGR flow and the overall
furnace cross-sectional area. This velocity is assumed constant at all
levels of the furnace, with the bulk gas flow area assumed to spread
(Section Z, Figure 2-Z) in proportion to the bulk gas flow until all flows
are in the furnace and the bulk gases fill the total cross-sectional area.
A-4
-------
Cooling times for the flow coming from the burners are the sum of all
of the intermediate mixing zone stay times. Cooling times for the
bulk gases are the stay times for one bulk gas mixing zone. The gases
from both sources are first mixed to determine an average temperature
at the beginning of the bulk gas zone and are then allowed to cool for
one half of the bulk gas zone stay time, when a mean uniform tempera-
ture for that zone is calculated.
A. 7 NO PORT MIXING ZONE
At the end of the primary furnace zone (not the burner
primary mixing zone), the bulk gases are at some calculated tempera-
ture and of some composition. If there are NO ports in the furnace
Ji
which are open, the composition of the bulk gases will be that resulting
from substoichiometric A/F combustion in the furnace primary region.
As the air flow from the NO ports is mixed into this bulk gas flow
ji
the A/F ratio of the bulk gases will pass through stoichiometric,
ending at the final overall furnace A/F ratio, which always contains
excess (above stoichiometric) air. The rate of this mixing is not
known. In this model, the mixing rate is assumed to be rapid and,
therefore, the air-fuel ratio in the NO port mixing zone is always
JL
taken at the overall furnace air-fuel ratio, regardless of the presence
of NO port,flow. In the four furnaces studied which have NO ports,
X X
the length of this zone was arbitrarily set at 10 ft. The velocity
through the zone is again taken to be the constant furnace velocity.
A. 8 FINAL MIXING ZONE
In all probability, by the time any NO port flow is
Ji
mixed into the bulk gases, a large amount of heat will have been
rejected to the water walls, and the gas temperature should be low
enough that NO formation rates are negligible. A final zone was
established, however, reaching from the top of the NO port mixing
a
zone to the first major convective cooling tubes (end of the radiant
A-5
-------
sectio of the boiler). The air-fuel ratio is that of the overall furnace.
The velocity (for stay time) is that of the bulk gases. The mean
temperature is taken as that calculated at the end of the NO port
mixing zone (or top level of bulk gases if no NO port flow is involved),
Ji.
further cooled by the constant cooling rate for one half of the final-
zone stay time.
A. 9 SIMPLIFICATION
Much of the complicated reasoning, assumptions, and
simplifications are to establish what important furnace geometry,
burner array, and operating condition variables can be significant to
the overall generation of NO . The detailed calculations of NO
X X
generated in each and every possible mixing zone are relatively un-
important. Of primary importance is the fact that the sum of the NO
X.
contributions of these individual mixing zones provides at least one
rational means of explaining or accounting for the effects on overall
NO generation of such primary variables as: (a) the overall boiler
Ji
air-fuel ratio, (b) NO ports, (c) BOOS, (d) combustion air and fuel
temperatures, (e) any combination of adjacent burners (vertically,
horizontally, and opposite), (f) furnace cooling factors, (g) furnace
geometry, and (h) the FGR.
The sum of these 104 mixing zones having been found,
as a function of these multiple independent parameters, the sum
can now be simplified for easier handling as a correlation equation.
Much of the error and gross simplification will be corrected in the
data correlation. The remaining errors will determine the goodness
of the fit.
A separate A/F ratio must be used for each of the bulk
gas mixing zones, the NO port, and the final mixing zones. For the
Ji
remaining 96 burner mixing zones, however, only 4 separate A/F
calculations are necessary. These result from certain (F+A) and (Air)
A-6
-------
burner combinations. No A/F calculations are necessary for (Air)
burners alone or for mixing between (Air) burners since no NO can
Ji
be generated in these pure air zones.
From the standpoint of cooling times (mean times to a
given zone) and stay times in a zone, the bulk gas, NO port, and final
Ji
mixing zones require separate calculations. Only 6 separate calcula-
tions of cooling and stay times are necessary for the 96 burner mixing
zones. These result from the four main mixing zones, defined at
the lowest level and the secondary and adjacent opposite zones shortened
by the bulk gases at the fifth and second levels, respectively.
The possible combinations of the 4 A/F and the 6 time
calculations result in 14 separate terms to describe the NO contri-
Ji
buttons of the 96 burner mixing zones. Together with 6 terms to
describe the 6 bulk gas mixing zones and 1 each for the NO port and
.X
final mixing zones, a total of 22 terms to describe the NOX contri-
tions from all of the 104 possible separate mixing zones. These 22
terms are designated by a code as follows:
where
ANO = C(NABXX) (DABXX)
N = the numbers of zones of the type ABXX
D = the NOX contribution from all zones of the type
ABXX
A = the type of adjacent burner combinations yielding a
given type of A/F in the zone (for convenience, six
numbers are used although they represent only four
A/F)
where
1 = primary (F+A) burner
2 = an (F+A) burner vertically above an (F+A) burner
or a zero flow burner (same A/F as 1).
A-7
-------
3 = an (F-KA) burner horizontally adjacent to or opposite
an (F+A) burner or a zero flow burner (same A/F
as 1).
4 = an (F+A) burner vertically above an (Air) burner
5 = an (Air) burner vertically above an (F+A) burner
6 = an (F+A) burner horizontally adjacent to or opposite
on (Air) burner
B = the type of mixing zone
where
P = primary
R = re circulation
S = secondary
O = horizontally adjacent or opposite
XX = numbers describing the appropriate burner levels
where the zones are similar. Where only one level
is described the code is OX.
For example, the term describing the NO contributions
ji,
from all recirculation mixing zones where an (F+A) burner is directly
above an (Air) burner would be
ANO = C (N4R16) (D4R16) (A-2)
3t I
The six bulk gas mixing zone contributions are arbitraily
designated by the code DNFX, where X = 1 to 6, The NO port and
X,
finaly mixing zone contributions are simply DNST and DNT, respectively.
The 22 parameters used to describe the total thermal NO generated
Jt
in the entire boiler are listed below:
1. (N1P16) (D1P16) 5. (N3001) (D3001)
2. (N2R16) (D2R16) 6. (N3O02) (D3O02)
3. (N2S14) (D2S14) 7. (N4R16) (D4R16)
4. (N2S05) (D2S05) 8. (N2S14) (D4S14)
A-8
-------
9. (N4S05) (D4S05) 16. DNF2
10. (N5R16) (D5R16) 17. DNF3
11. (N5S14) (D5S14) 18. DNF4
12. (N5S05) (D5S05) 19. DNF5
13. (N6O01) (D6O01) 20. DNF6
14. (N6O02) (D6O02) 21. DNST
15. DNF1 22. DNT
Each of the 22 terms in the correlation equation,
describing the mixing zones, involves one of these parameters and an
unknown coefficient to be determined by the linear regression analysis
of data. For correlation purposes, a simple constant term is also
included in the sum. Finally, a term is rather arbitrarily included in
the sum to account for the NO contribution from the bound nitrogen
in oil fuels. This bound-nitrogen term will be discussed later.
The initial correlation equation used consisted of 24
terms, 22 for thermal nitrogen conversion, and one constant. These
were considered adequate to correlate any case of wall-fired boilers
with gas or oil fuels and with or without staged combustion or BOOS.
Correlations involving furnaces with less than 24 total burners or
burner arrays of different width or height were handled by deleting or
expanding some of these terms.
A. 10 CALCULATION PROCEDURE FOR THERMAL NO..
^«^^^^^^^^^^^^^^^^^^^^^^^^^ Jt
Calculation of the ANO terms of .the type shown in
ji
Eq. (A-l) thermal fixation involves the number of times a given com-
bination of (F+A) and (Air) burners occurs in burner array (NABXX)
and the increment of NO formed in zones of that type. The calcula-
A.
tion of NABXX terms is straightforward, involving only the summing
those combinations. All of the complexity is in the DABXX terms.
The total ANO produced in the furnace is the sum of
all of the NO increments in each zone that a given burner flow passes
x
A-9
-------
through. Because of the way that the burner mixing zones were set
up, the flow from each burner passes, in series, through primary,
recirculation, secondary and combined adjacent-opposite mixing zones.
The ANO generated in each of these zones is represented by the first
14 terms of the correlation equation. Each of these terms is the pro-
duct of (a) the average total weight flow from a single burner, (b) the
stay time in the zone, and (c) the mean rate of change of NO concen-
Ji
tration within each zone. But the stay time in a given zone is calculated
from the length of the zone divided by the average flow velocity in the
zone. Therefore, the stay time is inversely proportional to the total
weight flow from a burner, and the product of (a) and (b) is independent
of the flow rate through the zone.
The vertical flow velocities in the bulk gas, the NO
JL
port, and the final mixing zones were assumed to be equal. This
velocity is inversely proportional to the total furnace throughput. The
product of the total flow through these zones and the stay times in the
zones, therefore, are independent of total flow.
This does not imply that the total NO emitted from
the furnace is independent of the furnace load. At reduced load or
throughput flow, the cooling times become longer, and the mean
temperature in a given zone and the combustion air temperature will be
lower. Since the effect of cooling times appears in the argument of the
exponential in Section 2, Eq. (2-1), the rate of NOx formation will be
reduced. Total NO generated, then, will decrease with decreasing
load, as was consistently observed in many boilers.
In the calculation of the DABXX terms, however, this
constant product is a convenient simplification in that the product of
flow rate and stay time can be combined into a pseudo-stay time, which
is almost strictly a function of geometry alone. Since most of the NO
ji
is generated in the burner mixing zones and burner geometry is not
greatly different between the boilers to be correlated, this constant
A-10
-------
product should not be greatly different between boilers. For purposes
of correlation, this relatively constant product term is multiplied by a
coefficient which is determined by the data correlation. For simplicity,
the product of flow rate and stay time is incorporated into the correla-
tion coefficient and is not calculated as such. The parameter DABXX
in Eq. (A-1),therefore, is simply the rate of change of the NO mole
ji
fraction, as shown in Section 2, Eq. (2-1). It must be kept in mind,
however, that the correlation coefficients [C in Eq. (A-l)] of each
term have the dimensions of a weight flow rate times a time.
Estimation of the stay times themselves cannot be
eliminated entirely because the appropriate sums of stay times repre-
sent the mean cooling times to the zones. The stay times are all
calculated from the mixing zone lengths and the flow velocities. From
the general descriptions of the mixing zones, the burner mixing zone
lengths (in Fortran notation) were determined as follows:
For the primary and recirculation zones:
ZLP = 2.* HWB (A-3)
For the adjacent-opposite zones in opposed-firing
boilers:
ZLA0 =[(HDF/2.) - ZLP]/1.5 (A-4)
For single wall-firing boilers:
ZLA0 = (HDF - ZLP)/1.5 (A-5)
For the secondary zones:
ZLS = ZLA0/2. (A-6)
A-ll
-------
The lengths for the bulk gas, NO port, and final
Jt
mixing zones are defined as HBB, ZLNP, and ZLF, respectively.
Again from the assumptions concerning the mixing
zones, the velocities in the zones are as follows:
For the primary and recirculation zones:
VP = WB/(RH0 * AB) (A-7)
For the secondary and adjacent/opposite zones:
VSA0 = 2.* WB/(RH0 * HWF * HWB) (A-8)
For vertical furnace velocities:
VF = WPT/(RH0 * HWF * HDF} (A-9)
Equations (A-8 and (A-9) can be written in terms of the
primary zone velocity:
VSA0 = VP * if * HWB/(2.* HWF) (A-10)
and
VF = VP * WPT * AB/(WB * HWF * HDF) (A- 11)
By use of (A-7), (A-10), and (A-11) and the appropriate
mixing zone lengths, all of the zone stay times can be calculated in
terms of the primary zone stay time. The primary zone stay time is
calculated from
A-12
-------
TUP = .01369 * ZLP * AB/WB (A-12)
The remaining mixing zone stay times are:
TUR = TUP (A-13)
TUS = TUP * K21/2 (A-14)
TU0 = TUP * K21 (A-15)
TUF = TUP * KF0 (A-16)
where
K21 = (2. * HWF * ZLA0)/(TT * HWB * ZLP) (A-17)
and
KF0 = (WB * HWF * HDF * HBB)/(WPT * AB * ZLP) (A-18)
Finally, the mean cooling time to a given zone can be
calculated by summing the stay times in the preceding zones and one
half of the stay time in the given zone. For proper coordination of
cooling times with the corresponding values of NABXX and DABXX,
a similar code was assigned such that cooling times to the burner mixing
zones are designated by TUBXX terms, where TU designates a cooling
time and B and XX have the same meaning as in NABXX and DABXX.
Bulk gas cooling times are designated by TUFX terms. Recalling that
the secondary zones are shorter at level 5 and the adjacent-opposite
zones shorter at level 2, the calculations for cooling times to the
burner and bulk gas mixing zones are as follows:
A-13
-------
TUP16 = .5 * TUP (A-19)
TUR16 = 1.5 * TUP (A-ZO)
TUSH = (1. + .25 *K21) * TUP (A-21)
TUS05 = (!.+. 125 * K21) * TUP (A-22)
TU001 = (1. + K21) * TUP (A-23)
TU002 = (1. + .75 * K21) * TUP (A-24)
TUF(l) = (1. + 1. 5 * K21) * TUP (A-25)
TUF (2) = (1. + 1.25 * K21 +0.5 * KF0) * TUP (A-26)
For boilers H5/6, the burners are separted between
burner levels 3 and 4 by more than the separation between other levels
(because of the vertical grouping of burners into cells of three burners
each). Other LADWP boilers maintain equal spacing and have no more
than four levels. Therefore, for the H5/6 boilers, it was found:
TUF(3) = (1. + K21 + 2. * KF0) * TUP (A-27)
TUF(4) = (1. + .875 * K21 + 3. * KF0) * TUP (A-28)
TUF(5) = (1. + .75 * K21 + 3.2 * KF
-------
and for all other LADWP furnaces:
TUF(3) = (1. + K21 + KF0) * TUP (A-31)
TUF(4) = (1. + .875 *K21 + 1.5 * KF0) * TUP (A-32)
The combustion product temperature in each of the
combined zones where the A/F ratio in the zone and the cooling time
to the zone are the same is then calculated from the sum of the initial
reactant temperature, the temperature increase due to reaction at the
A/F ratio in the zone (shifting equilibrium) and the temperature decrease
due to cooling enroute to the zone. The initial reactant temperature was
calculated as that resulting from mixing of the air flow through the
operating (F+A) burners at temperatures measured at the outlet of the
air preheater and the fuel at ambient temperature (300 degrees K)
(80 degrees F). No data were available to estimate combustion air
temperature changes between the air heater outlet and the individual
burners. The temperature increase due to reaction was calculated
using an Aerospace equilibrium combustion program and read into the
correlation program in tabular form as a function of A/F ratio alone.
The calculation was performed under constant conditions of l-.atm
pressure and 533 degrees K (500 degrees F) reactant temperatures.
This same calculation yielded the equilibrium concentrations of mole-
1/2
cular nitrogen and oxygen, and the product of [^ftC^] was also
entered in the correlation program in tabular form. Cooling enroute
to the zones under the constant cooling rate assumption discussed in
Section 2 was calculated simply from the product of that constant
cooling rate and the cooling times shown above in Eqs. (A-19) through
(A-32).
A-15
-------
The complete correlation equation, consisting of the
sum of 24 linear terms, was used in some preliminary correlations
to establish those terms and combinations of terms that were significant
to the correlation, yielded acceptable correlation coefficients, and
gave the most meaningful results.
None of the terms involved in this correlation equation
are independent of the others. Rather, the independent variables are
buried in each of the terms in complex nonlinear fashion. Thus, it
was expected that the resulting correlation equation would not consist
of terms describing directly the individual contributions of NO from
X
each zone described and that terms would be negative as well as
positive. The correlation program evaluated the significance of each
of the terms to the correlation by determining whether the term was
important in obtaining a good fit of the data. It quickly became clear
that the final zone, downstream of the NO port mixing zone, was
jt
insignificant to the correlation, as were some of the bulk gas mixing
zones and, of course, the bound-nitrogen term in the case of natural
gas fuel. Only the final zone term was deleted on this basis. The
bulk gas terms were instead combined into one term consisting of the
sum of the six bulk gas zone terms DNF 1 throughDNF6. These terms
were still calculated individually but were summed into one term in
the correlation equation. After the initial confirmation, the nitrogen
in the natural gas fuel was considered not as bound nitrogen, and that
term was set to zero in all further correlations of natural gas-fired
data.
Further correlations using an equation consisting of the
remaining 18 terms indicated that good correlations could be obtained,
but some very large coefficients resulted for some of the terms. Ex-
ercising the resulting equation with input conditions not close to the
data actually correlated resulted in the calculation of NO emissions
A-16
-------
that were far from reality. This appeared to result primarily from
allowing too many degrees of freedom in the correlation and from the
fact that the terms in the correlation equation were not independent.
Combining certain closely related terms not only reduced the degrees
of freedom for correlation but increased the degree of independence
of the remaining terms. Consequently, the correlation coefficient
was reduced somewhat, but the resulting equation was quite stable.
As always with empirical data correlations, however, care must still
be taken in interpreting the results of interpolations and extrapolations
that are far removed form existing data. Large data samples
including wide ranges of the independent variables are necessary to
improve confidence in interpolations and extrapolations. Further
discussion on simplification of the correlation equations is contained
in Section 2 . The final result of the simplifications yielded the 10-
term correlation equation briefly described in Section 2, Table 2-1.
A-17
-------
APPENDIX B
FUEL ANALYSES AND COMBUSTION CALCULATION RESULTS
The equation and method used in this study to calculate
the NO generation rate in each zone required inputs for the equilibrium
values of the concentrations of nitrogen and oxygen in the combustion
products and the temperature rise due to combustion. For this pur-
pose, nominal natural gas and low-sulfur oil fuel analyses were
obtained from the utility. These are listed in Table B-l. The Aero-
space N-element chemistry system was used to perform the combus-
tion calculations. Air temperatures were specified at 533 degrees K
(500 degrees F) and fuel temperatures at ambient. Pressure was con-
stant at 1 atm. For purposes of this calculation, equivalent fuels were
defined as follows:
Natural gas = C^ 1103H4 118OQ 0427N0.0174
Heat of formation = -19.71 kcal/mole of the above
equivalent fuel
Stoichiometric A/F ratio = 15.91 by weight
Low-sulfur oil = CQ 5647HQ 8622O0. 0034N0. 0014S0. 0007
Heat of formation = -1. 25 kcal/mole of the above
equivalent fuel
Stoichiometric A/F ratio = 13.89 by weight
B-l
-------
Table B -1. FUEL ANALYSES
Natural gas
Constituent
CH,
C2«6
C3H8
C4H!0
C5H12
co2
N2
«2°
Total
Volume, %
88.17
6.04
2.49
0.27
0.02
2.13
0.87
0.01
100.00
Low- sulfur oil
Constituent
C
H2
S
°2
N2
H20
Ash
-
Total
Weight, %
87.30
11.20
0.28
0.70
0.24
0.26
0.02
-
100.00
B-2
-------
From the equilibrium calculation, the molar
concentrations of nitrogen and oxygen were obtained. These concen-
trations enter the NO generation equation, Section 2, Eq. (2-1), in
Ji
the form:
SON =[N2][02]1/Z (B-l)
where
[N_] = the molar concentration of molecular nitrogen
(N2) in the combustion products
[O2] = the molar concentration of molecular oxygen
The factor SON and the equilibrium combustion tempera-
tures were calculated over a range of A/F ratios and entered into the
correlation program in tabular form. These data are listed in Table
B-2. The nitrogen in the two fuels was also entered in the program as
a fraction of the fuel weight. Initially, until the correlations confirmed
that it was not so, the nitrogen in the natural gas, as well as that in the
low-sulfur oil, was entered and treated as bound or fuel nitrogen.
These values were:
Natural gas: 0.01326
Low-sulfur oil: 0.002532
Early in the correlation studies, the bound nitrogen in
the natural gas was changed to zero.
B-3
-------
Table B-2. EQUILIBRIUM COMBUSTION INPUTS FOR
THE NO GENERATION RATE EQUATION
Natural gas
A/F ratioa
9.55
11. 14
12.73
14.32
15.91 (st.)
19.09
22.27
25.46
28.64
31.82
SON
1.348 X 10"4
1.031 X 10"3
5.501 X 10"3
2.395 X 10"2
6.329 X 10"2
0. 1303
0. 1717
0. 1994
0.2194
0.2347
Temperature,
°K
1926
2101
2239
2331
2332
2187
2022
1877
1757
1655
Low- sulfur oil
A/F ratioa
8.33
9.72
11. 11
12. 50
13.89 (st.)
15. 28
16.67
19.45
22. 22
25.00
27.78
SON
3.498 X 10"4
2. 678 X 10"3
1.319 X 10"2
4. 181 X 10"2
0.07899
0. 1120
0. 1393
0. 1798
0.2074
0.2273
0. 2423
Temperature,
°K
2048
2225
2355
2418
2401
2341
2264
2098
1946
1817
1708
w
I
By weight.
-------
APPENDIX C
INTERMEDIATE DATA CONVERSION AND CORRELATIONS
C. 1 INTRODUCTION
At the outset of this study it seemed clear that all of
the primary data necessary for the subsequent correlation effort would
not be available in all tests and that some primary data would have to
be generated from other related data. In many cases, some of the
primary data were not recorded in the data logs because other data
indicated that the primary data would be in a certain known range. For
example, combustion air temperatures might not be recorded when
operating at full load because the full load combustion air temperature
is known within an acceptable range. For the purposes of this study,
however, values had to be filled in for all necessary primary data.
It was also recognized that in exercising the final cor-
relation equations, certain realistic relations between some of the de-
pendent input variables would have to be preserved. For example,
when studying the effect of reduced load on NO emissions, the perfor-
mance of the air preheater at reduced load could be simulated by an
empirical relation between load and combustion air temperature.
As a result, seven intermediate correlations were
established in the data conversion section of the correlation model.
Where all inputs to a given relation were available, an empirical
equation was established, and the standard deviation of the data from
the equation calculated. In preparing the primary data for the final
C-l
-------
correlation, the data conversion program first evaluated for missing
primary data. Where primary data were found missing the available
alternative intermediate correlations were evaluated for missing
secondary data. Where secondary data were available for more than
one alternative intermediate correlation, that one with the smallest
standard deviation of the data was selected, and the missing primary
data were calculated from the appropriate equation. In actual practice,
either sufficient secondary data were available for only one intermediate
correlation, or the primary data were rarely missing. Some of the
intermediate correlations were useful, however, in other applications.
Finally, some other empirical correlations and theoretical
relations were generated for use in data conversion and in exercising
the correlation equations. For the purpose of presenting the empirical
correlations used and providing information which might be of value to
other users, these intermediate correlations are presented in this
Appendix.
C.2 FUEL FLOWS
One of the primary input parameters in this study was
the fuel flow. Because of its importance, two intermediate correla-
tions were attempted. Correlations of fuel flow with load were obtained
for both gas and oil fuels. These equations and the empirical coeffi-
cients are listed in Tables C-l and C-2.
Attempts were also made to generate fuel flows from
data on pressure drops across the gas spuds, or rings, and from the
oil gun tips and the appropriate flow resistances. In few cases,
however, could the gas spud orifice areas or the oil gun tip configura-
tions be positively established, and this approach was abandoned.
The correlation of fuel flows with load were used in a
few cases to fill in missing primary data. One area of significance of
these data, however, lies in the observation that load divided by the
C-2
-------
Table C-l. CORRELATIONS OF NATURAL GAS FLOW WITH LOAD
FGF = APQGF + BPQGF * LOAD
FGF = Natural gas flow, K-SCFHa
LOAD = Electrical output, MW
a = Standard deviation, % of full load fuel flow
Boiler
HI
H2
H3
H4
H5
H6
SI
S2
L3
Full load,
~MW
240
240
240
240
350
350
180
180
82
Data in
correlation
11
20
39
12
88
23
7
33
8
APQGF
103.5
84.02
24.20
59.00
128.4
202.6
56.61
21.06
16.23
BPQGF
8.923
8.765
8.748
8. 158
7.765
7.634
8.263
8.436
10.71
a, %
1. 1
0.9
1.2
2.7
2.6
1.9
2.7
2.0
2.2
lTo convert K-SCFH to m /hr multiply by 28. 32
C-3
-------
Table C-2. CORRELATIONS OF LOW SULFUR
OIL FLOW WITH LOAD
FOF = APQOF + BPQOF * LOAD
FOF = Low-sulfur oil flow, k-lb/hra
LOAD = Electrical output, MW
a = Standard deviation, % of full load fuel flow
Boiler
HI
H2
H3
H4
H5
H6
SI
S2
L3
Full load,
~MW
240
240
240
240
350
350
180
180
82
Data in
correlation
12
33
6
8
39
14
18
23
14
APQOF
6.403
4.076
9.467
2.990
13.69
0. 1347
3.971
5.508
3.062
BPQOF
0.4654
0.4615
0.4630
. 0.4777
0.4138
0.4592
0.4596
0.4281
0.6082
d, %
2.6
1.0
3.6
3.2
1.2
2. 1
2.5
2.3
3.7
lTo convert k-lb/hr to kg/hr multiply by 453. 59.
C-4
-------
fuel flow is a measure of plant efficiency, in megawatts per unit of fuel
flow. Thus, the effects of varying load on plant efficiency can be ob-
tained directly from these empirical equations. In terms of the con-
stants listed in Table C-l, this efficiency expression is
Load _ 1 L APQXF \ ,_. ..
Fuel flow ~ BPQXF \ ~ Fuel flowj V
Since for the boilers studied and for both natural gas and
oil fuels the APQXF values are positive, the efficiency described by
Eq. (C-l) always decreases with load (fuel flow). Values of APQXF
and BPQXF are contained in Tables B-l and B-2.
C.3 AIR FLOWS
Air flows as such were not measured. Air flow could
be obtained from an air foil flow indicator, calibrated in terms of per-
cent of rated flow, and an estimate could be made from the forced
draft fan electrical current, in amperes. Both of these methods, how-
ever, are calibrated using flows calculated from the fuel gas analysis
for the A/F ratio and measured fuel flows. This latter method was
chosen as the primary calculation of air flow data in this study. It
seemed highly likely, and it turned out to be so, that if NO measure-
Jt
ments were made at all, O2 and CO_ measurements would also have
been made. Theoretical expressions were derived from the fuel
analyses and equilibrium combustion calculations to relate the overall
furnace A/F ratio to the O_/CO_ ratio measured in the flue gases:
AFRG = 15.91 + 3.501 * O2/CO2 for natural gas fuel (C-2)
and
AFRO = 13.89 + 4.22 * O2/CO2 for oil fuel (C-3)
C-5
-------
Air flow was calculated from the product of the A/F
ratio, from Eq. (C-2) or (C-3), and from the measured fuel flow.
Fuel flow measurement accuracy was considered to be within 1 percent.
In case the O_ and CO_ data were lacking in a significant amount of the
data, however, the correlations shown in Tables C-3 and C-4 were
generated. They were little used.
C.4 COMBUSTION AIR TEMPERATURE
The only measurement of combustion air temperature
available, except in a few special cases, was the air preheater outlet
temperature. The combustion air must travel from the air" preheater
through considerable air ducting and the windbox before reaching the
the burners, but no data were found to correct the air preheater
out temperature to burner air temperature. The few windbox tempera-
ture measurements available appeared to show contradictory trends.
Since the air flows through uninsulated ducts but also along the outside
of the boiler water walls, no simple method using available data was
discovered to generate a correction method. Throughout this study,
then, it was assumed that there was no change in the combustion air
temperature from the air preheater outlet to the burners.
Even this temperature, however, was often not recorded.
It was also clear that some method of predicting this temperature as a
function of load would be necessary in later exercising of the correlation
equation. Correlations of combustion air temperature (air preheater
outlet) with load were generated, therefore, and are listed in Table C-5.
These correlations were useful both in filling out missing primary data,
and in performing later studies of the correlation equations.
C-6
-------
Table C-3. CORRELATIONS OF COMBUSTION AIR FLOW WITH
AIRFOIL PERCENT FLOW INDICATOR
FDFP = 2116. + 5. 2 * AFDP
WAT = FDFP * (APQAM + BPQAM * WATAF)/TPA
WAT = Combustion air flow rate, Ib/sec
AFDP = Forced draft fan pressure, in. of water
TPA = Combustion air temperature, degrees Rankine
WATAF = Airfoil indicator, % of full air flow
a = Standard deviation, % of full load air flow
Boiler
HI
H2
H3
H4
H5
H6
i
Data in
correlation
9
19
35
11
45
19
APQAM
- 3.018
18.80
-12.74
-15.42
-17.07
- 2.583
BPQAM
2.769
2.259
2.497
2.641
3.347
3.072
a, %
1.8
2.7
4.6
1.5
3.2
1.0
To convert from
to
Multiply by
Ib
/ 2
gm/cm
°K
0.4536
in. of water 0.3939
°R 1.80
C-7
-------
Table C-4. CORRELATIONS OF COMBUSTION AIR FLOW WITH
FORCED DRAFT FAN CURRENT
QAFA = APQRA + BPQRA *AFAA + CPQRA * AFAA #AFAA
QAFB = APQRB + BPQRB * AFAB + CPQRB *AFAB * AFAB
WAT = 1.225 * (QAFA + QAFB)
AFAA, AFAB = Forced draft fan current in the A and B fans, amp
WAT = Combustion air flow rate, Ib/sec
0- = Standard deviation, % of full load airflow
Boiler
HI
H2
H3
H4
H5
H6
Boiler
HI
H2
H3
H4
H5
H6
Data in
correlation
11
ZO
39
12
64
21
APQRB
-24.80
6.068
-71.57
-85.65
-15.23
-98.61
APQRA
-27.75
18.20
-73.98
-79.86
-61.90
-54.39
BPQRB
1.977
1.584
2.487
2.854
1.519
2.602
BPQRA
2.042
1.329
2.535
2.781
1.406
Z.066
CPQRB
-0.004382
-0.003150
-0.005713
-0.007709
' -0.002037
-0.004757
CPQRA
-0.004607
-0.002100
-0.005981
-0.007385
-0.001801
-0.003349
a, %
1.9
3.2
2.2
3.5
2.5
2.8
lTo convert Ib/sec to kg/sec multiply by 0.4536.
C-8
-------
Table C-5. CORRELATIONS OF COMBUSTION AIR
TEMPERATURE WITH LOAD
TPA = APQRT + BPQRT * LOAD + CPQRT * LOAD * LOAD
TPA = Combustion air temperature (air heater air outlet
temperature), °F
LOAD = Electrical output, MW
a = Standard deviation, % of full load temperature
Boiler
HI
H2
H3
H4
H5
H6
SI
S2
L3
Data in
correlation
9
18
35
12
70
19
7
33
8
Full load
temperature
°F
587
569
536
531
575
572
552
557
627
APQRT
487.3
486.3
479.0
466.6
361.8
501.9
422.8
504.7
377.7
BPQRT
0.3883
0.4822
0.02746
0.2480
0.6709
0.03082
1.079
0.007385
2. 147
CPQRT
0.0001151
-0.0005756
0.0008668
7.950 X 10"5
-1.717 X 10~4
4.813 X 10"4
-0.002005
0.001574
0.01093
a, %
0.5
1.3
1.7
0.04
3.5
2.6
0.8
1.4
0.9
aTo convert from °F to °K: TR = (5/9)(TF + 460).
C-9
-------
APPENDIX D
INPUT DATA FOR CORRELATIONS
This Appendix contains the significant input data
derived from the 267 tests with natural gas fuel and the 161 tests with
low-sulfur oil fuel in 6 of the 7 boilers listed in Section 2, Tables 2-1
and 2-2. The data from 8-gas and 16-oil tests in boiler L3 were
omitted because burner admittance data could not be derived from
that boiler and that data was not used in this study.
The operating data presented here are largely self-
explanatory. The "Air Fuel Ratio" data are A/F ratios by weight.
The following conversion rates are given for convenience:
a. To convert from Ib/sec to kg/sec multiply by
0.4536
b. To convert from °F to °K: TK = (5/9)(TF + 460).
In the burner configuration data, the burners are
identified by a numbering system which starts with burner 1, located
on the bottom level of the burner array on the extreme left side, as
seen when facing the front of the boiler. Subsequent numbering
proceeds around the boiler at the bottom level, as seen from the out-
side of the boiler, until the entire bottom level is accounted for.
The next number, is assigned to the burner just above burner 1 and
again proceeds around to the right. Figure 2-3 shows this numbering
system for the H5/6 boilers.
D-l
-------
The individual burners are considered operating fully
under one of three conditions:
a. 0 = No flow of either fuel or air
b. 1 = Air flow only (Air)
c. 2 = Both fuel and air (F+A)
While fuel can be shut off reliably, the air flow can only be shut off by
poorly sealing air registers. Nevertheless, in this study, when a
burner is indicated "0", the air flow is taken as zero. Obviously, if
a boiler has less than 24 burners, then burners listed for numbers
larger than the maximum for that boiler do not exist, and the fuel and
air flow for these burners are both truly zero.
The "NO ports" operational designation is the same
ji
as that for the burners except, of course, they cannot be designated
"2". Again, NO port flow is shut off with a damper, which could
3t
leak air, but the NO port flow is considered zero when designated
"0".
For both the burners and the NO ports, when a definite
effort was made and noted in the data to partially throttle the air
flow to some burners or NO ports, these tests were omitted from the
sample listed in this Appendix and were not used in this study. It was
considered impossible, with existing instrumentation, to estimate the
degree of throttling. In all cases used in this study, burner registers
and NO ports are either fully open or fully closed, as best as this
Jt
can be determined.
Flue gas samples were drawn from the flue gas ducts
between the economizers and the air preheaters in each boiler.
Primary data used in this study, when available, were from grab
samples analyzed by the chemical laboratory of the utility. Labora-
tory personnel conducted standard analyses of these samples by
accepted procedures. NO measurements are per ASTM D-1608.
D-2
-------
The majority of the emissions data used in this study
were from analyses conducted in the Air Quality Control Mobile
Laboratory (AQCML). This on-site laboratory consists of a 32-ft
trailer containing all necessary monitoring equipment and a complete
sample conditioning system. It is capable of sampling 30 different
points in the boiler ducts. The AQCML instrumentation includes the
following:
a. CO,: MSA model 300, NDIR, 0 to 20 percent
b. CO: Beckman model 315(s), NDIR, 0 to 400,
0 to 1000, and 0 to 2000 ppm
c. NO: Beckman model 315 (L), NDIR, 0 to 200,
0 to 500, and 0 to 1000 ppm
d. O~: Beckman model 742, electrochemical,
1, 5, 10, and 25 percent
In addition, some data were from tests conducted on-
site by ESSO Research and Engineering Company. The emissions
instrumentation used in these tests are described in a report by
W. Bartok, A.R. Crawford, and G. J. Piegari, Systematic Field
Study of NO Emission Control Methods for Utility Boilers, Final
Report No. GRU.4G. NOS. 71, NTIS Report No. PB 210-739
(December 1971).
Most of the boilers were equipped for continuous
monitoring of CO and NO , using MSA Model 300 CO analyzers, Theta
Model LS800 ANXR 1 sensors, and EnviroMetrics Model N-76DF
NO analyzers. Data from these sources were rarely used.
The following pages contain input operating conditions
and burner configuration data.
D-3
-------
NATURAL GAS FUEL
OPERATING DATA FOR BOILER HI
NO.
1
2
3
it
5
b
7
d
9
10
0 "
*».
UATE
061»71
061571
061571
Oo2*»71
Ot>*78.
181.
266.
395.
<>66.
145.
207.
..87.
<»66.
AIR FUEL
RATIO
15.6
16. <»
16.3
17.1
16.5
lo- 8
17.0
19.7
Id. 7
16.8
16.9
COMB. AIR
TEMP.
550.
569.
587.
519.
539.
56U.
573.
507.
519.
582.
586.
FOR FLOW
(LBS/SEO
106.
62.
0.
117.
98.
6<».
*»7.
135.
102.
62.
66.
-------
NATURAL GAS FUEL
BURNER COMFIGJRATIONS FOR BOILER HI
NO. OATE TEST
3*5
8 3 10 11 12 13 1*. 15 16 17 18 19 20 21 22 23 2U NOX P09TS
0
1
in
1
2
3
>*
5
6
7
a
9
10
11
061571
061571
061571
062U71
Ob2<»71
062«»71
062<»71
102071
102071
111671
111971
1
2
3
1
2
3
i»
1
2
2
2
?
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
9
2
2
2
2
2
2
2
5
2
2
2
2
2
2
2
2
2
2
^
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
1
I
1
1
1
1
1
2
2
2
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
1
1
1
1
1
1
1
0
0
0
0
0
0
0
0
Q
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
a
0
0
0
G
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
c
0
0
0
0
0
0
0
0
G
0
D
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
a
0
0
0
a
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
-------
NATURAL GAS FUEL
OPERATING DATA FOR BOILER H2
NO.
12
13
1*
15
16
17
18
19
20
21
D 22
^ «
2*
29
26
27
28
29
30
31
DATE
0*0671
0*0671
0*0671
0*1971
0*1971
0*2071
0*2071
0*2071
0*2171
0*2171
0*2171
0*2171
0*2171
0*2171
0*2171
0*2171
060871
060871
060871
06G871
TEST NO
(PPMI
1
2
3
1
2
3
*
5
1
2
3
*
5
6
7
8
1
2
3
*
111.
152.
353.
175.
120.
2*0.
219.
130.
71.
78.
63.
11*.
112.
11*.
116.
8*.
51.
*6.
a*.
92.
CO
(P°M>
*0.
-.0.
uj.
100.
120.
80.
30.
500.
80.
i*a.
ao.
130.
so.
50.
30.
100.
20.
93.
277.
275.
02
-------
NATURAL GAS FUEL
BURNER CONFIGJRATIONS FOR BOILER HZ
NO. OATE TEST
5 b
3 9 10 11 12 13 !< 15 16 17 18 19 20 21 22 23 2<» NOX PORTS
12
13
lH
15
16
17
18
19
20
21
22
23
2*
25
26
27
28
29
30
31
0*0671
0*0671
0*0671
0*1971
0*1971
0*2071
0*2071
0*2071
0*2171
0*2171
0*2171
0-2171
0-2171
0*2171
0*2171
0*2171
OoU371
060871
060871
060971
1
2
3
1
2
3
*
5
1
2
3
u
5
6
7
e
i
i.
3
H
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
1
2
2
1
2
2
2
2
1
1
1
1
2
2
2
2
2
2
2
1
2
2
2
2
1
1
1
1
1
1
1
1
1
1
1
1
2
2
2
1
2
2
2
2
1
1
1
1
1
1
1
1
1
1
1
1
2
2
2
1
1
2
2
1
2
2
2
2
1
1
1
1
2
2
2
2
2
2
2
2
1
1
1
2
1
1
1
1
2
2
2
?
1
1
1
1
2
2
2
2
2
2
2
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
2
2
2
2
2
2
2
2
2
2
2
2
2
?
2
2
1
1
1
2
1
1
1
1
2
2
2
2
1
1
1
1
0
0
0
a
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
3
0
0
0
0
0
0
3
0
0
0
0
0
0
0
0
0
G
0
0
0
0
a
0
0
0
0
0
0
0
0
0
0
a
0
0
0
0
0
0
0
a
a
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
a
0
0
0
0
0
d
0
0
G
0
0
0
G
0
G
0
I
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
g
0
0
0
0
0
0
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
-------
NATURAL GAS FUEL-
OPERATING DATA FOR BOILER H3
0
1
00
NO.
32
3,5
34
35
36
37
38
39
»0
Hi
*2
*3
<»«»
45
tb
47
48
t9
50
51
52
53
5 it
DATF
061171
Q61171
061171
061171
091371
091371
091571
091571
091571
091571
091571
091571
091671
091671
0917M
091771
092171
092171
092171
092171
092271
092271
092271
TEST NO
tPPHI
1
2
3
%
1
2
1
2
3
4
5
f>
1
2
1
2
1
2
3
<»
I
2
3
130.
163.
132.
144.
130.
173.
378.
403.
319.
108.
122.
270.
240.
267.
LfcQ.
134.
111*.
150.
210.
267.
134.
171.
303.
CO
tPPH»
-0.
-a.
-0.
-3.
15.
6.
St.
93.
73.
53.
!<*:>.
71.
67.
1 + 0.
6<+.
56.
19.
23.
60.
38.
13.
20.
66.
02
tPCH
2.05
2.<*0
1.70
1.95
1.86
1.92
1.76
2.78
3.17
1.58
2.90
*.i.3
3.10
1.73
2.78
3.20
3.30
3.faS
3.70
3.J.S
6.95
2.25
1.86
C02
tPCT>
9.15
1Q.23
10.90
10.95
11.43
11.57
11.55
10.65
10.33
11. 51
10.<*7
9.73
10.67
11.50
10.95
10.60
9.60
9.95
9.80
9.85
a,<+a
10.65
11.25
LOAD FUEL FLOW
IHVt> tLBS/SECl
SD.
130.
183.
2<«0.
211).
210.
21*2.
237.
235.
213.
233.
239.
2<»2.
2*1).
2
-------
NATURAL GAS FUEL
OPERATING DATA FOR BOILER H3
NO. DATE TEST NO CO
(PPM)
02 002 LOAO FUEL FLOW AIR FLOW AIR FUEL
(PCT) (PCT) MWI (LBS/SECI (LBS/SEC) RATIO
55
56
57
56
59
60
61
62
63
64
be.
66
67
68
b9
/O
092271
092371
092371
092371
092371
092371
092871
092671
042871
0*2871
092871
092871
093071
100171
100171
100171
4
2
3
it
6
7
1
2
3
4
5
6
1
1
2
3
169.
53.
5<*.
10W.
10i*.
155.
81.
38.
79.
59.
104.
135.
193.
95.
120.
109.
62.
<«6.
28.
136.
iti.
72.
22.
117.
65.
84.
32.
67.
175.
23.
25.
39.
2.93
t.*7
3.13
2.49
2.26
2.21
9.33
5.48
4.83
3.92
3.8?
3.65
1.85
3.76
5.50
5.25
11.07
4.93
10.25
10.97
11.25
11.33
7.03
9.17
9.1*5
10.37
10.63
10.60
10.63
7.35
9.25
9.15
2<*1.
65.
140.
200.
200.
2*0.
45.
8
-------
NATURAL GAS FUEL
BURNER CONFIGJRATIONS FOR BOILER H3
NO. DATE TEST 1 2 3 <* 5 & 7 8 4 10 11 12 13 It 15 16 17 18 19 20 21 22 23 2«f NOX PORTS
32
33
3t
35
36
37
38
39
to
tl
<*2
a **3
0 ^5
<&
t7
-------
NATURAL GAS FUEL
BURNER CONFIGURATIONS FOR BOILER H3
NO. OATE TEST 1 2
5 6 7 S ? 10 11 12 13 1<» 15 16 17 18 19 20 21 22 23 2* NOX PORTS
d
1
H*
55
56
57
58
59
60
61
t>2
63
b<»
65
66
67
68
69
70
092271
092371
J92371
092371
092371
092371
092(571
092871
092871
0*2871
092871
092371
093071
100171
100171
100171
<
2
3
»
6
7
1
2
3
<4
5
b
1
1
2
3
2
2
2
2
^
2
0
0
2
2
2
2
2
0
0
2
2
2
2
2
2
2
0
0
2
2
2
2
2
0
0
2
2
2
2
2
2
2
2
2
2
2
2
2
2
0
c
2
2
2
2
2
2
2
2
2
2
2
2
2
2
3
0
2
2
2
2
2
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
1
2
2
2
2
2
2
2
2
2
2
1
1
1
1
2
i
2
2
2
2
?
2
2
2
2
2
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
1
1
1
1
2
2
1
1
1
1
1
1
2
2
2
2
1
1
1
1
2
2
1
1
1
1
1
1
2
2
2
2
2
2
2
2
1
1
1
1
1
1
1
1
1
1
1
1
2
2
2
2
1
1
1
1
1
1
1
1
1
1
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
c
0
a
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Q
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
a
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
a
0
a
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
a
0
0
0
0
0
0
0
0
0
0
1.
1.
1.
1.
1.
1.
0.
0.
0.
1.
1.
1.
1.
0.
0.
1.
-------
NATURAL GAS FUEL OPERATING DATA FOR BOILER HI.
NO.
71
72
73
7*.
75
76
77
78
79
00
O ai
i
82
OATt
051171
051171
051171
051171
052771
052771
052771
061171
061171
061171
061171
051172
TEST NO
(PPM)
1
2
3
<»
2
3
H
1
2
3
1
99.
!««<*.
106.
126.
153.
230.
373.
330.
200.
161.
125.
203.
CO
(PPHI
-0.
-0.
-0.
-0.
-0.
-0.
-0.
-0.
-0.
-0.
-0.
-0.
02
(PCTI
6.95
3.10
1.85
2.10
2.67
2.1.3
£.1.0
2.05
1.60
2.05
5.65
3.75
C02
(PCT)
7.50
10.15
10.70
10.1.5
9.90
10. 35
10.50
10. *5
10. HO
10.20
6.-.0
9.35
LOAD FUEL FLOW
(MH) <_QS/SF.C>
30.
130.
130.
2*0.
130.
ISO.
2wO.
2*3.
181.
133.
50.
231.
9.2
11*. 8
20.1
27.3
15.2
20. e
2T.7
26. H
20.3
1 -t 6
23.7
AIR FLOW
(LBS/SECI
177.
252.
332.
<»5i».
255.
3<»7.
«.60.
<*39.
33<».
Z*3.
171.
AIR FUEL
RATIO
19.2
17.0
16.5
16.6
16.9
16.7
16.6
16.6
16.5
16.6
IS. 3
17.3
COMB. AIK
TEMP. JFI
1.87.
500.
51<*.
531.
500.
51<».
531.
531.
51U.
501.
<»87.
528.
FGR FLOW
(L3S/SEC1
50.
<»6.
t*3.
0.
*3.
38.
0.
0.
H3.
7<*.
0.
-------
NATURAL GAS FUEL
BURNER CONFIGJRfiTIONS FOR BOILER Hi.
NO.
DATE TEST
5 o 7 8 9 10 11 12 13 it. 15 16 17 18 19 20 21 22 23 2«» NOX PORTS
O
1
71
72
73"
7"»
75
76
77
76
79
60
SI
82
051171
l)»1171
Osll71
U31171
OJ2771
052771
052771
061171
061171
061171
Obll7l
051172
1
2
3
b
2
3
<
1
2
3
te
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
i
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
9
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
7
2
2
2
2
2
7
2
2
1
?
2
2
?
?
2
2
2
2
2
2
1
0
0
0
0
0
0
0
0
0
a
a
0
i)
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
a
0
0
0
0
0
0
0
0
a
0
0
0
0
0
0
0
0
0
0
0
0
c
0
0
0
0
0
0
0
a
n
C
C
J
0
0
0
0
J
0
0
0
0
0
a
0
0
0
0
0
0
c
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
G
a
0
0
0
0
0
0
0
0
0
0
0
a
0
0
a
.1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
1
1
0
1
-------
NATURAL GAS FUEL
OPERATING OAT4 FOR BOILER H5
NO.
DAT-
O
i
03 041569 1
3* 081569 2
65 031569 3
86 061569 4
47 041S69 5
3d 061E69 6
39 Od2069 7
90 062069 fi
91 012069 9
9<> 002069 10
93 0820*9 11
9
-------
NATURAL GAS FUEL
OPERATING DATA FOR BOILER H5
NO.
106
107
108
109
110
111
112
113
IK.
115
O 116
£n ll7
118
119
120
121
122
123
124
125
126
127
128
OATE
082989
012071
012071
012071
012071
012171
012171
012171
012271
012271
012271
.112571
012571
012571
012571
012571
012571
C12671
012o71
012671
012671
012671
C12671
TEST
2.
3L1
BL2
1A
13
2A
20
3A
3B
4A
4B
5A
50
6A
6B
7A
78
8A
30
9A
9B
10A
103
NO
CPP1I
443.
437.
919.
572.
343.
361.
203.
583.
361.
?22.
223.
380.
222.
28-).
205.
516.
293.
+35.
205.
«23.
217.
45<4.
262.
CO
CP»M)
-0.
-0.
-1.
3C.
3C.
20.
-7.
il.
50.
35.
35.
&T.
98.
5 J.
133.
35.
3*.
19.
15.
20.
20.
...
21.
02
(PCTI
3.09
2.65
2.30
1.92
2.38
2.J5
1.45
1.75
2.05
2.00
2.13
1.93
2.00
2.23
2.15
2.15
2.30
i.eo
2.35
2.18
2.15
l.BO
2.10
C02
(PCTI
ID. 09
10.30
10. -.5
10.58
10.43
10.43
10. bQ
10.68
10.43
10. 73
10.53
10.73
10.30
10.43
10.43
10.73
10.53
10.23
10.20
11.73
10.60
10.53
10.63
LOAD FUEL FLOW
MW» (LBS/SEC1
351.
350.
350.
350.
350.
350.
350.
350.
350.
3?0.
350.
350.
350.
3.9.
350.
350.
350.
350.
350.
350.
350.
350.
350.
36.8
37. b
3^.8
37.8
38.0
35.9
37.4
37.7
33.0
37.8
37.8
38.9
38.5
33.2
33.2
33.0
33.0
38.?
33.4
33.0
33.0
37.8
37.8
AIR FLOW
(LBS/SEC)
625.
632.
631.
626.
634.
600.
o20.
b21.
630.
627.
629.
643.
639.
637.
633.
631.
633.
629.
6*1.
629.
631.
625.
628.
AIR FUEL
RATIO
17.0
16.8
16.7
16.5
16.7
16.7
16.6
16.5
16.6
16.6
16.6
16.5
16.6
16.7
16.7
16.0
16.7
1^.5
16.7
16.6
lb.6
16.5
16.0
COMB. AIR
TEMP. (F»
585.
585.
535.
588.
586.
580.
580.
575.
580.
575.
57&.
570.
580.
"580.
580.
580.
580.
580.
580.
578.
580.
580.
580.
FOR FLO
tLBS/SE
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
-------
NATURAL GAS FUL
OPERATING DATA FOR BOILE9 H5
NO.
129
130
131
132
133
134
135
136
137
133
139
140
141
1*2
1*3
14*
1*5
14b
147
1*6
1*9
150
151
DATE
020171
020171
0202*1
020271
02G271
020271
0
261.
236.
186.
313.
225.
245.
194.
345.
199.
372.
b40.
980.
477.
241.
246.
54.
65.
703.
172.
60.
136.
105.
43.
CO
(DPH)
213.
-0.
*6.
5't.
70.
113.
78.
70.
73.
20.
?0.
20.
20.
20.
20.
20.
20.
29.
20.
21.
20.
20.
-0.
02
IPCT)
2.53
3.65
3.11
2.33
2.45
2.63
3.69
1.50
2.20
1.85
1.35
2.40
2.30
3.60
1.30
2.40
2.70
2.10
2.20
2.10
2.10
1.90
1.65
C02
IPCTl
10.1.9
9.79
9.9?
10.33
10.23
10.23
9.74
10. ao
10.40
10.60
10.80
10.40
10.30
9.80
10.70
10.40
11.83
12.23
12.11
12.10
12.00
12.30
13.80
LOAO FJEL FLOW
(MM) (LBS/SECI
350.
350.
350.
350.
350.
350.
350.
350.
350.
350.
350.
350.
350.
350.
153.
151.
152.
152.
153.
152.
151.
153.
147.
39.5
33.6
T9.7
39.6
39.6
39.6
34.7
33.6
13.9
33.9
39.7
39.9
3d. 9
39.9
19.6
19.5
19.5
19.5
14.3
19.3
19.1
17.9
16.6
AIR FLOH
(LBS/SEC)
646.
664.
659.
644.
646.
649.
b&8.
632.
647.
642.
63*.
650.
b56.
669.
304.
309.
308.
305.
303.
503.
299.
295.
273.
AIR FUEL
RATIO
16.8
17.2
17.0
16.7
16.8
16.8
17.2
16.4
16.7
16.5
16.3
16.7
16.9
17.2
16.3
16.7
16.7
16.5
1*.5
lo. 5
16.5
16.5
IS. 4
COMB. AIR
TEMP. (Fl
575.
575.
575.
575.
575.
575.
575.
575.
575.
580.
580.
575.
580.
sec.
415.
420.
385.
385.
495.
495.
459.
500.
493.
FGR FLOH
(LBS/SECI
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
113.
112.
0.
0.
86.
85.
96.
86.
"84.
-------
NATURAL CAS FUEL
OPERATING DATA FOR BOILER H5
NO. OATE
TEST NO CO 02 C02
(PPM) (PPM) (PCTI 0.
220.
3-0.
245.
96.
106.
166.
315.
-0.
-3.
-0.
-0.
-0.
-0.
-0.
55.
125.
S3.
90.
t.
1.0.
SO.
-0.
1*7.
15.
-0.
-0.
-3.
-0.
-0.
-0.
1.55
i.a=>
2.15
'.40
3.10
3.20
2.60
l.oO
1.90
1.90
2. 2 D
2.00
1.90
2.20
2.90
3.10
3. CO
3.00
2.65
2.65
2.65
2.53
3.00
10.75
10. 1.5
10.40
10.15
9.90
9.83
10. 2d
11.30
11.80
11.80
11. 5J
11.60
11.90
11.80
Id. 17
10.60
11.00
10.13
11.23
11.20
11.20
11.30
11.00
199.
?50.
298.
350.
150.
250.
350.
300.
300.
300.
300.
300.
300.
350.
330.
300.
350.
350.
250.
250.
250.
250.
250.
21
26
30
16
15
25
35
32
32
3?
3?
32
32
3*
32
32
37
37
27
27
27
27
27
.6
.3
.<
.9
.8
.0
.6
.3
.<*
.it
.
-------
NATURAL GAS FUEL
NO.
175
I7o
177
178
179
130
181
182
183
18*
0 «5
,L 136
00 187
lad
189
190
191
192
193
19V
195
196
isr
DATE
080571
030571
080571
080571-
080571
030571
080571
080571
080571
080571
080571
040571
102E*2
102672
102572
102572
102572
020173
020173
030173
0^0173
020173
320173
TEST NO
2
3
0.
90.
1<*9.
185.
387.
265.
CO
(PPM)
-0.
25.
53.
0.
-0.
-0.
-a.
-0.
-u.
-0.
-0.
-a.
-a.
-0.
-3.
-3.
-0.
15.
li.
15.
25.
25.
14.
02
(PCT»
3.
3.
2.
2.
3.
2.
2.
3.
2.
2.
;..
.
3.
it.
3.
3.
3.
5.
5.
5.
u.
If.
it.
20
2d
80
90
00
70
60
IP
60
60
50
00
75
21
90
80
90
10
70
25
85
20
90
OPERATING 3ATA Fl
C02
(PCTI
11.
11.
11.
11.
11.
11.
12.
11.
12.
12.
9.
9.
9.
9.
10.
10.
10.
9.
9.
9.
9.
10.
9.
00
<.a
80
60
80
80
00
90
63
»3
52
72
30
63
00
20
20
HO
00
23
30
00
30
LOAO FJEL FL(
(MM) (LOS/SEC
250.
250.
350.
325.
300.
265.
2<*0.
210.
175.
150.
150.
150.
155.
200.
252.
305.
305.
181.
131.
251.
30*,.
351.
3<*9.
27.
27.
37.
35.
37.
23.
25.
23.
19.
17.
17.
17.
13.
2<*.
28.
33.
33.
19.
20.
29.
33.
37.
37.
3
3
5
0
u
6
3
2
f
0
0
0
1
1
3
9
9
8
»
0
0
6
6
628.
586.
(.82.
438.
390.
326.
AIR FLOW AIR FUEL
(LBS/SEC) RATIO
16.9
16.9
16.7
16.9
16.8
16.7
16.7
16.8
16.6
16.6
17.6
17.if
17.3
17. 4
17.3
17.2
17.3
17.B
lo.1
17.9
17.7
17. <»
17.8
299.
296.
313.
1.20.
<*90.
584.
352.
370.
519.
58d.
653.
667.
COMB. AIR
TEMP. (F)
519.
519.
576.
562.
SU8.
528.
513.
It 95.
i«59.
i»59.
93.
<*93.
508.
515.
565.
565 r
FGR FLOW
UBS/SEC)
a.
o.
a.
a.
o.
o.
0.
0.
0.
1*0.
Ml.
ItO.
0.
0.
0.
0.
0.
0.
0.
0.
'o.
0.
-------
NATURAL GAS FUEL
BURNE* CONFIGURATIONS FOR BOILER H5
d
i
»-*
NO
NO. DATE TEST
33 031569 1
b<* 001569 2
69 16
99 002669 17
100 062669 14
101 05?'969 19
102 Oo296S 20
103 Od2969 21
10<» Oo2969 22
105 UB2969 23
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
?
2
2
2
2
2
2
2
2
2
2
?
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
.1
2
2
2
2
<;
2
2
2
2
2
2
2
?
2'
2
2
2
2
2
2
2
2
2
*«
2
?
7
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
3
2
2
2
2
2
2
2
«;
2
2
?
2
2
2
2
2
2
2
2
2
2
2
2
6
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
^
2
7
2
7
2
2
2
2
2
2
2
2
2
2
2
2
2
7
2
2
2
2
2
2
2
2
2
3
2
2
2
2
2
2
7
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
*
2
2
2
2
2
2
2
2
2
2
7
2
2
7
2
2
2
2
7
2
2
2
2
10
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
11
2
2
2
?
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
12
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
7
7
2
7
2
Z
2
2
13
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
H.
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
15
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
16
2
2
2
2
2
2
2
2
2
2
2
2
7
2
2
2
2
2
2
2
2
2
2
17
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
Id
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
19
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
20
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
21
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
22
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
23
2
2
2
2
2
2
7
2
2
7
2
7
2
2
2
2
2
2
2
2
2
2
7
21.
2
2
2
?
2
2
2
2
2
2
2
2
2
2
2
?
2
2
2
2
2
2
2
NOX P
0.
0.
0.
1.
1.
1.
0.
0.
0.
1.
1.
1.
0.
0.
0.
1.
1.
1.
0.
1.
0.
1.
0.
-------
N
O
NATURAL GAS FUEL
NO. DATE TEST 1 2 J
106 082969 2k
107 012071 8L1
108 012071 BL2
109 012071 1A
110 012071 IB
111 012171 eft
112 01217L 23
113 012171 3A
11-* 012271 38
115 012271 «»A
116 012271 i*0
117 312571 5A
118 012571 5B
119 012571 6A
120 012571 61
121 012571 7A
122 012571 78
123 012671 6&
12* 012671 EB
125 01Z671 9A
126 012671 93
127 312671 10A
129 012671 10B
BURNER CONFIGJR ATIONS FOR BOILER H5
7 3 9 10 11 12 11 It 15 16 17 Ib 19 20 21 22 23 2«» NOX PORTS
2
2
?.
Z
z
z
2
i
i
2
2
2
2
5
c
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
1
1
I
7
2
2
2
9
2
2
2
2
2
2
2
2
2
2
2
2
1
1
1
1
2
2
»
c
2
2
2
2
2
2
2
2
2
2
1
1
2
2
2
2
2
2
2
2
1
1
1
1
1
2
2
2
2
2
1
1
2
2
2
2
2
2
2
2
1
2
2
2
2
1
1
2
2
2
1
1
2
2
2
2
2
2
1
1
£
9
1
1
1
1
1
2
2
2
2
2
1
1
2
2
2
2
2
2
1
1
1
2
2
2
2
1
1
2
2
2
2
2
1
1
2
2
1
1
2
2
p
r.
1
1
1
2
2
1
1
2
2
2
2
2
1
1
2
2
1
1
2
2
1
2
2
1
1
2
2
2
2
2
2
2
1
1
2
2
2
2
I
1
1
1
1
2
2
1
1
9
2
2
2
2
1
1
2
2
»
2
1
1
1
2
2
1
1
2
2
2
2
2
2
2
2
2
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
1
i
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
1
1
2
2
2
2
2
2
1
1
2
1
1
2
2
2
2
2
2
2
1
1
2
2
2
2
2
2
1
1
2
2
2
1
1
1
1
Z
2
2
1
1
2
2
2
2
2
2
2
2
2
1
1
2
2
2
2
2
2
2
1
1
2
2
2
2
2
2
2
2
2
2
2
1
1
1
1
2
2
2
2
Z
1
1
Z
Z
1
1
1
1
2
1
1
1
1
1
1
Z
Z
Z
Z
Z
1
1
Z
2
1
t
1
1
2
2
2
2
2
2
2
2
2
2
2
2
1
1
2
2
2
2
2
»
2
1
1
1
1
1
1
2
2
2
2
2
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
1.
1.
0.
0.
1.
0.
1.
0.
1.
0.
1.
0.
1.
0.
0.
1.
0.
1.
0.
1.
-------
NATURAL GAS CUEL
BURNER CONFIGJRATIOMS FOR BOILER H?
NO. DATE TEST
8 9 10 11 12 13
15 16 17 lo 19 20 21 22 23
NOX »ORTS
123
130
131
132
1 J C
133
"*
136
137
138
139
d 1*3
N 1*1
h-^
1-J
!»*»
1«.5
1U6
1«.7
1<»3
1<»9
150
151
020171
320171
U b U 1 X
020271
n? n 771
u C u C r A
020271
?
022571
022571
022571
022571
022571
022571
022571
022671
J22571
022571
022671
022671
022671
0^2671
022671
030871
11A
1 7A
A f, **
12E
1 3S
133
1 <4&
A ^ H
1 ItR
A t n
1
2
3
<.
5
b
11
17
13
1*»
16
15
13
19
20
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
C.
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
if
2
2
2
2
2
2
2
»
2
2
2
»
u
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
C.
2
p
C
2
2
C
2
2
2
2
2
2
2
2
2
2
2
2
2
1
1
2
2
2
C
2
2
2
2
2
2
2
2
2
2
2
2
2
2
?
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
1
2
2
2
1
1
2
2
2
2
1
1
1
1
2
2
1
2
2
1
2
2
C
2
1
1
2
2
2
2
1
1
1
1
2
2
1
2
2
1
2
1
2
2
2
2
1
1
1
1
2
2
1
2
2
1
2
2
1
2
?
?
2
1
1
1
1
2
2
1
2
2
1
1
1
2
2
2
2
2
2
2
?
?
2
?
2
2
2
2
2
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
|
X
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
1
1
2
1
2
1
1
2
2
2
2
1
1
1
1
2
2
1
1
1
1
2
1
2
2
2
2
2
2
2
2
2
2
2
2
2
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
1
2
1
1
1
1
2
2
2
2
1
1
1
1
2
2
1
2
2
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
1
1
?
2
2
2
1
1
1
1
2
2
1
2
2
1
2
1
1
1
2
2
2
2
1
1
1
1
2
2
. 1
2
2
1
0.
1.
0.
1.
1.
0.
0.
1.
1.
0.
1.
1.
0.
1.
0.
1.
0.
1.
-------
NATURAL GAS FUEL BURNER CONFIGJRATIONS FOR BOILER H5
NO. DATE TEST 1 2 3 * 5 o 7 9 9 10 11 12 13 1«» 15 16 17 1? 19 20 21 22 23 2<» NOX 0QRTS
d
1
ts)
C\J
152
153
19*
155
156
157
159
159
160
161
« C 9
IOC
1 ft 1
±Q «J
1 feu
A U *V
i A*;
1O 7
166
167
163
169
170
171
172
173
17 4
030871
030471
030871
030871
032671
032671
OJ2671
072771
072771
072771
f!7 ?7 7 1
U f C. f f L
077771
U f C f i
117^771
U f £ r f J.
072371
072871
072371
072871
030<*71
0301+71
030U71
OdO<*71
080571
2
3
.
5
1
2
3
1
2
3
7
f
1
3
i.
5
2
3
<»
5
1
2
2
2
2
2
2
7_
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
>
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
,2
2
7
c
2
7
2
2
2
2
2
^
2«
2
2
Z
2'
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
2
2
2
2
2
2
2
2
1
1
1
1
2
2
2
1
1
1
4
i.
4
J.
2
2
2
2
2
1
1
1
2
1
1
1
1
2
2
2
1
1
1
1
2
2
2
2
1
1
1
2
1
1
1
1
2
2
2
1
1
1
2
1
1
1
2
1
1
1
2
1
1
1
1
2
2
2
1
1
1
1
1
1
1
2
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
1
1
2
2
2
2
2
1
1
1
1
2
2
2
2
1
2
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
2
2
2
2
2
2
2
2
1
1
1
1
2
2
2
1
1
1
2
1
1
1
2
1
1
1
2
2
2
2
2
2
2
2
1
2
1
1
1
1
1
2
1
1
1
2
1
1
1
1
2
2
2
1
1
1
2
1
1
1
2
1
1
1
2
1
1
1
1
2
2
2
1
1
1
1
1
1
1
2
1
1
1
2
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
n
u
0.
0.
1.
0.
1.
1.
1.
1.
1.
-------
NATURAL GAS FUEL
BURNER CONFIGJRftTIONS FOR BOILER H5
NO
DATE TEST
5 6 7 9 9 10 11 12 13 Hi 15 16 17 16 19 20 21 22 23 2fc NOX PORTS
175
176
177
173
179
180
181
loZ
133
16<»
185
O 186
ro 137
OJ
188
189
190
191
192
t Q T
1?O
19«*
195
196
197
080571
080571
030571
080571
030571
030571
360571
050571
080571
080571
080571
030571
102572
102572
102572
102572
102572
02017J
il 9 n 1 7 T
J c U 1 f . 1
U20173
C20173
020173
020173
2
3
#
5
6
7
8
9
10
11
12
13
1
2
3
(»
t
1
3
l*
t
6
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
I
2
2
2
2
2
£
2
2
2
2
2
2
2
2
2
^
2
2
2
2
2
2
2
2
2
2
?
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
?
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
?
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
1
1
1
1
1
1
1
1
2
2
1
1
1
1
1
1
1
1
1
1
2
1
1
1
1
1
1
1
1
1
2
2
1
1
1
1
1
1
1
1
1
1
2
1
1
1
1
1
1
1
1
1
2
2
1
1
1
1
1
1
1
1
1
1
2
I
1
1
1
1
1
1
1
1
2
2
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
3
Z
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
1
1
1
1
1
1
1
1
2
2
1
1
1
1
1
1
1
1
1
1
2
1
1
1
1
1
1
1
1
1
2
2
1
1
1
1
1
1
1
1
1
1
2
1
1
1
1
1
1
1
1
1
2
2
1
1
1
1
1
1
1
1
1
1
2
1
1
1
1
1
1
1
1
1
2
2
1
1
1
1
1
1
1
1
1
1
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
0.
1.
1.
1.
0.
0.
1.
1.
0.
1.
-------
NATURAL GAS FUEL
OPEPATTNG DATA F0
572.
520.
5<*0.
543.
555.
580.
575.
572.
572.
585.
580.
572.
510.
530.
550.
588.
590.
550.
545.
550.
565.
575.
572.
FGR FLOW
(LBS/SEC)
0.
84.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
82.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
-------
NATURAL GAS FUEL
OPERATING OATft FOR DOILER H6
NO.
221
222
223
22<»
225
226
227
DATE
120271
120271
120271
120271
1202^1
1202'!
071972
TEST NO
(PPMI
2
3
b
5
6
7
1
910.
795.
780.
590.
610.
309.
208.
CO
(PPM)
15.
19.
17.
20.
20.
20.
10.
02
(PCTI
3.00
2.70
2.50
2.50
2.30
2.50
2.11
C02
(PCT)
10.70
10.83
10.83
10.80
10.80
10.60
11.00
LOAO FUEL FLOH
(MM) (LGS/SEC)
350.
350.
350.
350.
350.
350.
348.
37.9
37.9
37.9
37.9
37.9
17.9
38.6
AIR FLOW
(LBS/SEC)
6«*0.
636.
634.
63<*.
631.
63i*.
b<*<*.
Alit FUEL
RATIO
lb.9
16.3
16.7
16.7
16.7
16.7
16.7
COMB. AIR
TE1P.CFI
572.
572.
572.
572.
572.
572.
523.*
FGR FLOW
LBS/SEC)
0.
0.
0.
0.
0.
0.
0.
o
I
Some Air Preheater Baskets Removed
-------
NATURAL OAS FUEL 9URNER CONFIGJRATIONS FOR BOILER H6
NO. JATE TST 1 2 3 «» 5 6 7 8 9 10 11 12 13 1<» 15 16 17 18 19 20 21 22 23 2U NOX PORTS
198
199
200
201
202
203
20 <»
206
207
?n A
cU O
209
III
212
213
21<»
215
216
217
218
219
220
051371
051U71
OsH.71
Oal-*71
051471
051471
031971
lllZl
051971
J- 9 n 7 1
*> t U f \
052071
o ty e ^ *
U?co r 1
052871
C52871
052871
05»2871
081871
081371
031871
061871
081871
120271
1
1
2
3
U
5
1
3
it
2
2
3
i*
5
1
?
3
i*
5
1
2
2
2
?
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
?
2
2
2
2
2
2
2
2
2
2
2
2
2
2
9
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
?
2
2
2
. 2
2
2
2
2
2
2
2
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
2
2
2
2
2
2
2
?
2
2
2
2
2
2
2
2
2
2
2
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
2
2
2
2
2
2
2
2
2
'2
2
2
2
2
2
2
2
1
1
2
2
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
1
1
2
2
1
1
1
1
1
1
1
1
1
2
2
2
y
2
2
2
1
1
2
2
A
1
1
1
1
1
1
1
1
2
7
2
2
2
2
2
1
1
2
2
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
2
2
2
2
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
2
2
?
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
2
2
2
2
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
1
2
1
1
1
1
2
1
1
1
1
1
2
2
2
2
2
2
2
1
2
2
2
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
1
1
2
1
1
1
1
2
1
1
1
1
1
2
2
2
2
2
2
2
1
2
2
2
1
1
1
1
1
1
1
1
1
2
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
-------
NATURAL GftS FUEL
BURNER CONFIGJRATIONS FOR BOILER Hb
NO.
JATE TEST
9 10 11 13 13 it 15 16 17 1
-------
NATURAL bAj F'JtL
NO. Oare T~ST NO
OPERATING 06TA FOR BOILER SI
o
1
228
229
230
231
232
233
23w
111273 1-
1112^C 2 a
Illt7r. 3C
11117-C i.".
11137C -.
11137C 6
111-371 H
61.
-54.'
63.
112."
in.
1TC.
131.
CO 32 C32 LOOT
(PCT) (PCT)
2.93 !".££ 5S.
i.?3 liJ.96 90.
1.53 II. C2 12'1.
1.-.U 11.02
1.^5 11.2C
2. .u 1C. SC ISC.
L FLOW
S/S£CI
6.6
6.5
ir.c
13. f
IE.. 5
19.9
21. t
AIR FLOW
(L8S/SECI
112.
1^3.
165.
?25.
269.
336.
155.
AIR c-JEL
RATIO
16. 9
15. 7
16. t*
16. !»
Ib.t-
15. H
16.5
COIR. AIR
TE1P. «F)
*70.
+ 90.
335.
525.
535.
561.
51.5.
FGR FLOW
(LBS/SEC)
39.
50.
to.
12.
10.
0.
9.
oo
-------
N4TIKAL ".fli F'JJL
N3. DftTi TEST 1 2 3 ", * 5
2?3 111273- 1 =
229 11127. 2-?
23i 11127: 3CT
£31 11137: -d
232 11137" Z '
23
4£x CONCI5J:?4TIOMS FOR TOILER Si
g 9 1C 11 1Z 13 1* 15 16 17 1» 19 20 21 22 23 2* NOX PORTS
u
t-
~2
2
2
2
' 2
"3
2
2
7
t
£
-?-
j
^
2
>
7
2
t
r
\.
i
2
c.
2
?
2"
"^
u
2
2
'
2
-?
L
C
-1
1
1
1
1
G
U
1
1
1
1
1
0
2
2
2
2
2
2
1
1
1~
1
1
1
1
~2
2
2
2
2
'
2
2
2
~?
2
2
2
?
1
1
1
1
1
1
1
2
>
2
2
2
2
?
2
2
2
3
2
2
2
2
2
>--
2
>
2
2
2
2
7
2
2
2
2
r
>
C
r
0
:,
c
c
3
0
3
3
C
-c
""
0
0
J
:
-
t>
0
0
c
0
c
0
c
0
c
0
0
r,
M
J
3
- o
c
o-
0
0
J
fl
0
0
0
c
0
0
0
0
G
0
0
0
0
0
0.
-c.
-0.
0.
0.
0.
0.
SO
-------
MATUR.AL r7fiS FUrL
,0.
235
236
237
2 ib
239
2-C
2*1
2
2ts
CfiTE TuST
IPC.TC 16
ict^rt 2ft
lUd70 3A
1CL97L Vi" '
-in* 70 B "
liu,7t 1
lluWO 3
1IC57C T
lll»7: *
NO CO
(PP1) (??1)
132. ?:.
iTir BI; "
l«»s. 7j.
IT 8. 5C.
306. ?;.
~gr.- --:.- -
bl. -J.
fc j. -J
6 ^ . j
lT9"i *"9 . "
52. ].
32
(3CI1
?.7t
^l.2V
2.15
3.1?
1. 11
7.ZJ-
3.43
2. BG
1.9J
l.'7B~
1.85
"PEitATIMG 3ATA FO* BOILER S2
C02
1 tPCTI
1J.-3
" 11. 40
ib.Lj
11. 1-
' 9.9?
10.15
11.13
rc;^c'"
11. C1
2«»6 Tid^T"*?" BT; =iir~" ITTI n.?7~
2V7
2t?
2«t9
«ibL'
257
2*1
>,S
d»b
11C57P 7
'I1TV7C "fa ~~
11.6'C fcp
IIL'^'O luA
"T?u*71"2
3Zl«»'l 3
j 2 e 1 7 1 >
JJJL4."! 6
'J2L>»71 7
iir. -;.
-ITT.' ~=!T.
U.. -C.
PC. >U.
72. -J.
ir3T" ~77.
2CO. 3?.
157. -;.
377. -0.
i.;:
irnr
1.25
i. ji
1.90
I7'2V
1.05
i, -8
1.67
2.'i
11. 33
1 1. Ta-
ll. 2C
1C. 3s
13.85
rrvrg-
11. 3<*
11..9
lu.9*
LOAD
160.
181.
182.
175.
172.
Ii3.-
<»*.
82.
' 8T.
on.
rcrr
15C.
T.6r~. -
131.
t>u.
ol.
-70;-
119.
182.
rsi."
180.
-U£L PLOh
CL3S/SEC)
2C.,
7.0. it
2C.«
19.9
19. ^
5.C
3.3
7. it
9.8
~ 'g;6
9.5
i3.e
17.*
«.-T
2C.O
7.2
9.3
BW
13.8
13.2
2C.it
20.3
"20.3
1 AIR FLCM AIA FUEL
(L9S/S:3) PATIO
3tU.
377.
3<*0.
' 317.
316.
«5.
91.
151.
' 15TJ .
157.
2TF.
29C.
7237
32*.
121.
161.
177-i
225.
223.
335.
"*'"*"
IS. 8
Io.7
16.3
17 .'I
16.3
~ 17 . 1
17.2
16.5
re-,5
16.5
re.T
16. t
T6.2 -
IS. 3
TO.T-
16.5
-1BY3
16.2
lb.9
16. i»
16.8
CO«B. AIR FOR FLOW
TP^O. (F> (LBS/SEC)
565.
550.
950.
' ' 5^0.
5*2.
522. '
,95.
505.
575.
513.
~5"27Y "
5<*2.
55s.
550.
5 1C.
525.
-525.
525.
510.
56C.
562.
38.
G.
0.
0.
0.
56.
Mt.
25.
J5\
38.
TT;
0.
- -jr.
0.
58.
5?. -
52.
0.
0*
0.
-------
U)
GAi F'JIL
NO. DATE TtST 1 2 3
FOR TOILER 52
6 7 e 9 1C 11 12 13 ic» 15 16 17 18 19 2C 21 22 23 24 NOX PORTS
233
23j
237
233
233
2-.J
2-1
2*2
2U3
2,,
2*5
2i*7
2,9
2«,9
250
2»i
252
2,,
256
25 T
ICC. 57;
i f e t ? .
1.C97;'
ut-j':
iiD?--
us-.?:
lib -,7-
110 -7,.
113^7-
iicsr-.
TlDaTT
IITST!"
11J57-
inroZ11-
11097C
ill 97C '
c«.?l
02.; -.71
b £u t7 i
1A
£A
3A'
«A
T " "
r
2
1
~
*
0
7"
jA
oli -
1CA
2
-
c
/
2
2
' 2
2
- - --
C
-
C
L
-
1.
2
2
Z
2
3
C
Z
2
d
2
£
. j_.
»
2
2
"2
j
~i
C
n~
J
J
2
2-
2
-,-
2
C
9
2
2
2
£
~ 2
2
2
2
-2
2
2
J
j
-
2
"2
2
"2-
2
«d
2
2
2
2
£
2
"T
2
'2
2
t!
2
2
2
i
i
2
2
1
2
2
2
2
2
2
i?
2
~Z
2
~2
2
-------
NATURAL jAS F'J^L OPERATING 3ATA FOR BOILER S2
O
i
OJ
ts*
MO. UAT- TcST NO 33 32 C02 LOAD -JEL FLO«
(po (PCTJ (PCD
-------
NATURAL 'AS FUEL
NO. OATf T£ST 1231*
=!Nr:? CONFISCATIONS p OR 30ILE1? S2
b * 3 9 1C 11 12 13 1U 13 16 17 Id 19 20 21 ?2 23 2<» NOX PO*TS
253
i39
25Q
2:- 1
ZoZ
263
26:.
265
256
O
OJ
C?1'-7I
L?C:>71
u20s71
CZ0571
uZC571
& 2057i.
020571
u£Q371
;2T57T
t
)
"T" '
11
12" '
U
-ru
it
~15
22
L
-"22- '
2 ?.
~ Z 2 "
2 2
C -
2 2
. ..? ,
2 £
j
2~ '2
2 2
2 2
2 2
22
2 2
? ~z
'2
2
c-
2
3 -
C.
2
3
r
2
2
-j
1
>
2
2
1
' I
2
2
" C
1
2
2
2
1
-I
--
"2
2
u
2
2
2
2
2
2
2"
E
1
1
2
2
2
1
I -
2
2
?
2
2
2
2
2
2
'2
2
7
2
- 7
2
2
2
2
2
2
1
1
2
2
2
1
-1
2
2
~2
2
2
2
-2
2
2
"2
2
.. 2 -
2
2
2
2
2
-2-
-
2
2
-2_.
2
2
2
2
2
7
2
2
2
2
2
2 -
2
C
J
C
c
n
r
G-
;
r -
3
2
C '
3
2
C
u "
c
r
0
C
3
"0
C
u
0
0
n
"C
c
-o -
0
0
c
c-
G
- c
C
0
c
0
p
3
C
0
T.
,1
C
C
0
3
n
0
0
c
- -J
0
0
-3 -
a
0
c
3
3
c-
0
0
0
0
3
C
a
0
0
0.
0.
0.
0.
0.
0.
0.
0.
-------
_JH SULFoK OIL FU£. OPERATING OATA FOR 30ILER HI
3. DATE: TcST NO
1
2
3
«
~
3
J
e
9
1C
11
12
«c22M 1
,n,«»2ri 2
Jo227l 3
Jfc227l v
Joc.3/1 1
,.b£37i 2
ufc£.J7l 5
-62J71 u
111671 1
111971 1
J127T3 1
ult7T3 2
27?.
1:8-3.
333.
230.
Itfe.
153.
177.
13£.
211.
2iC.
2iv.
ZC>9.
1..
23.
15.
15.
22.
3).
li.
13.
25.
25.
*j .
-t j .
3.35
1. >j
1.70
1. 8d
3.3d
2.63
2.05
2.7J
3.:ta
,.b3
>..9J
t. 23
1 1. 3j
It. 13
It.tJ
1-..28
13.31
13.5D
13.60
13tbC
13. Cu
12. Ui
11. 9C
13. 4L
90.
1-.0.
190.
230.
9d.
12:.
18C.
225.
232.
217.
227.
22*.
13.1
19.4
26.1.
31.9
13.6
18.9
25.6
31.1
31.7
30.o
30.6
29.7
19f.
?<33.
390.
*ol.
2.%.
273.
378.
V&8.
*71.
V7*.
<*78.
»i3.
14.9
1-.I.
14. i»
Kt.it
13. C
14.7
1^.3
It. 7
1>*.9
15.5
15.7
15.2
323.
5*5.
574.
598.
513.
5*2.
3&S.
59».
a3C.
;9G.
597.
595.
151.
1*1.
122.
137.
137.
129.
126.
137.
62.
102.
151.
157.
-------
uOrf :>U_FU* (Jit FUE. 3URNER
NO.
1
2
i
#
j
o
7
3
3
iC
11
d "
OJ
Ul
uATi
u u227 1
Ct.2271
Lo2 ^7 1
- L 2^/ 1
Co237i
OL2o71
Uc237l
-t237l
Ilio71
111971
Q12773
blei'7 J
T£31
i
2
j
.
1
e.
j
-
i
i
i
2
r i
?
£.
2
2
2
c
2
i.
2
2
C
2
2
2
2
2
2
2
2
2
2
?
2
2
2
3
2
2
2
2
2
2
a
9
2
2
2
2
^
2
2
2
2
2
2
2
2
2
2
2
2
:
2
2
2
2
?
2
2
2
2
2
C
2
o
2
2
2
2
1
1
1
1
1
1
1
1
7
2
2
2
2
1
1
1
1
1
1
1
1
8
2
2
2
2
2
2
2
2
2
2
2
2
CONFISCATIONS FO
9
2
2
2
2
1
1
1
1
1
1
1
1
1C
2
2
2
2
2
2
2
2
2
2
2
2
11
2
2
2
2
2
2
7
2
2
2
2
2
1?
2
t
2
2
1
1
1
1
1
1
1
1
13
u
C
u
u
C
C
C
0
u
u
L
C
l
-------
LOM SULFUR OIL FUiL
OPERATING DATA FOR BOILER HZ
d
I
13 04*12/1 1 ~
II, dt!271 2
15 Ot!271~5 ~"
16 utl371 1
17 0»7l *
26 i)tit71 &
21 Dtl<*7l 6 "
22 01*1571 7
23 Otl571 8
2t 01,1671 10
25 Ot 1671 "11
2b Otl.671 12
27 3t267l 1
28 Jt*>71 2
29 j<*267l 3~
30 a i»2&71 «
31 0*2671 5
32 Ot2671 6
33 u 1,2671 7
3* Ot2771 1
35 fll»2771 ?
333;
273.
23t.
178.
"2*6.
229.
28*.
295.
2o2.
210.
21 2.-
228.
"237.
231.
166.
165.
162.
162.
2TJU.
186.
309.
172.
16 r.
~"U.~
30.
~3j-.-
30.
"ti.
?0.
»0.
»o.
25.
30.
i».
50.
"60.
63.
S5.
35.
*o.
35.
-3TJ.
35.
33.
35.
*D.
l.'TU
2.1; a
».S3
2.73
3. oli"
2.63
2.90
2.12
2.93
2.50
2.98
3.05
3.10"
».as
3i IT
3.10
- j;oa"
3.00
3.05"
2.75
l.TO~
*.05
3.10"-
13.20
12.90
11. 713
13.50
13.C'0
13.95
16.03
15.98
is. aa
13.68
13.30
12. tO
12.50
12.00
12. 80
L2.80
12". 90
12.90
rs.oc
13.10
"IT; eo
12.53
T3.00
2tO.
165.
89.
230.
227.
232.
2*0.
2*0.
232.
2tO.
23D.
228.
232.
222.
195.
200.
ZOO'.
200.
ZOO.
200.
2 CO".
90.
138T.
31.9 ~
21. t
IZY5" ~
30.6
30~.0
31.1
31.9
31.7.
31.1
31.9
31.1
30.6
31.1"
29.7
' 26. li
26. t
26.7
26. t
26. t
26. t
26; 7 -
13.1
ia.^9
" t61.
311.
19I». -
tso.
»52~.
t57.
t&8.
t5B.
»57.~
t68.
*51.
t56.
t65. "
*55.
- "393." '
393.
r9T.~-
392.
" 39T.
390.
TVki ' "
199.
rffir
"It.t
It. 5
15.5
It. 7
~15.l
It. 7
It .7
It.t
"It. 7
It. 7
I*-;B
It. 9
It. 9
15.3
It. 9
It. 9
-IV. 9
It. 9
1*.?
It. 8
IF. »
15.3
"it. 9
~58ir.
568.
53W
582.
578T
579.
578~.
581.
5T5.
582.
578.
57 1.
JT3".
569.
56*.
966.
565.
56S.
' 569".
56S.
5T3.
532.
F«.
?8i"
153.
IB1.
129.
- 1Z9.
133.
-IZ9.
129.
ITS.
122.
12Zi
lit.
IK.~
110.
It7.
It3.
It7.~
It7.
151.
It9.
135;
181.
- -I57T
-------
_DH aULFU*. OIL FL'E.
OPERATING DATA FOR BOILEK H2
JO. iJATi T;S7 NO CD 3?
(FP1) (P3*) (PCT)
33U*27i*l 3 1 ? c . 35. 3.J3
37 J-,2771 +
39 ,3*23/1 1
39 u-,2371 2
.0 v-rtbil 3
<! <: dTl -r
*2 111171 1
»3 111171 2
*» lllli'l 3
0 1111/1 -
O
-J
18t. Vj. 2.
2;1. 33. 3.
£ui. 3J. 2.
'7... 3j. 1.
31*. 5u. 1.
193. 23. 3.
219. 25. 3.
ifc-. 23. 3.
22>. . 23. ;>.
7i
3a
Id
aJ
08
oO
dtl
9U
.-0
C32 LOAD -UEL FLOH
(PCT) (MM) (LdS/SEC)
13. 2C 18?. 2*. 2
13.
13.
13.
1*.
lu.
13.
13.
13.
12.
66 2*3.
as 90.
9> ItC.
3C 19C.
23 2«,-.
6? 19*.
6. 23u.
73 1«.7.
20 63.
12.
12.
2C.
2&.
31.
25.
31.
20.
it.
2
2
:
0
9
C
1
3
3
AIR FLO* AIR FUEL
(L3S/SZO PATIO
359. 1*.9
*75. 1*
18^. 15
291. 1*
361. 1*
!»62. 1*
376. 15
*68. 1?
3 u5. 15
2b2. 15
.7
.0
. y
m ^
m 5
. u
.1
.1
.3
C313. AIR
THM°. (F)
373.
j3T.
551.
569
579.
56i.
57b.
352.
53*.
FG1? FLOM
(L8S/SEC)
1*5.
98.
169.
161.
1*1.
102.
1*5.
65.
Io9.
lol.
-------
.OH SOuF'Ji
NO. uATEl
13 I.-.1271
it U-.1271
15 u-1271
lo b»-loi'l
17
13
19
20
21
22
23
d 2*
w *5
00
27
23
29
3j
31
32
3J
3-,
3>
b-r!37i
0*1371
Q-»l,71
utl-*/.
5-.1-.71
I...1&71
LHl371
-H1671
C-rl671
b-ib7i
b-,2671
t=.2c/l
ut3o7l
u.2o7l
C-.2671
C-r2o71
b-r2c71
is277l
b-2771
t Olu
TEST
1
1
2
3
-
j
3
7
&
lu
11
12
1
2
3
"r
j
6
7
1
2
FU:
*
2
j
2
2
2
2
2
2
2
2
1
2
2
2
2
2
£
2
2
2
2
2
2
2
J
b
2
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
?
2
3
2
J
J
2
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
3
2
*
2
0
u
2
2
2
2
£
2
2
2
1
2
2
2
2
2
2
2
2
2
2
2
2
C
1
2
2
2
1
1
1
2
2
1
2
2
2
2
i
1
2
2
2
t>
2
2
3
1
2
2
2
2
2
1
2
2
1
2
1
1
1
1
1
1
2
1
1
3U«
7
2
2
1)
1
2
2
2
2
2
1
2
2
1
2
1
1
1
1
1
1
2
1
1
d
2
2
b
2
1
2
2
2
1
1
1
2
2
i
2
2
2
2
1
1
2
?
2
CO
4
2
2
1
2
1
1
2
2
2
2
2
2
1
1
1
1
1
2
2
2
1
1
!MrI
1(1
2
2
2
2
2
2
2
1
2
2
1
1
j
2
2
2
2
2
2
2
2
2
2
f U 9
11
2
2
2
2
2
2
1
2
£
1
1
2
2
2
2
2
2
2
2
2
2
2
ATI
12
2
2
2
1
2
i
1
2
2
2
2
2
2
1
i
1
1
1
2
2
2
1
1
13
C
0
ll
J
G
C
C
b
C
L
u
j
G
0
G
S
3
C
0
i
.-
q
c
FOR BOILEk
1C 19 Ib 17
G a B a
COOP
C 0 3 b
o o a o
a
G
u
a
B
0
Q
0
C
0
U
0
0
0
b
c
c
c
0
3
B
0
0
0
C
0
3
0
0
b
G
0
C
C
3
C
il
C
B
B
C
G
C
3
0
a
0
a
a
0
a
0
3
0
G
B
0
c
(l
'
L
L
c
r
0
C
0
e
c
i
i
c
c
c
c
c
H2
li
J
u
j
i
a
0
a
G
0
3
J
a
a
i
c
3
b
J
0
3
3
a
a
19
c
b
w
J
3
5
2
r
J
3
j
L
3
j
C
,,
a
L
J
r
20
0
0
B
0
Q
B
a
0
0
c
0
0
9
0
I
0
0
c
0
0
0
0
0
21
0
0
c
g
0
B
0
0
0
0
c
0
J
3
G
a
j
a
0
c
j
j
8
22
3
0
Q
0
0
0
0
a
B
B
0
d
j
3
a
0
B
c
0
0
D
C
23
C
C
0
0
9
G
a
0
u
i)
u
D
0
d
J
B
0
a
u
c
0
0
0
2* N3K P
B B.
a c.
0 B.
0 0.
0
G
B
3
B
3
U
3
0
3
0
0
0
a
3
Q
0
0
0
B.
0.
G.
0.
0.
B.
B.
B.
C.
3.
0.
B.
0.
a.
c.
0.
a.
0.
0.
-------
a
i
10
NO
LOW jULFJk OIL FU;_ dll«NET.
J'j
37
33
39
-J
+ 1
-2
..J
»*
« 2
Ci.2, -71 3 2
0^2771 ..
u-,2471 i.
J-,2671 2
3-,2o7l 3
L-,2371 »
111171 1
11117: 2
111171 3
1111/1 -
2
b.
2
2
2
2
t
2
?
2
2
2
2
2
2
2
2
7
^
2
2
2
2
2
p
2
2
2
2
2
2
2
2
2
2
2
2
2
2
<:
2
2
2
i.
2
?
1
1
2
2
2
2
1
1
1
1
1
2
2
2
2
1
1
1
2
2
2
2
2
2
2
2
2
COMF1
9« r
1
1
2
i.
2
2
1
1
1
J. W
2
2
2
2
2
2
2
2
2
2
jLKATIONS
11 1 3 1 t
L X
2
2
2
2
2
c
2
>
2
2
A C.
1
1
2
2
2
2
1
1
1
1
* «»
G
C
C
C
J
u
L
0
G
C
FOR 3
1 u < ;
x »
0
G
C
C
b
0
0
0
G
u
A *
G
C
b
G
0
0
0
3
3
OIL
16
^ W
3
G
a
0
C
a
j
3
G
i r
A f
C
r
C
C
:
c
c
C
C
r
H2
13
A U
3
0
G
0
0
J
L
0
C
3
19
j
J
b
r
T
J
'
J
J
u
20
0
0
0
G
C
C
G
0
l
21
G
:
u
u
C
u
G
j
n
22
C
3
0
3
3
j
-
3
J
3
23
G
C
G
b
0
e
j
j
0
2I» NIK PD9TS
0
0
J
0
0
3
0
u
0
0
0.
G.
0.
J.
C.
0.
3.
G.
0.
0.
-------
bULFJR OIL
DATA FOR BOILED 43
d
*>
o
NO. £AT£
T^ST ND CD
(PP1I (PP1)
*5 u71
m2.
Iu2.
17:.
170.
225.
22o.
21..
21.4
23.9
23.6
31.9
31.9
FLOM
135.
3o5.
359.
<*35.
AIR FUEL
RATIO
15. &
15. &
15.3
15.2
19.2
15.1
COM3. AIR
TEMP. IF)
498.
300.
505.
505.
J18.
319.
FG* FLOW
(LBS/SEC)
SI.
59.
5*.
55.
27.
27.
-------
a
i
SULFUR OIL FUI. 3UkN£9 CONFISCATIONS FOR 30ILER M3
N3. LATt TcST 1 2 3 * a 6 7 8 9 10 11 12 13 It 13 16 17 IB 19 20 21 22 23 2l» NOX PO^TS
40
,7
»3
49
&0
si
0/1+71 o
671,71 -.
0/1,71 ,
L/l-71 &
C?l471 7
C/1,71 1
L
b
2
2
£
i
J
3
>
2
2
2
L
J
j
U
i
>
I
E
;
i
2
2
2
2
2
i
2
t
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
<:
2
«!
c.
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
0
G
C
^
L
C
0
0
c
\.
0
c
c
L
0
U
u
0
c
:
a
0
L
c
c
-
u
w
c
c
3
0
0
W
3
U
n
L
*
-
C
J
c
9
U
G
C
0
J
C
3
J
0
C
3
0
a
j
a
0
0
0
a
u
C
a
0
0
j
0
3
0
1.
0.
0.
1.
1.
a.
-------
.Out bJLFLU OIL
OATA FOR BOILER HU
10, DAT£ TC.ST NO
(PP1)
52 J7uS7l 1 Ibt.
93 -. IJ lu. 05
-3. *.8J 12. IS
-J. >*.;> 12.30
-I. «.*5 12. 2i
>.. T.o? 12. 1Q
-J. -r.ti 12.30
-j. 3.95 12.75
kOAD
(MH)
93.
?j.
142.
1-3.
192.
192.
235.
235.
ru£t FLOH
(Las/SEO
12.5
12. J.
20.0
2C.O
26.7
2b.7
11.9
31.4
AIR FJEL
RATI'
2CQ.
206.
311.
IO
16.0
16.5
15.&
15.it
la.5
15.1*
IS.2
01).
*91.
532.
502.
515.
515.
526.
FG3 FL3H
(LaS/SECl
120.
120.
82.
32.
i»7.
5.
3.
ro
-------
I.J-* bULF'JK Oil.
to.
52
5 J
5*
i,
>&
a7
7 3
j3
DAT£ T»_ii
07j971 1
.7..971 i
o/u*ri.
o/-,.9ri ,
luu J7l ?
C/J371 j
d?u9n -
-70971 *
. FU£.
r i 2
L 3
*. *
W J
J
2
2
2
2
2
2
2
2
0
2
2
2
2
2
2
2
2
T
-j
»
2
2
2
2
2
2
8
C
0
2
2
2
2
2
g
9
2
-------
bJLF')fc Olu FJ£.
OPERATING DATA ?OR BOILER H5
40.
Ot
61
62
63
0+
65
bi
o7
6b
a9
0 7i
* 71
*" 72
n
7,
*5
7-,
77
76
7j
.571
TeST
1A
14ft
1 J
2A
til
3A
3 A
b
j
1
2
3
,
7
b
9
1.
11
12
5A
9a
3 **
NO
(PP1I
2i...
226.
27e.
262.
32*.
21*.
13 Si.
li.2.
c7u.
^..
262.
2"*..
£7L.
192.
222.
132.
ICo.
12k.
15 b.
32i-.
3CT1
2u. -C.Cu
-c. -j.-:j
2W. 3. Oil
21. 3.7J
2j. 3.73
2j. 3.oJ
2?. ». 1C
23. 3. 3t)
2>. 2.vu
?.. l.bO
5 j. 1. e5
»J. 2.53
1 u . » . t b
» u. 3. w j
3 j. 2 . SJ
5J. i..7j
7j. 2.20
6J. 2. 50
i;. *. u
5u. 3 . -»5
5 j. 2. '.j
»=.. 2.ec
t5. 2.60
C02
-li. LC
-C.OJ
13. S3
13. CO
12. 2t
12. bj
12.70
12. 6C
l+.b.
1 + .6C
13.90
13. 1C
13. OC
13.2C
13. 6C
13. fC
13.3G
12. 6U
13. GC
13. s:
13.50
13.33
LOAO FUEL FLOM
(MM) (L4S/SEC)
25C.
2&C.
35J.
353.
3»C.
3&0.
15S .
1&3.
3tS.
350.
350.
152.
352.
J&l.
15,7.
lb-».
15*.
If*.
194.
ItJ.
3^3.
3&3.
3»3.
32.5
31.9
91.
AIR F'JEL
RATIO
16.3
16.5
1S.1
15.1
15.2
15.1
15.3
IS. 2
11*. 3
1H.3
1-..S
15.2
15.3
1<».7
l*.*.3
i<*. a
15.3
15.;
1«.6
1<*. 8
i«*.r
COM3. AIR
TE1P. (F»
320.
323.
363.
565.
5o5.
363.
515.
510.
363.
363.
355.
56-3.
563.
5 o5.
505.
»05.
310.
900.
503.
308.
565.
373.
363.
FGR FLOM
ILBS/SECI
0.
0.
0.
0.
3.
0.
111.
lit.
C.
0.
0.
0.
0.
c.
121.
120.
119.
nr.
115.
117.
0.
q.
0.
-------
LOrf SULFUR OIL FUE.
SO. JATi T£5T NO
(PP-O
63 330571 60 3t2.
0
1
Ul
6-*
65
63
87
08
89
9J
91
92
93
95
95
97
93
.31771
J31771
03*771
u3l77l
-31771
w32371
^32371
J32371
J32371
J12772
1LJ172
113172
1C3172
10317*
11! 31 72
1
2
3
4
I-
1
2
3
,
1
1
2
3
(*
5
1£*.
lo-..
18L.
21-.
26;.
2,3.
3c r.
36o.
wl9.
239.
213.
196.
221.
2-a.
CO
(PPM)
45.
35.
35.
Uj.
-I..
-J.
-3.
-j .
-C.
1G.
1C.
13.
U.
32
("CD
3.QJ
i.10
3.-0
2 .90
2.9u
2.ao
3.30
2.70
2.70
3.10
3.0J
9.13
». " f
<~.Sy
t.15
-.. «t a
OPERfiTISS
C02 LOS3 :
(PCD (M4) i
13.20 3t>6.
13.
13.
13.
13.
12.
13.
13.
13.
13.
11.
12.
12.
12.
12.
3C
00
40
i>3
90
55
50
2C
1C
60
3C
3L
85
60
1E>J.
198.
3C3.
352.
15C.
3CO.
JEO.
3&0.
350.
195.
its.
.255.
JC2.
3C2.
DATA FOR
rU;L FLOW
tLBS/SEC)
20.
25.
31.
37.
tJ.
21.
37.
<*<».
Hi*.
<*3.
26.
21.
31.
36.
36.
8
7
i*
5
9
i*
5
i*
W
6
I*
1
9
1
1
FOR BOILER H5
AIR FLOW
660.
311.
<»00.
465.
555.
320.
552.
655.
661.
i*16.
551.
555.
AIR PJEL
RATIO
1<*.3
!*.?
15. u
l-».8
1-..3
15.3
14.7
1-..9
14.9
la.8
15.4
15.4
15.3
15.4
CO'13. AIR
TEMP. CF»
563.
503.
323.
523.
538.
565 .
510.
9O9.
565.
563.
503.
500.
515.
328.
539.
FGR FLOW
(L3S/SZC)
0.
116.
1L<*.
7k.
0.
0.
115.
D.
n
U .
0.
0.
82.
110.
0.
0.
0.
-------
.OH SULFJR OIL F\)=.
NO. UATE TEST 1 2 3
3vJRN£ft CONFISCATIONS FOR BOILER »5
5 6 7 a 9 10 11 1Z 13 1* 15 16 17 1« 19 20 21 22 23
NOX PO*TS
ia
"
52
33
3<»
b->
= 6
37
3d
39
7J
71
72
73
7i*
79
76
77
76
9 Q
f ">
3u
31
82
022371
i 22 371
C 2237 1
022371
u 22*71
022-71
L22../-1
ii22-»7l
U 30171
C 30171
I3i,171
031)171
ooil71
C.3J171
1.3J271
u3Ct7l
0 j i,271
L 00271
,i 30271
n "» i 9 T 1
U <5U C.I 1
U 30 ? 7 1
u 3^ } 7 1
u
1A
1AA
IS
2A
26
3A
2b
fa
5
1
i
3
-
7
d
9
10
11
i 9
it
5 A
LA
2
1
1
£
2
£
2
2
2
2
2
2
i
2
2
2
2
2
2
2
2
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
?
2
7
2
2
1
1
1
2
2
2
2
2
2
2
7
2
2
?
2
2
2
2
2
2
2
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
1
1
»
2
2
1
*
1
2
2
2
2
2
2
2
2
2
2
2
2
1
1
1
1
2
2
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
1
1
1
1
2
2
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
1
1
1
1
2
2
1
1
1
1
1
2
2
2
2
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
£
2
2
2
2
2
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
1
1
2
2
2
2
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
2
2
2
2
2
2
2
2
?
2
2
2
2
2
2
2
2
2
2
2
2
1
1
1
2
2
2
2
2
2
1
1
1
1
i
2
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
1
1
1
1
2
2
1
1
1
1
1
1
1
1
1
2
2
2
2
2
?
2
2
2
2
2
2
2
2
2
2
2
2
1
1
1
2
2
2
2
2
2
1
1
1
1
2
2
1
1
1
2
2
2
2
2
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
1
1
2
2
2
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
1
1
2
2
2
1
1
1
1
2
2
1
1
1
1
2
2
1
1
1
2
2
2
2
2
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
1.
0.
1.
6.
1.
0.
1.
1.
0.
0.
1.
1.
b.
1.
0.
0.
1.
1.
0.
0.
1.
0.
-------
VJ'-.-uft OIL FUE- DJRNE1 CONFISCATIONS FOR 30lL£R H5
NO. CftTc TCsT i. 2 a « <-. b 7 9 9 1C 11 12 13 !» li 16 17 18 19 2C 21 22 23 2«» NOX PORTS
9*
a
8 J
80
87
id
39
90
91
92
93
o 9*
i. 9j
-4 ,
9 j
97
», 3U a f 1
"1 / T 1
b J lr / i
031771
U51771
ljol/71
LJ2371
jJ2J7l
i.3237:
.52371
w 12 / 7 2
103172
1.317.'
1C3172
DO
1
c.
j
.
=
1
2
j
T
*
I
2
-
e.
2
2
t
2
2
2
2
2
2
2
2
;!
5
2
2
2
2
2
2
?
*
a
2
2
7
7
7
2
2
2
2
2
7
>
2
2
?
2
2
2
?
2
2
2
4
1
A
.
2
2
2
*
1
1
1
1
2
1
1
1
1
2
2
2
2
1
1
2
1
1
1
1
2
2
2
2
1
1
i.
1
2
1
1
1
1
2
2
2
2
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
£
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
2
2
2
2
2
2
2
2
2
y
2
2
2
2
2
2
2
?
2
2
2
2
a
C
2
2
2
1
I
1
1
1
2
2
2
2
i
A
1
1
1
1
1
i
&
1
1
1
2
2
2
2
1
1
1
2
1
1
1
i
2
2
2
2
i
A
1
1
1
1
i
A
1
1
1
2
2
2
2
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1.
1.
j. .
1.
1.
1.
1.
1.
1.
0.
1.
G.
i m
X
1.
0.
-------
00
LO* i.'J,.FJi OIL FUI-
T.,7
CftTA FOrt BOILiR Hb
EC)
49 Iil570
lbj 12iS,~L
1C1 I2l!>ru
Iu2 lcli7C
1U 12A57J
1.* iiilpfu
luD 1 2 1 a 7 u
1 CO 121^' b
1L7 J32*?l
IGo J o 2 * i* 1
1 C3 bd^*/l
11. u s2t "1
111 -ottTi
112 do2*."l
113 Jd.2f.ri
11* uo<:t/l
113 i9i.l7l
1-t j9tl71
11,' i39tl71
lid 'j9tl71
119 09tl7l
12L -71972
1
c.
3
*
7
b
7
a
1
14
2
2A
3
J H
*
,A
1
2
3
w
5
1
32-.
256.
3i,,.
3co.
*71.
.C2.
-tOl.
35-.
22o.
2?b.
2bb.
£23.
2dc.
317.
382.
3ij.
190.
Clb.
t6i..
1&=.
IbC.
£1^.
-j.
-3.
" J
-J.
-L'.
u
~ J .
-J.
-3.
lj.
13.
li.
Ib.
2J.
14.
li.
13.
il.
2C.
23.
20.
It,.
i.-J
3.53
2.7-.
2.3J
2.4d
3. C 5
2. C3
2.25
J.C.S
J.il
3.UQ
3.C5
J.^3
5.22
2.75
3. .3
2.23
2.23
2.*1
2.2J
2.AQ
3.U5
12.65
13. IS
13.68
13.7.-
13.29
13.*-:
It. 1C
It. 1C
13. 3C
1*.G2
1*.8L
13. 2C
13.35
1*.->C
It. 3D
13. *0
l».6u
It. 83
It. HO
13.60
13. 3C
13.15
1JC.
150.
253.
-------
LJrf jULFJ* OIL F'JI.
NO. L.ATE T£3T 1 2
E* CON-I3U^ATIONS FOR 90ILEP H5
> 7 3 9 1J 11 12 13 1. 13 16 17 Id 19 20 21 22 23 2* NOX PO*TS
99 I21?7u
1:0 i2i5'.
Ul lc.l;7.
ii2 i2i>r:
l.J 121t7j
U, 12137C
1C 3 Itl57j
10 1> 12137J
107 ic2«7l
1" a ' bi "1
U9 cc2-»n
O llJ :o2«71
i
fc- 111 \.i2til
vO
112 Dc2wl
113 0«!2-.1'1
HH C52-.71
11 3 b^ul?*.
llo C90171
117 U9I.171
lid ii>»Gl71
lli U90171
l2u u/'1972
1
2
3
,
*
o
7
d
1
e.
2A
3
JA
<-
.A
i
i
j
,
;
1
a
e.
2
c
2
2
2
2
c
£
2
^
2
<:
e.
2
2
£
2
2
2
2
2
2
2
2
2
2
2
2
i
2
2
2
2
2
2
?
2
2
2
2
2
2
2
2
2
2
2
i
2
2
2
2
2
2
2
-
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
i
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
£
2
2
2
2
2
2
2
2
2
c
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
1
1
1
1
1
2
2
7
2
2
2
2
2
2
2
2
2
2
2
2
2
1
1
1
1
1
1
2
2
i
2
2
2
2
2
2
2
2
2
2
2
2
2
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
*»
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
7
2
2
2
2
2
2
2
2
2
?
2
I
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
i
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
1
1
1
1
1
1
3.
1.
1.
3.
9.
1.
3.
1.
1.
C.
0.
1.
1.
a.
c.
i.
i.
i.
i.
i.
i.
i.
-------
LOU SULF'JR OIL FU£.
OPIATING DATA FOR BOILER SI
0
Ul
o
^o. a AT-:
121 11177C
" rznnTTo
123 11177C
rzt riTBfrc
125 Illd72
TZE. 112C7C
127 112070
123 112070
129 112370
i3D~rrrj7D
131 1123'u
"i~32 1TIT7?
133 il2&"0
135 11257C
135" 11257u
I3t 121170
" T3g~ 1TI "S7T
Tr.S
1
z
3
5
~T~
6
"9"
Hi
- 5Z~
A3
"A% '
82
C2A
DIB
Z
r NO
159.
195.
174.
"Tl'9.
ltd.
I-.C.
2U.
~ 1*2."
165.
1*2.
269.
129.
163.
C3 32
(PCT)
-3. 3.13
~0 . 2. ?D
-;. 2.29
-;. 3.9..
-0. 3.35
-C. i.l
C32
CPCT)
12.76
~T3.'72'
13.69
12.bC
1Z. RATIO
15. C
1ft. 5
!<*.:»
15.2
15 .~D
15. B
r*t.9 '
I*.J~~
1<».9
1..9
T5.C
1<*.3
~ " rif.a
15.2
la. 5
L C013. AIR
T£1P.m
522.
~" 33Bi
.75.
512.
51?.
' -'aTSr
518.
53i.
5T5.
53*.
528.
' 537. '
FGR FLOW
(LB5/S£C>
67.
16.
61.
. 71>-
S3.
35.
52.
- *T; -
*'
F5.~
<»T.
*7.
,1.
IT.
-------
L3H SULF'J* OIL FU:_
«URNEB
FOR 301',-R 51
NO. cftTE TtiT 1 2
3 9 13 11 12 13 !«. 15 16 17 18 19 20 ?1 22 23 2i» NOK PORTS
121 11177:
122 11177:
123 11'177'C
12* 11137:
i2o iizor:
127 112i7j
126 112C7,
12-9 112 T7r
13J 11237j
131 1I23TT
132 112371
133 112T7TT
13, 1123?,
r3"5 llZyfM
13a H257i.
131 110*71
i "
2
3-
»
3"
r
"o"
9
-rrr
i2
~S3
Aw
"37
=13
OZl
* 1
2
?.
2
_. -j-
-
"T
c-
2
- T
2
r
£
_. -£..
2
H -j-
J 2
r 2"
2
?
?
z
n
j
T
2
»-
2
2
2
1
2
?
7
TT
2
2
U
^^~
j
-5T
2
7
2
2
--?-
2
~~2 *
2
2"
2
' 2
2
Z
C
T "
0
a
2
"2 ~
2
2'
2
-r-
2
-?
2
2
2
~7"
2
-2 -
L
2
2
-2
2
?-
2
Z"
2
r""
2
£
2
I
1
r
Q
1
r
2
r
i
T
1
T
1
T~
1
T"
1
T
1
"1-
<:
"T
~~r
2
~1-
1
~1'
1
1
1
r
i
2 ~r ~
2 1
0 1
~C 1
? i
2 1
2 2
-? -1-
2 1
? 1"
2 1
"2 1
2 1
~z r~ "
2 1
2 1
2" ' Z
2 2
7" "2
2 2
2 "2
2 ?
2 2
2 ?
2~ 2
2 2
2 2
2 2
Z~"Z
2 2
2 2
2 2
1
1
T
1
"1
A.
~l
2
~~"I
1
r
i
i
i
i
i
- 2-~
2
?~
2
2 "
2
' 7
1
"T
2
?-
2
. 7_
2
2
2
2
2
2
2'
2
2
?-
1
2-
2
2 -
2
"2
2
2'~
2
2
2
2
"2
2
2"
2
2
1
2
2
"2
2
2
2
"2"
2
2
- 2 "
2
2 -
2
2
2
2~
1
~ 2 ~
2
2-
2
2
2
2
2
C
n
~TT
r
C
-r
u
T-
c
c-
Q
w
c
c
~c~
(.
u
0
0
3
0
J
'I?
1
- o
3
-o-
0
--u
3
~a-
c
"U"
3
t.
;
r
j
u
c
3 -
3
U
J
-ff-
0
c
Q
0
n
c
T> "
g
r -
0
0
c~
a
-o-
9
o- -
0
"V '
c
-o
c
n
0
^
3
C
0
U
3
TJ
u
C
C
r
(!
u
J
T
G
"C
0
U
0
c
G
"0"
3
0
- o-
C
o
3
-TJ
0
~ "3"
Q
0""
3
D
0
tT
0
T
0
0"
0
-D-
0
~o-
0
0"
0
'0
0
"0" "
c
o
0
3-
0
u
0
TJ
0
0
0
tj
0
0
0
9
0
9"
0
- or
c.
c.
c.
0.
- o.
0.
c.
0.
--IT.
3.
ITi
C.
0.
0"."
0.
-------
LOrf jULF'JR OIL FU=:L
0«»>:*ATIMG DATA FOP 90IL-R 52
so.
139
1-1T
"14?"
1-.3
1-5
DAT:.
11JQ7C
ll"JC^u
11307C"
Ullf.
IZul'C
TEST NO
(PPI*
1 12o.
2 136.
3 lt>6.
I. ' ' 152.
5 132.
b 1*-G .
7 153.
It6 3* 7.
1 1&5.
2 191.
J S 1 '! .
* 179.
9 3D .
6 31b .
1 1 J8.
2 IS*.
J lists.
i, 232.
C3
-J.
~ -3. '
- j .
-T".
-:.
-3.
"37."
- j .
TC.
3J.
ij.
--.
" 27.
~ J
-3.
- j.
* 3.
.U.
33.
32
("CTJ
..S3
I*. 35
». Ji
3. S3
2.63
3.d3
*. t J
_'.r.
«. 6T
1.72
i. J J
1.79
3. *a
<.. 76
3.^1
r. *S
2.1.3
~~T. 91"
i. 33
1. fa
3.09
C02
i (»CT>
11.76
17.13"
i2.(.-e
12. 2B
12. 3e
I J. Ct
12.13
1 ii . b r*
IN,. 79
1^. ^2
1..9C
1J.1C"-
It. 71
1^. J^
12. &t
12. 58
13. »o
1C.,*
1?« 3^
11.52
1«.!»J
13.39
LOAD CU£L FLOW AI
(HM) (L8S/SEC) (L
vl.
81.
S^.
IcO.
179.
76
77.
i£i .
119.
131.
iei.
1< 'S.
178.
179.
119.
119.
12C.
18J.
5.3
8.F-
11.1
TT."3
15.8
IB. P' '
22.8
11'. B
1C. 8
lb.1
15.6
23.1
22.3
23.1
22.8
22.8
15.8
15.8
15. b
22.8
:% FLOW
98.
133.
171.
205.
237.
279.
167.
157.
"T*97
22U.
3*»9.
323.
<"»».
353.
356.
231.
35 u.
251.
339.
AIR FJEL C0^3. AIR FGR FLOW
) RATIO TE1P. .0
15.5
15 .3
15.1
ln.6
__
15.9
in. *»
,92.
5 XV.
517.
3l?.
525.
334.
50?.
323.
529.
»50.
553.
»4)B.
3-*2.
551.
7?9
»5s.
537.
527.
5<»2.
8U.
81.
50.
67.
2G.
35~.
8?.
35.
7*.
8.
25.
Q .
9.
19.
9.
65.
3?.
9.
-------
OIL
8URN£F CON-I3U*4TIONS FOR 30ILER 52
d
T
2
T-
'-
7
3
0
j
j
0
? ""
2
7-
3
.,_.
2
7" '
2
2
2
"2~
2
._T_
2
-2~-
2
2
if 5
0""C
L C
G 2"
: 2
2~7~
? 2
?- -2-
r ?
L" 2~
2 2
7" ~2'
2 2
?'" 2"
2 2
2 ?
z r
2 2
2" 7
2 2
Z 2
2 2
7~ 2"
6
0
Q
C
C
~T~
1
T
2
"2"
2
' T~
2
2
0
u~
1
T~
1
I"
l
l
2
7
L'
C
'J
G
T
1
2
~r
2
"2"
2
2
0
"C
1
r
1
T
1
1
2
2
a
j
0
2
2
j-
2
--2-
2
~T
2
2
2
2
2
2
T"
2
2~
2
2
2
2
9
1
1
1
~1
1
~i-
2
2
""2"
2
i!
0
" tr
i
~r
i
"i~
i
i
2
^
1C
2
<»
2
2~
>
- z
2
T
2
T
2
y
2
._2-
2
T
2
~r
2
x PO^TS
c.
0.
0.
a.
"r.
G.
-o.
a.
T. - -
a.
o. -
0.
TJ.
G.
- o;
G.
-c.- -
G.
a.
0.
TJ. "
a.
U
-------
APPENDIX E
EFFECT OF PARTIAL REACTION WITHIN A BURNER ON
THE RESISTANCE TO AIR FLOW THROUGH THE BURNER
The data shown in Section 3, Figure 3-2 indicate that
the admittance to air flow of an (F+A) burner is a function of the air
flow velocity through the burner, presumably from the effect of the
air velocity on the fraction of combustion completed within the burner.
As a means of substantiating this assumption, it was considered
necessary, however, to independently calculate the fraction of com-
bustion which must occur within the burner to result in the measured
admittance values. The values of ADMNP and ADMA estimated from
the data of Figure 3-2 must be at least sufficiently correct that the
independent calculations of ADMFA match the experimental ones with
fractions of combustion within the burner ranging from zero to one.
As in Figure 3-2, a value of 2.05 metric (10, English) was assigned
to both ADMNP and ADMA. Independent analyses by engineers
associated with the utility confirm that these two admittances are
approximately equal.
For this analysis, an operating (F+A) burner was
divided into three sections: (a) the inlet section, consisting largely of
air register and inlet resistance to pure air flow (Rj), (b) a cold mixing
section, including all of the unreacted air-fuel mixing and momentum
interchange (R?), and (c) a constant-area heat-addition section, extend-
ing to the exit of the burner, where some fraction C, of the reactants
E-l
-------
are consumed (R,). As mentioned in Section 3. 2. 1, the effect of
momentum interchange in the cold gas mixing section was found to be
negligible and, therefore, R, was set equal to zero. The air flow
resistance of an (F+A) burner then was defined:
(E-l)
With no heating due to reaction, R_ also becomes zero, and all resis-
tance is assumed to be due to air register and inlet losses, identical
with those in an (Air) burner. Therefore, it is seen that
, o
1 (ADM A)
-------
From the perfect gas law, it is found:
(E-6)
In this case, the entire pressure drop (dp) through the burner is small
compared with the mean pressure level (p), while the temperature
change due to heating (dT) can be very large compared with the mean
temperature (T). Therefore, dp/p can be neglected compared to dT/T,
and density variations become a function of temperature variations
only:
dp = - ^ dT (E-7)
The energy equation can now be written:
but since
(E-8)
M (E-9)
o
and variations in p are small within the burner, we can let
vT = v T (E-10)
1 'a a
E-3
-------
Then, it is seen that
( ^ )dT = 0 (E-ll)
WAX/
dp
VcV- "a
and across the heat release section of the burner that
/ -2 \
(T_-T)
a'
Define the heat release due to combustion:
q=wfqfCh
and constant pressure heating:
Then, it is found that
w
and
. 2
E-4
= wCp(T2 - Tt) (E-14)
_w 1 _,£ _L c, (E-16)
1 A2T /\W/ Cp h
'a a' r
-------
Equation (E-16) can be expressed in terms of the air flow by noting:
w=w
a
and
D
Finally,
Pi -
Noting that
g\a
Using the values:
(E-18)
w2 (E-20)
(E-21)
V = 0.5241 kg/m3 (0.03972 lb/ft3)
cL
A = 0.6567 m2 (7.069 ft2)
qf = 11, 700 kg-cal/kg (21,000 Btu/lb) of fuel
Cp = 0.24 kg-cal/kg-°K (0.24 Btu/lb-°R) (air)
E-5
-------
Ta = 567 °K (1020 °R)
Using the English units, Eq. (E-21) becomes:
(E-22)
Values of C, (the fraction of combustion completed in the burner)
were calculated from Eq. (E-22) by the following procedure:
a. Calculate total admittance ADMT.
b. Subtract ADMNP and the sum of ADMA for the
(Air) burners.
c. Calculate ADMFA for the (F+A) burners.
d. Calculate R~ from ADMFA and R..
e. Calculate air flow through the (F+A) burners from
the admittance ratios.
f. Calculate fuel flow in the (F+A) burners.
g. Calculate r, .
h. Calculate C, .
n
The resulting values of C, for all available gas-fired
data from the H3 boiler are shown in Figure 3-3. The C, values for
gas-fired data lie between zero and one. Calculations for the oil data
shown in Figure 3-3 also indicate C, values near zero, as expected.
The values of ADMNP and ADMA appear to be approximately
correct. Values of ADMA, then were estimated for all other boiler
types, where data were available, from a procedure such as shown
in Figure 3-2.
E-6
-------
NOMENCLATURE FOR APPENDIX E
The following is a list of terms used in the equations
of this Appendix. Other terms used in this Appendix can be found in
the "Nomenclature" of the text.
2 2
A = cross-sectional area of a burner, m (ft )
C = specific heat of reaction products, kg-cal/kg-°K
p (Btu/lb-°R)
2 2
g = acceleration of gravity, m/sec (ft/sec )
2 2
p = static pressure, kg/m (abs) (Ib/ft )
2 2
AP = pressure drop, kg/m (diff) (Ib/ft )
q = heat of reaction, kg-cal/kg (Btu/lb) of fuel reacted
q = rate of heat generation due to reaction, kg-cal/sec
(Btu/sec)
R, = overall air flow resistance of an (F+A) burner,
la sec2/kg-m2 (sec 2/lb-ft2)
R., R_, R- = air flow resistance in (a) the air-only inlet section,
(b) the cold air-fuel mixing section, and (c) the
reaction section of an (F+A) burner, sec2/kg-m2
(sec2/lb-ft2)
r = A/F ratio, by weight
Y = weight density, kg/m (Ib/ft )
Subscripts
1 = conditions at the upstream end of the reaction section
of an (F+A) burner
2 = same as (1), at the burner exit
E-7
-------
TECHNICAL REPORT DATA
(Please read Inunctions on the reverse before completing)
1 REPORT NO.
EPA-650/2-75-012
2.
3. RECIPIENT'S ACCESSION>NO.
4. TITLE AND SUBTITLE
Analysis of Test Data for NOx Control in Gas- and
Oil-Fired Utility Boilers
5. REPORT DATE
January 1975
6. PERFORMING ORGANIZATION CODE
7. AUTHORIS)
Owen W. Dyke ma
8. PERFORMING ORGANIZATION REPORT NO.
ATR-75(7487)-l
9 PERFORMING ORdANIZATION NAME AND ADDRESS
The Aerospace Corporation
Urban Programs Division
El Segundo, CA 90245
10. PROGRAM ELEMENT NO.
1AB014; ROAP 21ADG-089
11. CONTRACT/GRANT NO.
Grant R-802366
12 SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
NERC-RTP, Control Systems Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final: 5/1/73-8/31/74
14. SPONSORING AGENCY CODE
IS. SUPPLEMENTARY NOTES
16. ABSTRACT
describes analyses of a large number of emissions, operating
conditions, and boiler configuration data from full-scale, multiple -burner electric
utility boilers using natural gas and low-sulfur oil fuels. Objectives of the study in-
cluded: evaluation of the effects of combustion modifications on NOx emissions, in
fundamental combustion terms; evaluation of techniques for further reductions in NOx
emissions; and determination and substantiation of general mechanisms for observed
combustion and flame stability problems. The report includes: (1) a discussion of the
major combustion process medications resulting in NOx emission reductions due to
two-stage combustion, burners out-of-service, combustion air temperature reduc-
tion, load reduction, and excess air variations; (2) estimates of NOx minima achie-
vable in the boilers studied with current hardware; (3) estimates of most probable
longer term hardware and operating condition modifications likely to yield ultimate
NOx reductions with these fuels; (4) identification and verification of general mech-
anisms for the combustion and flame instabilities observed; and (5) a list of all of the
hardware configurations, operating conditions, and NOx, CO, O2, and CO2 emissions
data for 428 tests in eight full-scale, multiple -burner, face-fired electric utility
boilers using natural gas and low-sulfur oil fuels.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Air Pollution
Combustion
Emission
Boilers
Nitrogen Oxides
Carbon Monoxide
Hydrocarbons
Smoke
Natural Gas
Fuel Oil
Combustion Stability
Air Pollution Control
Stationary Sources
Utility Boilers
Emission Characteris-
tics
Low-Sulfur Res id
13B, 07C
21B
21D
13A
07B
B. DISTRIBUTION STATEMENT
Unlimited
19 SECURITY CLASS (This Report)
Jnclassified
21. NO. OF PAGES
265
20 SECURITY CLASS (This page I
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
------- |