EPA-600/2-77-025
January 1977
Environmental Protection Technology Series
REDUCTION OF NITROGEN OXIDE EMISSIONS
FROM FIELD OPERATING PACKAGE BOILERS:
Phase III of III
Industrial Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into five series. These five broad
categories were established to facilitate further development and application of
environmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL PROTECTION
TECHNOLOGY series. This series describes research performed to develop and
demonstrate instrumentation, equipment, and methodology to repair or prevent
environmental degradation from 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 REVIEW NOTICE
This report has been reviewed by the U.S. Environmental
Protection Agency, and approved for publication. Approval
does not signify that the contents necessarily reflect the
views and policy of the Agency, nor does mention of trade
names or commercial products constitute endorsement or
recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-77-02 5
January 1977
REDUCTION OF NITROGEN OXIDE EMISSIONS
FROM FIELD OPERATING PACKAGE BOILERS
PHASE III OF III
by
M.P. Heap, C. McComis, andT.J. Tyson (Ultrasysterns)
and
R. E. McMillan, R.E. Sommerlad, andF.D. Zoldak
(Foster Wheeler Energy Corporation)
Ultrasystems, Inc.
2400 Michelson Drive
Irvine, California 92664
Contract No. 68-02-0222
ROAPNo. 21ADG-043
Program Element No. 1AB014
EPA Project Officer: G. Blair Martin
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, NC 27711
Prepared for
American Petroleum Institute U.S. Environmental Protection Agency
1801 K Street, NW Office of Research and Development
Washington, DC 20006 Washington, DC 20460
-------
ABSTRACT
This report describes the final phase of a program .to determine
the optimum methods of applying both flue gas recirculation and
staged combustion to control NOV emissions from residual oil-fired
A
package boilers. Experimental investigations were carried out in
a laboratory firetube boiler simulator and an application program
was conducted on two boilers operating in the field. The ultimate
goal of the program was to determine if package boilers could
operate in the field after modification to control nitrogen oxide
s
emissions without encountering practical problems.
A 12 x 10 Btu/hr firetube boiler and a 25 x 106 Btu/hr heat out-
put watertube boiler were modified to extract cooled combustion
products from the stack and add them to the combustion air in the
windbox. The effectiveness of flue gas recirculation as a method
of controlling NO emissions was found to be dependent upon boiler
A
type. It was most effective in the firetube boiler; approximately
30 percent reduction in emissions was obtained with 40 percent
recirculation. NO reductions achieved by staged combustion were
A
greater in the field tests than in the laboratory investigation.
Forty-five percent reductions were achieved without undue smoke
emissions when 70 percent of the stoichiometric air requirements
were applied to the burner.
Based upon the results of these investigations it is doubted
whether flue gas recirculation is a cost-effective NO control
A
technique for residual oil-fired package boilers; however, staged
combustion techniques show significant promise for pollution
control.
-------
TABLE OF CONTENTS
Page
ABSTRACT
LIST OF FIGURES iii
LIST OF TABLES vi
1.0 SUMMARY 1
1.1 Scope of the Program 1
1.2 Laboratory Investigations 2
1.3 Investigations Involving Practical Boilers 4
1.4 NOX Control Techniques for Package Boilers 7
2.0 INTRODUCTION 11
3.0 SUMMARY OF PHASE II LABORATORY INVESTIGATIONS 13
4.0 LABORATORY INVESTIGATIONS 21
4.1 The Influence of Fuel Oil Type 21
4.2 Atomization Parameters and Pollutant Formation 26
4.3 Staged Combustion Investigations 33
5.0 FIELD INVESTIGATIONS 43
5.1 Selection of the Test Boilers 43
5.2 Equipment Used in the Field Investigations 48
5.3 Result of the Field Investigations 58
5.4 Operational Experience . 77
6.0 DISCUSSION AND CONCLUSIONS 81
6.1 Typicality of Field Test Units 81
6.2 Comparison of Laboratory and Field Test Results 84
6.3 Cost of Emission Control 88
6.4 Implication of Results on New Design 91
REFERENCES 93
APPENDIX 1 TABULATED FIELD TEST DATA 95
APPENDIX 2 BOILER PERFORMANCE BEFORE AND AFTER MODIFICATION 107
IV
-------
LIST OF FIGURES
Page No.
1.1 Comparison of FGR Effectiveness as a Control Technique for
Watertube and Firetube Package Boilers (20 Percent Excess
Air) 6
3.1 Laboratory Combustor - Schematic 14
3.2 Sketch of Modified Commercial Burner Used in the Laboratory
Investigations 16
3.3 The Influence of Flue Gas Recycle Injection Location on NO
and Smoke emissions ( No. 6 Fuel Oil, 17 Percent Excess Air,
Load 3.4 x 106 Btu/hr and Baseline Emission 273 ppm) 18
4.1 Comparison of Oil Types - the Effect of Excess Air on Emission 23
4.2 Comparison of Oil Types - the Effect of Primary/Secondary
Air Ratio 24
4.3 Operation of the Unmodified Burner 25
4.4 The Influence of Atomization Method 27
4.5 The Influence of Atomization Medium 28
4.6 Reproducibility Tests - Duplicate Oil Nozzles 29
4.7 The Influence of Nozzle Capacity 30
4.8 The Influence of Atomizing Medium 31
4.9 The Influence of Atomizatton Medium 32
4.10 The Influence of Swirl and Excess Air 34
4.11 The Influence of Atomization Pressure on Staging Performance
(P/S 15/85) 36
4.12 The Influence of Atomization Pressure on Staging Performance
(P/S 85/15) 37
4.13 The Influence of Primary Air Percentage and Swirl Level on
NO Emissions (17 Percent Excess Air 3.4 x 104 Btu/hr Heat
Input) 38
4.14 The Influence of Primary Air Percentage and Swirl Level on
Smoke Emissions (17 Percent Excess Air 3.4 x 104 Btu/hr
Heat Input) 39
-------
LIST OF FIGURES (CONT'D)
Page No.
5.1 Register Burner Installed in the Watertube Boiler 49
5.2 Layout of Flue Gas Recirculation System for the Watertube
Boiler 50
5.3 Sketch Showing the Windbox Burner Arrangement of the
Firetube Burner 52
5.4 Layout of Flue Gas Recirculation System for the Firetube
Boiler 53
5.5 Details of Equipment Used in the Staging Investigation 55
5.6 Schematic of Automatic Controls for Flue Gas Recirculation 56
5.7 The Influence of Load and Excess Air on Pollutant Emissions
From the Firetube Boiler (Boiler Performance Tests) 60
5.8 The Influence of Load and Excess Air on Pollutant Emissions
From the Watertube Boiler (Boiler Performance Tests) 61
5.9 The Influence of Register Position on Pollutant Emissions
Watertube Boiler (Boiler Performance Test Figures Besides
Symbols Denote 02 Percentage) 62
5.10 The Influence of Fuel Oil Temperature on Pollutant
Emission from the Watertube Boiler (Boiler Performance
Tests) 64
5.11 The Influence of Atomizing Steam Pressure on Pollutant
Emission from the Watertube Boiler (Boiler Performance
Tests) 65
5.12 The Influence of Load and Excess Air on NO Emission from
the Watertube Boiler (Boiler Performance Tests) 66
5.13 The Influence of Register Setting on NO Emission from
the Watertube Boiler Fired by Natural G$s (Boiler Per-
formance Tests) 67
5.14 The Influence of FGR on NO and Smoke Emissions Firetube
Boiler, 4,000 Ibs Steam per hour 70
5.15 The Influence of FGR on NO and Smoke Emissions Firetube
Boiler, 6,200 Ibs Steam per hour 71
5.16 The Influence of FGR on NO and Smoke Emissions Firetube
Boiler, 10,300 Ibs Steam pงr hour 72
VI
-------
LIST OF FIGURES (CONT'D)
Page No.
5.17 The Influence of FGR on NO and Smoke Emissions Water-tube
Boiler, 6,500 Ibs Steam peฃ hour 74
5.18 The Influence of FGR on NO and Smoke Emissions Watertube
Boiler, 10,000 Ibs Steam pงr hour 75
5.19 The Influence of FGR on NO and Smoke Emissions Watertube
Boiler, 16,000 Ibs Steam p@r hour 76
5.20 The Influence of Staging on NO Emissions from the Fire-
tube Boiler (6,000 Ibs of steaffl per hour) 78
5.21 The Influence of Staging on Smoke Emissions from the
Firetube Boiler (6.000 Ibs of steam per hour) 79
6.1 Comparison of Boiler Performance Data with the of Cato
et al(7) 82
6.2 Relationship of Fuel Nitrogen Content and NO Emissions
from Industrial Boilers 83
6.3 Fractional Reduction of NO Achieved by FGR Comparison
of Field and Laboratory Data 86
A2-1 Comparison of Boiler Performance Data Before and After
Modification (Firetube Boiler) 108
A2-2 Comparison of Boiler Performance Data Before and After
Modification (Watertube Boiler) 109
vii
-------
LIST OF TABLES
.. Page No.
4.1 Properties of the "EPA" and "Ultrasysterns" No. 6
Fuel Oil 22
4.2 The Influence of Nozzle Capacity on Stagine Performance 40
4.3 The Influence of Method of Staged Air Injection Upon
Pollutant Emissions 42
5.1 10 Year Survey of Package Firetube Boiler Sales (In-
cluding High and Low Pressure"Steam and Hot Water Units) 45
5.2 10 Year Survey of Package Watertube Boiler Sales 46
5.3 Possible Candidate Plans 47
5.4 Flue Gas Recirculation Tests 69
6.1 Cost Breakdown for Fitting FGR to the Two Boilers at
ECCC (1975 Dollars) 89
6.2 Approximate Cost Breakdown for Application of Flue Gas
Recirculation to New Boilers 90
6.3 Breakdown of Costa for Staged Combustion Investigation 90
Al-1 Boiler Performance Data ECCC Firetube Boiler No. 5 Fuel Oil 96
Al-2 Boiler Performance Data ECC Firetube Boiler Natural Gas 97
Al-3 Boiler Performance Data ECCC - Watertube No. 5 Fuel Oil 98
Al-4 Boiler Performance Data ECC - Watertube Natural Gas 99
Al-5 Flue Gas Recirculation Watertube No. 5 Oil 100
Al-6 Flue Gas Recirculation Firetube No. 5 Oil 102
Al-7 Staged Combustion Firetube No. 5 Fuel Oil 104
A2-1 Fuel Oil Analyses 111
A2-2 Independent Analysis 112
vm
-------
1.0 SUMMARY
1.1 Scope of the Program
There have been a considerable number of investigations detailing the application
of flue gas recirculation or staged combustion techniques to control nitrogen
oxide emissions from Utility Boilers. In comparison, very little is known con-
cerning the practical aspects of applying these same techniques to package
boilers.
A package boiler is described in the ABMA Lexicon as:
"a boiler equipped and shipped complete with fuel burning equipment,
mechanical draft equipment, automatic controls and accessories.
Usually shipped in one or more major sections."
Although boilers with a capacity up to 250,000 Ibs of steam per hour can be
shipped as a single unit by rail or truck, larger units (250,000 to 350,000 Ibs
of steam per hour) must be modularized. Thus, the term packaged encompasses a
wide range of equipment, size range, design type and fuel capability. The
investigations described in this report are somewhat more limited in scope since
they are mainly concerned with firetube boilers and practical experiments with
equipment in the lower size range (up to 25,000 Ibs of steam per hour).
This report describes the final phase of a program jointly supported by the
API and EPA. These phases were:
Phase I - Construction of a versatile combustor (EPA Report
R2-73-292a)
Phase II - Experimental investigations in that versatile com-
bustor to determine the optimum method of applying
both flue gas recirculation and staged combustion
to control NO emissions (EPA Report R2-73-292b)
A
Phase III - Demonstration of the applications of these techniques
to operating boilers in the field and extension of the
laboratory experiments.
-------
The ultimate goal of the total program was to demonstrate that package boilers
could be operated in the field without practical problems after modification to
control nitrogen oxide emissions. The techniques used to control emissions were
identified after an extensive series of laboratory investigations with flue
gas recirculation and staged combustion as the prime control candidates.
1.2 Laboratory Investigations
Equi pment
An axisymmetric calorimetric combustor constructed to simulate the combustion
chamber of a firetube boiler was used in the laboratory investigations. All
the results were obtained with a modified commercial burner in which the com-
bustion air supplied to the primary and secondary streams could be controlled
separately. The combustor was designed to allow the addition of flue gas ,
recirculation or staged combustion air at various locations. The investigations
were restricted to measurements of combustion product composition; the major
emphasis being nitrogen oxides and smoke.
Results
All attempts to control nitrogen oxide emissions from fuel oil flames were
generally limited by excessive smoke formation. The results of the laboratory
investigations can be summarized by:
1. NOV emissions were found to be lower when steam was used as the
A
atomizing medium rather than air.
2. The modified burner, when operated with a primary/secondary ratio
of 50:50, did not duplicate the results from an unmodified burner
burning the same fuel.
3. As expected, NOX emissions were found to be higher when burning oil
of a higher nitrogen content. However, trends with excess air and
load were found to be different for the two oils tested.
4. Emission characteristics were found to be dependent upon oil nozzle
capacity.
-------
5. The method of injection and the location of staged air addition
was found to have an influence on smoke emission during staging.
Radial staged air injection was found to be superior to tangential
injection.
6. The effectiveness of staging as an NO control technique was improved
A
by modifying conditions at the burner to improve mixing in the early
stages of heat release. In this way, lower NO concentrations were
A
obtained before smoke emissions became excessive.
The results of the laboratory investigations can be considered encouraging but
not representative of the ultimate expectation in NO control for oil fired pack-
X
age boilers. However they gave a strong indication of the direction that future
work in this area should take. With one exception all other attempts to control
NO emissions from residual oil fired systems have yielded similar results - a
J\
maximum of 50" reduction in NO emissions with increased particulate emissions.
/\
The problem of NO control for nitrogen containing liquid fuels cannot be mini-
A
mized. Although more is known now of the fate of fuel nitrogen during residual
oil combustion than was known when this program began more attention must be paid
in the future to minimizing the tradeoff between reduced fuel NO formation
and increased soot production.
Limitations of the Laboratory Investigations
It is generally accepted that the formation of nitrogen oxides in the type
of combustors studied in this investigation is mainly controlled by turbulent
transport which dictates the rate at which fuel and air are mixed. In the
present investigation, only a limited series of experiments were carried out in
which changes were made to the burner to influence the fuel/air mixing process.
It was found that staging performance could be improved significantly by varying
the axial and tangential velocity distribution at the burner throat.
-------
The commerical burner used in the laboratory investigation was modified and
operated in an unnatural mode with a fixed air distribution between the primary
and secondary streams. Apparently this distribution varies with load in the com-
mercial burner which may never operate with an equal flow in the primary and
secondary streams. Very little information is available on the influence of fuel
type or the atomization method on the effectiveness of control techniques. The
investigations were restricted to Input-Output (I/O) parametric studies; informa-
tion has not been generated to allow an explanation to be given for the observed
phenomena. Consequently, it is very difficult to generalize these results to the
many different situations likely to be encountered in the field. As the commer-
cial burner was not operated as it would be in the field, it is even difficult to
claim that these investigations relate to one class of practical firetube boiler
burners.
Perhaps the most serious limitation of the experimental investigation is the re-
striction to firetube boiler conditions. It may well be that emission control
techniques optimized for firetube boilers will not be optimum for watertube
boilers. The combustion chamber of a firetube boiler is characterized by a large
length to diameter ratio which imposes different requirements for flame shape
than those for watertube boilers. The general control principles for the two
boiler types will be the same but their method of implementation could be very
different.
1.3 Investigations Involving Practical Boilers
Equipment
The choice of boilers tested during the field investigation was dictated in part
by convenience to the Foster Wheeler Corporation. The other major criterion
was that the tests should be made with two boilers of different design burning
the same oil. The two boilers tested were:
a watertube boiler, 25,000 Ibs/hr steam, and
a firetube boiler, 12,000 Ibs/hr steam.
-------
Staged combustion investigations were conducted only in the firetube boiler on
an experimental basis. However, both boilers were modified to accept flue gas
recirculation to the windbox. These modifications included a fan, ductwork and
an automatic control system.
Results
Although only tested in the firetube boiler, control of NO emissions by staged
A
combustion techniques proved more successful in the field than in the laboratory
investigations. Forty-five percent reductions in NO emissions were achieved
/\
without undue smoke emissions when approximately 70 percent of the stoichiometric
air requirement was supplied through the burner. The improved performance in
the field could be attributed to changes made to the secondary air injection
system based upon laboratory experience and differences between the laboratory
combustor and the field test boiler.
The effectiveness of flue gas recirculation as a method of controlling NO
X
emissions was found to be dependent upon boiler type. Figure 1-1 allows a
comparison to be made between the influence of F6R on NO emissions from the
)\
two boilers tested at three loads and 20 percent excess air. The results
strongly suggest that FGR is not a cost effective control technique for residual
oil fired watertube boilers in this size range. Comparison between the laboratory
results and the results of other workers suggests that the results for the fire-
tube boiler are typical and approximately 30 percent reductions in emissions can
be expected with up to 40 percent recirculation. This is because reductions in
flame temperature (achieved by FGR) have only a minimal effect upon the oxidation
of fuel nitrogen to nitric oxide.
The difference in the design of the two boilers and burners probably results
in differences in the amount of fuel NO contained in the total emission. As
FGR will only be effective in reducing thermal NO formation, it can be con-
cluded that the watertube with its low level of combustion intensity, pro-
duced very little thermal NO and boilers of this type would be poor candidates
for NO control through FGR.
-------
Fi retube
200
-งSLioad
CM
o
o
o
s_
o
u
ง.
ex
Watertube
_ 15% Load
"***
-------
Prob1 ems
The field investigations had two objectives, not only were they planned to
serve as a demonstration of NO control techniques, but they were also
J\
intended to uncover potential practical problems associated with the applica-
tion of these techniques. The boilers chosen for testing were done so with a
knowledge of the requirements of the control system to be installed and yet
problems were uncovered that could not have been anticipated. Changes in the
geometry of the firetube boiler windbox to accommodate the flue gas recircula-
tion inlet resulted in the initiation of severe pulsations at several boiler
loads. Successive modifications succeeded in alleviating the problem but not
in eliminating it. Flame instability occurred in the watertube boiler with
the addition of greater than 30 percent recirculation at all boiler loads.
Without prior direct experience, neither of these practical problems could be
predicted. There is no reason to believe that these two boilers represent
special cases and it must be expected that similar problems could occur with
other units. The age of equipment and the lack of available space in the
vicinity of most package boilers will tend to extend the problems of retrofit,
particularly with respect to flue gas recirculation even if it were to be
shown to be an effective control technique.
On the basis of a field test of two package boilers it is difficult to draw
definite conclusions concerning NO control techniques. Consequently, the most
/\
serious limitation of the field investigation is their restriction to two
boilers. No influence of scale can be established (e.g., does a 6,000 Ibs/hr
of steam firetube behave differently when fitted with FGR or staged combustion
equipment than an 18,000 Ibs/hr of steam unit).
1.4 NO Control Techniques for Package Boilers
A
The scope of the present study was too limited to establish the absolute
cost and the effectiveness of conventional NOV control techniques as applied
X
to package boilers. However, cost information when considered in relation to
the results summarized earlier, does give a strong indication of the area in
which future work should be directed.
-------
Summary of Control Technique Costs
The application of NO control techniques to package boilers can be considered
A
from three viewpoints:
1. Retrofit of existing equipment already installed.
2. Shop retrofit of a new boiler of existing design before installation.
3. Redesign of the boiler/burner combination.
The direction of this investigation was heavily biased towards the first two
alternatives. The cost of equipping the two boilers for flue gas recircula-
tion was approximately $20,000 (1.6 to 0.8 dollars per Ib of steam). This
cost included design, equipment purchase, fabrication and installation. The
cost of installation of a new boiler would be approximately $3.1 per Ib of
steam. It could be expected that experience would allow economies to be
made, thus reducing unit cost. However, based upon this cost comparison,
retrofit of medium-sized boilers with FGR would appear to be expensive. This
statement is enforced by a recognition that each retrofit in the field would
be in some way unique. Consequently, for nitrogen containing fuels alterna-
tive, more cost effective control techniques should be sought.
Theoretically, staged heat release offers the possibility to control both
thermal and fuel NO and appears to offer more promise. The major drawback
appears to be the strong possibility of an emissions tradeoff between parti-
culate and NOX- The staged combustion investigations described in this study
can only be considered as experimental; this is reflected in the cost incurred
of $28,000. If a separate air injection were necessary some distance along
ttie firetube, then this could be accomplished in a new boiler without this
expense.
8
-------
Although the cost of the boiler modifications to control NO could be reduced
A
somewhat in the future they are still high. Particularly when compared to the
initial capital cost. A complete new burner system including fuel nozzle, oil
pump, blower controls, etc. could be purchased for the same as the cost of the
flue gas recirculation system. Thus, any control technique which could be de-
veloped requiring some modification to the existing burner would appear to be
the most promising from an econimic viewpoint. Two areas of burner redesign are
suggested.
t Modification of the fuel injection system, i.e., atomizer
characteristics; and
Redistribution of the combustion air to prevent rapid mixing of all
the air and fuel. This could be accomplished by injecting some of
the air around the periphery of the firetube or around the exit of a
watertube register burner.
Areas Requiring Further Study
Based upon experience gained during the present investigations, several areas
requiring more detailed investigation can be defined.
Optimization of control technique for fuel-type - if package
boilers are to have dual or multi-fuel capability, then the pollution
control technique must be optimized for all fuels. This study was
too restricted in this manner.
Optimization of the total combustion system for efficiency and
pollutant control - if fuel/air mixing is controlling pollutant
formation the total system (e.g., burner and staging equipment)
should be optimized. It may not be sufficient to add staging equip-
ment to a boiler without modifying the burner.
Investigations of staged preheated air addition to reduce soot
formation.
-------
In retrofit situations combustion pulsation and ignition instabilities
may limit the application of various control techniques. A basic
understanding of these phenomena could allow these problems either
to be avoided or to be overcome more easily.
Efforts must be made to allow the results of this type of investigation
to be generalized to a wider class of equipment.
Conclusion
Flue gas recirculation does not appear to be a cost effective NO control tech-
/\
nique for fuels containing bound nitrogen burning with low intensity in cold
wall combustors. Staged combustion has shown greater promise for NO control;
J\
however, further work is necessary to establish the optimum method of applying
staged combustion techniques to package boilers. This work should be directed
toward using the burner as the staging device because this will probably be the
most economic approach for liquid fuels.
10
-------
2.0 INTRODUCTION
Steam and hot water boilers with heat inputs ranging from 3 to 400 x 106 Btu/hr
presently account for approximately 50 percent of the oil fired in stationary
boilers and emit 16 percent of the nitric oxides attributed to all stationary
boilers. It remains to be proven whether these sources have a more serious
impact upon urban pollution problems because they are usually situated at the
centers of population. In view of the national energy problems, it is neces-
sary that all attempts to reduce combustion-generated pollutants do not increase
fuel usage.
The two previous phases of the present program were concerned with the con-
struction of laboratory simulator and the definition of promising NO control
(1 2)
techniques^ '. These techniques have been applied to two boilers operating
in the field. This report deals mainly with the reduction of pollutant emissions
associated with oil firing. Further work is in progress (EPA Contract 68-02-
1498) to provide more information on the use of staged combustion and flue gas
recirculation to control pollutant emissions from packaged boilers. The scope
of the work with residual fuel oil will be extended and comparison will be made
with natural gas and alcohol fuels.
11/12
-------
3.0 SUMMARY OF PHASE II LABORATORY INVESTIGATIONS
During the second phase of this three-phase program investigations were carried
out in a specially constructed laboratory combustor to establish NO control
- X
techniques suitable for oil-fired package boilers. Four control options were
investigated:
Burner Modifications. The commercial burner had been modified to
independently vary the swirl level and the air distribution in the
primary and secondary ducts. The other parameters investigated were
associated with the atomization of the fuel oil, viz. oil tempera-
ture, atomization air pressure and the use of nitrogen as an
atomizing fluid.
Flue Gas Recirculation. The influence of the addition of cooled
combustion products to the combustion air was investigated. The
combustion products could be mixed with the primary, secondary or
total air streams, injected separately through the gas ring or
through ports in the refractory burner throat.
t Staged Combustion. Second stage air was added through sidewall
injectors or from a rear boom to allow the influence of staged
heat release to be investigated.
Combined Flue Gas Recirculation and Staging. The influence of
simultaneous addition of cooled combustion products and staged
heat release. It was intended that these investigations would
define the optimum NOV control technique which could then be
A
tested in the field.
The experimental system used in the laboratory investigations has been des-
cribed in detail elsewhere The axisymmetric calorimetric combustor
was custom-built to enable recirculated products and second stage air to be
injected at various locations. A schematic layout of the combustor and the
associated air and flue gas supply system is presented in Figure 3-1. The total
combustion air supply could be divided into two variable streams, referred
to as first and second stage air. As indicated in Figure 3-1, the first
stage air was supplied through the burner and the second stage air could be
13
-------
Coolant
Refractory
Exit Divergent-
Atomizing
Air
Primary
Air
Throa t
Injec-
tors.
w ฎ <
Water Cooled
2nd Stage Air Injection
Lance
Flue
Sidewall
Injectors
Secondary Air
i i 11 i 11 1st Stage Air
.. 2nd Stage Air
i Recycled Flue Gas
Figure 3-1. Laboratory Combustor - Schematic
14
-------
injected either through throat injectors, sidewall injectors or through a
lance inserted from the rear of the combustor. Axial movement of this lance
allowed the influence of the position of second stage air injection to be
investigated. The combustor was fired by a modified commercial burner, the
details of which are presented in Figure 3-2. The major modification allowed
the total air supplied to the burner (first stage air) to be divided into
primary and secondary air streams. The primary air flow was essentially axial;
some rotation could be imposed upon the secondary stream by closing the inlet
air damper while maintaining the mass flow rate constant. Recycled combustion
products could be added to the primary, secondary or the total air streams or
injected separately through the gas ring or the throat injectors.
(3)
Muzio et alv have discussed the complete laboratory results in detail and
only those results which have implications for the field demonstrations will
be summarized in this report.
Burner Modifications
Smoke and NO emissions from No. 6 fuel oil were found to be very sensitive to
A
the primary-secondary air ratio. NO emissions were reduced and smoke emissions
X
were increased as the proportion of air in the primary stream was reduced. In
retrospect, these results are not compatible with the field investigations since
the burners used in the field tests did not have separate primary and secondary
air streams. Emissions were insensitive to variations in secondary swirl at
50 percent primary air flow. Increasing atomizing air pressure from 14 to
36 psig caused a reduction in the NO emission on the order of 20 percent. How-
ever, changes in oil temperature were found to have minimal effects.
Flue Gas Recirculation
Investigations with three fuels, natural gas, No. 2 and No. 6 fuel oil, indi-
cated that the effectiveness of flue gas recirculation as an NO control techni-
A
que was limited by the nitrogen content of the fuel. Flue gas recirculation
has only a minor influence on the conversion of fuel bound nitrogen to nitric
oxide. Thus, its effectiveness as a control technique is minimal if the major
portion of the NO emission is attributable to fuel nitrogen oxidation. During
15
-------
Primary Air
Gas Injector
Oil Nozzle
>':.: Refractory
>V''A Throat
Secondary Air
Inlet With
Swirl Damper
Figure 3-2. Sketch of Modified Commercial Burner Used in
the Laboratory Investigations
16
-------
the Phase II investigations flue gas recirculation was added at five separate
locations and the results indicated that only three of these locations were
effective and only two could be considered as having practical value. Figure 3-3
gives an example of the influence of percent recirculation on NO and smoke emis-
sions for fixed load, excess air, and primary/secondary air distribution. The
maximum reduction was observed when the combustion products were injected through
the gas ring. This reduction was accompanied by a visual degeneration of com-
bustion conditions and an increased smoke emission and cannot, therefore, be
considered suitable for practical equipment. The results suggest that the
addition of cooled combustion products to the total air supply will provide the
most effective use of FGR as an NO control technique. Under certain circum-
/\
stances an optimum recirculation rate was found; if high recirculation rates
were added to the primary air stream, emissions tended to increase. This can
be attributed to increases in the rate of air/fuel mixing in the early stages
of heat release.
Staged Combustion
The effectiveness of staged heat release as a control technique was limited by
the direct tradeoff between reduced NO and increased smoke emissions. Only
modest reductions were obtained before smoke emissions became excessive. The
optimum location for staged air injection was approximately two combustor diam-
eters downstream from the fuel nozzle. These results were disappointing, parti-
(4)
cularly in the light of the results reported by Siegmund and Turnerv ;; how-
ever, it should be noted that these workers had to significantly downrate the
boiler to achieve these results and the staged air was available at 50 psig.
Smoke emissions could be reduced somewhat by lining the inside of the combustor
with refractory, but this did not improve the effectiveness of staging since
baseline emissions were increased. It should be noted that during these
investigations no changes were made to the burner during the staging process.
Combined Methods^
Combining staged combustion with flue gas recirculation was found to be an
effective method of reducing NO emissions without producing excessive smoke.
A
17
-------
o
w
fl
ฃt
O
o
2
0.8
0.7
0.6
5 O.S
.K).4
0.3
0.2
0.1
h NO Smoke
J*L
O
a
O +
A A
0
Recycle
Injection
Total Air
Secondary Air
Primary Air
Gas Ports
Throat
10
20
% Recirculation
30
10
9
8
7
6
5
4
3
2
1
0
2
tj
"g
3
o
CO
Figure 3-3. The Influence of Flue Gas Recycle Injection Location
on NOX and Smoke Emissions (No. 6 Fuel Oil, 17 Percent
Excess Air, Load 3.4 x 10& Btu/hr and Baseline Emission
Us ppm)
% Recirculation =
mass of ^circulated products \
mass of air + mass of fuel x 10ฐ,
18
-------
A major question concerning the results of these investigations and one which
will be returned to later in this report is their applicability to practical
systems. Although a commercial burner had been used in the laboratory investi-
gations, questions arise as to its typicality and to what extent the minor
modifications, carried out to provide experimental versatility, had influenced
its performance. It is doubtful whether this particular burner would operate
in the field with the combustion air divided equally between the primary and
secondary streams and almost certainly the relative air flows will change with
load. Also, the majority of firetube boilers are fired by burners which do
not have separated primary and secondary air streams.
19/20
-------
4.0 LABORATORY INVESTIGATIONS
The API-EPA Steering Committee expressed concern over the modest NO reductions
achieved in the Phase II laboratory investigations. It could not be established
whether these results were attributable to fuel oil properties or to peculiarities
in the experimental combustor. Consequently, it was decided that the third phase
of the project should proceed along two parallel paths.
Additional laboratory experiments to provide further information
in an attempt to explain the Phase II observations.
i Continue with the field demonstration as originally intended
although it was recognized that the staged combustion investi-
gations would necessarily be experimental and could not be con-
sidered as a demonstration of a practical system.
The laboratory investigations carried out during Phase III were planned to
investigate the influence of fuel oil properties, the method of fuel oil
atomization and to provide further information on the application of staged
combustion control techniques to oil field package boilers.
4-1 the Influence of Fuel Oil Type
The EPA-Combustion Research Section operates an almost identical experimental
combustion system to that described in Section 3^ '. The EPA combustor is fired
by the same model commercial burner as used at UHrasystems, but in an unmodified
state, which represents the only significant difference between the two systems.
However, the performance of the two combustors was found to be very different;
in particular, smoke emissions were considerably lower from the EPA combustor.
Tests were carried out at Ultrasystems with the oil used in the EPA combustor
to provide a direct comparison of the influence of oil type on pollutant emission.
Table 4-1 compares the properties of the two oils.
21
-------
Table 4-1. Properties of the "EPA" and "Ultrasystems" No. 6 Fuel Oil
Characteristics
Gravity, ฐAPI
Flash Point, PMCC ฐF
Pour Point, ฐF
Viscosity SSF at 122 ฐF, sec
Heat of Combustion, gross Btu/lb
Ash %
Sulphur %
Nitrogen % (by Keldahl)
Carbon %
Hydrogen %
Ultrasystems
16.7
165
80
97
17,746
0.02
0.42
0.36
87.68
11.61
EPA
o>
JD
5
to
-------
320
300
g 280
ฐ 260
8 24ฐ
220
200
180
160
140
10
o
IO
i.
IO
I
CO
s_
01
-Q
OJ
_*
O
to
Load
3.4xl06 Btu/hr
2.8xl06 Btu.hr
2.0x106 Btu/hr
Oil
EPA
D
o
o
Ultrasys.
10 15 20- 25 30. 35
Excess Air (%)
40
45
50
Figure 4-1. Comparison of Oil Types - the Effect of Excess Air on Emission
23
-------
140
120
100
^xC'
-
' 1 1 1
10
i.
8
o
-------
OJ
o
o
o
o
o
a.
CL
Excess Primary/
Air Secondary
Commercial
Burner
Commercial
Burner
Ref. 3.1
This
Investigation
This
Investigation
250
200
Load 10" Btu/hr
Figure 4-3. Operation of the Unmodified Burner
25
-------
higher primary air flows at low loads (P/S of 80/20) and at a P/S of approximately
65/35 at maximum load. This trend is consistent with common practice as the
increase in primary air flow would act to reduce smoke emissions at low load.
In general, the tests with the EPA oil were inconclusive. It was thought that
the "Ultrasysterns" oil had a higher smoking potential than the EPA oil; however,
the results did not substantiate this hypothesis. Perhaps the most confusing
set of results refer to the influence of primary/secondary ratio on NOX emis-
sions; no explanation can be given for the completely different NO emissions
A
characteristics of the two oils.
4.2 Atomization Parameters and Pollutant Formation
The larger size ranges of package boilers frequently utilize steam as the atomi-
zation medium and the oil nozzle fitted to the commercial burner used in this
investigation was equally suitable for either air or steam. Figures 4-4 and 4-5
compared the emissions characteristics of the laboratory combustor using air and
steam as atomizing agents. NO emissions were generally lower with steam and
A
the smoke emissions higher. Trends with variation of the primary/secondary ratio
were similar for both air and steam. However, there appears to be an optimum
steam pressure of approximately 35 psig for minimum NO emissions.
A
All of the Phase II data had been obtained with one 80 gallon per hour nozzle.
A duplicate nozzle was obtained to determine whether or not the emission data
could be duplicated for two nozzles of the same capacity. Similar trends were
observed for both nozzles (see Figure 4-6).
The characteristics of a 100 gallon per hour nozzle are compared with those of
the 80 gallon per hour nozzle at a fixed atomizing air pressure (20 psig) in
Figure 4-7. N0x emissions are similar for the two nozzles at high and low primary
air ratio. However, the larger nozzle produces 80 ppm more NO than the smaller
nozzle at a primary/secondary ratio of 60/40. Figures 4-8 and 4-9 compare the
performance of the 100 gallon nozzle with either air or steam as the atomizing
medium. In general, smoke emissions were lower with the larger nozzle than with
the smaller nozzle which was the specified size for the nominal burner capacity.
26
-------
350
300
CM
O
O
O
NO
Smoke
Steam (Phase III Data)
Air (Phase II Data)
P/S 50/50
3.4 x 106 Btu/hr
17% Excess Air
O
u
Q.
Q.
250
200
150
20 30 40
Atomization Pressure, psig
Figure 4-4. The Influence of Atomization Method
50
27
-------
350
Steam Phase III Data
Air Phase II Data
150
20/80
40/60 60/40
Primary/Secondary Ratio
Figure 4-5. The Influence of Atomization Medium
28
-------
340
330
320
310
cvj
CD
o
4J
J-
O
o
300
290
280
270
NO Smoke
/\
10
O
Old Nozzle
New Nozzle
3.4 x 10ฐ Btu/hr
17% Excess Air
80 Gal/hr nozzle
Air atomized
o
rO
t-
IC
CQ
s-
01
jQ
3
0)
JM:
o
20 40 60
Percent Primary Air
80
100
Figure 4-6. Reproducibility Tests - Duplicate Oil Nozzles
29
-------
Cxi
o
o
o
17% Excess Air
3.4 x 106 Btu/hr
260
40 60
Primary/Secondary Ratio
Figure 4-7. The Influence of Nozzle Capacity
30
-------
360
CM
O
O
O
S-
o
O
100 gal Nozzle
17% Excess Air
3 A x 106 Btu/hr
Atomiz. F*ress. 22 psig
20 40
Percent Excess Air
Figure 4-8. The Influence of Atomizing Medium
31
-------
3.4 x 10" Btu/hr
P/S 50/50
100 gallon nozzle
16 20 24
Atomization Pressure, psig
Figure 4-9. The Influence of Atomization Medium
32
-------
The influence of swirl level and excess air on emissions for the large nozzle
are compared in Figure 4-10. The "swirl" level had only a minor influence on
emissions with the 100 gallon nozzle. This was similar to observations made
during the Phase II investigations.
No attempt was made to vary the spray angle of the nozzle; all the results
reported to date refer to an included cone angle of 70ฐ. Muzio et al '
suggested that the decreased NO emission produced with increasing atomization
/\
air pressure resulted from the narrowing of the spray angle. This effect had
been observed when the oil was sprayed in the open air. If the spray angle is
reduced, then NO emissions would tend to be reduced. Heap et al ' have
A
demonstrated that for fixed air flow conditions NO emissions from heavy fuel
A
oil flames are reduced as the spray angle is reduced when either mechanical or
steam atomization was employed. Reduced emissions can be explained by postula-
ting that the narrower spray angles produce more rich conditions in the early
stages of combustion, and therefore, less fuel NO is formed.
4.3 Staged Combustion Investigations
The majority of the Phase II staged combustion investigations involved a fixed
set of burner conditions (primary/secondary ratio, atomizer size, atomizing air
pressures, swirl damper setting). In the Phase II investigations the sidewall
staging injectors were positioned close to the burner. In an attempt to provide
further information on staging prior to the field tests, the location of the
first series of sidewall injectors was changed. The first and last coolant
sections in the experimental combustor were exchanged, thus the sidewall staging
injectors 1, 2 and 3 were situated 2.2, 2.6 and 3.5 combustor diameters from
the fuel injector. No changes were made to the burner, and the conclusions with
this new location of the staging injectors were essentially the same as those
reported in Section 3.0 - modest reductions in NOX emissions were achieved with
attendant increases in smoke emissions.
It is generally accepted that reductions in NOX emissions from No. 6 fuel oil
flames by staging the heat release are primarily due to reducing the net rate of
fuel nitrogen oxidation. This is accomplished because the percentage conversion
33
-------
CM
O
O
O
fc.
i.
O
u
a.
Air Atomizer
100 gallon nozzle
6
3.4 x 10" Btu/hr
P/S 50/50
O Low Swirl
D Medium Swirl
A A High Swirl
Phase II
40
% Excess Air
Figure 4-10. The Influence of Swirl and Excess Air
34
-------
of fuel nitrogen compounds to NO is dependent upon the stoichiometry of the heat
release zones. The general requirements of any staging system are:
a primary region which allows complete reaction of all fuel nitrogen
intermediates under fuel rich conditions;
a secondary region in which the staged air is rapidly mixed with the
products of the primary region, thus providing the maximum opportunity
for carbon burnout since those conditions in the primary region which
restrict NO formation also promote soot formation.
Since the burner was originally designed to satisfy criteria which did not include
staged air addition, it could be expected that the results with staging would not
be optimum unless attention was paid to the fuel/air mixing process in the primary
region. A series of exploratory investigations were carried out in which burner
conditions were varied in an attempt to improve staging performance.
Primary/Secondary Ratio. Muzio et al report that changing the
primary/secondary ratio during staging had a negligible effect upon
NO emissions. Since it has been shown that unstaged NO and smoke
' J\
emissions are lower than baseline emissions (50/50/ P/S ratio) with
primary/secondary ratios less than 20/80, it is reasonable to expect
that some effect of primary/secondary ratio would be evident during
staging. Exploratory measurements confirmed that there was an effect
of primary/secomdary air flow during staging. Lower absolute NOV
X
levels could be achieved at lower smoke numbers with low primary air
flows (see Figures 4-13 and 4-14).
Atomization Pressure. The influence of both atomization pressure and
primary/secondary ratio can be judged from the results presented in
Figures 4-11 and 4-12. With a total of 40 percent of the combustion
air divided equally between staging injectors 2 and 3, (i.e., 70 percent
of the stoichiometric air requirement at the burner) lower NOV emissions
A
at lower smoke numbers was observed at a primary/secondary ratio of
15:85 than at a ratio of 85:15.
Nozzle Capacity,. Table 4-2 compares NOX and smoke emissions for various
staging conditions with three oil nozzle sizes for steam and air atomiza-
tion. Under these conditions, it appears that steam atomization allows
35
-------
300
280
260
240
CM
O
s-s
O
O
u
O
-------
380
CM
O
a*
o
3
u
i
Excess!
3.4 x 106
P/S 85/15
Staged 70% of Stoichiometric air at the
Burner
Ol
XI
OJ
^6
- 4
o
ro
CD
Figure 4-12.
18 20 22
Atomization Pressure, psig
The Influence of Atomization Pressure on
Staging Performance (P/S 85/15)
37
-------
350 r
300
CM
o
ป*
o
s_
i.
o
(J
Q.
CL
250
200
150
a
o
A
Burner
Modified
Modified
plus -15ฐ
vanes
Burner
Stoichi-
ometry
1.17
0.94
1.17
0.94
0.76
D
O
A
i i
20
40 60
Percent Primary Air
80
Figure 4-13. The Influence of Primary Air Percentage and
Swirl Level on NO Emissions (17 Percent
Excess Air 3.4 x 104 Btu/hr Heat Input)
38
-------
10 r
o
CO
o
(O
o
n
Burner
Modified
Modified
plus -15ฐ
vanes
Burner
Sto'ichi-
ometry
1.17
0.94
1.17
0.94
0.76
D
O
A
o
to
CO
20
40
Percent Primary Air
60
80
Figure 4-14.
The Influence of Primary Air Percentage
and Swirl Level on Smoke Emissions (17
Percent Excess Air 3.4 x 104 Btu/hr Heat
Input)
39
-------
Table 4-2. The Influence of Nozzle Capacity on Staging Performance
Percentage of Total Air Added
Through Sidewall Injectors
1
-
20
-
-
20
-
2
-
-
20
-
20
20
3
-
-
-
20
-
20
NO ppm (dry) corr. to 0% 0,
A L.
Air
60 gal.
278
230
216
198
Air
80 gal.
285
242
226
194
194
188
Air
100 gal.
292
258
236
219
225
193
Steam
100 gal.
269
245
216
221
196
Smoke Number
Air
60 gal.
4 1/2
5
8
8 1/2
Air
80 gal.
3 1/2
4
5
8
7
8
Air
100 gal.
4
4
5 1/2
6 1/2
7 1/2
7 1/2
Steam
100 gal.
3
4 1/2
6
6
-------
higher staging levels, and therefore, low NO levels with lower smoke
emissions to be obtained.
The objective of these tests was to demonstrate that staging could be an
effective technique for NOX control provided due attention was paid to burner
conditions. The best results (i.e., low NOY and low smoke) (see Figures 4-13
/\
and 4-14) were achieved by fitting a 15 degree vane swirler to the oil nozzle
to rotate the primary air in the opposite sense to the secondary air swirl.
Forty-five percent reductions in NO emissions from "baseline" conditions were
obtained with an increase in the smoke level of 3.5 to 5.5 on the Bacharach
scale.
Species concentration measurements carried out under another EPA-sponsored
study (EPA Contract 68-02-1500) indicated that rapid mixing of the second
stage air with the partially oxidized products of the primary region was not
being achieved. The sidewall injectors were designed to promote swirling
(3)
second stage air injection* '. This construction directed the air jet away
from the axis of the combustor. Thus, the second stage air jets did not
penetrate to the combustor axis and mixing with the bulk of the partial
oxidation products was delayed.
New sidewall injectors were constructed to inject the staged air directly toward
the combustor axis and Table 4-3 compares emissions with the old and new injector
designs for various staging levels and injection locations. With a burner
stoichiometry of 0.94, the design of the injector had no effect upon the NO
emission. However, for the two cases where the injection took place close to
the burner, smoke emissions were reduced. When 40 percent of the total air
was staged (burner stoichiometry of 0.70) both the smoke and NO emissions were
influenced by the injector design. With the new injector design, NO emissions
were higher when 20 of the 40 percent staging air was added at x/D = 2.2.
After consideration of the results of these laboratory investigations, the staging
system for the field tests was designed with radial staged air injection. The
number of injection points was also increased from four to eight in order to
promote mixing.
41
-------
Table 4-3. The Influence of Method of Staged Air Injection Upon Pollutant Emissions
Percentage of Total Air Added
Through Sidewall Injectors
= 9 7
D
-------
5.0 FIELD INVESTIGATIONS
The laboratory investigations carried out during Phase II did not provide
sufficient information to allow the design of the optimum NO control system
_ A
for package boilers. Nevertheless, the decision was taken to continue with
Phase III of the project - a demonstration of pollutant control techniques in
the field, although it was recognized that the field investigations would be
on a more experimental basis than was originally intended. Two boilers, one
firetube and the other of watertube construction were modified to allow vitia-
tion of the combustion air with cool recirculated flue gases. Staged com-
bustion control techniques were investigated only in the firetube boiler.
5.1 Selection of the Test Boilers
The boilers tested during the field demonstration were both located in the same
boiler house and their choice represents an inevitable compromise between the
ideal and the practically attainable. In selecting the field units the following
criteria were considered to be of particular importance:
information should be provided on both watertube and firetube
designs;
the units to be tested should be typical of modern practice. The
value of the demonstration would be negated if the data were to be
obtained on equipment of outmoded design;
the units to be tested should reflect the bulk of the population
of package boilers both with respect to type and size;
the units tested should be capable of burning both natural gas
and heavy fuel oil;
the same oil supply should be burned in both units;
it must be possible to investigate both flue gas recirculation
and staged combustion techniques in the units;
the owners of the units must be cooperative since the tests could
not be carried out without some interruption of the normal routine;
the cost of the demonstration could not exceed the budget, this
criteria limited both the size and the location of the units which
could be considered.
43
-------
The aid of the American Boiler Manufacturers Association was solicited in order
to determine the type, size and characteristics of the "typical package boiler".
The ABMA were most helpful and provided survey data on sales of both watertube
and firetube boilers. This data was reviewed and is presented in Tables 5-1
and 5-2 respectively. The firetube boiler data is based upon orders placed
within the stated calendar year on high pressure (>15 psig steam) boilers, low
pressure boilers and hot water heaters. Table 5-1 indicates that the major
portion of the firetube population lies in the 100 to 200 hp range (3,450 to
6,900 Ibs of steam per hour). In recent years the bulk of the watertube units
o
ordered lies in the 21 - 40 x 10 Ibs steam per hour range (see Table 5-2).
Several steps were taken to locate units which could be tested and which satis-
fied the criteria discussed earlier. The ABMA, state and local air pollution
regulatory agencies were contacted in an attempt to locate candidate units.
Possible test sites were visited at the Bell Laboratories in Whippany, New Jersey,
and Passaic Pioneer Properties in Passaic, New Jersey. Following these inquires
a series of possible plans were drawn up which had four different approaches:
Test units owned and operated by the Foster Wheeler Energy Corporation;
Rent a firetube boiler for installation near a Foster Wheeler-owned
unit;
Test units located in the vicinity of the Foster Wheeler Energy
Corporation;
Rent both a firetube and a watertube boiler.
The candidate plans which were prepared based upon a survey of the various test
sites are presented in Table 5-3. Certain of these plans were rejected because
the units normally burned natural gas and the cost of conversion to fire fuel oil
was prohibitive. When the two boilers were not located in the same physical
plant the same oil supply could not be guaranteed for both units. The typicality
of the units were also considered, units with rotary cup atomizers and a water-
tube boiler with a water cooled front wall were rejected because these designs
were not typical of the major portion of package boilers. The expense associa-
ted with renting units eliminated those possibilities from consideration.
44
-------
Table 5-1. 10 Year Survey of Package Firetube Boiler Sales
(Including High and Low Pressure Steam and Hot Water Units)
(Supplied by ABMA)
Year
Unit Capaicty
HP (Less than or
equal to)
15
20
25
30
40
50
60
70
80
100
125
150
200
225
150
300
350
400
500
600
>601
Total
No. of Companies
1972
27
63
26
150
164
176
269
83
221
410
350
517
462
13
280
261
150
150
167
195
81
4215
15
1971
31
54
32
142
153
208
301
110
235
458
299
494
479
43
286
290
175
169
190
198
75
4422
15
1970
55
67
54
191
175
222
288
106
264
488
441
490
501
15
301
279
189
171
181
198
56
4732
15
1969
365
112
81
291
257
316
416
177
337
670
518
671
689
40
306
419
224
202
178
293
0
6562
15
1968
42
102
51
235
255
249
348
136
329
645
445
520
514
8
323
337
163
173
190
227
0
5292
13
1967
41
135
53
290
226
274
346
193
366
692
557
483
600
16
340
307
135
132
149
180
0
5517
10
1966
85
150
57
352
363
426
480
169
427
823
587
571
664
40
283
335
173
188
157
197
0
6602
10
1965
82
215
71
409
383
392
474
214
440
785
560
559
629
30
321
328
169
127
163
141
28
6520
11
1964
100
198
113
300
365
341
459
179
426
749
469 .
489
500
5
249
258
151
110
91
116
7
5675
10
1963
83
190
80
345
364
373
450
197
391
676
475
476
534
9
277
219
132
90
86
125
4
5576
10
tn
-------
Table 5-2. 10 Year Survey of Packaged Watertube Boiler Sales
(Supplied by ABMA)
Year
Unit Capacity
103 Ib/hr
10
11-20
21-30
31-40
41-50
51-60
61-70
71-80
81-90
91-100
101-150
151-250
250+
Total
Less than 250 psig,
percent
Sat. Steam, percent
1972
2
72
142
120
78
83
21
25
10
43
48
30
3
677
79
67
1971
3
64
116
101
65
64
19
42
13
29
73
28
2
619
82
90
1970
3
81
149
117
65
91
24
51
14
31
110
38
3
111
75
84
1969
16
140
145
129
90
80
32
51
18
38
114
17
4
873
75
85
1968
7
121
145
98
72
72
33
54
8
43
66
754
76
85
1967
10
118
155
98
63
62
15
34
15
36
39
675
77
79
1966
6
161
199
113
114
69
38
58
9
44
66
921
76
82
1965
6
180
161
101
83
59
29
32
5
51
50
759
76
82
1964
19
159
166
97
101
76
20
21
7
64
27
757
75
83
1963
25
178
132
89
50
38
24
21
7
16
13
593
85
87
1962
23
177
138
89
47
52
5
17
2
17
7
574
80
90
cn
-------
Table 5-3. Possible Candidate Plans
Plan
1. F.W. -Owned Units
2. Rent Firetube
2a. Rent Firetube
3. Test Watertube only
4. Area Location
5. Area Location
6. Area Location
7. Area Location
8. Rent FT and MR
Code
Firetube (FT)
Capacity
103 Ib/hr
5
6.9
6.9
12
20
6.9
Range
6.0
B&W - Babcock & Wilcox
C - Cyclotherm
CB - Cleaver Brooks
ER&E - Esso Research and Engineering
*
Natural Gas Available
Owner
FW*
CB or C*
Uabash*
Essex County*
Correction Center
Passaic*
Pioneer Prop.
Bell Labs
Sandoz
Uabash
Nfg. Location
S Livingston. NJ
CB Dansyille, NY
CB Dansville, NY
S Caldwell, NJ
CB Passaic, NJ
S Whippany, NJ
CB E. Hanover, NJ
CB Anywhere
Watertube (WT)
Capacity
103 Jb/hr
50
50
50
25
25
25
25
75
Owner
FW*
FW*
FW*
FW*
ECCC*
PPP*
BL
ER&E
Wabash
Mfg.
FW
FW
FW
FW
S
CB
N
B&W
M or N
Location
Dansville, NY
Dansville, NY
Dansville, NY
Dansville. NY
Caldwell. NJ
Passaic, NJ
Whippany, NJ
Florham Park, NJ
Anywhere
FW - Foster Wheeler
M - Hurray
N - Nebraska
S - Superior
-------
The units ultimately selected for the field demonstration were located at the
Essex County Correction Center (ECCC) which is plan No. 4 in Table 5-3. This
site had several advantages over other locations, the same oil supply was
assured for both units, the test site was close to the FWEC Research Center and
the units were both of modern design. Although the watertube units fall into
the most popular size range for boilers of this type, comparison with Table 5-1
indicates that the capacity of the firetube boiler is higher than the bulk of
the population; however, it is believed that there will be a shift towards
larger sizes in the future. Unfortunately, during the advanced stages of
preparation it was learned that changes in the hosts operating procedure
would result in a considerable decrease in the steam demand. Maximum fore-
seeable steam demand appeared to be less than 20,000 Ibs of steam per hour and
this would be strongly dependent upon climatic conditions. Recognizing that
this would cause certain difficulties, the API-ERA Steering Committee recom-
mended that the tests be carried out as planned.
5.2 Equipment Used in the Field Investigations
5.2.1 The Watertube Boiler
The watertube boiler tested at ECCC was built by Superior Combustion Industries
and was designed to fire No. 5 fuel oil or natural gas. Natural gas is avail-
able on an interruptible basis. The boiler is fired by a register burner, a
schematic of which is shown in Figure 5-1. The natural gas is injected from a
gas ring and the oil gun utilizes steam atomization. The front wall of this
"D" type forced draft unit is refractory lined and the gas passage is horizontal.
The combustion gases exit the radiant section, turn through 180ฐ before passing
through the convective section. The unit is rated at 25,000 Ibs of steam per
hour at 250 psig, and no provision if available for air preheat.
The flue gas recirculation system was designed to recirculate 30 percent of the
full load combustion products through the wind box. The system installed in
the watertube boiler is shown in Figure 5-2. The recirculation fan was located
at grade level between the two watertube boilers adjacent to the stack breeching
for both boilers. The flue gases were withdrawn before the boiler breeching
dampers and the 12 in. ID ducts carried the gases over the boiler before enter-
ing the windbox normal to the downward flow of the combustion air. The ductwork
48
-------
Gas Ring
Diffuser
Atomizer Tip
Furnace Wall
Figure 5-1. Register Burner Installed in the Watertube Boiler
49
-------
PLAN VIEW
en
O
-Q
,O
:0 :
SIDE VIEW
FRONT VIEW
Figure 5-2. Layout of Flue Gas Recirculation System for the Watertube Boiler
-------
was fabricated from rolled No. 10 gauge carbon steel plate with longitudinally
welded seams. Fabricated ductwork was used because of the difficulty in
obtaining light wall pipe or tubing within the available time. The mass of
recirculating combustion products was controlled by a damper installed at the
immediate discharge of the fan and metered by an ASME standard orifice installed
in the horizontal duct passing over the boiler.
5.2.2 The Firetube Boiler
The application of both flue gas recirculation and staged combustion control
techniques were investigated in the firetube boiler which was also manufactured
by the Superior Combustion Industries and a sketch of the burner firetube
arrangement is shown in Figure 5-3. The boiler was designed to burn No. 5 fuel
oil with an air atomized tip. Natural gas can also be burned and it is injected
from a ring embedded in the refractory throat and the boiler is refractory lined
for the first 2.5 feet. The unit has four gas passes, one radiant and three
convective passes. The firetube and second pass are surrounded by water and
the third and fourth passes lie in the vapor space at the top of the boiler.
The forced draft unit is rated at 12,000 Ibs of steam per hour at 250 psig.
Details of the flue gas recirculation system installed in the firetube boiler
are presented in Figure 5-4. All ductwork was 10 in. ID and the control and
metering system was similar to that used on the watertube boiler. The recircula-
tion fan was placed on the grating above and to the rear of the boiler. Com-
bustion gases were withdrawn from the stack at this level and entered the wind-
box tengentially in the same flow direction as the combustion air from the
F.D. fan. When the windbox was breeched to accept the recirculation ductwork
a baffle plate, not detailed on any Superior drawings, was found in the windbox.
This baffle plate reduced the volume of the windbox and also obstructed the
recirculation gas entry. Since at that time no function could be attributed to
this plate, it was removed, although subsequent experience appeared to indicate
that this was an error.
The only practical entry for the second stage combustion air supply was through
the rear of the firetube unit, although this was made difficult because the
boiler backed up to a 3 ft thick wall leaving only limited space for access.
51
-------
Gas
Combustion
Air
~TF^
L Uvv-r-
.-*.';
; ป
ป ' *
f i
ป ' ' ' ป/
.-' , ป.
' ป '". .
*'"'"* '
"l ""* 1
"*ซ i ป"*! <
- V * '
/
"ป:.- vV.'J
t
Gas
Injector
Oil Nozzle
Stabilization Vanes
Figure 5-3. Sketch Showing the Windbox Burner Arrangement
of the Firetube Burner
52
-------
HI
PLAN
en
CO
>r
HI ^ ^" n
N_
^ -
/
t=fO
FRONT ELEVATION
SIDE ELEVATION
Figure 5-4. Layout of Flue Gas Recirculation System for the Firetube Boiler
-------
Penetration of the front wall was rejected because it would necessitate cutting
through 30 inches of refractory in the form of three cast refractory rings.
The general arrangement of the staging system installed in the firetube boiler
is shown in Figure 5-5. Ambient air was supplied by a separate fan to a distri-
bution ring at the rear of the boiler. This ring supplied eight 2-inch stainless
steel staging pipes which entered the firetube through the rear door. These
pipes were laid along the wall of the firetube and provision was made to allow
the axial location of the staging injectors to be varied in later experiments.
The staged air was injected radially through fishtail orifices. The configura-
tion of the staged air injectors and reasons for the design will be discussed
later. Burner stoichiometry was varied by throttling the combustion air supply
and maintaining a constant overall excess air by increasing the air flow through
the staging injectors.
5.2.3 Automatic Controls for Flue Gas Recirculation
Firetube Boiler
The control logic for automatic operation of the firetube boiler is a motor
driven mechanical linkage activated by a pressure signal. As both air and fuel
quantities are regulated directly by the linkage, the controls for the flue
gas recirculation system were also tied directly to the linkage. This was
accomplished by incorporating a cam-follower mechanism on the modulating motor
of the linkage. The resulting signal activates a ratio control which maintains
a constant ratio between air/fuel and FGR. Another signal from a differential
pressure call reading pressure drop across an orifice, in the FGR duct, is fed
to the ratio controller and compared with the signal from the cam-follower. If
these two signals do not balance, a signal is sent to another controller which
resulates a butterfly valve in the FGR duct to increase or decrease flow. In
this manner, a flue gas recirculation flow is established in proportion to the
air and fuel.flow as signaled by the cam-follower position. The ratio of
recirculated flue gas to air and fuel may be regulated by the ratio controller.
A schematic of this system is shown in Figure 5-6.
54
-------
3=-
PLAN
DETAILS OF STAGING NOZZLE
en
en
ELEVATION
Figure 5-5. Details of Equipment Used in the Staging Investigation
-------
en
en
Ratio Con.
Input
Out.
Process
Input
Input ซ
> Gauge
Refer. <
Output <
h
H^
\
s s
J J
Nullmatic
(Stack)
Controller
Ratio
Controller
i Input Signal from Firetube
-I <
l_
Input Signal from Watertube
a
DP Cell
Auto-
Manual
Station
Solenoid Valves
FGR Duct
\ Butterfly Valve
^ Orifice
Figure 5-6. Schematic of Automatic Controls for Flue Gas Recirculation
-------
VJatertube Boiler
The control system for the watertube boiler is similar to that of the firetube
with the exception that the initial control signal originates from a pneumatic
source indicating total air flow to the boiler. In this case, if the boiler
load were to change, the fuel flow would change correspondingly causing an air
flow change and, thus, an F6R rate change. The schematic diagram for this
system is the same as for the firetube.
5.2.4 Flue Gas Measurement Systems
The sample was withdrawn from a port in the flue of each boiler via a 0.5 in.
O.D. x 0.049 in. wall stainless steel tube and passed to a condensate trap
immersed in an ice ba,th. The cooled sample gas was supplied via a 1/4 in. O.D.
teflon tube to the instrument manifold. The instruments used to continuously
monitor the concentration of several combustion products are listed below; in
some instances backup instruments were employed to ensure continuity if one
instrument failed during any particular test.
NO/NO . The primary instrument used to determine both NO and NOX
was a Thermo-Electron Chemiluminescence Analyzer, backup measurements
were made with a Theta Sensor US-6000 analyzer.
0?. Oxygen concentrations were measured by both the Theta Sensor
anialyzer and a Teledyne Model 320 AX Portable Analyzer with a Class
A-3 cell, the former instrument being used as the primary reference.
CO. An infrared absorption analyzer (MSA Model LIRA-303) was used to
determine the carbon monoxide content of the flue gases.
S0?. The Theta Sensor US-600 was used to determine the sulfur dioxide
coritent of the flue gases.
Smoke. Smoke readings were taken with a Bacharach smoke tester.
57
-------
Certified .zero and calibration gases were used to calibrate these monitors
throughout the investigation. Frequent calibration checks were used to ensure
the reliability of the data.
The instrumentation available in the boiler house was used to establish the
boiler load. Although this was adequate for normal operation, more precise
information of fuel and steam flow would have been desirable for this
investigation.
5.3 Result of the Field Investigations
The field investigations were carried out in three separate periods punctuated
by the need for equipment modification. These three different test periods
correspond to
Boiler Performance Tests;
* Flue Gas Recirculation Tests; and
t Staged Combustion Tests in the Firetube Boiler.
Although a rigid test matrix had been agreed upon for the baseline tests,
inability to control excess air level and boiler demand required some relaxation
of the test matrix. The exploratory nature of the FGR and Staged Combustion
tests necessitated that the test program depend largely upon the initiative and
the experience of the test supervisor. All the results for the three sections
of the field investigation are presented in Appendix 1.
5.3.1 Boiler Performance Tests
The boiler performance tests were carried out to determine the influence of
operational parameters on pollutant emissions from the two test boilers. These
tests not only established a baseline against which to judge the effectiveness
of the various control techniques, but also allowed an assessment to be made of
whether or not their performance was typical of the total set of package boilers.
The operational parameters investigated were:
ง Fuel type - natural gas and No. 5 fuel oil;
Load - the range was dependent upon demand and boiler characteris-
tics (limited to 70 percent full load for the watertube boiler by
maximum possible demand);
58
-------
Excess air - a wide variation dependent upon fuel type and load; and
Burner parameters - register setting, oil temperature, steam
pressure.
Firetube Boiler
The emission characteristics of the firetube boiler fired with natural gas and
fuel oil can be judged from Figure 5-7. A11 the test data is presented in
Tables 1 and 2 in Appendix 1. With liquid fuel, NO emissions appear to be
X
insensitive to load at low and medium loads but emissions increase as the load
is increased to maximum. Smoke emissions appeared to be insensitive to load
because of load demand, however, at low and medium loads NO emissions increase
X
with increasing load. The difference in behavior of the firetube boiler with
the two fuels is probably due to small fraction of the total emission attributable
to thermal NO for oil firing at low load. Emissions from fuel oil flames showed
a stronger dependence upon excess oxygen than did emissions from gas flames,
although it should be noted that the oil was burned satisfactorily with a lower
level of excess air which probably reflects the less effective fuel/air mixing
obtained with natural gas firing. Only one test was carried out to investigate
the influence of oil temperature variation (20ฐF variation); virtually no effect
upon either NO or smoke was observed.
X
Watertube Boiler
The results for the watertube boiler fired by oil tend to exhibit more scatter
than those reported for the firetube boiler as shown in Figure 5-8. Contrary
to the trends found in the firetube boiler, NO emissions increase with
/\
increasing load for low and medium loads and then show a slight decrease as
the load is increased still further. Several examples can be seen in Figure 5-8
where data is plotted for the same nominal load and 50 ppm difference in emis-
sions can be seen for almost the same oxygen concentration.
Figure 5-9 shows the influence of register setting on NOX and smoke emission
for a fixed load. Closing the register causes an increase in the rotational
motion of the combustion air flow, but it also increases the burner pressure
59
-------
C\J
o
o
o
o
u
Q.
Q.
260
240
220
200
180
160
140
120
100
Fuel
Oil
Gas
Steam flow
mlb/hr
5
7.5
11.0
1.0
4.0
6.5
8.5
N0x
O
D
A
V
0
O
0
Smoke
A
10
.G
U
a
CO
H
D
o
10
12
O
2 % by Vol
Figure 5-7. The Influence of Load and Excess Air on Pollutant Emissions
from the Firetube Boiler (Boiler Performance Tests)
60
-------
260
240
Watertube Boiler
No. 5 Oil
Base Line Tests
Fuel
Oil
Steam Flow, Mlb/hr
7.5
10
12.5
14.5-15
18.5
NO
X
o
n
V
A
O
Smoke
V
A
OJ
o
o
o
s-
o
o
o
Q.
Q.
220
200
180
160
140
120
100
80
60
40
20
0
0
Figure 5-8.
4 6
O2, % by Volume
10
The Influence of Load and Excess Air on Pollutant
Emissions from the Watertube Boiler (Boiler
Performance Tests)
10
9
o
cu
1
1 ง
61
-------
Watertube Boiler
140
120
Load: 14 ,500 Ib Steam/hr
Fuel: No. 5 Oil
NO Smoke
X
A A
1.5
CSJ
O
o
o
J_
i-
o
u
ฃ
Q.
Q.
100
-i 10
80
60
40
20
40
Figure 5-9.
60 70 80
Air Register Setting, % Open
90
The Influence of Register Position on Pollutant
Emissions - Watertube Boiler (Boiler'Performance
Test Figures Beside Symbols Denote 02 Percentage)
O
HJ
PQ
6
3
_4 oj
o
ra
100
62
-------
drop which tends to reduce the total air flow. Thus, as can be seen in
Figure 5-9, the flue gas oxygen content drops as the register closes and the
resulting change in emissions cannot be attributed to one effect. Decreasing
the excess oxygen content from 3 to 2 percent at fixed register setting
reduces NO emissions by approximately 40 ppm at 14,500 Ibs of steam per hour
A
(see Figure 5-8). Thus, it appears that closing the register at a constant
excess air level could cause a slight increase in NO emissions since the
A
decrease in emissions caused by the reduction of excess air alone is greater
than that produced by the combined effect of excess air and register setting.
Wide variations in oil temperature and steam pressure were not possible due to
operational limitations. Also, variations in these two parameters were
accompanied by unexplained changes in flue gas oxygen concentration. It
appears that reductions in fuel oil temperature cause a reduction in NO emis-
fc A
sions (Figure 5-10). Interpolating information from Figure 5-7 suggests that
NO emissions are reduced by reduced atomizing steam pressure (Figure 5-11).
/\
The influence of load on emissions from the watertube boiler when fired with
natural gas is rather erratic (see Figure 5-12). At 4.5 percent oxygen, maxi-
mum emissions occur at medium loads. Emissions at low and maximum load are
almost the same. Certainly the peak emission occurs at different excess air
levels for different loads. Closing the register and increasing the swirl
causes an increased emission with natural gas even though the excess air was
reduced (see Figure 5-13). This result can be attributed to an improvement in
fuel/air mixing which causes an increase in the NO emission. Visual observa-
** X
tions tend to support this argument since under normal operating conditions the
flame could be described as "soft", indicative of slow air/fuel mixing.
63
-------
200
175
ISOJk
125
o
o
S-
o
o
Q-
Q.
100
-c
75
50
25
A.5
110
Watertube Boiler
Load: 10, 000 Ib Steam/hr
Fuel: No. 5 Oil
NO Smoke
x
a
i.6
120
130
140
150
160
Fuel Temperature, F
10
Si
o
10
CD
u
cu
.0
e
3 6
CO
170
Figure 5-10. The Influence of Fuel Oil Temperature on Pollutant
Emissions from the Watertube Boiler (Boiler
Performance Tests)
64
-------
CM
O
0
o
i.
o
o
260
240
220
200
180
160
140
120
100
80
60
40
20
0
Watertube Boiler
Load: 10,000 Ibs Steam/hr
Fuel: No. 5 Oil
Approx. Emission
at 7.0% O
from Figure 5-8
'.0
' Approx. Emission at 3.1%
from Figure 5-8
3.1
D Measured NO
[J Estimated NO
Measured Smoke
_L
J.
-L
20 30 40 50
Atomizing Steam Pressure, psig
60
10
5 "5
(D
CQ
E
3
X
0)
JK!
o
s
co
1
0
Figure 5-11.
The Influence of Atomizing Steam Pressure on
Pollutant Emissions from the Watertube Boiler
(Boiler Performance Tests)
65
-------
120
110
CM 100
o
o
o
o.
a.
x
o
90
$_
o 80
o
-a 70
60
50
40
30
20
10
Watertube Boiler
Natural Gas
Baseline Tests
Legend
Steam Flow Mlb/hr
8.5
10.5
13.0
16.0
20.0
NO
X
k
D
O
A
O
Smoke
A
10
11
10
9
c "0
5 03
cu
ft
4 6
3
cn
O0 , % by Volume
Figure 5-12.
The Influence of Load and Excess Air on NOX
Emissions from the Watertube Boiler (Boiler
Performance Tests)
66
-------
Watertube Boiler
Natural Gas
Baseline Tests.
Numbers Beside Symbols Indicate O, Value
CO
o
o
o
O
o
Q.
Q.
X
O
100r
80
60
40
20
0
40
1.1
_L
50 60 70 80
Air Register Position, % Open
_ 10
90
u
(0
-------
5.3.2 Flue Gas Recirculation Tests
The extent of the test matrix for the flue gas recirculation tests was dictated
by the range of load and excess air levels that were practically attainable.
Flue gas recirculation tests were performed on both boilers firing oil and the
extent of the tests can be judged from the matrix presented in Table 5-4. As
part of the flue gas recirculation tests, emission data was obtained without
recirculation, thus allowing a comparison to be made of the boiler performance
before and after modifications. A change in the emission characteristics of
both boilers was observed. Emission levels were, in general, found to be
lower after modification. A more detailed discussion of these baseline
changes is presented in Appendix 2, which are confused because of contradictory
fuel analyses for nitrogen content. In most instances measured NO and smoke
/\
emissions were lower after modifications had been made to the boilers. No
explanation is available for the observed shift in baseline emissions, although
several possibilities are listed in Appendix 2:
changes in fuel properties;
errors in analyses; and
real changes due to modifications carried out to the boilers.
Firetube Boiler
The influence of flue gas recirculation on both NO and smoke emissions from the
A
firetube boiler at nominal loads of 4,000, 6,200 and 10,000 Ibs steam per hour
can be seen in Figures 5-14, -15 and -16. It can be seen that the addition of
flue gas to the combustion air had a significant influence on NO emissions at
/\
all loads. Smoke emissions were low (<2 Bacharach) for most conditions.
Excessive smoking conditions were only observed at high load and low excess air
levels. Flue gas recirculation did not reduce smoke emissions; in general,
smoke emissions tended to increase slightly. This effect was also observed in
the laboratory investigations.
68
-------
Table 5-4. Flue Gas Recirculation Tests
Boiler
Firetube
Watertube
Nominal
Load
MLB/Hr
4.0
(Low)
6.2
(Medium)
10.0 + 1.0
(High)
6.5
(Low)
10.0
(Medium)
15.5 t 0.5
(Medium)
Nominal
02
% Vol.
2
4
6
4
7
2.5 + 0.5
3.5
6
3.0
4.6
2.0
3.3
2
3
4.1
5.2
Nominal Flue Gas Recirculation Percentages
0, 20, 30, 40, 50
0, 20, 30, 40
0, 20, 20, 40
0, 20, 20, 35
0, 10, 30, 40
0, 10, 15, 20
0, 10, 20, 25
0, 10, 20
0, 10, 20, 25, 30
0, 10, 20, 25
0, 10, 20, 25, 30
0, 10, 20
0, 10, 20, 30
0, 10, 15, 20
0, 10, 15, 20, 25
0, 10
CTl
VO
-------
280
240
200
C\J
O
s-a
o
I
u
a.
0.
NO Smoke Nominal 0?
A t
40
20 30 40
% Flue Gas Recirculation
50
-C
o
s-
10
(J
CO
I
s_
Ol
-Q
-------
280
240
NO Smoke Nominal Q
O
A A 00 = 7.
CM
O
o
o
o
u
Q.
Q.
10
10
20 30
Flue Gas Recirculation
o
to
i.
to
x:
o
to
CD
5-
O)
-Q
E
HI
J*ฃ
o
CO
Figure 5-15.
The Influence of FGR on NOX and Smoke Emissions,
Firetube Boiler, 6,200 Ibs Steam per Hour
71
-------
280
240
200
CM
O
O
O
o
O
160
ex
Q-
NO Smoke Nominal 0?
Jv ^
120
10
20 30
Flue Gas Recirculation
40
50
Figure 5-16.
The Influence of FGR on NOX and Smoke Emissions,
Firetube Boiler, 10,3000 Ibs Steam per Hour
72
-------
One interesting feature of these results is that under certain boiler loads and
excess air levels the influence of flue gas recirculation does not appear to be
tailing off at high recirculation rates. This effect is contrary to the observa-
tions in experimental combustors and of most other workers. This effect could
be attributed to reduced ignition stability. The ignition zone could be moving
downstream as the amount of recycled flue gases is increased.
Hatertube Boiler
As stated previously, reductions in the total steam demand due to procedural
changes limited the extent of the testing with the watertube boiler. Recircula-
tion rates in the watertube tests were limited by ignition instability which
occurred around 25 percent recirculation. Significant NO reductions were not
A
obtained by adding cooled combustion products to the combustion air in the
windbox of the watertube boiler. Low recirculation rates often produced an
increased emission. The results for three boiler loads are presented in Fig-
ures 5-17, -18 and -19. Smoke emissions were not increased by flue gas
recirculation. In some instances smoke emissions were reduced by the addition
of small quantities of recirculation which also caused a reduction in carbon
monoxide emissions. NO emissions from the watertube boiler are somewhat lower
A
than might be expected. This can be attributed to the relatively low combustion
intensity. Visual observations suggest that fuel/air mixing rates are low,
producing a "loose soft" flame. The increased burner pressure drop due to the
addition of recirculation improved the oxygen/fuel mixing as indicated by the
reduced smoke and carbon monoxide emissions.
5.3.3 Staged Combustion
The original concept for the staged combustion investigations was that they
would represent a test of a commercial system. The laboratory results did not
provide sufficient data for such a system to be designed with any reasonable
probability of success. Consequently, the staging air delivery system was
constructed to include considerable flexibility by allowing the axial location
of the injectors to be varied. Time and funds restricted the tests in this
program to a single location; however, the investigation will be extended in
the future under EPA Contract 68-02-1498.
73
-------
140
120
CM
O
O
O
S-
o
u
a.
ฃ3.
100
80
60
40
20
NO_ Smoke Nominal O
D
A A O = 4.6%
Figure 5-17.
20 30
% Flue Gas Recirculation
The Influence of FGR on NOX and Smoke Emissions,
Water-tube Boiler, 6,500 Ibs Steam per Hour
74
-------
140
120
o0" 100
o
o
o
o
ฃ
o.
Q.
80
60
40
20
NO Smoke Nominal O
X
Figure 5-18.
o
to
its
co
O
20 30
Flue Gas Recirculation
The Influence of FGR on NOX and Smoke Emissions,
Watertube Boiler, 10,000 Ibs Steam per Hour
75
-------
140
120
100
C\J
o
o
o
S-
s_
o
o
a.
o.
80
60
40
20
NOx Smoke Nominal
20 30
% Flue Gas Recirculation
Figure 5-19.
The Influence of FGR on NOX and Smoke Emissions,
Watertube Boiler, 16,000 Ibs Steam per Hour
76
-------
The experimental data obtained during the staging investigations is tabulated
in Appendix 1. The staged air was added 1.5 firetube diameters downstream
from the burner tip. This location was chosen to prevent excessive smoke forma-
tion based upon the results of the laboratory investigations. NOX data for one
boiler load (6,000 Ibs steam per hour) and several excess air levels are pre-
sented in Figure 5-20 where it can be seen that the overall excess air has only
a slight effect upon total emissions (c.f., 2 percent 02 and 5 percent Oj.
Almost 50 percent reduction in emissions were obtained without any attempt to
optimize the system. Laboratory investigations had shown that optimum burner
conditions for unstaged operation (e.g., air distribution, atomization condi-
tions) might not be optimum for staged operation. Although smoke emissions
increased with reduced burner stoichiometries, they were only excessive (i.e.> 4)
in two of the tests (see Figure 5-21).
The staged combustion investigations were limited to one load because
combustion instabilities were encountered at high loads as the air flow through
the windbox was reduced.
5.4 Operational Experience
One of the objectives of this field demonstration was to try to identify some
of the problems which must be solved before retrofit of package boilers for
NO control can be considered. Since the boilers tested were "handpicked", it
J\
is reasonable to assume that the problems associated with modification would
not be exaggerated even though the suitability for control was not a major
criteria in this selection. Problems which might be encountered during retrofit
for NO control can be divided into two groups: those which can be solved by
A
adequate planning, and those which will only be uncovered during operation of
the system. Naturally, this is an oversimplification since it will also depend
upon the definition and the extent of the planning task. Those problems which
can be included in the former group are:
t Limitations of Available Space. The equipment associated with
the control technique must be designed to operate in an existing
and often confined space. Thus, the siting of fans and ductwork
are crucial, also the installation of the control system should
not hinder the normal operation of the boiler.
77
-------
Firetube Boiler
00
CVJ
o
O
O
4J
*
i.
O
o
13
Q.
CL
X
o
zzu
200
180
160
140
120
100
80
60
40
20
0
1!
w
A
A
-
ฐ2
1.4 - 2.4 O
2.5 - 3.4 ^7
3.5 - 4.4 D
4.5 - 5.4 O
5.5 - 6.4 O
6.5 - 7.4 A
M
A
^^
1 1
SO 140 130
QLayeu rumy
A
ฐ" -<8ฐ
Nominal O9 Cone. Q ^
^ -^ r 1
LJ_-/
2% D >y
0 900nCb
4% O
O
5%
6%
7%
l 1 I I I I I
120 110 100 90 80 70
% Stoich. Air at Burner
Figure 5-20. The Influence of Staging on NOX Emissions from the Firetube Boiler (6,000 Ibs of Steam per Hour)
-------
Firetube Boiler
Staged Firing
10 r-
JS
u
2
<0
o
<0
CQ
ft
ฃ
3
o
E
CO
1.4 - 2.4
2.5-3.4
3.5 - 4.4
4.5 - 5.4
5.5 - 6.4
6.5 - 7.4
O
V
D
O
O
A
D
V O
O
I
D
OD
COO D D D
O
o
150
140
130
120 110 100 90
Stoichiometric Air to Burner, %
80
70
60
Figure 5-21. The Influence of Staging on Smoke Emission from the Firetube Boiler (6,000 Ibs of Steam per Hour)
-------
0 Minimum Downtime. Boiler downtime can be minimized by adequate
planning, but for some period of time the boiler must be taken
out of service.
Since the boilers at ECCC were chosen because of their suitability for the
project, it is safe to assume that the problems which could be overcome by
planning would be minimized. It was stipulated initially that there would
be no cutting and welding of pressure parts at ECCC, and therefore, problems
associated with this subject could not be uncovered.
Several problems which fall into the latter group, operational problems, were
found during the investigations at ECCC. A major problem with the operation
of the firetube boiler was believed to be associated with the removal of the
baffle plate which was found when the windbox was opened to install flue gas
recirculation ducting. After removal of this baffle, serious instability
problems occurred during normal boiler operation with FGR. The instability
problem was only alleviated when almost the whole of the baffle plate was
replaced. Flow straightening vanes placed in the FGR entry had no beneficial
effect. Finally,'the boiler vibrations were reduced to an acceptable level
at most loads when an opening, equal to the area of the FGR duct, was left in
the baffle. However, there were still certain conditions under which vibra-
tions became excessive. It should be noted that it cannot be stated with
absolute certainty that instability problems were unknown before the boiler
was modified to accept flue gas recirculation.
High speed movie films taken during staging indicated that an intermittent
"flashback condition" occurred at certain loads and excess airs. It could not
be ascertained as to whether this was due to fluctuations in the oil supply
pressure or to the reduced air flow through the burner throat. Ignition
stability problems were encountered with the watertube boiler at flue gas
recirculation rates in excess of 25 percent at most loads.
Potential long-term problems due to equipment deterioration were not found;
however, it should be noted that none of the equipment was used continuously
for a sustained test.
80
-------
6.0 DISCUSSION AND CONCLUSIONS
6.1 Typicality of Field Test Units
Cato et an have provided a considerable body of data on the pollutant
emission characteristics of industrial boilers. Considerable effort was
expended during the present investigation to ensure that the units tested in
the field were typical of the whole class of package boilers. Figure 6-1
presents the baseline data obtained by Cato et al for boilers similar in size
to those tested at the Essex County Correction Center. Baseline results from
the laboratory combustor are also included. The boilers tested in this study
appear to have similar characteristics to a wide range of boilers. Emissions
from the ECCC firetube appear to be in the higher range for both No. 5 fuel
oil and natural gas; whereas those from the watertube boiler appear to be in
the lower range.
One of the difficulties associated with comparing the< data from liquid fuel
(8)
fired equipment is the nitrogen content of the fuel. Studies by Barrett et ar
and Cato et ar ' give regression equations relating NO emissions and fuel
nitrogen content. These relationships can be compared with the baseline emis-
sions measured at ECCC in Figure 6-2. As noted previously, emissions from the
watertube boiler appear to be low. Also, flue gas recirculation had only a
minor influence on NO emissions, indicating that thermal NO formation was
X A
probably low.
The general conclusions of Cato et al which relate to this present study are:
NO emission from natural gas fired boilers are weakly dependent
upon excess air and normally range from 50 to 120 ppm dry corrected
to 3 percent 02;
There does not appear to be any significant difference between
NO emissions from firetube and watertube boilers in their common
/\
size range;
Decreased oil temperatures tend to increase NOX emissions.
81
-------
300
CM
O
4J
c
O
s_
ฃ 200
n
E
a.
O.
100
J
t
,
1
1
(
ฐ
"
o
Z x^ _
I ฐ
1 0 0
!"ปป
'JZ.
Boiler
Construction
Water-
tube
D
O
m
ฉ
Fire-
tube Fuel Source
0 Gas Ref. 6.1
0 Oil Ref. 6.1
@ Oil Phase II
(U Gas ECCC
ฉ Oi 1 ECCC
S
C
2
o
i
\
i r 0
1 000
0
0
(5) j^k
ฉ
i
*imif .. 1 ^^^
00
S 1 8
z m H 0
U^v ^^1
0 12
ra
LL
ffii ^^
*gr -
0
!ป u
0 0
o
3 0 ^
0 0 g
O
,
D
3 a
3
10
20
Ibs steam/hr x 10
-3
30
Figure 6-1. Comparison of Boiler Performance Data with that of Cato et al
(7)
82
-------
CVl
o
o
o
o
u
240
220
200
180
160
140
120
100
/
ECCC Firetube /
at
T I
/
/
: 3% 0? 7 /
J// .
s.
1
1
/ /
//
/
r
v_
/ /
V /
A. "0
ECCC
)
/
f
,
/____ NO ซ 104
/""""" + 834.7 N^6'2)
/
/
.
46% Fuel
T Range reported
by Cato
. !et al(6-D -
N Conversion
+ 105 ppm Thermal NO
: Watertubi
a
0.1 0.2
Percent of Nitrogen Content
Figure 6-2. Relationship of Fuel Nitrogen Content and NO Emissions
from Industrial Boilers
83
-------
Experience in both the field tests and the laboratory investigations appears
to be in direct conflict with the last conclusion.
As discussed in Appendix 2, the nitrogen content of the fuel oil used in the
field tests is open to question. However, comparison with the available data
suggests that the emission characteristics of the firetube boiler are typical
of that class of equipment. Even recognizing that the watertube boiler was not
tested at full capacity, it appears to be a naturally low NO emitter. It is,
/\
of course, not possible to ascertain as to whether the conversion problems
encountered at ECCC are likely to be representative. The boilers chosen for
testing were done so with a knowledge of the requirements of the control system
to be installed and yet problems were uncovered which could not have been antici-
pated. There is no reason to believe that the two boilers tested in the field
represent special cases, and it must be expected that similar problems would
occur with other units.
6.2 Comparison of Laboratory and Field Test Results
A detailed discussion of the mechanisms of nitric oxide formation in turbulent
diffusion flames is outside the scope of this report. Nitrogen oxides are formed
from two sources of nitrogen during the combustion of fossil fuels, molecular
nitrogen, and nitrogen compounds which occur naturally in both liquid and solid
fuels. The reactions controlling the rate of oxidation of molecular nitrogen,
producing thermal NO are strongly temperature dependent and only proceed at
significant rates above 1600ฐC. It was originally thought that the reaction
between nitrogen molecules and oxygen atoms was mainly responsible for NO
production in flames. However, it is now known that hydrocarbon radicals formed
in flame zones also provide a path for thermal NO formation. The conversion of
fuel-bound nitrogen producing fuel NO depends upon the nitrogen content of the
fuel and upon oxygen availability. The amount of both fuel and thermal NO is
strongly dependent upon the rate of fuel/air mixing and bulk gas temperatures
which are functions of the combustion system. The rate at which the fuel and
air are mixed is controlled by burner design parameters and the bulk gas
temperature in the region of interest is dependent upon the volumetric heat
release rate and the temperature of the enclosure. Consequently, it is readily
84
-------
apparent that NO formation in turbulent diffusion flames is system-dependent and
detailed comparisons between the laboratory and the field is difficult.
The laboratory combustor was designed as a firetube simulator. Consequently,
it would be expected that similarities could be found between the results
obtained with the firetube boiler in the field and the laboratory combustor.
However, the watertube boiler has several important characteristics which dif-
ferentiate it from the laboratory combustor:
it is three dimensional and not axisymmetric;
the burner has a register and is very different from either
of the firetube boiler burners;
the flame is less confined by the boiler walls; and
visually, the flame in the watertube is of low intensity.
Thus, it would be expected that the emission characteristics of the watertube
boiler would not be simulated by the laboratory combustor.
Figure 6-3 compares the fractional reduction in NOV emission achieved by flue
X
gas recirculation. The data presented includes that taken in the field test,
the laboratory investigation as well as comparative data from other works:
Curve 1 presents Phase II laboratory data for a No. 2 fuel oil
(nitrogen content 0.05 percent).
Curve 2 presents Phase II laboratory data for a No. 6 fuel oil
(nitrogen content 0.36 percent).
(9)
Curve 3 is taken from the work of Armento and Sage and relates
' to experiments conducted in a circular tunnel furnace with a
register burner and No. 6 fuel oil (nitrogen content 0.23 percent)
but with preheated air.
(10)
Curves 4a and 4b are taken from the work of Turner et al and
were obtained in a 50 HP Cleaver Brooks Boiler for high nitrogen
and low nitrogen oils (curve 4a 0.77 percent nitrogen, curve 4b
0.03 percent nitrogen).
85
-------
O Essex County Watertube
D Essex County Firetube
Curve 3
\
Curve 4a
Curve 1
Curve 4b
D
0.6
20
40
Percent FGR
Figure 6-3. Fractional Reduction of NO Achieved by FGR Comparison
of Field and Laboratory Data
86
-------
Although Martin and BerkaiT11' found that FGR slightly reduced fuel NO
emissions, its primary effect is the reduction of thermal NO. It is difficult
to draw general conclusions based upon the results of one field test. How-
ever, the comparisons presented in Figure 6-3 strongly suggest that for fire-
tube boilers burning No. 5 or No. 6 fuel oil, a 30 percent reduction in
emissions could be expected with approximately 40 percent FGR. The absolute
reduction would depend upon the amount of refractory in the firetube- Larger
reductions could be expected for No. 2 fuel oil.
It is a gross oversimplification to state the FGR will only eliminate thermal
NO since the increased burner throat velocity will influence the rate of fuel/
air mixing which could also influence fuel NO formation. The addition of FGR
to the windbox of the watertube boiler had very little effect upon NO emissions.
Indeed, it actually increased NO emissions under certain circumstances. It is
contended that this particular burner/boiler combination has low NO characteris-
tics by virtue of the slow rate of fuel/air mixing (confirmed by visual observa-
tions) and generous furnace volume. Virtually the total emission can be attri-
buted to fuel NO and the increased emission with the addition of FGR is due to
improved mixing caused by the increased burner pressure drop. When comparing the
ECCC data with that of Armento and Sage it should be remembered that the experi-
mental tunnel was partially refractory covered and the air was preheated. Both
of these factors would tend to increase the amount of thermal NO formation.
Therefore, in watertube boilers without preheat, it is reasonable to expect that
a 15 percent reduction in NO could be obtained by recycling 40 percent of the
flue gases when-firing fuel oil.
As discussed earlier, the performance of the laboratory system during staging
could be improved by optimizing the burner conditions. It is encouraging that
the field tests were successful even without making changes to the burner. The
improved performance can be due to:
a burner system which was initially more suitable for staged
combustion;
improved design of the staged air injectors;
the second stage was heated before injection which would tend to
prevent chilling and help carbon burnout.
87
-------
6.3 Cost_of_Emiss i on Con trol
It is most difficult to accurately assess the cost of NO emission
X
control for residual oil-fired packaged boilers based solely upon the experi-
ence gained during this program. Four different situations can be envisaged
in which the cost of additional pollution control equipment will be different;
these are:
1. The modification of field operating boilers in a similar way
to the exercise carried out at ECCC.
2. Shop retrofit of a new or used boiler prior to use in the
field.
3. Manufacturer incorporation of the additional equipment in a
new boiler.
4. A new boiler design which is dictated by the requirements of
the pollution control equipment.
It is not possible to assess the costs associated with the fourth situation
and there is a considerable degree of uncertainty associated with the esti-
mates for the other three possibilities. Presented below are the actual
costs for the modifications carried out during this program. These costs
are naturally high because they reflect a necessary learning experience.
Having gained this experience, future work of a similar nature would be
less costly. An attempt has been made to estimate the costs associated
with the third alternative listed above and the most uncertain figure in
these estimates is that associated with design costs since this will depend
upon the frequency of the exercise and the sales volume of boilers with
additional pollution control equipment.
Flue Gas Re circulation
Retrofit of an existing unit to accept flue gas recirculation involves
both the addition of new equipment as well as alteration of the existing
plant. Depending upon the boiler house layout, a considerable design effort
may also be required. Table 6-1 presents an approximate breakdown of design,
installation and equipment costs associated with the retrofit of FGR systems
to the two boilers at ECCC.
-------
Table 6-1. Cost Breakdown for Fitting FGR to the Two
Boilers at ECCC (1975 Dollars)
Design (including drafting)
Fabrication of Duct
Installation of Duct
Blower
Butterfly Valve
Flanges
Electrical Hardware
Installation (other than duct)
Controls System
Total
Firetube
$ 5,000
2,900
1,800
1,500
490
600
450
5,000
2,600
$20,340
Watertube
$ 5,000
2,900
1,800
1,900
540
900
550
5,000
2,600
$21,190
If a flue gas recirculation system were to be applied to a new boiler,
costs should be considerably less than for a retrofit. The cost of design
would be small, as the system would be an integral part of the boiler. Per-
haps the blower could be eliminated entirely by upgrading the forced draft
fan and utilizing this as both the FD and FGR fan, although this would impose
more severe restraints on the control system. The amount of duct work could
also be reduced, thus lowering the cost. Table 6-2 is an estimate of the cost
of including an FGR system in a new package boiler by the manufacturer.
89
-------
Table 6-2. Approximate Cost Breakdown for Application of
Flue Gas Recirculation to New Boilers
Design (In excess of normal)
Fabrication and Installation
of Duct work
Blower (use FD fan)
Fittings
Electrical Hardware
(No additional)
Installation (In excess or normal)
Automatic Controls
$ 500
2,000
800
Total
1,000
2,600
$6,900
Staged Combustion
Although the staged combustion equipment was not considered as a practical prop-
osition, it is instructive to examine an approximate cost breakdown (Table 6-3).
Table 6-3. Breakdown of Costs for Staged Combustion Investigation
Design
Control and
Measurement
Materials
Installation
Total
As Carried Out At
Essex County
$ 7,000
700
2,680
18,000
$28,280
Estimated for
New Boiler
M4B4ปซซH^ซ^H^H^MMM
$2,000
3,400
500
3,700
$9,600
In addition to these costs, approximately $1,000 would be required for a blower.
However, in this case, the blower was supplied by Foster Wheeler.
90
-------
If a staging system were to be included in the original design of a
firetube boiler, cost savings over a retrofit could be realized in design,
materials and possibly installation. Instead of penetrating the rear of the
boiler, side penetration would be utilized. This would increase boiler costs
due to additional pressure welds and the possibility of rearranging tubing
locations. However, the forced draft fan could be used as the air supply,
and there would be less need for high temperature alloys. Automatic con-
trols would be an additional cost.
The costs given in Table 6-3 refer to a new package boiler of existing design
which would be modified before delivery to the customer. If a new class of
boiler were to be offered for sale whose design had been altered to more readily
include these additional facilities, the cost of this new class would not con-
tain all those items listed in Tables 6-2 and 6-3. Design costs, for instance,
would be minimal and additional fabrication costs could be reduced considerably.
6.4 Implication of Results on New Design
The results of the F6R tests on the two boilers indicate that, with the fuel
fired, FGR has a definite effect on lowering NO in the firetube boiler and an
A'
insignificant effect in reducing it in the watertube boiler. Therefore, FGR is
not recommended as a method for controlling NO emissions on a watertube boiler
A
of the size tested and firing the fuel tested. However, it is very difficult
to draw completely general conclusions on the basis of one series of labora-
tory investigations and tests on two field boilers. Combustion systems giving
rise to considerably thermal NO formation (e.g., firetube with refractory front
section, refractory firebox watertube boiler) will probably be responsive to
FGR as a control technique, but emissions are not expected to be lowered by
greater than 30 percent for No. 5 or No. 6 fuel oil.
The FGR system on the firetube boiler appeared satisfactory except for the
experience with pulsations described earlier. However, the FGR fan was over-
sized, as higher FGR rates at high boiler loads did not cause further
reductions in NO . In a new design, the system could be made more compact
91
-------
incorporating air and FGR fans in one combined function. This would
facilitate air/flue gas mixing and lower the temperature of the gas entering
the windbox. It may be advantageous to rearrange the location of boiler com-
ponents which presently pass through the windbox. This would avoid any diffi-
culty with respect to obstructions inside the windbox and problems associated
with pressure fluctuations, and perhaps temperature sensitive components. Addi-
tional consideration must be given to the rearrangement of external equipment
associated with boiler operation to enable ease in the fitting of FGR components.
Finally, testing should be performed to determine the causes of combustion
instability so that the severe vibrations previously experienced would not occur.
92
-------
REFERENCES
1.
2.
Muzio, L.J., and Wilson, R.P., Jr., "Experimental Combustor for Development
of Package Boiler Emission Control Techniques," Phase I of III, EPA Report
R2-73-292-a, 1973.
Muzio, L.J., Wilson, R.P., Jr
Modifications for Reducing Nitric
EPA Report R2-73-292-b, 1974.
and McComis, C., "Package Boiler Flame
Oxide Emissions," Phase III of III,
Muzio, L.J., Wilson, R.P.. Jr., and McComis, C., "Package Boiler Flame
Modifications for Reducing Nitric Oxide Emissions," Phase II of III,
EPA Report R2-73-292-D, 1974.
Siegmund, W.C., and Turner, D.W.,
Potential Control Methods," ASME Paper 73 IPWRIO, 1973.
"NOX Emissions from Industrial Boilers:
5.
6.
7.
8.
10.
11.
Lachapelle, D., Private communcation.
Heap, M.P., Lowes, T.M. and
Design Variables to Control
Flames", Fluid Mechanics of
Mechanical-Engineers (1974).
Martin, G.B., "The Optimization of Aerodynamic
the Formation of Nitric Oxide in Fossil Fuel
Combustion, p. 75, The American Society of
Cato, G.A., Buening, H.J., DeVivo, C.C., Morton, B.G., and Robinson, J.M.,
"Field Testing: Application of Combustion Modifications to Control
Pollutant Emissions from Industrial Boilers - Phase I". EPA Report
650/2-74-078-a.
Barrett, R.E. and Miller, S.E., "Field Investigation of Emissions from
Combustion Equipment for Space Heating, Final Report". Battelle-
Columbus Laboratories, Columbus, Ohio. Prepared fo.r the U.S. Environmental
Protection Agnecy and the American Petroleum Institute. EPA Report No.
R2-73-084a. NTIS No. PB 223-148, or API Publ. 4180, June 1973.
9. Armento, W.J., "Effects of Design and Operating Variables on NOX from
Coal-fired Furnaces
NTIS PB 229-986/AS,
, Phase I". Babcock
January 1974.
and Wilcox, EPA 650/2-74-002a,
Turner, D.W., Andrews, R.L., and Siegmund, C.W., "Influence of Combustion
Modification and Fuel Nitrogen Content on Nitrogen Oxides Emissions from
Fuel Oil Combustion". Presented at 64th Annual AIChE Meeting, San
Francisco, November 1971.
Martin, G.B. and Berkau, E.E., "Evaluation of Various Combustion Modifi-
cation Techniques for Control of Thermal and Fuel-Related Nitrogen Oxide
Emissions". 14th Symposium (International) on Combustion, Pennsylvania
State University, August 1972.
93/94
-------
APPENDIX 1
TABULATED FIELD TEST DATA
Table Al-1. Boiler Performance Tests, Firetube, No. 5 Fuel Oil
Table Al-2. Boiler Performance Tests, Firetube, Natural Gas
Table Al-3. Boiler Performance Tests, Watertube, No. 5 Fuel Oil
Table Al-4. Boiler Performance Tests, Watertube, Natural Gas
Table Al-5. Flue Gas Recirculation, Watertube, No. 5 Fuel Oil
Table A1-6. Flue Gas Recirculation, Firetube, No. 5 Fuel Oil
Table Al-7. Staged Combustion, Firetube, No. 5 Fuel Oil
95
-------
Table Al-1. Boiler Performance Data ECCC Firetube Boiler No. 5 Fuel Oil
Test
A
E
F
G
H
I
J
K
L
M
N
O
P
N(R)
Q
R
Steam Flow
Ib/hr x 10~3
4
5
5.5
5.5
5
7.5
7.25
7.5
7.5
7.5
11
11.5
^^^^^^^^^^^^^^w
11
11
11
8
NOX
ppm dry
170
180
170
180
170
165
155
173
175
163
193
190
202
195
197
175
NOX
ppm dry corr.
to 0% O2
230
199
178
228
244
257
268
241
206
171
237
215
^^^^^^^^^M--* Ml ป
250
231
239
260
CO
ppm dry
by vol .
120
>1000
^^W***^^^^^^^BW^^W^>-i^BIM^^
3
Smoke
Bacharach
Scale
6.5
9.5
3
1.5
1.5
1.5
2
5
(10)
3
6
Fuel Oil
Temp.
ฐF
155
150
155
155
150
155
155
155
155
160
160
160
160
140
150
SO 2
ppm dry
by vol .
450
568
665
500
438
383
343
433
527
615
512
550
; -
ฐ2
% By
vol . dry
5.5
2.1
0.9
4.4
6.4
7.5
8.9
6.0
3.2
1.0
3.9
2.4
- M- '
4.1
3.3
3.7
6.9
vo
-------
Table Al-2. Boiler Performance Data ECCC Firetube Boiler Natural Gas
Test
A-l
B-l
C-l
D-l
E-l
F-l
G-l
H-l
1-1
J-l
K-l
L-l
M-l
N-l
Steam Flow
IbArx 10-3
1
1
1
0.75
3.75
4
4
4
7
6.5
6.5
8.5
8
8.75
NOX
ppm dry
42
44
47
38
61
64
59
56
73
71
68
68
72
66
NOX
ppm dry corr.
to 0% O2
115
87
87
130
107
103
120
123
105
111
119
102
105
105
CO
ppm
dry
20
20
20
20
15
15
17
5
5
15
5
15
40
40
02
%dry
by vol .
13.3
10.4
9.6
14.9
9.1
8.0
10.7
11.5
6.5
7.6
9.0
7.1
6.6
7.8
-------
Table Al-3. Boiler Performance Data ECCC Watertube No. 5 Fuel Oil
Test
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Steam FlowQ
(Ib/hrx 10~3)
10
10.5
10
10
10
10
10.25
15
15
15.5
15
15
15.25
15
15
15
19
18.5
18.5
18.5
18.5
18.5
18.5
12.5
7.5
7.5-10
14.5
14.75
14.5
14.5
14.5
14.5
14.5
NO
ppm dry
139
118
142
155
93
117
140
123
135
143
140
110
87
78
120
102
130
120
130
135
110
88
117
98
107
66
113
105
110
115
118
120
120
N0x
ppm dry corr.
to 0% O2
177
144
199
232
109
148
187
140
220
202
175
126
98
90
170
131
163
215
203
182
132
110
148
124
178
78
133
113
120
129
131
135
137
CO
ppm dry
_
-
_
-
-
0
25
-
-
-
100
>3000
3000
-
-
-
-
-
-
100
1150
-
-
-
2500
-
>3000
2200
500
-
25
"~
02
% dry
by vol.
4.6
3.8
6.0
7.0
3.1
4.5
5.3
2.6
8.1
6.1
4.2
2.7
2.3
2.8
6.2
4.6
4.3
9.3
7.6
5.5
3.6
4.2
4.5
4.5
8.4
3.2
3.1
1.5
1.8
2.3
2.1
2.4
2.6 "
S02
ppm dry
531
565
488
422
572
530
502
607
408
466
528
584
788
653
478
529
535
365
416
486
575
579
520
517
371
585
518
675
630
600
590
603
600
Smoke No.
Bacharach
_
9
3
3
8
3.5
2.5
3.5
2
2
2
4
-
9.5
15
3
4
3
3
3
4
7
3
3
3.5
8.5
3
10+
8
3.5
3.5
3.5
"""
Atom.
Steam
Press.
psig
39
39
40
52
20
36
39
50
52
52
51
50
52
51
63
50
61
60
61
60
60
59
60
43
32
37
48
49
49
48
47
47
48
Fuel
Oil
Temp.
152
152
152
152
152
110
165
148
147
147
150
150
150
150
150
150
151
150
149
149
150
150
150
149
147
149
152
152
152
152
152
150
152
Air
Regis,
% Open
100
100
100
100
100
100
100
100
100
100
100
-
-
-
-
_
-
-
-
-
-
-
-
-
-
-
-
-
-
81
88
94
~~
CO
-------
Table Al-4. Boiler Performance Data ECCC Watertube Natural Gas
UD
10
Test
1-A
2 -A
3 -A
4-A
5-A
6-A
7-A
8-A
9-A
10-A
11-A
12-A
13-A
14-A
15-A
16-A
17-A
18-A
19-A
20-A
21-A
22-A
23-A
24-A
25-A
26-A
27-A
28-A
Steam Flow
(IbAr x 10~3)
13.5
13.0
13.0
13.0
10.0
10.25
10.5
10.5
13.0
13.5
13.5
13.0
12.75
12.5
20.0
20.0
19.75
19.5
19.25
15.75
16.0
16.25
15.5
8.25
8.0
8.5
8.5
8.5
NO
ppm dry
62
55
62
59
62
57
62
63
64
67
70
73
77
64
50
38
44
54
53
51
44
49
50
44
39
42
48
47
NOX
ppm dry corr.
to 0% O2
70
78
79
62
77
81
80
71
72
77
81
82
81
72
78
75
79
76
69
72
76
77
64
74
79
76
68
60
CO
,ppm dry
30
15
0
690
15
30
30
50
30
15
15
15
400
30
20
15
20
30
785
5
5
20
925
0
0
0
0
470
Smoke No.
Bacharach
0
0
0
2
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0.5
0
0
0
0
0
Air Register
% Open
100
100
100
100
100
100
100
100
100
94
88
81
75
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
-------
Table Al-5. Flue Gas Recirculation Watertube No. 5 Oil
Test
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
%FGR
0
10.1
18.8
21.9
22.2
27.2
23.5
19.9
8.9
0
0
11.4
17.3
24.1
0
9.0
17.4
26.5
33.0
0
0
8.0
19.4
24.1
7.5
9.2
10.7
0
0
0
8.1
11.1
21.2
Steam Flow
(Ib/hr x 10~d)
9.75
10.0
10.0
10.5
10.0
10.2
10.2
9.8
10.2
10.3
6.5
6.7
6.5
7.0
6.5
6.5
6.0
6.4
6.5
4.8
14.6
16.2
15.0
14.9
15.0
15.5
15.8
15.6
15.5
16.0
15.8
15.8
16.0
NO
ppm ary
94
99
95
93
82
78
82
84
82
83
110
102
96
91
89
92
89
84
80
91
112
113
106
108
113
110
97
91
105
114
106
103
101
NOX
ppm dry corr.
to 0% O2
112
118
113
109
91
86
91
93
91
92
142
132
123
117
104
106
105
96
92
157
139
141
130
135
135
148
107
100
122
141
123
120
119
CO
ppm dry
10
38
52
53
110
112
110
110
111
210
30
27
30
30
37
30
30
30
30
30
20
20
20
20
20
20
47
365
22
20
27
20
20
02
% dry
by vol.
3.3
3.4
3.4
3.0
2.1
1.9
2.1
2.1
2.0
2.1
4.7
4.7
4.6
4.6
3.0
2.8
3.1
2.7
2.8
8.8
4.0
4.2
3.9
4.2
3.4
5.4
2.0
1.9
2.9
4.0
2.9
3.0
3.1
SO2
ppm dry
215
209
210
215
225
225
225
223
225
235
182
175
180
175
170
176
179
180
177
141
178
175
176
173
181
170
183
183
198
192
192
198
196
Smoke No,
Bacharach
2
2
2
2
3
2.5
2.5
1
1
3
1.5
2
1.5
2.5
2.5
2
2
3
-
-
-
-
-
-
-
-
1
0
0
2
2
o
o
-------
Table Al-5. Flue Gas Rec ire illation Watertube No. 5 Oil (Cont.)
Test
34
35
36
37
38
39
40
41
42
43
44
45
%FGR
28.4
24.0
16.8
16.4
19.0
7.6
8.8
0
0
0
0
0
Steam Flow
(Ib/hr x 10~3)
15.6
16.3
16.0
16.2
16.8
16.8
16.5
16.7
16.0
16.0
15.2
15.5
N0x
ppm Hry
90
103
100
113
92
112
102
115
89
110
113
93
NOX
ppm dry corr.
to 0% O2
102
130
117
141
102
143
112
146
98
131
149
103
CO
ppm dry
20
15
20
20
45
20
32
20
213
-
-
"
ฐ2
% dry
by vol.
2.4
4.3
3.0
4.2
2.0
4.5
1.9
4.4
1.9
3.4
5.0
2.1
S02
ppm dry
200
191
192
187
200
196
204
197
200
197
131
215
Smoke No.
Bacharach
3
1
2
2
4
1
3
1
4
-
-
"
-------
Table Al-6. Flue Gas Recirculation Firetube No. 5 Oil
Test
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
18a
19
20
21
22
23
24
25
26
27
28
29
% FGR
0
0
22.7
31.5
40.2
0
22.5
32.0
38.6
0
19.3
27.8
38.8
52.7
0
9.8
19.8
39.8
28.9
34.8
28.2
17.8
0
0
0
9.4
18.0
22.2
19.9
14.7
Steam Flow
(lb.hr x 10~3)
4.0
4.1
4.0
4.0
4.0
3.9
4.0
3.9
4.1
4.2
4.2
4.3
4.0
3.8
6.0
6.0
6.0
6.4
6.5
6.2
6.3
6.3
6.3
11.25
11.0
11.0
11.0
11.0
10.5
10.75
NOX
ppm dry
167
162
160
145
118
160
152
129
123
175
168
153
139
118
145
134
123
113
120
120
123
133
162
174
175
145
127
119
125
135
NOX
ppm dry corr.
to 0% O2
180
204
198
179
148
223
209
176
171
197
188
171
154
132
224
203
186
170
171
150
152
166
202
215
200
170
144
128
138
149
C02
ppm dry
-
-
-
-
-
-
-
-
-
-
-
-
-
-
40
40
45
40
40
40
40
40
35
38
50
58
225
78
85
02
% dry
by vol .
1.5
4.3
4.0
4.0
4.2
5.9
5.7
5.6
5.9
2.3
2.2
2.2
2.0
2.2
7.4
7.1
7.1
7.0
6.2
4.1
4.0
4.2
4.1
4.1
3.0
3.1
2.5
1.5
2.0
1.5
SO2
ppm dry
248
203
197
204
205
181
180
178
174
225
224
225
230
230
162
160
160
160
170
198
200
200
191
210
223
210
220
228
223
228
Smoke No.
Bacharach
4
1
1
2
2
1
1
2
2
2
2
2
2.5
3.5
1
1
0.5
1.5
1.5
1.5
1.5
1.5
1
0.5
1
2.5
5
7
5
6
o
rv>
-------
Table Al-6. Flue Gas Recirculation Firetube No. 5 Oil (Cont.)
Test
30
31
32
33
34
35
36
37
% FGR
9.5
0
8.9
17.3
25.6
19.4
9.5
0
Steam Flow
(lb.hr x 10~3)
11.25
9.25
9.0
9.0
9.8
10.0
9.9
9.9
NOX
ppm dry
141
161
140
130
125
127
141
162
NOX
ppm dry corr.
to 0% O2
158
224
194
168]
150
152
169
195
C02
ppm dry
55
56
45
60
70
70
70
70
02
% dry
by vol.
2.3
5.9
5.8
4.7
3.5
3.4
3.5
3.5
SO2
ppm dry
223
200
198
210
224
228
228
231
Smoke No.
Bacharach
5
0.5
0.5
1
2
2
1
o
o
co
-------
Table Al-7. Staged Combustion Firetube No. 5 Fuel Oil
Test
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Load
4,000
6,000
6,000
7,000
6,000
6,000
6,000
6,000
6,000
6,000
10,000
13,000
11,000
11,500
11,000
7,000
8,000
7,000
8,000
7,100
7,000
7,000
7,000
5,500
5,000
5,000
5,500
6,000
6,000
6,000
Fuel Flow
380
380
380
380
380
380
380
380
380
380
380
380
380
380
380
380
380
380
380
380
380
380
380
380
380
380
380
380
380
380
Burner
Stoich .
133
117
109
96
128
97
96
85
83
74
119
121
121
113
117
127
145
98
103
95
91
91
87
112
110
92
94
87
76
77
ฐ2
% dry
by vol.
6.8
7.3
6.8
5.4
6.6
5.3
6.3
5.1
5.1
5.2
3.6
4.8
4.0
5.1
5.7
8.0
8.9
5.0
6.5
7.5
8.0
8.4
8.3
2.4
4.4
1.5
4.9
3.8
2.0
3.6
NOX
ppm dry
129
115
115
113
125
116
110
100
97
90
172
157
170
155
150
130
116
126
119
112
106
104
103
146
135
129
123
105
92
94
NOX
ppm dry corr.
to 0% O2
191
176
170
152
182
155
157
132
135
120
208
204
210
205
206
210
201
165
172
145
171
173
170
165
171
139
160
128
102
113
SO2
ppm dry
120
118
122
138
122
132
125
140
130
138
163
148
152
142
137
110
100
135
121
111
110
102
103
158
140
179
135
150
170
152
CO
ppm dry
20
20
20
30
30
30
30
40
30
40
50
50
50
50
60
60
50
50
52
50
50
50
50
60
58
185
65
70
145
60
Smoke No.
Bacharach
-
-
-
0.5
1.5
0.5
3.0
3.0
3.5
3.0
-
1.5
1.0
0
-
0
0
0
0
0
0
0
2.0
1.0
2.5
2.0
4.0
7.0
3.0
-------
Table Al-7. Staged Combustion Firetube No. 5 Fuel Oil (Cont.)
Test
31
32
33
34
35
36
37
38
39
40
41
42
43
Load
6,000
5,500
6,000
5,800
5,500
6,000
5,500
5,500
5,750
6,000
5,750
6,000
6,000
Fuel Flow
380
380
380
380
380
380
380
380o
380
380
380
380
380
Burner
Stoich.
85
76
65
65
70
66
77
72
69
81
73
84
101
ฐ2
% dry
by vol.
5.9
5.9
4.5
4.2
4.8
3.2
4.8
4.2
3.7
5.3
1.4
3.3
3.0
NOX
ppm dry
90
87
85
86
86
91
89
88
88
87
87
108
135
NOX
ppm dry corr.
to 0% O2
125
121
108
108
111
107
115
110
107
116
95
128
158
SOฃ
ppm dry
140
130
145
150
140
160
140
149
150
134
161
155
160
CO
ppm dry
65
50
55
58
55
60
55
56
60
55
250
75
6"0
Smoke No.
Bacharach
1,5
1.0
1.5
2.0
1.5
2.0
2.0
3.0
3.0
3.0
3.5
6.0
1.5
o
en
-------
APPENDIX 2
BOILER PERFORMANCE BEFORE AND AFTER MODIFICATION
Operational problems could be attributed to the modifications made to either
of the two boilers which were tested in the field have been discussed earlier.
Any influence of these modifications on the emission characteristics of the
bonlers ought to be able to be assessed by comparing boiler performance data
obtained before and after the modification. The relevant data is presented
in Figures A2-1 and -2 for the firetube and watertube boilers, respectively.
It can be seen that in most instances the reported smoke and NO baseline emis-
sions are lower after modification. There are three possible explanations which
could account for this variation in baseline performance:
Errors in flue gas concentration measurement.
The influence of boiler modifications on combustion conditions.
Changes in fuel oil properties.
Of these possibilities, the first two can almost certainly be discounted, no
maintenance was carried out between tests which would influence the results.
The experimental procedure included frequent calibration checks of all flue gas
analytical equipment. Provided the calibration gases were certified correctly,
systematic measurement errors are unlikely. As both smoke and NO emissions
^
were lower after modification, an error in flue gas oxygen concentration
determination could not explain the results. All physical alterations to the
boiler to allow recirculation of flue gases were associated with either the
stack or the windbox. When no flue gases are being recirculated (even though
the duct work is in place), it is unlikely that the slight modification to the
windbox would drastically change combustion conditions. Variations in fuel
properties offers the most plausible explanation for the difference in emission
characteristics observed before and after modification.
A fuel oil sample was taken during each set of tests (performance, flue gas
recirculation and staging) for both boilers. These samples were analyzed by
the FWEC Analytical Laboratory according to ASTM (or equivalent) standards.
Samples were resubmitted for analysis when it became apparent that certain
107
-------
NO and Smoke Number vs % O2 NOx Sniok*
X /"N ^ T
Firetube boiler
CM
O
0
0
s_
o
(J
s;
TJ
Q.
a.
X
o
280
260
240
220
200
180
160
140
120
100
80
60
40
20
0
-
-
-
- oO
a
-
.
-
-
-
-
^^^^^^^^^iiiiiiiiii
O Load 4,000 Ib steam/hr
D Load 6,200 Ib steam/hr
A A Load 10,300 Ib steam/hr
Large symbols refer to data after
modifications
Small symbols refer to baseline data
D
D
o
A
D
-l 10
Figure A2-1.
O2 , % by Volume
Comparison of Boiler Performance Data Before
and After Modification (Firetube Boiler)
108
-------
and Smoke Number vs % CX
Baseline Data and Data
After Modifications
Watertube
NO Smoke
X.
04
O
0
O
. ป
*-'
c
0
o
-a
e
a.
X
o
O Load 5,000-6,500 Ib steam/hr
280 r
260
240
220
200
180
160
140
120
100
80
60
40
20
0
U Load 10,000 Ib steam/hr
O Load 12,000 Ib steam/hr
A A Load 15,000-16,000 Ib steam/hr
O + Load 18,500 Ib steam/hr
Small symbols refer to baseline data
Large symbols refer to data taken after
modification
Mk A
/\
0
& o
o
AD ฐ
O
o
A ty$y
A/J$> ^
^A O
y \ ^^
D<>
^P I10
- 8 _
o
(0
c CO
_ _ b ซ- *
o3
A * - 4 i
> A * ฎ
Tl AA A A - 2 1
A A*
' A ' ' ฐ
O2 , % by Volume
Figure A2-2
Comparison of Boiler Performance Data Before
and After Modification (Watertube Boiler)
109
-------
discrepancies existed. Results of the oil analyses for the various samples
are presented in Table A2-1. The oil analysis data contains serious anomalies
which makes interpretation of the data difficult. The following anomalies are
readily apparent:
The carbon concentration in the original firetube boiler
performance oil is low and the reproducibility is poor compared
to other samples.
Nitrogen concentrations show a wide spread both in original
analyses and reanalysis. However, the anomalies are inconsistent.
The reproducibility of samples taken from the watertube boiler is
excellent. However, variations in the other analyses cast doubt
on the authenticity of all the results.
Values of 2 to 4 percent CL are most improbable and suggest errors
in the values of the other elements.
Carbon, hydrogen and nitrogen concentrations were determined with a Perkin-
Elmer Model 240 Elemental Analyzer which has been shown to give results
equivalent to those obtained using ASTM methods for carbon, hydrogen and
nitrogen in coal. Evidence exists showing that oil analyses change with
time and certain trace metal compounds can be lost (see T.F. Yen*).
Measured flue gas sulfur dioxide concentrations confirmed that the oil burned
during the performance tests had a higher sulfur content than that used in
the subsequent tests. Thus, it could be inferred that the nitrogen content of
oil used in the performance tests would be higher which could account for the
change in NO emissions before and after boiler modifications. A 1 percent
/\
sulfur fuel would normally contain more nitrogen than a 0.3 percent sulfur
fuel.
The original fuel analysis did not provide evidence in support of the above
hypothesis (see Table A2-1) and a further analysis was carried out by an
independent laboratory. In this instance the nitrogen content was determined
by the Kjeldahl method. The results of the third analysis are presented in
The Fate of Trace Metals in Petroleum, T. F. Yen, Ann Arbor Science, 1975.
110
-------
Table A2-1.
FUEL OIL ANALYSES
Performance
t->
c
01 -D
U 01
i- >
Q> *r*
0. (U
u
0)
10
C VI
-------
Table A2-2. Inspection of the sets of analyses reveals several notable
differences. In the independent analysis:
oxygen contents obtained by difference are more consistent and
lower than the original analysis;
carbon contents are higher and hydrogen contents are lower
than the original analysis;
nitrogen contents are considerably higher than the original
analysis.
In view of the earlier discussions, the differences in the nitrogen content
are most disturbing. If values of 0.2 percent nitrogen are correct, then
emissions from both boilers appear to be low compared to measurements reported
by other workers (see Figure 6-2).
Variations in fuel properties are the most probable reason for the difference
in emission levels before and after modifications. The lower sulfur content
probably suggests a lower nitrogen content, although the various fuel analyses
do not show this trend consistently.
Table A2-2. Independent Analysis
Component
Sulfur, %
Carbon, %
Hydrogen, %
Nitrogen, %
Oxygen, %
(by difference)
Sample
Firetube FGR
0.44
86.79
12.17
0.22
0.38
Watertube FGR
0.40
86.75
12.29
0.19
0.37
Firetube Staging
0.32
86.84
12.31
0.26
0.27
112
-------
2.
,/>/., TECHNICAL REPORT DATA
p. REPORT NO ~ f^fซซ'ซป
------- |