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
-------