November 1977                                                      EPA-600/8-77-003b
  REFERENCE  GUIDELINE FOR INDUSTRIAL BOILER MANUFACTURERS

     TO CONTROL POLLUTION WITH COMBUSTION MODIFICATION
                       Guidelines intended for use as a reference source by:

                       —  industrial boiler equipment manufacturers for
                          pollution control on existing and new boilers

                       -  service organizations responsible for coordinating
                          industrial pollution control activities

                       -  architects and engineers responsible for boiler
                          specifications

                       —  personnel responsible for pollution testing and
                          control

                       -  personnel responsible for air quality management
                          and planning
                                              \
                      U.S. ENVIRONMENTAL PROTECTION AGENCY
                      OFFICE OF RESEARCH AND DEVELOPMENT
                 INDUSTRIAL ENVIRONMENTAL RESEARCH LABORATORY
                         RESEARCH TRIANGLE PARK, NC 27711

-------
               RESEARCH REPORTING SERIES

Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology  Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:

      1.   Environmental  Health Effects Research
     2.   Environmental  Protection Technology
     3.   Ecological Research
     4.   Environmental  Monitoring
     5.   Socioeconomic Environmental Studies
     6.   Scientific and Technical Assessment Reports (STAR)
     7.   Interagency Energy-Environment Research and Development
     8.   "Special" Reports
     9.   Miscellaneous  Reports
                     EPA REVIEW NOTICE

    This document 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.

-------
                                  CONTENTS

Section                                                             Page
          LIST OF FIGURES                                            iv
          LIST OF TABLES                                              v
  1.0     INTRODUCTION                                                1
  2.0     SOURCES OF COMBUSTION POLLUTANTS                            3
          2.1   Oxides of Nitrogen                                    3
          2.2   Pollutants Other Than NO                              6
                                        x
  3.0     METHODS FOR NO  CONTROL                                    10
                        x
          3.1   Fuel Type                                            10
          3.2   Excess Air                                           11
          3.3   Staged Combustion                                    13
          3.4   Burner Adjustments                                   22
          3.5   Reduced Combustion Air Preheat                       23
          3.6   Flue Gas Recirculation (FGR)                          25
          3.7   Water Injection                                      39
          3.8   Furnace Characteristics                              39
          3.9   Burner Design                                        44
          3.10  Fuel-bed Firing                                      47
          3.11  Thermal DeNO  Process                                49
                            x
  4.0     IMPACT OF THE COMBUSTION MODIFICATIONS ON                  52
          BOILER EFFICIENCY
          4.1   Effect of Excess Air                                 52
          4.2   Effect of Staged Air                                 52
          4.3   Effect of Combustion Air Temperature                 55
          4.4   Effect of Flue Gas Recirculation With                55
                and Without Staged Combustion
                                                               6001-49

-------
                             CONTENTS (Continued)
Section                                                            Paqe
 5.0     COMPARISON OF TECHNIQUES ON TWO SPECIFIC                   60
         INDUSTRIAL BOILERS
         5.1  Combustion Modifications on a Boiler Rated            60
              at 17,500 Ib/hr of Steam Flow
         5.2  Combustion Modification on a Boiler Rated             61
              at 45,000 Ib/hr of Steam Flow
         5.3  Conclusions                                           64
 6.0     COST OF THE CONTROL TECHNIQUES                             67
 7.0     SUMMARY                                                    72
         7.1  Excess Air                                            75
         7.2  Staged Combustion                                     75
         7.3  Air Preheat                                           76
         7.4  Flue Gas Recirculation                                76
         7.5  Burner Design and Adjustments                         76
         7.6  Thermal DeNO  Process                                 76
                          x
         7.7  Limitations on Applying NOX Control                   77
 8.0     REFERENCES                                                 78
 9.0     CONVERSION FACTORS                                         80

 APPENDIX A - EMISSION CORRECTION CALCULATIONS FOR FLUE GAS         81
              MOISTURE CONTENT
                                     ii                       6001-49

-------
                                   FIGURES

Number                                                                 Paqe
 1       Effect of excess oxygen on nitrogen oxides emissions
         Single lines for watertube boilers, shaded areas represent
         firetube boiler data                                           12

 2       Schematic diagram of staged air system installed on a
         40 k Ib/hr watertube boiler                                    14

 3       Reduction in total nitrogen oxides due to staged com-
         bustion air, natural gas fuel (40 k Ib/hr, watertube
         boiler                                                         15

 4       Reduction in total nitrogen oxides emissions due to
         staged combustion air, No. 6 oil fuel (40 k Ib/hr
         watertube boiler)                                              17

 5       Reduction in total nitrogen oxides due to taking burners
         out of service, mixture natural and refinery gas fuel          21

 6       Effect of air register setting on NO  for a specific
         boiler                              x                          24
 7       Effect of air preheat at normal excess air levels              26

 8       Effect of air preheat at high excess air                       26

 9       Effect of combustion air temperature on total nitrogen
         oxides emissions, gas and oil fuels                            27

 10      Relationship between two definitions of "percent flue
         gas recirculation"                                             29

 11      Reduction in total nitrogen oxides emissions by flue gas
         recirculation with constant excess air                         30

 12      Secondary flue gas recirculation                               32

 13      NO dependence on gas recirculation for normal (air-
         rich) and two-stage combustion of pulverized coal              33

 14      Flue gas recirculation with natural gas firing; 320 Mw
         tangentially fired unit.                                       34

 15      Effect of flue gas recirculation rate on NO  emissions         36
                                                    x
 16      NO  emissions as a function of percent flue gas
         recirculation  (No. 2 oil)                                      37

 17      Effect of flue gas recirculation on NO  emissions
         (No. 6 oil)                           X                        38
                                      ill                    6001-49

-------
                              FIGURES (Continued)

Number                                                                 Page

 18      Effect of heat release on NO emissions                         40
 19      EPA ring burner test results for watertube boilers             42
 20      EPA ring burner test results.  Firetube boilers,
         ambient-temperature air                                        42
 21      Comparison of EPA industrial boiler ring burner
         test results to utility boiler ring burner test
         results                                                        43
 22      Operating limits of a spreader stoker burning a
         western subbituminous coal                                     45
 23      NO  emissions from a spreader stoker burning a
         western subbituminous coal                                     46
 24      Water-cooled vibrograte stoker, 45,000 Ib/hr steam.             48
 25      Results of the Thermal deNO  process                           51
                                    X
 26      Effect of reducing the excess combustion air on
         boiler efficiency                                              53
 27      Effect of staged combustion air on boiler efficiency           54
 28      Effect of operating with burners out of service on
         boiler efficiency                                              56
 29      Effect of reduced combustion air preheat temperature
         on boiler efficiency                                           57
 30      Effect of flue gas recirculation and staged combustion
         air on boiler efficiency                                       58
                                    TABLES
  1      Effect of Terminating Fuel Flow to Selected Burners            19
  2      Summary of NO  Reduction as a Function of Combustion
         Modification Technique for Various Fuels                       62
  3      Summary of Change in Boiler Efficiency Due to
         Combustion Modifications                                       63
  4     NO   Reduction and Efficiency Change for Combustion
        Modification Techniques Test at Location 39                     65
  5     Approach  to NOX Control Costs                                   68
  6     Costs  of  NO  Control by Combustion Modification                 70
  7     Summary of NOX Reduction Techniques                             73
                                                              6001-49

-------
                                  SECTION 1.0
                                 INTRODUCTION

        This manual is the second of two manuals on industrial boiler emis-
sions controls.   The first manual (Ref.  1)  was directed to boiler operators
and contains recommended procedures for  improving boiler performance by
methods available to boiler operators.   The purpose of this second manual
is to describe current combustion modification techniques that are available
to boiler manufacturers for controlling  air pollutant emissions from in-
dustrial-size fossil-fuel-fired steam boilers.  The main emphasis will be
placed on controlling emissions of oxides of nitrogen (NO ) from gas, oil and
coal fired boilers in the size range 10,000 to 500,000 Ib/hr steam.*  How-
ever, the impact of these techniques on  efficiency and on other combustion-
generated pollutant emissions (such as particulates and carbon monoxide) that
may be affected by NO  control will also be discussed.
                     x
        Recent estimates of the NO  contribution from the industrial boiler
                                  x
size category range from 18 to 28% of the national total for stationary
sources (Refs. 2, 3, 4).  Increased emphasis on control of stationary source
emissions is anticipated and may result  in requiring further controls on in-
dustrial boilers.
        The pollution control methods covered in this manual can be placed
into three classes according to the complexity of their application:
           change operating parameters of existing boilers
           retrofit combustion modifications to existing boilers
           incorporate advanced pollution control designs into
           new boilers
This manual will define the concepts of low-NO  combustion modifications.
These concepts then can be used as basic guideposts for their application in
hardware design.  Each individual method of combustion modification may not
*It is the policy of the U.S. Environmental Protection Agency to use the
 metric  (SI) system of units.  However, for the purposes of this report,
 English units were considered necessary.  For those who prefer metric units,
 a conversion table is provided in Section 9.0.
                                      1                     6001-49

-------
find wide applicability for all boiler types in the industrial category.  For
this reason, control techniques should be reviewed on a unit-by-unit basis.
        A significant amount of data has been collected on the various
control techniques and they are reported in detail in a number of publica-
tions  (Refs. 2 and 5-10).   These studies cover both field investigation and
smaller-scale laboratory work.  The discussion in this manual will draw
heavily from these studies.
        Certain control techniques have not been tried on industrial-size
boilers; this is the case  for  flue gas recirculation  (FGR) on pulverized
coal  (PC) firing.  In these instances, guidance can be obtained from large
utility boiler studies,- where, for example, it has been shown that FGR  is
not a viable NO  control technique on PC units  (Ref . 11) .
               X
        Emissions of NO  are expressed in this manual as concentrations
on a dry basis  (that is, moisture-free flue gas) corrected to 3% excess 0  .
This correction eliminates the effect of dilution of the flue gas by excess
air.  The measured concentration,  (NO )  is converted to the value at 3%
                                     x m
0_ using the  formula
              (N°x>3%  =
where 0  is measured oxygen concentration in volume percent on a dry basis.
If the measurement of NO  is on a wet basis, the value must first be corrected
                        x
to a dry basis  as discussed in Appendix A.
        The purpose of performing the correction to 3% 0  is to provide a
more meaningful way of comparing emissions at different  levels of excess
air.  It is important to understand  the difference between the above
procedure for correcting NO  to a common level of excess air and the actual
                           X
effect that changing excess air has  on NO  emissions.  For instance, burning
a high-nitrogen heavy oil with 7% excess oxygen might produce a measured
level of NO  of 390 ppmV.  Correction of this value to 3% O  results in
           X                                               £.
501 ppmV.  An actual reduction of excess air to 3% O  might result  in NO
                                                    2                   x
emissions of 425 ppmV.  This value is higher than that actually measured at
7% 02/- but when compared with the value of 501 ppm that  was corrected for
dilution, it is seen that NO  was actually lower at the  lower O  condition,
                            x                                  2
and NO  emission per unit heat input was reduced by lowering excess O .
      x                                                              2
                                                              6001-49

-------
                                  SECTION 2.0

                       SOURCES OF COMBUSTION POLLUTANTS
2.1     OXIDES OF NITROGEN
        The oxides of nitrogen (NO )  emitted by boilers are mostly nitric
oxide (NO) with a few percent of nitrogen dioxide (NO ).  Nitric oxide is
colorless whereas nitrogen dioxide is reddish brown.   However,  N02  is seldom
present in large enough concentrations to be visible in  a boiler stack
plume.
        Ambient air quality standards are set in terms of NO  concentrations
only.  However, both NO and NO  need to be controlled from emission sources
because NO is converted fairly rapidly to N02 in the  atmosphere.
        Oxides of nitrogen emissions arise from two sources in a flame:
           thermal fixation of atmospheric N_   (Thermal NOV)
                                            f.             A
           conversion of fuel bound nitrogen  (if present)   (Fuel NOX)
The NO , once formed, is "frozen" above its chemical equilibrium concentra-
tion in the flue gas because there are no available kinetic routes that can
destroy it within the residence time between the flame and stack exit.
        The formation of thermal NO  is determined by highly temperature-
dependent chemical reactions, the so-called Zeldovich reactions:
                   N  + 0 ->- NO + N
                   N + 02 -»• NO + 0
The rate of these reactions is significant only at high temperatures  (greater
than 3000 °F) and doubles for every 70 °F increase in flame temperature.  At
these temperatures, the rate of formation increases with the square root of
                                                               6001-49

-------
the 0  concentration except in a small region of the flame where oxygen
atoms (O) are present in super-equilibrium amounts.  In these regions,
the initial rate of formation of NO is higher.
        Natural gas, potentially a very clean fuel, can produce well over
1,200  parts per million by volume (ppmV) of thermal NO  in a boiler with
                                                       X
air preheat.  This occurs when the fuel is intensively mixed and burned
completely before much heat loss or much dilution by cooled combustion
products has occurred within the furnace.  Oil and coal typically give
thermal NO  concentrations between 60 and 210 ppmV.
          x
        The nitrogen contained in fuels such as distillate oil, residual
oil and coal may be partially converted to NO  during combustion.  If
                                             X
completely converted to nitric oxide, one percent by weight nitrogen will
produce about 1300 ppmV of NO  in oil firing and roughly 2100 ppmV in
                             X
coal firing.  The difference between the two is mainly due to the lesser
heating value of coal.  Both fuels produce roughly the same volume of flue
gas per million Btu.  But for coal more pounds of fuel, and thus more pounds
of bound nitrogen, must be burned to liberate a million Btu's.   The  theoretical
maximum flue gas content of NO from fuel nitrogen conversion is therefore
higher for a lower-Btu fuel of the same nitrogen content.
        The maximum possible NO levels from fuel nitrogen conversion are
listed below for some typical fuels of interest.
                            Typical N content,        Fuel-derived NO
        Fuel                  weight percent       (100% conversion), ppmV
        Natural gas                 0                        0
        No. 6 oil                   0.3                    390
        Bituminous coal             1.4                   2940
        Fortunately, conversion of all fuel bound nitrogen seldom occurs.
(Values of 25 to 50% are typical for coal and No.  6 oil in normal operation.)
Also, this conversion is not nearly as temperature-sensitive  as the  thermal
formation process. The fractional amount of fuel nitrogen converted  is greatly
influenced by the amount of oxygen which is available when the fuel
                                                            6001-49

-------
molecule decomposes.  If the surrounding atmosphere is fuel-rich, the
fuel molecule will crack and much of the nitrogen will form N .   On the
other hand, if 0_ is available then a large percentage of the N in the fuel
will form NO .  Theoretical analyses indicate that the percent conversion
of fuel N to NOx decreases as the ratio N/0  increases and as the tempera-
ture decreases.  (The dependence on temperature is slight.)
        High fuel volatility and intensive fuel-air mixing also encourage
conversion of fuel N to NO .  Conversion to NO  can range from 10% to over
                          x                   x
99%.  In residual oil- and coal-fired boilers, 10% to 60% conversion has
been observed.
        Many regulations set NO  limits based on pounds of NO  as NO  per
                               X                             X      £.
million Btu and fuel nitrogen content can be converted to this basis.
Assuming complete conversion of fuel nitrogen to NO  and taking, for example,
                                                   X
a coal of 1.4% nitrogen content with a heating value of 13,500 Btu/lb, the
N content can be expressed as

   1.4 Ib N       46 Ib NO?     Ib coal         10  Btu                 .
  100 Ib coal  X   14 Ib N  X  13,500 Btu  X  million Btu    3   ...  2704_
                                                                million Btu
If the actual conversion were only 50% then the emission rate would be
one-half the above value or 1.7 Ib/million Btu.  This is of course in
addition to any thermal NOx generated.
        Fuel-N-derived NO  is normally much more sensitive to excess O
                         X                                            ^
than is thermal NO  .  In general, at a given burning rate the higher the
                  X
burner airflow, the higher the fuel-N-derived NO  will be.  In contrast,
                                                X
thermal NO  theoretically maximizes at a fuel/air ratio corresponding to
          X
about 1% excess 0_.  In practice, vaporization, cracking, and mixing play
important roles so that the actual excess 0  for maximum NO  can not be
                                           £,               X
predicted.  However, fuels containing little or no chemically bound nitrogen
(natural gas and some oils) often exhibit a maximum in NO  emissions as
excess O  is varied.  This means that on any given boiler a decrease in
                                                             6001-49

-------
excess O  may either increase or decrease NO  emissions for low nitrogen
        fc                                   X
fuels, depending on where the maximum lies.  For fuels with high nitrogen
contents a reduction in excess 0_ nearly always produces a reduction in
NO .
  x
2.2     POLLUTANTS OTHER THAN NO
                                x
        Particulates, sulfur oxides, and smoke are the boiler emissions other
than NO  which are most generally limited by regulations.   Carbon monoxide
and unburned hydrocarbons are also of interest, although their concentrations
are seldom high in a well-tuned boiler.
2.2.1   Particulates
        In the EPA Method 5 particulates determination  (Ref. 12), the
Particulates content of the flue gases is determined by sampling isokinetically
and passing the sample through a probe to a heated filter.  Afterwards the
probe is rinsed with acetone.  The acetone is boiled off and the weight of
the probe washings and the weight gain of the filter are added.  This weight
is then divided by the fuel heat release represented by the volume of sampled
flue gas to obtain the emissions in Ib/million Btu.
        Isokinetic sampling means drawing the sample into the probe without
changing the velocity of the gas.  In this way the streamlines are not
disturbed (see sketch below).
          a.  Overspill
b.  Underspill
c.  Isokinetic
     sampling
        If the streamlines were disturbed, the sample would not represent
the true particulates content of the flue gas.  With too low a sampling rate
(overspill), too many large heavy particles would be captured because their
                                                            6001-49

-------
high inertia would to some extent prevent their being spilled along with
the gases.  With too high a sampling rate, on the other hand, not enough
of the large, heavy particles would be captured per unit of gas sampled.
However, there is less error from using a sampling rate which is too high,
than a rate which is too low.
        Although EPA Method  5 is widely used, local regulations may require
different methods of determining particulates.  In some localities, for
example, sulfuric acid mist  is included with particulates.
        When EPA Method 5 is used,  the two most important sources of parti~
culates are  (a) ash and (b) unburned condensed-phase combustibles like char,
soot, tar, or unburned oil or coal particles.   (Occasionally, bits of corroded
metal from the boiler or heavy dust loading in the inlet air can cause high
particulates readings.)
        Particulates emissions in gas firing are usually negligible.  When
firing heavy oil in a well-tuned boiler, combustible particulates are often
of the order of 0.05 Ib/million Btu, but can range from 0.01 to 0.05 Ib/million
Btu.  If the particulates limit is 0.10 Ib/million Btu, then the ash content of
the oil must usually be kept lower than 0.05 Ib/million Btu  (about 0.1% ash in
oil) so that the total of ash and combustibles is below 0.10 Ib/million Btu.
In oil firing, virtually all of the ash is emitted.
2.2.2   Smoke
        A very small amount of unburned combustible in the flue gases can
produce a noticeable opacity in the stack plume. " The presence of a visible
plume does not necessarily mean that particulates concentrations is above
0.10 Ib/million Btu.
        Official opacity measurements by EPA Method 9  (Ref.  12) must be made
visually by an observer who  has taken a training course and passed a certi-
fication test.  Recertification of an observer is required every six months.
        A typical opacity limitation is a maximum of 20% steady-state, with
a 40% maximum  for no more than two minutes in any one hour.
                                                                 6001-49

-------
        Opacity depends on many variables such as size of particles, N0_
content, weather conditions, precise time of observation while smoke
fluctuates, etc.  The best way to avoid protracted debates over the accuracy
of opacity measurements is to ensure that the boiler emits no visible
smoke.
2.2.3   Carbon Monoxide (CO)
        Carbon monoxide emissions from a well-tuned boiler are normally
negligible.  However, some adjustments which reduce NO  emissions have a
tendency to increase CO emissions, and the latter can be a sensitive indicator
of how far to go in reducing NO .  The following levels are of interest as
rough guides:
         CO,  ppmV        CO, volume percent,
        dry @ 3% 02          dry @ 3% 0=	
           0-50                0-0.005        Usual range for normal operation
         100-200               0.1-0.2        Good level for minimum NOX ,-
                                              CO may increase very rapidly if
                                              02 is lowered further
           400                    0.5          Lower limit of resolution of some
                                              control-room instruments.
                                              Heat loss due to CO no longer
                                              negligible.
2.2.4   Unburned Hydrocarbons
        As long as CO emissions are below 400 ppmV, emissions of unburned
hydrocarbons from boilers are usually very small:  20 ppmV or less.  Normally
it is not  considered necessary to measure unburned hydrocarbons from a
boiler, even during NO  reduction testing.
                      X
2.2.5   Oxides of Sulfur
        The total amount of sulfur oxides (SO ) emitted from the stack is
                                             X
largely determined by the sulfur content of the fuel, although in coal
firing some of the sulfur may be retained in the ash.
                                                            6001-49

-------
        Sulfur in various coals should be evaluated in terms of mass per
unit heating value; a fuel which is low in weight percent sulfur can have
high SO  emissions if its heating value is also low.
        The heating values of oils vary less than those of coals, and one
percent sulfur in oil produces about 580 ppmV  (dry, corrected to 3% excess
0 ) of SO  in the flue gas.
of SO .
     x
        Pipeline quality natural gas in the U.S. gives negligible emissions
        Normally about 98 to 99% of the SO  from fossil-fuel combustion in
boilers is emitted as sulfur dioxide  (SO ).  However, the remainder is emitted
as sulfur trioxide (SO ), which when combined with the water vapor in the
flue gases or the ambient air forms sulfuric acid  (H SO.) and its hydrates.
Because these have a higher dew point than does water, a persistent sulfur
acid mist plume may be observed in cold weather when the SO  concentration
is high.  Sulfuric acid condensation can also contribute to boiler cold-end
heat exchange surface corrosion and acid smut fallout.
        Sulfur trioxide is formed by oxidation of  sulfur dioxide.  Some of
this oxidation occurs in the flame, and some is due to catalysis by solid
materials such as vanadium on boiler tubes in the  convective section.  For
a given fuel, the simplest way to minimize SO  formation in both places is
to minimize oxygen concentration by operating the boiler with low excess air.
                                                            6001-49

-------
                                  SECTION 3.0
                            METHODS FOR NO  CONTROL
                            	 x 	

        Methods for reducing NO  and other combustion-generated pollutants
can be divided into three categories:
        1.   Changing the fuel or its properties
        2.   Modifying the combustion system
        3.   Flue gas cleanup.

        This report is mainly concerned with Category 2.   The Thermal DeNO
                                                                          X
Process,* described in a later section, might be considered a flue gas clean-
up method but it is included in this report because the process is one of
selective combustion and takes place only at temperatures well above normal
boiler flue gas exit temperatures.
3.1     FUEL TYPE
        From the discussion in the previous sections, it is evident that
fuels such as natural gas and low-nitrogen oils might be used in place of
residual oil and coal to lower NO  emissions.  However, the fuel situation
                                 x
in the U.S. is such that more coal and heavy oil will be used as boiler fuels
in the future.
        The effect of coal rank on fuel nitrogen conversion is largely
unexplored.  However, some recent data published (Ref. 13) show that western
(U.S.) subbituminous coals have about 15 to 20% lower NO  emission than
                                                        x
eastern bituminous coals with comparable fuel nitrogen contents on a heat-
value basis.  The difference is believed due to the lower flame temperature
associated with the lower heating values and higher moisture contents of
the western coals.
*Patented process of the Exxon Corporation.
                                      10                    6001-49

-------
3.2     EXCESS AIR
        A detailed discussion of the effects of excess air on emissions and
efficiency is contained in a guideline manual for boiler performance improve-
ment (Ref. 1).  That manual contains step-by-step procedures for optimization
of excess air.
        Figure  1  shows  the NO trends of a number of industrial boilers as
a  function of excess 0  in the flue gas for natural gas, oil, and coal.
Generally, the  slopes of the NO versus O  lines decrease from coal to oil
to gas.  This makes excess 0_ control critical in terms of NO  control for
                            2                                x
nitrogen-containing fuels.
        Some of the curves exhibit a maximum.  This results because although
the  oxygen content of the flue gas increases with increasing excess air,
the  flame temperature decreases.  As a result of these two opposing effects,
thermal NO  emissions pass through a maximum.  The fuel-derived NO , being
          X                                                       X
rather temperature-insensitive, increases monotonically with increased
excess air.
        Coal shows the  highest sensitivity to excess air, with an average
reduction of 50 ppm for each 1% decrease in excess oxygen.
        Low-excess-air  firing of a boiler requires additional control equipment
and  instrumentation to  monitor emissions; however, this mode of operation
generally results in an increase in system efficiency which translates
directly into fuel savings.  Thermal efficiencies of 77 to 87% are typical
of normally operating industrial boilers.  A reduction in excess oxygen of 1%
will result in  an average efficiency improvement of 0.5%.  As the price  of
fuel rises, investment  in such improvements becomes more feasible economically.
        Low excess air  can be an effective NO  reduction technique for oil
                                             x
and  coal, providing that care is taken not to produce excessive amounts of
CO and smoke.
        A point may be  reached where further reductions in excess air will
result in severe  flame  instability.  This limit will usually occur in the
range of 2 to 5% excess O  for pulverized coal and residual oils, 1-3% ex-
cess O  for distillate  oils, and 0.1-0.5% excess 0  for natural gas.
                                       11                   6001-49

-------
  500-
  400-
  300.
  200
              800
              600
              400
              200 I	1	L
o
2
   30CU
w  200,
w
Q
H
X
O
05
E-i
H
Z
   10Q.
    600
              400
6   200
A
o<
   200,
   10Q
    400
              200
                                       Coal

                                       Fuel
        0246
                                   8  10   12   14
I     I    I
                         I	I
                                       Oil

                                       Fuel
       0246
                                   8   10   12   14
                                       J	I
                                       Natural

                                       Gas  Fuel
                  02   4    6    8   10  12  14


                 FLUE GAS EXCESS OXYGEN,  %


                 -t-
         	1	1	1	1	1	1-
           10   20  30   40   50  60

          FLUE GAS EXCESS AIR
                                                70
  Figure   1.
     Effect of excess oxygen on nitrogen oxides

     emissions.   Single lines for watertube boilers,

     shaded areas represent firetube boiler data.
                                               6001-49
                               12

-------
        Low-excess-air firing can be very beneficial from the emissions and
efficiency standpoint; however, such systems must also be sophisticated
enough that a minimum of operator interaction is required.   Oxygen analyzers
must be available and properly utilized.   It is also important to provide
for measurement of combustibles and NO .

 3.3     STAGED COMBUSTION
         Staged combustion involves introducing some of the  combustion air
 into the furnace through overfire air ports (called NO  ports)  or through
                                                       X
 burners to which the fuel flow has been  terminated.   The active burners may
 thus be operated at air-to-fuel ratios which are in some cases  well  below
 stoichiometric.   The average 0  concentration and average temperature at
 the burner are reduced,  and this reduces formation rates of both thermal
 and fuel-N-derived NO .   Most of the nitrogen-bearing organic molecules
 are cracked to form N_ in this first stage.
         That part of the fuel which is not completely burned at the  burner
 combines with the air introduced at other points in the furnace to complete
 the combustion.   In this "second-stage"  combustion,  the temperatures are
 relatively low because the fuel and air  are at this point diluted by
 combustion products which have lost heat to the boiler walls.
         Few existing small industrial-size boilers are amenable to this NO
 reduction approach without significant modifications.  Most industrial
 boilers have only one or two burners, and no capability to  introduce
 additional combustion air through overfire or NO  ports. Figure 2 shows
 the schamatic of a 40,000 Ib/hr industrial watertube  boiler  as  modified to
 allow staged air injection at various points in the  furnace  (Ref.  2).  Figure
 3 gives the results of the staging tests on natural  gas for  this unit.   The
 data show that greater NO  reductions occur as:
                          x
         .  the burner becomes more fuel-rich,  and
         .  the separation from the burner of the staged air
           addition is increased.
                                        13                   6001-49

-------
Port
Nos.  |W|7 IVV9   ^11
t
183 cm
1

s
Jindbox
^^^•i
^m

mm
mmmm

Furnace
^

4 219
"f 166 cm
«- 86 .J
cm |
                                              z?;
                                           13,15
                                             II
                        310 cm
                                             ll

                                             II
            36  cm dia
            Manifold
                       6 1X18  rVllO  rV112,14
                        (a) TOP VIEW      Sidefire Air Fan
Windbox


Fur
86 on
(
\.
rPort 6,
Noe.
cm

nace
14,15
80 cm i 83 cm. 61 cmC^T —
T I
1 1 1 on
^ O O C\ *
J \J \~S \J
7 8,9 10,11 12,13




w.


l<
366 <


                        (b)  SIDE VIEW
                                              Dividing Wall
 Figure  2.  Schematic diagram of staged air system installed
             on a 40000 Ib/hr watertube  boiler.
                            14
                                             6001-49

-------
     120
     110..
     100
w

3   9
•H
X  (N
O O
  z

0) to
z c

•H
m
4->

8
      70..
      60. .
      SO..
                 240
                 220
                 200
                ^80 __
               (2)
                 160
                 140
                 120 	
                 100 	
                                                                         120
                       Theoretical Air  at  Burner,  % of Stoichiometric
Figure 3.  Reduction in total nitrogen oxides  due to staged combustion air,

           natural gas fuel  (40000 Ib/hr, watertube  boiler).



                                                                    6001-49
                                       15

-------
At 95% theoretical air through the burner, the NO  could be reduced from
                                                 X
160 ppm to 120 ppm by admitting the staged air through ports 14 and 15
rather than any ports closer to the burner.  It was found that higher
overall excess air had to be used in order to control combustible  (CO)
emissions at high degrees of staging.  The baseline data (the open circles
of Figure 3) were run at 1.9% 0  and the staged test through ports 14 and 15
(the inverted triangles) were run at 3.4% 0 .  Improvements in low excess
0  control in highly staged configurations will come only with advanced
designs.
        Figure 4 contains the data for staged combustion while firing No. 6
oil on this same 40000 Ib/hr watertube  boiler.   The conclusions  to be  drawn
from the oil tests are the same as from the gas tests; however,  the minimum
NO  levels attainable are not as low (100 ppmV for natural  gas and 150 ppmV
  X
for oil).   This is believed due to fuel nitrogen conversion to NOX-
        Design of a NO  port configuration for a new boiler is done largely
on an empirical basis? there are no set rules.   Water-model studies may be
helpful because the flow pattern is so complex in most furnaces.
        For a given secondary air supply pressure, the principal variable
the designer can work with is the number of NO  ports.  The minimum total
flow area of the NO  ports is set by the air supply pressure.
        For too large a number of NO  ports, the air jets issuing  from them
will be small and too easily deflected by the burner combustion product
stream  (the crosswind), and will not penetrate the stream.   For too few NO
                                                                          X
ports, their air jets may penetrate well but may be too large to be
effectively mixed.  The NO  port air stream would persist beyond the end of
                          X
the furnace.
        In general, delaying the mixing of second-stage air and first-stage
combustion products allows the products to lose more heat and lowers  NO .
However, the flow pattern in the furnace may in some cases be such that
moving the NO  ports themselves downstream has little effect.
                                        16                  6001-49

-------
  en
  0)

  •H
  x
  O
    in
  0 (0
  M
  4J
  s
      225.
       20CL .
       175- -
       150- .
       125- -
100. -
        75. .
        50- •
        25-.
                  400
                  350
                  300
         £250

         <*>
         ri
                -a 200
                a
                  150
                  100
                   50
                                  I
                              ^FUEL RICH
                    BASELINE NO

                    (3.0% 02)
                                            Symbol  Port Open

                                             O  None (1.6<02<6.2)

                                                 6 & 7
                                             Q  8 & 9

                                             Q  10 & 11

                                             O  12 & 13

                                             0  14 & 15

                                             Q  10,11,14 & 15

                                             A  8,9,10 & 11
m

v


6^

v
Figure
                   80       90      100      110    120     130      140


                   Theoretical  Air at Burner, % of Stoichiometric


4.  Reduction  in  total  nitrogen oxides emissions due to staged
    combustion air, No.  6 oil fuel (40000  Ib/hr watertube boiler).


                                                           6001-49
                                        17

-------
        The downstream limit on NO  port location is determined by emissions
                                  X
of smoke and combustibles.  The temperature of the combustion product and
air mixture must be above the ignition point.
        In multiburner furnaces, staged combustion can be achieved by shutting
off the fuel to one or more of the burners while maintaining the same air flow
through all of the registers.  Operating a boiler with the fuel terminated
to one or more of the  burners while  maintaining approximately  the  same  over-
all excess air level,  is not difficult for units with four or more burners.
The difficulty lies in empirically selecting the best burner to turn off and
the best settings of excess 0 , air register, etc. which will allow satis-
factory operation at reduced NO  levels without smoke or carbon monoxide
formation.
        Burners-out-of-service operation has been evaluated for a number of
multiburner industrial boilers (Ref. 2).   Significant NOX reductions (18-49%)
were obtained with a variety of burner patterns for both oil and gas fuels.
Operation in the region below 100% theoretical air at a burner was limited
by smoke and CO formation.  NO  reduction is most significant in this region
if proper fuel/air mixing occurs.  In most cases when burners were taken out
of service, the overall excess oxygen level had to be increased above the
normal level in order to maintain satisfactory CO and smoke levels.
        Results of the study are summarized in Table 1 and indicate the
importance of burner pattern, position of the particular burner removed
from service, and burner theoretical air.  These results do not represent
ultimately achievable lower limits  (i.e., low NO  emission and low excess
0  level) since the burner settings were not optimized in each configuration.
In some cases, taking burners out of service forces the unit to a lower
load since the fuel system on the remaining in-service burners cannot handle
the increased fuel flow necessary to maintain full load.  This problem can
usually be solved simply by using larger capacity oil gun tips or drilling
out gas orifices.
                                      18                    6001-49

-------
                        TABLE 1.  EFFECT OF TERMINATING FUEL FLOW TO  SELECTED  BURNERS


Burner Pattern


1 2
O 0

o o
3 4



138
s O O O

o2o4 o6



O1
O2O3

o o oo
1234



Fuel
No. 6
Oil
No. 6
Oil

Nat.
Gas

Nat.

Gas
P.S.t
300


P.S.
300

No. 6
Oil
No. 5
Oil



Capacity
103 Ib/hr
160

105

60

300


59


65


90

85




Test Load
10 3 Ib/hr
112
118
80
76
76
46
41

200

204
46
43
47
43
50
51
52
71
60
58
51
50
50
50

Burner Out
of Service
ABIS§
2
ABIS
2
1
ABIS
2

ABIS

1
ABIS
1
3
5
ABIS
2
4
ABIS
1
ABIS
1
2
3
4
Excess
Oxygen
(%)
7.8
8.3
7.0
6.2
6.55
2.6
2.6

2.7

3.8
2.9
6.1
5.5
8.1
5.8
5.0
5.0
7.4
8.2
7.3
8.1
8.4
8.4
8.2
Burner
Theoretical
Air (%)
157
123
149
105
108
114
82

114

91
117
118
113
135
139
109
109
115
109
152
122
124
124
123

NOX


NOX/ Normal O2 comments
(281)*
0.6
(289)
.83
.74
(242)
.84

(178)

.57
(619)
1.05
.83
1.1
(466)
.98
.94
(245)
.71
(338)
.65
.72
.8
.85
Normal O_
2
Normal O

Normal O

Normal O_
2

Normal O


Normal O


Normal O

Normal 0



*Numbers in parentheses indicate NOx concentrations  for  normal  operation.
t-rhe API gravity of PS 300 oil  is typical of No.  5  oil.
§"ABIS"  means all  burners in service
                                                                                      6001-49

-------
        Figure 5 presents data in which burners-out-of-service operation
was investigated for natural gas and refinery gas on a horizontal, three-
burner boiler (Ref. 2).  The open square data points are for all three
burners in service at two different excess air settings.  The NO  decreases
                                                                X
with decreasing excess air.  Further reductions are obtained when the fuel
to the middle burner is terminated.  During this off-stoichiometric firing,
the overall flue gas excess 0. concentration had to be increased from a
low-excess-air configuration of 3.7% with all burners in service to 5.3%
O_ for the greatest degree of staging.  This mode of staged combustion
reduced the NO  emissions by 30%.
              x
        The horizontal burner arrangement is not the most responsive to a
burners-out-of-service modification.  A four-burner boiler with a square
burner pattern (two rows and two columns) was tested with pulverized coal
(Ref. 2).  A 40% reduction in nitrogen oxides emissions from 1011 to 618 ppmV
was achieved by removing the top two coal burners from service while main-
taining the same load (75% of design) and excess O » using the burner ports
for second-stage air injection.  The carbon monoxide emissions were essentially
zero for all tests and boiler efficiency did not change from the baseline
condition of 88%; however, smoke increased from a No. 2 to an unacceptable
6.5 Bacharach number when the top burners were out of service.  In this con-
figuration one must sacrifice NOX reduction for smoke reduction.  A trade-
off between NOX and smoke might, for example, be achieved at a 30% NOX reduc-
tion and an acceptable smoke number.
        In general, the results on industrial boilers showed:
           Burners-out-of-service operation is seldom more effective
           than just low-excess-air operation with all burners in
           service.
           Operating with burners out of service on a boiler with
           a square burner pattern  (rows and columns) is more
           effective than with a staggered pattern.  This is
           probably a result of better mixing between burners in
           the regular pattern than in the staggered arrangement.
                                       20                   6001-49

-------
                             Tests  123,  124,  146,  147,  and 151
                             Fuels:  Natural  and Refinery Gas
         150
         125. _
         100. _
     •H
     X
     O  C

     ci
     Q)
     tr> en
     0 nJ
     lH
     -U ^
     •H \
     Z DI
     4J
     o
75. .
          50- .
          25- -
••/ W V I
250
200
(N
O
<*>
n
 150
•o
§, 100
a
50
0
1 1 1 1 1 1
FUEL RICH I AIR RICH
COMBUSTION 1 COMBUSTION
.£*'"' ^P-
*' .^ 0X0' -
^^^ ^^H ^^B
.^r ^^H .^^^
^""^ ^^B^^^ ^ ^^^^
Q Natural Gas
__ O Refinery Gas _
—, ^s Burners Out of
U ^ Service
B 9 No Burners Out
A | | | | of Service
OV 80 90 100 110 1'
                            Theoretical  Air at the Burner, % of Stoichiometric
Figure 5.  Reduction in total  nitrogen  oxides due to taking burners out of
           service, mixture  of natural and refinery gas  fuel.
                                                                       6001-49
                                       21

-------
            Removing a  burner  from service which  is  close  to  the
            convective  section is  more effective  than  removing  a
            burner  from the  row farthest  from  the convective
            section.

            Removing inner rather  than outer burners is more
            effective;  operation at lower excess  oxygen levels
            is possible due  to better mixing of the  air from
            the  burner  out of  service with the combustion  pro-
            ducts from  the operating burners.  Also, the outer
            flames  are  more  effectively cooled.

         These results  indicate that industrial boilers with  multiple  burners
 can be  operated with a fuel injector out of service in order to  achieve  sub-
 stantial NO reductions  (30-40%).   The lowest level of NO achievable is
            x                                             x
 primarily limited  by smoke, particulates or CO increases  that  can  occur.
 Therefore,  the  specific  NO  reduction obtained varies with boiler  type and
                          X
 design.   However,  successful  application requires judicious  selection of
 which burners to take  out of  service.  In order  to  implement combustion  stag-
 ing in  new boiler  designs,  several factors should be  considered:

            To control  NO and provide good fuel  burnout,  suf-
            ficient space should be provided within  the burner
            matrix  so that air-only ports could be installed.
            Staged  combustion  implementation by taking burners
            out  of  service is  more effective if there  are  several
            burners, or NO   ports,  in each vertical  column in which
            the  top row is air-only, the  next  row is in service,
            the  next row  air-only,  and the next row  in service.
            This arrangement can provide  good  NO   reduction while
            maintaining acceptable combustible losses.
            If burners  are to  be taken out of  service  in a retrofit
            situation,  then  additional fuel delivery capability should
            be provided to the in-service burners to avoid load re-
            duction.  Such items as oil gun nozzles, fuel  pump
            capacity and  pressure,  and gas pressure  and orifice
            size, should  be  considered in the  design.

3.4     BURNER ADJUSTMENTS

        Adjustments to the burner fall into the category of modification of
the fuel and air mixing patterns.   These  mixing patterns are modified to
provide as much aerodynamically staged mixing as possible without having to
provide additional  air injection locations  as discussed in the section on
                                       22                    6001-49

-------
 staged combustion.  The low-NO  configuration that one hopes to achieve
                              X
 is a long, narrow flame where the fuel and air mix gradually over its
 entire length.  Such flames can be achieved by reducing the swirl of the
 secondary air and by changing the angle at which the fuel is injected into
 the secondary air stream.  For oil and pulverized coal, it may help to
 reduce the spray angle of the fuel injector.  This technique has been found
 to be especially effective for fuels containing bound nitrogen (Ref. 14).
 However, these low-intensity flames alter the heat transfer characteristics
 of the boiler and may produce increased amounts of particulate emissions.
        In the normal operating range of a burner with variable air swirl,
 decreasing the swirl usually decreases the NO .  This is illustrated by
 Figure 6; in this case NO  could be varied from 210 ppmV down to less than
 140 ppmV at constant excess 0  merely by changing air swirl.
        At very high swirl settings the NO  may again decrease if the flame
                                          X
 changes to a widely flaring shape which is more effectively cooled by the
 walls and by entrainment of cooled combustion products from within the
 furnace.
        With multiburner boilers, experience has shown the most important
 effect of air register adjustments to be in air flow rate to individual
 burners .   This controls  the  air distribution  and  air/fuel  mixture  ratio
across the burner front.

 3.5     REDUCED COMBUSTION AIR PREHEAT
        Reducing combustion air temperature reduces peak flame temperatures
 and to this extent lowers thermal NO  formation.  However, there is a direct
 trade-off with unit efficiency, which drops as preheat temperature is
 lowered.
        Reduced air preheat temperature for NO  control is most effective
                                              X
 with natural gas.  Gas-fired industrial boilers without air preheat typically
 have NO  levels between 50 and 150 ppmV, while those with air preheat have
       X
 NO  levels between 90 and 400 ppmV.
                                       23
                                                            6001-49

-------
           250
          

           150
         T)
         -H
         X
         O

         o
         •H
           100
                                   Natural Gas
  High Negative

Furnace Pressure
                        Decreasing Swirl
                                 Normal

                                Baseline

                                Setting
2.4


2.0
                                                    I
                                 30                60

                          Register Vane Position  (degrees open)
                                                  90
Figure 6.  Effect of air register setting on NO  for a specific boiler.



                                                                  6001-49
                                      24

-------
        The effect of preheated air was investigated in a 5 million Btu/hr
lab scale combustor (Ref. 15).  Data for three fuels and two excess air
levels are shown in Figures 7 and 8.  At normal excess air levels,  (Figure 7),
NO  emissions with coal and oil exhibited little or no dependence on air pre-
  X
heat, but with natural gas NO  emission was increased at higher air preheat.
                             X
At high excess air levels, (Figure 8), the effect of preheat with natural
gas remains about the same as for normal excess air.  NO  emissions with oil
are slightly higher but the increment with increased preheat is unchanged.
With coal fuel there is a significant increase in NO  with increased air pre-
heat.
        Additional field test data have been obtained (Ref. 2)  which support
the above results.  Figure 9 contains these data for three boilers firing
oil and natural gas over a range of preheat temperatures.

3.6     FLUE GAS RECIRCULATION (FGR)
        Applying flue gas recirculation requires that a fraction of the flue
gas be withdrawn from the stack and ducted to the burners.  During the com-
bustion process, the recirculated flue gas acts primarily as a diluent,
lowering the combustion temperatures.  This reduction in bulk gas temperature
suppresses the "thermal NO" formation.  In addition, the presence of the
recirculated flue gas reduces the oxygen concentration, which will tend to
reduce the conversion of fuel bound nitrogen to NO .
        There are two bases commonly used for expressing flue gas recircula-
tion quantitatively.  In the first, the mass flow rate, m, of recirculated
flue gas (FGR) is divided by the total flow of combustion products plus
recirculated flue gas:

where CA signifies combustion air and f signifies fuel.  This formula gives
the percent flue gas in the total oxidizing gas fed to the burners.
                                      25
                                                           6001-49

-------
               1000
                500
              O
              z
                          Coal
                              Gas
                                       500
                                   Preheat,  °F
1000
Figure 7.  Effect of air preheat at normal excess air levels.
                1000
              a  500
              o
                          Coal
                          Oil
                             Gas
                                        i
                                       500
                                    Preheat,  °F
1000
Figure 8.  Effect of air preheat at high excess air.
                                                                6001-49
                                       26

-------

200.




150-
U) rH
0) -H
-0 O
-H
X M
O 0
4-1
c
0) -
o* 0s*
4->
•1-1 in
Z it)
r-4 .. 100.
,200
it
•0
»
B
Q^
a,


100




0
1 1 1 1
,
°d*
— ./V Boiler rated —
>-xXQ ^ at 44500 Ib/hr
^r O^ steam flow
\
Boiler rated at 40000 1
Ib/hr steam flow J""l
1
— J[^^ ^J~"^ —
^B^^^^\**^^\]^^L* ^^i
^^^p3[^^ ,— /^
>«^i 1 ' ta
^•— * /^— '
^^^^^r S
\\ *^
D Q /
Boiler rated at r- j
•250000 Ib/hr LJ —
steam flow
^ 0 Baseline Air Temp.
Q Natural Gas
O No. 6 Oil
i 1 1 1
                                    100
200
300
400
500
1

300



350

1 ,

400
K
1

450

I

500

                                         Combustion Air Temperature

Figure 9.  Effect of combustion air temperature on total nitrogen oxides
           emissions with gas and oil fuels for three watertube boilers.
                                        27                         6001-49

-------
         In  the  second method  the  amount  of recirculated flue  gas  is  divided
 only by  the mass of  flue  gas  leaving  the stack:

         (%FGR)=
                2     (mcA  +  mf)

 This  represents the  ratio of  gas  recirculated  to  gas  emitted  from the  stack.
         Formulas  (1)  and  (2)  both convey  essentially  the  same information,
 but the numerical  values  are  different.   Formula  (2)  can  theoretically give
 values  approaching infinity,  while numbers  from formula  (1) can  only approach
 100%.   Formula (1) is used  in this report,  but it is  necessary when examining
 data  from other sources to  make sure  the  data  are compared on a  consistent
 basis.   The  mathematical  relationship between  the two is  plotted in Figure  10.
         Nitrogen oxides reductions obtained with  flue gas recirculation in  an
 industrial boiler  are plotted in  Figure 11  for both air-  and  steam-atomized
 No. 6 oil and for  natural gas (Ref. 2).   By careful operation, it was  possible
 to  reduce the excess oxygen from  3% to 1.3%.   With the reduced excess  oxygen,
a reduction in NO  of about 20% was realized at 20% circulation for No. 6 oil.
                 x
To achieve this reduction  at such  a low excess air level  required constant
 monitoring of the  exhaust gases for carbon  monoxide and smoke.  Reductions
 on  natural gas were  much  higher,  up to 70%.
         Since the  reduction with  natural  gas was  so successful,  the question
 was raised as to the effectiveness with a distillate  oil  that contained little
 or  no fuel bound nitrogen.  An attempt was  made to simulate a low-nitrogen
 fuel  by firing a combination  of  natural gas and No. 6 oil.  The  results
 are shown by the diamonds in  Figure 11.   The NO  reduction  is somewhat higher
 than  was possible  with residual oil alone.
         Other studies have  also indicated that the NO reductions obtained
 with  flue gas recirculation depend very strongly  on the type  of  fuel burned.
 Reductions of approximately 22% were obtained in a laboratory  combustor fired
                                        28                  6001-49

-------
                70
                60
                50
           o
           o
          a
          o
                40
             14-1
             f(  30
                20
                             I          I           I          I
                            10        20



                                  '"FGR


                             (ACA + V + *
 30




x 100
40
50
Figure 10.  Relationship between two definitions of  "percent flue gas

            recirculation. "
                                                                 6001-49
                                       29

-------
               100,
                80
             3
             5
                60
             I  40
             0
             •o
                20
 s
Q No. 6 Fuel Oil, Air Atomized

Q No. 6 Fuel Oil, Steam Atomized

Q Natural Gas Fuel

   Natural Gas Fuel & No. 6 Fuel
   Oil, Air Atomized
                                   D
                              10           20          30
                                   Flue Gas Recirculation, %
                                              40
Figure 11.  Reduction in total nitrogen oxides emissions by flue gas
            recirculation with constant excess air.

                                                              6001-49
                                       30

-------
with a distillate oil doped with 1% pyridine, while 85% reductions were
realized with the distillate fuel alone (Ref. 16).   A reduction of 33%
was obtained in a 50 hp Cleaver Brooks boiler fired with a residual fuel
oil (Ref. 7).  Reductions of 25%, 38%, and 68% when firing residual No. 6
oil, distillate No. 2 oil, and natural gas, respectively (30% flue gas
recirculation) have been observed (Ref. 9).  The differences were attributed
to  (1) the different combustion characteristics of the fuels, and (2) the
bound nitrogen content (natural gas - 0%,  No. 2 oil - 0.05%, No. 6 oil -
0.35% by weight).
        Figure 12 shows results obtained for flue gas recirculation into
the secondary combustion air for coal, oil and natural gas  (Ref. 15).  The
greatest reductions were achieved on natural gas, with coal and oil following
in that order.  These data indicated that FGR was not very effective for NO
                                                                           x
reduction on coal or oil.
        It was also observed that FGR is not a very effective NO  reduction
                                                                X
technique on large pulverized-coal-fired boilers (Ref. 11 j.  Figure 13 shows
NO data for FGR with all burners in service at two different loads and for
staged firing at two loads.-  The gas recirculation reduced NO  emissions
                                                             X
by 16% with all burners in service but had only a small effect when combined
with staged combustion.
        Data from a natural-gas-fired utility boiler, Figure 14, shows the
effect of gas recirculation on a 320 MW tangentially  fired boiler.  At 30%
recirculation, NO  emissions dropped below 100 ppm from a baseline of 270 ppm,
for a reduction of 66%.
        The method or location of injecting the flue  gas has also been found
to be very important.  There are several methods for  adding flue gas to the
burner.  It can be mixed with the combustion air, or  a separate passage can
be provided.
        Flue gas recirculation at three different locations was evaluated
in an industrial burner firing into a 39" diameter tunnel:  total air,
secondary air, and injectors located at the exit of the burner throat  (Ref. 17)
                                       31
                                                            6001-49

-------
   100,
<#>
o
z
0
•ft

4-1

u
                 Flue Gas Recycled,  %
 Figure 12.  Secondary flue gas  recirculation.
                                                    6001-49
                        32

-------
     1000
  ct
                                          525

                      ALL BURNERS IN SERVICE
          475MW,
RICH)  FIRING
                                                           . 5% 0
                                                            O 350 MW,
                                                               5.0% 02"
      300
     200
     100
                 J	I	I	L__L
                   J	I	L
                             10                   20
                               PERCENT GAS  RECIRCULATION
                               30
Figure  13.   NO dependence on gas recirculation for normal (air-
              rich)  and two-stage combustion of pulverized coal.
                               33
            6001-49

-------
                       % Recirculation,  W  /W  + W
                                         UK  r    a
Figure 14.  Flue gas recirculation with natural gas firing;  320 M

            tangentially-fired unit.  (Bagwell,  et.  al, J.A.P.C.A.,


             11, November  1971, p.  704).
                                                                6001-49
                                      34

-------
 Typically,  50% reductions were obtained when the flue gas was added to the
 primary or secondary air.  When the flue gas was added through the injectors
 at the exit of the throat, no reductions in NO  were obtained.
                                               x
         In a 90 hp firetube facility (Ref.  9; flue gas could be added to the
 primary air, secondary air, total air,  throat injectors, or through the
 natural gas ring when firing oil fuel.   These results showed:
         1.   No NOX reductions obtained when the flue gas was added
             through the throat injectors or secondary air stream.
         2.   Adding the flue gas to the total air stream or primary
             air stream was effective in reducing NO  emissions.
                                                    X
         3.   Adding the flue gas to the total air stream is probably
             the most effective overall.  When large quantities of
             flue gas were added through the primary, the NOX emissions
             tended to rise.  This was probably due to the increased
             mass flow through the primary air passage with associated
             increased fuel air mixing rates.
         4.   NOX reductions were obtained with only small increases
             in smoke emissions, typically one Bacharach number.
         In a recent study, discussed more fully in Section 5, flue gas was
 added to the combustion air just a few inches upstream of the throat. For
 natural gas, it was necessary to use high air swirl and a central gas gun
 rather than a ring in order to stabilize the flame.
         The results are shown in Figures 15 for natural gas, 16 for No.  2
oil of negligible nitrogen content, 17 for No. 6 oil (0.20% N).  Substantial
reductions in NO  were achieved with gas and No. 2 oil at 3% excess 0_, and
                x                                                    2
a significant reduction was achieved with No. 6 oil when excess oxygen was
minimized while using FGR.
         In considering the use of FGR for NO  control on new industrial
boilers, several factors should be evaluated.  In light of the preceding
discussion, the type of fuel to be used should be of paramount concern:
for some boilers FGR is only effective with natural gas fuel.  Therefore,
                                        35                    6001-49

-------
     120
     100"
13
 *
 o
O

of
      60
      40
  o
  z
      20
                                         Location 19
                                         Load:   83%  of Rated
                                         Fuel:   Natural Gas
                                                         x 100
                      (176)
             Excess  Oxygen
             (  )  Test  Number
                                                   Stability Limit
                            10        15         20

                            Recirculated Flue Gas,  %
                                                         25
30
Figure 15.  The effect of flue gas recirculation rate on NO  emissions
            (natural gas).                                  x
                                                          6001-49
                                       36

-------
125
                                         Location 19
                                         Load = 83% of Rated
                                         Fuel:   No. 2 Oil (0.00% N)
                                         Air Atomization
                                                        x 100
       (Test No.)
        Excess O
                     10       15       20        25

                       Recirculated Flue Gas, %
 Figure 16.    NOX emissions as a function of percent flue gas
              recirculation (No. 2 oil).
                                                     6001-49
                               37

-------
         250
     <*>
     n
     I
     a,
     o
     z
         200
         150
         100
          50
                     Normal 0  (3.2%)
Location 19

Load = 83% of Rated

Fuel:  No. 6 Oil (0.20% N)

Air Atomized
                 % FGR
                                               100
                               10        15        20        25


                            Recirculated Flue Gas, % of Total
                                                      30
Figure 17.  The effect of flue gas recirculation on NO  emissions  (No. 6 oil)



                                                                     6001-49
                                        38

-------
if gas is in limited supply as an industrial boiler fuel, FGR may not be
a useful control technique.  The cost per unit of NO  reduction for FGR versus
other NO  reduction techniques should be weighed before selecting a NO  control
        X                                                             X
combustion modification.  Some factors that make up the cost-to-reduction
quotient are:
           impact on overall energy efficiency (auxiliary loads)
           system maintenance
           flame stability (safety)
           amount and method of injection
        No overall recommendations can be set for FGR application to industrial
boilers due to the large number of designs covered by the category.  Individual
cost estimates must be made according to unit size and type.

3.7     WATER INJECTION
        Water injection into preheated combustion air can be thought of as a
simpler alternative to flue gas recirculation.  To be effective in reducing
NO , the water must be evaporated before passing through the flame.  Water in-
jection involves high stack gas heat loss and is not very effective on fuel-N-
derived NO .  The method has been used on boilers, but normally is considered
          x
unacceptable because of the associated efficiency loss.
3.8     FURNACE CHARACTERISTICS
        As the rate of heat release of a particular furnace increases, the
NO  emissions increase only slightly for gas and oil; however, the increase
is rather more marked for coal.  The coal flame radiates less effectively to
the walls due to its large optical density; this lowers the heat removal
from the core of the flame whence the liberated heat goes to increase the
average temperature of the flame and also the NO  emissions.

        Figure 18 presents data on NOX emissions as a function furnace load
for gas and coal fuels (Ref. 15).
                                      39                    6001-49

-------
    1000
                             20                   40
                         Rate of Heat Release,  10   Btu/ft  /hr
Figure 18.  Effect of heat release rate on NO emissions.
           60
                                                                 KVB 6001-49
        Liberal furnace volume makes it easier to use certain NO  reduction
                                                                X
techniques which involve distributing the combustion throughout a larger
volume.  These include low-excess-air firing and staged combustion.  However,
large furnaces in general can cool the flame more effectively.  For this
reason, another  useful parameter  is  the  furnace  heat-absorbing  area.
        For gas-fired industrial boilers it was found that by using the furnace
heat absorbing area as the basic parameter and then correcting for the furnace
excess air level, a reasonable correlation of NO  data could be achieved
(Ref.20).   The test data for watertube boilers for which the necessary furnace
information was available are shown in  Figure 19.   The band shown includes
load variations, air preheat effects, different burner ring designs, and air
swirl effects.  The very low N0x values measured for test 67 are seen to be
largely due to the high furnace heat-absorbing area for the amount of heat
released.
                                       40
6001-49

-------
        Since these data were taken from boilers having various types of
wall construction, it was necessary to estimate the actual heat-absorbing
surface area using the available information.  It was assumed for boilers
with construction of water tubes separated by refractory that only half of
the area on the furnace sidewalls, ceiling, and the end wall opposite the
burner was effective heat absorbing area.  The furnace floor and burner
wall (as well as the refractory between the tubes) were assumed to be inef-
fective as heat-absorbing areas.  The boilers with either tangent-tube or
flanged-tube construction were assumed to have wall and ceiling areas, as
described above, 100% effective as heat-absorbing surfaces; the burner
wall and the floor were assumed to be ineffective.
        The data for firetube boilers do not fall into agreement with data
for the watertube boilers, but appear to form another band of data with a
much shallower slope than for the watertube boilers; this is shown in
Figure 20.  The lower levels of NO  are due in part to the lack of air
preheat on all the firetube units and may be partly due to the differences
in the heat-transfer characteristics of the furnace.

        Data for much larger (utility) boilers are compared to the industrial
boiler data in Figure 21.  A significant aspect of the utility boiler data
is the spread seen in the NO values around the band of the industrial boiler
data.  A possible explanation of this spread is the influence of the gas
ring orifice design.   Boiler No.  6 has a large number of small  gas  orifices,
while boilers No.  4 and No.  5 have approximately half the number of gas
orifices of boiler No.  6,  but of a larger orifice diameter.   The remaining
boilers shown have rings varying in number and size of gas orifices between
those of No. 4,  No. 5,  and No.  6.
         Heat release per unit heat-absorbing area can of course be lowered
 by reducing the firing rate.   Emissions of NO  almost always decrease as
 boiler load is  lowered.
         The reduction of total steam load to control NO  emissions would not
 be an acceptable control strategy in any but the most drastic  of circumstances.
 However, the change in NO  emissions with unit firing rate might be utilized
                                       41                    6001-49

-------
        500
       .400
      T)
      a 300
      a
      * 200
      o
      z
        100
                                                                    Ambient
       Fraction Theoretical Air x
100,000         200,000        300,000

           	Furnace Heat Input
                                   Furnace Heat Absorbing Area
                                      ,  Btu/hr/ft
                                                                         400,000

                                                                          2
Figure 19.  EPA ring burner test results for watertube boilers.
        200

|l50
I
.. 100
rvj
O
m
® 50
x
o
z
0
J | | Test Number
O 37
D
_^~ •

A 1 1 ^^^^*"^*^^..^ l_J ^*
^C^^» — •• "*^^"^^^f ^t^P /^
^y "*" y B 	
	 M— — ^~O" —
__ 	 	 	 v
\ \ \
0 100,000 200,000 300,000
1 Furnace Heat Input 1
38
39 —
40
41
48 	
KVB 80-hp
Boiler



400








,000

        Fraction Theoretical Air x
                                   Furnace Heat Absorbing Area
Figure 20.  EPA ring burner test results.  Firetube boilers, ambient-temperature
            air.
                                      42                        6001-49

-------
    T3

       600
       500
       400
       300
    £  200
    0*
    2  100
           0           100,000         200,000

             Fraction Theoretical Air x
           300,000
400,000
    Furnace Heat Input	I
Furnace Heat Abosrbing Area! '   .  2
Figure 21.  Comparison of EPA industrial boiler ring burner test  results to
            utility boiler ring burner test results.

                                                              ;•-••••• 6001-49
                                       43

-------
in a plant containing sufficient excess capacity such that the required steam
load could be produced using all boilers at partial load, as opposed to using
fewer boilers at full load.  A minimum-NO  firing strategy might then be possible
without limiting steam production.
         Field tests of both firetube and watertube industrial boilers indicate
that there is no significant difference in emissions between the two types of
units when firing the same type of fuel without air preheat (Refs. 2, 4).
3.9      BURNER DESIGN
         There are several studies in process to develop new burner designs
for improved NO  control.  The design approaches vary, but there are several
common burner characteristics which are sought:

         1.   Good atomization and mixing,  allowing operation at low
             excess  air without emissions  of smoke or combustibles,
         2.   Good flame stability and absence of pulsations under
             conditions of low or negative excess air (substoichio-
             metric  air flow)  or flue gas  recirculation at the burner,
         3.   Flexibility in the form of adjustment mechanisms for
             altering flame shape and local air velocities and/or
             the pattern of fuel distribution.
         Good flame  stability with flue gas recirculation may in some cases
 be aided by dividing the air flow into two parts, keeping the flue gas out
 of the part which stabilizes the flame.  However, the flue gas will be
 ineffective if it is completely removed from the flame.
         Flexibility can be provided by having multiple variable-swirl air
 registers,  movable  diffusers, variable oil gun-diffuser relationship, etc.
         If a furnace is so long and narrow that flame impingement on the
 wall is not a problem, a narrow oil spray atomizer  (perhaps 30°) could help
 to lengthen the flame and thus reduce the NO .
                                        44                          6001-49

-------
   12. 0
   11.0
   10.0
    9.0
c
o
X

0   8.0
ui
in
0)
u
x
w
    7.0
    6.0
                  T
           T
           T
            T
              Percent Carbon in Outlet.Fly Ash  is Given

              in Brackets  (%C  in Ash)
                                Normal Operation
                                          ID Fan Limited
                                                  [19.9]
             CO,  Clinker, or

             Excessive  Smoke  Limited
                                                                  125.5J
       70
80
90
100
110
                             Unit Load, 10
                                                             120
130
       Figure  22.   Operating limits of a spreader stoker burning a

                     western subbituminous coal.
                                                             6001-49
                                     45

-------
       600
 (Nl
o
<#>
+J
(0 -v
  (A
>i Q)
M -Q
•O O
p( 4J
CX 0>
  .H
 - c
(/) M

O Q)
•H C
U) O

-H U
e >,
w u
o  o
       500
                                Medium Load
                                 (MOB x 103 Ib/hr)
       400
       300
       200
•a M
•H 0)
x >
u
•H
       100
          50
                      60
70
80
90
100
110
                                   Excess Air, %
   Figure   23.    NO   emissions from a spreader stoker burning a western
                  subbituminous coal (coal  fuel N = 0.68%).
                                        46
                                                              6001-49

-------
3.10    COAL FIRING
        There are two methods used to fire coal in industrial boilers:
suspension firing of pulverized coal, and stoker firing of lump coal.
Due to the nature of the combustion in each of these firing types, the NO
                                                                         X
emissions are radically different.  The high intensity of the pulverized
firing, if not properly controlled, can lead to much higher NO  emissions
than from the low-intensity stoker combustion.
        In a stoker the fuel, for the most part, devolatilizes under fuel-
rich conditions in a bed.  The volatiles mix with air and burn above the
grate.  After the volatiles are gone, the remaining carbon burns in a
diffusion flame around the coal particle.  This mode of combustion represents
a naturally staged configuration.  Figure 22 is a plot of excess O  as a
function of unit load and shows the limits of staging that were achieved
with a particular spreader stoker of 160,000 Ib  steam/hr capacity.  The
lower dashed line represents the limit defined by excessive combustible
losses and/or fuel bed clinker formation.  The upper dashed line is the
limit of the induced draft fan capacity.   Normal operation lies between
these two dashed lines and is defined, in this case for a western coal, by
the solid line.   The lower the excess air requirements at any load the lower
will be the NO  emisisons, as shown in Figure 23 for the same fuel and unit.

        An illustrative example of the effect of stoker firing type on NO
emissions can be made by comparing the data in Figure 24 with the data in
Figure 23.  The data in Figure 23 are for a spreader stoker, in which about
half of the combustion takes place in suspension because the coal is fed to
the furnace by a rotating paddle wheel.  Combustion is completed after the
coal particles fall onto a grate.  On this unit, NO  emissions were in the
300-400 ppm range, depending on excess 0 .  Figure 24, on the other hand,
contains NO  data (for two coals)gathered on a watercooled vibrating grate
stoker where more than 90% of  the combustion  takes place  in  the  fuel  bed.
NO  emissions on this unit were in the 150-250 ppm range, depending on
excess air.
                                      47                    6001-49

-------
     250
 rg
O

 w
 U)
 0)
 u
 X
 0)
(*)
 a
     225 -
      200
     175
     150
     125
                    ?
                      /
                   /
                 /
                                                                15000 Ib/hr
      oL
        0
                                                       Eastern Coal

                                                 _ —  (Kentucky Vogue)



                                                 /\  14000^15000 Ib/hr

                                                 ^  25000  Ib/hr

                                                 D  40 Ib/hr

                                                 	 Western Coal

                                                       (Wyoming Bighorn)

                                                 Q  25000Ib/hr

                                                  •   15000-18000ib/hr
                                                            10
                                                                     11
                                                                              12
Figure 24.
                                      Excess  O ,  %



            Water-cooled vibrating grate  stoker, 45,000  Ib/hr steam.

            Comparison  of nitric oxide  emissions from  western and eastern

            coal.
                                         48
                                                                     6001-49

-------
        Other  factors considered when choosing a  stoker design  are  listed
below:
            flexibility with  regard  to coal type and  size
            turndown  ratio
            unit  size/steam demand
            efficiency
For  example, it  is factors such as  these  that dictate  the  use of a  spreader
stoker  furnace instead of  an underfed stoker for  applications above about
50000 Ib/hr steam.   The spreader stoker has excellent  turndown  ratio character-
istics, can burn a wide range of fuel from municipal waste to coal, and is
more efficient than  the underfed stoker.  The older  stoker designs  such as
the  traveling  grate  and underfed stokers  cannot develop the heat release
rates necessary  for  th§ larger industrial boilers.   For these reasons  a
stoker  type cannot be chosen merely on the basis  of  its NO  emissions.
                                                           X
3.11    THERMAL DeNO  PROCESS*
                     X
         A recent development which has  performed successfully in commercial
 demonstrations is the Thermal DeNO  Process  (Ref.  18).   Ammonia (NH.)  is
 injected into the combustion products at a point where  their temperature is
 1750 ± 50 °F.  Despite the presence of large quantities of O  , the NO
 is a more powerful oxidizer and a large percentage of the NH, reacts with
 NO:
        4 NH  +6 NO  -»•  5N+6HO.
            J               fi      £,
 A competing reaction which uses some of the .NH  is

        4 NH  + 3 0   -*•  2 N  + 6 HO.
However, under favorable conditions  (optimum temperature, good mixing, low
ratio of NH  to NO), tests indicate that fr
ammonia molecules  react with NO molecules.
ratio of NH  to NO),  tests indicate that from 50 to nearly 100% of the
*Patented processes of Exxon Corporation
                                       49                   6001-49

-------
        If the temperature of the combustion products is lower than optimum,
a small amount of hydrogen (not enough to appreciably raise the temperature)
can be injected with the ammonia.  This has a beneficial effect in that it
lowers the temperature at which optimum NO  destruction takes place.  However,
it is of course less costly to inject pure ammonia where the temperature is at
its optimum.  A moving injector  (similar to a traversing sootblower) with a
thermocouple attached might be most suitable for this process.
        If the temperature is too high, the ammonia acts like other nitrogen-
bearing fuels:  NO  is formed rather than destroyed.  If the temperature is
too low, the ammonia does not react and is emitted to the atmosphere.
        The Thermal DeNO  Process was first demonstrated commercially in a
                        x
refinery boiler in Japan in 1974.  Results are shown in Figure 25.  The process
has since been used in several other boilers in that country.  There has as
yet been no commercial demonstration in the U.S.
        The only potential problem seen thus far is possible cold-end corrosion
due to formation of ammonium bisulfate when the Thermal DeNO  Process is used
                                                            x
with sulfur-bearing fuels.  Ammonium bisulfate is a corrosive liquid in the
temperature range 400-500 °F.  To date, however, there is apparently no con-
firmation that such a problem occurs.
        For a normal retrofit application the cost of the Thermal DeNOx  Process
was estimated by Exxon Research and Engineering Company to be 7-15$ per million
Btu fired  (Ref.  18).
        An obvious advantage to the process is that modifications of the
boiler's main combustion process are not necessary.  However, for boilers
with high firing rates it is still desirable to minimize NO  production
in the  furnace in order to minimize the cost of the ammonia and hydrogen used.
                                        50                   6001-49

-------
OP
c
0
• M
-P
u
3
-O
S
o
z

                       • Boiler
                       A Furnace
                                   a) Up to 70% reduction in NOX  emissions was
                                      achieved in a Kawasaki refinery  boiler in
                                      1974.
                                               b)   NOX versus temperature without
                                                   hydrogen injection
       NH  /NO  Ratio-Mole/Moles
                                               300

                                               250
                                               too

                                               SO
                                                                    NOppm
                                                 1400   1600  1800  2000
                                                         Temp.  ° F
                                 c)  NOX  versus  temperature with hydrogen
                                     injection.
    1200  1400  1600  1800
              Temp-°F
                Figure 25.   Results for the Thermal  DeNOx Process.
                                        51
                                                             6001-49

-------
                                  SECTION 4.0
           IMPACT OF THE COMBUSTION MODIFICATIONS ON BOILER EFFICIENCY
        Trade-offs between NO  emissions and boiler efficiency for a range
                             X
 of combustion  modifications  have been  examined  (Ref.  2).  The data for  low
 excess  air,  overfire  air, burners out  of service, reduced air preheat and
 flue  gas  recirculation are presented below.
 4.1    EFFECT OF EXCESS AIR
        In general, lowering the excess air in  order  to reduce NO  emissions
                                                                 X
 will  also result in lower dry gas losses and improved unit efficiencies.
 However,  a point is reached  at which combustible losses in the form of  CO
 and carbon become unacceptable from both an emissions and an efficiency
 standpoint.
        The  measured  effect  of low-excess-air firing  on boiler efficiency is
 illustrated  in Figure 26.  The majority of the  data points are located  in
 the best  quadrant, i.e., where a reduction in emissions is accompanied  by
 an increase  in efficiency.   Efficiency was bettered by as much as 2.5%  in
 two cases.   On one coal-fired boiler,  reducing  excess oxygen resulted in a
 44% reduction  in nitrogen oxide emissions along with  a 2.0% increase of
 boiler  efficiency.  In three cases  with gas fuel, lowering the excess oxygen
 resulted  in  an increase  in emissions at a higher efficiency.
        The  efficiency of most of the  boilers tested  was improved and NO
                                                                         J*
was reduced as the  excess oxygen was reduced.

 4.2    EFFECT OF STAGED AIR
        The  effect of NO  ports on  boiler efficiency  is illustrated in  Fig-
 ure 27.  Generally, the  reduction of nitrogen oxide emissions by using  NO
                                                                         X
 ports had an adverse  effect  on efficiency.  This behavior was to be ex-
 pected  since NO ports normally requires that the level of excess oxygen
 be maintained  at higher  than baseline  levels to assure complete  combustion.

                                       52                   6001-49

-------
    •H
    u
    •H
    w

    c
    •H

    0)
    CT>
    C
     o &
,QQ§d
1 6*
-50 -30 WA

••




_ +2

_ +1
•t— |
D
1 I
1 1
+10 +30 +50
- -1
- -2
- -3
                  - •«-  Change  in Total Nitrogen Oxides, % -»• +



         ^    Coal Fuel


         0    Oil Fuel


         D    Natural Gas  Fuel




Figure  26.  Effect of reducing the  excess  combustion air on

             boiler efficiency.
                                                   6001-49
                                53

-------


+
t
Of
>t
O
§
•H
Effic
c
•H
1
m
u
^

i



Most Desirable Quadrant




-
D
i 9 AJ i -
1 \r^ V Ir+l
-so Qi-30 Dir1
TJ u
0 A ^

D
A


D
D D
D

__ +3
_. +2


-d1


\
+10 +30 +50

^.-1



	 2


	 3

                    - •*•  Change  in Total Nitrogen Oxides. % -+- +

                A Coal Fuel
                O Oil Fuel
                Q Natural  Gas Fuel
Figure 27.  Effect of NO  ports on boiler efficiency.
                                      54
                                                              6001-49

-------
This greater quantity of heated air being exhausted through the stack
contributes significantly to the lowering of unit efficiency.   When excess
0  is kept constant, staged combustion can increase efficiency slightly.
        A few boilers exhibited increases in efficiency when staged air
was used.  These boilers had staged air ports which were part of the
original boiler design and were therefore more appropriately sized and
located; excess 0  could be kept low.
        Figure 28 presents the effect which taking burners out of service
has on boiler efficiency.  The efficiency changes were generally small,
0.6% or less, but were mostly in the positive direction.  One would expect
the effect of burners out of service to be similar to that of NO  ports since
                                                                x
both techniques involve staging combustion.  The quantity of test data from
burners out of service on industrial boilers is small, making it difficult
to'draw any concrete conclusions.

4.3     EFFECT OF COMBUSTION AIR TEMPERATURE
        The effect of varying the combustion air preheat temperature is
shown in Figure 29.  As expected, lowering the temperature to reduce emissions
resulted in a degradation of boiler efficiency, because a reduction in air
preheat was accompanied by an increase of flue gas temperature.  The tests
where the efficiency increased were tests where the air temperature was
raised, rather than lowered.
        If reduced air preheat is to be adopted as a permanent nitrogen oxide
emissions reduction technique for a particular boiler, the stack losses can
be recouped by redesigning the steam side of the boiler for more heat
absorption.  An example of this would be the installation or enlargement of
an economizer.
4.4     EFFECT OF FLUE GAS RECIRCULATION WITH AND WITHOUT STAGED COMBUSTION
        The effects of flue gas recirculation on boiler efficiency are shown
in Figure 30.  Also illustrated are the effects of flue gas recirculation
combined with staged air.  Flue gas recirculation, alone, had only small
                                        55                   6001-49

-------
      t
      *
      u
      c
      a)
      •H
      u
      •H
      <4-l
      14-1
      U

      C
      •H
      A
      U

      \

       I
               Most Desirable Quadrant
 D
-10
                                           +2
        +1
+10     +30     +50
                                           -1
        -2
                                           -3
               - -*- Change in Total  Nitrogen  Oxides,  %  -»- +




               A  Coal Fuel

               Q  Oil Fuel

                  Natural Gas  Fuel
Figure  20.    Effect of operating with  burners  out of service on

              boiler efficiency.

                                     56                6001-49

-------
u

0)
•H
U
•r-l
u

C
0)
en
u
       Most Desirable Quadrant
                        Open  Symbols Represent
                         Reduced Preheat Tempera-
                       3   ture


                        Solid Symbols Represent
                        Increased  Preheat Temp.
-50
-30
+10
ff
n +30
+50
                                 -3
              Change in Total Nitrogen Oxides, %
         ^ Coal Fuel

         Q Oil Fuel

         • Natural Gas Fuel
Figure  29.   Effect of combustion air preheat temperature on boiler
            efficiency.
                              57
                                       6001-49

-------
  O
  c
  a)
 •H
  o
 •H
 M-1
 IM
 W

  C
  
-------
effects on efficiency; however, it was quite successful in reducing NO
                                                                      X
on gas fuel but generally not as successful on oil.  (It should be noted
that the ASME boiler efficiency does not account for the auxiliary load
of running the recirculating gas fan system.  In some cases this added
penalty in terms of total energy use may be significant.)   The efficiency
changes were 0.6% or less and varied from positive to negative.  When
staged combustion was combined with FGR, the efficiency dropped by about
1.5%, partly due to the necessary increase in the overall excess oxygen
level for complete combustion.
        In general, the nitrogen oxide reduction techniques of low excess
oxygen firing and burners out of service resulted in boiler efficiencies
equal to or better than baseline levels.  NO  ports and reduced combustion
                                            X
air preheat produced a degradation of efficiency.  Flue gas recirculation
gives a small degradation of efficiency.
                                       59                    6001-49

-------
                                  SECTION 5.0

                           COMPARISON OF TECHNIQUES
                     ON TWO SPECIFIC INDUSTRIAL BOILERS
        Although exhaustive pilot-scale tests of combustion modifications
have been carried out on a number of laboratory-scale boilers, comparisons
of techniques on working industrial boilers have been comparatively rare.
Two such comparisons were recently completed (Ref. 22) and the results will
be summarized here.  Most of the important NO  reduction techniques were used
                                             X
except for the Thermal DeNO  Process.
5.1     COMBUSTION MODIFICATIONS ON A BOILER RATED AT 17,500 LB/HR OF
        STEAM FLOW
        The boiler was a single-burner 17,500 Ib/hr oil- and gas-fired
watertube unit with no economizer or air preheater.  The tests were run at
83% of full load.  Second-stage combustion air was injected through four
steel lances which were inserted through the windbox and the refractory firing
face.  Flue gas was withdrawn from the base of the stack and was added to
the burner air through wide slots just upstream of the throat.
        Baseline emissions of NOX (corrected to 3% excess 02, dry) were:
           Fuel                            NO , ppm  (dry)
         Natural gas  (Ring burner)               96
                      (Gun burner)                92
         No. 2 oil (less than 0.01% N)          110
         No. 6 oil (0.20% N)                    221
        Oil atomization was by high-pressure (28 psig) air.  A gas ring
normally used for natural gas injection was found to be unstable with the
particular flue-gas recirculation pattern used.  Therefore a gas gun was
used, rather than a ring, for most of the gas fuel tests.
                                      60                    6001-49

-------
        The percentage reductions in NO  are shown in Table 2 and the
resulting changes in efficiency are given in Table 3.  The No. 2 oil results
are unusual in that a 68% NO  reduction was achieved with flue gas recircula-
tion alone.  This may be because the high-intensity mixing and atomization
by the air atomizer, and the relatively high volatility and low nitrogen
content of the fuel, combined to give a gas-like flame.
        It was noted that boiler efficiency was reduced by flue-gas recircu-
lation even though no increase in excess 0  was necessary.  This was because
the average temperature the combustion process was reduced, lowering the
temperature-difference potential for heat transfer.  The flow per unit area
increased, tending to increase heat transfer, but the reduced temperature
difference resulted in a greater offsetting tendency for reduced heat transfer.
        The results, shown in Tables 2 and 3, were achieved with CO concen-
trations of less than 500 ppmV and particulates emissions of less than 0.10
Ib/million Btu.
5.2     COMBUSTION MODIFICATION ON A BOILER RATED AT 45,000 LB/HR OF
        STEAM FLOW
        The boiler was a single-burner oil- and gas-fired packaged water-
tube unit rated at 45,000 Ib per hour of steam flow.  The rated heat input
was about 55 x 10  Btu/hr.  This unit was equipped with an air preheater pro-
viding a combustion air temperature of up to 350 °F.
        This boiler was modified to provide ten ports in the furnace side
walls, five on each side, for introduction of staged combustion air.  The
configuration of the ports was previously shown in Figure 2.  Staged air at
ambient temperature was provided by a separate fan.  Therefore the staged
air was not preheated.  The air preheater for combustion air had provision
for bypass of combustion air around the preheater so that combustion air
temperature could be varied.
                                       61                   6001-49

-------
                TABLE 2.  SUMMARY OF NOX REDUCTION AS A  FUNCTION
                OF COMBUSTION MODIFICATION TECHNIQUE FOR VARIOUS
                                 FUELS — LOCATION 19
Lowered 0 19 *
Staged Combustion Air, 32
Normal O^
Staged Combustion Air, 42
Low 02
Flue Gas Recirculation,
Normal O_
2
Flue Gas Recirculation,
Low Oj
FGR + SCA, Normal O2
FGR + SCA, Low 0
3.3
46

—

77


79

76
t
20
30

44

68


73

69
77
30
29

42

11


40

53
55
* Reduction occurred at increased O  level in this case.
t Stability limits prevented lowering 0_.
                                      62
6001-49

-------
             TABLE 3.  SUMMARY OF CHANGE IN BOILER EFFICIENCY
                       DUE TO COMBUSTION MODIFICATIONS
                                LOCATION 19
Natural Gas
Mode Ring* Gun*"
Baseline efficiency 78.2 82.6
Low O - 1.2 +0.9
Staged Combustion Air, +0.1 +0.3
Normal ©2
Staged Combustion Air, +0.5 	
Low 02
FGR, Normal O 	 - 0.8
FGR, Low O 	 - 0.4
FGR + SCA, Normal O 	 - 0.5
FGR + SCA, Low 02 (Note 1)
No. 2 Oil
82.5
+ 1.5%
+ 0.9%
+ 1.1%
- 1.9%
+ 0.9%
- 1.2%
- 0.8%
No. 6 Oil
82.5
+ 1.5%
+ 0.1%
+ 0.8%
- 0.7%
+ 0.6%
- 0.8%
+ 0.1%
*  Standard  ring burner

   Optimized gas gun
NOTE (1):   Unable to lower 0  due to stability limits.
                                                            6001-49
                                      63

-------
        Baseline emissions of NO  (corrected to 3% excess 0 ,  dry)  were as
follows:

                 Fuel                        NO ,  ppm (dry)
              Natural gas                          161
              No. 6 oil (0.31%N)                   286

        The effect of each combustion modification on both NO  emissions and
                                                             x
boiler efficiency is compared in Table 4.   This comparison shows that lowered
oxygen was more effective for NO  reduction with No.  6 oil while lowered air
preheat was more effective with natural gas.  Lowered oxygen level increased
efficiency slightly for both fuels.   Lowered air preheat reduced efficiency
by 1% with gas and 2.8% with oil.  Staged combustion  at normal oxygen was
equally effective for both fuels with only slight changes in efficiency.
Staged combustion with low excess oxygen produced the largest NOX reduction
observed with No. 6 oil.  A maximum reduction in NO  of 70% was achieved with
                                                   x
natural gas by the use of lowered air preheat in combination with staged
combustion.  However, this change resulted in a reduction in efficiency of
3.2%.
5.3     CONCLUSIONS
        Reductions in NO  emissions of between 70% and 80% with natural gas
                        x                                               ^
fuel were observed for the two modified industrial boilers tested.  These
reductions resulted in NO  emission levels of between 20 and 50 ppm at 3%
O , dry, and were achieved with the use of various combined combustion modifi-
cations including lowered excess air, staged combustion, lowered combustion
air preheat, and flue gas recirculation.  With the exception of lowered air
preheat, these reductions were achieved with only slight changes in efficiency
and no apparent detrimental changes in other emissions.  With lowered air
preheat, efficiency was reduced about 3%.
        On the one boiler tested with No. 2 oil, NO  reductions of about 75%
                                                   X
were observed with the use of flue gas recirculation and reduced excess air.
This reduction resulted in NO  levels of the order of 25 ppm at 3% O , dry,
                             X                                      £•
and was obtained with an increase in efficiency of about 1% (neglecting fan
power requirements).
                                       64                        6001-49

-------
           TABLE 4 .   NOX REDUCTION AND EFFICIENCY CHANGE
              FOR  COMBUSTION  MODIFICATION TECHNIQUES
                       TESTED AT  LOCATION 38



Baseline NO , ppmV at 3% O
X £•

Combustion Modification
Technique
Lowered O_
Lowered Air Preheat (LAP)
Staged Combustion Air (SCA) ,
Normal 0_
SCA, Low O
LAP + SCA
Baseline efficiency, %

Combustion Modification
Technique
Lowered 0
LAP
SCA
LAP + SCA

Natural Gas
161
!0 Reduction ,
6
24
31
—
70
81.2

Change in
+0.2 .
-1.0
-0.4
-3.2
No. 6 Oil
(0.31%N) (a)
286
Percent of Baseline
21
9
33
42
—
84.8

Efficiency, Percent
+0.8
-2.8
+0.2
—

(a)
   Fuel oil nitrogen content, percent by weight.
                                                         6001-49
                                 65

-------
        Reductions in NO  with No. 6 oil were not as dramatic as for the other
                        X
fuels.  However, it was significant that a 55% reduction could be achieved
with the use of combined flue gas recirculation,  staged combustion, and
lowered excess air.  It is believed that reductions of this magnitude have
not been achieved previously with the heavier fuels of the No. 6 type.
        The use of flue gas recirculation can result in flame instability
but it was found that these problems could be resolved by relocating the fuel
injection point to areas where flue gas is not fully mixed with air.  This
technique requires that the flue gas be only partially mixed  (and stratified)
in the burner throat as opposed to the more common method of complete mixing
in the windbox.
        The degree of the NO  reductions achieved with natural gas and oil
                            X
fuels is very encouraging for the prospects of controlling NO  emissions from
industrial size boilers.  However, the differences between the two boilers
tested in NO  emissions and reductions achieved indicate that important unit-
            x
to-unit variations exist in terms of the relative effectiveness of these NO
control methods.  Therefore, the specific reductions obtained should not be
interpreted as achievable on all industrial boilers.
                                      66                   6001-49

-------
                                  SECTION 6.0
                        COST  OF THE CONTROL TECHNIQUES

        Building an industrial boiler which incorporates modified combustion
techniques involves considerations of control, compactness, and customer
acceptance which are beyond the scope of this report.   It is not practical
to give plausible estimates of the cost of equipment needed to achieve a
certain result on the wide variety of industrial boiler types.  However,
some order-of-magnitude estimates of the cost per unit reduction for compari-
son of various techniques have been made (Ref. 19)  and will be reviewed as
a general guide.
         The cost components  are  shown in Table 5.   The first term represents
the size-independent or fixed capital costs for such things as design,
instrumentation, etc. while the second term represents the size-dependent
capital costs such as for equipment,  installation,  etc.  These two capital
costs are annualized at a uniform 20% per year.  The third term represents
the operating costs, which depend on the boiler size and the operating hours
per year and accounts for fuel penalties or savings, as well as power or
chemical consumption for the  control.  The denominator represents the emis-
sion reduction which depends  on the boiler size, its emission factor, the
reduction factor for the control, and the annual operating hours.  For very
small boilers the first term dominates and the cost/reduction ratio rises
in inverse proportion to the device size.  For very large devices or controls
with high operational costs,  the  third term dominates and the cost/reduction
ratio asymptotes to a value independent of either device size or operating
hours per year.
                                       67                   6001-49

-------
       TABLE 5.   APPROACH TO NO  CONTROL COSTS
                               x
                         A -f
      COST/REDUCTION  =
                             QFRT
                                          TON
  A = Size Independent Annualized Capital Cost, $



 BQ = Size Dependent Annualized Capital Cost, $



CQT = Annual Operating Costs:  Size, Time Dependent, $



  Q = Heat Input Rate, 10  Btu/hr


                                      6
  F = NO  Emission Factor, Tons NO /iO  Btu
        x                         x


  R = Fractional NO  Reduction by Control Method
                   X


  T = Annual Operating Time, Hr/Year
                                           6001-49
                           68

-------
        In developing cost data with which to compare costs on a generalized
basis, rather than for a specific combustion device, one is forced to deal
with only the generic costs.  Unusual site-specific or application-specific
costs cannot be readily factored in.
        Table 6 shows estimated cost/reduction ratios for NO  control by each
of the options mentioned earlier plus an additional one, reburnering with
"low-NO  burners."  Capital costs were annualized at 20% per year.  Operating
       X                                                       6
time was assumed  to 6000 hr/year, and fuel cost to be $2.00/10  Btu.
        Lowering excess air by 10%, in those applications where present
operation permits such a reduction, results in a cost/reduction ratio of over
$1800/ton of NO  (as NO )  reduced for devices smaller than 10 x 10  Btu/hr.
This was estimated assuming that the excess O  could be lowered by 2%, with a
                                      c
reduction of 30 ppmV (about 0.04 lb/10  Btu)  in NO  emissions.  Oxygen instru-
                                                  X
mentation, its installation and checkout were assumed to cost $8,500 per unit.
Because low-O  operation results in reduced fuel use, the initial fixed
costs for instrumentation and burner adjustments become comparatively smaller
for larger boilers.
        For staged combustion a minimum capital cost for setup on small units
was assumed to be $10,000.  For a 50 x 10  Btu/hr boiler a capital cost of $30,000
was assumed.  A 1/4% loss in fuel efficiency, and a 33% reduction from an
                                               /-
initial emission rate of 0.3 Ib NO   (as NO»)AO  Btu were used for the calcu-
                       6          x       2
lations.  For a 50 x 10  Btu/hr boiler, then, the cost/reduction ratio CE would be
       (0.2)($30,000)/yr +  (6,000 hr/yr)(0.0025)($2AQ6 Btu)(50 x 1Q6 Btu/hr)
CE -  		
                    (50 x 10  Btu/hr)(0.1 lb/10  Btu)(6,000 hr)
   =  $0.25/lb or $500/ton.
For staged combustion the cost/reduction ratio drops sharply  for larger sizes
as the fixed costs for implementing this mode are divided by  larger emissions
reductions.  Opportunities for cost ratios as low as shown for the largest
boilers are primarily limited, however, to those with forced draft,  multiple
burners, and air preheat.
                                       69                  6001-49

-------
TABLE 6.   COSTS OF NO  CONTROL BY COMBUSTION MODIFICATION  (Ref.  19)
                     x
(Units of $/Ton NO
Firing Rate
"""^--^Million Btu/Hr
Control ^^^~--^^
Option ^~"~"--^^^
Low Excess Air (-10%)
Staged Combustion (25%)
Flue Gas Recirculation (20%)
Reduced Air Preheat (-100 °F)
Water Injection (1 Ib/lb)
Ammonia Injection (90%)
Reburnering - Low NO Burners
10
1800
1000
2300
1650
2500
1000
800-1000
Prevented)
100
300
300
900
1580
2200
390
450-550

1000
-300
100
700
1500
2000
370
200-300
                                                      6001-49
                              70

-------
        The minimum cost of installing a flue gas recirculation ^system on a
small boiler was assumed to be $30,000.  A fuel efficiency decrement of
1% and a 40% NO  reduction from an initial level of 0.3 lb/10  Btu were
               "                                                   6
used.  For larger boilers the capital cost was taken to be $1000/10  Btu, so
that for a 100 x 10  Btu/hr ratio the cost effectiveness ratio was

„_ _  (0.2/yr)($100,000) + (6,000 hr/yr)(0.1)($2/lQ6 Btu)(100 x 1Q6 Btu/hr)
CE -                      g                  r
                 (100 x 10  Btu/hr)(0.2 lb/10  Btu)(0.4)(6000 hr/yr)
   =  $0.44/lb or $890/ton
The asymptotic cost/reduction ratio at large size stays higher than
for staged combustion because of the efficiency penalty that it causes.
        Reduced air preheat and water injection have comparatively low
capital costs but comparatively high operational costs, resulting in a
uniformly high cost/reduction ratio for all sizes.
        Water injection has the highest cost per ton of NO  reduced for all
                                                          X
boiler sizes and is therefore not considered as a practical technique in
view of the other options available.
        The Thermal DeNO  Process injection is seen to have attractive cost/
reduction ratios for all sizes, leveling off beyond that size where a minimum
installation cost dominates.   At sizes 100 x 10  Btu/hr or larger the costs are
dominated by the chemical costs.  The costs shown are for $200/ton of ammonia
delivered on site.  Other costs are based on proprietary data.  Lest one
jump to the conclusion that this is the universal cost-effective solution,
it must be pointed out that not all combustion devices are configured to
provide access for the injection at the proper temperatures or to provide
the subsequent mixing and reacting zone required.
        The final NO  control costed is that of reburnering with "low-NO
                    X                                                   X
burners."  The unit costs shown here are seen to be attractive compared with
the other alternatives—but again the range of applicability has limitations.
For this option the costs are almost totally for hardware, installation,
adjustment, and checkout with negligible operating costs.  The major variable
in the costs is the degree of change of burner controls, igniters, and flame
detectors necessitated by the new burners.  A NO  reduction of 60 ppmV (0.08
                                                X
lb/10  Btu) was assumed.
                                       71                   6001-49

-------
                                  SECTION 7.0
                                    SUMMARY

        Field tests and laboratory studies indicate that the most promising
techniques for oxides of nitrogen control from industrial-sized boilers
include:
           reduced excess air levels
           staged combustion
           reduced air preheat
           flue gas recirculation (gas and oil fuels only)
           burner redesign and adjustments
           Thermal DeNO  Process
                       x
Table 7 is a summary of these specific techniques and others, indicating their
applicability for NO  control on boilers firing natural gas, fuel oil, and coal.
                    X
        On the basis of the results of the field measurements reported, it
appears to be possible and practical to reduce the total nitrogen oxides
emissions by up to 50% with certain selected combustion modification methods
which are appropriate for the particular unit design.  However, with only two
of the methods is the boiler efficiency unimpaired: excess air reduction,
and burner-out-of-service.
        The burner-out-of-service modification is not as attractive as
reduction of excess air because there is generally a penalty in the maximum
load that is possible unless the in-service burners are enlarged to handle
the increased fuel flow.
        Industrial boilers exhibit a large amount of variability from boiler
to boiler in terms of both design and operating characteristics.  Thus, a given
industrial boiler may respond in a totally different manner compared to another
                                       72                    6001-49

-------
                                                    TABLE 7.   SUMMARY OF NOV REDUCTION TECHNIQUES
COMBUSTION
MODIFICATION
Excess air
level





Staged
Combustion



















Air register
adjustment


Atooization
technique





MECHANISM OF
KOx REDUCTION
Varies the overall
fuel/air nixture
ratio




Creates local
fuel/air ratio
stratification by
bypassing air from
the burner (s) .


Burners -Out -of
Service^ Fuel
flow stopped to
selected burn-
ers.






NO. Ports;
Special ports
provided for air
addition down-
stream of the
burner.
Controls the
swirl Level and
local rate of
fuel/air oixing
Controls fuel/air
nixing rates





OPERATIONAL EXPERIENCE
GENERAL TRENDS
o For high-nitrogen
fuels. KOy emis-
sions decrease with
decreasing excess
air level.


o NOg emissions de-
crease as the air/
fuel ratio at the
burner is decreased



o Successful appli-
cation of these
techniques requires
sufficient test
tide to empirically
determine boiler ad-
justiaents and burner
patterns to allow
operation at low
with a burner out of
service.






o NOX emissions de-
crease as swirl is
decreased

o When properly ato-
mized, the type of
oil atomizer had
little effect on
NOX emissions (air
steam, or nechani-
cal)
EFFECT OF FUEL TYPE
o Excess air level has
the greatest effect
on coal, followed by
No. 6 and S, then No.
3 fuel oil. Little
effect is seen with
natural gas.
o For fuel oil, conver-
sion of fuel nitrogen
to NOX is reduced by
staged coobustion.
Nitrogen conversion
can be reduced to the
range of 20 to lot
compared with normal
levels of 401 to SO*.
Thermal NOx is also re-
duced for all fuels.













o Effects similar for
all fuels


a Only applicable to
oil-fired units






GENERAL APPLICABILITY
Feasible nodification for
both new and retrofit.












Burners -Out-of-Service
Applicable only to mjltiple-
burnez boilers. Not appli-
cable to stoker-fired units.

















Tip changes on existing
units feasible






IMPACT ON IWIT EFFICIENCY
o Improves efficiency
due to lower dry gas
losses




o Often lowers efficiency
since increased excess
Oj levels mst be used
to prevent nooka

















o Little effect on efficiency



o Efficiency will improve if
lower excess Oj operation
results from better
atoaization



o
o

-------
      TABLE 7.   Continued
COMBUSTION
MODIFICATION
Combustion air
temperature
decrease
Firing rate
(\ load)
Flue gas re-
circulation
(FG30
Burner heat
release rate
(Btu/hr/bumer)
Fuel type
change
Boiler tune-up
Thermal deNOx
MECHANISM OF
NOx REDUCTION
Decreased cocbus-
ture decreases
peak gas tempera-
tures
Affects the heat
release rate per
unit volume in
the furnace and
the heat transfer
rates
Reduces peak gas
temperatures and
O- concentrations
by adding combus-
tion products to
the combustion
air (diluents)
Affects local
ccobustion rates
and gas heat
> transfer rates
in the furnace
Reduces fuel -bound
nitrogen content
Assures performance
according to design
specifications
Selective non-
catalytic reaction
OPERATIONAL EXPERIENCE
GENERAL TRENDS
o NO emissions de-
crease with decreas-
ing coabustion air
temperature.
o NOx emissions for
larger burners
seeoi to decrease
with decreasing air
temperature than for
smaller burners.
o Little effect of
firing rate on NOX
emissions on units
without air pre-
heat.
o With air preheat,
NOx cmissions in-
crease with firing
rate.
o NOx emissions de-
crease as the
amount of FGR in-
creases
o Greatest reduc-
tions occur with
first 10* recircu-
lated.
o NOX emissions
increase with
increased burner
loading
o KOX emissions in-
crease with in-
creasing fuel
o The average conver-
gen to NOX is typi-
cally 40* for fuel
oils.
o NOx can increase
or decrease follow-
'ing a boiler tune-
• up depending on the
•as found" condition
o Reduction of NOX
when correct amount
of NHa or H2 is used
EFFECT OF FUEL TYPE
o Effect is mainly seen
on natural gas-fired
boilers, not an import-
ant consideration for
oil and coal-fired
units.
o Effects similar for all
fuels.
o Greatest reductions for
natural-gas-f ired units
which involve all "therm-
al NOX"
o FGR does not greatly
affect conversion of
fuel nitroaen to NOx
o For natural gas, the
NOX relationship de-
pends on the combus-
tion air temperature;
the higher the air
temperature, the nore
important the burner
loading.
o Not applicable to gas
fuels (except in seme
waste fuel streams.)
o Effect similar for all
fuels.
o None known to date.
GENERAL APPLICABILITY
Air preheat presently used
only on the larger indust-
rial boilers

Little field experience
with industrial boilers
available. Feasible for
new boilers, moderately
difficult for existing
units .
Applicable to multiple
burner units.
Depends on suitability and
costs of the fuels due to
constraints on availability
of fuels. Not a generally
applicable technique.

Experience limited.
IMPACT ON UNIT EFFICIENCY
o Lovers unit efficiency

o Lowers overall efficiency
due to extra fan power and
increased stack gas tempera-
ture.


o Should icprove efficiency
o None, but more expensive
•fuels' (NH3, H2> arc usad.
O
O

-------
boiler when a particular NO  reduction technique is used.   Thus while overall
trends can be illuminating with regard to NO  reductions to be expected,  each
unit should be examined individually in terms of its design characteristics
and the choice of NO  reduction techniques to be applied.   It is also true
                    X
that boiler designs may limit the type of combustion modification that is
applicable to certain boiler types.  For example, sufficient furnace space
may not be available for staged combustion implementation.
7.1     EXCESS AIR
        In order to operate industrial boilers at lower excess air levels
the designer should provide controls, alarms, and combustion quality monitors
of the type that are used to operate utility boilers.  Recent advances in
instrumentation technology make combustion monitors very reliable.  Dependable
in-situ O  analyses can be made using the solid state ZrO  monitor.  These
systems have inherently low maintenance requirements.  The closer control
necessary with reduced excess air operation probably precludes boiler opera-
tion by hand, as most industrial boilers are operated;  therefore,  the controls
should be automatic, again following the lead of the utility boiler control
designs.  Projected future fuel costs make these control systems more attrac-
tive because of their fuel savings as well as pollution reduction potential.
7.2     STAGED COMBUSTION
        Staged combustion can be an effective means of reducing NO  emissions;
                                                                  X
however, this modification requires extra air passages and fans in addition
to the combustion controls discussed under the low excess air section.  In
some cases, staging the combustion air works at cross purposes to low excess
air firing since the overall excess air must be increased to control combustible
losses.
7.3     AIR  PREHEAT
        Reducing the combustion air preheat  temperature as a method of NO
                                                                         X
control on existing units  can only be considered as  a  last-resort measure due
to the large efficiency penalties involved.  New boilers would have to be
redesigned to recover most of the excess  flue gas enthalpy involved if no
preheat were used in future designs.  Presently, few industrial size boilers
use preheated air.
                                       75                   6001-49

-------
7.4     FLUE GAS RECIRCULATION
        The use of FGR on gas fuel is a very effective NO  reduction tech-
nique and can be coupled with staged combustion and low excess air to substan-
tially reduce NO  emissions.  The effectiveness is more limited on oil fuel
                X
and very limited with coal.  FGR use does not come without penalties since
an additional high-temperature fan and duct work are required.  Typically,
a unit will be more sensitive to overall excess 0  and its smoke level will
be reduced while using FGR.
7.5     BURNER DESIGN AND ADJUSTMENTS
        Improvements in thermal efficiency as well as reduced pollutant
emissions can be obtained by proper maintenance of burners if, for example,
the overall excess O? can be lowered by improved fuel/air mixing.  New
burner designs are available which advertise low emission operation as well
as versatile operation.  These new designs should be considered when a unit
retrofit is anticipated.
7.6     THERMAL DeNO  PROCESS
                    x
        This process involves injection of ammonia into combustion products
at a point where the temperature is 1750 +_ 50 °F.  The ammonia reacts with
NO, reducing NO  emissions.  Although insufficient operating experience with
this process has been accumulated to date, it may prove to be the most suit-
able technique for industrial boilers.  It is apparently not sensitive to
fuel or boiler characteristics, and to a certain extent is "tunable" to suit
individual boiler characteristics and local regulations.   The equipment
needed apparently is simple, with low initial cost, and the operating
expense is reasonable.   In order to minimize the operating expense, how-
ever, it is desirable that NO  production in the furnace be relatively low.
                             X
                                      76                       6001-49

-------
7.7     LIMITATIONS ON APPLYING NO  CONTROLS
                                  x
        Any modification or redesign of a boiler to incorporate controls
for NO  can affect the combustion process in a way that causes significant
      X
increases in other pollutants.  Monitoring and control of carbon monoxide,
unburned hydrocarbons, particulates, smoke, and opacity must be an integral
part of any program to control NO .  In addition, it is necessary to eval-
                                 X
uate flame scanner performance, flame stability, steam temperature, and
load-following performance.  Long term effects on tube life and corrosion
should also be monitored.  Each of these problems will not necessarily
occur on every boiler.  However, it is important to be aware of the poten-
tial occurrence and to establish limits for NO  control based on adequate
margin to avoid such problems.
                                      77                       6001-49

-------
                                SECTION 8.0

                                REFERENCES

 1.     McElroy,  M. W.  and Shore, D. E.,  "Guidelines  for  Industrial Boiler
       Performance Improvement," EPA-600/8-77-003a,  NTIS No. PB 264  543/OWP,
       January 1977.

 2.     Cato,  G.  A.,  et al.,  "Field Testing:   Application of Combustion
       Modifications to Control Pollutant  Emissions  from Industrial  Boilers,
       Phase  I," EPA Report  650/2-74-078a,  (PB 238 920/AS), October  1974.
       Phase  II, EPA Report  650/2-76-086a,  (PB 253 500/AS), April 1976.

 3.     Lachapelle, D.  G. , et al.,  "Overview  of Environmental Protection
       Agency's NOX  Control  Technology for Stationary Combustion  Sources,"
       presented at  67th Annual AIChE Meeting, Washington,  D.C.,  December  1974.

 4.     Bartz,  D. R., et al., "Staged  Combustion in Combined-Cycle Supplementary
       Fired  Boilers," presented at EPA-ERDA Conference  on  Environment and
       Energy Conservation,  Denver, CO,  November 3-6, 1975.

 5.     Levy,  A., et  al.,  "A  Field  Investigation of Emissions from Fuel Oil
       Combustion for Space  Heating," API  Publication 4099, 1971.

 6.     Barrett,  R. E.  and Miller,  S.  E., "Field Investigation  of  Emissions
       from Combustion Modification Equipment for Space  Heating," API
       Publication 4180, EPA Report R2-73-094a, June 1973.

 7.     Turner, D. W. and Siegmund, C. W.,  "Staged Combustion and  Flue Gas
       Recycle:   Potential  for Minimizing  NOX from Fuel  Oil Combustion,"
       presented at  the American Flame Research Committee Flame Days,
       Chicago,  IL,  September 6-7, 1972.

 8.     Baldwin,  J. D.  C.  and Long, C. H.,  "Effect of Swirl  on  NO   Emissions
       from a Gas-Fired Burner," ASME Paper  74-WA/FU-3,  1974.

 9.     Muzio,  L. J., et al., "Package Boiler Flame Modifications  for Reducing
       Nitric Oxide  Emissions, Phase  II of III," EPA Report R2-73-292b
       (API Publication 4208), NTIS No.  PB 236-752b, 1974.

10.     Sarofim,  A. F.  and Flagan,  R.  C., "NOX Control for  Stationary Sources,"
       Report to National Research Council,  1975.

11.     Thompson, R.  E., et  al., "Effectiveness of Gas Recirculation  and
       Staged Combustion in  Reducing  NOX on  a 560 MW coal Fired Boiler,"
       EPRI NO  Control Seminar, San  Francisco, CA,  February 1976.

12.     code of Federal Regulations, 40CFR60, "Protection of the Environment,
       Parts  60 to 99", Appendix A -  Reference Methods,  U.  S.  Government
       Printing Office, July 1976  (or latest issue).
                                      78                   6001-49

-------
13.     Maloney, K. L.,  "Western Coal Use in Industrial Boilers," presented
        at the Central States Section of the Combustion Institute, Columbus,
        OH, 1976.

14.     Lowes, S. C. et al., "Reduction of Pollution by Burner Design,"
        International Flame Research Foundation,  IJmuiden,  Holland, Doc. No.
        K20/2/74.

15.     Armento, W. S., "Effects of Design and Operating Variables on NO
        from Coal-Fired Furnaces Phase II," EPA Final Report 650/2-75-OC)2b,
        NTIS No. PB 241-283/AS, February 1975.

16.     Martin, G. B. and Berkau, G., "Evaluation of Various Combustion
        Modification Techniques for Control of Thermal and Fuel Related
        Nitrogen Oxide Emissions," presented at 14th Symposium  (International)
        on Combustion, Pennsylvania State University, 1972.

17.     Bell, 0., "Nitric Oxide Formation in Stationary Oil Flames at
        Atmospheric Pressure," Brennst-Warme-Kraft 2&_  (7),  pp. 291-297, 1974.

18.     "A Way to Lower NO  in Utility Boilers," Environmental Science and
        Technology,  p 226, March 1977.

19.     Bartz, D. R., et al., "Control of Oxides of Nitrogen from Stationary
        Sources in the South Coast Air Basin," prepared for California Air
        Resources Board, NTIS PB 247 688/7WP, September 1974.

20.     Setter, J. G. and Arand, J. K., "Staged Combustion for NOX Reduction
        in Industrial Boilers," KVB Report No. 7600-408, submitted to the
        American Gas Association, 1975.

21.     Hunter, S. C., Hall, R. E., Sotter, J. G., and Nazimowitz, W.,
        "Evaluation of Two Industrial Boilers with Combustion Modifications
        for Reduced Pollutant Emissions," American Society of Mechanical
        Engineers, Paper 77-WA/APC-l, August 1977.
                                      79                     6001-49

-------
          SECTION 9.0
        CONVERSION FACTORS
SI  UNITS TO METRIC OR ENGLISH  UNITS

To Obtain
g/Mcal
106 Btu
2
MBH/ft
•>
MBII/ft
Btu
103 Ib/hr*
Ib/MDtu
ft
in
n
ft2
ft3 '
Ib
Fahrenheit

psig
psia
iwg (39.2
6
10 Btu/lir
GJ/hr

From
ng/J
GJ
-1 -2
GJ-hr -m
-1 -3
GJ-hr -m
gm cal
or MBII GJ/hr
ng/J
m
cm

m
3
m
Kg
Celsius
Kelvin
Pa
Pa
«F) Pa

HW
MW

Multiply By
0.004166
0.948

0.08806

0.02684
3.9685 x 10~3
0.948
0.00233
3.261
0.3937

10.764
35.314
2.205
tp • 9/5(tc)+32
t » 1.8K - 460
* A
V - (P 1(1.450x10 1-J4.7
Vpoi»°  U.450xlO )
P1 • (P a)(4.014xlO~3)

3.413
3.60
•Ib/lir of equivalent saturated steam
conversions between ng/J and ppm are approximate and based on
To Obtain ppm. Multiply
at 3% 03 of in

Natural Gas Fuel

CO

HC
NO or NOx
SO or SOx
Oil Fuel
CO

HC
NO or NOx
SO, or SOx
Coal Fuel
CO
KC
MO or NOx
S02 or SOx




typical fuel compositions
Concentration
ng/J by



3.23

5.65
1.96
1.41

2.93

5.13
1.78
1.28

2.69
4.69
1.64
1.18





ENGLISH AND METRIC UNITS TO SI UNITS

To Obtain
ng/J
ng/J
-1 -2
GJ-hr -m
GJ-hr'^ra"3
GJ/hr


n
en
„'
mj
Kg
Celaiu*
Mlvln
Pt
P«
P.
MW
MM

Prom
Ib/MBtu
g/Mcal
2
MBH/ft
MBH/rt3
103 lb/hr«
or 10* Btu/hr


in
ft2
ft3
Ib
Fahrenheit

piig
p*i*
iwg (39.2T)
I06 Btu/hr
OJ/hr

Multiply By
430
239

11.356
37.257
1.055


'
2.54
0.0929
0.02832
0.4536
t - 5/9 
-------
                                 APPENDIX A
                      EMISSION CORHECTION CALCULATIONS
                        FOR FLUE GAS MOISTURE CONTENT

        This appendix provides a procedure for correcting emission measure-
ments to a dry basis.  Some measurement systems provide for condensation
and removal of moisture from the gas being sampled prior to measurement of
the pollutant.  Those systems give the measurement on a dry basis directly
and no moisture correction is necessary.  However, some pollutants, such
as NO , SO., and hydrocarbons would be removed from the sample gas by mois-
ture condensation.  It is therefore necessary to measure these pollutants
on a wet basis, then correct the reading to a dry basis for reporting.
        Each method or instrument used for each pollutant should be exa-
mined to clearly understand whether each compound measured is being indi-
cated on a wet basis or a dry basis.  All data recording forms should
record that basis.
        The equation for correction of a volumetric measurement (in ppm
or % by vol.)  from wet to dry basis using NO  as an example, is:
                          NO
        NO             =    x, meas., wet
        N°x, dry basis      (100 - % H,,0)   X 10°                    (A l)
The values for NO              and % H.,0 should be at actual stack condi-
                 x, meas., wet        2
tions as measured.
        Use of Equation A-l requires knowing the moisture content  (% HO)
of the flue gas as a percentage by volume of the total gas.  The EPA has
provided a standard method for determining moisture content.*  The moisture
content of the stack gas should be measured by this method or a comparable
method for most accurate results.
*Code of Federal Regulations, Title 40, Protection of Environment, Part
 60, Appendix A, Reference Method 4, Determination of Moisture in Stack
 Gases, U.S. Government Printing Office.
                                     81                     6001-49

-------
        When the moisture content cannot be measured, the following equation
can be used to calculate moisture content:
                  (20.95 - % 0      )  x 100
        %H2°  =     (2095/A - i O? „  )                            (A-2)
                                  2, dry
To use this equation, the oxygen content must be known, expressed on a dry
basis.  The factor "A" is a function of the fuel burned and is the percent
moisture that would be present if the fuel was burned with a theoretical
amount of air (zero excess air) .  Typical values of A are given below for
common fuel types:                            .  „ „ _ ,   „ ,
                                              A, % HO by Vol.
                                                at zero % CL
                   Fuel                                    2
                   Natural Gas                      18.4
                   No. 2 Oil                        12.0
                   No. 6 Oil                        11.5
                   Coal                              7.0
A more accurate value of A can be obtained from the ultimate analysis of
the fuel, if available, by!the following equation:

              	1.92 (%H)	
              5.53(%H) + 1.53(%C) + 0.57(%S)  + 0.14(%N) - 0.46(%O)    l ~ '

where %H, %C, %S, %N and %O are the content by percent weight in the fuel
of hydrogen, carbon, sulfur, nitrogen, and oxygen, as determined by the
ultimate analysis.
        When the moisture content is measured, Equations A-2 and A-3 are
not needed.  However, it is a good practice to use these equations anyway
and compare the calculated moisture content with the measured value.  If
the comparison is poor, the measurement equipment should be checked for
proper operation.
                                       82                    6001-49

-------
                                TECHNICAL REPORT DATA
                          (Please read Inuructivns on the revcne before completing)
\. REPORT NO.
 EPA-600/8-77-003b
                           2.
                                                       3. RECIPIENT'S ACCESSION NO.
4. TITLE ANDSUBTITLE
Reference Guideline for Industrial Boiler Manufactu-
   rers to Control Pollution with Combustion
   Modification
                                 5. REPORT DATE
                                 November 1977
                                  i. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
                                                       8. PERFORMING ORGANIZATION REPORT NO.
G.A. Cato, K. L. Maloney,  and J.G. Setter
9. PERFORMING ORGANIZATION NAME AND ADDRESS
KVB, Inc.
17332 Irvine Boulevard
Tustin, California 92680
                                  10. PROGRAM ELEMENT NO.

                                  1AB014; ROAP 21BCC-046
                                  11. CONTRACT/GRANT NO.
                                  68-02-1074
12. SPONSORING AGENCY NAME AND ADDRESS
 EPA, Office of Research and Development
 Industrial Environmental Research Laboratory
 Research Triangle Park, NC  27711
                                  13. TYPE OF REPORT AND PERIOD COVERED
                                  Final; 6/73-9/77      	
                                  14. SPONSORING AGENCY CODE
                                   EPA/600/13
is. SUPPLEMENTARY NOTES
Drop 65,  919/541-2477.
                              project officer for this report is Robert E.  Hall,  Mail
  . ABSTRACTThe repopt degcribes combustion modification methods that are available
to boiler manufacturers for controlling air pollutant emissions from industrial size
fossil-fuel-fired steam boilers.  The methods discussed include reduction of excess
air, staged combustion, air register adjustment, fuel oil atomization, combustion
air temperature adjustment, flue gas re circulation, burner heat release rate, fuel
type, burner tune-up, and ammonia injection. The report summarizes results of tests
on specific industrial boilers and discusses the applicability of combustion modifica-
tions.  Specific guidelines for  boiler design to achieve specified emission levels could
not be given because of the variability in emissions and  responsiveness to control
among the various  industrial boiler types.
17.
                              KEY WORDS AND DOCUMENT ANALYSIS
                 DESCRIPTORS
                     b.IDENTIFIERS/OPEN ENDED TERMS C.  COSATI Field/Group
Air Pollution
Efficiency
Combustion
Boilers
Burners
Nitrogen Oxides
Fuel Oil
Dust
Smoke
Carbon Monoxide
Sulfur Oxides
Fossil Fuels
Natural Gas
Coal	
Air Pollution Control
Stationary Sources
Industrial Boilers
Combustion Modification
Particulates
Boiler Design
Burner Desie
13B
14B
21B
13A

07B
11G
2 ID
18. DISTRIBUTION STATEMENT

 Unlimited
                     19. SECURITY CLASS (This Report)
                      Unclassified
                         21. NO. OF PAGES
                              87
                     20. SECURITY CLASS (This page)
                      Unclassified
                                               22. PRICE
EPA Form 2220-1 (9-73)
                    83

-------