United States
 Environmental Protection
 Agency
 Industrial Environmental Research
 Laboratory
 Research Triangle Park NC 27711
 Research and Development
EPA-600/S7-81 -126  August 1982
Project  Summary
Industrial   Boiler  Combustion
Modification  IMOx  Controls
K. J. Lim, C. Castaldini, R. J. Milligan, H. I. Lips, R. S. Merrill, P. M. Goldberg,
E. B. Higginbotham, and L. R. Waterland
  Volume I of the report gives results
of an environmental assessment of
combustion modification NO, control
techniques for coal-,  oil-, and gas-
fired industrial boilers, with focus on
NO, control effectiveness, operational
impacts, thermal efficiency impacts,
capital and annualized operating
costs, and  effects on  emissions of
pollutants other than NO,.  Major
industrial boiler design types are
characterized  and equipment trends
are reviewed. Currently available con-
trol techniques can achieve 10-25
percent NO, reductions for coal- and
residual-oil-fired boilers and 40 - 70
percent  reductions for distillate-oil-
and  gas-fired units with minimal
adverse operating impacts. Controls
should increase steam costs by only 1
- 2 percent, but the initial investment
required could be significant; up to 20
percent of the boiler cost on a new
boiler and up to 40 percent of the
boiler cost for a retrofit. Volumes II
and III of the report give  results of
detailed Level 1 tests on two stoker-
coal-fired boilers,  indicating that
combustion modification reduces the
source potential environmental hazard
by lowering NO, emissions, leaving
the emissions of other pollutants
largely unaffected.
  This Project Summary was devel-
oped by EPA's Industrial Environmen-
tal Research Laboratory, Research Tri-
angle Park, NC. to announce key find-
ings of the research project that is fully
documented in a separate report of the
same title (see Project Report ordering
information at back).
Introduction
  With the increasing extent of NO,
control application in the field, and
expanded NO, control development
anticipated for the future, there is cur-
rently a need to: (1) ensure that the cur-
rent and emerging control techniques
are technically and environmentally
sound and compatible with efficient and
economical operations of systems to
which they are applied, and (2) ensure
that the scope and timing of new control
development programs are adequate to
allow stationary sources of  NO, to
comply with potential air quality stand-
ards. With these needs as background,
EPA's IERL-RTP initiated an "Environ-
mental Assessment  of Stationary
Source NO, Combustion Modification
Technologies  Program" (NO, EA) in
1976. This program has two main ob-
jectives: (1) to identify the multimedia
environmental  impact of  stationary
combustion sources and NO, combus-
tion modification controls applied to
these  sources, and (2) to identify the
most cost-effective, environmentally
sound NO, combustion modification
controls for attaining and maintaining
current and projected NO2 air quality
standards to the year 2000.
  The NO, EA's assessment activities
have placed primary emphasis on major
stationary fuel combustion NO, sources -
utility boilers, industrial boilers, gas
turbines, internal combustion engines,
and commercial and residential warm
air furnaces; conventional gaseous, liq-
uid, and solid fuels burned in these
sources; and combustion modification
controls applicable to  these  sources

-------
with potential for implementation to the
year 2000.
  Volume I of the report summarizes the
EA of combustion modification controls
for industrial  boilers. It outlines the
environmental, economic, and  opera-
tional impacts of applying combustion
modification controls to  this source
category. Volumes II and III summarize
results of two field test programs aimed
at providing data to support the environ-
mental  and  operational impact evalu-
ation.


Conclusions
Source Characterization
  Industrial boilers are defined here as
coal-, oil-, or  natural-gas-fired steam
generators with heat input capacities of
2.9 - 73 MW (10 - 250 X 106 Btu/hr).
The boilers provide electrical or mechan-
ical power, process heat, or a combina-
tion of these in a wide variety of indus-
tries. This capacity range  does  not
encompass all steam and hot water gen-
erators used in industry. In fact boilers in
this size category represent about 60
percent of the installed capacity used in
the industrial sector. In addition, indus-
trial boilers fire fuel  other than coal, oil,
or natural gas.
  However,  industrial  boilers larger
than 73 MW (250 X 106 Btu/hr)  heat
input are generally similar in design and
controllability to utility boilers. Boilers
smaller than 2.9 MW (10 X 106 Btu/hr),
generally used for hot water and space
heating, can be grouped with  commer-
cial heating  units.  Both utility  boilers
and  commercial  heating units  are
treated  in other NOX EA reports. 1~6
  The  industrial boiler category as
defined here represented the third  larg-
est contributor to stationary source NOX
emissions in the U.S. in 1977, contribu-
ting about 14 percent, as shown in Fig-
ure 1. This share  is expected to remain
high, given incentives to switch to coal
firing in the future. Thus, this same
category represents a priority category
for control evaluation in the NOX EA.
  Coal-fired  industrial boilers are  gen-
erally of  the  watertube design.  Two
major design categories are: pulverized-
coal-fired units and stokers. Pulverized-
coal-fired units  accounted  for  only
about 8 percent of the installed coal-
fired population. But since these units
are almost entirely greater than 29 MW
(100 X 106 Btu/hr) capacity,  they
account for  almost 20 percent of the
coal-fired capacity.  Characteristic de-
signs are similar to those in the utility
                        Noncombustion 1.9%

                  Warm Air Furnaces 2.0%

                    Gas Turbines 2.0%

            Others 4.1%
       Industrial Process
       Heaters 4.1%
                                   — Incineration 0.4%
                         Industrial
                         Boilers 14.4%
                           14.4%
                                Reciprocating
                                1C Engines
                                  18.9%
Figure  1.
                                Total: 10.5 Tg/yr (11.6 x 106 tons/yr)
Distribution of stationary anthropogenic NO* emissions for the yeai
1977 (controlled NO, levels).
sector; tangential and single- and oppos-
ed-wall designs predominate.
  Stoker-fired boilers account for nearly
all the remaining coal-fired installa-
tions. These boilers are classified by the
method  of  introducing fuel to the fur-
nace: spreader,  underfeed,  and  over-
feed. Spreader stokers are most popular
in newer installations.
  Oil- and gas-fired boilers can be clas-
sified as either watertube or firetube.
Both shop-assembled, or packaged, and
field-erected watertube boilers  exist;
however virtually all firetube boilers are
packaged  units  since firetube boilers
are generally limited in size to about 8.7
MW(30X  106Btu/hf).
  There are two  major types of  pack-
aged watertube boilers: horizontal
straight-tube and bent-tube. Newer
boilers are exclusively bent-tube, further
classified by tube configuration. A-, D-,
and  O-tube  configurations are  most
common.
  Industrial firetube boilers can be clas-
sified as horizontal return tube (HRT),
                            scotch, or firebox. The HRT is a two-pass
                            boiler and was the most popular into the
                            1960s. Scotch boilers, in two, three, or
                            four  passes, have since become the
                            most popular. Firebox boilers are shod
                            and compact, employing three passes a!
                            most, finding use in installations where
                            floor  space is limited.

                            Control Alternatives
                              NOX is the primary flue gas pollutam
                            from  industrial boilers amenable to con-
                            trol by combustion  modification. The
                            major combustion modification  tech-
                            niques which have been shown to be
                            effective in reducing NOX emissions
                            from  industrial boilers are: low excess
                            air firing, staged combustion usinc
                            overfire air ports or burners-out-of
                            service,  low NOX  burner designs, flue
                            gas recirculation,  reduced air preheat
                            load  reduction  or  reduced combustior
                            intensity, and homogeneous  reductior
                            of NO* using ammonia injection.
                              Typical baseline (uncontrolled) NO
                            emission factors for  industrial boiler;

-------
are given in Table 1. Note that these are
average  values;  NOX emissions from
individual units can vary  significantly
within a design/fuel category. Sim-
ilarly, the effectiveness of the above NOx
control techniques varies with boiler
design and fuel fired, as well as within a
given design/fuel category.  Thus,  the
following  discussion is organized  by
design and fuel.

Pulverized Coal-Fired  Boilers
  Combustion  modification  NOx con-
trols have been successfully applied to
only a limited number of  coal-fired
industrial boilers.  Those considered
most promising on pulverized coal-fired
units  are low excess air, staged com-
bustion, low NOx burners, and ammonia
injection.
  Low excess air (LEA) operation is rela-
tively simple to implement. It applies to
all boilers and requires only reducing
airf low to the burner windbox. However,
in a multiburner unit, the windbox may
have to be modified to improve air distri-
bution to individual burners during LEA
operation.  Lowering excess air can
reduce the safety margin for complete
combustion. Hence, an oxygen trim sys-
tem may have to be added, in addition to
the normal airflow  controllers. Never-
theless, boiler efficiency gains with LEA
should offset any additional hardware
            costs,  making LEA the most attractive
            NOx control technique for first imple-
            mentation (5 - 25 percent NOX reduc-
            tion). Figure 2 shows results of LEA
            tests on representative coal-fired indus-
            trial boilers. The  slopes of the data
            bands indicate the  relative effective-
            ness of LEA on each equipment cate-
            gory. LEA is about equally effective for
            each.
             Staged  combustion with overfire air
            (OFA) and LEA is the best demonstrated,
            available  control  option for pulverized
            coal-fired industrial boilers, potentially
            reducing NO« emissions by up to 30 per-
            cent. The LEA and OFA control system
            has an advantage over  other control
            systems  because of its commercial
            availability and effectiveness. The cost
            of the system is  not prohibitive  when
            OFA ports are designed as part of new
            boilers. In addition, careful operation of
            staged air injection is not  expected to
            affect emissions of other criteria pollu-
            tants seriously. Burner stoichiometries
            of 100 - 110 percent would achieve a
            20-percent NOx reduction. At these stoi-
            chiometry levels, oxidizing atmospheres
            would prevail in the furnace, thus min-
            imizing concern over possible furnace
            slagging and boiler tube wastage. How-
            ever, achieving more stringent NOX con-
            trol with combined LEA and OFA may
            require burner stoichiometries  below
              100 percent in some cases. This low
              burner stoichiometry level would cause
              reducing  atmospheres in parts of the
              furnace, creating the potential for corro-
              sion of water tubes, especially when fir-
              ing  high-sulfur coal.  Generally, boiler
              manufacturers do not  recommend
              burner operation with stoichiometry
              below 100 percent, primarily because of
              increased corrosion potential. Another
              potential  adverse  impact  is that  addi-
              tional  excess air  may be required to
              ensure complete combustion, resulting
              in a decrease in boiler efficiency. How-
              ever, experience with utility boilers indi-
              cates that these potential problems can
              be overcome with proper design  and
              implementation. Indeed,  30-day, con-
              tinuous monitoring tests of staged com-
              bustion with LEA, at varying reduced
              boiler loads, demonstrated a 30-percent
              NOx reduction with no adverse opera-
              tional impacts.
                Burners-out-of-service (BOOS), the
              other technique that  can be used for
              staged combustion, is primarily consi-
              dered  for retrofits. However, it  is not
              favored for several reasons:
                • Extensive engineering and testing
                 on an  individual boiler basis is
                 required to determine the optimal
                 BOOS pattern.
                • An effective BOOS pattern is some-
                 times not possible because pulver-
Table  1.    Representative Industrial Boilers and Typical Baseline /V0X Emission Levels
Fuel
Pulverized Coal
Stoker Coal
Residual Oil*
Distillate Oil
Boiler Type
Single Wall and Tangential
Spreader
Underfeed
Chain Grate
Firetube
Watertube
Firetube
Watertube
Without air preheater
With air preheater
Typical Size
(Heat Input Capacity)
MW
(10s Btu/hr)
59
44
9
22
4.4
44
4.4
29
(200)
(150)
( 30)
( 74)
( 15)
(150)
( 15)
(100)
Average
/VOx Baseline
Emission Level
ng NOz/J
(lb/WsBtu)
285
265
150
140
115
160
70
55
90
(0.663)
(0.616)
(0.349)
(0.0326)
(0.267)
(0.372)
(0.163)
(0.128)
(0.208)
 Natural Gas
Firetube
Watertube
  Without air preheater
  With air preheater
 4.4  ( 15)

29    (100)
 40  (0.093)

 45  (0.105)
110  (0.255)
 'Includes No. 5 and No. 6 fuel oils.

-------
  400—
\

I
i
  300—
  200 —
  ;oo-
                                 'Cyclone
            Pulverized Coal Boilers
                                     'Spreader
                                     Stokers
                                                              Underfeed and
                                                              Traveling
                                                              Grate Stokers
                   I     I
                  2    3
 I     I     I     I     I     I
456789

   Excess Oxygen, percent
 I     I     I     I
10   11   12   13
Figure 2.    Effect of excess oxygen on N0t emissions from coal-fired boilers.
     \zeis  may serve burners on  two
     levels. The most effective  BOOS
     pattern often  involves the top level
     of burners on air only.
     Burners/pulverizers that operate
     during BOOS often cannot handle
     increased coal flow, necessitating a
     significant reduction  in the boiler
     steam rating (e.g., 20-percent).
               • Potential  increased slagging and
                corrosion.
               Several low NOX burner (LNB) designs
             are under development by commercial
             firms, with 40 - 60 percent NOX control
             projected.  In addition, an  advanced
             design under study by EPA is the distrib-
             uted fuel/air mixing concept. Field test-
             ing and application is scheduled for late
1982, with a target NOX level of 86 ng/J
(0.2 lb/106 Btu).
  In some applications, LNBs may have
several advantages over other combus-
tion  modifications such  as  OFA and
BOOS. For example, one utility boiler
manufacturer claims that LNBs will
maintain the furnace in an  oxidizing
environment, minimizing  slagging and
reducing the potential for furnace corro-
sion  when firing high-sulfur coal. Also,
more complete carbon utilization maybe
achieved due to better coal/air mixing
in  the furnace. Finally,  lower oxygen
levels may be obtained with all the com-
bustion air admitted through the burners.
  Since the burners generally alter the
flame configuration, care must be taken
when applying  the burners to existing
boilers. For instance, some LNBs have
longer flames.  Such  burners can be
installed only in boilers large enough tc
avoid cold-wall  impingement. Once
developed, however, low NOX, coal-firec
burners  for  industrial  boilers coulc
become the best control system because
of  the expected lower cost, higher NO,
reduction capability, and  other opera-
tional advantages.
  If additional control, over and above
boiler/burner modifications,  is needec
(e.g.,  to  meet  stringent  local regula-
tions), ammonia injection is offeree
commercially. The technique has yettc
be demonstrated on coal-fired boiler;
and is several times more costly thar
conventional combustion modifica
tions. In addition, as a developing tech
nology,  several potential implementa
tion and operational problems need t(
be resolved:
  • Optimal effectiveness for  noncata-
    lytic reduction of NO by NH3 occurs
    over a  very  narrow temperature
    range;  hence, the precise locatior
    of NH3 injection ports.
  • Since the temperature profile in t
    boiler changes with load, NOX con-
    trol with  NHs may restrict load.
  • Emissions of NH3 and by-products
  • Possible  boiler equipment foulinc
    by ammonium sulfates.
However, the major strengths of ammo
nia injection are its potential for moder
ate NOX removal (40 -  60 percent), anc
its applicability as an additional contro
that can be combined with conventiona
combustion  techniques  for  mcreasec
NOX reductions.

Stoker-Coal-Fired  Boilers
  NOX emissions from stokers are gen
erally lower than those from pulverize!
coal.  These  lower emissions can bi

-------
attributed  to  the  lower combustion
intensity and to the partial staged com-
bustion that naturally occurs during
combustion on fuel beds.
  As shown in Figure 2, NO* emissions
from spreader stokers tend to be higher
than those from other stokers. The coal
in a spreader stoker boiler burns partly
in a suspended state  and partly on a
moving or vibrating grate. The combus-
tion of coal  in the suspended state
apparently causes generally higher NOX
emissions  than for  other stokers that
feed and combust  coal  directly on a
moving grate. In addition,  the  higher
heat release  rates of spreader stokers
probably contribute  to  high NOX emis-
sions.
  Four methods have been used to mod-
ify stoker coal combustion to reduce NOX
emissions: reduced undergrate air or
LEA,  OFA, reduced heat input, and
reduced  air preheat (RAP). Of these
methods, only LEA firing has been dem-
onstrated to be widely effective.
  EPA field tests of  17  stokers indicate
that the excess oxygen levels at base-
line operating conditions averaged
about 9 percent. During LEA tests, the
average  excess oxygen  level  was
reduced to 6.4 percent  by reducing the
undergrate airflow  while maintaining
the OFA flow close to normal operation.
Such reduction lowered NOX emission
levels approximately  10 percent for
each 1 percent reduction in excess oxy-
gen. Additional data from an EPA-DOE-
ABMA field test program, involving 11
relatively new design stokers operating
near the lower excess air level, support
this conclusion.
  The minimum achievable excess air is
limited by several factors. Except for the
water-cooled vibrating  grate, the grate
is cooled only by airflow. If this air is cut
back too much, the grate can overheat.
There is also the danger of creating local
reducing zones and  of forming harmful
corrosion products as the air is cut back.
Another problem during field tests was
the formation of clinkers and increased
CO emissions as the excess oxygen was
reduced. However, test results indicate
that, if excess oxygen levels are main-
tained above 5 percent, CO emissions
will stay  below 150 ppm.
  Fuel combustion with  lowest possible
levels of excess air ensures maximum
boiler  efficiency unless the  air is
decreased to the point where unburned
carbon losses are  greatly  increased.
Limited available data indicate that, if
airflow is maintained for an excess oxy-
gen level above about 6 percent, no seri-
ous operational or emission problems
should result. NOX emission reductions
of about 5-25 percent and increases in
boiler efficiency of 1 percent can be ex-
pected with LEA. if fuel burnout does
not change during the process.

Residual-Oil-Fired Boilers
  As with coal-fired boilers, combustion
modification NOX controls have been
applied only to a limited number of oil-
fired boilers.
                              This  experience  indicates  that  low
                            excess  air firing is the only demon-
                            strated  universally applicable control
                            technique for all oil-fired boilers. Figure
                            3 shows excess air test results.
                              Baseline NOX emissions from residual-
                            oil-fired firetube boilers are  relatively
                            low, averaging 115 ng/J, as noted in
                            Table  1. Low excess air operation
                            should  lower emissions by about 20
                            percent and  also increase boiler effi-
                            ciency.  The same possibility of increased
  300—
  200—
.o
£
i
  700-
                                                     PS 300 (High Nitrogen
                                                     No. 5 Oil)
                        \
                       3
                     I     I     I
                     56789

                  Excess Oxygen, percent
 I
JO
 I
//
T
 12
Figure 3.
Effect of excess oxygen /VOX emissions from distil/ate and residual
oil-fired boilers.

-------
CO and  hydrocarbon emissions  dis-
cussed for coal firing under low excess
air applies here also. Low NOx burners
and  staged  combustion are the  pre-
ferred alternatives for additional con-
trol. However, neither has been demon-
strated for firetube boilers. Developing
low NOx  burners may become the first
control choice  after LEA  because of
their potential for  high NOx reduction
with the lowest  boiler operational
impact.
  The generally larger watertube boil-
ers with higher NOX emissions(160 ng/J
average) can also benefit from the same
controls: low  excess air,  low  NOx
burners, and staged combustion. Staged
combustion has been demonstrated for
large multiburner watertube boilers.
However, if developing low NOX burners
are successful and achieve 40 - 60 per-
cent reduction,  down to 86 ng/J (0.2
lb/106 Btu),  they  should  prove more
cost effective. The only other alternative
for stringent control is ammonia injec-
tion. Although demonstrated and in
limited commercial operation for oil and
gas firing in Japan, this system  is a
severalfold more costly alternative for
NOx reduction than the other  two. In
addition, operational problems and
potential emissions of NH3 and by-
products  are of environmental concern.


Distillate-Oil- and Gas-Fired
Boilers
  NOx emissions from distillate oil and
natural gas  combustion are primarily
from thermal NOX formation. The rela-
tively low uncontrolled  baseline  NOx
emissions of these boilers (see Table 1)
should permit very low controlled NOx
levels. These control levels can be met
in most cases with commercially avail-
able combustion  modification tech-
niques. The  preferred control systems
are low excess air, reduced air preheat,
flue  gas recirculation,  and low  NO*
burners  (under development),  in  that
order, lowering NOX down to about 65
ng/J (0.15  lb/106 Btu). Distillate oil-
and natural-gas-fired boilers not equip-
ped with air preheaters (all firetubes,
some watertubes) generally  exhibit
significantly lower average NOx emis-
sions than those with air preheaters, re-
gardless  of boiler heat input capacity, as
shown in Figure 4. Figure 4 shows that
bypassing an existing  preheater  sub-
stantially reduces NOx (shown  for nat-
ural gas, though similar behavior is ex-
pected for distillate oil). Those boilers
without  air  preheat should be able to
    175
  (0.41)

 ^  /50
    125
  >
    100

     50
   0.72,
 Uj
 ^.Distillate Oil
 D Natural Gas
  P Boiler With Preater
NP Preheater Bypassed

                                                a
                        1    1   1  1 t-1 1 1
                                         1
                                1    1
        1(3.4)
              10(34)

Boiler Capacity, MW(106 Btu/hr)
                                                         100(3<
Figure 4.    Effect of combustion air preheat and boiler capacity of /VOX emissions
             from distillate-oil- and gas-fired industrial boilers.
reach 43 ng/J (0.1 lb/106 Btu) with just
flue gas recirculation; air-preheater-
equipped boilers may require combined
reduced air preheat and flue gas recir-
culation. Figure 5 shows the high effec-
tiveness (40  - 75 percent NOX reduc-
tion) of flue gas recirculation for distil-
late oil and natural gas firing.

Cost of Controls
  The primary contributions of combus-
tion modification controls to steam cost
changes are  the equipment modifica-
tion costs and changes in thermal effi-
ciency  and  fan power demand.  In
general, combustion modification con-
trols should be cost-effective for indus-
trial boilers, raising steam costs only 1 -
2 percent in most cases.  However, the
initial investment, especially for smaller
boilers, may  be  a large fraction of the
cost of the boiler itself, up to 25 percent
when controls are installed  on a new
boiler. Retrofit control costs,  highly site
specific,  could be two to three times
higher.
  Table 2 summarizes costs and cost
effectiveness of controls to attain var-
ious control levels for the various boiler
design  and fuel categories. Costs  in
Table 2 reflect annualizing capital costs
and adding these to annual operating
costs.
                              LEA, in many cases, will actually lowei
                            steam costs due to the increase in ther-
                            mal efficiency. In general, LEAisrecom
                            mended with other control technique:
                            to lessen their  cost impact and to give
                            higher NOX reductions. Staged combus
                            tion causes an estimated small increase
                            in steam cost; but, with careful desigr
                            and operation,  this estimated cost  car
                            probably be reduced. Flue gas recircula
                            tion, although costly, is the most effec
                            tive technique for the clean  fuels
                            distillate oil  and  natural gas. Again
                            optimal design and operation will proba
                            bly lower  the cost.  Low NOX burners
                            promise to be the most cost-effective
                            However, they  are still under develop
                            ment.
                              Post-combustion control, because o
                            higher capital equipment, raw material
                            and energy  requirements, is signifi
                            cantly more  costly. Ammonia injectioi
                            is several times more costly than con
                            ventional  combustion  modifications
                            Flue gas treatment costs are about ai
                            order of magnitude higher than com
                            bustion modifications.

                            Incremental Emissions Due to
                            Controls
                              Combustion  modifications, used t
                            control  NOx  emissions from  Industrie

-------
             70
 2
 QQ

-------
OFA ports. The reinjection air is sup-
plied by the OFA fan.
  During  this program, two furnace
operating  conditions were tested:  (1)
under normal operating conditions, and
(2)  with increased  OFA  (at constant
overall  excess air)  to  determine the
effect on NOX as well as paniculate and
trace element emissions.
  Increased OFA caused apparent oper-
ating efficiency to increase from 77.8 to
80.8 percent. The largest  contributing
factor to this increase was a decrease in
the combustible content of the flyash.
  Table 3 summarizes  flue gas emis-
sions at the ESP outlet for all compo-
nents analyzed. The table shows that
NOx emissions were reduced 18 percent
under  OFA  firing;  however, emission
levels of CO, SO2, and SO3 increased. The
increases  in  sulfur  species emissions
were probably due to measurement dif-
ficulties or nonhomogeneities  in coal
composition, rather than changes in fir-
ing mode. Particulate load  increased at
the ESP outlet, as did flue gas organic
species OC7) emissions under low NOX
firing. Infrared analyses of flue gas sam-
ple extracts  indicated the  presence of
carboxylic acids and some aromatics in
the baseline extract, and aromatics and
possibly an amide in the low N0«extract.
Emission  levels of  the trace  element
species remained unchanged, within
analytical accuracy,  with firing mode.
  The bottom ash,  mechanical hopper
ash, and ESP hopper ash were also ana-
lyzed for trace elements, ionic species,
and Level  1 organic  content.
  Results indicated that concentrations
of the  trace elements and most ionic
species were unchanged with firing
mode.  Interestingly, however,  bottom
ash nitrate content  (oxidized nitrogen)
apparently decreased while ammonium
content (reduced nitrogen) apparently
increased  for the low NOx test.  For the
organic species, levels  were higher in
the low NOx bottom ash  than  in the
baseline bottom ash, although the base-
line mechanical hopper ash had higher
organic content than the low NOX me-
chanical hopper ash.
  Infrared spectrometry  analyses of
ash sample  extracts showed that car-
boxylic  acids, esters, and ethers were
present in both  bottom ash samples,
and  that  only aliphatic hydrocarbons
were  present in  ESP hopper ash
samples.

Site B
  The unit tested at  Site B was a Riley
single-pass boiler with a Riley spreader
Table 3.    Flue Gas Composition at the ESP Outlet (ug/dscm): Site A
                                   Baseline
                               Low /VOX
/VO,
SOZ
SO3
CO

Particulate
6.3 x 10s
8.2 x 10s
5.8 x 103
2.5 x 10s

4.0 x 10"
5.2 x 10s
1.1 x 706
3.5x10*
4.8 x 10s

6.3 x 10*
Antimony
Arsenic
Barium
Beryllium
Bismuth
Boron
Cadmium
Chromium
Cobalt
Copper
Iron
Lead
Manganese
Mercury
Molybdenum
Nickel
Selenium
Tellurium
Thallium
Tin
Titanium
Vanadium
Zinc
Organics OCj)
Methane
<19
3.2
420
3.8
<24
910
6.1
17
9.2
29
1. 1 x 103
<57
32
<0.037
10
34
32
1.1
<34
8.4
1.2
8.7
63
1.0x W3
4.9 x 1O*
<3.2
<7.0
70
<0.097
<6.7
350
<17
150
<48
<98
1.9x 103
<180
<77
<0.016
<8.0
<140
45
<5.5
<37
<220
580
<17
35
1.8x 103

stoker. The unit had a continuous rating
of 25 kg/s steam (200,000 Ib/hr) with a
2 hour maximum capacity of 28 kg/s.
The steam is produced at a pressure of
1.38 MPa (185 psig) and temperature of
260°C (500°F). This boiler also had a
flyash reinjection system. All the ash
from the boiler hopper is reinjected and
part of the ash from the mechanical col-
lector is also reinjected. An ESP is used
for final flyash control.  The boiler was
also equipped with an economizer.
  The NO, EA  field tests at Site B con-
sisted of a baseline (normal operation)
test and a test with low excess air. Low
NOx operation  reduced  NOx by 37 per-
cent at the economizer outlet.
  Table  4 summarizes the flue  gas
emissions at the economizer outlet for
all species  analyzed in  the  test.  As
     shown, CO levels remained relatively
     unchanged while particulate emissions
     increased slightly under low NOx firing.
     Results of sulfur species analyses were
     inconclusive; errors are suspected in the
     values obtained by the Shell-Emeryville
     method used. Flue gas organic emis-
     sions  increased with low  NOX firing.
     Infrared spectrometry  indicated  the
     presence of aliphatic hydrocarbons,
     ethers, esters, and carboxylic  acids in
     the extracts from both tests. Trace ele-
     ment emissions were  not  affected by
     firing mode, within analytical accuracy.
     The bottom ash,  mechanical  hopper-
     ash, and ESP hopper ash were analyzed
     for trace elements, ionic species, and
     Level 1 organic content.
       Concentrations of these did  not vary
     significantly with firing mode.  Infrared

-------
Table 4.    flue Gas Composition at ESP Outlet f/jg/dscm): Site B

                                  Baseline
                              Low /VOx
A/0,
SO2
S03
CO
CO2

Paniculate
4.7x10*
5.2 x 10s
2.4 xW3
5.4 x 10*
2.1 x 10*

1.9x JO4
3.4 x 10*
3.3x10*
2.2 xlO3
3.5 x 10*
2.4x10e

2.9 x 10*
Antimony
Arsenic
Barium
Beryllium
Bismuth
Boron
Cadmium
Chromium
Cobalt
Copper
Iron
Lead
Manganese
Mercury
Molybdenum
Nickel
Selenium
Tellurium
Thallium
Tin
Titanium
Vanadium
Zinc
Organics OCj)
Methane
3.5
35
170
1.2
2.6
<6.8 x 103
<1.2
61
23
64
280
<1.2
12
2.4
320
2.6 x 103
13
2.2
2.0
9.5
<55
130
9.7
920
1.27 x 10*
3.5
39
2.7
0.25
3.2
7.8 x 103
5.2
20
5.8
0.050
<71.2
<0.50
11
1.3
210
2.4 x JO3
13
<3.0
<2.2
<8.8
<49
18
<0.058
1.37 xW3
5.72x10*
spectrometry of  ash sample  extracts
indicated the presence of aliphatic and
aromatic hydrocarbons, esters, and car-
boxylic acids in bottom ash samples; ali-
phatic hydrocarbons in mechanical
hopper ash  samples; and  possibly an
ester or carboxylic acid in  ESP hopper
ash samples.


Environmental Impact
Evaluation
  The data obtained in the field test pro-
gram discussed above were evaluated by
a Source Analysis Model (SAM), specifi-
cally SAM/IA,8 to give quantified mea-
sures of the potential hazard posed by
emissions from stoker-fired' industrial
boilers and to evaluate how low NOX fir-
ing affects the potential hazard. SAM/IA
was developed by IERL-RTP for use in
     EA projects to estimate the potential
     hazards of source discharge streams.
       The SAM/IA model defines two in-
     dices of potential  hazard: Discharge
     Severity (DS) and Weighted Discharge
     Severity (WDS). DS is the  ratio of the
     pollutant discharge  concentration to  its
     Discharge Multimedia Environmental
     Goal  (DMEG).  (DMEGs, within  IERL-
     RTP's EA program for a large number of
     species, represent maximum pollutant
     concentrations desirable in discharge
     streams to preclude adverse effects on
     human health or ecological  systems.) A
     DS exceeding unity flags the existence
     of a potential hazard.  A stream's Total
     Discharge Severity  (TDS) is the sum  of
     the DSs calculated for the discharge
     stream.
       WDS is the product of DS times the
     discharge stream mass flowrate. Total
Weighted Discharge Severity (TWOS) is
the sum of WDSs calculated for the dis-
charge stream. Thus WDS  and TWOS
indicate the magnitude of  a potential
hazard and can be used to rank the rela-
tive hazard posed by different discharge
streams.
  Table 5 gives results of the SAM/IA
evaluation of the flue gas composition
data from Site A, for uncontrolled (base-
line) and controlled operation (low NO,).
The  table shows MEG category DS for
each firing condition for components
with DS greater than 1 in either test. DS
values shown were calculated from
air/health-based DMEGs. Table 5
shows that NOX and SOz are potentially
the  most hazardous  flue species.  The
sum of theDS values for these two spe-
cies comprises over 70 percent of the
stream TDS. The DS for S02 fluctuates
with day-to-day fuel sulfur content,
masking the effects of the reduction of
the NOX DS with the application of  NOX
control, so that stream TDS remains rel-
atively constant with NOX control appli-
cation. Other species of potential
concern  include CO, SOs (vapor),  and
several trace  elements. In this test the
S03 increased sixfold with NOX control.
However, this increase may be within
the accuracy of the analytical technique
used to measure SOs. Carboxylic acids
are flagged here also.
  Table 6 shows stream TWOS for each
stream under each firing mode tested.
The table shows that thef lue gas stream
dominates the potential hazard of the
source, with a WDS over two orders of
magnitude larger than any other stream.
In addition, the TWDSs for the  ash
streams remain relatively constant with
firing mode. Thus, changes  in flue gas
TWOS will elicit corresponding changes
in total source TWOS.  It is concluded
that  the NOX control tested does not, of
itself, increase total  source potential
hazard.
  Table 7 gives results of the SAM/IA
evaluation of the flue gas from the Site B
unit, for  the  two modes of operation
(baseline and low NOX). MEG category
DS values for compounds with  DS
greater than 1 in either test are given.
Again, these  values were  calculated
with air/health DMEGs.
  Conclusions from Table 7 are analo-
gous to those discussed for Table 5. NOX
and  SOa remain the potentially most
hazardous species, accounting for over
half  the  flue  gas  stream TDS. Here,
though, the DS for SO2 (which varies
exclusively with fuel sulfur content)
drops in  the  low NOX mode, so that

-------
  Table 5.    Flue Gas Discharge Severity: Site A
                                               Discharge Severity
Component
/vo,
SOZ
CO
SOa {vapor)
Beryllium
Arsenic
Iron
Carboxylic acids
Titanium
Cadmium
Lead
/wco
Category
47
53
42
53
32
49
72
8
41
82
46
Baseline
70
63
6.3
5.8
1.9
1.6
1.1
1.0
0.34
0.61
0.38
Low /VOX
58
85
12
35
0.049
3.5
1.9
1.8
3.7
1.7
1.2
  Total Stream
                               154
                     206
Table 6.     Total Weighted Discharge
          Severity: Site B


                    Total Weighted
                  Discharge Severity
                       (kg/s)
    Stream
     Baseline  Low N0t
Flue gas           6,600   7,900
Bottom ash            15      16
Mechanical
collector hopper ash     1.7     1.8
ESP hopper ash        5.3     2.6
 Table  7.
Flue Gas Discharge Sever-
ity: Site B


           Discharge
           Severity	
             MEG
Component  Category Baseline Low /VOx
/vo,
SOz
COi
Arsenic
Boron
SOa (vapor)
CO
47
53
42
49
68
53
42
68
39
35
18
2.2
2.4
1.4
38
23
27
20
2.5
1.6
0.88
Total stream
         168    114
stream TDS decreases with NOx control
application.
  Other  flue  gas species  with DS
greater than  1  include COz, CO, S03
(vapor), and the trace elements arsenic
and  boron.  The  DS for SOa shows
essentially no change with firing mode.
In addition, no organic category had a DS
value greater than 1.
  Table 8 shows stream TWOS values
for each stream at Site  B, under both
firing modes tested. The  Site A conclu-
sions hold here as well. Specifically, the
flue gas stream dominates the potential
source hazard, and NOX control reduces
total  source  potential hazard in the
absence of 562 considerations which
depend on the fuel sulfur content and
not on the NOx control.
  Bioassays of the bottom ash and ESP
hopper ash from the Site B low NOx test
indicated negative mutagenicity  and
nondetectable  toxicity.  Only the  ESP
hopper ash elicited a positive response,
giving a low toxicity result in the RAM
cytotoxicity assay.

Recommendations
  NOx controls  have been  applied to
industrial boilers only  to  a limited
extent. An exception is low excess air,
which is often used to increase boiler
efficiency. Thus, there is a general need
for data on the effectiveness, costs, and
operational impacts of NOx combustion
modification  control applied to indus-
trial steam raising equipment. The gen-
eral trends highlighted in this report are
meant to be only guidelines; there will
certainly be exceptions, and much
research and development work remains
to be done before NOx control technol-
ogy is well characterized for the wide
diversity of industrial boiler design and
equipment types.
  However,  it can be generally con-
cluded that currently available combus-
                                                                   Table 8.    Total Weighted Discharge
                                                                               Severity: Site B
                                                                                                   Total Weighted
                                                                                                 Discharge Severity
                                                                                                      fkg/s)

                                                                                    Stream     Baseline   Low NO,
Flue gas
Bottom ash
Mechanical
collector hopper
ash
ESP hopper ash
7.160
a


50
1.9
3,640
7.7


5.5
3.7
^Bottom ash sample not taken.

tion modification technology is capable
of moderate reductions (10 - 25 percent)
for coal- and residual-oil-fired  boilers,
while major reductions (40 - 70 percent)
are possible for distillate-oil- and gas-
fired units with minimal adverse  opera-
tional or environmental  impacts. Ad-
vanced techniques under development,
such as low NOx burners and ammonia
injection, are potentially  capable of
more efficient operation  and/or  addi-
tional reductions, and the development
of these techniques should  continue.
   EPA is currently sponsoring  several
field test programs demonstrating com-
bustion modification NOX controls  for
industrial  boilers. Results from  these
studies should help fill some of the data
gaps identified in this study. In addition,
several  other field tests of these and
other combustion controls  are under-
way,  including  30-day continuous
monitoring programs. The results of
these and other test programs should
be monitored and incorporated in future
updates of the assessment of combus-
tion modification NOX controls.

References
1.  Lim.  K. J., et al., "Environmental
    Assessment  of Utility Boiler Com-
    bustion Modification NOX Controls:
    Volume 1. Technical  Results, and
    Volume 2.-Appendices," EPA-
    600/7-80-075a,b (NTIS  PB80-
    220957 and 80-212939), Industrial
    Environmental Research Laboratory,
    Research Triangle Park, NC,  April
    1980.

2.  Higginbotham,  E.  B., and P.  M.
    Goldberg,  "Combustion Modifica-
    tion NOx Controls for Utility Boilers:
    Volume I. Tangential  Coal-Fired
                                 10

-------
    Unit  Field Test," EPA-600/7-81 -
    124a, Industrial Environmental
    Research Laboratory, Research Tri-
    angle Park, NC, July 1981.

3.  Sawyer,  J. W., and E.  B. Higginbo-
    tham, "Combustion Modification
    NOx  Controls for Utility Boilers:
    Volume  II. Pulverized-Coal Wall-
    Fired Unit Field Test," EPA-600/7-
    81-124b, Industrial Environmental
    Research Laboratory, Research Tri-
    angle Park, NC, July 1981.

4.  Sawyer,  J. W., and E.  B. Higginbo-
    tham, "Combustion Modification
    NOx  Controls for Utility Boilers:
    Volume  III. Residual Oil Wall-Fired
    Unit  Field Test,"  EPA-600/7-81 -
    1 24c, Industrial Environmental
    Research Laboratory, Research Tri-
    angle Park, NC, July 1981.

5.  Castaldini, C., et al., "Combustion
    Modification Controls for Residen-
    tial and  Commercial Heating  Sys-
    tems: Volume  I. Environmental
    Assessment," EPA-600/7-81 -123a,
    Industrial Environmental Research
    Laboratory, Research Triangle Park,
    NC, July 1981.

6.  Castaldini, C., et al., "Combustion
    Modification Controls for Residen-
    tial and  Commercial Heating  Sys-
    tems: Volume II. Oil-Fired Residen-
    tial Furnace Field Test," EPA-600/
    7-81-123b,  Industrial Environ-
    mental  Research  Laboratory, Re-
    search  Triangle Park,  NC,  July
    1981.

7.  Lentzen,  D. E., et  al., "IERL-RTP
    Procedures Manual: Level 1 Envi-
    ronmental Assessment (Second Ed-
    ition),"  EPA-600/7-78-201 (NTIS
    PB293795), Industrial Environ-
    mental  Research  Laboratory, Re-
    search Triangle Park, NC, October
    1978.

8.  Schalit,  L.  M.  and K.  J. Wolfe,
    "SAM/IA: A Rapid Screening Meth-
    od for Environmental Assessment
    of Fossil  Energy Process Effluents,"
    EPA-600/7-78-015 (NTIS PB
    277088), Industrial Environmental
    Research Laboratory Research Tri-
    angle Park, NC, February 1978.
K. J. Urn, C. Castaldini, R. J. Milligan, H. I. Lips, R. S. Merrill, P. M. Goldberg, E.
  B. Higginbotham, and L. R. Water/and are with Acurex Corporation, Energy
  and Environmental Division, Mountain View, CA 94042.
Joshua S. Bowen is the EPA Project Officer (see below).
The complete report consists of three volumes, entitled "Industrial Boiler Com-
  bustion Modification /VOX Controls,"
    "Volume I. Environmental Assessment," (Order No. PB 82-231 077; Cost:
    $28.50, subject to change)
    "Volume II. Stoker-Coal-Fired Boiler Field Test—Site A," (Order No. PB
    82-231 085; Cost $16.50, subject to change)
    "Volume III.  Stoker-Coal-Fired Boiler Field  Test—Site B," (Order No. PB
    82-231 093; Cost: $18.00, subject to change)
The above reports will be available only from:
       National Technical Information Service
       5285 Port Royal Road
       Springfield, VA 22161
        Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
       Industrial Environmental Research Laboratory
       U.S. Environmental Protection Agency
       Research Triangle Park, NC 27711
                                                                               11
                                                                               U. S. GOVERNMENT PRINTING OFFICE: I982/559-092/0484

-------
                                                    o •

                                                    o
                                                    5'

                                                    I.
                                                    O
                                                    01
                                                    O)
                                                    00
CO

 3D fr
   c
   fr
   2T
   o
                                               m > TJ m

                                               fill
                                               CO O O O
                                               W< A 3
                                               w   03

-------