-------
Volatile fractions
(Hydrocarbons. RN etc.)
Ash
virtually
nitrogen
free
Figure 3-3.
Possible fate of fuel nitrogen contained in coal
particles during combustion (Reference 3-26).
3-9
-------
Nitrogen retained 1n the char may also be oxidized to NO, or reduced to N. through
heterogeneous reactions occurring 1n the postctmbustlon zone. However, it is clear that the
conversion of char nitrogen to NO proceeds mucd more slowly than the conversion of devolatllized
nitrogen. In fact, based on a combination of experimental and empirical modeling studies, It 1 now
believed that 60 to 80 percent of the fuel NOX results from volatile nitrogen oxidation (References
3-27, i-33). Conversion of the char nitrogen to NO 1s in general lower, tv faccors of two to three,
than conversion of total coal nitrogen (Reference 3-30).
Regardless of the precise mechanism of fuel NO formation, several general trends are
evident, particularly for coal combustion. As expected, fuel nitrogen conversion to NO is highly
dependent on the fuel/air ratio for the range existing 1n typical combustion equipment, as shown in
Figure 3-4. Oxidation of the char nitrogen is relatively insensitive to fuel/air changes, but
volatile NO formation Is strongly affected by fuel/air ratio r.Manges.
In contrast to thermal NO. fuel NO production is relatively Insensitive to small changes in
K X
combustion zone temperature (Reference 3-30). Char nitrogen oxidation appears to be a very weak
function of temperature, and although the amount of nitrogen volatiles appears to Increase as
temperature increases, this 1s believed to be partially offset by a decrease ii. percentage
conversion. Furthermore, operating restrictions severely limit the magnitude of actual temperature
changes attainable in current systems.
As described above, fuel NO emissions are a strong function of fuel/air mixing, In general,
any change which Increases the mixing between the fuel and a'r during coal devolatilization will
dramatically increase volatile nitrogen conversion and increase fuel NO . In contrast, char NO
formation is only weakly dependent on initial nixing.
From the above modifications, it appears that, in principle, the best strategy for fuel NO
abatement combines low excess air (LtA) firing, optimum burner design, and two stage ombustion.
Assuming suitable stage separation, low excess air may have little effect on fuel NU , but it
Increases syster, efficiency. Before using LEA firing, the need to gt t good caibon burnout and low
CO emissions rrust be considered.
Optimum burner design for coal ensures locally fuel-Hch conditions Juring devolati 1 ization,
which promotes reduction of devolatllized nitrogen to N,. Two-stage combustion produces overall
fuel-rich conditions during the first 1 to 2 seconds and promotes the reduction of NO to N. through
reburning reactions. Higfi secondary <".ir preheat may also be desirable, because it promotas more
3-10
-------
I/)
QJ
X
O
c
a
c
o
+J
c
o
>
V.
01
c
o
(J
ijali temp 1500°K
Flame temp 1600"K
A Lignite 75-90 Min
te 38-45 ,>m
Bituminous 38-45
234
Fuel equivalence ratio
(Inverse of stoichiometric ratio)
Figure 3-4. Conversion of nitrogen In coal to NO (Reference 3-25).
J\
-------
complete nitroyen uevo'aLlMiotlor. ir. the fuel rich initial cofbustlon stage. This leaves less char
nitrogen to be subsequently oxidized in the fuel-le.in second stage. Unfortunately, it also tend: to
favor thermal NO formation, and at present there is .10 general agreemtnt on which effect dominates.
3.1.1.3 Summary of Process Modification Concepts
In summary of the above discussion, both thermal and fuel NOX are Mnetically or
aerodynamically limited in that their emission rates are far below the levels which would prevail at
equilibrium. Thus, the rate of formation of both thermal and fuel NO is dominated by combustion
conditions and is amenable to suppression througi combustion process modifications. Although the
mechanisms are different, both thermal and fuel NO are promoted by rapid mixing of oxyren with the
fuel. Additionally, thermal NO is greatly increased by lono residence time at high temperature.
The modified combustion conditions and control concepts which have been tried or suggested to combat
the formation mechanisms are as follows:
(1) Decrease primary flame zone 0* level oy
-- Decreased overall 0,, levsl
-- Controlled mixing of fuel dnd air
-- Use of fuel-rich primary fleme zone
(2) Decrease time of exposure at nigh temperature by
-- Decreased peak temperature:
-- Decreased idiabatic flame temperature through dilution
-- Decreased combustion intensity
Increased flame cooling
-- Controlled mixing of fuel and air or use of fuel-rich primary flame icne
-- Decreased primary flame zone residence time
Table 3-2 relates these cont.ol concepts to applicable combustion proce1:: modifications and
equipment types. The process modifications are categorized according to their role 1t> the control
development sequence: operational adjustments, hardware modifications of existing eouipment or
through factory installed controls, and major redesigns of new equipment. The controls for
decreased Oj are also generally effective for peak temperature reduction but have not been repeated.
The following subsections briefly review the status of each of the applicable control techniques.
3-12
-------
TABLE 3-2. SUMMARY OF COMBUSTION PROCESS MODIFICATION CONCEPTS (Reference 3-1)
Combustion
Conditions
Decrease
primary
**ame zone
02 level
Decrease
peak
flame
temperature
Control
Concept
Decrease overall
02 level
Delayed nixing
of fupl and air
Increased fuel/
air mixing
Urinary fuel-
rich flame
zone
Decrease
adlabatlc flaw
temperature
Decrease com-
bust Ion
intensity
Increased flame
zone cooling/
reduce resi-
dence time
Applicable
Equipment
Boilers, furnaces
Boiler, furnaces
L-is turbines
Boilers,
furnaces, JC
Boilers, fur-
naces. 1C.
ga$ turbines
Boilers, furnaces
Boilers, furnaces
Effect on
Thermal HO
Rsduc>s O^-rich,
h1gh-NOs pockets
In the flame
Flame cooling and
dilution during
'leleyed mix re-
duces peak temp.
Reduces local hot
itctchlometrlc
regions IP over-
all fuel lean
combustion
Plane cooling In
low-02, I CM- temp.
primary zone re-
duces peak temp.
Direct sup.-es-
slon of thermal
NOx mechanism
Increased flame
zone cooling
yields lowe*
peak temp.
Increased flame
zone cooling
yields lower
peak tewp.
Effect on
Fuel N0x
Reduces exposure
of fuel nitrogen
Intermediaries
to 02
Volatile fuel N
reduces to N; In
tne absence <>f
oxygen
Increases
Vo'.atllo fuel K
reduces to «2 In
the absence of
oxygen
Ineffective
HI nor direct
effect; Indirect
effect on mixing
Ineffective
Primary Applicable Controls
Operational
Adjustments
low excess air
firing
Burner
adjustments
Burners out of
service; biased
burner firing
Reduced air
preheat
Lead reduction
Burner tilt
Hardware
Modification
Flue gas reclrcu-
latlon (FGR)
Low NOX burners
Over fire air
ports, stratified
charge
Mater Injection,
FGR
Major
Redesign
Optimum burner/
firebox design
New can design;
premix, prev*p.
Burner/firebox
design for two-
stage combustion
Enlarged firebox
Increased burner
spec Ing
Redes t 99 heat
transfer sur-
faces, firebox
aerodynamics
-------
3.1.2 Modification of Operating Conditions
The modification techniques described in this subsection include low excess air, off
stoichiometric or staged combustion, flue gas recirculation, reduced air preheat, reduced firing
rate, and steam or water injection.
3.1.2.1 Low Excess Air Combustion (LEA)
Reducing the total amount of excess air supplied for combustion is an effective demonstrated
method for reducing NO emissions from utility and industrial boilers, residential and commercial .
A
furnaces, warm air furnaces, and process furnaces. Low excess air (LEA) firing reduces the local
flame zone concentration of oxygen, thus reducing both thermal and fuel NO,, formation. LEA firing
A
is furthermore easy to implement and increases boiler efficiency. It is, therefore, used
extensively in both new and retrofit applications, either singly or in combination with other
control measures. The ultimate level of excess air is generally limited by the onset of smoke or
carbon monoxide emissions which occurs when excess air is reduced to levels far below the design
conditions. Fouling and slagging may also increase in heavy oil- or coal-fired applications at very
low levels of excess air, thus limiting the potential of this technique.
Low excess air firing is usually the first NO control technique applied to utility boilers.
It may be used with virtually all fuels and firing methods. It was initially implemented to
increase thermal efficiency and reduce stack gas opacity due to acid mist, and it is now often
considered more of an energy conservation measure than a NO control technique. A number of studies
have shown LEA firing to be effective in reducing NOX emissions without significantly increasing CO
or smoke levels (References 3-34 through 3-39). As shown in Section 4, Table 4-1, numerous tests of
low excess air firing on utility boilers have indicated NO emission reductions averaging between 16
X
and 21 percent for coal, oil, and natural gas firing compared to earlier baseline levels.
The minimum practical level of excess air which can be achieved i,n existing boilers and
process heaters, without encountering operational problems, depends upon factors in addition to the
type of fuel fired. These factors include low load operation, nonuniformity of air/fuel ratio, fuel
and air control lags during load swings, use of upward burner tilt to increase steam superheat (for
tangentially-fired boilers), and coal quality variation and ash slagging potential (for coal-fired
boilers). They tend to increase the minimum excess air level at which the boiler can operate
safely.
*
3-14
-------
Other factors such as secondary air register settings and steam temperature control
flexibility also affect the excess air levels. The boiler combustion control system must be
modified so that the proportioning of fuel and air is adequate under all operating conditions.
Uniform distribution of fuel and air to all burners is increasingly important as excess air is
lowered. Excess air levels are also affected if other NO control techniques are employed. Staging
and operating at reduced load increases the minimum excess air levels.
As discussed in Section 4.2, LEA firing is also a very effective method for controlling NO
in industrial boilers. For residential and commercial furnaces, however, while LEA is a potentially
feasible NOX control technique, the trend in NO control for these sources has been in improved
burner design in order to obtain low NO levels without extensive CO emissions.
LEA is not a very promising technique for 1C engines and gas turbines. When the air/fuel
ratio is reduced, CO and HC emissions increase sharply for 1C engines. In gas turbines, the overall
air/fuel ratio cannot be modified to control NO , since the ratio is determined by the turbine inlet
temperature.
In summary, changing the overall air/fuel ratio to control NO emissions is a simple,
feasible, and effective technique for utility and industrial boilers but is less applicable for
other stationary sources of combustion. For certain applications such as utility boilers, LEA
firing is presently considered a routine operating procedure and is incorporated in all new units.
Also, more and more industrial boilers are incorporating this techniques as well. Since it is often
efficient and easy to implement, LEA firing may see increasingly widespread use in other
applications. Most sources will require additional control methods, in conjunction with LEA, to
bring NOV emissions within statutory limits. In such cases, the extent to which excess air can be
A
lowered will depend upon the other control techniques employed. However, virtually all
developmental programs for advanced NO controls are placing maximum emphasis on operation at
minimum levels of excess air. LEA will thus be an integral part of nearly all combustion
modification NO controls, both current and emerging.
3.1.2.2 Off-Stoichiometric or Staged Combustion (OSC)
Off stoichiometric or staged combustion seeks to control NO by carrying out initial
X
combustion in a primary, fuel-rich combustion zone, then completing combustion at lower temperatures
in a second, fuel lean zone. In practice OSC is implemented through biased burner firing (BBF),
burners out of service (BOOS), or overfire air injection (OFA).
3-15
-------
Biased Burner Firing (BBF). Burners Out of Service (BOOS)
Biased burner firing consists of firing the lower rows of burners more fuel-rich than the
upper rows of burners. This may be accomplished by maintaining normal air distribution to the
burners while adjusting fuel flow so that a greater amount of fuel enters the furnace through the
lower rows of burners than through the upper rows of burners. Additional air required for complete
combustion enters through the upper rows of burners which are firing air rich.
In the burners out of service mode, individual burners, or rows of burners, admit air only.
Correspondingly the total fuel demand is supplied through the remaining fuel admitting or active
burners. Thus the active burners are firing more fuel-rich than normal, with the remaining air
required for combustion being admitted through the inactive burners.
These methods reduce NOV emissions by reducing the excess air available in the firing zone.
A
This reduces fuel and thermal NOX formation. These techniques are applicable to all fuels and are
particularly attractive as control methods for existing units since few, if any, equipment modifica-
tions are required (References 3-40 and 3-41). In some cases, however, derating of the unit may be
required if there is too limited extra firing capability with the active burners. This is most
likely to be a problem with pulverized coal units without spare pulverizer cetpacity.
Monitoring flue gas composition, especially 0« and CO concentrations, is very important when
employing these combustion modifications for NO control. Local reducing atmospheres may cause
X
Increased furnace slagging when burning coal because of the lower ash fusion temperature associated
with reducing atmospheres (References 3-42 and 3-43). In addition, it is important to closely
monitor flue gas, excess air, and CO to avoid reducing boiler efficiency through flue gas heat and
unburned combustible losses, and to prevent unsafe operating conditions caused by incomplete
combustion. For these reasons, accurate flue gas monitoring equipment and increased operator
monitoring of furnace conditions are required with these combustion modifications.
As shown in Section 4.1, Table 4-2, emission tests of burners out of service firing on
utility boilers have indicated average NOV reductions of 31 to 37 percent for coal, oil, and natural
A
gas firing compared to earlier baseline levels. A typical burners out of service pattern is shown
in Figure 3-5(a).
3-16
-------
OVIRfllt All
NOZZlii
O Active burners
)SC Burners admitting air only
a. Typical burners out of
service arrangement
opposed fired unit
WIMSSOX
MCONOAIV All CAMPUS
UCOMOAtY AH
BAMNI MlVf UNIT
COAIMOZZUS
OIL OWN
b. Typical overfire air system for
tangential fired unit (Reference 4-21)
Burners
Apportioning
dampers
Air
Forced draft fan
Flue gas recirculating
fan
c. Typical flue gas recirculation system for NO control
A
Figure 3-5. Typical arrangements for (b) overfire air,, (a) burners out
of service, and (c) flue gas recirculation (Reproduced from
Reference 3-1,'p. 4-26).
3-17
-------
Overfire Air (OFA)
The overfire air technique for NO control involves firing the burners more fuel rich than
normal while admitting the remaining combustion air through overfire air ports or an idle top row of
burners.
Overfire air is very effective for NO reduction and may be used with all fuels. However,
there is an increased potential for furnace tube wasteage due to local reducing conditions when
firing coal or high sulfur oil. There is also a greater tendency for slag accumulation in the
furnace when firing coal (References 3-23, 3-41, 3-43, and 3-44). In addition, with reduced airflow
to the burners, there may be reduced mixing of the fuel and air. Thus, additional excess air may be
required to ensure complete combustion. This may result in a decrease in efficiency
(References 3-41 and 3-44).
Overfire air is more attractive in original designs than in retrofit applications because of
cost considerations. Additional duct work, furnace penetrations, and extra fan capacity may be
required. There may be physical obstructions outside of the boiler setting making installation more
costly. There may also be insufficient height between the top row of burners and the furnace exit
to permit the installation of overfire air ports or to allow sufficient residence time for the
completion of combustion (Reference 3-41).
Also, overfire air is more easily implemented without large efficiency or cost penalties on
large units than on small ones. As unit size is decreased, furnace volume decreases faster than the
available wall surface. Hence, furnace residence times available for fuel combustion tend to be
shorter in small units. Staged combustion techniques such as overfire air serve to dalay or prolong
the combustion process. Thus, on small units larger proportional increases in furnace size and cost
may be required to assure complete fuel combustion with the application of these techniques. Or
alternatively, increased excess air rates through the overfire air ports may be required thus
leading to decreased unit efficiency. For these reasons, staged combustion techniques are commonly
applied to utility size boilers; but their application to smaller units is more limited.
As shown in Section 4.1, Table 4-3, some emission tests of overfire air on utility boilers
have indicated average NO reductions of about 25 to 60 percent for coal, oil, and natural gas
firing compared to earlier baseline levels. A typical overfire air system is shown in
Figure 3-5(b).
3-18
-------
An off stoichiometric or staged combustion technique similar to overfire air is also
applicable to control of NOV from natural draft process heaters (Reference 3-45). The staged
A
combustion system that has been demonstrated in process heaters uses a number of air lances arranged
around each burner. The burner is operated under fuel rich conditions with about 65 percent of the
air required for combustion entering through the burner air registers. The remainder of the air
required for complete combustion is injected into the flame zone some four feet above the burner.
This technique, in conjunction with low excess air operation, reduced NO emissions by 21 percent
below a baseline of 66 ng/J for a unit firing refinery gas. Fuel consumption was reduced by nearly
5 percent. - ,
*
3.1.2.3 nue£a? Recirelation (FGR)
Flue gas recirculation for NO control consists of extracting a portion of the flue gas,
A
usually from the economizer outlet with utility boilers, and returning it to the furnace. The flue
gas may be admitted through the furnace hopper or through the burner windbox or both. Flue gas
recirculation lowers the bulk furnace gas temperature and reduces the oxygen concentration in the
combustion zone (References 3-41 and 3-44).
Flue gas recirculation through the furnace hopper and near the furnace exit has long been
used for steam temperature control. Flue gas recirculation through the windbox and, to a lesser
degree, through the furnace hopper is very effective for NOX control on gas- and oil-fired units
(References 3-40 and 3-44). However, it has been shown to be relatively ineffective on coal-fired
units (Reference 3-20).
Flue gas recirculation for NO control is more attractive for new designs than as a retrofit
A . »
application. Retrofit installation of flue gas recirculation can be quite costly. The fan, flues,
dampers, and controls as well as possibly having to increase existing fan capacity due to increased
draft loss, can represent a large investment. In addition, the flue gas recirculation system itself
may require a substantial maintenance program due to the high temperature environment and potential
erosion from entrained ash. Thus the cost-effectiveness of this method of NO control has to be
examined carefully when comparing it to other control techniques.
As a new design feature, the furnace and convective surfaces can be sized for the increase in
mass flow and change in the furnace temperatures. In contrast, in retrofit-applications the
increased mass flow increases turbulence and mixing in the burner zone, and alters the boiler heat
absorption profile. Erosion and vibration problems may result (References 3-44 and 3-46). Flame
3-19
-------
detection can also be difficult with flue gas recirculation through the windbox. In addition,
controls must be employed to regulate the proportion of flue gas to air so that a sufficient concen-
tration of oxygen is available for combustion (Reference 3-47).
On utility boilers, flue gas recirculation has most often been used in combination with other
low NO combustion techniques. Test data for these types of applications are discussed in
Section 4.1. A typical flue gas recirculation system in a utility boiler application is shown in
Figure 3-5(c).
Flue gas recirculation has also been applied to a few process heaters. However, no perfor-
mance data are available for these units. The technique may not be applicable to all types of
heaters because it lowers flame temperature and can cause problems with flame stability
(Reference 3-48). Flue gas recirculation is therefore unlikely to be used in high temperature
applications such as ethylene pyrolysis heaters.
3.1.2.4 Reduced Air Preheat Operation (RAP)
Reducing the amount of combustion air preheat lowers the primary combustion zone peak
temperature, generally lowering thermal NO production as a result. Because of the energy penalty
associated with this technique, it has been used only sparingly in utility and industrial applica-
tions. It is applicable to utility steam generators and large industrial boilers which employ heat
exchangers to impart about 280°K (500°F) incremental heat to the combustion air.
With present boiler designs, reducing air preheat would cause significant reductions in
thermal efficiency and fuel penalties of up to 14 percent. This technique would be feasible for
thermal NOV control if means other than air preheat were developed to recover heat from 420°K to
X
700°K (300°F to 800°F) gases. For example, in new industrial boilers it is often practical to
replace the air preheater with an economizer thereby reducing or eliminating the energy penalty
associated with this technique. However, this technique appears relatively ineffective in
suppressing fuel nitrogen conversion (References 3-49, 3-50).
This technique is also applicable to turbocharged internal combustion engines and regenera-
tive gas turbines. The turbocharged 1C engines normally have an intercooler to increase inlet
manifold air density permitting higher mean flowrates, and consequently higher power output. The
reduced air temperature also reduces NO emissions. However, regenerative gas turbines recover some
of the thermal energy in the exhaust gas where temperatures range from 700°K to 870°K (800°F to
3-20
-------
1100°F) to preheat the combustion air. Any reduction in air preheat would cause severe fuel
penalties unless other means of recovering the heat in the exhaust could be implemented.
3.1.2.5 Reduced Firing Rate
Thermal NO formation generally increases as the volumetric heat release rate or combustion
intensity increases. Thus, NO can be controlled by reducing combustion intensity through load
reduction (or derating) in existing units, and by enlarging the firebox in new units. The reduced
heat release rate
(Reference 3-51).
heat release rate lowers the bulk gas temperature which in turn reduces thermal NOX formation
The heat release rate per unit volume is generally independent of unit rated power output.
However, the ratio of primary flame zone heat release to heat removal increases as the unit capacity
is increased. This causes NO emissions for large units to be generally greater than for small
units of similar design, firing characteristics, and fuel.
The increase in NO emissions with increased capacity is especially evident for gas-fired
boilers, since total NOV emissions are due to thermal NOV. However, for coal-fired and oil-fired
A A
units the effects of increased capacity are less noticeable, since the conversion of fuel nitrogen
to NO for these fuels represent a major component of total NOV formation. Still, a reduction in
A A
firing rate will affect firebox aerodynamics which may, consequently, affect fuel NO emissions.
But such effects on fuel NOX production are less significant.
Reduced firing rate often leads to several operating problems. Aside from the limiting of
capacity, low load operation usually requires higher levels of excess air to maintain steam
temperature and to control smoke and CO emissions. The steam temperature control range is also
reduced substantially. This will reduce the operating flexibility of the unit and its response to
changes in load. The combined results are reduced operating efficiency due to higher excess-air and
reduced load following capability due to a reduction in control range.
When the unit is designed for a reduced heat release rate, the problems associated with
derating are largely avoided. The use of an enlarged firebox produces NO reductions similar to
load reduction on existing units.
3-21
-------
3.1.2.6 Steam and Water Injection (MI)
Flame temperature, as discussed above, is one of the important parameters affecting the
production of thermal NO . There are a number of possible ways to decrease flame temperature via
thermal means. For instance, steam or water injection, in quantities sufficient to lower flame
temperature to the required extent, may offer a control solution. Water injection has been found to
be very effective in suppressing NOV emissions from internal combustion engines and gas turbines;
A
Figure 3-6 shows NOV emission reductions from a gas turbine as high as 80 percent (Reference 3-52).
A
Since steam and water injection reduce NO by acting as a thermal.ballast, it is important
that the ballast reach the primary flame zone. Combustion equipment manufacturers vary in their
methods of water or steam introduction. The ballast may be injected into the fuel, combustion air,
or directly into the combustion chamber.
Water injection may be preferred over steam in many cases, due not only to its availability
and lower cost, but also to its potentially greater thermal effect. In gas- or coal-fired boilers,
equipped for standby oil firing with steam atomization, the atomizer offers a simple means for
Injection. Other installations will require special rigging so that a developmental program may be
necessary to determine the degree of atomization and mixing with the flame required, the optimum
point of injection, and the quantities of water or steam necessary to achieve the desired effect.
The use of water injection may entail some undesirable operating conditions, such as
decreased thermal efficiency and increased equipment corrosion. This technique has the greatest
operating costs of all combustion modification schemes with a fuel and efficiency penalty typically
of about 10 percent for utility boilers and about 1 percent for gas turbines. It has therefore not
gained much acceptance as a NOX reduction technique for stationary combustion equipment except for
gas turbines (References 3-49 and 3-50). Gas turbines, in addition to having the lowest efficiency
losses with water injection, also showed no major operational problems or reduced equipment life
with this technique. Water injection for NO reduction does not appear to have a significant effect
on stack opacity and emissions of CO and HC.
3.1.2.7 Combinations of Techniques
Since 1969 it has been demonstrated that several of the previously discussed modification
techniques can be effectively utilized in combination since they reduce NO by different mechanisms.
X
Most often, off stoichiometric combustion is used with low excess air or load reduction on all
fuel-boiler type configurations. For oil- and gas-fired units flue gas recirculation is used in
conjunction with the above techniques. Flue gas recirculation and load reduction lower peak
3-22
-------
§
.Ol
,P
CM
O
CO
03
2
O
.o
o
UJ
I
CM
LO
t
(0 O)
10 S-
S- CD
o o:
tt XIM ^
o
O O
CD -Z.
r- O
i. S-
O) (O
00 -»->
£= 01
O =
ir^ O
V 3^
XQ
O OO
s: en
o
a s-
s_ en
O n3
cj cr>
i
CO
o
CM
(SISV8 ABO **O %Sl) -ON Wdd
3-23
-------
combustion temperatures, while off-stolchiometric operation reduces the amount of fuel burned at
peak temperature under oxidizing conditions. For the most part, combining control techniques has
been shown to be complementary but not additive for NOX reduction (Reference 3-49).
3.1.3 equipment Oe-.lgn Modification
3.1.3.1 Burner Com 1gurat1 on
Bur-ner or combustor modification for NOX control 1s applicable to all stationary combustion
equipment categories. The specific design and configuration of a burner has an Important bearing on
the amount of NOX formed. Certain design types have been found to give greater omissions than
others. For exarcple, the spud-type gas burner appears to give a higher emission rate than the
radial spud type., which, 1n turn, produces more NO than the ring type.
During the early 1970's specially designed "low-NOx" burnsrs were produced for thermal N0x
control. For the most part, they were designed for utility and Industrial boilers and employ
inflame LEA, OSC, or FSR principles. The aim 1s to strike a balance between "inimum NOX formation
and acceptable combustion of carbon and hydrogen in the fuel.
There are currently several commercial low-NOx gas and ..11 burner resigns for borers and
process heaters in operation and under development (References 3-53 through 3-57). Full scale test
results for boilers in Japan show reduction: a). Some of the more innovative
methods for oil burners include: flame splitting distributor tips which cause a flower petal flamr
arrangenqnt, and atomizers with fuel injection holes of different diameters which create fuel-rich
and fuel-lean combustion zones (References 3-53, 3-56, 3-58). Up to 55 percent reductions in NO
emissions are reported with the use of these nozzle tips. However, the change in flame shape may
cause problems due to Impingement on walls, and effectiveness may be reduced as flames Interact in
multiburner furnaces.
Other air-fuel modifications Include a low-NOx burner (offered by at least one company in the
U.S.) for oil- and gas-f1rad package boilers. This burner uses shaped fuel injection ports »nd
controlled air-fuel mixing to create a thin stuhby ring-shaped flame (References 3-53, 3-55). With
this modification, reductions 1n NO from 20 to 50 percent are claimed. The most extensive air-fuel
3-24
-------
modifications Involve tne self-red rculating and staged combustion chamber type of burners, used in
Industrial process furnaces. These burners are equipped with a prevaporizatlon or a precombustior
chamber 1n the wlndbox. In the chamber the fuel 1s vaporized and premixcd with part of the
combustion air, or is allowed to undergo partial combustion under oxygen def^ciert conditions before
being discharged Into the furnace. NO reductions of about 55 percent are typical for these
devices.
Low N0x '.turners are the most common typo of NO, emission control technique applied to process
heaters. The most common type of low NO butner tr oe best described as & two-stage combustion
burner which 1s fired fuel rich In the first stagt. The burner 1s designed to Inject tertiary air
after sufficient time has elapsed 1n the 'educing zone of the fla.ne. Controlled introduction of the
tertiary air provides reduction of NOX emissions In the reducing zone of the flame without
significant changes 1n flame pattern and burner operation.
A second klrd of staged combustion burner Is the staged fuel burner. This technique Involves
combustion of a fuel with high excess air. The remainder of the fuel 1s injected In the second
stage of' the reaction and combustion is completed at low excess air. The high excess air permits
the first stage of the combustion to occur at a low temperature. Depending en the amount of excess
air, the theoretical temperature may be as low as 1030°C (2000°F). As the combustion reaction goes
to completion in the first zone, the additional fuel 1s Injected. The second reaction begins with a
reduced partial pressure of oxygen which tends to limit the formation of NO . Other vendors offer
low NO burners based on flue gas recirculitlon and two stage combustion. The self-reelrcu^ating
gasification (SRG) burner has been designed to employ flue gas rec1rcui?tion, two-staged combustion,
gasification reactions, and low excess air. The key feature 1s the creation of an exceptionally
strong recirculation eddy in the burner throat tile. This draws combustion reaction products from
the furnace to gasify the fuel stream. The primary air flow 1s between ten and thirty percent
stlochiometric depending upon the design. The result 1s that the gases leaving the burner throat
are very rich in hL and CO.
Several utility boiler manufacturers nave also beer active in the development of new burners
designed to reduce NO emissions from coal-fired units. Most low NO burners designed for utilitv
boilers control NOX by reducing flame turbulence, delayirg fueVair mixing, and establishing fuel-
rich zones where combustion initially takes place. This represents a departure from the usual
burner design procedures which prcr.jte high tL-bulence, high intensity, rapid combustion flames.
The longer, less intense flames produced with low NO burners result 1n lower flame temperatures
3-25
-------
which reduce thermal NO generation. Moreover, the reduced availability of oxygen 1n the Initial
combustion zone Inhibits fuel NOV conversion. Thus, both thermal and fuel NO, are controlled by the
A A
low NOX burners.
The Babcock and Mil cox Company Is currently Installing the dual register pulverized coal-
fire^ burner 1n all Its not utility bo'lle»-s in order to meet current NSPS (Refsrences 3-46 and
3-59). TV. limited turbulence, control1:*! di^efus1on flame burner Is designed to minimize *uel and
air mixing at the burner to that retired to obtafn Ignition and sustain iv.able combustion of the
coal. A VentuH mixing device, located in the coal nozzle, provides a uniform coal/primary air
mixture at the burner. Sscondary air 1s Introduced through two concentric zone* surround!r-g the
coal nozzle, each of which 1s Independently controlled by Inner and outer air zone registers.
At least seven dual register burner-equipped utility boilers have been tested for NO
emissions (Reference 3-53), Tests on four bituminous coal-fired units showed NO emissions ranging
fro*! 190 to 260 ng/J (0,45 to 0.6 lb/106 Btu, 320 to 420 ppm). Tests on three subbltuminous coal-
f1r«d unit:, showed NOX emissions In the range of 130 to 150 ng/J (0.3 to 0.35 lb/106 Btu. 210 to
250 ppm). Comparisons with NO emissions from similar unite equipped with tne high turbulence older
burners show reductions in "iOx levels from 40 to 60 percent due to the new burner design.
In another recently reported test of the dual register burner on a bituminous coal unit,
EPA collected approximately 68 days of continuous monitoring data (Reference 3-60). Dally average
emissions were consistently below 200 rvyj (0.47 lb/10 Btu} and 30 day rolling average emissions
ranged froai 160 to 170 ng/J (0.37-0.39 lb/10G Btu).
BftW claims that NOX control through Its dual register burners 1s superior to staging as It
maintains *he furnace ' Injected above the burner zone.
Although the du«l register burners were developed for use in new boilers, they can also be
retrofitted to older units, However, the new boilers are also designed to provide airflow control
en a per pulverizer basis. This may not be possible in some of the older units, or the cost
involved In retrofitting a compartmented windbox and making the necessary changes 1n pulverizer
burner piping may be high. If careful control of fuel and air tc each burner is not feasible, the
3-26
-------
burners will not be as effective in reducing NOX emissions. Nevertheless, the new burners should
reduce NO levels below those obtained with the older high turbulence burners. They may still be
considered for retrofit arplication, perhaps 1n conjunction with other NOX control techniques, but
much development work remains.
Foster Wheeler Energy Corporation has developed a dual register rosl burner for Installation
In Its new boilers (References 3-43 and 3-61). The new burner reduces turbulence as compared to the
older designs and causes controlled, gradual mixing of fuel and air at the burner. This 1s achieved
using a dual throat with two registers which splits the secondary air Into two concentric streams
with Independently variable swirl. The mixing rate between the primary and secondary air streams
and the rate of entrapment of furnace gases can thus be varied.
Test results for the new Foster Wheeler burners are reported 1n References 3-43 and 3-61.
Reductions in NO emissions of about 40 percent were observed on a four-burner ste^m generator when
operated at full load with the new burners. Three utility steam generators, two 265 MW opposed
fired units and one 75 MU front wall unit, have been retrofitted with the new burners and tested for
NOX emissions. Controlled NOX emissions were In the 170 to 220 ng/J (0.4 to 0.5 lb/10S Btu, 280 to
350 ppm) range, about 40 to 50 percent lower than similar units with older design burners.
In addition to NOX control in new units, the Foster Hhaeler dual register burner 1s
technically we'll »uited for retrofit application. The airflow to the new burners is controlled
individually at each burner by means of a perforated hood. !!ence, precise air/fuel control at each
burner is possible without Incurring major hardware changes besides burner replacement.
RHey Stoker Corporation 1s currently modifying the burners used 1n Its turbo furnace to
lower NO emissions (Reference 3-62). The new burners are designed to be more flexible and to
control fuel/air mixing to reduce thermal and fuel NOX> Uith the new burners and changes 1n furnace
design Riley Stoker expects to meet current NSPS requirements without Increased carbon or unburned
hydrocarbon losses. The new burners can be used with coal, oil, and gas fuels but are not being
considered for retrofit application. No tes.*, data are available on the performance of the new
burners at present.
In summary, low NOX burners appear very attractive, with potential NOX reductions of the
order of 50 percent. Data from long term, full scale demonstrations a;-e imminent, and commercial
application 1s well underway.
3-27
-------
3.1.3.2 Burner Spacing
The Interaction between closely spaced burners, especially in the center of a multiple-burner
installation, Increases flame temperature at these locations. There is a tendency toward greater
NO emissions with tighter spacing and a decreased ability to radiate to cooling surfaces. This
effect 1s Illustrated by the higher NOX emissions from larger boilers with greater multiples of
burners and tighter spacing.
in most new utility boiler designs, vertical and horizontal burner spacing has been widened
to provide more cooling of the burner zone area. In addition, the furnace enclosures are bunt to
allow sufficient time for complete c ^ustlon with slower and more controlled heat release rates,
such as that associated with the oft-st -iometric operating mode. Furthermore, furnace plan areas
have been Increased to allow for larger heat transfer to the cooling walls. This increase in the
burner zone dimensions creates more wall area thus increasing the distance between evenly spaced
burners.
Horizontal burner spacing is largest for tahydntially fired boilers with the burners being
located at each corner of the furnace. Flames in a corner-fired unit Interact only at the center of
the furnace in the well know spiral configuration. As a result the flames radiate widely to the
surrounding cooling surfaces before Interacting with one another. Also, the tangential firing
configuration results in s'ow mixing of fuel with the combustion air. For these reasons,
tangent1ally-f1red boilers show baseline, uncontrolled emissions beloi; those for other utility
boilers firing configurations. It has been observed, however, that for many tangentially-fired
units, reductions 1n NOX balow the naturally low uncontrolled levels is more difficult than reduci^
NO on units with higher uncontrolled emissions (References 3-39 and 3-49).
3.1.3.3 Advanced Burner/Furnace Designs
A number of advanced burner designs are being developed and tested to reduce NO emissions
from coal- and o1l-f1red utility and Industrial boilers. Advanced burners, as compared to low NO
burners, are defined as those devices still under experimental or pilot scale development for
lowering NOX emission. Burner modification has the potential of lowering N0y emissions well below
levels attainable by conventional combustion modification techniques. Burner modification also has
the advantage of requiring minimal changes 1n current boiler design and operation and is suitable
for retrofit application. A few of the techniques under development are discussed below.
3-28
-------
Some manufacturers of oil-firing equipment ait ',r. +*< process of developing burners capable
of operating at very low levels of excess air. The low excess air requirements Increase boiler
efficiency and reduce fan power" corsumption while decreasing NO emissions. The low excess air nay
also reduce SO* conversion. The Peabody Engineering Company has designed the Air Pressure Recovery
(APR) burner designed to operate at excess oxygen levels down to 1/2 percent without Increase 1n
partlculate and unbumed hydrocarbon emissions. The Coen Company 1s developing the LEA burner which
uses a tip swlrler to operate down to 0.1 percent excess oxygen (Reference 3-63). Both burners are
currently undergoing testing and no data on NOX emissions are available.
For coal-fired utility boilers, Foster Wheeler 1s currently testing an advanced dual register
split flame burner design. A device id-led at the burner no.z1e splits the primary air-coal flow
Into several distinct streams. Coal particles become concentrated within each stream and, hence
diffuse more slowly Into the secondary air. This further Inhibits NOX formation by extending the
slow-burning characteristics of the dual register burner. Results from an Industrial size test
boiler are promising with a NOX level of approximately 130 ng/J (0.3 lb/10 Btu) for subbltumlnous
coal (Reference 3-61). However, the burner tested on a 375 MU electrical output boiler produced
approximately 215 ng/J (0.5 lb/10 Btu). A further modification of this burner with a variable
velocity split flame nozzle will be Installed, and a NOX level of 15C to 170 ng/J (0.35 to
0.4 lb/106 Btu) 1s expected (Reference 3-64).
Babcock & W11cox and Energy and Environmental Research, under EPA sponsorship, are developing
an advanced low NOX coal burner, the distributed fuel/air mixing burner, for field testing
(Reference 3-65). The burner if designed to control both thermal and fuel NOX. Coal and prlmiry
air are Injected from the center of the burner with a moderate axial component. This stream 1s
surrounded by a divided secondary alrstream with a swirl component for stabilization. Tertiary air
for burnout 1s added axlally around the periphery of the burner. The arrangement results In t. hot,
rich redrculatlon zone at the center or the 'flame. Time 1n the rich zone helps maximize ev./lutlon
of nitrogen from the char and reduce most of the fuel NO that may be formed. Also, axial addition
of the tertiary air leads to a large flame zone. Heat extraction prior to completion of burnout,
along with dilution of the tertiary air by combustion products, lowers the peak flame temperature,
thus reducing thermal NO . Although experimental prototypes have achieved NO emissions below
86 ng/J (0.2 lb/10 Btu), actual field testing has not yet been conducted.
Advanced burners are also under development for tangentlally-fired systems (Reference 3-67).
Combustion Engineering and Acurex Corporation, the latter under sponsorship of the EPA, are
3-29
-------
developing a burner called the "fuel-rich fireball." .This burner achieves a fuel-rich, relatively
low temperature flame by diverting some of the normal air flow along the furnace walls. The air
rejoins the combustion zone higher up in the furnace for complete fuel burnout. Demonstration
testing is underway. Another advanced burner for tangentially-fired systems is under development by
Mitsubishi Heavy Industries of Japan. The test program for this burner has also involved Japan's
Electric Power Development Company and Combustion Engineering of the United States. Pilot scale
testing has yielded emissions as'low as 86 ng/J (0.2 lb/10 Btu).
Babcock & Wilcox Company is developing a primary combustion furnace concept for coal-fired
utility boilers in a program sponsored by the Electric Power Research Institute (Reference 3-66).
The fundamental NO control process in this furnace is conversion of fuel nitrogen to N2 through
fuel-rich combustion. Pulverized coal is introduced into an extended combustor with substoichio-
roetric air, so that combustion occurs under fuel-rich conditions isolated from the rest of the
furnace. The length of the combustor is sufficient to provide the necessary residence time to
partially oxidize the coal and permit the desirable N2 producing reactions to occur. Heat is
removed along the combustion chamber to prevent slagging. Secondary air is added at the exit of the
primary combustion furnace to bring the combustion products to oxidizing conditions before they
enter the secondary furnace. Pilot scale testing of a 1 MW (4 x 10 Btu/hr) heat input prototype
has achieved the targeted NO level of below 86 ng/J (0.2 lb/10 Btu). Commercial offering of a
full scale furnace is not expected until at least 1983 (Reference 3-66).
In summary, several promising advanced burner/furnace concepts are under development and may
available in the next few years. These techniqt
lower than current combustion modification techniques.
become available in the next few years. These techniques may yield NO emissions substantially
3.1.4 Fuel Modification
While not necessarily thought of as NC'X control techniques, some additional methods either
are, or potentially will be, available for reducing NOV emissions in unique situations. These
, A
include various fuel modification techniques. Three candidate techniques are fuel switching, fuel
additives, and fuel denitrification.
3-30
-------
3.1.4.1 Fuel Switching
In the past, the predominant reasons for fuel switching were related to fuel cost or SCL
control. However, conversion to a fuel with a reduced nitrogen content or to one that burns at a
lower temperature may also result in a reduction in NOX emissions. NOX emission factors for various
types of fuels were presented in Section 2.
Natural gas firing is an attractive NO control strategy because of the absence of fuel NO
X X
in addition to the flexibility it provides for the implementation of combustion modification
techniques. However, for large fuel burning applications, fuel choice decisions tend to depend on
fuel costs and other regulatory constraints, such as the Powerplant and Industrial Fuel Use Act of
1978. Indeed, the trend is toward the use of coal for electric power generation and larger
industrial processes. Fuel switching to natural gas or distillate oil is not a promising option for
widespread implementation.
Western coals constitute one abundant alternate source of potentially low-NO fuels. The
direct combustion of western subbituminous coals in large steam generators may produce lower NO
emissions than with combustion of eastern bituminous coals. Three mechanisms could result in lower
NO emissions: first, western coals in general contain less bound nitrogen than eastern coals on a
unit heating value basis; second, the excess Op in a steam generator burning western coal can be
maintained at very low levels; and third, the high moisture content of western coal produces lower
flame temperatures. However, some studies have indicated that these factors may be offset by the
higher fuel 02 levels in western coals. These higher levels may lead to increased conversion of
fuel bound nitrogen to NO (Reference 3-68).
\
Some potential problems associated with burning low sulfur, high moisture content coals in
combustion equipment designed for higher quality coals are listed below (Reference 3-69):
- Poor i-gnition;
- Reduced boiler load capacity;
- Increased carbon loss;
- Boiler slagging/fouling; .
- High superheat steam temperature;
- Flame instability;
- Increased boiler maintenance;
- Reduced boiler efficiency; and
- Reduced collection efficiency of electrostatic precipitator (ESP).
3-31
-------
However, most of these operational problems can be solved with current boilers specifically designed
to bum these lower grade coals.
Formerly, a major incentive for switching to western coals was the low sulfur content of
these fuels. Economic conditions made fuel switching from high sulfur eastern bituminous coals to
low sulfur western subbituminous coal competitive with the cost of gets scrubbing for S02 removal.
Therefore, low sulfur, low nitrogen western coals represented a promising short-range option in fuel
switching for large industrial and utility boilers. However, the 1977 Clean Air Act Amendments
.'
require that NSP5 be based on a percentage reduction in the pollutant emissions which would have
resulted from the use of fuels which are not subject to treatment prior to combustion. This reduced
or eliminated the advantage of fuel switching in many applications.
A potential long-range option is the use of clean synthetic fuels derived from coal.
Candidate fuels include low to high Btu gas (3.7 to 30 MJ/Nm3, or 100 to 800 Btu/scf) and synthetic
liquids and solids. Two alternatives for utilizing low- and intermediate-Btu gases (up to 26 MJ/m ,
or 700 Btu/scf) are firing in a conventional boiler or in a combined gas and steam turbine power
generation cycle. The NO emissions from lower-Btu gas-fired units are expected to be low due to
reduced flame temperatures corresponding to the lower heating value of the fuel. The effects on NO
formation of the molecular nitrogen and the intermediate fuel nitrogen compounds, such as ammonia,
in the lower-Btu gas have not yet been fully determined and require further study.
The synthetic fuel oils or solid solvent refined coal (SRC) may be expected to be high in
fuel nitrogen content even though some denitrification may occur in the desulfurization process.
This high nitrogen content, carried over from the parent coal, would promote high NOX emissions.
Other/potential alternate fuels that might be considered and their potential for fuel or thermal NOX
are listed in Table 3-3.
TABLE 3-3. NOX FORMATION POTENTIAL OF SOME ALTERNATE FUELS
FUEL
Shale Oil
Coal -Oil Mixture
Coal -Liquid Mixtures
Methanol
Water-Oil Emulsion
Hydrogen
THERMAL NOX
Moderate
Moderate
Low
Low
Low
High
FUEL NOX
High
Moderate
Unchanged
Low
Unchanged
Low
aFuel NO is probably unchanged unless a significant amount of low
nitrogen oil or methanol replaces part of the coal on a heating
value basis.
3-32
-------
Besides coal, shale oil is another abundant source of nonpetroleum fossil fuel in the United
States. However, current shale oil production is very limited due to economics. The combustion of
shale oil will cause higher levels of fuel NO because this fuel generally contains bound nitrogen
X
in excess of 2 percent. Distillation of shale oil would reduce fuel nitrogen content, however.
Coal-oil mixtures have recently become of interest as an alternate fuel which could reduce
fuel costs for existing oil-fired boilers. NO from combustion of this fuel will depend on the
quantity of nitrogen present in the coal and oil and the percentages of coal and oil used to make
the mixture. However, NO emissions are expected to be lower than emissions obtained from
combustion of coal only.
Other types of coal liquid mixtures are also under development, including coal-water, coal-
oil-water, and coal-alcohol. The primary incentive for development of these fuels is to obtain a
relatively cheap coal-based fuel for .combustion in existing oil-fired boilers. However, some boiler
modifications are expected to be required to burn these fuels. These fuels should burn at a lower
temperature than the parent coal thereby reducing thermal NOV formation.
X
Methanol is currently produced from the synthesis of methane from natural gas. Some future
production may also come from synthetic gas generated from coal and biomass. Baseline NO emissions
X
from the combustion of methanol in an experimental hot wall furnace system were reported at 50 to
70 ppm, compared to 240 to 300 ppm for distillate oil; With flue gas recirculation, the NO
emissions from methanol combustion were reduced to 10 ppm, or 15 percent of the baseline level
(Reference 3-70).
In gas turbines 74 percent less NO was produced using methanol, compared to distillate oil.
The hot wall experimental furnace showed a 20 percent increase in stack heat loss compared to a loss
of 14 percent for distillate oil (based on 115 percent theoretical air at a 473°K (390°F) stack
temperature). For natural gas, turbine efficiency levels increased by 6 percent due to higher inlet
temperatures.
Since water-oil emulsions affect only thermal NO these alternate fuels have a definite NO
X X
reduction potential when distillate oil is used (Reference 3-71). NO emission levels from
emulsions with approximately 50 mass percent water in distillate oil approached the, levels, obtained
from methanol combustion (Reference 3-72).
3-33
-------
Hydrogen as a fuel is used in high energy production concepts such as rocket engines. The
high levels of thermal energy released make this fuel attractive for other energy conversion
systems. Thermal NO levels are, however,.high when hydrogen meets with oxygen in the presence of
atmospheric nitrogen.
3.1.4.2 Fuel Additives
For purposes of this document, a fuel additive is a substance added to any fuel to inhibit
formation of NO when the fuel is burned. The additive can be liquid, solid, or gas. For liquid
fuels, the additive should preferably be a liquid soluble in all proportions in the fuel, and it
should be effective in very small concentrations. The additive should not in itself create an air
pollution hazard nor be otherwise deleterious to equipment and surroundings.
In 1971, Martin, et al., tested 206 fuel additives in an oil-fired experimental furnace, and
four additives in an oil-fired packaged boiler. None of the additives tested reduced NO emissions
but some additives containing nitrogen increased NO formation (Reference 3-73). Fuel additives
reduced NOX emissions from gas turbines by an average of 15 to 30 percent but are not attractive due
to added cost, serious operational difficulties and the presence of the additives, as pollutants, in
the exhaust gas (Reference 3-74). Average NO reductions of 15 to 18 percent have been recorded
using fuel additives in diesel engines (Reference 3-75).
3.1.4,3 Fuel Denitrification
Fuel denitrification of coal or heavy oils could in principle be used to control the
components of NO emission due to conversion of fuel bound nitrogen. The most likely use of this
concept would be to supplement combustion modifications implemented for thermal NO control.
X
Current technology for denitrification is limited to the side benefits of fuel pretreatment to
remove other pollutants. There is preliminary data to indicate that marginal reductions in fuel
nitrogen result from oil desulfurization (Reference 3-76) and from chemical cleaning or solvent
refining of coal for ash and sulfur removal (Reference 3-77). The low denitrification efficiency of
these processes does not make them attractive solely on the basis of NO control. They may prove
cost effective, however, on the basis of total environmental impact.
3-34
-------
3^1.5 Alternate Processes
In the future, some alternate processes may be available which result in lower NO emissions
than conventional stationary combustion technologies. Some candidate processes include fluidized
bed combustion, catalytic combustion, repowering, and combined cycles.
3.1.5.1 Fluidized Bed Combustion
In a fluidized bed combustor (FBC) combustion occurs in an air-supported bed of relatively
large (~3.2mm) coal ash and sand or limestone particles. The temperature- in the bed is generally in
the range of 1,070°K to 1,270°K (1,500°F to 1,900°F) which makes the combustion process self-
sustaining. The combustor may be at atmospheric pressure (A.FBC) or it may be pressurized (PFBC).
A 30 MW AFBC pilot plant began operation in late 1976 (Reference 3-78). Pressurized systems
are still being tested, with a pilot plant planned for the early 1980's. Results of recent work in
FBC, the status of FBC development, and EPA, DOE, and EPRI FBC programs can be found in
Reference 3-78.
Suggested advantages for fluidized bed combustion compared to conventional boilers are:
(1) compact size yielding low capital cost, modular construction, factory assembly and low heat
transfer area, (2) higher thermal efficiency, (3) lower combustion temperature resulting in less
fouling and corrosion and reduced NO formation, (4) potentially efficient sulfur oxides control by
A
direct contact of coal with an SOg acceptor, (5) fuel versatility, (6) applicable to a wide range of
low-grade fuels including char from synthetic fuels processes, and (7) adaptable to a high
efficiency gas-steam turbine combined power generation cycle. The principle disadvantages of FBC
are: (1) potential large amounts of solid waste (the sulfur acceptor material) and (2) heavy
particulate loading in the flue gas.
The feasibility of FBC for power generation and utility boilers depends in part on the
following: (1) development of efficient methods for regeneration and recycling of the dolomite/
limestone materials used for sulfur absorption and removal, (2) obtaining complete combustion
through fly ash recycle or an effective carbon burnup cell, (3) development of a hot-gas particulate
removal process to permit use of the combustion products in a combined-cycle gas turbine without
excessive blade erosion.
Nitrogen oxides emissions from fluidized bed combustors have been shown to be predominately
fuel-derived. Seven to ten percent of fuel nitrogen is converted to NO (References 3-79 and 3-80).
X j
Experiments with nitrogen-free fuels resulted in NO concentrations in agreement with equilibrium
3-35
-------
values at the bed temperature. However, coal-fired experiments resulted in NO concentrations in
A
excess of the equilibrium values. Furthermore, experiments using nitrogen-free gases with coal
yield substantially similar NOX levels as combustion in air (Reference 3-79).
NOV emissions have been found to be slightly dependent on coal particle size, the type and
X
amount of sulfur acceptor, the amount of excess air and the design of the combustor itself.
Emission levels from pressurized fluidized bed combustors are significantly less than from
atmospheric combustors. This is probably a result of greatly increased NO decomposition rates at'
elevated pressures. Even at 100 percent excess air, NO emissions from a PFBC are well below the
current NSPS. Results of 160 ng/J (0.37 lb/10 Btu) have been reported (Reference 3-78).
In general, NO control in F5C is a matter of good management of the normal process
variables. If more stringent standards are enacted, conventional NO controls, such as flue gas
X '
recirculation and off-stoichiometric combustion, may be used. Exploratory results indicate that
two-stage combustion could be advantageous for both NO and SO control.
A X
3.1.5.2 Catalytic Combustion
Catalytic combustion refers to combustion occurring in close proximity to a solid surface
which has a special (catalytic) coating. A catalyst accelerates the rate of a chemical reaction, so
that substantial rates of burning should be achieved at low temperatures, avoiding the formation of
NO . Moreover, the catalyst itself serves to sustain the overall combustion process, thereby
A
minimizing the stability problems (References 3-81 and 3-82). However, the overall success of a
catalytic combustion system in reducing, CO and unburned hydrocarbons (UHC) to low levels is a
function of both heterogeneous and gas phase reactions; surface reactions alone appear to be unable
to achieve the desired low levels.
Emissions from catalytic experiments have typically been: NO <2 ppm, UHC e4 ppm, and CO =
10 to 30 ppm. Both gaseous and distillate fuels have been used and combustion efficiencies above
95 percent have been obtained (Reference 3-82).
At high temperatures, above 1,270°K (1,830°F), catalyst degradation can be significant.
Excess air can be used to lower the bed temperature; but except for gas turbines, excess air is
unattractive since it also reduces thermal efficiency. Further research is underway to consider
other systems, such as catalyst bed cooling, exhaust gas recirculation and staged combustion to
maintain a low bed temperature.
3-36
-------
Recent tests evaluated the applicability of catalytic combustors for gas turbines. Test
fuels used were No.2 distillate oil and low Btu synthetic coal gas, for a range of pressure,
temperature, and mass flow conditions. Test results show that the catalyst bed temperature profile
at the bed exit was very uniform for low Btu gas, but not as uniform for No.2 oil. Exceptionally
low emissions (2 to 3 ppm NO , 20 to 30 ppm CO) were achieved for both fuels, and unburned hydro-
A
carbons were less than 1 ppm (Reference 3-83). However, much additional work is needed before
catalytic combustion can be applied to gas turbines in the field*
Catalytic combustion has been demonstrated to be effective in removing pollutants such as
NOV, CO, and UHC, but at present, catalytic combustors are limited by the catalyst bed temperature
X
capability. Various government agencies and private industries are developing catalysts that will
withstand high temperatures, retain high catalyst activity, and last longer. Catalytic combustion
systems are'also under development; it appears that in the future catalytic combustion concepts may
be incorporated into new gas turbine and residential, commericial, and industrial heating designs.
»
3.1.5.3 Repowering
Repowering adds a combustion turbine to an existing steam plant, providing additional
capacity at lower initial costs and lower energy costs than other spare capacities available to a
utility.
Repowering includes: (1) steam turbine repowering, in which gas turbines and new heat
recovery boilers are added to an existing steam electric generating plant; (2) boiler repowering in
which gas turbines are added to the existing steam generating facilities for power generation,
requiring the conversion of existing conventional boilers to heat recovery type boilers; and (3) gas
turbine repowering in which a steam generating plant is added to an existing gas turbine plant
(References 3-84 and 3-85).
Depending on the system and power needs, repowering of existing facilities offers the
following advantages:
- There is no need to acquire and develop a new plant site;
- Repowering generally requires smaller increments of investment, saving on fixed
charges since major investment on new plants is deferred;
- Repowering improves heat rate, which lowers fuel consumption;
3-37
-------
- The environmental impact is reduced, with improving schedules for environmental
and site related approvals;
- For boiler and steam turbine repowering, there is no increase in cooling water
requirements; and
- Gas turbines may be operated independently as peaking units, which provides
greater plant flexibility.
References 3-84 and 3-85 describe in detail the application of repowering to boiler, gas
turbine, and steam generating plants; savings in capital and operating costs are anticipated.
Repowering of two steam turbine units in the City of Glendale, California increased power output by
75 MM and reduced power cost to the consumer by 8 percent (Reference 3-86). Under contract from the
Electric Power Research Institute, Westinghouse Electric Corporation is evaluating repowering
conventional steam power plants without replacing the boiler. Earlier pilot scale work for EPRI by
KVB Inc. shows a low NOV potential for repowering. The boiler is fired fuel-rich using approxi-
A
mately 85 percent of the NOX bearing gas turbine exhaust as the combustion air. The remaining gas
turbine exhaust provides the boiler second stage air which is injected through overfire air ports
above the fuel-rich primary stage. Up to 55 percent of the NO in the gas turbine exhaust is
chemically reduced by the fuel rich primary stage of the boiler. 'Also, the use of overfire air
reduces the NO formed in the boiler by up to 50 percent.
X
3.1.5.4 Combined Cycles
Combined cycles may, in the long term, reduce emissions of sulfur oxide, nitrogen oxide,
particulate matter, and waste heat while generating power at efficiencies higher than conventional
fossil-fueled steam stations (Reference 3-87). The combined gas and steam turbine system could
consist of a gas turbine firing a coal-derived fuel., which exhausts into an unfired waste-heat
recovery boiler. In this system, a portion of the power would be generated by the gas turbine and a
portion by the steam boiler system. Combined cycle efficiency improves significantly as the gas
turbine inlet temperature is increased. At turbine inlet temperatures of 1,480°K (2,200°F), an
efficiency improvement of 2 percentage points per 55°K (100°F) increase in turbine inlet temperature
is found.
3-38
-------
3.2 COMBUSTION FLUE GAS TREATMENT
Historicany4 the major NOX,control emphasis in the United States has been on combustion or
process modification. However, in Japan where NO emission standards are more stringent, flue gas
X
treatment (FGT) technologies have undergone extensive development and implementation. Recently, in
the U.S. several pilot and demonstration scale units have been built and operated.
Flue gas treatment consists of any of several technologies designed to, remove or eliminate
NO in the flue gas downstream of the combustion zone. Several FGT processes are potentially
capable of very high (greater than 80%) NO removal efficiencies. And since FGT processes are
located downstream of the combustion zone, this NO control can occur in addition to any control
X
already achieved with combustion modifications. The application of FGT will usually involve
considerably more construction and expense than the comparatively simple combustion modifications.
But FGT may prove to be a viable control alternative for situations where very high levels of
control are needed.
These postcombustion processes can be divided into dry or wet types. Some dry processes are
designed to control NO alone while others are designed to control SOV and NOV. The wet processes
A A X
employ a wet scrubber to control both pollutants. This subsection briefly describes some of the
major FGT processes under development and how they work to control NO . Most of the development of
FGT processes has been undertaken with application to utility boilers in mind. Consequently, more
extensive discussions of FGT technology development status and process impacts are reserved to the
section on NOX control for utility boilers, Section .4.1. Much of the material on FGT was taken from
References 3-88 and 3-89; the reader is referred to these and other documents (e.g. References 3-90,
3-91, and 3-92) for additional information.
3.2.1 Dry Flue Gas Treatment
The dry processes can be categorized into four subdivisions: catalytic reduction,
noncatalytic reduction, adsorption, and irradiation. The majority of the dry processes are of the
reduction type. These catalytic and noncatalytic reduction processes can also be classified as
selective or nonselective processes based on the type of reducing agent used. The majority are
selective and usually use NH3 as the reducing agent. If the NH, is injected after the boiler
economizer, where'temperature of the flue gas is about 370°C to 450°C (700°F to 800°F), a catalyst
is necessary. These processes are described as selective catalytic reduction (SCR) processes. If
NH3 is injected into the secondary superheater region of the boiler, where temperature of the flue
gas approaches 980°C (1,800°F), a catalyst is not necessary. These processes are described as
3-39
-------
selective noncatalytic reduction (SNR) processes. This subsection briefly describes SCR, SNR, and
other dry F6T processes.
3.2.1.1 Selective Catalytic Reduction (SCR)
The SCR method is the most advanced FGT method, and the one on which the overwhelming
majority of existing FGT units are based. As with the majority of all FGT processes, most of the
SCR processes were developed in Japan.
SCR systems use ammonia to selectively reduce nitrogen oxides. The chemical mechanisms can
best be summarized by the following gas phase reactions.
4NO + 4NH3 + 02 * 4N2 + BHgO (3-6)
2N02 + 4NH3 + 02 * 3N2 + 6H20 (3-7)
The first reaction predominates since flue gas NOX consists primarily of NO. Oxygen is in large
excess in the flue gas and does not limit the extent of reaction.
A process flow diagram is shown for a NO only SCR process in Figure 3-7. Flue gas is taken
from the boiler between the economizer ^and air preheater. Ammonia., taken from a liquid storage tank
and vaporized, is injected and mixed with the flue gas prior to the reactor. The flue gas passes
through the catalyst bed where NO,, is reduced to N9. The flue gas then exits the reactor and is
A ^
sent to the air preheater and, if necessary, additional pollutant control devices (e.g. FGD system,
ESP).
With SCR systems it is desirable to treat flue gas exiting the economizer at 300 to 400°C,
prior to any air preheater, since it is at this temperature range that the catalysts show the
optimum range of reactivity and selectivity (Reference 3-88). Research and development on catalyst
formulations and shapes during recent years has resulted in some standardization among the catalyst
types offered. A catalyst formula consisting primarily of oxides of titanium and vandium appears to
be universally used (Reference 3-92). This formulation has proven to be resistant to poisoning.by
sulfur compounds in the flue gas.
Reactor designs tend to vary depending on the application. Catalyst pellets in a fixed bed
are commonly used for gas-fired applications. For oil- or coal-fired applications where the flue
gas contains particulate matter, reactor designs usually incorporate honeycomb, pipe, or parallel
3-40
-------
Boiler
Economi zei
i
NH.
Air
Preheater
Air in
Reactor
Gas to stack
or additional
pollution control
Figure 3-7. Flow diagram for typical NO -only SCR process.
X
3-41
-------
plate shaped catalysts which allow the flue gas to pass in parallel along the catalyst surface.
Another design uses a moving bed arrangement.
NOV emissions from boilers using SCR processes in Japan are generally reduced by 80 percent.
X
Higher reductions are possible but costs are greater for these units (Reference 3-92). As discussed
1n Section 4.1, there are presently over 60 full scale SCR units operating on gas- or oil-fired
boilers in Japan. Also, two commercial units are operating on coal-fired boilers. Construction is
scheduled to be completed during 1981-1984 on at least 14 additional SCR units on coal-fired boilers
in that country. In the U.S., SCR applications have been limited to a few pilot scale units on
coal-fired boilers and a demonstration scale unit under construction on an oil-fired utility boiler.
3.2.1.2 Selective Noncatalytic Reduction (SNR)
Exxon Research and Engineering Corporation developed the SNR process in which NH, is injected
into the boiler where proper flue gas temperatures (about 900 to 1010°C, 1650 to 1850°F) allow the
reduction of NOV by reaction with NH, to proceed without a catalyst. Generally, 40 to 60 percent
A «J
NO., reduction is achieved with NH,:N() molar ratios of 1:1 to 2:1. SNR may be more attractive than
X OX
SCR in cases where only 40 to 60 percent NO is needed since SNR is simple and does hot require
expensive catalysts.
The general disadvantage of SNR is the limited NO control achievable, especially with larger
/>
boilers. This limited control generally results from the difficulty of achieving rapid uniform
mixing of NHj with the flue gas and from the variations of flue gas temperature and composition
usually present within the boiler region where SNR occurs. NH, consumption and unreacted NH, levels
can also be high.
There are several large SNR units installed in Japan, between 30- and 100-MW capacity,
* »
mostly supplied by Tonen Technology (a subsidiary of Toa JJenryo) which has a license from Exxon.
These units are operated on gas- and oil-fired boilers or furnaces. Practically all are only for
emergency use during a photochemical smog alert or when total plant emissions exceed the regulation.
There are presently two commercial SNR plants operating in the United States. One is on a
glass melting furnace and the other a petroleum refinery, both located in California. The construc-
tion of five other industrial-scale units is planned. The SNR process is also being installed by
Exxon at the No.4 oil-fired unit of the Haynes Station of the Los Angeles Department of Water and
Power.
3-42
-------
3.2.1.3 Other Dry FGT Processes
In addition to SCR and SNR, other dry NOV control processes are being developed which will
i A
allow simultaneous control of SO . These include:
(1) Activated carbon processes where NH, reduces NO,, to N,;
«j X &
(2) Copper oxide processes where NH. reduces NO to N^; and
(3) Electron beam irradiation processes in which NH, is added
to produce ammonium sulfate and nitrate.
The optimum temperature range for simultaneous SO and NO control with activated carbon
X A
processes is 220°C to 230°C (430°F to 445°F). Although NOV may be adsorbed below 100°C (212°F), for
* A
treating large quantities of flue gas above 100°C the carbon is mainly useful as an NO reduction
catalyst. Therefore, while NO is converted to N2 by reaction with NH3 in the presence of the
activated carbon catalyst, S02 is simultaneously adsorbed by the carbon to form H2SO,. The H^SO,
may also compete for NH, in forming ammonium sulfate or bisulfate. The formation of these ammonium
salts increases NH, consumption and also lowers catalyst activity. The carbon must be regenerated,
either by washing or thermal regeneration. Washing produces a dilute solution. Concentration of
the solution to produce a fertilizer requires much energy. Therefore, thermal regeneration seems to
be preferred. A concentrated S02 gas is recovered, which can be used for sufuric acid or elemental
sulfur production.
The major drawback of the activated carbon processes is the enormous consumption of activated
carbon, which is more expensive than ordinary carbon used only for SO removal. Since carbon and
A
ammonia consumption increase with the S02 content of the flue gas, the process is best suited for
flue gases relatively low in S02« In Japan' Sumitomo Heavy Industries and Unitika Company have
operated activated carbon pilot plants of 0.6 MW and 1.5 MW capacity respectively.
The Shell Flue Gas Treatment process may simultaneously remove SO and NOX. SOX reacts with
the copper oxide acceptor to form copper sulfate. The copper sulfate and copper oxide are SCR
catalysts for the NOV reduction by NH,. Regeneration of the multiple catalyst beds by a reducing
X «'
gas, such as H-, yields a S02-rich stream that can be used to produce liquid S02> elemental sulfur,
or sulfuric acid. By eliminating NH3 injection, the process is strictly an FGD process, whereas,
eliminating regeneration of the catalyst beds allows the process to be used for only NO control.
The major disadvantages are the large consumption of fuel for making hydrogen and the catalyst
expense.
3-43
-------
In addition to the EPA-sponsored pilot plant mentioned earlier, the process has been installed
in Japan on a 40-MW oil-fired boiler. The unit has demonstrated 90% SOX removal and 70% NOX
reduction.
Another process for simultaneous SO and NO control is the electron beam process developed by
Ebara Manufacturing Company in Japan. NH, is added to the flue gas, after which the gas stream is
irradiated with an electron beam in a reactor, promoting the conversion of SO . NOV, and ,NH, to
A A O
ammonium sulfate and ammonium nffrate. The ammonium sulfate and aranonium nitrate may be collected
downstream in an ESP or baghouse and potentially sold as a fertilizer. The most economically
practical removal efficiency range appears to be 80% to 90% for each of NO and SO , though higher
removals can be achieved with much greater electron beam energy input. The optimum temperature
range is 70°C to 90°C (160°F to 195°F).
Ebara has worked on the process since 1971. It has been tested at a 0.3 MW and 3 MW scale in
Japan. Avco Corporation in the United States has also examined this technique and has a
cross-licensing agreement with Ebara in sharing of technology and in marketing of the process.
Although the process appears attractive because of simplicity, simultaneous SOV and NO control, and
' A A
byproduct formation, there are still many questions concerning costs and byproduct quality which
must be determined.
Development of an alternate electron beam scrubbing process was begun in 1979 by ,
Research-Cottrell under contract to the Department of Energy (DOE). With this process a lime spray
dryer is located upstream of the reactor. Calcium sulfate and calcium nitrate are produced in the
reactor and caught in a downstream baghouse. Some bench scale testing has been done with this
process. DOE plans proof-of-cohcept scale testing of both the ammonia injection and lime slurry
Injection electron beam processes on real coal-fired slip streams (Reference 3-93).
3.2.2 Wet Flue Gas Treatment
The wet F6T processes normally involve simultaneous removal of SO and NO . The major problem
X A
associated with wet NO control processes is the absorption of NO by the scrubbing solution. NO
in the flue gas is predominantly NO, which is much less soluble than NOp, whereas, N0~ is even less
soluble than S0£. The two common methods of removing the NO in flue gas by wet processes are:
(1) direct absorption of the NO in the absorbing solution or (2) gas-phase oxidation to convert the
3-44
-------
relatively insoluble NO to NOp, followed by absorption of NOg. Presently, development of the wet
NO FGT processes has practically ceased because of the complexity and unfavorable economics of
these processes in comparison with the dry processes (Reference 3-94).
3.3 NONCOMBUSTION GAS CLEANING
Emissions from noncombustion sources as industrial or chemical processes are small relative
to the total emissions from stationary sources (1.7 percent). Nationwide NO emissions from nitric
acid manufacturing are estimated for the year 1980 at 100 Gg (110,000 tons) uncontrolled emissions,
which is about 1.0 percent of the total stationary source emissions. The Environmental Protection
Agency issued standards (under the authority of the Clean Air Act) that new nitric acid plants
constructed after December 23, 1971, have a maximum permitted nitrogen oxide effluent of 1.5 kg
(measured as N02) per Mg of acid (100 percent basis) produced (3 Ib/ton). This is equivalent to
approximately 210 ppm NO . For existing plants the maximum nitrogen oxides permitted has been set
at 2.75 kg/Mg (5.5 Ib/ton) of acid or approximately 400 ppm NO in several states. These standards
X
were established in consideration of the then available technology, which was catalytic reduction of
NO to N£ and water using methane or hydrogen.
Several economic factors,, discussed in Section 3.3.2.4 have stimulated development of
improved processes for tail gas cleaning and improvements in the nitric acid process itself. One of
the major considerations is that much of the residual oxides of nitrogen formed in the manufacture
of nitric acid can be recovered and converted into nitric acid, thus increasing the plant yield.
Also, new plants can be designed to have low NO emissions without add-on control equipment. These
designs will be described in Section 3.3.1'. Techniques suitable for retrofit abatement for older
plants or add-on controls for plants built using old technology include catalytic reduction,
extended absorption with and without refrigeration, wet chemical scrubbing, and molecular sieve
adsorption. These techniques will be descriBed in Section 3.3.2. The techniques used , for other
noncombustion sources, such as explosive plants and adipic acid plants, are basically the same as
those used for nitric acid plants, but vary with choice depending on economies of scale and
throughput.
3.3.1 Plant Design for NOX Pollution Abatement at New Nitric Acid Plants
Nitric acid is manufactured in the United States by the catalytic oxidation of ammonia over a
platinum catalyst with the subsequent absorption of the product gases, primarily N02 and NO, by
3-45
-------
water to make nitric acid. A more detailed discussion of the chemical process is given in
Section 6. Each of these two catalytic processes have optimum conversions at different operating
conditions. Moderate pressures of 300 to 500 kPa allow longer catalyst life by lowering operating
temperatures in the initial oxidation reaction. Higher pressures in the range of 800 to 1100 kPa
(116 to 160 psia) allow higher absorption rates in the absorption columns with smaller equipment
sizes and lower costs. The higher conversions of M02 to HNO, allow for smaller equipment for both
the main process plus any tail gas treatment required to meet emission standards. Currently most
existing plants operate at low or moderate pressures throughout the process. Sections 3.3.1.1 and
3.3.1.2 will discuss how the design of new nitric acid plants has taken these factors into account
to increase conversion and decrease emission control costs.
3.3.1.1 Absorption Column Pressure Control
By designing a new plant so that the inlet pressure at the absorber is 800 to 1000 kPa
(116 to 145 psia), the efficiency of the absorber can be increased so that an effluent of less than
200 ppm NO is emitted. A high inlet gas pressure at the absorber can be achieved either by running
X
the ammonia-oxygen reaction at high pressure, or by running the ammonia-oxygen reaction at low
pressure, with compression of the gas stream before introduction to the absorber. Higher absorption
pressures will increase the conversion of NO, to nitric acid and minimize NO emissions. However,
fc A
there are economic penalties in the form of increased equipment cost, thicker walls and compressors,
and increased maintenance costs.
3.3.1.2 Strong Acid Processes
Nitric acid is usually produced at strengths of 50 to 65 percent by weight in water due to
azeotrope limitations. Azeotropic conditions result in a constant composition in both vapor and
liquid phases. With higher pressures nitric acid up to 68 percent can be obtained. Further
concentration is sometimes accomplished by dehydration of the acid or further distillation with
sulfuric acid addition.
However, nitric acid of high strength can be made directly from ammonia by the Direct Nitric
Acid (DSNA) process. Ammonia is burned with air near atmospheric pressure, and the nitrogen oxides
are oxidized to nitrogen dioxide in a contact tower. The nitrogen dioxide is then separated from
the gas stream by physical absorption in chilled high-concentrated nitric acid, stripped by
distillation and then liquified as NO.
3-46
-------
The liquid dinitrogen tetroxide is pumped to a reactor together with aqueous nitric acid.
Pure oxygen is added and the dinitrogen tetroxide reacts at a pressure of approximately 5200 kPa
(760 psig) directly to highly concentrated nitric acid. Variations on the process can produce both
strong (98 to 99 percent) nitric acid and weak (50 to 70 percent) nitric acid at the same plant
(Reference 3-95). Tail gas emissions from this process are within the 1.5 g/kg (3 Ib/ton) NO
X
regulation. This occurs primarily by ensuring oxidation to NO- and physical absorption with the
concentrated nitric acid at low temperature.
Concentrated nitric acid has also been made by the SABAR (Strong Acid By Azeotropic Reactivi-
cation) process. Ammonia combustion occurs at near atmospheric pressure and at 1,120°K (1,560°F)
with the usual waste-heat boiler, tail gas preheater, cooler/condenser effluent train. By mixing
the combustion gases with feed air and recycled nitrogen dioxide, and compression, nearly all the NO
is converted to NO-. Chemical absorption with an azeotropic mixture of about 68 percent (by weight)
nitric acid produces a superazeotropic mixture. A 99 percent (by weight) overhead product is
produced by vacuum distillation.
3.3.2 Retrofit Design for NOX Pollution Abatement at New or Existing Nitric Acid Plants
Most existing nitric acid plants were not designed with the present NO emission standards in
A
mind. Abatement methods for these plants are installed on a retrofit basis. The available abate-
ment methods include chilled absorption, extended absorption, wet scrubbing, catalytic reduction,
and molecular sieve adsorption. In this section, these various control techniques for NOX are
described. These same procedures are also used on new nitric acid plants using the earlier low or
moderate operation pressure design where the abatement facility is designed to process the tail gas
to meet the 1.5 g NO^/kg of acid product (3 Ib/ton) emission standard.
3.3.2.1 Chilled Absorption
The basic principle involved is that the amount of NO that can be removed from the process
gas by the absorber (water) increases as the water temperature decreases. Therefore, this method of
retrofit provides for chilling of the water prior to entry into the absorption tower or by direct
cooling of the absorption trays. This method of NOV reduction has only provided marginal results
A
and has had problems in continuously meeting the NSPS, especially in warm weather. Refrigeration
requirements can prove costly, both in equipment and energy use.
3-47
-------
3.3.2.2 Extended Absorption
One of the most commonly used retrofit processes, which has been used effectively to meet the
HSPS, 1s extended absorption. Figure 3-8 shows the flow diagram of a nitric acid plant after
addition of the extended absorption system, which consists of an additional absorber and a pump.
This method is offered by several licensors both with and without other features such as compression
*
of the tail gas before entry to the additional tower or a supply of chilled water to the absorption
column trays. Because of the additional pressure loss in the second column an inlet pressure of at
least 700 kPa (100 psia) is preferred to make the economics of this method attractive.
3.3.2.3 Wet Chemical Scrubbing
Wet chemical scrubbing removes NO from nitric acid plant tail gases by chemical reaction.
Liquids such as alkali hydroxide solutions, ammonia, urea, and potassium permanganate convert NO* to
nitrates and/or nitrites. These techniques produce a liquid effluent which needs disposal. For
three recent techniques - urea scrubbing, ammonia scrubbing and nitric acid scrubbing - the effluent
1s a valuable byproduct which can be reclaimed and sold as fertilizer.
Caustlc Scrubbi ng
In this process, NO in the tall gas reacts with sodium hydroxide, sodium carbonate, or
ammonium hydroxide to form nitrite salts. Although caustic scrubbing removes NO from the tail gas,
it has not found extensive use in the industry because of the difficulties encountered in disposing
of the spent solution. The Alkali metal nitrite and nitrate salts contained in the spent solution
become a serious water pollutant if released as a liquid effluent, and their concentrations are too
dilute for economic recovery.
Urea Scrubbing
Urea can be used to treat all gases for NO control since it reacts rapidly with nitrous
acid. Nitrogen dioxide, N02 reacts with water to form both nitric acid (HN03) and nitrous acid
(HN02) in equal proportions. Nitrous acid will rapidly decompose to form NO and N02- Urea
(CO (NHpJo) when contacted with the tail gas will absorb N0« indirectly as nitrous acid to form
ammonium nitrate, NH.NOg and free nitrogen, N^. By depleting .the liquid phase of nitrous acid the
equilibrium conversion of nitric oxide, NO, to nitrogen dioxide occurs to remove NO also. The
3-48
-------
wroonla
compressor
converter
*»
ID
absorber
j-^i»
heat condenser
recovery
A
I
I
I
extended P~
X" ~ ~ T absorber I
A | J '
tall gas.^
I
I
^-jEyip;--^
product Lfll
pump
power
recovery
product
acid
Figure 3-8. Extended absorption system on existing nitric acid plant.
-------
result is conversion of N02 to either free nitrogen which is vented to the atmosphere or ammonium
nitrate which is sold as a fertilizer.
Ammonia Scrubbing
Ammonia, a weak base, can be used to scrub the oxides of nitrogen (weak acids) from the
nitric acid plant tail gas. The product of this scrubbing reaction is an ammonium nitrate solution
'(NH.NO-) which can be recovered and sold as fertilizer. This process can be applied to tail gas
*r O
concentrations up to 10,000 ppm and requires 1 to 1.5 percent excess oxygen.
Nitric Acid Scrubbing
Nitric acid scrubbing of tail gas has been commercially applied by one licensor. The process
uses both physical absorption and stripping and chemical oxidation absorption. The process uses
only water and nitric acid and converts nitrogen oxides in the tail gas to nitric acid at concentra-
tions which can be commercially utilized (Reference 3-96).
Potassium Permanganate Scrubbing
A potassium permanganate scrubbing process has been used.to reduce NO emissions from
1800 ppm to 49 ppm at a nitric acid concentration,plant in Japan. The process reacts potassium
permanganate with nitrogen oxide and sodium hydroxide to form potassium sodium manganate, sodium
nitri'te, and potassium nitrite. The potassium permanganate is regenerated by oxidizing the
potassium sodium manganate electrolytically (References 3-97 and 3-98).
3.3.2.4 Catalytic Reduction
There are three types of catalytic reduction processes used for NO control: nonselective
reduction, which removes both NOV and oxygen; selective reduction, which removes only NOV; and
X A
heterogeneous catalysis used in conjunction with wet scrubbing. Each of these will be discussed in
the following paragraphs.
Nonselective Catalytic Reduction
The nonselective reduction process reacts NO with Hp or CH^ to yield Ng, COp and HpO. The
process is called nonselective because the reactants first deplete all the oxygen present in the
3-50
-------
tail gas, and then remove the NO . Prior to the large increases in natural gas prices the excess
fuel required to reduce the oxygen did not impose a heavy economic penalty. The reactions were
exothermic, and much of the heat could be recovered with a waste heat boiler.
The nonselective reduction process is used for decolorization and energy recovery, as well as
for NO abatement. Decolorization and power recovery units reduce NOx and NO and react part of the
oxygen, but their capacity to reduce NO to elemental nitrogen is limited. The nonselective'
abatement units carry the process through to NO reduction as well. In nonselective reduction, the
tail gases from the absorber are heated to the necessary catalyst ignition temperature, mixed with a
reducing agent, such as hydrogen or natural gas, and passed into the reactor and through the
catalyst. The main chemical reactions that take place are:
CH4 + 4N02 -> 4NO + C02 + 2H20 (3-8)
CH4 + 202 + C02 + 2H20 (3-9)
CH4 + 4NO - 2N2 + C02 + 2H20 . (3-10)
Similar equations can be written substituting hydrogen for methane, in which case two moles
of hydrogen are needed to replace one mole of methane. The reaction kinetics are such that
reduction reaction (3-8) is faster than reduction reaction (3-9), but abatement reaction (3-10) is
much slower than reaction (3-9). Thus, decolorization can be accomplished by adding just enough
fuel for partial oxygen burnout. If NO abatement is required, however, sufficient fuel must be
x N
added for complete oxygen burnout.
Both catalyst and nitric acid manufacturers report satisfactory performance for decoloriza-
tion units. The reduction-of total .NO., is limited, but ground-level NO, concentration in critical
A £
areas near the plant is reduced substantially.
NOX abatement using nonselective catalyst is more difficult technically than decolorization,
and commercial results have been less satisfactory. Provisions must be made to control the heat
released in reacting all the tailgas oxygen. The thermal control must be done before extensive NO
reduction proceeds.
3-51
-------
In Section 6 the success of the various types of catalytic abaters in coping with the
problems of temperature rise and high space velocities will be discussed. In general, nonselective
catalytic reduction is not likely to be used in the future for NO control. The availability and
cost of natural gas, increasing catalyst cost and poor performance have led to a decline in interest
1n this process.
Selective Catalytic Reduction
In selective catalytic reduction, ammonia is reacted with the NO to form N2- The oxygen in
the tall gas does not react with the ammonia, so stoichiometric amounts of ammonia are used.
In contrast to nonselective techniques, selective .catalyst abatement must be carried out
within the narrow temperature range of 483°K to 544°K (410°F to 520°F). Within these limits,
ammonia will reduce N02 and NO to molecular nitrogen, without simultaneously reacting with oxygen.
The overall reactions are shown in the following equations:
8NH3 + 6N02 * 7N2 + 12 HgO (3-11)
4NH3 + 6NO * 5N2 + 6H20 (3-12)
Above 544°K, ammonia may oxidize to form NOX; below 483°K, it may form ammonia nitrate.
Selective oxidation with ammonia has several advantages over nonselective reduction:
- The reducing agent, ammonia, is usually readily available since it is
consumed as feed stock in the nitric acid process;
- Temperature rise through the reactor bed is only 20°K to 30°K (36°F to
54°F) so that energy recovery equipment, such as a waste heat boiler or
high temperature gas turbine, is not required; and
- Lower raw material costs since the amount of ammonia required is
approximately equal to the molar equivalent amount of NOV abated.
X
Heterogeneous Catalysis
One wet scrubber process uses heterogeneous catalysis in a packed column to oxidize NO to N02
(References 3-99 and 3-100). This system is currently in the development stage.
3-52
-------
3.3.2.5 Molecular Sieve Adsorption
One method of NO control involves the adsorption of NO onto a solid followed by regenera-
tion of the adsorbent. Materials such as silica gel, alumina, charcoal, and commercial zeolites or
molecular sieves have been employed for this method. Molecular sieves have been found to be the
most effective medium for this method of control, since they adsorb N02 selectively. Special sieves
have been developed which incorporate a catalyst to simultaneously convert NO to N02. This process
operates best only when low concentrations of oxygen are present, which is true of most tail gas
streams. The abatement bed is usually provided with a dehydration section prior to contact with the
abatement sieve to improve overall performance.
The adsorbent bed is regenerated by thermally cycling the bed after it is loaded with N02-
The required regenerating gas is obtained by using a portion of the treated tail gas stream to
desorb the adsorbed N02 from the bed. This gas stream is then recycled to the nitric acid plant
absorption tower. No other liquid, solid or gaseous effluents are produced by this process.
Two plants using this system were in operation and had experienced difficulties. The process
has become unattractive for future installations because of the cost of the catalyst bed, the energy
cost of thermal cycling, and the operational difficulties of using a cycling adsorption process with
a steady state nitric acid plant.
3-53
-------
REFERENCES FOR SECTION 3
3-1 Lira, K.J., et al., "Environmental Assessment of Utility Boiler Combustion Modification NO
Controls: Wlume 1. Technical Results," EPA-600/7-80-075a, April 1980. pp. 4-1 to 4-56.
3-2 Zeldovich, J., "The Oxidation of Nitrogen in Combustion and Expolsions," Acta Physiochem
URSS, (Moscow), Vol. 21, p. 4, 1946 (as cited in Reference 3-1, Section 4~T
3-3 Bowman, C.T. and Seery, D.J., "Investigation of NO Formation Kinetics in Combustion
Processes: The Methane-Oxygen-Nitrogen Reaction," In Emissions from Continuous Combustion
Systems, Cornelius, W. and Agnew, W.6., eds., Plenum, 1972 (as cited in Reference 3-1,
Section 4).
3-4 Bartok, W., et al., "Basic Kinetic Studies and Modeling of NO Formation in Combustion
Processes," ftlCnT Symposium Series No. 126, Vol. 68, 1972 (as cited in Reference 3-1,
Section 4).
3-5 Halstead, C.J. and Munro, A.J.E., "the Sampling, Analysis, and Study of the Nitrogen
Oxides Formed in Natural Gas/Air Flames," Company Report, Shell Research, Egham, Surrey,
U.K., 1971, (as cited in Reference 3-1, Section 4).
3-6 Thompson, D., et al., "The Formation of Oxides of Nitrogen in a Combustion System,"
presented at th~e TUth National AIChE Meeting, Atlantic City, 1971, (as cited in
Reference 3-1, Section 4).
3-7 Lange, H.B., "NO Formation in Premixed Combustion: A Kinetics Model and Experimental
Data," presented at the 64th Annual AIChE Meeting, San Francisco, 1971 (as cited in
Reference 3-1, Section 4).
3-8 Sarofim, A.F. and Phol, J.H., "Kinetics of Nitric Oxide Formation in Premixed Laminar
Flames," 14th Symposium (International) on Combustion, The Combustion Institute,
Pittsburg, 1973 (as cited in Reference 3-1, Section 4).
3-9 Iverach, D., et al., "Formations of Nitric Oxide in Fuel-Lean and Fuel-Rich Flames,"
ibid., 1973 (as cfted in Reference 3-1, Section 4).
3-10 Wendt, J.O.L. and Ekmann, J.M., "Effect of Fuel Sulfur Species on Nitrogen Oxide Emissions
from Premixed Flames," Comb. Flame, Vol. 25, 1975 (as cited in Reference 3-1, Section 4). .
3-11 Malte, P.C. and Pratt, D.T., "Measurement of Atomic Oxygen and Nitrogen Oxides in Jet-
Stirred Combustion," 15th Symposium (International) on Combustion, The Combustion
Institute, Pittsburgh, 1975 (as cited in Reference 3-1, Section 4).
3-12 Mitchell, R.E. and Sarofim, A.F., "Nitrogen Oxide Formation in Laminar Methane Air
Diffusion Flames," presented at the Fall Meeting, Western States Section, The Combustion
Institute, Palo Alto, California, 1975 (as cited in Reference 3-1, Section 4).
3-13 Bowman, C.T., "Non-Equilibrium Radical Concentrations in Shock Initiated Methane
Oxidation," 15th Symposium (International) on Combustion, The Combustion Institute,
Pittsburg, 1975 (as cited in Reference 3-1, Section 4).
3-14 Fenimore, C.P., "Formation of Nitric Oxide in Premixed Hydrocarbon Flames," 13th Symposium
(International) on Combustion, The Combustion Institute, Pittsburgh, 1971 (as cited in
Reference 3-1, Section 4).
3-15 MacKinnon, D.J., "Nitric Oxide Formation at High Temperatures," Journal of the Air
Pollution Control Association, Vol. 24, No. 3, pp. 237 to 239, March 1974 (as cited in
Reference 3-1, Section 4).
3-16 Heap, M.P., e_t al_., "Burner Criteria for NO Control; Volume I ~ Influence of Burner
Variables on NO in Pulverized Coal Flames,8 EPA-600/2-76-061a, NTIS-PB-259-911/AS,
March 1976, (as cited in Reference 3-1, Section 4).
3-17 Bowman, C.T., ejt al_., "Effects of Interaction Between Fluid Dynamics on Chemistry or
Pollutant Formation in Combustion," In: Proceedings of the Stationary Source Combustion
Symposium; Volume I Fundamental Research, EPA-600/2-76-152a. NTIS-PB-256-320/AS.
June 1976, (as cited in Reference 3-1, Section 4).
3-54
-------
3-18 Shaw, J.T. and Thomas, A.C., "Oxides of Nitrogen in Relation to the Combustion of Coal,"
presented at the 7th International Conference on Coal Science, Prague, June 1968 (as cited
in Reference 3-1, Section 4).
3-19 Pershing, D.W., et al., "Influence of Design Variables on the Production of Thermal and
Fuel MO from Residual Oil and Coal Combustion," AIChE Symposium Series, No. 148, Vol. 71,
pp. 19 to 29, 1975 (as cited in Reference 3-1, Section 4).
3-20 Thompson, R.E. and McElroy, M.W., "Effectiveness of Gas Recirculation and Staged
Combustion in Reducing NO in a 560-MW Coal-Fired Boiler," EPRI FP-257, NTIS-PB-260-582,
September 1976 (as cited Sn Reference 3-1, Section 4).
3-21 Sorofim, A.F., et al., "Mechanisms and Kinetics of NO Formation: Recent Developments,"
presented at the 6lTfh Annual AIChE Meeting, Chicago, November 1976 (as cited in
Reference 3-1, Section 4).
3-22 Martin, 6.B. and Berkau, E.E., "An Investigation of the Conversion of Various Fuel
Nitrogen Compounds to Nitrogen Oxides in Oil Combustion," presented at the 70th National
AIChE Meeting, Atlantic City, August 1971 (as cited in Reference 3-1, Section 4).
3-23 Hablet, W.W. and Howe11, B.M., "Control of NO Formation in Tangentially Coal-Fired Steam
Generators," In: Proceedings of the NO Control Technology Seminar, EPRI SR-39,
NTIS-PB-253-661, February 1976 (as cited in Reference 3-1, Section 4).
3-24 "Air Quality and Stationary Source Emission Control," U.S. Senate, Committee on Public
Works, Serial No. 94-4, March 1975 (as cited in Reference 3-1, Section 4).
3-25 Pohl, J.H. and Sarofim, A.F., "Fate of Coal Nitrogen During Pyrolysis and Oxidation," In:
Proceedings of the Stationary Source Combustion Symposium; Volume I -- Fundamental
Research, EPA-600/2-76-152a, NTI5-PB-256-320/AS. June 1976 (as cited in Reference 3-1',
Section 4).
3-26 Heap, M.P., et jil_., "The Optimization of Burner Design Parameters to Control NO Formation
in PulverizecTCoal and Heavy Oil Flames," In: Proceedings of the Stationary Source
Combustion Symposium; Volume II Fuels and Process Research and Development,
EPA-600/2-76-152b, NTIS-PB 256 321/AS, June 1976 Reference 3-1, Section 4).
3-27 Pohl, J.H. and Sarofim, A.F., "Devolatilization and Oxidation of Coal Nitrogen," presented
at the 16th Sympsoium (international) on Combustion, Cambridge, Massachusetts, August 1976
(as cited in Reference 3-1, Section 4).
3-28 Blair, D.W., et^ al_., "Devolatilization and Pyrolysis of Fuel Nitrogen from Single Coal
Particle Combustion," 16th Symposium (International) on Combustion, Cambridge,
Massachusetts, August 1976 (as cited in Reference 3-1, Section 4).
3-29 Brown, R.A., et al., "Investigation of Staging Parameters for NO Control in Both Wall and
Tangentially UoaT^fired Boilers," In: Proceedings of the Second Stationary Source
Combustion Symposium: Volume III. New Orleans, EPA-600/7-77-073C, NTIS-PB-271-757/AS,
July 1977 (as cited in Reference 3-1, Section 4).
3-30 Pershing, D.W., "Nitrogen Oxide Formation in Pulverized Coal Flames," Ph.D, Dissertation,
University of Arizona, 1976 (as cited in Reference 3-1, Section 4).
3-31 Axworthy, A.E., Jr., "Chemistry and Kinetics of Fuel Nitrogen Conversion to Nitric Oxide,"
AIChE Symposium Series. No. 148, Vol. 71, pp. 43 to 50, 1975 (as cited in Reference 3-1,
Section 4).
3-32 Axworthy, A.E., e_t jil_., "Chemical Reactions in the Conversion of Fuel Nitrogen to NO ,"
In: Proceedings of the Stationary Source Combustion Symposium, Volume I, x
EPA-600/2-76-152a, NTIS-PB-256-320/AS, June 1976 (as cited in Reference 3-1, Section 4).
3-33 Pershing, D.W. and Wendt, J.O.L., "The Effect of Coal Combustion on Thermal and Fuel NO
Production from Pulverized Coal Combustion," presented at Central States Section, The
Combustion Institute, Columbus, Ohio, April 1976 (as cited in Reference 3-1, Section 4).
3-55
-------
3-34 Habelt, W.W. and B.H. Howell, "Control of NO Formation in Tangentially Coal-Fired Steam
Generators," In: Proceedings of the NO Control Technology Seminar, EPRI SR-39,
February 1976. ;*'
3-35 Barr, W.H., and D.E. James, "Nitric Oxide Control - A Program of Significant Accomplish-
ments," ASME 72-WA/Pwr-13.
3-36 Barr, W.H., et^l_., "Retrofit of Large Utility Boilers for Nitric Oxide Emissions
Reduction - "Experience and Status Report."
3-37 Crawford, A.R., et a±., "Field Testing: Application of Combustion Modifications to
Control NO Emissions from Utility Boilers-," Exxon Research and Engineering Co.,
EPA-650/2-74-066, June 1974.
3-38 Crawford, A.R., et £l_l, "The Effect of Combustion Modification on Pollutants and Equipment
Performance of Power Generation Equipment," Exxon Research and Engineering Co.,
EPA-600/2-76-152c, prepared for the Stationary Source Combustion Symposium,
September 24-26, 1975.
3-39 Blakeslee, C.E., and H.E. Burbach, "Controlling NO Emissions from Steam Generators,"
C.E. Inc., APCA 72-75, 65th Annual Meeting of the Air Pollution Control Association,
June 18-22, 1972.
3-40 Norton, D.M., et al., "Status of Oil-Fired NO Control Technology," In: Proceedings of
the NO ControTTic'hnology Seminar, EPRI SR-3$, NTIS-PB 253 661, February 1976 (as cited
1n Reference 3-1, Section 4).
3-41 Campobenedetto, E.J., Babcock & Wilcock Co., letter to Acurex Corp., November 15, 1977 (as
cited in Reference 3-1, Section 4).
3-42 Durrant, O.W., "Pulverized Coal New Requirements and Challenges," presented to ISA
Power Instrumentation Symposium, Houston, Texas, May 1975 (as cited in Reference 3-1,
Section 4).
3-43 Vatsky, J., "Attaining Low NO Emissions by Combining Low Emission Burners and Off-
Stoichiometric Firing," presented at the 70th Annual AIChE Meeting, New York,
November 1977 (as cited in Reference 3-1, Section 4).
3-44 Rawdon, A.H. and Johnson, S.A., "Control of NO Emissions from Power Boilers," presented
at Annual Meeting of the Institute of Fuel, Adelaide, Australia, November 1974 (as cited
in Reference 3-1, Section 4).
3-45 Tidona, R.J., W.A. Carter, and H.J. Buening, (KVB) "Application of Advanced Combustion
Modifications to Industrial Process Equipment Process Heater Tests" (Preliminary draft
report prepared for Office of Research and Development, U.S. Environmental Protection
Agency.) Research Triangle Park, North Carolina. EPA Contract No. 68-02-2645.
November 1981.
3-46 Campobenedetto, E.J., "The Dual Register Pulverized Coal Burner Field Test Results,"
presented to Engineering Foundation Conference on Clean Combustion of Coal, New Hampshire,
August 1977 (as cited in Reference 3-1, Section 4).
3-47 Barr, W.H., et aj.., "Modifying Large Boilers to Reduce Nitric Oxide Emissions," Chemical
Engineering "Progress, Vol. 73, pp. 59 to 68, July 1977 (as cited in Reference 3-1,
Section 4).
3-48 U.S. Environmental Protection Agency. Control Techniques for Nitrogen Oxides Emissions
from Stationary Sources Second Edition. (Prepared for* U.S. Environmental Protection
Agency.) Research Triangle Park, N.C. Publication No. EPA-450/1-78-001. January 1978.
pp. 3-19 (as cited in Reference 3-1, Section 4).
3-49 Brown, R.A., H.B. Mason, and R.J. Schrfeber, "Systems Analysis Requirements for Nitrogen
Oxide Control of Stationary Sources," Environmental Protection Technology Series
EPA-650/2-74-091, September 1974.
3-56
-------
3-50 U.S. Environmental Protection Agency, "Draft - NO SIP Preparation Manual Volume II -
Support Sections," Office of Air Quality Planning and Standards, Research Triangle
Park, N.C., April, 1976.
3-51 Bell, A.W., et a\_., "Nitric Oxide Reduction by Controlled Combustion Processes,"
KVB, Inc., Western States Section/Combustion Institute, April 20-21, 1970 (as cited in
Reference 3-1, Section 4).
3-52 Crawford, A.R., E.H. Manny and U. Bartok, "Field Testing: Application of. Combustion
Modifications to Power Generating Combustion Sources," In: Proceedings of the Second
Stationary Source Combustion Symposium. Volume II, Utility and Large Industrial Boilers,
EPA-600/7-77-073b."
3-53 Ando, J. and T. Heiichiro, "NO Abatement for Stationary Sources in Japan," Environmental
Protection Technology Series, £PA-600/2-76-013b, January 1976.
3-54" Shoffstall, D.R., "Burner Design Criteria for Control of Pollutant Emissions from Natural
Gas Flames," Institute of Gas Technology, EPA-600/2-76-152b, June 1976.
3-55 Koppang, R.R., "A Status Report on the Commercialization and Recent Development History of
the TRW Low NOV Burner," TRW Energy Systems Group.
X
3-56 Tsuji, S., et al., "Control Technique for Nitric Oxide - Development of New Combustion
Methods," IHI Engineering Review, Vol. 6, No. 2.
3-57 Ando, J., et al., "NOV Abatement for Stationary Sources in Japan," August 1976
(Preliminary Draft). x
3-58
Heap, M.P., et al., "The Optimization of Burner Design Parameters to Control NO Formation
in PulverizecTCoal and Heavy Oil Flames," In: Proceedings of the Stationary Source
Combustion Symposium, EPA-600/2~76-152b, June 1976.
3-59 Barsin, J.A., "Pulverized Coal Firing NO Control," In: Proceedings: Second NO Control
Technology Seminar, Electric Power Research Institute, Report No. FP-1109-SR, Pafo Alto,
California, July 1979 (as cited in Reference 3-1, Section 4).
3-60 Peduto, E.F., R.R. Hall, and G. Tucker, "Characterization of the NO and S02 Control
Performances: Southern Indiana Gas and Electric Co., A.B. Brown Unit No. 1, Volume 1 ~
Program Results," Prepared for U.S. Environmental Ptoection Agency, Washington, D.C., EPA
Contract No. 68-02-3168, March 1982.
3-61 Vatsky, J., "Experience in Reducing NO Emissions on Operating Steam Generators," In:
Proceedings: Second NO Control Technology Seminar, Electric Power Research Institute,
Report Nol FP-1109-SR, Palo Alto, CA, July 1979 (as cited in Reference 3-1, Section 4).
3-62 "NO Control Review," Vol. 2, No. 4, EPA Industrial Environmental Research Laboratory,
RTP, North Carolina, Fall 1977 (as cited in Reference 3-1, Section 4).
3-63 Stavern, D.V., "The Coen Low Excess Air Burner," presented at the NO Control Technology
Workshop, Pacific Grove, California, October 1977 (as cited in Reference 3-1, Section 4).
3-64 Vatsky, J., "Larger Burners and Low NO ," Heat Engineering, Vol. 49, No. 2, pp. 17-25,
April-June 1979 (as cited in Reference 3-1, Section 4).
3-65 Martin, G.B., "Field Evaluation of Low NO Coal Burners on Industrial and Utility
Boilers," In: Proceedings of the Third Stationary Source Combustion Symposium, Volume I,
EPA-600/7-79-050a, February 1979 (as cited in Reference 3-1, Section 4).
3-66 Johnson, S.A., et al., "The Primary Combustion Furnace System -- An Advanced Low-NO
Concept for Pulverized Coal Combustion," In: Proceedings: Second NO Control Technology
Seminar, Electric Power Research Institute, Report No. FP-1109-SR, PaTo Alto, California,
July 1979.
3-67 Whitaker, R., "Trade-offs in N0x Control," EPRI Journal, January-February 1982, pp. 18-25.
3-68 Habelt, W.W. The Influence of the Coal Oxygen to Coal Nitrogen Ratio on NO Formation.
Presented at the 70th Annual AIChE Meeting. New York. November 13-17, 1977.
3-57
-------
3-69 Ctvrtnicek, I.E., et aT_., "Evoluation of Low-Sulfur Western Coal Characteristics,
Utilization and Combustion Experience," Monsanto Research Corp., EPA-650/2-75-046,
Hay 1975.
3-70 Martin, 6.B., "Evaluation of NO Emission Characteristics of Alcohol Fuels in Stationary
Combustion Systems," Presented at Joint Meeting, Western and Central States Sections, The
Combustion Institute, April 21 and 22, 1975, San Antonio, Texas.
3-71 Hall, R.E., "The Effect of Water/Residual Oil Emulsions on Air Pollutant Emissions and
Efficiency of Commercial Boilers," ASME 75-WA/APC-l, July 14, 1975.
3-72 Martin, 6.B., "Environmental Considerations in the Use of Alternate Clean Fuels in
Stationary Combustion Processes."
3-73 Martin, 6.B., D.W. Pershing, E.E. Berkau, "Effects of Fuel Additives on Air Pollutant
Emissions from Distillate Oil-Fired Furnaces," EPA, Office of Air Programs, AP-87,
June 1971.
3-74 Shaw, H., "Reduction of Nitrogen Oxide Emissions from a Gas Turbine Combustor by Fuel
Modifications," ASME Transactions, Journal of Engineering for Power, 95, 4, October 1973.
3-75 Altwicker, E.R., et al., "Pollutants from Fuel Oil Combustion and the Effects of
Additives," Paper No7"71-14, 64th Annual APCA Meeting, Atlantic City, N.J., June 1971.
3-76 Barrett, R.E., et al., "Field Investigation of Emissions from Combustion Equipment for
Space Heating," EP/PR2-73-084a, June 1973.
3-77 Frey, D.J., "De-Ashed Coal Combustion Study," Combustion Engineering, Inc., October 1964.
3-78 Energy Research and Development Agency, "Proceedings of the Fourth International
Conference on Fluidized Bed Combustion," McLean, VA., December 1975.
3-79 Jonke, A.A., et a\_., "Pollution Control Capabilities, of Fluidized Bed Combustion," AIChE
Symposium Sertes No. 126, Vol. 68, 1972.
3-80 Chronowski, R.A., and B. Molayem, "NOV Emissions from Atmopsheric Fluidized-Bed Boilers,"
ASME-75-PWR-4, October 1975. x
3-81 Pfefferle, W.C., e£ al., "CATATHERMAL Combustion: A New Process for Low-Emissions Fuel
Conversion," presented at the 1975 ASME Winter Annual Meeting, Houston, Texas, ASME Paper
No. 75-WA/FU-l.
3-82 Kesselring, J.P., et al., "Catalytic Oxidation of Fuels for NOV Control for Area Sources,"
EPA Report, EPA-60^72^76-037, February 1976. x
3-83 DeCorso, S.M., et al., "Catalysts for Gas Turbine Combustors - Experimental Test Results,"
paper presentecTaY~A"SME Gas Turbine Conference and Products Show, Mew Orleans, March 1976,
ASME Paper I76-GT-4.
3-84 Gerstin, R.A., "A Technical and Economic Overview of the Benefits of Repowering," paper
presented at the Gas Turbine Conference and Products Show, Houston, Texas, March 2-6,
1975, ASME Paper #75-GT-16.
3-85 Ahuja, A., "Repowering Pays Off for Utility and Industrial Plants," Power Engineering,
pp. 50-54, July 1976.
3-86 Stambler, I., "Repowering Gives Glendale Extra 75 MW and Lower Rates," Gas Turbine World,
September 1977.
3-87 Robson, F.L., and A.J. Giramonti, "The Use of Combined-Cycle Power Systems in Nonpolluting
Central Stations," JAPCA, Vol. 22, pp. 177-180, 1972.
3-88 Jones, G.D. and K.L. Johnson, "Technology Assessment Report for Industrial Boiler
Applications: NOX Flue Gas Treatment," EPA-600/7-79-178g, December 1979.
3-89 Maxwell, J.D. and L.R. Humphries, "Evaluation of the Advanced Low-NO Burner, Exxon, and
Hitachi Zosen DeNOx Processes," EPA-600/7-81-120, TVA/OP/EDT-81-28, 3u1y 1981.
3-58
-------
3-90 Ando, J., "NOX Abatement for Stationary Sources in Japan," EPA-600/7-79-205, August 1979.
3-91 Faucett, H.L., J.D. Maxwell, ,and T.A. Burnett, "Technical Assessment of NO Removal
Processes for Utility Application," TVA Bulletin Y-120, EPA-600/7-77-127 (NTIS
PB-276-637/6WP), EPRI FP-1253, 1977.
3-92 Jones, G.D. "Selective Catalytic Reduction and NO Control in Japan, A Status Report,"
EPA-600/7-81-030, January 1981. x
3-93 Trexler, E.C. "DOE's Electron Bream Irradiation Developmental Program," prepared for
Seventh Symposium on Flue Gas Desulfurization, May 17-20, 1982, p. 11.
3-94 Reference 3-89, p. 11.
3-95 "Nitric Acid from Ammonia," Hoechst-Uhde Corp. brochure (FWC 11 619),
Englewood Cliffs, N.J.
3-96 Mayland, B.J., "The CDL/VITOK Nitrogen Oxides Abatement Process," Chenoweth Development
Laboratory, Louisville, KY.
3-97 "New System Knocks NOX Out of Nitric," Chemical Week, September 3, 1975, pp. 37-38.
3-98 "NO Removal System Now Available," Wet Scrubber Newsletter, September 30, 1973, pp. 3-4.
3-99 Mayland, B.J., "Application of the CDL/VITOK Nitrogen Oxide Abatement Process," presented
at Sulfur and Nitrogen Symposium, Salford, Lancashire, U.K., April 1976.
3-100 Mayland, B.J., and R.C. Heinze, "Continuous Catalytic Absorption for NO Emission
Control," Chemical Engineering Process, Vol. 6, May 1973, pp. 75-76.
3-59
-------
-------
SECTION 4
LARGE FOSSIL FUEL COMBUSTION PROCESSES
Fossil fuel combustion by utilities and industry accounted for about 85 percent of NO
emissions from stationary sources in 1980 (see Section 2). This combustion occurred principally in
s
boilers and large internal combustion (1C) engines. The large boiler category encompasses applica-
Jion to utility and industrial power generation and industrial process steam generation. Large 1C
engines are used predominantly for power generation and for pipeline pumping and encompass large
bore reciprocating engines as well as continuous combustion gas turbine engines. This section
summarizes the effectiveness, cost, user experience, and energy and environmental impacts of the
implementation of NO controls on these equipment categories.
4.1 UTILITY BOILERS
Most of the nation's electricity is generated in large fossil-fueled central station power
plants, which primarily consist of high-pressure watertube boilers in the 100 to 1300 MW* range
serving turbine-generators. Firing capacities of individual burners in utility boilers commonly
have thermal inputs as high as 29-58 MW* (100-200 x 106 Btu/hr). A 1000 MW* opposed wall-fired unit
may require as many as 60 separate burners.
Although there are differences among utility boiler designs in furnace volume, operating
pressure, and configuration of internal heating transfer surface, the principal distinction is
firing mode. Firing mode is characterized by the type of firing equipment, the fuel handling
system, and the placement of the burners on the furnace walls (see Section 2.3.1).
The total NOX emitted in 1980 by the electric utility industry was 6.4 Tg (7.1 x 106 tons) or
58 percent of the total stationary source emissions of NO . Coal-fired boilers accounted for
approximately 81 percent of the total utility emissions. A more detailed emission breakdown is
presented in Section 2. Ranges of .uncontrolled NO emissions for three principal types of coal-
X
fired boilers are presented in Figure 4-1. These emission rates essentially represent baseline
*Utility boilers are commonly described in terms of electrical output rating rather than in terms of
thermal input. Burners are commonly described in terms of thermal input to the boiler as noted by
the subscript "t". This convention will be used throughout Section 4.1 unless otherwise noted in
the text.
4-1
-------
1600
1400
1200
6 1000
Cu
a.
CM
o
en
O
z
800
600
400
200
0
I I
Wet bottom
Wall
I
200 .400 600
MW
800
1000
Figure 4-1. Baseline HO emissions - coal-fired utility
boilers (Reference 4-1).
4-2
-------
emissions from units designed prior to the early 1970s, without the combustion modification controls
often designed into later units. As shown, cyclone-fired boilers typically have the highest
uncontrolled emission rates, tangentially-fired units the lowest, with one wall- and opposed
wall-fired units at an intermediate level. The old wet bottom (slag tap) units are exceptions and
have very high NO emissions. All of these boilers fire pulverized coal except for the cyclone
units which fire crushed coal. Very few stokers are used in utility applications.
4.1.1 Control Techniques
Techniques for controlling NOX emissions from utility boilers fall into two broad
classes; combustion modifications and flue gas treatment (F6T). Considerable experience exists in
the U.S. with various combustion modification controls. Flue gas treatment techniques, on the other
hand, have primarily been developed and applied in Japan, with some testing performed in the U.S.
This section discusses the status of application of these techniques to utility boilers, the NO
control performance achieved, and the process impacts associated with their application.
4.1.1.1 Combustion Modifications
The general concept of combustion modification as potential NO control techniques for
stationary sources was discussed in Section 3.1. These techniques have been developed and refined
in numerous laboratory test installations and in many successful field applications to commercial.
utility boilers.
Utility boilers, due to their importance as NO sources and their control flexibility, are
the most extensively modified stationary equipment type. The selection and implementation of
effective NOV controls for a specific utility boiler are generally dependent on several variables.
X
These include the furnace characteristics (i.e., geometry and operational flexibility), fuel/air
handling systems and automatic controls, and the potential for operational problems which may result
from combustion modifications. The following discussion is, therefore, not intended to provide
application guidelines, but rather to give a broad overview and evaluation of tested procedures.
Combustion modifications which have been applied or proposed for utility boilers include:
- Low Excess Air Firing (LEA);
- Off Stoichiometric Combustion (OSC) or Staged Combustion (SC),
including Biased Burner Firing (BBF), Burners Out of Service (BOOS),
and Overfire Air (OFA);
- Flue Gas Recirculation (FGR);
- First Generation Low NOX Burners (LNB);
- Enlarged Furnace Design/Reduced Firing Rate;
4-3
-------
- Reduced Air Preheat (RAP);
- Water Injection (WI); and
- Advanced Burner/Furnace Designs.
Some of these techniques, such as enlarged furnace designs, first generation low NOV burners, OFA,
X r-
and LEA, have become a standard part of new unit design. Techniques such as LEA, F6R, and various
methods of OSC have also been extensively applied to existing units. Other techniques have not
gained wide acceptance because of adverse impacts on unit efficiency or capacity. These include
reduced firing rate, RAP, and WI. Also, various OSC techniques may result in reduced unit capacity
1n some retrofit applications. Advanced burner and furnace designs represent techniques which are
stm under development but may become available in the next few years.
Retrofit NO control implementation by combustion modification usually proceeds in several
stages depending on the emission limits to be reached. First, fine tuning of combustion conditions
by lowering excess air and adjusting the burner settings and air distribution is employed. If NO
emission levels are still too high, the minor modifications, such as BBF or BOOS are implemented.
If further reductions of NO are necessary, these minor modifications are followed by the more major
X
retrofits, Including OFA ports, FGR systems, and new burners.
The feasibility, effectiveness, and method of applying the modifications within each stage of
control depend heavily on the fuel and firing type. For example, testing has shown that FGR does
not significantly reduce fuel NOV, so this technique is usually not cost-effective for NOV control
A X
on coal-fired units. Also, such techniques as BOOS or OFA are implemented differently on wall-fired
and tangentially-fired units due to burner configuration and hardware differences. Tables 4-1
through 4-8 summarize available test data by fuel and firing type for various combustion modifica-
tions applied to utility boilers. These data were obtained from Reference 4-2 and the reader is
referred to this reference for further information about these tests.
The practical limits on combustion modifications are based initially on three subjective
criteria: emission of other pollutants (i.e., CO, smoke, 'and carbon in fly ash), onset of slagging
or fouling, and incipience ot flanie instability at the burner.
The remainder of this section describes recent combustion modification experience on coal-,
oil-, and gas-fired boilers. The material in this section was taken from Reference 4-3. The reader
1s referred to this document for additional information and for the original sources of this
material.
4-4
-------
TABLE 4-1. AVERAGE NO REDUCTION WITH LOW EXCESS AIR FIRING (LEA)a
x (Reference 4-2)
Equipment
Type
Tangential
Opposed Hall
Single
Hall
All Boilers*
Fuel
Coal
Oil
Nat
Gas
Coal
Oil
Nat
Gas*
Coal
Oil
Nat
Gas
Coal
Oil
Nat
Gas
All
Fuels
Number
of
Boilers
Tested
11
~
1
5
4
6
7
(2)
4
3
(1)
23
8
10
41
Baseline
Stoichionetry
to Active Burners
(percent)
124
-'
117
126
120
115
123 u
(134)"
120
117
(124)
. 124
120
116
120'
NOX Emissions
(ppm dry Q 3X 02)
459
-- -
340
746
357
717
624
(1338)
409
418
(992)
609
383
492
495
Low Excess Air (LEA)
Stoichiometry
to Active Burners
(percent)
116
-
113
US
113
110
114
(118)
112
108
(112)
116
115
110
114
NO Emissions
[ppm dry 9 38 03)
373
" --
245
660
290
600
522
(1325)
315
356
(931)
522
302
400
408
Average
NO. Reduction
(percent)
19
..,- 28
12
19
16
16
(1)
23
15
(6)
16
21
20
19
Maximum NOX
Reduction
Reported
(percent)
42 .
-
28
23
30
33
25
(3)
26
15
(6)
30
28
25
28
I
en
jjBoiler load at or above 80 percent HCR^ For individual tests, corresponding baseline and controlled loads were nearly identical.
"Numbers in parentheses refer to boilers originally designed for coal firing with wet bottom furnaces.
-------
TABLE 4-2. AVERAGE NO REDUCTION WITH BURNER OUT OF SERVICE (BOOS)3
(Reference 4-2)
en
E
16
<24)<=
16
(12-24)l»
K .
(12-16)>»
28
(16-56)0
20 h
(12-24)"
16
(6-36)0
20
(a-ffijb
StotchlOMtry
to Active
Burners
(percent)
121
-
112
122
107
US
123
(134)<=
119
11?
122
113
115
117
NO,
Emissions
(pp* dry
»M02)
462
--
146
670
442
674
610
(1196JC
425
A\Q
S83
433
412
4/6
lurners Out of Service (MIOS)
Percent
Burners
on Air
Only
17
--
HA
16
33
28
19
<33)C
ia
22
17
25
25
n
Stolchlowtry
to Active
Burners
(percent)
98
--
06
102
73
84
97
(a9)<=
95
as
99
84
86
90
HOX
Caisslons
(ppmdnr
»3X02)
293
--
146
522
292
290
412
(577 )C
256
214
409
274
217
300
Average NO.
Reduction
(percent)
37
--
0
22
34
57
33
{52)£
40
4$
31
37
35
34
Maxima HO.
Reduction
Reported
(percent)
56
--
0
46
34
61
48
(52)C
48
S3
SO
41
43
45
'Boiler load at or above 80 percent HCR. For individual tests, turrespondlnu, baseline and controlled loads were nearly identical.
"Range in number of burners firing
Numbers in parentheses refer to boilers originally designed for coal firms with *et bottom furnaces.
-------
TABLE 4-3. AVERAGE NO - REDUCTION WITH OVERFIRE AIR (OFA)a
. (Reference 4-2)
Equipment
Type
Tangential
Opposed Wall
Single
Wall
Fuel
Coal
Oil
Nat
Gas
Coal
Oil
Nat
Gas
Coal
Oil
Nat
Gas
Number
of
Boilers
Tested
6
~
* 5
2
-
« ,
~
Baseline
Stoichiometry
to Active
Burners
(percent)
129
~
--
~
118
114
..
-
N°x
Emissions
(ppra dry
0 3X 02)
454
~
-
376
928
--
-
-
Over fire Air (OFA)
Stoichiometry
to Active
Burners
(percent)
105
---
~
96
99
--
-
Furnace
Stoichiometry
(percent)
122
118
112
--
NOX Emissions
(ppm dry
3 3X 02)
311
"
287
378
-
Average NOX
Reduction
(percent)
31
~
~
--
24
59
--
--
--
Maximum NOX
Reduction
Reported
(percent)
41
-
--
30
66
--
-
-
aBoi1er load at or above 80 percent NCR. For individual tests, corresponding baseline and controlled loads were nearly identical.
-------
TABLE 4-4. AVERAGE NO REDUCTION WITH FLUE GAS RECIRCULATION (FGR)a
(Reference 4-2)
I
00
Equipment
Type
Tangential
Opposed Wall
Single
Hall
Fuel
Coal
Oil
Nat
Gas
Coal
Oil
Nat
Gas
Coal
Oil
Nat
Gas
Number
of
Boilers
Tested
1
* 1
1
«
1
Baseline
Stoichionetry
to Active
Burners
(percent)
-
117
128
122
~
~
106
>»x
Emissions
(ppa dry
@ 3* 02)
-
340
855
304
-
470
Overfire Air (Of A)
Stoichionetry
to Active
Burners
(percent)
-
115
127
126
107
FGH
(percent)
--
23
15
11
--
11
NOX Emissions
(ppra dry
3X 02)
135
735
263
. ~
307
Average NOX
Reduction
(percent)
..
60
17
13
--
35
Maximum I(0X
Reduction
Reported
(percent)
60
17
13
- ~
35
afloi)er load at or above SO percent HCR. For individual tests, corresponding baseline and controlled loads were nearly identical.
-------
TABLE 4-5. AVERAGE NO REDUCTION WITH REDUCED FIRING RATE0
(Reference 4-2)
Equipment
Type
Tangential
Opposed Wall
Single
Hall
All
Boilers
Fuel
Coal
Oil
Nat
Gas
Coal
Oil
Nat
Gas
Coal
on
Nat
Gas
Coal
011
Nat
Gas
All
Fuels
Number
of
Boilers
Vested
7
1
4
4
5
2
(2)
3
2
(1)
13
7
8
28
Baseline (BOX HCR or Above)
Firing
Rate
(percent
HCR)
93
100
93
98
98
92
(90)
98
97
(98)
93
98
99
97
Stolchiometry
to Active
Burners
(percent)
112
117
131
118
115
125
(133)»
119
118
(115)
126
119
117
120
NOX
Emissions
(ppra dry
9 3* QZ)
462
340
825
362
651
651
(1338)
425
442
(992)
646
393
478
506
Reduced Load
Firing
Rate
(percent
HCR)
64
75
70
61
57
67
(54)
53
35
(59)
67
57
55
60
Stolchlometry
to Active
Burners
(percent)
127
135
136
121
US
130
(138)
119
117
(131)
131
120
122
124
NO),
Emissions
(ppm dry
9 3t 02)
408
..
332
758
249
269
496
(990)
296
125
(522)
554
272
242
356
Average NQX
Reduction
(percent)
12
2
8
31
59
24
(26)
30
72
(47)
14
31
44
30
Maximum NOX
Reduction
Reported
(percent)
25
-
32
18
48
64
25
(33)
45
82
(47)
23
47
59
43
'Numbers 1n parentheses refer to boilers originally designed for coal firing with wet bottom furnaces.
-------
TABLE 4-6
AND
. AVERAGE N0₯ REDUCTION WITH OFF STOICHIOMETRIC COMBUSTION
FLUE GAS RECALCULATION (OSC AND FGR)a (Reference 4-2)
Equipment
Type
Tangential
Opposed Hall
Single
Wall
Fuel
Coal
Oil
Nat
Gas
Coal
Oil
Nat
Gas
Coal
Oil
Nat
Gas
Number
of
Boilers
Tested
1
1
1
-
2
i
Baseline
Stolchlometry
to Active
Burners
(percent)
-
117
128
122
-
-
.118
106
HOX
Emissions
(ppm dry
9 3X Oz)
.-
340
781
304
-
355
470
OSC and FGft
Type of
OSC
BOOS
BOOS
OFA
-
BBF
BOOS
BOOS
Stoichiowetry
to Active
Burners
(percent)
~
-
75
99
97
~
~
91
75
FOR
(percent)
..
21
19
11
-
14
12
KOX
Emissions
(ppm dry
9 3X 02)
-
105
453
247
--
154
115
Average HOX
Reduction
(percent)
~
-
69
42
19
--
--
57
76
Maximum NOX .
Reduction
Reported
(percent)
--
69
42
19
59
76
toiler load at or above 80 percent MCR. For individual tests, corresponding baseline and controlled loads were nearly identical.
-------
TABLE 4-7. AVERAGE NO REDUCTION WITH REDUCED FIRING RATE AND
OFF STOICHIOMETRIC COMBUSTION (Reference 4-2)
Equipment
Type
Tangential
Horizontally
Opposed Wall
Single
Wall
All
Boilers
Fuels
Fuel
Coal
011
Nat
Gas
Coal
Oil
Nat
Gas
Coal
011
Nat
Gas
Coal
011
Nat
Gas
All
Number
of
Boilers
Tested
8
3
4
6
4
(2)
3
2
(1)
15
7
8
30
Baseline
Firing
Rate
(percent
NCR)1
93
-
-
93
99
100
90
98
97
(98)
92
99
99
97
Stoichlometry
to Active
Burners
(percent)
122
~
--
129
118
115
124
(133)4
120
118
(125)
125
119
117
120
NOX
Emissions
(ppro dry
» 3% 02)
453
-
820
362
717
663
(1338)
426
442
(992)
645
.394
579
b39
Low Load and OSC
Firing
Rate
(percent
NCR)
61
~
73
64
58
73
(59)'
56
35
(71)
69
60
31
bJ
Type of
OSC
BOOS
OFA
--
BOOS
BOOS
OFA
BOOS
OFA
BOOS
BBF
BOOS
BBF
BOOS
BOOS
OFA
BBF
BOOS
OFA
BOOS
OFA
DBF
noos
OFA
Stolen iometry
to Active
Burners
(percent)
95
-
102
117
88
99
(91)
97
93
(102)
99
107
91
99
1
NO.
Emissions
(ppra dry
9 3X02)
248
~
634
177
148
381
(887)
228
78
(641)
421
202
- 113
245
Average HOX
Reduction
(percent)
45
~
-
23
51
79
43
(34)
46
82
(35)
37
49
80
S5
Maximum
NOX
Reduction
Reported
(percent)
62
-
--
32
67
89
50
(55)
59
87
(351. '
48
63
8S
66
'Numbers in parentheses refer to boilers originally desigmn) for coal firing with wot .but turn furnaces.
-------
TABLE 4-8. AVERAGE NO REDUCTION WITH LOAD REDUCTION, OFF STOICHIOMETRIC
COMBUSTION AND FLUE GAS RECIRCULATION (Reference 4-2)
Equipment
Type
Tangential
Opposed Hall
Single
Hall
All
Boilers
Fuel
Coal
Oil
Nat
Gas
Coal
Oil
Nat
Gas
Coal
Oil
Nat
Sas
Coal
Oil
Nat
Gas
All
Fuels
Hunter
of
Boilers
Tested
-
--
--
-
3
2
--
2
2
5
4
9
Baseline
Firing
Rate
(percent
NCR)
...
~
99
100
--
98
100
99
100
100
Stoichlonetry
to Active
Burners
(percent)
«
118
113 .
--
118
110
118
112
115
HOX
Emissions
(pp* dry
9 3* 02)
--
~
398
945
--
355
421
~
376
683
530
Controlled/Low Load and OSC and FGR
Firing
Rate
(percent
HCR)
~
46
43
62
65
54
54
54
Type of
OSC
--
-
BOOS
OFA
BOOS
OFA
BOOS
BOOS
-
BOOS
OFA
BOOS
OFA
BOOS
OFA
Stoichioroetry
to Active
Burners
(percent)
--
87
90
92
81
90
86
88
FGR
(percent)
--
-
-
--
39
27
30
20
-
35
23
29
NOX
Emissions
(pp* dry
9 3% 02)
.. .
--
194
130
--
152
171
173
ISO
162
Average HOX
Reduction
(percent)
-
56
87
57
59
57
73
66
Maximum HOX
Reduction
Reported
(percent)
-
--
59
90
-
57
83
58
87
73
-------
Application to Coal-Fired Boilers . . _
Currently the most commonly applied low NO technique for coal-fired boilers is off stoichio-
X
metric or staged combustion through the introduction of oyerfire air (OFA). This technique has been
used in both retrofit and new unit applications. Application of burners out of service (BOOS), an
alternate staging technique with retrofit application, is limited because it is often accompanied by
a 10 to 25 percent load reduction. Average NO reductions of 30 to 50 percent from the levels shown
in Figure 4-1 (controlled emissions of 210 to 300 ng/J, 0.5 to 0.7 lb/10 Btu) can be expected with
either technique in dry bottom boilers. Flue gas recirculation (FGR) has been, tested but was found
to be a relatively ineffective control for coal firing, giving only about 15 percent NO reductions.
A
More recently, new low NO burners (LNB) have been installed on some units and found to be at least
as effective as OFA. In fact,- new wall-fired units currently being brought online generally rely on
LNB to meet the 1971 New Source Performance Standard (NSPS). The combination of OFA with LNB has
resulted in 40 to 60 percent NO ' reductions (controlled emissions of 170 to 260 ng/J, 0.4 to
A
0.6 lb/106 Btu).
There has been a steady improvement in combustion modification control technology over recent
years. Current demonstrated technology is capable of 40 to 60 percent NO reductions from the
X
levels shown in Figure 4-1, easily meeting the 1971 NSPS of 300 ng/J (0.7 lb/106 Btu) and in most
cases the 1979 revised standard of 210 to 260 ng/J (0.5 to 0.6 lb/10 Btu). Current R&D programs,
such as the development and demonstration of the EPA advanced low NO burner and the EPRI primary
combustion furnace, should result in combustion modification techniques capable of meeting
NOX emission levels of 86 to 130 ng/J (0.2 to 0.3 lb/106 Btu).
The effects of low NO operation on coal-fired boilers are summarized in Table 4-9. The
»
table describes experience obtained through field test programs on the several units listed. The
major concerns regarding low NO operation on coal-fired boilers have been the possible adverse
effects on boiler efficiency, load capacity, water wall tube corrosion and slagging, carbon loss,
heat absorption profile, and convective section tube and steam temperatures.
In most past experience with staged combustion, optimal excess air levels have been
comparable to levels used under baseline conditions. In these cases the efficiency of the boiler
remains unaffected if unburned carbon losses do not increase appreciably. However, in some cases
when, due to nonuniform fuel air distribution or other causes, the excess air requirement*increases.
under staged firing, a significant decrease in efficiency occurs. From Table 4-9, it is seen that
efficiency decreases of up to 1 percent have been experienced. It is also seen that the same boiler
4-13
-------
TABLE 4-9. EFFECT OF LOW NO OPERATION ON COAL-FIRED BORERS*
(Reference 4-3)
Boiler
Tanqentlai
Barry No. 2
Columbia
No. 1
Huntington
Canyon No. 2
Barry No. 4
Navajo No. 2
Ccaanche No. 1
Kingston No. £
Opposed Uall
Harllce Branch
No. 3
i
Four Corners
He «
Low NOX
Technique
BOOS
OFA
OFA
OFA
LEA. BOOS
LEA. BOOS. OFA
OFA
BBF
BOOS
LEA, BOOS
LEA. BOOS
Boiler
Efficiency
Unaffected
Unaffected
Unaffected
Unaffected
Unaffected
Unaffected
Unaffected
Unaffected
0.5X average
Increase
0.6X average
decrease
0.6X increase
Corrosion
Measured 75X
increase, but
within nora. )
range
Measured 701
increase, but
within normal
range
No change
Measured 25X
decrease, but
within noraal
range
No significant
change
No significant
change
No significant
change
~
Slight increase
No significant
change
Load
Capacity
20* derate
Unaffected
'Jnaffected
Unaffected
20X AT more
derate with
BOOS
Unaffected
Uni^fected
Unaffected
202 derate
Up to 17X
derate
with BOOS
Up to 25X
derate
with BOOS
Carbon Loss
in Fly*sh
Slight increase
Slight increase
Slight increase
Slight increase
~50JC average
decrease
No change
~3W average
decrease
Unaffected
Unaffected
~130;i average
~50X average
decrease
Oust Loading
~100X increase
100* ircrease
~SOX average
increase
~40t average
increase
~20* average
decrease
Unaffected
Unaffected
~1W average
increase
~15X average
decrease
Part. Sue
Distribution
--
--
No change
NO significant
change
Unaffected
Unaffected
"
Other Effects.
C omen Is
Minor changes in heat
absorption profile
SH atteaperatton
increased by 7M
Minor charges in heat
absorption prof lie
SH atteaperation
increased over 20QX
Xinor changes in heat
absorption profile
SH *ttMper«ttoii
increased hy 70t
Minor changes in heat
Absorption profile
No SH atteaperation
required
-------
TABLE 4-9. (Continued)
Boiler
Kstfteld No. 3
F..C. Gaston
Ko. 1
Crist No. 7
Slnole Hall
Widows Creek
Ho. 5 (TVA
Test)
Widows Creek
No. 5 (Exxon
test)
Widows Creek
No. 6
Mercer Station
No. 1 (wet
bOttOM)
Crist Station
No. 6
Low NOX
Technique
BOOS
FGR
LNB. LEA. COOS
BOOS
BOOS
LEA, BOOS
LEA, BOOS
LEA. Biased
firing
LEA. BOOS
Boiler
Efficiency
O.jl decrease
0.4t decrease
in bailer
efficiency.
Sow decrease
In cycle
efficiency
due to RH.
atteaperatlon.
0.31 decrease
on average
(LNB baseline)
Unaffected
It decrease
It average
Increase
Unaffected
Unaffected
0.4t decrease
Corrosion
_.
So significant
Increase
Increase
Results of
tests
Inconclusive
No significant
increase
No significant
increase
| Load
Capacity
10X derate
Unaffected
Up to 30t
derate
(LNB with
BOOS)
Unaffected
UnaffecteJ
Unaffected
Unaffected
U(vaff(>cted
Up to 15*
derate
1
Carbon Loss
in Flyash
~30X average
increase
~120t average
ircrease
~13(Jt average
ifK.ftS~t (L"*
baseline)
Sllghi increase
30t increase
30t average
decrease
70S average
Increase
80t average
Increase
SOU increase
Oust Loading
Unaffected
Unaffected
~15t average
increase (LNB
base line}
Unaffected
No significant
increase
}*)t average
dec ease
20t average
decrease
lOt average
increase
tflff 4«u"i»«kACtt
Part. Size
Dlstr .lion
--
Shift towards
smaller par-
ticles (LNB.
with or with-
out BOOS)
Unaffected
_ _
_.
No significant
change
Other Effects,
Cements
No slagging cr fouitno.
No sitwiiflcart Increase
In tube temperatures.
Stable flatus and
uniform coobusvi;^.
Increase In RH
atteaperatton. No
significant Increase
in tube temperatures.
Urit retrofitted
wi^h low NO. burners.
Baseline. LEA and BOOS
tests with LNB compared
to baseline tests on
sister boiler with no
1
denotes tni: pameter MIS not lavestlgat-d.
-------
(Widows Creek No. 5) tested at-a different time with BOOS showed an average increase in efficiency
of 1 percent.
Many new boilers now come factory equipped with OFA ports. Older boilers can be retrofitted
with OFA ports, or can operate with minimal hardware changes under BOOS firing. BOOS firing is
normally accomplished by shutting off one or more pulverizers supplying coal to the upper burner
levels. If the other pulverizers cannot handle the extra fuel to maintain a constant total fuel
flow, boiler derating will be required. From Table 4-9 it is seen that boiler derating of 10 to
25 percent is not uncommon with BOOS firing.
The possibility of increased corrosion has been a major cause for concern with staged
operation. Furnaces fired with certain Eastern U.S. bituminous coals with high sulfur contents may
be susceptible to corrosion attack under reducing atmospheres. Local reducing atmosphere pockets
may exist under staged combustion operation even when burner stoichiometry is slightly over
100 percent. The problem may be further aggravated by slagging as slag generally fuses at lower
temperatures under reducing conditions. The sulfides in the molten slag may then readily attack
tube walls. Still, as noted in Table 4-9, experience with most short-term tests has generally been
that no significant acceleration in corrosion rates occurs under staged firing. Recent 12-month
corrosion measurements by Exxon during low NO firing in a 500 MW boiler showed significantly higher
corrosion rates with low NO operation (Reference 4-4). Nevertheless, the issue cannot be
considered resolved until further results are available from long-term corrosion tests with
measurements on units equipped with factory equipped NO controls. Insofar as slagging is
concerned, short-term tests performed to date generally indicate no significant increase in slagging
or fouling of tubes under staged combustion.
Increased carbon loss in fly ash may occur with staged firing if complete burnout of the
carbon particles does not occur in the furnace. 'High carbon loss will result in decreased boiler
efficiency and may also cause electrostatic precipitator (ESP) operating problems. From Table 4-9,
it is seen that increases in carbon los% vary over a wide range and can be as high as 70 to
130 percent of baseline values in some cases. However, -increased carbon loss levels have generally
been considered acceptable and not a major problem associated with staged combustion.
Extension of the combustion region to higher elevations in the furnace may result in
potential problems with excessive steam and tube temperatures in retrofit applications. However,
among the numerous short term combustion staging tests conducted, no such problems have been
reported. In some tests where furnace and, convective section tube temperatures were measured
4-16
-------
directly, no significant increase was found. Changes in heat absorption profiles were also found to
be minor. Superheater attemperator spray flowrates tripled in one case under retrofit OFA
operation, but in all cases they were well within spray flow capacities of the units. Reheater
attemperator spray flowrates did not show any increase due to staged operation, thus cycle
efficiencies were not affected.
Many new wall-fired coal boilers are,being fitted with low NOX burners (LNB). These burners
are designed to reduce NOV levels either alone, or in some cases, in combination with OFA ports.
X
Using the new burner designs has the advantage of eliminating or decreasing the presence of reducing
or near reducing conditions near furnace walls that could cause corrosion. Although low NO
A
burner flames are generally less turbulent and hence longer than flames from normal burners, the
combustion zone will probably not extend any farther up the furnace than with overfire air
operation. Potential changes in heat absorption profile and excessive steam and tube temperatures
are, therefore, less likely to occur.
As fuel and air flows are controlled more closely in LNB equipped systems, nonuniform
distribution of fuel/air ratios leading to excessive CO generation or high excess air requirements
should be reduced. Boiler efficiencies should, therefore, not be affected, especially in new unit
designs where adequate volume for carbon burnout is available. However, Table 4-9 shows that the
efficiency of one boiler decreased slightly (0.3 percent) when retrofitted with low NO burners.
*
The decrease in efficiency was mainly due to the large increase in unburned carbon loss. However,
such problems noted in retrofit applications can be avoided in units specifically designed for low
NOX burners. Corrosion rates inferred from tests with corrosion coupons showed no significant
increase with the new burners. . Some BOOS tests were also carried out on the LNB equipped boiler. A
substantial decrease in NOX emissions resulted, below those already achieved with the new burners
alone. However, the boiler was derated by up to 30 percent. Other potential problems noted above
as being associated with staged combustion could also arise with this type of firing.
It should be emphasized that the operational effects of NO control, in many cases, will be
critically dependent on boiler operating conditions. Still, with proper design of retrofit systems
and adequate maintenance programs, low NO operation should not result in a substantial increase in
operational problems over normal boiler operation. Moreover, when NO controls are designed into
new units, potential problems can be anticipated and largely avoided.
4-17
-------
Other advanced NO control techniques under development, including the EPA advanced low NO
burner and EPRI primary combustion furnace, are designed to further minimize some of the potential
problems with conventional' combustion modifications, such as losses in boiler efficiency and
Increases in corrosion rates.
Application to Oil-Fired Boilers
For cost reasons, most oil-fired utility boilers fire residual oil. The most commonly used
low NOX techniques for these boilers are staged combustion and flue gas recirculation (FGR), both
employed in combination with low excess air firing. These techniques have generally been employed
on a retrofit basis since few new oil-fired units are being installed. Other techniques which have
been tested are water injection (WI) and reduced air preheat (RAP). However, these have found
little application due to associated efficiency losses.
Staged combustion has been applied through the use of overfire air ports (OFA) and by
removing burners from service (BOOS). Typical NO reductions using OFA are 20 to 30 percent from
baseline conditions (controlled emissions of 150 to 170 ng/J, 0.35 to 0.4 lb/106 Btu). BOOS has
been slightly more effective, giving 20 to 40 percent reductions (controlled levels of 130 to
170 ng/J, 0.3 to 0.4 lb/10 Btu). A load reduction may accompany the use of BOOS on some units.
Flue gas recirculation also typically gives 20 to 30 percent NOV reductions, but requires
A
more hardware modifications. The combinations of BOOS or OFA with FGR have been most effective,
resulting in 30 to 60 percent reductions (controlled emission of 86 to 170 ng/J, 0.2 to
0.4 lb/106 Btu). With FGR, OFA is preferred over BOOS because flame stability seems to be more of
a problem with the combination of FGR + BOOS.
t
There have been some R&D efforts by EPA and private industry to develop low NO emission
burners for oil firing. Although no detailed test reports have been released yet, the Babcock &
WUcox dual register low NO oil-fired burner has been successfully retrofitted on a utility boiler,
achieving emissions below 130 ng/J (0.3 lb/10 Btu) (Reference 4-5). The combination of overfire
air and low NO burners may potentially achieve emissions below 86 ng/J (0.2 lb/10 Btu). An
oil-fired low NO burner is also under development by Southern California Edison to meet the
stringent NOV regulations for utility boilers in the Los Angeles area.
X
4-18
-------
The effects of low NO operation on oil-fired boilers are summarized in Table 4-10. The
major concerns regarding low NO operation on oil-fired boilers are effects on boiler efficiency,
load capacity, vibration and flame instability, and steam and tube temperatures.
Staged combustion operation generally increases the minimum excess air requirements of the
boiler, which may result in a loss in boiler efficiency. In extreme cases when the boiler is
operating close to the limits of its fan capacity, boiler derating may be required. Derating of as
much as 15 percent has been required in some cases due to the lack of capability to meet the
increased airflow requirements at full load. In addition, under BOOS firing the fuel flow to the
active burners must be increased if load is to remain constant. In many cases, it has been
necessary to enlarge the burner tips in order to accommodate these increased flows.
Other potential problems associated with applying staged combustion on a retrofit basis in
oil-fired boilers have concerned flame instabilities, boiler vibrations, and excessive convective
section .tube temperatures. However, in past experience, none of these problems has been
significant. Staged operation can result in hazy flames and obscure flame zones. Thus new flame
scanners and detectors are often required in retrofit applications. In addition, because staged
combustion produces an extended flame zone, flame carryover to the convective section may
occasionally occur. However, in one case where intermittent flame carryover occurred, no excessive
tube temperatures were recorded.
Similarly, there are a number of potential problems which can occur in retrofit F6R
applications. The most common problems, such as F6R fan and duct vibrations, can usually be avoided
by good design. Other problems such as flame instability, which can lead to furnace vibrations, are
caused by the increased gas velocity at the burner throats. Modifications to the burner geometry
and design such as enlarging the throat, altering the burner tips, or adding diffuser plates or
flame retainers, may then be required.
Another potential problem sometimes associated with FGR is high tube and steam temperatures
in the convective section. The increased mass velocities which occur with FGR reduce furnace heat
transfer, but increase convective heat transfer in the convection section. The result is increased
convection section flue gas, steam and tube temperatures. Increased mass flowrates in the furnace
may also cause furnace pressures to increase beyond safe limits.
The combination of staged combustion and FGR is very effective in reducing NO emissions.
However, the potential problems associated with each technique are also combined. Tube and steam
4-19
-------
TABLE 4-10. EFFECT OF LOW NO OPERATION ON OIL-FIRED BOILERS2
(Reference 4-3)
Boiler
Tangential
South Bay No. 4
Pittsburg Ho. 7
SCE tangential
boilers
Opposed Wall
Moss Landing
Nos. 6 and 7
(Early
experience)
Moss Landing
Ho. 6 (NO, EA
tests)
Ormond Reach
Nos. I and ?
Low MOX
Technique
LEA
BOOS
RAP
OFA and FGR
BOOS and FGR
BOOS and FGR
FGR
BOOS and FGR
BOOS and FGR
Water Injection
Efficiency
5< Increase
«
Decrease in efficiency
compared to LEA due to
Increased excess air
requirements
Unaffected due to
special preheater
design
"
Increased excess air
requirements resulting
In decreased efficiency
Unaffected
Unaffected
Increased excess air
requirements resulting
in decreased efficiency
Increased sensible and
latent stack losses
Load
Capacity
Slower startups
and load changes
Unaffected
Unaffected
10 to 15X derate
due to maxed FD
fan capacity
Vibration and
Flame Instability
--
--
FGR fan vibration
problems
FGR fan and duct
vibration, furnace
vibration problems.
Associated flame
instability.
None
None
Flame instability
and associated
furnace vibration
Stew ind Tuhe
Te«per«tures
--
-
High water wall tube
temperatures
--
No attemper at Ion
changes required
No attemperation
changes required
""
Other Effects, Comments
No adverse effects reported.
Fan power consumption
reduced.
No other adverse effects
reported
Limited tests. NOX
control effectiveness not
demonstrated.
No adverse effects reported
High furnace pressures.
Increased FGR and forced
draft fan power consumption.
Unit currently operating
under low NOX continuously
Flame detection problems
due to change In flame
characteristics
Limited tests carried out
with WI at oartial loads.
Excess a'*- requirements
increased.
-------
TABLE 4-10. (Continued)
Boiler
SCE 6&W Units
Sewaren Station
No. 5
Single Mall
Enctna Hos. 1,
2 and 3
Turbo
South Bay No. 3
Potrero No. 3-1
Low NOX
Technique
BOOS and FGR
LEA. BOOS
LEA and BOOS
(2 burners
.on air only)
BOOS
(3 burners on
air only)
Airflow
adjustments
Water injection
Reduced air
preheat
OFA and FGR
Efficiency
FGR reduced minimum
excess air requirements
increasing unit
efficiency
Increased unit effi-
ciency. Some adverse
effect on cycle eff i-
ciency. due to lower
steam temperatures.
increased excess air
requirements resulting
In reduced efficiency
Slight reduction in
EA resulting in slight
increase in efficiency
65! decrease at full
load
Reduction in effi-
ciency greater than
that with water
injection
Higher excess air re-
quirements, hut addi-
tion of economizer
surface expected to
improve efficiency
Load
Capacity
_ M
53 derate due to
maxed 10 fan
capacity
-.
^ »
_
51 derate due to
excessive tube
temperatures
Vibration and
Flame Instability
Boiler vibration
problems
__
_
In most tests no
flame instability
or blowoff noted
_.
No flame Instability
noted even at high
rates of MI
^
Side to side
wind box oxygen
cycling
Steam and Tube
Temperatures
....
Decrease In SH & RH
steam temperature
*£
Intermittent flame
carryover to SH
inlet but tube
temperature limits
not exceeded
~» -
»«>
_.
Tube and steam tem-
perature limits ap-
proached. Increased
SH tube failures.
Other Effects, Comments,
Flame detection problems
due to change in flame
characteristics
Tests carried out at partial
loads. No adverse effects
reported. Particulate load-
ing and size distribution
unaffected.
No other adverse effects
reported
"No abnormal tube fouling.
corrosion or erosion noted.
Increased tendency to smoke
and obscure flame zone.
No adverse effects reported
No other adverse effects
reported
Limited tests
Increased tendency to smoke
required higher minimum
excess Op levels., RH
surface ranoved to avoid
excessive RH steam attem-
peratlon. Larger economizer
installed to compensate for
RH surface removal.
I
ro
denotes not investigated.
-------
temperature problems in the upper furnace are an area .of concern, as both combustion staging and F6R
tend to increase upper furnace temperatures and convective section heat transfer rates. In
addition, boiler efficiencies usually decline slightly with combined staged combustion and FGR
firing due-to higher excess air requirements and greater fan power consumption. However, potential
adverse effects for retrofit NO control systems can be minimized by proper design and installation.
Many of the problems experienced in the past can now be avoided.
Application to Gas-Fired Boilers
The most commonly applied NO control technique for gas-fired boilers, as with oil-fired
X
boilers, is staged combustion through the use of OFA or BOOS. These prove more effective when
combined with FGR; however, flame stability may be of greater concern when BOOS is combined with
FGR. Typical NO reductions under either OFA, BOOS, or FGR are 30 to 60 percent from baseline
conditions (controlled emissions of 86 to 150 ng/J, 0.2 to 0.35 lb/10 Btu). The combination of
staged combustion and FGR is capable of 50 to 80 percent reduction (controlled levels of 43 to
110 ng/0, 0.1 to 0.25 lb/106'Btu).
There are no major efforts toward developing a low NO burner or other new combustion
A
modification techniques for gas firing because NO emissions under current control techniques are
X
already relatively low, and no new gas-fired utility boilers are currently being sold.
The effects of low NO operation on gas-fired boilers are summarized in Table 4-11. The
effects of low NO firing on gas-fired boilers are very.similar to those for oil-fired boilers.
X
Usually, there is no distinction between oil- and gas-fired boilers as they are designed to switch
from one fuel to the other according to availability. Since boiler design details, NOX control
methods, and the effects of low NOX operation are similar for gas- and oil-fired units, most of the
above discussion of applicable NOX control measures for oil-fired boilers and potential resulting
problems applies here as well. Some effects specific to gas-fired boilers alone are treated briefly
below.
NO emissions are often difficult to control after switching from oil to gas firing.
Residual oil firing tends to foul the furnace due to the oil ash content. Thus, NO control
X
measures which have been tested on a clean furnace with gas may be found inadequate after oil firing
due to the changed furnace conditions.
4-22
-------
TABLE 4-11. EFFECT OF LOW N0y OPERATION ON GAS-FIRED BOILERS0
(Reference 4-3)
I
M
CO
Bo Her
Tangential .
Smith Bay No. 4
Pittshiirg Ko. 7
Horizontally
Opposed
Hoss Landing
Not. 6 and 7
Pittsburg
Has 5 and 6
Contra Costa
Has. 9 and 10
Single Hall
Incina Nos. 1,
2 and 3
LowHOx
Technique
BOOS
OFA and FGR
OFA and FGR
OFA and FGR
OFA and FGR
BOOS
(? and 3
burners out
of service)
Efficiency*
Slight decrease in
efficiency due to
increased excess air
requirements
O.BJC decrease in cycle
efficiency due to BH
steam atlemperation
"
low EA levels were
possible even with
BOOS, resulting in
increased efficiency
Lnada
Capacity
25X derate
-------
Table 4-11. Concluded.
Boiler
Turbo
South Bay No. 3
Potrero Ho. 3-1
Low NOX
Technique
Water Injection
OFA and FGR
Efficiency
10% decrease at full
load
Installation of larger
economizer expected to .
improve efficiency
Load
Capacity
--
5X iterate due to
prohlens with high
terrperatures
Vibration .ind
Flame Instability
No flame instability
noted even at high
rates of HI
Side to side
trindbox oxygen
cycling
Steam and Tune
Temperatures
Tube metal and steam
temperature limits
reached at high
loads
Other Effects, Coaments
No other adverse effects
"reported
Hardware modifications
included partial RH surface
removal to avoid excessive
RH steam attemperation.
Larger economizer fion
installed to compensate for
smaller RH surface.
I
rv>
£»
-------
Boilers tired with gas usually have higher gas temperatures at the furnace outlet than those
fired with oil. The upper furnace and convective section inlet surfaces are thus subject to higher
temperatures with gas firing. These temperatures may increase further under staged firing or F6R.
Upper furnace and convective section tube failures and excessive steam temperatures are therefore
more likely to occur with staged firing and FGR applied to gas-fired boilers. The situation may be
aggravated further if switching to gas fuel occurs after an oil burn, as fouling will further reduce
furnace absorption and, hence, increase gas temperatures. Excessive tube temperatures will usually
require derating of the system. These problems could be minimized on new units but no new gas-tired
utility boilers are being built.
4.1.1.2 Flue Gas Treatment
Historically, the major NOX control emphasis in the United States has been on combustion or
process modification. However, in Japan where NO emission standards are more stringent, flue gas
treatment (FGT) technologies have undergone extensive development and implementation. Recently, in
the U.S. several pilot and demonstration scale units have been built and operated.
As discussed in Section 3.2, flue gas treatment consists of any of several technologies
designed to remove or eliminate NO in the flue gas downstream of the combustion zone. Since FGT
X
technologies are distinct from the combustion process, their performance capabilities'are usually
described in terms of NO percentage reduction in the flue gas rather than in terms of achievable
emission levels. NO reductions with I-GT can occur over and above the reductions attributable to
combustion modifications.
These postcombustion processes can be divided into dry or wet types. The dry processes can
be further categorized into four subdivisions: catalytic reduction, noncatalytic reduction,
adsorption, and irradiation. Ihe majority of the dry processes are of the reduction type. These
catalytic and noncatalytic reduction processes can also be classified as selective or nonselective
*
processes based on the type of reducing agent used. The majority are selective and usually use NH,
as the reducing agent. If the NH, is injected after the boiler economizer, where temperature of the
flue gas is about 370°C to 430°C (700°F to 800°F), a catalyst is necessary. These processes are
described as selective catalytic reduction (SCR) processes. If NH3 is injected into the secondary
superheater region of the boiler, where temperature of the flue gas approaches 980°C (1,800°F), a
, catalyst is not necessary. These processes are described as selective noncatalytic reduction (SNR)
processes.
4-25
-------
The remainder of this section describes the development status of flue das treatment technolo-
gies and experience in applying them to boiler flue gases. The discussion is organized by fuel
type. Much of the material was obtained from Reference 4-6; the reader is referred to this and
other documents (e.g. References 4~7, 4-8, and 4-9) for additional information.
Selective Catalytic Reduction (SCR)
The SCR method is the most advanced FGT method, and the one on which the overwhelming
majority of existing NOX FGT units are based. As with the majority of all types of NOX FGT, most of
the SCR processes were developed in Japan. The Japanese have found that with the optimum reaction
temperature, usually 300°C to 450°C (570°F to 840°F), an NH3:NO molar ratio of 1:1 typically reduces
NOV emissions by 90% with residual NH, concentrations of 10 to 20 pprn or higher. It should be noted
A O
that the Japanese seem to prefer 80% NOX removal in which NH3:NO molar ratios range from 0.81:1 to
0.9:1. Under these conditions the unreacted NH3 concentration is usually less than 5 ppm. This
reduces capital and operating costs as well as effects on downstream equipment from ammonium salt
deposition.
Presently, there are over 60 full scale SCR units successfully operating on gas- or oil-fired
boilers in Japan. Over 10 percent of these units are larger than 330 MW. Two commercial SCR units
began operating in 1980 on coal-fired boilers in Japan. Construction is scheduled to be completed
during 1981-1984 on at least 14 additional SCR units for coal-fired boilers ranging in capacity from
75 to 700 MW (References 4-10, 4-11, and 4-12). The results of the Japanese experience on these
units must be tempered by the dissimilarities with U.S. facilities including maintenance practices,
load cycling, and overhaul practices.
In the United States, EPA and the Electric Power Research Institute (EPRI) are evaluating SCR
on coal-fired pilot scale units. EPA sponsored two 0.5 MW size tests. The Shell Flue Gas Treatment
process for simultaneous NOV and SOV control was evaluated at Tarnpa Electric Company's Big Bend
. * x
Station. This process controls NO by selective catalytic reduction although it uses a different
X
catalyst than other SCR processes. The Hitachi Zosen SCR process was tested at the Plant Mitchell
Station of Georgia Power Company. Test results showed that both processes are technically capable
of achieving significant NO reductions from coal-fired boilers in the U.S.
One conclusion of the testing was that SCR test work is needed when considering SCR process
applications for untested coals. Another was that a prototype scale test on a 10 to 100 MW facility
would be useful for demonstrating the technology for coal-fired sources in the U.S. For further
details on the results of these tests, the reader is referred to Reference 4-13.
4-26
-------
In other U.S. testing, EPRI has operated a 2.5 MW pilot unit on a coal-fired boiler at the
Arapahoe Station of Public Service Company of Colorado using the Kawasaki Heavy Industries process.
Also, testing of the first large scale SCR demonstration unit in the U.S. is planned by Southern
California Edison Company at the Huntington Beach Station. The unit will be 107.5 MW in capacity
and applied to a oil- and gas-fired boiler.
Since the SCR reactor is located downstream of the boiler economizer, its process impacts are
also largely limited to this region. These include potential problems with the SCR reactor itself,
the air preheater, and downstream emission control systems.
Dr. Oumpei Ando (Reference 4-14) reports that early ,in its development SCR had several
serious problems. These included:
(1) Catalyst poisoning by SO in the flue gas;
(2) Plugging of the catalyst by dust;
(3) Deposition of ammonium bisulfate on the catalyst at reduced
boiler loads;
(4) Deposition of ammonium bisulfate on the air preheater;
(5) Promotion of the oxidation of SOg to S03 in the flue gas; and
(6) Erosion of the catalyst by fly ash from the coal.
These problems have been largely resolved as follows:
Problem (1) - Development of SOX resistant catalysts based on TiOg
rather than Al20g or Fe203;
Problem (2) - Using parallel-flow catalysts such as honeycomb,
plate, and tube shapes, parallel passage reactors, and
sootblowing when needed;
Problem (3) - Keeping the reactor temperature above 300°C by using
economizer bypass gas when needed;
Problem (4) - Keeping NH, leakage at the reactor outlet at a low
level (e.g. below 5 ppm);
Problem (5) - Development of low oxidation catalysts which also
helps to solve problems 3 and 4; and
4-27
-------
Problem (6) - Using moderate gas velocities, hard catalysts, and
a device (e.g. a dummy spacer) tor erosion prevention.
Jones (Reference 4-15) reports that a major area of research and development involves
minimizing the impacts of SCR systems on downstream equipment such as air preheaters, particulate
collection devices and SCL removal equipment. Problems with the air preheater occur when ammonium
blsulfate (NH.HSO.) deposits plug and corrode the elements. NH.HSO. is the product of a
\
condensation reaction between NHj, SO, and H-0, which can occur when the flue gas temperature drops
below about 210°C. Japanese pilot unit tests have shown that the plugging problem is most severe in
units which fire coal or high sulfur oil and also remove fly ash upstream of the NOX reactor. When
fly ash 1s removed downstream, plugging problems are significantly reduced. It is felt that the fly
ash produces a sandblasting effect that cleans the air preheater elements and also that some of the
NH^HSO^ condenses on the fly ash particles rather than the elements. Plugging problems are reduced
or eliminated by installing soot blowers on both sides of the air preheater and increasing both the
frequency and pressure of the soot blowing operation. In some cases, special air preheater designs
will be used in which the intermediate and low temperature zones are manufactured as a single
element. These have been tested on pilot unit equipment. A full scale installation is scheduled at
the Electric Power Development Company's Takehara Power Station.
NHg from an SCR reactor apparently does not impair F6D system performance although, in some
cases, the wastewater must be treated to remove nitrogen compounds. It is not known if SCR systems
will affect dry S02 removal systems (e.g., spray drying) since these techniques are not used in
Japan. The one apparent adverse impact that may occur is NH.HSO. affecting the performance of the
downstream baghouse. The effect of an SCR .system on baghouses is under investigation. Pilot unit
tests are underway.
Several of the coal-fired SCR applications that are under construction utilize hot-side ESPs
tor upstream particulate removal, and there are a variety of reasons for selecting hot-side
particulate removal. These reasons include:
(1) To eliminate fly ash from entering the NO system and
potentially causing plugging or erosion problems;
(2) To obtain the capability to remove particulates from a
wide range of coals with varying characteristics; and
4-28
-------
(3) To avoid ammonia compounds in the ash that can result
when a cold-side ESP is used.
However, cold-side ESPs also have unique advantages such as:
(1) Lower capital and operating costs; and
(2) Allowing the fly ash to reduce or eliminate NH4HS04
deposits on air preheaters.
Jones further discusses the fact that .there has been concern that the catalyst and reactor may
plug with ash when applied to coal-fired boilers. Pilot unit tests have indicated that plugging is
not a problem when honeycomb' or pipe shape catalyst is used in a vertical, downflow arrangement.
However, soot blowers will be installed in the reactors of current full scale applications as a
conservative design measure. Another concern in the U.S. has been catalyst poisoning by flue gas
components. While it is true that certain alkali metals, such as sodium and potassium, will slowly
poison the catalyst, the concentrations are low enough that catalyst activity will not be affected
during the guarantee period. Catalyst life guarantees are usually one year for coal, one to two
years for oil, and two to three years for gas although the experience on gas- and oil-fired boilers
has been that actual catalyst life exceeds the guarantee.
The labor requirements of the operating, full-scale systems are small. No additional
operating personnel are required and maintenance labor consists* primarily of NH, and catalyst
loading and cleaning the air preheater during the annual outage. Since there have been no catalyst
changes to date, the labor estimates for this work vary widely. Operators indicate that the SCR
processes themselves are very reliable, essentially 100%. However, in some cases, a boiler shutdown
has been necessary where air preheater plugging has occurred. In most cases steps have been under-
taken to reduce the plugging rate to the extent that cleaning is only required during normal boiler
outages.
Some additional process impacts have been reported for retrofit applications (Reference 4-16).
hor example, in many cases all equipment and ducting downstream of the economizer, including the
stack, will have to be relocated to make room for the reactor and the ammonia flue gas mixer.
Existing structures, equipment, and other site constraints may interfere with the required expansion
of the back end of the boiler, thus requiring major site rearrangements*. This requirement could
cause long construction periods and extended unit outages. Also, larger units may require a booster
4-29
-------
fan to overcome the added system pressure drop, causing a conversion from forced draft to balanced
draft operation. This may result in somewhat greater risk of boiler implosion.
Selective Noncatalytic Reduction (SNR)
Exxon Research and Engineering Corporation developed the SNR process in which NHg is injected
Into the boiler where proper flue gas temperatures allow the reduction of NO by reaction with NH,
* x 3
to proceed without a catalyst. Generally, 40% to 60% NOV reduction is achieved with NH,:NOV molar
X OX
ratios of 1:1 to 2:1. SNR may be more attractive than SCR in cases where only 40% to 60% NO
control 1s needed since SNR is simple and does not require expensive catalysts.
The general disadvantage of SNR is the limited NO control achievable, especially with larger
boilers. This limited control results from the difficulty of achieving rapid uniform mixing of NH,
with the flue gas and from the variations of flue gas temperature and composition usually present
within the boiler region where the SNR is operated. NH3 consumption and unreacted NH3 levels can
also be high because of the high NH,:NO molar ratios needed with this process. Also, many of the
same problems of ammonium salt formation, previously discussed for SCR processes, occur with SNR as
well.
There are several large SNR units installed in Japan, between 30- and 100-MW capacity, mostly
supplied by Tonen Technology (a subsidiary of Toa Nenryo) which has a license from Exxon. These
units are operated on gas- and oil-fired boilers or furnaces. Practically all are only for
emergency use during a photochemical smog alert or when total plant emissions exceed the regulation.
There are presently two commercial SNR plants operating in the United States. One is on a
glass melting furnace and the other a petroleum refinery, both located in California. The
construction of five other industrial-scale units is planned. The SNR process is also being
Installed by Exxon at the No.4 oil-fired unit of the Haynes Station of the Los Angeles Department of
Water and Power.
Other Flue Gas Treatment Techniques
In addition to SCR and SNR, dry processes which are being developed for simultaneous SO and
NO control include:
(1) Activated carbon processes where NH, reduces NO to N-;
4-30
-------
(2) Copper oxide processes where NH3 reduces NO to N2; and
(3) Electron beam irradiation processes in which NH, is added
to produce ammonium sulfate and nitrate.
Also, work has been conducted on various wet processes for simultaneous SO and NO control.
X X
The optimum temperature range for simultaneous SO and NO control with activated carbon
processes is 220°C to 230°C (430°F to 445°F). Although NOV may be adsorbed below 100°C (212°F), for
A
treating large quantities of flue gas above 100°C the carbon is mainly useful as an NO reduction
catalyst. Therefore, while NO is converted to N9 by. reaction with NH, in the presence of the
X £ -3
activated carbon catalyst, S02 is simultaneously adsorbed by the carbon to form H2SO.. The H-SO.
may also compete for NH., in forming ammonium sulfate or bisulfate. The formation of these ammonium
salts increases NH3 consumption and also lowers catalyst activity. The carbon must be regenerated,
either by washing or thermal regeneration. Washing produces a dilute solution. Concentration of
the solution to produce a fertilizer requires much energy. Thus, thermal regeneration seems to be
preferred. A concentrated SO- gas is recovered, which can be used for sulfuric acid or elemental
sulfur production.
The major drawback of the activated carbon processes is the enormous consumption of activated
carbon, which is more expensive than ordinary carbon used only for SOV removal. Since carbon and
X
ammonia consumption increase with the S02 content of the flue gas, the process is best suited for
flue gases relatively low in SO,,. In Japan Sumitomo Heavy Industries and Unitika Company have
operated activated carbon pilot plants of 0.6 MW and 1.5 MW capacity respectively.
Ihe Shell Flue Gas Treatment process may simultaneously remove SO and NO . SO reacts with
the copper oxide acceptor to form copper sulfate. The copper sulfate and copper oxide are SCR
catalysts for the NO reduction by NHg. Regeneration of the multiple catalyst beds by a reducing
gas, such as H2, yields a S02-rich stream that can be used to produce liquid S02, elemental sulfur,
or sulfuric acid. By eliminating NH3 injection, the process is strictly an F6D process, whereas,
eliminating regeneration of the catalyst beds allows the process to be used for only NO control.
The major disadvantages are the large consumption of fuel for making hydrogen and the catalyst
expense.
In addition to the EPA-sponsored pilot plant mentioned earlier, the process has been
installed in Japan, on .a 40-MW oil-tired boiler. The unit has demonstrated 90% SO removal and 7056
NO reduction.
4-31
-------
Another process for simultaneous SO and NO control is the electron beam process developed
X A
by Ebara Manufacturing Company in Japan. NH3 is added to the flue gas, after which the gas stream
is irradiated with an electron beam in a reactor, promoting the conversion of SOV, NOV, and NH, to
XX O
ammonium sulfate and ammonium nitrate. The ammonium sulfate and ammonium nitrate may be collected
downstream in an ESP or baghouse and potentially sold as a fertilizer. The most economically
practical removal efficiency range appears to be 80% to 90% for each of NOV and SOV> though higher
X X
removals can be achieved with much greater electron beam energy input. The optimum temperature
range is 70°C to 90°C (160°F to' 195°F).
Ebara has worked on the process since 1971. It has been tested at a 0.3 MW and 3 MW scale in
Japan. Avco Corporation in the United States has also examined this technique and has a
cross-licensing agreement with Ebara in sharing of technology and in marketing of the process.
Although the process appears attractive because of simplicity, simultaneous SOV and NOV control, and
X X
byproduct formation, there are still many questions concerning costs and byproduct quality which
must be determined.
Development of an alternate electron beam scrubbing process was begun in 1979 by
Research-Cottrell under contract to the Department of Energy (DOE). With this process a lime spray
drier is located upstream of the reactor. Calcium sulfate and calcium nitrate are produced in the
reactor and caught in a downstream baghouse. Some bench scale testing has been done with this
process. DOE plans proof-of -concept scale testing of both the ammonia injection and lime slurry
injection electron beam processes on real coal-fired slip streams (Reference 4-17).
The wet F6T processes normally involve simultaneous removal of SO and NO . The major
problem associated with wet NO control processes is the absorption of NO by the scrubbing solution
in which it can be concentrated and converted into other forms. NO in the flue gas is predomi-
nantly NO, which is much less soluble than NOg, whereas, NO- is even less soluble than SOg. The two
common methods of removing the NO in flue gas by wet processes are: (1) direct absorption of the
NO in the absorbing solution or (2) gas-phase oxidation to convert the relatively insoluble NO to
NOp. followed by absorption of NOp. Presently, development of the wet NO F6T processes has
practically ceased because of the complexity and unfavorable economics of these processes in
comparison with the dry processes.
4-32
-------
4.1.2 Costs .
In this section some available estimates of the capital and annualized costs of alternative
NO control techniques are presented. The section is divided into two major subsections -
combustion modifications and flue gas treatment. The costs and discussion for combustion modifica-
tions were largely taken from two reports prepared for EPA. by Acurex Corporation (References 4*18
and 4-19). Much of the costs and discussion for flue gas treatment techniques were taken from a
report prepared for the EPA by TVA (Reference 4-20). However, as noted in the section, several
other references were also used.
The reader should be careful in attempting to compare the costs presented in this section.
Some differences in methodologies and bases were used in developing these costs. For detailed
discussions of how these costs were developed, the reader is referred to the original references.
4.1.2.1 Combustion Modification
Recently, as discussed in the above references, Acurex Corporation conducted an environmental
assessment for EPA of various utility boiler combustion modification NO controls. As part of this
study costs were developed for a number of cases representing retrofit and new boiler applications.
The following discussion is largely taken from this study.
The use of accepted estimation procedures for costing NOX control implementation in current
dollars was employed in this study, with heavy reliance on discussions with boiler manufacturers,
equipment vendors, and utilities. For the case of retrofit control costs, preliminary design work
was performed to allow estimation of hardware and installation needs, as well as engineering
requirements. The analysis was applied to a number of cases to give a range of retrofit control
costs. For the cost of NOX controls in new boilers, the services of two major suppliers, the
Babcock & Mil cox Company and the Foster WheeleY Energy Corporation, were enlisted.
For the analysis of the cost of controls, regulated public utility economics were adopted.
Based on a revenue requirement approach, an annualized cost methodology was developed, adapted from
that used by the Tennessee Valley Authority in evaluating the cost of power plant projects for EPA
(Reference 4-21) and EPRI (Reference 4-22). This procedure has been generally accepted in the
industry (References 4-23 through 4-25).
4-33
-------
For the present application, the additional revenue requirement represents the incremental
cost of operating a boiler under controlled conditions over and above the cost of operating the same
boiler uncontrolled. In other words, the revenue requirement takes into account the initial-
investment, the annual capital charges resulting from that investment, and all direct operating
costs such as operation and maintenance. This methodology is described in detail in Reference 4-18.
Retrofit Control Costs
Based on this cost analysis methodology, typical retrofit control costs (1978 dollars) are
summarized in Table 4-12. The costs shown in the table should be considered only representative of
retrofit costs. They apply to retrofitting relatively new boilers, approximately 5 to 10 years old
with at least 25 years of service remaining. With the exception of BOOS for coal-fired units, and
FGR/OFA for oil- and gas-fired units, annualized control costs generally fall in the $0.50 to
0.70/kW-yr, based on a 7000-hour operating year. For comparison, the cost of operating a power
plant is approximately $175/kW-yr.
BOOS was not treated in the cost analysis as a recommended control technique for coal firing.
Rather it was included to show the extremely high cost of derating. This high cost was due
principally to the need to purchase make up power and to account for the lost capacity of the system
through a capital charge.
Tables 4-12 and 4-13 present projected retrofit control requirements for alternative NO
emission'levels. Based on favorable process analysis results, it is evident from an examination of
these tables that OFA and LNB are the preferred, cost-effective NO controls for coal firing. For
high levels of NO control for coal-fired units (170 ng/J), both OFA and LNB may be appropriate.
For more moderate levels of control, LNB are less expensive and more cost-effective than OFA in
reducing NO in wall-fired units. However, retrofit of low NO burners may not be widely
X X
applicable.
As far as moderate control for oil- and gas-fired units, staged combustion via BOOS appears
to be the preferred route, as indicated in Tables 4-12 and 4-13. Initial investment is minimized
since there are no associated major hardware requirements, only engineering and startup costs. To
reach the next level of NOX control (86 ng/J for oil, 43 ng/J for gas), FGR with OFA would seem to
be in order. However, this alternative results in a cost increase from $0.52/kW-yr for BOOS to
$3/kW-yr for FGR and OFA.
4-34
-------
TABLE 4-12. SUMMARY OF UTILITY BOILER COMBUSTION MODIFICATION
CONTROL COSTS (1978 DOLLARS) (Reference 4-19)
I
c*>
U1
Boiler/Fuel Type
Tangential /Coal -Fired
OFA
Opposed Wall /Coal -Fired
OFA
LNB
BOOS
Single Wall/Oil- and Gas-Fired
BOOS
FGR/OFA
MCRa
(MW)
225
540
540
540
90
90
Initial
Investment
($/kW)
0.96
0.66
2.17
0.09
0.32
6.09
Annual ized
Indirect
Operating
Cost
($/kW-yr)
0.22
0.17
0.36
5.70
0.05
1.22 .
Annual ized
Direct
Operating
Cost
($/kW-yr)
0.34
0,55
0.06
26.40
0.47
2.04
Total Cost To
Control .
($/kW-yr)D
0.57
0.73
0.42
32.10
0.52
3.26
aMaximum continuous rating in MW of electrical output.
Based on 7000-hour operating year. Typical costs only.
-------
I
CO
-------
Control Costs tor New Boilers
Estimating the incremental costs of NOX controls for new boilers is in some respects an even
more difficult task than costing retrofits. Certain modifications on new units, though effective in
reducing NO emissions, were originally incorporated due to operational considerations rather than
X
from a control viewpoint. For example, the furnace of a typical new unit has been enlarged'to
reduce slagging potential and allow the burning of poorer quality fuels. But this improvement also
reduces NO due to the lowered heat release rate. Thus, since the design change would have been
implemented even without the anticipated NO reduction, the cost of that design modification should
**
not be attributed to NO control.
X
Babcock & Wilcox has estimated the incremental costs of NO controls on a coal-fired boiler
X
designed to meet 19/1 NSPS (Reference 4-26). Units designed prior to 1971 did not include NOX
controls. NO emissions from coal-fired units designed at that time were on the order of 430 ng/J
(1.0 lb/106 Btu). The 1971 NSPS required that these emissions be limited to 300 ng/J
(0.7 lb/10 Btu). The two units used in the comparison by Babcock & Wilcox were identical except
for NO controls on the NSPS unit which included:
X ,
- Replacing the high turbulence, rapid-mixing cell burner
with the limited turbulence dual register (low NO ) burner;
X
- Increasing the burner zone by spreading the burners
vertically to include 22 percent more furnace surface; and
- Metering and controlling the airflow to each row of
burners using a compartmented windbox.
To provide these' changes for NO control, the price increase was about $1.87 to $2.67/kW (1978
dollars). If these costs are annualized, they translate to $0.30 to 0.43/kW-yr.
In addition, hoster Wheeler has performed a -detailed design study aimed at identifying the
=ntal costs of NO control to meet
X
unit designs with the following results:
incremental costs of NO control to meet 1971 NSPS (Reference 4-27). Foster Wheeler looked at three
X , "
Boiler Design Relative Cost
Unit 1: Pre-NSPS base design 100
Unit 2: Enlarged furnace, no 114
active NOV control
A
4-37
-------
Unit 3: NSPS design; enlarged . 115.5
furnace, new burner design,
perforated hood, overfire
air, boundary air
For a pre-NSPS coal-fired boiler costing about $192/kW In 1978 construction costs, the
Incremental cost of active NOX controls (LNB plus OFA) is $2.97/kW, or about $0.47/kW-yr annualized.
The Foster Wheeler estimate, which includes both LNB and OFA, thus agrees quite well with the
Babcock & Mil cox estimate, which includes only LNB and associated equipment.
Comparing these costs with the retrofit costs (0.40 to 0.70 S/kW-yr for LNB or OFA) presented
in Table 4-13 and considering the better NO control anticipated with NSPS units, it is certainly
X
more cost effective to implement controls on new units. Furthermore, fewer operational problems are
expected with units specifically designed for these controls.
Advanced combustion modification concepts under development, such as the EPA advanced low NOX
burner (Reference 4-28) and EPRI primary combustion furnace (Reference 4-29), are targeted to
achieve NOX emission levels around 86 ng/J (0.2 lb/106 Btu) on a commercial basis in the 1980's.
Projected cost for the EPRI furnace is $4/kW or $0.80/kW-yr (Reference 4-30). The EPA advanced
burner costs should fall in the same range. These developing advanced combustion modifications
should eventually prove much more cost-effective than the developing post combustion techniques.
discussed next. However, the latter techniques are currently closer to commercialization.
In conclusion, conventional combustion modifications are indeed a cost-effective means of
control for NOX, raising the cost of electricity less than 1 percent in most cases. Furthermore,
the initial capital investment required should also only be of the order of 1 percent or less of the
installed cost of a boiler. Advanced techniques such as advanced low NO burners and advanced
boiler/furnace concepts have projected costs of the same order as conventional combustion modifica-
tions. Therefore, preferred current and projected combustion modification techniques are not
expected to have a substantial adverse economic impact.
4.1.2.2 Flue Gas Treatment
As discussed in Reference 4-20, TVA, under contract to EPA, conducted a preliminary economic
analysis in 1980 to compare several flue gas treatment (F6T) processes. The following discussion is
largely taken from this reference. Some additional cost information and references are also
presented.
4-38
-------
1980 TVA Study
The TVA study developed preliminary economics, comprising total capital investment and annual
revenue requirement estimates, for seven NO F6T processes. The economics were calculated based on
a consistent set gf design and economic premises that have formed the basis for many previous flue
gas desulfurization (F6D) studies done by TVA. The reader should be cautioned that at the time
these costs were developed, most of these systems were at an early stage of development for
coal.-fired applications. Thus, actual systems could vary significantly from the costs presented
here.
The FGT processes evaluated are shown'in Table 4-14. These include one dry and three wet
processes for simultaneously controlling SOV and NOV and three dry processes for controlling NOV
A ' A X
alone. The dry processes all employ selective catalytic reduction (SCR) for controlling NO . Among
the dry processes, the Hitachi Zosen process employed a fixed bed reactor with "honeycomb" shaped
catalyst cells through which the flue gas passes in parallel flow; the Kurabo Knorca process
employed a moving bed reactor with spherically shaped catalyst located downstream of a hot ESP; and
the Shell Flue Gas Treatment process employed a fixed bed, parallel passage reactor in which the
catalyst is contained in unit, cells and the flue gas is forced across the face of the catalyst
layer.
The power plant assumed as a basis for this study was a new, 500 MW coal-fired boiler. The
coal had a heating value of 10,500 Btu/lb and contained 3.5 percent sulfur and 16 percent ash. The
plant efficiency was 9000 Btu/kWh and the boiler on-stream time was 7000 hr/yr.
In addition to the design premises for the.NO FGT process itself, the design premises for an
overall FGT system, including PM removal and FGD, were developed to allow for comparisons of the
various dry and wet processes. The design premises for the FGT system included the following
removal efficiencies:
- particulate: 99.5%;
- SOX: 90%; and ;
- NOX: 90% from a baseline of 600 ppm.
The capital investment estimates were based on mid-1979 construction costs. The annual
revenue requirements were based on mid-1980 operating costs using average capital charges with a
'7000 hr/yr on-stream time. Additional details on the design and economic premises are provided in
Reference 4-20.
4-39
-------
TABLE 4-14. NOY FGT PROCESSES SELECTED FOR EVALUATION
BY TVAXIN 1980 STUDY (Reference 4-20)
Process
Type
Dry NOV -only
A
Hitachi Zosen
Kurabo Knorca
UOP Shell Flue Gas
Treatment (SFGT-N)
Selective catalytic reduction
Selective catalytic reduction
Selective catalytic reduction
Dry SOX -
UOP Shell Flue Gas
Treatment (SFGT-SN)
Sorption of SO and selective
catalytic reduction of NO
A
Wet SOY -NOY
A > A
Asahi
Ishikawajima-Harima
Heavy Industries (IHI)
Moretana Calcium
Absorption-reduction
Oxidation-absorption-reduction
Oxidation-absorption-reduction
4-40
-------
The capital investment and annual revenue requirement estimates by TVA for these systems >are
presented in Tables 4-15 and 4-16, respectively. The capital investments for the various NO -only
FGT processes ranged from $38/kW to $48/kW. For the combined SCR-FGD-ESP systems, the capital
investments ranged from $165/kW to $175/kW. For the wet SO -NOV processes the capital investments
A X
ranged from $205/kW to $482/kW. The dry. SO -NO -ESP was $169/kW.
A A
The annualized costs of a dry SCR system ranged from 2.1 to 3.6 mills/kWh. The total system
costs of SCR combined with FGD and ESP ranged from 7.1 to 8.6 nrills/kWh. In comparison, the costs
of a dry SOX-NOX-ESP system were 7.5 mills/kWh and the costs of the wet SOX-NOX systems ranged from
12 to 20 mills/kWh.
The results of this study can be summarized as follows. The wet SO -NO processes do not
*
appear economically attractive for new power plant applications when compared with either the dry
SOX-NOX process or the SCR-FGD systems. Comparisons between the dry SFGT-SN process and SCR-FGD
systems were close enough to be inconclusive considering the state of development of these systems
at that time. Comparisons within the SCR-FGD systems, i.e., moving versus parallel flow, fixed bed
reactors were also inconclusive. However, recent trends in Japan indicate the fixed bed reactor
systems are preferred.
0_thjir; Stud1 es
Some other studies have also presented cost information for various FGT systems as summarized
below. '
A 1981 TVA study (Reference 4-6) made a preliminary economic evaluation of three control
methods for obtaining 50 percent NO reduction and three methods for obtaining 90% reduction. The
base case power plant was a new 500 MW coal-fired unit emitting 0.6 Ib NO-XIO Btu in the flue gas.
Capital investment estimates were based on projected mid-1982 construction costs. The revenue
requirements were based on projected 1984 costs.
The three 50 percent NO reduction processes evaluated were the EPA sponsored advanced low
NOX burner (ALNB), the Exxon Thermal DeNOx process, and the Hitachi Zosen SCR process. For
90 percent NO reduction the ALNB was combined with the Hitachi Zosen process, the Exxon process was
combined with the Hitachi Zosen process, and the Hitachi Zosen process was used alone. Capital
*
investments and annual revenue requirements for these processes are presented in Tables 4-17 and
4-18, respectively.
4-41
-------
TABLE 4-15. CAPITAL INVESTMENT DEVELOPED BY TVA FOR
ALTERNATIVE FGT SYSTEMS (Reference 4-20)
Total Capital Investment
Process
FGT
M$
FGD ESP
Total
$/kW
Total
$/aft3/nri n
Total
Dry NO -only
x\
UOP SFGT-N
Kurabo Knorca
Hitachi Zosen
Dry SOX-NOX
UOP SFGT-SN3
Wet SO -NOV
A /\
Moretana Calcium
Asahi
IHI
18.4
21.2
23.3
67.2
88.0
104.9
203.6
50.4 10.8
50.4 12.1
50.4 10.8
14.6
7.2
_ -
-
79.6
83.7
84.5
81.8
95.2
104.9
203.6
165
174
175
169
205
233
482
37.1
39.0
39.4
38.1
44.4
48.9
94.9
aBased on hydrogen production from naphtha and H9SOA byproduct from S09.
TABLE 4-16. ANNUAL REVENUE REQUIREMENTS DEVELOPED BY TVA FOR
ALTERNATIVE FGT SYSTEMS (Reference 4-20)
Annual Revenue
Requirements,
Process
FGT
M$
FGD
ESP
Total
Equivalent Unit
Revenue Requirements,
mills/kWh
Total
Dry NO -only
A
UOP SFGT-N
Kurabo Knorca
Hitachi Zosen
Dry SOX-NOX
UOP SFGT-SNa
Wet SO -NO
Mofetafia Calcium
Asahi
IHI
7.2
9.3
12.2
22.5
38.1
39.8
58.6
14.7
14.7
14.7
-
_
-
2.2
2.7
2.2
3.0
1.5
-
-
24.1
26.7
29.1
25.5
39.6 .
39.8
58.6
7.13
7.91
8.60
7.53
12.20
12.63
19.82
Based on hydrogen production from naphtha and
byproduct from
4-42
-------
TABLE 4-17. SUMMARY OF CAPITAL INVESTMENTS DEVELOPED
IN 1981 TVA STUDY (Reference 4-6)
Process
Capital Investment
(projected nrid-1982$)
M$
$/kW
50% NO Reduction
y\
ALNB
Exxon
Hitachi Zosen
90% NOV Reduction
A
ALNB/Hitachi Zosen
Exxon/Hitachi Zosen
Hitachi Zosen
2.4
9.9
15.7
25.9
32.1
25.5
4.8
19.7
31.4
51.8
64.2
50.9
TABLE 4-18. SUMMARY OF ANNUAL REVENUE REQUIREMENTS DEVELOPED
IN 1981 TVA STUDY (Reference 4-6)
Annual Revenue Requirements
(projected 1984$)
Process
M$
First Year
Mills/kWh
Mi
Level ized
Mills/kWh
50% NOV Reduction
X
ALNB
Exxon
Hitachi Zosen
90% NOV Reduction
X
ALNB/Hitachi
Zosen
Exxon/Hitachi
Zosen
Hitachi Zosen
0.45
3.4
8.0
11.5
14.2
13.3
0.17
1.2
2.9
4.2
5.2
4.9
0.54
5.1
13.0
18.4
22.6
21.9
0.20
1.9
4.7
6.7
8.2
7.9
4-43
-------
As expected, the ALNB, a combustion modification, is projected to be the least expensive NO
control method among those studied. Also, the costs for obtaining high levels of NO reduction
(90 percent) are significantly greater than for more moderate levels (50 percent).
In November 1982 the Electric Power Research Institute completed a detailed technical and
economic review of four FGT processes (Reference 4-31). The economic analysis was based on
preliminary designs for a 500 MW (net) capacity power plant located in the midwest and burning low
sulfur (0.5%) Powder River Coal. The results of the economic analysis are presented in Table 4-19.
As found in previous studies, non-catalytic processes are less expensive than catalytic
processes from both a net cost and a cost per ton of pollutant removed basis. Also, costs increase
significantly for higher levels of control and for, treating flue gases with higher NOX
concentrations.
In a recent paper (Reference 4-32), Ando reports that most coal-fired boilers in Japan will
use a flue gas cleaning system consisting of SCR using a honeycomb or plate catalyst, an ESP, and
FGD by the limestone-gypsum process using a prescrubber with a separate liquor loop. The costs of
generating power by coal, including, gas cleaning and wastewater treatment, is less than that by low
sulfur oil without gas cleaning. SCR is much less costly than FGD but is more costly than
combustion modification. Simultaneous SOV/NOV removal processes and fluidized bed combustion have
A A
not yet proved to be better alternatives to combined cleaning.
4.1.3 Energy and Environmental Impact
In addition to affecting the cost of operating electrical generating combustion equipment,
Implementing NO control techniques can also impact overall plant efficiency and emissions levels of
pollutants other than NO . These energy and environmental impacts are discussed below.
4.1.3.1 Energy Impact
This section discusses the energy impacts resulting or expected with various combustion
modification and flue gas treatment NO control techniques.
4-44
-------
TABLE 4-19a ESTIMATED TOTAL CAPITAL REQUIREMENT OF POST COMBUSTION' MOX CONTROL PROCESSES
(S/fcW, 1979) ...
NON-CATALYTIC
inlet NOx, PPm
NOx Reduction, %
Outlet NOx. PPm
Exxon Dual Grid
Exxcn - With «2
33
200
11
12
300
50
150
11
13
55
135
12
13
600
55
270
14
17
CATALYTIC
Inlet NOx. ppm
t4Qx Seduction, %
Outlet NOx. opm
Kawasaki H!
Hitachi Zosen
Shell FQT*
300
67
100
41
58
161
S3
50
42
85
169
90
30
46
68
163
600
§5
30
56
73
17S
TABLE 4-19b ESTIMATED LEVELIZED COST REQUIREMENT OF POST COMBUSTION NOX CONTROL PROCESSES
(MUls/kWh 1979)
NON-CATALYTIC
Inlet NOx. PPm
NOx Reduction, %
Outlet NOx, ppm
Exxon - Dual Grid
Exxon - With »2
300
33
200
0.96
0.97
50
150
1.12
1.18
55
135
1.19
1.24
600
55
270
1.68
1.85
CATALYTIC
Inlet NOx, pom
NOx Reduction, %
Outlet NO», ppm
Kawasaki HI
Hitachi Zosen
Shall FGT*
300
67
100
5.20
5.05
11.00
83
50
6.56
7.30
12.00
90
30
7.12
8.00
12.10
600
95
30
9.09
9.75
12.90
Accuracy; +30% - »0%
"Include* 20% Removal of 500 PPm SO2
(Reference 4-31)
4-4S
-------
Combustion Modifications
The following discussion of combustion modification energy Impacts nas taken largely from
Reference 4-33. The largest potential energy Impact of combustion :ioJi fixations is their effect
upon boiler thermal efficiency. Another significant source of energy "impact 1s the change 1n fan
power requirements caused by these '.ontrols. Boiler con'rol sy:tems Installed for low NO operation
also Increase electricity and instrument air requirements, but the energy Impact Is usually minimal.
Some discussion of the energy Impacts ot applied NOX controls was already presented on a boiler-by-
boner bas^s earlU.r In Section 4. As noted thore, with proper engineering and Implementation,
there should be no major adverse energy impacts with preferred combustion modifications. A review
of that analysis follows.
Applying low excess air (LEA) firing not only results In a small decrease 1n NO emissions
but also results in an increase 1n boiler efficiency through reduced sensible heat loss out the
stack. For this reason the technique has gained acceptance and das become more of a standard
operating procedure than a specific NO control method 1n both old and new units.
The other commonly applied combustion modifications, OSC and FGR, often lead to decreases in
bci'ler efficiency when implemented on a retrofit basis. OSC usually increases excess air
requirements res'iltinq In decreases in efficiency of up to 0.5 parcent. Unbu^ned fuel losses either
due to OSr. or FGR may cause a decrease in efficiency of up to O.b percent. If a substantial
increase in reheat steam attemperation is required due to OSC or FGR, cyclt; efficiency losses of up
to 1 percent may occur. Increased tan power requirements due to OSC or FGR will also i.npact
efficiency, resulting 1n losses of up to 0.? to 0.3 percent. No significant energy impact is
expected with LNB, either retrofit or new Installation.
Other combustion modification techniques, WI and RAP, ciir Impose quite significant energy
penalties on boiler operation, with decreases in efficiency from b to 10 percent. As a consequence,
these techniques have found little acceptance.
These decreases in aoiler etticiency (increases 1n energy consumption) discussed above for
the preferred NOX control techniques (OSC, FGR, and LNB) are mo t likely to occur when the
techniques are applied 01. a retrofit basis. These same combustion modifications are not expected to
adversely affect unit efficiency when included 1n the design of a new unit. Thus, with proper
engineering and development, combustion modification NO, controls can be Incorporated into new unit
designs with no significant adverse energy Impacts.
4-46
-------
A related problem that may occur witr retrofit application of some techniques 1s derating of
the unit. Loss 1n boiler load capacity due tc limited coal pulverizer capacity will occur 1n nwny
coal-fired boilers operated with burners-out-of-serv1ce (BOOS). Derating of 10 to 25 percent may
occur For o1l-f1ret boilers on OSC, higher excess air requirements nay causa fan capacity limits
to be reached 1n some cases. Although derating due to fan capacity is not common, reductions of up
+o 15 percent have been reported. With OSC and flue gas redrculation (FGR), excessive tube anu
steam temperatures may lead to derating, especially for gas fired boilers, and 1n some cases for
oil-fired boilers.
Flue Gas Treatment
Estimates of the energy Impacts of various dry NOX control systems i»i:d of dry and wet SOX/NOX
control systems have been presented in several sources (see References 4-31 and 4-34 through 4-38).
The following discussion focuses on the direct energy Impacts of FGT systems. However, many of
these systems utilize chem.cals such as ammonia which are energy intensive to manufacture
themse'ves.
The energy requirements of dry SCR NOX control systems result primarily from the e'ectrical
energy required to overcome the reactor pressure drop and the compressed air and steam usijd for soot
blowing. These requirements range from about 0.2 percent 01 the holler capacity for gas-fired units
to about 0.3 percent for coal-fired units.
When 5CR systems are used 1n combination with wet FGO systems and ESPs, total flue gas
cleaning energy rsquireiKnts for coal-fired boilers range from about 3.5 to 4 percent of the boiler
capacity. By comparison, recent estimates of the energy requirements of the Shell FGT process
achieving similar control levels (90 percent control of SOX and N0x) are about 5 percent of the
boiler capacity (Reference 4-36). Most of this energy consumption is in tne fo>m of fuel reqiired
by the process.
Fnergy requirement estimates of various wet SOX/NO control processes are considerably higher
- about 8 percent of the boiler capacity for the Moretana Calcium process, 11 percent for the Asahi
process, a d 19 percent for the IHI process (Reference 4-35). For the latter process, most of the
anetgy Is consumed in gt-neratlng ozone.
Enc "gy requirements for the £xxon Thsrmal OeNOx process have been estimated at about
0.4 percent of tne boiler capacity with most of the energy consumed by the large air compressors in
the ammonia storage and injection section (Reference 4-34). However, these estimates are for a
lower level of NOX control (about 50 percent) than the processes described above.
4-47
-------
Estimated energy requirements for achieving large reductions of SOX and NOX with the electron
beam process range from about 3 to 4 percent of the boiler capacity (References 4-32 and 4-371.
However, this process 1s still relatively far from commercial application.
4.1.3.2 Environmental Impact
This section discusses the environmental Impacts resulting or expected with various
combustion modification and flue gas treatment NOX control techniques,.
Combustion Modifications
The Environmental Protection Agency recently sponsored a throe year evaluation of combustion
modification controls for emissions of NOX and other pollutants from stationary combustion sources.
Some of the results of this program Included field tests of gaseous, liquid, and solid effluents
from seven stationary combustion sources and estimates of environmental effects of using combustion
modification control. Detailed results are presented 1n Reference 4-38. A few highlights
pertaining to utility and Industrial boilers are presented below.
Environmental assessment field testing was conducted on five boilers. Table 4-20 summarizes
the key aspects of these field tests. Test results were evaluated by comparing effluent stream
pollutant concentrations to discharge stream compositions desirable to preclude adverse effects to
human health. In addition to NO and SO^, potentially hazardous flue gas streair pollutants include
vapor phase SO, and condensed sulfate. organic acids, and trace elements such is As, Be, Cd, and V.
Coil-fired sources are generally more hazardous due to these pollutants. Potentially hazardous ash
stream pollutants from coal-fired sources are the trace elements Fe, Nn, Cr, N1, Be, Ba, Pb, and
occasionally As, Se, Tl, and Sn.
Conclusions evident from the field testing and analysis program were as follows.
(1) For the sources tested, the flue gas stream presents the
greateit potential environmental hazard.
(2) NOX and SCU are the potentiaMy most hazardous flue gas,
pollutants.
4-48
-------
TABLE 4-20. PARTIAL SUMMARY OF EPA/ACUREX COMBUSTION MODIFICATION
ENVIRONMENTAL ASSESSMENT FIELD TESTING (Reference 4-38)
Source Category
Coal-fired
Utility Boiler
Description
Kingston #6; 180 MW
tangential; twin
furnace, 12 burners/
Test Points
(Ur.it Operation)
Baseline
Biased Firing (2)
BOOS (2)
Sampling Protocol
Continuous NO , SO,,
CO, CO, 0 * i
Inlet to^lst ESP:
lest
Collaborator
TVA
furnace, 3 elevations;
cyclone, 2 ESP's for
particulate control
SASS
Method 5
Method 8
-- Gas grab (C.-C,
HC)
.-,
Outlet ot 1st ESP0
-- SASS
-- Method 5
- Method 8
Gas grab (C.-C, HC)
Button ash *
Hopper ash (1st ESP,
cyclone)
Fuel
Operating data
Coal -fired
Utility Boiler
Oil-fired
Utility Boiler
Crist 17; 500 MW Baseline
opposed wall-fired; DOOS (2)
24 burners, 3 eleva-
tions; ESP for
partlculate control
Moss Landing 16 aselfne
740 MW Opposeo wall- .UR
fired; 48 burners, FGfi * OF A
6 elevations
Continuous NO., CO, Exxon
co2, o x
ESP tnlet
SASS
Method 5
- Method 8
Gas grab (C.-C, HC)
ESP Outlet l °
SASS
Method 5
Method 8
Gas grab (C,-C HC)
Botton ash l
FSP hopper ash
Fuel
Operating data
Bioass^y
Continuous NO , CO, ' None
co2. o, x
- SASS
Method b
Method 8
Gas grab (C,-CK HC)
Fuel '- 5
Operating Data
Bioassay
Key to ?crony»s: (see end of table).
-------
(ABLE 4-20. (Continued)
Source Category
Description
Test Points
(Unit Operation)
Sampling Protocol
Test
collaborator
Coal-fired
Industrial Boiler
Traveling grate
spreader stoker,
38 kg/s (300,000 Ib
steam/hr)
Baseline
LEA + high OFA
Continuous HU , CO,
C0?, 0? x
Boilfer exit:
SASS
~ Method 5
Shell-Emeryville
-- Gas grab (C,-C,. HC)
ESP outlet l °
SASS
~ Method 5
Shell-Emeryville
Gas grab (C,-Cfi HCJ
Bottom ash x °
Cyclone hopper ash
Fuel
Operating Data
KVB
Coal-fired
Industrial Boiler
01
o
Traveling grate
spreader stoker,
25 kg/s (200,000 Ib
steam/fir), ESP for
particulate control
Baseline
LEA
Continuous NO , CO,
co2, o2 x
Boiler exit:
~ SASS
Method 5
Shell-Emeryville
Gas grab (C.-C,- HC)
ESP outlet * °
~ SASS
~ Method 5
-- Shell-Emeryville
Gas grab (C^C,- HC)
Bottom ash x °
ESP hopper ash
Fuel
Operating Data
Bioassay
KVB
Key to acronyms:
BOOS: btaged combustion technique with burners out ot service.
ESP: Electrostatic precipitator.
FGR: Flue gas recirculation through burner windbox.
LEA: Low excess air firing.
OFA: Staged combustion with overfire air injection.
SASS: Source assessment sampling system for organic and inorganic emission collection and measurement.
-------
(3) The total flue gas hazard is decreased or, at worst, does
not increase with applying the combustion modifications
tested. Changes in emissions due to day-to-day fuel
composition changes are often of greater magnitude than
those attributable to NO,, control.
X
(4) The effluent streams from the sources tested are not
mutagenic, and, in general, elicit nondetectable toxicity
in bioassay testing.
(5) The combustion modifications tested:
Have no effect, or increase only slightly, emissions
of CO and vapor phase hydrocarbon;
~ Have no effect on particulate mass emissions;
Have no effect, or tend to increase slightly, emitted
particle size distribution;
~ Have no measurable effect on inorganic trace element>
emissions or on trace element partitioning tendencies;
Have no effect, or decrease slightly S03 and particulate
sulfate emissions;
Have little effect on total higher molecular weight
organic emissions; and
Marginally increase POM emissions; however, emission
levels remained on the order of the detection levels .
of the instrument.
(6) Emissions of many organic priority pollutants were below the
detection limit for the sources tested.
It was further concluded in the program that the NO control methods investigated are
environmentally sound since the potential adverse impact of the controlled source is either
decreased or unchanged with NOV control applications. However, it should be pointed out that these
A
conclusions are based on short term tests conducted under steady operating conditions, and that the
controls tested were the relatively straight forward, current technology combustion modifications.
4-51
-------
Flue Gas Treatment
Since F6T processes have been less extensively applied than combustion modifications, their
environmental impacts have been less extensively tested. However, the principal impacts expected
can be qualitatively assessed. The following discussion presents some assessments of these impacts
and the results of some pilot unit testing sponsored by EPA. The discussion focuses primarily on
SCR processes since these are the most extensively developed and applied F6T processes.
The principal impacts expected to result from the application of SCR processes are as
follows. NOV emissions in the flue gas are reduced considerably, but some emissions of ammonia and
A
ammonium compounds will result and some increases in SO, or other compounds are possible. Also, in
some cases wastewater treatment and disposal may be complicated by the addition of nitrogen
containing compounds. An additional impact with catalytic systems is the need for catalyst
disposal.
By way of comparison, SNR processes are less efficient at removing NO from the flue gas.
X
And the increased ammonia injection levels will result in greater emissions of ammonia and ammonium
compounds than the SCR processes, potentially leading to greater wastewater impacts as well.
However, these processes will not likely increase flue gas SO, levels and they are not faced with
the need for catalyst disposal.
In a recent report covering a survey of SCR systems in Japan (Reference 4-39), Jones
evaluates potential SCR environmental impacts as follows. NO emissions from boilers utilizing
catalytic de-NO are generally reduced by 80 percent. Higher reductions are possible, but costs
will be greater for these units. NH, emissions for oil- and gas-fired applications are reported to
be about 3 to 10 ppm. Emissions from full-scale, coal-fired facilities will not be known until
after the start-up of several units currently under construction. In addition to concern about NH,
emissions, there is some concern that other compounds, such as cyanides, nitrosoamines and nitrates,
may also be emitted. SCR system vendors and operators, however, were not aware of any instances
where compounds such as these were emitted. The possibility of a visible ammonium sulfite plume
resulting from NH, emissions entering a downstream, wet FGD system may be a problem if the slip NH,
is high (>50 ppm). Plumes of ammonium sulfite are known to occur during certain atmospheric
conditions when NH_ based FGD systems are used. However, the Japanese have experience with
situations in which 10 ppm of NH3 enters the scrubber and, based on this experience, do not feel
that SCR systems will cause visible plume formation.
4-52
-------
Jones goes on to discuss other potential environmental impacts including nitrogen compounds
in the wastewater and catalyst disposal. NH, can enter the wastewater through the F6D system or
from air preheater washwater. In locations where discharge of this water will cause problems, an
activated sludge technique can be used to treat the wastewater. In other locations it may be
possible to blend wastewater containing NH, with other water discharges. The catalyst disposal
issue has not been fully addressed. While the process vendors indicate that they will dispose of
spent catalyst, the specific method of disposal had not been identified at the time of the survey.
This is partially due to the fact that none of the full scale systems had required a catalyst
change.
As discussed earlier in Section 4, EPA recently sponsored pilot plant tests of two SCR
processes applied to coal-fired utility boilers. These included the Hitachi Zosen and Shell Flue
Gas Treatment processes. As part of the pilot plant tests, EPA conducted a sampling program
designed to quantify process emissions of pollutants other than NO and S0«. The test and
evaluation program is discussed in Reference 4-13. Parts of that discussion, pertaining to the
environmental impacts of these processes, are presented below.
The objective of the sampling program was to determine if any adverse flue gas concentration
impacts can result from application of SCR technology to a coal-fired source. Samples were
collected at the inlet and outlet of the pilot plant reactors and analyzed for selected pollutants.
Inlet and outlet samples were collected simultaneously, so that measured differences in pollutant
concentration were due to changes across the reactors.
Table 4-21 identifies the specific pollutants measured and presents the results of the
sampling program. As shown, the concentrations of some pollutants changed across the reactor while
others did not. And in some cases, the concentrations of pollutants were below the detection limits
of the analytical techniques employed during the sampling program.
As illustrated in Table 4-21, concentrations of hydrogen cyanide and nitrosoamines at the
outlet of the pilot plant reactors were below the detection limits of the analytical techniques
employed. For hydrogen cyanide, the detection limit is equivalent to 10 ppb and for nitrosoamines a
maximum of 2 ppb (the actual concentration level depends on the nitrosoamine compound(s) present).
4-53
-------
TABLE 4-21. MEASURED CHANGE IN THE CONCENTRATIONS OF SELECTED POLLUTANTS
ACROSS THE EPA PILOT PLANT SCR REACTORS (Reference 4-13)
I
in
Pollutant
Hydrogen Cyanide
Nitrosoaminesa
Carbon Monoxide
Hydrocarbons (C, - Cg)
Sulfur Tri oxide
Ammonia
Hitachi Zosen
Reactor Inlet
Concentration
<10 ug/m
3
<5 yg/m
<0.017*
<1.0 ppmv
8.4 ppmv
0
Pilot Plant
Reactor Outlet
Concentration
<10 yg/m
<5 ug/m
<0.017%
<1.0 ppmv
20.7 ppmv
54.8 ppmv
Shell
Treatment
Reactor Inlet
Concentration
<10 yg/m3
<5 yg/m
0.13%
28.5 ppmv
11.4 ppmv
0
Flue Gas
Pilot Plant
Reactor Outlet
Concentration
<10 yg/m
<5 yg/m
<0.017«
21.0 ppmv
0.1 ppmv
15.3 ppmv
Gas volume taken at i5.5°C, 1 atm, dry basis.
-------
At the Hitachi Zosen plant, the concentrations "f both CO and hydrocarbons were also below the
detection limits of the analytical techniques. But at the Shell FGT pilot plant, the concentrations
of these compounds were found to decrease across the reactor. This decrease represents an environ-
mental benefit for the Shell process, and 1t 1s believed to result from oxidation of these compounds
1n the reactor. It should be noted that the Hitachi Zosen reactor may also oxidize CO and hydro-
carbons, but no conclusions can ue drawn since those compounds were present 1n such low
concentrations.
One of the most significant results of the emissions sampling program was the effect of the
Hitachi loicn and Shell processes on the concentration of SO* In the flue gas. As shown In
Table 4-21, SO, was produced In the Hitachi Zosen pilot plant reactor while the Shell FGT process
removed SO, from the gas stream. These results are significant because of the effects SO- can have
on equipment 'located downstream of an SCR reactor, especially the air preheater.
Another significant result of the er Mons sampling program is the measured NH~ emissions from
the processes. As shown in Table 4-21, NH. emissions from the Hitachi Zosen process were over three
times greater than the emissions from the Shell FGT process even though the NH./NO Injection ratio
was higher at the Shell pilot plant. The higher NH, emissions are due to the fact that the catalyst
used 1n the Shell process promotes NH., oxidation while essentially i.b NH, oxidation occurred 1n
Hitachi Zosen process.
4.2 INDUSTRIAL BOILERS
Industrial boilers are a /ery common piece of equipment in industrial plants. These boilers
typically range 1n size from 3 to 250 MW thermal inpi.'t (10 to 850 x 10 Btu/hr) am< Include a wide
variety of firing types and fuels. In 1980 industrial boilers represented the second largest
stationary source of NOX emissions, preceded only by electrical utility boilers (See Chapter 2).
Nationwide NOX emissions from Industrial boilers were estimated to be 3 Tg (3.3 x 10 tons)
annually.
Industrial boilers are topically classified by the type of firing mechanism employed, the
heat transfer mechanism, and Che type of fuel fired. Firing mechanisms Include either burners,
spreader-fed, or mass-fed. With burners, the fuel is Injected Into the boiler through a nozzle anO
burns while suspended within the boiler combustion chamber. Mass-fed arid spreader-fed boilers are
4-55
-------
used for most solid fuel industrial boilers. They combust the fuel on a grate in the boiler.
Watertube is the most common mechanism used for heat transfer in industrial boilers. In
watertube boilers the water for steam generation is contained in banks of tubes suspended in the
boiler combustion chamber and flue. Firetube boilers invert this configuration and pass hot flue
gases through tubes suspended in a water drum. Firetube boilers are seldom sized larger than 7.3 MW
(25 x 10 Btu/hr) thermal input (Reference 4-40).
Industrial boilers,are fired with a wide variety of fossil and nonfossil fuels. Most common
among the fossil fuels are natural gas, accounting for 43 percent of the fossil fuel-fired
Industrial boiler capacity, followed by fuel oil and coal comprising 32 percent and 25 percent,
respectively. Nonfossil fuels fired in industrial boilers include wood, bark, agricultural waste,
municipal waste, and industrial waste; the most common being wood and bark. Nonfossil fuels account
for less than 5 percent of the industrial boiler capacity.
The following discussion on NO emission control techniques for industrial boilers will focus
on fossil fuel-fired boilers below 73 MW (250 x 10 Btu/hr) in size. Those boilers greater than
approximately 73 MW are essentially identical to utility boilers and are able to apply the same NO
emission control technologies that are addressed in Section 4.1. Additionally, non-fossil fuels
generally exhibit relatively low NOV emissions with respect to solid fuels resulting in a lack of
A
demonstrated NO emission control technologies for non-fossil fuels (Reference 4-41). The
population distribution of U.S. fossil-fired watertube boilers by size range is presented in
Table 4-22. The corresponding material for U.S. firetube boilers is presented in Table 4-23.
4.2.1 Control Techniques
Currently, the most promising NOV control options for industrial boilers include combustion
X
modification and post combustion techniques. The former has been the most successful and widely
used option, and is described below for gas-, oil- and coal-fired units. The post combustion
techniques are discussed after combustion modification and are currently being demonstrated for some
industrial boiler applications.
4-56
-------
TABLE 4-22. INSTALLED CAPACITY OF U.S. HATERTUBE INDUSTRIAL BOILERS
BY UNIT SIZE AND FUEL TYPE (IN 1977)
(MW Thermal Input (106 Btu/hr))
Fuel
(0 to 2.9
(0 to 10)
2.9 to 14.7
(10 to 50)
Capacity by
14.7 to 29.3
(50 to 100)
unit size
29.3 to 73.3
(100 to 250)
>73.3
(>250)
Totals
ulverized coal
Spreader-stoker coal
Inderfeed-stoker coal
Dverfeed-stoker coal
Fotal Coal
Residual oil
Distillate oil
Total Oil
Natural gas
Total all fuels
0
70
(240)
680
(2,300)
85
(290)
835
(2,830)
3,960
(13,500)
2,560
(8,700)
6,520
(22,200)
4,475
(15,300)
11,830
(40,330)
0
(0)
4,650
(15,900)
14,105
(48,000)
3,470
(11,800)
22,225
(75,700)
48,190
(164,000)
8,280
(28,200)
56,470
(192,200)
57,900
(197,500)
136,595
(465,560)
0
(0)
6,175
(21,060)
17,265
(58,900)
4,455
(15,200)
27,895
(95,160)
36,640
(122,000)
4,295
(14,600)
39,935
(136,600)
53,585
(182,800)
121 ,415
(414,560)
19,895
(67,800)
20,295
(69,000)
7,000
(24,200)
3,555
(12,100)
50,825
(173,100)
44,790
(153,000)
6,370
(21,700)
51,160
(174,700)
63,320
(216,000)
165,305
(563,800)
40,180
(,137,000)
11,010
(37,600)
5,230
(17,800)
3,510
(12,000)
59,930
(204,400)
43,570
(148,600)
4,085
(13,900)
47,655
(162,500)
95,935
(327,200)
203,520
(694,100)
60,075
(204,800)
42,200
(143,800)
44,360
(151,200)
15,075
(51,390
. 161,710
(551,190)
176,150
(601,100)
25,590
(87,100)
201,740
(688,200)
275,215
(938,800)
638,665
(2,178,190)
I
in
Reference 4-40.
-------
TABLE 4-23. INSTALLED CAPACITY OF U.S. INDUSTRIAL FIRE-TUBE BOILERS
BY UNIT SIZE AND FUEL TYPE (1977)
(MM Thermal Input (106 Btu/hr))
Fuel
Coal
tesidual Oil
)i still ate Oil
latural Gas
Total All Fuels
Unit
0.1 to 0.4
(0.4 to 1.5)
1,690
(5,700)
8,960
(30,600)
4,160
(14,200)
15,420
(52,600)
30,230
(103,100)
Capacity, MW
0.4 to 2.9
(1.5 to 10)
3,960
(13,500)
26,320
(89,800)
13,610
(46,400)
43,700
(149,100)
87,590
(298,800)
Thermal (106
Btu/h)
2.9 to 7.3 7.3 to 14.7
(10 to 25) (25 to 50)
4,950
(16,900)
19,280
(65,800)
11,760
(40,200)
37,270
(127,200)
73,260
(250,100)
2,830
(9,600)
6,580
(22,500)
4,010
(13,700)
9,230
(50,700)
28,280
(96,500)
Totals
13,430
(45,700)
61,140
(208,700)
33,540
(114,500)
111,250 .
(379,600)
219,360
(748,500)
I
in
00
Reference 4-40.
-------
4.2.1.1 Combustion Modification
Combustion modification control techniques reduce the formation of NOX emissions by altering
the combustion conditions present in the combustion chamber. These techniques include modifications
to the fuel and combustion air feed systems, and modifications to the combustion chamber design.
One of the most extensive sets of data on combustion modification technique performance was
derived in an EPA-sponsored study involving the testing of 65 boilers. Ten different combustion
modification techniques were implemented resulting in a total of 116 test runs. The effects of
these techniques on NO emissions and boiler efficiency are summarized in Figure 4-2 for 73 separate
boiler tests (Reference 4-42).
The graph is divided into quadrants. The criterion for the best quadrant is that the
modification technique should simultaneously reduce NOX and increase efficiency. In general, the
study showed that total NO emission reductions of up to 47 percent were possible by using one or a
A I
combination of five different methods. These methods were: excess air reduction, burner out of
service, flue gas recirculation, overfire air addition, and reduced air preheat. In the first three
methods boiler efficiency was generally unimpaired.
Since the original combustion modification study, the EPA has continued to study all of the
above techniques and several newly developed techniques. The most promissing of the recently
developed NO combustion modification techniques is low !M burners for oil- and gas-fired boilers.
" X
Tables 4-24 and 4-25 summarize the results of EPA's studies on the performance of combustion
modification techniques on gas-, oiU, and coal-fired boilers. The remainder of this section is
devoted to the discussion of combustion modification experience on these boilers, including control
efficiency, operational problems, and applicability.
Gas- and Oil-Fired Boilers
Combustion modification controls have been most successful in the reduction of NOV emissions
X
from gas- and oil-fired industrial boilers. In large part, this success can be attributed to the
greater flexibility of fluid fuels with respect to alterations in fuel firing conditions.
4-59
-------
+75
WORST QUADRANT
+50
~ +25
x
O
a
o
oc
2
o
tu
o
4
-25
so
75
-1C
Figure 4-2.
COMBUSTION MODIFICATION METHOD
O FLUE GAS RECIRC.
O AIR REGISTER ADJ.
A OIL VISCOSITY
O BURNERTUNEUP
V ATOMIZATION PRESSURE
ATOMIZATION METHOD
REDUCED EXCESS AIR
A OVERRRE AIR
REDUCED AIR PREHEAT
T BURNI-R-OUT-OF-SERVICE
a
BEST QUADRANT
+10
-50+5
CHANGE IN EFFICIENCY, percent
Effect of combustion modification methods on total nitrogen
oxides emissions and boiler efficiency (Reference 4-42.)
4-60
-------
TABLE 4-x/l. N0x EMISSION CONTROL TECHNIQUES FOR GAS-FIRED AND OIL-FIRED INDUSTRIAL BOILERS
Effectiveness
X N0x Reduction)
Operational Impact
Applicability
Commercial Availability/R & D Status
Low Excess Air (LEA)
(Gas) 0-30
(d. oil) 0-30
(r. oil) 0-20
Increased boiler efficiency
Applicable to all gas and oil
industrial boilers. Generally
stack 0« can be reduced to 1-2%
for gas, ?% for dist. oil, and
yt> for resid. oil. Developing
burners will allow lower stack
V
Method well demonstrated and control
equipment commercially available for
all boiler types.
Over-Fire Air
Ports (OFA)
(Gas) 25-45
(d. oil) 20-40
(r. oil) 20-50
Slight decrease in boiler
efficiency of 0 to 3%. The
decrease can be mitigated
in part with the combined
use of LEA controls.
Applicable to all gas and oil
industrial boilers. Generally
70%-90% burner stoichiometries
can be used with proper instal-
lation of secondary air ports.
Best implemented on new units. Not
commercially available for all design
types especially fire tubes. Retrofit
not feasible for most units, especially
packaged units.
Burners Out of
Service (BOOS)
(Gas) 25-45
(d. oil) N/A
(r. oil) 10-40
Perhaps slight decrease in
boiler efficiency. The
decrease can be mitigated
with the combined use of
LEA controls. May require
derating unless fuel
delivery system is
modified.
Applicable only to multi-burner
boilers. Best suited to square
burner pattern.
Commercially available. Retrofit
application only. Not demonstrated for
distillate oil.
Low-N0x Burners (LNB)
20-50
May potentially require
larger fire box area.
Retrofit application may
require derating due to
larger flame pattern.
New burners described generally
applicable to all boilers.
More specific information
needed.
Commercially offered for burner sizes
up to 150 x 10 Btu/hr. Only demon-
strated for a limited range of boiler
types at this time.
Flue Gas Recirculation
(FGR)
Gas) 45-75
d. oil) 40-70
r. oil) 15-20
Possible flame instability
and fan errosion problems
which can be reduced
with proper engineering.
Applicable to all design types
except gas ring burners.
Commercially available for most boiler
types but best suited for new boilers
because retrofit would result in possi-
bly extensive modifications.
Deduced Air Preheat
(RAP)
(Gas) up to 55%
(d. oil) up to
453;
(r. oil) UD to
20%
Depending on
amount of pre-
heat.
Significant loss in boiler
efficiency unless compen-
sated for by use of feed
water economizer.
Applicable to all design types.
Comraerically available but best suited
to new boilers where designs can be
modified to include feed water
economizers.
Hmmonia Injection
40-70%
Possible implementation
problems: fouling and
corrosion problems with
high sulfur oils. Close
operator attention required.
Appears most applicable to gas
and low sulfur oils.
Commercially available but very limited
demonstration (in Japan only.) Best
suited to base load boilers.
selective Catalytic
deduction (SCR)
70-90
No impact on efficiency.
Fouling and corrosion
problems with high sulfur
oils due to ammonium
sulfate salts. Close
operator attention required.
Appears applicable to all gas-
and oil-fired boilers, although
high -sulfur oils may pose
greater operational problems.
Commercially available but
limited demonstration (in Japan only.)
-------
TABLE 4-25. NOX EMISSION CONTROL TECHNIQUES FPR COAL-FIRED STOKER INDUSTRIAL BOILERS
Low Excess Air (LEA)
Staged Combustloi Air
(SCA)
Reduced Air Preheat
(RAP)
Aaino.ila Injection
Selective Catalytic
Reduction (SCR)
Effectiveness
(t Nt>x Reduction)
5-25
not well
defined
8
40-60
70-90
Operational Impact
Increased boiler efficiency.
Close operator attention
required to prevent grate
overheating and clinker
formation.
Possible SOM; decrease 1n
efficiency. Close operator
attention required to prevent
grate overheat and cUnker
formation.
Significant loss in boiler
efficiency unless compen-
sated for by use of feed-
Mater ecoiunizers.
Possible i^jlementaMon
difficulties, fouling
problems with high sulfur
fuels, ioac restrictions.
Close operator attention
required.
No Impact on efficiency.
Fouling and corrosion
problems with high sulfur
coals due to iinnonlum sulfate
salts and fly ssh. Close
operator attention required.
Applicability
Applicable to all stokers.
Generally stack 0. can be
reduced to 4-61. L
Most stokers already equipped
with OFA ports.
Applicable to boilers with
combustion air preheat ers.
Appears most applicable to
low sulfur fuels.
Appears applicable to coal-
fired boilers.
Commercial Avallablllty/R 1 0 Status
Commercially available and performance
well demonstrated. Minimum 0, level
ay not > Identified for some boiler
types.
Commercially available but further
research required to identify optimum
Of A port position and performance
levels.
Commercially available but not signifi-
cantly effective. Best suited for new
boilers where design can be modified to
Include feedwater economizers.
CoMerclally available but not demon-
strated on stoker boilers.
Commercially available, but very
limited demonstration in Japan only.
Significant questions persist about
operational problems on coal -ft red
boilers.
-------
Low Excess Air - Low excess air (LEA) operation has proven to be ex'i-emely effective 1n
lowering NO emissions from gas and distillate oil-fired Industrial boilers. As discussed in
Chapter 3, LEA controls are most effective 1n reducing thermal-NO emissions, which compose the
major fraction of the NO emissions from these two fuels.
An EPA study of 213 gas-fired, 60 distillate-fired, and 148 residual-fired boilers revealed
DX emissions from these sources could t
the following correlations (Reference 4-431:
that NO emissions from these sources could be related to the quantity of excess air present using
Eg - 0.079 «0.20T1.!1,0.17
EdQ - 0.32 H°-46 T°'31 A0'29
Ero . 24.2 T°'34 A0'24 0.055 N1'06
,1? E » total NO emissions for gas cnnbustlon adjusted to 3 percent
9 02 and dry basis (ppm)
E. » total NO emissions for distillate oil combustion adjusted
to 3 percent 02 and dry basis (ppm)
E * total NOV emissions from residual oil combustion adjusted
to 3 percent 02 and dr/ basis (ppm)
H * combustion zone heat release rate (10" Btu/hr - ft")
T * combustion air preheat temperatura °R
A « flue gas oxygen concentration (mole fraction)
N * fuel nitrogen content (Ib N/10" Btu)
These correlations are presented in Figure 4-3 as a plot of flue gaj 02-vs-NOx for natural gas,
distillate oil, and a 0.3 wt percent nitrogen residual fuel oil.
Based on these correlations, reducing excess air from a typical flue gas 02 level of
5 percent down to 1 percent 1n a gas-fired boiler will result in a NO emission reduction of
24 percent. Correspondingly, a reduction 1n excess air from a flue gas 0- level of 5 percent down
to 2 percent in a distillate oil-fired boiler will reduce NO emissions by 23 percent. Residual
oil-fired boilers with their characteristically high fuel NO emissions exhibit a much lower
reduction with LEA controls. Reducing excess air from a typical flue gas 0,, level of 6 percent down
to 3 percent 1n a residual oil-fired boiler combusting a 0.3 wt percent, nitrogen oil will result in
a N0x emission reduction of only 6 percent. Table 4-26 presents both the typical and the minimum
flue gas 02 levels applicable to gas- and oil-fired boilers.
4-63
-------
0.4
0.3
0.1
RESIDUAL OIL
0.3 wt * N
No A1r Preheat
DISTILLATE OIL
No Air Preheat
Full Load
NATURAL GAS
No Air Preheat
Full Load
Flue Gas Oxygen (;)
Figure 4-3. Effects of flue gas oxygen on the formation of NO
emissions from gas- and oil-fired boilers (Reference 4-43)
4-54
-------
TABLE 4-26. SAFE OPERATING LEVELS FOR LEA (Reference 4-45)
MINIMUM FLUE GAS 0,
FUEL/FIRING TYPE (percent) z
Natural Gas 0.5-3
Distillate and 2-4
Residual 011
TYPICAL FLUE GAS 0«
(percent)
4 - 8
4 - 8
EPA studies of LEA controls concluded that these controls can be applied to some degree on
all gas- and oil-fined boilers. However, the lowest excess air levels can be achieved with newer
LEA burners which Incorporate design feature* 'emit complete fuel combustion at very low air
levels. These studies also recommend the use of v. gen trim systems as an Integral part of all LEA
control system to maintain a minimum but safe excess air level. If air levels are allowed to drop
too low. Incomplete combustion can occur, resulting 1n Increased emissions of hydrocarbons, carbon
monoxide, and smoke (Preference 4-44).
Staged Combustion A1r (SCA) - These controls reduce NO emissions by selectively staging tha
Introduction of combustion air Into the combustion zone. On gas- and oil-fired boilers staged
combustion air (SCA) controls can be employed by two techniques, burners-out-of-service vBOOS) and
overfire air ports (OFA). These techniques are described in Chapter 3.
EPA conducted a study of five watsrtube gas-fired boilers, four of which were equipped with
combustion air preheat. Without SJA controls the NO emissions from these boilers ranged from
103 to 142 ng/J (24 to 33 lb/106 3tu) and averaged 120 ng/J (28 lb/106 Btu). With the application
of SCA controls, the NOX emissions from these boilers ranged from 60 to 129 ng/J (14 to
30 lb/10 Btu) and averaged 82 ng/J (19 lb/10 Btu). These data demonstrate an average emission
reduction of 33 percent (Reference 4-46).
An EPA study of a firetube boiler combusting natural gas documented a 25 percent reduction in
NOX emissions using SCA controls. This reduction was achieved with a burner air stoichicmtetry of
90 percet.t ano a flue gas oxygen concentrecion of 2.9 percent (Reference 4-47).
4-65
-------
SCA controls demonstrate their greatest NO emission reduction efficiencies on residual
X
oil-fired boilers because of their high effectiveness on fuel-NO emissions. EPA studies on two
packaged watertube boilers firing residual oil documented NOX emission reductions of 40-45 percent
using OFA ports to achieve SCA controls (Reference 4-48). Numerous tests on field-erected watertube
boilers firing residual oil documented NO emission reductions of 25-40 percent (References 4-42 and
4-49). This latter set of boilers used BOOS techniques to achieve SCA controls.
Very little data is available on the performance of SCA controls on distillate oil-fired
boilers, or on residual oil-fired firetube boilers.
Operational Impacts of SCA controls include possible flame stability problems and boiler
derating for retrofit applications. Flame stability problems can be mitigated by the proper
location of OFA ports and by the use of commercially available air flow controls which maintain the
required staged air injection and burner combustion air flowrates throughout the boiler load range
(Reference 4-50).
With retrofit applications where BOOS controls are being applied, some boiler derating may
occur due to the size limit of the burners which remain in service. This problem can generally be
eliminated by the installation of larger burners and by modification of the air registers to allow
greater air supply to the burners remaining in service (Reference 4-44).
Flue Gas Recirculation - A third technique for NOX control by combustion modification is flue
gas redrculation (FGR). This technique involves extracting a portion of the flue gas and returning
it to the furnace through the burner windbox. FGR suppresses NO formation by diluting the 0, level
X £*
in the combustion zone and by reducing peak flame temperatures with the heat absorptive capacity of
the recirculated gas. FGR primarily reduces-thermal NO and is, consequently, most effective when
applied to gas- and distillate-fired boilers.
EPA sponsored FGR tests on two gas-fired watertube industrial boilers. On a 73 MW (250 x 10
Btu/hr) boiler NO emissions were reduced an average of 70 percent at a flue gas recirculation rate
of 45 percent (Reference 4-47). The second test was conducted on a 5 MW (17 x 10 Btu/hr) boiler
and achieved an average 75 percent NO emission reduction with a 20 percent flue gas recirculation .
rate (Reference 4-48).
4-66
-------
F6R controls demonstrated similar NO emission reduction capabilities on two long-term tests
conducted by EPA, reaching a maximum control efficiency of 70 percent (Reference 4-48).
The results of a multi-fuel test on a 5.1 MW (17.5 x 10 Btu/hr) packaged watertube boiler
are presented in Figure 4-4. This series of tests demonstrates both the relative performance of F6R
controls on various fuels as well as the impact of varying the flue gas recirculaton rate. The
greater impact of FGR on NOV emissions from gas and distillate oil combustion is clearly
A
demonstrated in these results. More specifically, these test data showed that when firing natural
gas, FGR reduced NO emissions from 28 ng/J (0.07 lb/10 Btu) at normal operating conditions to an
X
average of 13 ng/J (0.03 lb/10 Btu). These data for gas combustion represent a NO emission
reduction of 53 percent. When the same boiler was tested combusting a 0.14 percent nitrogen
residual fuel oil, FGR only reduced NO emissions from a LEA emission level of 95 ng/J
(0.22 lb/106 Btu) to a FGR emission level of 78 ng/J (0.18 lb/106 Btu). (References 4-42 and 4-46).
Ther.e are two operational problems associated with the application of FGR controls; flame
stability problems and errosion of the recirculation fan blades.. Flame stability problems will
necessitate the use of different burner configurations and the use of flame sensors to detect the
onset of instability problems. Proper fan design will reduce fan'problems. It has also been
observed that FGR controls may not be easily retrofitted to small industrial boilers due to space
limitations on the extra ducting requirements (Reference 4-44).
Reduced Air Preheat - By reducing the amount of preheating applied to combustion air, a
reduction of the peak flame temperature can be achieved in the combustion zone. This reduction in
turn lowers thermal NO production. Most industrial watertube boilers with design heat input
capacities greater than 15 MW (50 x 10 Btu/hr) recover some flue gas heat in combustion air
preheaters or feedwater economizers to maximize thermal efficiency (Reference 4-42). The
installation of an economizer to replace or reduce the use of a combustion air preheater will result
in lower peak temperatures while still allowing for effective flue gas heat recovery
(Reference 4-51). Lowering peak temperature is primarily effective for reducing thermal NO , but
has little effect on fuel NO . Hence, the technique of reduced combustion air preheat will result
in higher percent reductions for low nitrogen fuels distillate oil and natural gas
(Reference 4-52). .
.4-67
-------
100
80
-------
The impact of combustion air temperature on a typical gas-, distillate oil-, and residual
oil-fired boiler is presented in Figure 4-5. These correlations were developed from the study of
over 400 short-term tests on industrial watertube boilers. The results of this-study conclude that
a"222°C (400°F) reduction in combustion air preheat would reduce NO emissions from natural gas
combustion by 44 percent, from distillate oil combustion by 33 percent, and from residual oil
combustion by 7 percent. Within the data base used in this study, combustion air preheat
temperatures typically ranged from 149°C (300°F) up to 360°C (680°F) (Reference 4-43).
Although RAP is applicable to all gas- and oil-fired boiler types it may'pose a significant
energy penalty for certain applications. New boiler applications can generally recover waste flue
gas heat by use of feed water economizers. However, situations do exist where feed water
temperatures are returned to the boiler at temperatures too high to make the use of feed water
economizers practical. In many retrofit applications, feed water economizers are not already in
place, and the retrofit of feed water economizers is not practical for most industrial boilers
(Reference 4-43).
Low NO Burners (LNB) - New burner designs are being developed for industrial boilers which
alter the mixing of air, fuel, and combustion products within the burner flame zone to reduce NO
emissions. At this time LNB controls have seen very limited application, and are commercially
available for only a limited range of industrial boiler types.
In a 30-day test of a 30 MW gas-fired boiler, LNB controls reduced NO emissions from
113 ng/J (0.26 lb/106 Btu) down to 33 ng/J (0.08 lb/106 Btu) under full load conditions. When
tested at partial load conditions, LNB controls reduced NO emissions from 95 ng/J (0.22 lb/10 Btu)
down to 44 ng/J (0.10 lb/10 Btu). These results demonstrate a 70 percent NOV emission reduction at
X
full load and a 54 percent NOX emission reduction at partial loads (Reference 4-53).
Preliminary vendor test results indicate that in typical boiler applications LNB controls
will achieve a 30-50 percent reduction in NOV emissions from distillate oil- and gas-fired boilers
A
and a 17-23 percent reduction in NOX emissions from heavy residual oil combustion (.3 wt % N)
(Reference 4-54).
4-69
-------
0.4
0.3
3
±J
ca
RESIDUAL OIL
0.3 wt % N
3% 00
1
0.2
0.1
NATURAL GAS
Full Load
DISTILLATE OIL
2% 02
Full Load
600 800
Air Preheat Temnerature ("P.)
1000
Figure 4-5.
Effects of air preheat temperature on NO emissions from
gas- and oil-fired boilers (Reference 4-43).
4-70
-------
Operational problems with LNB controls have not yet been well identified. However,
preliminary information indicates that some LNB controls increase either the length and/or the width
of the burner flame. In retrofit applications where adequate firebox space is not available some
degree of derating may be required (Reference 4-55).
Coal-Fired Boilers
The baseline NO emissions from coal-fired boilers are generally higher than those from gas
and oil-fired units. Emissions range from 100 to.550 ng/J (0.23 to 1.3 1b/10 Btu). Although the
fuel nitrogen contents of the test coals are high, ranging from 0.8 to 1.5 percent by weight, field
studies indicate no strong dependence of NOX emissions upon fuel nitrogen content. Other factors
are apparently more important in,determining NOX production, such as furnace geometry, excess air,
firing rate, burner type, and other fuel properties (Reference 4-56).
Coal-fired boilers are generally classified by their coal feeding mechanism: pulverized
coal, spreader stoker, and mass fe'd stoker. Studies have found that pulverized coal-fired boilers
produce the highest uncontrolled NOX emissions -- approximately 328 ng/J (0.76 lb/10 Btu). In
these units finely pulverized coal is blown into the boiler through burners. The pulverized coal
burns relatively rapidly in suspension, resulting in a very high combustion intensity.
Spreader stoker boilers, in which the coal is thrown onto the grate from above, exhibited
intermediate NO emission rates ~ approximately 274 ng/J (0-.63 lb/10 Btu) for uncontrolled
full-load firing. In these units some of the fuel is burned in suspension with air supplied by over
fire air ports, and the remainder is combusted on the grate with under fire air. The resulting
combustion is therefore partially staged. Due to design characteristics and the partial staging
characteristics, the combustion intensities of stoker boilers are less than of pulverized coal-fired
boilers, contributing to a decrease in NO emissions.
Mass fed stokers had the lowest emissions ~ approximately 145 ng/J (0.34 lb/10 Btu). In
these units the combustion air fed up through the grating is insufficient for complete oxidation, so
additional air is introduced above the grating through over fire air ports. Combustion is,
therefore, effectively staged, and the NOV emissions are quite low.
A
4-71
-------
The following discussion of combustion modification controls for coal-fired boilers will
focus primarily on spreader stoker boilers. Pulverized coal-fired boilers are predominantly
constructed in sizes greater than 73 MW (250 x 10 Btu/hr). Large industrial boilers that are
greater than this size are very similar to utility boilers which are discussed in Section 4.1. Mass
fed stokers will generally not be discussed because there is very little data available on the
control of NO emissions from these sources. This lack of data can be attributed to the smaller
size of these units ^generally less than 29 MW (100 x 10 Btu/hr)] and the naturally lower NOX
emissions from this source (Reference 4-57).
Low Excess Air - Low excess air (LEA) controls are the most effective of the combustion
modification techniques for the control of NO emissions from stoker-fired boilers. At full load,
stoker boilers typically exhibit a flue gas oxygen concentration .of 6 to 11 percent. Studies
conducted by the American Boiler Manufacturers Association have demonstrated that excess air rates
for spreader stoker boilers can be safely reduced to a flue gas oxygen concentration of 4 to
6 percent at full load conditions (Reference 4-56). At full load conditions, the heat release rate
is the highest and consequently, so is the potential for NO emissions.
%
The EPA conducted a series of tests on 17 spreader stoker boilers to determine the
performance characteristics of LEA controls. Under normal operating conditions, these stokers
exhibited a flue gas oxygen concentration of 9 percent. When LEA controls were applied the average
flue gas oxygen level was reduced to 6.4 percent with an accompanying 26 percent reduction in NO
emissions (References 4-42, 4-46, 4-49, 4-58,'4-59).
A more indepth study was conducted on four of these boilers to determine the specific
onship between excess air levels and NO emissions.
X
in Figure 4-6 for full load conditions (Reference 4-60).
relationship between excess air levels and NO emissions. The results of this study are presented
X
The EPA also conducted 30-day studies on two spreader stoker boilers. On a 36 MW (125 x
10 Btu/hr) boiler, NOV emissions under LEA controls averaged 170 ng/J (0.4 lb/106 Btu)
/\
(Reference 4-61). For a 55 MW (190 x 106 Btu/hr) boiler the NOV emissions under LEA controls
A
averaged 208 ng/J (0.48 lb/106 Btu) (Reference 4-62).
4-72
-------
0.3
CM
O
0.7
0.6
0.5
S3
o
,! 0.4
X
o
0.3
0.2
20 40
60 80
Percent Excess Air
11-1
100 120
Figure 4-6. Effects of excess air on NO emissions from stoker
coal-fired boilers. (Reference 4-60).
4-73
-------
The American Boiler Manufacturers Association (ABMA) has also recently completed a series of
tests on stoker boilers. The ABMA study found that for five out of six spreader stokers studied,
the HO emissions decreased from 9.0 to 15.5 ng/J (0.021 to 0.036 lb/106 Btu) for each 10 percent
decrease in excess air (approximately 1 percent decrease in flue gas 02). The sixth spreader stoker
boiler exhibited a 28.8 ng/0 (0.067 lb/10 Btu) decrease in NO for the same decrease in excess air.
For these spreader stokers, LEA controls achieved an average NO emission reduction of 24 percent
over high excess air operation at full load (Reference 4-56).
The ABMA results from testing several underfed stokers exhibited NO emission reductions of
6.9 to 11.6 ng/J (0.016 to 0.027 lb/10 Btu) for each 10 percent reduction in excess air
(approximately equivalent to 1 percent reduction in flue gas Og). The average NO reduction
achieved by LEA controls on these boilers was 20 percent over high excess air operation
(Reference 4-56).
LEA controls are applicable to all types of stoker boilers. The major potential problem with
LEA controls is insufficient combustion air. If adequate combustion air is not supplied to the
grate, there is an increase in carbon monoxide, VOC, and unburned carbon emissions, in addition to
the formation of clinkers on the grate. Oxygen trim systems are commercially available which
monitor 0« and/or CO concentrations in the stack and adjust combustion air flow appropriately to
insure good combustion (Reference 4-45).
Staged Combustion Air - Staged combustion air (SCA) has not proved to be as effective in
reducing NO emissions from stoker boilers as it was for other types of boilers. This
ineffectiveness has been attributed to the observation that stoker boilers generally achieve some
degree of staged combustion by their inherent design. Fuel is burned relatively slowly on a grate
supplied by overfire air (OFA) and undergrate air. Using OFA tends to reduce undergrate air,
creating a locally oxygen difficient zone at the fuel bed (Reference 4-44).
The ABMA conducted a study of the impact of further reducing undergrate air and compensating
with an increase in OFA on 11 stoker boilers. The conclusion from this study was that SCA controls
had an insignificant impact on NOX emissions from both spreader- and mass-fed stoker boilers
(Reference 4-56).
4-74
-------
Similar results were also obtained from a SCA study of a 41 MW (140 x 10 Btu/hr) spreader
stoker boiler, a 22 MW (75 x 106 Btu/hr) mass-fed stoker boiler, and a 85 MW (290 x 106 Btu/hr)
mass-fed boiler (Reference 4-46).
The major operational problem associated with the implementation of SCA controls is
insufficient undergrate air. This problem results in increased emissions of VOC, carbon monoxide
and particulate emissions, in addition to grate slagging and corrosion problems. These operational
problems can be avoided by the use of commercially .available oxygen trim systems which monitor stack
02 and CO concentrations, and make combustion air adjustments appropriately (Reference 4-45).
Flue Gas Recirculation - The partial recirculation of flue gas (F6R) to the combustion chamber
has not been demonstrated on stoker boilers. However F6R has been tested on other high nitrogen
fuels such as residual oil-fired boilers and pulverized coal-fired boilers. The results of these
tests have shown that recirculation rates of up to 15 percent decreased NO emissions by only
17 percent, whereas similar recirculation rates decreased NOX by as much as 50 percent for gas- and
distillate oil-fired boilers (Reference 4-47, 4-63, 4-64).
Reduced Air Preheat - The,technique of reduced air preheat (RAP) attempts to reduce NO
emission formation by reducing the temperature of preheated combustion air. This technique has been
studied on only a very limited scale on stoker boilers. A comparison of the typical emissions from
six spreader stoker boilers equipped both with and without combustion air preheaters showed no
significant difference in NOX emissions which can be associated with combustion air preheat
(Reference 4-60). Some researchers claim that the coal bed preheats the combustion air before the
combustion occurs and thus defeats the purpose of the method.
This technique is of course limited to stokers equipped with combustion air preheaters. Only
larger stokers, greater than 29 MW (100 x 10 Btu/hr) tend to have air preheaters. In addition,
significant losses in boiler efficiency will occur if flue gas temperatures leaving the stack are
increased as a concequence of bypassing the preheater. Economizers can be added to avoid these
efficiency losses.
4-75
-------
4.2.1.2 Post Combustion Techniques
Post combustion techniques promise to be more effective than combustion modification
techniques for control of NOX emissions from industrial boilers. However, the demonstration of post
combustion techniques currently lags a significant distance behind combustion modification.
The two most developed post combustion techniques are ammonia injection and selective
catalytic reduction. Both of these techniques are based on the reaction of ammonia with nitrogen
oxide to form elemental nitrogen and water. These two techniques will be discussed in this section.
Other post combustion techniques such as wet scrubbing and electron beam irradiation are still in
the research and development stage, and will not be discussed in this section. Accurate information
is not yet available on these emerging technologies with respect to their application, performance,
cost, and operating problems for industrial boilers.
Awnonia Injection
Ammonia injection is the best demonstrated of the post combustion NO control techniques. In
X
this process, ammonia (NHg) is injected into the flue gas downstream of the firebox where it reacts
in a gas phase reaction with NO to produce N2 and H20. One advantage of ammonia injection is that
it can be used alone, or else in conjunction with combustion modification to achieve an additive NO
control effect.
In Japan, ammonia injection was applied-on four industrial boilers; one oil-fired an.d three
gas-fired ranging in size from 16 to 79 MW (55 - 270 x 106 Btu/hr) heat input. The Japanese tests
demonstrated a 40 to 65 percent reduction in NOX emissions. The most important variable in
determining performance was the flue gas temperature at the point of ammonia injection
(Reference 4-65).
In another test conducted on an oil field steam generator in California, ammonia injection
d a 50 to 1
(Reference 4-66).
achieved a 50 to 70 percent reduction in NOX emissions. This unit was burning a heavy crude oil
The performance of ammonia injection has not been tested on firetube boilers, residual
oil-fired boilers or stoker boilers. However the performance of ammonia injection is primarily
dependent on the flue gas temperature at the point of injection and is relatively independent of
fuels and combustion conditions. Therefore, the performance of this technique should be very
similar for all common types of fossil fuel-fired boilers.
4-76
-------
There are two operational problems of concern with the application of ammonia injection;
maintaining the optimum injection location and ammonium sulfate deposits. As the boiler load
fluctuates, there is a corresponding fluctuation in the temperature profile in the flue gas ducts.
Because of the sensitivity of the ammonia - no reaction to temperature - load fluctuations also
result in a fluctuation of the optimum injection location. For this reason ammonia injection is
best suited for constant load boilers. Alternatively, ammonia injection systems may be accompanied
with multiple injection points and as'sociated controls so that ammonia can be injected at the proper
point for the corresponding boiler load (Reference 4-66).
Ammonium sulfate problems are associated with burning high sulfur fuels. Unreacted ammonia
reacts with sulfur oxide in the flue gas to form ammonium sulfate salts. These salts create
plugging and corrosion problems for down stream preheaters and boiler parts. Ammonium sulfate ;
problems are best mitigated by applying ammonia injection only to very low sulfur fuels such as
natural gas and clean fuel oils and by the use of ammonia analyzers/controllers to detect and
control the presence of excess ammonia (Reference 4-66).
Selective Catalytic Reduction - Selective catalytic reduction (SCR) is a technique involving
removal of the flue gas NO by reacting the NO with ammonia in a catalytic reactor to form
A X
elemental nitrogen. With the exception of the use of a catalyst, it is similar to the ammonia
injection NO control technique just discussed.
Although not demonstrated in the U.S., SCR is thought to be applicable to all types of fossil
fuel-fired industrial boilers. Greater'than 90 percent NOV reduction is achieved at ammonia to NO
X X
ratios of 1:1 on commercial systems applied to industrial boilers in Japan. These systems have been
applied to a variety of gas- and oil-fired boilers in Japan, and appear to be viable, techniques of
attaining up to 90 percent NOV control on all industrial boilers (Reference 4-67).
A
Two operational problems associated with SCR are the formation of ammonium sulfate salts «and
the loss of unreacted ammonia. The ammonium sulfate salts pose plugging and corrosion problems for
down stream equipment and the loss of unreacted 'ammonia poses an emission problem. Both of these
problems are the focus of ongoing EPA and Electric Power Research Institute studies
(Reference 4-67).
4-77
-------
4.2.2 Cost Iiapact
NOX emission control costs are a very Important concern to the owners and operators of
Industrial boilers. These costs can vary greatly with control technique, degree of control, and
boiler size and type. This section discusses the capital and operating cost of various NOX control
technologies, and projects the Impact of these controls on the cost of producing steam.
The discussion on combustion modification and ammonia Injection controls was summarized from
a single study conducted for the Industrial Environmental Research Laboratory (IERL) of EPA 1n July
1981 (Reference 4-54). This study provides a comprehensive analysis of NO, control costs an a common
cost basis. The reader 1s referred to this report for detailed Information on the assumptions and
basis applied 1n developing these costs.
The discussion on SCR costs was summarized from a study conducted for the IERL on N()x flue
gas treatment technologies, completed December 1979. The reader 1s referred to this study,
Reference 4-67, for further Information on the assumptions and basis applied In developing these
costs.
Section 4.2.2.1 discusses the costs associated with the application of combust'on
modification controls for NOX emissions. And Section 4.2.2.2 discusses the costs associated with
the application of post combustion controls for NOX emissions.
4.2.2.1 Combustion Modification Techniques
Basis
Two cost components are studied
-------
Detailed Information on the cost basis applied to jenerate the following combustion modifi-
cation costs 1s presented 1n Reference 4-54. In genera1 the equipment, labor, fuel, and utility
costs are based on 1978 prices and are presented 1n 19'8 dollars. A load factor of 45 percent was
assumed for boilers smaller than 7000 kg steam/hr (15,000 Ib/hr) and a load factor of 60 percent was
assumed for all boilers larger than that size. A capital recovery factor of 0.2 was applied to
convert capital costs to annual 1zed capital charge'*.
In Che following analysis, the costs to irodlfy ths boiler are given as a percentage of the
boilers Installed cost. The boiler's Installed cost does not Include costs for auxllllary support
equipment such as water treatment and fuel hardllng. However, the Impacts on steam costs do
consider the to*ii cost to produce steam, 1nOud1ng the costs for auxiliaries.
Information on retrofit costs are not available in the same detail. However the costs for
applying combustion modification controls to existing boilers are estimated to be twice the cost of
appHc tlon to new boilers (Reference 4-68).
Results
The estiiwted capital and operating costs associated with applying LEA, SCA, and FGR controls
on industrial boilers are presentee1 1n Tables 4-27, 4-28, and 4-29 respectively.
For the reduced air preheat (RAP) control option, the cost: are negligible for new boilers
when a feed water economizer is substituted for the combustion fir preheater. Feed water econo-
mizers are estimated to recover the equivalent amount of waste heat for the equivalent cost of a
combustion air preheater. The retrofit of RAP controls on an existing boiler where c; feedwater
economizer is not feasible can sasily result in a boiler efficiency loss of up to 3 percent and a
stea.n price increase of as much as 2 percent (Reference 4-69, 4-70, 4-71). Efficiency drops of
twice this amount could result from significant preheat reductions.
Since low NOX burners are still in the developmental stage, costs ''re noc yet firmly
established. For new boilers, costs would be effected by the incremental cost difference between a
low NOX burner and a standard burner plus the cost of an oxygen trim system, if needed. Windbox
modifications may also be needed. Finally, 1t 1s expected that LNB will allow LEA operation which
would keep the economic impact of LNB at a minimum, thus, it seems likely that LNB operation wll1
cost no more than SCA operation and, in fact, may have an even smaller effect on steam costs than
SCA (Reference 4-72 and 4-73).
4-79
-------
I ABIE 4-27. ESTIMATED COST OF LOU EXCESS AIR OPERATION FCft NEW BOILERS (1976 Dollars)
00
o
AHNUALIZED COSTS
TYPE
Pulverized Coa1
Spreader Stoxer
Chain Grate
Stoker
Underfeed
Stoker
Residual Oil
WatertL-be
Residual Oil
Flretube
Distillate Oil
Uatertube
Distillate Oil
Flretube
Natural Gas
Vatertube
Natural Gas
Fire lube
»EAT,IHPUT
MU (10° Btu/hr)
59 (200)
44 (IbO)
22 (75)
9 (30)
4« (ISO)
4.4 (15)
29 (100)
4.4 (15)
29 (?00)
4.4 (Ib)
PERCENT NO
REDUCTION"
15
15
15
15
20
15
10
10
5
5
CAPITAL
COST
1000 $
27
22
17
14
17
9
14
9
14
9
HXEO
, Kills/
10J kg/steam
1J 16)
16 (7)
26 (12)
54 (24)
12 (5)
91 (41)
16 (7)
91 (41}
16 (7)
91 (41)
OPERATING*
, Bills/
10J kg steam
-28 (-13)5
-31 (-14)
-28 (-13)
-23 (-10)
-75 (-34)
-59 (-27)
-41 (-19)
-26 (-12)
-23 (-11)
-11 (-5)
TOTAL
, Bills/
10J kg/stea*
-15 (-7)
-15 (-7)
-2 (-1)
21 (M)
-63 (-29)
32 (14)
-25 (-11)
65 (29?
-9 (-4)
80 (36)
PERCENT CHANGE IN COST OF
BOILER
0.4
0.5
1.2
1.5
1.9
9.0
3.4
9.0
4.0
9.0
STEAM
-D.I
-0.1
0
0.1
-0.5
0.1
-0.2
0.2
-0.2
0.3
dNuirt>ers In parentheses are in units of rail Is/103 Ib stream
''Negative values represent cost savings
cReference 4-54.
-------
TABLE 4-?K. ESTIMATED COST OF STM&D COMBUSTION OPERATION FOR NEW BOILERS (1978 Dollars)
ANNUAL I ZED COSTS1
TYPE HEAT,INPUT PERCENT MO CAPITAL FIXED OPERATING
MU (10b Btu/hr) REDUCTION" CUM
Pulverized loai M>
Spreader Stoker 44
Chain Grate 22
Stoker
Underfeed 9
Stoker
Residual Oil 44
Uatertube
Residual Oil 4.
Flretube
Distillate Oil 29
Uatertube
Natural Gas 29
Uatertube
1000 t
, Bills/ , Bills/
Mr kg/steaa 10J kg stCM
(200) 25 47 . 24 (11) (21)
(150) 20 22 16 (7) 4 (2)
(75) 15 17 26 (12) 6 (3)
(30) 15 14 54 (24) 12 (5)
(150) 35 3? 23 (10) 83 (36)
4 (15) 25 IS 193 {88} 117 (53)
(100) 30 26 29 (13) 84 (38)
(100) 30 27 29 (13) 69 (31)
10! AL
103 kg/stea*
71 (32)
20 (9)
32 (15)
66 (29)
103 (46)
310 (141)
'13 (51)
» (**)
PEKttNl CHANGE IN COST OF
BOILER STEAM
0.8 0.6
O.b 0.3
1.2 0.2
1.5 0.3
5.S 1.2
21.0 1.2
7.0 1.2
7.0 1.5
* NuBbers In parentheses jre In units of Ills/Hi3 Ib strew.
Reference 4-54.
i
4
-------
TABLE 4-29. ESTIMATED COST OF aUE GAS RECIRCUtATlON OPERATION FOR HEM BOILERS (1978 Dollars)
oo
PO
AHNUALIZED COSTS3
TYPE
Distillate Oil
Water-tube
Distillate Oil
Flretube
Natural Gas
Hatertube
Natural Gas
Flretube
HEAT,INPUT PERCENI NOV
HW (10b Btu/hr) REDUCTION55
zy (i(10) 40
4.4 (15) 40
29 (1UO) 40
4.4 (15) 40
CAPITAL
COST
1000 $
26
19
26
19
FIXED
, mills/
10J kg/steam
29 (13)
192 (87)
29 (13)
192 (87)
OPERA 1 ING
, mills/
10 kg steam
135 (blj
175 (80)
120 (54)
159 (72)
TOTAL
, mills/
10J kg/steam
164 (74)
367 (167)
150 (67)
351 (159)
PtRCENT CHANGE IN COS 1 Oh
BOILER STEAM
7.0 1.7
21.0 1.2
7.0 2.4
21.0 1.7
a Numbers In parentheses are in units of mills/10 Ib stream.
Reference 4-54.
-------
Based on the limited data available on combustion modification costs for industrial boilers,
some tentative conclusions can be made.
- Low excess air operation, in many cases, will actually lower steam costs due to the
increase in thermal efficiency. In general, LEA operation is recommended for use with
other control techniques to lessen their cost impact and to give higher NO reductions.
X i
- Staged combustion causes an estimated small increase in steam cost but with careful design
and operation this estimated cost increase can probably be reduced.
- Flue gas recirculation, though cost'ly, is the most effective for the low nitrogen fuels,
distillate oil and natural gas. Again, optimal design and operation will probably lower
the cost.
- Low NO burners hold the promise of being the most cost-effective technique for oil and
gas boilers. However, they are still under development.
- Reduced air preheat is recommended for those boilers where an economizer may be installed
in place of an air preheater.
4.2.2.2 Post Combustion Techniques
This section presents the costs for installing and operating the two most advanced post
combustion techniques for industrial boiler NOV control: ammonia injection and selective catalytic .
f A
reduction.
Ammonia injection is not a demonstrated control technology for industrial boilers within the
U.S. The few sources available on ammonia injection costs are based on large utility boilers.
Extrapolating down to an industrial-size boiler, the capital cost for installing ammonia injection
on a 59 MW (200 x 10 Btu/hr) pulverized coal boiler is estimated to be about $236,000. This cost
represents about a 4 percent increase over the cost of an uncontrolled boiler. The incremental
steam cost is estimated to be 220 mills/10 kg steam (100 mills/10 Ib steam) or a 2.5 percent
increase in steam price. Because of the large error in extrapolating from a utility- size boiler
down to an industrial-size pulverized coal boiler or to other boiler types, the above price
estimates have very large uncertainties (References 4-74, 4-75, 4-76).
4-83
-------
Selective catalytic reduction (SCR) Is another post combustion technique which has not been
applied to industrial boilers in the U.S. The most extensive cost estimates for SCR controls is
available from a detailed engineering cost analysis of proposed plant designs (Reference 4-77). The
results of this analysis are presented in Tables 4-30, 4-31, and 4-32. These cost estimates are
based on mid-1978 costs. Annualized costs are calculated from capital costs using a capital
recovery factor of 0.13, reflecting a 10 percent interest rate and a 15 year recovery period. The
load factors for coal, residual oil, distillate oil, and natural gas were 60 percent, 55 percent,
45 percent, and 45 percent respectively. Finally, the most stringent control level reflects
90 percent NOV control and the moderate control level reflects 70 percent N0y control.
J\ A
4.2.3 Energy and Environmental Impact
4.2.3.1 Energy Impact
The energy impacts of applying NO emission controls are very important to the owner/operator
of industrial boilers. By far, the major cost of operating an industrial boiler is the fuel cost.
Combustion inefficiencies created by NOV emission control techniques translate directly into higher
A
steam costs. This section discusses the energy impacts associated with applying NO emission
control technologies. These energy impacts are a combination of changes in boiler thermal
"\
efficiency, and direct energy consumption by the control technology. The energy impacts presented
here were summarized in Tables 4-24 and 4-25.
Low Excess Air
Low excess air (LEA) is the simplest of all NOX control techniques, and one which saves fuel.
Virtually all boilers tested show an increase of about 0.5 percent in efficiency for each 1 percent
decrease in flue gas oxygen. However, in a few isolated cases, LEA did not increase efficiency for
natural-gas-fired units. The average energy savings was approximately 1 percent of the thermal
Input (Reference 4-42, 4-46, and 4-49).
Staged Combustion Air
Staged combustion air (SCA) can be achieved by use of BOOS and by use of OFA. In the limited
number of studies conducted on BOOS controls, very little impact on thermal efficiency has been
observed. The thermal efficiency has ranged from an 0.5 percent increase to an 0.5 percent decrease
(Reference 4-42, 4-49, and 4-68).
4-84
-------
TABLE 4-30. ANNUAL COST OF NO CONTROL SYSTEMS
APPLIED TO COAL-FIRED BOILERSX(Reference 4-67)
Boiler
Underfeed Stoker
Chaingrate
Spreader Stoker
Pulverized Coal
Size,
MBtu/hr
30
75
150
200
Annual Cost, $1000/yr
Control System
Parallel Flow SCR
Parallel Flow SCR
Parallel Flow SCR
Parallel Flow SCR
Moderate
Control
108
153
221
254
Stringent
Control
130
197
291
351
TABLE 4-31. ANNUAL COST OF NOV CONTROL SYSTEMS
APPLIED TO OIL-FIRED BOILERS (Reference4-67)
Boiler
Distillate Oil
Distillate Oil
Residual Oil
Residual Oil
Residual Oil
Residual Oil
Size,
MBtu/hr
15
150
30
30
150
150
Annual Cost, $1000/yr
Control System
Fixed Packed Bed SCR
Fixed Packed Bed SCR
Parallel Flow SCR
. Moving Bed SCR
Parallel Flow SCR
Moving Bed SCR
Moderate
Control
64
137
96
120
181
168
Stringent
Control
67
176
108
130
223
204
TABLE 4-32. ANNUAL COST OF NOV CONTROL SYSTEMS
APPLIED TO NATURAL GAS-FIRED BOILERS (Reference 4-67)
Boiler
Package,
Package,
Fi retube
Watertube
Size,
MBtu/hr
15
150
Annual Cost, $1000/yr
Control System
Fixed Packed Bed SCR
Fixed Packed Bed SCR
Moderate
Control
64.4
129
Stringent
Control
67.6
' 175
4-85
-------
Tests on OFA controls have shown a wider thermal efficiency impact, ranging from a 1 percent
efficiency gain to a 3 percent efficiency loss. With OFA controls there is an additional energy
loss associated with increased fan power requirements. This additional fan power is required to
overcome the pressure drop in the air ducts leading to the OFA ports,, and amounts to less than
0.1 percent of the thermal input (Reference 4-42 and 4-72).
Both SCA controls are sensitive to burner stoichiornetry and to location of air injection. The
use of oxygen trim systems will help offset possible boiler efficiency losses associated with either
SCA control.
F1ue Gas Reelrculati on
In tests run to date flue gas recirculation (F6R) had only a small effect on boiler thermal
efficiency. Thermal efficiency impacts ranged from a 0.5 percent increase to a 1.0 percent
decrease, while most impacts were less than a 0.5 percent decrease (Reference 4-42, 4-48, 4-49,
4-68, 4-77). According to utility boiler data, an increase of 0.25 percent of boiler heat input
could be required to power the F6R fan (Reference 4-72).
Low HO Burners
At present, there is very little data available on the energy impacts of low NO burners (LNB).
LNB are not expected to have a significant impact on thermal efficiency. In fact some improvement
1n efficiency may be possible due to their expected use of lower excess air (Reference 4-78).
Reduced A1r Preheat
The use of reduced air preheat (RAP) has the potential for significantly impacting thermal
efficiency. Tests have demonstrated thermal efficiency decreases of up to 3 percent with the use of
RAP. Even greater energy penalties are incurred with the full reduction of air preheat. This
efficiency decrease is associated with the loss of valuable waste heat when the preheater is
bypassed. However, on new units where feedwater economizers can be substituted for preheaters in
the boiler design, this thermal efficiency loss can be entirely mitigated (References 4-42, 4-46,
and 4-49).
4-86
-------
Amnonfa Injection
Very little information is available on the energy impacts associated with ammonia injection
controls. However, since ammonia injection is a post combustion control technology, its primary.
energy impact will be associated with the energy usage of ammonia handling and injection equipment.
For SCR controls, .the ammonia handling and injection equipment consumed steam and electricity
totaling less than 0.1 percent of the thermal energy input (Reference 4-67).
Selective Catalytic Reduction
Selective catalytic reduction (SCR) is a post combustion control technology and therefore is
not expected to have an impact on the direct thermal efficiency of the boiler. However, SCR
controls do consume electricity and steam. The electrical demand is required to overcome the
pressure drop across the catalytic reactor and to transfer ammonia. Steam is consumed in vaporizing
and diluting the ammonia. The combined electrical and steam demand required to operate SCR is
consistently less than Q.64 percent of the boiler thermal input and generally less than 0.3 percent
(Reference 4-67);
4.2.3.2 Environmental Impact
Very little research has been conducted on the environmental impacts associated with the use of
NOX emission controls on industrial boilers. Much more extensive research has been conducted on
utility size boilers, the results of which are presented in Section 4.1.3.2. Because of the
similarities between industrial and utility boilers with respect to emission characteristics, the
reader is referred to the above section for information on the environmental impacts of applying NO
X
controls to industrial boilers.
In general, combustion modification controls were found to generate very slight or no increase
in carbon monoxide, VOC, particulates, or trace elements. However there was a slight increase in
POM emissions.
, Post combustion techniques including ammonia injection and SCR are expected to increase
emissions of ammonia, amonia salts, and S03. In some cases, wastewater treatment and disposal may
be complicated by the addition of nitrogen compounds. Finally SCR systems will require waste
catalyst disposal. The full extent of these environmental Impacts is not known and is the subject
of ongoing EPA research.
4-87
-------
4.3 PRIME HOVERS
4.3.1 Reciprocating Internal Combustion Engines
Stationary reciprocating engines account for nearly 20 percent of the NOX from stationary
sources, or 2.4 Tg per year (2.66 x 10s tons). There are presently no Federal regulations for gase-
ous emissions from these engines. Some local areas, such as the South Coast Air Pollution Control
District of Southern California, have set standards for internal combustion engines.
A 1973 study by McGowin (Reference 4-79) provides a good overview of emissions from station-
ary engines, particularly the large bore engines used 1n the oil and gas industry and for electric
power generation. An EPA-sponsored Standards Support and Environmental Impact Study (SSEIS) for
these engines (Reference 4-80) will be completed in 1978 and will be the most comprehensive study of
stationary reciprocating-engines to date.
4.3.1.1 Control Techniques
The NO control techniques for 1C engines must be effective in reducing emissions over a broad
A
range of operating conditions from continuous operation at rated load to lower utilization appli-
cations at variable load. In general, large natural gas spark ignition engines running at rated
loads have the highest NOX emission factors. Gasoline engines, in contrast, frequently operate at
lower loads (less than 50 percent of rated) and produce substantially higher levels of CO and HC. The
NOX control techniques for these engines often involve HC and CO control since these emissions fre-
quently increase as NOX is reduced. Divided chamber diesel-fueled engines produce low levels of
NO (accompanied by greater fuel consumption than open chamber designs). In general, all diesel-
X ,
fueled engines have relatively small HC and CO emissions (less than 4 g/kWh*).
The following paragraphs will discuss NOX control techniques in general followed by a tabula-
tion of specific HOX reductions, by engine group. A lack of emission data precludes any discussion
of natural gas engines less than 75 kW/cylinder (100 hp/cylinder).
Table 4-33 summarizes the principal, combustion control techniques for reciprocating engines.
These methods may require adjustment of the engine operating conditions, addition of hardware, or a
combination of both. Retard, air-to-fuel ratio change, derating, decreased inlet air temperature,
or combinations of these controls appear to be the most viable control techniques in the near term.
Nevertheless, there is some uncertainty regarding maintenance and durability of these techniques
because, 1n the absence of regulation, very little data exists for controlled engines outside of
laboratory studies, particularly for large stationary engines. In general, increases in fuel
consumption, as much as 10 percent, are the most immediate consequence of the application of these
shaft output
4-88
-------
Table 4-33.
SUMMARY OF NOX EMISSION CONTROL TECHNIQUES FOR
RECIPROCATING INTERNAL COMBUSTION ENGINES
CONTROL
RETARD
injection (CD*
ignition (SI)b
CHANGE AIH-TO-FUa (A/F)
RATIO
DERATE
INCREASE SPEED
DECREASE INLET HANIFOU
AIR TEMPERATURE
EXHAUST GAS RECIRCUUTION
(SR)
External
Internal
valve overlap
or rttird
exhaust back
pressure
CHAJQER MODIFICATION
Preeombustlon (CD
Stratified charce (SI)
HATER INDUCTION
CATALYTIC CONVERSION
PRINCIPLE OF REDUCTION
Reduces peak temperature
by delaying start of
combustion during the
combustion stroke.
Peak conbustlon tenpera-
ture 1i reduced by off-
stolchlomtHc operation.
Reduces cylinder pres-
sures and temperatures.
Decreases residence tine
of gases at elevated
temperature and pressure.
Reduces peak temperature. '
Dilution of 1ncon1ng com-
bustion charge with Inert
gases. Reduce excess
oxygen and lower peak
coabustlon temperature.
Cooling by increased
scavaglng, richer
trapped air-to-fuel
ratio.
Richer trapped air-to-
fuel ratio.
Combustion In ante-
chamber permits lean
combustion fn main cham-
ber (cylinder) with less
available oxygen.
Reduces peak combustion
temperature.
Catalytic reduction of
NO to «2.
APPLICATION
An operational adjustment. Delay
can or Injection punp tilting (CI);
delay Ignition spark (SI).
An operational adjustment. In-
crease or decrease to operate on
off-»to1chiometr1c mixture. Reset
throttle or Increase air rate.
An operational adjustment. Units
maximal bmepc (governor setting).
Operational adjustnent or design
change.
Hardware addition to Increase
aftercoiHng or add aftercoollng
(larger heat exchanger, coolant
punp).
Hardware addition; plumbing to shunt
exhaust to Intake; cooling nay be
require! to be effective;' controls
to vary rate with load.
Operational hardware modification:
adjustamt of valve can timing.
Throttling exhaust flow.
Hardwam modification; requires
different cylinder head.
Hardware! addition: Inject water Into
Inlet onnifold or cylinder directly;
effective at water-to-fuel ratio
1 (kg H;>0/kg fuel).
Hardware) addition: catalytic con-
verter Installed In exhaust plumbing
or reducing agent (e.g. amonla)
Injected Into exhaust stream.
BSFCd
' INCREASE
Yes
Yes
Yes
Yes
No
No If EGR
rates not
excessive
Yes
Yes
Yes
No
HO
COWENTS -- LIMITATIONS
1
Particularly effective with moderate amount i
of retard; further retard causes high exhaust
temperature with possible valve damage and
substantial 3SFC Increase with smaller NOX
reductions per successive degree of retard.
Particularly effective on gas or dual-fuel
engines. Lean A/F effective but limited by
misfiring and poor load response. Rich A/F
effective but substantial 8SFC, HC, and CO
Increase. A/F less effective for diesel-
fueled engines.
Substantial Increase In BSFC with additional
units required to compensate for less power.
HC and CO emission increase also.
Practically equivalent to derating because
bmep 1s lowered for given power requirements.
Compressor applications constrained by vibra-
tion considerations. Not a feasible tech-
nique for existing and oast new facilities.
Ambient temperatures limit maximum reduction.
Raw water supply may be unavailable.
Substantial fouling of heat exchanger and f1»
passages; anticipate increased maintenance.
Hay cause fouling in turfiocharged, aftercooleal
engine. Substantial Increases in CO and smote
emissions. Maximum recirculation limited by
smoke at near rated load, particularly for
naturally aspirated engines.
Not applicable on natural gas engine due to
potential gas leakage during shutdown.
Llnlted for turbocharged engines due to
choking of turboconpressor.
5 to 10 percent Increase in 8SFC over open-
chamber designs. Higher heat loss implies
greater cooling capacity. Major design
development.
Deposit buildup (requiring denlnerallzation);
degradation of lube oil, cycling control
problems.
Catalytic reduction of NO Is difficult in
oxygen-rich environment. Cost of catalyst
or reducing agent high. Little research
applied to large-bore 1C engines.
'Compression Ignition
"Spark Ignition
ebnep brake mean- effective pressure .
dSSFC brake specific fuel consumption
4-89
-------
techniques (excluding Inlet air cooling). All control techniques involve only operational adjust-
ments with the exception of (1) derating which may require additional installed capacity to compen-
sate for the decreased rating, (2) inlet manifold air cooling which involves the addition of a heat
exchanger and a pump, and (3) catalytic conversion, which requires adding a catalytic reactor.
While exhaust gas recirculation (EGR) yields effective reduction of NOX, this technique -
requires additional development to overcome fouling of flow passages and increased smoke levels. In
general, recirculated exhaust is cooled in order to be effective. This practice promotes fouling.
EGR has not been field tested for large engines, and has been rejected by one manufacturer of heavy-
duty diesel truck engines and limited by another manufacturer. EGR has potential application in
naturally aspirated engines if full load EGR cutoff is provided to prevent excessive smoke (<10
percent opacity). EGR, however, has been applied successfully in combination with other techniques,
such as retard, in gasoline-fueled automobile engines (References 4-80, 4-81)..
Hater injection, similarly, has serious maintenance and durability problems associated with
mineral deposit buildup and oil degradations. Despite use of demineralized water and increased
oil changes, the control problems associated with engine startup and shutdown persist. This
factor, coupled with the need for a water source, has led manufacturers to reject this technique
(Reference 4-80).
Combustion chamber modifications such as precombustion and stratified chambers have demon-
strated large NOV reductions, but also produce substantial fuel consumption increases (5 to 8 per-
A
cent more than open chamber designs). With the rapid increases in the price of diesel fuel and
gasoline, manufacturers have been reluctant to implement this technique. In fact, one manufacturer
of divided chamber engines is vigorously pursuing development of low emission open chamber engines
(Reference 4-80).
Table 4-34 summarizes emission reductions achieved with large .bore'engines by use of retard,
air/fuel ratio changes, derating, and reduced inlet manifold air temperature (MAT). This table
Includes only those techniques from Table 4-33 which could be readily applied by the user. The
cited emission reductions are based on results obtained from engines tested in manufacturers'
laboratories. Therefore, some uncertainty exists concerning durability and maintenance over longer
periods of operation. In general, the greatest NO reductions are accompanied by the larger
Increases in fuel consumption. This is a direct result of reducing peak combustion temperatures
and, thus, decreasing thermal efficiency.
4-90
-------
Table 4-34. EFFECT OF NOX CONTROLS ON LARGE-BORE
INTERNAL COMBUSTION ENGINES
a. Normalized percent reductions of NO
Fuel
Number Cylinders
Baseline3
Retard
Alr-to-Fuel
Derate
MAT
Gas
2
BSb
20
2.5
0.19
6.2
0.9
TC
17
3.1
4.5
2.6
1.3
4«
NA
24-29
1.5
1.8
0.25-1.3
TC
17-30
4.1-0.6
3.3
0.34-1.9
0.4-0.9
Dual Fuel
2
TC
12
9.1
1.7
1.3
4
TC
10-17
1.5-6-3
2.4-2.5
0.01-0.94
0.6-0.8
Diesel
2
BS
18-26
6.9
0.84-0.92
0
TC
14-19
5.3-5.7
0.2-0.4
4
TC
13-15
2.7-4.4
0.17
0.1-0.3
b. Percent increase in brake specific
fuel consumption
Retard
A1r-to-Fue1
Derate
MAT
5.2
2.0
2.6
1.3
4.3
1.5
6.1
0.5
3.6
1.0
8.2C
1.2
2.3
c
1.1
0
3.4
.2.6
7.0C
0.4
1.0C
1.9
+0.5
C
3.4
3.3C
«.
1.6
2.2C
9.6
0
Baseline data in gm/kWh shaft output, all other data in percent NOX reduction/unit control. Unit
control is 1 retard, 1 percent air flow increase, 1 percent derating, or 1.8K (l.OF) air
.temperature decrease.
|!BS - blower scavenged, TC - turbo-charged, NA - naturally aspirated.
Average value
-------
Numerous investigators have studied control techniques to reduce NOX in diesel-fueled auto-
motive truck applications. Many of these studies are summarized in Reference 4-81. Retard, turbo-
charging, aftercooling, derating and combinations of these controls are techniques that are current-
ly utilized by manufacturers to meet California heavy-duty vehicle (>2700 kg, or 6000 1b) emission
limits for diesel-fueled engines.
Table 4-35 lists,five samples of NOX control techniques currently implemented by truck
manufacturers to meet the 1975 California 13.4 g/kWh* (10 g/hp-hr) combined NOX and HC emission level.
Manufacturers indicate that greater reductions will require (1) increasing degrees of application
of these controls (and incurring additional fuel penalties) or, (2) application of techniques that
need further development to overcome maintenance, control, and durability problems. Controls in
this second category include EGR, water injection, and MOX reduction catalysts.
Gasoline engine manufacturers, in response to Federal and State regulations, have also con-
ducted considerable research of emission control techniques to reduce NO , as well as HC and CO,
levels. Efforts in this area have been directed at reducing emissions to meet (1) Federal and
California heavy-duty vehicle limits, and (2) Federal and California passenger car emissions limits.
Table 4-36 lists Federal and State emission limits, and Table 4-37 lists the various controls that
are used in several combinations by manufactures to meet these limits. Table 4-38 gives specific
examples of control techniques recently applied to meet Federal light duty vehicle emission limits.
Based on the preceding discussion, potential NO emissions reductions for stationary recipro-
cating engines can be summarized as follows:.
Controls such as retard, air-to-fuel ratio change, turbocharging, inlet air cooling (or
increased after cooling), derating and combinations of these controls have been demon-
strated to be effective and could be applied with no required lead time for development.
Fuel penalities, however, accompany these techniques and may exceed 5 percent of the
uncontrolled consumption.
t Exhaust gas recirculation, watar injection, catalytic conversion and precombustion or
stratified charge techniques involve some lead time to develop as well as time to address
maintenance and control problems.
rated shaft output
4-92
-------
Table 4-35.
CONTROL. TECHNIQUES FOR TRUCK SIZE
DIESEL ENGINES [<375 kW (500 HP)]
TO MEET 1975 CALIFORNIA 13.4 G/KWHR .
(TO G/HP-HR) COMBINED N0₯ AND HC LEVEL'
/\
Control
Percent
bsfc" Increase
Retard, modify fuel
system and turbocharger
Retard, modify fuel
system and turbocharger,
add aftercooler
Add turbocharger and
aftercoolerc
Retard0 (naturally
aspirated version)
Precombustion chamber
3
3
5-8
Based on Federal 13 mode composite cycle
bsfc = brake specific fuel consumption
Stationary versions of this engine would
require a cylinder head with four exhaust
valves rather than existing two valves.
4-93
-------
TABLE 4-36. 1975 VEHICLE EMISSION LIMITS
Passenger Car,
g/Wh (g/m1)a
California
Federal
Light duty truck,
g/kWn (g/ral)
California
Federal
Heavy duty vehicles,
g/kHh
California
Federal
MX
6 (2.0)
9 (3.1)
6 (2.0)
9 (3.1)
HC
3 (0.9)
4 (1.5)
6 (2.0)
6 (2.0)
13
21
CO
26 ( 9)
44 (15)
59 (20)
59 (20)
40
53
"Emissions limits are estimated In g/kWh from g/mi assuming an average of 38.4
lun/hr requiring 8195 W (11 bhp) for the '/-mode composite cycle.
TABLE 4-37, EMISSION CONTROL TECHNIQUES FOR AUTOMOTIVE GASOLINF ENGINES
CoMtrol
NOX:
Rich or lean A/F ratic,
Ignition timing retard
Exhaust gas redrculatlon
(5 to 10 percent)
Catalytic converters
(reduction)
Increased exhaust back pressure
Stratified combustion
HC. CO:
Thermal reactor
Catalytic converter (oxidation)
Exhaust manifold air Injection
Positive crankcase ventilation
Comment
Increased bsfc, HC, and CO
Increased bsfc, HC, and CO, amount
of control limited by potential
exhaust valve damage
Increase bsfc and maintenance
related to fouling, smoking limits
de
-------
TABLE 4-33. EMISSION CONTROL SYSTEMS FOR CONVENTIONAL GASOLINE INTERNAL COMBUSTION ENGINES
(ADAPTED FROM Reference 4-81 i
Number
0
1
2
3
Year
1972
1973 Federal
1975 Federal
1975 Calif.
Syste»
EM-1
EM* + El + FC * AI + E6R
EM' + El * 1C + QHI + AI + EGR
« + El + 1C + QKi + EfiR * AI + OC
Fuel Penalty v
_
7*3
5 ± 2
8 t 2
Reduction Factors1*
HCC
1 ± 0.375
1.35 i 0.30
0.65 i 0.15
0.18 ± 0.05
COC
1 t 0.375
1.0 t 0.23
0.55 t 0.1S
0.15 t 0,03
NO/
1 i 20
0.6 t 0.10
0.06 t 0.10
0.06 t 0.10
Syste.
Oc erloratlon
L
L
L
M ;HC. co)
L (NOX)
vo
in
aH972 baseline engine: andiflcatlons Included fn tht baseline engine configuration are retard, lera air-to-fuel, and reduced
compression ratio.
Cc^ponent Identification
EM - Engine Modifications; retard, air-to-fuel, compression ratio
El - Electronic ignition
FC - Fast choke
QHI - Quick heat intake
AI - Exhaust manifold air injection
EGR - Exhaust gas reclrculatlon
1C - Improved carbureticn
OC - Oxidizing catalyst.
Deduction factor defined as: "S^baseffneHgssloBS basea on LA~4 dr1vi"9 c>cle
CA11 enissions dcta taken using or corrected to 1975 CVS-CH test procedure
Deterioration of present systems; L - 10X. M = 10 - 301, ' = 301
-------
NO control technology for automotive applications can be adapted to stationary engines;
»
however, NQX reductions and attendant fuel penalties for automotive applications are
closely related to the load cycle, which in some cases nay differ from stationary
applications
Viable control techniques may involve an operational adjustment, hardware addition, or
a combination of both
Additional research is necessary to
Establish controlled levels for gaseous-fueled engines (<75 kW/cylinder, or
100 hp/cylinder)
Establish controlled levels for roediurn-powered diesel and gasoline engines based
on stationary application load cycles
Supplement the limited emissions data available for large bore engines
4.3.1.2- Costs
As discussed earlier, stationary engines are unregulated for gaseous pollutants. Consequently,
few data are available for field-tested controlled engines, particularly for large (>375 kW or 500
hp) engines. Sufficient data exist, however, to give order or magnitude NOX control costs for the
following engine categories:
Natural gas-, dual-, and diesel-fueled engines above 75 kW/cylinder
(100 hp/cylinder)
Small to medium (<75 kW/cylinder) diesel-fueled engines
Gasoline-fueled engines (10 kW to 375 kW)
Costs for large stationary engines can be estimated based on Reference 4-82 and information
supplied by Reference 4-80 (1974 costs). These costs, however, relate to emission reduction
achieved by engines tested in laboratories rather than to field installations. Reference 4-83'
indicates, nevertheless, that these data are representative (1972 costs).
In contrast to the large stationary engines, more published cost data exists for smaller
(<37S kW, 500 hp) gasoline and diesel engines which must meet State (California) and Federal
emission limits for mobile applications. Stationary engines in this size range are versions of these
mobile engines. Therefore, costs can be estimated based on a technology transfer from mobile appli-
cations to stationary service, keeping in mind that in some cases mobile-duty cycles (variable
4-96
-------
load) can differ from stationary-duty cycle's (rated load). Hence, costs (e.g., fuel penalties)
associated with a control technique used in a stationary application may vary from the mobile case.
Control costs for the three categories discussed above may include:
Initial cost increases for control hardware and/or equipment associated with a particular
control (e.g., larger radiator for manifold air cooling or more engines as a result of
derating)
Operating cost increases Which consist of either increased fuel consumption and/or in-
creased maintenance associated with NO control system
Combinations of initial and operating cost increases
Control Costs for Large Bore Engines '
TABLE 4-39 lists cost impacts for control techniques available to users of large stationary
engines. These cost impacts may be related to actual installations using baseline data presented in
TABLE 4-40 (1974 costs). In practice, these figures vary depending on the application, but, in
general, they are representative of the majority of applications. Basically, these controls involve
an operating adjustment with the exception of derating and manifold air cooling, which would require
hardware additions. Derating is not a viable technique for existing installations un'iess additional
units can be added to satisfy total power requirements.
The impact of the above control costs may vary considerably given the following considera-
tions:
Standby (<200hr/yr) application control costs are primarily a result of initial £
cost increases due to the emission control, whereas continuous service (>6000 hr/yr)
control costs are largely a function of fuel consumption penalties
Controls which require additional hardware with no associated fuel penalty (e.g.,
manifold air-cooling) may be more cost effective in continuous service (>6000 hr/yr)
than operating adjustments which impose a fuel penalty (e.g., retard, or air-to-
fuel change)
The price of fuel can affect the impact of a control which incurs a fuel penalty.
For example, a control which imposes a fuel penalty of 5 percent for both gas and
diesel engines has more impact on the diesel fueled engine because diesel oil costs
about 40 percent more per Joule than natural gas. This impact will diminish if gas
prices increase more rapidly than oil prices.
4-97
-------
TABLE 4-39. COST IMPACTS OF NOX CONTROLS FOR LARGE-BORE ENGINES
Control
Cost Impact
Retard
Air-to-fuel changes
Derate
Manifold air cooling
Combinations of above
Control techniques
Increased fuel consumption, more frequent '
maintenance of valves
Increased fuel consumption, more frequent
maintenance of turbocharger
Fuel penalty, additional hardware, and increased
maintenance associated with additional units
Increased cost to enlarge cooling system, and
increased maintenance for cooling tower water
treatment
Initial, fuel, and maintenance
Increases as appropriate
TABLE 4-40. TYPICAL 1974 BASELINE COSTS FOR
LARGE (>75 KW CYLINDER) ENGINES3
Costs
1. Initial,5 $/kW
2. Maintenance,
$/kWh
3. Fuel and lube,
$/kWh
Total Operating,
2 + 3
Gas
174
4 x 10'3
10 x 10-3
14 x 10-3
Dual Fuel
174
4 x 10"3
10 x 10-3
14 x 10-3
Diesel
174
4 x 10-3
23 x 10"3
27 x lO'3
Based on Reference 4-82 and information supplied to
Reference 4-80 by manufacturers.
'includes basic engine and cooling system.
4-98
-------
Control Costs for Small and Medium Gasoline- and Diesel-Fueled Engines
Control costs for these engines can be characterized by analogy to those incurred to meet
- <
State and Federal emission limits for automotive vehicles. Again, these costs consist of initial
purchase price increases for control hardware and increased operating costs (fuel and maintenance
.cost increases).
Table 4-41 lists typical costs for techniques implemented for 1975 diesel-fueled truck
engines (1974 costs). These costs are presented to indicate order of magnitude effects. More
research is required to relate specific emission'control reductions to initial and operating cost
increases for stationary engine applications.
Table 4-42 gives control hardware costs to meet gasoline-fueled passenger vehicle emission
limits through 1976 (1973 costs). Note that cost increases correspond to increasingly more complex
controls to meet more stringent emission limits.
TABLE 4-41. TYPICAL CONTROL COSTS FOR DIESEL-FUELED ENGINES USED IN HEAVY-DUTY
VEHICLES (>2700 kg OR 3 tons)
engine
cooling system
turbocharger
aftercooler
EGR
$40-$67/kW ($30-$50/hp)
8%-14% engine
$4/kW ($3/hp)
6%-10% engine
$3-$4/kW ($2-$3/hp)
Operating
Fuel: Fuel penalties range from 3 to 8 percent for various techniques.
Typical present fuel cost: $0.095/1iter ($0.36/gallon) #2 diesel
or $2.13-$2.37/GJ ($2.25-$2.50/106 Btu).
Maintenance: EGR system will require periodic cleaning. Note that turbo-
charged, aftercooled engines require additional maintenance for
the turbocharger and aftercooler compared to a similarly rated
naturally aspirated version.
aBased on information supplied to Reference 4-80 by manufacturers (1974 costs).
4-99
-------
TABLE 4-42. ESTIMATES OF STICKER PRICES FOR EMISSIONS HARDWARE FROM
.966 UNCONTROLLED VEHICLES TO 1976 DUAL-CATALYST SYSTEMS
(Reference 4-81, Costs Taken From A 1973 Report)
Model
Year
1966
1968
1970
t
1971-
1972
1973
Configuration
PCV-Crank Case
Fuel Evaporation
System
Carburetor Air/Fuel Ratio
Compression Ratio
Ignition Timing
Transmission Control
System
Total 1970
Ant1-D1ese11ng
Solenoid
Thermo A1r Valve
Choke Heat Bypass
Assembly Line Tests,
Calif (1/10 vol)
Total 1971-1972
OSAC (Spark Advance
Control)
Transmission Changes
(some models)
Induction Hardened Valve
Seats (4 and 6 cyl)
EGR (11 - 14S)
Exhaust Recirculation
Air Pump Air
Injection System
Quality Audit, Assembly
Line (1/10 vol)
Total 1973
Typical Hardware
Value
Added
1.90
9.07
0.61
1.24
0.61
2.49
3.07
2.49
2.74
'0.18
0.48
0.63
0.72
5.48
27.16
0.23
Price
2.85
14.25
0.95
1.90
0.95
3.80
4.75
3.80
4.18
0.57
0.95
0.95
1.90
9.50
43.32
0.38
Excise
Tax
0.15
» 0.75
0.05
0.10
0.05
0.20
0.25
0.20
0.22
0.03
0.05
0.05
0.10
0.50
2.28
0.02
Sticker
Price
3.00
15.00
1.00
2.00
1.00
4.00
8.00
5.00
4.00
4.40
0.60
14.00
1.00
1.00
2.00
10.00
45.60
0.40
60.00
4-100
-------
TABLE 4-42. ESTIMATES OF STICKER PRICES FOR EMISSIONS HARDWARE FROM 1966 UNCONTROLLED
VEHICLES 70 1976 DUAL-CATALYST SYSTEMS (Reference 4-81) (Concluded)
Model
Year
1974
1975
1976
Configuration
Induction Hardened
Valve Seat V-8
Some Proportional EGR
(1/10 vol at $52)
Precision Cams, Bores,
and Pistons
Pretest Engines -
Emissions
Calif. Catalytic Con-
verter System (1/10 vol
at $64)
Total 1974
Proportional EGR
(acceleration-
deceleration)
New Design Carburetor
with Altitude
Compensation
Hot Spot Intake Manifold
Electric Choke (element)
Electronic Distributor
(pointless)
New Timing Control
Catalytic Oxidizing-
Converter
Pellet Charge (6 Ib at
$2/1b)
Cooling System Changes
Underhood Temperature
Materials
Body Revisions
Welding Presses
Assembly Line Changes
End of Line Test So/No--Go
Quality Emission Test
Total 1975
2 NOX Catalytic Converters*
Electronic Control3
Sensors3
Total 1976
Typical Hardware
Added
0.72
3.21
2.44
1.80
4.02
20.07
7.52
2.87
2.67
4.35
1.40
18.86
12.00
. 1.17
0.63
0.67
0.13
1.85
1.22
22.00
28.00
3.00
List
Price
1.90
4.94
3.80
2.85
6.08
30.02
14.25
4.75
4,75
9.50
2.85
34.20
20.52
1.90
0.95
1.90
0.95
2.85
1.90
37.05
47.50
5.70
Excise
Tax
0.10
0.26
0.20
0.15
0.32
1.58
0.75
0.25
0.25
0.50
0.15
1.80
1.08
0.10
0.05
0.10
0.05
0.15
0.10
1.95
2.50
0.30
Sticker
Price
2.00
5.20
4.00
3.00
6.40
20.60
31.60
15.00
5.00
5.00
10.00
3.00
36.00
21.60
2.00
1.00
2.00
1.00
3.00
,2.00
138.20
39.00
50.00
6.00
134.00
*1976 most common configuration
4-101
-------
Figure 4-7 illustrates the effect of various control techniques on fuel economy. Fuel
cost increases can be easily derived from typical gasoline costs, presently $0.55-0.75/gallon.
In addition to this operating expense, control techniques utilizing catalysts and EGR require peri-
odic maintenance.
Manufacturers, in addition, incur certification costs for gasoline and diesel-fueled engines
which must meet State and Federal regulations. These costs are passed on to the user in the form of
increased initial costs. Manufacturers of diesel-fueled engines report these costs range from
$50,000 to $100,000 for a particular engine family (1972 costs). This can result in a $125 cost per
engine based on a low sales volume family.
4.3.1.3 Energy and Environmental Impact
The energy impacts of applying NOX controls to stationary reciprocating 1C engines are mani-
fested almost exclusively through corresponding increases in fuel consumption (bsfc). Typical
percentage increases as a function of applied control were discussed in detail previously in
Sections 4.3.1.1 and 4.3.1.2.
Potential adverse environmental impacts.occur through increases in emissions of combustion-
generated pollutants other than NOX attendant to applying a NOX control. Since 1C engines emit
only an exhaust gas effluent stream, impacts through liquid and solid effluents need not be con-
sidered. In addition, since 1C engines fire "clean" fuels (natural gas and distillate oil) incre-
mental effects on the emissions of such pollutants as SOX and trace ruetals are relatively unimpor-
tant. Thus, the following discussion will focus on the measured effects of specific NOX control
techniques on incremental emission of CO, HC, and particulate (smoke). Again, all available data
were obtained in tests on laboratory engines. Nevertheless, such data should be representative.
Carbon Monoxide
<
As discussed in Section 4.3.1.1, the most common NOX reduction techniques applied to 1C
engines Include derating, ignition retard, altering air/fuel (A/F) ratiols, reducing manifold air
temperatures (MAT), and water Injection. The effects of each of these NOX controls on engine CO
emission levels are summarized in Table 4-43.
As Indicated, baseline CO emissions from two-cycle engines are generally lower than those
from four-cycle engines. However, derating two-cycle engines increases CO emissions 50 to 100 per-
ctnt, while derating four-cycle engines actually gives a 60 to 100 percent decrease in CO levels.
4-102
-------
1.5
1.0
0.5
VARYING EGR AND
SECONDARY AIR RATES
3.0
2.5
2.0 §
a
LU
1.5"
cc
a.
CO
X
1.0?
0.5
10 20
INCREASE IN BSFC, purcent (OVER UNCONTROLLED VEHICLE)
30
GENERAL CORRELATION
ESTIMATED FOR ADDITION OF NOX CATALYST
BED AT 75 PERCENT EFFICIENCY
VARYING DRIVING CYCLES
AND CONTROL TECHNIQUES
"I 1
5 10 15 20 25
INCREASE IN BSFC, percent (OVER UNCONTROLLED VEHICLE)
30
a
ai
o
o
cc
CO
o
X
a
Figure 4-7. Effect of NOx emissions level on fuel penalty for light duty trucks
(Reference 4-83).
4-103
-------
1^
o
Table 4-43. REPRESENTATIVE EFFECTS OF NOX CONTROLS ON CO EMISSIONS FROM INTERNAL COMfftfSTIGN ENGINES3
(Reference 4~80).
Fuel
Natural Gas
Diesel
Dual Fuel
Engine Type
2-cycle
4-cycle
2-cycle
4-cycle
2-cycle
4-cycle
Baseline
Emissions
(ng/J)
15 - 40
75 - 3350
72 - 325
114 - 546 '
165
200 - 670
NOX Control CO Emissions (ng/J)
Derate
40 - 94
54 - 150
89
100 - 180
244
289
Retard
Ignition
35 - 45
80 - 1000
140 - 628
260 - 654
244 - 267
679 - 1070
Increase
A/F
29 - 31
__h
439
288
Decrease
A/F
117
675
244
296
Reduce
MAT
29 - 45
131
71
142 - 550
67
632
Mater
Injection
194
464
460 - 606
503 - 507
This table Is Included 1n Appendix A 1n English-units.
Denotes no data reported.
-------
Retarding ignition generally causes increased CO output for all engines. This is somewhat
expected, though, since retarding ignition decreases both peak combustion temperature and combustion
gas residence time, which can lead to incomplete combustion. Both increasing A/F ratios and reduc-
ing manifold air temperature (MAT) has little effect on CO levels. However, decreasing A/F causes
50 to 100 percent increases in CO emissions. Water injection seems not to affect CO emissions from
gas and dual fuel engines, but increases diesel engine CO emissions by 60 to 130 percent.
Hydrocarbons
The use of NO combustion controls on 1C engines can also have significant effects on HC
emissions, with different NOX reduction techniques eliciting different effects.
As shown in Figure 4-8,, derating causes HC emissions to increase, with the increase becom-
ing more pronounced as load is further reduced. As the figure illustrates, derating can cause a 20
to 130 percent increase in HC emissions. Figure 4-9 shows the effect of ignition retard on incre-
mental HC emissions. In contrast to the effects of engine derating, ignition retard tends to
decrease slightly or not affect emissions of HC. However, in cases where retarding ignition initi-
ally reduces HC emissions, increasing the degree of ignition retard seems to have little further
effect. The data in the figure indicate that HC emissions decrease on the average of 30 percent
when ignition is retarded 3 to 8 degrees.
Changing the air-to-fuel (A/F) ratio, decreasing manifold air temperature (MAT) and water
injection can all result in increased HC emissions. As shown in Figure 4-10,. both increasing and
decreasing the A/F ratio by 10 percent increases HC levels 20 to 65 percent. Larger percentage
increases occur in engines with high baseline emissions. Figure 4-1H shows analogous effects when
MAT is decreased. Decreasing 10 to 20 K (20 to 40 F) increases HC emissions 5 to 50 percent. HC
levels increase as MAT is further reduced. Turbocharged engines exhibit the largest percentage
emissions increases. Water.injection also increases HC emissions from 1C engines regardless of the
baseline HC level, as shown in Figure 4-12. Average increases of 16 to 25 percent have been experi-
enced for water/fuel (W/F) ratios of 0.1 to 0.25.
Particulates
Virtually no data are available specifically on particulate emission rates from stationary
1C engines because it is difficult, time consuming, and expensive to measure particulate emissions
from these engines directly. Instead, exhaust gas opacity readings have been used as a substitute
measure of particulate emissions. These readings effectively measure particulate since a relation-
ship between visible smoke and particulate mass emissions, has been established for medium power
4-105
-------
100
O 2 CYCLE, BLOWER SCAVENGED
Q 2 CYCLE, TUREIGCHARGED
A 4 CYCLE, NATURALLY ASPIRATED
04 CYCLE, TURBOCHARGED
G NATURAL GAS
OF DUAL FUEL
0 DIESEL
§
20 30
POWER DERATE, percent
40
50
Figure 4-8. Effect of derating on 1C engine HC emissions (Reference 4-80),
1000
O 2 CYCLE, BLOWER SCAVENGED
Q 2 CYCLE. TURBOCHARGED
4 CYCLE. NATURALLY ASPIRATED
O4 CYCLE, TURBOCHARGED
G NATURAL GAS
OF DUAL FUEL
D DIESEL
8
4 6
TIMING RETARD, degrees
Figure 4-9. Effect of retarding ignition timing on 1C engine HC emissions (Reference 4-80)
10
4-106
-------
1000
CO
o
53
S2
100
" 20
30
20
O 2 CYCLE, BLOWER SCAVENGED
Q 2CYCLE.TUR80CHARGED
O 4 CYCLE, TURBOCHARGEO
G NATURAL GAS
OF DUAL FUEL
D DIESEL
10 . 0 10
CHANGE IN A/F RATIO, percent
20
30
Figure 4-10. Effect of air-to-fuel ratio on 1C engine HG emissions (Reference 4-80)
4-107
-------
1000
O 2 CYCLE, BLOWER SCAVENGED
Q 2 CYCLE,TUR80CHARGEO
O4CYCLE TUR80CHARGED
G NATURAL GAS
OF DUAL FUEL
D DIESEL
10 15
MAT DECREASE, °K
20
25
Figure 4-11. Effect of decreased manifold air temperature (MAT) on 1C engine
HC emissions (Reference 4-80).
1000
>DF
O 2 CYCLE, BLOWER SCAVENGED
O4CYCLE.TURBOCHARGED
G NATURAL GAS
OF DUAL FUEL
0 DIESEL
0.4 0.6
W/F RATIO
0.8
1.0
Figure 4-12. Effect of water injection in 1C engine HC emissions (Reference 4-80),
4-108
-------
diesel engines (Reference 4-84 and 4-85). Therefore, 1C engine smoke emissions are generally re-
ported as percent plume opacity, or as Bosch or-Bacharach smoke spot numbers.
The plumes from most large-bore engines; are nearly invisible when the engine is operating
at steady-state. However, applying NOX combustion controls can significantly affect smoke emissions.
Figure 4-13 shows the relationship between smoke emissions and NOX reduction as a function of NOX
control for those engines where data were reported on both pollutants. As the figure shows, NO
controls, other than derating, generally increase smoke emissions, while derating decreases smoke
levels. Ignition retard and exhaust gas recirculation (E6R) cause the most significant increases
fn smoke level.
Since NOX controls which caused smoke levels to exceed 10 percent opacity were considered
unacceptable in the tests summarized in figure 4-13 none of the data points for controlled engines
are above this value. However, the effect qf progressively applying ignition retard and ESR on
smoke emissions is best demonstrated by data v/hich include higher smoke levels. Such data are pre-
sented in Table 4-44 for two-cycle diesel engines, and clearly show that smoke emissions increase
progressively as percentage EGR or degree of retard is increased.
In summary, experimental data have shown that applying conventional combustion modification
NOX controls to 1C engines can cause increase!! in CO, HC, and particulate (smoke) emissions. This
is so because the combustion conditions required to prevent NOX formation generally lead to less,
complete combustion.
4.3.2 Gas Turbines
Gas turbines contributed only 2 percent of the annual stationary source NO emissions in
1974, or 236 Gg (2.6 x 10s tons). They do, however, comprise a very rapidly growing source with
increasing application in intermediate and base load power generation, pipeline pumping, natural gas
compression, and onsite electrical generation. The increasing application of gas turbines carries
with it the potential for increasing the NOV emission contribution from these sources. In response
A -
to this, the frequency of control technique demonstration and implementation has increased in the
past several years.
Uncontrolled NOX emissions are a function of turbine size (or efficiency) and fuel type. In-
creasing the turbine size (or efficiency) increases the NO concentrations primarily due to higher
combustion temperatures and to increased residence time at high temperatures. Oil-fired turbines
generally have higher NO concentrations than gas-fired units. Typical uncontrolled NO, emissions
A . X
from gas turbines are illustrated in Figures 4-14 and 4-15 for large and small units, respectively.
4-109
-------
NOX LEVEL, j/kWh
12 16
20
24
a
A
C
0
E
H
u
s
CONTROL CODE
AIR-TO-FUEL RATIO I
REDUCE COMP. RATIO M
DERATE R
EXTERNAL EGR
H20 INDUCTION S
INTERNALEGR
MANIFOLD AIR TEMP.
RETARD IGNITION
TIMING
INCREASE SPEED
R
~"\. FUEL
TYPE^^s^
2 STROKE
BLOWER
SCAVENGED
2 STROKE
TURBO-
CHARGED
4 STROKE
NORMALLY
ASPIRATED
4 STROKE
TURBO-
CHARGED
DIESEL
o
a
A
DUAL
FUEL
0
ENGINE CODE NUMBER (*) DENOTES INITIAL POINT.
CONTROL CODE DENOTES LEVEL AFTER CONTROL.
BACHARACH AND BOSCH METERS ARE FILTER-TYPE
INSTRUMENTS.
SMOKE LEVELS FOR ENGINES *8-12 WERf MEASURED
WITH A 30SCH METER.
FINAL SMOKE LEVEL IS AY END OF LINE HAVING
CONTROL CODE.) I I
10
<
#34
D
10 12
NOX LEVEL, j/hp-hr
14
16
18
20
Figure 4-13. Smoke levels versus NOX levels *or large bore diesel engines.
4-110
-------
TABLE 4-44. RELATIONSHIP BETWEEN SMKE,
EGK, AND RETARD
(Reference 4»80).
Engine Type
2-cycle, Blower
Scavenged Diesel
2-cycle,
Turbocharged Diesel
Control a
None
10% EGR
20% EGR
39% EGR
4° advance
None
48 retard
None
4.9% EGR
8.4% EGR
12. U EGR
Opacity, %
4.7
12
27.5
59
2.7
4.6
10
7.5
10.0
11.5
14.8
'All EGR data based on hot EGR.
Injection advance 1s not a control; data included to show trend,
-------
250
200
o
£ 150
100
u
M
SO
O GAS-FIRED UNITS
D OIL-FIRED UNITS
NOTE: DATA NOT ADJUSTED FOR
GASTURBINE EFFICIENCY
a
PROPOSED NSPS
a
o
a
to
15 20
TURBINE SIZF, MW
25
30
35
Figure 4-H. NOX emissions from large gas turbines without NOX controls (Reference 4-86).
4-112
-------
200
I
ISO
O GAS-FIRED UNITS
DOIL-FiRED UNITS
NOTE: DATA NOT ADJUSTED FOR
GAS TURfclNE EFFICIENCY
a:
u<
ci
100
o
u
50
PROPOSED NSPS
a
a
o.s
1.0
2.5
1.5 2.0
TURBINE SIZE, MW
Figure 4-15. NOX emissions from small gas turbines without HOX controls (Reference 4-86).
3.0
4-113
-------
Imposed on these figures is the proposed NSPS of 75 ppm for these sources. Very few units meet'these
standards in the uncontrolled state.
4.3.2.1 Control Techniques
Combustion modification techniques for gas turbines differ from those of boilers since turbines
operate at a lean air/fuel ratio with the stoichiometry determined primarily by the allowable turbine
inlet air temperature. The turbine combustion zone is nearly adiabatic and flame cooling for NO
A
control is achieved through dilution rather than radiant cooling. The majority of NOX formation in
gas turbines is believed to occur in the primary mixing zone, where locally hot stoichiometric flame
conditions exist. The strategy for NOX contra! in gas turbines is to eliminate the high temperature
stolchioroetric regions through water injection, premixing, improved'primary zone mixing, and down-
Streamed dilution.
Combustion modifications for gas turbines are classified into wet and dry techniques. Wet
methods, such as water and steam injection, presently provide substantial reductions. As yet, no
combination of dry methods has been successful on field units in reducing emissions below a typical
standard of 75 ppm NQX at 15 percent oxygen. Presently available wet and dry methods for NO reduc-
tion are aimed at either reducing peak flame temperature, reducing residence time at peak flame
temperature, or both. These techniques, along with their reduction potential and future prospects,
are shown in TABLE 4-45.
Wet techniques are the irost effective methods yet implemented with reduction potentials as
high as 90 percent for gas and 70 percent for oil fuels; With wet control, water or steam is intro-
duced into the primary zone either by premixing with the fuel prior to injection into the combustion
zone, by injection into the primary airstream, or by direct injection into the primary zone. The
effectiveness of each method is strongly dependent on atomization efficiency and primary zone resi-
dence time. In the case of water injection, peak flame temperatures are reduced further through
vaporization of the water.
Although NOX reduction is quite effective, numerous difficulties offer incentive to the
developmant of dry controls. If dry controls are developed as expected, the long-term future of
wet control does not appear promising based on the following inherent problems of wet controls:
Requirements for "clean" water or high-pressure steam
Hardware requirements which increase plant size
t Delivery system hardware which results in increased failure potential and overhaul/
maintenance time
4-114
-------
TABLE 4-45. GAS TURBINE - SUMMARY OF EXISTING TECHNOLOGY - COMBUSTION MODIFICATIONS
Modification
Wet Controls
Water Injection
Stead Injection
Dry Controls
Lean Out Prlnary
Zone
Increase Mass
FloMrate
Earlier Quench
with Secondary
Air
Air Blast or
Air Assist
Atomlzation
Reduce Inlet
Preheat
(Regenerative)
Other Minor
Combustor Modi-
fications and
Retrofit
Exhaust Gas
Recirculatlon
Catalytic
Combustion
Approach to NOX Reduction
Control Potential
Lower peak flame temp 50-90+ t
by utilization of
heat capacity and
heat of vaporization;
most effective when
Injected Into primary
flame zone
Same as above 50-90+ %
Lower peak flame temp 10-201
Reduce residence time To 159E
at peak temperatures
Reduce residence tine To 1SX
Reduce peak flame temp To -90%
by Increasing mixing
thereby reducing local
A/F ratio
Reduce peak flame temp Not Available
Reduce peak flame temp To 60S
through premlxlng, Combined
secondary air Injec-
tion, primary zone
flow reclrculatlon
Reduce peak flame To 38!
temperatures
Complete combustion To 98%
at lower peak
temperatures
Hear Term
To date, most effective
measure, reasonable cost
To date, most effective
measure after water
Injection
Attractive option, re-
quires additional con-
trols to meet standards
Attractive option If
feasible
Minor combustor modi-
fication used pre-
sently with wet
controls
Limited application
for retrofit
Not attractive due to
thermal efficiency
reduction
Attractive near tern
as an Interim solu-
tion
Option has seen use in
minor combustor modi-
fications
Technology not avail-
able
Far Term
Not as attractive as
dry controls, but
Is adequate If dry
not developed
Same as above
»
Generally seen as an
option to be incor-
porated Into new low
NOX designs
Not an attractive long
term option due to In-
flexibility
An attractive concept
to be employed in
advanced combustors
Promising method to be
Incorporated Into new
low NOX design
Not attractive for long
term solution
Unknown at this
time
An attractive option
for future design
with Internal com-
bustors
Attractive method for
new combustor designs
Additional Comments
Reduces efficiency. Increases capital
costs up to 10%. Operating costs as
low as It depending on usage. Hin-
dered by requirement for "clean"
water supply. Ineffective In reducing
fuel NOX.
Increases overall efficiency by In-
creasing flowrate. Installation and
operating costs same as water Injec-
tion. Requires high pressure steam.
Ineffective In reducing fuel NOX.
Possible decrease In power output,
less control over flame stabilization
Possible Increase In shaft speed
constant torque
An attractive option both for near
tern minor combustor modifications
and for Incorporation Into new de-
signs. Limited by flowrates and
Incomplete combustion.
Generally considered a new design
concept
Reduces efficiency.
With proper design, efficiency
unimpaired
Reduced efficiency requires additional
controls.
Current research aimed at reducing
reliability, maintainability, life-
time and start-up
Refs.
4-81
4-86
4-87
4-88
4-81
4-86
4-23
4-86
4-81
4-86
4-89
4-86
4-86
4-86
4-86
4-86
-------
Uncertainty regarding long-term control effects on turbine components.
Although few combinations of presently available dry controls have the NOX reduction poten-
tial of the wet methods, many dry techniques are used in conjunction with water or steam injection,
particularly on larger units. On the smaller units, dry controls may be sufficient to meet stan-
dards. The dry controls now available are:
Lean out primary zone Reduces NO levels up to 20 percent by Towering peak flame
temperatures. This option allows less control over flame stabilization and reduces
power output but is an attractive control to be built into future low NO combustors.
t Increase mass flowrates Possible NOV reductions up to 18 percent by reducing
J\ '
residence time at peak flame temperature. This control essentially increases
the turbine speed at constant torque and is not feasible in many applications.
Earlier quench with secondary air This is a minor combustor modification which
«
entails upstream movement of the dilution holes to reduce residence time at peak
temperatures. This is a promising control which is generally employed in advanced
combustor research.
Reduce inlet air preheat A control applicable only to regenerative cycle units.
It is not attractive due to reduction in efficiency.
A1r blast and air assists atomization Use of high-pressure air to improve atomiza-
tion and mixing requires replacement of injectors and addition of high-pressure air
equipment. This control is considered an excellent candidate for incorporation into
new low NOX design combustors.
Exhaust gas recirculation Possible NOX reduction of 30 percent. A candidate dry
control for future design, though it has limited application in some online units.
EGR requires extensive retrofit relative to other dry controls and also requires a
distinct set of controls for the EGR system.
Other minor combustor modifications are generally aimed at providing favorable interval flow
patterns in the primary zone and fuel/air premixing. The bulk of these modifications are combus-
tor-specific and are being investigated by the manufacturer. In general, dry controls available
for immediate implementation have not exceeded 40 percent NOX reduction and as such may be insuffi-
cient controls for the larger units at present. Since dry techniques approach NOX reduction dif-
ferently than do wet controls, their effects are complementary and, consequently, can be used toge-
ther. Figure 4-16 illustrates the effect of dry and wet controls used separately and in combination
4-116
-------
IIU
100
s
a.
-------
for liquid fuels (Reference 4-86). The figures show dry controls to be not sufficiently developed
to meet the standards, whereas wet controls are sufficient.
Future NOX control in gas turbines is directed toward dry techniques with emphasis on com-
bustor design. Medium term (1979-1985) combustor designs incorporate improved atomization methods
or prevaporization and a premixing chamber prior to ignition. These 'developmental combustors are
projected to attain emission levels of 20 ppm NOX at 15 percent oxygen,. A possible long term option
1s catalytically supported combustion. Laboratory tests have given NOX reductions of up to 98 per-
cent while maintaining stable, complete combustion. This concept - described in Section 3.1.5.2 of
this report-will probably require a new combustor design to accomodate it (Reference 4-81 and
4-86).
4.3.2.2 Costs
The most recent cost study of NOX controls for gas turbines has been performed by the EPA
.(Reference 4-86). Based on information presented in this study, the best available system of emis-
sion reduction considering costs are the wet systems. Wet systems can be applied to turbines imme-
diately and their cost impact is minimal. Although dry control techniques may be preferable because
of their minimal impact on efficienty, their complete development and application to large produc-
tion turbines is still several years away. Cost considerations for dry methods are, therefore, not
discussed. All costs in Reference 4-86 are stated in terms of 1975 dollars.
Table 4-46, derived from Reference 4-86, shows the expected increase in installed turbine cost
that will result from using water injection to control NOX to the proposed standard of 75 ppm. The
impact varies from 0.8 percent in the case of the 820 kW (1100 hp) standby unit to 7.1 percent for
the unit requiring extensive water treatment equipment.
Table 4-47 presents a summary of the costs in mills/kWh which would be incurred for 11
simple cycle turbine plants to meet the 75 ppm standard. This analysis was part of a cost model
developed in an EPA report (Reference 4-86). The results for each case are explained below.
Standby Units
The first two cases, S-l and S-2, differ only in the^number of hours operated annually. Unit
S-l operates 80 hours and S-2 200 hours per year. These units show the highest percentage impact in
terms of the incremental costs per net kWh of power generation. The low number of hours operated
each year tends to increase the cost of producing power because fixed costs are spread over a rela-
tively small base. The estimated impact in both cases was roughly 2.4 percent.
4-118
-------
TABLE 4-46. IMPACT OF NOX EMISSION CONTROL ON THE INSTALLED CAPITAL COST
OF GAS TURBINES (Reference 4-86, 1975 Costs)
I
-J
to
Application
A. Standby
1 . 260 kwa (350 hp)
2. 820 kWa (1100 hp)
B. Industrial
1. 3 MWa (4000 hp) -typical
2. 3 MWa - offshore
C. Utility
1 . 66 MWb
Installed Cost (1000$)
Uncontrolled
55.6
177.9
352.8
352.8
9900.0
Controlled
58.0
179.3
366.8
379.8
10070.0
% Increase
2.4
0.8
4.0
7.1
1.7
ashaft output
electrical output
-------
Table 4-47. WATER INJECTION COSTS, MILLS/kWh (Reference 4-86, 1975 Costs)
Item
Unit Size
Hours of Operation
Per Year
Annual 1 zed Fixed
Costs
Operating Cost of
Hater Treatment
Hater for Injection
Energy Penalty
Hater Transport
Costs
Output Enhancement
Total
Baseline Costs
Percent Impact
Standby
S-l
260 kH
80
13.65
0.46
0.01
O.S1
-
-
14.63
611.29
2.39
S-2
260 kU
200
5.46
0.46
0.01
0.51
-
-
6.44
264.79
2.43
S-3
820 kU
80
4.32
0.37
0.01
0.51
- «
-
5.21
611.29
0.85
S-4
820 kU
200
1.73
0.37
0.01
0.51
-
-
2.62
264.79
0.99
Industrial
1-1
3HU
2000
0.48
0.10
0.01
0.43
-
-
1.02
43.77
2.32
1-2
3HU
2000
0.12
0.10
0.01
0.43
-
0.09
0.57
32.53
1..75
1-3
3 KU
3000
0.12
0.10
0.01
0.43
0.64
0.09
1.21
35.53
3.71
Utility
U-l U-Z U-3 U-4
66 KU
200
2.58
0.11
0.01
0.34
-
-
3.04
180.00
1.69
66 MM
500
1.03
0.11
0.01
0.34
-
-
1.49
85.45
!.7S
66 HU
2000
0.26
0.11
0.01
0.34
-
-
0.72
38.20
1.88
6HH
8000
0.06
0.11
0.01
0.34
-
0.11
0.41
26.39
1.5S
Offshore
Drilling
Platform
11 GJ/hr
8000
0.23
0.35
-.
0.43
-
0.09
0.92
32.53
2.83
f\J
o
9Ut111tv turbine size (is electrical output, others «s shaft output
-------
Cases S-3 and S-4 are 820 kW (1100 hp) units operating the same number of hours, respectively,
as the smaller 260 kW units. These units can use exactly the same water purification system as the
smaller units. Since the costs of producing power independent of the water injection system (the
baseline cost) are identical between cases S-l and S-3 and S-2 and S-4, the percentage impact of
NOX control is decreased to less than one percent.
Industrial Units
Case 1-1 represents a normal, single shaft gas turbine application. The unit is operated
2000 hours per year and is slightly oversized. This negates any benefits that might be derived from
improved unit output. For Case 1-2, also a baseload turbine, a credit was taken for the improved
capacity of the unit.
The highest cost impact was recorded in Case 1-3, which represents a remote turbine applica-
tion in an arid climate in which water must be transported fifty miles at a cost of 2i per gallon.
The impact in such cases, including water storage facilities, is approximately a 3.7 percent in-
crease in the average cost of generating power. Since water injection results in a slight increase
in the power output capacity of the unit, a credit of 0.05 mills per kWh. was taken for the output
enhancement.
Utility Applications
The first unit operated 200 hours, the second 500 hours, the third 2000 hours, and the fourth
8000 hours annually. A credit for enhanced output was taken in the last case, since the unit is
baseloaded. In all four cases, the impact is less than 2 percent.
Offshore Drilling Platform
Initially, it was thought that this case would evidence the highest cost impact. The unit
was assumed to use sea water to fuel the water purification system, resulting in a substantial
increase in the capital and operating cost of the system. The installed cost of the water treat-
ment equipment was $27,000, compared to $14,,000 for an onshore application. Despite these higher
costs, the availability of water offset the costs associated with transporting water to the remote
gas compressing station application (1-3). The total cost of water injection for the offshore plat-
form was 0.92 mills/kWh compared to 1.21 mills/kWh for the remote site.
In the EPA cost model, no attempt was made to provide detailed estimates of the control costs
for regenerative and combined cycle gas turbines. The cost impacts, in abso?ute terms, are not
expected to be much greater than for simple cycles. Indeed, the percentage impacts will be less,
given the higher cost per kW of generating capacity of these units.
4-121
-------
In summary, the resulting estimates showed that, except for standby units, the total change
in costs will probably fall in the range of 0.4 to 1.5 mills per kWh for turbines used in industrial
and utility applications. This cost is equivalent to about a 2 percent increase in operating costs.
Control costs for standby units were much higher, ranging from 2 to 14 mills per kWh. This is pri-
marily due to their low use factor. This cost is equivalent to approximately a 2.5 percent increase
1n operating costs.
4.3.2.3 Energy and Environmental Impact
As was the case for reciprocating 1C engines, the energy impacts of applying NOX controls to
gas turbines occur almost solely through effects on unit fuel consumption, which were noted in the
foregoing discussion. Dry controls, except for reduced air preheat applied to regenerative cycle
turbines, have insignificant effects on unit efficiency. On the contrary, wet controls can impose
energy penalties. Water injection at the rate of 1 kg HgO/Kg fuel reduces turbine efficiency by
about 1 percent. If waste steam is available, steam injection can increase turbine efficiency by
increasing turbine power output at constant fuel input. But, if a fuel debit is taken for heat
needed to raise injection steam, overall plant efficiency losses comparable to those experienced
with water injection will occur.
Again, as with 1C engines, gas turbines emit only an exhaust gas effluent stream and fire
"clean" fuels. Thus the potential environmental impacts of NOX controls applied to gas turbines
will occur through incremental effects on emissions levels of exhaust gas CO, HC, and particulate
(smoke). Effects through liquid and solid effluents need not be considered, and incremental
impacts on SO , trace metal, and, to some extent, higher molecular weight organic emissions are
insignificant.
The effects of some comnonly applied NOX control techniques on CO emissions from gas turbines
are summarized in Table 4-48. From the table," it is apparent that dry controls, notably leaning
the primary zone and air blast (or air-assist atomization) reduce CO levels. This is expected since
the additional air introduced into the combustor when applying these techniques allows more complete
fuel combustion. On the other hand, wet control techniques, such as water injection, tend to quench
combustion and give lower combustor temperatures. This leads to incomplete combustion and increased
CO levels as shown in Table 4-48..
The very limited data on incremental hydrocarbon emissions due to NOX combustion controls
applied to stationary gas turbines are summarized in Table 4-49. As the table shows, the effects
of dry NOX controls are mixed. Air blast tends to increase HC emissions while leaning the primary
zone tends to decrease HC levels. Increased combustion efficiency due to higher combustion
4-122
-------
TABLE 4-48. REPRESENTATIVE EFFECTS OF NOX CONTROLS'ON CO EMISSIONS FROM GAS TURBINES (Reference 4-86).
NOX Control
Lean Primary Zone
Air Blast/
Piloted Air Blast
Water Injection
Fuel
Natural Gas
Kerosene
"Diesel
Kerosene
Diesel
Natural Gas
(*
Diesel
CO Emissions (ppm)a
Basel i ne
102
102
53
195
969
53
147
252
99
135
93
NOX Control
51
96
99
59
no
36
1134
1512
144
162
30
02> dry basis. Emission:; levels at full load.
TABLE 4-49 SUMMARY OF THE EFFECTS OF NOX CONTROLS ON VAPOR PHASE HYDROCARBON
EMISSIONS FROM GAS TURBINES (Reference 4.36).
NOX Control
Air Blast
Lean Primary Zone
Water Injection
Fuel
Jet-A
Natural Gas
Diesel Fuel
Kerosene
Natural Gas
Diesel Fuel
HC Emissions (ppm)a
Baseline
18
9
33
30
3
27
234
141
36
24
NOX Control
41
11
9-12
12
7
12
372
246
27
12
Comment
Idle
Full load
( Full load
/ W/F = 0.5
a3% 02, dry basis.
4-123
-------
temperatures tends to support this latter observation. The effects of applying wet controls are
also mixed. As indicated in the table, with water injection at a water-to-fuel (W/F) weight ratio
of 0.5, HC emissions increased for turbines having high baseline HC emissions, but decreased for
turbines which emitted low baseline HC levels.
The data on particulate emissions from gas turbines resulting from applied NO controls are
also very limited and are as inconclusive regarding the increment in particulate emissions from NO
X
controls as those for incremental CO and hydrocarbon emissions. For example, the effect of water
injection on particle emissions seems to be related to the specific injection method used
(Reference 4-86). Some tests show smoke level reduction of 1.5 to 1.75 smoke sopt numbers when
water injection is used. Other tests, however, indicate particulate emissions with water injection
at peak load.
4.4 SUMMARY
Table 4-50 summarizes current and emerging NOX control technology for the major source
categories (1974 costs). As shown, most of the- current NO control technologies rely on combustion
process modification. Emerging technologies include various combustion modification and flue gas
treatment processes. However, combustion process modifications will likely predominate in the U.S.
except for situations where very stringent emission levels are required.
Historically, utility boilers have been the most extensively regulated source category, and
accordingly, NO control technologies for utility boilers are the most advanced. Available combus-
tion modification technology ranges from simple operational adjustments such as low excess air,
biased burner firing, and burners out of service to application of overfire air ports, flue gas
recirculation, low NO burners, and enlarged furnace designs. Some adverse operational impacts have
been experienced with the use of combustion modification on existing equipment. In general these
have been solved through combustion engineering or by limiting the degree of control application.
With factory-installed controls on new equipment, operational problems have been minimal.
The technology for other sources is less well developed. Control techniques shown effective
for utility boilers are being demonstrated on existing industrial boilers. Here, as for utility
boilers, the emphasis in emerging technology is on development of controls applicable to new unit
design. Advanced low NO burners and/or advanced off-stoichiometric combustion techniques are the
X
most promising concepts. This holds true for the other source categories as well. The R&D emphasis
for gas turbines and reciprocating 1C engines is on developing optimized combustion chamber designs
matched to the burner or fuel/air delivery system.
4-124
-------
TABLE 4-50. SUMMARY OF NOV CONTROL TECHNOLOGY
A
Equipment/
Fuel Category
Existing coal -fired
utility boilers
New coal -fired
utility boilers
Existing oil-fired
utility boilers
Existing gas-fired
utility boilers
Oil-fired industrial
watertube boilers
Stoker-fired,
industrial water-
tube boilers
Gas-fired industrial
watertube boilers
Current Technology
Available
Control
Technique
LEA + OSC
(OFA, BOOS,
BBF); new
burners
LEA + OFA;
new burners;
enlarged
furnace
design
LEA + OSC *
FGR;
LEA + OSC
+ FGR;
LEA + OSC
(OFA, BOOS,
BBF); FGR;
new burners
LEA + OFA
LEA + OSC
(OFA, BOOS,
BBF); FGR;
new burners
Achievable
NO Emission
l.8vel,ng/J
(lb/10° Btu)
215 - 300
(0.5 - 0.7)
170 - 260
(0.4 - 0.6)
86 - 170
(0.2 - 0.4)
43 - 100
(0.1 - 0.25)
85 - 130
(0.2 - 0.3)
150 - 190
(0.35-0.45)
86 - 130
(0.2 - 0.3)
Estimated
Differential
Capital Cost
$0.7 - 2.2/kVJ
1978$ e
$2 - 3/kW
1978$ e
$6/kH
1978$e
$6/kW
1978$e
$1 - 1.5/kW.
1978$ *
$1 - 3/kW.
1978$ l
$1.5/kU.
1978$ fc
Operational Impact
Possible increase
in corrosion &
slagging & carbon
in fly ash
No major problems
Possible flame
instability;
boiler vibration
Possible flame
instability;
boiler vibration
Possible IX
increase in fuel
consumption; flame
instability;
boiler vibration
(retrofit)
Possible 1%
increase in fuel
consumption; .
corrosion; slagging
of grate (retrofit)
Possible 1%
increase in fuel
consumption; flame
instability;
boiler vibration
(retrofit)
Emerging
Technology
Advanced low-NO
burners; SCR; SRR
Advanced low-NO
burners; advanced
furnace designs;
SCR; SNR;
fluidized bed
combustion
Low-NO burners;
oil defiitrifica-
tion; SCR; SNR
Low-NO burners;
SCR; SNR
Low-NO burners;
OFA in new unit
designs, optimized
burner/ firebox
designs; oil
denitrifi cation;
SCR; SNR
Inclusion of OFA
in new unit design;
fluidized bed
combustion; SNR
Low-NO burners;
OFA in new unit
design; optimized
burner/firebox
designs; SNR
Comments
Flue gas treatment techniques are
potential supplements to combustion
modifications if needed
Flue gas treatment techniques are
potential supplements to combustion
modifications if needed
No new units; emission levels are
limit of current technology
No new units; emission levels are
limit of current technology
Current technology still undergoing
development
Current technology still undergoing
development
Current technology still under-
going development
ro
tn
-------
TABLE 4-50. SUMMARY OF NOV CONTROL TECHNOLOGY (Continued)
3\
Equipment/
Fuel Category
Industrial flretube
boilers
Gas turbines
1C engines
Current Technology
Available
Control
Technique
LEA + FGR;
LEA + OSC
Water, steam
injection
Fine tuning;
changing A/F
Achievable
NOX Emission
Level fing/J
(lb/100 Btti)
65 - 110
(0.15 - 0.25)
110 - 150
(0.25 - 0.35)
1,070 - 1,290
(2.5 - 3.0)
Estimated
Differential
Capital Cost
$6/kVL
1978$c
$1 - 2/kW
1975$
$0.70 - 2.00/kW
($0.50 - 1.50/BHP)
1975$
Operational Impact
Possible IS
Increase 1n fuel
consumption; flame
instability;
(retrofit
Possible 1%
increase in fuel
consumption;
affects only
thermal NOX
5 - 1Q% increase
in fuel consump-
tion; misfiring;
poor load response
Emerging
Technology
Low-NO burners;
OFA or FGR in new
unit design;
optimized burner/
firebox design
Advanced combustor
designs for dry
NO controls;
catalytic combus-
tion; advanced
can designs
Include moderate
control in new
unit design;
advanced head
designs
Comments
Development continuing on current
technology
Current technology widely used
Technology still being tested
I
t-»
01
-------
REFERENCES FOR SECTION 4
4-1 Maxwell, J.D. and L.R. Humphries, "Evaluation of the Advanced Low-N0v Burner, Exxon, and
Hitachi Zosen DeNOx Processes," EPA-600/7-81-120, TVA/OP/EDT-81/28, July 1981. p. 5.
4-2 Lim,, K.J., e_t aQ_., "Environmental Assessment of Utility Boiler Combustion Modification NO
Controls; Volume 1. Technical Results," EPA-600/7-80-075a, April 1980. pp. 4-1 to 4-56.
4-3 Waterland, L.R., et al., "Environmental Assessment of Stationary Source NO Control
Technologies FTnaTReport," EPA-600/7-82-034, May 1982. pp. 133-196. x
4-4 Manny, E.H., and P.S. Natanson, "Fireside Corrosion and NO Emission Tests on Coal-Fired
Utility Boilers," In: Proceedings of the Joint Symposium on Stationary Combustion NO
Control, Volume I, Utility Boiler NO Control by Combustion Modification. IERL - RTP-I083,
October 1980 (as cited in Reference 4-3).
4-5 Barsin, J.A., "Pulverized Coal Firing NO Control," In: Proceedings Second NO Control
Technology Seminar, EPRI FP-1109-SR, Jul$ 1979 (as cited in Reference 4-3).*
4-6 Reference 4-1, 111 pp.
4-7 Ando, J., "NOX Abatement for Stationary Sources in Japan," EPA-600/7-79-205, August 1979.
4-8 Faucett, H.L., J.D. Maxwell, and T.A. Burnett, "Technical Assessment of NO Removal
Processes for Utility Application," TVA Bulletin Y-120, EPA-600/7-77-127 (fiTIS PB 276
637/6WP), EPRI FP-1253, 1977 (as cited in Reference 4-6).
4-9 Jones, G.D., "Selective Catalytic Reduction and NO Control in Japan, A Status Report,"
EPA-600/7-81-030, January 1981. ' x
4-10 Reference 4-9, p. 10.
4-11 Reference 4-1, p. 9.
4-12 Ando, J., "SO, and NO Removal for Coal-Fired Boilers in Japan," Prepared for the Seventh
Symposium on Flue Gas Desulfurization, May 17-20, 1982. p. 5.
4-13 Mobley, J.D. and J.M. Burke, "EPA's Pilot Plant Evaluations of NO and NO /SO Flue Gas
Treatment Technology," prepared for the Seventh Symposium on Flue Gas Desulfurization,
May 17-20, 1982.
4-14 Reference 4-12, p. 10.
4-15 Reference 4-9, pp. 1-5.
4-16 Johnson, L.W., C.L.W. Overduin, D.A. Fellows, "Status of SCR Retrofit at Southern
California Edison Huntington Beach Generating Station Unit 2." In: Proceedings of the
Joint Symposium on Stationary Combustion NO Control, Volume II Utiliby Boiler NO Control
by Flue Gas Treatment, IERL-RTP-1084. Octobir 1980.pp. 32, 33. *
4-17 Trexler, E.G., "OOE's Electron Beam Irradiation Developmental Program," Prepared for the
Seventh Symposium on Flue Gas Desulfurization, May 17-20, 1982. p. 11.
4-18 Reference 4-2, pp. 7-1 to 7-49.
4-19 Reference 4-3, pp. 149 to 154.
4-20 Maxwell, J.D., T.A. Burnett, H.L. Faucett, "Preliminary Economic Analysis of NO Flue Gas
Treatment Processes," EPA-600/7-80-021, EPRI FP-1253, TVA ECDP B-6, February 1980.
4-21 McGlamery, G.G.,-et al., "Detailed Cost Estimates for Advanced Effluent Desulfurization
Processes," EPA-60072^75-006, January 1975 (as cited in Reference 4-18).
4-22 'Waitzman, D.A., et al., "Evaluation of Fixed-Bed Low-Btu Coal Gasification Systems for
Retrofitting Power Plants," EPRI Report 203-1, February 1975 (as cited in Reference 4-18);
4-127
-------
4-23 Ponder, W.H., R.D. Stern, and G.6. McGlamery, "SO Control Methods Compared," The Oil and
Gas Journal, December 1976. pp. 60-66 (as cited tn Reference 4-18).
4-24 Engdahl, R.B., "The Status of Flue Gas Desulfurization," ASME Air Pollution Control
Division News, April 1977 (as cited in Reference 4-18).
4-25 Princiotta, F.T., "Advances in SO Stack Gas Scrubbing," Chemical Engineering Progress,
February 1978. pp. 58-64 (as cited in Reference 4-18).
4-26 Campobenedetto, E.J., Babcock & Wilcox Co., Barberton, OH, Letter to K.J. Lim, Acurex
Corporation, Mountain View, CA, November 15, 1977 (as cited in Reference 4-19).
4-27 Vatsky, J., "Effectiveness of NO Emission Controls on Utility Steam Generators," Foster
Wheeler Report to Acurex Corporation, FW Contract 2-43-3245, Livingston, NJ, November 1978
(as cited in Reference 4-19).
4-28 Martin, G.B., "Field Evaluation of Low NO Coal Burners on Industrial and Utility
Boilers," In: Proceedings of the Third Stationary Source Symposium; Volume I,
EPA-600/7-79-050a, February 1979 (as cited in Reference 4-19).
4-29 Johnson, S.A., et al., "The Primary Combustion Furnace System An Advanced Low-NO
Concept for Pulverized Coal Combustion," In: Proceedings: Second NO Control Technology
Seminar. EPRI FP-1109-SR, July 1979 (as cited in Reference 4-19).*
4-30 Teixeira, D., "NO Control Technology." EPRI Journal. V3(9):37, November 1978 (as cited
1n Reference 4-19J.
m
4-31 "Technical and Economic Feasibility of Ammonia-Based Post Combustion NO Control",
EPRI CS-2713, Electric Power Research Institute, Palo Alto, California, November 1982.
4-32 Reference 4-12, p. 19.
4-33 Reference 4-2, pp. 8-39 to 8-41.
4-34 Reference 4-6, pp. 81-85.
4-35 Reference 4-20, pp. xxv to xxviii.
4-36 Reference 4-13, p. 8.
4-37 Reference 4-17, pp. 5,6.
4-38 Reference 4-3, 448 pp.
4-39 Reference 4-9, pp. 6,7.
4-40 Devltt, T., et al., "The Population and Characteristics of Industrial/Commercial Boilers,"
EPA-600/7-79^I75a", August 1979.
4-41 "Non-fossH-fuel Industrial Boilers, Background Information," EPA-450/3-82-007,
Environmental Protection Agency, Research Triangle Park, NC, March 1982.
4-42 Cato, G. A., jet al_., "Field Testing: Application of Combustion Modification to Control
Pollutant Emissions from Industrial BoilersPhase II," EPA-600/2-76-086a, April 1976.
4-43 Burklin, C. E., and W. D. Kwapil, "Regressions for NO Emissions from Oil- and Gas-Fired
Industrial Boilers," EPA Contract No. 68-02-3058, Radtan Corporation, Durham, NC,
May 27, 1982.
4-44 "Fossil Fuel-Fired Industrial Boilers, Background Information, Volume I: Chapters 1-9,"
EPA-450/3-82-OQ6a, Environmental Protection Agency, Research Triangle Park, NC,
March 1982.
4-45 McElroy, M. W., and D. E. Shore, "Guidelines for Industrial Boiler Performance
Improvement," EPA-600/8-77-003a, January 1977.
4-128
-------
4-46 'Hunter, S. C., and H. J. Buening, "Field Testing: Application of Combustion Modifications
to Control Pollutant Emissions from Industrial Boilers, Phase I and II, Data Supplement,
"EPA 600/2-77-122, June 1977.
4-47 Chicanowica, J. E., et jil_., "pollutant Control Technique for Packaged Boilers. Phase I.
Hardware Modifications and Alternate Fuels," Ultrasystems 'Draft Report, EPA Contract
No. 68-02-1498, Ultrasystems Corporation, Irvine, CA, November 1976.
4-48 Carter, W. A., et al., "Emissions Reduction on Two Industrial Boilers with Major
Combustion ModiTTcations," EPA-600/7-78-099a, June 1978.
4-49 Cato, G. A., et al., "Field Testing: Application of Combustion Modifications to Control
Pollutant Emission's from Industrial BoilersPhase I," EPA-600/2-74-078a, October 1974.
4-50 Lim, K. J., et al., "Environmental Assessment of Utility Boiler Combustion Modification
NO Control sT1" ETA Contract No. 68-02-2160, Acurex Corporation, Research Triangle
Park, NC, April 1978.
4-51 Palazzolo, M. A., "Air Preheat vs. Economizers," Technical Note, Radian Corporation,
Durham, NC, September 2, 1981.
4-52. Mason, H. 8., et al., "Preliminary Environmental Assessment of Combustion. Modification
Techniques: VoTuie II, Technical Reports," EPA-600/7-77-119b, October 1977.
4-53 Langsjoen, P. L., ejb al., "Field Tests of Industrial Stoker Coal-Fired Boilers for
Emissions Control and~E~fficiency ImprovementSite K," EPA-600/7-80-138a, May 1980.
4-54 Lim, K. J., C. Castaldini, and H. I. Lips, "Industrial Boiler Combustion Modification NO
Control: Volume \, Environmental Assessment," EPA-600/7-81-126a, July 1981.
4-55 Burklin, C. E., "'Application of Staging Burners to Industrial Boilers,""Draft Technical
Note, EPA Contract No. 68-02-3058, May 1982.
4-56 American Boiler Manufacturers Association, "Emissions and Efficiency Performances of
Industrial Coal Stoker-Fired Boilers, Volume 1," DOE/ET/10386-T1 (Vol.lj, Arlington, VA,
August 1981.
4-57 "Environmental Assessment of Stationary Source NO Control Technologies: Final Report,"
EPA-600/7-82-034, Environmental Protection Agency, Research Triangle Park, NC, May 1982.
4-58 Maloney, K. L., et al., "Low Sulfur Western Coal Use in Existing Small and Intermediate
Size Boilers," EPA"-600/7-78-l53a, July 1978.
4-59 Gabrielson, J. E., et al., "Field Test of Industrial Stoker Coal-Fired Boilers for
Emissions Control and "ETficiency Improvement - Site A," EPA-600/7-78-136a, July 1978.
4-60 Keller, L. E., and M. S. Jennings, and W. D. Kwapil, "Regressions for NO Emissions from
Coal-Fired Spreader Stoker Industrial Boilers," EPA Contract No. 68-02-3058, Radian
Corporation, Durham, NC, July 22, 1982.
4-61 Carter, W. A., et al., "Thirty-day Field Tests of Industrial Boilers: Site 1Coal-Fired
Spreader Stoker"?1' EPA-600/7-80-085a, April 1980.
4-62 Carter, W. A. and J. R. Hart, "Thirty-day Field Tests of Industrial Boilers: Site 4
Coal-fired Spreader Stoker," EPA-600/7-80-085d, April 1980. ,
4-63 Carr, R. C., "Effectiveness of Gas Recirculation and Staged Combustion in Reducing NO on
a 560-MW Coal-Fired Boiler," Electric Power Research Institute, NTIS No. PB-260-282, x
September 1976.
4-64 ' Carterm W. A., et. al_., "Emission Reduction on Two Industrial Boilers with Major Combustion
Modifications," EPA-600/7-78-099a, June 1978.
4-65 Exxon Research and Engineering Company, "Exxon Thermal DeNo Process," New Jersey, Exxon
Technology, April 1978.
4-129
-------
4-66 "Exxon Corp. Stationary NO Emissions Significantly Reduced at Plant," Air and Water
Pollution Report, February 20, 1978.
4-67 Jones, 6. D., and K. L. Johnson, "Technology Assessment Report for Industrial Boiler
Applications: NO Flue Gas Treatment, Final Report," EPA-600/7-79-178g, December 1979.
4.-6S Heap, M. P., et al., "Reduction of Nitrogen Oxide Emissions from Field Operating Package
Boilers, Phase~~lTT," EPA-600/2-77-025 , January 1977.
4-69 Schwieger, R. , "Industrial Boilers - What's Happening Today," Power, February 1977.
4-70 Schwieger, R., "Industrial Boilers - What's Happening Today, Part II," Power,
February 1978.
4-71 Personal communication, K.J. Lim of Acurex Corporation with J. Lindsay, Zurn Industries,
San Francisco, CA, August 17, 1978.
4-72 Lira, K. J., et al,, "Environmental Assessment of Utility Boiler Combustion Modification
NOV Control s7rlPA-600/7-80-075a and b, April 1980.
A
4-73 Personal communication, K.J. Lim of Acurex Corporation with B. Morton, E. Keeler, Co.,
Williamsport, PA, August 8, 1978.
4-74 Lyon, R. K. , and J. P. Longwell, "Selective, Non-Catalytic Reduction of NOV by NH,,"
Proceedings of the NO Control Technology Seminar, EPRI SR-39, NTIS-PB 2532661, J
February 197FT
4-75 Wong-Woo, H., and A. G. Goodley, "Observation of Flue Gas Desulfurization and Denitrifica-
tion Systems in Japan," California Air Resources Board, SS-78-OOK, March 7, 1978.
4-76 Vargo, G. M. , Jr., et al . , "Applicability of the Thermal DeNO Process to Coal-Fired
Utility Boilers," EPA-600/7-79-079 , March 1979. x
4-77 Cato, G. A., et al., "Reference Guideline for Industrial Boiler Manufacturers to Control
Pollution witFCoinbustion Modification," EPA-600/8-77-003b, November 1977.
4-78 Koppang, R. R. , "A Status Report on the Commercialization and Recent Development History ,
of the TRW Low NO Burner," TRW Internal Report, TRW, Inc., Redondo Beach, CA,
January 1976.
4-79 McGowin, C.R., "Stationary Internal Combustion Engines in the United States,"
EPA-R2-73-210, April 1973.
4-80 "Standard Support and Environmental Impact Statement - Stationary Reciprocating Internal
Combustion Engines," (Draft Report). Acurex Corp. /Aero therm Division, Mountain View,
California, Project 7152, March 1976.
4-81 Aerospace Corporation, "Assessment of the Applicability of Automotive Emission Control
Technology to Stationary Engines," EPA-650/2-74-051, July 1974.
4-82 The American Society of Mechanical Engineers (ASME), "Power Costs, 1974 Report on Diesel
and Gas Engines," March 1974.
4-83 Calspan Corporation, "Technical Evaluation of Emission Control Approaches and Economics of
Emission Reduction Requirements for Vehicles Between 6000 and 14000 Pounds GVW,"
EPA-460/73-005, November 1973.'
4-84 Bascom, R.C., et al., "Design Factors that Affect Diesel Emissions," SAE Paper 710484, .
July 1971.
4-85 Hills, F.J., et al., "CRC Correlation of Diesel Smokemeter Measurements," SAE
Paper 690493 ,~Hay~1969.
4-86 "Standards Support and Environmental Impact Statement, Volume I: Proposed Standards of
Performance for Stationary Gas Turbines," EPA-450/2-77-017a, September 1977.
4-130
-------
4-87 Shaw, H., "The Effects of Water, Pressure and Equivalence Ratio on Nitric Oxide Production
in Gas Turbines," ASME Paper 73-WA/GT-l.
4-88 Hilt, M.B. and Johnson, R.H., "Nitric Oxide Abatement in Heavy Duty Gas Turbine Combustion
by Means of Aerodynamic and Water Injection," ASME Paper 72-GT-53.
4-89 Stern, R.D., "The EPA Development Program for NO Flue Gas Treatment," In: Proceedings of
the National Conference on Health, Environmental Effects, and Control Technology of Energy
Use, EPA-600/7-76-002, February 1976.
4-131
-------
-------
SECTION 5
OTHER COMBUSTION PROCESSES
Significant amounts of the total fuels burned and NO emissions released in the United States
A
are associated with small-scale combustion processes. These include important nonindustrial uses in
domestic and commercial heating, hot water supply, a wide variety of incinerators, and open burning
of solid wastes. The contribution to ambient N02 can be significant, particularly in localized,
residential5 areas. Control techniques, costs, and energy and environmental impacts are discussed
for those systems where data are available.
5.1 SPACE HEATING
Emissions from stationary source fuel combustion in 1980 were present in Section 2. As shown
in Table 2-2, commercial and residential combustion sources contributed about 6 percent to the
national NO emissions. A study of 1977 NO, emissions showed that 56 percent of the total emissions
A A
from residential and commercial combustion sources were from residential space heating systems and
28 percent were from commercial space heating systems (Reference 5-1).
Natural gas and distillate oil are the primary fuels burned in residential heating systems.
The combination of gas and oil accounted for 90 percent of the fuel burned for domestic heating in
1976. Gas and oil are also the primary fuels consumed in commercial heating systems. The
predominance of gas and oil is also reflected in the distribution of heating equipment types in the
residential sector. Electrical heaters accounted for nearly 14 percent of residential heating
equipment in 1976; LPG, coal, and wood-fired units accounted for 8 percent; and the remaining
equipment was gas- and oil-fired (Reference 5-1).
Residential.heading units are characterized by thermostatically controlled heating cycles
(on/off cycles). Natural gas-fired residential heating equipment generally employs single port
upshot or tubular multiport burners. A standing pilot flame is often used to ignite the burner;
however, there is a rapid trend toward interrupted pilots which are turned off in the standby mode.
Oil-fired residential heating systems usually use high pressure atomizing gun-type burners. Nearly
all new oil-fired systems use the flame retention burner because of its high efficiency. Residen-
tial coal-fired furnaces are generally stoker fed and wood-fired heaters are usually hand fed
(Reference 5-1).
5-1
-------
Commercial heating systems can be divided Into three general categories: space heaters, warm
air furnaces, and hot water or ot^am systems. Warm air furnaces may be direct or indirect-fired.
Qiriic".-fired neoters use clean gaseous fuels and exhaust combustion products directly into the
heated soace. Indirect fired heaters vent to the outdoors a>id an? simitar to residential warm air
furnace designs. Commercial steam and hot water units include wtitertubes, firetubes, and cast iron
Doi'ieis. Gas- and uii-fired boilers normally employ single powe" burner designs. Some atmospneric
gas burners are used for snail units. Coal-fired steam and hot water units are underfed stoker
units (Reference 5-1).
5.1.1 Emissions
Hall, e_t a_K (Reference 5-2) studied the factors that affect emission levels from residential
heaters. This project, which concentrated or. an oil-fired warm air furnace, showed thet excess air,
resiij,.ce time, flame retention devices, and maintenance are major factors in the control of
emissions.
As shown in Figure 5-1, emissions of CO, HC, smoke, and purticulate matter pass through a
ninimum as excess air is increased from stoichiometric conditions. By contrast, both therm?1
efficiency and NO '"-missions pass through maximum points as excess air is increased. The experi-
Tiental results showed that increased residence time of the combustion products reduces emissions of
CO gaseous HC, and smoke but has no ffeu on N0x emissions. Combustion chamber material was found
to affect i':1 emissions. Fu-riaces wltt steel-linecl chambers requirsd higher excess air levels to
rearfi optimum emission levels, thus reducing efficiency. The shape of the combustion chamber had
little effect on pollutant generation. A specially designed flame retention device designed to
decrease participate emission* was found to increase NOX emissions, but such a dtvice also increased
furn?i;e efficiency. Poorly maintained f^rnac^s also yielded higher NOX emissions.
In another study rf spa-e heating equipment (Reference 6-3), emission levels 'vere found to be
dependent upon boiler size, design, b'"-ner tyoe, burner age, and operating conditions. The type of
fuel used i-i the combustion equipment for space heating al^o affect; NOX emissions because of fuel
trogen convt-sion (Reference b-1). Ot tha fuels typically used for space heating, coal and
residual oil <.:antain significant quantities of nitroyen which may be converted to NO emission,
Conversion rates lor coal rarje from 20 to 60 percent and conversion rates in residential and
coirtr.e.'cia 1 heatiincj systems . « on tne hiqh sicie of the range (Reference S-l), Because natural gas
any nial oi'i i:"9 ve/' low in fuel nitrocfcii, N0x ^reduction from natural gas ana distillate oil
-------
.2.0
OPTIMUM SETTING FOR MINIMUM
EMISSIONS AND MAXIMUM
SMOKE EFFICIENCY
(10TH
8
I
X
>-
a:
sc
a
1.4 1.6 1.8 2.0
STOICHIOMETRIC RATIO
2.4
2.6
Figure 5-1. Stneral trend of smoke, gaseous emissions, and effi-
ciency versus stoichiometric ratio for residential
heaters (Reference 5-4).
5-3
-------
systems is due primarily to tnerml NO which is promoted by high temperatures and long residence
times (Reference 5-1).
NO emission factors for domestic heatinc reported in AP-42 (Reference 5-5) are summarized in
Table 5-1. Some additional emission rates reported in the literature are summarized in Table 5-?..
Allen (Reference 5-6) found that although NO emissions from wood stoves have not been
recognized as serious they can be significant. He found that even though the fuel nitrogen content
in wood is low, conversion can be significant. He further found fiat temperature: within wood
stoves are not expected to reach the level where atmospheric nitrogen fixation would occur.
5.1.2 Control Techniques
Currently available emission reduction techniques for space haatir.j units are: (1) tuning:
the bust adjustment in terms of the smoke-COj relationship that can be achieved by normal cleanup,
nozzle replacement, simple scaling and adjustment with tha benefit of Vielo instruments, (2) equip-
ment replacement: installation of a new, advanced low-NOx uriit, and installation of a new
low-emission burner.
5.1.2.1 Tuning
Reference 5-3 indicates that tuning has a beneficial effect on all pollutants with the
exception of NO. In the field program, oil-fired un'ts considered in "poor" condition were
replaced and all others were tuned, resulting in reductions in smoke, CO, HC, and filterable
particulate matter by 59, 81, 90, and 24 percent respectively, with no significant change in NO
levels. Table 5-3 shows mean emission levels prior to and after replacement or tuning. Although
tuning or replacement '?as been snown to have little effect on NO levels, yearly inspection
accompanied by one of these techniques is highly recommended since other pollutant levels are so
greatly reduced.
As an aid to controlling emission levels from residential and commercial space heating
systems, EPA has made available guidelines * oil and gas burner adjustments (References 5-7, 5-2
5-9). These guidelines are intended for the jse of skilled technicians and for training service
personnel. The reroronended adjustment guidelines provide for efficient fuel utilization and
rimmize air pollution with reliable automatic operation.
5-4
-------
TABLE 5-1. EMISSION FACTORS FOR RESIDENTIAL AND COMMERCIAL
HEATING SYSTEMS*
FUEL
EMISSION FACTOR
Gas
Distillate Oil
Anthracite
Bituminous
Lignite
1280 - 1920 kg/106 m3
(80 - 120 lb/105 ft3)b
2.3 kg/10J liter
18 lb/103 gal)
1.5 - 9 kg/Mg
(3 - 18 Ib/ton)
3 kg/Mg
(6 Ib/ton)
3 kg/Mg
(6'lb/ton)
a AP-42 (Referenrj 5-5).
Lower value for residential systems. Higher value for commercial systems,
5-5
-------
TABLE 5-2. UNCONTROLLED N0v EMISSIONS FROM RESIDENTIAL SPACE HEATING SYSTEMS
A
FUEL
N0x EMISSION RATE (as N02)
SOURCE
Natural gas (blue flame)
Natural gas (yellow flame)
.42 ng/J (0.098 Ib m/10b Btu)
38 ng/J (0.088 Ib m/106 Btu)
(Reference 5-10)
Laboratory tests
in
I
Oi
Natural gas (mostly boilers)
Natural gas (mostly.hot air
furnaces)
42 ng/J (0.098 Ib m/10b Btu)
49 ng/J (0.114 Ib m/106 Btu)
(Reference 5-11)
Field tests: measurements
of emissions in chimneys of
natural gas heated homes.
Fuel oil
fri.5 ng/J "(0.143 lb/10b Btu)
1.8 g/kg fuel
(Reference 5-3)
Field tests
Anthracite
Bituminous
3.9 g/kg fuel
0.9 g/kg fuel
(Reference 5-12)
Composites of many emission
rate measurements.
-------
TABLE 5-3. COMPARISON OF MEAN EMISSIONS FOR CYCLIC RUNS ON RESIDENTIAL
OIL-FIRED UNITS (REFERENCE 5-3)
Units
All Units
All units,
except those
in need of
replacement
Condition
As Found
Tuned
As Found
Tuned
Units
In
Sample
32
33
29
30
Mean
Smoke
No.
-
-
3.2
1.3
Mean.,Emission Factors
kg/rrf (lb/1000 gal)
CO
>2.65
(>22.1)
>1.96
(>16.4)
0.93
(7.8)
0.52
(4.3)
HC
0.68
(5.7)
0.36
(3.0)
0.09
(0.72)
0.07
(0.57)
NOX
2.32
.C19.4)
2.34
(19.5)
2.35
(19.6
2.34
(19.5)
Filterable
Parti cul ate
0.35
(2.9)
0.28
(2.3)
0.29
(2.4)
0.26
(2.2)
Ul
I
-4
-------
5.1.2.2 Equipment Replacement
Control equipment options for controlling NO from gas- and oil-fired residential furnaces
are listed in Tables 5-4 and 5-5. Each of the tables also shows a conventional unit for purposes of
comparison.
Gas-Fired Equipment
The American Gas Association Laboratories developed radiant screens capable of reducing NO
emissions from 36 to 76 percent with an average reduction of 58 percent. Incandescent radiant
screens in a natural gas flame radiate heat to the surroundings and cool the flame. Test results
showed that performance was more significant for multiport burners than for single port burners.
The screens were also found to increase steady-state furnace efficiency slightly because of
increased flame and burner radiation (Reference 5-1). The Gas Appliance Manufacturer Association
has reported installation and performance problems with the screens. Some of the potential problems
cited were performance sensitivity to screen location, effect on CO emissions, and deterioration due
to thermal shock (Reference 5-1).
Secondary air baffles control secondary air flow into the flame front. Decreasing the
concentration of excess oxygen at peak temperatures with reductions in secondary air was found to
decrease NOX emissions from 10 to 40 percent (Reference 5-10). However, NQX reductions without
increases in CO emissions were generally limited to about 15 percent; (Reference 5-1). The Gas
Appliance Manufacturer Association expressed concern over reliability and performance of secondary
air baffles. Furthermore, it is not clear that secondary air baffles can be applied to all types of
residential furnaces (Reference 5-1).
A surface combustion burner employs surface combustion or premixed natural gas and air on a
refractory material. The burner radiates heat to an air-cooled firebox, and the combustion zone is
maintained below about 1,250 K (1,790°F). NOX emissions of about 7 ng/J (0.016 lb/106 Btu) have
been reported with a prototype furnace using the surface combustor (Reference 5-13). The surface
combustor is similar to some larger commercial systems which employ surface combustors.
Another manufacturer's design incorporates a perforated burner. Natural gas mixes with air
through the perforations and combustion occurs at the burner surface. Heat is transferred to a
glycol solution in small tubes imbedded in a fin arrangement surrounding the burner. The
5-8
-------
TABLE 5-4. PERFORMANCE SUMMARY OF LOW-NOV CONTROL EQUIPMENT FOR
NATURAL GAS-FIRED RESIDENTIALxHEATERSa
CONTROL
Conventional
Units
Radiant
Screens
Secondary Air
Baffles
Surface Combus-
tion Burner
Perforated
Burner
Modulating
Furnace
Pulse
Combustor
Catalytic
Combustor
AVERAGE
OPERATING
EXCESS AIR
(percent)
40-120
40-120
60-80
10
HA
NA
NA
HA
CYCLIC POLLUTANT EMISSIONS
ng/J HEAT INPUT
N0xb
28-45
15-18
22
7.5
7.7
25
10-20
< 5
CO
8.6-25
6.4
14
5.5-9.6
26
NA
NA
NA
UHCC
3.3-33
NA
NA
NA
, Nft
NA
NA
NA
STEADY STATE
EFFICIENCY
(percent)
70
75
NA
NA
85
75
95
90
CYCLE
EFFICIENCY
(percent)
60-65
70
NA
NA
80
70
.95 .
85
1978
INSTALLED
CONTROL
COST
d
NA
NA
$100-$200
$100-$300
over
conventional
furnace
$50-$250
over con-
ventional
furnace
$300-$600
$100-$250
COMMENTS
Costs include installation.
Emissions of CO and HC can
increase significantly if
screen is not placed properly
or deforms.
Requires careful installation.
Suited for single port upshot
burners.
IF
Not commercially available.
Still under development.
Commercially available design.
Spark ignited thus requires no
pilot.
Furnace 1s essentially derated.
It requires longer operation
to deliver a given heat load.
Currently being Investigated
by AGAL.
Still at the R&D stage.
Reference 5-1.
b Sum of NO + N02 reported as N02.
c Unburned hydrocarbons calculated as methane (CH4).
Typical costs of uncontrolled unit $500-$800.
NA = not available.
-------
TABLE 5
-5. PERFORMANCE SUHHARY OF LOW-NOr COHTROL EQUIPMENT FOR
DISTILLATE OIL-FIRED RESIDENTIAL HEATERS3
CONTROL
Conventional
Units
Flame Reten-
tion Burner
Head
Controlled
Mixing
Burner Head
Integrated
Furnace Sys-
tem
"Blue flaw"
Burner/Furnace
System
Internal
Redrculatlon
AVERAGE
OPERATING
EXCESS AIR
(percent)
50-85
20-40
10-50
fc
20-30
20
10-15
"
CYCLIC POLLUTANT EMISSIONS
ng/0 HEAT INPUT
"*
37-85
26-88
34
19
10
10-25
CO
15-30
11-22
13
20
4.5-7.5
<30
UHCC
3.0-9.0
0.2-1.8
0.7-1.0
1.2
1.5-2.5
NA
Smoke
ttaufcer
3.2
2.0
<1.0
<1.0
zero
<1.0
Participate
7.6-30
NA
NA
NA
NA
NA
STEADY STATE
EFFICIENCY
(percent)
75
80-83
also depends
on heat
exchanger
. 80
also depends
on heat
exchanger
84
84
85
CYCLE
EFFICIENCY
(percent)
65-70
NA
NA
74
74
NA
1978
INSTALLED
COHTROL
COST
d
$52e
$43e
$250 over
conven-
tional
furnace
$100 over
conven-
tional
furnace
NA
COMMENTS
Range In NO emissions Is
for residential systems
not equipped with flame
retention burners.
Emissions for other
pollutants are averages.
If a new burner 1s needed
as well as a burner head,
the total cost would be
$385.
Cost of mass produced
burner head only about
$1.50. Combustible
emissions are relatively
low because hot firebox
was used.
Uses optimized burner
head. For new furnace
only. Combustible
"emissions are higher than
with burner head because
of quenching in air
cooled firebox.
New installation only.
Furnace is commercially
available.
Both for retrofit or new
Installations. Not yet
commercially available in
U.S.
(X
I
o
Reference 5-1.
Sum of NO and N02 reported as NOj.
c Unburned hydrocarbons calculated as methane.
Typical costs of uncontrolled unit $650-$!,000.
e Original costs reported for years other than 1978 were-corrected for Inflation using
Gross National Product (GNP) Implicit price inflators (Reference 1-9).
NA - Not available
-------
glycol solution then transfers the heat to room air. Ignition occurs by means of a spark rather
than by means of a standing pilot; therefore, seasonal fuel consumption is reduced compared with
pilot ignition furnaces. Reported NO emissions averaged 7.7 ng/J (0.018 lb/10 Btu)
(Reference 5-10).
The modulating furnace is commercially available and differs from conventional units in that
the firing rate responds to heating load demand instead of cycling on and off. The reported
emission rate for this furnace design is 25 ng/J (0.058 lb/10 Btu). The American Gas Association
Laboratories attributes the lower emissions to the fact that the furnace is .essentially derated,
yielding a cooler flame and thus low NO emissions. However, the Gas Appliance Manufacturer
Association attributes the decreased emission rate to the single-port inshot burner that was used
(Reference 5-10).
Pulse combustion involves combustion in a chamber fitted at one end with flapper valves and
at the other end with an open exhaust pipe. Fuel and air entering through the flapper valves are
ignited with a spark. The pressure resulting from combustion forces the flapper valves closed and
forces the product gases out the exhaust pipe. The valves open again as the exhaust gases leaving
the chamber create a negative pressure. Preliminary measurements of NO emissions from pulse
X
combustion of natural gas are about 20 ng/J (0.047 lb/10 Btu) (Reference 5-14). There are no
marketed pulse combustion systems in use today.
Catalytic residential combustors, although not available commercially, offer the potential
for very low NO emissions with good combustion efficiency in residential furnaces. The catalyst
promotes combustion at low temperatures so that thermal NO formation is significantly reduced.
Catalytic combustors require large amounts of combustion air. The use of a condensing residential
system in which the latent heat of vaporization is recovered is being considered for minimizing heat
losses due to high excess air levels (Reference 5-2).
Oil-Fired Equipment
Essentially all new residential and commercial sized oil-fired furnaces and boilers are
equipped with flame retention burners. Flame retention devices are generally desireable for all
conventional high pressure atomizing gun burners because they allow for operation at low excess air
levels and stay tuned longer. However, laboratory experiments have shown that most flame retention
burners increase NOV emissions (References 5-15, 5-3). In a test of ten commercially available high
A *
pressure atomizing oil burners, only one burner was found to reduce NO emissions while .
5-11
-------
also reducing smoke emissions. NO emissions were reduced from 37 to 26 ng/J (0.087 to
0.061 lb/106 Btu) (Reference 5-15).
An advanced residential warm air oil furnace has been developed in an EPA-funded program
(References 5-16 and 5-17). The integrated furnace system is said to increase the fuel utilization
efficiency by up to 10 percent. In addition, a 65 percent reduction in NO emission levels was
realized. The advanced oil furnace design consists of an optimized oil burner and firebox
combination. The system has completed a 500 hour laboratory performance test. The tests evaluated
the effects of combustion air swirl angle, nozzle spray angle, and axial injector placement on NOX
emissions levels for various oil flowrates and overall excess air combinations. The optimum burner
was a nonretention gun-type with six swirl vanes set at a 26 degree angle. The firebox design
selected was a cylindrical fin-cooled firebox. The optimum burner/firebox combination yielded
emissions of 0.6 g NO/kg of fuel (1.2 Ib/ton) at 10 percent excess air compared to 2-3 g/kg
(4-6 Ib/ton) for the baseline commercial burners.
In a study related to the development of the integrated furnace system, a controlled mixing
burner head for retrofit application to residential oil heating equipment was developed
(Reference 5-16). Laboratory testing indicated that the burner design is feasible for
commercialization and can be retrofitted on existing residential space heating equipment. The
burners operated successfully with long life potential. Retrofit of the burner into standard
o1l-f1red furnaces would result in an estimated 20 percent decrease in NOV emissions with an
A
accompanying increase in thermal efficiency of up to 5 percent (Reference 5-17).
Another advanced burner/furnace design consists of a "blue flame" oil burner integrated with
the firebox of a warm air furnace package (Reference 5-18). Two sizes are currently available:
0.63 cm oil/sec (0.6 gph), and 0.79 cm3 oil/sec (0.75 gph). The efficiency of the burner is
reported to be about 84 percent and the NOV emission level is about 20 ppm. This is a significant
X
Improvement over conventional systems for which typical efficiencies are 75-80 percent, and NOV
X
emissions range from 70 to 90 ppm. In the blue flame system combustion air and gases are
redrculated throughout the combustion chamber; the recirculation zone is designed such that
blue-flame burnout of CO and organics results. These systems are available as a single unit
(burner/furnace combinations) for new installations. Retrofits to existing burners are not
practical since the blue flame burner must be matched to the firebox geometry and heat transfer
characteristics (Reference 5-19). The blue flame furnace system is the only commercially available
low NOX system in the U.S. (Reference 5-1).
5-12
-------
Another oil burner recirculates combustion gases in a manner similar to the blue flame*
system. However, this burner recirculates combustion gases Internally; whereas, the blue flame
system uses external gas recirculatlon and requires an air tight combustion chamber to prevent air
leaks. This Internal recirculatlon feature imiy permit retrofit Installation on oil-fired furnaces.
Estimates of NOX emissions from this burner are 10-25 ng/J (0.059 lb/10 Btu) (Reference 5-2).
Coal and Wood-Fired Equipment
NO emission control techniques for residential coal- and wood-fired heaters and boilers have
A
not been widely investigated, primarily because of the declining use of this type of equipment in
the past. Also, small coal-fired equipment is; not amenable to extensive modification for
controlling NO emissions. Excess air and ov«rf1re air Injection in some units are the only
feasible control alternatives which have some impact on overall NO emissions. However, excess air
A
reduction in residential coal-fired equipment is very limited due to Increases in carbonaceous
emissions. Overfire air injection, only available with larger commercial stokers, 1s only
moderately effective in reducing NOV (Reference 5-1).
A
Commercial Equipment
Application of control technology to commercial heating equipment has been very limited.
However,'the potential for applying some of the control techniques applied to residential systems
exists. Compared to residential gas-fired equipment, a greater percentage of commercial warm air
heaters or duct heaters utilize power burners instead of naturally aspirated burners. Power burners
generally have more flexibility for excess air control while maintaining low CO and VOC emissions
(Reference 5-1). Furthermore, theoretical considerations indicate that the flame quenching and
surface combustor concepts of gas-fired residential burners could be Implemented for commercial
systems. Application of control techniques similar to those for residential oil-burners may also be
possible for commercial oil-fired furnaces. The EPA controlled mixing burner head design was found
to minimize NOX emissions from burners with oil flow capacity up to 13 ml/s (12 gph)
(Reference 5-1).
Although NOX control techniques for small firetube boilers have not been widely investigated,
the similarity in equipment design between small firetube boilers and larger Industrial firetube
boilers may allow application of similar combustion modification control techniques. NOX control
5-13
-------
techniques investigated for industrial stoker coal-fired boilers could also be potentially
applicable to commercial size stokers (Reference 5-1).
5.1.3 Costs »
Table 5-6 summarizes estimated cost data for the most effective NOV control alternatives for
A
residential heating systems. The use of advanced low-NO burner-furnace units for new sales appears
to be the most attractive option for NOX control in space heating equipment. The perforated
burner is presently considered the best available control technology for gas-fired residential and
commercial heaters (Reference 5-1).
The blue flame furnace and the integrated furnace system are options for new oil-fired
furnace installations. The blue flame unit has been commercially available since 1974 and has been
widely tested in field installations. The integrated furnace system (Reference 5-16) is undergoing
field demonstration preparatory to certification and potential commercialization. With proper
maintenance, both units offer a NOV reduction potential of 50 percent or greater compared to
A
conventional units. Fuel savings of 5 percent or more, relative to standard units, are achievable
with these units. Use of these low-NOv units in new houses and for replacement of obsolete
A
conventional units in existing installations would yield a nationwide decrease in residential NO
emissions which would more than offset the potential emissions increase due to population growth for
several decades.
For long term application to NOX control in new residential units, there is the additional
possibility of utilizing the alternate design concept of catalytic combustion. This concept,
discussed in Section 3.5, offers the potential for extremely low levels of NO (1-10 ppm) when
firing natural gas or distillate oils. Catalytic combustion is still in the exploratory stage of
development and no reliable cost estimates are available for residential heating systems.
As indicated in Table 5-6, retrofit of the controlled mixing burner head (EPA/Rocketdyne) for
existing residential oil-fired furnaces primises to be the most cost-effective approach for
achieving lower NO emission levels. However, these burners are not currently commercially
available (Reference 5-1).
Furnace tuning and, if required, burner head replacement (conventional burner head) are
strongly recommended for reduction of carbon monoxide and smoke and for improving unit efficiency.
The impact on NOX is negligible, however. Furnace tuning (cleaning, leak detection, sealing and
burner adjustment) costs a minimum of $40 for the average residential unit. Burner head retrofit
5-14
-------
TABlf 5-6. COST 1HPACT OF HO, COHTitOl ALTERNATIVES*
V
«*l*
(ft
CONTROL
Perforated 8um«r
Htidtriattng Furnace
Surface cw&ugtlon
Burner
pulse Combustion
Rumen *
Catalytic
Couinistion
fl«ri>er
flame Retention
Burner Head
Claud Retention
Burner
Controlled Mixing
Burrurf (lead
Integrated Furnaca
System
Blue Plane
FUEL
«»tural Gas
Natural Gas
Natural Gas
Natural GM
datura! Cas
Distillate Oil
distillate 01 1
Distillate OH
IHstltlate Oil
Distillate Oil
ACHIEVABLE HO LEVEL
ng/J USEfW. HEAf
12
(339 ng/w3 fuel)
i 36
(920 ng/«3 fuel)
12
(3WI tig/a? g«S
21
(603 ng/»3 UBS)
Estimate S
(153 ng/M3 g*$)
50
(l.B g/Sg fuel)
SO
(1.8 k/kg fualj
45
(l.« 9/H fuel)
29
(0.7 B/kg/fuel)
SO
(0.7 a/fca fual)
1978
IHCflEtttHTM.
IHtfESnUNT CAST
$100-J300 ovur cost of
conventional furnace
$BO-$Z5U over cost of
conventional funitca
$100-$200 itver cost of
conventional furnace/
heater
$30fl-$600 aver cost of
conventional furnace/
heater
fl6Q-$2SO over vost of
conventional furnace/
heater
«2 -- retrofit
including instillation
|385 retrofit of
reduced capacity burner
t« - retrofit
including installation
9250 over cost of
conventional furnace
$100 over cost of
conventional furnace
COST EFFECTIVENESS
l/Bfl/J1'
(Ib/UT Btu)
1.1 - 6.2
1.4 - /.O
1.7 - 3.4
6.1 - li.Z
2.3 - 3.9
2.6
12.8
1.3
4.2
1.7
ttHf&tCK PERIOD
BASED ON ANNUAL
FUEL BILL OF $500
1-3 years
1 ' 3.9 years
3.S - 8.U
1.7 - 3.5
1.4 - 2.3
Less than 1 .year
3.S years
Less titan 1 yettr
2.S years
1 year
OEVELOFff.HI
STATUS
Counerclally
available
CoflMrciolly
«vi liable
Hot coranercially
available
Not Rocwercialty
available
Hot coHiwrclally
available
Cowiercltflly
available
Cwnnerclally
available
Hot cowerc tally
available
Mot commercially
available
Ciwffierc tally
availaftle
(Reference 5-1}.
b Based oa uncontrolled Mix emissions of 'IS ng/J host output for natural ass-fired heaters and flO ng/.) heat output
for distillate oil-fired heaters. tost-e?fectlv«n«ss Is based on the differential investment cost of the control.
c Based on installation of » condensing system whtsr» se&swta! efficiencies can be as high as 55 percent.
Only one fla«e retention burner tested lowered fiOx emissions.
-------
replacement costs an additional $25 less installation. These control measures are usually cost
effective in view of the fuel savings and increased safety derived from the maintenance.
5.1.4 Energy and Environmental Impact
5.1.4.1 Energy Impact
Both of the NOX emission reduction techniques (tuning, equipment) result in improved system
efficiencies and, consequently, reduced fuel consumption. The exact amount of improvement varies
widely depending on the type of equipment. The most promising method, unit replacement, appears to
offer in excess of 5 percent fuel savings. On a national basis, this represents a potential savings
of 0.6 percent of annual fuel consumption if all space heating equipment were replaced with new
designs.
5.1.4.2 Environmental Impact
The effect of lower excess air on CO, VOC, and parti oil ate emissions was discussed previously
and is illustrated in Figure 5-1. By constraining incremental emissions during control development,
however, it has been possible to achieve low-NO combustion conditions without increasing adverse
emissions of other species (Reference 5-17). Table 5-7 shows a comparison of the emissions from
typical uncontrolled units and from a prototype unit with an optimized burner/firebox. Incremental
emissions were held constant or reduced when using the low-NO furnace. Table 5-7 also shows
incremental emissions with a commercially available oil emulsifier burner. Again, low-NO operation
was achieved with no adverse effects on incremental emissions (Reference 5-20).
Over 90 percent of residential and commercial warm air furnaces fire either.natural gas or
distillate oil. Emissions of sulfates and trace metals from these units are thus of minor concern
compared to coal-fired boilers. About 3 percent of U.S. warm air furnaces still fire coal. For
these furnaces, sulfates, trace metals and especially POM's could cause severe localized environ-
mental problems. However, except for fuel switching, it is doubtful that NOV controls will be
X
developed and implemented for these sources, and they will not be considered further here.
An additional factor in evaluating incremental emissions for warm air furnaces is the cyclic
nature of operation. Warm air furnaces typically undergo two to five on/off cycles per hour.
Studies of emissions without NO controls show that the starting and stopping transients have a
strong, sometimes dominant, effect on total emissions-, of CO, HC and pSr-ticulate matter (smoke)
5-16
-------
TABLE 5-7. EFFECT OF LOW-NO OPERATION ON INCREMENTAL EMISSIONS
AND SYSTEM PERFORMANCE FOR RESIDENTIAL WARM AIR FURNACES
,
Typical uncontrolled
field units
(References 5-2,5-3)
Optimum low-NO unit
J\
(Reference 5-17)
Water/distillate oil
emulsifier burner:
(Reference 5-20)
Excess
Air
90%
15%
32%
Thermal
Efficiency
(Steady-State)
70*
802
802
NO
g/kg fuel
1.1 - 2.7
0.6
0.85
CO
g/kg fuel
1.05
1.0
0.3
HC
g/kg fuel
0.1
0.1
iM«l
Smoke
Bacharach
3.2
1
-1
01
I
-------
(References 5-2 and 5-3). The effect of NOV controls on transient emissions has-not been widely
A
studied. Incremental steady-state emissions must eventually be weighed against the transient
emissions for this significance to be shown.
Comparative data on warm air furnace POM emissions under low-NOx operation are apparently
nonexistent. Data on both transient and steady operation with and without NOX controls are needed
to form a general conclusion on the total incremental impact of NOX controls. Additionally, it
should be emphasized that the incremental emissions data shown in Table 5-7 are for well-maintained
laboratory operation. Data are needed on long-term field operation with NO controls.
5.2 INCINERATION AND OPEN BURNING
5.2.1 Municipal and Industrial Incineration
According to a Public Health Service survey conducted in 1968 (Reference 5-21), an average of
2.5 kg (5.5 pounds) of refuse and garbage is collected per capita per day in the United States. An
additional 2 kg (4.5 pounds) per capita per day are generated by incineration of industrial wastes,
wastes burned in commercial and apartment house incinerators, and backyard burning. The total per
capita waste generation rate is conservatively estimated at about 4.R kg (10 pounds) per day
(Reference 5-21).
Incineration is economically advantageous only if land is unavailable for sanitary landfill.
Incineration requires a large capital investment, and operating costs are higher than for sanitary
landfill.
The most common types of incinerators use a refractory-lined chamber with a grate upon which
refuse is burned. Combustion products are formed by contact between underflre air and waste on the
grates in the primary chamber. Additional air is admitted above the burning waste to promote
burnout of the primary combustion products.
Incinerators are used in a variety of applications. The main ones are municipal and
industrial solid waste management. Municipal incinerators consist of multiple chamber units that
have capacities ranging from 23 kg (50 pounds) to 1,800 kg (4,000 pounds).
5-18
-------
5.2.1.1 Emissions
Nationwide NO emissions from incineration in 1974 amounted to 39 Gg per year (43,400 tons
per year) which is 0.3 percent of the total NOX emissions from stationary sources. Fifty-five
percent of these emissions result from industrial incineration with the remainder due to municipal
Incineration. A number of other multimedia effluents from incineration may be of greater concern
than NOX. These include metallic compounds in the particulate flyash and hopper ash and chlorinated
\ '
organic and inorganic gaseous compounds. Incinerator effluent rates .are strongly dependent on the
composition of the solid waste, the incinerator design and specific operating variables such as
excess air and firing rate. The effluent rates can vary considerably from day to day because of
variations in refuse composition. An average emission factor for incineration of 1.5 g N02/ kg
refuse (3 Ib/ton) was reported by Niessen (Reference 5-22). AP-42 (Reference 5-23) specifies the
same value for multichamber industrial and municipal incinerators. For single chamber industrial
incinerators, a lower factor.of 1 g NOg/ kg refuse (2 Ib/ton) is.specified.
Stenberg, et al., conducted field tests to study the effects of excess combustion air on NO
'" --" X
*
emissions from municipal incinerators (Reference 5-24). The nitrogen oxide emissions ranged from
0.7 g/kg (1.4 Ib/ton) to 1.65 g/kg (3.3 Ib/ton) of refuse charged for a 45.3 Mg (50 ton) per day
batch-feed incinerator and a 227 Mg (250 ton) per day continuous-feed incinerator. As shown in
Figure 5-2, NOX emissions increase with increasing amounts of excess air. The amount of underfire
air also has a significant effect on NOX production and is shown in Figure 5-3.
In general, nitrogen oxide emissions from incineration are not a primary source of air pollu-
tion; however, particulate emissions are significant. It is for this reason that incinerator air
pollution control equipment is adopted to the removal of particulate matter rather than NO .
A
Activity in pollution abatement for incinerators to date has focused on particulate control rather
than NOX.
5.2.1.2 Control Techniques
The use of waste disposal methods other than combustion may be the most likely means for
reducing NO emissions, since the methods normally used for control of other emissions from inciner-
ation, such as particulate matter, organics, and carbon monoxide, tend to increase emissions of
5-19
-------
0.6
0.5
0.4
I
ec*
o
x
01
LU
O
cc
11J
cu
g
o
111
si! 0.3
cc
a
x-
a
st
0.2
0.1
O FURNACE - BEFORE SCRUBBER
NOX = 0.081 + 0.00144 (PERCENT EXCESS)
O STACK - AFTER SCRUBBER
NOX = 0.093 + 0.00156 (PERCENT EXCESS)
100
200
300
EXCESS AIR, percent
Figure 5-2. Effect of excess air on NOX emissions from a
45.3 Mg (50 ton) per day batch-feed incinerator
(Reference 5-24).
5-20
-------
NOX = 0.365 0.00183 (PERCENT UNDERF1RE AIR)
40 60
UNOERFIRE AIR, percent
Figure 5-3.
Effect of underfire air on NO emissions from a
227 Mg (250 ton) per day continuous feed incinerator
(Reference 5-24).
5-21
-------
NO. Other disposal methods Include dumping, sanitary landfill, composting, burial at sea, disposal
A
In sewers and hog feeding.
One of the first refuse disposal methods used was open dumping of refuse on land. This
method 1s obviously very Inexpensive, but extremely objectionable and offensive in and near popu-
lated areas.
Sanitary landfills nay be alternatives, to the extent that land usable for this purpose
Is available. Approximately 1233 m3 (1 acre-foot} of land 1s required per 1000 persons per year
of operation for a waste production of 2 kg (4.5 pounds) per day per capita (Reference 5-25). In
addition, cover material approximating 20 percent by volume of the compacted waste is required;
the availability of cover material may limit the use of sanitary landfill.
5.2.1.3 Costs
i
At present, gaseous emission controls are not applied to Incinerators. As described earlier,
only particulate emission controls are employed. Reference 5-26 presents estimated construction
costs in 1966 and operating costs for particulate pollution control.
5.2.2 OpenBurning
Open burning includes forest wildfires, prescribed burning, coal refuse fires, agricultural
burning, and structural fires. Open burning for solid waste management is usually done in large
druas or boskets, in large-scale open dumps or pits and on open fields. Commonly, municipal waste,
landscape refuse, agricultural field refuse, wood refuse, and bulky Industrial refuse are disposed
of by open burning.
5.2.2.1 Emissions
Emissions from open burning are affected by many variables including wind, ambient tempera-
ture, composition and moisture content of the'debris burned, and compactness of the pile. Nitrogen
oxides emissions depend mainly upon the nitrogen content of the refuse. Generally, due to the low
temperatures associated with open burning, nitrogen oxides emissions are low.
Annual emissions fronr open burning vary from year to year, and the data for the various
sources are not entirely consistent. Table 5-a shows the estimated NOX emissions from open burning
5-22
-------
sources for 1971 as reported in Reference 5-27. More recent estimates from the 1976 NEDS data file
+
and Reference 5-28 are also given in Table 5-8. Increasing awareness of air pollution problems
has contributed to a general decline in the quantity burned (and thus the emissions) from tnose
categories-which can be controlled. For example, despite the continuing growth in crop harvest,
NOx emissions from agricultural open burning has declined from an estimated 29 So (32,000 tons) in
1969 to 13 Sg (14,300 tons) for 1973 {Reference 5-23).
TABLE 5-8. ANNUAL EMISSIQHS OF NITROGEN OXIDES FROM OPEH BURNING
Source
Solid Waste Disposal
Forest Wildfires
Prescribed Burning
Agricultural Surning
Coal Refuse Fires
Structural Fires
NOX Emissions
1971, .
Reference 5-27
So.
ISO
138
19
29
31
6
1C3 Tans
16S
152
21
32
34
7
1S7S NEDS
Ss
55
48
30
13a
S3
5
10s Tons
105
53
33
14a
58
6
a1973 estimate from Reference 5-28.
5.2.2.2 Control Techniques
Solid Waste Disposal
* '
Frora the standpoint of air pollution, sanitary landfills are alternatives ta open burn-
Ing. In addition i» the land necessary for sanitary landfill. cover material approximating 20
percent by volume of tfte compact waste is required. The availability of caver material nay limit
the use of the sanitary landfill method.
Unusual local community factors may lead to unique approaches to the landfill site problem.
For example. Reference 5-29 reports that in a pilot project the refuse Is shredded and baled for
loading on rail cars for shipment to abandoned strip urine landfill sites.
Other noncoabustion alternatives may have application fn some localities. Composting is
now being tested on a practical scale (Reference 5-30). Hog feeding has been used for disposal
of cartage. Dumping at sea has been practiced by some seaesast cities, but is now extensively regulated.
S-23
-------
Elsewhere, refuse has been ground and compressed into bales, which are then wrapped in chicken
wire and coated with asphalt. The high-density bales sink to the bottom in the deeper ocean areas
and remain intact. The practice of grinding garbage in kitchen units and flushing it down the sewer
has been increasing. This in turn increases the load of sewage disposal plants and the amount of
sewage sludge (Reference 5-31).
Forest Wildfires
12 '^
In the United States, forests comprise approximately 3.2 x 10 m'' (786 million acres), or
34.4 percent, of the land area. Seasonal forest fires are quite prevalent in dry western regions.
Considerable activity has been and is being directed toward reducing the frequency of occurrence
and the severity of these fires. These activities include publishing and advertising information
on fire prevention and control, surveillance of forest areas where fires are likely to occur, and
various firefighting and control activities. Additionally, prescribed burning is being used to
reduce the loading of combustible underbrush and thereby decrease the fire hazard and potential
fire spread rate.
The U.S. Forest Service estimated that 2.06 x TO10 m2 (5.11 million acres) of land were burned
1n 1976 (the World Almanac, 1978). A similar estimate for 1971 (Reference 5-27) was 1.73 x TO10 m2
(4.28 minion acres) burned, producing 138 Gg (152,000 tons) of nitric oxides emissions. Emissions
from forest fires are dependent on the local combustion intensity, the overall scale of the fire,
and, to some extent, the nitrogen content of the fuel. These in turn are related to the topography
of the forest, the composition and dryness of the underbrush, the local meteorological conditions,
and the elapsed time since a previous fire. The topography of the forest, the composition and dry-
ness of the underbrush, the elapsed time since a previous fire and the meteorological condition are
all Interrelated and dictate the burn rate and spread, Intensity of the burn, and the size of the burn.
»
Prescribed Burning
Prescribed burning is the use of controlled fires in forests and on ranges to reduce the pos-
sibility of wildfire and for other land management goals. Four classes of open burning operations
are traditionally practiced by the Forest Service (Reference 5-32):
Slash disposal resulting from forest harvesting operations
Forest management operations for forest floor fuel reduction, seedbed preparation, pest
control, forest thinning and undergrowth control
» Public works construction operations to clear reservoir and dam-sites, utility and high-
way rights-of-way and building and structure site areas
5-24
-------
« Public works maintenance operations for the disposal of reservoir driftwood and of
rights of way and storm damage debris
In addition, controlled burning is used to reduce unwanted quantities of waste and to improve land
utilization.
Because collection and incineration of these materials would tend to increase NO emissions,
the only current way to control emissions is to avoid combustion. In the future it may be possible
to develop incineration processes that can control NO and other emissions such as particulate
matter, organics, odorous compounds, and carbon monoxide; or it may be possible to develop equipment
that cart burn these materials as substitutes for fossil fuels.
Other alternatives to incineration are abandonment or burying at the site, transport to and
disposal in remote areas, and utilization. Abandonment or burning at the site is practical in cases
where no other harmful effects will ensue. Abandoned or buried vegetation can have harmful effects
up6n plant life by hosting harmful insects or organisms, for example. Agricultural agencies such as
the U.S. Department of Agriculture, or state and local agencies should be consulted before these
techniques are employed.
Agricultural Burning
Agricultural burning includes the burning of residues of field crops, row crops, and fruit
and nut copes for at least one of the following reasons (Reference 5-28):
r Removal and disposal of residue at low cost
Preparation of farmlands for cultivation
« Clearing to facilitate harvest
Control of disease, weeds, insects, or rodents
Mitigation of the environmental impact of agricultural open burning is possible by proper
fire and fuel management (for example, single-line backfiring), meteorologically scheduled burning
to optimize dispersion, or by the substitution of other alternatives, such as mobile incineration,
incorporation into the soil, and mechanical removal. Care must be exercised in the choice of alter-
nate methods of disposal since a change In method may have significant adverse effects. For ex-
ample, in situ burning can provide thermal treatment to the soil which raises the production yield
substantially, incorporation of the residue into the soil may restrict rapid replanting, and residue
decomposition may deplete the soil nitrogen.
Coal Refuse Fires
An estimated 53 Gg (58,000 tons) of NOX is emitted each year from burning coal refuse.
Extinguishing and preventing these fires are the techniques used for eliminating these emissions.
5-25
-------
These techniques involve cooling and repiling the refuse, sealing refuse with impervious material, in-
jecting slurries of noncombustibles into the refuse, minimizing the quantity of combustibles in refuse,
and preventing ignition of refuse. The NO emissions from coal refuse fires are highly dependent on
the nitrogen content of the coal.
Structural Fires
There were almost one million buildings attacked by fire during 1971 with losses estimated at
$2.21 billion (Reference 5-27). An estimated 6.3 Gg (7,000 tons) of NOX were emitted during 1971.
Prevention is the best control technique to reduce these emissions. Use of fireproof construction,
proper handling, storage, and packaging"of flammable materials, and publishing and advertising infor-
mation on fire prevention are some of the techniques used to prevent structural fires.
Fire control techniques include the various methods for promptly extinguishing fires: use of
sprinkler, foam, and inert gas systems; provision of adequate firefighting facilities and personnel;
provision of adequate alarm systems. Information on these and other techniques for prevention and
control are available from agencies such as local fire departments, National Fire Protection Asso-
ciation, National Safety Council, and Insurance companies.
5.3 INDUSTRIAL PROCESS HEATING
Fossil fuel derived heat for industrial processes is supplied in two1 ways: (1) by direct
contact of the raw process material to flames or combustion products in furnaces or specially-
designed vessels, and (2) by heat transfer media (e.g., steam, glycol or hot water) from boilers and
I.C. engines. NOX emissions and control techniques for. the latter equipment types have been de-
scribed in previous sections of this document. The former equipment types are described 1n the
present section. Industries covered include petroleum and natural gas, metallurgical, glass, cement,
and coal preparation plants. Much of this section is taken directly from a recent study of indus-
trial process heating performed by the Institute of Gas Technology (Reference 5-33).
There is currently very little application of NOX control to industrial process heating equip-
ment. Consequently there are very few data on MOX control costs or energy and environmental impact,
and separate sections for these topics will not be included. EPA's Industrial Environmental Research
Laboratory (RTP) is sponsoring a field test program to identify the potential for NOX control in a
"diversity of process furnaces, ovens, kilns, and heaters. Partial results from that study are given
in Reference 5-34 and are discussed, as appropriate, in the following subsections. The complete
results of that program (scheduled for 1978) will provide a broad data base on which to evaluate
alternate control options.
5-26
-------
5.3.1 Petroleum, Chemicals, and Natural Gas;
Petroleum refining is the process of converting crude oil into salable products. This con-
version into salable products is accomplished in various operations which require the feedstock to
be at elevated temperatures. Typically, feedstocks are raised to the required temperature in a
fired heater. Many chemical manufacturing plants, natural gas plants, and pipeline stations have
fired heaters using the same general design principles. Methods used to control NOX emissions from
fired heaters used in these industry groups are similar and therefore will be discussed together.
Another source of NO emissions at petroleum refineries that will be discussed in this section is
catalytic crackers and CO boilers.
5.3.1.1 Fired Heaters \
5.3.1.1.1 .Process Description
Petroleum Refining and Chemical Manufacturing - Process emissions of nitrogen oxides from the
petroleum refining and chemical, manufacturing industries are produced primarily by fired heaters.
Fired heaters transfer heat, that is liberated by the' cumbustion of fossil fuels, to fluid contained
in tubular coils. Industrial processes usually use fired heaters when fluid temperature require-
ments are above 204°C (400°F). The fluids include any gas or liquid, with the exception of liquid
water when it is used for the generation of hot water or steam. Industry also refers to fired
heaters as process heaters, process furnaces, and direct fired heaters.
Typical applications of fired heaters include heating of oil or other heat transfer fluid,
steam superheating, distillation, thermal cracking, coking, pyrolysis, and reforming. Although
there are some applications of fired heaters in other industries, the dominant use is in the
petroleum refining and chemical manufacturing industries.
Petroleum refineries (SIC 2911) process crude oil to form products like transportation fuels,
heating fuels, lubricating oils, and chemical feedstocks. In 1981, there were 303 refineries
9 6
operating in the U.S. with a capacity to process 2.93 x 10 liters/d (18.45 x 10 barrels/day) of
crude oil (Reference 5-35). These refineries range in size from 6.36 x 10* liters/day (400 b/d) to
8
1.02 x 10 . liters/day (640,000 b/d)(Reference 5-35). Forty-one states had at least one refinery.
Approximately 56 percent of the industry capacity is located in Texas, California, and Louisiana.
5-27
-------
TABLE 5-9. MAJOR REFINERY PROCESSES REQUIRING A FIRED HEATER
Process
Distillation
Atmospheric
Vacuum
Themal Processes
Theresa! Cracking
Cokl ng
₯i streaking
Catalytic Cracking
Fluldlred
Catalytic
Cracking
Catalytic
Hydrocracfcing
Hydroprocessing
Hydrodesulfur-
Izatlon
Hydrotr eating
Hydroconverslon
Alkylation
Catalytic
Reforming
Process Description
Separate light hydrocarbons
from crude in a distillation
column under atmospheric
conditions.
Separates heavy gas oils
from atmospheric distillation
bottoms under vacuum.
Thermal decomposition of
large molecules into lighter,
more valuable products.
Cracking reactions allowed to
go to completion. Lighter
products and coke produced.
Mild cracking of residuals
to Improve their viscosity
and produce lighter gas oils.
Cracking of heavy petroleum
products. A catalyst is used
to aid the reaction.
Cracking heavy feedstocks
to produce lighter products
in the presence of hydrogen
and a catalyst.
Remove contaminating metals,
sulfur, and nitrogen from
the feedstock. Hydrogen is
added and reacted over a
catalyst.
Less severe than hydrodesulfur-
Izatlon. Removes metals,
nitrogen, and sulfur from
lighter feedstocks. Hydrogen
is added and reacted over a
catalyst.
Combination of two hydrocarbons
to produce a higher molecular
weight hydrocarbon. Heater
used on the fractionator.
Low octane napthas are
converted to high octane,
aromatic napthas. Feedstock
Is contacted with hydrogen
over a catalyst.
Process Heatd
Heaters Requirements
Used kj/liter (103 3tu/b) feed
Preheater, 511 (77)
Reboiler
Preheater, 418 (63)
Reboiler
Fired 4,648 (700)
Reactor
Preheater 1,460 (220)
Fired 1,328 (200)
Reactor
Preheater 385 (58)
Preheater 1,262 (190)
Preheater 465 (70)a
Preheater 239-498 (36-75)
Reboiler 1,992-7,304 (300-1, 100)*
Preheater 2,258 (340)
Feedstock Temperature
Outlet of Heater
°C (°F)
371 (700)
399-443 (750-830)
454-538 (850-1000)
482-524 (900-975)
454-510 (850-950)
e, f
316-474 (600-385)
204-454 (400-850)
199-427 (390-800)
316-427 (600-800)
-204 (-400)
454-538 (850-1000)
'heavy gas oils and middle distillates
b!1ght distillate
of total alfcylate
'erence 5-36
dRefi
Reference 5-39
Reference 5-40
5-28
-------
In general, each process in a refinery requires at least one fired heater. All refineries have
atmospheric and vacuum distillation units to separate the lighter products from crude oil. The
number and type of downstream processes depend on the type of crude processed and the products
desired. For example, a complex refinery that takes heavy crude oil and produces unleaded gasoline
and lighter oils will require processes such as cracking, hydrotreating, reforming, and alkylation.
A refinery of this type can require as many as 100 heaters (Reference 5-36). A small, simple
refinery (topping refinery) whose main operation is to separate the crude oil into its major
fractions may have 10 heaters. Table 5-9 presents the major refining processes along with a brief
description of each process, the types of heaters it requires, and its fired heat and temperature
requirements.
The heaters for the refinery processes are usually either preheaters, fired reactors, or
. reboilers. Hot oil furnaces can be used as preheaters or reboilers. Circulating heaters that heat
the heavy crude oils to decrease the viscosity and improve flow through pipes are also used. In
refinery fired heaters, the feedstock usually flows through the radiant tube coils only once. The
heaters require an even heat distribution to prevent coking wi-thin the tubes and to control feed-
stock temperature.
The total refining industry fired heater energy requirements can be estimated by assuming 75
percent of the total fuel consumed by the refining industry is used by fired heaters (Reference
5-37). For 1979, the fired heater energy requirements are estimated to be 2300 PJ/yr (2.18 x 1015
Btu/yr). This calculated value compares well with other reports which show a range of fired heater
energy requirements of 1910 to 2330 PJ/yr (1.131 - 2.21 x 10 Btu/yr)(Reference 5-37). In a report
published by the American Petroleum Institute, the estimated number of fired heaters used in the
petroleum refining industry in 1977 was 3240(Reference 5-38).
Chemical Manufacturing Industry - In the chemical manufacturing industry (SIC 28), tubular
fired heaters are used to drive endothermic reactions such as natural gas reforming and thermal
cracking. They are also used as preheaters to raise the feedstock temperature to a certain
temperature to control a reaction and as reboilers in some distillation processes. When a narrow
temperature range is required, hot oil furnaces are preferred because they allow better temperature
control.
5-29
-------
The chemical industry fired heater applications are similar to those used in the refining
industry. Fired heaters are used when the feedstock temperature requirements are above 177-204°C
(350-400°F). The predominant heater application is as fired reactors.
Ten chemicals have been identified in the chemical manufacturing industry as major users of
fired heaters (Reference 5-37). Table 5-10 presents these chemicals along with their manufacturing
process, temperature requirements, and the type of heaters they use. Of these ten chemicals, all
but one, ammonia, are organic chemicals. In 1980, about 38.5 Gg (85,000 x 103 Ibs) of these nine
organic chemicals were produced. Seven of these nine chemicals are among the top 50 produced
chemicals in the U.S. Also in 1980, 17.3 Gg (38,100 x 10 Ibs) of ammonia were produced. Ammonia
1s the second largest volume chemical in the chemical industry (Reference 5-41). Other smaller
volume organic chemicals may use tubular fired heaters. However, the ten chemicals presented
represent a large portion of the chemical industry fired heater energy requirements. In addition,
all of the basic heater applications expected to be used in the chemical industry are represented by
these ten chemicals. Therefore, it is believed that these ten chemicals are representative of the
fired heater population for the chemical industry. As in refineries, a major portion of the
production capacity is located in Texas and Louisiana.
An estimate of the total chemical industry heater population can be based on process heat
requirements. Thirty percent of the total fuels consumed by the chemical industry is required for
process heating (Reference 5-37). From this the process heat energy requirements are estimated as
approximately 779 PJ/yr (7.38 x 10 Btu/yr). However, this estimate of process heat requirements
may include other types of process heat besides tubular fired heaters such as dryers, kilns, and
roasters. Based on the available data, these other types of process heat account for approximately
14 percent of the total process heat energy requirements (Reference 5-42). Therefore, 86 percent of
the total process heat energy requirements is assumed to be met by fired heaters, or 670 PJ/yr (6.35
x 1014 Btu/yr).
Fired Heater Design - The design of fired heaters can vary depending on the heater application
and client preference. Industry uses fired heaters in a variety of applications. Table 5-11
describes the common applications of fired heaters used in the chemical manufacturing and refining
industries. The following design classifications distinguish the various heater designs from each
other:
5-30
-------
TABLE 5-10. TYPICAL FIRED HEATER APPLICATIONS IN THE CHEMICAL INDUSTRY
01
I
Chemical
Ethyl ene/.
Propyl ene
Butadiene
Ethanolf
Benzene
Dimethyl
Styrene6
b e
Ammonia '
Methanol9
Process
Naptha Cracking
Dehydrogenation
Ethyl ene Hydra -
tion
Extraction of
Reformate
Formed from
D-Xylene and
Methanol
Dehydrogenation
of Ethylbenzene
Natural Gas'
Ref ormi ng
Hydrocarbon
Reforming
Heater Type
Fired Reactor-
Thermal Cracking
Preheater and
Reboiler
Preheater
Reboiler
Preheater-llot
Oil Furnace
Fired Reactor-
Steam Hydro-
carbon Reformer
Fired Reactor-
Steam Hydro-
carbon Reformer
Fired Reactor-
Steam Hydro-
carbon Reformer
Feedstock
Temperature
°C (°F)
800-900 (1,472-1,652)
425 (800)
750 (1,382)
375 (707)
250-280 (482-536)
630-710 (1,166-1,310)
800 (1,472)
840-900 (1,544-1,652)
Heat Requirements8 >c
PJ/yr 10" Btu/yr
(205.16)
(3.58)
(1.76)
. (47.30)
(4.11)
(20.93)
(146.65)
(17.54)
194.46
3.39
1,67
44,83
3.90
19.84
139.0
16.63
b
All 1971 data, except ammonia
1975 data
Reference 5-42
^Reference 5-43
Reference 5-44
Reference 5-45
^Reference 5-46.
-------
TABLE 5-11. BASIC FIRED HEATER APPLICATIONS
en
i
to
CO
Type
Fired Reactor
Description
Outlet Feedstock
Temperature *C (*F)
Firebox Temperature6
Beboller He»ts liquid charge stock from
» fractionating column before
returning the liquid to the
column.
Fractionating- Heats charge stock before it
colum feed enters a distillation colum.
preheater
Reactor feed Raises the temperature of the
preheater feedstock to control a reaction
taking place in an adjacent
process.
Circulation Heats feedstocks to lower their
Heaterviscosity thus improving pumping
and flow through' pipes.
Prevents condensation when gases
are transferred.
Steam/hydrocarbon Heats the feedstock along with
reformer heater steam to drive an endotheraic
reaction. The reaction usually
takes place over a catalyst.
Pyrolysls heater Keats the feedstock for thermal
decomposition within the heater
tubes.
ileats a recirculating medium which
transfers heat to a feedstock.
Used when a narrow temperature range
is required.
204-288 (400-550)
371-443 (700-830)* '
538 (1000)U
399 (750)b
< 260 (<500)
< 260 (<500)
d
788-899 (1.450-1,650)
815-899 (l,500-l,650)d
177-399 (350-750)e
538-982 (1000-1800) Charge stock from ataospherlc or
vacuum distillation columns In a
refinery or a fractionating colum In
organic chemical Manufacturing.
871-1,093 (1600-2000) Heats feedstock before it enters
ataospheric or vacuum distillation
columns.
871-1,093 (1600-2000)
Refinery catalytic reforming
Ethylene hydration to fom ethanol.
538-982 (1000-1800) Crude oil heating in a refinery to
improve flows.
538-982 (1000-1800) Natural gas processing plants.
1,093-1,427 (2000-2600) Natural gas reforming to yield
hydrogen for ammonia synthesis.
1.093-1.427 (2000-2600) Thermal cracking In refineries to
produce lighter petroleum products
and cracking of natural gas liquids
and petroleum feedstocks in organic
chemical manufacturing to produce
olefins.
538-982 (1000-1800) Used as a reboiler In benzene
extraction from refinery feedstocks.
Reference 5-39
Reference 5-46
cReference 5-48
Reference 5-49
"Reference 5-50
-------
Radiant and convective tube coil orientation
Draft type
Use of preheated combustion air
Feedstock temperature
Fuel types used
Burner location
Burner type
Many designs of fired heaters are available. Some examples of the variety of designs available
are presented in Figure 5-4. All fired heaters have a radiant section and the majority have a
convection section. The radiant section is located within the firebox and contains the burners and
a single row of tubular coils. The primary heating of the feedstocks occurs within the radiant
.section. As the name implies, radiation is the primary method of heat transfer.
The tube coil in the rad.iant section consists of a number of tubes connected in series by 180
degree return bends (Reference 5-47). Each set of consecutive tubes is considered a "pass" or
parallel stream.' The inlet feedstock stream can make one pass or can be separated into a number of
passes. Tube diameter can vary, but an average diameter is around 10.2 cm (4 inches)(References 5-52,
5-53). The spacing between the tubes and the distance of the tube coil from the refractory walls
depends on the tube diameter. Spacing between the tubes usually ranges from 1.5 to 3 diameters
(Reference 5-53). Increasing the tube to wall clearance improves heat flux to the tube until the,
distance reaches about two tube diameters (Reference 5-54)., The walls are lined with an insulated
material such as insulated firebrick, castable refractory, or ceramic fiber. The insulation
protects the steel structure from overheating and flue gas corrosion. In addition, the insulation
minimizes heat loss and keeps the firebox at.a high temperature by reradiating heat to the tube
coils (Reference 5-47).
The convection section is located after the radiant section and also contains a set of tubes.
The convection section recovers the residual heat of the flue gas before it goes to the stack. The
temperature of the flue gas leaving the radiant section usually ranges from 816- 982°C (1,500-
1,800°F)(Reference 5-52). The first few rows of tubes, called shield tubes, are .subject to some
radiant heat transfer. In most heaters, the feedstock flows through the convection section to
preheat it before flowing to the radiant section (Reference 5-52). Some convection sections are
also used to generate steam. Convection sections can improve heater efficiency particularly if
5-33
-------
Canwedon coil
llint coil'
Convection coil-.
n«dtant ceil
Artor or wickwtyp*
V«ro'«aJ-tub«, linglarow, douWe-fir»d
Vertical-cylindrical with eranflow-
convection saction
. Sw*.ffr»d box Cabin Tw-xail box
Figure 5-4. Examples of fired heater designs (Reference 5-51).
5-34
-------
they have extended surface area. Many older heaters have bare convection section tubes and operate
at only 65 to 70 percent efficiency (References 5-37, 5-55). New extended surface tubes can improve
efficiency by 10 percent and reduce flue gas temperature by 149°C (300°F). Fins or studs are
usually used to extend the tube surface area. Newer heaters can operate at around 90 percent
efficiency. The lower limit of exit flue gas temperatures is around 149-177°C (300-350°F)
(Reference 5-55). Because many flue gases contain SO,, temperatures below this lower limit will
cause corrosion problems due to sulfuric add condensation. Typical exit temperatures are approxi-
mately 371°C (700°F).
Two basic draft types are available to supply combustion air and to remove flue gas. These are
natural-draft and mechanical-draft. Approximately 90 percent of all gas-fired heaters and 76
percent of all oil-fired heaters have natural-draft (Reference 5-51). Natural-draft heaters rely on
the natural stack effect to remove flue gas and induce the flow of. combustion air into the firebox.
The natural stack effect, maintains a negative pressure within the fired heater. The negative
pressure prevents leaking of flue gases and overheating of the heater structure.
In a mechanical-draft heater, a fan supplies the combustion air and removes flue gas. A
mechanical-draft heater can use either an induced- draft, forced-draft, or induced-draft/
forced-draft (balanced draft) design. An induced-draft heater uses an induced-draft fan located
above the convection section and before the atack to induce the flow of combustion air and remove
flue gas. The fan also maintains a negative pressure in the fired heater. A forced-draft heater
uses a forcedrdraft fan to supply combustion air under positive pressure. Although the combustion
air is under positive pressure, the firebox still remains under negative pressure. The negative
pressure occurs because the flue gas is removed, as in a natural-draft heater, by the stack effect.
A balanced draft heater uses a forced-draft fan to supply combustion air and an induced- draft fan
to maintain a negative pressure within the heater and remove flue gas.
Typical heaters have a negative pressure of 0.007 - 0.01 kPa (0.03 - 0.05 in H20)(References
5-52, 5-54). Typical stack flue gas velocities and mass flowrates range from 7.6 to 12.2 m/sec
(25-40 ft/sec) and 3.7 to 4.9 kg/s m2 (0.75 -1.0'lb/s ft2), respectively (References 5-52, 5-54,
5-56). .
Combustion air preheaters are often-used to improve the efficiency of a fired heater. The
maximum thermal efficiency obtainable with current air preheat equipment is 92 percent
5-35
-------
(LHV)(Reference 5-55). In the preheater, heat is transferred from the flue gas to the combustion
air. Therefore, less heat is required to heat the combustion air which allows a greater proportion
of the total heat released to be absorbed in the radiant section. And less fuel is required to
reach the required combustion temperature. In addition, the preheater raises the adiabatic flame
temperature above that of ambient air heaters. The trend in the 1980's will be to apply air
preheaters to larger sources because of improved fuel efficiency (Reference 5-57). Because of the
lower density of the heated combustion air, all heaters using air preheaters will also require a
fan.
The temperature requirement of the feedstock is an important factor in heater application and
design. Typical feedstock temperature requirements for refinery and chemical manufacturing
processes are included in Tables 5-9 and 5-10. Feedstock temperature can influence the number and
spacing of tubes and the mass velocity of the feedstock within the tube. The mass velocity of the
feedstock determines the tube'size and number of passes. Mass velocities usually range from 222 to
1,972 kg/s m2 (45 - 400 Ib/s ft2)(Reference 5-52) In addition, heat transfer rate and firebox
temperature will be determined by the feedstock temperature requirement. Heat transfer rates range
from 85 to 160 MJ/hr m2 (7,500 - 14,000 Btu/hr ft2)(Reference 5-52). Firebox temperatures are
included in Table 5-11 and range from 538 to 1427°C (1000-2600°F)..
The lowest temperature fired heaters, less than 260°C (500°F) feedstock temperature, are used
in natural gas processing plants to prevent condensation. These heaters average only about 4.4 MW
(15 x 106 Btu/hr) fuel input and can be as low as 0.15 MW (0.51 x 106 Btu/hr)(Reference 5-38). The
lowest temperature heaters in the refining and chemical manufacturing industries are the circulation
heaters, while the highest are fired reactors.
Fired heaters can use a variety of fuels. In general, the chemical manufacturing and refining
industries use oil or gas. These heaters also burn a wide variety of waste liquids or gases that
are not usually considered as fuels. Other fuels such as coal and petroleum coke can be used, but
they are not expected to make a significant contribution to the fuels used during the 1980's. Both
industries use more off gases and natural gas than oil. However, the refining industry burns more
off gases than the chemical industry. Off gases are a by-product from manufacturing processes.
They can be made up of a variety of components whose relative composition can vary considerably over
time. Heating values of refinery off gases can range from under 100 Btu per cubic foot to over 3000
Btu per cubic foot (Reference 5-58). It is expected that in the future, refineries will use more
5-36
-------
refinery off gas and heavier fuels because of the higher market demands for lighter fuels. The
chemical industry will probably continue to use primarily natural gas.
The burners in a fired heater can be arranged to fire from the top, bottom, or sides of the
heater. Most heaters fire from the bottom or sides because of the design simplicity and efficiency.
Some high temperature specialty units such as pyrolysis heaters and steam-hydrocarbon reformer
heaters are designed to use many small radiant-wall burners to heat the refractory surface. Other
designs of these specialty units include the use of bottom fired burners near the wall to heat the
refractory, a combination of bottom fired and radiant-wall burners to heat the refractory, and
bottom fired burners midway between the refractory and the process tubes.
Many different types of burners are available for fired heaters. Burners can be differentiated
by their flame shape, method of mixing of fuel and air, atomization type, and draft type. The
primary objective of a burner is to mix the fuel and oxygen before and during ignition.
The application of the heater and the temperature requirements affect the type of burner
selected. For example, some processes will require a more even heat distribution or more intense
heat. One of the important operating features of a burner that is considered in the heater design
is the flame type. The type of flame will determine heat intensity and heat distribution.
Important flame properties are shape and heat flux rate. A burner must be able to maintain a
flame-stable operation during a wide range of operating variables and provide a reasonable flame
shape when fuel.and air input varies (Reference 5-58).. Flame shape is affected by the manner in
which the fuel and air are introduced. The length of the flame is affected by the type and amount
of fuel, combustion air temperature, register draft loss, and excess air rate (Reference 5-59). The
flame shape must also comply with the mechanical configuration of the radiant section so that the
flame or hot gasses do not impinge on the tubes. It should also provide an even heat distribution.
Typical flame shapes are flat and conical. Flat flame burners can produce short, wide flame
patterns or longer patterns with less spread. Conical shaped flames can have short flames or
extremely long flames. Short flames usually are high intensity while long flames provide a uniform
heat flux rate throughout the radiant section. The heat flux rate is usually determined by the
process and heater design (Reference 5-58).
Burners can be designed to burn gas, oil, or a combination of gas and oil. Gas-fired burners
are simpler in design and operation than oil-fired burners. The basic gas burner classifications
5-37
-------
are prenrfx insoirating and raw gas burning. In the prenrix burner (Figure S-5a), primary air 1s
mixed with the fuel gas prior to ignition at the burner tip. Induction and mixing of primary air
occurs due to the kinetic energy of the fuel gas as it expands through orifices in the burner.
Secondary air is usually required to allow for swings in fuel calorific values and to complete
combustion. If fuel calorific value is very constant, linearity of excess air levels can be
maintained from 33 percent to 100 percent of firing capacity by inspiratlng all primary air
(Referencs 5-56). As swings in calorific value of the fuel increase, the use of secondary air must
be increased to permit excess air linearity. However, as secondary *ir use increases, the range of
firing rates where excess air levels can ba maintained without burner register adjustments
decreases. When the firing rate is below the range the burner is capable of maintaining excess air
linearity, excess air levels Increase significantly as firing rate is decreased. Generally, anout
50 to 60 percent of the combustion air is mixed as primary air in the burner. Premix burners
rsquire a fuel gas pressure of greater than 170 kPa (10 psig) to maintain combustion air mixing. An
advantage of prenrix burners is that they can operate with high or low excess air rates
(Reference 5-47).
Raw gas burners (Figure 5-5b) receive fuel gas without any prenrlxing of combustion air. Mixing
occurs in the combustion rone at the burner tip. The tip has a series of small ports to aid mixing.
The raw gas burner can handle large turndown ratios for a given combustion condition; however,
combustion air adjustments must be made over the full burner operating range (Reference 5-47).
Oil-fired burners (Figure 5-5) are classified by the type of fuel atomization. Oil is
atomized to improve the mixing of fuel and combustion afr. Atomization methods commonly used are
Htechanfcal, air, or steam. Steam atomization is usually preferred because it is mor« economical,
controls the flame better,' and can handle larger turndown ratios. Typical steam requirements are
0.07 - 0.16 kg steam/kg of oil (0.07 - 0.16 Ib steam/'lb of oil). The main requirements for the
steam are that it must be dry and available at a constant pressure, approximately 791 kPa (100 psig).
As in premix burners, cil-Fired burners utilize primary and secondary air to improve heater efficiency
(Reference 5-47).
Combination burners (Figure S-5d) are designed to bum all gas, all oil or any combination of
oil and gas. A typical arrangement is to have a single oil gun in the center of an array of raw gas
nozzles. The air for oil or gas firing can be controlled separately in these units. Therefore,
5-38
-------
.Burner Tip
Primary Air
Register
fcfractnry
Register-
Assembly
Inspiriting
Qas Pilot
Full Gas
:ntet
a. Premix Burner
b. Raw Gas Burner
HUfft*
Block
Air Shutter
Sas Burner
Ajieably
Oil ami Stea Inlet 'Primary Air
InUt Control
Fuel 011
Burner
.^luffl
t Block
Secondary Air
Control
.Rax tnlet
. 1 And Steam
Inlet
c. Oil Burner d. Combination Oil and Gas Burner
Figure 5-5. Typical burners by type of fuel burned.
5-39
-------
these burners can control primary air for oil combustion and secondary air for fuel-gas combustion
(Reference 5-47).
High intensity burners are generally used when a heavy fuel,or when a fuel that has a large
quantity of inerts, is to be burned. High intensity burners usually have a large, cyclindrical .
refractory lined combustion chamber. The mixing of fuel and combustion gases is very intense in
this chamber. Combustion is fully established in the chamber but not completed. Because of cir-
culation patterns present in the chamber, flames with stable and controlled size and shape are
f
produced at relatively low excess air. The fuel gases from the high intensity combustion are
expelled at a high velocity and temperature. This produces a uniform temperature profile in the
firebox. High intensity burners require forced-draft air supplied at relatively high pressures.
*
These air pressures can range from 1.5 to 5 kPa (5-20 in HgOHReference 5-47).
The burner type and design are affected by the type of draft. In the past few years the trend
has been towards forced-draft burners because they are more fuel efficient. Forced-draft burners
can control airflow, therefore, they can be operated at lower excess air, control flame shape
better, control fuel/air ratio, and be used with air preheaters (Reference 5-54). They can also
cleanly combust fuels with low or varying heat values at low excess air rates. Having the
capability to control or use each of these features allows the heater to be operated more
efficiently. In addition forced-draft burners have a larger capacity per burner; therefore, less.
burners are required which reduces cleaning, maintenance costs, and makes them more amenable to
automatic control (Reference 5-60).
5.3.1.1.2 Emissions and Control Techniques
Presently there are a wide range of control techniques that have been studied or demonstrated
on fired heaters. These techniques include, operational changes, combustion modification,, flue gas
treatment, or a combination of these techniques since they are usually more effective than flue
gas treatment alone. Combustion modification techniques have been the most widely used methods to
control NO emissions. Because an uneven distribution of heat to the process tubes could cause
X
coking of the feedstock, combustion modification techniques have been limited more than in other
combustion processes.
Since combustion modifications are the predominant technologies being used their descriptions
and capabilities will dominate the following discussion. Prior to'discussing the various combustion
5-40
-------
modification techniques, emissions from uncontrolled furnaces, and operational control methods and
their emissions will be discussed. Flue gas treatment will be discussed last.
Operational Control Techniques - Reducing emissions of NOX from standard burners used In fired
heaters 1s sometimes possible using methods available to the heater operator. These methods Include
monitoring stack oxygen and combustibles, and then reducing average excess air levels, reducing the-
amount of oil combusted, and adjusting burner air registers for minimum NO emissions levels.
Typically, heater operators adjust -the stack damper and burner air registers for a compact, well
defined and stable flame. Additionally, the stack damper 1s adjusted to assure the operator that
sufficient combustion air is available during all operating conditions. At these settings the stack
oxygen level usually averages 5 to 9 percent. Recently, because of energy conservation programs
within some plants, stack oxygen levels are being monitored and controlled to lower levels.
Emissions from a patrolewa refinery reformer with 70 preraix burners were analyzed using over
2000 hourly measurements (Reference 5-61). Using an auto regress ion technique, NQX emissions as a
function of stack oxygen level were predicted. Figure 5-6 shows the results and the 95 percent
confidence limits of this predictive procedure for firing rates that were typical during the test
period. The maximum design firing rate for the heater was 20.7 HW (70 x 10 Btu/hr). The actual
operating conditions for this heater was as follows:
Fuel Firing Rate Stack 0, , NOV Emissions
KM (106 Btu/hr) % £ ng/Q (Ib/lpS Btu)
14.3 (48/9) 3.15 39 (0.09)
3.8 (13) 0,,1 0 (0.0)
19.0 (65) 17.0 120 (0.28)
5-47
-------
01
i
fvj
170
160
150
140
130
120
55 110
n 100
5
jj 90
] 80
x
: 70
60
50
40
30
20
10
Naptha Reformer: Prenrix
Burner, 70 percent load
Btu/hr)
3 4 56
Stack Oxygen Level, Percent
8
0.4
0.35
0.3
0.25
V)
0.2 §
in
cr
_j
o
0.15
CO
<*
0.1
0.05
Figure 5-6. NOW emissions from standard premix burners.
-------
This heater was closely monitored and adjusted to maintain NO emissions within permit limits.
The NO emissions are controlled by manual adjustment of the stack damper to control excess air
levels. The plant estimates that NOV monitoring and control require 4 manhours/day for six heaters.
A
Although it appears that NO emissions per unit of heat input can be reduced by about 20 percent by
reducing stack oxygen levels from 6 percent to 3 percent, in actuality average NO emissions may be
/*
reduced less. The reason is that NO emissions per unit of heat input tend to maximize at some
A
firebox oxygen level above 6 pe'rcent to 8 percent oxygen due to the cooling effect of the-air on the
flames. By operating for some periods of time above this point of maximum emissions, the average
emissions would be reduced from the emissions produced at the average stack oxygen level. It should
be noted, however, that because of the increased usage of fuel, the NOX emissions per unit of time
may continue to increase above this stack oxygen level.
Less comprehensive data are available on a standard combination raw gas and oil burner
(Reference 5-62). The data were collected on a 108 cubic meter per hour (16,250 barrels per day)
natural draft, vertical cylindrical, crude oil heater with six 2.7 MW (9.2 x 10 Btu/hr) maximum
capacity burners. Although 89 one hour NOX emission tests were conducted, there were six variables
that could be controlled by the operator that may have affected emission levels. These variables
included the process rate, heat input rate, fuels combusted, secondary air register setting, steam
injection rate and stack oxygen level. Due to the large 'number of variables, the effect of excess
air levels on NO emissions could not be quantified with precision. However, a trend can be
observed when data are arranged in an order of increasing stack oxygen concentrations in groups
where the other variables are similar. .Table 5-12 contains selected data from these 89 tests where
the other variables were similar and there were at least three different values of stack oxygen
level. A definite trend of increasing NOX emissions with increasing stack oxygen levels is apparent
within each of the groups.
Combustion Modification Techniques - As stated previously, combustion modification techniques
are the most common and widely used method to control NO emissions. Techniques that have been used
on burners designated by the vendor as 1ow-NO burners include low excess air, high swirl burners;
staged addition of combustion air; combustion gas self recirculation; and staged addition of fuel.
These techniques are effective on a number of raw gas and dual fuel burner types. Because of design
5-43
-------
TABLE 5-12.
SUMMARY OF NOX TEST DATA ON A STANDARD DUAL FUEL
BURNER
Fuel* Process Rate
MG-FG/AG/16 (m3/hr)
100/0/0 55
100/0/0 74.9
100/0/0 94.1
0/100/0 76.2
56/0/44 55
37/0/63 76.2
54/0/46 94.1
63/0/37 95.4
Heat Input
(MM)
7.4
7.9
7.9
8.6
10.9
10.8
11.3
11.8
15.4
15.8
' 15.9
16.3
11.5
11.0
11.1
11.8
11.1
11.8
11.5
11.6
8.2
11.4
14.1
14.3
Secondary Air Stack Op
Register /,%
(% open) (%>
502 2.3
4.2
5.6
8./
bO% 0.9
2.0
3.9
5.6
50% 1.9
3.3
4.2
4.5
50% 1.3
1.7
2.0
3.6
3.9
4.0
4.1
4.1
50% 1.7
2.9
4.8
6.7
50% 1.25
2.2
3.8
50% 1.9
2.6
3.4
50% 1.2
1.6
2.8
2.9
NO Emissions
ng/J (1b/106 Btu)
29
44
48
47
' 32
44
46
52
35
41
41
42
42
47
51
54
56
50
54
56
75
88
90
63
64
82
101
73
71
87
64
73
79
85
(.Ob8)
(.102)
(.111)
(.110)
(.07)
1.102)
(.107)
(.121)
(.081)
(.096)
(.096)
(.098)
(.100)
(.110)
(.120)
(.130)
(.130)
(.120)
(.130)
(.130)
(.175)
(.205)
(.210)
(.147)
(.149)
(.191)
(.235)
(,170)
(.165)
(.203)
(.149)
(.170)
(.184)
(.198)
Percentage ot natural gas-fuel gas/adsorber gas/number 6 fuel oil.
5-44
-------
requirements, premix burners incorporate tha intimate mixing of fuel and air and can combine this
with combustion air staging and low excess ciir firing. This may explain why the heater with
standard premix burners had significantly lower NOV emissions than the standard dual fuel burner
A «
discussed previously. Therefore, the discussions that follow will be limited to raw gas or liquid
fuel burners that incorporate one or more of these technologies.
Low Excess Air, High-Swirl Burners - Burners that depend on low excess air high-swirl
conditions to reduce NO emissions have been used on a number of petroleum refinery process
heaters. Two of these heaters were tested to determine emissions of NO. CO, and SO, and to
' X t
determine the effects of heater operation on emissions (Reference 5-63). The first test was of a
3
four cell, balanced draft refinery gas fired reformer heater with a maximum capacity of 134 M /h
(20,000 bbl/day) of reformate. Each cell was separated from adjacent cells with a common wall and
had four burners mounted in each end wall. The combustion air for all burners was preheated to
approximately 544 K (520°F). NOX emissions for this heater ranged from 27.4 ng/J (0.06 lb/10
Btu) to 77.4 ng/J (0.18 lb/10 Btu). NO emissions exhibited a strong dependence on excess air
levels and a moderate dependence on heat input rate. Figure 5-7 is a presentation of the multiple
regression of NOX emissions as a function of heat input rate and stack oxygen concentrations. For
simplicity the five curves were plotted at the average heat input rate five groups of four process
conditions of the four cells tested.
The second heater tested was a vertical cylindrical debutanizer bottoms reboiler. A single
floor mounted burner was used with ambient temperature combustion air supplied by forced draft. The
o
process rate for the second heater was 187 M /hr (28.3 bbls/day) with the burner firing at a heat
input rate of 7.62 MW (26.1 x 10 Btu/hr). This was about 95 percent of the process capacity of
the heater. Compared to the first heater at near maximum load, NO emissions from the second
heater were less. Figure 5-8 presents the NO emissions during the four tests of the second
heater. '
Staged Combustion Air Burners - Burners that stage combustion air to reduce NOX emissions
have been the predominant design used in U.S. petroleum refinery heaters. Most of the burners
appear similar to standard dual fuel burners except that a third air register exists to provide
additional control. Other burner designs incorporate the gas and oil gun within the primary tile
case. Air to the primary tile case for initial combustion of the fuel is regulated by a single air
register with secondary (staged) combustion air controlled by a second air register and admitted
between the primary and secondary tile case. Non-vendor published data are available only on
burners of the three air register design.
5-45
-------
(0.2
80
70
(0.15)
60
3
CD
50
tn
o
r*
I/I
in
O
as
40
30
(0.05)
20
10
Heat Input Rate
10.2 MW (35 x 106 Btu/hr)
9.5 MW (32 x 106 Btu/hr)
8.27 MW (28.3 x 106 Btu/hr)
6.45 MW (22.1 x 106 Btu/hr)
4.56 MW (15.6 x 10G Btu/hr)
Heater Operating Conditions
Preheat Air Temperature: 544 K
(520°F)
1 23 456
Stack Oxygen Level, Percent
Figure 5-7. NO emissions from a refinery gas-fired reforming heater
with low excess air burners.
5-46
-------
(0.15)
60
3
4->
CO
50
g(O.l)
01
VI
30
(0.05)
20
10
345
5itack Oxygen Level, Percent
Figure 5-8. NOX emissions from a single burner, forced draft
debutanizer bottoms reboiler.
5-47
-------
NO emission data for the three air register design burner type firing gaseous fuels include
one set of short terra emission tests of a forced draft heater and two sets of long term data from a
natural draft heater and a balanced draft heater with preheat.
The forced draft heater was a vertical cylindrical heater rated for 92.7 M /h (14,000 bbl/day)
crude oil throughput (Reference 5-63). Three burners were floor mounted in the heater and fired a
combination of natural gas and refinery gas. During the test the process rate varied from 57
percent to 80 percent of rated capacity and the firing rate varied from 3.48 MW (11.9 x 10 Btu/hr)
to 4.85 MW (16.5 x 10 Btu/hr). NOX emissions for this heater were as low as 24 ng/J (.056 lb/10
Btu) to as high as 44 ng/J (0.10 lb/10 Btu). Although the air register settings were adjusted to a
number of different settings, 13 tests were conducted with air registers set to approximately the
same openings that existed prior to the start of the tests. Of the other air register settings
evaluated during the test, a setting of primary and tertiary registers 100 percent open and
secondary registers closed was designed as optimum for low NOV operation. Figure 5-9 presents the
X
results of the 13 tests and three tests conducted at the approximate setting designated as optimum.
Additionally, the two tests conducted with the air registers set at openings that may be considered
as simulating a standard burner are also included. Although not conclusive, it appears that the air
register setting could increase or decrease NO emissions from 3 to 20 percent from the levels
produced with the original register settings. Also it appears that excess air levels have a greater
impact on NO emissions than either air register settings or load.
The two sets of long term data for gas fired heaters were collected by the plants as a result
of local permit requirements (Reference 5-61). The first was a naphtha reformer that consisted of
four separate vertical cylindrical heaters with a common'convective section. The combined capacity
of the four heaters is 25.4 MW (87 x 10 Btu/hr) with four burners in three of .the heaters and three
burners in the remaining heater. The outlet temperature of the naphtha from the heaters was approx-
imately 770 K (930°F). Over 1000 hourly data points of stack oxygen level, fuel firing rate, and
NO emissions were collected. The plant monitors oxygen level and adjusts the stack damper for
energy conservation considerations. Adjustments are not performed to the burners or stack damper to
control NOV emissions because the plant is well within its permit limits. Fuel firing rate for
A
these heaters averaged 20.9 MW (71.3 x 106 Btu/hr) and ranged from 1.37 to 24.5. MW (4.7 to 83.8 x
10 Btu/hr). Stack oxygen levels averaged 5.5 percent and ranged from 2.7 to 8.4 percent for this
same period. The resulting NOX emissions for these heaters during this period averaged 42.9 ng/J
(0.10 lb/106 Btu) and ranged from 21.5 ng/J (0.05 lb/106 Btu) to 68.7 ng/J (0.16 lb/106 Btu).
5-48
-------
(.12
50
01
-fc.
lO
CO
VD
O
(.10)
C(-08)
30
c?(.06)
I 20
« (.04)
i
O
10
(.02)
Trend Line
3.SWW (12 x KT Btu/hr)
4.7 MW (16 x 106 Btu/hr)
OProcess Rate, 53 m3/hr (8000 bbl/day), Fuel Firing Rate 3.48-4.16 MW
£ AProcess Rate, 72.87-73.13 m^/hr (11,000-11,040 bbl/day), Fuel Firing Rate 4.48-4.85
"SO Optimum Low-N0x Register Adjustment (High Process Rate)
D. 1-2
-------
Because there was a significant amount of missing data in the middle of the data, both parts of
the data were analyzed to determine the effects of load and stack excess air levels on NOV
A
emissions. Load effects on the NOV emissions could not be determined to be of significance with
A
either data set analyzed. However, stack oxygen level did have a significant effect on NO
emissions. Figure 5-10 shows the predicted average NO emissions as a function of stack oxygen and
the 95 percent confidence limits for these values. Only the data set with the highest predicted NO
emission rate is shown. The other data set predicted approximately 20 percent less NO emissions.
A
The second set of data collected was on a balanced draft vertical cylindrical crude heater that
used preheated combustion air (Reference 5-61). The heater was retrofitted with nine staged air
burners with a total heat input rating of 25 MW (86.3 x 10 Btu/hr). The preheat temperature was
between 485 K (414°F) and 50C K (550°F). Almost 3,000 hourly data points of stack oxygen level,
fuel firing rate, and NO emissions were collected. The plant monitors and controls NO emissions
because of local permit limits. Adjustments to NO emissions are accomplished by adjusting the
stack damper to control excess air levels. Burner registers are adjusted to control flame impinge-
ment on the process tubes. Fuel firing rate for this heater averaged 16.4 MW (56 x 10 Btu/hr) and
ranged from 2 to 20.3 MW (6.9 to 69.3 x 10 Btu/hr). Stack oxygen levels averaged 2.11 percent and
ranged from 0.1 to 19.0 percent. NOX emissions ranged from 3.4 ng/J (.008 lb/10 Btu) to 68.8 ng/J
(0.16 lb/106 Btu) and averaged 34.4 ng/J (0.08 lb/106 Btu).
The data were analyzed statistically to determine the effects of load and stack excess air
levels on NO emissions. Increases in both load and stack excess air levels increased emissions of
NO . Figure 5-11 shows the predicted NO emissions as a function of stack oxygen level for two
A A
different loads. The 95 percent confidence limits for these values are also indicated.
Limited NO emission data from staged combustion air burners firing 100 percent liquid fuels
are available. As a result of a study to determine emission levels from process heaters with staged
combustion air burners three emission tests were conducted (Reference 5-63). The first test was on
a natural draft vertical cylindrical crude heater which fired a 0.15 percent nitrogen distillate
oil. A single burner of 4.4 MW (15 x 10 Btu/hr) maximum rated firing capacity was installed in the
floor of the heater. Actual firing rate varied from about 4.03 MW (13.8 x 10 Btu/hr) to 4.49 MW
(15.3 x 10 Btu/hr). Stack oxygen levels varied every few seconds resulting in a variation of from
5 to 8 percent 02 in the "as found" condition. This variation was attributed to the low overall
pressure drop of the heater (which did not have a convective section) and local variation in wind
5-50
-------
70
en
i
60
50
?40
VI
c
to
VI
30
20
10
0.15
0.10
CO
0.05
1 234 5 6 7 8
Stack Oxygen Level, Percent
Figure 5-10. NOX emissions from a natural draft Naptha reformer with staged combustion
air burners.
-------
60
50
at
i
in
cn
c
w"40
c
o
20
10
0.15
18.8
0.10
o
Ol
(43
0.05
I
1 2345678
Stack Oxygen Level, Percent
Figure'5-11. NOX emissions from a balanced draft crude preheater with staged
combustion air burners, combustion air preheater to 485 to 500 K.
-------
conditions at the site. Stack oxygen levels and burner air register settings were adjusted to
determine their effects on NO emissions. Reducing stack oxygen levels decreased NO emissions
A X
significantly as shown in Figure 5-12. Minor adjustments in.burner air register settings did not
seem to affect NO emissions significantly. However, as the two tests indicated in Figure 5-12 by
the hexagonal points, closing the tertiary air register to about 15 percent open and the primary air
register to 50 percent open increased emissions moderately. The complete closure of the tertiary
air register would simulate a standard burner.
The second test was on a natural draft horizontal heater firing 0.81 percent nitrogen residual
oil (Reference 5-63). Twelve 0.97 MW (3.3 x 10 Btu/hr) staged air burners were mounted on the long
walls of the heater and fired toward the center. The heater was fired at about 9.8 MW (33 x 10
Btu/hr) during the test. The stack damper on this heater was stuck, therefore, the burner air
registers were adjusted to attempt to reduce excess air levels. Figure 5-13 shows the results of
the six tests on this heater. The data points, connected by lines represent tests where air register
settings were the same. Although the tests with more open air register settings had higher
emissions, the differences are probably not significant.
The third test was on a natural draft vertical cylindrical heater firing 0.81 percent nitrogen
oil (Reference 5-63). Three 2.24 MW (7.64 x 10 Btu/hr) staged combustion air burners were mounted
in the floor of the heater. Although most of the tests were conducted at about 6 MW (20 x 10
Btu/hr) heat input, one test was conducted at about 6.5 MW (22 x 10 Btu/hr). The "as found" stack
oxygen level was between 8.1 and 11.4 percent. This was reduced to as low as 6.9 percent by
partially closing one of the two stack dampers and by minor adjustments of the burner air registers.
No clear indication of the effect of stack oxygen on NOX emissions can be discerned from the data.
Possible reasons for this are that too few tests were conducted and the stack oxygen levels may not
have been reduced past the level of maximum emissions above which flame cooling by the combustion
air begins to reduce emissions. The results of the seven tests are shown on Figure 5-14.
The simultaneous firing of both liquid and gaseous fuels in the same heater is a more typical
occurrence in the chemical and petroleum industries. This is accomplished either by firing only
liquid fuels in some burners and only gaseous fuels in the remainder of the burners or by base loadint
all of the burners on liquid fuel and supplementing the heat requirements with gaseous fuels. Three
tests are available on combined firing of liquid and gaseous fuel in staged combustion air burners.
5-53
-------
Ul
I
01
130
120
no
100
90
80
70
60
50
40
30
20
10
0
O
Burner Type: Staged Combustion Air
s Fuel: Distillate Oil
Load: 33.12 m3/h (5000 bbl/d)
Q All registers 100X Open (Baseline)
Reduced Tertiary Air and Primary Air
Reduced Primary Air Only
Primary and Secondary Air Registers 502 Open
Primary Register 50% Open,
Secondary Register 70% Open
I I 1 I I
4 5
Stack Oxygen, % Dry
8
0.30
0.25
0.20
o
CO
rt-
C
0.15
0.10
0.05
Figure 5-12. NO emissions as a function of stack oxygen for a distillate oil-fired,
natural draft process heater.
-------
250
(Q.S8)
Z40
(0.56)
03
, 230
UQ.53)
. 220
2(0.5.1}
210
(0.49)
200
(0.47)
Firing Rate 9.8 MW (33 x 10b Btu/hr)
35/86/81
Mr Register Setting:
Primary/secondary/tertiary
Stack Oxygen, % cry
Figure 5-13. Residual oil-fired horizontal heater with
staged combustion air burners.
5-55
-------
290
(0.6$
CO
IO
O
280
'(0.65;
-------
The first of these tests consisted of 14 short term emissions evaluations (Reference 5-63).
The heater was a vertical rectangular natural draft heater with 10 staged combustion air burners
installed. Each burner is rated at 1.2 MW (4 x 10fi Btu/hr).- The firing rate for the heater during
the tests varied from 8.79 MW {30 x'106 Btu/hr) to 11.98 MW (40.9 x 106 Btu/hr)'. The liquid fuel
fired was 0.81 percent nitrogen oil. All but three of the tests were conducted with all burner air
registers completely open. A curve for the tests with over 80 percent oil firing and all burner air
registers open shows a definite lowering of NOX emissions with reduced excess air levels. The other
three tests were conducted with only minor adjustments to the burner air registers. Considering the
small number of tests and variations in firing rate and oil-to-gas ratios, a significant change in
NOX emissions with changes in burner air register settings is not noticeable. Figure 5-15 shows the
results of these tests. Tests with different burner air register settings are indicated with
different shape data point markers. The percentage of oil fired is also indicated in Figure 0.
Based upon the two tests with lower rates of oil firing it appears that the percentage or amount of
fuel oil fired affects th« emissions of NO^.
The other two tests of staged combustion air burners are from long term continuous emission
data on heaters firing a combination of refinery gas and residual oil (average nitrogen content of
about 0.65 percent)(Reference 5-61). Both heaters were operated by the same refinery. The refinery
continuously monitored NOX emissions to show compliance with a permit requirement that they easily
achieved. Therefore, heater adjustments were not usually made to reduce NO emissions. The plant
did, however, monitor stack oxygen level and adjustments were made when stack oxygen levels were
outside control limits.
The first heater was a natural draft vertical cylindrical atmospheric crude heater
(Reference 5-61). Eight floor fired burners with a total firing capacity of 377.5 MW (110.6 x 106
Btu/hr) are installed 1n the heater. The crude temperature to the atmospheric distillation column
was between 636 K and 639 K (685°F and 690"F). The oil fraction averaged 31 percent and varied from
20 to 45 percent of the total heat requirement during the almost 1000 hours of data collection. The
total fuel firing rate averaged 416.7 MW (122.1 x 10° Btu/hr) and ranged between 67.6 and 466.6 MW
(19.8 x 10 and 136.7 x 10 3tu/hr). The stack oxygen levels were maintained between 1.0 and 9.0
percent and averaged 3.5 percent. NOX emissions for this heater averaged 51.5 ng/J (.12 Ib/lQ6 Btu)
and ranged from a low of 30.0 ng/J (.07 lb/105 Btu) to as high as 68.7 ng/J (.16 lb/10S Btu).
Mathematical modeling of the data could not detect a relationship between oil-to-gas ratio or heater
load and NO emissions. However, stack oxygen levels did correlate with emissions of KQX> Figure
5-37
-------
*->
CO
3
C
O
(0.5)
210
200
(0.45)j
190
180
(0.4)
170
160
(0.35)
150
140
130
. (0.3)
Burner Air Register Openings
O « 100/100/100 (Curve drawn for this data)
D * 65/72/100
A - 60/59/76
(Oil firing rate in MW/percent of
heat input by oil)
(10.0/85%J
O
(8.9/82%)
(9.4/84%)
O/ D(9.5/85%)
0(9.4/83%)
0(9.1/82%)
O(l 0.0/84%)
(8.2/81%)
O
39.4/84%)
'(9.4/843!)
A(9.4/84%)'
(7.9/81%)
O
(5.9/61%)
O
(5.0/57%)
O
I
I
I
8
4567
Stack Oxygen Level, Percent
Figure 5"-15.Residual oil/refinery gas-fired natural draft heater.
5-58
-------
5-16 shows the NO emissions as a function of stack oxygen and also the 95 percent confidence limits
on the NO emissions.
The second heater was a natural draft vertical cylindrical vacuum distillation column heater
(Reference 5-61). Four floor fired burners can provide a total heat input capacity of 126.3 MW
(37.0 x 106 Btu/hr). The process fluid is heated to approximately 650 K (710°F). The oil fraction
fired in this heater varied more (0 to 40 percent) but averaged less (6 percent) than the previous
heater. The total firing rate averaged 93.2 MW (27.3 106 Btu/hr) varying from 67.6 to 151.9 MW
(19.8 x 10 to 44.5 x 10 Btu/hr). The stack oxygen levels were maintained between 0.9 and 11.4
percent and averaged 3.9 percent. NOX emissions during these conditions ranged from 17.2 to 51.5
ng/J (0.04 to 0.12 lb/106 Btu) and averaged 30.0 ng/J (.07 lb/106 Btu). Mathematical modeling of
the data indicated that emissions increased with increasing oil fraction and decreasing load. One
possible explanation for this is that the heater was base loaded on oil. Therefore, as load was
reduced the oil fraction increased, however, emissions rates were not reduced or reduced only
slightly, resulting in an increased emission factor. Figure 5-17 includes the NO emissions with
95 percent confidence limits for two different loads as stack oxygen level varies. As with the other
heaters NO emissions decrease with decreasing stack oxygen level.
All of the above data on staged combustion air burners were collected on cylindrical flame
burner designs where the oil gun was in the center of the recirculating primary oil tile, the gas
tips were between the primary and secondary tile cases and a tertiary air port surrounds the
secondary tile. Data on other burner designs and on other burner types that stage the combustion
air are available. At least one manufacturer claims that their burner is designed to incorporate
flue gas self-recirculation in addition to staged combustion and the ability to operate at low excess
air levels (Reference 5-64). Emissions reported by the vendor for this burner installed in a 6000
barrels/day hydro desulfurization unit show reductions in ,NOV emissions while firing gaseous fuels
A
of from 60 to 80 percent compared to a burner considered by this manufacturer to be standard. Data
collected by a refinery on a heater with these burners installed show about a 50 percent reduction
of NOV emissions (Reference 5-65). Test furnace data on another manufacturers burner of a somewhat
A
similar appearance (gas tip'and oil gun within primary tile and and secondary air port around
primary tile) show NO emissions as a function of primary air register opening at. similar excess air
levels from 40 to 70 percent less than a burner considered to be standard by this manufacturer
(Reference 5-58). Data for both of these manufacturers' burners show reductions in NO emissions
during liquid fuel firing of about 30 to 40 percent compared to a burner considered by each to be
5-59
-------
in
k
70f
60
50
'-3
O1
VI
c
o
30
20
10
0.15
345
Stack Oxygen Level, Percent
o
o>
0.10
0.05
Figure 5-16. NOX emissions from a natural draft atmosphere crude heater with staged
combustion air burners, oil and gas fuel combustion.
-------
60
0.15
01
I
en
50
V)
c
o
r-
V)
l/>
30
x
o
20
0.10
cr
CO
H-
0.05
10
I
1 2 3 4 56.7 8
Stack Oxygen Level, Percent
Figure 5-17.NOX emissions from a natural draft vacuum distillation column heater with
staged combustion air burners, oil and gas firing.
-------
standard. Data for the two different burners cannot easily be compared because of differences in
test conditions.
At least one manufacturer has two burner designs incorporating staged combustion air that can
be substituted for standard high intensity burners and standard flat flame burners (Reference 5-58).
Although these burners have been successfully used in full scale turnovers, emission data from full
scale furnaces are not available. Test furnace data for the high Intensity burners show reductions of
NOX emissions during gas or oil firing of about 60 percent from a burner the manufacturer considers
standard. Data presented by the manufacturer show that as fuel nitrogen increases the amount of
reduction of NOV emissions increases. Test furnace data for a low NO., flat flame burner firing next
A X
to a radiant wall shows a reduction of between 40 and 50 percent by changing the staged air port from
the side away from the radiant wall to the side of the burner nearest the radiant wall. According to
the Manufacturer, emissions for both of these flat flame burners are less than the standard flat
flame radiant burner.
Staged Air Lances - As a result of a research project to develop a method to reduce NOX
emissions from refinery process heaters, an air lance system was developed, installed, and tasted on
an operating natural draft vertical cylindrical crude heater (Reference 5-62). The design consists
of four 3.18 CM (1.25 1n) stainless steel lance tubes inserted through the furnace floor around each
of the six 2.68 HW (9.14 x 10S Btu/hr) dual fuel burners. A 45° elbow is welded on the end of each
lance and a fan was used to supply air to the tubes. An optimum staging height of 0.3 m (1 ft)
above the furnace floor was established for gas firing and 1.22 m (4 ft) was established for oil/gas
firing. Since burner rile tops were about 0.23 m (.75 ft) above the furnace floor, staging heights
of less than this impinged on the burner tile. An optimum' burner equivalence ratio (the ratio of
air drawn through the burner to the stoichiometric requirement) was-not determined, however,
emissions declined with decreased equivalence ratio (increased staging). A leveling off of
emissions was not apparent at the maximum air injection rate.' During short term tests, when fan
capacity limited the air injection rate, an NO emission reduction of 35 to 45 percent was
demonstrated during gas firing and a reduction of about 35 percent was demonstrated during
combination oil/gas firing. Short term tests that were conducted after increasing the fan capacity,
thereby, allowing increased staging through the air lances, demonstrated lower NOV emissions during
i X
combination oil/gas firing. Additionally, by combining staged air injection with low overall excess
air reduction, a further reduction of NO emissions could be achieved.
A long term continuous emission test was conducted on this heater. During the test, the crude
throughput was maintained at 71.2 m /hr (10,717 bfals/day) or 66 percent of capacity. The fuel fired
3 "?
airing the test was refinery gas with an average calorific value of 61,030 kj/m (1638 Btu/ft ).
Fifteen minute average values for stack oxygen, carbon dioxide, carbon monoxide, and nitrogen oxide
-5-62
-------
were continuously recorded for a 34-day period. The stack oxygen level averaged 2.8 percent during
the test with a range from 0.1 percent to 10.8 percent. The first 15 days of the test the heater
was operated at an average of 1.9 percent stack oxygen. The average NO emissions for this period
were 18.5 ng/J (0.04 lb/10 Btu). The last 16 days of the test the heater was operated at an
average of 3.4 percent stack oxygen. The average NO emissions for this period were 25.5 ng/J (0.06
lb/10 Btu). Minimum NOV emissions were 2 ng/J, however, CO emissions of over 400 ng/J (1 lb/10
X
Btu) occurred at the same time. Minimum NO,, emissions without an increase of CO emissions were 12
A
ng/J (.03 lb/10 Btu) when stack oxygen levels were 1.0 percent. Maximum NO emissions were 59 ng/J
(0.14 lb/10 Btu) where stack oxygen levels were at about 7.3 percent. Although baseline emissions
(at 4 percent stack oxygen level) of 66 ng/J (0.15 lb/10 Btu) were established on this heater prior
»
to the long term test,' other tests of the standard burners without air lances being used (see Table
5-12) at about the same load documented emissions as low as 54 ng/J (0.13 lb/10 Btu).
Staged Fuel Burners - A burner that is being installed in at least two high temperature steam
hydrocarbon reformers, one at a U.S. methanol plant, and one at a Canadian methanol plant, offers
substantial reductions of NO emissions compared to presently used standard burners and possibly
many of the low NOV burners. The burner requires mechanical draft because of the high pressure drop
A
required across the burner (Reference 5-58). At the present time, only gaseous fuels have been used
in the burner, however, the manufacturer believes that with proper nozzle design liquid fuels may be
able to be fired. The burner can be made in round or conical flame and flat flame designs. .Test
furnace data by the vendor show a 60 to 70 percent reduction in NOV emissions from a baseline of 90
X
ppm at 3 percent Og (approximately 46 ng/J or 0.11 lb/10 Btu)(Reference 5-58). Actual emissions
data from an operational furnace would probably be higher because of higher furnace temperatures and
the use of a high degree of preheat.
Flue Gas Treatment - Two methods of treating process heater flue gases to reduce NOX emissions
are available. Both methods use ammonia as a reactant with NO to form nitrogen gas and steam
(Reference 5-65). The difference in the methods is the use of a catalyst to reduce the reaction
temperature of 1250 to 1370 K (1800 to 2000°F) required in the non-catalytic method. The two
methods are called selective catalytic reduction (SCR) and selective non-catalytic reduction (SNCR)
or "Thermal DeNO ." The "Thermal DeNOx method, patented by Exxon, can incorporate the use of hydrogen
to lower the reaction temperature to approximately 980 K (1300°F).
5-63
-------
There are at least 24 refinery heaters that have had an SNCR system installed (Reference 5-65).
Emissions reductions of 35 to 50 percent have been achieved in new and retrofit situations. It is
expected that with Improved ammonia injection methods, increased reductions can be achieved. Since
one of the disadvantages of SNCR is that the fraction of NOX reduced decreases as the concentration of
NO., decreases, this technology may not be as effective when used with low NO burners firing gaseous
rC , A
fuels. However, some refineries use this as an interim technique to meet permit limits during times
of high NOX emissions such as during oil firing or during high production periods.
At least one U.S. refiner has retrofitted an SCR system on a balanced draft crude topping
heater. The heater has been operating since May 1979 without difficulty (Reference 5-65). The
catalyst activity has not deteriorated nor the pressure drop increased in spite of seven emergency
shutdowns unrelated to the heater and several occasions when the furnace operated under a temporary
upset condition. Emissions have been reduced by 90 percent from an average of 100 ppm to an average
of 10 ppm.
Costs - The cost of controlling NOX emissions from refinery heaters has been reported in at
least two studies. The most comprehensive study of costs was recently accomplished by the South
Coast A1r Quality Management District and the Stationary Source Control Division of the California
Air Resources Board. This study, which will be referred to as the CARB reoort, was developed as
part of the background information for a public meeting considering NOX emissions control from
boilers and process heaters at refineries. The major technologies for which cost data were
developed were the retrofit Installation of low NOX burners, selective non-catalytic reduction and
selective catalytic reduction. A less comprehensive study evaluated the cost of retrofitting a
staged air lance system and automatic oxygen trim control on a process heater. These two studies
art the source of cost Information presented in this section (References 5-65, 5-66). The costs of
NO control will be presented in the following order: automatic oxygen trim control, low NO
burners, selective non- catalytic reduction and selective catalytic reduction.
Automatic Oxygen Trim Control - Although automatic oxygen trim systems are common on large
chtsrical industry heaters, roost refineries that control stack oxygen levels do this manually by
using plant operators to adjust stack dampers. The daily time required to monitor and adjust stack
oxygen level varies with the--level of oxygen control desired and the variabilities in process heat
requirements and fuel composition. One plant has estimated that about two hours per day are
required to adjust stack oxygen levels on six heaters in order to control NO emissions
5-64
-------
(Reference 5-61). Much of this time could be reduced by installing an automatic oxygen trim system.
Additional benefits that may be gained include tighter control of the excess air levels and
decreased potential of upset conditions occurring. The installed cost of retrofitting an automatic
draft controller to a natural draft heater of about 16 MW (55 x 10 Btu/hr) thermal input capacity
Was expected to be about $40,000 (Reference 5-62). It was estimated that the initial cost of an
automatic draft controller would be proportional to the heat input rate to the 2/3 power. Annual
maintenance costs for the automatic systems were estimated to be about $200. Because of the tighter
control achievable with an automatic draft controller, additional fuel savings could be achieved
resulting in an overall cost savings. These fuel savings were estimated at about $32,000 per year
for a 16 MW (55 x 106 Btu/hr) heater; $59,000 per year for a 29.3 MW (100 x 106 Btu/hr) heater; and
c
$296,000 per year for a 147 MV1 (500 x 10 Btu/hr) heater (Reference 5-62). The annualized savings
for these heaters were estimated to be about $17,000, $35,000 and $230,000 respectively
(Reference 5-62).
Low NO Burners - Low NO burners are capable of being installed in most process heaters.
However, because of the differences in installation requirements and heat distribution requirements,
the cost to retrofit low NO burners in two different heaters of the same size may be substantially
different. Data collected from eight refineries on the cost of retrofitting low NO burners in nine
separate heaters is presented in Table 5-13 (Reference 5-65). As can be seen in this table," a wide
variation in burner costs and retrofit costs exists, and the variation would not be estimated by
adjusting for the size of the furnace. In the document where these data are presented, it is stated
that the annual cost of retrofitting low NO burners can be estimated using a scale up factor of 0.6
n
from the cost of $11,795 for retrofitting a 9.4 MW (32 xlO Btu/hr) furnace. A contingency factor
of 35 percent was added to this estimation in a later supplement to the report. Therefore, the cost
of retrofitting low NO burners using this procedure could be estimated using the formula I = 4160
(Q)0'6, where Q is in MW (or I = 1990 (Q) ' , where Q is in 10 Btu/hr). However, by regression
analysis of the retrofit cost and heater size presented in the table, a scale up factor of 0.96 is
calculated with a correlation coefficient of 0.85. Using this analysis with a contingency factor of
* 0 96
35 percent, the retrofit cost could be estimated using the formula I = 1500 (Q) .* where Q is in MW
(or I = 456 (Q)0'95 where Q is in 106 Btu/hr). It should be realized that because of the vari-
abilities in the cost of retrofitting low NOX burners, either one of these estimation methods may.be
off by a factor of two or three.
5-65
-------
TABLE 5-13. RETROFIT COSTS FOR LOW NOX BURNERS
Refinery
Beacon
Chevron
Coastal
Petroleum
ECO Petroleum
Golden Eagle
Kern County
Newhall
Newhall
Texaco
Average
Size of
Furnace
(106 Btu)
31.29
90
21.7
27
25.7
.20.6
5
5
40
.286.29
9
= 31.8
Cost of
Burners-
Dollars
(Year)
10,400
(1978)
120,000
(1981)
10,000
(1979)
4,360
(1977)
4,800
(1979)
5,800
(1979)
4,800
(1979)
2,220
(1979)
8,640
*
Retrofit
Cost-
Dollars
(Year)
4,000
(1978)
80,000
. (1981)
11,000
(1979) .
17,660
(1977)
10,000
(1979)
20,000
(1979)
9,320
(1979)
4,660
(1979)
13,820
Total
Cost-
Dollars
(Year)
14,400
(1978)
200,000
(1981)
21 ,000
(1979)
22,020
(1977)
14,800
(1979)
25,800
(1979)
14,120
(1979)
6,880
(1979)
22,460
Total
Cost
(1981
Dollars)
19,645
200,000
26,261
32,204
18,508
32,263
17,657
8,604
32,848
Total
Annual 1 zed
Cost (1981
Dollars)
5,375
54,720
7,185
8,811
5,064
8,827
4,831
2,354
8,987
106,154
9
= 11,795
-------
Selective Non-catalytic Reduction - The use of selective non-catalytic reduction requires the
installation of the system and controls to inject the ammonia at the proper location in the heater
or the ammonia and hydrogen if the temperatures in the heater are lower than required, and a
hydrogen and/or an ammonia storage tank. The cost of one refinery to retrofit the system, controls
and a large ammonia storage tank to control NO emissions from a 9.4 MM (32 xlO Btu/hr) heater were
$83,000 in 1980. Of this $83,000 installation cost, $20,000 was attributed to installing an ammonia
tank. To estimate the retrofit costs for other facilities, the authors of the CARS report made the
assumption that an ammonia storage tank would not be required at all facilities and that the system
and controls can be estimated using a scale-up factor of 0.6. The resulting formula to estimate-the
capital cost of a SNCR system is I = 16420 (Q)0-6, where Q is in MW (or I = 7860 (Q)0'6, where Q is
in 10 Btu/hr.) The annual cost of controlling NOX with SNCR was made by assuming a capital
recovery factor of 0.274; maintenance, labor and spare parts cost of 3 percent of capital cost; and
a plant overhead cost of 25 percent of maintenance and labor. Therefore, excluding annual costs of
ammonia, steam and hydrogen, the annual costs would be about 0.31 of the total capital cost. The
annual cost of ammonia, steam and hydrogen were estimated using the following additional
assumptions: (!) an ammonia to NOX mole ratio of 1.5 to 1.0 at a retail cost.of ammonia of $0.25 per
pound (wholesale costs of $0.125 per pound could be used'if 2 large storage tanks were installed),
(2) a steam requirement of 12 pounds per 10 Btu of fuel used at a cost of $3.50 per 1000 pounds of
steam, and (3) a hydrogen to ammonia mole ratio of 1.0 to 1.0 at a cost of $1.10 per pound of
hydrogen.
Selective Catalytic Reduction - Estimating the costs of retrofitting an SCR system to reduce
NO emissions is more difficult than predicting the costs of SNCR or low NO burner retrofits. Some
of the reasons for these difficulties include greater site characteristic dependencies; whether
reheat before the catalyst bed is required; and other contingencies such as furnace down time
requirements, frequency, type, and amount of liquid fuel fired in the furnace, start up costs, and
vendor guarantee costs. .
One refinery installed a selective catalytic reduction sytem using in-house labor on a new 14.2
MW (48.5 x 106 Btu/hr) boiler. The capital cost of this installation was $333,000 in 1980. By
using a factor of 1.25 to account for contracting the installation, adding a 35 percent contingency
factor, and including a site factor of 2.2 when reheat is required or 1.5 when reheat is not
necessary, a formula to estimate the capital cost of an SCR system for heaters of different sizes
was developed. The formuVa for a unit requiring reheat is: I = 255,400 (Q) + 3170 Q, where Q is
5-67
-------
the rated heat Input 1n MW (or I = 122,000(Q)°'6 + 9300 Q, where Q is in 106 Btu/hr). The formula for
a unit that does not require reheat is: I - 174,000(Q)°'6 + 31700 Q, where Q is the rated heat input
1n MW (or I « 83,000(Q)0>6 + 9300 Q, where Q is in 106 B.tu/hr). The second term in these equations is
the'cost of the catalyst, which is estimated to last about two years, Annual operating costs can be
estimated by multiplying the capital cost of the system by the factor specified before, which covers
the capital recovery factor; maintenance, labor and spare parts requirements; and plant over-head
requirements. In addition, the cost of ammonia which must be added at a NO,, mole ratio of 1 to 1,
> x
catalyst replacement every two years, steam costs at a rate of 0.5 Ib per 10 Btu fired, and
electrical costs at 0.25 KM per 106 Btu fired must be included.
5.3.1.2 Catalytic Crackers and CO Boilers
5.3.1.2.1 Process Description
A fluid-bed catalytic-cracking unit is often an integral part of a modern refinery. Preheated
gas oil is charged to a moving stream of hot regenerated catalyst while it is being transferred from
the regenerator to the reactor. The gas oil" is cracked in the reactor or the tube inlets to the
reactor; the products then pass through cyclone separators for removal of entrained catalyst and are
cut into products in a fractionator. Coke forms on the catalyst during the reaction.
Spent catalyst is withdrawn from the bottom of the reactor and transferred to the regenerator
where coke is burned off. The regenerator flue gas passes through cyclone separators for catalyst
removal and is discharged through the stack. The hot, regenerated catalyst flows back to the
reactor, supplying heat and catalyzing the cracking reaction.
The regenerator flue gas contains from 6 to 12 percent carbon monoxide. This gas is sometimes
fed to a CO boiler where it is burned in preheated air to generate steam. Auxiliary fuel is
required to maintain satisfactory combustion conditions and to allow variable firing rates to meet
the refinery steam demands.
5.3.1.2.2 Emissions and Control Technology
KOX is also released from the catalytic-cracking regenerator and from CO boilers, which are
fired by the catalytic cracker off gas. Emission testing in CO boiler stacks, summarized in Table
5-14, has shown results ranging from 100 ppm to 230 ppm of NO . Hunter (Reference 5-34) reported
5-68
-------
TABLE 5-14. NOX EMISSIONS FROM PETROLEUM
REFINERY CO BOILERS (REFERENCE 5-33)
Investigator
NOX
(ppm as measured)
Schulz, et. al., (Reference 5-67)
Schulz, et. al., (Reference 5-68)
Shea (Reference 5-71)
Shea (Reference 5-69)
Cowherd (Reference 5-70)
104-116
(average 106)
70-89
(average 78)
96-233
(average 163)
101-159
(average 135)
108-162
(average 129)
5-69
-------
testing one CO boiler that was equipped with staged air ports. Baseline emissions were 126 ppm.
Lowering excess oxygen from 2.1 to 1.8 percent reduced NOX by 8 percent. Adjustment of the air
ports and BOOS has negligible effect on NO emissions. CO emissions, however, were very sensitive
to excess air and increased rapidly below about 2 percent excess oxygen, the lack of response of
N(L to combustion modifications is attributed to NOV that is formed from ammonia in the CO gas feed
x x
acting similarly to fuel nitrogen*in oil or coal.
The average emission factor for NO from fluid catalytic cracking units is estimated in
Reference 5-23 as 0.24 kg N02/liter feed (84.0 lb/103 bbl feed). The total nationwide annual
emissions from fluid bed and thermal cat crackers was estimated to be 45 Gg (50,000 tons) in 1974.
If the regenerator exhaust is burned in a CO boiler , the resulting NOX emissions can presumably be
controlled by the classical methods discussed in Section 4.2 of this report.
5.3.2 Metallurgical Processes
5.3.2.1 Process Description and Control Techniques
The iron and steel industry is the predominant source of NO emissions derived from metallur-
gical processes. Other industries, such as aluminum production, extensively use electric melting
furnaces or operate the process equipment at temperatures below the minimum required for formation
of significant quantities of NOX. Copper, lead, and zinc smelting require combustion operation in
the reverberatory furnaces and converters (copper) and in sintering machines (lead and zinc). These
combustion emissions are deemed insignificant relative to the emissions from the iron and steel
Industry. Emissions from these other industries may become'significant as a result of the trend
toward higher melting rates in new equipment designs. This section reviews the equipment types and
available NOX control technology for the major sources of NO within the iron and steel industry.
Section 5.3.2.2 summarizes NO emission factors for these equipment types. Major portions of this
section are taken from a 1976 I6T study (Reference 5-33) which uses 1971 steel industry data as a
source for fuel consumption and NO emissions estimates.
Palletizing
Pelletizing of extremely fine low grade iron ore occurs in a specially designed furnace at
or near the iron mine. The cost of shipping the unbeneficiated ore would be almost double that of
the palletized product.
Previous studies by the Institute of Gas Technology have shown that pelletized ore production
will be about 54 Tg per year (60 million tons/yr) by 1985. The fuel consumed by the pelletizing
*
furnaces has remained about constant at 0.7 MJ/kg (600,000 Btu/ton). This indicates that annual NO
emissions from palletizing furnaces will reach about 7.65 Gg (8,500 tons) by 1985. The steel industry
5-70
-------
and equipment builders are considering coal firing the palletizing furnace combustion chambers. If
this is done, it will probably bring about an increase of about 50 percent in NO emissions. There
is no information available concerning NO control techniques for pelletizing furnaces (Reference 5-33).
Sintering
Some of the iron ore and flue dusts are available in particle sizes too small to be charged
directly to the blast furnace. These particles are mixed with flux and coke breeze and loaded onto a
traveling grate-sintering machine. An auxiliary fuel such as natural gas, coke oven gas, or oil is
used to initiate combustion on the surface of the mixture and is referred to as ignition fuel. Com-
bustion is continued over the length of travel by forcing air through the mixture on the grates.
The mixture is heated to a fusion temperature, which causes agglomeration of the iron-bearing par-
ticles. The discharged sinter is cooled, crushed, and screened prior to transfer to the blast fur-
nace charging oven.
The major source of energy used in th« production of sinter is the carbon content of coke
breeze and flue dust. The amount of ignition fuel required is about 140 J/g (0.12 million Btu per
ton) of sinter. The total fuel requirement, including coke breeze, is about 1.74 kJ/g (1.5 million
Btu per ton) of sinter.
The use of sinter machines to agglomerate ore fines, flue dust, and coke breeze has been
declining since 1966 and amounted to 39 Tg (43 x 106 tons) in 1971. If the present rate of decline
continues, the 1985 production of sinter would be about 24.3 Tg (27 x 106 tons). The attitude of
the steel industry is mixed because many steel plants are phasing out sinter lines, while at least
one major producer has replaced several small sinter lines with a large machine designed to meet
pollution control regulations. On the other hand, the use of sintering for recycling iron has
simultaneously been increasing. Therefore, the projected decrease in the number of sinter machines
may not occur. In any case, the I6T estimates (Reference 5-33) show that NO will continue to be a
major pollutant. There is no information available concerning NOX emission control techniques for
these furnaces.
Blast Furnace
The blast furnace is the central unit in which iron ore is reduced, in the presence of coke
and limestone, for the production of pig iron. The blast furnace itself is normally a closed unit
and therefore has no atmospheric emission. A preheated air blast is supplied to the furnace from
the blast furnace stove, through nozzle-like openings called tuyeres. The subsequent reactions in
the blast furnace are not pertinent to this discussion. Excellent descriptions are available, how-
ever, such as the complete discussion of the process of changing raw ore to finished steel published
by the United States Steel Corporation (Reference 5-72).
5-71
-------
The hot blast reacts with the coke to produce heat and more carbon monoxide than is needed to
reduce the ore. The excess CO leaves the top of the blast furnace with other gaseous products and
particulates and is known as blast furnace gas. This gas is cleaned to remove the particulates, which
could .later cause plugging. It is then available for heating purposes. Blast furnace gas contains
about one percent hydrogen and 27 percent carbon monoxide; it has a heating value of approximately
3600 kJ/Nm1, or, 92 Btu/ft* (Reference 5-72).
Coke Ovens
Coke is an essential component in making pig iron and steel; coke ovens are generally an
Integral part of the steel plant complex. One-sixth of the total bituminous coal produced is charged
to coke ovens. On the average, 1.4. kg of coal is required for each kilogram of coke produced.
Conventional coking is done in long rows of slot-type ovens into which coal is charged
through holes in the top of the ovens. The sidewalls, or liners, are built of silica brick, and the
spaces between the chambers are flues in which fuel gas bums to supply the required heat. Each
kilogram of coal carbonized requires 480 to 550 kJ (450 to 520 Btu). Flue temperatures are as high
as 1.753K or 2.700F (Reference 5-73). Much of the remaining heat in the partially spent combustion
gases 1s accumulated in a brick checkerwork, which releases it to preheat the combustion air when
the cycle is reversed. This is a typical regenerative cycle to conserve fuel and give a higher flame
temperature.
The coal in the coking chambers undergoes destructive distillation during a heating period of
about 16 hours. The noncondensable gaseous product is known as coke oven gas and on a dry basis
has a heating value of about 22 MJ/Nn3 (570 Btu/ft3). Approximately 35 percent of the coke oven gas
produced is used in heating the oven.
The major sources of emissions from coke ovens are the rapid evolution of steam and other
gases when moist coal is charged, the discharge of gases and particulates from the charging openings
during charging, and the emissions during the coke push and subsequent quencing. Recent coke-oven
battery designs have reduced the emissions from charging and pushing by using advanced engineering
features and improved operating procedures. During the coking process, leakage from the push side
and coke side door seals can account for most of the emissions during the coking process itself.
Improved door sealing techniques reduce door leakage substantially.
Although the current practice of firing coke ovens with a mixture of blast furnace gas and
coke-oven gas and slow mixing in the combustion chambers should tend to minimize NOX production,
the estimated total is substantial because of the large quantity of fuel consumed.
The reduction in the coke required per kilogram of hot metal achieved during the 1960's will
continue, but steel mills are currently installing new coke ovens because of the increased need for
5-72
-------
hot metal due to the high BOF*hot metal-scrap ratio. It is believed that the decline in coke rate
may have been stopped by the increased cost of fossil fuels used as injectants. The 1985 projection
for coke-oven underfiring fuel is 485 PJ (458 trillion Btu). If the NOX concentration remains con-
stant, the resulting total, emissions of NOX will reach 57.8 Gg (64,120 tons) per year.
Although it may be reasonable to assume that substitution of form coke may result in a sub-
stantial reduction in NO production, the general -opinion in the steel industry is that form coke
will not be a significant factor in 1985 (Reference 5-33).
Blast Furnace Stove
Between 2.2 and 3.5 kg of blast furnace gas is generated for each kilogram of pig iron pro-
duced. Some 18 to 24 percent of this gas is used as fuel to heat the three stoves which are usually
associated with each blast furnace. Two are generally on heat while the third is on blast.
The blast furnace stove is a structure about 8 to 8.5 m (26 to 28 feet) in diameter and
about 36 m (120 feet) high. A roughly cylindrical combustion chamber extends to the top of the
structure and the hot combustion gases pass through a brick checkerwork to the bottom by reverse
flow and then to the stack. The checkerwork usually contains 25,500 m2 (275,000 ft2) of heating
surface and has about 85 percent thermal efficiency. Unlike the conventional regenerators, which
extract heat from the waste combustion gases, the blast furnace stove is heated by burning fuel.
The stored heat is then used to preheat air for the combustion of fuel in the furnace to be served.
As in the case of coke oven underfiring, the blast stoves require very large quantities of
fuel for heating. However, since the stoves are heated primarily with blast-furnace gas (3.0,to
3.5 MJ/Nm3, or 80 to 95 Btu/ft3) the NOX concentration is lower due to the presence of diluents and
a low flame temperature.
The projected need for hot metal in 1985 is 112 Tg (124 million tons). This amount of hot
metal will require 295 PJ (280 trillion Btu) for blast-stove heating. Assuming no reduction in NOX
stack-gas concentration, the NOX emission.in 1985 will be 17.7 Gg/yr (19,600 tons/yr). Because of
the low estimated NO concentration and the presence of inerts in the fuel gas, equivalent to flue-
gas recirculation, the potential for NO reduction is probably small (Reference 5-33).
Open Hearth Furnace
Steel making by the open hearth process has been decreasing since it reached a peak in 1956,
when it represented 90 percent, or 92.7 Tg (103 million tons), of the total production. The use of
open hearth furnaces is expected to continue to decline and will probably amount to about 10 percent
of total steel production by 1985. Regardless of this dramatic decline due to the inroads of the
basic oxygen furnace (BOF) and electric arc furnace steelmaking processes, its NOX emission poten-
tial deserves consideration.
Basic Oxygen Furnace 5-73
-------
The optn hearth furnace 1s both reverberator? «nd regenerative, like tht glass melting fur-
naces It 1s reverberator? In that the charge is wlted 1n « shallow hearth by heat fro* a flaw
passing over the charye and by rac'latlon frcw the hefted dome. It Is regenerative 1n that the
remaining heat 1n the partially spent combustion gases from the reverberatory chamber 1s accumulated
In a brick filled chamber, or "checker", and released to preheat the Incoming com&ustlon air when
the cycle 1s reversed. Fuel of low calorific value such as blast furnace g«s as well as the com-
twitlor. air may be preheated by the checkers in order to obtain the high temperatures required.
Hot metal from the blast furnace, pig Iron, scrap Iron, and 11me are the usual materials
charged to an or>en hearth furnace. These are heated over a period averaging 10 hours, at a tempera-
ture as high as the refractories will permit. Fuel of! Is the preferred fuel and 1s burned with
excess *1r to provide an oxidizing Influence on the charge.
NOX emissions from open hearth furnaces ere very high because of the high combustion air- pre-
heat temperature, high operating temperature, and the use of oxygen Unces to Increase production
rates. The data avullable Indicate that NOX concentrations will be In tie 1000 to 2000-ppm range.
Although many open hearths ere being phased out because of emission control difficulties and better
economics of steel production with the BOF process, several steel mills are modernizing open hearth
shops, Including pollution control equipment to provide flexibility 1n the hot metal-scrap ratio,
particularly those mills with a hot-metal deficiency. Therefore, predictions that the open hearths
will be phased on entirely by 1985 are unrealistic, and 1t 1s anticipated that about 13.5 Pa. (15
million tons) will still be trade by the open hearth process in 1985. Fuel consumption has been
decreasing and may reach 2.9 MJ/kg (2.5 million Btu/ton) 1n 1985. This will require a fuel con-
sumption of 40 PJ (37.5 trillion Btu) for open hearth steel production and result in an NO emis-
sion level of 14 Gg (15,750 tons) (Reference 5-33).
Basic Oxygen Furnace
In the basic oxygen furnace (BOF), oxygen 1s blown downward through a water-cooled lance Into
a bath containing scrap and hot metal. Heat produced by oxlditlon of carbon, silicon, manganese, and
phosphorous Is sufficient to bring the metal to pouring temperature and auxiliary fuel 1s not required.
The furnace 1s an open top, tillable, refractory-lined vessel shaped somewhat like the cld-fashloned
glass nrilk bottle. Furnace capacities range up to 309 Mg (340 tors). The time required per cycle Is
very short - from 45 to 60 minutes.
The BOF has displaced the open hearth as the major steel production process, but is much less
flexible because of the Inherent limitation of 25 percent to 30 percent scrap in the charge. The
5-74
-------
mount of BOF capacity 1n an Integrated $te«l plant 1s, therefor*, closely associated with hot Mtal
availability. Acdltlonol flexibility In scrap use can be obtained by preheating the scrap with an
oxygen-fuel burner. In many steel plants, the open hearth shop is modernized and equipped with
appropriate pollution control equipment so that 1t can be used In conjunction with BOF shops to
provide the required flexibility to accommodate variations 1n hot metal-scrap ratio. A combination
of BOF shops and eiectrlc furnace shop't provides the maximum 1n flexibility and nay re^esent the
makeup of future steelmakliig facilities.
Excluding fuel use for scrap preheating, other uses are for refractory dryout and to keep
the BOF vessel from cooling between heats. Their us&s amount to about 232 kg per kg (200.000 Btu
per ton) of steel produced.
DecarbuHzatlon of the Iron charged to the BOF produces tbout 467 kJ of carbon monlxlde per
kilogram of steel (400,000 Btu/ton). The off-gases also contain large amounts of participates,
wMch must be removed before discharge Into the atmosphere. Typical American practice 1s tc bum
the combustible gases 1n w«ter-cooled hoods mounted above the BOF vessel, cool with excess air or
water sprays, and pass the cooled gases through h
-------
Existing fuel conservation measures in soaking-pit heating include improved scheduling so as
to charge at a higher ingot temperature, programmed input control, improved burner designs, air/
fuel ratio control responsive to stack-gas oxygen content, addition of recuperators to existing
cold combustion air installations, and use of recuperators designed to give higher preheat tempera-
ture. Of these, the use of high-mixing-rate burners and an increase in combustion air preheat are
likely to increase the NO- emission level. At the present time, only experimental information is
available concerning the effect of these parameters on NO levels.
Soaking-pit and reheat-furnace operating temperatures are such that the estimated NO levels
should fall in the 250 to 350-ppm range. However, the very large amounts of fuel used result in a
total NOX output estimated at 97 Gg (107,000 tons) in 1971.
A major factor that will reduce consumption of purchased and in-plant fuels and thereby de-
*
crease NOV output is the trend toward use of continuous casting to replace some ingot casting. In
A , ' ' '
this process, billets and slabs which are hot-rolled prior to cooling are produced from molten
steel, thus eliminating soaking-pits and most of the reheat requirement. About 20 percent of total
steel production, or 36 Tg (40 x 106 tons), is estimated to be produced by continuous casting in
1985. In spite of this, soaking-pit and reheating"furnace steel capacity will have to be increased
during the 1975 to 1985 period to provide for the expected growth in steel production and for the
steel which for process reasons will have to be cast in ingots. According to the I6T projection,
conventional steel processing will account for 144 Tg (160 x 10s tons) in 1985. At present fuel
consumption of 5.4 MJ/kg (4.7 x 10s Btu/ton), the total fuel consumed for soaking-pits and reheat
furnaces in 1985 will be 795 PO (750 x 1012 Btu). This fuel consumption will result in estimated
HOV emissions of 143 Eg (157,900 tons).
" ,
Heat Treating and Finishing Operation
This category includes annealing, hardening, carburizing and normalizing of some of the
steel industry cold-rolled products, as well as production of coated products. Fuel consumption
In 1971 was about 632 PJ (600 x 101! Btu) for the production of cold-rolled products (about 25
percent of total steel production). NOX emission levels are assumed to be in the 150 to 250-ppm
range. On this basis, total NOV emission in 1971 for this category will be about 7.6 Gg (8,400
A
tons). Assuming that production of cold-rolled products remains at about 25 percent of total steel
production, the 1985 NOX emission will amount to 10 Gg (11,200 tons) per year. There is no informa-
tion available concerning NOX control techniques for these sources (Reference 5-33).
5-76
-------
Electric Furnaces
Production of steel in electric-arc furnaces has grown rapidly since World War II and is
currently estimated to be about 20 percent of total steel production. Because of the phase out of
open hearth steelmaking, the increase in BOF steel production,-and the associated scrap-use limita-
tion, the amount of steel produced in electric-arc furnaces is expected to increase even more.
The combustion of fossil fuels currently plays a very small role in electric steelmaking. ,
This may change in the future as advances in technology permit the increased use of scrap preheating.
Most authorities agree that scrap preheating will be accomplished outside the electric-arc furnace
in a specially designed charging bucket, probably equipped for bottom discharge. Many of the designs
use excess air burners to limit flame temperature and minimize oxidation of the scrap. Associated
air-pollution problems include particulates from dirty scrap, iron oxide, and oil vapors. The
requirement for both incineration at or above 1.033K (1.400F) and particulate removal has caused
shutdown of several scrap preheating installations because of economic considerations.
The use of electricity for heat in steel production transfers the NOX emissions to the
utility plant where the problem is easier to control. Electric furnaces are, in any case, a very
minor source of NO from the steel industry (Reference 5-33).
5.3.2.2 Emissions
Emissions in the steel industry and its related processing have historically consisted of
fumes, smoke, -and dust or particulates. The gases usually considered obnoxious have been SOg, CO, '
and odors. The presence of oxides of nitrogen has been obscured by the heavy emission of particu-
lates and a resulting lack of physical evidence. The NO emissions observed can be traced largely
to the combustion of fuel oils and gas and, in part, to the burning of carbon monoxide, which is a
product of the processing operations.
The emission of nitrogen oxides from iron and steelmaking and processing equipment does not
appear to have been extensively investigated. However, reasonable estimates can be made by assuming
a relationship between known operating temperatures and NOX concentrations in stack gases (Reference
5-33). This relationship is affected by other variables, such as combustion air preheat temperature
and oxygen enrichment of combustion air.
Table 5-15shows the estimated NO concentrations for the major energy-intensive processes
or
energy consumption data (Reference 5-33).
and the resulting total annual combustion-related NOX production based on 1971 steel production
5-77
-------
TABLE 5-15. ESTIMATED NOX EMISSIONS FROM STEEL HILL PROCESSES AMD
EQUIPMENT (Reference 5-28 except where noted)
Process
or
Equipment
Pelletizlng
Sintering
Blast Furnace
Coke Oven
Blast Furnace Stove
Open Hearth Furnace
Basic Oxygen Furnace
Soaking Pit and
Reheat Furnaces
Heat Treating and Finishing
Electric Furnaces
Annual Fuel Consumption
PJ
29
98
; ndb
225
i 212
135
; nd
541
64
; nd
10" Btu
27
93
nd
212
200
127
nd
510
60
nd
NO Emission Factors3
A
ppm
(as measured)
300
500
' (230)e
negc
200
(10-485)6
100
600
nd
300
(92)e
200
6-25
ng/J
180
300
neg
120
60
360
nd
180
120
nd
Ib/lO* Btu
0.42
0.70
neg
0.28
0.14
0.84
nd
0.42
0.28
nd
Annual NO^ Emissions3
Gg tons
5.1 5670
29.6 32550
neg neg
26.9 24680
12.6 14000
48.3 53340
nd nd
96.4 107100
7.6 8400
.02-. 09 26-110
Notes: expressed as NO,
l L-
nd = no data
neg « negligible emissions
Reference 5-75
American Iron and Steel Institute test data provided by Dr. Walter Jackson (U.S. Steel), (Reference 5-74)
-------
Other test results provided by the American Iron and Steel Institute (Reference 5-74} indi-
cate different emission factors as shown in parentheses in Table 5-15-. The emission levels for the
coke ovens are the result of three separate tests (10, 186, and 485 ppm).
Results of recent tests reported by Hunter, e_t a_l_. (Reference 5-34) are summarized in Table
5-16. The open hearth furnace was tested while operating on natural ;as and Number 6 fuel oil
(50/40). The wide fluctuations in NO and CO observed as various operations were performed are
A
shown in Figure'5-18. Large changes in excess air occurred as the operators opened doors to look at
the steel and to add material or adjust fue'l flow to change heating rate. NOX emissions varied
from 100 to 3500 ppm and averaged about 1800 ppm or about 950 ng/J (2.2 Ib/MMBtu). NO increased
somewhat linearly with excess 02- ParticuTate emissions were 2200 ng/J (5.02 IbMMBtu), measured
upstream of the precipitator. Following baseline tests the furnace was overhauled to repair refrac-
tory and fix leaks. A second test cycle was observed on the repaired furnace and the average NO
*v
was 1094 ng/J (1250 ppm), a reduction of about 40 percent. During baseline tests, N0x frequently
exceeded 2000 ppm but with the excess air controlled, excursions over 2000 ppm occurred only twice.
One steel -billet reheat furnace was tested while firing natural gas at heat rates between 13
and 30 MW. Baseline NOX emissions at 24 MW (82 million Btu/hr) were 56 ng/J (110 ppm) and particu-
lates were 17 ng/J (0.04 Ib/MMBtu). This furnace had two heating zones with 13 and 14 burners,
respectively. The row with 13 burners released about 80 percent of the heat input. Combustion
modifications included reduced excess air, resulting in a 24 percent NO reduction, and burners out
of service which produced a 43 percent NO reduction with three burners out of service in the row
of 13 burners.
One steel ingot soaking pit was tested (site 16/2) while firing natural gas at about 2.9 MW
(10 MHBtu/hr) through a single burner. Baseline NOX emissions at 2 MW were 52 ng/J (101 ppm).
Reduction of excess air reduced NOX fay 69 percent with no adverse effect on the steel.
5.3.3 Glass Manufacture
5.3.3.1 Process Description
.The glass manufacturing industry is made up of several basically different types of opera-
tions. They are:
5-79
-------
TABLE 5-16. EFFECTS OF NOX CONTROLS ON STEEL INDUSTRY
NOX EMISSIONS (Reference 5-34)
in
&>
O
Device
Type
Steel Open Hearth
Furnace.
Steel Reheat
Furnace
Steel Soaking Pit
Fuel
Nat. Gas +
No, 6 Oil
Nat. 'Gas
Nat. Gas
Average Baseline
N0x
ng/J
1094
56
52
ppm
@ 3$ 02
2070
no
101
Max.
Ctf
/o
Reduction
40
43
69
Combustion
Modification
Low 00-
3/27 BOOS
Low 02
-------
4000
3500
3000
2500
cs
9
»
CM
a
&
Ui
u
cc
vu
a.
2000
1500
1000
BASELINE TEST
N0.6QILAND GASFliEL
SOO
1200
130G
1400
1500 1600 1700
TIME OF DAY, hr
1800
1SQO
2000
Figure 5-18. NOX emissions as a function of time for an
op£n hearth furnace (Reference 5-34}.
s-ai
-------
1. Glass container manufacture
2. Fiberglass manufacture
3. Flat glass manufacture
4. Specialty glass manufacture
The largest type is the glass container Industry, which produces about 45 percent of the total
amount of glass (by weight) produced by the entire industry.
While the specific processes used within each segment of the industry vary according to the
product being manufactured, glass manufacturing involves three major energy-consuming processes:
melting the raw materials, refining the molten glass, and finishing the formed products. Typically,
about 80 percent of the energy consumed by the glass industry is for melting and refining, 15 per-
cent is for finishing, and 5 percent is for mechanical drives and conveyors. The primary differ-
ences in processes used among the various segments occur in the refining and finishing operations.
The raw materials used in glass manufacture consist'primarily of silica sand, soda ash, lime-
stone, and cullet (crushed waste glass). In the production of window and plate glass, for example,
temperatures in the range of 1.783K to 1,838K (2.750F to 2.850F) may be required to melt these raw
materials into a viscous liquid.
The furnaces used are of the pot type if only a few tons of a specialty glass are to be pro-
duced, or of the continuous tank type for larger quantities. By far the larger amount of glass is
melted in furnaces, and only these will be considered in connection with NOV control.
A
*
Continuous reverberatory furnaces have a hoi ding'capacity of up to 1.27 Gg (1,'400 tons) and a
daily output of as much as 270 Hg (300 tons). Reverberatory furnaces in this industry are broken
into two classifications according to the firing arrangement used: end-port and side-port melters.
In the operation of a side-port-fired furnace, the preheated combustion air mixes with the fuel in
the port, resulting in a flame that burns over the glass surface. The products of combustion exit
via the opposite port, down through the c'heckerbricks, and. out through the reversing valve to the
exhaust stack. Typically, there are several ports situated along each side of the furnace. In
contrast, there are only two ports in an end-port-fired furnace, located on the rear wall of the
furnace. The flame is ignited in one port, travels out over the glass toward the bridgewall, and
"horseshoes" back to the exit port - the other port in the rear of the furnace. In both types of
furnaces, the firing pattern is reversed every 20 to 30 minutes, depending upon the specific furnace.
During this reversal period, the flame is extinguished, the furnace is purged of combustion gases by
reversing the flow of combustion nir and exhaust gases passing through the reversal valve, and
5-82
-------
combustion 1s then reestablished in what was previously the exhaust port. Both types of me Hers are
operated continuously throughout a campaign that normally lasts 4 to 5 years, at sustained tempera-
tures up to 1.867K (2,900F).
In addition to the reverberatory-type melters, day tanks, unit melters, and pot melters are
used, mostly in the pressed and blown glass industry. Many of these melters are batch-type, as
opposed to continuous, resulting in a substantial reduction in fuel-utilization efficiency. Much
of the fuel that is wasted is due to the antiquated methods of operation and associated equipment
used with these melters (Reference 5-33).
The combustion gases, on leaving the melting zone, retain a considerable amount of heat. This
is reclaimed in a regenerator or brickchecker chamber. When the firing cycle is reversed; combus-
tion air is preheated by being passed through the brick work. Preheating saves fuel but increases
the flame temperature which promotes NOX formation.
Coal is not used in glass melting. Since molten glass is conductive, electrical heating is
used as a booster to supplement fuel firing whenever technically and economically practical. Gas
and, to a lesser extent, fuel oil are the preferred fuels.
5.3.3.2 Emissions
The flue gas from glass-melting furnaces 1s the major source of NO emission 1n the glass
industry. The operation of these furnaces is similar to that of open hearth furnaces used in steel-
making; regenerative checkerwork sets absorb heat from the combustion gases for subsequent release
to the incoming combustion air. This is accomplished by a reversing valve which puts each checker-
work set through its heating and cooling cycle In turn. The sequence of intense high-temperature
combustion and quenching in the checkerwork sometimes raises NO emissions to levels higher than
those experienced in a steam boiler of equivalent heat release. For example, during a recently
completed experimental program, NO emissions were measured during a complete firing cycle of a
glass meltef. NOX emissions were highest at the beginning of the firing cycle and then, as the
cycle continued, decreased by about 30 percent. At the beginning of the firing cycle, the combus-
tion air is preheated to a higher temperature, which results in a hotter flame than at the end of
the cycle when the checkerbrick and hence the air have cooled considerably. Other major factors in
NOX formation in a glass melter, such as flame velocity and recirculation patterns of flue gases,
are being studied.
Table 5-17 summarizes the emissions from several glass melters as measured by a number of
investigators.
5-83
-------
CO
CO
s:
o
a: co
u. LU
CO et
z: z
o a:
HH Z3
co u.
LU l-H
xti
O LU
t
LA
LU
CQ
C/)
4«^
C
g
Is
o
0
'a"
cu
s_
3
CO
. (O
xai
f^ S
*s^
CO
re
g
Q.
o.
i-
o
(O
Cft
1
M
CO
CU
C
1 1
&-
re
CO
CO
a;
o
X
cu
^s>
oo
««i
o
cu
s.
*^»
«*-
CO
re
en
o
o
'"
0
4-»
0
CT>
*+
o>
1
LO
(U
a
c:
£
.«
fr
a
02
j
3
O
c
C
3
S-
'Jo
CO
co
cu
u
X
LO
x_
?
LO
0}
^~+ U
vo c
r- a>
1 y^
LO CM
o
re
ai
>
'
^^.
oo
r*.
i
LO
CU
u
c
CD
S-
lf"»
CU
oc
c
o
CO
*cu
c
(O
o
5-84
-------
5.3.3.3 Control Techniques
According to representatives of the glass Industry, the efforts of the Industry to reduce
air pollutant emissions are severely hampered by the variations in regulations that exist from
state to state. This lack of uniformity requires that different solutions.to the problem be sought,
depending on the location of the specific plant. This, in turn, adds substantially to the cost of
pollution control. In addition, not only are,the regulations variable from one location to another,
but these regulations are constantly changing. As a result, very few air pollution control equip-
ment installations have been made on glass furnaces, and there is very little data available on
the effectiveness and cost of these devices.
In general, SOX, NOX, and particulates are the primary air pollutants from the glass manu-
facturing processes. The concern is primarily with the melting process because this is the largest
energy consumer and the major contributor to air pollutant emissions. The major pollution problem tit
the combustion process is NO emissions.
While the formation of NOV in the combustion process is not entirely understood, it is clear
A
that the goals of reducing NOY emissions and reducing energy consumption are seemingly at odds. NO
A A
formation is a temperature-related phenomenon; as temperature increases, NOX emissions increase.
On the other hand, increasing available heat to a process may result in increases in efficiency and
in temperature, which in turn increase NO emissions. Analysis of the process modifications under
.. *
consideration in the glass industry shows that there is a possibility of increasing NO emissions.'
If the implementation is carried out properly, however,'this need not occur.
Six recommended modification programs are listed in Table 5-18. The order of listing is
according to programs that afford the greatest potential for solving the problems in the shortest
period of time. The table also presents estimates of improvements that may be obtained, where such
estimates can reasonably be made (Reference 5-33). Cost data for these programs are not available
at this time. Two of the six recommendations are currently, being pursued by EPA/1ERL-Cincinnati.
5.3.4 Cement Manufacture
5.3.4.1 Process Description
The cement industry includes, all establishments engaged in the manufacture of hydraulic cement
(generic name: port!and cement), masonry, natural, and pozzuolana cements. This discussion is
limited to the production of Portland cement because it accounts for 95 percent of the total
5-85
-------
TABLE 5-18. RECOMMENDED PROGRAMS FOR REDUCING EMISSIONS AND ENERGY
CONSUMPTION IN THE GLASS INDUSTRY (REFERENCE 5-33)
Program
Expected Improvements in
Energy Consumption, %
Expected Improvements in
Air Pollutant Emissions
oo
.en
1. Develop current emission data
2. Raw batch pretreatment i.e.,
preheating and agglomeration
3. Oxygen enrichment
4. Augmentation of heat transfer
from flames e.g., burner
positioning
5. Use of low-temperature heat
for driving compressors
6. Development of submerged
combustion process
25-50
5-15
10-20
50
25% to 50% potential NOX reduction,
may reduce particulate in form of
batch carry-over
No effect on NOX, SOX, or particulates
Proportional NOX reduction
Will substantially reduce NOX
emissions
-------
cement manufactured in the United States, with the remaining 5 percent split among the other
types.
Raw materials used in the manufacture of portland cement consist of limestone, chalk or marl,
and seashells. These are combined with either clay, shale, slate, blast furnace slag, iron ore,
or silica sand, the end product is a chemical combination of calcium, sil'icon, aluminum, iron,
and other trace materials. The raw materials are first ground and blended together. Depending
upon which of the two processes is used, water may be added during blending (the wet process) or
the ingredients can be mixed on a dry basis (the dry process). In general, the moisture content
of the raw materials determines the process used. If the moisture content is greater than 18 per-
cent, by weight, the wet process will be used. If the moisture content is less than 18 percent,
the dry process will be used. The next step is the calcining or burning of the mixed raw material
in a rotary kiln. During this step, the material is heated to approximately 1,755K (2.700F) and
transformed into clinker, which has different chemical and physical properties than the raw
materials had initially. The clinker is discharged from the kiln and cooled. The last step is to
grind the clinker to the desired fineness and add gypsum to control the setting time of the concrete
(Reference 5-33).
5.3.4.2 Emissions
The major air pollutant emission problem in the manufacture of portland cement is particu-
lates, which occur in all phases of cement manufacturing from crushing and raw material storage to
clinker production, clinker grinding, storage, and packaging. However, emissions also include the
products of combustion of the fuel used in the rotary kilns and. drying operations; these emissions
are typically NO and small amounts of SO . For both the wet and dry kiln processes, the limited
data shows that nitrogen oxides are emitted at a rate of about 1.3 g per kg (2.6 Ib per ton) of
cement produced.
The largest 'source of emissions in cement plants is the kiln operation. At present, about
56 percent of the cement kilns in operation use the wet process, and 44 percent use the dry process.
Based on this information, estimates of total MQX emissions from cement plants in 1972 are 42.7 Gg
(4.7 x 10* tons) for the dry process and 54.5 Gg (6 x 10* tons) for the wet process. These estimates,
because of a lack of data, assume the use of no controls by the industry. Without an inventory of
control equipment in use, they cannot be refined.
Future efficiency-improving process modifications that increase flame temperature without
improving heat transfer to the process load will almost certainly result in increased NO emissions.
5-87
-------
Conversely, adequate removal of the additional heat resulting from the applicable process modifica-
tions should maintain HO emissions at their current level.
J\ :
Of the process modifications deemed to be near-term, only the use of oxygen enrichment has
any great potential of increasing air pollutant emissions, primarily NOX- In some applications in
other industries, for example, glass melting, oxygen enrichment can be used without increasing NO
emissions. However, due to the different type of load in the cement industry and the different
patterns of heat transfer, it is suspected that NO would increase with the implementation of oxygen
enrichment (Reference 5-33).
5.3.4.3 Control Techniques
/
There is very little information in the literature regarding commercial installation of equip-
ment for removing NOX from kiln waste gas or of modifications to kiln operations to reduce NOX
production. Water scrubbing is sometimes used for particulate removal from waste gas from lime
kilns. In this operation, the gas contacts a slurry of calcium hydroxide, which should remove a
50/50 mixture of NO and NO? and reduce NO up to 20 percent. Flue gas recirculation, which is used
Ct " i
to control temperature in some lime kilns, should reduce NO emissions by lowering flame temperature.
Reference 5-34 reports NO emission test results for both a dry process kiln and a wet pro-
cess kiln. The dry process kiln was tested at full capacity while firing a 68/32 mixture of coke
and natural gas. Data for the same kiln firing natural gas and oil separately were also available
for comparison. Emissions of NO while firing natural gas were 1,050 to 1,800 ng/J (1680 to 2900 ppm).
Operation on oil resulted in a 60 percent reduction (400-710 ng/J). Operation on combined coke
and natural gas produced emissions of 655 to 710 ng/J,. a 50 percent reduction.
I
Lower NOX emissions on solid and liquid fuels compared to gas are attributed to the highly
adiabatic nature of the process. Many cement kilns are currently being converted from gas to solid
fuels. This conversion wil-1 be beneficial in reducing NO and could be pursued as an NO control
A A
method that is consistent with the reduction of industrial gas consumption.
The wet process cement kiln was tested only while firing natural gas and had baseline
emissions of 1319 ng/J (2250 ppm). Combustion modifications Investigated included variation of
combustion air inlet temperature and excess oxygen. Increase of combustion air temperature from
644K (700F) to 767K (920F) increased NO emissions to 1518 ng/J, and 15 percent increase. Reduction
of excess oxygen at baseline air temperature reduced NO to 846 ng/J, a 36 percent reduction. The
I
Independent reductions of either excess air or air temperature caused unacceptable reduction of
kiln temperature that can result in a process upset. The NO emissions were found to be a strong
5-88
-------
function of kiln temperature, as shown in Figure 5-19. It was found that simultaneous reduction of
excess air and increase in air temperature could produce a reduction in NO of about 14 percent while
maintaining kiln temperature.
Electric heating eliminates all the pollutants associated with combustion sources, but its use
in kiln operation is very limited. Another means of emission control in kiln operation is the choice
of kiln type. Some NO reduction in limestone calcining is obtained by using a vertical instead of
a rotary kiln. The mechanism of operation is such that heat transfer to the load is very high, and
peak temperatures are lower than required to obtain the formation of NOX in large amounts.
5.3.5 Coal Preparation Plants
Coal in its natural state contains impurities such as sulfur, clay, rock, shale, and other
inorganic materials, generally called ash. Coal mining adds more impurities. Coal preparation plants
serve to remove these impurities. Coal cleaning processes utilized by coal preparation plants may be
wet, dry, or a combination of both. Wet processes are a minor source of oxides of nitrogen.
After the coal is wetted by the cleaning process, primary drying is done mechanically by
dewatering screens followed by centrifugal driers. When lower surface moisture is desired (3 to 6
percent) with finer coal sizes, secondary drying is required. Such low moisture levels can best be
accomplished by thermal drying. It appears that new coal preparation plants that install thermal
dryers will use a fluidized-bed type.
In the fluidized bed drier, hot combustion gases from a coal-fired furnace are passed upward
through a moving bed of finely-divided wet coal. As the bed fluidizes, the coal is dried as the
fine particles come into intimate contact with the hot gases.
The major po-llutant evolved from the thermal dryer is particulate. Well-controlled thermal
driers emit only minor quantities of NOX> Concentrations of 40 to 70 ppm (0.16 to 0.28 kg/MJ, or
0.39 to 0.68 lb/106 Btu) have been measured (Reference 5-81). These emission rates are below the
NSPS of 300 ng/J (0.7 lb/108 Btu) for large steam generators. In any case, no NOX standards have yet
been proposed since the thermal dryer capacities are generally less than the smallest power plants
required to control NOX emissions: 73.2 HW, .or 250 x 10s Btu/hr (Reference 5-81>.
5-89
-------
1422
30QO,
1500
KILN TEMPERATURE, °K
1600 1700
1800
1867
2500
^2000
ec
CM
O
1500
u
ec
1000
500
0
2100
SHADED AREA SHOWS EFFECT OF INLET AIR TEMPERATURE VARIATIONS
NUMBERS REPRESENT EXIT 02 CONCENTRATIONS IN PERCENT ~
ROTARY CEMENTKILN, WET PROCESS
2300
2500
KILN TEMPERATURE, °F
2700
2900
Figure 5-19. The effect of cement kiln temperature on NOX
emissions .(Reference 5-34).
5-90
-------
REFERENCES FOR SECTION 5
5-1 Castaldini, C., e£al., "Combustion Modification Controls for-Residential and Commercial
Heating Systems: VFTume I. Environmental Assessment," EPA-600/7-81-123a, July 1981.
5-2 Hall, R. E., 0. H. Wasser, and E. E. Berkau, "A Study of Air Pollutant Emissions from
Residential Heating Systems," EPA-650/2-74-003, January 19.74.
5-3 Barrett, E. R., S. E. Miller, and 0. W. Locklin, "Field Investigation of Emissions from
Combustion Equipment for Space Heating," Battelle-Columbus Laboratories, EPA R2-73-084a,
June 1973.
5-4 Hall, R. E., e£ aj_., "Status of EPA's Combustion Research Program for Residential Heating
Equipment," presented at the 67th APCA Annual Meeting, June 1974.
5-5 U.S. Environmental Protection Agency, "Compilation of Air Pollutant Emission. Factors,"
AP-42, February 1980.
5-6 Allen, J. M., "Control of'Emissions from Residential Wood Combustion by Combustion Modifi-
cation" Proceedings of the Joint Symposium on Stationary Combustion NO Control. Vol.
III.. IERL-RTP^1085, October 1980. x
5-7 Locklin, 0. W. and R. E. Barrett, "Guidelines for Residential Oil Burner Adjustments,"
EPA-600/2-75-069a, October 1975.
5-8 Locklin, D. W. and R. E. Barrett, "Guidelines for Burner Adjustments of Commercial
Oil-Fired Boilers," EPA-600/2-76-088, March 1976.
5-9 DeWerth, D. W., R. L. Himmel, and 0. W. Locklin, "Guidelines for Adjustment of Atmospheric
Gas Burners for Residential and Commercial Space Heating and Water Heating,"
EPA-600/8-79-005, February 1979.
5-10 Thrasher, W. H. and D. W. DeWerth, "Evaluation of the Pollutant Emissions from Gas- Fired
Air Furnaces," Research Report No. 1503, American Gas Association, Cleveland Laboratories,
Cleveland, Ohio, May 1975 as cited in Okuda, A. S. and L. P. Combs, "Design Optimization
and Field Verification of an Integrated Residential Furnace - Phase 1," Rockwell Inter-
national, Rocketdyne Division, EPA-600/7-79-037a, February 1979.
5-11 Kalika, P. W., G. T. Brookman, and J. E. Yocum, "A Study on Measuring the Environmental
Impact of Domestic Gas-Fired Heating Systems," Fipal Report, The Research Corporation of
New England, June 1974 as cited in Okuda, A. S. and L. P. Combs, "Design Optimization and
Field Verification of an Integrated Residential Furnace - Phase 1," Rockwell
International, Rocketdyne Division, EPA-600/7-79-037a, February 1979.
5-12 De Angelia, D. G. and R. B. Reznik, "Source Assessment: Residential Combustion of Coal,"
Monsanto Research Corporation, EPA-600/2-79-019a, January 1979.
5-13 Martin, G. B., "Evaluation of a Prototype Surface Combustion Furnace," Proceedings of the
Second Stationary Source Combustion Symposium, Volume III, EPA-600/7-77-073c,
NTIS-PB 271 757, July 1977.
f
5-14 Personal communication with Chuck Mueller, Amana, Inc., Amana, 10, April 14, 1979 as cited
in Reference 5-2.
5-15 Hall, R. E., et al., "Study of Air Pollutant Emissions from Residential Heating Systems,"
EPA-650/2-74-WTNTIS-PB 229697, January 1974.
5-16 Dickerson, R. A., and A. S. Okuda, "Design of an Optimum Distillate Oil. Burner for Control
of Pollutant Emissions," EPA-640/2-74-047, June 1974.
5-17 Comb, L. P. and A. S. Okuda, ".Residential Oil Furnace System Optimization - Phase I,"
Rocketdyne Division, Rockwell International, EPA-600/2-76-038, February 1976.
5-18 National Petroleum News, January 1975, pp. 34-35.
5-91
-------
5-19 Personal Communication, Lenney, R. J., Blueray Systems, Inc., Weston, Massachusetts,
September 1975.
5-20 Hall, R. E., "The Effect of Water/Distillate Oil Emulsions on Pollutants and Efficiency of
Residential and Commercial Heating Systems," Air Pollution Control Association Paper
75-09.4, June 1975.
5-21 Black, R. J., H. L. Hickman, Jr., A. J. Muchick, and R. D. Vaughan, "The National Solid
Wastes Survey: An Interim Report," Public Health Service, Environmental Control
Administration, Rockville, Maryland, 1968.
5-22 Niessen, W. R., et jil_., Systems Study of Air Pollution from Municipal Incineration, Report
to NAPCA under contract CPA 22-69-23, Arthur D. Little, Inc., Cambridge, Mass., 1970.
5-23 McGraw, J. J. and R. L. Duprey, Compilation of Air Pollutant Emission Factors (Revised),
AP-42, EPA, February W72.
5-24 Stenburg, R. L., et cil_., "Field Evaluation of Combustion Air Effects on Atomspheric
Emissions from Municipal Incinerators," J. Air Pollution Control Assoc., Vol. 12,
pp. 83-89, February 1962.
5-25 Kirsh, J. B., "Sanitary Landfill," In: Elements of Solid Waste Management Training Course
Manual Public Health Service, Cincinnati, Ohio, 1968, p, 1-4.
5-26 Fife, J. A., and R. H. Boyer, Jr., "What Price Incineration Air Pollution Control?,"
Proceedings of 1966 National Incinerator Conference, American Society of Mechanical
Engineers, flew York, 1966.
5-27 OAQPS Data File of Nationwide Emission, 1971. National Air Data Branch Monitoring and
Data Analysis Division, May 1973.
5-28 Chi, C. T., and D. L. Zanders, "Source Assessment: Agricultural Open Burning,
State-of-the-Art," EPA-600/2-77-107a, July 1977.
5-29 Air Pollution Problems from Refuse Disposal Operations in the Delaware Valley, Department
of Public Health, Air Management Services, Philadelphia, Pa., February 1969.
5-30 Wiley, J. S. et al_., "Composting Developments in the U.S.," Combust. Sci. 6(2)-.5-9, 1965.
5-31 Kurkey, C., "Reducing Emissions from Refuse Disposal," J. Air Pollution Control Assoc.,
19: 69-72, February 1969.
*
5-32 Personal Communication, Mr. Peter L. Cook, Office of Federal Activities, U. S.
Environmental Protection Agency, November 1977.
5-33 Ketels, P. A., J. D. Nesbitt, and R. D. Oberle, "A Survey of Emissions Control and
Combustion Equipment Data in Industrial Process Heating," Final Report by Institute of Gas
Technology for EPA, IGT Project No. 8949, June 1976.
5-34 Hunter, S. C., e£ al., "Application of Combustion Modifications to Industrial Combustion
Equipment, "ProceedTngs of the Second Stationary Source Combustion Symposium. Vol. Ill,
EPA-600/7^77-0736, July Wf!
5-35 Cantrell, A. Annual Refining Survey. Oil and Gas Journal. 79(13): 110-153. March 30,
1981. ~
5-36 Hunter, S. C. and S. C. Cherry. (KVB) NO Emissions from Petroleum Industry Operations.
(Prepared for the American Petroleum Institute.) Washington, D. C. API Publication
No. 4311. October 1979.
5-37 Radian Corporation, unpublished data. Source Category Survey Report on. NO Emissions From
Fired Heaters. (Prepared for Emission Standards and Engineering Division, U. S.
Environmental Protection Agency, Research Triangle Park, North Carolina.)
DCN 81-231-372-19-04, EPA Contract No. 68-02-3058. July 1981.
5-92
-------
5-38 Cherry, S. C. and S. C. Hunter. (KVB) Cost and Cost Effectiveness of NO Control in
Petroleum Industry Operations. (Prepared for the American Petroleum Institute.)
Washington, D. C. API Publication No. 4331. October 1980.
5-39 Radian Corporation. Assessment of Atmospheric Emissions From Petroleum Refining:
Volume 5, Appendix F. (Prepared for the U. S: Environmental Protection Agency.)
Publication No. EPA-600/2-80-075E. Research Triangle Park, N-.C. July 1980.
5-40 Murcia, A. A., et £]_., Add Flexibility to FCC's. Hydrocarbon Processing. 58_:134.
September 1979.
5-41 Production by the U. S. Chemical Industry. Chemical and Engineering News. 59(23): 33-35.
June 8, 1981.
5-42 Hoover, J. R., J. R. Blacksmith, and P. W. Spaite. (Radian Corporation) Energy Use
Patterns and Environmental Implications of Direct-Fired Industrial Processes. (Prepared
for U. S. Environmental Protection Agency.) Cincinnati, Ohio. EPA Contract No.
68-01-4136. August 1980.
5-43 Parsons, T. B., C. M. Thompson, and G. E. Wilkins. (Radian Corporation) Industrial
Process Profiles for Environmental Use: Chapter 5. Basic Petrochemicals Industry.
(Prepared for U. S. Environmental Protection Agency.) Washington, 0. C.
EPA-600/2-77-023e. January 1977.
5-44 Kent, J. A. (ed.) RiegeVs Handbook of Industrial Chemistry. Seventh edition. New York.
Van Nostrand Reinhold Company, 1974.
5-45 1979 Petrochemical Handbook Issue. Hydrocarbon Processing. 5_9_:154. November 1979.
5-46 Considine, D. M. (ed.). Chemical and Process Technology Encyclopedia. New York,
McGraw-Hill Book Company. 1974.
5-47 Berman, H. L. Fired Heaters - II: Construction materials, mechanics! features,
performance monitoring. Chemical Engineering. 85_(17):89. July 31, 1978.
5-48 Smith, T. M. Applying Ceramic Fib«r Furnace Linings. Hydrocarbon Processing.
60:169-172. April 1981.
5-49 Berman, H. L. Fired Heaters - I: Finding the Basic Design for Your Application.
Chemical Engineering. 85_(14):99. June 19, 1978.
5-50 Safety in High-Temp Heat Transfer Ruid Systems. Chemical Engineering. 88(9): 14-15.
May 4, 1981. ~~
5-51 Ayraud, S. L;, and R; J. Schreiber. (Aerotherm/Acurex) Refinery Heater Screening Study.
(Prepared'for U. S. Environmental Protection Agency.) Research Triangle Park, North
Carolina. Publication No. EPA-450/3-79-007. March 1979.
5-52 Berman, H. L. Fired Heater III: How Combustion Conditions Influence Design and
Operation. Chemical Engineering. 85_(18): 138.
5-53 Evans, F. L. Equipment Design Handbook for Refineries and Chemical Plants: Volume 2.
Houston, Gulf Publishing Company, 1980. p. 9.
5-54 Goyal, 0. P. Guidelines Help Combustion Engineers. Hydrocarbon Processing. 59: 209.
November 1980.
5-55 Berman, H. L. Fired Heaters - IV: How to reduce your fuel bill. Chemical Engineering.
85_(20): 166-7. September 11, 1978.
5-56 Reed, R. D. Furnace Operations. Houston, Gulf Publishing Company, 1976.
5-57 Waterland, L. R., et al_., (Acurex/Aerotherm) Environmental Assessment of Stationary
Source NO Control~Technologies: Second Annual Report. (Prepared for U. S. Environmental
Protection Agency.) Research Triangle Park, North Carolina. Publication No.
EPA-600/7-79-147.
5-93
-------
5-58 Martin, R. R. Burner Design Parameters for Flue Gas NO Control. John Zink Technical
Publication No. 4010. Tulsa, Oklahoma, John Zink Company, 1981.
5-59 Coe, W. W. How Burners Influence Combustion. Hydrocarbon Processing. 60: 181. May
~~
5-60 Cox, N. D., and W. Paul Jensen. Improving Fired Heaters Saves Fuel. Oil and Gas Journal.
77 (49): 73. December 3, 1979.
5-61 Radian Corporation, unpublished data. Interim Data Collection and Analysis for
Development of New Source Performance Standards for NO Emissions from Fired Heaters.
(Prepared for Emission Standards and Engineering Division, U. S. Environmental Protection
Agency.) Research Triangle Park, North Carolina. DCN 82-231-372-19-15. Contract
No. 68-02-3058. May 1982.
5-62 Tidona, R. J., W. A. Carter, and H. J. Buening. (KVB) Refinery Process Heater NO Control
Using Staged Combustion Lances. (Prepared for U. S. Environmental Protection Agency.)
Research Triangle Park, North Carolina. EPA Contract No. 68-02-2645. November 1981.
5-63 Tidona, R. J., H. J. Buening, and J. R. Hart. (IVB) Emissions From Refinery Process
Heaters Equipped with Low-NO Burners. Publication No. EPA-600/7-81-169. Research
Triangle Park, North Carol in!. October 1981.
5-64 Chapman, Kirk S. (Coen Company, Inc.) NO Reduction on Process Heaters With a Low NO
Burner. Presented at the 72nd Annual Meeting of the Air Pollution Control Association.
Cincinnati, Ohio. June 24-29, 1979.
5-65 State of California Air Resources Board. Public Meeting to Consider a Suggested Control
Measure for the Control of Emissions of Oxides of Nitrogen from Boilers and Process
Heaters in Refineries. (Prepared by the staff of the South Coast Air Quality Management
District and Stationary Source Control Division Air Resources Board.) October 1981.
5-66 State of California Air Resources Board. A Suggested Control Measure for Emissions of
Oxides of Nitrogen from Boilers and Process Heaters in Refineries. (Prepared by the staff
of the South Coast Air Quality Management District and Stationary Source Control Division
Air Resources Board.) March 1982.
5-67 Schultz, E. J., L. J. Hellenbrand, and R. B. Engdahl, "Source Sampling of Fluid Catalytic
Cracking Plant of Standard Oil of California, Richmond, California," Batelle-Columbus
Labs, July 1972.
5-68 Schultz, E. J., L. J. Hellenbrand, and R. B. Engdahl, "Source Sampling of Fluid Catalytic
Cracking, CO Boiler and Electrostatic Precipitators at the Atlantic Richfield Company,
Houston, Texas," Battelle-Columbus Labs, July 1972.
5-69 Shea, E. P., "Source Testing, Standard Oil Company, Richmond, California," Midwest
Research Institute, Kansas City, Missouri, 1972.
5-70 Gowherd, C., "Source Testing, Standard Oil of California Company, El Segundo, California,"
Midwest Research Institute, Kansas City, Missouri, 1972.
5-71 Shea, E. P., "Source Testing, Atlantic Richfield Company, Wilmington, California," Midwest
Research Institute, Kansas City, Missouri, January 1972.
5-72 McGannon, H. E., The Making, Shaping and Treating of Steel, 8th ed. , Pittsburgh, United
States Steel Co.,~T567T ~~
5-73 Russel, C. C., "Carbonization." In: Kirk-Othmer Encyclopedia of Chemical Technology.
Standen, A. (ed.), Vol. 4, 2d ed., New York, Interscience PublilFers, Inc., 1964, p.
400-423.
5-74 Personal communication, Dr. Walter Jackson, U. S. Steel, November 1977.
5-75 Personal Communication, Mr. Andrew Trenholm of the Office of Air Quality Planning and
Standards, U. S. Environmental Protection Agency, Durham, North Carolina, May 1976.
5-94
-------
5-76 , Stockham, J. D., "The Composition of Glass Furnace Emission," presented at the 63rd Annual
meeting of the Air Pollution Control Association, St. Louis, June 1970.
5-77 Mills, J. L., e£ a]_., "Emissions of Oxides of Nitrogen from Stationary Sources in
Los Angeles County; Oxides of Nitrogen Emitted by Medium and Large Sources," Joint
District, Federal, State, and Industry Project, Los Angeles County Air Pollution Control
District, Los Angeles, Calif., Report Number 3, April 1962.
5-78 Air Pollution Engineering Manual, Daniel son, J. A. (ed.). National Center for Air
"FoTlution Control, Cincinnati, Ohio, PHS Publication No. 999-AP-40.
5-79 Nesbitt, J. D., D. H. Larson, and M. Fejer, "Improving Natural Gas Utilization in a
Continuous End Port Glass-Melting Furnace," In: Proceedings of the Second Conference on
Natural Gas Research and Technology, Session IV, Paper 9, Chicago, 1972.
5-80 Ryder, R. J., and J. J. McMackin, "Some Factors Affecting Stack Emissions from a Glass
Container Furnace," Glass Ind., 50, June-July 1969.
5-81 "Background Information for Standards of Performance: Coal Preparation Plants; Volume 1:
Proposed Standards," EPA 450/2-74-021 a, October 1974.
5-95
-------
-------
SECTION 6
NONCOMtJUSTION PROCESSES
The problem of NO emissions has been researched in the chemical industry more intensively
than anywhere else because it may represent the loss of a valuable raw material. The following sec-
tions of this report discuss commercial processes developed for NOX control in the manufacture and
uses of nitric acid.
The NO released in vent gases from the manufacture and industrial uses of nitric acid, dif-
fers markedly from that emitted from a combustion flue gas in concentration, total amount, and the
ratio of N02 and NO present. The N0x-containing chemical gas is commonly a process stream which
must be recycled with maximum NO recovery in order to have an economical process. Vent gas is re-
leased only because it is too impure to recycle or too low in concentration for economic recovery.
The economic limit with a pure gas, as in nitric acid manufacture, is about 0.1 to 0.3 percent NOV,
A
or 1,000 to 3,000 ppm. The limit is higher in organic nitrations, such as the manufacture of nitro-
glycerine, where NOX content of the vent gas may approach 1 percent NOX, or 10,000 ppm.
The total amount of NO emitted from all chemical manufacturing is about 1.7 percent
(203 Gg or 2.2 x 10s tons/yr) of all NOX from manmade sources in the United States. These pro
cesses present problems only in special local areas. The problems have been most serious in military
ordance works, which manufacture large volumes of nitric acid and use it in organic nitrations. A
single plant like the Volunteer Ordance Works has produced, for example, emissions of NO equal to
all nonmilitary uses of nitric acid in the United States.
A high ratio of NO^/NO at high concentrations causes the gases to be visible as a brownish
plume. The visibility limit depends on the total amount of N02 present in the gas volume or layer
observed. A convenient rule of thumb is that a stack plume or air layer will have a visible brown
color when the N0~ concentration exceeds 6,100 ppm divided by the stack diameter in centimeters (Ref-
erence 6-1). This means that the threshold of visibility for a 5 cm-diameter stack is about 1,200 ppm
*
of N02 and for a 30 cm-diameter stack, 200 ppm of N02 (or 2,000 ppm of NOX at a 1:10 ratio of N02:NO).
6-1
-------
The distinction between N0» concentrations and total amount can be quite important in chemi-
cal vent gases, since a short burst of N0« at 10,000 ppm may be visible but less hazardous than
many tiroes- as much NO emitted from a large stack at a lower concentration. The total amount in a
A
short, concentrated emission may be too small to have a detectable effect on NOV levels in ambient
f X
air.
A large amount of research with varying degrees of success has been carried out on the devel-
opment of processes for the removal of-NO from the off-gas resulting from the manufacture and uses
of nitric acid. The abatement processes are discussed in detail in Section 6.1.3.
6.1 NITRIC ACID MANUFACTURE
Nitric acid plants are divided into two types: those that make dilute nitric acid (50-68 per-
cent nitric acid) and those that make strong nitric acid (over 95 percent nitric). Nitric acid and water
form an azeotropic (constant-boiling) mixture at about 58 percent nitric acid content; this is the limiting
factor in the nitric acid concentration available through distillation and absorption methods.
The add is concentrated to 98 percent in an acid concentration unit using extractive distillation.
The direct process for making strong nitric acid usually depends on direct formation of nitric acid
in an autoclave where nitrogen oxides react with oxygen and water to form nitric acid. Most (> 95
percent) nitric acid plants presently in operation are of the first kind.
6.1.1 Dilute Nitric Acid Manufacturing Processes
Nitric acid in the United States is made by the catalytic oxidation of ammonia. . Air and
ammonia are preheated, mixed, and passed over a catalyst, usually a platinum-rhodium complex. The
following exothermic reaction occurs:
4NH3 + 502 + 4NO + 6H20
(AH = -906 J/mole)
The stream 1s cooled to 311K (100F) or less, and the NO then reacts with oxygen to form nitrogen
dioxide and Its liquid dlmer, nitrogen tetroxide.
2ND + 02 * 2 N02 * N204
(AH - -113 J/mole)
The liquid and gas then enter an absorption tower. Air is directed to the bottom of the
tower and water to the top. The N02 (or N204) reacts with water to form nitric acid and NO, as
follows:
6-2
-------
3N02 + H20 -* 2HN03 + NO
(6-3)
(AH 3 -135 J/mole) ,
The formation of 1 mole of NO for each 2 moles of HN03 makes it necessary, to reoxidize NO after each
absorption stage since the gas rises up the absorber and limits the level of recovery that can be
economically achieved.
Acid product is withdrawn from the bottom of the tower irrconcentrations of 55 to 65 percent.
The air entering the bottom of the tower serve;; to strip N02 from the product and to supply oxygen
for reoxldizlng the NO formed In making nitric add (Equation 6-3).
The oxidation and absorption operations can be carried out at low pressure (- 100 kPa, 1 atm),
at medium pressure (400 to 800 kPa or 58 to 11(5 psia) or at high pressure (1000 to 1200 kPa or 145 to
174 psia). Both operations may be at the same pressure or at different pressures.
Before corrosion-resistant materials were developed, the ammonia oxidation and absorption
operations were carried out at essentially atmospheric pressure. This also had advantages compared
to the higher pressure processes of longer catalyst life (about 6 months), and increased efficiency
of ammonia combustion. However, because of th« low absorption and NO oxidation rates, much more
absorption volume is required, and several large towers are placed in series. Some of these low
pressure units are stm in operation, but they represent less than 5 percent of the current U.S.
nitric acid capacity.
Combination pressure plants carry out the ammonia oxidation process at low or medium pres-
sure and-the absorption step at medium or* high pressure. The higher combustion temperature and gas
velocity at an increased pressure for the oxidatjon reaction shortens catalyst's lifetime (1 to 2
months) through increased erosion and lowers the ammonia oxidation conversion efficiency (Reference
6-2). Thus lower pressures in the oxidation.process are preferred. On the other hand, higher pres-
sures in the absorption tower increase the absorption efficiency and reduce NO levels in the tail
gas. Of course these advantages must be weighed against the cost of pressure vessels and compressors.
The choice of which combination of pressures to use is very site specific and 1s governed by
the economic trade-offs between the costs of raw materials, energy and equipment, and process effi-
ciency; and local emissions limits. In the 1960's, combination low pressure oxidation/medium pressure
absorption and single pressure (400 to 800 kPa) plants were preferred. Since the early 1970's, the
trend has been toward medium pressure oxidation/high pressure absorption plants in Europe and single
pressure plants (400 to 800 kPa) in the United States.
6-3
-------
6.1,1.1 Single Pressure Processes
In the single pressure process, both the oxidation and absorption processes are carried out
at the same pressure either low pressure (100 kPa or - 1 atm.) or medium pressure (400 to 800 kPa).
Single pressure plants are the most common type. Figure 6-1 is a simplified flow diagram of a
single pressure process (Reference 6-3). A medium pressure process will be described below.
Air is compressed, filtered, and preheated to about 573K by passing through a heat exchanger.
The air is then mixed with anhydrous ammonia, previously vaporized in a continuous-stream evaporator.
The resulting mixture, containing about 10 percent ammonia by volume, is passed through the reactor,
which contains a platinum^rhodium (2 to 10 percent rhodium) wire-gauze catalyst (for example", 80-mesh
and 75-ym diameter wire, packed in layers of 10 to 30 sheets so that the gas travels downward through
the gauze sheets). Catalyst operating temperature is about 1,023K. Contact time with the catalyst
_*
is about 3 x 10 sec.
The hot nitrogen oxides and excess air mixture (about 10 percent nitrogen oxides) from the reac-
tor are partially cooled in a heat exchanger and further cooled in a water cooler. The cooled gas is
introduced into a stainless-steel absorption tower with additional air for the further oxidation of
nitrous oxide to nitrogen dioxide. Small quantitites of water are added to hydrate the nitrogen
dioxide and also to scrub the gases. The overhead gas from the tower is reheated by feed/effluent
heat exchangers and then expanded through a power recovery turbine/compressor used to supply the
reaction air. The" tail gas is then treated by the tail-gas treater for NO abatement. The bottom
of the tower yields nitric acid of 55 to 65 percent strength.
6.1.1.2 Dual Pressure Processes
In order to obtain the benefits of increased absorption and reduced NOX emissions from high-
pressure absorption, dual-pressure plants are installed. Recent trends favor moderate-pressure
oxidation and high-pressure absorption.
A process flow diagram for a dual-pressure plant by Uhde is shown in Figure 6-2. Liquid
ammonia is vaporized by steam, heated and filtered before being mixed with air from the air/nitrous
oxide compressor at from 300-500 kPa (44 to 72 psia). The ammonia/air mixture is catalytically
burned in the reactor with heat recovery by an integral waste heat boiler to generate steam for use
in the turbine driven compressor. The combustion gases are further cooled by tail gas heat exchange
and water cooling before compression to the absorber pressure of 800-1400 kPa (116 to 203 psia). The
6-4
-------
WASTE GASES TO
POWER RECOVERY
AND TAIL GAS
'TREATMENT
AMMONIA
(ANHYD
ROUS)
COMPRESSOR
EVAPORATOR
F
W
WEAK k ^
ACID ] *"
iiTcn AIR au
kTERj
I
ABSORBER
NITRIC ACID
Figure 6-1. Single pressure nitric acid manufacturing process (Reference 6-3).
6-5
-------
1. AMMONIA EVAPORATOR 9.
2. AMMONIA PREHEATER ; 10.
3. AMMONIA FILTER j 11.
4. AIR FILTER i 12.
5. AIR COMPRESSOR 13.
6. AMMONIA-AIR MIXER 14.
7. BURNER WITH LA MONTE BOILER 15.
8. TAIL GAS PREHEATER I
GAS COOLER
TAIL GAS PREHEATER II
ABSORPTION COLUMN
DEGASSING TOWER
MIXING JET
NO COMPRESSOR
TAIL GAS TURBINE
Figure 6-2. Dual pressure nitric acid plant flow diagram (Reference 6-2).
-------
absorption tower is internally water cooled to increase absorption by water. Nitric acid up to 70
percent concentration is withdrawn from the bottom of the column and degassed with the air feed to
remove unconverted NO before being sent to storage. The air/NO mixture is combined with reactor
effluent to form the absorber feed. High yields of up to 96 percent conversion and tail-gas emissions
as low as 200 ppm NO- can be obtained by this process.
6.1.1.3 Nitric Acid Concentration
Figure 6-3 illustrates a nitric acid concentration unit using extractive distillation with
sulfuric acid. A mixture of strong sulfuric acid and 55 to 65 percent nitric acid is introduced at
the top of a packed column, and flows down the column counter-current to the ascending vapors.
Nitric acid leaves the top as a 98 percent nitric acid vapor containing small amounts of NO and
oxygen, which result from the dissociation of nitric acid. The vapors pass to a bleacher and a
condenser to condense nitric acid and separate NO and oxygen, which pass to an absorber column for
conversion to, and recovery of, nitric acid. Air is admitted to the bottom of the absorber. Dilute
sulfuric acid is withdrawn from the bottom of the dehydrating tower and is sent to be concentrated
further to be used for other purposes. The system usually operates at essentially atmospheric
pressure.
6.1.1.4 Direct Strong Nitric Acid Processes
Nitric acid of high strength can be made directly from ammonia in direct strong nitric acid
processes. These processes depend upon the formation of nitric acid by reaction of N0« or N-O. with
oxygen and water forming 95 percent to 99 percent nitric acid.. -In this direct process, the composition
of the product nitric acid is not restricted by the azeotropic limit.
The principal licensors of these direct processes are Uhde and Davy Powergas. Uhde has built
two plants in this country using their direct strong nitric acid process. The Uhde process will be
described in detail below. Davy Powergas has two direct strong nitric acid processes; the CONIA
process and the SABAR process. Davy has not built any plants utilizing these processes in the
United States, but there is a CONIA plant recently constructed in Sweden and a SABAR plant recently
constructed in Spain. How these processes differ from the Uhde process will also be described below.
Figure 6-4 shows a process flow diagram for a direct strong nitric acid plant. Air and
gaseous ammonia are mixed and reacted where steam is generated in a combination burner/waste heat-
boiler by the heat of reaction. The reaction products are cooled, and a weak nitric acid condensate
removed. The remaining gases are put through two oxidation columns where the NO is converted to NO.
6-7
-------
TAIL GAS TO
ATMOSPHERE (VOLUME %)
COUNTERCURRENT
CONDENSER
VAPOR
98XHN03
FEED
HN03
DEHYDRATING
COLUMN
VAPOR
i
1
CONDENSATE
BLEACHER
NCIN-
CONDEIUSABLE
GASES
T
95-99% HN03
TO COOLER AND STORAGE
VAPOR
ILIQUID
STEAM-
COIL
74.3 N2
20.4 02
1.0NO + N02
4.2 H20
\
ABSORPTION
COLUMN
BOILER
70% H2S04
TO COOLER
I
.AIR
55% HN03
Figure 6-3. Nitric acid concentrating unit.
6-8
-------
oc
3
1. AIR FILTER
2. AMMONIA-AIR MIXER
3. AMMONIA GAS FILTER
4. BURNER WITH WASTE-HEAT BOILER
5. GAS COOLER I
6. BLOWER
7. GAS COOLER II
8. OXIDATION COLUMN
9. ACID COOLER 17.
ID. CIRCULATING ACID TANK 18.
11. ACID COOLER 19.
12. FINAL OXIDATOR 20.
13. BRINE GAS COOLER 20.
14. ABSORPTION COLUMN 21.
15. TAIL GAS SCRUBBER 22.
RAW ACID COOLER
LIQUEFIER
AGITATOR VESSEL FOR
RAW MIXTURE
REACTOR VESSEL
BLEACHING COLUMN
FINAL ACID COOLER
16. HEAD TANK FOR RAW ACID 23. MIXING JET
Figure 6-4. Process flow diagram for direct production of highly concentrated nitric !acid (Reference 6-2).
-------
The overhead vapors are compressed to a pressure that allows the equilibrium di-nltrogen tetroxide
'ii :, i" ' .
(N20.) to be liquefied with the use of cooling water alone. The liquid NgO^ is converted to nitric
add of about 95 to 99 percent by reacting the N204 with oxygen at a pressure of 5000 kPa (50 atm).
The conversion reaction is: 2N20. + 2H20 * 4HNO,. Tail gases from the absorption column are
scrubbed with water and condensed N^O* in a tall-gas scrubber before being released. The liquid
from the tail-gas scrubber is mixed with the concentrated add from the absorption column, which
has been bleached and liquefied. The combined product liquid (containing N20. as well as HNO,) is
reacted with oxygen in the reactor vessel, cooled, and bleached to produce the concentrated nitric
acid.
Both Uhde plants using this process were built in 1973 for the U.S. Government: one in
Ooliet, Illinois makes 236 Gg/d (260,000 tons/day) of 98.5 percent nitric acid; the second in
Chattanooga, Tennessee, makes 313 Gg/d (345,000 tons/day) of 98 percent nitric acid. Neither are
1n operation at present, although, both were designed to meet the New Source Performance Standards.
In the Davy Powergas SA6AR process (Reference 6-4), like the Uhde process, ammonia and air
art reacted at atmospheric pressure, and a 2-3 percent nitric acid is condensed and removed as a
byproduct. Davy Powergas estimates 0.3 kg of this weak acid byproduct is produced per kg of con-
centrated add. As in the Uhde process, N02 1s then produced from the product gases and absorbed
1n concentrated nitric acid. However, whereas Uhde forms N204 from this liquid, and reacts the N20.
»
with oxygen, the SABAR process takes the concentrated HN03 to a vacuum rectification column, where
concentrated HNO, comes off overhead and azeotropic nitric acid is collected at the bottom. Atmos-
pheric emissions are less than 500 ppm nitric oxides, which would not meet the new source standards
1n the United States without further treatment. '
Davy Powergas developed the CONIA process to meet the more stringent environmental regulations
for Its site in Sweden. The CONIA process also depends on the ammonia-air reaction, followed by re-
moval of the water which is generated. The plant produces both 99.5 percent nitric acid and 54 per-
4 '
cent nitric add, with less than 200 ppm NOX in the stack gases, and no other solid or liquid waste
streams. However, Davy Powergas considers this particular plant design to be over-designed and hence
too costly for most applications unless lower emissions limits must be met (Reference 6-5).
6.1.2 Emissions
Absorber tail gas is the principal source of NOX emissions from nitric acid manufacturing.
Minor sources include nitric acid concentrators and the filling of storage tanks and shipping con-
tainers. Nitrogen oxide emissions from nitric acid manufacturing are estimated at 127 Gg
6-10
-------
(140,000 tons) during 1974, which is about 1.0 percent of the NOX emissions from stationary sources.
It is estimated that 7.4 Tg (8.2 million tons) of nitric acid (100 percent) were produced in 1974
(Reference 6-6). AP-42 (Reference 6-7) cites an average emission factor for uncontrolled plants of
25 to 27.5 kg/Mg of acid. Typical uncontrolled tail-gas concentrations are on the order of 3000 ppm
of NO with equal amounts of NO and N02 (Reference 6-8). Unde cites emission levels in excess of 800
ppm for low-pressure plants, 400 to 800 ppm for medium-pressure plants, and less than 200 ppm for high-
pressure plants (Reference 6-2). The extent of control for these plants is not known, although, Uhde
did state that all three processes could be designed in such a way as to meet State and Federal emis-
sion limits.
In any nitric acid plant, the NOX content of tail gas is affected by several variables. Abnormally
high levels may be caused by insufficient air supply, high temperature in the absorber tower, low-
pressure, production of acid at strengths above design, and internal leaks, allowing gases with
high nitrogen oxide content to enter the tail-gas streams. Careful control and good maintenance are
required to hold tail-gas nitrogen oxide content to a minimum.
6.1.3 Control Techniques for N0y Emissions from Nitric Acid Plants
Nitric acid plants can be designed for low NOX emission levels without any add-on processes.
Such plants are usually designed for high absorber efficiency; high inlet gas pressures and effec-
tive absorber cooling can lead to low NOX emissions. However, many new plants, and all existing
plants, are not designed for NO emission levels low enough to meet'present standards. For these
plants, add-on abatement methods are riecessary. >
The available abatement methods suitable for retrofit include chilled absorption, extended
absorption, wet chemical scrubbing, catalytic reduction, and molecular sieve adsorption. In this
section, these various control techniques for NOX are described. These techniques may also be
appropriate for retrofit of explosive and adipic acid plants. .
Many of the retrofit processes are offered by more than one licensor, and many licensors
(such as Uhde) offer more than one process. Table 6-1 lists the major processes, the types of
plants for which the processes are most suitable, and examples of nitric acid plants where the
processes have been applied. (The examples of nitric acid plants are not meant to be inclusive.)
The selection of a control method depends on such things as the degree of control required,
the operating pressure of the plant, and the cost and availability of fuel. For example catalytic
reduction was used to establish the NSPS-originally. Since that time fuel costs have risen to the
point where catalytic abatement is not economically attractive for new nitric acid plants but can
be used as an effective secondary treatment to meet the NSPS.
6-11
-------
TABLE 6-1. HOX ABATEMENT METHODS OH NEH OR EXISTING HITRIC ACID PLANTS
Process
Chilled
Absorption
Extended
Absorption
yet Chemical
Scrubbing
Method
Increased solu-
bility of NOX 1n
chilled water
Increased absorp-
tion of NOX by
additional ab-
sorption equip-
ment
Scrubbing tail
gases with urea
solution or
ammonia to
recover NOX
*
Comments
Usually cannot meet
NSPS without other add-
on technology or low-
ered acid product
concentration
Inlet pressure of 760
kPa required (addi-
tional compressors
may be required)
Requires additional
compressor
Performs better at high
pressure but operable at
lower pressures. Recovers
ammonium nitrate and urea
solution. Requires re-
.friqeration.
May require an evaporator
to produce a concentrated
ammonium nitrate by prod-
uct. No refrigeration
required.
Licensors
COL-VITOK
TVA
J. F. Pritchard
(Grande Paroisse)
P. M. Weatherly
Chemico
Uhde
C&I Girdler
CoFAZ
MASAR {urea
scrubbing)
Norsk Hydro
(urea scrubbing)
Goodpasture
(ammonia
scrubbing)
Examples
2-318 Mg/d (350 tons/day) (with Gulf cata-
lytic reduction add-on). Nitram, Taiupa, Fla.
2-50 Mg/d (55 tons/day) plants at Muscle
Shoals, (1972)
327 Mg/d (360 tons/day) plant, Miss.
Chemicals, Yazoo City, Miss. 1973.
272 Mg/d (300 tons/day) Holston Army
Ammunition Plant, Kingsport, Tenn.
Cominco Plant, Beatrice Neb.
Kaiser. Tampa. Fla. and Bainbridae, Ohio
9 U.S. plants, 1 Japan plant (employs
chilled absorption process)
908 Mg/d (1000 tons/day) Monsanto,
Pensacola, Fla. 1977.
250 Mg/d (275 tons/day) plant, Allied
Chemical, Oman, Neb. 1975
None built to date
111. Nitrogen Pit., Marsalles, 111.
Air Products & Chem. , Pace, Fla.
Norsk Hydro, Proggunn, Norway
90 Mg/d (100 tons/day) Goodpasture pit.,
1974. Dimmitt, Texas
Chevron Oil Co., Richmond, Calif. 1976
C.F. Industries, Fremont, N.D.
2 scrubbers for 7 plants totalling 544
mg/d (600 tons/day).. Cyanamid, Wei land, Ont.
I
_J
ro
-------
TABLE 6-1. NOX ABATEMENT METHODS ON NEW OR EXISTING NITRIC ACID PLANTS (Concluded)
Process
Catalytic
Nonselectlve
Catalytic'
Selective
Heterogeneous
Catalysis
Chemical
Absorption
Molecular
Sieve
Method
Burns NOX and 02
with CH4 or H2 to
form N2, HzO, C02
Burns NOx with
ammonia to form
Hy and HyQi 02
not affected ~
Oxidation of NO *
N0£ catalyzed by
heterogeneous ca-
talysis before
absorption
Oxidation with
KMn04 (KHn04
electrolytlcally
reclaimed
Absorption by
molecular sieve,
regeneration of
the sieve by
thermal cycling
Comments
Consumes natural gas,
uneconomical If high
NOX or 02 content
(also reacts with 02)
Hay be used In con-
junction with extended
absorption
Energy recovery
possible
Works at low or high
pressure
Uses ammonia, can be
expensive to operate .
Often used with ex-
tended absorption
Works at low or high
pressure
Energy recovery
usually not possible
Can achieve very low
emission If desired
Limited success
Uneconomical not pres-
ently offered
High energy and capi-
tal demands
Hard to fit cycling
of sieve Into, con-
tinuous plant opera-
tion
Licensors
C&I Girdler
D. M. Weatherly
Cherolco
Gulf
Uhde (BASF cata-
lysts)
Mitsubishi
CDL/VITOK
Cams Chemical
Puraslv N
(Union Carbide)
Examples
Olln, Lake Charles, La. (also, Weatherby
plants)
IMC Corp., Strellngton, La. (1976) (with
extended absorption). 817 Mg/d (900 tons/
day ) . 1 977 . Col umbl a N1 trogen ,- Augusta ,
Ga.
Location not available
Nltram plants In Tampa, Fla., Installed
after CDL/VITOK process.
10 plants In U.S.
Plants 1n Europe and Japan
Under development
2 plants 1n Japan, not currently offered
1n U.S.
SO mg/d (55 tons/day) Hercules, Inc.
Bersemer, Ala. 1974
50 mg/d (55 tons/day) U.S. Army, Hols ton,
Kingston, Tenn. (Inoperable, dismantled)
-------
The Inlet pressure at the absorber is an important factor in the selection of NO control
equipment. In general, extended absorption equipment cannot be economically installed where the
equipment will have inlet pressures of less than 758 kPa (110 psia). Consequently, extended adsorp-
tion is toot usually chosen for older, low pressure nitric acid plants. Wet scrubbing and molecular
sieve absorption are also not as effecive at low pressures. Catalytic reduction, however, does not
require high pressures. *
6.1.3.1 Chilled Absorption -
This method is used primarily for retrofit of existing plants. Chilling the water used in
a nitric acid absorption tower leads to higher yields of nitric acid and lower NOX concentrations
in the tail gas. Both water and brine solutions have been used in.a closed loop system to provide
local cooling to the liquid on the trays of the absorption tower. Absorption may be further en-
hanced by heterogeneous catalytic oxidation of NO to N02 upstream of the absorption tower. ,
CDL/VITOK Process
Figure 6-5 shows a CDl/VITOK process flow diagram. Tail gas enters the absorber, where the
gases are contacted with a nitric acid solution to both chemically oxidize and physically absorb
nitrous oxides. The reaction of NO to N0« may be catalyzed in the main absorber. The upper portion
of the absorber is water cooled to improve absorption. The nitric acid solution from the absorber
1s sent to a bleacher where air removes entrained gases and further oxidation occurs. The bleached
nitric add solution is then either sent to storage or^recirculated to the absorber after the addi-
tion , of make-up water. The process employes a closed loop system to chill the recirculated acid
solution and tower cooling water by ammonia evaporation.
One variation in this system proposed by CDL/VITOK includes the addition of an auxilliary
bleacher operating in parallel with the primary unit. Another variation uses a secondary absorber
with its own bleacher.
At the Nitram, Tampa, Florida location two 318 Mg/d plants were fitted with the CDL/VITOK
process. NO tail gas concentrations were reduced from 1500 to 1800 ppm to 600 to 800 ppm. With
*
the addition of a gulf catalytic abatement system the plant meets local regulations. A second plant
at Nitram fitted with the process showed promise but was shut down and replaced with a new nitric
add plant.
TVA Process
The Tennessee Valley Authority, at their nitric acid plant in Muscle Shoals, Alabama, designed
and Installed refrigeration for NOV abatement pruposes in 1972, in order to meet State standards of 2.75
^»
6-14
-------
PURIFIED
TAIL GAS
COOLING
WATER
RETURN
FEED GAS/
LIQUID'
FROM HEAT
EXCH.
ABSORBER
BREACH AIR
RECOVERED ACID-**
COOLING
WATER
MAKE-UP WATER
PUMP
Figure 6-5. Schematic diagram of the CDL/V1TOK NOX removal process(Reference 6-8).
&-15
-------
kg/Kg of nftric add (5.5 Ibs/ton). A flow diagram of their abatement equipment 1s shown in Figure
6-6 (Reference 6-10). It consists of a cooler attached to the nitric acid absorption tower, and a
bleacher from which any effluent gases are recycled to the absorption tower. As a result of adding
the NO control process the concentration of the product acid dropped from 65 percent to 51 to 57
percent.
6.1.3.2 Extended Absorption
The extended absorption process basically consists of a second absorption column to which the
tall gas from the nitric acid plant is sent. The NOX is absorbed by water and forms nitric acid,
which increases the acid yield. Extended absorption can be added in conjunction with pressurizing
the tall gas upstream of the tower or chilling the absorbent in the tower. However, neither of these
options is a necessary part of the absorption process.
This process is offered by several licensors, including J. F. Pritchard (Grande Paroisse
process), D. H. Weatherly, Chemico, Uhde and C and I Grtdler (CoFaz process). The economics of the
process generally require the inlet pressure at the absorber to be at least 758 kPa (110 psia). Also,
cooling is usually required if the inlet NO concentration 1s above 3000 ppm (Reference 6-6). There is nc
A
liquid or solid effluent from extended absorption; the weak acid from the secondary absorber is
recycled to the first absorber, increasing the yield of nitric acid. In some cases, extended absorp-
tion can be used in conjunction with catalytic tail gas treatment (see Section 6.1.3.4).
Figure 6-7 shows a process flow diagram for the Grande Paroisse process, which is representa-
tive of extended absorption processes. Off-gas from the existing absorber flows into the secondary,
or Grande Paroisse absorber. The tail gas from the secondary absorber goes to an existing tail-gas
heater before being vented'to the atmosphere or passing through a catalytic reduction unit. The
, l
liquid effluent is returned to the primary absorber to become part of the acid product.
More 'than 15 extended absorption plants (by various licensors) are operating in the United
States. In cases where the off-gas must be compressed before going to the secondary absorber, or
where refrigeration Is used, maintenance requirements are increased. Power recovery by an air
compressor/tail-gas expander is usually employed when a pressurized absorber is used.
6.1.3.3 Met Chemical Scrubbing
Met chemical scrubbing uses liquids, such as alkali hydroxides, ammonia, urea and potassium
permanganate to convert NOg to nitrates and/or nitrites by chemical reaction. Also, scrubbing may
be done with water or with nitric acid. Several of these processes are described below.
6-16
-------
ABSORPTION TOWER AND BLEACHER
_ DETAIL
COILS SUBMERGED
WATER
(
COOLING
WATER HEADER
r£-
r
7
COOLER CONDENSER
NITRIC ACID GAS FROM AMMONIA
I
OXIDIZER
Figure 6-6. TVA chilled absorption process (Reference 6-10).
6-17
-------
TO EXISTING
TAIL GASPREHEATER
I
r___L_
1
i EXISTING [.,,
ABSORBER r*1l
1 1
1
I
|
NITROUS j
1
L --.
NITRIC ACID }
1
1
i
1
1
|
1
j
"L
LEVEL
CONTROL
i Ml
i
SECONDARY
ABSORBER
TOWER
RECOVERED
NITRIC
AC
ACID TRANSFER
PUMPS
(ONE SPARE)
^1 ' *v
" W ^
^
S1 ^1
ID
,
^J \
FLOW
CONTROL
'vijL
i
i
i
.. i
PROCESS
WATER
ivn i bit
|j»^"°^ ^^^
STARTUP^ STARTUP PROCESS WATER
ONLY PUMP RUMP
(EXISTING)
Figure 6-7, Grande Paroisse extended absorption process for NOx treatment.
,6-18
-------
Urea Scrubbing
This process is offered by two licensors: MASAR, Inc. and Norsk Hydro. The.mechanisms
given below have been proposed for this process (Reference 6-11).
HN02 + CO(NH2)2 £ N2 + HNCO + 2H20 (6-4)
HNCO + HN02 * N2 + C02 + HgO (6-5)
HNCO + H20 + H+ t NH^ + COg (6-6)
When the concentration of nitric acid is low, reaction (6-6) predominates so that the overall
reaction is
HN02 + CO(NH2)2 + HN03 * N2 + C02 + NH4N03 + ti£Q (6-7)
As shown in reaction (6-7), half the nitrogen in the reaction will form NH^NOj, a valuable byr
product, and half will form N2, a nonpolluting species.
The MASAR process is shown in Figure 6-8. A three-stage absorption column is used with gas
and liquid chillers on the feed gas and recirculated solvents. The process as described by MASAR,
Inc., (Reference 6-12) is given below.
The MASAR process, as applied to nitric acid plants, takes the tail gas from the exit of the
absorption tower and passes it to a gas chiller where it is cooled. During this cooling operation,
condensation occurs with the formation of n'ltric acid. This chilled gas and condensate passes into
Section A of the MASAR absorber. Meanwhile, the norma.1 feedwater used in the nitric acid plant
absorption tower is chilled in Section C of the MASAR absorber and ,is then fed to Section A of the
MASAR absorber, where it flows down through the packing countercurrent to the incoming chilled tail
gas to scrub additional NOX from the tail gas. This scrubbing water is recirculated through a
chiller to remove reaction heat and then this weak nitric acid stream is fed to the nitric acid
plant absorber to serve as its feedwater.
. The tail gas then passes into Section B of the MASAR absorber where it is scrubbed with a
circulating urea-containing solution. A urea/water solution is made up in a storage tank and
metered into the recirculating system at a rate necessary to maintain a specified minimum urea
residual content. As the solution scrubs the tail gases, both nitric acid and nitrous acids are
formed, and the urea in the solution reacts with the nitrous acid to form COfOOg). N2, and HgO. As
the solution is circulated, the nitric acid content rises and some of the urea present hydrolyzes
and forms some ammonium nitrate. To maintain the system in balance, some of the circulated solution
is withdrawn. The recirculated solution is also pumped through a chiller to remove the heat of
reactions and to maintain the desired prdcess temperature in Section B.
6-19
-------
TAIL GAS
- TO ^
NITRIC ACID
PLANT
SPENT MASAR
A (BLOW DOWN)
LIQUID
CHILLER
/ 1
A i ^
CONCENTRATED MASAR f V^V
TAIL GASES
FROM PLANT
*
TAIL GAS
.,. I7HHIFR
fiHIl 1 FR' ««. .__ _.,.. .1 V
* PUMP
FEED WATER
SECTION
x~x
l"1 L~^
'^
SECTION
XX
_n_
/"'^
SECTION
XX
r
^ y
MASAR
ABSORBER
*g .
TO NITRIC ACID PLANT
ABSORBER COLUMN
Figure 6-8. Flow diagram of the MASAR process (Reference 6-12).
6-20
-------
The tail gases then pass Into Section C where they are again scrubbed by the feedwater
stream that is used, ultimately, as the nitric acid plant absorption tower feedwater. The tail
gases then leave the MASAR absorber and pass on to the normally existing mist eliminator and neat
exchanger train of the nitric acid plant. The cooling medium used in the gas chiller can be liquid
ammonia. The vaporized ammonia is subsequently used as the feed to the plant ammonium nitrate
neutralizes For non-ammonia nitrate producers, mechanical refrigeration could be used or the
ammonia vapor can be used in the nitric acfd converter directly.
The MASAR process has been reported to reduce NO emissions from 4000 ppm to 100 ppm. ' The
process could technically be designed for no liquid effluent. In practice, however, liquid blowdown
of 16 kg/h (35 Ib/hr) of urea nitrate in 1£!0 kg/h (396 Ib/hr) of water is estimated for a 320 Mg of
acid/day (350 tons/day) plant (Reference 6-12).
A MASAR unit installed in 1974 for Illinois Nitrogen Corporation on a 320 Mg/d plant regu-
larly operates with between 100 and 200 ppm of NOX in the tail gas. According to the Illinois Nitro-
gen plant manager (Reference 6-13), inlet NO concentrations to the MASAR unit are approximately 3500
ppm and outlet concentrations are between 200 and 400 ppm. The Illinois Environmental Protection
Agency has tested this unit, using Method 7, with reproducible results of 57 ppm average emissions.
The unit is reported to operate with good reliability and has increased the net product recovered.
The Norsk Hydro process was developed by Norsk Hydro A/S, the Norwegian state-owned power
generating authority and fertilizer and chemical manufacturer, to reduce NOX emissions from 1525 ppm
to 850 ppm. The modifications were made to an older, atmospheric pressure plant and two more recent
medium-pressure plants (300 and 500 kPa) (44 to 72 psia). Basically, the last absorption towers in the
process streams of the older plant were modified to contact the tail gases from all three plants with urea
solution and nitric acid. The result was a net 44 percent reduction in NOX emissions, as given
above. On a plant-wide basis, 10.4 kg of ammonium nitrate are produced per Mg of nitric acid (20.8 Ib/ton)
(Reference 6-11).
Norsk Hydro has also used urea addition on three plants producing a total of 5 Gg/d (5500 tons/
day) of prilled NPK fertilizers. This method was used to control NOV emissions for lower-grade phosphate
A *
rock. Nitrous oxide is evolved when nitric and nitrous acid oxidizes impurities in the rock such
as sulphides and organic material. The. addition of urea to the phosphate rock digester tends to
reduce NO emissions to 2.5 kg/Mg (5 Ib/ton) phosphate from levels as high as 40 kg NO per Mg phosphate
(80 Ib/ton) by adding 5 to 10 kg urea per Mg phosphate rock (10 to 20 Ib/ton) (Reference 6-11).
6-21
-------
Ammonia Scrubbing (Goodpasture Process)
Goodpasture, Inc. of Brownfield, Texas is the licensor of a process developed in 1973 in
order for its Western Ammonia Corporation nitrogen complex in Dimitt, Texas to meet a 600 ppm
maximum NO effluent imposed by the Texas Air Control Board. The process which was developed is
suitable to retrofit existing plants for reduction of an inlet concentration of 10,000 ppm to
within the 1.5 kg NOo/Mg acid (^210 ppm) standards set for new nitric acid plants.
The process flow diagram fpr this process is shown in Figure 6-9. Feed makeup streams to
this process are .ammonia and water with ammonium nitrate produced as a byproduct. The total process
is conducted in a single packed contact absorption tower with three sections operated in a co-
current flow. Goodpasture states that the key to successful operation is the process' capability
|
to minimize the formation of ammonium nitrite and to oxidize the ammonium nitrite which does form to
ammonium nitrate.
The Goodpasture process consists of three distinct sections. The first is a gas absorption
and reaction section operating on the acidic side, the second is a gas absorption and reaction
section operating on the ammoniacal side, and the third is principally a mist collection and ammonia
recovery step.
In the first section, a significant portion of the oxides of nitrogen react to form nitric
acid which maintains the acidic condition in this section. The nitric acid formed reacts with the
free ammonia content of the solution from the ammoniacal section to form ammonium nitrate a portion
reacting in the acidic section, and a portion reacting in the ammoniacal section. The feed solution
to the acidic section is the product solution from the ammoniacal section. The ammonium nitrite
content of this solution-is oxidized to ammonium nitrate by the acidic conditions existing in this
first section. The product solution from the Goodpasture process is withdrawn from this acidic
section.
In the second, or ammoniacal contacting, section the remainder of the oxides of nitrogen react
to form ammonium nitrate and ammonium nitrite; the proportion of each being dependent on the oxida-
tion state of the oxides of nitrogen in the gas phase. Ammonia is added to the circulating solution
within this section to maintain the pH at a level of 8.0 to 8.3. The liquid feeds to this section
are the product solution from the mist collection section, and a portion of the acidic solution from
the first section.
The third section is incorporated principally to collect the mist, and any ammonium nitrate
or ammonium nitrite aerosols which form in the first two sections. In addition, any free ammonia
6-22
-------
TAIL
GAS IN"
pH
RECORD.
CONTROL.
'
TREATED
TAIL GAS
OUT
-^
LEVEL
ACIDIC
SECTION
»> A /N
STEAM
CONOENSATE
AMMONIACAL
SECTION
COLLECTION
PRODUCT
AMMONIUM
NITRATE
SOLUTION
LEVEL
" CONTROL
pH
RECORD. ;
AMMONIA-
LEVEL
CONTROL.
tS ^
HYDRAULIC
CONTROL
VALVE
pH
RECORD.
CON-TROL.
4XH"^-i^
Figure 6-9. Process flow diagram for trie Goodpastiire process (Reference 6-14),
6-23
-------
stripped from the solution in the amnoniacal section is also recovered in this third section.
Process water or steam condensate is fed to this section in quantities sufficient to maintain the
product ammonium nitrate solution in the 30 to 50 percent concentration range. A small amount
of the acidic solution is also fed to this section in order to control the pH to approximately 7.0.
The product solution from the abatement process is withdrawn at about 35 to 40 percent
ammonium nitrate concentration, and contains approximately 0.05 percent ammonium nitrite. At the
Dimraitt plant, this solution is heated to 390K (240F), which completes the removal of the ammonium
nitrite, before further processing. Other users have discovered that if the solution sits for a
day in a day-tank, without heating, the ammonium nitrite is removed.
The Goodpasture process has been installed at CF Industries' Fremont, Nebraska plant and
.Chevron Chemical's Richmond, California plant. In addition, American Cyanamid Company is installing
the process at one high-pressure and six low-pressure plants in Canada. Existing systems have
given reliable operation and have met the emissions requirements for which they were designed.
« .. I
One particular advantage of this process is that the pressure losses in the process are only
6.8 to 13.0 kPa (1-2 psi) which allows its application to low-pressure plants. One older, 340 kPa
(49 psia) plant has consistently met its required 400 ppm outlet concentration. Another advantage
of the low-pressure drop is that reheat and power recovery of the effluent train in moderate-
pressure plants is usually economical. However, special precautions must be taken to eliminate
deposition of ammonium nitrate on the turbine blades.
Energy requirements of the process have been less than expected, the original design speci-
fied heating the ammonium nitrate scrubbing solution to facilitate oxidation of ammonium nitrite
to nitrate. However, it has been found that this reaction occurs spontaneously if the solution
is allowed to stand for a day in a holding tank.
The retrofit of a Goodpasture unit may require some additional process modifications beyond
the abatement equipment. For example, modem fertilizer plants use ammonium nitrate solutions in
excess of 85 percent. The Goodpasture byproduct solution is only 35 to 55 percent ammonium nitrate;
therefore, additional evaporators may be needed to concentrate the Goodpasture effluent. Chevron;
however, reports significant overall steam savings without additional evaporators.
Caustic scrubbing
Sodium hydroxide, sodium carbonate and other strong bases have been used for nitric acid
scrubbing. Typical reactions for this process are:
6-24
-------
2NaOH + 3NO£ * 2NaN03 + NO + HgO (6-8)
2NaOH + NO + N02 ' * 2NaN02 + HgO (6-9)
A caustic scrubbing system was installed at a.Canadian nitric acid plant in the late 1950's
(Reference 6-15). However, disposal of the spent solution is a serious water pollution problem,
and the concentrations of the salts are too low for economic recovery. There have been no recent
installations of this process.
v
Potassium Permanganate Scrubbing
Another potential chemical for scrubbing solutions is potassium permanganate. The Carus
Chemical Company (a large producer of potassium permanganate) has developed a process for potassium
permanganate solution scrubbing of NO . However, in the process, permanganate is reduced to manganate,
which must be electrolytically oxidized. The cost for the electrolysis, as well as the permanganate
make up cost, makes the process uneconomical. This process has not been installed at any nitric acid
plant in this country. , Two plants are in operation in Japan, but no cost or user information is
available.
6.1.3.4 Catalytic Reduction
This section describes two different catalytic reduction processes. They are nonselective
catalytic reduction and selective catalytic reduction.
Nonselective Catalytic Reduction
In nonselective catalytic absorption, methane or hydrogen reacts with the NO and oxygen in
the tail gas to form N2, H-0, and C02. A schematic of a typical catalytic reduction unit is shown
in Figure 6-10. The reactions (given in Section 3.3.2.4) in the abater are exothermic; and careful
temperature control is necessary for effective operation. The controls needed for operation as
a decolorizer are much less stringent.
Catalytic reduction units for decolorization and power recovery are used 1n about 50 nitric
acid plants in the United States. Many plants use natural gas for the reducing agent because of its
easy availability and low cost.* Some plants use hydrogen. When natural gas is used, the tail gas
must be preheated to about 753K (900F) to ensure ignition. A preheat temperature as low as 423K
(300F) is .sufficient to ignite hydrogen.
Catalytic reduction is highly exothermic. The temperature rise for the reaction with methane
is about 128K (230F) for each percent oxygen .burnout; with hydrogen it 1s about 150K (270F). For
*
Relative to hydrogen this is still true; however, the economics of using natural gas have greatly
changed in the last three years.
6-25
-------
TAIL GAS PREHEATER
SET POINT
CH4 OFF WHEN
NHa'TO CONVERTER
IS OFF
ON-OFF
SET POINT-OPEN
A HIGH 02
CH4/02
CONTROLLER
I
02
ANAILYZER
CONTROLLER
Qfl.QFF
HIGH TEMP.
SET POIWT
CH4
ANALYZER
CONTROLLER
TEMP.
RECORDER/
CONTROLLER
MOLECULAR SIEVE ~
DESULFURIZER
POKIER RECOVERY
TURBINE
COMPRESSOR
PROCESS
AIR
Figure 6-10. Nonselective catalytic reduction system (Reference 6-16).
6-26
-------
decolonization, the outlet temperature is ordinarily limited to 923K (1.200F), the maximum tempera-
ture limit of turboexpanders with current technology. Increased power recovery may justify adding
sufficient methane to reach the temperature limit of the turbine.
The tail gas must be preheated to 753K (9QOF) ,to insure ignition when methane is used as the
reducing agent. Outlet temperatures would reach 1.088K and 1.138K (1.500F and 1,590F) for 2 and 3
percent oxygen burnout, respectively. These temperatures compare to the 923K (1.200F) maximum
temperature limit for single-stage operation. The oxygen in the tail gas cannot exceed 2.8 percent
to remain within the temperature limit of the catalyst. Cooling must therefore be provided to meet
the turboexpander limit. Older turbines may have even lower temperature limitations.
A somewhat cheaper but less successful alternative is two-stage reduction for abatement.
One system involves two reactor stages with interstage heat removal (Reference 6-17). Another
two-stage system for abatement involves preheating 70 percent of the feed to 753K (900F), adding fuel,
and passing the mixture over the first-stage catalyst. The fuel addition to the first stage is
adjusted to obtain the desired outlet temperature. The remaining 30 percent of the tail gas, pre-
heated to only 393K (250F), is used to quench the first stage effluent. The two streams plus the
fuel for complete reduction are mixed and passed over the second-stage catalyst; the effluent
passes directly to the turboexpander. This system avoids high temperatures and the use of coolers
and waste heat boilers (References 6-18, 6-19, and 6-20). >
Honeycomb ceramic catalysts have been employed in two-stage abatement, with hourly gas-space
velocities of about 100,000 volumes per hour per volume in each stage (Reference 6-21).
Nonselective catalyst systems are offered by D. M. Weatherly, C & I Girdler and Chemico.
These systems are not as popular as NO control methods because of rising fuel costs.
Two or three plants are known to have installed single-stage nonselective abaters. They are
believed to have been designed for natural gas. As noted above, oxygen concentration cannot exceed
about 2.8 percent. The reactors must be designed to withstand 1.088K to 1,118K (1.500F to 1.550F)
at 790 to 930 kPa, which requires costly refractories or alloys. Ceramic spheres are used as cata-
lyst supports, at hourly gas space velocities; up to 30,000 volumes per hour per volume. One company
reports that they have been able to maintain NOX levels of 500 ppm or less over an extended period of
time. Operation close to 300 ppm might be attainable. On a plant scale, the effluent gas must be
cooled by heat exchange or quenched to meet the temperature limitation of the turbine. It may be
practical to use a waste heat boiler to generate steam.
Commercial experience with single-stage catalytic abaters has been modestly satisfactory,
but two-stage units operating on natural gas have not been as successful. Two-stage units designed
6-27
-------
for abatement have frequently achieved abatement for periods of only a few weeks, at which point
declining catalyst activity results in increasing NO levels. Recent data indicate that successful
abatement can be maintained for somewhat longer periods. Units that no longer abate NO emissions
can, however, continue to serve for energy recovery and decolorization.
*
The success of single-stage abaters compared to the limited success of two-stage units may
result from the following factors: the catalyst is in a reducing atmosphere, the temperatures are
higher, and spherical rather than honeycomb catalyst supports are used. It has not been practical $o
change catalyst type in two-stage units because the reactors designed for a space velocity of
100,000 volumes per hour per volume would be too small to accommodate a spherical catalyst, which
effectively removes NO at a space velocity of about 30,000. The failure of the honeycomb catalyst
1n NO reduction compared to its success in decolorization may reflect that reaction kinetics make
1t much more difficult to reduce NO than NOg.
Fuel requirements for nonselective abatement with methane are typically 10 to 20 percent
over stolchiometric. Some hydrocarbons and CO appear in the treated tail gas. Furthermore, not
all methane is converted in decolorization reduction units. Less surplus fuel is required when
hydrogen is used.
Selective Catalytic Reduction
In selective catalytic reduction, ammonia is reacted with the NO to form N~. No large
temperature rise occurs for ordinary operating conditions, so no waste heat or steam is generated.
The catalyst used in selective abatement units is platinum on a honeycomb support. Many catalytic
systems are installed between the expander and the economizer heat exchanger, and operate at
ambient pressure. This lack of pressure sensitivity is an advantage for retrofitting older low-
pressure nitric acid plants. It is important to control the temperature between 483K and 543K
(41OF and 518F) because above 543K, ammonia may oxidize to form NO ; below 483K, it may form ammonium
nitrate.
Gulf Oil Chemicals is the main licensor of selective catalytic abaters in North America.
They have eight systems onstream, and two more planned. Of these systems, nine operate at
ambient pressure, and one operates at 590 kPa (86 psia). Many of these catalyst beds also use a
molecular sieve for N02 adsorption to promote the reaction with ammonia.
Uhde licenses the BASF selective catalyst process and recommends it for tail gas treatment
of 600 kPa (87 psig) nitric acid plants.
/
User experience with these processes has been good. Catalyst lifetimes of over 2 years have
been reported, and expected lifetime is 5 to 10 years. Catalytic processes have also been used to
supplejnent chilled absorption units when they have failed to meet emission limits.
6-28
-------
6.1.3.5 Molecular Sieve Adsorption
The main equipment in a molecular sieve adsorption system is in the form of a two-section
packed bed. The first section is packed with a desiccant, since the NOX adsorption sieve material
works best on a dry gas. The second section contains the material which acts as nitrous oxide
oxidation catalyst and NO adsorber.
Figure 6-11 is a schematic of a molecular sieve system added to an existing nitric acid
plant. NO removal is accomplished in a fixed bed adsorption/catalyst system. The water-saturated
nitric acid plant absorption tower overheat stream is chilled to 283K (50F), the exact temperature
level being a function of the NOX concentration in the tail-gas stream. It is then passed through
a mist eliminator to remove entrained water and acid mist. The condensed water, which absorbs
some of the NOg in the tail gas to form a weak acid, is collected in the mist eliminator and either
recycled to the absorption tower or sent to storage. The tail gas then passes through a molecular
sieve bed where the special properties of the NOX removal bed material results in the catalytic
conversion of nitric oxide (NO), to nitrogen dioxide (N02). This occurs in the presence of the low
concentrations of oxygen typically present in the tail-gas stream. Nitrogen dioxide is then selec-
tively adsorbed.
Regeneration is accomplished by thermally cycling (or swinging) the adsorbent/catalyst bed
after it completes its adsorption step and while it contains a high adsorption loading of NO-- An
oil-fired heater is used to provide heat for regeneration. The required regenerator gas is obtained
by using a portion of the treated tail-gas stream for desorption of the NOg. This N02-loaded gas
is recycled to the nitric acid plant absorption tower. The pressure drop in the molecular sieve
averages 34 kPa (5 psi) and NOX outlet concentration averages 50 ppm (Reference 6-22).
This process has been applied to three plants in the United States (Reference 6-6). Tables
6-2 and 6-3 show the performance of the system at two installations. The commercial name for
the process is the Purasiv N process. The unit at the 50 Mg/d (55 tons/day) acid plant of;.
Hercules, Inc. started up in 1974. Abatement ranged from 95.9 to 98.7 percent averaged over indivi-
dual cycles and was highest at the beginning of a cycle (Reference 6-23). The U.S. Army Holston
Purasiv N unit was started up in August 1974, but has been inoperable for several years.
Both plants have dual-unit NO adsorbers, operating on a 4 hour adsorption, 4 hour regenera-
tion cycle (Reference 6-22). Initial reports on the operation were very favorable; the effluent
standards were met, and the sieve showed no noticeable deterioration after 6 months. One sieve was
damaged by accidental acid back-up, however, and did not achieve the expected 50 ppm outlet concen-
tration.
6-29
-------
ABSORBER
(EXISTING)
TAIL GAS
CONTAINING NOX
I *-
1
i
I
! i
j i-.
HOT GAS
CONTAINING
DESORBEDN02
1
S
^,
/
^,
NO OXIDATION TO N02
AND N02[,~H20
ADSORPTION
REGENERATION
X.
CLEAN DRY /~~~^
TAIL GAS f ^\
-*-. HEAT
EXCHANGER
j (EXISTING)
1
POWER
RECOVERY
(EXISTING)
FGAS
Figure 6-11. Molecular sieve system (Reference 6-22).
6-30
-------
TABLE 6-2. PERFORMANCE OF HERCULES PURASIV N UNIT DURING THREE.DAY
RUN (REFERENCE 6-24)
o>
NOV in Effluent
A
Average, ppm
Range, ppm
NOV in Feed
A
Average, ppm
Range, ppm
Average Gas Flow
Tail gas flow, Nm3/sa
Recycle gas flow,
NraVs
Total gas flow, Nm3/s
May 16
2
0-6
2,600
2,000-3,000
2.29
0.45
2.74
May 27
2
0-7
2,400
2,300-2,500
2.19
0.45
2.64
May 28
5
0-25
2,450
2,300-2,500
2.17
0.45
2,62
alNmVs = 2118.9 scfm
-------
Ol
I
UI
ro
TABLE 6-3. PERFORMANCE OF U.S. ARHY-HOLSTON PURASIV N UNIT DURING
THREE DAY RUN (REFERENCE 6-24)
(PLANT LARGELY DISMANTLED}
NOX 1n Effluent
Average, ppm
Range, ppm
NOV in Feed
A
Average, ppm
Range, ppm
Average Gas Flow .
Tail gas flow, Nm3/sa
Recycle gas flow,
Nms/s
Total gas flow, Nm3/s
August 17
<1
0-2
4,100
3,000-5,700
2.08
0.52
2.60
August 18
6
0-30
3,700
3,500-4,400
2.08
0.52
2.60
August 19
7
0-30
3,900
2,500-4,700
2.08
0.52
2.60
Nm3/s =2118.9 scfm
-------
The process has been successful in meeting emission standards. The principal criticisms have
been high capital and energy costs, and the problems of coupling a cyclic system to a continuous
acid plant operation. Furthermore, molecular sieves are not considered as state-of-the-art
technology. . ...
6.1.4 Costs
The most recent cost and energy utilization comparisons of the various abatement processes
are given in Tables 6-4 and 6-5 (Reference 6-6, 1975 costs). Direct comparison of these data is
rather difficult since not all the side effects, such as changes in plant yield, and the degree of
abatement, are described.
Chilled Absorption
The cost figures in Table 6-4 for the CDL/VITOK process are in agreement with data provided
by Reference 6-25 (1976 costs). According to Reference 6-10, the bottom line costs for the chilled
absorption process used by the TVA is $2.09/Mg ($1.90/ton) of acid, which includes $0.14/Mg
($0.13/ton) credit for additional product, 8 kWh and 85 kg steam per Mg acid (170 Ib/ton). This
cost is higher than the $1.74 Mg ($l.i>8/ton) given in Table 6-4 and does not include the reduction
in capacity caused by the reduction in the nitric acid concentration.
Extended Absorption
The Grande Paroisse process is capital intensive; therefore, costs may be dominated by the
assumptions made to calculate return on investment and depreciation. The figures in Table 6-4
reflect a 20 percent return on capital. The Grande Paroisse literature shows a cost of $0.98 to
$1.13 per Mg ($0.89 to $1.03/ton) but does not consider a return on investment cost.
Even with high capital cost and unfavorable cost of capital, extended absorption is competi-
tive with other processes. It has low maintenance costs and low energy requirements.
Wet Chemical Scrubbing
The economics and energy use of two wet scrubbing processes, MASAR, urea scrubbing, and Good-
pasture, ammonia scrubbing, are given in Tables 6-4 and 6-5. Costs for the Norsk Hydro process
would be similar if applied to a new plant. Capital costs for the Goodpasture process are estimated
as $425,000 for a 270 Mg/d (300 tons/day) plant (Reference 6-26). No costs estimates are available
for potassium permanganate and caustic scrubbing since they are not in general use.
6-33
-------
TABLE 6-4. CAPITAL AND OPERATING COSTS FOR DIFFERENT MOX ABATEMENT
SYSTEMS IS A 270 Mg/d NITRIC ACID PLANT (References 6-6 and 6-26 P
o>
to
Capital Investment, ($)
Royalty
Operating Labor, (hr/yr)
($/yr)
Maintenance Labor,
(*/yr)
Labor Overhead (Incl. fringe
benefits & supervision, $/yr)
Catalyst or Molecular Sieve
Cooling Water, (l/m1n)
tt/yr)
Steam, (kg/hr)
($/yr credit)
Electricity, (kWh)
($/yr)
. Boiler Feed Mater, (1/mln)
($/yr)
Fuel, m
($/yr)
Nitric Add, (Hg/day)
($/yr)
Urea, (Mg/day)
($/yr)
Ammonium Nitrate, (Hg/day)
($/yr)
Depreciation (11-yr. life)
Return on Investment (@ 202)
Taxes & Insurance (@ 2%)
Total Annual Cost, ($/yr)
Annual Cost, ($/Mg)
Catalyst
Reduction
1,384,000
360
2,200
315
2,200
4,400
77,800
--
(7,182)
(387,590)
128
20,890
132
12,850
8.3
465,120
--
125,900
276,800
27,700
628,270
6.79
Molecular
Sieve
1,200,000
360
2,200
315
2,200
4,400
45,600
1,892
7,330
113
6,120
322
52,550
0.6
32,640
(6.0)
(112,200)
109,090
240,000
24,000
413,930
4.48
Grande
Paroisse
1,000,000
Included
360
2,200
315
2,200
4,400
1,135
4,420
__
__
90
14,690
__
__
(5.4)
(102,000)
90,910
200,000
20,000
236,780
2.56
CDL/
Vitok
575,000
None
360
2,200
315
2,200 "
4,400
3,861
14,980
324
17,500
265
43,250
__
__
(5.4)
(102,000)
52,300
115,000
11,500
161,330
1.74
Hasar
663,000
Fee
360
2,200
315
2,200
5,975
.- -
594
32,070
20
3,260
__
__
(4.8)
. (89,760)
1.24C
74,528
(1.13)
(24,500)
60,300
132,600
13,260
195,708
2.12
Goodpasture
425,000
51 ,000
360
2,200
315
2,200
4,400
*
114
440
--
45
7,340
--
--
~
--
--
01.8)
(422,000)
38,640
85,000
8,500
(42,290)
(0.46)
aThis table is given in Appendix A 1n English units.
Investment estimates exclude Interest during construction, owners expenses, aid land costs.
clnclude credit for 0.0017 Hg of urea/Hg of nitric acid produced present in the spent
solution (D.SITPD).
Parenthesis indicate credit taken.
-------
TABLE 6-5. ANNUAL ENERGY REQUIREMENTS (TJ) FOR NOX ABATEMENT SYSTEMS
FOR A 270 Mg/d NITRIC ACID PLANT (Reference 6-6 and 6-26 )a
Steam (Credit)
Electrical
Natural Gas
Oil
Basic Nitric
Acid Plant
(75.2)
172.0
96.8
Catalyst
Reduction
(136.18)
11.56
245.12
108.94
Molecular
Sieve
2.15
29.08
-
17.20
48.43
Grande
Paroisse
-
8.13
-
-
8.13
CDL/
Vitok
6.14
23.94
-
-
30.08
Masar
11.27
T.80-
-
-
13.07
Goodpasture
' -
1.38
-
U38
This table is given in Appendix A in English units.
TABLE 6-6. BASIS FOR TABLES 6-4 AND 6-5 (Reference 6-6)a
(Plant Capacity 270 Mg/day and 92 Gg/yr)
(March 1975 Dollars, ENR Index = 2.126)
1. Operating Labor
2. Maintenance Labor
3. Overhead
4. Cooling Water
5. Boiler Feedwater
6. Natural Gas
7. Oil
8. Depreciation
9. Return on Investment
10. Taxes and Insurance
11. Nitric Acid
12. Urea
13. Ammonium Nitrate
14. 1 kWh * 11.07 MJ
15. Electricity
16. Ammonia
6 $6.1/hr
9 $7.0/hr
@ 100% of labor (including fringe
. benefits and supervision)
@ $0.008/1000 1
9 $0.20/1000 1
@ $1.90/GJ
@$1.90/GJ
@ 11 yr straight line
@ 2035 of capital cost
@ 2% of capital cost
@ $ 99/Mg
@ $176/Mg
9 $lJO/Mg
@ $0.02/kWh
@ $173/Mg
aThis table is given in Appendix A in English units.
6-35
-------
Capital and operating costs for these processes are very low and are aided by credit for the
byproducts (ammonium nitrate). In the Goodpasture process approximately 75 percent of the ammonia
is reclaimed as ammonium nitrate.
Catalytic Reduction
The cost and energy data given in Tables 6-4 and 6-5 are for a natural gas-fired nonselective
catalytic reduction unit. The process is considerably more expensive than the other processes. Not
only does a catalytic combustor have a high capital cost, but fuel costs are large (and will
probably increase).
Costs for selective catalytic reduction are not included in Table 6-4. Capital costs are
estimated as $100,000 to $125,000 for a 270 Mg per day unit by Gulf (Reference 6-27, 1976 costs).
Operating and maintenance costs are expected to be low except possibly for catalyst replacement.
The major operating expense is the cost of ammonia for reaction with NO .
Molecular Sieve
Both capital and operating costs for the molecular sieve process are high. Fuel for the
regeneration phase, high maintenance costs, and catalyst replacement are the primary contributors to
the operating costs. Not included in the cost figures are any extra costs which may result from
upsets or process alterations in the nitric acid plant as a result of the cyclic operation of the
abatement unit.
5.2 NITRIC ACID USES
Important uses of nitric acid and the estimated quantities consumed in each are listed in
Table 6-7. Approximately 65 percent of the nitric acid produced In the United States is consumed in
making ammonium nitrate, of which approximately 80 percent is used for fertilizer manufacturing.
Adipic acid manufacture, the second largest use, consumes only about 7 percent. Other uses include
metal pickling and etching, nitrations and oxidations of organic compounds, and production of
metallic nitrates.
6.2.1 Ammonium Nitrate Manufacture
6.2.1.1 Process Description
Ammonium nitrate is produced by the direct neutralization of nitric acids with ammonia:
, NH3 + HN03 + NH4N03 (6-10)
6-36
-------
TABLE 6-7. ANNUAL NITRIC ACID CONSUMPTION IN THE UNITED STATES, 1974
(Reference 6-3 and 6-6)
Product
Ammonium Nitrate
Adipic Acid
Nitrobenzene
Potassium Nitrate
Miscellaneous Fertilizers .
Military, other than
NH4N03
Isocynates
Steel Pickling
Other
Total Nitric Acid
Production
Quantity of HNO, used in manufacture
Gg
4830
520
74
37
371
258
m
37
1193
7431
TO3 tons
5324
573
82
40
409
286
122
41
1315
8192
6-37
-------
About 735 kg (1600 Ib) of nitric acid (TOO percent equivalent) and 190 to 205 kg (420 to 450 Ib) of
anhydrous ammonia are required to make 909 kg (1 ton) of ammonium nitrate. In actual practice, 100
percent nitric acid is not used, and typical feed acid contains 55 to 60 percent HNO,. The product
is an aqueous solution of ammonium nitrate, which may be used as liquid fertilizer or converted
into a solid product. The heat of reaction is usually used to evaporate part of the water, giving
typically a solution of 83 to 86 percent ammonium nitrate. Further evaporation to a solid may be
accomplished in a falling-film evaporator (Reference 6-28), in a disk-spraying plant (Reference
6-29), or by evaporation to dryness in a raked shallow open pan (graining). The graining process
1s no longer used due to hazardous conditions.
A majority of the solid ammonium nitrate produced in the United States is formed'by "prilling",
a process in which molten ammonium nitrate flows in droplets from the top of a tower countercurrent
to a rising stream of air, which cools and solidifies the melt to produce pellets or prills (Refer-
ence 6-3).
6.2.1.2 Emissions
No significant amount of NOX is produced in this process; the most likely source of nitric
acid emissions would be the neutralizer. The vapor pressure of ammonia, however, is much higher
than the vapor pressure of nitric acid, and the release of nitric acid fumes or NO is believed to
be negligible (Reference 6-30), especially since a slight excess of NHg is used to reduce product
decomposition.
6.2.2 Organic Oxidations
6.2.2.1 Process Description
Nitric acid is used as an oxidizing agent in the commercial preparation of adipic acid,
terephthalic acid, and other organic compounds containing oxygen. The effective reagent is probably
KOg. which has very strong oxidizing power.
Adipic acid (COOH-(CH2)4*COOH) is a di-basic acid used in the manufacture of synthetic fibers.
It is an odorless white crystalline powder which is manufactured by the catalytic oxidation of cyclo-
hexane, with cyclohexanone and cyclohexanol as intermediates. About 618 Gg (681,000 tons) of adipic
add were manufactured 1n 1975 (Reference 6-31). Approximately 90 percent of adipic acid is consumed
in the manufacture of nylon 6/6.
In the United States, adipic acid is made in a two-step operation. The first step is the
catalytic oxidation of cyclohexane by air to a mixture of cyclohexanol and cyclohexanone. In the
6-38
-------
second step, adipic acid is made by the catalytic oxidation of the cyclohexanol/cyclohexanone mix-
ture using 45 to 55 percent nitric acid. The product is purified by crystallization (Reference 6-32).
The whole operation is continuous. The chemistry of the reactions in the two steps is:
*
cyclohexanone + nitric acid * adipic acid + NOX + H-0 (6-11)
cyclohexanol + nitric acid + adipic acid + NO, + H-0 (6-12)
The main nitrogen compounds formed in the above reactions are NO, NO-, and N^O. The dissolved
oxides are stripped from the adipic acid/nitric acid solution with air and steam. The NO and N02
are recovered by absorption in nitric acid. The off-gas from the NO absorber is the major contri-
butor to NOX emissions from the adipic acid manufacturing process.
Nitric acid is used for the oxidation of other organic compounds in addition to the adipic
acid, but none approaches the adipic acid product volume.
Terephthalic acid is an intermediate in the production of polyethylene terephthalate, which
is used in polyester, films, and other miscellaneous products. Terephthalic acid can be produced
in various ways, one of which is by the oxidation of paraxylene by nitric acid (Reference 6-33).
In 1970, the process was used for about a third of terephthalic acid production and accounted for
approximately 20 percent of NO emissions from nitration processes. Since 1975, however, the use
of nitric acid as a feedstock in the production of terephthalic acid has been discontinued (Reference
6-34). No NOX is now generated in terephthalic acid plants.
6.2.2.2 Emissions
The off-gases leaving the adipic acid reactor after nitric acid oxidation of organic materi-
als may contain as much as 30 percent NO before processing for acid recovery (Reference 6-35).
One of the principal compounds of the off gas, N20, is not counted as NOX, since it is not oxidized
to NOX in the atmosphere and is considered harmless. The seven adipic acid manufacturing plants in
the United States generated about 14.5 Gg (16,100 tons) of - NOX in 1975 (Reference 6-31) from a total
acid production of 618 Gg (681,000 tons). This gives an average emission factor of 23.7 kg N02/Mg
(47.4 Ib N02/ton) compared to the nominal value 6 kg N02/Mg (12.0 Ib N02/ton) specified by AP-42
(Reference 6-36).
6.2.2.3 Control Techniques
In commercial operations, economy requires the recovery of NO as nitric acid. It is recov-
ered by mixing the off-gas with air and sending the stream to an absorbing tower, where nitric acid
is recovered as the stream descends and unrecoverable N20 and nitrogen pass off overhead.
6-39
-------
If the resulting emission rates are too high, further reduction could be attempted by stan-
dard techniques such as extended absorption or wet chemical scrubbing. These techniques are
described in Section 6.1.3. A potential, long-range control for eliminating NOX from organic oxi-
dation processing is the replacement of nitric acid as an oxidant by catalytic processes using air
oxygen. Tht laboratory catalytic oxidation of cyclohexanol and cyclohexanone by air to adipic acid
has also been reported, but no commercial process is known (Reference 6-37).
6.2.2.4 Costs - -
Economy requires that nitric acid be recovered from reactor off-gas in large-scale organic
oxidations using nitric acid as the oxidizing agent. For example, the incentive for acid recovery
for a 270 Kg/d (300 tons/day) adiplc acid plant would be about $2.48 x 10 per year. This figure is
based on recovering 0.3 kg of HN03 per kg of adipic acid at a nitric acid cost of $8.6 per 100 kg (Ref-
erence 6-38). The optimum economic recovery level depends upon economic factors at each installation.
6.2.3 Organic Nitrations
6.2.3.1 Process Description
Nitration is the treating of organic compounds with nitric acid (or N0«) to produce nitro
compounds or nitrates. The following equations illustrate the two most common types of reaction:
RH + HON02 - RN02 + HgO (6-13)
ROH + HON02 * RON02 + HgO (6-14)
Examples of products of the first reaction (C-nitration) are compounds such as nitrobenzene, nitro-
toluenes, and nitromethane. Nitroglycerin (or glycerin trinitrate) and nitrocellulose are examples
of compounds produced by the second reaction (0-nitration).
Nitrating agents used commercially include nitric acid, mixed nitric and.sulfuric acids
(mixed acids), and N02. Mixed nitric and sulfuric acid is most frequently used. The sulfuric acid
functions to promote formation of N02 ions and to absorb the water produced in the reaction.
i-
Nitrations are carried out in either batch or continuous processes. The trend is toward
continuous processes, since control is more easily maintained, equipment is smaller, system holdup
is smaller, and hazards are reduced. A multiplicity of specialty products such as dyes and drugs,
which are'produced in small volumes, will continue, however, to be manufactured by small batch
nitrations.
6-40
-------
Batch nitration reactors are usually covered vessels provided with stirring facilities and
cooling coils or jackets. The reactor bottom is sloped, and product is withdrawn from the lowest
point. When products are potentially explosive, a large tank containing water (drowning tank) is
provided so that the reactor contents can be discharged promptly and "drowned" in case of abnormal
conditions.
When the reaction is completed, the reactor contents are transferred to a separator, where
the product is separated from the spent acid. The product is washed, neutralized, and purified;
spent acids are'processed for recovery. Figure 6-12 illustrates a batch nitration process for
manufacturing nitroglycerin (Reference 6-39).
Continuous nitration for nitroglycerin is carried out in many types of equipment. Two widely
employed processes are the Schmid-Meissner process (illustrated in Figure 6-13) and the Biazzi pro-
cess (illustrated in Figure 6-14)., Both processes provide for continuous reaction, separation,
water washing, neutralization, and purification. The Biazzi process makes greater use of impellers
for contacting than the Schmid-Meissner, which uses compressed air to provide agitation during
washing and neutralizing. Both types of equipment can be used for nitrating in general.
When mixed acid is used, the spent acid is recovered in a system similar to that shown in
Figure 6-15. The mixed acid enters the top of the denitrating tower. Superheated steam is admitted
at the bottom to drive off the spent nitric acid and NOX overhead. The gases are passed through a
condenser to liquefy nitric acid, which is withdrawn to storage; the uncondensed gases are then .
sent to an absorption tower. Weak sulfuric acid is withdrawn from the bottom of the denitrator
tower and concentrated or disposed of by some convenient arrangement.
When nitric acid alone is used for nitration, the weak spent acid is normally recovered by
sending it to an absorption tower, where it replaces some of the water normally fed as absorbent.
Nitrobenzene and dinitrotoluenes are produced in large volumes as chemical intermediates.
Explosives such as TNT, nitroglycerin, and nitrocellulose are produced in significant but lesser
volumes.
Nitrobenzene is manufactured in both continuous and batch nitration plants. Mixed acids
containing 53 to 60 percent H2SO., 32 to 39 percent HNOj, and 8 percent water are used in batch
operations, which may process 3.785 m* (1000 gallons) to 5.678 m* (1500 gallons) of benzene in 2 to
4 hours. Continuous plants, as typified by the Biazzi units (Figure 6-14) also use mixed acids.
The major use of nitrobenzene is in the manufacture of aniline. It is also used as a solvent.
Nitrobenzene production in 1970 was an estimated 188 Gg (207,500 tons). Nitric acid requirements
6-41
-------
NITRATING HOUSE
wATrn t ^ nnnwtVFR «_-»i
(ED
;iu 1 ,
r " ~ /
^ MIXED-ACID ^ MIXED-ACID . NiTRATOR » SEPAI
^ STORAGE, SCALE TANK * NITRATOR » SEPAf
> i
1
LYCERIN -L GlYCERIN 1 \ GLYCERIN
LYCERIN j GI.Y.fcKlN *SCALETABrK J
GLYCERIN ' y^
HEATER X
HOUSE y^
SPENT ACID WATER
1 L
SPEKT-ACID i L
SO
WA'
UR.
fER
WE! NG
1ATOR -NG-*.
WATER
HOT
WATER
SODA
WATER
PREWASH
*
S i-
N
t ' >
STORAGE SODA ASH V
SODA
WATER
N
JL_
T
NG
5
1 <-^-
CATCH
TANK
1
WASTE
WATER
J NEUTRALIZING
, HOUSE
EUTRALIZER *J
/WATER
_! '_ ^ _^^_^ . " / X
, j WASIE
|
NITRIC ACID TOWER
RECOVERY HOUSE |
J_l-1ZJ__^.l_iJ N
' 1 1
BUGGY
NG NG
CATCH
TANK
i r_
- .
CATCH
TROQLYCERIN TANK
TO POWDER
1
1
RECOVERED WEAK
NITRIC SULFURIC
ACID ACID
WASTE
WATER
Figure 6-12. Batch process for the manufacture of nitroglycerin (NG) (Reference 6-39).
642
-------
GLYCERIN
FEED
AGITATOR
NITROGLYCERIN
M
WASH
WATER
TO BAFFLED
SETTLING TANKS
MIXED
Adl
WATER
SEPARATOR
J
A
(
NITROGLYCERIN
SODA WATER*»
eg.
<0
3"
AIR
V
AIR
TO DROWNING
TANK
WATER
AIR
L
STORAGE
TANK
.NITROGLYCERIN.
WATER
(PASSED THROUGH ADDITIONAL NITROGLYCERIN
WASH COLUMNS IF NECESSARY)
Figure 6-13. Schmid-Meissner continuous-nitration plant (Reference.6-39).
6-43
-------
AGITATOR
GLYCERIN
FEED
i
ACID
AGITATOR
WATER
i
NITROGLYCERIN
NITRATOR
f
TO-
DROWNING
LINE OF
SEPARATION
SPENT ACID
TO
LEVELING DEVICE
ACID X
SEPARATOR 1
SEPARATOR
LEVELING
DEVICE
NITRO- )
GLYCERIN'
ACID
MECHANICAL
WASHER
TO DROWNING
TANK
SPENT ACID
TO AFTER-SEPARATOR
AND STORAGE
~L
WASTE
-WATER1
WATER
SEPARATOR
AGITATOR
NITROGLYCERIN
LEVELING
DEVICE
SODA ASH
SOLUTION
t_
MECHANICAL
NEUTRALIZER
i
NITROGLYCERIN
EMULSION
TO ADDITIONAL
WASHERS AND
SEPARATORS
T
WASH WATER TO SEPARATOR
Figure 6-14. Biazzi continuous-nitration plant (Reference 6-39).
6-44
-------
GASES TO
ABSORPTION TOWER
S-BEND
CONDENSER
SPENT-ACID
FEEDTANK
CHEMICAL-WARE
BV.OCKCOCK
NITRIC ACID
TO STORAGE
NITRIC
DISTILLATE
SAMPLER
SULFURICACID
COOLING TUB
Figure 6-15. Recovery of spent acid (Reference 6-39).
6-45
-------
are approximately 0.54 kg per kg of nitrobenzene (Reference 6-39). On this basis, nitric acid used
1n nitrobenzene synthesis was estimated at 126 Gg (139,000 tons) for 1970.
Dinitrotoluene is manufactured in two stages in both continuous and batch units. The first
stage'is the nitration of toluene to mononitrotoluene, which is nitrated to dinitrotoluene in the
second stage. For making mononitrotoluene in the batch process, mixed acid consisting of 28 to 32
percent HNO,, 52 to 56 percent H2S04, and 12 to 20 percent water is used in equipment sized to
handle up to 11.4 m* (3000 gallons^. Operating temperature ranges from 298K (77F) to 313K (104F).
Mononitrotoluene yields of 96 percent are typical (Reference 6-40). The second step, the production
of dinitrotoluene, 1s carried out separately because it requires more severe conditions.
Dinitrotoluene is made from mononitrotoluene using stronger mixed acid containing 28 to 32
percent HN03, 60 to 64 percent HgSO^, and 5 to 3 percent water. Temperatures are increased to 363K
(194F) after all the acid has been added. Dinitrotoluene yields are about 96 percent of theoretical
(Reference 6-41).
The principal use of dinitrotoluene is as intermediate in making toluene diisocyanate (TDI)
for use in polyurethane plastics. It is usually supplied as mixtures of the 2,4 and 2,6 isomers.
6.2.3.2 Emissions
Relatively large NOX emissions may originate in nitration reactors and in the denitration of
the spent acid. NOX is also released in auxiliary equipment such as nitric acid concentrators,
nitric add plants, and nitric acid storage tanks.
Nitration reactions per se do not generate N0₯ emissions. NOV is formed in side reactions
~ * . A i
Involving the oxidation of organic materials. Relatively little oxidation and NOV formation occur
X
when easily nitratable compounds, such as toluene, are processed. Much more severe conditions are
required 1n processing compounds that are difficult to nitrate, such as dinitrotoluene; more oxida-
tion takes place and, thus, more NOX is formed.
Limtied data are available on actual NO emissions from nitrations. For continuous nitra-
tions, one company has reported emissions of 0.06 to 0.12 kg N02 per Mg of nitric acid (0.12 to 0.24
Ib/ton), with a mean of 0.09 kg N02/Hg (0.18 Ib/ton 1 at a stngle location CReference 6-40). At the
same location, emissions averaging 7 kg of N02 per Mg of acid were reported in manufacturing specialty
products 1n small batch-type operations. According to Reference 6-42, 0.25 kg of N02 per Mg of nitric
acid (0.5 Ib/ton) are generated in'the production of nitrobenzene. In the manufacture of dinitrotol-
uencs, 0.135 kg of N02 is estimated to be generated for every Mg of nitric acid used (0.27 Ib/ton).
6-46
-------
Using the Reference 6-42 emission factors as a lower limit, and 7 kg NOX per Mg HN03 (14 Ib/ton),
(Reference 6-40) as upper limits for nitrations, the NOX emissions in 1970 would have the range indi-
cated in Table 6-8. Even using the upper limit, NOX emissions from nitrobenzene and dinitrotoluene
synthesis are relatively small but may present local nuisance problems. Since the upper limit
represents specialty batch operations on a small scale, the emissions are. probably much higher than
would be encountered in large volume production of these products in either batch or continuous
equipment.
6.2.3.3 Control Techniques
In large batch or continuous nitrations, operations are carried out in closed reactors.
Fumes are conducted from the reactor, air is added, and the mixture enters an absorption tower for
recovery of nitric acid. If too much NOg remains in the residual gas from the absorber, it may be
further reduced by techniques such as wet chemical scrubbing. Details of the control techniques
are discussed in Section 6.1.3. ..
Noncondensable gas from acid denitration is treated.in the same manner as reactor gas. A
common absorber is sometimes employed.
Small batch nitrators used in manufacturing specialties such as drugs and dyes are small-
volume, high-intensity NOX emitters. In ones plant-, reaction times ranged from 3 to 12 hours, depend-
ing on the product made. From 3 to 850 batches of each product were made each year. Emissions
ranged from 0,7 to 130 kg of NOX per Mg of nitric acid (0.14 to 260 Ib/ton) with a median of 21 kg per
Mg of nitric acid (21 Ib/ton). The median emission was 7 kg per Mg (14 Ib/ton) when one product was
excluded from the calculations. The emissions, which are vented in individual stacks, are brown in
color for a few hours per batch.
Caustic scrubbing and NOX incineration are regarded as the most plausible controls for
specialty batch nitrations. Catalytic reduction is usually ruled out because of organic and other
impurities in the gas. Neither control is considered highly efficient in this application.
The intermittent character of emissions makes them difficult to control and contributes to
very high pollution abatement costs per ton of nitric acid consumed. According to DuPont, operating
costs for such equipment would render approximately half of the small batch nitrations so unecono-
mical that the manufacture of these products would be terminated (Reference 6-40). Large batches
may be suitable for conversion to continuous operating, but small batches are not.
-------
TABLE 6-8. ESTIMATED. NOv EMISSIONS FROM ORGANIC NITRATIONS
IN 1970 (REFERENCE 6-42).
O)
Product
Nitrobenzene
Blnitrotoluene
1970 Production
Mg
1 (tons)
233,600
(257,000)
131,542
(145,000)
Estimated HNOa
consumption
Mg
(tons)
126,099
(139,000)
101,151
(111,500)
NOX Emissions
Mg
(tons)
Lower Upper
limit limit
883
(973)
708
(780)
-------
6.2.3.4 Costs
Fume Incinerator investments are quoted at $10,000 to $20,000 by one source (Reference 6-43,
1966 costs). Another suggests that investments of $75,000 to $150,000 are necessary for flame
abatement facilities for existing small batch nitrators and $75,000 to $250,000 for existing large
nitrators. Annual operating costs were estimated at $25,000 to $85,000 per product for small batch
nitrators and $25,000 to $40,000 for continuous nitrators (Reference 6-40).
6.2.4 Explosives: Manufacture and Use
6.2.4.1 Process Description
An explosive is a material that, under the Influence of thermal or mechanical shock, decom-
poses rapidly and spontaneously with the evolution of large amounts of heat and gas. Explosives
fall into two major categories: high (industrial) explosives and low explosives.
Industrial explosives in the United States consist of over 80 percent by weight of ammonium
nitrate and some 10 percent of nitro organic compounds. During 1975, an estimated 1.4 Tg (3.1 x .
10* pounds) of industrial explosives were manufactured, which is about 13 percent higher than the
1974 productions (Reference 6-44). High explosives are less sensitive to mechanical or thermal
shock, but explode with great violence when set off by an initiating explosive (Reference 6-45).
Low explosives, such as nitrocellulose, undergo relatively slow autocombustion when set off and
evolve large volumes of gas in a definite and controllable manner.
Production and consumption data for military explosives are classified. Some of the more-
important ingredients in military explosives are known, however: trinitrotoluene (TNT), penter-
ythritol tetranitrate (PETN), cyclotrimethylene-tri-nltramine (RDX), and trinltrophenylmethyl-
nitramine (Tetryl). Nitration is an essential step in the manufacture of each of these.
PETN is most commonly used in conjunction with TNT in the form of pentolites, made by incor-
porating PETN into molten TNT. RDX is used 1in admixture with TNT, or compounded with mineral jelly
to form a useful plastic explosive. Tetryl is most often used as a primer for other less sensitive
explosives.
TNT (symmetrical trinitrotoluene) may be prepared by either a continuous process or a batch,
three-stage nitration process using toluene, nitric acid, and sulfuric acid as raw materials. In
the batch process, a mixture of oleum (fuming sulfuric acid) and nitric acid that has been concen-
trated to a 97 percent solution is used as the nitrating agent. The overall reaction may be
expressed as:
Ot + H,S04* °,N | Oj NO^
CH + 3HONO + HS0* °N O NO + 3H 0 +' H SO (6-15)
6-49
-------
Spent acid from the nitration vessels is fortified with makeup 60 percent nitric acid before
entering the next nitrator. Fumes from the nitration vessels are collected and removed from the
exhaust by an oxidation-absorption system. Spent acid from the primary nitrator is sent to the ac^'d
recovery system in which the sulfuric and nitric acid are separated. The nitric acid is recovered
as a 60 percent solution, which is used for refortification of spent acid from the second and third
nitrators. Sulfuric acid is concentrated in a drum concentrator by boiling water out of the dilute
add. The product from the third nitration vessel is sent to the wash house at which point asym-
inetrical isomers and incompletely nitrated compounds are removed by washing with a solution of
sodium sulfite and sodium hydrogen sulfite (Sellite). The wash waste (commonly called red water)
from the purification process is discharged directly as a liquid waste stream, is collected and sold,
or is concentrated to a slurry and incinerated in rotary kilns. The purified TNT is solidified,
granulated, and moved to the packing house for shipment or storage. A schematic diagram of TNT pro-
duction by the batch process is shown in Figure 6-16.
Nitrocellulose is prepared by the batch-type "mechanical dipper" process. Cellulose, in the
form of cotton linters, or specially prepared wood pulp, is purified, bleached, dried, and sent to
a reactor (niter pot) containing a mixture of concentrated nitric acid and a dehydrating agent such
as sulfuric acid, phosphoric acid, or magnesium nitrate. The overall reaction may be expressed as:
C6H?02(OH)3 + 3HON02 + H2S04 * C6H702(ON02}3 + 3 HgO + H2S04 (6-16)
When nitration is complete, the reaction mixtures are centrifuged to remove most of the spent acid.
The spent acid is fortified and reused or otherwise disposed. The centrifuged nitrocellulose under-
goes a series of water washings and boiling treatments for purification of the final product.
6.2.4.2 Emissions
The major emissions from the manufacture of explosives are nitrogen oxides and nitric acid
mists. Emissions of nitrated organic compounds may also occur from many of the TNT process units.
In the manufacture of TNT, vents from the fume recovery system, and nitric acid concentrators are
the principal sources of emissions. Emissions may also result from the production of Sellite
solution and the incineration of "red water". Many plants now sell the red water to the paper
industry where it is of economic importance.
Principal sources of emissions from nitrocellulose manufacture are from the reactor pots and
centrifuges, spent acid concentrators, and boiling tubs used for purification.
The most important factor affecting emissions from explosives manufacture is the type and
efficiency of the manufacturing process. The efficiency of the acid and fume recovery systems for
6-50
-------
TOLUENE-
EXHAUST
H
ELECTROSTATIC
PRECIPITATOR
SULFURIC ACID !
CONCENTRATOR
HOT GAS
1
FURNACE
H2S04
TANKS
t
FUEL:
H2S04
,GO%HN03 BI-AC10 60%HN03 TRI-ACID OLEUM
I f
MONO-HOUSE
1
r t i *
BI-HOUSE
1 | f
TRI-HOUSE
1 41
.SPENT MONO-OIL FUMES BI-OIL TRl'oiL
j FUMES *,
DENITRATOR
r FUMES
OXIDATION
CHAMBER
4
1 ' '
OXIDATION
TOWERS AND!
SEPARATORS
PURIFI
TNT
f
BUBBLE CAP
TOWER
VE
1 '
BUBBLE CAP
TOWER
1
VENT
HN03
TANKS
WASH
HOUSE
ED RED WATER i«
93
H2£
97% HN03 6Q%
| HN03
VENT
1 ,
NITRIC ACID
CONCENTRATOR
i 1 Na2C
LLITE EXHAUST 1 1
LUTION 1 ,BEsinUAL 1
f^ 1 i K2SG4 i
EVAPORATORS
WASTE LIQUOR
MT HOT GAS 1
1 1 t
FURNACE
ROTARY
KILNS
%
.04
03
SELLITE
PLANT
H2fl
STE/
OLEUM EXHAUST
GAS
,
IM
FUEL Na2S04 EXHAUST
Figure 6-16. Trinitrotoluene (batch process) manufacturing diagram (Reference 6-45).
-------
TNT manufacture will directly affect the atmospheric emissions. In addition, the degree to which
acids are exposed to the atmosphere during the manufacturing process affects the NO emissions. For
nitrocellulose production, emissions are influenced by the nitrogen content and the desired quality
of the final product. Operating conditions will also affect emissions. Both TNT and nitrocellulose
are produced in batch processes. Consequently, the processes may never reach steady state and emis-
sion concentrations may vary considerably with time. Such fluctuations in emissions will influence
the efficiency of control methods. Table 6-9 presents the emission factors for the manufacture of
explosives and the effects of various control devices upon emissions (Reference 6-45). Although the
manufacture of explosives is a very small source of NO emissions nationwide, explosions could 'be an
Intense source in confined underground spaces. Precautions should be taken to avoid chronic exposure.
6.2.4.3 Controls
Explosives manufactured by the commercial industry use ammonium nitrate extensively as the
base material. The ammonium nitrate production process is reviewed in Section 6.2.1. Nearly half
the plants use the catalytic reduction technique for control of NOV emissions.
A
The military explosives which are produced in large amounts include nitroglycerin, nitrocellu-
lose, TNT, and RDX. The molecular sieve abatement system is used at Holston Army Ammunition Plant
1n Klngsport, Tennessee. Another Army Ammunition Plant at Radford, Virginia, is constructing two
molecular sieve units to treat vent gas streams from their nitrocellulose plant. The description
of the molecular sieve control technique is included in Section 6.1.3.5.
6.2.4.4 Costs
Costs for controlling NOX from explosives manufacture by tail gas treatment process were
covered in Section 6.1.4.
6.2.5 Fertilizer Manufacture
Sulfuric and phosphoric acids are the principal acids used, in the United States, in acidu-
lating phosphate rock. A few manufacturers produce "nitric phosphate" fertilizers by acidulating
phosphate rock with nitric acid to form phosphoric acid and calcium nitrate. In subsequent steps,
ammonia is added with either carbon dioxide or sulfuric or phosphoric acid, and "nitric phosphates"
are formed. Dibasic calcium phosphate and ammonium nitrate are the useful compounds produced
(Reference 6-48).
U.S. Department of Agriculture statistics do not segregate nitric phosphate fertilizers made
by acidulation of phosphoric rock; but private sources indicate that nitric phosphate fertilizer
6-52
-------
TABLE 6-9. EMISSION FACTORS FOR MANUFACTURE
OF EXPLOSIVES (REFERENCE 6-45)
Type of process
TNT - batch process b
Nitration reactors
Fume recovery
Acid recovery
Nitric acid concentrators
Red water incinerator
Uncontrolled0
Wet scrubber
Sellite exhaust
TNT continuous process6
Nitration reactors
Fume recovery
Acid recovery
Red water incinerator
Nitrocellulose6
Nitration reactors
Nitric acid concentrator
Nitrogen oxides3
(N02)
kg/Mg
12.5(3-19)
27.5(0.5-68)
18.5(8-36)
13(0.75-50)
2.5
«
4(3.35-5)
1.5(0.5-2.25)
3.5(3-4.2)
7(1.85-17)
7(5-9)
Ib/ton
25(6-38)
5^(1-136)
37(16-72)
26(1.5-101)
5
%
8(6.7-10)
3(1-4.5)
7(6.1-8.4)
14(3.7-34)
14(10-18)
For some processes considerable variations in emissions have been reported.
The average of the values reported is shown first, with the ranges given
in parentheses. Where only one number is given, only one source test was
available.
Reference 6-46
cUse low end of range for modern, efficient units and high end of range for
older, less efficient units.
Apparent reductions in NOX and particulate after control may not te sig-
nificant because these values are based on only one test result.
6Reference 6-47
For product with low nitrogen content (12 percent), use high end of range.
For products with higher nitrogen content, use lower end of range.
6-53
-------
made in this manner was estimated at 450 Gg (500,000 tons) in 1967, and nitric acid consumptions
at 135 Gg (150,000 tons) (Reference 6-49).
NOX emissions are dependent on the quantity of carbonaceous material in the rock, since NO
is formed as nitric acid oxidizes the carbonaceous matter. The use of calcined rock avoids the
production of NOX.
Air pollution abatement by fertilizer manufacturers' efforts has centered on reducing particu-
lates and fluorides emissions, which are severe problems. The water scrubbing used to.reduce these
pollutants would be expected to reduce NOX emissions to only a minor degree. Although no measure-
ments of NOX emissions are available, brown plumes are said to occur.
One company has found that the addition of urea to the acidulation mixture reduces NOV emis-
X
sions and eliminates the brown plume (Reference 6-49). Urea, as discussed in Section 6.1.3.3
reacts with nitric and nitrous acids to form N2-
6.2.6 Metals Pickling
The principal use of nitric acid in metals pickling is 1n treating stainless steel. Mill
scales on stainless steels are hard and are difficult to remove. Pickling procedures vary; some-
times a 10 percent sulfuric acid bath at 333K (140F) to 344K (160F) is followed by a bath at 328K
(130F) to 339K (150F) with 10 percent nitric acid and 4 percent hydrofluoric acid. The first bath
I
loosens the scale, and the second removes it. A continuous system for stainless steel strip con-
sists of two tanks containing 15 percent hydrochloric acid, followed by a tank containing 4 percent
hydrofluoric and 10 percent nitric acid-at 339K (150F) to 350K (170F). One effective method is the
use of molten salts of sodium hydroxide to which is added some agent such as sodium hydride. This
may be followed by a dilute nitric acid wash (Reference 6-50).
No measurements were found of emission rates from nitric acid pickling of stainless steel.
Treating equipment should be properly hooded and ventilated and the fumes scrubbed to protect
workers. Urea would probably control the NOX emissions.
Nitric acid is also used in the chemical milling of copper or iron from metals that are .not
chemically attacked by nitric acid, and for bright-dipping copper. In the latter operation, a cold
solution of nitric and sulfuric acid has been customarily used. It has been reported that copper
can be bright-dipped on cold nitric acid alone when urea is added. A highly acceptable finish is
obtained, and NOX fumes are eliminated.
6*54
-------
Sulfuric acid should not be used with the nitric acid-urea mixture since nitrourea, an explo-
sive, can form. Not more than 62 ml of urea per liter should be added, and satisfactory operation
can be obtained with only 15 ml per liter.
In chemical milling, the addition of 46 to 62 ml of urea per liter of 40 percent nitric acid
will reduce N02 emissions from 8,000 ppm to levels below 10 ppm, provided a bubble disperser is
used (Reference 6-51).
A small, but intense, source of NOV occurs in the manufacture of tungsten filaments for
"
lightbulbs. Tungsten filaments are wound on molybdenum cores, and after heat-treating, the cores
are dissolved in nitric acid.
Reference 6-43 describes air pollution equipment for reducing the dense NO, fumes given off
periodically when trays of the filaments are dissolved. The fumes pass over a charcoal adsorber
bed, which adsorbs NOX as fumes are generated and desorbs when no fumes are being generated. This
smooths out peaks and valleys in NOX content in off-gases, which are then heated and combined with
carbon monoxide and hydrogen from a rich combustion flame. The mixture is then passed through a
bed of noble metal catalyst. A colorless gas is released from the equipment.
REFERENCES FOR SECTION 6
6-1 Manney, E.H. and S. Skopp, "Potential Control of Nitrogen Oxide Emissions from Stationary
Sources," Presented at 62nd Annual Meeting of the Air Pollution Control Association,
New York. June 22-26, 1969.
6-2 Freithe, W. and M. W. Packbier, "Nitric Acid: Recent Developments in the Energy and Environ-
mental Area," presented at AICHE Symposium, Denver, Colorado, August 28, 1977.
6-3 Lowenheim, F.A. and M.K. Morgan, ed., Faith, Keyes, and Clark's Industrial Chemicals, 4th
edition. New York, Wiley Interscience Publication, 1975.
6-4 "Strong Nitric Acid, Process Features Low Utility Cost," Chemical Engineering, December 8,
1975, p. 98-99.
6-5 Personal communication. Mr. Dave Kirkbe, Davy Powergas, Houston, Texas, November 1977.
6-6 "Environmental Considerations of Selected Energy Conserving Manufacturing Process Options,
Volume XV, Fertilizer Industry Report," EPA-600/7-76-0340, December 1976.
6-7 "Compilation of Air Pollution Emission Factors (Second Edition)," Publication No. AP-42,
Environmental Protection Agency, Research Triangle Park, North Carolina, April 1973.
6-8 Gerstle, R.W. and R.F. Peterson, "Atmospheric Emissions from Nitric Acid Manufacturing
Processes," National Center for Air Pollution Control, Cincinnati, Ohio, PHS Publication
Number 999-AP-27, 1966.
6-9 Mayland, B.J., "Application of the CDL/VITOK Nitrogen Oxide Abatement Process," Presented
to Sulfur and Nitrogen Symposium, Salford, Lancashire, U.K. April 1976.
6-10 Barber, J.C. and N.L. Faucett, "Control of Nitrogen Oxide Emissions from Nitric Acid Plants,"
Third Annual Air Pollution Control Conference, March 1973.
6-11 "NOX Abatement in Nitric Acid and Nitric Phosphate Plants," Nitrogen, No. 93, Jan/Feb 1975.
6-.12 "MASAR Process for Recovery of Nitrogen Oxides," Company brochure, MASAR, Inc.
6-55
-------
6-13 Personal communication, Mr. Feaser, Plant manager, Illinois Nitrogen Plant, Marseilles, 111.
November 1977.
6-14 Service, W.J., R.T. Schneider, and D. Ethington, "The Goodpasture Process for Chemical Abate-
ment and Recovery of NO ," Conference on Gaseous Sulfur and Nitrogen Compound Emissions,
Salford, England, Aprilx1976.
6-15 Streight, H.R.L., "Reduction of Oxides of Nitrogen in Vent Gases'," Chem. Eng., Vol. 36, 1958.
6-16 Gillespie, G.R., A. A. Boyum, and M.F. Collins, "Nitric Acid: Catalytic Purification of Tail
Gas," Chemical Engineering Progress, Vol. 68, 1972.
. ,, in , |, i , "
6-17 Decker, L, "Incineration Technique for Controlling Nitrogen Oxides Emissions," Presented at
the 60th Annual Meeting of the Air Pollution Control Association, Cleveland, Ohio, June 1967.
6-18 Andersen, H.C., W.J. Green, and D.R. Steele, "Catalytic Treatment of Nitric Acid Tail Gas,"
Ind. Eng. Chem., 53:199-204, March 1961.
6-19 Anderson, G.C. and W.J. Green, "Method of Purifying Gases Containing Oxygen and Oxides of
Nitrogen (Englehard Industries, Inc., U.S. Patent No. 2, 970, 034), Official Gazette U.S.
Patent Office, 752(5} -.969, January 3\,
6-20 Newman, D.J. and L.A. Klein, "Apparatus for Exothermic Catalytic Reactions," (Chemical Con-
struction Corp., U.S. Patent NO. 3, 443, 910), Official Gazette U.S. Patent Office, 862
(2):514, May 1969.
6-21 Andersen, H.C., P.L. Romeo, and W.J. Green, "New Family of Catalysts for Nitric Acid Tail
Gases," Nitrogen 50:33-36, November-December 1967.
6-22 Rosenberg, H.S., "Molecular Sieve NO Control Process in Nitric Acid Plants," Environmental
Protection Technology Series, EPA-600/2-76-015, January 1976.
6-23 Chehaske, J.I. and J.S. Greenberg, "Molecular Sieve Tests for Control of NO,. Emissions from
a Nitric Acid Plant," Volume 1, EPA-600/2-76-048a, March 1971. x
6-24 Rosenburg, H.S., "Molecular Sieve NO Control Process in Nitric Acid Plants," EPA-600/2-76-
015, January 1976.
6-25 May! and, B.J. , The CDL/VITOK Nitrogen Oxides Abatement Process," Chenoweth Development
Laboratory, Louisville, Ky.
6-26 Personal communication, Mr. Don Ethington, Goodpasture, Inc., Brownfield, Texas, November 1977,
and February, 1978, and D. F. Carey, EPA-IERL, February 1978.
6-27 "New Unit for Nitric Plants Knocks Out NOX," Chemical Week, July 28, 1976.
6-28 "Ammonium Nitrate," Hydrocarbon Process. 46:149, November 1967.
6-29 Miles, F.D., Nitric Acid - Manufacture and Uses. London, Oxford University Press, 1961.
6-30 Private communication with Esso Research and Engineering Co.
6-31 Durocher, D.F. , P.O. Spawn, and R.C. Galkiewicz, "Screening Study to Determine Need for
Standards of Performance for New Adipic Acid Plants," draft report GCA-TR-76-16-G GCA Cor-
poration, Bedford, Massachusetts, June 1976.
6-32 Goldbeck, M., Jr., and F.C. Johnson, "Process for Separating Adipic Acid Precursors," (E.I.
DuPont de Nemours and Co., U.S. Patent No. 2, 703, 331). Official Gazette U.S. Patent
Office. 692(1):110, March 1, 1955.
6-33 Burrows, L.A., R.M. Cavanaugh, and W.M. Nagle, "Oxidation Process for Preparations of
Terephthalic Acid," (E.I. DuPont de Nemours and Co., U.S. Patent No. 2, 636, 99). Official
Gazette U.S. Patent Office. 669(4): 1209, April 28, 1953.
6-5S
-------
6-34 Durocher, D.F. et al_., "Screening Study to Determine Needs for Standards of Performance for
New Sources of Dimethyl Terephthalate and Terephthalic Acid Manufacturing" Draft Final Report,
6CA-TR-76-17-6. Submitted to EPA/OAQPS by GCA Corp., Bedford, Massachusetts, June 1976.
6-35 Lindsay, A.F., "Nitric Acid Oxidation Design in the Manufacture of Adipic Acid from Cyclohex-
anol and Cyclohexanone," Special Suppl. to Chem. Eng. Sci. 3:78-93, 1954.
6-36 Compilation of Air Pollutant Emission Factors, Environmental Protection Agency, AP-42,
February 1972.
6-37 "Process for Oxidation of Cyclohexanft and for the Production of Adipic Acid (British Patent
No. 956, 779) and Production of Adipic Acid," (British Patent No. 956, 780). Great Britain
Office. J. No. 3918:814, March 19, 1%4.
6-38 Oil, Paint, and Drug Reptr. 195(6):l-48, April 21, 1969.
6-39 Crater, W. deC. Nitration. In: Kirk-Othmer Encyclopedia of Chemical Technology. Standen,
A. (ed.). Vol. 9. New York, Interscience Publishers, 1952.
6-40 Private communication with E.I. DuPont de Nemours and Co., March 1969.
6-41 Urbanski, T., "Chemistry and Technology of Explosives," Jeczalikowa, I. and S. Laverton,
(Trs.). Vol. I. New York, MacMillan Co., 1964.
6-42 Processes Research, Inc., "Air Pollution from Nitration Processes," Cincinnati, Ohio.
APTD-1071, 1972.
6-43 Decker, L., "Incineration Technique for Controlling Nitrogen Oxides Emissions," Presented
at the 60th Annual Meeting of the Air Pollution Control Association, Cleveland. June Tl-16,
1967. .
6-44 Nelson, T.P., and Pyle, R.E., "Screening Study to Determine the Need for New Source Perfor-
mance Standards in the Explosives Manufacturing Industry," Draft Report, Radian Corporation,
Austin, Texas, July 1976.
6-45 EPA, Compilation of Air Pollutant Emission Factors, AP-42, Supplement No. 5, December 1975.
6-46 Air Pollution Engineering Source Sampling Surveys, Radford Army Ammunition Plant. U.S.
Army Environment Hygiene Agency, Edgewood Arsenal, Md.
6-47 Air Pollution Engineering Source Sampling Surveys, Volunteer Army Ammunition Plant and Joliet
Army Ammunition Plant. U.S. Army Environmental Hygiene Agency, Edgewood Arsenal, Md.
6-48 McVickar, M.H. et al.., "Fertilizer Technology'and Usage," Madison, Wisconsin, Soil Science
Society of America, 1963.
6-49 Consumption of Commercial Fertilizers, U.S. Dept. of Agriculture. Statistical Reporting
Service.
6-50 McGannon, H.E., The Making, Shaping and Treating of Steel. 8th ed. Pittsburgh, United States
Steel Co., 1964.
6-51 Kerns, B.A., "Chemical Suppression of Nitrogen Oxides," Ind. Eng. Chem. Process Design Develop.
Vol. 4, pp. 263-265, 1965.
6-57
-------
-------
APPENDIX A
SELECTED TABLES IN ENGLISH UNITS
This appendix contains the, English engineering unit version of three large cost tables
\
presented in Section 6. The tables are arranged sequentially in the order in which they appear in
the section and have the same table number as their counterparts except for the prefix "A".
A-l
-------
TABLE A6-4. CAPITAL AND OPERATING COSTS FOR DIFFERENT NOx ABATEMENT SYS-
TEMS IN A 300 TPD NITRIC ACID PLANT (Reference 6-6 and 6-26)
Capital Investaent,* ($)
Royalty
Operating Labor, (hr/yr)
(5/yr)
Maintenance Labor,
($/yr)
lawsr Overhead (1nc1. fringe
benefits & supervision, S/yr)
Catalyst or Molecular Slave
Cooling Water, (gps»)
(S/yr)
Stew, (1b/yr)
(5/yr credit)
Electricity, (kW)
($/yr)
Boiler Feed Water, (gpm)
(5/yr)
Fuel, (10* 8tu/hr)
(5/yr)
Hltrlc Acid, (tpd)
(S/yr)
Urea, tpd
(S/yr)
AnoonltM Nitrate, (tpd)
($/yr)
Depreciation (11-yr Ufa)
Return on Investment (8 20!1)
Taxes & Insurance, (9 22)
Total Annual Cost, ($/yr)
Annual Cost, ($/ton)
Catalyst
Reduction
V.384,000
360
2.200
315
2.200
4,400
77,800
_
MM
(15,8335°
(387,590
128
20,890
35
12,850
28.5
465,120
_
125.900
276,800
27.700
628,270
. 6.16
Molecular
Sieve
1,200,000
360
2,200
315
2,200
4,400
45,600
500
7,330
250
6,120
322
52.550
__
.
2.0
32.640
(6.6)
(112,200)
109.090
240,000
4f|;^§
4.06
Grands
Parolsse
1 ,000,000
Included
360
2,200
315
2.200
4,400
_
. 300
4,420
_
90
14,690
--
_
M
(6.0)
(102.000)
90,910
200,000
20.000
2ld,?80
2.32
COL/
V1tok
575,000
none
360
2,200
315
2,200
4,400
..
1,020
14,980
715
17,500
265
43,250
..
_
_
(6.0)
(102,000)
52.300
115,000
lUSOO
T81T33?
1.53
Masar
663,000
fee
360
2,200
540
3,775
5,975
._
i ^
1,310
32,070
20
3,260
.-
_
~
..
(5.28)
(89,760)
1.370
74,528
1.25
(42,500)
60,300
132.600
13.260
193,/od
1.92
Goodpasture
425,000
51 ,000
360
2,200
315
2,200
4,400
~
30
440
-
«
45
7340
-
-
-
-
-
-
-
-
(13.0)
(422,000)
38,640
85,000
3,500
(42,290)
(Q-.42)
'investment estimates exclude Interest during construction, owners expenses, and land costs.
Includes credit for 0.0017 tons of urea/ton or nitric acid produced present 1n tne spent
solution (O.SCTPD).
Parenthesis Indicate credit taken.
A-2
-------
tABLE A6-5. ANNUAL ENERGY REQUIREMENTS (109 Btu) FOR NOX ABATEMENT SYS-
TEMS FOR A 300 TPD NITRIC ACID PLANT (Reference 6-6 and 6-26)'
Steam (Credit)
Electrical
Natural Gas
011
Basic Nitric
Acid Plant
(71.4)
-
163.2
-
91.8
Catalytic
Reduction
(129.20)
10.97
232.56
-
114.33
Molecular
Sieve
2.04
27.59
-
16.32
45.95
Grande
Paroisse
-
7.71
-
-
7.71
CDL/
Vitok
5.83
22.71
-
-
28.54
Masar
10.69
1.71
-
-
12.40
Goodpasture
-
1.38
-
-
1,38
TABLE A6-6. BASIS FOR TABLES A6-4 AND A6-5 (Reference 6-6)
(Plant Capacity 9 300 tpd and 102.000 tons/yr)
(March 1975 Dollars, 0JR Index =* 2.126)
1. Operating Labor
2. Maintenance Labor
3. Overhead
4. Cooling Water
5. Boiler Feedwater
6. Natural Gas
7. Oil
8. Depreciation
9. Return on Investment
10. Taxes and Insurance
11. Nitric Acid
12. Urea
13. Antnbnium Nitrate
14. 1 kWh - 10,500 Btu
15. Electricity
16. Amrania
@ $6.1/hr
9 $7.0/hr
@ 100% of labor (including fringe benefits
and supervision)
9 $0.03 1,000 gaL
9 $0.75 1,000 gal
9 $2.00/10* Btu
9 $2.00/10* Btu
9 11 yr straight Line
9 Z0% of capital cost
9 2% of capital cost
9 $90/ton
9 $160/ton
9 $100/ton
9 $0.02/kWh
9- $1577ton
*o.s. GOVEia«m
OPF1CK: 1983-639-ooi/30o2
A-3
-------
-------
APPENDIX B
PREFIXES FOR SI UNITS
The names of multiples and submultlples of SI units may be formed by application of these
prefixes:
Factor by Which
Unit 1s Multiplied
1019
1015
1012
10»
10s
103
102
10
10-1
10-2
10-3
io-s
10-9
1Q-12
10-is
io-18
Prefix
exa
peta
tera
glga
mega
kilo
hecto
deka
deci
centl
mini
micro
nano
p1co
femto
atto
Symbol
E
P
T
G
M
k
h
da
d
c
m
U
n
P
f
a
8-1
-------
-------
APPENDIX C
GLOSSARY
Biased Firing An off-stoichiometric combustion technique in which the burners of a wall-fired
utility boiler are operated either fuel- or air-rich in a staggered configuration.
Boiler Efficiency - Heat Output x 100.
Heat Input
The overall figure reflects combustion efficiency, radiation and convection losses from the boiler,
and heat lost in exhaust gases.
Burners Out Of Service (BOOS) - An off-stoiehiometric combustion technique in which some burners
are operated on air only.
Combustion Modification An alteration of the normal burner/firebox configuration or operation
employed for the purpose of reducing the formation of nitrogen oxides.
Derating - Reducing the heat input and power or steam output of a boiler below the level for which
it was designed.
Excess Air - Any increment of air greater than the stoichiometric fuel requirement. With gas-, oil-,
and coal-fired boilers, some excess air is used to assure optimum combustion.
Field-Erected Boiler -All components of a boiler are delivered to the site and assembled in the
field. Mainly pertains to utility and largft industrial boilers.
Firetube Boiler - Steam or hot water generator with heat transfer surface consisting of steel tubes
surrounded by water and carrying hot combustion gases.
Flue Gas Recirculation (FGR) -A combustion modification in which a portion of the Roller exhaust
gases are recirculated to the burners to inhibit NO formation.
Flue Gas Treatment -A proce'ss which treats tail gases chemically to remove NO before release to
the atmosphere.
Fuel Nitrogen - Nitrogen that is chemically bound in the fuel.
Heat Input - The product of the fuel feedrate and the higher heating value, e.g., 10 Mg per hour
of coal with a higher heating value of 29 MJ/kg provides a heat input of 80.5 MM (290GJ/h).
0-1
-------
Beat Release Rate The rate of combustion per unit volune of firebox, typically In terms of MH/mJ.
Higher or Gross Heating Value (HHV) The heat generated by complete combustion of a fuel, always
referenced to baseline temperature, e.g., 16°C. Heat available at the reference temperature 1s
Included In- the higher heating value even if it 1s not practically available, 1:e.<- heat of con-*
dtnslng water vapor.
la* Excess Air A combustion modification 1n which NOX formation Is Inhibited by reducing the excess
air to less than normal ratios.
Lower or Met Heating Value (LHV) The heat that Is practically available from a fuel to generate
steas or otherwise raise the temperature of the media receiving energy. The net heating value assumes
complete combustion. It differs from the higher heating value in that heat of vaporizing* water of
ccB&ustion is considered a recoverable loss.
Off-Stoichloflietric Combustion (OSC) A combustion modification technique in which burner stoichi-
ctnetry is altered to inhibit HOX formation. Types of OSC Include biased firing, burners out of
service, and two-stage combustion.
Packaged toilers - These are usually boilers that are smaller and more economically assembled at
the plant, then shipped to the boiler site as one integral unit ready for operation after connection
to water, stream, and power.
Pplyeycljc Organic Hatter (POM) - Organic compounds which exists in condensed phase at ambient tem-
ptraturt and are emitted as either "carbon on particular" or condensed onto emitted particulate.
Polynuclear Aromatic Hydrocarbons (PKA) - Same as POM.
* *
Stoichloroetric Air - That quantity of air which supplies only enough oxygen-to react with the com-
bustible portion of the fuel.
Two-Stage Combustion - A type of off-stoichiometric combustion 1n which the burners are operated
fuel-rich and the remainder of the required combustion air is Introduced through separate ducts
located above the burner. This is also called "overfire air" or MNOX port operation.
Vatsrtube Boiler A steam generator with hea^ transfer surface consisting of steel tubes carrying
water that are exposed to hot combustion gases.
C-2
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-450/3-R3-nn?
2.
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Control Techniques for Nitrogen Oxides Emissions from
Stationary Sources - Revised Second Edition
5. REPORT DATE
January 1983
6. PERFORMING ORGANIZATION CODE
. AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-02-3513
Work Assignment 24
12. SPONSORING AGENCY NAME AND ADDRESS
United States Environmental Protection Agency
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
EPA 200/04
15. SUPPLEMENTARY NOTES
This document is issued in accordance with the requirements of Section 108 of the
Clean Air Act, as amended, 1977.
16. ABSTRACT
As required by Section 108 of the Clean Air Act, this revised second edition
compiles the best available information on NO emissions; achievable control levels
and alternative methods of prevention and control of NO emissions; alternative
x
fuels, processes, and operating methods which reduce NO emissions; cost of NO
x
control methods, installation, and operation; and the energy requirements and x
environmental impacts of the NO emission control technology.
x
Each stationary source of NO emissions is discussed along with the various
control techniques and process modifications available to reduce NO emissions.
Various combinations of equipment process conditions and fuel types are identified
and evaluated for NO emission control.
x
This revised second edition of Control Techniques for Nitrogen Oxides Emissions
from Stationary Sources updates the second edition (EPA-45,0/1-78-001) published in
January 1978. The changes are limited to revisions of information on emissions and
emission factors (Chapter 2), combustion modifications (Section 3.1), combustion
flue gas treatment (Section 3.2), utility boilers (Section 4.1), industrial boilers
(Section 4.2), space heating (Section 5.1), and industrial process heating (Section 5.
3)
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
Nitrogen Oxides Emissions
Control Techniques
Fossil Fuel Combustion
Nitric Acid Manufacturing
Costs
Photochemical Oxidants
13b
18. DISTRIBUTION STATEMENT
Release Unlimited
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
428
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 22201 (Rev. 4-77) PREVIOUS EDITION is OBSOLETE
-------
-------