-------
Figure 5-18 shows a lean premixed combustor design used by
another manufacturer for an annular combustor. The air and fuel
are premixed using a very lean A/F, and the resultant uniform
mixture is delivered to the primary combustion zone where
combustion is stabilized using a pilot "flame. Using one or more
mechanical systems to regulate the airflow delivered to the
combustor, the premix mode is operable for output loads between
50 and 100 percent. Below 50 percent load, only the pilot flame '
is operating, and NOX emissions levels are similar to those for
conventional combustors.4°
Another manufacturer's production low-NOx design uses a silo
combustor. Unlike the can-annular and annular designs, the silo
combustor'is mounted externally to the turbine and can therefore
be modified without significantly affecting the rest of the
turbine design, provided the mounting flange to the turbine is
unchanged. In addition, this large combustion chamber is fitted
with a ceramic lining that shields the metal surfaces from peak
flame temperatures. This lining reduces the requirement for .
cooling air, so more air is available for the combustion
process.17
This silo low-No., combustor design uses six burners, as
Jv
shown in Figure 5-19. For operation on natural gas, each burner
serves to premix the air and fuel to deliver a lean and uniform
mixture to the combustion zone. To achieve the lowest possible
NOX emissions, the A/F of the premixed gases is kept very near
the lean flammability limit and a pilot flame is used to
stabilize the overall combustion process. This burner design is
shown in Figure 5-20. Like the can-annular design, the burner in
the silo combustor cannot operate over the full power range of
the gas turbine in the premix mode due to inability of the premix
mode to deliver suitable A/F's at low power output levels. For
this reason, the burners are designed to operate in a
conventional diffusion burning mode at startup and low power
outputs and switch to a premix burning mode at higher power
output levels.
5-43
-------
tn
-H
CO
0)
-o
a
o
•H
4J
a)
o
u
(TS
T)
-H
e
(U
^
a
c
OS
OJ
M-l
o
o
-H
4-)
O
0)
in
i
GQ
CO
O
M
U
CO
H
i
IT)
0)
Cn
-H
5-44
-------
SILO COMBUSTOR
Figure 5-19. Cross-section of a low-NOx silo
5-45
-------
oo
O
4J
CO
d
o
u
•H
CO
0)
o
o
CN
in
Cn
-H
5-46
-------
For operation on distillate oil with the current burner
design, combustion occurs only in a diffusion mode and there is
no premixing of air and fuel.
5.2.2.2 Applicability. As discussed in Section 5.2.2.1,
lean premixed combustors apply to can-annular, annular, and silo
combustors. This combustion modification is effective in
reducing thermal NOX emissions for both natural gas and
distillate oil but is not effective on fuel NOX. Therefore, lean
premixed combustion is not as effective in reducing NOX levels if
high-nitrogen fuels are fired.4^
The multiple operating modes associated with the percent
operating load results in "stepped" NO^. emission levels. To
J\.
date, low NOY emission levels occur only at loads greater than 40
Jv
to 75 percent.
Lean premixed combustors currently are available for limited
models from three manufacturers contacted for this study.6'17'24
Two additional manufacturers project an availability date of 1993
or 1994 for lean premixed combustors for some turbine
models.11'50 All of these manufacturers state that these lean
premixed combustors will be available for retrofit applications.
5.2.2.3 Factors Affecting Performance. The primary factors
affecting the performance of lean, premixed combustors are A/F
and the type of fuel. To achieve low NO., emission levels, the
Jl*
A/F must be maintained in a narrow range near the lean
flammability limit of the mixture. Lean premixed combustors are
designed to maintain this A/F at rated load. At reduced load
conditions, the fuel input requirement decreases. To avoid
combustion instability and excessive CO emissions that would
occur as the A/F reaches the lean flammability limit, all
manufacturers' lean premixed combustors switch to a
diffusion-type combustion mode at reduced load conditions,
typically between 40 and 60 percent load. This switchover to a
diffusion combustion mode results in higher NCL. emissions.
Jt
Natural gas produces lower NOX levels than do oil fuels..
The reasons for this are the lower flame temperature of natural
gas and the ability to premix this fuel with air prior to
5-47
-------
delivery into the second combustion stage. For operation on
liquid fuels, currently available lean premixed combustor designs
require water injection to achieve appreciable NOV reduction.
Ji.
5.2.2.4 Achievable NOX Emission Levels. The achievable
controlled NOX emission levels for lean premixed combustors vary
depending upon the manufacturer. At least three manufacturers
currently guarantee NOX emission levels of 25 ppmv, corrected to
15 percent 02 for most or all of their gas turbines for operation
on natural gas fuel without wet injection.^'17/^4 Each of these
three manufacturers has achieved controlled NO., emission levels
A
of less than 10 ppmv at one or more installations in the
United States and/or Europe and guarantee this NO,, level for a
Jv
limited number of their gas turbine models.^1 All three
manufacturers offer gas turbines in the 10+ MW (13,400 hp+) .range
and anticipate that guaranteed NOX emission levels of 10 ppmv or
less will be available for all of their gas turbines for
operation on natural gas fuel in the next few years. These
low-.NOx combustor designs apply to new turbines and existing
installation retrofits.
For gas turbines in the range of 10 MW (13,400 hp) and
under, one gas turbine manufacturer offers a guarantee for its
lean premixed combustor, without wet injection, of 42 ppmv using
natural gas fuel for two of its turbine models for 1994 delivery.
This manufacturer states that a controlled NOX emission level of
25 ppmv has been achieved by in-house testing, and this 25 ppmv
level firing natural gas fuel is the goal for all of its gas
turbine models, for both new equipment and retrofit
applications.50
These controlled NOX emission levels of 9 to 42 ppmv
correspond to full output load; at reduced loads, the NO... levels
Jt
increase, often in "stepped" fashion in accordance with changes
in combustor operation from premixed mode to conventional or
diffusion-mode operation (see Section 5.2.2.3). Figure 5-21.
shows these stepped NOX emissions levels for a can-annular
combustor for natural gas and oil fuel operation. Figure 5-22
5-48
-------
NATURAL GAS FUEL
-I 100
I I I I I I I I I I
20
30 40 SO 60 70 80
% GAS TURBINE LOAD
90
100
200 r
OIL FUEL
10 20 30 40 50 60 70
% GAS TURBINE LOAD
SO
90
100
Figure 5-21. "Stepped" NOX and CO emissions for a low-NO
can-annular combustor burning natural gas and distillate oil
fuels.47
5-49
-------
ui
O
O
I
O Diffusion Burner Operation
0 Premix Burner Operation
with 9* Pilot Flame
-CO Emission
NOx Emission
Maximum
Dilution <>
Air
Min.
Max.
Compressor Mass Flow
(Adjustable Inlet Guide Vanes)
' In Dry Exhaust Gas with 15H 0» by Volume
Figure 5-22. "Stepped" NO and CO emissions for a low-NO,
silo combustor burning natural gas.35
5-50
-------
shows the emissions for a silo combustor operating on natural gas
only. The emission levels shown in Figures 5-21 and 5-22
correspond to full-scale production turbines currently available
from the manufacturers.
Reduced NOX emissions when burning oil fuel in currently
available lean premixed combustor designs have been achieved only
with water or steam injection. With water or steam injection, a
65 ppmv NOX level can be achieved in the turbine with a can-
annular combustor design; a 65 ppmv level can also be met with
water injection in the turbine with a silo combustor at a WFR of
1>4_48,52 This 65 ppmv level for lean premixed combustors is
higher than the controlled NOX levels achieved with water
injection in oil-fired turbines using a conventional combustor
design.
Modification of- the existing burner design used in the silo
combustor to allow premixing of the oil fuel with air prior to
combustion is under development. Tests performed using a 12 MW
(16,200 hp) turbine achieved NOX emission levels below 50 ppmv.
without wet injection, corrected to 15 percent C>2, compared to
uncontrolled levels of 150 ppmv or higher. The NOX levels,
without wet injection, as a function of equivalence ratio are
shown in Figure 5-23. The design equivalence ratio at rated load
is approximately 2.1. As shown in this figure, NOX emissions
below 50 ppmv were achieved at rated power output at pilot fuel
flow levels of 10 percent of the total fuel input.52
Site test data for two turbines using silo-type lean
premixed combustors, as reported by the manufacturer, are shown
in Table 5-12. As this table shows, NOV emission levels as low
Jv
as 16.5 ppmv were recorded for using natural gas fuel without
water injection. Subsequent emission tests have achieved levels
below 10 ppmv.51 Corresponding data for operation on oil fuel
using only the pilot (diffusion) stage for combustion, and with
water injection, is shown in Table 5-13. Levels of NOY emissions
J^
at base load for No. 2 fuel oil are between 50 and 60 ppmv.
Based on information provided by turbine manufacturers, the
potential NOX reductions using currently available lean premixed
5-51
-------
ppm
300-
250-
200-
C
o
'35
X
O
100-1
50-
Pilot Fuel Oil Flow:
O = 100%
A = 20%
n = 10%
0 = 0%
1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8*
Equivalence Ratio
' In Dry Exhaust Gat with 15% 0, by Volume
Figure 5-23. Nitrogen oxide emission test results from a lean
premix silo combustor firing fuel oil without wet injection.53
5-52
-------
TABLE 5-12. MEASURED NOX EMISSIONS FOR COMPLIANCE TESTS
OF A NATURAL GAS-FUELED LEAN PREMIXED COMBUSTOR
WITHOUT WATER INJECTION22
Turbine No.
1
1
2
2
. 1
2
Output, percent of
baseline
107
100
100
75
50
50
NOX emission level,
ppmva
17.7
16.5
24.1
20.4
22.3
22.2
lln dry exhaust with 15 percent 02, by volume.
5-53
-------
TABLE 5-13. MEASURED NO EMISSIONS FOR OPERATION OF A LEAN
PREMIXED COMBUSTOR DESIGN OPERATING IN DIFFUSION MODE
ON OIL FUEL WITH WATER INJECTION22
Turbine No.
1
2
1
2
1
2
2
Output, percent of
baseload
Peak
Peak
100
100
75
75
50
NOX emission level,
ppmva
69.3
53.6
59.9
51.6
54.3
49.2
54.8
LIn dry exhaust with 15 percent 02, by volume.
5-54
-------
combustors are shown in Table 5-14. As this'table indicates, NOX
emission reductions range from 14.7 tons/yr for a 1.1 MW
(1,480 hp)~ turbine to 10,400 tons/yr for a 204 MW (274,000 hp)
turbine for operation on natural gas without wet injection.
Corresponding NOX emission reductions for operation on oil fuel,
with water injection, range from 620 tons/yr for a 22.6 MW
(30,300 hp) turbine to 7,360 tons/yr for an 83.3 MW (112,000 hp)
turbine.
Limited data from two manufacturers showing the impact of
lean premixed combustor designs on CO emissions are shown in
Table 5-15. For natural gas-fueled turbines with rated outputs
of 10 MW (13,400 hp) or less, controlled NO,, emission levels of
x *
25 to 42 ppmv result in a rise in CO emission levels from 25 ppmv
or less to as high as 50 ppmv.4^ For turbines above 10 MW
(13,400 hp), controlled NOX emission levels of 9 ppmv result in a
rise in CO emissions from 10 to 25 ppmv for natural gas fuel.
Conversely, for controlled NOX emission levels of 25 ppmv, the
CO emissions drop from 25 to 15 ppmv. 1 For one manufacturer's
lean premixed silo combustor design, CO emissions at rated load
are less than 5 ppmv, as shown previously in Figure 5-21. This
limited data suggest that the effect of lean premixed combustors
on CO emissions depends upon the specific combustor design and
the controlled NOX emission level.
The emission levels shown in Table 5-15 correspond to rated
power output. Like NOX emission levels, CO emissions change with
changes in combustor operating mode at reduced power output. The
"stepped" effect on CO emissions is shown in Figures 5-21 and
5-22, shown previously.
Operation on oil fuel with wet injection, shown previously
in Figure 5-21, shows CO emission levels of 20 ppmv. Additional
CO emission data were not available for operation on oil fuel
with water injection in lean premixed combustors. Developmental
tests for operation on oil fuel without wet injection in a silo
combustor are presented in Figure 5-24. At rated load, shown in
this figure at an equivalence ratio of approximately 2.1,
CO emissions are less than 10 ppmv, corrected to 15 percent 0,
^ /
5-55
-------
TABLE 5-14.
POTENTIAL NO REDUCTIONS FOR GAS TURBINES USING
LEAN PREMIXED COMBUSTORS
Turbine model
Saturn0
Centaur T-
4500C
Centaur "H"c
Taurusc
Mars T-12000C
Mars T- 14000°
MS6001B
MS7001E
MS7001F
MS9001E
MS9001F
GT10
GT11N
V84.2
V94.2
V64.3
V84.3C
V94.36
Power
output,
MW
1.1
3.3
4.0
4.5
8.8
10.0
39.0
84.7
161
125
229
22.6
83.3
105
153
61.5
141
204
NOX emissions
Uncontrolled
Gas fuel,
ppmv
99
130
105
114
178
199
148
154
210
161
210
150
390
212
212
380
380
380
Oil fuel,
ppvm
150
179
160
168
267
NAd
267
228
353
241
353
200
560
360
360
530
530
530
Controlled
Gas fuel,
ppmv
42
42
42
42
42
42
25/9e
25/9e
25
25/9e
25
25
25/9e
25/9e
96
42
42
42
Oil fuel,
ppmv
NAd
NAd
NAd
NAd
NAd
NAd
65
65
65
65
65
42
42
NAf
NAf
NAd
NAd
NAd
NOY reduction
Gas fuel,
toos/yr*
14.7
59.5
49.8
62.4
212
270
* 829/937
1,820/2,050
4,540
2,740/3,060
6,500
476
5,070/5,290
3,030/3,290
4,410/4,780
3,210
7,230
10,400
Oil fuel,
tons/yr* b
NAd
NAd
NAd
NAd
NAd
NAd
1,139
2,360
5,190
3,490
7,250
620
7,360 .
NAf
NAf
NAd
NAd
NAd
fBased on 8,000 hours operation per year.
"Requires water or steam injection.
°Scheduled availability is 1994 for natural gas fuel.
NA = Data not available.
eStandard NOX guarantee is 25 ppmv. Manufacturers offer guaranteed NOX levels as low as 9 ppmv for these
turbines.
^Scheduled availability 1993 for oil fuel without water injection. Reference 17.
5-56
-------
TABLE 5-15. COMPARISON OF NO AND CO EMISSIONS FOR STANDARD
VERSUS LEAN PREMISED COMBUSTORS FOR
TWn MaTJTTParTTTR'RPG' TTTPRTTVraS^O r 54
TWO MANUFACTURERS' TURBINES
GT Model
Centaur H
Mars T-14000
MS6001B
MS7001E
MS9001E
MS7001F
MS9001F
Emissions, ppmv, referenced to 15 percent C^*
Power
output,
MW
4.0
10.0
39.0
84.7
125
161
229
Standard combustor
NOX
105
199
148
154
161
210
210
CO
15
5.5
10
10
10
25
25
Lean premixed combustor
NOX
25-42
25-42
9
9
9
25
25
CO
50b
50b
25
25
25
15
15
aFor operation at ISO conditions using natural gas fuel.
"Maximum design goal for CO emissions. Most in-house test configurations have achieved CO emission levels between 5
and 25 ppmv.
5-57
-------
350-,
300-
250-
200-
*0 15
-------
and are in the range of 0 to 2 ppmv for a pilot oil fuel flow of
10 percent (representing 10 percent of the total fuel flow).53
This 10 percent pilot fuel flow corresponds to controlled NOX
emission levels below 50 ppmv, as shown previously in
Figure 5-22. No data for HC emissions were available for lean
premixed burner designs.
5.2.3 Rich/Ouench/Lean Combustion
5.2.3.1 Process Description. Rich/quench/lean (RQL)
combustors burn fuel-rich in the primary zone and fuel-lean in
the secondary zone. Incomplete combustion under fuel-rich
conditions in the primary zone produces an atmosphere with a high
concentration of CO and hydrogen (H2). The CO and H2 replace
some of the oxygen normally available for NOX formation and also
act as reducing agents for any NOX formed in the primary zone.
Thus, fuel nitrogen is released with minimal conversion to NOX.
The lower peak flame temperatures due to partial combustion also
reduce the formation of thermal NOV.55
-A.
. As the combustion products leave the primary zone, they pass
through a low-residence-time quench zone where the combustion
products are rapidly diluted with additional combustion air or
water. This rapid dilution cools the combustion products and at
the same time produces a lean A/F. Combustion is then completed
under fuel-lean conditions. This secondary lean combustion step
minimally contributes to the formation of fuel NOX because most
of the fuel nitrogen will have been converted to N2 prior to the
lean combustion phase. Thermal NOX is minimized during lean
combustion due to the low flame temperature.55
5.2.3.2 Applicability. The RQL combustion concept applies
to all types of gas turbines. None of the manufacturers
contacted for this study, however, currently have this design
available for their production turbines. This may be due to lack
of demand for this design due to the current limited use of
high-nitrogen-content fuels in gas turbines.
5.2.3.3, Factors Affecting Performance. The NO,, emissions
Jt
from RQL combustors are affected primarily by the equivalence
ratio in the primary combustion zone and the quench airflow rate.
5-59
-------
Careful selection of equivalence ratios in the fuel-rich zone
will minimize both thermal and fuel NO., formation. Further NOV
Jt JC
reduction is achieved with increasing quench airflow rates, which
serve to reduce the equivalence ratio in the secondary (lean)
combustion stage.
5.2.3.4 Achievable NO.. Emissions Levels Using
•*» .
Rich/Ouench/Lean Combustion. The RQL staged combustion has been
demonstrated in rig tests to be effective in reducing both
thermal NOX and fuel NOX. As shown in Figure 5-25, NOX emissions
are reduced by 40 to 50 percent in a test rig burning diesel
fuel. At an equivalence ratio of 1.8, the NOX emissions can be
reduced from 0.50 to 0.27 Ib/MMBtu by increasing the quench
airflow from 0.86 to 1.4 kg/sec. Data were not available to
convert the NOX emissions figures to ppmv. The effectiveness of
rich/lean staged combustion in reducing fuel NOX when firing
high-FBN fuels is shown in Figure 5-26. Increasing the FBN
content from 0.13 to 0.88 percent has little impact on the total
NOV -formation at an operating equivalence ratio of 1.3 to 1.4.
Jt —
Tests on other rich/lean combustors indicate fuel nitrogen
conversions to NOX of about 7 to 20 percent.58'59 These fuel
nitrogen conversions represent a fuel NOX emission reduction of
approximately 50 to 80 percent.
One manufacturer has tested an RQL combustor design in a
4 MW (5,360 hp) gas turbine fueled with a finely ground coal and
water mixture. The coal partially combusts in a fuel-rich zone
at temperatures of 1650°C (3000°F), with low O2 levels and an
extremely short residence time. The partially combusted products
are then rapidly quenched with water, cooling combustion
temperatures to inhibit thermal NOX formation. Additional
combustion air is then introduced, and combustion is completed
under fuel-lean conditions. In tests at the manufacturer's
plant, cosponsored by the U. S. Department of Energy, a NOX
emission level of 25 ppmv at 15 percent Q^ was achieved. This
combustor design can also be used with natural gas and oil fuels.
Single-digit NOX emission levels are reported for operation on
5-60
-------
0.70
0.60
03
0
O
IS)
C
O
0.50
O
•M
»
J3
O
0.40
0.30
0.20U
0.86
t
n
v
Uncontrolled
Controlled
/
' Quench Air
1.4
JL
1.5 1.6 1.7 1.8 1.9 2.0 • 2.1
Primary Zone Equivalence Ratio
Figure 5-25. Nitrogen oxide emissions versus primary zone
equivalence ratio for a rich/quench/lean combustor firing
distillate oil.56
5-61
-------
0.3
I/I
I/I
0.2
1.0
Fuel FBN, wt.%
Distillate Oil .,&-- 0.13
Residual Oil —O— 0-27
Middle distillate-Q- 0.88
coal-derived fuel
1.1 1.2 1.3 1.4
Rich Zone Equivalence Ratio
1.5
Figure 5-26. Effects of fuel bound nitrogen (FBN) content of NOX
emissions for a rich/quench/lean combustor. '
5-62
-------
natural gas fuel. This combustor design is not yet available for
production turbines.60
5.3 SELECTIVE CATALYTIC REDUCTION
Selective catalytic reduction (SCR) is an add-on NOX control
technique that is placed in the exhaust stream following the gas
turbine. Over 100 gas turbine installations use SCR in the
United States.61 An SCR process description, the applicability
of SCR for gas turbines, the factors affecting SCR performance,
and the achievable NOX reduction efficiencies are discussed in
this section.
5.3.1 Process Description
The SCR process reduces NOX emissions by injecting ammonia
into the flue gas. The ammonia reacts with NOX in the presence
of a catalyst to form water and nitrogen. In the catalyst unit,
the ammonia reacts with NOX primarily by the following
equations:62
NH3 + NO + 1/4 02 -» N2 + 3/2 H20; and
. NH3 + 1/2 N02 + 1/4 O2 •* 3/2 N2 + 3/2 H20.
The catalyst's active surface is usually either a noble
metal, base metal (titanium or vanadium) oxide, or a
zeolite-based material. Metal-based catalysts are usually
applied as a coating over a metal or ceramic substrate. Zeolite
catalysts are typically a homogenous material that forms both the
active surface and the substrate. The geometric configuration of
the catalyst body is designed for maximum surface area and
minimum obstruction of the flue gas flow path to maximize
conversion efficiency and minimize back-pressure on the gas
turbine. The most common catalyst body configuration is a
monolith, "honeycomb" design, as shown in Figure 5-27.
An ammonia injection grid is located upstream of the
catalyst body and is designed to disperse the ammonia uniformly
throughout the exhaust flow before it enters the catalyst unit.
In a typical ammonia injection system, anhydrous ammonia is drawn
from a storage tank and evaporated using a steam- or
electric-heated vaporizer. The vapor is mixed with a pressurized
carrier gas to provide both sufficient momentum through the
5-63
-------
Figure 5-27. Cutaway view of a typical monolith catalyst
. body with honeycomb configuration.62
5-64
-------
injection nozzles and effective mixing of the ammonia with the
flue gases. The carrier gas is usually compressed air or steam,
and the ammonia concentration in the carrier gas is about
5 percent.62
An alternative to using the anhydrous ammonia/carrier gas
system is to inject an a aqueous ammonia solution. This system
is currently not as common but removes the potential safety
hazards associated with transporting and storing anhydrous
ammonia and is often used in installations with close proximity
to populated areas.61'62
The NH3/NOX ratio can be varied to achieve the desired level
of NO,, reduction. As indicated by the chemical reaction
x *
equations listed above, it takes one mole of NH3 to reduce one
mole of NO, and two moles of NH3 to reduce one mole of N02- The
NO,, composition in the flue gas from a gas turbine is over
./t
85 percent NO, and SCR systems generally operate with a molar
NH3/NOX ratio of approximately l.O.6^ Increasing this ratio will
further reduce NOX emissions but will also result in increased
unreacted ammonia passing through the catalyst and into the
atmosphere. This unreacted ammonia is known as ammonia slip.
5.3.2 Applicability of SCR for Gas Turbines
Selective catalytic reduction applies to all gas turbine
types and is equally effective in reducing both thermal and fuel
NOX emissions. There are, however, factors that may limit the
applicability of SCR.
An important factor that affects the performance of SCR is
operating temperature. Gas turbines that operate in simple cycle
have exhaust gas temperatures ranging from approximately 450° to
540°C (850° to 1000°F). Base-metal catalysts have an operating
temperature window for clean fuel applications of approximately
260° to 400°C (400° to 800°F). For sulfur-bearing fuels that
produce greater than 1 ppm S03 in the flue gas, the catalyst
operating temperature range narrows to 315° to 400°C (600° to
800°F). The upper range of this temperature window can be
5-65
-------
increased using a zeolite catalyst to a maximum of 590°C
(1100°F),64
Base metal catalysts are most commonly used in gas turbine
SCR applications, accounting for approximately 80 percent of all
U.S. installations, and operate in cogeneration or combined cycle
applications. The catalyst is installed within the HRSG, where
the heat recovery process reduces exhaust gas temperatures to the
proper operating range for the catalyst. The specific location
of the SCR within the HRSG is application-specific; Figure 5-28
shows two possible SCR locations. In addition to the locations
shown, the catalyst may also be located within the evaporator
section of the HRSG.
+
As noted above, zeolite catalysts have a maximum operating
temperature range of up to 590°C (1100°F) , which is compatible
with simple cycle turbine exhaust temperatures. To date,
however, there is only one SCR installation operating with a
zeolite catalyst directly downstream of the turbine. This
catalyst, commissioned in December 1989, has an operating range
of 260° to 515°C (500° to 960°F) and operates approximately
90 percent of the time at temperatures above 500°C (930°F).65
Another consideration in determining the applicability of
SCR is complications arising from sulfur-bearing fuels. The
sulfur content in pipeline quality natural gas is negligible, but
distillate and residual oils as well as some low-Btu fuel gases
such as coal gas have sulfur contents that present problems when
used with SCR systems. Combustion of sulfur-bearing fuels
produces S02 and S03 emissions. A portion of the S02 oxidizes to
SO3 as it passes through the HRSG, and base metal catalysts have
an S02-to-S03 oxidation rate of up to five percent.64 In
addition, oxidation catalysts, when used to reduce CO emissions,
will also oxidize S02 to S03 at rates of up to 50 percent. °
Unreacted ammonia passing through the catalyst reacts with
S03 to form ammonium bisulfate (NH4HS04) and ammonium sulfate
[(NH4)2 S04] in the low-temperature section of the HRSG. The
rate of ammonium salt formation increases with increasing levels
of SO3 and NH3, and the formation rate increases with decreasing
5-66
-------
Steam
Superheater SCR Evaporator
Catalytt
Economizer
Steam
Superheater Evaporator
SCR Economizer
Cataiyet
Figure 5-28. Possible locations for SCR unit in HRSG.62
5-67
-------
temperature. Below 200°C (400°F), ammonium salt formation occurs
with single-digit ppmv levels of SO3 and NH3.66
The exhaust temperature exiting the HRSG is typically in the
range of 150° to 175°C (300° to 350°F), so ammonium salt
formation typically occurs in the low-temperature section of the
HRSG.66 Ammonium bisulfate is a sticky substance that over time
corrodes the HRSG boiler tubes. Additionally, it deposits on
both the boiler and catalyst bed surfaces, leading to fouling and
plugging ..of these surfaces. These deposits result in increased
back pressure on the turbine and reduced heat transfer efficiency
in the HRSG. This requires that the HRSG be removed from service
periodically to water-wash the affected surfaces. Ammonium
sulfate is not corrosive, but like ammonium bisulfate, it
deposits on the HRSG surfaces and contributes to plugging and
fouling of the heat transfer system.33
Formation of ammonium salts can be avoided by limiting the
sulfur content of the fuel and/or limiting the ammonia slip. Low
SC>2-to-S03 oxidizing catalysts are also available. Base metal
catalysts are available with oxidation rates of less than
1 percent, but these low oxidation formulas also have lower NOX
reduction activity per unit volume and therefore require a
greater catalyst volume to achieve a given NOX reduction level.
Zeolite catalysts are reported to have intrinsic S02-to-S03
oxidation rates of less than 1 percent.64'66 As stated above,
pipeline-quality natural gas has negligible sulfur content, but
some sources of natural gas contain ^S, which may contribute to
ammonium salt formation. ' For oil fuels, even the lowest-sulfur
distillate oil or liquid aviation fuel contains sulfur levels
that can produce ammonium salts. According to catalyst vendors,
SCR systems can be designed for 90 percent NOY reduction and
J\*
10 ppm or lower NH3 slip for sulfur-bearing fuels up to 0.3
percent by weight.64 Continuous emission monitoring equipment
has been developed for NH3, and may be instrumental in regulating
ammonia injection to minimize slip.67
To date, there is limited operating experience using SCR
with oil-fired gas turbine installations. One combined cycle
5-68
-------
installation using oil fuel, a United Airlines facility in
San Francisco installed in 1985, experienced fuel-related
catalyst problems and now uses only natural gas fuel. 3 In the
past, sulfur was found to poison the catalyst material.
Sulfur-resistant catalyst materials are now available, however,
and catalyst formulation improvements have proven effective in
resisting performance degradation with oil fuels in Europe and
Japan, where catalyst life in excess of 4 to 6 years has been
achieved, versus 8 to 10 years with natural gas fuel.64 A
zeolite' catalyst installed on a 5 MW (6710 hp) dual fuel
reciprocating engine in the northeastern United States has
operated for over 3 years and burned approximately
600,000 gallons of diesel fuel while maintaining a NOX reduction
efficiency of greater than 90 percent.3
In its guidance to member states, NESCAUM recommends that-
SCR be considered for NOX reduction in dual-fueled turbine
applications. There are four combined cycle gas turbines
installations operating with SCR in the northeast United States
burning natural gas as the primary fuel with oil fuel as a
back-up.3 These installations, listed in Table 5-16, began
operating recently and have limited hours of operation on oil
fuel. As indicated in the table, two of these installations shut
down the ammonia injection when operating on oil fuel to prevent
potential operating problems arising from sulfur-bearing fuels.
Permits issued more recently in this region for other dual-fuel
installations, however, require that the SCR system be
operational on either fuel.3
A final consideration for SCR is catalyst masking or
poisoning agents. Natural gas is considered clean and free of
contaminants, but other fuels may contain agents that can degrade
catalyst performance. For refinery, field, or digester gas fuel
applications, it is important to have an analysis of the fuel and
properly design the catalyst for any identified contaminants.
Arsenic, iron, and silica may be present in field gases, along
with zinc and phosphorus. Catalyst life with these fuels depends
upon the content of the gas and is a function of the initial
5-69
-------
TABLE 5-16. GAS TURBINE INSTALLATIONS IN THE NORTHEASTERN
UNITED STATES WITH SCR AND PERMITTED FOR
BOTH NATURAL GAS AND OIL FUELS3
Installation
Altresco-Pittsfield
Cogen
Technologies
Ocean State
Power
Pawtucket Power
State
MA
NJ
RI
RI
Gas turbine
model
MS6001
MS6001
MS7001E
MS6001
Output,
MWa
38.3
38.3
83.5
38.3
NO emissions, ppmv (gas fuel/oil fuel)
Uncontrolled11
148/267
148/267
154/277
148/267
Wet
injection"
42/65
42/65
42/65
42/65
Wet
injection
+ SCRC
9/18d e
!5/65f
9/42f
9/18d
aPower output for a single gas turbine. Installation power output is higher due to multiple units and/or
combined cycle operation.
Per manufacturer at ISO conditions.
°Operating permit limits.
This installation requires the SCR system to be operational when burning oil fuel.
eThis installation operated 185 hours on oil fuel in 1991, burning approximately 354,000 gallons of oil fuel.
'Ammonia injection is shut down during operation on oil fuel.
5-70
-------
design parameters. With oil fuels, in addition to the potential
for ammonium salt formation, it is important to be aware of heavy
metal content. Particulates in the flue gas can also mask the
catalyst.64
Selective catalytic reduction may not be readily applicable
to gas turbines firing fuels that produce high ash loadings or
high levels of contaminants because these elements can lead to
fouling and poisoning of the catalyst bed. However, because gas
turbines are also subject to damage from these elements, fuels
with high levels of ash or contaminants typically are not used.
Coal, while not currently a common fuel for turbines, has a
number of potential catalyst deactivators. High dust
concentrations, alkali, earth metals, alkaline heavy metals,
calcium sulfate, and chlorides all can produce a masking or
blinding effect on the catalyst. High dust can also erode the
catalyst. Erosion commonly occurs only on the leading face of
the catalyst. Airflow deflectors and dummy layers of catalyst
can be used to straighten out the airflow and reduce erosion. .
There is currently no commercial U.S. experience with coal. In
Japan, which burns low-sulfur coal with moderate dust levels,
catalyst life has been 5 years or more without replacement. In
Germany, with high dust loadings,-the experience has also been
5 years or more.64
Masking agents deposit on the surface of the catalyst,
forming a barrier between the active catalyst surface and the
exhaust gas, inhibiting catalytic activity. Poisoning agents
chemically, react with the catalyst and render the affected area
inactive. Masking agents can be removed by vacuuming or by using
soot blowers or superheated steam. Catalysts cleaned in this
manner can recover greater than 90 percent of the original
reduction activity. The effects of poisoning agents, however,
are permanent and the affected catalyst surface cannot be
regenerated.64
Retrofit applications for SCR may require the addition of a
heat exchanger for simple cycle installations, and replacement or
extensive modification of the existing HRSG in cogeneration and
5-71
-------
combined cycle applications to accommodate the catalyst body.
For these reasons, retrofit applications for SCR could involve
high capital costs.
5.3.3 Factors Affecting SCR Performance
The NOX reduction efficiency for an SCR system, is influenced
by catalyst material and condition, reactor temperature, space
velocity, and the NH3/NOX ratio.63 These design and operating
variables are discussed below.
Several catalyst materials are available, and each has an
->.
optimum NOX removal efficiency range corresponding to a specific
temperature range. Proprietary formulations containing titanium
dioxide, vanadium pentoxide, platinum, or zeolite are available
to meet a wide spectrum of operating temperatures. The NOX
removal efficiencies for these catalysts are typically between 80
and 90 percent when new. The NOX removal efficiency gradually
decreases over the operating life of the catalyst due to
deterioration from masking, poisoning, or sintering.63 The rate
of catalyst performance degradation depends upon operating
conditions and is therefore site-specific.
The space velocity (volumetric flue gas flow divided by the
catalyst volume) is an indicator of gas residence time in the
catalyst unit. The lower-the space velocity, the higher the
residence time, and the higher the potential for increased NOX
reduction. Because the gas flow is a constant determined by the
gas turbine, the space velocity depends upon the catalyst volume,
or total active surface area. The distance across .the opening
between plates or cells in the catalyst, referred to as the
pitch, affects the overall size of the catalyst body. The
smaller the pitch, the greater the number of rows or cells that
can be placed in a given volume. Therefore, for a given catalyst
body size, the smaller the pitch, the larger the catalyst volume
and the lower the space velocity. For natural gas applications
the catalyst pitch is typically 2.5 millimeters (mm) (0.10 inch
[in.]), increasing to 5 to 7 mm (0.20 to 0.28 in.) for coal-fuel
applications.64
5-72
-------
As discussed in Section 5.3.1, the NH^/NO^. ratio can be
varied to achieve the desired level of NOX reduction. Increasing
this ratio increases the level of NOX reduction but may also
result in higher ammonia slip levels.
5.3.4 Achievable NOX Emission Reduction Efficiency Using SCR
Most SCR systems operating in the United States have a space
velocity of about 30,000/hr, a NH3/NOX ratio of about 1.0, and
ammonia slip levels of approximately 10 ppm. The resulting NOX
reduction efficiency is about 90 percent.41 Reduction efficiency
is the level of NOX removed as a percentage of the level of NOX
entering the SCR unit. Only one gas turbine installation in the
United States was identified using only SCR to reduce NOX
emissions. This installation has two natural gas-fired 8.5 MW
gas turbines, each with its own HRSG in which is installed an SCR
system. A summary of emission testing at this site lists NOX
'emissions at the inlet to the SCR catalyst at 130 ppmv.
Controlled NOX emissions downstream of the catalyst were 18 ppmv,
indicating a NOX reduction efficiency of 86 percent. Maximum
ammonia slip levels were listed at 35 ppmv.68 .
All other gas turbine installations identified as using SCR
in the United States use this control method in combination with
wet injection and/or low-NOx combustors. The emission levels
t that can be achieved by this combination of controls are found in
Section 5.4.
5.3.5 Disposal Considerations for SCR
The SCR catalyst material has a finite life, and disposal
can pose a problem. The guaranteed catalyst life offered by
catalyst suppliers ranges from 2 to 3 years.64 In Japan, where
SCR systems have been in operation since 1980, experience shows
that many catalysts in operation with natural gas-fired boilers
have performed well for 7 years or longer.63'64 In any case, at
some point the catalyst must be replaced, and those units
containing heavy metal oxides such as vanadium or titanium
potentially could be considered hazardous wastes. While the
amount of hazardous material in the catalyst is relatively small,
the volume of the catalyst body can be quite large, and disposal
5-73
-------
of this waste could be costly. Some suppliers provide for the
removal and disposal of spent catalyst. Precious metal and
zeolite catalysts do not contain hazardous wastes.
5.4 CONTROLS USED IN COMBINATION WITH SCR
With but one exception, SCR units installed in the United
States are used in combination with wet controls or combustion
controls described in Sections 5.1 and 5.2. Wet controls yield
NOX emission levels of 25 to 42 ppmv for natural gas and 42 to
110 ppmv for distillate oil, based on the data provided by gas
turbine manufacturers and shown in Figures 5-10 and 5-11. A
carefully designed SCR system can achieve NOX reduction
efficiencies as high as 90 percent, with ammonia slip levels of
10 ppmv or less for natural gas and low-sulfur (<0.3 percent by
weight) fuel applications.^4
As discussed for wet injection in Sections 5.1.4 and
5.2.2".4, controlled NOX emission levels for natural gas range
from 25 to 42 ppmv for natural gas fuel and from 42 to 110 ppmv
for -oil fuel. Applying a 90 percent reduction efficiency for .
SCR, NOX levels can be theoretically reduced to 2.5 to 4.2 and
4.2 to 11.0 ppmv for natural gas and oil fuels, respectively.
For oil fuels and other sulfur-bearing fuels, a reduction
efficiency of 90 percent requires special design considerations
to address potential operational problems caused by the sulfur
content in the fuel. This subject is discussed in Section 5.3.2.
The final controlled NOX emission level depends upon the NOX
level exiting the turbine and the achievable SCR reduction
efficiency.
Test reports provided by SCAQMD include three gas turbine
combined cycle installations fired with natural gas that have
achieved NOX emission levels of 3.4 to 7.2 ppmv, referenced to
15 percent oxygen. The NOX and CO emissions reported for these
tests are shown in Table 5-17. Ammonia slip levels were not
reported. Ammonia slip levels were reported, however, in a
summary of emission tests for 13 SCR installations and are
presented in Table 5-18.68 For these sites, operating on natural
gas fuel, the NOX reduction efficiency of the catalyst ranges
5-74
-------
TABLE 5-17.
EMISSIONS TESTS RESULTS FOR GAS TURBINES USING
STEAM INJECTION PLUS SCR69"71
Test
No.
1
2
3
Gas turbine
model
MS7001E
MS7001E
MS6001B
Output,
MW
82.8
79.7
33.8
Fuel
Natural gas 4-
refineiy gas mixture
Natural gas +
refinery gas +
butane mixture
LPG + refinery gas
mixture
NOX emissions, ppmv (Ib/hr)
Uncontrolled
154
148
148
Wet
injection
42
42
42
Wet injection
+ SCR
5.66
(25.2)
7.17
(31.7)
3.36
(5.82)
CO, ppmv
<2.00
<2.00
<2.00
5-75
-------
TURAL
00
1
1
-
0
s
00
"*
r-
o
^
?}
LM2500 1 1
S
-
(S
S
q
§
o
£
0
MS7001E
a
o
*
•*
00
o\
oo
^
O
i
o
MS6001 1
§
->
oo
§
en
S
-<
Z
1
p.
MS6001
U
o
s
^J
BO
c
ea
1
o
1
3
e
*
t
o
GB
'S,
1
>n
•8
3
3
E
o
«
«^
CO
1
Ui
5
<*5 ea)
(l '§ § |
C ^ 3 '1
U CO g
^ . i g*
^ 1> *» M
C «S i^ - C
§3 » § 1
g ^ .2 2 *
|l!l|
jSoTI-1
BllJS
iilli
ffl go B •••) 3
I § Jl^ §*
HiL
?-i- 1 o-a
i 1 « « § 'a
J - a S J 2
? 1 2 "8 t «
1 1 a a I 8
•S "• 5 'S ° ii
o « » op n "
•S K B | a <
U Z H_P Z Z
«r^ o-o «
-------
from 60 to 96 percent, with most reduction efficiencies between
80 and 90 percent. Ammonia slip levels range from 1 to 35 ppmv.
The site with the 35 ppmv ammonia slip level is unique in that it
is the only site identified in the United States that uses only
SCR rather than a combination of SCR and wet injection to reduce
NOX emissions. With the exception of this site, all NHg slip
levels in Table 5-18 that are based on test data are less than
10 ppmv. Based on information received from catalyst vendors, it
is expected that an SCR system operating downstream of a gas
turbine without wet injection could be designed to limit ammonia
slip levels to 10 ppmv or less. 4 No test data are available for
SCR operation on gas turbines fired with distillate oil fuels.
5.5 EFFECT OF ADDING A DUCT BURNER IN HRSG APPLICATIONS
A duct burner is often added in cogeneration and combined
cycle applications to increase the steam capacity of the HRSG
(see Section 4.2.2). Duct burners in gas turbine exhaust streams
consist of pipes or small burners that are placed in the exhaust .
gas.stream to allow firing of additional fuel, usually natural. .
gas. Duct burners can raise gas turbine exhaust temperatures to
1000°C (2000°F), but a more common temperature is 760°C (1400°F).
The gas turbine exhaust is the source of oxygen for the duct
burner.
Figure 5-29 shows a typical natural gas-fired duct burner
installation. Figure 5-30 is a cross-sectional view of one style
of duct burner that incorporates design features to reduce NOY.
«A.
In this low-NOx design, natural gas exits the orifice in the
manifold and mixes with the gas turbine exhaust entering through
a small slot between the casing and the gas manifold. This
mixture forms a jet diffusion flame that causes the recirculation
shown in Zone "A." Due to the limited amount of turbine exhaust
that can enter Zone A, combustion in this zone is fuel-rich. As
the burning gas jet exits into Zone "B," it mixes with combustion
products that are recirculated by the flow eddies behind the
wings of the stabilizer casing. The flame then expands,into the
turbine exhaust gas stream, where combustion is completed.
5-77
-------
u.
3
CN
r-
CO
(3
"t*
i i i
01 00
lit
-------
GAS
TURBINE
EXHAUST
GAS
GAS
MANIFOLD
STABILIZER
CASING
Figure 5-30. Cross-sectional view of a low-NO,, duct burner.73'74
5-79
-------
For oil-fired burners, the design principles of the burner
are the same. However, the physical layout is slightly
different, as shown in Figure 5-31. Turbine exhaust gas is
supplied in substoichiometric quantities by a slip stream duct to
the burner. This slip stream supplies the combustion air for the
fuel-rich Zone A. The flame shield produces the flow eddies,
which recirculate the combustion products into Zone B.7^
Most duct burners now in service fire natural gas. In all
cases, a duct burner will produce a relatively small level of NOX
emissions during operation (See Section 4.2.2), but the net
impact on total exhaust emissions (i.e., the gas turbine plus the
duct burner) varies with operating conditions, and in some cases
may even reduce the overall NOX emissions. Table 5-19 shows the
NO., emissions measured at. one site upstream and downstream of a
Jv
duct burner. This table shows that NO., emissions are reduced
JL
across the duct burner in five of the eight test runs.
The reason for this net NOX reduction is not known, but it
is believed to be a result of the reburning process in which the
intermediate combustion products from the duct burner interact
with the NOX already present in the gas turbine exhaust. The
manufacturer of the burner whose emission test results are shown
in Table 5-19 states that the following conditions are necessary
for reburning to occur:
1. The burner flame must produce a high temperature in a
fuel-rich zone;
2. A portion of the turbine exhaust containing NOX must be
introduced into the localized fuel-rich zone with a residence
time sufficient for the reburning process to convert the turbine
NOX to N2 and 02; and
3. The burner fuel should contain no FBN.7^
In general, sites using a high degree of supplementary
'firing have the highest potential for a significant amount of
reburning. In practice, only a limited number of sites achieve
these reburning conditions due to specific plant operating
requirements.7{*
5-80
-------
tn
r-
rn
t>
fl
-H
•H
o
O
14-1
T)
(1)
CJ
&1
-H
CO
0)
0)"
4J
U
=3
•O
i
i
o
H
m
0)
$-1
3
Cn
•H
5-81
-------
>
r-
PA
fa
&
§
E-i
B
Q
rfj
rt/
K
W
£
<;
i
w
Di
2
«
Q
111
Oi
D
rrt
3
!
rrt
I
H
co
r/\
w^
H
w
>
o
z
CT>
1
ID
j
m
1
EH
i
*O
4rt
O
-o
§
&
u
2
0
&
|
•O
**
Q
u
.1
is
JO
4^
Q
60
ll
* i
X) a.
t3
3
a
1
§
1
§.
a
o
o
o
a
• 3
^
a
0
1
X
O
Z
la
. BQ
nx
O
z
3
J3
o"
z
s
s
> 3
£ •—
. BQ
-X
o
Z
!
0s1
Z
S
S
la' *
£ 2
- BQ
0*
z
t3 §
•3 1
-" J3
1^
.9 oa
8*^"J
2
S %
"* e>&
<2 .2 £
I -^>-2*
M -9 '2
£
2
A
o
2:
ts
f2
r~
"?
i
9
r~
in
in
^
Oh
o
6
•«t
«
ON
2
0*
•— •
es
oo
00
en
en
S
O
00
en
en
-"
t-N
6
8
6
ON
00*
•n
en
*i^
6
00
00
in
CM
1-
»•*
6
en
p»
>n
en
rn
ON
5
o
q
in
en
«s
ci
-"
8
o
o
r-
00*
•n
00
*•"*
"^
6
in
£
Tf
en
0
q
in
M
00
o
^
8
o
in
^t
en
en
ON
t
•<*•
O
9
ON
en
00
_
•n
i^
0
oo
in
00
^_
-S
0
•*
•^
00
in
K
en
*"*
^
O
0
rl
en
^
n
•n
1
9
es
NO
«n
V-M
o
t5.
es
6
r»
ON
•n
C4
ON
en
6
ON
en
00
C";
VO
en
*"*
3
o
xn
^^
en
vO
O
o
1
00
en
r*
9
C1^
4
en
i-^
en
6
r-
4
Tj
00
en
6
00
•n
es
q
e~4
5=
3
o
q
en
en
t^
q
rn
VM
6
H
•t
es
en
6
ON
«s
f^K
in
6
•n
vd
00
ON
es
^
f-^
3
6
—
*
_^
00
5-82
-------
5.6 ALTERNATE FUELS
Because thermal NOX production is an exponential function of
flame temperature (see Section 4.1.1), it follows that using
fuels with flame temperatures lower than those of natural gas or
distillate oils results in lower thermal NOX emissions.
Coal-derived gas and methanol have demonstrated lower NOX
emissions than more conventional natural gas or oil fuels. For
applications using fuels with high FBN contents, switching to a
fuel with a lower FBN content will reduce thermal NO... formation
Jv
and thereby lower total NOX emissions.
5.6.1 Coal-Derived Gas
Combustor rig tests have demonstrated that burning
coal-derived gas (coal gas) that has been treated to remove FBN
produces approximately 30 percent of the NOX emission levels
experienced when burning natural gas. This is because coal gas
has a low heat energy level of around 300 Btu or less, which
results in a flame temperature lower than that of natural gas.79
The.cost associated with producing coal gas suitable for
combustion in a gas turbine has made this alternative
economically unattractive, but recent advances in coal
gasification technology have renewed interest in this fuel.
A coal gas-fueled power plant is currently operating in the
United States at a Dow Chemical plant in Placquemine, Louisiana.
This facility operates with a subsidy from the Federal
Government, which compensates for the price difference between
coal gas and conventional fuels. Several commercial projects
have been recently announced using technology developed by
Texaco, Shell, Dow Chemical, and the U.S. Department of Energy.
Facilities have been permitted for construction in Massachusetts
and Delaware.80
A demonstration facility, known as Cool Water, operated
using coal gas for 5 years in Southern California in the early
1980's. The NOX emissions were reported at 0.07 lb/MMBtu.80
Fuel analysis data is not available to convert this NOX emission
level to a ppmv figure. No other emissions data are available.
5-83
-------
5.6.2 Methanol
Methanol has a flame temperature of 1925°C (3500°F) versus
2015°C (3660°F) for natural gas and greater than 2100°C (3800°F)
for distillate oils. As a result, the NOX emission levels when
burning methanol are lower than those for either natural gas or
distillate oils.
Table 5-20 presents NOY emission data for a full-scale
J^
turbine firing methanol. The NO., emissions from firing methanol
j\*
without water injection ranged from 41 to 60 ppmv and averaged
49 ppmv. This test also indicated that methanol increases
turbine output due to the higher mass flows that result from
methanol firing. Methanol firing increased CO and HC emissions
slightly compared to the same turbine's firing distillate oil
with water injection. All other aspects of turbine performance
were as good when firing methanol as when the turbine fired
natural gas or distillate oil.^^ Turbine maintenance
requirements were estimated to be lower and turbine life was
estimated to be longer on methanol fuel than on distillate oil.
fuel because methanol produced fewer deposits in the combustor
and power turbine.
Table 5-20 also presents NOV emission data for methanol
Jv
firing with water injection. At water-to-fuel ratios from
O.ll to 0.24, NO., emissions when firing methanol range from 17 to
J\f
28 ppmv, a reduction of 42 to 65 percent.
In a study conducted at an existing 3.2 MW gas turbine
installation in 1984, a gas turbine was modified to burn
methanol. This study was conducted at the University of
California at Davis and was -sponsored by the California Energy
Commission. A new fuel delivery system for methanol was
required, but the only major modifications required for the
turbine used in this study were new fuel manifolds and nozzles.
Tests conducted burning methanol showed no visible smoke
emissions, and only minor increases in CO emissions. Figure 5-32
shows the NOX emissions measured while burning methanol and
natural gas. Reductions of up to 65 percent were achieved, as
NOX emissions were 22 to 38 ppm when burning methanol versus
5-84
-------
TABLE 5-20.
'NO EMISSIONS TEST DATA FOR A
FIRING METHANOL AT BASELOADa'81
GAS TURBINE
Test
A
B
C
D
E
F
G
H
I
J
K
L
M
AVERAGE
N
0
P
Q
W/F ratio,
Ib/lb
0
0
0
0
0
0
0
0
0
0
0
0
0
0.11
0.23
0.23
0.24
NOY emissions
ISO
conditions,
ppm at 15% 02
41
45
48
49
60
47
53
48
51
52
41
47
48
49
28
17
18
18
N0x;
reduction,
percent-"
0
0
0
0
0
0
0
0
0
0
0
0
0
42.2
65.2
62.7
62.7
^Baseload =• 25 MW output
^Calculated using the average of the uncontrolled emissions
5-85
-------
120
100
o
4»
I/I
rt
8
-------
62 to 100 ppm for natural gas. In addition to the intrinsically
lower NOX production, water can be readily mixed with methanol
prior to delivery to the turbine to obtain the additional NOX
reduction levels achievable with wet injection. Gas turbine
performance characteristics, including startup, acceleration,
load changes, and full load power, were all deemed acceptable by
the turbine manufacturer.83
The current economics of using methanol as a primary fuel
are not attractive. There are no confirmed commercial
methanol-fueled gas turbine installations in the United States.
5.7 SELECTIVE NONCATALYTIC REDUCTION
Selective noncatalytic reduction (SNCR) is an add-on
*
technology that reduces NOX using ammonia or urea injection
similar to SCR but operates at a higher temperature. At this
higher operating temperature of 870° to 1200°C (1600° to 2200°F),
the following reaction occurs:84
NOX + NH3 + 02 + H2O + (H2) •* N2 + H20.
• This reaction occurs without requiring a catalyst, . -
effectively reducing NC- to nitrogen and water. The operating
J\,
temperature can be lowered from 870°C (1600°F) to 700°C (1300°F)
by injecting hydrogen (H2) with the ammonia, as is shown in the
above equation.
Above the upper temperature limit, the following reaction
occurs:84
NH3 + 02 -» NOX + H2O.
Levels of NOY emissions increase when injecting ammonia or
Jt,
urea into the flue gas at temperatures above the upper
temperature limits of 1200°C (2200°F).
Since SNCR does not require a catalyst, this process is more
attractive than SCR from an economic standpoint. The operating
temperature window, however, is not compatible with gas turbine
exhaust temperatures, which do not exceed 600°C (1100°F).
Additionally, the residence time required for the reaction is
approximately 100 milliseconds, which is relatively slow for gas
turbine operating flow velocities.85
5-87
-------
It may be feasible, however, to initiate this reaction in
the gas turbine where operating temperatures fall within the
reaction window, if suitable gas turbine modifications and
injection systems can be developed.**5 This control technology
has not been applied to gas turbines to date.
5.8 CATALYTIC COMBUSTION
5.8.1 Process Description
In a catalytic combustor, fuel and air are premixed into a
fuel-lean mixture (fuel/air ratio of approximately 0.02) and then
pass into a catalyst bed. In the bed, the mixture oxidizes
without forming a high-temperature flame front. Peak combustion
temperatures can be limited to below 1540°C (2800°F), which is
below the temperature at which significant amounts of thermal NOX
begin to form.^^ An example of a lean catalytic combustor is
shown in Figure 5-33.
Catalytic combustors can also be designed to operate in a
rich/lean configuration, as shown in Figure 5-34. In this
configuration, the air and fuel are premixed to form a fuel-rich
mixture, which passes through a first stage catalyst where
combustion begins. Secondary air is then added to produce a lean
mixture, and combustion is completed in a second stage catalyst
bed.89
5.8.2 Applicability
Catalytic combustion techniques apply to all combustor types
and are effective on both distillate oil- and natural gas-fired
turbines. Because of the limited operating temperature range,
catalytic combustors may not be easily applied to gas turbines
subject to rapid load changes (such as utility peaking
turbines).9^ Gas turbines that operate continuously at base load
(such as industrial cogeneration applications) would not be as
adversely affected by any limits on load following capability.91
5.8.3 Development Status
Presently, the development of catalytic combustors has been
limited to bench-scale tests of prototype combustors. The major
problem is the development of a catalyst that will have an
acceptable life in the high-temperature and -pressure environment
5-88
-------
QC
i—•
<
-HH-
_J OS
UJ >->
£ <
o
oo
o
JJ
CD
O
U
U
-H
OJ
fd
U_
m
m
i
in
OJ
i-i
en
-H
5-89
-------
X
00
CO
O
-U
CQ
O
O
U
•H
4J
01
4J
rfl
U
(U
u
-H
m
i
in
OJ
^
d
Cn
5-90
-------
of gas turbine combustors. Additional problems that must be
solved are combustor ignition and how to design the catalyst to
operate over the full gas turbine operating range (idle to full
load).92
5.9 OFFSHORE OIL PLATFORM APPLICATIONS
Gas turbines are used on offshore platforms to meet
compression and electrical power requirements. This application
presents unique challenges for NOX emissions control due to the
duty cycle, lack of a potable water source for wet injection, and
limited space and weight considerations. The duty cycle for
electric power applications of offshore platforms is unique.
This duty cycle is subject to frequent load changes that can
instantaneously increase or decrease by as much as a factor of
10.93 Fluctuating loads result in substantial swings in turbine
exhaust gas temperatures and flow rates. This presents a problem
for SCR applications because the NOX reduction efficiency depends
upon temperature and space velocity (see Section 5.3.3) .
The lack of a potable water supply means that water must be
'shipped to the platform or sea water must be desalinated and
treated. The limited space and weight requirements associated
with an SCR system may also have an impact on capital costs of
the platform.
A 4-year study is underway for the Santa Barbara County Air
Pollution Control Board to evaluate suitable NOY control
Jt
techniques for offshore applications. The goals of the study are
to reduce turbine NOX emissions at full load to 9 ppmv, corrected
to 15 percent O2, firing platform gas fuel and to achieve part
load reductions of 50 percent. The study consists of two phases.
The first phase, an engineering evaluation of available and
emerging emission control technologies, is completed. The second
phase will select the final control technologies and develop
these technologies for offshore platform applications. Phase I
5-91
-------
of this study concludes that the technologies with the highest
estimated probability for success in offshore applications are:
- Water injection plus SCR (80 percent);
- Methanol fuel plus SCR (70 percent);
- Lean premixed combustion plus SCR (65 percent); and
- Steam dilution of fuel prior to combustion plus SCR
(65 percent).
A key conclusion drawn from Phase I of this study is that
none of the above technologies or combination of technologies in
offshore platform applications currently has a high probability
of successfully achieving the NO., emission reduction goals of
JS»
this study without substantial cost and impacts to platform and
turbine operations, added safety considerations, and other
environmental concerns. These issues will be further studied in
Phase II for the above control technologies.
5.10 REFERENCES FOR CHAPTER 5
1. National Archives and Records Administration. Code of
•Federal Regulations. 40 CFR 60.332. Subpart GG.
Washington, D.C. Office of the Federal Register. July
1989.
2. South Coast Air Quality Management District. Emissions of
Oxides of Nitrogen from Stationary Gas Turbines. Rule 1134.
Los Angeles. August 4, 1989.
3. Letter and attachments from Conroy, D. B., U.S. EPA Region
I, to Neuffer, W. J.„ EPA/ISB. January 15, 1992. Review of
draft gas turbine ACT document.
4. Northeast States For Coordinated Air Use Management.
Recommendation On NOX RACT for Industrial Boilers, Intern
Combustion Engines and Gas Turbines. September 18, 1992.
5. Letter and attachment from Leonard, G. L., General Electric
Company, to Snyder, R. B., MRI. February 1991. Response to
gas turbine questionnaire.
6. Letters and attachments from Schorr, M., General Electric
Company, to Snyder, R. B., MRI. March, April 1991.
Response to gas turbine questionnaire.
7. Letter and attachments from Gurmani, A., Asea Brown Boveri,
to Snyder, R. B., MRI. February 4, 1991. Response to gas
turbine questionnaire.
5-92
-------
8. Letter and attachment from Swingle, R., Solar Turbines
Incorporated, to Snyder, R. B., MRI. February 1991.
Response to gas turbine questionnaire.
9. Letter and attachment from Kimsey, D. L., Allison Gas
Turbine Division of General Motors, to Snyder, R. B., MRI.
February 1991. Response to gas turbine questionnaire.
10. Letter and attachment from Kraemer, H., Siemens Power
Corporation, to Snyder, R. B., January 1991. Response to
gas turbine questionnaire.
11. Letter and attachments from Antos, R. J., Westinghouse
Electric Corporation, to Neuffer, W. J., EPA. September 11,
1991. Review of Draft Gas Turbine ACT document.
12. Letter and attachment from Bogus, A. S., Garrett Turbine
Engine Company, to Dalrymple, D., Radian Corporation.
April 13, 1983. Stationary gas turbines, p. 7.
13. General Electric Company. General Electric Heavy-Duty Gas
Turbines. Schenectady, New York. 1983. Section 6.
14. Letter from Dvorak, United Technologies Corporation, Power
Systems Division, to Goodwin, D. R., EPA. April 7, 1978.
Limits on water used for injection into the FT4 gas turbine
combustion chamber to control emissions. '"_
15. Letter and attachments from Solt, J. C., Solar Turbines
Incorporated, to Noble, E., EPA. August 23, 1983. NSPS
review.
16. General Motors. General Motors Response to Four-Year Review
Questions on the NSPS for Stationary.Gas Turbines.
Submitted to U. S. Environmental Protection Agency.
Research Triangle Park, NC. July 5, 1983. 144 Federal
Register 176. September 10, 1979. 52 pp.
17. Letter and attachments from Rosen, V., Siemens AG, to
Neuffer, W. J., EPA/ISB. August 30, 1991. Review of Draft
Gas Turbine ACT document.
18. Letter and attachments from Sailer, E. D., General Electric
Marine and Industrial Engines, to Neuffer, W. J., EPA/ISB.
August 29, 1991. Review of Draft Gas Turbine ACT document.
19. Letter and attachments from Mincy, J. E., Nalco Fuel Tech,
to Neuffer, W.J., EPA/ISB. September 9, 1991. Review of
draft gas turbine ACT document.
5-93
-------
20
21,
22
23
U. S. Environmental Protection Agency. Background
Information Document, Review of 1979 Gas Turbine New Source
Performance Standard. Research Triangle Park, NC. Prepared
by Radian Corporation under Contract No. 68-02-3816. 1985.
p. 4-36.
Letters and attachments from Leonard, G. L., General
Electric Company, to Snyder, R. B., MRI. May 24, 1991.
Response to gas turbine questionnaire.
Telecon. Snyder, R., MRI, with Rayome, D., U.S. Turbine
r*^T-n<^T-a t" i i-in Ma^r O "3 1QQ1 (Tla a *-n •vl-i •{ in ev MO n,ont-v^l 3n<^
Corporation. May 23, 1991
maintenance impacts.
Gas turbine NO control and
Letter and attachment from Gurmani, A., Asea Brown Boveri,
to Snyder, R. B., MRI. May 30, 1991. Response to gas
turbine questionnaire.
24. Letter and attachments from van der Linden, S.,, Asea Brown
Boveri, to Neuffer, W. J., EPA/ISB. September 16, 1991.
Review of draft gas turbine ACT document.
25. Wilkes, C., and R. C. Russell (General Electric Company).
The Effects of Fuel Bound Nitrogen Concentration and Water •
Injection on NOX Emissions from a 75 MW Gas Turbine.
. Presented at the Gas Turbine Conference & Products Show.
London, England. April 9-13, 1978. ASME Paper No.
78-GT-89. p. 1.
26. Reference 20, pp. 4-33, 4-34.
27. Reference 20, pp. 4-39 through 4-47.
28. Letter and attachments from Valentine, J. M., Energy and
Environmental Partners, to Neuffer, W. J., EPA/ISB.
April 26, 1991. Control of NOX emissions using water-in-oil
emulsions.
29. Reference 20, pp. 4-48 thru 4-50.
30. Sailer, E. D. NOX Abatement With Steam Injection on
Aircraft Derivative Gas Turbines. General Electric Marine
and Industrial Engines. Presented to the American
Cogeneration Association. Scottsdale, AZ. march 13, 1989.
5 pp.
31. Becker, E., M. Kosanovich, and G. Cordonna. Catalyst Design
for Emission Control of Carbon Monoxide and Hydrocarbons
From Gas Engines. Johnson Matthey. Wayne, PA. For
presentation at the 81st Annual Air Pollution Control
Association meeting. Dallas. June 19-24, 1988. 16 pp.
32. Reference 20, p. 4-51.
5-94
-------
33. Schorr, M. NOX Control for Gas Turbines: Regulations and
Technology. General Electric Company. Schenectady, NY.
For presentation at the Council of Industrial Boiler Owners
NOX Control IV Conference. February 11-12, 1991. 11 pp.
34. Reference 20, pp. 4-2 thru 4-5.
35. Maghon, H., and A. Krutzer (Siemens Product Group KWU,
Muelheim, .Germany) and H. Termuehlen (Utility Power
Corporation, Bradenton, FL). The V84 Gas Turbine Designed
for Reliable Base Load and Peaking Duty. Presented at the
American Power Conference. Chicago. April 18-20, 1988.
20 pp.
36. Meeting. Barnett, K., Radian Corporation, to File.
February 6, 1984. Discuss Rolls-Royce Emission Testing
Procedures and Low-N0x Combustors. p. 3.
37. U. S. Environmental Protection Agency. Standards Support
and Environmental Impact Statement. Volume 1: Proposed
Standards of Performance for Stationary Gas Turbines.
Research Triangle Park, NC. Publication No.
EPA 450/2-77-017a. September 1977. pp. 4-48 - 4-83.
38. Touchton, G. L., J. F. Savelli, and M. B. Hilt (General
Electric Company, U.S.A.). Emission Performance and Control
Techniques for Industrial Gas Turbines. Schenectady,
New York. Gas Turbine Reference Library No. GER-2486H.
1982. p. 351.
39. Johnson, R. H. and C. Wilkes (General Electric Company).
Emissions Performance of Utility and Industrial Gas
Turbines. Presented at the American Power Conference.
April 23-25, 1979. Schenectady, New York. p. 5.
40. Reference 20, p. 4-5.
41. Angello, L. (Electric Power Research Institute, Palo Alto,
CA) and P. Lowe (InTech, Inc., Potomac, MD). Gas Turbine
Nitrogen Oxide (NOX) Control. Current Technologies and
Operating Combustion Experiences. Presented at the 1989
Joint Symposium on Stationary NOX Control. San Francisco.
March 6-9, 1989. 18 pp.
42. Guthan, D. C. and C. Wilkes (General Electric Company,
U.S.A.). Emission Control and Hardware Technology.
Schenectady, New York. Gas Turbine Reference Library
No. GERP3125. 1981. p. 4.
5-95
-------
43. Letter and attachments from Malloy, M. K., Rolls-Royce
Limited, to Jennings, M., Radian Corporation. May 12, 1983.
8 pp. Response to questionnaire concerning emission levels
of Rolls-Royce gas turbines and of emission control
techniques offered.
44. McKnight, D. (Rolls-Royce Limited). Development of a
Compact Gas Turbine Combustor to Give Extended Life and
Acceptable Exhaust Emissions. Journal of Engineering for
Power. UQ1(3):101. July 1979.
45. Reference 36, Attachment 1.
46. Smith, K. 0., and P. B. Roberts. Development of a Low NOX
Industrial Gas Turbine Combustor. Solar Turbines Inc. San
Diego, CA. Presented at the Canadian Gas Association
Symposium on Industrial Application of Gas Turbines. Banff,
Alberta. October 16-18, 1991. 18 pp.
47. Letter and attachments from Cull, C., General Electric
Company, to Snyder, R. B., MRI. April 1991. Response to
request for published General Electric Company presentation
materials.
48. Maghon, H., and L. Schellhorn (Siemens Product Group KWU,
Muelheim, Germany); J. Becker and J. Kugler (Delmorva
Power & Light Company, Wilmington, DE); and H. Termuehlen -
(Utility Power Corporation, Bradenton, FL). Gas Turbine
Operating Performance and Considerations for Combined Cycle
Conversion at Hay Road Power Station. Presented at the
American Power Conference. Chicago. April 23-25, 1990.
12 pp.
49. Reference 20, p. 4-10.
50. Letter and attachments from Swingle, R., Solar Turbines
Incorporated, to Snyder, R., MRI. May 21, 1991. Low-N0x
gas turbine information.
51. Smock, R. Utility Generation Report - Gas turbines reach
9 ppm nitrogen oxide emissions dry. Power Engineering.
9_£(3) :10. March 1992.
52. Davis, L. Dry Low NOX Combustion Systems for GE Heavy-Duty
Gas Turbines. General Electric Company. Schenectady, NY.
Presented at 35th GE Turbine Sate-of-the-Art Technology
Seminar. August 1991. 10 pp.
5-96
-------
53. Magnon, H., and P. Berenbrink (Siemens KWU) and
H. Termuehlen and G. Gartner (Siemens Power Corporation).
Progress in NOX and CO Emission Reduction of Gas Turbines.
Presented at tne Joint American Society of Mechanical
Engineers/Institute of Electronic and Electrical Engineers
Power Generation Conference. Boston. October 21-25, 1990.
7 pp.
54. Letter and attachments from King, D., General Electric
Industrial Power Systems Sales, to Snyder, R. B., MRI.
August 25, 1992. Performance and emission levels for
industrial gas turbines.
55. Cutrone, M., and M. Hilt (General Electric Company,
Schenectady, NY); A. Goyal and E. Ekstedt (General Electric
Company, Evandale, OH); and J. Notardonato (NASA/Lewis
Research Center, Cleveland, OH). Evaluation of Advanced
Combustors for Dry NOX Suppression With Nitrogen Bearing
Fuels in Utility and Industrial Gas Turbines. Journal of
Engineering for Power. 104;429-438. April 1982.
56. Stambler, I. Strict NOX Codes Call for Advanced Control
Technology. Gas Turbine World. 13 (4):58.
September-October 1983. p. 58.
57. Novick, A. S., and D. L. Troth (Detroit Diesel Allison) and
J. Notardonato (NASA Lewis Research Center.) Multifuel
Evaluation of Rich/Quench/Lean Combustor. ASME Paper No.
83-GT-140. p. 6.
58. Lew, H. G. (Westinghouse Electric Company) et al. Low
and Fuel Flexible Gas Turbine Combustors. Presented at
Gas Turbine Conference & Products Show. Houston, TX.
March 9-12, 1981. ASME Paper No. 81-GT-99. p. 10.
59. McVey, J. B., R. A. Sederquist, J. B. Kennedy, and L. A.
Angello (United Technologies Research Center). Testing of a
Full-Scale Staged Combustor Operating with a Synthetic
Liquid Fuel. ASME Paper No. 83-GT-27. p. 8.
60. Allison-DOE Run Gas Turbine Directly on Pulverized Coal. Gas
Turbine World. 21(6):39. November-December 1991.
61. Minutes of meeting dated February 5, 1992, among
representatives of the Institute of Clean Air Companies
(formerly Industrial Gas Cleaning Institute), U.S.
Environmental Protection Agency, and Midwest Research
Institute. December 10, 1991. Review of draft gas turbine
ACT document.
5-97
-------
62. Radian Corporation. Evaluation of Oil-Fired Gas Turbine
Selective Catalytic Reduction (SCR) N02 Control. Prepared
for the Electric Power Research Institute, Palo Alto, CA,
and the Gas Research Institute (Chicago). EPRI GS-7056.
December 1990. pp. 4-7.
63. Benson, C., G. Chittick, and R. Wilson. (Arthur D. Little,
Inc.). Selective Catalytic Reduction Technology for
Cogeneration Plants. Prepared for New England Cogeneration
Association. November 1988. 54 pp.
64. Letter and attachments from Smith, J. C., Institute of Clean
Air Companies, to Neuffer, W. J., EPA/ISB. May 14, 1992.
Response to EPA questionnaire regarding flue gas treatment
processes for emission reductions dated March 12, 1992.
65. Letter and attachments from Craig, R. J., Unocal Science and
Technology Division of Unocal Corporation, to Lee, L.,
California Air Resources Board. July 24, 1991. Gas turbine
SCR installation experience and information.
66. May, P. A., L. M. Campbell,, and K. L. Johnson (Radian
Corporation). Environmental and Economic Evaluation of Gas
Turbine SCR NOX Control. Research Triangle Park, NC.
Presented at the 1991 Joint EPRI/EPA Symposium for
Stationary Combustion NOX Control. March 1991. Volume 2.
18 pp.
67. Durham, M. D., T. G. Ebner, M. R. Burkhardt, and F. J.
Sagan. Development of An Ammonia Slip Monitor for Process
Control of NHj Based NOX Control Technologies. ADA
Technologies, Inc. Presented at the Continuous Emission
Monitoring Conference, Air and Waste Management Association.
Chicago. November 12-15, 1989. 18 pp.
68. Field Survey of SCR Gas Turbine Operating Experience.
Prepared for the Electric Power Research Institute. Palo
Alto, CA. May, 1991.
69. Harris, B., and J. Steiner (Pope and Steiner Environmental
Services). Source Test Report. South Coast Air Quality
Management District. Los Angeles. PS-90-2107., April 11,
1990. . '
70. Harris, B., and J. Steiner (Pope and Steiner Environmental
Services). Source Test Report. South Coast Air Quality
Management District. Los Angeles. PS-90-2108. April 12,
1990.
71. Harris, B., and J. Steiner (Pope and Steiner Environmental
Services). Source Test Report. South Coast Air Quality
Management District. Los Angeles. PS-90-2148. May 1,
1990.
5-98
-------
72. Reference 20, pp. 3-20.
73. Letter and attachments from Brown, R., Coen Company, Inc.,
to Dalrymple, D., Radian Corporation. August 16, 1983.
Duct Burner Emissions in Turbine Exhaust Gas Streams.
74. Reference 20, p. 3-21.
75. Reference 20, p. 3-22.
76. Reference 20, pp. 3-19, 4-79, 4-80.
77. Podlensky, J., et al. (GCA Corporation). Emission Test
Report, Crown Zellerbach, Antioch, CA. March 1984.
78. Backlund, J., and A. Spoormaker. Experience with NO
Formation/Reduction Caused by Supplementary Firing or
Natural Gas in Gas Turbine Exhaust Streams. The American
Society of Mechanical Engineers. New York. 85-JPGC-G7-18.
1985. 5 pp.
79. Reference 36, pp. 3-93, 3-94.
80. Smock, R. Coal Gas-fired Combined Cycle Projects Multiply.
Power Engineering. 13_5_(2) :3'2-33 . February 1991.
81.' Weir, A., Jr., W. H. von KleinSmid, and E. A. Danko
(Southern California Edison Company). Test and Evaluation
of Methanol in a Gas Turbine System. Prepared for Electric
Power Research Institute. Palo Alto California.
Publication No. EPRI AP-1712. February 1981.
pp. A-76 through A-78.
82. Reference 81, pp. 5-1, 5-2.
83. Shore, D., and G. Shiomoto (KVB, Incorporated, Irvine, CA)
and G. Bemis (California Energy Commission, Sacramento, CA).
Utilization of Methanol as a Fuel for a Gas Turbine
Cogeneration Plant. Prepared for Electric Power Research
Institute. Chicago. CS-4360, Volume II, EPA Contract
No. 68-02-3695. January 1986.
84. Fellows, W. D. Experience with the Exxon Thermal DeNOx
Process in Utility and Independent Power Production Exxon
Research and Engineering Company. Florham Park, NJ. August
1990. 5 pp.
85. Bernstein, S., and P. Malte (Energy International, Inc.).
Emissions Control for Gas Transmission Engines. Prepared
for the Gas Research Institute. Chicago. Presentation
No. PRES 8070. July 1989. 17 pp.
5-99
-------
86. Krill, W. V., J. P. Kesselring, and E. K. Chu (Acurex
Corporation). Catalytic Combustion for Gas Turbine
Applications. Presented at the Gas Turbine Conference &
Exhibit & Solar Energy Conference. San Diego, CA.
March 12-15, 1979. ASME Paper No. 79-GT-188. p. 4.
87. Reference 58, p. 6.
88. Reference 86, p. 8.
89. Washam, R. M. (General Electric Company). Dry Low NOX
Combustion System for Utility Gas Turbine. Presented at the
1983 Joint Power Generation Conference. Indianapolis, IN.
ASME Paper No. 83-JPGC-GT-13. p. 1.
90. Reference 86, p. 7.
91. Reference 20, p. 4-23.
92. Reference 37, p. 4-88
93. Little, A.D. Offshore Gas Turbine NO- Control Technology
Development Program. Phase I--Technology Evaluation.
Prepared for Santa Barbara County Air Pollution Control
Board. August 1989. 130 pp.
5-100
-------
6.0 CONTROL COSTS
Capital and annual costs are presented in this chapter for
the nitrogen oxide (NOX) control techniques described in
Chapter 5.0. These control techniques are water and steam
injection, low-NOx combustion, and selective catalytic
reduction (SCR) used in combination with these controls. Model
plants were developed to evaluate the control techniques for a
*
range of gas turbine sizes, fuel types, and annual operating
hours. The gas turbines chosen for these model plants range in
size from 1.1 to 160 megawatts (MW) (1,500 to 215,000 horsepower
[hp]) and include both aeroderivative and heavy-duty turbines.
Model plants were developed for both natural gas and distillate
oil-fuels. For offshore oil production platforms, cost
information was available only for one turbine model.
The life of the control equipment depends upon many factors,
including application, operating environment, maintenance
practices, and materials of construction. For this study, a
15-year life was chosen.
Both new and retrofit costs are presented in this chapter.
For water and steam injection, these costs were assumed to be the
same because most of the water treatment system installation can
be completed while the plant is operating and because gas turbine
nozzle replacement and piping connections to the treated water
supply can be performed during a scheduled downtime for
maintenance. Estimated costs are provided for both new and
retrofit low-NOx combustion applications. No SCR retrofit
applications were identified, and costs for SCR retrofit
applications were not available. The cost to retrofit an
existing gas turbine installation with SCR would be considerably
higher than the costs shown for a new installation, especially
for combined cycle and cogeneration installations where the
6-1
-------
heat- recovery steam generator (HRSG) would have to be modified or
replaced to accommodate the catalyst reactor.
This chapter is organized into five sections. Water and
steam injection costs are described in Section 6.1. Low-NOY
-J\.
combustor costs are summarized in Section 6.2. Costs for SCR
used in combination with water or steam injection or low-NOx
combustion are described in Section 6.3. Water injection and SCR
costs for offshore gas turbines are presented in Section 6.4, and
references are listed in Section 6.5.
6.1 WATER AND STEAM INJECTION AND OIL-IN-WATER EMULSION
Ten gas turbines models were selected, and from these
turbines 24 model plants were developed using water or steam
injection or water-in-oil emulsion to control NO... emissions.
Jt
These 24 models, shown in Table 6-1, characterize variations in
existing units with respect to turbine size, type (i.e., aero-
derivative vs. heavy duty), operating hours, and type of fuel.
A total of 24 model plants were developed; 16 of these were
continuous-duty (8,000 hours per year) and 8 were intermittent-
duty (2,000 or 1,000 hours per year). Thirteen of the
continuous-duty model plants burn natural gas fuel; 6 of the
13 use water injection, and 7 use steam injection to reduce NOX
emissions. The three remaining continuous-duty model plants burn
distillate oil fuel and use water injection to reduce NOX
emissions. Of the eight intermittent-duty model plants, six
operate 2,000 hours per year (three natural gas-fueled and three
distillate oil-fueled) , and two operate 1,000 hours- per year
(both distillate oil-fueled). All intermittent-duty model plants
use water rather than steam for NO... reduction because it was
Jv
assumed that the additional capital costs associated with steam-
generating equipment could not be justified for intermittent
service.
Costs were available for applying water-in-oil emulsion
technology to only one gas turbine, and insufficient data were
available to develop costs for a similar water-injected model
plant for this turbine. As a result, the costs and cost
6-2
-------
CO
w
s
^^
9
•s 3 .S
S S
o s
•g §
0 o
0.
.f
§•!
Q QQ QQ Q
as « as a
00 00 08 00 00 CO
!<=8
! aa «~
666666
i i i t i i
O O O O U O
Q Q Q Q Q QQ
Q
3 <
<33 <5B EEC
00 OO OO 09 OO 00 OO
f*k oo i
r* cs i
i = oHffltt-
o a o o
Io >n CQ r* r^
51
- p ^- >o
> OO 00 —
6666666
Q Q Q
XXX
III
000
H a. tu
<" r-
CO CO
CO CO
-------
effectiveness for the water-in-oil emulsion model plant should
not be compared to those of water-injected model plants.
Capital costs are described in Section 6.1.1, annual costs
are described in Section 6.1.2, and emission reductions and the
cost effectiveness of wet injection controls are discussed in
Section 6.1.3. Additional discussion of the cost methodology and
details about some of the cost estimating procedures are provided
in Appendix B.
Fuel rates and water flow rates were calculated for each
model plant using published design power output and efficiency,
expressed as heat rate, in British thermal units per
kilowatt-hour (Btu/kWh)-1 The values for these parameters are
presented in Table 6-2 for each model plant. Fuel rates were
estimated based on the heat rates, the design output, and the
lower heating value (LHV) of the fuel. The LHV's used in this
analysis for natural gas and diesel fuel are 20,610 Btu per pound
(Btu/lb) and 18,330 Btu/lb, respectively, as shown in Table 6-3.2
Water (or steam) injection rates were calculated based on
published fuel rates and water-to-fuel ratios (WFR) provided by
manufacturers. "1 According to a water treatment system
supplier, treatment facilities are designed with a capacity
factor of 1.3.13 An additional 29 percent of the treated water
flow rate is discarded as wastewater.2 Consequently, the water
treatment facility design capacity is 68 percent (1.. 30 x 1.29)
greater than the water (or steam) injection rate.
6.1.1 Capital Costs
The capital costs for each model plant are presented in
Table 6-4. These costs were developed based on methodology in
Reference 2, which is presented in this section. The capital
costs include purchased equipment costs, direct and indirect
installation costs, and contingency costs.
6.1.1.1 Purchased Equipment Costs. Purchased equipment
costs consist of the injection system, the water treatment
system, taxes, and freight. All costs are presented in
1990 dollars.
6-4
-------
CN
I
I a
It
t- 5
It
o! 1!
"
11
o *
§2
— O
vo co
r- S oo oa
O O O O — O
r-" o o o o o
n ts
m 0 <-•
r- ts « >n ""
t~ ^
99999
z z z z z z z
o o o o o o o
0 O 0 0 0 0 0
vi o
• —
o o o
r- o o
f- a. id
g 0> 09
<3 ss
CON-O-W-3.3
CON-O-W-26.3
CON-O-W-83.3
322
>O t-- 00
o° o o'
f- a, w
S S
KR-Q-W-3.3
KR-Q-W-26.8
KR-Q-W-84.7
-o o
^ r-'
3
000
•
-------
TABLE 6-3. FUEL PROPERTIES AND UTILITY AND LABOR RATESa
Fuel properties
Natural gas
Diesel fuel
Factor
20,610
930
18,330
7.21
Utility rates
Natural gas
Diesel fuel
Electricity
Raw water ,
Water treatment
Waste disposal
3.88
0.77
0.06
0.384
1.97
3.82
Labor rate
Operating
Maintenance
25.60
31.20
Units
Btu/lb
Btu/scfc (LHV)
Btu/lb (LHV)
Ib/gal
Reference
Ref. 3
Ref. 3
Ref. 2
Ref. 2
$/scf
$/gal
$/kW-hr
$/l,000 gal
$/l,000 gal
$/l,000 gal
Ref. 4
Ref. 5
Ref . ' a 6 and 7
Ref. 2, escalated ® 5% per
year
Ref. 2, escalated @ 5% per
year
Ref. 2, escalated ® 5% per
year
$/hr
$/hr
Ref. 2, escalated ® 5% per
year
Ref. 2, escalated ® 5% per
year
aAll .costs are average costs in 1990 dollars.
"Natural gas and electricity costs from Reference 4 are the average of the
costs for industrial and commercial customers.
cscf = standard cubic foot.
6-6
-------
(TJ
CO
O
Q
CO
O
B
H
S
O
H
B
H
H
Oi
£
B
CO
8
EH
H
§
VO
W
I
|So
5-E=- +
§"00
= ll£
- «
•jr o
Q «
J= <*.
ap o
17^
iH
00 vo
00 m
^ - O l^
oo r- o >o o o o
O oo Ov en m o en
00 >o O> O
C4 Q tn oo
m * o «i
oo r» e o >n o
i— vo en * — m
0. p W H.
; tn t~- —
9999999
z z z z z z z
8888888
25! P
T
P
E
Cen
MS
MS
H vo m
« «N 00
=*£?
999
z z z
888
00 js_ io_
— of
m •& r-
in «n en
-------
6.1.1.1.1 Water injection system. The injection system
delivers water from the treatment system to the combustor. This
system includes the turbine-mounted injection nozzles, the flow
metering controls, pumps, and hardware and interconnecting piping
from the treatment system to the turbine. On-engine hardware
(the injection nozzles) costs were provided by turbine
manufacturers. ' Flow metering controls and hardware,
pumps, and interconnecting piping costs for all turbines were
calculated using data provided by General Electri^: for four
heavy-duty turbine models.17 No relationship between costs and
either turbine output or water flow was evident, so the sum of
the four costs was divided by the sum of the water flow
requirements for the four turbines. This process yielded a cost
of $4,200 per gallon per minute (gal/min), and this cost, added
to the on-engine hardware costs, was used for all model plants.
6.1.1.1.2 Water treatment system. The water treatment
process, and hence the treatment system components, varies
according to the degree to which the water at a given site must
be treated. For this cost analysis, the water treatment system
includes a reverse osmosis and mixed-bed demineralizer system.
The water treatment system capital cost for each model plant was
estimated based on an equation developed in Reference 2:
WTS = 43,900 X (G)0<5°
where
WTS = water treatment system capital cost, $; and
G » water treatment system design capacity, gal/min.
This equation yields costs that are generally consistent
with the range of costs presented in Reference 18.
6.1.1.1.3 Taxes and freight. This cost covers applicable
sales taxes and shipment to the site for the injection and water
treatment systems. A figure of 8 percent of the total system
cost was used.2'7
6.1.1.2 Direct Installation Costs. This cost includes the
labor and material costs associated with installing the
foundation and supports, erecting and handling equipment,
electrical work, piping, insulation, and painting. For smaller
6-8
-------
turbines, the water treatment system is typically skid-mounted
and is shipped to the site as a packaged unit, which minimizes
field assembly and interconnections. The cost to install a skid-
mounted water treatment skid is typically $50,000, and this cost
is used for the direct installation cost for model plants less
than 5 MW (6700 hp).19 For larger turbines, it is expected that
the water treatment system must be field-assembled and the direct
installation costs were calculated as 45 percent of the injection
and water treatment systems, including taxes and freight.2
6.1.1.3 Indirect Installation Costs. This cost covers the
indirect costs (engineering, supervisory personnel, office
personnel, temporary offices, etc.) associated with installing
the equipment. The cost was taken to be 33 percent of the
systems' costs, taxes and freight, and direct costs, plus
$5,000 for model plants above 5 MW (6,700 hp).2 The indirect
installation costs for skid-mounted water treatment systems are
expected to be less than for field-assembled systems; therefore,
for.model plants with an output of less than 5 MW (6,700 hp) , the
cost percentage factor was reduced from 33 to 20 percent:
6.1.1.4 Contingency Cost. This cost is a catch-all meant
to cover unforeseen costs such as equipment redesign/
modification, cost escalations, and delays encountered in
startup. _ This cost was estimated as 20 percent of the sum of the
systems, taxes and freight, and direct and indirect costs.2
6.1.2 Annual Costs
The annual costs are summarized in Table 6-5 for each model
plant. Annual costs include the fuel penalty; electricity;
maintenance requirements; water treatment; overhead, general and
administrative, taxes, and insurance; and capital recovery, as
discussed in this section.
6.1.2.1 Fuel Penalty. The reduction in efficiency
associated with water injection varies for each turbine model.
Based on data in Reference 2, it was estimated that a WFR of
1.0 corresponds to a fuel penalty of 3.5 percent for water
injection and 1.0 percent for steam injection. This percentage
was multiplied by the actual WFR and the annual fuel cost to
6-9
-------
in
W
E-«
lb
as
®
5
Je
b
1
•3
•8
"Om OO
mr- — oo
•
(§8!
oor-oo*
— (SCO 00
O O
••- r^
a^f o ^ "^
O^—lOO
»" m — r»oo
or-oor-ci f
999999
zzzzzz
888888
JOOOO
^o >o m c* M
^* *O p^ ^^ Ov f
>OO OOO9
—• -^oo oooo
TT m
»'
TT f •» WS
rn op op—
««en « «>«>«>
9999999
zzzzzzz
8888888
185
*l
n<
^- OO fO
flOOv
H&.U
S
999
888
HO.UJ
OSS
999
8OO
fO ~"
m o\ 01
?5
O —00
-sa
HO.U
S
JMC
>s:
999
oo v
•»>.
v^vS
!E
ll
3 M
-§ 2
£3
a a
a 8
o-S
ll
!4!
it
•s o
8 »
«j
II
14
— .a
§!
en &
S4)
o.
IS
£S
II
«•«
S «
2=3
S.
^o
38 •-§§»?»£
<* a K. O.S.5 3
•=s 1^22-S a
II alS.S.2? 2
6-10
-------
determine the fuel penalty for each model plant. The fuel flow
was multiplied by the unit fuel costs to determine the annual
fuel costs. As shown in Table 6-3., the natural gas cost is
$3.88/1,000 standard cubic feet (scf) and the diesel fuel cost is
$0.77/gal.4'5
An increase in output from the turbine accompanies the
decrease in efficiency. This increase was not considered,
however, because not all sites have a demand for the available
excess power. In applications such as electric power generation,
where the excess power can be used at the site or added to
utility power sales, this additional output would serve to
decrease or offset the fuel penalty impact.
6.1.2.2 Electricity Cost. The electricity costs shown in
Table 6-5 apply to the feedwater pump(s) for water or steam
injection. The pump power requirements are estimated from the
pump head (ft) and the water flow rate as shown in the following.
equation:2
power pump (kWe) = ** x H x (S.G.) x * X °'7457 kW x * '
3,960 0.6 hp 0.9
where:
FR = feedwater flow rate, gal/min (from Table 6-2);
H = total pump head (ft);
S.G. = specific gravity of the feed water;
0.6 = pump efficiency of 60 percent;
0.9 = electric motor efficiency of 90 percent;
3,960 = factor to correct units in FR and H to hp; and
0.7457 = factor to convert hp to kW.
For water injection, the feedwater pump(s) supply treated water
to the gas turbine injection system. For steam injection, the
feedwater pump(s) supply treated water to the boiler for steam
generation. This cost analysis uses a feedwater temperature of
55°C (130°F) with a density of 61.6 lb/ft3 and a total pump head
requirement of 200 pounds per square inch, gauge (psig)
6-11
-------
(468 ft) .2 Based on these values, the pump electrical demand for
either water or steam injection is calculated as follows:
pump power (M,.) =
|il|
x 0.7457 x
0.161 x FR
The electrical cost for each model plant is the product of
the pump electrical demand, the annual hours of operation, and
the unit cost of electricity. The unit cost of electricity,
shown in Table 6-3, is $0.06/kWH.6'7
Maintenance costs were developed based on information from
manufacturers, and water treatment labor costs were estimated
based on information from a water treatment.vendor. Other costs
were developed based on the methodology presented in Reference 2.
No backup steam or electricity costs were developed for
water or steam injection because it was assumed that no
additional downtime would be required for scheduled inspections
and repairs. Maintenance intervals could be scheduled to
coincide with the 760 hr/yr of downtime that are currently
allocated for scheduled maintenance. If this were done, the
annual utilization of the backup source would not increase.
6.1.2.3 Added Maintenance Costs. Based on discussions with
gas turbine manufacturers, additional maintenance is required for
some gas turbines with water injection. The analysis procedures
used to develop the incremental maintenance costs are presented
in Appendix B.
The incremental maintenance cost associated with water
injection for natural gas-fueled turbines was provided by the gas
turbine manufacturers.10,20-24 ^^ gag turbine manufacturers
contacted stated that there were no incremental maintenance costs
for operation with steam injection. Two manufacturers provided
maintenance costs for natural gas and oil fuel operation without
water injection.1^/20 using an average of these costs,
incremental maintenance costs for water injection are 30 percent
higher for plants that use diesel fuel instead of natural gas.
6-12
-------
Costs were prorated for model plants that operate less than
8,000 hr/yr.
6.1.2.4 Water Treatment Costs. Water treatment operating
costs include the cost of treatment (e.g., for chemicals and
media filters), operating labor, raw water, and wastewater
disposal. The raw water flow rate is equal to the treated water
flow rate (the water or steam injection rate) plus the flow rate
of the wastewater generated in the treatment plant. As noted in
Section 6.1, the wastewater flow rate is equal to 29 percent of
the injection flow rate. The annual raw water, treated water,
and wastewater flow rates were multiplied by the appropriate unit
costs in Table 6-3 to determine the annual costs. Water
treatment labor costs were calculated at $0.70/1,000 gal for
water injection. 5 This cost was multiplied by the total annual
treated water flow rate to determine the annual water treatment
labor cost for water injection. Labor costs for steam injection.
were assumed to be half as much as the costs for water injection
because it was assumed that the facility already has a water -
treatment plant for the boiler feedwater. Therefore, the
operator requirements would be only those associated with the
increase in capacity of the existing treatment plant.
6.1.2.5 Plant Overhead. This cost is the overhead
associated with the additional maintenance effort required for
water injection. The cost was calculated as 30 percent of the
added maintenance cost from Section 6.1.2.3.2
6.1.2.6 General and Administrative. Taxes, and Insurance
Costs (GATI). This cost covers those expenses for administrative
overhead, property taxes, and insurance and was calculated as
4 percent of the total capital cost.2
6.1.2.7 Capital Recovery. A capital recovery factor (CRF)
was multiplied by the total capital investment to estimate
uniform end-of-year payments necessary to repay the investment.
The CRF used in this analysis is 0.1315, which is based on an
equipment life of 15 years and an interest rate of 10 percent.
6.1.2.8 Total Annual Cost. This cost is the sum of the
annual costs presented in Sections 6.1.2.1 through 6.1.2.7 and is
6-13
-------
the total cost that must be paid each year to install and operate
water or steam injection NOX emissions control for a gas turbine.
6.1.3 Emission Reduction and Cost-Effectiveness Summary for
Water and Steam Injection
The uncontrolled and controlled NO... emissions and the annual
A.
emission reductions for the model plants are shown in Table 6-6.
The emissions, in tons per year (tons/yr), were calculated as
shown in Appendix A.
The total annual cost was divided by the annual emission
reductions to determine the cost effectiveness for each model
plant. For continuous-duty natural gas-fired model plants, the
cost-effectiveness figures range from approximately $600 to
$2,100 per ton of NOX removed for water injection, and decrease
to approximately $400 to $1,850 per ton for steam injection. The
lower range of cost-effectiveness figures for steam injection is
primarily due to the greater NOX reduction achieved with steam
injection. For continuous-duty oil-fired model plants, the cost
effectiveness ranges from approximately $675 to $1,750 per ton. of
NOX removed, which is comparable to figures for gas-fired model
plants. The cost-effectiveness figures are higher for gas
turbines with lower power outputs because the fixed capital costs
associated with wet injection system installation have the
greatest impact on the smaller gas turbines.
Cost-effectiveness figures increase as annual operating
hours decrease. For turbines operating 2,000 hr/yr, the cost-
effectiveness figures are two to nearly three times higher than
those for continuous-duty model plants, and increase further for
model plants operating 1,000 hr/yr. For the oil-in-water
emulsion model plant, the cost effectiveness corresponding to
1,000 annual operating hours is $l,840/ton of NOX removed. No
data were available to prepare a conventional water injection
model plant for this turbine to compare the relative cost-
effectiveness values.
6-14
-------
o
O. NO_
wf -*"
999999
O O O O O O
O O O O O W
oo — r~ ?s o o
vo «*
3 ss
2?3S
sf 3f^
999
o g o
o> m
U1 —
ci r-
2
"
S
r~ — IN
— r< ts
H 0. W
1 NO •*
m ff oo
^ >?
c^^9
OO
I
6-15
-------
6.2 LOW-NOX COMBUSTORS
Incremental capital costs for low-NOx combustors relative to
standard designs for new applications were provided by three
manufacturers for several turbines.3'14'26 Based on information
from the manufacturers, the performance and maintenance
requirements for a low-NOx combustor are expected to be the same
as for a standard combustor, and so the only annual cost
associated with low-NOx combustors is the capital recovery. The
capital recovery factor is 0.1315, assuming a life of 15 years
and an interest rate of 10 percent.
Table 6-7 presents the uncontrolled and controlled emission
levels, the annual emission reductions, incremental costs for a
low-NOY combustor over a conventional design, and the cost
Jt
effectiveness of low-NOx combustors- for all gas turbine models
for which sufficient data were available. Cost-effectiveness
figures were calculated for 8,000 and 2,000 hours of operation
annually, using controlled NOX emission levels of 42, 25, and
9 parts per million, by volume (ppmv), referenced to 15 percent
oxygen, which are the achievable levels stated by the turbine
manufacturers. The cost effectiveness varies according to the
uncontrolled NOX emission level for the conventional combustor
design and the achievable controlled emission level for the
low-NOx design. For continuous-duty applications, cost
effectiveness for a controlled NOX emission level of 42 ppmv
ranges from $353 to $1,060 per ton of NOX removed. The cost-
effectiveness range decreases to $57 to $832 per ton of NOX
removed for a controlled NOX emission level of 25 ppmv and
decreases further to $55 to $137 per ton of NOX removed for a
9 ppmv control level. In all cases, the cost effectiveness
increases as the operating hours decrease. In general, the cost
effectiveness is higher for smaller gas turbines than for larger
turbines due to the relatively higher capital cost per kW for
low-NO^ combustors for smaller turbines.
Ji
The cost-effectiveness range is lower for low-NO.. combustors
J\,
than for water or steam injection because the total annual costs
are lower and, in some cases, the controlled emission levels are
6-16
-------
L
i
a
o
o
C--
W
J
I
ssasa
0» ov r* rliS
10 •»\o^"N
ooooo
^pt^oa^
9 M* «e f"; 3t
o «o ••* M g*
tno —(~3>
e ee OD oe no
tn o m oo O
• HH
23322
33333
illii
SSSSl
aaaa
W>
a, p,
oOOv
-------
also lower. According to two turbine manufacturers, retrofit
costs are 40 to 60 percent greater than the incremental costs
shown in Table 6-7 for new installations.^'14
6.3 SELECTIVE CATALYTIC REDUCTION
The costs for SCR for new installations were estimated for
all model plants. Retrofit costs for SCR were not available but
could be considerably higher than the costs shown for new
installations, especially in applications where an existing heat
recovery steam generator (HRSG) would have to be moved, modified,
or replaced to accommodate the addition of a catalyst reactor.
To date, most gas turbine SCR applications use a base metal
catalyst with an operating temperature range that requires
cooling of the exhaust gas from the turbine. For this reason,
SCR applications to date have been limited to combined cycle or
cogeneration applications that include an HRSG, which serves to
cool the exhaust gas to temperatures compatible with the
catalyst. The introduction of high-temperature zeolite
catalysts, however, makes it possible to install the catalyst
directly downstream of the turbine, and therefore feasible to
use SCR with simple-cycle applications as well as heat recovery
applications. As discussed in Section 5.3.2, to date there is at
least one gas turbine installation with a high-temperature
zeolite catalyst installed downstream of the turbine and upstream
of an HRSG. At present, "no identified SCR systems are installed
in simple-cycle gas turbine applications.
An overview of the procedures used to estimate capital and
annual costs are described in Sections 6.3.1 and 6.3.2,
respectively; a detailed cost algorithm is presented in
Appendix B. The emission reduction and cost-effectiveness
calculations are described in Section 6.3.3.
6.3.1 Capital Costs
Five documents in the technical literature contained SCR
capital costs for 21 gas turbine facilities. Most of these
documents presented costs that were obtained from vendors, but
some may have also developed at least some costs based on their
own experiences.27"31 Most of the documents presented only the
6-18
-------
total capital costs, not costs for individual components, and
they did not provide complete descriptions of what the costs
included. These costs were plotted on a graph of total capital
costs versus gas turbine size. To this graph were added
estimates of total installed costs for a high-temperature
catalyst SCR system for installation upstream of the HRSG for
four turbine installations ranging in size from 4.5 to 83 MW
(6,030 to 111,000 hp). These high-temperature SCR system
estimates include the catalyst reactor, air injection system for
exhaust temperature control, ammonia storage and injection
system, instrumentation, and continuous emission monitoring
equipment. These SCR costs were estimated by the California Air
Resources Board (CARS) in 1991 dollars and are based on NOX
emission levels of 42 ppmv into and 9 ppmv out of the SCR. 5
These estimated costs, shown in Appendix B, fit well within the
range of costs from the 21 installations discussed above, and the
equation of a line determined by linear regression adequately
fits the data (R2 = 0.76) for all 25 points. Based on this
graph, the total capital cost for either a base-metal SCR system
installed within the HRSG or a high-temperature zeolite catalyst
SCR system installed directly downstream of the turbine can be
calculated using the equation determined by the linear
regression. This equation is shown in Table 6-8 and was used to
calculate the total capital investment for SCR for each model
plant shown in Tables 6-9 and 6-10.
6.3.2 Annual Costs
Total annual costs for SCR control were developed following
standard EPA procedures described in the OAQPS-Control Cost
Manual for other types of add-on air pollution control devices
(APCD's). Information about annual costs was obtained from the
same sources that provided capital costs.27"31 Total annual
costs consist of direct and indirect costs; parameters that make
up these categories and the equations for estimating the costs
are presented in Table 6-8 and are discussed below. The annual
6-19
-------
TABLE 6-8. PROCEDURES FOR ESTIMATING CAPITAL AND
ANNUAL COSTS FOR SCR CONTROL OF NO,, EMISSIONS FROM GAS TURBINES3
A. Total capital investment, $b
B. Direct annual costs, $/yr
1. Operating labor0
2. Supervisory labor
3. Maintenance labor and materials
4. Catalyst replacement
5. Catalyst disposal
6. Anhydrous ammonia®
7. Dilution steam^
8. Electricity^
9. Performance loss"
10. Blower (if needed)
11. Production loss1
C. Indirect annual costs, $/yr
1. Overhead
2. Property taxes, insurance, and
administration
3. Capital recovery^
= (49,700 x TMW) + 459,000
(1.0 hr/8 hr-shift) x ($25.60/hr) x (H)
(0. IS) x (operating labor)
(1,250 x TMW) + 25,800
(4,700 x TMW) + 37,200
(V) x (SIS/ft3) x (.2638)
(N) x ($360/ton)
(N) x (0.95/0.05) x (MW H2O/MW NH3) x ($6/1,000
Ib steam) x (2,000 Ib/ton)
N/A
(0.005) x (TMW) x ($0.06/KWH) x (1,000 KW/MW)
x(H)
0.1 x (Performance Loss)
None
(0.6) x (all labor and maintenance material costs)
(0.04) x (total capital investment)
(0.13147) x [total capital investment - (catalyst
replacement/0.2638)]
aAll costs are in average 1990 dollars.
= turbine output in MW for each model plant.
annual operating hours are represented by the variable H. The labor rate of $25.60/hr is from Table 6-3.
dThe catalyst volume in fr is represented by the variable V. The catalyst volume for each model plant is
estimated as V = (TMW) x (6,180 fP/SS MW).
eThe ammonia requirement in tons is represented by the variable N and is calculated using a NH-j-to-NOx
molar ratio of 1.0.
The annual tonnage of NOX is taken from the controlled levels shown hi Tables 6-11 and 6-12.
MW of NH, = 17.0
N = annual tonnage of NO, x (_____)
f "
The ammonia is diluted with steam to 5 percent by volume before injection.
SThe amount of electricity required for ammonia pumps and exhaust fans is not known, but is expected to be
small. The electricity cost comprised less than 1 percent of the total annual cost estimated by the South Coast
Air Quality Management District (SCAQMD) for SCR applied to a 1.1 MW turbine.
"Based on information from three sources, the backpressure from the SCR reduces turbine output by an average
of about 0.9 percent.
'No production losses are estimated because it is assumed that all SCR maintenance, inspections, cleaning, etc.
. can be performed during the 760 hours of scheduled downtime per year.
Jibe capital recovery factor for the SCR is 0.13147, based on a 15-year equipment life and 10 percent interest
rate. The catalyst is replaced every 5 years. The 0.2638 figure is the capital recovery factor for a 5-year
equipment life and a 10 percent interest rate.
6-20
-------
8
CO
o
a
o
—
3.1|,
^r!
cs >n oo >n o
•* v/i vo ^ wi
cs cs >n « vo
§888
*
oo
oo oo
•* « cs m
^ ^ ^ ^
§2 fi S fi
o o o o
oC o jj" oC o"
OO OO OO OO OO
" " "
88
z
a. 5 w
m- °. CN vd « •»
co f cs cs oo so
>***^>
999999
llllil
wi >o >n r; o O
vn in •«• * <5 cs
cs n vo m m CM
— — cs
00_ 00_
' m m
cs cs "
t~; oo * 0 rr
m cs oo
=f * ^
999
vn ci cs
*~ cs
\ o
'
-28
8.
88
-• (S
28
8
00
'S
a. u
— o
99.
Q P
ju
3
I
•s
8
so
1)
2
a
|
1
!
I
*s
s
1
•o
u
(O
a
I
6-21
-------
o
H
EH
s
8
i*
Dn
O
a
CO
o
Q
Q
W
8
CO
o
ft.
CO
o
u
H
O4
6
o
H
I
•3 fr
.3 «
JS'
s g.s
2 § S
(2 I-a
SI-
I -
s-t
.3
j j
3
Sv* « m in
c* r< fi m
S? S 2 S S
S 5?
vT oo" o" oo* m
a' a" R s R"
*" * 2 s" a"
o o o o
r- o >o «n
•*f oo oo »o
* "1 "I "I
-* -« r*
8
8
C4 ^ OO oo" Tf
•o « in r- o»
II
_
SO i— ^ OO
en rt M m
o
lll
"
a a a" a
>o «o « oo
PP
(M O O O -O O O
w> 3 <»i v» w »n c-
M n m « ce v» m
8 SS 88
w")ooQoomv).*r»
v*)v)or^ooo^^
a a" R a s * s
8888888
•« >n r- * * r- *
§ S§ 28SS
r-Ao»«^mo
- . - 6 rf, A wi,
33333333
S8SSSI8S
s ss as
_ (N ,N — CH
§11
' oC 01 o"
§0000
t- «n i-* oo
SIS
8
w» r* * o* en
vi o o -^
Illll
33333
§
S S S S 8 I 8
«*» m eo rs o o oo
8 R S R* S 3 8
8
O Q O
M «^ vT
"
§2
S8SS2SSS
ssss
'
R 2 « S :
o o oe ON
u
.s
a
8
tn
i
a
•3
t
i
I
I
•s
l
6-22
-------
costs are shown in Tables 6-9 and 6-10 for injection and dry low-
NOV combustion, respectively, for each of the model plants.
jfc
6.3.2.1 Operating and Supervisory Labor. Information about
operating labor requirements was unavailable. Most facilities
have fully automated controls and monitoring/recording equipment,
which minimizes operator attention. Therefore, it was assumed
that 1 hr of operator attention would be required during an 8-hr
shift, regardless of turbine size. This operating labor
requirement is at the low end of the range recommended in the
OAQPS Control Cost Manual for other types of APCD's.7 Operator
wage rates were estimated to be $25.60/hr in 1990, based on
escalating the costs presented in Reference 2 by 5 percent per
year to account for inflation. Supervisory labor costs were
estimated to be 15 percent of the operating labor costs,
consistent with the OAQPS Control Cost Manual.
6.3.2.2 Maintenance Labor and Materials. Combined
maintenance labor and materials costs for 14 facilities were
obtained from four articles, but almost half of the data
(6 facilities) were provided by one source.27"30 The costs were
escalated to 1990 dollars assuming an inflation rate of 5 percent
per year. All of the data are for facilities that burn natural
gas. Provided that ammonium salt formation is avoided by
limiting ammonia slip and sulfur content, the cost for operation
with natural gas should also apply for distillate oil fuel.32
Therefore, it was assumed that the cost data also apply to SCR
control for turbines that fire distillate oil fuel. The costs
were plotted versus the turbine size, and least-squares linear
regression was used to determine the equation of the line through
the data (see Appendix B). This equation, shown in Table 6-8,
was used to estimate the maintenance labor and materials costs
shown in Table 6-9 for the model plants.
6.3.2.3 Catalyst Replacement. Replacement costs were
obtained for nine gas turbine facilities, and combined
replacement and disposal costs were obtained for another six gas
turbine facilities.27"30 The disposal costs were estimated for
the six facilities as described below and in Appendix B. The
6-23
-------
replacement costs for these six facilities were then estimated by
subtracting the estimated disposal costs from the combined costs.
A catalyst life of 5 years was used. All replacement costs were
escalated to 1990 dollars assuming a 5 percent annual inflation
rate.
The estimated 1990 replacement costs were plotted versus the
turbine size, and least-squares linear regression was used to
determine the equation of the line through the data (see
Appendix B). This equation is shown in Table 6-8 and was used to
estimate the catalyst replacement costs shown in Table 6-9 for
the model plants.
6.3.2.4 Catalyst Disposal. Catalyst disposal costs were
estimated based on a unit disposal cost of $15/ft3, which was
obtained from a zeolite catalyst vendor.32 This cost was used
for each model plant, but the disposal cost may in fact be higher
for catalysts that contain heavy metals and are classified as
hazardous wastes. The catalyst volume for each model plant was
estimated based on information about the catalyst volume for one
facility and the assumption that there is a direct relationship
between the catalyst volume and the turbine output (i.e., the
design space velocity is the same regardless of the SCR size).
At one facility, 175 m3 (6,180 ft3) of catalyst is used in the
SCR with an 83 MW (111,000 hp) turbine.33 The disposal cost for
this catalyst would be $92,700, using a cost of $15/ft3.
6.3.2.5 Ammonia. The annual ammonia (NHj) requirement is
calculated from the annual NOY reduction achieved by the SCR
J^
system. Based on an NH^/NOjj. molar ratio of 1.0, the annual
ammonia requirement, in tons, would equal the annual NOX
reduction, in tons, multiplied by the ratio of the molecular
weights for NH3 and NOX. Anhydrous ammonia with a unit cost of
$360/ton was used.34'35 The equation to calculate the annual
cost for ammonia is shown in Table 6 - 8.
6.3.2.6 Dilution Steam. As indicated in Section 5.3.1,
steam is used to dilute the ammonia to about 5 percent by volume
before injection into the HRSG. According to the OAQPS Control
6-24
-------
Cost Manual, the cost to produce steam, or to purchase it, is
about $6/1,000 Ib.
6.3.2.7 Electricity. Electricity requirements to operate
such equipment as ammonia pumps and ventilation fans is believed
to be small. For one facility, the cost of electricity to
operate these components was estimated to make up less than
1 percent of the total annual cost, but it is not clear that the
number and size of the fans and pumps represent a typical
installation.27 , This cost for electricity is expected to be
minor, however, 'for all installations and was not included in
this analysis.
For high-temperature catalysts installed upstream of the
HRSG, a blower may be required to inject ambient air into the
exhaust to regulate the temperature and avoid temperature
excursions above the catalyst design temperature range. The cost
to operate the blower is calculated to be 10 percent of the fuel
penalty.35
. 6.3.2.8 Performance Loss. The performance loss due to
backpressure from the SCR is approximately 0.5 percent of the
turbine's design output.34-36 To n^^g Up for this lost output,
it was assumed that electricity would have to be purchased at a
cost of $0.06/kWH, as indicated in Table 6-3.
6.3.2.9 Production Loss. No costs for production losses
were included in this analysis. It was assumed that scheduled
inspections, cleaning, and other maintenance will coincide with
the 760 hr/yr of expected or scheduled downtime. It should be
recognized that adding the SCR system increases the overall
system complexity and the probability of unscheduled outages.
This factor should be taken into account when considering the
addition of an SCR system.
6.3.2.10 Overhead. Standard EPA procedures for estimating
annual control costs include overhead costs that are equal to
60 percent of all labor and maintenance material costs.
6.3.2.11 Property Taxes. Insurance, and Administration.
According to standard EPA procedures for estimating annual
control costs, property taxes, insurance, and administration
6-25
-------
costs are equal to 4 percent of the total capital investment for
the control system.
6.3.2.12 Capital Recovery. The CRF for SCR was estimated
to be 0.13147 based on the.assumption that the equipment life is
IS years and the interest rate is 10 percent.
6.3.3 Cost Effectiveness for SCR
As indicated in Section 5.4, virtually all gas turbine
installations using SCR to reduce NOX emissions also incorporate
wet injection or low-NOx combustors. The NOX emission levels
into the SCR, therefore, were in all cases taken to be equal to
the controlled NOX emission levels shown for these control
techniques in Tables 6-6 and 6-7. The most common controlled NOX
emission limit for gas-fired SCR applications is 9 ppmv,
referenced to 15 percent oxygen. The capital costs used in this
analysis are expected to correspond to SCR systems sized to
reduce controlled NO... emissions ranging from 25 to 42 ppmv from
Jt •
gas-fired turbines to a controlled level of approximately 9 ppmv
downstream of the SCR. Based on the controlled NOX emission
limits established by the Northeast States for Coordinated Air
Use Management (NESCAUM), shown in Table 5-3, these SCR systems
would reduce NOX emissions to 18 ppmv for oil-fired applications.
Cost-effectiveness figures for SCR in this analysis are therefore
calculated based on controlled NOX emission levels of 9 and
18 ppmv, corrected to 15 percent oxygen, for gas- arid oil-fired
SCR model plants, respectively.
Cost effectiveness for SCR used downstream of wet injection
or dry low-NOx combustion is shown in Tables 6-11 and 6-12,
respectively. For continuous-duty, natural gas-fired model
plants using water or steam injection,the cost effectiveness for
SCR ranges from approximately $3,500 to $10,800 per ton of NOX
removed.
The cost-effectiveness range for SCR installed downstream of
continuous-duty, natural gas-fired turbines from 3 to 10 MW
(4,000 to 13,400 hp) using dry low-NOx combustion is $6,290 to
$10,800 per ton of NOX removed for an inlet NOX emission level of
42 ppmv. The cost-effectiveness range for SCR increases for an
6-26
-------
B
H
3=
CO
H
CQ
fl
\O 5f OO
O O1* VI CO
Q S?
' CT\
00 n oo in o
s a s s s
§1
a
X
ns/yr
OO ^ OO
.« ; 'O
^3 fi »*
A
<*> 0 2
r- 55
Prf H
o o P-
O oo «•> r-
6 -X 2 d
m ^t
O 00 fv| «1 t- —
o' .n v» 2 2 ?}
m •» •» "
2 S!
2 a
s
o p p q p p
O O O O O O O
pop
oo' oo oo
pop
O\ Ov Ov
pop
oo' 00 00
O p
oo oo
_; —
33S
° >
§i
m O (^ oo m i^
O t- oo ^ « r-
H
H
I
V£»
w
t- 0. Ul
H 0, W
g
u S
Sfe
Je
'a,
^J?SS
oo * m r~. _
49999?
888881
«?
9
99099
z z z z z z z
8
888888
m- vb
-------
_ -3
CO
S
\
VD
H
I
f
M
§o o o o
00 O m ON
. I"", ft «\ f*,
O ON ON *O *O
o o ^ oo A
to o — r- 3*
m o >n oo p
m •» *r oo o
500
T
'H
Cen
Tau
n T
n T
22233
z z z z z
88888
I
oooooooo
-p^oo-gSS
tnoxioooomr-
^^ ft OO OO
1 .
ff
= " H H§ g
1 1 5 sis-
33133333
liiliiii
S^ C"* (^
o o o\
o\ o
— tS
O O O O O
~ m. P >O 09
O (~ C~ f V
O u*l TT 00 O>
<»i o — r- 5\
m p w> oo p
v% ^ ^r oo o
H
m "* — m m
OOOOOOOO
o r~ r~- •*•
^ - g 5
OOim
^^ «*4 OO
O O ^? "v
8S
H? sfop
3 S "HHgO
I
1
I
S 1-nO
«"—<-"
ixli
z 5 js
|J«
w i2 *
J5 a 13
6-28
-------
inlet NOX emission level of 25 ppmv due to the lower NOX
reduction efficiency. For an inlet NOX level of 25 ppmv, the
cost effectiveness ranges from $12,800 to $22,100 per ton of NOX
removed for 3 to 10 MW (4,000 to 13,400 hp) turbines and
decreases to $6,940 to $7,660 per ton of NOX removed for larger
turbines ranging from 39 to 85 MW (52,300 to 114,000 hp). As
these ranges indicate, the cost effectiveness for SCR is affected
by the inlet NOX emission level and not the type of combustion
control technique used for the turbine. The cost effectiveness
for continuous-duty, oil-fired model plants ranges from
approximately $2,450 to $8,350 per ton of NOX removed. The SCR
cost-effectiveness range for oil-fired applications is lower than
+
that for gas-fired installations in this cost analysis because
the same capital costs were used for both fuels (capital costs
were not available for applications using only distillate oil
fuel). The percent NOX reduction for oil-fired applications is
higher, so the resulting cost-effectiveness figures for oil-fired
applications are lower. It should be noted that this higher NOX
reduction for oil-fired applications may require a larger
catalyst reactor, at a higher capital cost. As a result, the
cost-effectiveness figures may actually be higher than those
shown in Table 6-11 for oil-fired applications.
The cost-effectiveness figures are higher for smaller gas
turbines because the fixed capital costs associated with the
installation of an SCR system have the greatest impact on smaller
gas turbines. Cost-effectiveness figures increase as annual
operating hours decrease. For turbines operating 2,000 hours per
year, cost-effectiveness figures are more than double those for
continuous-duty model plants, and they increase even further for
model plants operating 1,000 hr/yr.
Because virtually all SCR systems are installed downstream
of controlled gas turbines, combined cost-effectiveness figures
for wet injection plus SCR and also dry low-NOx combustion plus
SCR have been calculated and are shown in Tables 6-13 and 6-14,
respectively. These combined cost-effectiveness figures are
calculated by dividing the sum of the total annual costs by the
6-29
-------
Sg'
i~
* *-"
S28838
•*f \n oo^ oo oo >/>
•»"«"•-"•- —
i
O O
P* m
° 2
oo - P
-• o t-
..; >n »n
« Mi-
imofSC
3
s
i
2
pj o oo
"! 6 >n
r~ m f •
IS
a.
o.
o o o o o o
o o o o o o o
o o o
o o o
c?v ON
p p q
od oo oo
p p
00 CO
"1
o*
at
u
O 00
S
S
2
— vo -» m
s c ~
\o —
o 2 * °
"*•"-
o r- TT
O* •-" oo
F- — «s
Ol —
m O "^ °°- **!*":
p- oo ^ « r— — •
_ p
m
vo
W
J
I
H a. W
0. U4
a. u
g
I
i
S S
??^2S
999999
Illsls
® o •»
m cs oo
SE S*SS
<>^4
_ p
S3
? Ul
6-30
-------
cr\
H
I
10
w
OX-SH.
-It
||!
f
plant
-H 00 O\ OO
o\ o — ts
r) m « f
a s 8
- - * °-
NO r-' r» 2
o o e> o o
vi
NO
00 00 00 00 00
m q vn ooi q
m ^ ^° oo o
33333
§§^§.?
O O XJ
U
c~ 2 ^J
OOOOOOOO
2 2 fj
-: «. <•: 2° = « a
5DS
oooooaoooaaooooo
oovm'*
— tn o» oo
illlll
M cs n cs
vn op ON vn ON
•••••- pi m
O O O O O
-« w> — t P
(^> oo ON ,2 ^^
o P- r- v f
O VI •* 00
0V
O*
m o
tn •v
oo q
oo o
H
2 B
«n O vn oo 1
O ON *
< I
x £
< °
•S g
e n
7 ca
O 3
•S Q.
i l
I if
irj
i^l
|||
Is*
ox||
•S § o
§• 8 *
1-8 1
-«^^
6-31
-------
sum of the annual reduction of NOX emissions"for the combined
emission control techniques. For continuous-duty, natural gas-
fired model plants, the combined cost-effectiveness figures for
wet injection plus SCR range from approximately $650 to $4,500
per ton of NOX removed. For continuous-duty, oil-fired model
plants, the combined cost effectiveness ranges from approximately
$1,100 to $3,550 per ton of NOX removed. The combined cost-
effectiveness figures for dry low-NOx combustion plus SCR for
continuous-duty, natural gas-fired model plants range from
approximately $350 to $3,550 per ton of NOX removed.
The combined cost-effectiveness figures increase with
decreasing turbine size and annual operating hours. Data were
not available to quantify the wet injection requirements and
controlled emissions levels for oil-fired turbines with IOW-NCL.
Jt
combustors, so cost-effectiveness figures were not tabulated for
this control scenario.
6.4 OFFSHORE TURBINES
The only available information about the cost of NOY
Jv _
controls for offshore gas turbines was presented in a report
prepared for the Santa Barbara County Air Pollution Control
District (SBCAPCD) in California.^7 The performance and cost of
about 20 NOX control techniques for a 2.8 MW (3,750 hp) turbine
were described in the report. Wet injection and SCR were
included in the analysis; low-'NOx combustors were not. The costs
from the report are presented in Table 6-15 without adjustment
because there is insufficient cost information to know what
adjustments need to be made. Additionally, insufficient
information is available to scale up these costs for larger
turbines. The water and steam injection costs and SCR costs for
offshore applications are discussed in Sections 6.4.1 and 6.4.2,
respectively.
6.4.1 Wet Injection
The report prepared for SBCAPCD assumed water injection
costs are the same as steam injection costs. The report did not
describe the components in the capital cost analysis for these
injection systems, but the results are much lower than those that
6-32
-------
TABLE 6-15. PROJECTED WET INJECTION AND SCR COSTS
FOR AN OFFSHORE GAS TURBINEa
Capital cost, $
Annual costs, $/yr
Ammonia
Catalyst replacement
Operating and maintenance^
Fuel penalty6
Capital recoveryf
Total annual costs, $/yr
Wet injection
costs
70,000
N/Ab
N/A
24,600
10,500
14,000
49,100
SCR costs
585,000
3,050C
28,000
18,000
5,000
117,000
171,000
aCosts are for a 2.8 MW gas turbine and are obtained from
Reference 37.
bN/A = Not applicable.
^Ammonia cost is based-on $150/ton and 0.4 Ib NH3/lb NOX.
dOperating and maintenance cost for SCR is estimated as 3 percent
of the total capital investment.
eFuel penalty is estimated as 2 percent of the annual fuel
consumption for wet injection and 1 percent for SCR.
fCapital recovery is estimated based on an equipment life of
8 years and an interest rate of 13 percent.
6-33
-------
would be estimated by the procedures described in Section 6.1.1
of this report. The authors may have assumed that the engine-
mounted injection equipment cost was included in the turbine
capital cost and that a less rigorous water treatment process is
installed. Annual costs are also much lower than those that
would be estimated by the procedures described in Section 6.1.2
of this report. There are at least three reasons for the
difference: (1) the low capital cost leads to a low CRF, even
though the turbine life was assumed to be only 8 years;
(2) overhead costs and taxes, insurance, and administration costs
are not considered; and (3) the capacity factor is only
50 percent (i.e., about 4,400 hr/yr, vs. 8,000 hr/yr, as in
Section 6.1.2). The turbine life was only 8 years, which may
correspond to a typical service life of an offshore platform.
6.4.2 Selective Catalytic Reduction
The total capital costs presented in the report for SBCAPCD
are similar to those that would be estimated by the procedures in
Section 6.2.1 of this report. However, it appears that $150,000
of the total in Reference 37 is for structural modifications to
the platform and $75,000 is for retrofit installation. When the
difference in the load factor is taken into account, some of the
annual costs are similar to those that would be estimated by the
procedures in Section 6.2.2 for a similarly sized turbine. The
catalyst replacement cost, however, is much lower; neither the
type of catalyst nor the replacement frequency were identified.
Ammonia costs are lower because the uncontrolled NOX emission
level was assumed to be 110 ppmv instead of 150 ppmv and because
a unit cost of $150/ton was used instead of $400/ton. The
reference does not indicate whether or not catalyst disposal,
overhead, taxes, freight, and administration costs were
considered. Capital recovery costs are higher because the
equipment life is assumed to be only 8 years on the offshore
platform.
6-34
-------
REFERENCES FOR CHAPTER 6
1. 1990 Performance Specifications. Gas Turbine World.
11:20-48. 1990.
2. U. S. Environmental Protection Agency. Background
Information Document, Review of 1979 Gas Turbine New Source
Performance Standards. Research Triangle Park, NC.
Prepared by Radian Corporation under Contract
No. 68-02-3816. 1985.
3. Letter and attachments from Swingle, R., Solar Turbines
Incorporated, to Neuffer, W. J., EPA/ISB. August 20, 1991.
Review of draft gas turbine ACT document.
4. Monthly Energy Review. Energy Information Administration.
March 1991. p. 113.
5. Petroleum Marketing Annual 1990. Energy Information
Administration.
6. Reference 3, p. 109.
7. OAQPS Control Cost Manual (Fourth Edition).
EPA-450/3-90-006. January 1990.
8. Letter and attachment from Leonard G., General Electric
Company, to Snyder, R., MRI. May 24, 1991. Response to gas
turbine questionnaire.
9. Letter and attachment from Swingle, R., Solar Turbines
Incorporated, to Snyder, R., MRI. February 8, 1991.
Maintenance considerations for gas turbines.
10. Telecon. Snyder, R., MRI, with Rayome, D., US Turbine
Corporation. May 6, 1991. Maintenance costs for gas
turbines.
11. Telecon. Snyder, R., MRI, with Schorr, M., General Electric
Company. May 22, 1991. Gas turbine water injection.
12. Letter and attachments from Gurmani, A., Asea Brown Boveri,
to Snyder, R., MRI. May 30, 1991. Response to gas turbine
questionnaire.
13. Letter and attachment from Gagnon, S., High Purity Services,
Inc., to Snyder, R., MRI. April 4, 1991. Water treatment
system design.
14. Letter and attachments from Gurmani, A., Asea Brown Boveri,
to Snyder, R., MRI. February 4, 1991. Response to gas
turbine questionnaire.
6-35
-------
15. Letter and attachment from Kimsey, D., Allison Gas Turbine
Division of General Motors, to Snyder, R., MRI.
February 19, 1991. Response to gas turbine request.
16. Letter and attachment from Leonard, G., General Electric
Company, to Snyder, R. MRI. February 14, 1991. Response to
gas turbine questionnaire.
17. Letter and attachments from Cull, C. General Electric
Company, to Snyder, R., MRI. May 14, 1991. On-engine costs
for water and steam injection hardware.
18. Bernstein, S., and P. Malte (Energy International, Inc.).
Emissions Control for Gas Transmission Engines. Prepared
for the Gas Research Institute. Chicago. Presentation
No. PRES 8070. July 1989. 17 pp.
19. Letter and attachments from Ali, S. A., Allison Gas Turbine
Division of General Motors, to Neuffer, W. J., EPA/ISB.
August 30, 1991. Review of draft gas turbine ACT document.
20. Telecon. Snyder, R., MRI, with Schubert, R., General
Electric Marine and Industrial Division. April 26, 1991.
Maintenance costs for gas turbines.
21. Letter and attachments from Swingle, R., Solar Turbines
Incorporated, to Snyder, R., MRI. May 21, 1991.
Maintenance considerations for gas turbines.
22. Walsh, E. Gas Turbine Operating and Maintenance
Considerations. General Electric Company. Schenectady, NY.
Presented at the 33rd GE Turbine State-of-the-Art Technology
Seminar for Industrial, Cogeneration and Independent Power
Turbine Users. September 1989. 20 pp.
23. Telecon. Snyder, R., MRI, with Pasquarelli, L., General
Electric Company. April 26, 1991. Maintenance costs for
gas turbines.
24. Letters and attachments from Schorr, M., General Electric
Company, to Snyder, R., MRI. March, April 1991. Response
to gas turbine questionnaire.
25. Kolp, D. (Energy Services, Inc.), S. Gagnon (High Purity
Services), and M. Rosenbluth (The Proctor and Gamble Co.).
Water Treatment and Moisture Separation in Steam Injected
Gas Turbines. Prepared for the American Society of
Mechanical Engineers. New York. Publication No. 90-GT-372,
June, 1990.
26. Letter from Cull, C., General Electric Company, to Snyder,
R., MRI. May 29, 1991. Low-NOx Combustor Costs.
6-36
-------
27. Permit Application Processing and Calculations by South
Coast Air Quality Management District for proposed SCR
control of gas turbine at Saint John's Hospital and Health
Center, Santa Monica, CA. May 23, 1989.
28. Prosl, T. (DuPont), and G. Scrivner (Dow). Technical
Arguments and Economic Impact of SCR's Use for NOX Reduction
of Combustion Turbine for Cogeneration. Paper presented at
EPA Region VI meeting concerning NOX abatement of combustion
turbines. December 17, 1987.
29. Sidebotham, G., and R. Williams. Technology of NOX Control
for Stationary Gas Turbines. Center for Environmental
Studies. Princeton University. January 1989.
30. Shareef, G., and D. Stone. Evaluation of SCR NOX Controls
for Small Natural Gas-Fueled Prime Movers. Phase I.
Prepared by Radian Corporation for Gas Research Institute.
July 1990.
31. Hull, R., C. Urban, R. Thring, S. Ariga, M. Ingalls, and G.
O'Neal. Nox Control Technology Data Base for Gas-Fueled
Prime Movers, Phase I.- Prepared by Southwest Research
Institute for Gas Research Institute. April 1988.
32. Letter and attachments from Henegan, D., Norton Company, to
Snyder, R., MRI. March 28, 1991. Response to SCR
questionnaire.
33. Schorr, M. NOX Control for Gas Turbines: Regulations and
Technology. General Electric Company. Schenectady, New
York. Paper presented at the Council of Industrial Boiler
Owners NOX Control IV Conference. Concord, California.
February. 11-12, 1991. . 11 pp.
34. Letter and attachment from Smith, J. C., Institute of Clean
Air Companies, to Neuffer, W. J., EPA/ISB. May 14, 1992.
Response to EPA questionnaire regarding flue gas treatment
processes for emission reductions dated March 12, 1992.
35. State of California Air Resources Board. Determination of
Reasonably Available Control Technology and Best Available
Retrofit Technology for the Control of Oxides of Nitrogen
From Stationary Gas Turbines. May 18, 1992.
36. Field Survey of SCR Gas Turbine Operating Experience.
Prepared for the Electric Power Research Institute. Palo
Alto, CA. May, 1991.
6-37
-------
37. Offshore Gas Turbine NOX Control Technology Development
Program. Phase I Technology Evaluation. Arthur D. Little,
Inc. for Santa Barbara County Air Pollution Control
District. August 1989.
38. Champagne, D. See SCR Cost-effective for Small Gas
Turbines. Cogeneration. January-February 1988. pp. 26-29
6-38
-------
7.0 ENVIRONMENTAL AND ENERGY IMPACTS
This chapter presents environmental and energy impacts for
t
the nitrogen oxide (NOX) emissions control techniques described
in Chapter 5.0. These control techniques are water or steam
injection, dry low-NOx combustors, and selective catalytic
reduction (SCR). The impacts of the control techniques on air
pollution, solid waste disposal, water pollution, and energy
consumption are discussed.
The remainder of this chapter is organized in five sections.
Section 7.1 presents the air pollution impacts; Section 7.2
presents the solid waste disposal impacts; Section 7.3 presents
the water pollution impacts; and Section 7.4 presents the energy
consumption impacts. References for the chapter are listed in
Section 7.5.
7.1 AIR POLLUTION
7.1.1 Emission Reductions
Applying any of the control techniques discussed in
Chapter 5 will reduce NOX emissions from gas turbines. These
emission reductions were estimated for the model plants presented
in Table 6-1 and are shown in Table 7-1. For each model plant,
the uncontrolled and controlled emissions, emission reductions,
and percent reductions are presented. The following paragraphs
discuss NOX emission reductions for each control technique.
Nitrogen oxide emission reductions for water or steam
injection are estimated as discussed in Section 6.1.3. The
percent reduction in emissions from uncontrolled levels varies
for each model plant ranging, from 60 to 96 percent. This
reduction depends on each model's uncontrolled emissions, the
7-1
-------
o
CO
o
H
CO
CO
H
§
OLLED
IQUES
f^ •
Di O
[^
O W
u ^
EH M
&«
04
w
1
*
H
w
s
§ s
i o<
1:?
i§£
a! "»
1
i
2
S
o
3""
X
O
"2
§
u
is
* CJ3
it of o ai
i 1 'f »
.s 1* +
i
0X2
z s §
ill
a 1
oX2 s
i!1
0*2 >
il^
t-i^
Q u
.!.£ NO
•l-si
J|3
* 2 ^
•^
, o «- w
"3
o
'3
J
|
*o °
« * a
|=J|
lie
§ § §
< 8--"
1
3
cs
0 •* JR
~ ts S
NO cs ^
Z ' '
ON tS JR
» ^ 00
v^ NO K%
S 5! S
vo NO a«
* i ON oo
00°
00
I f 4
ill
1
s
00
f
s«>
1 « s
CJ m O
oo
cs
n
S — ON
Z ' '
Z ' '
Z ' '
a 5 *
00
f f 4
1 1 ^
III
1
e
00
S S J
i
00 oo *
*o 3"
^f *^ O
< ., ,
Z ' '
Z ' '
Si.
s
ft 4
8 ** •§
ill
i
5
00
II"5
i S 3
S
*o S «
S! 2 P
g
f f 3
i 1 I
•3 ts c
1
O
oo
111
IS 3
00
»n
fj
(S ON SR
Z ' '
Z ' '
Z ' '
<^ {S *
m °* P
S
f 1 1
22-2
a" a -3
5 S "P
•i i *
s
1
09
a
oo
111
ts
O ?*" ""*
Z ' '
Z ' '
Z ' '
a i S
oo
** u tfi
f i i
a a =
1 -1 1
e
I
0]
8
00
III
5
oo £&
"^ — ON
Z ' '
Z ' '
< u ,
^ 9\ SR
— o o
cs 5s! ri
8
1* U, Kg
-s -s -
! ! 1
I 1 1
•S 3 ~S
S ^ 5
u as H
1
8
00
°" ^
2 5 «
S
-------
T)
0)
-H
JJ
fl
0
O
H
C-
I
% s
2 o.
;i£0_
I1
s
,>»
ca
X
O
TJ
u
*
*\$
Dry low-NOx
combustor to
9 ppmv
fll
Dry low-NOx
combustor to
42 ppmv
O Cfl \Q
:|1|
y «« *ss 2
C u S
M &
Annual emissions8
!*•!
ill
i
3
V)
ri o ">
^f 55 o\
Z 1 I
z ' '
z ' '
.11
o
m
Emissions, tons/yr
Reduction, tons/yr
Toul reduction, %
S
1
01
oo"
ill
S S 3
S
3 5 S
a S 5
— 5i o\
5 g «
z ' '
.11
o
wi"
Emissions, tons/yr
Reduction, tons/yr
Toul reduction, %
S
1
01
oo"
Z
Sal
VO
SI!
.11
R § 5
m 2 oo
Z ' '
IIS
g
ts"
Emissions, tons/yr
Reduction, tons/yr
Toul reduction, %
S
1
oo"
III
ii 3
r*.
r~
91!
Z ' '
1 1 I
z ' '
Hi
s
Emissions, tons/yr
Reduction, tons/yr
Total reduction, %
01
00*
u, _
2 ™ 3
5
2 gj g
z ' '
z ' '
z ' '
oo M tK
? S3 S
S
Emissions, tons/yr
Reduction, tons/yr
Total reduction, %
I
oo"
,
in
O m O
8
•o o aft
>p s !S
^ O >O
Z ' '
Z ' '
Z ' '
o yj
o\ £3 •"*
'I
en"
Emissions, tons/yr
Reduction, tons/yr
Toul reduction, %
i
oo"
2 ^
is 1
o
2*8
z ' '
z ' '
z ' '
^1 •& 00
r- Z >o
o
a
Emissions, tons/yr
Reduction, tons/yr
Total reduction, %
S
ts
3
fill
323
2
"*
-------
T3
(U
G
-H
JJ
S
O
U
a S
-i I
I 2%
II i
jf
2
S
•a
1
o
z
"8
6
NOX
emissions, wet
injection
+ SCRb
0*2
||l
o 3
°*! a
Z S S
if 8:
& -e ">
O S
°*2 >
=? = 1
jj M O.
v « \A
•If 2
| o (2
li|!
2 i
*
1
•g |
I1 * .'§"
lie
§ * S
< |x
1
TT
oi
238
a 5 J
s S 1
z ' '
oo vi fcR
Tr a\ m
g
i* h M»
•a -a -
1 1 .2
. ts
Ul Oi H
1
5
M
2 > ,
111
r-
O
— °! O
2 ' '
2 ' '
Z ' '
2 S S
p<
S. S. *
111
1 1 1
1
8
«
^
S^ ,
| s a
cj « 5
wi
a s s
z ' '
z ' '
z ' '
oo 2 2
s
bt h y^
f .1 1
III
1
g
rf
W 00 "3
*§ <> S
2 . >» °^
Jii
I
g
M
u ^
i s 5
o
° ^ s
Q ^ O^
z ' '
z ' '
z ' '
2 m s
3
wi
>. x *
|Jj
1
s
^
H
I ^ -3
= 2 1
oo — b
1 • •
2 a *
Z ' '
Z ' '
z ' '
5o O> ON
g
u u wg
?^
1 1
III
s §
IM
s
F—l
1 1 1
£ a 3
X
s-
•Is
2 8
• •al
111
g I»
.SgO
o > o
111
•so.—
«i?
o U x
Ifl
S-S&"
8-M
-s
?!
•s. 2
|9
1 «§
Is
•S o
QS S g
302
2 -i o
11
I!
-3 -S
o .a x
D U
I a
^z«2
7-4
-------
water-to-fuel ratio (WFR), and type of fuel and whether water or
steam is injected.
Achievable emission levels from gas turbines using dry low-
NOV combustors were obtained from manufacturers. Controlled NOV
X -A-
levels of 42, 25, and 9 parts per million, by volume (ppmv),
referenced to 15 percent oxygen, were reported by the various
turbine manufacturers, and each of these levels is shown in
Table 7-1, where applicable, for each model plant. The percent
reduction in NO,, emissions from uncontrolled levels for gas
-A.
turbines using these combustors ranges from 68 to 98 percent.
Virtually all SCR units installed in the United States are used
in combination with either wet controls or combustion controls.
For this analysis, emission reductions were calculated for SCR in
combination with water or steam injection. Using the turbine .
manufacturers' guaranteed NOX emissions figures for wet injection
and a controlled NOX emission level of 9 ppmv, referenced to 15
percent oxygen, exiting the SCR, the percent reduction in NOX
emissions for this combination of control techniques ranges from
93 to 99 percent.
Estimated ammonia (NHj) emissions, in tons per year,
corresponding to ammonia slip from the SCR system are also shown
in Table 7-1. These estimates are based on an ammonia slip level
of 10 ppmv, consistent with information and data presented in
Section 5.4. For continuous-duty model plants, the annual NH3
emissions range from approximately 3 tons for a 3.3 megawatt (MW)
(4,425 horsepower [hp]) model plant to 72 tons for a 160 MW
(215,000 hp) model plant.
7.1.2 Emissions Trade-Offs
The formation of both thermal and fuel NOX depends upon
combustion conditions. Water/steam injection, lean combustion,
and reduced residence time modify combustion conditions to reduce
the amount of NOX formed. These combustion modifications may
increase carbon monoxide (CO) and unburned hydrocarbon (HC)
emissions. Using SCR to control NOX emissions produces ammonia
emissions. The impacts of these NO., controls on CO, HC, and
J^
ammonia emissions are discussed below.
7-5
-------
7.1.2.1 Impacts of Wet Controls on CO and HC Emissions. As
discussed in Section 5.1.5, wet injection may increase CO and HC
emissions. Injecting water or steam into the flame area of a
turbine combustor lowers the flame temperature and thereby
reduces NOX emissions. This reduction in temperature to some
extent inhibits complete combustion, resulting in increased CO
and HC emissions. Figure 5-12 shows the impact of water and
steam injection on CO emissions for production gas turbines.2
The impact of steam injection on CO emissions is less than that
of water injection. As seen in Figure 5-12, CO emissions
increase with increasing WFR's. Wet injection increases HC
emissions to a lesser extent than it increases CO emissions.
Figure 5-13 shows the impact of water injection on HC emissions
for one turbine. In cases where water and steam injection result
in excessive CO and HC emissions, an oxidation catalyst (add-on
control) can be installed to reduce these emissions by converting
the CO and HC to water (H20) and carbon dioxide (C02).
• 7.1.2.2 Impacts of Combustion Controls on CO and HC
Emissions. As discussed in Section 5.2.1, the performance of
lean combustion in limiting NOX emissions relies in part on
reduced equivalence ratios. As the equivalence ratio is reduced
below the stoichiometric level of 1.0, combustion flame
temperatures drop, and as a result NOX emissions are reduced.
Shortening the residence time in the high-temperature flame zone
also will reduce the amount of thermal NOX formed. These lower
equivalence ratios and/or reduced residence time, however, may
result in incomplete combustion, which may increase CO and HC
emissions. The extent of the increase in CO and HC emissions is
specific to each turbine manufacturer's combustor designs and
therefore varies for each turbine model. As with wet injection,
if necessary, an oxidation catalyst can be installed to reduce
excessive CO and HC emissions by converting the CO and HC to C02
and H20.
7.1.2.3 Ammonia Emissions from SCR. The. SCR process
reduces NOX emissions by injecting NH3 into the flue gas. The
NH3 reacts with NOX in the presence of a catalyst to form H20 and
7-6
-------
nitrogen (N2)• The NOX removal efficiency of this process is
partially dependent on the NH3/NOX ratio. Increasing this ratio
reduces NOX emissions but increases the probability that
unreacted ammonia will pass through the catalyst unit into the
atmosphere (known as ammonia "slip"). Some ammonia slip is
unavoidable because of ammonia injection control limitations and
imperfect distribution of the reacting gases. A properly
designed SCR system will limit ammonia slip to less than 10 ppmv
(see Section 5.4).
7.2 SOLID WASTE DISPOSAL
Catalytic materials used in SCR units for gas turbines
include precious metals (e.g., platinum), zeolites, and heavy
metal oxides (e.g., vanadium, titanium). Vanadium pentoxide, the
most commonly used SCR catalyst in the United States, is
identified as an acute hazardous waste under RCRA Part 261,
Subpart D - Lists of Hazardous Wastes. The Best Demonstrated
Available Technology (BDAT) Treatment Standards for Vanadium P119
and.P120 states that spent catalysts containing vanadium
pentoxide are not classified as hazardous waste.1 State and
local regulatory agencies, however, are authorized to establish
their own hazardous waste classification criteria, and spent
catalysts containing vanadium pentoxide may be classified as a
hazardous waste in some areas. Although the actual amount of
vanadium pentoxide contained in the catalyst bed is'-small, the
volume of the catalyst unit containing this material is quite
large and disposal can be costly. Where classified by State or
local agencies as a hazardous waste, this waste may be subject to
the Land Disposal Restrictions in 40 CFR Part 268, which allows
land disposal only if the hazardous waste is treated in
accordance with Subpart D - Treatment Standards. Such disposal
problems are not encountered with other catalyst materials, such
as precious metals and zeolites, because these materials are not
hazardous wastes.
7-7
-------
7.3 WATER USAGE AND WASTE WATER DISPOSAL
Water availability and waste water disposal are
environmental factors to be considered with wet injection. The
impact of water usage on the water supply at some remote sites,
in small communities, or in areas where water resources may be
limited is an environmental factor that should be examined when
considering wet injection. The volume of water required for wet
injection is shown in Table 7-2 for each model plant.
Water purity is essential for wet injection systems in order
to prevent erosion and/or the formation of deposits in the hot
sections of the gas turbine. Water treatment systems are used to
achieve water quality specifications set by gas turbine
manufacturers. Table 5-4 summarizes these specifications for six
manufacturers.
Discharges from these water treatment systems have a
potential impact on water quality. As indicated in Section 6.1,
approximately 29 percent of the treated water flow rate
(22.5 percent of the raw water flow rate) is considered to be .
discharged as wastewater. The wastewater flow rates for each of
the model plants with a water or steam injection control system
are estimated using this factor, and the results are presented in
Table 7-2. The wastewater contains increased levels of those
pollutants in the raw water (e.g., calcium, silica, sulfur, as
listed in Table 5-4) that are removed by the water treatment
system, along with any chemicals introduced by the treatment
process. Based on a wastewater flowrate equal to 29 percent of
the influent raw water, the concentration of pollutants
discharged from the water treatment system is approximately three
times higher than the pollutant concentrations in the raw water.
The impacts of these pollutants on water quality are
site-specific and depend on the type of water supply and on the
discharge restrictions. Influent water obtained from a
municipality will not contain high concentrations of pollutants.
However, surface water or well water used at a remote site might
contain high pollutant concentrations and may require additional
pretreatment to meet the water quality specifications set by
7-8
-------
TABLE 7-2.
WATER AND ELECTRICITY CONSUMPTION FOR N03
CONTROL TECHNIQUES
Gas turbine
model*
Centaur T4500
501-KB5
LM2500
MS5001P
ABBOT11N
MS7001E
501-KB5
LM2500
MS5001P
LM5000
ABBGT11N
MS7001E
MS7001F
Centaur T4SOO
MSS001P
MS7001E
Cenuur T4500
MS5001P
MS7001E
Centaur T4500
MS5001P
MS7001E
SATURN
T1500
TPMFT4
Turbine
power
output,
MW
3.3
4.0
22.7
26.8
83.3
84.7
4.0
22.7
26.8
34.4 '
83.3
84.7
161
3.3
26.3
83.3
3.3
26.3
84.7
3.3
26.3
84.7
1.1
28.0
Annual
operating
hours
8,000
8,000
8,000
8,000
8,000
8,000
8,000
8,000
8,000
8,000
8,000
8,000
8,000
8,000
8,000
8,000
2,000
2,000
2,000
2,000
2,000
2,000
1,000
1,000
Fuel
type
Gas
Gas
Gas
Gas
Gas
Gas
Gas
Gas
Gas
Gas
Gas
Gas
Gas
Oil
Oil
Oil
Gas
Gas
Gas
Oil
Oil
Oil
Oil
Oil
Type of
emission
control
Water inj.
Water inj.
Water inj.
Water inj.
Water inj.
Water inj.
Steam inj.
Steam inj.
Steam inj.
Steam inj.
Steam inj.
Steam inj.
Steam inj.
Water inj.
Water inj.
Water inj.
Water inj.
Water inj.
Water inj.
Water inj.
Water inj.
Water inj.
Water inj.
Waterr
in-oil
emulsion
Total
water
flow,
gal/mina
2.5
3.94
14.8
22.2
154
69.2
7.38
29.5
33.3
50.8
178
104
199
2.76
26.7
63.8
2.50
22.2
69.2
2.76
26.7
63.8
0.81
21.7
Waste
water
flow,
gal/min"
0.73
1.14
4.29
6.44
44.7
20.1
2.14
8.56
9.66 .
14.7
51.6
30.2
57.7
0.80
7.74
18.5
0.73
6.44
20.1
0.80
7.74
18.5
0.23
6.29
Water
pump
power,
kW°
0.40
0.63
2.38
3.57
24.8
11.1
1.19
4.75
5.36
8.18
28.7
16.7
32.0
0.44
4.30
10.3
0.40
3.57
11.1
0.44
4.30
10.3
0.13
3.49
Wet injec-
tion power
consump-
tion,
kW-hr/yrd
3,220
5,070
19,100
28,600
198,000
89,100
9,510
38,000
42,900
65,400
229,000
134,000
256,000
3,550
34,400
82,200
3,220
28,600
89,100
3,550
34,400
82,200
1,040
27,900
SCR
power
penalty,
kW-hr/yre
132,000
160,000
908,000
1,070,000
3,330,000
3,390,000
160,000
908,000
1,070,000
1,380,000
3,330,000
3,390,000
6,440,00a
132,000
1,050,000
833,000
33,000
263,000
847,000
33,000
263,000
847,000
5,500
140,000
"From Table 6-2.
bCalculated as 29 percent of the total water flow.
cPower requirement for water pump is calculated as shown in Section 6.1.2.2.
dWet injection electricity usage = (water pump IcW) X (annual operating hours).
eSCR power penalty » (0.005 X turbine power output, kW) X (annual operating hours).
7-9
-------
manufacturers. This additional pretreatment will increase the
pollutant concentrations of the wastewater discharge. Wastewater
discharges to publicly-owned treatment works (POTW's) must meet
the requirements of applicable Approved POTW Pretreatment
Programs.
7.4 ENERGY CONSUMPTION
Additional fuel and electrical energy is required over
baseline for wet injection controls, while additional electrical
energy is required for SCR controls. The following paragraphs
discuss these energy consumption impacts.
Injecting water or steam into the turbine combustor lowers
the net cycle efficiency and increases the power output of the
turbine. The thermodynamic efficiency of the combustion process
is reduced because energy that could otherwise be available to
perform work in the turbine must now be used to heat the
water/steam. This lower efficiency is seen as an increase in
fuel use. Table 5-10 shows the impacts of wet injection on gas
turbine performance for one manufacturer. This table shows a 2
to 4 percent loss in efficiency associated with WFR's required to
achieve NOX emission levels of 25 to 42 ppmv in gas turbines
burning natural gas. The actual efficiency loss is specific to
each turbine model but generally increases with increasing WFR's
and is higher for water injection than for steam injection
(additional energy is required to heat and vaporize the water).
One exception to this efficiency penalty occurs with steam
injection, in which exhaust heat from the gas turbine is used to
generate the steam for injection. If the heat recovered in
generating the steam would otherwise be exhausted to atmosphere,
the result is an increase in net cycle efficiency.'
The energy from the increased mass flow and heat capacity of
the injected water/steam can be recovered in the turbine,
resulting in an increase in power output accompanying the reduced
efficiency of the turbine (shown in Table 5-10 for one manufac-
turer) . This increase in power output can be significant and
could lessen the impact of the loss in efficiency if the facility
has a demand for the available excess power.
7-10
-------
Water and steam injection controls also-require additional
electrical energy to operate the water injection feed water
pumps. The annual electricity usage for each model is the
product of the pump power demand, discussed in Section 6.1.2.2,
and the annual hours of operation. Table 7-2 summarizes this
electricity usage for each of the model plants.
For SCR units, additional electrical energy is required to
operate ammonia pumps and ventilation fans. This energy
requirement, however, is believed to be small and was not
included in this analysis.
The increased back-pressure in the turbine exhaust system
resulting from adding an SCR system reduces the power output from
the turbine. As discussed in Section 6.3.2.9, the power output
is typically reduced by approximately 0.5 percent. This power
penalty has been calculated for each model plant and is shown in
Table 7-2.
7.5 REFERENCE FOR CHAPTER 7
1. 55 PR 22276, June 1, 1990.
7-11
-------
APPENDIX A
Exhaust NCL- emission levels were provided by gas turbine
Jv
manufacturers in units of parts per million, by volume (ppmv) , on
a dry basis and corrected to 15 percent oxygen. A method of
converting these exhaust concentration levels to a mass flow rate
of pounds of NOX per hour (Ib N0x/hr) was provided by one gas
turbine manufacturer.1 This method uses an emission index
(EINOV) , in units of Ib NOY/1,000 Ib fuel, which is proportional
Jt Jv
to the exhaust NOX emission levels in ppmv by a constant, K. The
relationship between EINOX and ppmv for NOX emissions is stated
in Equation 1 below and applies for complete combustfbn of a
hydrocarbon fuel and combustion air having no C02 and an 02 mole
percent of 20.95:
NO Ref . 15% 02 = K Equation 1
where: NOX Ref. 15% 02 = N°x/ Ppnwd- @15% 02 (provided by gas
turbine manufacturers) ;
EINO-. = NOV emission index, Ib NO.,/1,000 Ib
Jv _ A_ _ .A.
fuel ; and
K = constant, based on the molar
hydrocarbon
ratio of the fuel.
The derivation of Equation 1 was provided by the turbine
manufacturer and is based on basic thermodynamic laws and
supported by test data provided by the manufacturer. According
to the manufacturer, this equation can be used to estimate NOY
Jx
emissions for operation with or without water/steam injection.
Equation 1 shows that NOX emissions are dependent -only upon
the molar hydrocarbon ratio of the fuel and are independent of
the air/fuel ratio (A/P) . The equation therefore is valid for
all gas turbine designs for a given fuel. The validity of this
approach to calculate NOX emissions was supported by a second
-------
turbine manufacturer.2 Values for K were provided for several
fuels and are given below:1'2
Pipeline quality natural gas: K = 12.1
Distillate fuel oil No. 1 (DF-1): K = 13.1
Distillate fuel oil No. 2 (DF-2): K = 13.2
Jet propellant No. 4 (JP-4): K - 13.0
Jet propellant No. 5 (JP-5): K = 13.1
Methane: K = 11.6
The following examples are provided for calculating NOX
emissions on a mass basis, given the fuel type and NOX emission
level, in ppmv, dry (ppmvd) , and corrected tt> 15 percent 02.
Example 1. Natural gas fuel
Gas turbine: Solar Centaur 'H'
Power output: 4,040 kW
. Heat rate: 12,200 Btu/kW-hr
NOX emissions: 105 ppmvd, corrected to 15 percent 02
Fuel: Natural gas
- lower heating value = 20,610 Btu/lb
- K » 12.1
Fuel flow:
4,040 kWx 12'2Q°BtU X * lb £uel = 2,391 lb/hr
kW-hr 20,610 Btu
From Equation 1:
105
EINOX
12.1
A-2
-------
NOX emissions, Ib/hr:
lb fuel
2/391
hr
8.68 lb NOX
1,000 lb fuel
= 20.8
lb NOS
hr~~
Example 2. Distillate oil fuel
Gas turbine: General Electric LM2500
Power output: 22670 kW
Heat rate: 9296 Btu/kW-hr
Nox emissions: 345 ppmvd, corrected to 15 percent 02
Fuel: Distillate oil No. 2
lower heating value = 18,330 Btu/lb
K = 13.2
Fuel flow:
22,670 kW X 9296 BtU X
, T o
kW-hr 18,330Btu
= 11,500 Ib/hr
From Equation 1:
345
EINO,
= 13.2
NOX emissions, Ib/hr:
lb
11 500
11,500
26 • 1 lb N°
300
300
REFERENCES FOR APPENDIX A:
1. Letter and attachments from Lyon, T.F., General Electric
Aircraft Engines, to Snyder, R.B., MRI. December 6, 1991.
Calculation of NOX emissions from gas turbines.
2. Letter and attachments from Hung, W.S., Solar Turbines, Inc.,
to Snyder, R.B., MRI. December 17, 1991. Calculation of NO
emissions from gas turbines.
X
A-3
-------
APPENDIX B. COST DATA AND METHODOLOGY USED TO PREPARE COST
FIGURES PRESENTED IN CHAPTER 6
-------
APPENDIX B. RAW COST DATA AND COST ALGORITHMS
The maintenance costs for water injection and several of the
SCR costs presented in Chapter 5 are based on information from
turbine manufacturers and other sources that required
interpretation and analysis. Information about additional gas
turbine maintenance costs associated with water injection is
presented in Section B.I. Information on SCR capital costs,
catalyst replacement and disposal costs, and maintenance costs is
presented in Section B.2. References are listed in Section B.3.
B.I WATER INJECTION MAINTENANCE COSTS
Information from each manufacturer and the applicable
analysis procedures used to develop maintenance cost impacts for
water injection are described in the following sections.
B.I.I Solar
This manufacturer indicated that the annual maintenance cost
for the Centaur is $16,000/year.1 The cost for the Saturn was
estimated to be $8,000.2 This $8,000 cost was then prorated for
operation at 1,000/hr/yr, and was multiplied by 1.3 to account
for the additional maintenance required for oil fuel.
B.I.2 Allison
Maintenance costs for water injection were provided by a
company that packages Allison gas turbines for stationary
applications. This packager stated that for the 501 gas turbine
model, a maintenance contract is available which covers all
maintenance materials and labor costs associated with the
turbine, including all scheduled and unscheduled activities. The
cost of this contract for the 501 model is $0.0005 to $0.0010 per
KW-hour (KWH) more for water injection than for a turbine not
using water injection.-^ For an installation operating
8,000 hours per year at a base-rated output of 4,000 KW, and
using an average cost of $0.00075 per KWH, the annual additional
maintenance cost is $24,000. By the nature of the contract
offered, this figure represents a worst case scenario and to some
extent may exceed the actual incremental maintenance costs that
would be expected for water injection for this turbine.
B-l
-------
B.I.3 General Electric
General Electric (GE) offers both aero-derivative type
(LM-series models) and heavy-duty type (MS-series models) gas
turbines. For the aero-derivative turbines, GE states that the
incremental maintenance cost associated with water injection is
$3.50 per fired hour. This cost is used to calculate the
maintenance cost for water injection for GE aeroderivative
turbines. No figures were provided for steam injection and no
maintenance cost was used for steam injection with these
turbines.4
Water injection also impacts the maintenance costs for the
heavy-duty MS-series models. Costs associated with more frequent
maintenance intervals required for models using water injection
have been calculated and summarized below. A GE representative
stated that the primary components which must be repaired at each
maintenance interval are the combustor liner and transition
pieces. Approximate costs to repair these pieces were provided
by GE.5 For this analysis, the maximum cost estimates were used
to calculate annual costs to accommodate repairs that may be
required periodically for injection nozzles, cross-fire tubes,
and other miscellaneous hardware. According to GE, a rule of
thumb is that if the repair cost exceeds 60 percent of the cost
of a new part, the part is replaced.5 The cost of a replacement
part is therefore considered to be 1.67 times the maximum repair
cost. If water purity requirements are met, there are no
significant adverse impacts on maintenance requirements on other
turbine components, and hot gas path inspections and major
inspection schedules are not impacted.5 Combustion repair
schedules, material costs, and labor hours are shown in
Table B-l. Scheduled maintenance intervals for models with water
injection were provided in Reference 6. Corresponding
maintenance intervals for models with steam injection were
assumed to be the same as models with no wet injection; these
scheduled maintenance intervals were provided in Reference 7.
Using the information in Table B-l, the total annual, cost is
B-2
-------
calculated and shown in Table B-2 for three -GE heavy-duty turbine
models.
B.I.4 Asea Brown Bpveri
This manufacturer states there are no maintenance impacts
associated with water injection.**
B.2 SCR COSTS
The total capital investment, catalyst replacement, and
maintenance costs are estimated based on information from the
technical literature. The cost algorithms are described in the
following sections.
B.2.1 Total Capital Investment
Total capital investment costs, which include purchased
costs and installation costs, were available for SCR systems for
combined cycle and cogeneration applications from five
sources.9"13 These costs were scaled to 1990 costs using the
Chemical Engineering annual plant cost indexes and are applicable
to SCR systems in which the catalyst was placed within the heat
recovery steam generator (HRSG). In addition, estimated capital
investment, costs were available from one source for SCR systems
in which a high temperature zeolite catalyst is installed
upstream of the HRSG.14 Both the original data and the scaled
costs are presented in Table B-3. The scaled costs were plotted
against the turbine size and this plot is shown in Figure B-l. A
linear regression analysis was performed to determine the
equation for the line that best fits the data. This equation was
used to estimate the total capital investment for SCR for the
model plants and was extrapolated to estimate the costs for model
plants larger than 90 MW.
B.2.2 Maintenance Costs
Maintenance costs for SCR controls were obtained from four
literature sources, although 6 of the 14 points were obtained
from one article.9'11"13 These costs were scaled to 1990 costs
assuming an inflation rate of five percent per year. All of the
data are for turbines that use natural gas fuel. Because there
are no data to quantify differences in SCR maintenance costs for
oil-fired applications, the available data for operation on
B-3
-------
natural gas were used for both fuels. Both the original data and
the scaled costs are presented in Table B-4. The scaled costs
were plotted versus the turbine size in Figure B-2. The equation
for the line through the data was determined by linear
regression, and it was used to estimate the maintenance costs for
the model plants.
B.2.3 Catalyst Replacement Costs
Catalyst replacement costs were obtained from three articles
for nine gas turbine installations.^'11'13 Combined catalyst
replacement and disposal costs were obtained for another six gas
turbine installations from one article.1^ The disposal costs for
these six gas turbine installations were estimated based on
estimated catalyst volumes and a unit disposal cost of $15/ft3,
given in Reference 15.
The catalyst volumes were estimated assuming there is a
direct relationship between the volume and the turbine size; the
catalyst volume stated in Reference 16 for one 83 MW turbine is
175 .m3. The resulting disposal costs for these six facilities .
were subtracted from the combined replacement and disposal costs-
to estimate the replacement-only costs. All of the replacement
costs were scaled to 1990 costs assuming an inflation rate of
5 percent per year. The original data and the scaled costs are
presented in Table B-5, and the scaled replacement costs were
also plotted versus the turbine size in Figure B-3. Linear
regression was used to determine the equation for the line
through the data. This equation was used to estimate the
catalyst replacement costs for the model plants.
B-4
-------
CQ
(U
c
•H
•8
-------
co
0)
fi
-H
CO
o
Og
8|
8
o
(0
O
O
DC
CM
8 S 8 S.
(spuBsnoqjJ
j//$ isoo aoireueiuiEoi
S
CD
.N
(O
nl
O
I
^
Cn
tn
G
O
•H
CO
CO
X
O
55
O
r-t
O
04
O
CO
i-i
O
M-l
4J
m
o
a)
o
a
m
c
a)
nj
03
(U
I
•H
B-6
-------
CO
o
O
CO
co
CD
o
o
eg
03
O
CO
0)
(spuBsnotji)
CQ
JJ
c
o
u
o
CO
o
4-1
4J
CQ
o
u
8
-------
CD
N
CO
0
H
0
TO
cc
<
CM
CQ
0)
c
-H
•8
4J
(TJ
O
JJ
0)
M
CD
OJ
M
en
-H
fa
S/q|'8iEJMO|JJIBJ8|U|
B-8
-------
EH
CO
O
CJ
H
&
CO
§
CO
H
Oi
O4
U
i*
g
m
w
I
Is
i
.o
!!
•H
V
Q
3
.
-1 H
Ul
ca
U
.S
1
I
^
t-; « g «i
(5 a< 05 oS
B-9
-------
S3
O
H
E-»
U
W
§
H
W
EH
OS
O
b
OS
H
D
O
W
as
w
o
H
CN
I
CQ
W
S
a
J
s;
V
i
1
oC
I-
*
-------
TABLE B-3. TOTAL CAPITAL INVESTMENT FOR SCR TO CONTROL
NO EMISSIONS FROM GAS TURBINES
Gas
turbine
size, MW
1.1
1.5
3
3.2
3.7
3.7
4
4.5
6
8.4
9
10
20
21
. 21
21
22
26
33
37
37
78
80
80
83
SCR capital costa
$
1,250,000
180,000
320,000
600,000
477,000
579,000
839,000
750,000
480,000
800,000
1,100,000
1,431,000
1,700,000
798,000
1,500,000
1,200,000
1,000,000
1,800,000
990,000
2,000,000
2,700,000
4,300,000
5,400,000
1,760,000
5,360,000
Year
1989
1986
1986
1989
1988
1989
1991
1988
1986
1986
1987
1991
1987
1988
1986
1986
1987
1991
1988
1986
1986
1986
1987
1988
1991
Refb
9
10
10
11
12
11
14
11
10
11
13
14
13
12
10
10
11
14
12
11
10
10
13
12
14
Scaling
factor^
357.6/355.4
357.6/318.4
357.6/318.4
357.6/3.554
357.6/342.5
357.6/355.4
1.0
357.6/342.5
357.6/318.4
357.6/318.4
357.6/323.8
1.0
357.6/323.8
357.6/342.5
357.6/318.4
357.6/318.4
357.6/323.8
1.0
357.6/342.5
357.6/318.4
357.6/318.4
357.6/318.4
357.6/323.8
357.6/342.5
1.0
1990 SCR
capital
cost, $
1,260,000
202,000
359,000
604,000
498,000
583,000
839,000
783,000
539,000
898,000
1,210,000
1,431,000
1,880,000
833,000
1,680,000
1,350,000"
1,100,000
1,800,000
1,030,000
2,250,000
3,030,000
4,830,000
5,960,000
1,840,000
5,360,000
continued
B-ll
-------
TABLE B-3. (Continued)
aTotal capital costs were provided by several sources, but it is
not clear that they are on the same basis. For example, it is
likely that the type of catalyst varies and the target NOX
reduction efficiency may also vary. In addition, some estimates
may not include costs for emission monitors; auxiliary equipment
like the ammonia storage, handling, and transfer system; taxes
and freight; or installation.
"Reference 12 also provided costs for SCR used with 136 MW and
145 MW turbines. All of the costs for this reference are lower
than the costs from other sources, and the differential
increases as the turbine size increases. Because there are no
costs from other sources for such large turbines, these two data
points would exert undue influence on the analysis; therefore,
they have been excluded. Costs for large model plants were
estimated by extrapolating with the equation determined by
linear regression through the data for turbines with capacities
less than 90 MW (see Figure B-l).
GCosts for years prior to 1990 are adjusted to 1990 dollars
based on the annual CE plant cost indexes. Costs estimated in
1991 dollars were not adjusted.
B-12
-------
TABLE B-4. MAINTENANCE COSTS-FOR SCR
Gas
turbine
size, MW
1.1
3.2
3.7
3.7
8.4
8.9
9
20
21
33
80
80
136
145
SCR maintenance costa
$/yr
52,200
50,000
43,000
15,500
22,000
18,000
25,000
50,000
37,900
63,700
124,000
60,000
184,000
205,000
Year
1989
1989
1988
1988
1986
1988
1987
1987
1988
1988
1988
1987
1988
1988
Ref
9
11 '
11
12
11
11
13
13
12
12
- 12
13
12
12
Scaling
factor15
1.050
1.050
1.103
1.103
1.216
1.103
1.158
1.158
1.103
1.103
1.103
1.158
1.103
1.103
1990 SCR
maintenance
cost, $
54,800
52,500
47,400
17,100
26,700
19,800
28,900
57,900
41,800
70,200
137,000
69,500
203,000
226,000
aAll of the maintenance costs are for turbines that are fired .
with natural gas. Although sulfur in diesel fuel can cause
maintenance problems, there are no data to quantify the impact.
Therefore, the maintenance costs presented in this table were
used for both natural gas and diesel fuel applications.
^Scaling factors are based on an estimated inflation rate of
5 percent per year.
B-13
-------
CO
o
o
o
o<
CO
H
w
E-»
CO
>
J
I
IT)
ffl
w
J
a
~
o
§
Ov
§
§
3
8
8
S
g
8
8
S
CM
VO
a
es
a
S
s
es
S
§
S
s
8
en1
cS
il
If
•^ m. r
2. % S
**
1
os...,
3|§3
! .r* 8
B-14
-------
B.3 REFERENCES FOR APPENDIX B
1. Letter and attachments from Swingle, R., Solar Turbines
Incorporated, to Snyder, R., MRI. May 21, 1991.
Maintenance considerations for gas turbines.
2. Letter and attachments from Swingle, R., Solar Turbines
Incorporated, to Neuffer, W.J., EPA/ISB. August 20, 1991.
Review of draft gas turbine ACT document.
3. Letter and attachments from Lock, D., U.S. Turbine
Corporation, to Neuffer, W.J., U.S EPA/ISB. September 17,
1991. Review of draft gas turbine ACT document.
4. Letter and attachments from Sailer, E.D., General Electric
Marine and Industrial Engines, to Neuffer, • W.J., EPA/ISB.
August 29, 1991. Review of draft gas turbine ACT document.
5. Telecon. Snyder, R., MRI, with Pasquarelli, L.,
General Electric Company. April 26, 1991. Maintenance
costs for gas turbines.
6. Letter and attachment from Schorr, M., General Electric
Company, to Snyder, R., MRI. April 1, 1991. Response
to gas turbine questionnaire.
7. • Walsh, E. Gas Turbine Operating and Maintenance
Considerations. General Electric Company.
Schenectady, NY. Presented at the 33rd GE Turbine
State-of-the-Art Technology Seminar for Industrial,
Cogeneration and Independent Power Turbine Users.
September, 1989. 20 pp.
8. Letter and attachments from Gurmani, A., Asea Brown
Boveri, to Snyder, R., MRI. May 30, 1991. Response to
gas turbine questionnaire.
9. Permit application processing and calculations by South
Coast Air Quality Management District for proposed SCR
control of gas turbine at Saint John's Hospital and
Health Center, Santa Monica, California. May 23, 1989.
10. Hull, R., C. Urban, R. Thring, S. Ariga, M. Ingalls,
and G. O'Neal. NOX Control Technology Data Base for
Gas-Fueled Prime Movers, Phase I. Prepared by
Southwest Research Institute for Gas Research
Institute. April 1988.
11. Shareef, G., and D. Stone. Evaluation of SCR NOX Controls
for Small Natural Gas-Fueled Prime Movers. Phase I.
Prepared by Radian Corporation for Gas Research Institute.
July 1990.
B-15
-------
12. Sidebotham, G., and R. Williams'. Technology of NOX Control
for Stationary Gas Turbines. Center for Environmental
Studies. Princeton University. January 1989.
13. Prosl, T., DuPont,and Scrivner, G., Dow. Technical
Arguments and Economic Impact of SCR's Use for NOX
Reduction of Combustion Turbine for Cogeneration.
Paper presented at EPA Region 6 meeting concerning NOX
abatement of Combustion Turbines. December 17, 1987.
14. State of California Air Resources Board. Draft Proposed
Determination of Reasonably Available Control Technology And
Best Available Retrofit Technology for Stationary Gas
Turbines. August, 1991. Appendix C.
15. Letter and attachments from Henegan, D., Norton
Company, to Snyder, R., MRI. March 28, 1991. Response
to SCR questionnaire.
16. Schorr, M. NOX Control for Gas Turbines: Regulations
and Technology. General Electric Company.
Schenectady, New York. Paper presented at the Council
of Industrial Boiler Owners NOX Control IV Conference.
Concord, California. February 11-12, 1991. 11 pp.
B-16
-------
TECHNICAL REPORT DATA
Please read instructions on me reverse oetore comoienngi
1 REPORT NO. 12. 3. RECIPIENT'S ACCESSION NO.
EPA-453/R-93-007 | i
4. TITLE AND SUBTITLE
Alternative Control Techniques Document-
NOX Emissions from Stationary Gas Turbines
7 AUTHORIS)
Robert B. Snyder
9. PERFORMING ORGANIZATION NAME ANO ADDRESS
Midwest Research Institute
401 Harrison Oaks Boulevard
Gary, NC 27513-2412
12. SPONSORING AGENCY NAME ANO ADDRESS
U. S. Environmental Protection Agency
Emission Standards Division (MD-13)
Office of Air Quality Planning Standards
Research Trianqle Park, NC 27711
3. REPORT DATE
January 1993
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO.
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-01-0115
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY COOE
15. SUPPLEMENTARY NOTES
EPA Work Assignment Manager: William Neuffer (919) 541-5435
16. ABSTRACT
This Alternative Control Techniques document describes available control"
technologies for reducing NOX emissions levels from stationary combustion gas
turbines. Information on the formation of NOX and uncontrolled NOX emissions from
gas turbines is included. Water injection, steam injection, and low-NOx combustors,
used independently or in combination with selective catalytic reduction (SCR), are
discussed. Achievable controlled NOX emissions levels, costs and cost
effectiveness, and environmental impacts are presented and applicability to new
equipment as well as retrofit applications is discussed. The application of these
technologies to gas turbines operating in offshore platform applications is
included. Information on the use of alternate fuels, catalytic combustion, and
selective noncatalytic reduction (SNCR) to reduce NOX emissions is also briefly
presented.
17. KEY WORDS ANO DOCUMENT ANALYSIS
a. DESCRIPTORS
Stationary Gas Turbines
Nitrogen Oxide Emissions
Water/Steam Injection
Selective Catalytic Reduction (SCR)
Control Techniques for NOX Emissions
Costs of Emissions Control
Dry Low-NOx Combustion
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI 1 ifid.CfOUp
'18. DISTRIBUTION STATEMENT
I
19. SECURITY CLASS IThu Report i
21. NO. OF PAGES
246
20. SECURITY CLASS /This page>
I22. PRICE
EPA Form 2220-1 (R.v. 4-77)
-BEVlOUS EDi TtON I S OBSOLETE
-------