SO3 Mitigation Guide and Cost Estimating Workbook
Paper #34
Katherine Dombrowski, Gary Blythe
URS Corporation, P.O Box 201088, Austin, TX 78720
Richard Rhudy
EPRI, 3412 Hillview Avenue, Palo Alto, CA 94304
ABSTRACT
SO3 or sulfuric acid in flue gas is an issue for plants that fire medium- to high-sulfur coals,
particularly those with SCR for NOX control, because of adverse effects such as duct corrosion,
air heater plugging and elevated plume opacity. EPRI has recently published an updated SO3
Mitigation Guide that discusses the state of the art for measuring and controlling SO3 emissions
from coal-fired plants. Included in the guide are cost estimates for ten different SO3 control
technologies as applied to an example plant, and a cost estimating workbook that allows the
report user to develop site specific estimates for candidate control technologies. This paper
presents and discusses the cost estimate data for the example plant. Capital, reagent, and utilities
costs are presented for each of ten technologies for comparison. Sensitivity case estimates are
also presented showing the effects of delivered reagent costs, SO3 control system operating
period (ozone season only versus year-round), and new versus retrofit installations.
INTRODUCTION
A wide range of SO3 or sulfuric acid controls have been tested and/or applied for coal-fired
power plants, such as low-sulfur-coal switches, injecting alkaline sorbents at various locations in
the furnace or ductwork, and installing wet ESPs. The costs of applying these technologies can
vary over a wide range, and each has positive and negative attributes that could make it more or
less favorable for application at a given power plant. This paper describes a number of
SO3/sulfuric acid control technologies and presents estimated costs for applying these
technologies at an example plant that is being retrofitted with SCR for NOX control.
Figure 1 shows the flue gas path of a typical coal-fired power plant, and illustrates where a
number of candidate SO3/sulfuric acid control technologies might typically be applied. The
following paragraphs briefly discuss each of the potential control technologies illustrated.
Fuel Switch/Blending
Low-sulfur coal blending or switching lowers the SO2 content of the flue gas, which reduces the
concentration of SO3 produced in the furnace and across an SCR catalyst. The ability to
implement this option and its cost effectiveness can be very site specific. If low-sulfur-coal firing
is feasible, it can result in other benefits such as making more SO2 allowance tons available and
allowing lower air heater outlet flue gas temperatures, the latter of which can result in plant heat
rate improvements.
Figure 1: Illustration of Potential SO3/Sulfuric Acid Control Options
- Byproduct
Mg Injection
- Commercial
Mg Injection
To
Stack
Wet ESP
SCR
- Fuel
Switch
- Fuel
Additive
- MgO Injection
Economizer
Air Heater
- SBS Injection
Outlet
Wet ESP
conversion
ESP
FGD Absorber
- SBS Injection
- Hydrated Lime Injection
ID/Booster Fans
- Ammonia Injection
- Humidification with Alkali Injection
Notwithstanding capital expenses that may be required, the economics of coal switching are
driven almost entirely by the delivered cost differential between the current coal and the lowsulfur coal. Changes in the delivered fuel price of a few cents per million Btu can change the
annual costs for this control option by several hundred thousand dollars or more.
Fuel Additive
Omni Materials, Inc., of Mt. Carmel, Illinois has developed a proprietary fuel additive for coalbased applications, OmniClearTM. Based on information provided in company literature, the
additive is apparently ground dolomite (75% minus 200 mesh) that is partially calcined, then
treated with a coating that helps prevent dead burning.1 The company literature reports the
results of a full-scale test on a plant with an SCR and firing approximately a 3% sulfur coal (5 lb
SO2/MM Btu). An additive rate of approximately 0.5 to 1.0 wt% of the coal feed rate reportedly
lowered SCR outlet SO3 concentrations from 50 to 65 ppm to about 5 ppm at the ESP outlet.
This equates to approximately 90 to 92% SO3 removal. However, since no ESP outlet
concentration data were reported for without the additive, it is not clear what percentage of the
reported removal was attributable to the additive (i.e., some removal would be expected across
the air heater and ESP without the additive).
Furnace Injection of Magnesium-based Alkalis
This technology involves injecting magnesium hydroxide (Mg(OH)2) aqueous slurry into the
upper furnace, where it reacts with SO3 to produce magnesium sulfate solids. These solids are
removed in the downstream particulate control device. Testing co-funded by EPRI focused on
the injection of magnesium hydroxide slurries into coal-fired furnaces to achieve 90+% removal
of furnace-formed SO3.2 For plants without SCR, furnace Mg injection is capable of high SO3
2
removal percentages, but SO3 removal may be limited by the performance of downstream coldside ESPs. At high SO3 removal percentages, the loss of sulfuric acid conditioning of fly ash
particles appears to cause high resistivity conditions that limit ESP performance.3 For plants
with SCR, furnace injection can lower SO3 concentrations in the SCR inlet flue gas, which can
allow SCR operation at lower load without capillary condensation of ammonium bisulfate in
catalyst pores. However, furnace Mg injection does not appear to be effective at removing SO3
formed across the SCR catalyst.3
With or without SCR, furnace Mg injection should lower the flue gas acid dew point, which may
allow the plant to operate at lower air heater outlet flue gas temperatures and realize heat rate
improvements. The potential benefits of SO3 removal upstream of the air heater on plant heat
rate are mentioned several places in this paper, and warrant explanation. Most coal-fired power
plants have steam or glycol heaters to preheat the ambient combustion air going to the air
heaters. The amount of preheat is adjusted to control the air heater outlet air and flue gas
temperatures. On bituminous coal-fired plants, the air heater outlet flue gas temperature is
typically controlled to minimize adverse effects of sulfuric acid condensation, such as plugging
and/or corrosion in the air heater, or downstream corrosion. Lowering the flue gas SO3
concentration upstream of the air heater, such as by furnace Mg injection, may allow the plant to
lower the air heater outlet temperature. This is typically implemented by reducing the amount of
steam used to preheat the combustion air to the air heater (or to heat the glycol solution).
Reduced steam consumption for combustion air preheat can improve the overall plant heat rate.
For this evaluation, cost estimates have been developed for both an FGD byproduct Mg(OH)2
slurry (byproduct Mg) and commercial Mg(OH)2 slurry (commercial Mg).
Alkali Injection into the SCR Outlet Duct
Costs were estimated for two injection technologies for downstream of the SCR and upstream of
the air heater: MgO powder injection and sodium bisulfite/sulfite solution (SBS) injection. These
alkalis react with SO3 to form salts that are removed in the particulate control device. Since these
technologies can remove SO3 upstream of the air heater, they offer a potential for lower air
heater gas exit temperatures, with corresponding plant heat rate improvements. They may also
reduce corrosion in the air heater and downstream equipment.
The reactive MgO powder is marketed by Martin Marietta, but other vendors may provide
similar products. Relatively few performance data are available in the literature for the MgO duct
injection process. Martin Marietta literature cites examples of high SO3 removal efficiency
(>80%) with near stoichiometric amounts of MgO being injected.4
The SBS InjectionTM process is patented by Codan Development LLC and available by license.
SBS solution, a byproduct from sodium-based FGD systems, or commercially available sodium
sulfite or bisulfite can be used as a feed material. SBS process performance data are available
from full-scale testing at Vectren Corporation’s A.B. Brown Station.5 High sulfuric acid removal
levels are possible (greater than 90%, or down to less than 2 ppmv at the ESP outlet) when
injecting at Na:SO3 mole ratios in the range of 1.5:1 to 2:1.
3
Alkali Injection into the Air Heater Outlet Duct
Two technologies involving alkali injection into the air heater outlet duct were evaluated in the
economic comparison, hydrated lime injection and ammonia injection. These, and all
technologies applied downstream of the air heater, offer no potential benefits to SCR low-load
turn down or plant heat rate.
EPRI evaluated duct injection of alkali powders such as hydrated lime at pilot scale at their
Environmental Control Technology Center (ECTC) in the early 1990s.6 Alkaline powders react
with flue gas sulfuric acid to form sulfate salts that are collected with the fly ash in the
particulate control device. EPRI found it was possible to achieve high sulfuric acid removal
percentages when injecting hydrated lime, sodium bicarbonate, or other dry alkaline powders
into the ductwork between the air heater and particulate control device, primarily a cold-side
ESP.
Ammonia injection has been tested by many and has been implemented at some plants for
sulfuric acid control. Ammonia can be injected between the air heater and cold-side ESP at
NH3:SO3 mole ratios in the range of 1.5:1 to 2:1 to achieve high sulfuric acid removal levels
(upwards of 95%). The biggest issue for ammonia injection is its effect on fly ash disposal/reuse.
Flue Gas Humidification, With or Without Alkali Injection
In this technology, also tested by EPRI at the ECTC, the flue gas is partially humidified and
cooled to below its acid dew point. The sulfuric acid appears to be removed by condensing onto
existing fly ash particles, which can be collected in a cold-side ESP or wet scrubber. Alkali can
also be injected, and serves several apparent purposes. One is to neutralize the sulfuric acid
droplets formed, so they are less corrosive to ductwork, ESP collecting plates, etc. Another
reason is to control the fly ash resistivity. When injecting upstream of a cold-side ESP, EPRI
found that humidification can result in low fly ash resistivity and a tendency for increased ash reentrainment emissions, which could be avoided by also injecting a calcium-based alkali. Of the
various possible configurations, humidification with separate hydrated lime powder injection
upstream of the ESP was selected for evaluation in the EPRI guide.
Wet ESP
Wet ESP technology has been in existence for nearly a century, but sulfuric acid control on coalfired power plants represents a relatively new application in the U.S. In this control technique,
sulfuric acid mist is collected by electrostatic forces in a wetted-plate ESP. The collected mist is
washed from the plates either periodically or continuously, and the blow down can go to an FGD
system or to a separate system for treatment. To achieve high sulfuric acid removal efficiencies
(>80%), the wet ESP could be installed downstream of the wet FGD absorber or could be
installed inside the absorber vessel as a replacement for the chevron-style mist eliminator. For
lower efficiencies (~60%) it is possible to convert the last field of a conventional dry ESP to wet
operation.7
4
By being installed at or near the end of the flue gas path, wet ESPs typically address only plume
opacity and particulate emissions, and provide no upstream benefits for air heater exit
temperatures or duct corrosion.
EXAMPLE PLANT FOR ECONOMIC STUDY
Cost estimates were prepared for applying a number of sulfuric acid control technologies to a
hypothetical or example power plant. The plant consists of a single, 500-MW unit located in the
Midwest, and firing a 3.5% sulfur bituminous coal. The plant is retrofitting an SCR system for
NOX control, and currently has a cold-side ESP for particulate control and a wet FGD system for
SO2 control. The FGD system uses limestone reagent and produces wallboard-grade gypsum,
which is sold as a byproduct. The fly ash is handled dry, and either sold as a byproduct or
disposed in a landfill. During the “ozone season,” all of the fly ash is typically sold. Table 1
summarizes the assumptions made for the hypothetical plant.
Table 1. Assumptions made for heat and material balances for hypothetical plant.
Parameter
Value
Unit Load (gross MW)
500
Gross Plant Heat Rate (Btu/hr/KW )
9200
Capacity Factor (%)
85
Flue Gas Flow Rate (acfm at economizer outlet)
2.07 x 106
Coal Sulfur Content (%)
3.5
Flue Gas SO2 Content (ppmv wet at economizer
2790
outlet)
Furnace Conversion of SO2 to SO3 (mole %)
1.0
Conversion of SO2 to SO3 Across SCR Catalyst
0.75
(mole %)
Stack Sulfuric Acid Concentration during SCR
21
Operation (ppmv)
NOx Season Duration (months/yr)
5
Target Stack Sulfuric Acid Concentration (ppmv, dry basis):
For lower SO3 removal percentage target
9.5 (return to pre-SCR conditions)
For higher SO3 removal percentage target
3.0 (assumed value corresponding to “clear” stack)
The plant has a wall-fired, pulverized coal boiler that typically converts 1% of the coal sulfur to
SO3. The SCR catalyst guarantee is for no more than 0.75% conversion of SO2 to SO3 across the
installed catalyst layers. Figure 2 summarizes the assumed baseline SO3/sulfuric acid
concentrations at various locations in the flue gas path, both with and without the SCR in line.
The concentrations in the figure assume approximately 6 ppmv of SO3/sulfuric acid removal
each across the air heater and ESP and 50% removal of sulfuric acid mist across the FGD
absorber for both cases.
5
Figure 2: Baseline and Post-SCR SO3/Sulfuric Acid Concentrations for Hypothetical Plant
9.5 ppm baseline
21 ppm w/SCR
Stack
SCR
53 ppm w/SCR
31 ppm
Economizer
Air Heater
Outlet
19 ppm baseline
42 ppm w/SCR
Absorber
ESP
25 ppm baseline
48 ppm w/SCR
ID/Booster Fans
METHODOLOGY FOR ECONOMIC CALCULATIONS
The objective of applying controls was to restore the stack sulfuric acid concentration to the preSCR value of 9.5 ppmv (dry basis) from the projected value of 21 ppmv with the SCR in service.
A sensitivity case also looked at costs for achieving a higher level of SO3 control, down to 3
ppmv (dry basis) or less at the stack. This lower concentration should be adequate to result in a
clear stack, or low stack plume opacity (at least due to the sulfuric acid mist contribution to
plume opacity) under most conditions. The SCR was assumed to operate only during the “ozone
season,” from May 1 through September 30, and it was assumed the sulfuric acid controls will
only be required to operate during that time period as well. Other sensitivity cases considered
both a new and retrofit plant installation where the SCR and SO3 controls operate year-round.
As a first step in developing cost estimates for the various potential sulfuric acid controls, heat
(enthalpy) and material balance calculations were conducted to estimate the expected SO3/
sulfuric acid control performance of each technology, and to estimate reagent and utility
consumption rates. Performance information for each technology was taken from data supplied
either by the vendor or by EPRI. The details of the heat and material balances are provided in the
EPRI SO3 Mitigation Guide Update.8
Cost estimates were not developed for ammonia injection at the lower SO3 removal percentage.
When the NH3:H2SO4 molar ratio is low, ammonia injection can cause operating problems via
the formation of a sticky ammonium bisulfate.
Two technologies, byproduct Mg injection and commercial Mg injection in the furnace, were not
estimated to be able to achieve the higher sulfuric acid removal percentage (down to 3 ppmv at
the stack) independently. Two U.S. power plants have operated or are planning to operate with a
combination of furnace Mg injection and hydrated lime injection upstream of the ESP to achieve
6
relatively high control efficiencies. Consequently, this combination of control technologies was
included for the economic evaluation for the higher removal percentage case.9
The performance estimates and heat and material balances formed the basis for the capital cost
estimates for eight of the ten technologies under consideration. Spreadsheets were then used to
develop an overall project cost estimate.
The two not estimated in this manner were wet ESP and fuel switch. Capital cost estimates were
developed for the wet ESP cases but were based solely on general budgetary cost factors. A
capital cost factor of $35/kW was used for a last-field dry ESP conversion for the lower SO3
removal case, and a factor of $55/kW for standalone wet ESP retrofit downstream of the wet
FGD absorber for the higher removal percentage. The capital costs for a wet ESP retrofit can be
very site specific, depending on the difficulty of the retrofit and materials of construction
selection, and could vary by ±50% or greater. No capital cost estimates were made for fuel
switching because the capital cost requirements for this technology can be very site specific.
The reagent and utility consumption estimates and the amortized capital cost estimates were
combined to generate annual cost estimates for each of the technologies. These estimates were
not “levelized” to reflect escalation and discount rates, but instead represent first-year costs. Nor
do they include operating labor, nor maintenance labor and materials. For most of the
technologies, there is not a good experience basis for estimating these values, so such an estimate
would be based on a percentage of capital, reagent, and/or utility costs, and would not help
differentiate the projected costs of the technologies.
The factors used to develop costs are shown in Table 2. The reagent costs are based either on
f.o.b. quotes from vendors and an assumed shipping distance at an average shipping cost, as
shown in the table, or typical delivered costs for the Midwest. Plant water, softened water,
auxiliary power and fuel costs are also “typical” values taken from previous economic
evaluations conducted by URS for power plants east of the Mississippi River. The values for
gypsum byproducts, fly ash byproduct sales and incremental landfill disposal costs represent the
nominal averages (rounded to the nearest whole dollar) for power plants with medium- and highsulfur coal limestone forced oxidation FGD systems, as reported on their 2001 Form EIA-767.
Capital cost recovery values are based on an annual recovery factor of 0.15, which nominally
represents an 8% discount rate and a 10-year book life.
BASELINE CASE ECONOMIC RESULTS
The first year cost estimates are summarized in Figure 3 for the lower removal target. Because
the cost estimated for ammonia injection was relatively low, ammonia injection was included in
Figure 3 even though it was estimated for achieving the higher removal percentage target.
7
Table 2. Factors used to generate annual reagent and utility costs.
Factor
OmniClearTM fuel additive, 75% minus 200 mesh, delivered from Omni Materials ($/ton)
Byproduct Mg(OH)2 slurry, delivered from Midwest supplier ($/dry ton of pure Mg(OH)2,
shipped at 18% solids, 65% purity in solids, assumed 50-mile delivery distance)
Commercial Mg(OH)2 slurry, delivered from Manistee, MI ($/dry ton Mg(OH)2, shipped at
58 wt% solids, 100% purity in solids, assumed 600-mile delivery distance)
Utilimag 40 MgO powder, delivered from Manistee, MI ($/dry ton MgO, assumed 600mile delivery distance)
Sodium Sulfite, delivered from Green River, WY ($/dry ton available Na as Na2SO3)
Ammonia, delivered from existing plant system ($/ton)
Hydrated Lime, delivered from Midwest location ($/ton)
Truck Transit Costs ($/ton-mile)
Plant Water Cost ($/1000 gal)
Plant Softened Water Cost ($/1000 gal)
Plant Auxiliary Power ($/kwh)
Plant Fuel Costs ($/MM Btu)
Plant Low-sulfur Fuel Cost ($/MM Btu)
Gypsum Byproduct Value ($/wet ton, f.o.b. plant)
Fly Ash Sales Value ($/ton, f.o.b. plant)
Incremental Landfill Disposal Costs ($/ ton)
Annual Capital Recovery Factor
Value Used
100
151
334
472
300
300
75
0.12
0.40
2.30
0.032
1.04
1.24
5.00
3. 00
4. 00
0.15
Figure 3: Summary of First-year Capital Recovery and Non-labor Operating Costs by
SO3/Sulfuric Acid Control Technology for Achieving the Lower Removal Percentage Target
2500
2,745
Capital Recovery
Reagent and Utility Costs
1500
1000
500
ES
P
W
et
e
yd
.
H
/L
i
m
id
.
Hu
Li
m
m
e
3
H
N
SB
S
M
gO
M
g
By
p.
om
m
.M
g
C
Fu
el
lB
le
Ad
di
ti
ve
nd
0
Fu
e
Annual Cost ($1000)
2000
8
For the lower sulfuric acid control percentage plot, five control options had projected reagent,
utility, and capital recovery costs in the range of $460,000 to $670,000 annually. The five
include the fuel additive, furnace injection of byproduct Mg, MgO powder injection, SBS
InjectionTM, and ammonia injection (at the higher removal level).
The differences between the projected costs of these four control options are not great relative to
the level of uncertainty for these estimates. For example, if these annual cost estimates are
assumed to have an uncertainty of ±25%, the “error bars” (not shown in Figure 3) for these five
lower cost options would substantially overlap. This suggests that a more detailed analysis would
be required to determine the most cost effective control option for this application.
Three other control options were estimated to be somewhat more expensive, in the range of
$760,000 to $820,000 annually. These include low-sulfur coal blending, commercial Mg
injection in the furnace, and humidification with hydrated lime injection between the air heater
and ESP. Recall that the costs for fuel blending with low-sulfur coal disregard any potential
capital modifications that may be required by the plant. Significant capital modification
requirements could make this option less attractive. Also, its ranking is based almost entirely on
an assumed value for the fuel cost differential. This cost differential could vary widely from the
assumed value of $0.20/MM Btu.
The remaining technologies for which costs were developed at the lower removal percentage
include dry hydrated lime powder injection upstream of the ESP and a wet ESP retrofit to the last
field of the existing dry ESP. These two options appear to be less cost effective, with annual
costs projected at $1.0 and $2.8 million, respectively. However, the wet ESP comparison is
skewed by the assumption that the control technologies will only operate five months out of the
year, and the relatively high capital recovery factor (10-year recovery). Both of these
assumptions favor low capital cost control options such as fuel blending or sorbent injection.
This effect is further discussed below.
The estimated costs for the higher sulfuric acid removal percentage are shown in Figure 4. These
estimates show ammonia injection as a potentially low cost technology, with projected first-year
costs of $640,000, followed by the OmniClearTM fuel additive at $790,000 annually. Three other
technologies, MgO powder injection, SBS injection and humidification with hydrated lime
injection, show somewhat higher annual projected costs between $920,000 and $980,000. Again,
if these annual cost estimates have an uncertainty of ±25%, the error bars for these five
technologies would overlap.
9
Figure 4: Summary of First-year Capital Recovery and Non-labor Operating Costs by
SO3/Sulfuric Acid Control Technology for Achieving the Higher Removal Percentage Target
Annual Cost ($1000)
2500
Capital Recovery
4,191
Reagent and Utility Costs
2000
1500
1000
500
ES
P
W
et
H
3
um
id
./L
H
yd
im
.L
e
im
e
+
Bi
ca
rb
N
H
S
SB
gO
M
e
yd
.L
C
om
m
.M
g
g
+
+
H
H
im
yd
.L
im
e
ve
dd
iti
By
p.
M
Fu
el
A
Fu
el
B
le
nd
0
Four other technologies show considerably higher first-year operating costs of $1.9 to $2.4
million. These include fuel switching; the two furnace injection technologies, which were
combined with hydrated lime injection between the air heater and ESP for this estimate; and
hydrated lime/sodium bicarbonate blend injection alone. Note that the latter three technologies
might adversely affect the base plant ESP performance due to the addition of relatively large
quantities (17 to 21% of the base fly ash loading) of particulate matter to the ESP inlet flue gas.
The wet ESP retrofit downstream of the wet FGD system shows the highest first-year cost of
over $4 million. However, as mentioned above, the wet ESP comparison is skewed by the
assumption that the control technologies will only operate five months out of the year and the
relatively high capital recovery factor (10-year recovery).
COST ESTIMATE SENSITIVITY CASES
The relative cost effectiveness of SO3/sulfuric acid control technologies is impacted by a number
of parameters. Comparison of the cost estimates presented in Figure 3 with those in Figure 4
illustrates the effect of one parameter, the required SO3/sulfuric acid removal percentage.
However, many of the assumptions made to develop these cost estimates also impact the results.
This subsection presents the results of sensitivity cases run to show how three parameters –
delivered sorbent costs (for sorbent additive/injection technologies), months per year of control
system operation, and new versus retrofit applications – impact the relative economics of SO3
control. Sensitivity cases for each of these parameters are presented and discussed below.
Of course, additional parameters impact the relative economics of SO3 control, such as the size
of the plant being controlled, what type of existing control device is used for particulate control,
10
whether the plant has an FGD system and what reagent is used, whether or not the plant sells its
fly ash and/or FGD byproduct, and many others. It is beyond the scope of the guide and this
paper to show cost sensitivities to all of these parameters, even though many of them can
significantly impact SO3 control economics. The reader can refer to the SO3 control cost
workbooks published in EPRI’s SO3 Mitigation Guide Update (Report 1004168) to develop
initial cost estimates that are more specific to a particular plant’s circumstances.
Effect of Delivered Reagent Cost
The estimated costs for most of the additive or injection technologies are largely driven by
delivered sorbent cost, as the capital costs for these technologies are relatively low (less than
$10/kW). There is uncertainty in the delivered costs of all of these sorbents. Since the sorbents
are typically available at a discrete number of locations and the distance between those locations
and the candidate plant could vary over a wide range, transportation costs could be quite
variable. Furthermore, transportation costs on a ton-mile basis will vary significantly depending
on whether truck, rail or barge delivery is employed.
The f.o.b. price of the sorbents could also vary significantly. The factors that drive the costs of
several potential sorbents are described below:
OmniClearTM fuel additive – This sorbent/fuel additive is currently supplied by only one vendor.
The f.o.b. price of this additive can therefore be influenced by market supply/demand pressures.
Byproduct Mg – This material is not sold commercially, so the market price per ton is not
established. Some plants would be able to produce this material internally from their FGD
byproduct, but many would have to purchase this material from a third party. Carmeuse has
estimated the production cost at $67/ton of pure magnesium hydroxide,9 but this does not include
capital recovery or any profit.
MgO Powder – The Utilimag 40 sorbent used as the basis for this evaluation is supplied by only
one vendor. The f.o.b. price can therefore be strongly influenced by market supply/demand
pressures.
SBS – Either sodium sulfite or sodium bisulfite, or a blend of the two, could be used as the
reagent for this technology. Either reagent could be procured as a byproduct or as a technical
grade chemical. This reagent could be shipped as a concentrated solution or as a dried powder.
Depending on which reagent source and form is employed, the f.o.b. and delivered cost of this
reagent could vary by hundreds of dollars per ton.
Ammonia – Both ammonia and urea are commodity chemicals used in high volumes in the U.S.
The costs for these reagents are driven by supply and demand pressures, and by the price of raw
materials, and can fluctuate over a wide range.
Hydrated Lime – Hydrated lime is available from a number of suppliers throughout the Midwest.
Depending on the plant size and hydrated lime requirements, and distances from suppliers, it
may be more cost effective to purchase quicklime and hydrate it on site. While a delivered price
11
of $75/ton was used as the basis for these cost estimates, quicklime can be purchased for about
$50/ton and contains about 30% more alkali on a molar basis, making it about half the cost of
hydrated lime on a molar basis. Plants that use lime or magnesium-enhanced lime as an FGD
reagent on site already may have a long-term contract for quicklime and have presumably
negotiated a competitive delivered price. However, using quicklime to produce hydrated lime on
site requires additional capital investment and operating and maintenance labor, which will offset
some of the savings.
Given these issues, it is quite possible that the costs of any of the sorbents considered could vary
over a range of ±50% from the values used as the basis for the cost estimates presented in
Figures 3 and 4. Since sorbent costs represent a majority of the injection technology costs, this
could greatly influence the annual reagent, utility, and capital cost recovery estimates. Figure 5
illustrates the effects of delivered sorbent cost, over the range of ±50% of the values presented in
Table 2, for the higher sulfuric acid removal percentage case.
For all of the sorbent injection technologies except ammonia injection and humidification with
hydrated lime injection, sorbent delivered cost variations produce a wide range in the technology
cost estimates. The ammonia injection costs are mostly driven by the assumption of lost fly ash
sales, and the humidification with lime injection costs are mostly driven by capital costs.
The results in Figure 5 show that among the five candidate sorbent injection technologies that
were estimated to have lower annual costs earlier in this paper, an increase in sorbent cost for
one coincident with a decrease for another could markedly change the ranking of injection
technology costs.
Figure 5: Effect of Sorbent Delivered Price Variations on Control Technology Cost Estimates
(includes reagent, utility and capital recovery costs)
3500
-50% Reagent Cost
Estimated Reagent Cost
+50% Reagent Cost
2500
2000
1500
1000
500
Bi
ca
rb
m
e
3
H
H
yd
.
H
Li
m
e
+
um
id
./L
i
N
SB
S
gO
M
im
e
H
yd
.L
m
.M
g
om
C
By
p.
M
g
+
+
H
yd
.
lA
dd
Li
itiv
e
m
e
0
Fu
e
Annual Cost ($1000)
3000
12
These results illustrates that, before selecting an SO3 control technology for a particular plant, it
would be best to have the delivered sorbent cost established to a relatively high confidence level.
Effect of Control Technology Operating Period
The control technology cost estimates presented earlier in this section were based on the
assumption that the SCR on the example plant would operate only during the 5-month ozone
season, and that the control technology would operate only during that period as well. However,
some plants will have to operate their SCR year-round, and/or will have to operate their SO3
control technology year-round once installed. A change to year-round operation will improve the
relative economics of technologies whose costs are driven mostly by capital recovery (e.g., wet
ESP) and could erode the economics of technologies whose costs are driven mostly by reagent
and/or utility costs (e.g., fuel switch/blend or fuel additive).
Two simplifying assumptions were made in generating the 12-month per year costs. One is that
the SCR operates year-round. The analysis would be more complex if the SCR operated 5
months while the control technology operated 12 months, as there would be different baseline
SO3 concentrations and different SO3 removal percentage requirements for the two periods of the
year. The second simplifying assumption was that the plant’s fly ash is sold year-round. Again,
the analysis would be more complex if the ash is only sold for part of the year, particularly for
the ammonia injection case where the cost estimate is largely driven by the impacts of lost fly
ash sales.
The results from Figure 4 are incorporated into Figure 6 to show the effect of 12-month versus 5month operation on the relative economics of the SO3 control technologies at the higher removal
percentage. As expected, the relative ranking of the fuel switch/blend case erodes for 12-months
operation and the relative ranking of the wet ESP case improves. For the five injection
technologies that were estimated at less than $1 million annually for 5-months operation,
ammonia injection remains the lowest estimated cost option (notwithstanding other balance-ofplant issues) but the relative rankings of the other four technologies reverse. For example, the
more capital intensive humidification with lime injection process goes from being the most
costly of the five to being the second lowest (behind ammonia injection) for 12-months
operation. Conversely, the fuel additive process goes from being the second least costly to the
most costly of the five at 12-months operation.
13
Figure 6: Comparison of First-year Capital Recovery and Non-labor Operating Costs by
SO3/Sulfuric Acid Control Technology for Achieving the Higher Removal Percentage Target (5month vs. 12-month operating period)
Annual Costs ($1000)
6000
5-mos. Operation
12-mos. Operation
5000
4000
3000
2000
1000
et
W
+
H
yd
.
Li
m
e
ES
P
rb
Bi
ca
im
e
./L
H
um
id
N
H3
SB
S
M
gO
m
e
H
+
om
m
.M
g
C
By
p.
M
g
+
H
yd
.
yd
.
Li
ve
Ad
d
iti
nd
Fu
el
lB
le
Fu
e
Li
m
e
0
Effects of New Versus Retrofit Installation
Another assumption used for this evaluation that impacts the relative economics of the candidate
SO3 control technologies is the assumption that the controls are being retrofit to an existing
scrubbed plant. If it is instead a new plant, there are at least two impacts on the cost estimates.
One is that a longer plant life would be assumed and hence a lower capital recovery factor would
be applied. The second is that, for the wet ESP case, the wet ESP design and implementation
could be integrated with the wet FGD absorber, possibly lowering the capital cost. For the other
technologies, the capital costs would not likely change as significantly for a new plant.
Figure 7 illustrates first-year annual cost estimates for candidate SO3 control technologies for the
higher SO3 removal percentage and for 12-months operation (the most likely scenario for a new
plant) with two changes from the estimates shown in Figure 4. One is that the capital recovery
factor has been lowered to 0.10 from 0.15 (same discount rate, 20-year versus 10-year recovery
period), and the capital costs for the wet ESP case have been reduced to $35/kW from $55/kW,
reflecting the wet ESP being integrated into the FGD module design. 10,11
14
Figure 7: Summary of First-year Capital Recovery and Non-labor Operating Costs by
SO3/Sulfuric Acid Control Technology for Achieving the Higher Removal Percentage Target
(New plant, 12-month operating period).
5000
5640
Annual Cost ($1000)
4500
Capital Recovery
Reagent and Utility Costs
4000
3500
3000
2500
2000
1500
1000
500
SP
W
et
E
SB
S
gO
M
N
H
3
H
um
id
./L
H
yd
im
.L
e
im
e
+
Bi
ca
rb
Fu
el
Bl
en
d
Fu
el
Ad
By
di
p.
tiv
M
e
g
+
C
H
om
yd
.L
m
.M
im
e
g
+
H
yd
.L
im
e
0
Figure 8 compares annual cost estimates for all three estimate bases (5-months operation, 12months operation, and 12-months operation for a new plant). The estimates illustrated in Figures
7 and 8 show that the relative economics of the wet ESP case greatly improve relative to the
sorbent injection technologies. Furthermore, the wet ESP offers two advantages relative to
sorbent injection. One is that the wet ESP can also lower the emissions of trace metals (e.g.,
beryllium, lead) present in the fine particulate emitted from the dry ESP or baghouse, which
could help in the permitting of the new coal plant. The second is that the primary cost driver for
the wet ESP case, capital recovery, would be fixed at the outset whereas the primary cost driver
for sorbent injection, delivered sorbent costs, could vary significantly over the life of the plant.
On the other hand, the sorbent injection technologies that remove SO3 from the flue gas upstream
of the air heater could be exploited to further improve the plant heat rate and to lower costs
associated with corrosion repair in the back end of the plant. Neither of these factors has been
considered here.
15
Figure 8: Comparison of First-year Capital Recovery and Non-labor Operating Costs by
SO3/Sulfuric Acid Control Technology for Achieving the Higher Removal Percentage Target
Annual Cost ($1000)
6000
5-mos. Operation
12-mos. Operation
12-mos. Operation, New Plant
5000
4000
3000
2000
1000
+
Li
m
e
ES
P
et
./L
i
um
id
H
yd
.
H
W
Bi
ca
rb
m
e
3
H
N
SB
S
gO
M
C
om
m
.M
g
M
g
+
+
H
yd
.
Hy
d.
L
Li
im
e
m
e
ve
iti
lA
dd
By
p.
Fu
e
Fu
el
Bl
en
d
0
SO3 CONTROL PROCESS ECONOMICS WORKBOOK OVERVIEW
The example economics shown above illustrate that the most cost effective SO3 control
technology can be very site specific, depending on the plant location, existing environmental
controls, disposal or reuse practices for coal combustion byproducts, utility-specific capital
recovery factors, etc. A workbook was designed to enable the user to perform a simplified,
preliminary investigation of the costs for implementing these technologies at the user’s coal-fired
plant. The workbook is part of the EPRI SO3 Mitigation Guide Update (TR-1004168). The
purpose of the workbook is to allow the reader to calculate the approximate capital and annual
reagent and utility costs associated with each of the possible SO3 control options for a particular
plant. The control options that are covered by this workbook include wet ESP, fuel blending, fuel
additives, and a number of sorbent injection technologies.
The workbook is oriented towards plants that fire eastern bituminous coals, because eastern
bituminous coals typically have higher sulfur content than western coals and there is little data
available for other coal types. The workbook considers the effect of Powder River Basin
subbituminous coal, but only as a low-sulfur blending coal, in the context of being implemented
as a mitigation technology. It is possible that some plants that fire Texas lignite could have high
enough flue gas SO3/sulfuric acid concentrations to cause plume opacity issues, as some could be
considered medium- to high-sulfur coals on the basis of pounds of SO2 produced per million Btu
of heat input. Consequently, the workbook also includes some calculations oriented towards
Texas lignite. However, the user is cautioned that no data were available for evaluating SO3
control technologies for Texas lignite applications, so there are some technical risks associated
with using the workbook for a Texas-lignite-fired plant.
Completion of the workbook involves three main steps:
16
Step 1: Enter specific data about plant operations.
Step 2: Estimate (or enter actual values for) flue gas flow rate, SO3 profile across the flue gas
path, and a target stack SO3 concentration.
Step 3: Choose a control option and complete its corresponding worksheet. Sorbent injection
technologies and fuel additives are covered by one worksheet, fuel switching by another, and wet
ESP by a third worksheet. Completing the specified worksheet will estimate capital and
operating costs for the selected technology.
Note that the calculations in the worksheets have been simplified from those used to generate the
relative economics presented in this paper. Even the calculations in this paper do not include all
of the details that should be considered in a site-specific engineering evaluation of control
technologies. The results from these worksheets are meant to develop a relative ranking of
control options for a particular plant, based on a limited set of variables that have been identified
as significantly impacting control costs. The initial estimates developed using these worksheets
should be useful for selecting a subset of control options that appear to be more cost effective for
a particular plant.
The use of these worksheets to develop initial estimates does not preclude the need to conduct a
thorough engineering evaluation of control options to make a final selection of a control
technology for implementation. Such an evaluation should involve developing site-specific
capital and delivered reagent cost estimates, and might involve applying risk factors based on the
extent of full-scale, commercial experience for each technology considered and potential sorbent
cost volatility.
CONCLUSIONS
The example economics presented show that ammonia injection was projected to be a low cost
technology, but balance-of-plant impacts may preclude its use. OmniClearTM fuel additive, MgO
injection and SBS injection appear to be cost competitive technologies at either control level,
although more data are needed to support the performance estimates used for the OmniClearTM
additive and MgO powder injection at high control efficiencies. Humidification combined with
lime injection may also be a cost-competitive control approach, particularly if the two steps can
be combined by injecting lime or limestone slurry. Variations in delivered sorbent cost from
those assumed for this evaluation could significantly change the relative ranking of the cost
effectiveness of these SO3 control technologies.
Byproduct Mg injection in the furnace can be a cost-effective technology for restoring stack
sulfuric acid mist concentrations to pre-SCR levels. For the example plant, commercial Mg
injection in the furnace was slightly less cost effective. Neither furnace injection technology was
estimated to be capable of achieving the higher sulfuric acid removal percentage target. Although
it would be possible to combine furnace injection with another technology downstream to
achieve less than 3 ppmv of sulfuric acid mist at the stack, the economics of installing two
control technologies in series did not appear to be attractive for the set of assumptions made in
this study.
17
Fuel switching or blending with low-sulfur coal could be cost effective depending on the cost
differential between the current and low-sulfur coal, and assuming only minimal capital cost
implications. Finally, although wet ESP technology was evaluated at the highest cost in this
comparison, its economics look better if SO3 control is required year-round rather than just
during the ozone season. A longer capital recovery period and wet ESP integration with a new
wet scrubber also make the economics for the wet ESP cases more competitive with the low
capital cost, sorbent injection or additive alternatives.
The example economics presented here underline two important observations about SO3/sulfuric
acid controls. One is that there is no one “best” control option available. The cost effectiveness
of the various options depend on site specifics such as the location of the plant relative to reagent
sources, coal combustion byproduct reuse and disposal practices, the sulfuric acid mist control
level required, and the extent to which other particulate matter contributes to plume opacity.
Depending on these specifics, one technology may be favored over the others for a particular
plant situation. The cost estimating workbook associated with EPRI’s SO3 Mitigation Guide
Update provides the user with the ability to estimate relative costs for a specific plant’s
circumstances.
The second observation is that there are several candidate SO3/sulfuric acid control technologies
for which there is not adequate full-scale demonstration and test data to serve as a basis for
power generators to evaluate control technologies with confidence. For example, the fuel
additive, MgO powder injection and humidification/lime injection technologies (combined or
separately) show promise as being cost effective for some situations, but a well-documented,
full-scale demonstration of each of these technologies is needed.
Finally, it should be noted that the candidate SO3/sulfuric acid control technologies evaluated in
this section should not be considered an all-inclusive list; the technologies evaluated here were
limited to those for which the authors had performance and cost data available. The retrofit of
SCR to plants that fire high-sulfur coal has created a need for retrofit SO3/sulfuric acid controls,
and has spawned the development of new processes. Some processes were not included in this
evaluation because not enough information was available to allow their costs to be estimated
(e.g., condensing heat exchanger, and trona injection) and there may be others that were not
included merely because the authors of this report are not yet aware of them.
REFERENCES
1. “OmniClear, A new alternative for sulfur trioxide control.” Document provided by Omni
Materials, Inc., Maysville, KY, 2003.
2. Blythe, G., R. McMillan, J. Davis, M. Lamison, R. Rhudy, J. Beeghly, L. Benson, and E.
Goetz. “Furnace Injection of Alkaline Sorbents for Sulfuric Acid Control,” paper presented at
the US EPA/DOE/EPRI Combined Power Plant Air Pollutant Control Symposium, The Mega
Symposium, Chicago, IL, 20-23 August 2001.
18
3. Sulfuric Acid Removal Process Evaluation: Long-term Results. EPRI, Palo Alto, CA, U.S.
Department of Energy, National Energy Technology Laboratory, Pittsburgh, PA, TVA,
Chattanooga, TN, American Electric Power, Columbus, OH, and FirstEnergy, Shippingport,
PA, 1004165, 2002.
4. Martin Marietta Specialty Chemicals promotional sheet for Utilimag 40.
5. Results of Full-Scale Testing of Sodium Bisulfite Injection for Flue Gas Sulfuric Acid Control.
EPRI, 1004167, December 2002.
6. Peterson, J.R. A.F. Jones, F.B. Meserole, and R.G. Rhudy. “SO3 Removal from Flue Gas by
Sorbent Injection, EPRI HSTC Phase II Test Results,” paper presented at the SO2 Control
Symposium, Boston, MA, 24-27 August 1993.
7. Monroe, L.S. and K.M. Cushing. Wet Electrostatic Precipitator Addition to the Outlet of a
Marginal Dry ESP. Final Report for EPRI Contract WO3453-04. Palo Alto, CA. 1996.
8. SO3 Mitigation Guide Update. EPRI, 1004168, March 2004.
9. Benson, L.B., K.J. Smith, and R.A. Roden. “Control of Sulfur Dioxide and Sulfur Trioxide
Using By-Product of a Magnesium-Enhanced Lime FGD System,” paper presented at the
Institute of Clean Air Companies (ICAC) Forum 2003: Multi-Pollutant Emission Controls &
Strategies. Nashville, TN, 12-15 October 2003.
10. Reynolds, J., W. Buckley, and I. Ray. “Advances in wet electrostatic precipitation
technology at fossil fuel power plants for multi-pollutant control,” paper presented at the
Institute of Clean Air Companies (ICAC) Forum 2003: Multi-Pollutant Emission Controls &
Strategies. Nashville, TN, 12-15 October 2003.
11. Staehle, Richard C., Ronald J. Triscori, K. Sampath Kumar, Gary Ross, and Randy Cothron.
“Wet Electrostatic Precipitators for High Efficiency Control of Fine Particulates and Sulfuric
Acid Mist,” paper presented at the Institute of Clean Air Companies (ICAC) Forum 2003:
Multi-Pollutant Emission Controls & Strategies. Nashville, TN, 12-15 October 2003.
19