CD /V U.S. Environmental Protection Agency Industrial Environmental Research
EPA-600/7-78-030b
Iff ice of Research and Development Laboratory __ . Hfto
Research Triangle Park, North Carolina 27711 MaTCn 1978
FLUE GAS DESULFURIZATION
SYSTEMS:
DESIGN AND OPERATING
CONSIDERATIONS
Volume II. Technical Report
Interagency
Energy-Environment
Research and Development
Program Report
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments of, and development of, control technologies for energy
systems; and integrated assessments of a wide range of energy-related environ-
mental issues.
EPA REVIEW NOTICE
This report has been reviewed by the participating Federal Agencies, and approved
for publication. Approval does not signify that the contents necessarily reflect
the views and policies of the Government, nor does mention of trade names or
commercial products constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/7-78-030b
March 1978
FLUE GAS DESUFURIZATION SYSTEMS:
DESIGN AND OPERATING CONSIDERATIONS
Volume II. Technical Report
by
C. C. Leivo
Bechtel Corporation
50 Beale Street
San Francisco, California 94119
Contract No. 68-02-2616
Task 2
Program Element No. EHE624
EPA Project Officers:
John E. Williams and Kenneth R. Durkee
Industrial Environmental Research Laboratory Emission Standards and Engineering Division
Office of Energy, Minerals, and Industry Office of Air Quality Planning and Standards
Research Triangle Park, N.C. 27711 Research Triangle Park, N.C. 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
and Office of Air and Waste Management
Washington, D.C. 20460
-------
ABSTRACT
A brief introduction is given to flue gas desulfurization technology.
The SO- removal mechanisms for wet scrubbing and dry processes are
described, along with the typical gas contacting equipment used.
Processes and their critical operating parameters are described for:
Lime and limestone scrubbing
Sodium carbonate scrubbing
Double alkali (soda lime) scrubbing
Magnesium oxide scrubbing
Sodium sulfite (Wellman Lord) scrubbing
The report describes the implications of requiring 90 percent or greater
S09 removal for new FGD installations. For lime and limestone scrubbing
data are presented from the EPA Shawnee Test Facility to show the effects
of process operating parameters on SO removal efficiency. Mathematical
correlations of the data are also presented. Additional test data is
presented for sodium carbonate and double alkali scrubbing.
Coal characteristics and their effects on FGD system design and oper-
ation are summarized. A qualitative description is given for the effects
of coal heating value and sulfur, ash, moisture, and chlorine content.
The reasons and methods for reheating scrubbed flue gases are described.
Energy consumption of the various methods is defined, and problems
experienced with reheater design, operation, and maintenance are
presented. Alternative approaches to eliminating reheat are discussed.
11
-------
FOREWORD
EPA has undertaken a review of the New Source Performance Standards
for S0_ emissions from coal-fired steam generators. As part of this
review, EPA is assembling information related to the SOo removal
capability of FGD systems and technical and economic data for the
design of FGD systems for higher S02 removal efficiencies.
In the conduct of this review the EPA has contracted for the services
of a number of supporting organizations, each of which will make specific
contributions to the total subject. In this respect, a substantial num-
ber of separate, but related, issues will be investigated, including
such broad issues as economic, energy, social, technological, and
environmental issues.
This report addresses itself to the following specific tasks:
(1) Describe FGD system parameters which are commonly
monitored to ensure proper operation and explain how
those parameters are varied to accommodate combustion
system variations such as changing boiler loads, changes
in S0,j inlet concentrations, and changes in other fuel
characteristics. Describe the parameters which are
controlled to limit scale build-up in the scrubber
system and mist eliminator and the ranges within which
they must be controlled and parameters which are con-
trolled to limit other problems.
(2) Where an installed full-scale FGD system has demon-
strated good operation, but at lower SCL removal
efficiencies (such as 60 to 85 percent), define the
changes that would be applied to a new facility to
obtain 90 percent or greater S0? removal and present
data showing the effects of these changes. For example,
iii
-------
SO removal efficiency may be increased by increasing
the liquid-to-gas ratio, changing pH, addition of MgO
to lime/limestone scrubbers, changes in scrubber gas
velocity, increasing bed heights in absorption towers,
and others. Any rationale, engineering calculations,
etc., necessary to extrapolate from an existing FGD
system operating well at lower percent SCL removal to
a 90 percent or greater SO- removal should be presented.
(3) Define the effects of changes in coal characteristics
on FGD systems and describe operating parameters or
design techniques which can be used to compensate for
local and nationwide variations in coals. For example,
coals with high chloride contents affect S02 removal
efficiency, may corrode FGD constituents, and may
require the purge of additional waste streams. As
another example, describe the effects of different
coals on the resultant mixture of combustion gases
(e.g., percent oxygen in exhaust) and the resulting
effects on the chemical reactions and efficiency of the
scrubber. Define the scrubber system parameters
which are controlled to compensate for these variations.
(4) Describe the purpose, need, and methods for reheat of
exhaust gases downstream of FGD systems. Define
energy consumption for reheat of these gases for typical
plants ranging in size from 25 to 1000 megawatts for
the .different FGD systems. Describe problems exper-
ienced with reheaters and design and operating or main-
tenance factors used to satisfactorily overcome these
problems. Assess approaches which have been or could
be used to eliminate reheat of these gases.
As used herein, the term "S02 removal efficiency" is limited to per-
formance obtained from a single module of removal equipment over
a short time, and tested over a period of several weeks to show
that scale accumulation is not likely. The term does not reflect the
longer periods of time necessary for commercial operations, nor any
implication concerning the equipment reliability or availability.
Inasmuch as increasing the percent S0_ removal above currently demon-
strated levels will require greater chemical work, it will inherently
cause some reduction in the equipment reliability and availability
unless suitable counteracting steps (such as the use of more modules)
iv
-------
are taken. It will also somewhat increase the energy consumption
requirements, the chemical feed requirements, and the waste disposal
requirements over those currently being used. It is understood that
these specific subjects, their quantification, and potential solutions
are being addressed by others and that their results will be integrated
with the results reported herein. Thus, the results contained in this
report should not be construed as a complete assessment of all factors
related to the judgment as to the establishment of S0_ removal New
Source Performance Standards.
-------
CONTENTS
Section Page
1 SUMMARY 1-1
2 FGD PROCESSES 2-1
FGD Technology Summary 2-1
SO Removal Mechanisms 2-7
3 SELECTED FGD PROCESS DESCRIPTIONS AND CRITICAL
SYSTEM OPERATING PARAMETERS 3-1
Lime/Limestone Scrubbing 3-1
Sodium Carbonate Scrubbing 3-17
Double Alkali (Soda Lime) Scrubbing 3-20
Magnesium Oxide Scrubbing 3-23
Sodium Sulfite (Wellman Lord) Scrubbing 3-27
4 IMPLICATIONS OF REQUIRING 90 PERCENT OR
GREATER S02 REMOVAL FOR NEW FGD INSTALLATIONS 4-1
Lime and Limestone Scrubbing 4-1
Sodium Carbonate Scrubbing 4-24
Double Alkali Scrubbing 4-28
Magnesium Oxide Scrubbing 4-32
Sodium Sulfite (Wellman Lord) Scrubbing 4-33
5 EFFECT OF COAL PROPERTIES ON FGD SYSTEMS 5-1
Coal Heating Value and Moisture Content 5-6
Sulfur Content of Coal 5-7
Ash Content of Coal 5-10
Chlorine Content of Coal 5-12
Nitrogen Content of Coal 5-14
VI
-------
Section Page
6 REHEAT OF SCRUBBED FLUE GASES 6-1
Purpose and Need for Reheat 6-2
Methods of Reheating 6-9
Energy Consumption 6-16
Problems Experienced 6-19
Design, Operation, and Maintenance of Reheaters 6-30
Alternatives to Reheat 6-33
Reheat Summary 6-40
Appendix
A DATA FROM EPA ALKALI SCRUBBING TEST FACILITY A-l
B MATHEMATICAL MODELS AND NOMOGRAPHS FOR S02 REMOVAL BY
LIMESTONE AND LIME WET SCRUBBING B-l
C CONVERSION TABLE C-l
vii
-------
ILLUSTRATIONS
Figure Page
2-1 Summary of FGD Processes Tested on Coal-Fired Boilers 2-2
2-2 Station Electrical Loss as a Function of Draft
, Requirements 2-13
2-3 Station Electrical Loss as a Function of L/G Ratio
and Nozzle Pressure 2-13
3-1 Typical Process Flow Diagram for Lime/Limestone
Scrubbing 3-7
3-2 Operating Data from Shawnee Test Facility TCA with
Limestone 3-11
3-3 Operating Data from Shawnee Test Facility Venturi/
Spray Tower with Lime 3-12
3-4 Operating Data from Shawnee Test Facility Venturi/
Spray Tower with Lime and Magnesium 3-13
3-5 Simplified Process Diagram for Sodium Carbonate
Scrubbing System 3-18
3-6 Simplified Process Diagram for Double Alkali System 3-21
3-7 Simplified Process Diagram for Magnesium Oxide
Recovery System 3-24
3-8 Simplified Process Diagram for Wellman Lord
Recovery System 3-28
4-1 Effect of Inlet S0_ Concentration on SC>2 Removal
Efficiency for Fixed Design and Operations Conditions 4-4
4-2 Effect of Liquid-to-Gas Ratio on S02 Removal
Efficiency with Low Sulfur Coal at the Mohave
Power Station 4-6
4-3 Effect of Liquid-to-Gas Ratio on SO- Removal
Efficiency TCA with Limestone 4-7
4-4 Effect of Gas Velocity on S02 Removal Efficiency
TCA with Limestone 4-9
4-5 Effect of Scrubber Inlet pH on S02 Removal
Efficiency TCA with Limestone 4-11
viii
-------
Figure Page
4-6 Effect of Bed Height on SO- Removal Efficiency
TCA with Limestone 4-13
4-7 Effect of Liquid-to-Gas Ratio on SC>2 Removal
Efficiency TCA with Limestone and Magnesium 4-15
4-8 Effect of Scrubber Inlet pH on SC>2 Removal Effi-
ciency TCA with Limestone and Magnesium 4-16
4-9 Effect of Magnesium on SCU Removal Efficiency
TCA (no spheres) with Limestone 4-17
4-10 S02 Absorption Efficiency for Two Scrubbers in Series 4-19
4-11 Effect of Scrubber Inlet Liquor pH on S02 Removal
Efficiency Lime and Limestone Derived Slurries 4-21
4-12 Effect of Pressure Drop on S02 Removal Efficiency
Venturi with Sodium Carbonate (10 MW size) 4-25
4-13 Effect of Liquid-to-Gas Ratio on S02 Removal Effi-
ciency TCA (no spheres) with Sodium Carbonate
(10 MW size) 4-26
4-14 Effect of Liquid-to-Gas Ratio on S02 Removal Effi-
ciency Marble Bed Scrubber with Sodium Carbonate
(10 MW size) 4-27
4-15 Effect of Feed Stoichiometry on Removal Efficiency
in the Venturi/2-Tray Tower Absorber for the EPA-ADL
Double Alkali Pilot Program 4-30
4-16 Effect of pH on S02 Removal for the CEA/ADL Double
Alkali Prototype System 4-31
5-1 Coal Fields of the United States 5-4
6-1 Example of Plume Behavior for Reheat Alternatives 6-7
6-2 Reheat Energy Sources and Transfer Methods 6-10
6-3 Direct Combustion Reheat 6-12
6-4 Reheat with External Waste Energy 6-12
6-5 Direct Inline Reheat with Steam 6-13
6-6 Indirect Hot Air Reheat 6-13
6-7 Bypass Reheat Useful System Energy 6-14
JX
-------
Figure P,age
6-8 Bypass Reheat Waste System Energy 6-14
6-9 Theoretical Reheat Energy Requirements for Typical
Coal-Fired Boiler 6-18
6-10 Frequency of Occurrence of Visible Plumes with
Direct Reheat 6-37
6-11 Frequency of Occurrence of Visible Plumes with
Indirect Reheat 6-38
x
-------
TABLES
Table Page
2-1 Summary of FGD Processes on U.S. Coal-Fired
Utility Boilers 2-5
3-1 Operational FGD Systems on U.S. Coal-Fired
Utility Boilers 3-2
5-1 Typical Range of Coal Analyses 5-3
5-2 Examples of Coal and Flue Gas Compositions 5-5
6-1 Reheat Using Heat Exchangers 6-20
6-2 Reheat Using Combustion Gas 6-21
6-3 Direct Steam Reheat of Scrubbed Flue Gases 6-23
6-4 Direct Hot Water Reheat 6-26
6-5 Steam Heated Air Reheat 6-27
6-6 Gas and Oil-Fired Reheat 6-28
xi
-------
Section 1
SUMMARY
FGD PROCESSES
All flue gas desulfurization (FGD) installations operating on U.S.
coal-fired utility boilers are wet scrubbing processes that contact
the flue gas with a water solution of soluble (or sparingly soluble)
alkali to absorb the SC>2. The SC>2 removal is carried out in devices
(scrubbers) designed to provide intimate contact between gas and liquid
over a large gas-liquid interface area. The S02 removal efficiency of an
FGD system is a function of the type of scrubber selected, the absor-
bent composition and circulation rate to the scrubber, and additional
parameters.
Of the potentially available processes, lime and limestone slurry
scrubbing have received the most attention and are presently operating
on a number of coal-fired boilers. Additional new units are under
construction. These processes achieve high S02 removal efficiencies,
use naturally occurring raw materials, produce relatively insoluble
wastes, and have lower capital and operating costs than many other
processes. However, they require careful design to prevent problems
such as scaling and plugging in the scrubbers and .erosion and corrosion
of equipment.
Sodium carbonate scrubbing is a clear liquor absorbent process that
is less susceptible to scaling than calcium-based systems. This pro-
cess has had limited application because it consumes an expensive
alkali and produces a soluble sodium waste product that must either be
1-1
-------
regenerated at a substantial cost or impounded in suitably lined reten-
tion ponds. To date, application has been limited to geographical areas
having a cheap source of low grade carbonate; these include three units
operating at a power station in Nevada and one under construction in
Wyoming. Both stations burn low sulfur coal and the FGD systems are
designed to realize about 90 percent removal efficiency.
The double alkali process retains the desirable characteristics of the
clear liquor sodium carbonate system. It minimizes the soluble waste
problem by reacting the scrubber effluent with lime or limestone to
precipitate an insoluble waste product and regenerate the costly sodium
absorbent. The process has achieved high SOo removal efficiencies on
industrial boilers and greater than 90 percent removal during prototype
testing on jp. coal-fired utility boiler. The first full-scale coal-
-: ''.**;,.,
fired utility-application is scheduled to start up in 1979. The system
is designed to meet performance guarantees of 95 percent removal on 5
percent sulfur coal. Construction of two additional units is planned.
The process operation is somewhat more complex than lime or limestone
scrubbing.
The magnesium oxide slurry scrubbing process features recovery of sul-
furic acid or sulfur. It has been demonstrated to be feasible on
full-scale coal-fired boilers and is capable of S02 removal efficiencies
greater than 90 percent, although the reliability of these installations
must be improved. A 120 MW unit is currently undergoing testing and
has been designed to achieve 90 percent SCL removal from 2.5 percent
sulfur coal. No additional systems are under construction.
The Wellman Lord process recovers either sulfuric acid or sulfur and
has demonstrated SC>2 removal efficiencies greater than 90 percent on
oil-fired boilers and industrial plants. The first coal-fired boiler
application, a 115 MW installation, is currently undergoing a compre-
hensive one-year demonstration program. The unit was designed to meet
1-2
-------
performance guarantees of 90 percent S02 removal from 3.5 percent sulfur
coal and has demonstrated greater efficiencies. Two additional Wellman
Lord installations are under construction. The process operation is
more complex and costly than lime and limestone scrubbing.
A number of FGD system parameters must be monitored and controlled to
ensure proper operation, to accommodate variations such as changing
boiler load and inlet SCL conditions, to ensure proper mist eliminator
operation, and to prevent scaling in scrubbers. These parameters
include a varying alkali feed rate to match inlet SCL and boiler load
changes, and a number of mist eliminator washing schemes and design
configurations to maintain trouble-free operation. Techniques employed
to prevent scaling typically include control of scrubber pH, adequate
hold tank residence time, control of suspended solids concentration in
slurry systems, and high liquid-to-gas ratios. Clear liquor scrubbing
processes such as the sodium carbonate, double alkali and Wellman Lord
are not susceptible to scaling problems to the extent that the lime
and limestone processes are.
IMPLICATIONS OF REQUIRING 90 PERCENT OR GREATER
S02 REMOVAL FOR NEW FGD INSTALLATIONS
For the FGD processes described in this report, S0» removal efficiencies
of 90 percent or greater as defined in the foreword to this report
are obtainable by suitably changing a number of the operating conditions.
For lime and limestone scrubbing, the process variables requiring change
may include scrubber type, liquid-to-gas ratio, gas velocity, scrubber
inlet pH, additives such as magnesium, and the number of scrubbing stages.
An FGD system is designed to achieve a specified SO- removal efficiency
for the most severe operating conditions anticipated, such as maximum
boiler load (equivalent to maximum gas volume to be treated) and maximum
sulfur content of the coal to be burned (equivalent to maximum SO- con-
centration in the gas). The removal efficiency is largely fixed by the
hardware and by the chosen level of operating parameters. Once these
1-3
-------
parameters are set, the system should achieve the design SC>2 removal at
maximum boiler load and SC>2 concentration. For lower inlet SC>2 con-
centration, performance is generally improved for lime and limestone
scrubbing. For some types of scrubbers, such as spray towers, effi-
ciency may be improved with a reduction in boiler load. For others,
such as tray towers, performance may be impaired and the FGD system
must be carefully designed to achieve the required removal efficiency
during boiler turndown.
For a given coal and FGD system, increasing the sulfur removal increases
the sulfur load on the system, and with it, the associated cost, com-
plexity, and energy consumption. On the other hand, for lime and lime-
stone scrubbing a higher SC>2 removal efficiency can normally be achieved
when using a low-sulfur coal, than when using a high-sulfur coal, with
substantially less complexity. The ease of achieving a specified SC^
removal efficiency depends not only upon the system or apparatus em-
ployed, but upon the coal sulfur content and heating value as well.
Raising the design efficiency of a lime or limestone FGD system from
60 or 85 percent to 90 percent or greater affects essentially all pro-
cess design parameters and equipment specifications. Problems related
to nonuniform gas and liquid distribution and scrubber turndown capa-
bilities are amplified as the design efficiency is increased. Bypass
reheat would be marginal or impossible for 90 percent S02 removal be-
cause all the flue gas must be treated.
EFFECT OF COAL PROPERTIES
The major coal properties affecting FGD system design and operation are
heating value, sulfur, ash, moisture, and chlorine content.
Lower heating value coals generally give rise to greater flue gas vol-
umes and SO emissions per megawatt of generated power. On a megawatt
basis, the FGD system is larger and capital and operating costs are
1-4
-------
higher for coal with lower heating value (for a given sulfur content
and absorption efficiency). A characteristic of lower heating value
coals and coals of high moisture content is a flue gas with a greater
proportion of water vapor. This leads to smaller amounts of water
evaporation in the scrubbers, which in turn leads to higher gas and
recirculating liquor temperatures. Elevated temperatures could affect
absorption efficiency, depending on the chemistry of the particular
process.
The sulfur content together with the allowable emission standards deter-
mines the required S0_ absorption rate. For a given absorption effi-
ciency, the sulfur content affects the entire FGD system design. The
high S0~ removal efficiency required for high sulfur coals leads to a
more complex and costly FGD system. Low sulfur coals tend to promote
higher sulfite oxidation, which may be desirable or undesirable depend-
ing on the process.
FGD processes using nonregenerable absorbents (lime, limestone, or
soda) can be used for combined removal of fly ash and S0~. The fly ash
is then disposed of together with the spent absorbent. However, it may
still be desirable to remove the fly ash upstream of the FGD system
since it can erode process equipment, interfere with process chemistry
(and cause scaling in the soda system), and contribute to the waste
volume. Some fly ash from low sulfur western coals has a high alkali
content that can be used in the absorption process.
The small amounts of chlorine in coal are converted to gaseous chloride
in the boiler. The chloride is absorbed from the gas by wet scrubbing
processes. Its presence provides the potential for chloride stress-
corrosion, requiring the use of high alloy equipment wherever rubber
or other protective coatings are not applicable. Alkali associated with
the chloride is lost as dissolved solids in the waste stream and, be-
cause of its solubility, the waste may require special disposal. Chlo-
ride may also affect system chemistry. Prescrubbing may be used to
absorb chlorides from the flue gas upstream of the FGD system.
1-5
-------
REHEAT OF SCRUBBED FLUE GAS
Flue gas normally enters a wet FGD system at a temperature between
120 and 150°C (250 and 300°F). When contacted with the scrubbing
liquor, the gas is cooled by evaporation to the water dew point of
about 50 C (125 F). Discharge of cool, wet gas to the stack can lead
to: condensation and the resulting acidic water corrosion and fly ash
fouling of ductwork, fans, and stacks; acid rain from the plume;
impaired plume rise and dispersion; and increased plume opacity.
Reheating the flue gas 15 to 30°C (about 30 to 50°F) or more is fre-
quently required to eliminate the above problems. Methods of reheating
include: steam or pressurized hot water heated tubular exchangers in
direct contact with the flue gas; steam heated finned exchangers used
to heat air for blending with the flue gas; blending hot combustion
gases from an oil or gas burner with the flue gas stream; and bypassing
a portion of the hotter untreated gas around the FGD system for mixing
with the scrubbed gas.
Reheat is undesirable because it consumes sizable amounts of energy.
Alternatives to reheat include tall stacks to enhance plume dispersion,
corrosion protection with high alloy steels or noncorrosive coatings
and linings, upstream warm gas fans, and efficient mist eliminators.
A number of FGD installations operate without reheat under "wet-stack"
conditions. Thus, reheat does not seem to be a universal requirement
for wet scrubbers.
1-6
-------
SECTION 2
FGD PROCESSES
FGD TECHNOLOGY SUMMARY
More than 50 processes are in various phases of development and
commercial application for the removal of S02 from industrial waste
gases and boiler flue gases. The processes differ in the manner in
which they remove S02 from the gas and in the physical form in which
the absorbed sulfur is discharged. Important classifications of
flue gas desulfurization (FGD) processes that have been tested on
coal-fired boilers are shown in Figure 2-1.
A major distinction is generally made between recovery processes,
which recover S02 in useful forms such as sulfuric acid or elemental
sulfur, and throwaway processes, which produce a solid or liquid
waste. In the throwaway case, the waste product must be stored in a
pond or treated and used for landfill.
Throwaway and recovery processes have been developed that operate in
either wet, dry, or semidry modes. Wet scrubbing processes contact
the flue gas with a large and generally recirculated flow of a water
solution or slurry. The scrubbing liquor absorbs the SO- and cools the
gas by evaporation of water to a temperature only slightly above the
water dew point. It may then be desirable to reheat the gas to mini-
mize condensation, fouling, and corrosion in ducts, fans, and stacks;
to avoid a visible plume; and to improve plume rise and dispersion.
Dry processes have an advantage in not requiring stack gas reheat
because the flue gas is not contacted by water. Use of a spray drier
2-1
-------
(1) FULL SCALE > 35 MW
(2) PROTOTYPE 10-35 MW
(3) PILOT PLANT 1-10 MW
MANY PROCESS VARIATIONS
AND SUPPLIERS
MANY PROCESS VARIATIONS
AND SUPPLIERS
1 LIMITED APPLICATION
DRY INJECTION WITH
1 SCRUBBING-ABANDONED
MANY SUPPLIERS
CEA/ADL, ZURN. FMC.
ENVIROTECH
SAARBERG-HOLTER
CHIYODACT-101
3 ATOMICS INTERNATIONAL
WHEELABRATOR- FRYE.
AMERICAN AIR FILTER
DRY INJECTION WITHOUT
SCRUBBING ABANDONED
CHEMICO/BASIC.
UNITED ENGINEERS
WELLMAN-LORD
CATALYTIWIFP.
BUREAU OF MINES.
3 PFIZER. PEABODY-MCKEE.
CHEMICO-STAUFFER
3 CONOCO COAL
ATOMICS INTERNATIONAL
8ERGBAU FORSCHUNG-
FOSTER WHEELER.
LURGI- SULFACID
UOP-SHELL,
B&W-ESSO
ABANDONED
Figure 2-1. Summary of FGD Processes Tested on Coal-Fired Boilers
2-2
-------
results in a semidry mode of operation in which the gas is contacted
by small quantities of aqueous absorbent. In this case, the absorbent
is dried, the flue gas is only partially cooled, and reheat can be
avoided.
Wet Processes
Some wet scrubbers are suitable for the removal of both fly ash and
SC>2 so that, for a new power plant, one of the options for the control
of these pollutants may be to let the FGD system remove both. In
practice, however, there are some valid reasons for collecting fly
ash separately with an electrostatic precipitator. For example,
possible interference with process chemistry is prevented, erosion of
process equipment by fly ash slurry is minimized, contribution to
additional throwaway sludge volume is avoided (although the fly ash
must still be disposed of), and, in the case of recovery processes,
contamination of by-products is prevented. Another consideration is
that if the power plant has an FGD system bypass and separate dust
removal equipment, particulate removal operation can be continued during
periods of FGD system maintenance (if such operation is allowed).
Wet FGD processes fall into two classifications - slurry scrubbing and
clear liquor. In slurry scrubbing the reactive absorbent and the
absorbent-SCL reaction products are present as suspended solid par-
ticles in water. During the absorption phase of slurry scrubbing,
the suspended alkali dissolves to react with the absorbed S02 and
to precipitate waste solids which are disposed of in throwaway pro-
cesses or treated to regenerate active alkali in recovery pro'cesses.
The circulation of saturated solutions and crystallization of reaction
products makes the control of process chemistry very important to
prevent scaling in the scrubber. Slurry system components are also
vulnerable to plugging and erosion. Clear liquor scrubbing processes
minimize the problems of the slurry systems by keeping the absorbent
and reaction products dissolved in water.
2-3
-------
Dry Processes
Dry FGD processes, in most cases, operate at low gas velocities, thereby
requiring large vessels, while wet systems operate at higher velocities
with smaller vessels. The dry carbon adsorption processes treat flue
gases at temperatures lower than 150°C (300°F), which means that they
can be located downstream from the power plant air heater and, therefore,
can be added to existing power stations. The copper-oxide and catalytic-
oxidation processes require flue gas temperatures above 370 C (700 F).
Therefore, they would require reheat as an add-on system, and their
full potential would be better realized by integrating them upstream
of the air heater in a new installation. Nahcolite injection can be
carried out at temperatures above or below 150 C (300 F), although
better efficiency is realized at higher temperatures.
Dry and Semidry Processes
The dry and semidry throwaway processes produce a dry waste product
that may be more easily disposed of than the sludge or moist waste
solids from wet throwaway processes.
Commercial Processes on U.S. Boilers
Table 2-1 summarizes by process the number and capacity of full scale
operational FGD systems and systems under construction on utility
boilers in the U.S. The lime and limestone processes have received
the most attention. These throwaway processes achieve high S02 re-
moval efficiencies and have capital and operating costs that are lower
than many other processes. On the other hand, they require careful
design to minimize or prevent problems with erosion, corrosion, sludge
disposal, and scaling and plugging in the scrubbers.
Alkaline fly ash scrubbing is of interest because it utilizes the
alkaline values in the fly ash of some western low sulfur coals, thereby
minimizing raw material requirements. Its application is, therefore,
limited to coals with the appropriate ash qualities.
2-4
-------
Table 2-1
SUMMARY OF FGD PROCESSES ON U.S. COAL-FIRED UTILITY BOILERS
Process
Limestone
Lime
Alkaline fly ash with
lime or limestone
Sodium carbonate
Magnesium oxide
Double alkali
Wellman Lord
Aqueous carbonate
Operational
No.
9
9
2
3
1
1
-
MW
3,767
2,702
720
375
120
-
115
Construction
No.
16
9
1
-
-
1
2
-
MW
6,750
4,158
450
-
-
575
715
-
Contract Awarded
No.
10
4
2
1
-
2
-
1
MW
4,911
2,350
1,400
509
-
527
-
100
Reference 1
June 1977
2-5
-------
Sodium carbonate scrubbing is a clear liquor absorbent process which
is less susceptible to scaling than calcium-based systems. This pro-
cess has had limited application because it consumes an expensive
alkali and produces a soluble sodium waste product which must either
be regenerated at substantial cost or impounded in suitably lined
retention ponds.
The double alkali process retains the desirable characteristics of the
clear liquor sodium carbonate system, but avoids the soluble waste
problem by reacting the scrubber effluent with lime or limestone
to precipitate an insoluble waste product and regenerate the costly
sodium absorbent. However, the process is more complex and requires
additional equipment. It has not yet been tested in full-scale
operation on a coal-fired boiler.
The magnesium oxide recovery process is less prone to scaling than
lime or limestone processes and features recovery of sulfuric acid
or sulfur. It has been demonstrated to be feasible on full-scale coal-
fired boilers, but the reliability of these installations must be improved.
The Wellman Lord system has been the most widely installed recovery
process for oil-fired boilers in Japan. It recovers either sulfuric
acid or sulfur and is currently undergoing full-scale demonstration on
a high-sulfur coal-fired boiler.
Modifications to existing processes and new processes with features
designed to overcome the drawbacks of current systems will continue
to be developed. In the U.S. full-scale coal-fired demonstration
units are planned for the semidry sodium carbonate recovery process and
for the clear liquor citrate recovery process. In Germany the Saarberg-
Holter clear liquor lime/chloride scrubbing process has been operating
successfully over 14,000 hours on a coal-fired boiler (Ref.2).
2-6
-------
SO REMOVAL MECHANISMS
Removal of SO from flue gas can be accomplished by absorption in
liquids and by sorption mechanisms on solids. In absorption, the SO
passes through the gas-liquid phase boundary to become distributed
throughout the liquid and undergo chemical reaction with the absorbent.
With the use of solid adsorbents, the SO may be retained on the sur-
face by mechanisms that are either physical or chemical in nature.
Wet Processes
In the absorption of S0? by a liquid, the SO- must diffuse out of the
gas phase and into the liquid phase. The driving force for the trans-
fer of SO- from the gas to liquid phase is provided by the concentra-
tion differences between the SO- in the flue gas and the equilibrium
vapor pressure, or back pressure, of the absorbed S0? in the liquid.
The SO , when absorbed in the liquid, has a finite back pressure that
increases in proportion to its concentration in the liquid. When
this back pressure equals the SO <
further absorption can take place.
this back pressure equals the SO concentration in the flue gas, no
In the selection of an absorbent for scrubbing flue gas, the object is
to find a liquid with capacity to absorb a large quantity of SO- with-
out building up an appreciable equilibrium back pressure. This is accom-
plished by using a solution with high soluble alkali content; e.g., lime,
limestone, sodium carbonate, sodium sulfite, magnesium sulfite. The wet
processes are designed to operate so that the SO back pressure over the
incoming absorbent is essentially zero (less than 1 ppm SO,,).
With clear liquor scrubbing, the absorbent and chemical reaction products
of absorbed SO with absorbent remain dissolved in water. In slurry
systems such as lime and limestone, however, the absorbents and reaction
products are of limited solubility, so that both are present as suspended
solids in the scrubbing slurry. During the absorption step of slurry
scrubbing, the suspended alkali dissolves in the scrubber to react with
the absorbed SO- and precipitate the reaction products.
2-7
-------
Design Parameters. The SO absorber has as its objective the intimate
contact between gas and liquid over a large gas-liquid interfacial sur-
*.
face area. Important design parameters in the selection of a scrubber
are:
Liquid-to-Gas Ratio. The liquid-to-gas ratio (L/G),
expressed as liters of liquid pumped through the
scrubber per cubic meter of gas passing through the
scrubber (l/m^)t Or gallons per thousand cubic feet
(gal/Mcf), is related to the gas-liquid interfacial
area in the scrubber and the quantity of absorbent
available for reaction with the SC^. The surface
area may be generated largely by production of liquid
spray through nozzles or by atomization by action of
the gas flow. The liquid-to-gas ratio strongly affects
the system energy losses. High L/G ratios introduce
major piping and structural design considerations
which are reflected in higher capital costs as
well
Gas Velocity. To minimize equipment cost, scrubbers
are designed to operate at maximum practicable gas
velocities, thereby minimizing vessel size. Maximum
velocities are dictated by gas-liquid distribution
characteristics and by the maximum allowable liquid
entrainment that the mist eliminator can handle. Gas
velocities may be 1.5 to 10 m/sec (5 to 30 ft/sec) in
tower-type scrubbers and more than 30 m/sec (100 ft/sec)
in the throat of a venturi scrubber. The lower the
velocity, the less the entrainment, but the more costly
the scrubber
Gas Turbulence. To promote rapid diffusion of S02
from the bulk gas to the liquid interface, a high
degree of turbulence is desirable. With a clear
liquor absorbent that is not limited by slow dissolu-
tion of suspended alkali particles, gas residence
times of a few hundredths of a second are adequate
to obtain high SCL absorption efficiencies such as
in a venturi scrubber
Gas Pressure Drop. The energy required to move the
flue gas through the scrubber is measured in terms of
gas pressure drop through the scrubber, expressed as
millimeters of mercury (mm Hg), or inches of water
(in. H20)
2-8
-------
Slurry Holdup. For slurry scrubbing processes, large
liquid holdup, or residence time, is desirable to
provide time for dissolution of alkali so that a maxi-
mum amount is utilized through the scrubber. Resi-
dence times in towers with packing may be 5 seconds
or more. Residence times in venturi scrubbers are a
few hundredths of a second, too small for high absorp-
tion efficiencies with lime or limestone slurry scrub-
bing (unless additives are provided see Section 4)
Gas Distribution. A major problem in the scaleup of
scrubbers from pilot to commercial size has been
maintaining uniform gas distribution. If the flow
is not uniform the scrubber will not operate at de-
sign SC>2 absorption efficiencies. In practice,
uniform flow has been difficult to achieve. Typi-
cally, turning vanes near the scrubber inlet duct
and compartmentalization have been employed.
Scrubber Internals. To promote maximum gas-liquid
surface area and liquor holdup a number of devices
are used. Common methods include stages of moving
spheres supported on grids, towers filled with pack-
ing, towers with closely spaced rods or grids, and
perforated trays. Fixed packings are generally not
used in slurry service because of a tendency to plug
or, in lime or limestone service, to scale. The
absorbent is often sprayed into the towers through
several stages of spray headers
Turndown. To follow changes in boiler load it is
necessary that the scrubber provide good gas-liquid
distribution, high liquid holdup for some processes,
and high gas-liquid interfacial area over varying
gas flow rates. Some scrubbers can be turned down
to 50 percent of design, while others must be divided
into sections that can be closed off. A venturi can
have a variable throat area to accommodate turndown.
In a big installation, individual modules can be
taken out of service
Gas-Liquid Flow Configuration. Scrubbers can be
counter-current (the gas flows upward and the liquor
downward), co-current (gas and liquor flow downward
together), or cross-current (gas flows generally
horizontally and liquor is sprayed at a right angle
2-9
-------
to the flow). Theoretically, counter-current or
cross-flow would be best because the scrubbed gas is
being continually contacted by the fresher absorbent,
but multiple spray stages in co-current flow can be
designed to achieve the same effect. The gas flow
configuration directly affects the mechanical layout
and duct work
Mist Elimination. A mist eliminator is installed in
the scrubber or downstream of the scrubber to sepa-
rate mist carried over in the cleaned gas. This is
done primarily to limit the emission of particulates
to the atmosphere; to prevent corrosion of downstream
ductwork, fans, and stack lining; to prevent plugging
of inline reheater tubes and solids deposition on fan
blades which could cause imbalance; and to reduce re-
heater energy requirements by removing water mist
from the flue gas stream
Wet Scrubbers. Leading scrubber types include the following:
Venturi. The venturi scrubber (Figure A-l) has been
used when both fly ash and sulfur dioxide must be re-
moved. The venturi has been able to remove fly ash
down to 0.05 gm/m (0.02 gr/scf) with pressure drops
of 20 to 30 mm Hg (10 to 15 inches H20) and liquid-to-
gas ratios of 1.5-4.0 l/m3 (10 to 30 gal/Mcf) for typi-
cal dust loadings and particle size distributions
from power plant stack gases. The venturi may contain
an adjustable throat area to permit control of pressure
drop over a wide range of flow conditions. The S0_
absorption efficiency is limited to 30 to 50 percent
per stage with lime-limestone because of the short
liquid residence time. It therefore requires additives
(such as MgO) or two stages of Venturis or an after-
absorber to achieve higher absorption efficiencies
than this with high sulfur fuels
Turbulent Contact Absorber (TCA). The TCA (Figure A-2)
is a counter-current multistage scrubber with screens
that both support and restrain the plastic spheres.
The spheres move in a turbulent fashion providing good
gas-liquid contact and scale removal. With additive-
free calcium systems, the number of stages are generally
between two and four for maximum SO absorption effi-
ciency with liquid-to-gas ratios of 6 to 8 1/m (40 to
60 gal/Mcf). The pressure drop per stage is approxi-
mately 4 to 5 mm Hg (2 to 2.5 inches HO)
2-10
-------
Spray Tower. The spray tower (Figure A-l) is a
counter-current type scrubber with a low gas pressure
drop. Absorbent is sprayed through several headers,
and liquid-to-gas ratios of 10 1/m (80 gal/Mcf) or
more are required for maximum S0~ absorption
with additive-free calcium systems
Packed Tower. The packed tower is a very efficient
type of scrubber, but has seen little use with slurry
scrubbing because of susceptibility to plugging.
With clear liquor scrubbing, high'absorption effi-
ciencies may be achieved with liquid-to-gas ratios
of 1.5 to 4 1/m (10 to 30 gal/Mcf)
Marble-Bed Absorber. The marble-bed absorber uses a
10-cm (4-inch) bed packing of glass spheres (marbles)
that are in slight vibratory motion. A turbulent
layer of liquid and gas above the glass spheres in-
creases mass transfer and particulate removal. Pres-
sure drop is generally 7.5 to 11 mm Hg (4«to 6 inches
HO). Liquid-to-gas ratios of 3 to 4 1/m (25 to 30
gal/Mcf) have been used with additive-free calcium
systems. Use of this device is no longer advocated
by its foremost developer because of scaling and dis-
tribution problems encountered over a number of years
Tray Column. The tray column offers high liquid hold-
up and absorption efficiency at relatively low pres-
sure drop. The main disadvantage is its vulnerability
to scaling. Liquid-to-gas ratios of 5 1/m (40 gal/Mcf)
are required for additive-free calcium systems. In
addition, undersprays are used to wash off soft scale
deposits, consuming further energy
Cross-Flow Absorber. The cross-flow absorber is in-
stalled in a horizontal position. It has a low pres-
sure drop and has been tested with packing or sprays.
It requires a high L/G ratio to realize high absorp-
tion efficiency with additive-free calcium systems
Screen or Grid Tower. The screen or grid scrubber
contains five to ten screens or grids (22 mm, or 7/8-
inch openings). A liquid-to-gas ratio of 7 or more
l/m3 (50 gal/mcf) is used with additive-free calcium
systems. Pressure drops are lower for this type of
absorber than for tray or packed tower types
2-11
-------
Venturi Rod Absorber. The venturi rod absorber uti-
lizes closely spaced parallel circular rods placed per-
pendicular to the gas flow in the absorber. The cir-
cular cross-section of the rods provides a converging-
diverging gas flow path (similar to a venturi). This
type of scrubber is marketed by Riley Stoker under the
trade name Ventri Rod.
Energy Requirements. Wet scrubbing systems have significant energy
requirements to overcome fan and pumping losses. Although there can
be considerable variation, Figures 2-2 and 2-3 (Ref. 3) show the order-
of-magnitude expected for typical scrubbers. Other energy losses in-
clude discharge pumps, small auxiliaries, lighting, instrumentation,
and additional process equipment, and also thermal energy consumed
by gas reheat (where used) and associated fan power to overcome added
pressure losses.
Wet Scrubber Efficiency.' For a given process, the design efficiency
is determined by knowledge of the system characteristics. Selection of
scrubber type can provide high or low liquid pumping or gas pressure
drop energy input, contacting area for the phases, and scrubber resi-
dence time. For example, a spray tower has low gas pressure drop and
moderate contact area and residence time, while a venturi has high
pressure drop and contact area, but very low residence time. In a
packed tower, increased packing (at the expense of higher pressure
drop) tends to increase SC>2 removal because of the increased gas-liquid
interfacial area provided. High L/G provides more liquid surface area
for absorption, and also provides a larger capacity for absorption.
Increased gas velocity can be detrimental to removal because of de-
creased gas-phase residence time and lower L/G for fixed liquor pumping
rate. However, these effects can be offset by the higher energy input,
which increases agitation and contact surface area for some types of
scrubbers (TCA and venturi scrubber, but not spray tower). The design
efficiency represents a trade-off between maximizing removal and mini-
mizing cost and, most importantly, must be consistent with demonstrated
high availability as well.
2-12
-------
Figure 2-2.
10 20 30 40 50 60
PRESSURE DROP, INCHES OF WATER
Station Electrical Loss as a Function
of Draft Requirements
L/G RATIO, GAL/1000 ACFM (INLET)
Figure 2-3. Station Electrical Loss as a Function
L/G Ratio and Nozzle Pressure
2-13
-------
Semidry Process
The use of a spray drier has been proposed (Ref. 4) for SO- liquid
absorption as a process whereby the flue gas is contacted by very small
quantities of sprayed sodium carbonate solution (L/G less than 0.13
1/m3, or 1 gal/Mcf). In a spray drier, droplets generated by a cen-
trifugal atomizer are much finer than, for example, those in a spray
tower. Because of the low L/G, there is not enough water to saturate the
flue gas. Only slight cooling takes place so that reheat is not needed.
Because all the water is evaporated, the spent reactants can be removed
as dry particles.
Dry Processes
The primary feature of dry FGD processes is removal of S09 without satura-
ting the flue gas or significantly lowering its temperature. Dry pro-
cesses involve several different mechanisms, including:
S0_-solid reaction with an alkali or metal oxide
dispersed on a high surface area carrier
Chemisorption with oxygen and water on the surface
of a catalytic solid
The ultimate disposal of the SO from a dry system may be by:
Discarding the reaction products
Regenerating the SO. acceptor solid by liberating a
concentrated stream of SO- for subsequent recovery
or processing
Nahcolite. Short-term pilot plant tests have been performed (Ref. 5)
in which a dry powdered alkali (sodium bicarbonate or nahcolite) was
injected directly into a slip stream of boiler flue gas. In this case
the S02 reacts directly with the nahcolite to form a dry sodium sulfate
product which can then be collected in a baghouse, or other particulate
collector, along with the fly ash. The process is somewhat more effi-
cient at higher temperatures (290 to 320°C, or 550 to 600°F), and up
to 90 percent S02 removal has been claimed.
2-14
-------
Activated Carbon Adsorption. A number of FGD processes employ the
concept of adsorption of the SO- onto the surface of a char or acti-
vated carbon. In this process, oxygen and water vapor are also
adsorbed onto the char, which catalyzes formation of sulfuric acid
from the SO , oxygen, and water vapor. The acid is adsorbed on the
carbon surface. When the charcoal bed becomes loaded with S0?, it
must be regenerated for further use. Suggested methods for regenera-
tion include:
Thermal Regeneration. The SO loaded char is heated
to about 650 C (1,200°F) at which temperature the
adsorbed acid in the pores reacts with the carbon
and is reduced to SO . A gas stream of concentrated
SO , CO , and water is exhausted from the generator
ana further treated to produce sulfur or sulfuric
acid. This is the method used in the Bergbau Forschung-
Foster Wheeler process (Ref.6)
Water Wash Regeneration. The charcoal bed is sprayed
with water, which removes the sulfuric acid formed in
the carbon pores. The dilute sulfuric acid is col-
lected and disposed of. This is the method piloted
in the Lurgi Sulfacid process and used in the Hitachi
process, which to date has been limited to oil-fired
boilers (Ref. 7)
Chemical Regeneration. In this method the loaded
carbon bed is brought into contact with hydrogen
sulfide to produce elemental sulfur. This is the
Westvaco process concept (Ref. 8)
The heart of the typical SO- adsorption process is the gas-solid
contactor. The objective is to maximize contact of gas and adsorbent
and minimize gas pressure drop, vessel size, and mechanical stress on
the adsorbent. The two common types of contactors are:
Moving-Bed Contactor. Charcoal is cycled continuously
through the adsorption vessel (usually counter-current)
and then through the regeneration vessel. The process
is continuous. This is typified by the Bergau
Forschung-Foster Wheeler method
2-15
-------
Fixed-Bed Contactor. The charcoal remains in place.
The bed receives gas until loaded with SO , after
which the flow is switched to a parallel contactor.
The idle charcoal bed is then regenerated and adsorbed
SO removed for further processing
A drawback of all carbon adsorption processes is the necessary low gas
velocity (0.3 to 0.6 m/sec, or 1 to 2 ft/sec) and, consequently, the
large vessels or vessel-ductwork combination required.
Copper Oxide. In the copper oxide process (Ref. 9), the flue gas is
blown through open passages parallel to a fixed bed of copper oxide
acceptor material at 370 to A30°C (700 to 800°F). The SO is removed
from the gas to form copper sulfate. When the acceptor is fully sul-
fated, the gas is switched to a parallel contacting vessel, and the
idle bed is contacted with a gas containing hydrogen. A concentrated
stream of SO,, is released for further processing. The copper sulfate
is reduced to copper and then reoxidized during the start of the next
adsorption cycle. This process has been applied to flue gas from
refinery heaters in Japan, but these are only small test installations.
Catalytic Oxidation. In the catalytic oxidation process, the flue gas
passes through a catalyst bed where the SO- is oxidized to S0«. The
SO^ reacts with water vapor in the flue gas to form sulfuric acid.
The gas is then cooled in a heat exchanger, and the sulfuric acid is
condensed in a packed absorbing tower. The'acid is collected and
disposed of (or marketed). This process has been installed at the
100 MW level in the U.S. (Ref. 10), but has been unable to achieve
sustained operation.
2-16
-------
REFERENCES FOR SECTION 2
1. Summary Report, Flue Gas Desulfurization Systems, May-June 1977,
Pedco Environmental Inc., 11499 Chester Road, Cincinnati, Ohio
45256.
2. Esche, M. , "Flue Gas Desulfurization and Stack Gas Purification
Using the Technology of the Saarberg Holter Process," 4th Inter-
national Congress on the Prevention of Atmospheric Pollution,
Tokyo, Japan, May 18-20, 1977.
3. Raben, I.A., "Status of Technology of Commercially Offered Lime
and Limestone Flue Gas Desulfurization Systems," EPA Flue Gas
Desulfurization Symposium, New Orleans, May 14-17, 1973.
4. Gehri, D. C. and R. D. Oldenkamp, "Status and Economics of the
Atomics International Aqueous Carbonate Flue Gas Desulfurization
Process," EPA Flue Gas Desulfurization Symposium, New Orleans,
March 1976.
5. "Evaluation of Fabric Filter as Chemical Contactor for Control of
Sulfur Dioxide from Flue Gas," Air Preheater Company, Inc., NAPCA
Report, December 31, 1969.
6. Strum, J. J., et. al., "BF Dry Adsorption System," EPA Flue Gas
Desulfurization Symposium, New Orleans, March 1976.
7. Slack, A. V., and G. A. Hollinden, Sulfur Dioxide Removal from
Waste Gases, Noyes Data Corporation, Park Ridge, New Jersey, 1975.
8. "Catalytic/Westvaco Desulfurization Process Prototype Demonstration
Program," EPA Flue Gas Desulfurization Symposium, New Orleans,
March 1976.
9. Pohlenz, J. B., "The Shell Flue Gas Desulfurization Process,"
EPA Flue Gas Desulfurization Symposium, Atlanta, November 1974.
10. Jamgochian, E.M., and W. E. Miller, "The Cat-Ox Demonstration
Program," EPA Flue Gas Desulfurization Symposium, Atlanta,
November 1974.
2-17
-------
Section 3
SELECTED FGD PROCESS DESCRIPTIONS AND
CRITICAL SYSTEM OPERATING PARAMETERS
This section presents brief descriptions of the following flue gas
desulfurization processes:
Lime/limestone scrubbing
0 Sodium carbonate scrubbing
Double alkali (soda-lime) scrubbing
Magnesium oxide scrubbing
Sodium sulfite (Wellman Lord) scrubbing
Of these processes, lime, limestone, sodium carbonate and Wellman Lord
systems are presently operating at full-scale coal-fired utility
boilers, and additional new units are under construction. The mag-
nesium oxide process is undergoing full-scale demonstration and a
full-scale double alkali system is also under construction. The opera-
tional FGD systems on U.S. coal-fired utility boilers are summarized
in Table 3-1.
LIME/LIMESTONE SCRUBBING
From the process standpoint, lime scrubbing and limestone scrubbing
are similar operations that differ both with respect to feed slurry
preparation and the internal chemistry. For a lime system, quick-lime,
CaO, is first slaked in water to form calcium hydroxide, Ca(OH)~.
For a limestone system, the raw calcium carbonate stone, CaCX^, is
finely ground in a wet mill to form a makeup slurry.
3-1
-------
Table 3-1. OPERATIONAL FGD SYSTEMS ON U.S. COAL-FIRED UTILITY
BOILERS (Sheet 1 of 4)
Plant Name and Company
Cholla 1
Arizona Public Service
Duck Creek 1A
Central Illinois Light
La Cygne 1
Kansas City Power & Light
Lawrence 4
Kansas Power & Light
Lawrence 5
Kansas Power & Light
Martin Lake 1
Texas Utilities
Sherburne 1
Northern States Power
Sherburne 2
Northern States Power
Southwest 1
Springfield City Utilities
Widows Creek 8
TVA
Will County 1
Commonwealth Edison
Bruce Mansfield 1
Pennsylvania Power
Cane Run 4
Louisville Gas & Electric
Conesville 5
Columbus & Southern Ohio Elec.
Process
Limestone
Limestone
Limestone
Limestone(c)
Limestone(c)
Limestone
Limestone
Limestone
Limestone
Limestone
Limestone
Lime with MgO
Lime
Lime with MgO
Vendor
Research Cottrell
Riley Stoker
Environeering
Babcock & Wilcox
Combustion Engineering
Combustion Engineering
Research Cottrell
Combustion Engineering
Combustion Engineering
UOP
TVA
Babcock & Wilcox
Chemico
American Air Filter
UOP
New or
Retrofit
R
N
N
R
N
N
N
N
N
R
R
N
R
N
Rating
(MW)
115
100
820
125
400
793
710
680
200
550
167
835
178
400
d ,ioval is about ~0/for two units.
(b) Shakedown operation only. . -nt will burn low sulfur -->al until 3 additional modules are built.
(r1 Converted from boiler injection to conventional scrubbingunit 4 in 1976, Unit 5 in 1977.
(d) Bypass reheat reduces net removal to about 60%.
3-2
-------
Table 3-1. OPERATIONAL FGD SYSTEMS ON U.S. COAL-FIRED UTILITY
BOILERS (Sheet 2 of 4)
% Sulfur
in Coal
0.4-1
2.5-3
3-6
0.5
0.5
1-2
(lignite)
0.8
0.8
3.5
3.7
4
4.7
3.5-4
4.5-4.9
Fly Ash
Collector
ESP
-
ESP
-
-
ESP
ESP
ESP
-
ESP
ESP
SOz Absorber
A-Venturi and packed
tower
B-Venturi
Ventri rod absorber
Venturi and sieve tray
tower
Marble bed converted to
venturi rod/spray
tower in 1976
Marble bed being
converted to venturi
rod/spray tower
in 1977
Spray quencher/wetted
film contactor
Venturi rod/marble bed
Venturi rod/marble bed
TCA
Venturi/spray tower
Venturi/two-stage
perforated plate
tower
2-stage venturi
Mobile bed
TCA
No. of
Modules
2
1
7
2
2
6
12
12
2
4
2
6
2
2
L/G
(gal/Mcf)
20&25
20
50
33
80
27
-
60
50
34
-
30-40
60
Design
Efficiency
%
> 90(a)
45
75-85
60-80
75-85
97{d)
50-60
50-60
80-87
80
85
92
90
90
Startup
10/73
9/76(b)
2/73
12/68
11/71
Mil
3/76
4/77
3/77
5/77
2/72
4/76
8/76
3/77
Reference
1.2
1,3
1,3
1
1
2
1.3
1
1,4
1
3
1
1,3
1.4
3-3
-------
Table 3-1. OPERATIONAL FGD SYSTEMS ON U.S. COAL-FIRED UTILITY
BOILERS (Sheet 3 of 4)
Plant Name and Company
Elrama
Duquesne Light
Green River 1,2,3
Kentucky Utilities
Hawthorn 3
Kansas City Power & Light
Hawthorn 4
Kansas City Power & Light
Paddys Run 6
Louisville Gas & Electric
Phillips
Duquesne Light
EPA Shawnee Test Facility
Tennessee Valley Authority
Colstrip 1
Montana Power
Colstrip 2
Montana Power
Reid Gardner 1
Nevada Power
Reid Gardner 2
Nevada Power
Reid Gardner 3
Nevada Power
Eddy stone 1A
Philadelphia Electric
D.H. Mitchell II
Northern Indiana Public Service
Process
Lime with MgO
Lime
Lime(a)
Lime(a)
Lime (Carbide)
Lime with MgO
Lime and
limestone
Alkaline fly ash
with lime
Alkaline fly ash
with lime
Sodium carbonate
Sodium carbonate
Sodium carbonate
Magnesium oxide
Sodium sulfite
(Wellman Lord)
Vendor
Chemico
American Air Filter
Combustion Engineering
Combustion Engineering
Combustion Engineering
Chemico
Chemico
UOP
ADL/CEA
ADL/CEA
ADL/CEA
ADL/CEA
ADL/CEA
United Engineers
Peco
Davy Powergas Allied
Chemical
New or Rating
Retrofit (MW)
R 510
R 64
R 140
R 100
R 65
R 410
R 10
10
N 360
N 360
R 125
R 125
N 125
R 120
R 115
(a) Converted from limestone boiler injection to conventional scrubbing in 1977
(b) Tests have been made with and without ESP
3-4
-------
Table 3-1. OPERATIONAL FGD SYSTEMS ON U.S. COAL-FIRED UTILITY
BOILERS (Sheet 4 of 4)
% Sulfur
in Coal
1-2.8
3.8-4
0.5-3.5
0.5-3.5
3.5-4
1-2.8
3-5
0.8
0.8
0.5-1
0.5-1
0.5-1
2.5
3.2-3.5
Fly Ash
Collector
ESP
Mechanical
-
-
ESP
ESP
(b)
-
-
Mechanical
Mechanical
Mechanical
ESP/
Venturi
ESP/
Venturi
SOz Absorber
Venturi
Venturi/mobile bed
Marble bed
Marble bed
Marble bed
Three 1 -stage venturi
trains
One 2-stage venturi
train
Venturi/spray tower
TCA
Venturi/spray tower
Venturi/spray tower
Venturi/sieve tray tower
Venturi/sieve tray tower
Venturi/sieve tray tower
Venturi rod absorber
Absorption tower
No. of
Modules
5
1
2
2
2
4
1
1
3
3
1
1
1
3
1
L/G
(gal/mcf)
30-35
35
-
-
-
30
Design
Efficiency
%
85
>80
70
70
80
83
90
Varies with
each test
15&18
15& 18
-
-
-
-
-
60
60
85
85
85
90
>90
Startup
10/75
9/75
11/72
8/72
4/73
7/73
4/72
10/75
7/76
4/74
4/74
7/76
9/75
8/77
Reference
3
1,3
1
1
1
1
1
1
1
1
1
1
1
5
3-5
-------
Process Description
A process flow diagram for a typical lime or limestone system is pre-
sented in Figure 3-1. Flue gas from the boiler enters the scrubber
where it is contacted by a recirculating slurry containing the sus-
pended alkali. The scrubbed gas leaves the absorber through a mist
eliminator. While passing through the scrubber the gas is cooled and
saturated with water vapor. The gas may have to be reheated to restore
buoyancy to the plume and to prevent water condensation or solids
deposition in the ductwork, fan, and stack (see Section 6).
During the absorption process the suspended alkali dissolves to react
with the absorbed SCL and precipitate calcium sulfite and sulfate
(gypsum) as solid waste. In Figure 3-1 the slurry from the scrubbers
drains to a hold tank where reactions go to completion. Makeup lime
is added to the tank and part of the slurry recycles back to the
scrubbers. If fly ash is not collected upstream, it will be captured
by the scrubber liquor. Typically, the continuously recycled scrubber
slurry may contain from 5 to 15 percent suspended solids consisting of
fresh alkali, reacted waste products, and fly ash.
To remove solid waste products from the system, a bleed stream is with-
drawn from the recycle loop and routed to the solids separation equip-
ment. In Figure 3-1, the bleed is routed to a clarifier (also called
a thickener or settling tank). The suspended solids settle to the
bottom of the tank, and clear liquor is drawn off the top and returned
to the scrubber loop. At the bottom of the clarifier, the solids may
be concentrated to 20 to 40 percent. This underflow then goes to a
filter for further thickening, where the separated liquor is also
returned to the scrubber loop and the filter cake, consisting of 60
percent or more solids, is discharged to a waste disposal area. The
greater volume of clarifier underflow sludge may be routed directly
3-6
-------
SETTLING POND
Gas Stream
Liquor Stream
Figure 3-1. Typical Process Flow Diagram for Lime/Limestone Scrubbing
-------
to a pond, thereby eliminating the filter. The clarifier may also be
eliminated by sending the still greater volume of bleed stream directly
to a pond. As the solids settle to the bottom of the pond, the clear
liquor on top may be returned to the scrubber loop.
Makeup water is added to the system to replace the water evaporated
into the flue gas and the water entrained in the waste stream. The
water is added as mist eliminator wash, through pump seals, and
through the lime slaker or feed makeup.
Process Status and Background
Operational lime and limestone FGD systems on U.S. coal-fired utility
boilers are included in Table 3-1. In addition to the systems currently
operating, there are 25 additional lime and limestone units under con-
struction, totaling about 11,000 MW. Several full-scale operational
systems have also been terminated. These include (Ref.l):
Meremac 2, Union Electric 140 MW limestone injection/
wet scrubbing experimental installation abandoned
Four Corners 5A, Arizona Public Service 160 MW
limestone scrubbing test program completed and
system shut down
St. Glair 6, Detroit Edison 133 MW limestone scrub-
bing switched to low sulfur coal
Valmont 5, Colorado Public Service 50 MW limestone
scrubbing experimental continuation of future FGD
operations is contingent upon regulatory decisions
Mohave 1A, Southern California Edison 170 MW lime-
stone scrubbing test program completed and system
shut down
Mohave 2A, Southern California Edison 160 MW lime
scrubbing test program completed and system shut
down
3-8
-------
Because they use naturally occurring materials and produce relatively
insoluble wastes, and because their capital and operating costs are
not as high as many other processes, lime and limestone scrubbing
processes have received the most attention of all flue gas desulfur-
ization methods. Significant improvements have been made over the
past several years in the areas of reliability, variable load opera-
tion, system control, and sludge disposal techniques.
Process Design and Control Parameters
A number of system parameters must be monitored and controlled to ensure
proper operation of lime and limestone FGD systems, to accommodate vari-
ations such as changing boiler load and inlet SC>2 conditions, to ensure
proper mist eliminator operation, and to prevent scaling in the scrubber.
Changing Inlet SO^ and Boiler Load. An FGD system design goal is to
achieve a specified S02 removal efficiency for the most severe operating
conditions anticipated, such as maximum boiler load (equivalent to maxi-
mum gas volume to be treated) and maximum sulfur content of the coal
to be burned (equivalent to maximum SC>2 concentration in the gas). The
removal efficiency is largely fixed by the hardware, as described in
Section 2, and by the chosen level of operating parameters (see
Section 4). Once these parameters are set, the system should achieve
a certain minimum SO removal at maximum boiler load and S0_ concen-
tration and even operate at higher absorption efficiencies for less
severe conditions.
Examples that show the response characteristics of 10 MW scrubbers
under long-term testing at the EPA Alkali Scrubbing Test Facility
are helpful in interpreting the effect of these process changes.
3-9
-------
Figures 3-2, 3-3, and 3-4 are presented to illustrate operating char-
acteristics with limestone using a TCA and lime with a venturi and
spray tower in series.
Figure 3-2 (Reference 6) shows operation of a small TCA with limestone.
The gas flow rate was constant at 2.6 m/sec, or 8.6 ft/sec (22,600 Mm /hr,
f^
or 20,500 acfm at 300°F) and the L/G was fixed at 9.8 1/m (73 gal/mcf).
Inlet S02 concentration varied from about 2,000 to 4,500 ppm with
the normal variation of coal sulfur content. The S02 removal efficiency
was held between 75 and 88 percent by varying the limestone feed rate
to compensate for the varying inlet S02« Removal efficiency tended
to fall off during periods of high inlet S02. This effect is described
in Section 4.
Figure 3-3 (Reference 7) shows lime scrubbing operation for a small ven-
turi and spray tower in series. During this test the gas flow rate
through the scrubbers varied with the changing boiler load while the
liquor flow rate remained constant. During operation the gas rate
varied from 18,700 to 38,600 Nm3/hr, or 17,000 to 35,000 acfm (1.4 to
2.9 m/sec, or 4.5 to 9.4 ft/sec in the spray tower), while the liquor
rates were 2,270 1/min (600 gpm) to the venturi and 6,060 1/min (1,600
gpm) to the spray tower. The effect of this type of operation is that
when the gas flow is reduced the L/G increases (up to 16 1/m^, or 117
gal/mcf for the spray tower) and absorption efficiency is enhanced for
these types of scrubbers (reducing the gas velocity in other types of
scrubbers may cause the efficiency to deteriorate see Section 4).
During operation the inlet S02 concentration ranged from 1,000 to 3,500
ppm. Therefore, the most severe operating conditions the FGD system
had to deal with occurred when gas rate (boiler load) was at a maximum
and inlet S02 was at 3,500 ppm. In this situation the S02 absorption
efficiency was about 70 percent. During less severe conditions higher
absorption efficiencies were obtained, and ranged up to 98 percent at
low boiler load and low inlet S02 concentration. This example of opera-
tion with changing gas flow rate and constant liquor rate is typical for
3-10
-------
1.500
I 10/3 I 10/4 | 10/5 I 10/6 I 10/7 I 10/8 I 10/9 I 10/10 I 10/11 I 10/12 f 10/13 I 10/M I 10/15 I 10/16 I 10/17 I 10/18 I 10/19 I 10/20 I 10/21 !
CALENDAR DAY (1974)
Figure 3-2. Operating Data from Shawnee Test Facility
TCA with Limestone
3-11
-------
200 24O
TEST TIME, Houn
8/16 I 8/17 I 8/18 I 8/19 I 1/20 I 8/21 I 8/22 I 8/23 I 8/24 I 8/2S I 8/26 I 8/27 I 8/28 I 8/29 I 8/30 I 8/31 I 9/1 I 9/2 I a/3 I
CALENDAR DAY (1975)
9/4 I
Figure 3-3.
Operating Data from Shawnee Test Facility
Venturi/Spray Tower with Lime
3-12
-------
j BEGIN RUN 643 1A
END RUN 643 1A !
CC I
s!
100
90
N c
si
3.000
2.SOO
2.000
1.500
40
^
100
90
30
70
GO
75
70
6.5
6.0
5.5
5.0
3.500
3.000
2.500
2.0OO
80
120
ISO
280
320
360
400
440
1.500
200 240
TEST TIME. Hours
I 9/28 I 9/29 I 9/30 I 10/1 I 10/2 I 10/3 I 10/4 I 10/5 I 10/6 I 10/7 I 10/8 I 10/9 I 10/10 I 10/11 I 10/12 I 10/13 I 10/14 I 10/1S I 10/16 I
CALENDAR DAY (1976)
480
Figure 3-4.
Operating Data from Shawnee Test Facility
Venturi/Spray Tower with Lime and Magnesium
3-13
-------
installations using constant speed pumps. During operation the alkali
feed rate is continuously adjusted to compensate for the varying amounts
of SC>2 absorbed.
Figure 3-4 (Reference 8) again shows lime scrubbing operation for the
venturi and spray tower in series. The previous two figures have
shown that SOo absorption efficiency may vary over a wide range during
normal operation. In Figure 3-4 the gas rate is 38,600 Nm3/hr, or 35,000
acfm (2.9 m/sec, or 9.4 ft/sec in the spray tower) and the L/G's are
2.8 1/m3 (21 gal/mcf) for the venturi and 15 1/m3 (50 gal/mcf) for
the spray tower. Magnesium was added to enhance the S02 removal effi-
ciency (see Section 4). The inlet S02 ranged from 2,000 to 3,000 ppm
while absorption efficiency ranged from 98 to 95 percent.
Control of Mist Eliminator Fouling. A mist eliminator is installed in
the scrubber or downstream of the scrubber to catch mist carried over in
the cleaned gas. This is done primarily to:
Limit the emission of particulates to the atmosphere
Prevent corrosion of downstream ductwork, fans, and
stack lining
Prevent plugging of inline reheater tubes and solids
deposition on fan blades
Reduce reheater energy requirements by removing
water mist from the flue gas stream
The chevron mist eliminator is a commonly used style for FGD scrubbers.
This device consists of a set of baffles that force the gas to flow in
a zigzag pattern. As the gas passes through the mist eliminator, en-
trained mist is impacted on the blades and coalesces to form larger
droplets which drain into the scrubber or into a collection trough.
Other concepts have been developed including radial-vane, wet precipi-
tation sections, two stages of chevrons, and many proprietary varia-
tions of the basic chevron design. In FGD scrubbers the mist elimin-
ators have experienced slurry solids and chemical scale buildup in
3-14
-------
the narrow passages with resulting restriction of gas flow. Radial
vanes may impart an undesirable swirling motion to the gas because
of the shear effect. Many techniques have been employed to improve
mist collection and minimize operational problems. These are sum-
marized in Reference 3 and include:
Washing. The mist eliminator can be washed with a
spray of process makeup water or a mixture of make-
up water and clarified liquor. It may be washed
from one or both sides, and washing may be continuous
or intermittent.
Gas Velocity Reduction. The gas velocity may be reduced
to 2 to 3 m/sec (7 to 10 ft/sec) by expanding the width
of the mist eliminator section of the scrubber. This re-
duces the amount of entrained mist that reaches the
mist eliminator.
Wash Tray. Wash trays have been installed ahead of
the mist eliminator to reduce the entrained slurry
solids. The gas passes through clear liquor that is
circulated across the tray, while entrained slurry
solids tend to be captured.
Horizontal Gas Flow. If the gas flow is reoriented
to a horizontal direction, the mist eliminator can
have a vertical configuration. This may offer supe-
rior drainage and be adaptable to a separate wash-
water recycle loop.
Freeboard Distance. The distance the entrained mist must
travel after leaving the final absorption stage before
entering the mist eliminator can affect performance. At
greater distances less entrained mist tends to reach the
mist eliminator. About 1.2 meters (4 feet) is a minimum,
but up to 5 or 6 meters (about 15 or 20 feet) has been employed.
Slanting. A compromise between horizontal and verti-
cal mist eliminator configurations is to use a slanted
one in a vertical gas flow. The object is to improve
drainage and reduce the potential for plugging.
Mist Eliminator Design Factors. Many variations of
the basic chevron design are in existence. Factors
include the number of vanes (passes), spacing between
vanes, angle of vanes, and various proprietary design
details.
3-15
-------
Improved Limestone Utilization. During tests at the Shawnee
facility, a correlation was found between alkali utilization
and the accumulation of mud-type solids. Above about 85 per-
cent alkali utilization the mist eliminators were kept clean:.
with very infrequent (once per 8-hour shift) washing with
fresh makeup water. An entirely clean system was maintained
over an extended period using only 25 percent of the available
makeup water. This correlation was confirmed using three
different methods for improving alkali utilization, including
reduced scrubber liquor pH, operating with 3 hold tanks in
series, and during forced oxidation tests using staged scrub-
bing. The correlation was also confirmed at TVA's Colbert
pilot plant.
Control of Chemical Scaling. The formation of hard scale on scrubber
surfaces, particularly in the narrow passages of mist eliminators and on
any packing material used, has been one of the main problems experienced
with lime/limestone scrubbing. Calcium sulfite (CaSO-.1/2H20) is pre-
cipitated from the chemical reaction of absorbed SC>2 with the slurry
alkali. Calcium sulfate (gypsum, CaSO,.2H20) is also precipitated
when dissolved sulfite is oxidated by oxygen absorbed from the flue
gas. With lime, calcium carbonate (CaCCL) may be precipitated if the
pH is too high when carbon dioxide (COo) in the flue gas reacts with
the lime. Normally these reaction products crystallize on the suspended
solids in the recirculating slurry. However, prevention of crystalli-
zation, particularly of calcium sulfate and calcium sulfite, on surfaces
within the scrubber is a recognized problem in the design of lime and
limestone processes.
Normally, as the scrubber liquor becomes supersaturated, crystallization
will take place on the seed crystals present in the slurry. If the de-
gree of supersaturation is allowed to increase above about 1.3 to 1.4,
the precipitation will also occur on scrubber surfaces.
The techniques employed to prevent scaling typically include:
Control of pH. If a limestone system is operated at
pH1s above"5.8 to 6.0, or if a lime system is oper-
ated above 8.0 to 9.0, there is a risk of sulfite
3-16
-------
scaling. The pH is controlled by adjusting the feed
stoichiometry. Automatic control of feed by online
pH sensors has been successful for lime systems but
has generally proved impractical for limestone instal-
lations due to insensitive pH changes with limestone
feed rate under the normal operating pH range. However,
in the limestone system the feed can be automatically
controlled by variations in flue gas flow rate and in-
let S0? concentration.
Hold Tank Residence Time. By providing retention time
in the scrubber hold tank, the supersaturation of
the liquor can be decreased before recycle to the
scrubber. Typical residence times of 5 to 15 minutes
have been demonstrated at Shawnee and used in some
full-scale systems.
Control of Suspended Solids Concentration. The degree
of supersaturation can be minimized by keeping an ade-
quate supply of seed crystals in the scrubber slurry.
Typical levels in newer installations range from 5
to 15 percent suspended solids. Solids are generally
controlled by regulating the slurry bleed rate, with
occasional reset. Magnetic flow meters are effective
monitoring devices, and slurry densities can be
measured by various devices.
Liquid-to-Gas Ratio. High liquid-to-gas ratios re-
duce scaling problems because the absorber outlet
slurry is more dilute with respect to absorbed SC>2.
Liquid rate is normally set for a constant flow,
the L/G varying inversely with gas load.
SODIUM CARBONATE SCRUBBING
The absorption of SO from flue gas may be accomplished by absorption
in a clear water solution of sodium carbonate and sulfite.
Process Description
The process is shown in Figure 3-5. A solution of soda ash (
reacts with S02 in the absorber to form sodium sulfite/bisulfite. Both
the reactant and the reaction products are completely soluble. If fly
ash is removed prior to S07 removal, the absorber can be either a packed
tower or a tray tower, high-efficiency devices with low pressure drops.
If fly ash is not removed a venturi scrubber may be used for both particu-
late and SO-, removal, with somewhat lower S02 absorption efficiency and
3-17
-------
MAKE-UP WATER
FLUE GAS
Na2CO3
1
SODA
LIQUOR
STORAGE
^ TO CHIMNEY
BLEED
ABSORBER
WASTE
LIQUOR
SURGE
TO
SEALED
DISPOSAL
POND _
Figure 3-5.
Simplified Process Diagram for Sodium Carbonate
Scrubbing System
3-18
-------
higher pressure drop. In either type of absorber, a recirculating liquid
stream as well as fresh soda makeup is required to improve gas-liquid
contact, which improves absorption efficiency. Because the sodium alkali
is very reactive, the design L/G ratios can be low (in the 1.3 to 3.4 1/m3,
or 10 to 25 gal/Mcf range). The system responds rapidly to changes in S02
loadings, and the soda feed rate can be varied to match a pH signal from
the absorber effluent, the pH changing with the S02 loadings.
Process Status and Background
A prototype unit, serving two industrial coal-fired boilers (equivalent
to 25 MW) at the General Motors assembly plant in St. Louis has been in
operation since 1972. Three 125 MW units at Nevada Power Company's Reid
Gardner Station are operating with low sulfur coal (see Table 3-1) and
a 520 MW installation is under construction at the Jim Bridger Station
in Wyoming.
Process Design and Control Parameters
A stream of spent alkali solution, containing sodium sulfite/bisulfite
equivalent to the amount of SO removed, must be purged from the system
to maintain chemical balance. The purge rate can be controlled by the
liquid density. This purge stream, usually slightly acidic, is neutra-
lized with more soda alkali before disposal. Process water makeup is
required to compensate for the water evaporated in the flue gas and
lost in the purge stream. Poor quality of this water (high calcium
content) can lead to calcium scaling.
The spent alkali purge stream is normally discharged to a sealed (lined)
evaporation pond for drying. Other proposed disposal methods include
precipitation of the sodium salts and the recovery of sodium sulfate
(salt cake) for sale.
The process consumes a premium chemical, either caustic soda or soda
ash; therefore, its application is limited to small industrial boilers
or utility boilers located near a source of the alkali. System
3-19
-------
advantages are simplicity, high absorption efficiency, and minimum
capital cost. However, disposal of the spent alkali requires careful
consideration.
DOUBLE ALKALI (SODA LIME) SCRUBBING
The double alkali process retains the high SO absorption efficiency
and nonscaling characteristics of the clear liquor sodium carbonate
system. It avoids the sodium salt waste disposal problem by insolu-
bilizing the scrubber effluent with lime precipitating calcium sulfite
and sulfate. This likewise regenerates the costly sodium absorbent
by displacing it with the less costly lime.
Process Description
The process is shown in Figure 3-6. As indicated, flue gas contacts
clear soda solution in an absorber. The spent absorbent is mixed
with lime slurry to precipitate the insoluble calcium salts and re-
generate the absorbent. The calcium sludge is filtered, washed
with water to recover the entrained soda solution, and discarded.
Sufficient washwater is used to balance the water evaporated during
the wet scrubbing of the hot flue gas.
Process Status and Background
Approximately 12 installations totaling 700 MW equivalent have been
operated on oil-fired and coal-fired industrial boilers in the U.S.
and Japan. Also, three installations totaling 1,050 MW on oil-fired
utility boilers have operated in Japan, and 20 MW prototype testing
has been carried out on a Gulf Power Company coal-fired utility
boiler in Florida. Three full-scale units totaling 1,100 MW are
scheduled for operation in the U.S. on coal-fired utility boilers
(Refs. 1,9).
3-20
-------
CHIMNEY
SCRUBBER
FLY ASH-FREE
FLUE GAS
H2O
Ca(OH)2
WASTE
CALCIUM
SALTS
1 1
MIXING
TANK
1
R
S
REACTOR
SYSTEM
Figure 3-6. Simplified Process Diagram for Double Alkali System
-------
Process Design and Control Parameters
One of the main problems associated with the double alkali process is
the difficulty in regenerating sodium sulfate because it does not re-
act well with hydrated lime in the presence of sodium sulfite. Sulfate
is formed by the oxidation of sulfite in the absorber. Oxidation can
be minimized by using a concentrated absorbing solution. It is
reported that as long as oxidation does not exceed 25 percent, the
sulfate ion formed can be coprecipitated as calcium sulfate in the
calcium sulfite crystals, thus releasing all the sodium for recycle.
Where oxidation cannot be controlled to below 25 percent (as with low-
sulfur coal and/or high 0 /SO ratios), a dilute absorbing solution
can be used to overcome the detrimental effect of excessive oxidation.
For dilute systems, in the absence of sulfite, the sulfate reacts
well, regenerating NaOH for recycle to the absorber. However, the dilute
mode has the following process disadvantages:
A higher L/G ratio
An extra softening step for the regenerated solution
to prevent scaling (added cost)
Higher lime requirement
The concentrated mode of operation, should oxidation exceed 25 per-
cent, has the following process disadvantages:
Soda makeup is increased to compensate for the purge
of unregenerable sodium sulfate
The sodium sulfate discarded in the waste sludge may
result in brackish groundwater
In any one application, these factors must be weighed to determine
the most suitable mode of operation. The concentrated mode is ex-
pected to be simpler and less costly and would be preferred for
3-22
-------
applications in which oxidation is expected to be relatively low. A
concentrated mode would be indicated for high sulfur coals.
MAGNESIUM OXIDE SCRUBBING
The magnesium oxide scrubbing process is a wet slurry scrubbing recovery
process. It differs from the similar calcium oxide (lime) process in
that: (1) magnesium sulfite produced can be readily calcined to
recover S02 in concentrated form, and (2) during calcining, the mag-
nesium oxide is regenerated and can be recycled to the process.
Thus, there is no sludge to dispose of, and only a small amount of
chemical makeup is required. This is in contrast to the calcium-
based scrubbing processes, which require a large quantity of chemical
makeup, and disposal of large volumes of sludge. Although it is a
slurry scrubbing process, scaling of absorbers in this system has
not been a problem because magnesium sulfite is 300 times more soluble
than the corresponding calcium sulfite, and magnesium sulfate is 120
times more soluble than calcium sulfate at the normal scrubbing
temperatures.
Process Description
A simplified flow diagram illustrating the basic process steps of the
magnesium oxide system offered by Chemico-Basic is shown in Figure 3-7.
For coal-fired boiler application, fly ash must be removed from
the flue gas before SO absorption to prevent buildup of inert solids.
Absorption of SO takes place in a suitable contactor (such as a ven-
turi) in which the flue gases are scrubbed with a circulating slurry
containing about 10 percent solids by weight. Magnesium oxide (MgO)
reacts with SO to form magnesium sulfite hydrates. Some of the
sulfite is oxidized to sulfate.
The magnesium sulfite/sulfate crystals are withdrawn in a bleed stream
from the absorber slurry loop to a centrifuge. The wet cake from the
3-23
-------
, »- CHIMNEY
SCRUBBER
RECOVERED SULFUR
FLY ASH-FREE
FLUE GAS
SULFUR OR ACID PLANT
Figure 3-7.
Simplified Process Diagram for Magnesium
Oxide Recovery System
3-24
-------
centrifuge is dried, and the remaining liquor is recycled to the
absorber. The dried crystals are conveyed to a calciner where they
are combined with a small amount of reducing agent (coke or carbon)
and calcined under a reducing atmosphere at about 760° to 870°C
(1,400 to 1,600 F) to produce SO and MgO. The reducing agent is
required to reduce the sulfate only. The calciner exit gases con-
tain between 10 and 16 percent SO by volume on a dry basis.
After dust removal, the S02-rich stream may be used for sulfuric acid
or sulfur production. The MgO, together with any required makeup, is
recycled to the scrubber slurry makeup system (slaker).
Process Status and Background
In the U.S., two full-scale installations have been operated, one on an
oil-fired boiler and the other on coal. A third full-scale unit is under-
going shakedown operation. These are summarized below.
A Chemico-Basic MgO scrubber at Boston Edison's Mystic Station went
onstream in 1972 (Ref.10); the power plant generally operated on 2 percent
sulfur fuel oil. The product magnesium sulfite and sulfate was
shipped to a chemical plant for regeneration of the magnesia and pro-
duction of by-product H7SO,. MgO was'delivered to the plant in a very
fine pulverized state (90 percent passing a 325-mesh screen). The
S02 removal efficiency was higher than 90 percent. MgO losses aver-
aged around 10 percent per cycle, although it was felt that with proper
system control this could be reduced to about 5 percent while operat-
ing satisfactorily at steady-state conditions with regenerated MgO.
The major problem areas were: (1) formation of fine trihydrate sulfite
crystals in the scrubber, which are more difficult to handle than the
coarser hexahydrate crystals; (2) dust emission from the dryer;
(3) erosion of pumps, valves, and piping; and (4) excessive wear on
3-25
-------
the internal parts of the centrifuge. The dryer dust problem was
solved by rerouting the dryer off-gas to the scrubber inlet rather
than the stack, but this prevented use of the dryer flue gas for
stack gas reheat. The unit was operated during 1972-1974. Since then
the utility has received a variance to burn higher sulfur oil.
Because Mystic Station was oil-fired, it did not provide information
on the effects of fly ash on system operability. Fly ash must be effi-
ciently removed ahead of the MgO scrubber to prevent buildup in the
recycled material. An MgO scrubbing installation at Potomac Electric
Power Company's Dickerson Station, started up in 1973, treating half
the flue gas from a 190 MW coal-fired boiler. A two-stage venturi
scrubber was used: the first stage removed the fly ash and the second
stage removed the S02« Reheat was accomplished by mixing scrubbed
flue gas with unscrubbed flue gas (Ref. 11).
The system operated intermittently from 1973 through 1975. Particulate
removal was in excess of 99 percent; S0~ removal efficiency ranged from
88 to 96 percent, depending upon gas flow rate. The MgO was operated
through only one cycle. Although scrubber availability was not all
that was desired due to problems with logistics in supplying raw
materials (MgO), and to mechanical problems mainly attributed to under-
design in such areas as piping, slurry pumps, and other auxiliary equip-
ment, it was felt that the basic concept and design have been proven
to be feasible (Ref. 12).
The United Engineers system differs from the Chemico-Basic system in
that: (1) the operating pH is lower (about 6) and the active absorbent
is magnesium sulfite instead of magnesium oxide; and (2) magnesium oxide
is added to the absorber effluent tank instead of the absorber (where it
reacts with magnesium bisulfite to form magnesium sulfite).
3-26
-------
The United Engineers MgO system at Philadelphia Electric Company's coal
fired Eddystone Station (120 MW) was operated briefly in 1975, then
shut down to relocate the MgO regeneration portion of the process nearby.
The system started up again in 1977 and is undergoing shakedown (Ref 1).
The scrubber train consists of a venturi particulate scrubber followed
by a venturi rod absorber. The system is reported to be designed to
realize 90 percent S02 removal from 2.5 percent sulfur coal (Ref. 13).
Process Design Considerations
The dried magnesium sulfite is quite stable and easy to transport.
For this reason the regeneration of the absorbent can be carried out
at any convenient location quite distant from the power plant, thus
permitting the use of a large central regeneration facility to serve
several flue gas cleaning locations.
SODIUM SULFITE (WELLMAN LORD) SCRUBBING
The Wellman Lord system is a clear liquor sodium alkali scrubbing
process with thermal regeneration and sulfuric acid or elemental
sulfur recovery.
Process Description
A flow diagram illustrating the basic process steps is shown in
Figure 3-8. In the absorber, sodium sulfite reacts with S02 to
form sodium bisulfite in solution. In the regeneration evaporator,
sodium bisulfite is thermally decomposed to sodium sulfite (solid) and
S02 (gas). The S02 is stripped with steam and passes from the evapo-
rator to a cooler where most of the steam is condensed. A concentrated
S02 stream (about 90 percent S02 and 10 percent H20) is fed to a sul-
furic acid plant or to a sulfur plant. The sodium sulfite crystals
are separated in a clarifier and redissolved in water for recycle to
the absorber.
3-27
-------
I
Ni
00
r
_^ TO
CHIMNEY
FLUE GAS
I
I
PRESCRUBBER
-TLfLTL
FLY ASH
PURGE
ABSORBER
SURGE
COOLING
WATER>
SODIUM CARBONATE
MAKEUP
ABSORBER
FEED
DISSOLVING
TANK
PURGE
STREAM
SODIUM
SULFATE
PURGE
TREATMENT
EVAPORATOR-
CRYSTALLIZER
SO2TO
SULFUR PLANT
Figure 3-8. Simplified Process Diagram for Wellman Lord Recovery System
-------
Sodium sulfate is formed by oxidation of sodium sulfite in the system.
This sulfate will not decompose in the regeneration evaporator, and
must be purged from the system.
Process Status and Background
The process was developed originally by Wellman Lord, Inc., and is
now licensed by Davy Powergas Company of Lakeland, Florida. A Well-
man Lord system has been in operation on two oil-fired industrial
boilers (35 MW each) in Japan since August 1971 with reported avail-
ability close to 100 percent. Two other full-scale installations on
oil-fired boilers were started up in Japan during 1973: one is a
220 MW oil-fired utility boiler, and the other is a 125 MW equivalent
industrial boiler. Both systems are reported to have been operating
successfully.
Thus far the Wellman Lord installations in operation in the United
States have been for treating tail gas for Glaus plants and sulfuric
acid plants. The first full-scale boiler install-ation in the United
States is a retrofit at the coal-fired Mitchell station of Northern
Indiana Public Service Company. This 115 MW unit is designed to re-
cover elemental sulfur, using natural gas as the reducing agent. The
unit has been completed and a comprehensive one year demonstration pro-
gram is underway (Ref. 14). Two additional units totaling 715 MW
are under construction for coal-fired utility boilers in New Mexico.
The same power plant has signed a letter of intent for 2 additional
units of 500 MW each.
Process Design Considerations
An important consideration in adapting the Wellman Lord process to
coal-fired plants is the removal of particulare matter ahead of the
absorber. Even with good particulate removal, the absorption liquor
may have to be filtered to facilitate subsequent processing steps.
3-29
-------
An additional concern is the oxidation of the sodium sulfite to
sodium sulfate. From 5 to 10 percent or more of the total incoming
sulfur is expected to be lost as soluble sodium sulfate in the purge
stream. This means higher makeup costs, using soda ash or caustic
soda. The purge stream must be disposed of in an environmentally
acceptable manner. Methods are offered for recovering salt cake
(Na SO.) from this stream.
3-30
-------
REFERENCES FOR SECTION 3
1. Pedco Environmental Specialists, Inc., Summary Report - Flue
Gas Desulfurization Systems, January-February 1977, Cincinnati,
Ohio.
2. Private communication, Mr. Peter Hall, Research-Cottrell, Palo
Alto, California, May 24, 1977.
3. Battelle Columbus Laboratories, Guidelines for the Design of Mist
Eliminators for Lime/Limestone Scrubbing Systems, for Electric
Power Research Institute, EPRI FP-327, December 1976.
4. Private communication, Mr. David Olson, Air Correction Division,
UOP, Los Angeles, telephone contact May 25, 1977.
5. LaKatos, S. F., W. M. Aubrey, and W. D. Hunter, "Status of Demon-
stration Wellman Lord/Allied Chemical FGD Systems NIPSCO D. H.
Mitchell Generating Station," EPA Flue Gas Desulfurization Sympo-
sium, New Orleans, March 1976.
6. Bechtel Corporation, EPA Alkali Scrubbing Test Facility: Advanced
Program First Progress Report, EPA Report 600/2-75-050, September
1975.
7. Bechtel Corporation, EPA Alkali Scrubbing Test Facility: Advanced
Program Second Progress Report, EPA Report 600/7-76-008, September
1976.
8. Bechtel Corporation, EPA Alkali Scrubbing Test Facility: Advanced
Program Third Progress Report, EPA Report EPA-600/7-77-105, Septem-
ber 1977.
9. Kaplan, N., "Introduction to Double Alkali Flue Gas Desulfurization
Technology," EPA Flue Gas Desulfurization Symposium, New Orleans,
March, 1976.
10. Quigley, C.P., and J.A. Burns, "Assessment of Prototype Operation and
Future Expansion Study Magnesia Scrubbing Mystic Station," EPA Flue
Gas Desulfurization Symposium, Atlanta, November 4-7, 1974.
11. Taylor, R.B., et.al, "Summary of Operations of the Chemico-Basic MgO
FGD System at the PEPCO Dickerson Generating Station," EPA Flue Gas
Desulfurization Symposium, New Orleans, March 1976.
3-31
-------
12. Magnesia FGD Process Testing on a Coal-Fired Power Plant, EPA
Report, EPA-600/2-77-165, August 1977.
13. Gille, J.A., "Magnesium Oxide Scrubbing at Philadelphia Electric's
Eddystone Station," EPA Flue Gas Desulfurization Symposium, New
Orleans, March 1976.
14. "Wellman-Lord S02 Recovery Process Flue Gas Desulfurization
Plant," EPA Capsule Report, EPA 625/2-77-011, 1977.
3-32
-------
Section 4
IMPLICATIONS OF REQUIRING 90 PERCENT OR GREATER
S02 REMOVAL FOR NEW FGD INSTALLATIONS
This section discusses the changes in a demonstrated FGD system design
of, for example, 80 percent S02 removal efficiency to obtain a capability
of 90 percent or more. In summary, these design changes could involve:
Increase in absorbent circulation rates
Increase in scrubbing (contacting) stages or in con-
tacting intensity (scrubber-type)
More stringent requirements for uniform flow
distribution
Elimination of gas by-pass if used for reheat
* Increases in absorbent concentrations which may in
turn require residence time or other changes to con-
trol enhanced scaling tendencies
If the increased SO^ removal requirement is also accompanied by an
increase in coal sulfur content, this change will intensify all of
the above requirements. Other fuel related factors such as coal
heating value, ash content and composition, must also be accommodated
and with increasing difficulty as removal requirement rises.
LIME AND LIMESTONE SCRUBBING
This section presents data to illustrate the effects of process operat-
ing parameters on S0~ removal efficiency for the lime and limestone
scrubbing process. Most of the data are taken from test results at the
EPA Alkali Scrubbing Test Facility located near Paducah, Kentucky
(Ref. 1). The test facility is integrated into the flue gas ductwork
of a 160-MW coal-fired boiler at the TVA Shawnee Power Station. Two
4-1
-------
parallel wet scrubber systems are in operation: a venturi/spray
tower system and a turbulent contact absorber (TCA) system. Each
system can treat approximately 10 MW (equivalent to 33,000 nm^/hr,
or 30,000 acfm at 300°F) of flue gas containing 1,500 to 4,500 ppm
of S0r
The bulk of the data in this section is for limestone scrubbing on the
TCA system. Additional data are presented in Appendix A for the TCA
with the lime and the venturi and spray tower with lime and limestone
operation. The Shawnee data are the result of short-term factorial
tests 6 to 8 hours in length and long-term tests about one week in
length. The short-term tests were aimed at determining S02 absorption
efficiencies without regard necessarily for possible long-term scaling
conditions that might be associated with the process conditions chosen.
Therefore, these particular data are not to be interpreted as evidence
of long-term reliable operation. However, during other portions of
the Shawnee program adequate demonstrations have been made that these
FGD systems can be operated for extended periods without scaling
problems.
Approximate equations and nomographs for predicting SOo absorption
efficiency for the Shawnee scrubbers are presented in Appendix B.
For the TCA Shawnee data in this section, the S02 removal curves
have been predicted by Equation B-4.
Effects of Process Variables on S02 Absorption Efficiency.
The effects of the following variables on S02 absorption efficiency
are summarized in this section:
Inlet S02 concentration
Liquid-to-gas ratio
Gas velocity
4-2
-------
Scrubber inlet pH
Scrubber internals
Magnesium addition
Number of scrubbing stages
Lime versus limestone
Inlet SQg Concentration. Figure 4-1 shows the effect of inlet S02 con-
centration on S02 removal for fixed design and operating conditions.
Data have shown (Refs. 2, 3, 4) that in lime and limestone scrubbing
over a wide range of inlet SC>2 concentrations, greater removal effi-
ciency is realized for low inlet SC^ concentrations than for high, for
fixed design and operating conditions. This happens because the actual
quantity of SC^ removed per volume of gas treated goes down, reducing
the loading on the absorbing liquor.
For high SOo concentrations, more S02 is actually absorbed by the liquor.
For example, an absorber with 80 percent reduction from a gas contain-
ing 3,000 ppm S02 might absorb 90 percent from a 1,000 ppm gas, all
things being equal. But in absolute terms, 2,400 ppm would actually
have been removed from the higher S02 case, and only 900 ppm from the
lower.
It was at one time thought that as inlet S02 concentration decreased,
the driving force for absorption would decrease and high removal effi-
ciencies would be more difficult. Clearly, as equilibrium S02 back
pressure over the bulk liquid approaches the SC>2 gas concentration,
this must pose a problem. In practice, however, the S02 back pressure
over fresh lime and limestone scrubbing slurries is very low (less than
1 ppm), and scrubbers can be designed so that the low S02 concentration
scrubbed gas exiting the scrubber is contacted by the freshest slurry
(counter current or cross flows). For these reasons, removal effi-
ciency greater than 95 percent has been achieved during tests from
4-3
-------
100
o
z
111
U.
UJ
EC
(M
8
I-
UJ
U
cc
Ul
a.
90 --
85 --
80 --
75 --
70
O EPAPILOTTCA
SPHERE HEIGHT - 7 INCHES/BED, 3 BEDS
LIQUID TO GAS RATIO = 85 gal/Mcf
TCA GAS VELOCITY = 7.5 ft/sec
TVA PILOT SPRAY TOWER
LIQUID TO GAS RATIO = 85 gal/Mcf
LIMESTONE SCRUBBING
1,000
2,000 3,000
INLET SO2 CONC., ppm
4,000 5,000
{Ref. 1)
Figure 4-1. Effect of Inlet SC>2 Concentration on S02 Removal
Efficiency for Fixed Design and Operating Conditions
4-4
-------
flue gas that averaged only 200 ppm at the inlet, as shown in Figure 4-2
for a TCA system operated at the Southern California Edison Mohave Sta-
tion (Ref. 5). The outlet SO,, concentration in this case is less than
10 ppm. It is important to recognize that to obtain equivalent effi-
ciency for 3,000 ppm S02 gas, an FGD system would need to absorb and
process about 15 times as much S02 (2,850 vs 190 ppm). Achievement of
comparable efficiencies at the high loading corresponding to 3,000 ppm
S02 gas cannot be considered to have been demonstrated by the Mohave
example. The gas S02 concentration, or coal sulfur content, must always
be accounted for when extrapolating the performance of different instal-
lations.
Liquid-to-Gas Ratio. Figure 4-3 illustrates the effect of L/G on SO-
removal efficiency in a TCA for limestone scrubbing. Similar effects
are observed for lime and in the venturi and spray tower scrubbers as
shown in Appendix A. Liquid-to-gas ratio is a major design variable
in scrubbing systems. As the liquid pumping rate is increased, addi-
tional alkali is added to the scrubber and the gas-liquid interfacial
area increases, resulting in greater absorption efficiency.
In most scrubber types, the benefits of increasing L/G for fixed
scrubber condition can be designed into the system up to the point
where problems such as flooding or nonuniform gas and liquor distri-
butions occur. It can be seen from the way the curves flatten out
at high L/G in Figure 4-3 that the incremental increase in pumping
to improve the absorption from, say, 80 to 90 percent (an increase
from about 50 to 80 L/G at a pH of 5.8) would be significantly greater
than that required to improve it from 70 to 80 percent (an increase
from about 35 to 50 L/G). Higher L/G could also be obtained by opera-
ting at lower gas velocity, at the expense of increased scrubber size.
An increase in the design L/G for an FGD system requires the use of
larger pumps, pipes, and spray headers. Scrubber recirculation tanks
4-5
-------
100-
99-
98-
u
Z
S 97-
o
u_
LL.
LU
96-
o
cc
CM 95-
O
CO
I-
LU
o
LU
CL.
94-
93-
92-
91-
90-
TCA 4 STAGES
LIMESTONE
INLETSO2 = 200PPM
GAS RATE = 450.000 SCFM
(Redrawn from Reference 5)
0
20
40
60
LIQUID TO GAS RATIO, gal/Mcf
Figure 4-2. Effect of Liquid-to-Gas Ratio on SO Removal Efficiency
with Low Sulfur Coal at the Mohave Power Station
4-6
-------
100
SCRUBBER INLET pH
pH = 5.8 LONG-TERM TEST
O pH = 5.7-5.9 FACTORIAL TESTS
90
n
A
80
O
uj 70 ..
O
UJ
cc
CM
8 60
H
UJ
O
DC
UJ
O_
50
pH = 5.4-5.6
pH = 5.1-5.3
FACTORIAL TESTS
FACTORIAL TESTS
40
30
SCRUBBER GAS VELOCITY = 10.4 ft/sec
TOTAL HEIGHT OF SPHERES = 15.0 in.
EFFECTIVE LIQUOR Mg++ CONCENTRATION = 0 ppm
INLET SO2 CONCENTRATION = 2,400-2.900 ppm
LIQUOR Cl~ CONCENTRATION = 3,000-7,000 ppm
-f-
4-
20
30
40 50 60
LIQUID - TO - GAS RATIO, gal/Mcf
70
80
(Ref.
Figure 4-3. Effect of Liquid-to-Gas Ratio on SO Removal
Efficiency TCA with Limestone
4-7
-------
must be larger to maintain design residence time for the increased L/G
and to provide residence time for the increased S0? absorption. The
energy penalty for increased L/G is shown in Figure 2-3. At higher
L/G's, gas pressure drop may also increase in some types of scrubbers,
causing additional penalty as shown in Figure 2-2.
Higher L/G's may tend to reduce scaling potential in the scrubber by
reducing the S0_ absorbed per gallon of absorbent.
Scrubber Gas Velocity. Figure 4-4 shows the effect of gas velocity on
SO- removal efficiency in a TCA. The small velocity effect at a
fixed slurry flow rate represents a trade-off between factors tending
to increase efficiency with velocity (e.g., superior sphere agitation
and increased liquid-gas contact) and the corresponding decrease in
L/G. On the other hand, a scrubber such as the spray tower (Figure
A-14) will have higher efficiencies at lower gas velocities, at a fixed
slurry flow rate. In the spray tower, there are no internals and little
liquid hold-up, so that the decrease in L/G with increasing gas veloc-
ity is not offset by other factors.
For cost reasons, scrubbers are designed to operate at maximum prac-
tical gas velocities, thereby keeping vessel diameters minimal. These
maximum velocities are dictated by gas-liquid distribution character-
istics of the scrubber and by the maximum allowable slurry entrainment
that the mist eliminator can handle. Designers generally achieve the
required L/G by increasing the slurry pumping rate rather than by de-
creasing the gas velocity. Designing for low gas velocities introduces
larger and costlier scrubber systems (modules).
In specifying high removal efficiency, it should be kept in mind that
some very efficient scrubbers (including certain tray towers) have
poor modular turndown characteristics, so that absorption efficiency
4-8
-------
100 +
SLURRY FLOW RATE
38 gal/min-ft2 LONG - TERM TESTS
O 38 gal/min-ft2 FACTORIAL TESTS
D 28 gal/min-ft2 FACTORIAL TESTS
A 19 gal/min-ft2 FACTORIAL TESTS
90 4-
z 80
UJ
ui
O
Ul
cc
M
O
w
H
UJ
u
DC
UI
Q.
70 +
60 +
50 4-
40
TOTAL HEIGHT OF SPHERES = 15.0 in.
SCRUBBER INLET pH = 5.7-5.9
EFFECTIVE LIQUOR Mg++ CONCENTRATION = 0 ppm
INLET SO2 CONCENTRATION = 2,000-3,000 ppm
LIQUOR Cl~ CONCENTRATION = 2,000-6,000 ppm
+
+
+
+
10 11
SCRUBBER GAS VELOCITY, ft/sec
12
13
(Ref, 1)
Figure 4-4.
Effect of Gas Velocity on.
TCA with Limestone
Removal Efficiency
4-9
-------
falls off significantly as gas velocity decreases. Although such
scrubbers may be designed to accommodate the normal boiler load turn-
down, they may lose efficiency below that range. One solution is to
take modules off-line as load decreases.
Scrubber Inlet pH. Figure 4-5 illustrates the effect of scrubber inlet
pH on absorption efficiency. Scrubber inlet pH is a measure of the lime
or limestone feed stoichiometry (moles Ca added/mole S0_ absorbed).
As the alkali feed is increased (i.e., as pH goes up), the additional
suspended alkali particles in the slurry lead to greater dissolution
rates in the scrubber. Since lime and limestone dissolution tends to
be limiting, increasing dissolution improves the efficiency.
If the lime or limestone processes are operated at too high a pH, calcium
sulfite scaling (and calcium carbonate scaling in the lime system)_ occur
in the scrubber. The actual maximum pH for reliable operation depends
on such variables as dissolved salts (calcium chloride, for example) and
liquid-to-gas ratio. For the systems explored at Shawnee, the maximum
safe pH values are about 5.9 for limestone (corresponding to a stoi-
chiometric ratio of about 1.4) and about 8.5 for lime (about 1.1 stoi-
chiometric ratio). Slightly lower pH values of 5.8 and 8.0 are more
comfortable and leave room for unintentional process excursions to
higher values. It is not good practice to provide design latitude to
achieve high efficiencies by operating at pH levels higher than these.
Scale formation caused by operation at such levels may plug scrubber
passages, preventing gas flows and increasing gas pressure drop. Dis-
lodged scale may damage slurry recirculation pumps and plug spray nozzles.
As alkali feed stoichiometries are increased, the excess alkali reacts
incompletely with absorbed SO- and is discarded, unused, with the waste
sludge. For limestone scrubbing at a stoichiometric ratio of 1.4, only
71 percent of the limestone is used. Additional increase in stoichiometry
4-10
-------
100
90
LIQUID-TO-GAS RATIO
FACTORIAL TESTS
O 60gal/mcf
O 45 gal/mcf
A 30 gal/mcf
80
UJ
O
iZ
ft
111
cc
V)
i-
Ul
o
cc
70 f
60 f
50
40 -
SCRUBBER GAS VELOCITY = 10.4 ft/sec
TOTAL HEIGHT OF SPHERES = 15.0 in.
EFFECTIVE LIQUOR Mg++ CONCENTRATION = 0 ppm
INLET SO2 CONCENTRATION = 2,300-2,700 ppm
LIQUOR Cl~ CONCENTRATION = 5,000-7,000 ppm
30
4.9
t
5.1
5.3 5.5 5.7
SCRUBBER INLET pH
5.9
6.1
{Ref. 1)
Figure 4-5.
Effect of Scrubber Inlet pH on S02 Removal Efficiency
TCA with Limestone
4-11
-------
wastes even more limestone and adds to the volume of waste sludge.
This is another drawback to the use of increased pH (and stoichiometry
to increase SO- removal.
Scrubber Internals. Figure 4-6 shows the effect of mobile packing height
on SO removal efficiency in a three-bed TCA. Increasing bed heights
by adding more spheres increases slurry holdup in the scrubber. This
has the combined beneficial effects of providing additional liquor
holdup time for alkali dissolution and a greater gas-liquid contact
area for improved efficiency. In the TCA, where sphere retaining grids
are about four feet apart, the static sphere height per bed can be adjusted
over a range of several inches. If greater removal is desired, the number
of beds can be increased. Increasing the bed height, however, increases
the gas pressure drop.
Other improvements in efficiency are achievable by increasing the
packing height in fixed packing towers and by adding more trays in
tray towers, although these types of scrubbers are not often used in
slurry service because of their susceptibility to plugging. These
improvements are made at the expense of increased gas pressure drop
across the scrubber and, therefore, increased fan power requirements (see
Figure 2-2).
Magnesium Addition. Adding soluble magnesium species to lime and lime-
stone process slurries is beneficial to scrubbing efficiency. The
alkaline magnesium compounds are highly soluble (in the form of magnesium
sulfite and sulfate) compared to calcium-based compounds and build up
to high concentrations in the scrubbing liquor. Since the absorbed SO
can react directly with the liquid-phase active magnesium species (mag-
nesium sulfite), the absorption rate is not limited by the slower dis-
solution of lime or limestone suspended solids, and efficiency is improved.
Depending on the magnesium concentration in the liquor, the absorption
can be accommodated almost entirely by the magnesium, or only partially
by the magnesium and the remainder by the lime or limestone.
4-12
-------
T
100
SLURRY FLOW RATE -
FACTORIAL TESTS
O 38 gal/min-ft2
D 28 gal/min-ft2
A 19 gal/min-ft2
90
40
SCRUBBER GAS VELOCITY = 10.4 ft/sec
SCRUBBER INLET pH = 5.8
EFFECTIVE LIQUOR Mg++ CONCENTRATION = 0 ppm
INLET SO2 CONCENTRATION = 2,300-2,700 ppm
LIQUOR Cl~ CONCENTRATION = 4,000-9,000 ppm
6 12 18 24
TOTAL HEIGHT OF SPHERES, inches
30
(Ref. 1)
Figure 4-6.
Effect of Bed Height on SC>2 Removal Efficiency
TCA with Limestone
4-13
-------
Figures 4-7 through 4-9 show the combined effects of dissolved mag-
nesium, L/G, and pH on SC>2 removal for the TCA with limestone.
Magnesium can be added to the process as magnesium oxide or magnesium
sulfate or, in a lime system, as dolomitic lime. In any case, the active
portion of the dissolved magnesium enters the scrubber as magnesium sulfite,
MgSCL, and leaves the scrubber after absorption as magnesium bisulfite,
Mg(HSO-) . In the slurry hold tank the lime or limestone continues to
dissolve and react with Mg(HSO-) to form the normal lime and limestone
calcium-based reaction products (waste sludge) and to regenerate the
active MgSO for recycle to the scrubber. Soluble magnesium compounds
are lost in the water portion of the sludge, requiring continued addi-
tion of makeup magnesium to the process. Also, the waste sludge must
be treated or impounded to prevent groundwater contamination by the
magnesium sulfate (Epsom salt).
The small amounts of chlorine in coal are converted to gaseous chloride
in the boiler. The chloride is absorbed from the flue gas by wet scrub-
bing processes, and the resulting dissolved chloride ion can build up
to appreciable concentrations in the scrubber liquor (2,000 to 16,000 ppm
at the Shawnee facility, where chlorine concentration in the coal ranges
from 0.03 to 0.3 percent and system water loop closure corresponds to
a discharge stream containing 35 to 65 percent solids). The dissolved
chloride neutralizes magnesium to an inactive form (MgCl_) relative to
the absorption process. To enhance absorption, magnesium concentrations
must be increased above that amount made inactive by the chloride. This
additional amount is referred to in Figures 4-7 through 4-9 as the effec-
tive liquor magnesium concentration.
The magnesium addition rate required to maintain a specified effective
concentration is, therefore, a function of the FGD process water loop
closure (amount of water lost in waste sludge) and the chlorine content
of the boiler feed coal. At Shawnee, in the TCA operating with lime-
stone, about 2 to 6.5 pounds of magnesium oxide per 100 pounds of
limestone are required to maintain a 9,000 ppm effective magnesium
concentration in the scrubber liquor.
4-14
-------
100
90
o
UJ
o
LL
U.
Ul
UJ
80
£ 70
H
UI
o
oc
Ul
60
50 -
40
EFFECTIVE LIQUOR Mg+
FACTORIAL TESTS
O 7,000-10,000 ppm
D 3,500-5,500 ppm
A 0-500 ppm
* CONCENTRATION
20
30
SCRUBBER GAS VELOCITY = 10.4 ft/sec
TOTAL HEIGHT OF SPHERES = 15.0 in.
SCRUBBER INLET pH = 5.4-5.6
INLET SO2 CONCENTRATION = 2,200-2,800 ppm
LIQUOR Cl~ CONCENTRATION = 6,000-16,000 ppm
1 1 1
40 50 60
LIQUID - TO - GAS RATIO, gal/Mcf
70
80
(Ref. 1}
Figure 4-7. Effect of Liquid-to-Gas Ratio on S02 Removal Efficiency
TCA with Limestone and Magnesium
4-15
-------
100
90
2 Removal Efficiency
TCA with Limestone and Magnesium
4-16
-------
100
90
1 r-
SCRUBBER INLET pH -
FACTORIAL TESTS
O pH = 5.7-6.0
D pH=5.4-5.6
A pH = 5.1-5.3
80
O
UJ
O
il
u_
UJ
UJ
EC
CM
8
U
DC
UJ
O.
70
60
50
40
30
SCRUBBER GAS VELOCITY = 10.4 ft/sec
LIQUID - TO - GAS RATIO = 45 gal/Mcf
TOTAL HEIGHT OF SPHERES = 0 in.
INLET S02 CONCENTRATION = 2,300-2,700 ppm
LIQUOR Cl~ CONCENTRATION = 5,000-14,000 ppm
1 1 1
2,000 4,000 . 6,000 8,000 10,000 12,000
EFFECTIVE LIQUOR MAGNESIUM CONCENTRATION, ppm (Ref. 1)
Figure 4-9. Effect of Magnesium on S02 Removal Efficiency
TCA (no spheres) with Limestone
4-17
-------
Number of Scrubbing Stages. If a single scrubber does not provide
sufficient SC^ removal efficiency, it is possible, with increased
cost, to design for two scrubbers in series. The net overall effi-
ciency is substantially higher than that of either stage.
Figure 4-10 shows the variation in two-stage absorption efficiency
as a function of individual scrubber efficiency. In this case the
first scrubber is a relatively inefficient (for lime or limestone
without additives) scrubber, such as a venturi. When the venturi and
spray tower have been operated together at the Shawnee Test Facility
(see Figure A-l), 862 removals greater than 90 percent have been
realized during short-term tests.
When two scrubbers are operated in series, the mechanical complexity
increases and additional energy penalties are incurred for the greater
pumping requirements and gas pressure drop (see Figures 2-2 and 2-3).
Japanese practice with oil combustion and recent U.S. developments in
coal combustion indicate that the extra capital cost of two-stage
scrubbing may be offset somewhat by the use of forced oxidation tech-
niques to bring down the high costs of sludge disposal and improve
alkali-utilization. In lime and limestone scrubbing, the waste prod-
uct is normally a slurry of calcium sulfite, calcium sulfate (gypsum)
and fly ash (if removed by the scrubber). The calcium sulfite can be
oxidized to gypsum by air-slurry contact (forced oxidation). The
resultant product has improved properties including higher settling
rates, improved dewatering characteristics, and reduced total waste
volume. One successful approach is to oxidize the slurry in the first
of two scrubbing stages. The cost offset of two-stage forced oxida-
tion does not apply to coals of lower sulfur content where oxidation
and settled sludge density are normally high anyway. The relationship
between sulfur content, oxidation, impounded sludge behavior, and cost
is not yet fully quantified.
4-18
-------
85
30
40 50
PERCENT SO2 REMOVAL FOR FIRST SCRUBBER
Figure 4-10. SO Absorption Efficiency for Two Scrubbers in Series
4-19
-------
Lime versus Limestone. The relative effects of limestone and lime
scrubbing on efficiency in the Shawnee TCA, spray tower and venturi
scrubbers at different pH's are shown in Figure A-11 for selected run
conditions. The performances of the three scrubber types are not
directly comparable because of differing L/G's. However, the lime and
limestone operating ranges for a given scrubber can be compared.
The fineness of the limestone particle size in the scrubbing slurry
is largely determined by the degree of grinding before addition to
the system. Lime, on the other hand, undergoes chemical reaction in
the hold tank to precipitate very fine particles of calcium sulfite
and calcium carbonate which then become the active alkali in the lime
system. The limestone system also precipitates calcium sulfite, but
the suspended alkali particles tend to be finer with lime. Because
the particle size is larger for limestone, there is less surface area
for dissolution so that, at equal stoichiometric feed ratios, the lime-
stone exhibits slower dissolution rates than lime and, therefore,
lower absorption efficiencies. To obtain similar SO removals, the
limestone system must be operated at higher stoichiometries.
In practice, when the limestone and lime processes are operated at
maximum reliable pH's (pH's above which scaling may occur) of about
5.8 and 8.0 at the Shawnee facility, they achieve similar absorption
efficiencies. In this case the limestone system would be operating
at a feed stoichiometry of about 1.4 while the lime system would be
at about 1.1.
In Figure 4-11, at a limestone pH of 5.8 and a lime pH of 8.0, the
TCA shows almost identical efficiency, the spray tower shows slightly
higher efficiency using lime, and the venturi shows better efficiency
using limestone. For these same pH values, it is possible to come up
with slightly different quantitative effects for limestone versus lime
than shown in the figure by altering such variables as L/G, liquid
holdup times, inlet S0_ concentration, and feed stoichiometry.
4-20
-------
100
90 --
80
UJ
u
£70
UJ
60
CM
8
g 50
UJ
a.
UJ
tc.
Ul
> 40
30
20
LIMESTONE -«*.
SPRAY TOWER
(51 gal/Mcf)
4-
678
SCRUBBER INLET LIQUOR pH
9 10
(Ref. 1)
Figure 4-11.
Effect of Scrubber Inlet Liquor pH on SC>2 Removal
Efficiency Lime and Limestone Derived Slurries
4-21
-------
It should be noted again that the S0? equilibrium back pressures are
very low, both when the limestone process slurry pH values range from
5 to 6 and when the lime process slurry pH values range from 7 to 8,
so that the higher lime pH values do not hold an advantage in this respect.
Extrapolation of Data
Approximate equations and nomographs for predicting absorption effi-
ciencies in limestone and lime wet scrubbing systems are presented in
Appendix B. The equations are semitheoretical in form and have been
fitted to spray tower and TCA SCL removal data from the Shawnee Test
Facility (Ref.l). The nomographs cover approximately the same ranges
of operation as the data to which the equations were fitted.
For a number of reasons the efficiencies predicted by these equations
may not agree completely with that of a commercial installation. Among
these are differences in scrubber internal configuration and liquid
holdup, nozzle pressures, limestone quality and grind, lime quality,
and variables that can affect liquor reactivity, such as sodium ion
concentration, sulfate saturation, and sulfite oxidation. Neverthe-
less, the equations and figures in this section and Appendix A arfe
significant in that they illustrate the relative effects of important
process variables, and contribute to the technical data which must be
available to permit design of full-scale units to meet specific local
and regional conditions.
Effect of Higher Absorption Efficiency on Process Design
Raising the design efficiency of a lime or limestone FGD system from
65 or 85 percent to 90 percent or greater affects a number of process
design parameters. Among possible required changes:
Scrubbers. More elaborate and equally reliable scrub-
ber modules must be engineered to provide greater
slurry holdup and gas-liquid interfacial area, or
existing design must be upgraded to include additional
internal devices. The number of vessels could be
doubled to provide operation with two scrubbers in
series. .
4-22
-------
Recycle Pumps and Piping. If L/G is increased to
improve absorption efficiency, pump size is increased.
Recycle slurry pipe diameters must be increased to
maintain design flow velocities
Limestone Preparation Equipment. To accommodate
greater feed rates in a limestone system, larger
grinding equipment, storage silos, transfer equip-
ment, slurry makeup tanks and feed pumps, and other
related equipment will be needed
Lime Preparation Equipment. To accommodate greater
feed rates in a lime system, the slaking equipment,
storage silos, transfer equipment, slurry tanks and
feed pumps, and other related equipment must be
larger
Magnesium Addition Equipment. If magnesium is added
to improve efficiency, the magnesium handling equip-
ment must be added to the process design, or dolo-
mitic lime used. The waste sludge must be treated
or contained to prevent groundwater contamination by
the dissolved compounds
Scrubber Recirculation Tanks. Tanks must be larger
to maintain design residence time for increased L/G
and to provide residence time for increased S02
absorption. These tanks may be directly under the
scrubber, integrated into the scrubber vessel design,
or offset. Slurry agitators will require redesign
Slurry Solids Separation Equipment. This equipment
must be larger to accommodate greater solids discharge.
This applies to clarifiers, filters, centrifuges, and
all related slurry pumps and piping
Fans. If greater gas pressure drop is incurred, the
fans must either be larger, or there must be more of
them with corresponding increases in the complexity
of the fan control system
Reheat. Bypass reheat would be marginal or impossible
because all the flue gas must be treated
Nonair Quality Environmental Effects. As S(>2 removal
is increased, provision must be made for handling the
increased waste products produced. If soluble addi-
tives are used to enhance absorption efficiency, steps
must be taken to prevent groundwater contamination
Energy Requirement. Power requirements would in-
crease with larger equipment loads
4-23
-------
This list illustrates the complexity of considerations that must be
addressed in attempting to design lime and limestone FGD systems for
90 percent or greater SO- removal efficiencies. Such an effort
inevitably involves higher cost and energy penalties. The energy
penalty will be in the form of increased auxiliary power consumption.
The cost penalties can be not only in terms of heavier equipment, but
of greater operating complexity and higher maintenance as well.
SODIUM CARBONATE SCRUBBING
The once-through sodium carbonate process is capable of high efficiency
over a wide range of inlet SO- concentrations. This clear liquor
process is gas-phase controlled rather than being limited by the
slow dissolution of suspended solids. However, the process con-
sumes a premium chemical and produces a soluble waste salt which is
normally discharged to a sealed (lined) evaporation pond for drying in
arid areas.
During 1972 a series of tests was carried out at the Shawnee Test
Facility using sodium carbonate (Ref. 6). Runs were made with air
containing injected SO and with flue gas from the boiler. The tests
were designed primarily to determine coefficients in mathematical
correlation models for predicting SO- removal efficiencies.
Figure 4-12 shows the effect of pressure drop on SO absorption effi-
ciency. Better than 90 percent efficiency was realized at a .pressure
drop of 15 mm Hg (8 in. HO) in these tests. The effects of L/G and
gas velocity on efficiency in the TCA with no internals (operated as
a spray tower) are shown in Figure 4-13. When the TCA was operated
in its normal three-bed configuration, efficiencies higher than 99 per-
cent were realized. Figure 4-14 shows the effects of L/G and marble
4-24
-------
100
95 --
90 --
U
UJ
O
iT
u.
W
o
UJ
oc
CM
8
8 75
DC
85 --
80 --
70 --
65 -
60
SCRUBBER INLET LIQUOR pH = 9.5
SCRUBBER LIQUOR SODIUM CONC. = 1.0 wt %
SO2 INLET CONC. = 600-3,300 ppm
(Air/SO2 & Flue Gat)
LIQUID-TO-GAS RATIO = 8-50 gal/Mcf
THROAT GAS VELOCITY = 41-105 ft/we
PLUG OPENING = 40-80 percent
4-
4-
468
PRESSURE DROP, in H2O
10
12
(Ref. 6)
Figure 4-12. Effect of Pressure Drop on S02 Removal Efficiency
Venturi with Sodium Carbonate (10 MW size)
4-25
-------
100
95 --
90 -
UJ
U
u.
U-
UJ
O
5
HI
DC
CM
O
CO
LU
O
cc
LU
O.
85 --
80 --
75 --
70 --
65
D
SCRUBBER INLET LIQUOR pH = 9.5
SCRUBBER LIQUOR SODIUM CONC. = 0.5 wt %
S02 INLET CONC. = 900 ppm (Air/SO2)
D GAS VELOCITY = 6.2ft/wc
O GAS VELOCITY = 9.2 ft/sec
A GAS VELOCITY = 12 ft/sec
20
40 60 80 100 120
LIQUID-TO-GAS RATIO, gal/Mcf (Ref. 6)
Figure 4-13. Effect of Liquid-to-Gas Ratio on S02 Removal Efficiency
TCA (no spheres) with Sodium Carbonate (10 MW size)
4-26
-------
100
95 ~
90 --
O
Ul
O
u.
85 --
ui
cc
CM
z
ill
O
cc
UJ
80 --
75 -
70 -
65
SCRUBBER INLET LIQUOR pH = 9.5
SCRUBBER LIQUOR SODIUM CONC. = 1.0 wt %
SO2 INLET CONC. = 1200 ppm (Air/SO2)
GAS VE LOCITY = 8 ft/sec
O MARBLE HEIGHT =2 in.
A MARBLE HEIGHT = 5 in.
4-
10 20 30 40
LIQUID-TO-GAS RATIO, gal/Mcf
50
(Ref. 6)
Figure 4-14. Effect of Liquid-to-Gas Ratio on SC>2 Removal Efficiency
Marble Bed Scrubber with Sodium Carbonate (10 MW size)
4-2-7
-------
bed height on SC>2 removal in the marble bed scrubber. Absorption
^
efficiencies of 98 percent were obtained for inlet S02 concentrations
of 1,200 ppm.
To date application for coal-fired utility plant FGD systems has been
limited to geographical areas having a cheap source of low grade car-
bonate. These include 3 units operating at a power station in Nevada
(Reid Gardner - 375 MW) and one under construction in Wyoming (Jim
Bridger - 520 MW). Both stations burn low-sulfur coal and the FGD sys-
tems are designed to realize about 90 percent removal efficiency.
DOUBLE ALKALI SCRUBBING
The clear-liquor sodium-based double alkali processes, like the once-
through sodium carbonate, are aimed at realizing high removal effi-
ciencies over a wide range of inlet concentrations. The process is gas-
phase rather than solid dissolution limited. More efficient scrubber
types .can be used, such as packed towers and tray towers. Venturi scrub-
bers provide excellent SO removal, although at substantial energy
penalty. The process is more complex than the lime and limestone
scrubbing processes.
Approximately 12 installations totalling 700 MW equivalent have been
operated in oil-fired and coal-fired industrial boilers in the U.S. and
Japan. Three installations totaling 1,050 MW on oil-fired utility
boilers have operated in Japan, and 20 MW prototype testing has been
carried out on a Gulf Power Company coal-fired utility boiler in Florida.
The double alkali process remains to be tested at full-scale on a coal-
fired boiler.
During 1973 an -8-month test program with a 3,400 m /hr (2,000 cfm)
double alkali pilot plant was carried out by Combustion Equipment
Associates and Arthur D. Little for the EPA to generate design data for
the prototype system at the Sholz Steam Plant of Gulf Power Company.
4-28
-------
Results of the laboratory program and pilot operations are reported in
the literature (Ref. 7). Figure 4-15 illustrates the high S02 removal
efficiency possible with the double alkali processes.
The 20 MW prototype system was operated at the Sholz plant during 1975
and 1976. The scrubber system consisted of a venturi followed by a
tray tower and was designed to remove a minimum of 90 percent of the
SO. in the flue gas from medium to high sulfur coal (up to 5 percent
sulfur, Ref. 7).
The system was operated with the venturi and absorber together in
series and with the venturi alone. The absorption efficiency for
each of these crrttfi'g-ura'.tions is shown in Figure 4-16 for an inlet S02
range of 1,050 to 1,250 ppm. According to Reference 8, the points
in the figure each represent at least a few hours of operation at
the condition shown. Achieving a given outlet SOj level (within the
efficiency limit of each configuration) was essentially a matter of
adjusting the operation pH of the scrubber system. The scrubber L/Gs
are not indicated in the figure. However, the 2.0 to 2.3 1/m (15 to 17
gal/Mcf) L/G operation in Figure 4-15 shows that the L/G of a double
alkali process is much lower than for lime and limestone processes.
Using the venturi and tray tower in series, outlet S0? levels below
50 ppm were realized at a venturi liquor pH above 5.2, correspond-
ing to better than 95 percent absorption efficiency. Above a pH of
6.0, outlet SO levels dropped to 20 ppm or less - greater than 98
percent removal.
Generally, when the venturi and tray tower were operated together,
the pH of the venturi bleed liquor was maintained between 4.8 and
5.9. Consequently, outlet SO levels ranged between 15 and 100 ppm
and, typically, between 30 to 70 ppm according to Reference 9.
During one 2-month period the outlet SO level ranged from 5 to 210
ppm and averaged 45 to 50 ppm. On five occasions the outlet S02
exceeded 100 .ppm.
4-29
-------
100
95
90
I
_c
8
2 85
"3
i
oc
N
w> on
75
70
Scrubber Operating Conditions
Bleed Liquor Temperature 140°F - 150°F
Active [Na+1 0.3 - 0.5 M
Total Dissolved Solids 5-15 wt°o
Sulfite Oxidation 250 ppm SOj E(iuivaliMit
Vpnturi: P ' 11-14 in. H2O
L/G - 15 - 17 gal/Macf saturated
o
0.7
Range of SO2 Inlet Levels
D 500-700 ppm
A 800-1,000 ppm
02,250-2,650 ppm
0.8 0.9 1.0 1.1 1.2 1.3
Scrubber Feed Stoichiometry (mols Na + capacity/mos SO2inlet)
1.4
1.5
(Ref. 7}
Figure 4-15.
Effect of Feed Stoichiometry on Removal Efficiency in
the Venturi/2-Tray Tower Absorber for the EPA^ADL Double
Alkali Pilot Program
4-30
-------
500
400
I. 300
^Q.
CN
200
100
0
4.
Operational Configuration
Venturi + 2 Trays
o Venturi + 2 Trays
Venturi (No Feed to Trays)
Venturi
R,(in, H20)
5-7
8-11
8-11
Inlet S02 = 1050- 1250ppm
Active
Na+, (M)
0.25-0.4
0.15-0.3
0.15-0.3
5.0
5.5 6.0
Scrubber Bleed Liquor pH
6.5
157
65
74 w
O
C/5
+J
I
I «
83 °-
.1
3
191
100
7.0
(Ref. 8)
Figure 4-16. Effect of pH on SO Removal for CEA/ADL Double
Alkali Prototype System
4-31
-------
Double alkali FGD systems are currently being offered and designed
to achieve high absorption efficiency in both industrial and utility
applications, although the process has yet to be demonstrated on a
full scale coal-fired utility boiler. A 227 MW installation at
Louisville Gas & ElectricTs Cane Run Station is scheduled to start up
in 1979. The system is designed to meet performance guarantees of
SC>2 emissions no greater than 200 ppm at coal sulfur contents up to
5 percent and 95 percent removal for coal containing over 5 percent
sulfur.
MAGNESIUM OXIDE SCRUBBING
The magnesium oxide slurry scrubbing process has seen limited full
scale operation in the U.S., although several systems operate in Japan
on various industrial plant off-gases. A 155 MW installation was oper-
ated intermittently during 1972 through 1974 at Boston Edison's oil-
fired Mystic Station. The system was designed for 90 percent SO
removal, but experienced mechanical operating difficulties. The system
was generally operated at more than 90 percent removal throughout the
program (Ref 10) an-d, during tests in 1974, averaged 91 percent removal.
A 95 MW magnesium oxide system was operated during 1974 and 1975 at
Potomac Electric's coal-fired Dickerson Station. Although scrubber
availability was not all that was desired due to problems with logist-
ics in supplying raw materials (MgO), and to mechanical problems mainly
attributed to under-design in such areas as piping, slurry pumps, and
other auxiliary equipment, it was felt that the basic concept and de-
sign have been proven to be feasible (Ref. 10). S02 removal efficiency
was reported to be consistently above 90 percent whenever operating at
design gas flow and inlet S02 concentrations above 1,000 ppm (Ref. 11).
During performance testing, removal efficiency ranged from 88 to 96 per-
cent, depending on the gas flow rate (Ref. 12).
4-32
-------
A 120 MW installation at Philadelphia Electric's coal-fired Eddystone
Station was operated briefly in 1975, then shut down to move the MgO
regeneration portion of the process to a nearby location. The unit
started up again in 1977. The scrubber train consists of a venturi
particulate scrubber followed by a venturi rod absorber. The system
was designed to achieve better than 90 percent SO removal from
2.5 percent sulfur coal (Ref. 13).
The magnesium oxide slurry scrubbing process has been demonstrated to
be feasible on full-scale coal-fired boilers and is capable of
S02 removal efficiency greater than 90 percent although the reliability
of these installations must be improved. The process is more complex
and costly (Refs. 14 and 15) than lime and limestone scrubbing processes.
SODIUM SULFITE (WELLMAN LORD) SCRUBBING
S02 removal efficiencies greater than 90 percent for this process are
well established. About 17 installations are in operation in the U.S.
and Japan on industrial plants and oil-fired boilers (Ref. 16). A
115 MW installation on the coal-fired Mitchell Station of Northern
Indiana Public Service has been completed and a comprehensive one year
demonstration program is underway. This is the first application of
the process on a coal-fired boiler.
All other Wellman Lord installations operating in the U.S. are in
refineries and display S02 removal efficiencies greater than 90 per-
cent. Like other clear liquor scrubbing processes, the absorption
efficiency is not limited by the slow dissolution of lime or limestone.
The process is scale free and does not recirculate a slurry. Present
scrubbers in this service are both efficient and reliable. The process
*
is more complex and costly (Refs. 14 and 15) than the lime and lime-
stone scrubbing processes.
4-33
-------
The installation at Mitchell Station was designed to meet performance
guarantees of a minimum of 90 percent S0~ removal from flue gas gen-
erated by 3.5 percent sulfur coal (Refs. 17, 18) and has demonstrated
greater than 90 percent removal.
4-34
-------
REFERENCES FOR SECTION 4
1. Bechtel Corporation, EPA Alkali Scrubbing Test Facility: Advanced
Program Third Progress Report, EPA Report EPA-600/7-77-105, September
1977.
2. R. H. Borgwardt, Limestone Scrubbing at EPA Pilot Plant, Progress
Report No. 6, EPA Report, January 1973.
3. J. M. Potts et al., "Removal of Sulfur Dioxide from Stack Gases
by Scrubbing with Limestone Slurry: Small-Scale Studies at TVA,"
Second International Lime/Limestone Wet Scrubbing Symposium, New
Orleans, November 8-12, 1971.
4. Gleason, J. R. , "Limestone Scrubbing Efficiency of Sulfur Dioxide
in a Wetted Film Packed Tower in Series with a Venturi Scrubber,"
Second International Lime/Limestone-Wet Scrubbing Symposium, New
Orleans, November 8-12, 1971.
5. Weir, A., et al, "Results of the 170 MW Test Modules Program
Mohave Generating Station, Southern California Edison Company,"
EPA Flue Gas Desulfurization Symposium, New Orleans, March 1976.
6. Bechtel Corporation, EPA Alkali Scrubbing Test Facility: Summary
of Testing Through October 1974, EPA Report 650/2-75-047, June 1975.
7. LaMantia, C.R. et al., "EPA-ADL Dual Alkali Program - Interim
Results," EPA Flue Gas Desulfurization Symposium, Atlanta,
November 4-7, 1974.
8. LaMantia, C. R. et al. , "Operating Experience - CEA/ADL Dual Alkali
Prototype System at Gulf Power/ Southern Services, Inc.," EPA Flue
Gas Desulfurization Symposium, New Orleans, March 1976.
9. Quigley, C. P., and J. A. Burns, "Assessment of Prototype Opera-
tion and Future Expansion Study - Magnesia Scrubbing Mystic
Station," EPA Flue Gas Desulfurization Symposium, Atlanta, Novem-
ber 4-7, 1974.
10. Magnesia FGD Process Testing on a Coal-Fired Power Plant, EPA
Report, EPA-600/2-77-165, August 1977.
4-35
-------
11- Taylor, R. B. et al., "Summary of Operations of the Chemico-Basic
MgO FGD System at the PEPCO Dickerson Generating Station," EPA
Flue Gas Desulfurization Symposium, New Orleans, March 1976.
12. Erdman, D.A., "Mag-Ox Scrubbing Experience at the Coal-Fired
Dickerson Station, Potomac Electric Power Company," EPA Flue
Gas Desulfurization Symposium, Atlanta, November 4-7, 1974.
13. Gille, J.A., "Magnesium Oxide Scrubbing at Philadelphia Electric's
Eddystone Station," EPA flue Gas Desulfurization Symposium, New
Orleans, March 1976.
14. Hollinden, G. A., and W. L. Wells, "Effects of Coal Quality on the-
Reliability and Economics of FGD Systems," NCA/BCR Coal Conference,
Louisville, Kentucky, October 19-21, 1976.
15. McGlamery, G. G., et al., "Detailed Cost Estimates for Advanced
Effluent Desulfurization Processes," EPA Report 600/2-75-006,
January 1975.
16. Pedroso, R. I., "An Update of the Wellman Lord Flue Gas Desulfuri-
zation Process," EPA Flue Gas Desulfurization Symposium, New
Orleans, March 1976.
17. LaKatos, S. F. et al., "Status of Demonstration Wellman Lord/
Allied Chemical FGD Systems NIPSCO D. H. Mitchell Generating Sta-
tion," EPA Flue Gas Desulfurization Symposium, March 1976.
18. "Wellman-Lord S02 Recovery Process - Flue Gas Desulfurization
Plant," EPA Capsule Report, EPA-625/2-77-011, 1977.
4-36
-------
Section 5
EFFECT OF COAL PROPERTIES ON FGD SYSTEMS
The major coal properties affecting FGD system design and operation are
heating value and sulfur, ash, moisture, and chlorine content.
The effects include:
Heating value of coal. Affects flue gas flow rate
generally higher for lower heating value coals which
also contribute a greater water vapor content to the
flue gas
Moisture content. Affects the heating value and con-
tributes directly to the moisture content and volume
of the flue gas
Sulfur content. The sulfur content together with the
allowable emission standards determines the required SO.
absorption efficiency, the FGD system complexity and cost,
and also affects sulfite oxidation
Ash content. May affect FGD system chemistry and
increases erosion. In some cases it may be desirable
to remove fly ash upstream from the FGD system
Chlorine content. May require high alloy metals or
linings for some process equipment and could affect pro-
cess chemistry or require prescrubbing
The importance of these factors is described in this section.
Coals are commonly classified according to rank, which refers to the
degree of geological transformation from lignite to anthracite. A num-
ber of classification systems have been developed, but in very broad
terms coal is ranked as anthracite, bituminous, subbituminous, and lignite.
Typical ranges of coal properties for these categories are presented
5-1
-------
in Table 5-1. Bituminous coals are largely used by the U.S. power in-
dustry, but subbituminous coals and lignite are increasing in importance
as supplies of high grade coal are depleted. Figure 5-1 indicates the
geographical distribution of the various classes of coal deposits.
The properties of power plant flue gas depend on the complexities of
the coal composition and on the equipment and conditions of the com-
bustion process. An overview of the coal-flue gas relationship is
possible by considering the apparent differences in the major combustion
products from coals of widely differing rank fired in boilers appropriate
for the efficient combustion of each. For example, Table 5-2 compares
the heating values, ultimate analyses, wet flue gas analyses, firing
rates, and S0? and fly ash emissions for a bituminous coal fired in a
pulverized coal boiler and for a lignite fired in a cyclone boiler, both
at 130% total air.
The above comparison should not be regarded as necessarily typical
because similarly-ranked coals may vary widely in composition.
The data show that the coals are significantly different with respect
to heating values (related to their differing ratios of carbon to oxygen)
and their sulfur, moisture, ash, nitrogen, and chlorine content. The
major significant differences between the combustion gas streams is the
volume per megawatt, moisture content, SO- and fly ash concentrations,
and to a lesser degree, NO and chloride content. In general, larger
A
flue gas volumes per megawatt result from coals with lower heating values.
The following sections summarize the major effects of coal and its com-
bustion gas characteristics on FGD processes.
5-2
-------
Table 5-1
TYPICAL RANGE OF COAL ANALYSES
Analysis :
Moisture, %
Volatile matter,
%
Fixed carbon, %
Ash, %
Heating value,
103 Btu/lb
Sulfur, %
Nitrogen, %
Anthracite
2-5
5-12
70-90
8-20
12-14.5
< 1
0.5-1
Bituminous
2-15
18-40
40-75
3-25
10-14
0.5-5
1-2
Subbituminous
15-30
30-40
35-45
3-25
8-10.5
0.5-3
1-1.5
Lignite
25-45
25-30
20-30
5-30
5.5-8
0.5-2.5
0.5-1.5
(Ref. 1)
5-3
-------
Ui
I
Anthracite and Semmnlhro
Low-volatile bituminous coal
(Ref. 2)
Medium and high-volatile bituminous coal
Figure 5-1. Coal Fields of the United States
-------
Table 5-2
EXAMPLES OF COAL AND FLUE GAS COMPOSITIONS
Characteristic
r-t
a
o
u
(0
OJ
o
1-1
t
Heating value, Btu/lb
Carbon
Hydrogen
Oxygen
Nitrogen
Sulfur
Chlorine
Ash
Moisture
Nitrogen
Carbon dioxide
Oxygen
Water vapor
Sulfur dioxide
Sulfur trioxide
Nitrogen oxides
Hydrogen chloride
S02 concentration, ppm
Fly ash, grain/scf (wet)
Coal firing rate, Ib/hr/MW
Gas rate, acfm/MH @ 300°F,1 a tin
S02 emission: Ib/hr/MH
lb/10& Btu
Fly ash emission, Ib/hr/MW
J * f
lb/106 Btu
Bituminous
11,700
64.2 wt %
4.4
6.5
1.3
4.0
0.1
11.5
8.0
100.0
74.29 vol %
12.66
4.52
8.17
0.30
0.003
0.05
0.007
100.00
3,000
4.3
795
3,100
64
6.9
78
8.4
Lignite
7,070
42.6 wt %
2.9
11.8
0.6
0.7
6.4
35.0
100.0
69.06 vol 7,
12.67
4.20
13.91
0.08
0.0008
0.07
100.00
800
1.3
1,460
3,760
20
1.9
28
2.7
Note: Similarly-ranked coals vary so widely in composition that these
examples should not tie regarded as typical.
Basis:
30% excess air; 100% sulfur overhead
Bituminous coal: pulverized coal boiler; 9,300 Btu/kWhr
boiler heat rate; 85% ash overhead
Lignite coal: cyclone boiler; 10,300 Btu/kWhr
boiler heat rate; 30% ash overhead
5-5
-------
COAL HEATING VALUE AM) MOISTURE CONTENT
Coal Firing Rate
Because a power plant using a low heating value coal must fire at a
higher burn rate to generate the same amount of power, such coals pro-
duce a larger volume of flue gas and greater SO emissions per unit of
generated power. The effect on the FGD system is twofold. First, the
flue gas handling equipment, including the scrubbers, must be of a
larger size to accommodate the greater gas volumes. Typically, power
3
plant flue gas volumes may range from 5,000 to more than 7,000 m /hr/MW
(about 3,000 to 4,000 acfm/MW), depending on the coal composition, boiler
heat rate, gas temperature, and power plant elevation (or gas pressure).
Secondly, the increased S0? emissions mean that on a megawatt basis the
FGD system must treat proportionally larger quantities of S0_. This is
illustrated with the examples presented in Table 5-2 above. Here the
higher heating value bituminous coal contained nearly 6 times as much
sulfur as the lignite (4.0% versus 0.7%). Yet on a megawatt basis the
bituminous coal S0_ emissions were only 3 times as large (20 versus
9 kg/hr/MW, or 64 versus 20 Ib/hr/MW).
On a megawatt basis, therefore, the FGD system (as well as the power
plant) equipment capacity is greater and capital and operating costs
are higher for coal with lower heating value for a given coal sulfur
content and S09 absorption efficiency (Ref. 3).
Water Vapor
A characteristic of lower heating value coals and coals of high moisture
content is a flue gas with a greater proportion of water vapor. In a
wet scrubber this means that less water is evaporated per unit volume of
gas and the adiabatic saturation temperature for the gas (the temperature
to which the gas is cooled in the scrubber by contacting the scrubbing
\
liquor) is higher. This also results in higher recirculating slurry
5-6
-------
temperatures. The overall effect Is that the S02 absorption process
may take place at somewhat higher tempertures (perhaps 6 to 8°C, or
-O
10 to 15 F, higher for lignite over bituminous coals). For the example
in Table 5-2 the bituminous coal flue gas has a saturation temperature
of 52 C (126 F), while the corresponding lignite gas saturation tempera-
ture is 59°C (138°F). Elevated temperatures could affect absorption
efficiency depending on the chemistry of the particular process.
SULFUR CONTENT OF COAL
Sulfur is present in coal in amounts ranging from traces to 5 percent
or more. Three forms of sulfur exist in the coal:
Organic sulfur which is combined with the coal
substance
Pyritic sulfur which is combined with iron
Sulfate in the form of calcium or iron sulfate
Sulfate sulfur accounts for less than one percent of the total sulfur
present, while the proportions of organic and pyritic sulfur vary widely
depending on the coal. Upon combustion most of the sulfur appears in
the flue gas as sulfur dioxide. The inorganic sulfur content of coal
can be reduced by washing, but no system has yet been commercialized
for this purpose, and not all coals are amenable to such treatment.
S00 Removal Rate
"~ JL
The sulfur content of the coal together with the allowable emission
standards determines the absolute removal rate of S02 in pounds per hour.
For a given absorption efficiency the sulfur content of the coal directly
affects the design of almost every piece of equipment in the FGD system.
5-7
-------
For example, a lime or limestone system designed for high rather than
low sulfur coal would have:
« Scrubbers with capacity for greater SO- removal
Higher L/G's and therefore bigger pumps and piping
and higher pumping energy requirements
e Bigger fans and greater energy requirements if the
improved scrubber design results in higher gas
pressure drop
e Larger sized alkali storage, preparation and feed
equipment
e Greater lime or limestone feed rates
Larger scrubber recirculation tanks to maintain
residence time for increased L/G's and to provide
additional time for increased S0_ absorption load
Greater capacity slurry solids separation equipment
Provision for disposing of the larger waste volumes
Increased power requirements for the larger equipment
loads
The result is that for higher sulfur coals the lime and limestone FGD
systems are more complex and have higher capital and operating costs
(Refs. 3, 4).
Other FGD processes are similarly affected by the sulfur content of
the coal. For systems using regenerable absorbents (double alkali,
magnesium oxide and Wellman Lord processes), the capacity of the regen-
eration section is directly proportional to the sulfur content. With
high sulfur coal the overall cost of these sections (capital and oper-
ating) represents a large portion of the total cost for the system.
With recovery processes there is relatively little waste but a large
by-product processing cost, although the greater amount of by-product
produced helps to offset these costs.
5-8
-------
Sulfite Oxidation
The amount of sulfite oxidized to sulfate in scrubber solutions is pro-
portional to the relative amounts of oxygen and S02 absorbed. It also
depends on the pH and temperature of the liquid as well as the composi-
tion of particulate emissions which may contain iron or copper that act
as catalysts for the oxidation reaction. In general, however, as the
gas S02 concentration becomes smaller, the fraction of sulfite oxidation
tends to increase. For this reason, when lime or limestone scrubbing
systems are used for low sulfur coal or for boilers operating on high
excess air they may experience high sulfite oxidation and produce waste
solids that are mainly gypsum. This can be a desirable feature since
gypsum solids are more easily dewatered due to faster settling rates
and higher final settled densities. Conversely the lower oxidation
observed with high sulfur coals can lead to solid wastes high in sulfite
and difficult to dewater.
In the double alkali process the formation of sodium sulfate by oxida-
tion is normally undesirable since in concentrated solution in the
presence of sodium sulfite it is relatively difficult to convert back
to an active form of sodium. For high sulfur coal applications in
which oxidation may be expected to be relatively low, a concentrated
double alkali system can be employed. With this mode of operation (see
Section 3) oxidation is controlled by using a concentrated absorbing
solution which minimizes absorption of oxygen from the gas. For low
sulfur applications where higher oxidation is unavoidable a dilute mode
of operation is required. In the dilute solution the sodium sulfate
reacts well with lime to precipitate gypsum and regenerate active sodium
for recycle to the absorber. Thus a high degree of oxidation is not
detrimental and forced oxidation is sometimes used to convert all the
sulfite to sulfate.
5-9
-------
Sulfur Trioxide
About one percent of the sulfur in the flue gas is present as sulfur
trioxide (SO ). When the gas is cooled below about 135°C (275 F), sul-
fur trioxide reacts with water vapor to form sulfuric acid. This acid
condenses at temperatures well above the water dew point. If the acid
droplets form in an atmosphere of water vapor, a fine mist is produced
which is difficult to remove, even by wet scrubbing. If the reaction
occurs by absorption of gas into the scrubbing liquor, the SO. is re-
moved and contributes to sulfate formation.
ASH CONTENT OF COAL
Most coals fired in U.S. utility boilers contain 5 to 30 percent ash.
After combustion, part of the ash falls to the bottom of the furnace
and the remainder is carried upward with the flue gas. The fraction
of the ash that is carried overhead is a function of the boiler design
and combustion parameters. With pulverized coal firing, 85 percent or
more of the ash appears as fly ash; in cyclone boilers about 20 to 30
percent of the ash goes overhead (Ref. 1).
Mineral Content
The principal inorganic constituents of fly ash are silicon, aluminum,
iron, and calcium. The oxides of these four elements comprise 95 to
99 percent of the composition of ash. Ash also contains smaller amounts
(0.5 to 3.5 percent) of magnesium, sulfur, sodium, and potassium as well
as very small quantities (parts per million) of from twenty to fifty
elements (Ref. 2). When fly ash containing potentially hazardous heavy
metals is removed together with S0_ the waste product may require special
disposal precautions to prevent contamination of groundwater.
Fly Ash Removal
Fly ash can be removed upstream of the FGD system by a precipitator,
fabric filter, or prescrubber, or integrally within the FGD system itself.
5-10
-------
Not all FGD processes are suitable for combined removal. Even when
fly ash is removed upstream, residual ash becomes entrained in the pro-
cess liquor.
Fly ash is invariably abrasive; some is chemically inert, and some is
highly acidic due to S0_ adsorption. Fly ash can cause excessive erosion,
scaling and plugging of equipment. It contributes to the waste volume of
throwaway processes, the loss of absorbent for regenerative processes,
and may contaminate the byproduct of recovery processes. Certain coals
from Wyoming, Montana, and North Dakota produce alkaline fly ash with
large amounts of reactive calcium, magnesium, sodium, and potassium
oxides. With combined removal of alkaline fly ash and SCL, major re-
duction in the alkali makeup requirement can be realized. The presence
of calcium alkali in the ash can, however, aggravate wet dry interface
problems by producing hard insoluble deposits.
In general, only processes using nonregenerable absorbents (lime, lime-
stone, soda) can be used for the combined removal of fly ash and S0_.
The fly ash is then disposed of together with the spent absorbent. Lime
and limestone processes can better tolerate this fly ash and are some-
times used for combined removal. Venturi scrubbers are often used for
this purpose at the expense of increased pressure drop over other absorp-
tion systems. However, the fly ash contributes to solids buildup at the
wet-dry interface and causes erosion of pipes, pumps, spray nozzles, and
scrubber internals.
Soluble clear liquor soda processes can be designed for combined removal,
but the FGD systems must then be designed for slurry handling and must
deal with special disposal considerations. For combined removal with a
soda system the fly ash must be free of calcium to retain scale-free
operation of such systems.
Processes using regenerable absorbents (double alkali, magnesium oxide,
and Wellman Lord) are not suited for combined removal of fly ash and S0~.
5-11
-------
A double alkali mode of operation, while theoretically feasible for
joint removal, would suffer from high cost and absorbent loss. Fly
ash could in some instances introduce a scaling problem. Similarly, the
Wellman Lord system would become complex, absorbent loss would increase,
and the system would produce a fly ash sludge contaminated by sodium
sulfate. The magnesium oxide process cannot be used for combined re-
moval because of contamination of the spent absorbent by the fly ash.
Alkaline Fly Ash Scrubbing
A significant characteristic of many western U.S. low sulfur coal ashes
is their high content of calcium oxide and magnesium oxide. The alkali
content tends to be highest in lignite and less prevalent in subbitumous
and bituminous coals. In lignite, calcium oxide concentrations may be
20 percent or higher in the fly ash, while magnesium and sodium oxides
may range up to 6 percent or more. These fly ash alkali values have
been successfully used for the absorption of S0? with a major reduction
in the raw alkali makeup requirement (Ref. 5).
The alkali content may be sufficient to react with and absorb a large
portion of the S0? produced by the coal combustion. The fly ash can
either be captured by the FGD scrubbers and recirculated with the absorp-
tion liquor, or it can be collected dry in a precipitator and then added
to the process. A key design factor for alkaline fly ash scrubbing is
that the alkalinity is generally released and usable only under low pH
absorption conditions (Ref. 6). This can limit absorption efficiencies,
but also tends to promote high oxidation of the waste product for improved
disposal characteristics. Due to the natural variability of fly ash
quantity and alkali content a supplemental feed of limestone or lime is
generally required for effective chemical control of the system.
CHLORINE CONTENT OF COAL
Chlorine varies from a trace to as high as 0.5 percent in U.S. coals.
It is rare in western coal, but tends to be common in the East. During
5-12
-------
combustion the chlorine is converted to hydrogen chloride or other
volatile chlorides. Nearly all of the hydrogen chloride in flue gas
is absorbed by the scrubbing liquor in a wet FGD system. Chloride
concentrations in the gas may be as high as 100 ppm for coal with high
chlorine content. Upon absorption, chloride concentrations build up
in the recirculating scrubber liquor. The concentration is a function
of the water loop closure of the FGD system, and levels may reach 5,000
to 10,000 ppm or more for high chloride coals and water loop closures
corresponding to the discharge of a waste stream containing 60 percent
solids or higher.
Stress Corrosion
The presence of chloride in the liquid circuits of an FGD system pro-
vides the potential for chloride stress-corrosion in alloys susceptible
to the phenomena. The potential is related to concentration, tempera-
ture, pH, alloy treatment and stress levels. The solution to the problem
requires the use of high alloy equipment wherever rubber or other pro-
tective coatings are not feasible.
Effect on Process Chemistry
In wet scrubbing processes, dissolved chloride replaces active calcium,
magnesium or sodium alkalis by their chloride salts, which are inactive
in the absorption process. The alkali associated with the chloride is
then lost as dissolved solids in the water portion of the waste sludge.
From a cost standpoint this is particularly objectionable for magnesium
and sodium based processes (or magnesium enhanced lime and limestone
processes), because these alkalis are relatively expensive. For such
processes, prescrubbing may be used to absorb chlorides from the flue
gas upstream of the FGD system. This minimizes both alkali loss and
chloride stress-corrosion problems. For lime and limestone processes,
an equivalent amount of calcium is used up by the chloride (the amount
is small relative to that used for SO- absorption), but the calcium is
relatively inexpensive.
5-13
-------
NITROGEN CONTENT OF COAL
Nitrogen oxides are formed by oxidation of nitrogen compounds in the
coal and by reaction of nitrogen and oxygen in the combustion air at
high temperature. In utility boilers nitrogen oxides are typically
about 95 percent nitric oxide (NO) and about 5 percent nitrogen dioxide
(N0?). Less than 0.1 percent of the nitrogen entering the boiler leaves
as NO . Cyclone boilers'operate at higher flame temperatures than other
X
boilers, generating somewhat higher concentrations (Ref. 1).
Although nitrogen oxides are present in small concentrations (500 ppm
or less in pulverized coal boilers), they may participate in SO- oxi-
dation phenomena and thereby affect the chemistry of FGD systems (Ref. 7)
Only small amounts of these oxides are absorbed, showing up as nitrates
in the scrubber liquor. In most installations to date the level of
nitrates has not been high and nitrogen oxides are normally not con-
sidered in designing a wet scrubbing FGD system.
FGD systems currently installed in the U.S. cannot effectively remove
nitrogen oxides. Control of their ground level concentrations often
establishes stack height and may contribute to requirements for reheat.
5-14
-------
REFERENCES FOR SECTION 5
1. McKnight, J.S., Effects of Transient Operating Conditions on
Steam-Electric Generator Emissions. EPA Report 600/2-75-022,
August 1975.
2. Ray, S.S., and F.G. Parker, Characterization of Ash from Coal-
Fired Power Plants, EPA Report 600/7-77-010, January 1977.
3. Hollinden, G.A., and W.L. Wells, "Effects of Coal Quality on the
Reliability and Economics of FGD Systems," NCA/BCR Coal Conference,
Louisville, Kentucky, October 19-21, 1976.
4. McGlamery, G.G., et. al., "Detailed Cost Estimates for Advanced
Effluent Desulfurization Processes," EPA Report 600/2-75-006,
January 1975.
5. Ness, H.M., et. al., "Status of Flue Gas Desulfurization Using
Alkaline Fly Ash from Western Coals," EPA Symposium on Flue Gas
Desulfurization, New Orleans, March 1976.
6. Grimm, C., et. al., "Particulate and S0£ Removal at the Colstrip
Station of the Montana Power Co.," The Second Pacific Chemical
Engineering Congress, Denver, Colorado, August 28-31, 1977.
7. Graefe, A.F., et. al., "The Development of New and/or Improved
Aqueous Process for Removing S02 from Flue Gases," NAPCA Report
No. APTD-0620, October 1970.
5-15
-------
Section 6
REHEAT OF SCRUBBED FLUE GASES
The reasons and methods- for reheating scrubbed flue gases are described
in this section. The energy consumption of the various methods is
defined, and problems anticipated or experienced with reheater design,
operation, and maintenance are described. Alternative approaches to
reheat are discussed.
Flue gases are normally discharged from the regenerative air heater of
a power plant at 120 to 150°C (250 to 300°F). The temperature is selected
to remain above the dew point of the traces of H«SO, normally present in
order to reduce corrosion and permit carbon steel to be used for fans,
ducting, and stack lining.
When a wet scrubber is inserted between the air heater and stack for
S09 removal, the flue gas exiting the scrubber is saturated with water
and cooled to the saturation temperature of about 50 C (125 F). Dis-
charge of the cool, wet gas to the stack produces:
Water condensation in the stack exit gas (acidic
rain or fallout from the plume) and possibly in
the exit duct and stack
Duct work and stack corrosion precluding use of
carbon steel
e Impaired plume rise and hence poorer dispersion
for a particular stack height
To correct the above undesirable aspects of wet scrubbing, the
treated gas may be reheated to a higher temperature before discharge.
6-1
-------
Although considerable attention has been focused on the importance of
proper scrubber design, there is perhaps less understanding of the
reheat function; yet, failure of a reheater can shut down the entire
system. Investment and operating costs for reheating can be high.
For these reasons the EPA has initiated a comprehensive reheat assess-
ment study.
PURPOSE AND NEED FOR REHEAT
The reheater is basically a heat transfer and/or gas mixing device
for heating the flue gas and restoring buoyancy by the addition of
sensible heat. This sensible heat partially compensates for the
latent heat losses that occur during adiabatic saturation within the
scrubbers. The amount of energy added and the final flue gas temper-
ature depend on the specific need for which the temperature increase
is being considered.
A number of installations are operating under "wet stack" conditions,
i.e., without reheat (see page 6-19). Such conditions approach the
ideal design goal of minimizing the thermal energy penalties associated
with flue gas cleaning. These installations provide substantial sulfur
dioxide removal, and evidence to justify reheat is scarce. Some of the
factors that could influence such evidence are discussed in the follow-
ing paragraphs.
Ground Level Concentrations of Pollutants
Scrubbers are installed to reduce emissions and thereby reduce the
exposure to pollutants at ground level. A stack provides dispersion
of the plume so that concentration of pollutants will fall within
accepted standards by the time any part of the flue gas reaches the
ground.
6-2
-------
An important aspect of this dispersion of the flue gas plume is the
height to which the plume rises after it leaves the stack. The plume
rise varies with the meteorological conditions, but is particularly
dependent on the temperature of the gas. The higher the plume rise,
the lower will be the peak ground-level concentrations downwind of
the stack.
Without reheat the oxides of nitrogen, which are not significantly
removed by the scrubber, may reach ground-level concentrations exceed-
ing those which would have prevailed had no scrubbing occurred. Suf-
ficient plume rise must be provided to prevent this.
When the design efficiencies for sulfur dioxide removal are realized,
it is unlikely that the ground-level concentration downwind of the
*
stack will be higher than without scrubbing.
Duct and Stack Corrosion
One of the features which limits boiler efficiency is the effect of
corrosion on heat recovery hardware as the temperature of the exit
flue gas approaches the critical region of 120 to 150°C (250 to 300°F).
Further heat recovery by cooling in a typical power plant requires
corrosion-resistant construction, the cost of which cannot be justified.
During combustion, while most of the sulfur is converted to sulfur
dioxide, a portion reacts further to form sulfur trioxide (SO ) .
Within the scrubber these acid gases react to form sulfurous and sul-
furic acids which are neutralized by the alkali. Beyond the scrubber
Such a situation was reported in England at Battersea. Under certain
conditions, the plume which was not reheated fell to ground immed-
iately after leaving the stack, causing complaints by nearby residents,
Battersea, however, employed once-through scrubbing and it seems
probable that outlet gas temperatures fell well below the adiabatic
saturation temperature characteristic of present day closed-loop scrub-
bers, thereby contributing to the severe plume drop.
6-3
-------
outlet, water vapor continues to condense, reacting with residual SC>2
to form sulfurous acid. In the absence of alkali, an acidic, corrosive
condition prevails.
Sulfur trioxide has a far more corrosive effect than sulfur dioxide.
Concentrations may range from 0.5 to 5 percent of the total sulfur
oxides present (Ref. 1) . Even this low concentration of SC>3 raises
the acid dewpoint temperature considerably. Condensation occurs,
and corrosive sulfuric acid is formed.
Only partial removal of SO, occurs in the scrubber. During tests with
a venturi scrubber at the EPA Shawnee Test Facility, inlet SO- concen-
tration ranged from zero to 22 ppm, while emission concentrations
ranged from zero to 14 ppm. Average S0~ removal was about 60 percent
(Ref. 2). For S0_ concentrations in the flue gas between 6 and 8 ppm,
theoretical acid dewpoints are between 127 and 132 C, or 260 and 270 F
(Ref. 3).
The water dewpoint of scrubbed flue gas is typically about 50 C (125 F).
The presence of as little as 0.1 ppm of SO- in the flue gas creates an
acid dewpoint of above 93°C, or 200°F (Ref. 3).
It is questionable whether the reheat of such gas to the acid dewpoint
can be justified. This may be further quantified by the EPA reheat
assessment study. The very slight condensation of SO- with water depletes
the acidity within the gas, thereby dropping the acid dewpoint down-
stream. In the final analysis, it is only necessary to reheat suffi-
ciently to prevent significant condensation of corrosive vapor. This
means heating the gas to a point intermediate between water dewpoint
and theoretical acid dewpoint. The exact point can only be determined
by empirical means, using judgment from experience. The severity of
corrosion depends on the materials used.
6-4
-------
The degree of reheat required to protect carbon steel has not been
confirmed by experience, but about 30°C (50°F) is often prescribed.
Fan Protection
When dry fans are installed ahead of an FGD system, reheat does not
affect fan selection. Fly ash collection must, of course, be installed
ahead of the fan. Although this may raise the cost of the flue gas
cleaning system, it removes the fans from the list of factors to be
considered in designing for reheat.
In some installations gas is drawn through the scrubber by an induced-
draft fan. To keep the fan in balance, scrubber residue must not be
allowed to deposit on fan blades. A wet gas (or high dewpoint) favors
such deposition. Reheat is needed to be sure that all dispersed
solids in the gas stream are dry before they enter the fan intake. A
reheat level of 30 C (50 F) has often been prescribed to attempt to
prevent such a condition.
Wet fans and washed fans have also been used; in such cases no reheat
is applied ahead of the fan. Wet fans are high in initial cost because
the fan wheel is fabricated from high alloy steels, while the casing is
rubber lined. The wet fan consumes less power because the wet gas is
somewhat lower in volume (Ref. 4). Experience with wet fans is limited.
Plume Appearance
When the gas leaves the stack, water vapor condenses in the cooler
atmosphere 'forming a visible plume. As the visible plume disperses,
condensed vapor evaporates at a rate depending on the ambient humidity
and temperature. With high external humidities early in the day, the
visible plume may travel long distances before disappearing. In a
desert environment at mid-day, on the other hand, the visible plume
6-5
-------
vanishes rapidly. If the appearance of such a plume is considered
undesirable, it can be reduced or eliminated by reheating so that
condensation does not occur.
The NSPS calling for 20 percent opacity as originally promulgated by
the EPA Administrator in 1971 specifically provided that the effect
of uncombined water vapor be excluded from the standard. In
October 1975 this provision was stricken from the standard; however,
Method 9, the procedure for visual determination, was not altered.
Method 9 identifies two types of plume: attached and detached. The
method specifies that the opacity be evaluated at a point following
moisture plume disappearance in the case of attached plumes, and before
the point of condensation in the case of detached plumes. As a prac-
tical matter this poses a difficult judgment for the observer who may
not be able to discern the boundary between steam plume and that
caused by fly ash. Reheat may therefore be used as a means for assur-
ing minimum opacity readings. In the case of attached plumes, however,
the provision of Method 9 favors having the maximum steam plume prac-
ticable since actual visible emissions may be masked until they are
adequately dispersed.
Figure 6-1 is a psychrometric chart showing the relationship between
moisture content, temperature, and plume visibility. A straight line
ties the condition of the flue gas to that of the ambient atmosphere.
When this line crosses the saturation curve, a visible plume will
form from condensation. The longer the line on the condensation side
of the curve, the more persistent the plume. Addition of heat will
displace the line to the right of the saturation curve, preventing
visible steam plume formation.
Stack Rain
Stack rain is a serious concern in the design of scrubber installations.
Condensation in the exit duct and stack may result in mist droplets that
6-6
-------
0.12
UNREHEATEDGAS
CONDITION
20
60
100
140 180
TEMPERATURE, °F
220
260
Figure 6-1. Example of Plume Behavior for Reheat Alternatives
6-7
-------
can settle around the power station. The situation is aggravated if
the mist eliminator is so inefficient that liquor blows out the stack
without condensation (Ref. 5). Typical phenomena are:
« Large, muddy droplets with a pH less than 8
Small, clean droplets, mistlike in nature, pH less
than 5
The first condition is characteristic of a high velocity stack coupled
with incomplete mist elimination. The second is less important and
may occur only intermittently.
With the double-wall construction that typifies modern power plant
stacks, the temperature loss in the stack may be as little as 3 to 6 C
(5 to 10°F) but sufficient to cause condensate nucleation without
agglomeration by the time the gas reaches the top of the stack. Heavy
droplets are more likely to result from impingement and coalescense on
the wall near the base of the stack, coupled with the drag effect of gas
flow on the resulting water film at velocities higher than 12 to 15 m/sec
(40 to 50 ft/sec). In the absence of reheat, wet fans and inefficient
scrubber mist eliminators lead to impingement and film buildup.
Little reheat is needed to prevent the stack rain that may be caused by
condensation due to stack temperature drop. Use of reheat to overcome
the effect of liquid entrainment from the scrubber or from the wet fans
is wasteful; use of more efficient mist eliminators would be the pre-
ferred alternative.
Stack Icing
Under freezing conditions liquid entrainment may lead to ice formation
on top of the stack. The possibility of ice falling from the stack
is a hazardous condition that can be alleviated by reheat.
6-8
-------
Opacity Monitoring
The appearance of a plume can be quantitatively expressed by opacity,
the plume's ability to reflect light and to interfere with light
transmissions. The opacity results from the scattering of light by
solid particles and condensed or entrained liquid droplets in the
plume. Persistent opacity as the plume disperses is related to solids
continuing after droplets have been evaporated.
Continuous optical monitoring of the flue gas opacity within the stack
is a current NSPS requirement. This provides a useful tool for esti-
mating emission and visual opacity values based on calibrated correla-
tions with measurements taken by Methods 5 and 9. The presence of
droplets because of cooling or entrainment interferes with the preci-
sion of the optical devices which do not distinguish between liquids
and solids. Reheat which effects vaporization of liquids serves to
overcome this drawback.
METHODS OF REHEATING
Flue gas can be reheated in many ways, and several approaches have
been developed. The basic differences in reheat methods are energy
sources and methods of transferring that energy to the flue gas. As
shown in Figure 6-2, several different energy sources and energy
transfer methods can be used.
Reheat methods currently in use include:
e Direct inline reheat using steam or hot water
heat exchangers
Direct combustion reheat using gas or oil in either
inline burners or external combustion chambers
Indirect hot air reheat using steam to heat air
to mix with the scrubbed gas
6-9
-------
ENERGY SOURCES
USEFUL SYSTEM
ENERGY
WASTE
SYSTEM
ENERGY
EXTERNAL
ENERGY
EXTRACTION
STEAM
HOT
FLUE
GAS
(>300°F)
COLD
UNSCRUBBED
FLUE GAS
( < 300°F)
DIRECT
INTERMEDIATE
HOT AIR
DIRECT
INTERMEDIATE
HOT AIR
DIRECT
INTERMEDIATE
WATER
CLEAN FUEL
(WITH ADDITIONAL AIR)
WASTE FROM
OTHER PROCESSES
ENERGY TRANSFER METHODS
HEAT EXCHANGER
TUBE
Figure 6-5
LJUNGSTROM
j
<
MIXING
Figure 6-6
Figure 6-7
Figure 6-8
f f
Figure 6-3
Figure 6-4
Not Practical
Figure 6-2. Reheat Energy Sources and Transfer Methods
6-10
-------
Bypass reheat bypassing a portion of the untreated
hot flue gas to mix with the scrubbed gas
Some typical approaches to reheat are shown in Figures 6-3 through 6-7.
Energy Sources
Energy may come from the power plant itself as useful or waste energy
or from an external source. Figure 6-2 shows energy sources that may
be used, and discussion of these sources is presented below.
Useful System Energy. Useful system energy refers to energy extracted
from the power plant system itself which might otherwise be used to
produce electricity. Two general streams of fluid contain such useful
energy: extraction steam and hot flue gas.
Extraction steam comes from the system either before the steam turbine
or at some intermediate stage of expansion before the last turbine
stage. Hot flue gas refers to flue gas that retains some useful energy;
for example, that between the economizer section and the air preheater
section at a temperature between 260 and 400 C (500 and 750 F). Each
source is available either to heat the scrubbed gas directly or to
heat air which can then be mixed with the scrubbed gas.
The choice depends on convenience, optimization, and cost. In either
case, extraction of energy from the system must be factored into its
design. A modification is to use flue gas recirculated from a point
downstream of the reheater. This would amount to bleeding off a por-
tion of the reheated gas in Figure 6-6 and cycling it through the steam
coil heat exchanger in place of the air. Since the gas is hotter than
the air, there could be a savings in the amount of energy used, but
costs for ducts would be higher. This method is being tested
experimentally at small scale by TVA (Ref. 6).
6-11
-------
AIR FAN
ELECTROSTATIC SCRUBBER REHEATER STACK
PRECIPITATOR
Figure 6-3. Direct Combustion Reheat
AIR FAN
ELECTROSTATIC SCRUBBER
PRECIPITATOR
BLENDER STACK
Figure 6-4. Reheat with External Waste Energy
6-12 .
-------
AIR FAN
ELECTROSTATIC
PRECIPITATOR
SCRUBBER REHEATER STACK
Figure 6-5. Direct Inline Reheat with Steam
AIR FAN ELECTROSTATIC SCRUBBER BLENDER STACK
PRECIPITATOR
Figure 6-6. Indirect Hot Air Reheat
6-13
-------
BOILER
COAL
HOT FLUE GAS
(NOT SCRUBBED)
AAAA
HOT
ELECTROSTATIC
PRECIPITATOR
HOT FLUE GAS
C7
AIR FAN
AIR PREHEATER
SCRUBBER BLENDER STACK
Figure 6-7. Bypass Reheat Useful System Energy
ELECTROSTATIC SCRUBBER BLENDER STACK
PRECIPITATOR
Figure 6-8. Bypass Reheat Waste System Energy
6-14
-------
Waste System Energy. In a typical power plant, low-level energy is pres-
ent in the warm stack gas and the condenser cooling water. Obviously,
the only source of these two which is hot enough for reheat is the
warm flue gas before it enters the scrubber. Some of the energy can
be released either by transfer of mass directly to the scrubbed flue
gas (i.e., bypass), or by using an intermediate loop with heat trans-
fer fluid in a closed system. This fluid would convey heat from the
120 to 150°C (250 to 300°F) flue gas downstream to the scrubbed flue
gas. The latter approach is being tested experimentally at small
scale by TVA (Ref. 6).
External Energy. When it is not feasible to use waste energy in the
fashion described above, supplemental energy sources can be used. This
may be particularly applicable for retrofit installations. One method
is to burn natural gas or low-sulfur oil, followed by direct injection
of the combustion products into the scrubbed flue gas. The main draw-
back of this type of reheat in the U.S. is the limited availability and
cost of the low-sulfur fuel required. Another method might use waste
energy from a separate system; for example, exhaust gases from gas
turbines which are generally at 315 to 370°C (600 to 700°F).
Heat Transfer Methods
The second factor to be considered, in addition to the energy source,
is the method of energy transfer. Either a heat exchanger or mixing
is employed.
Heat Exchanger. Two types of heat exchangers are used in power plants:
the tube type and the regenerative or Ljungstrom £ype. Both have been
used extensively, and their advantages and drawbacks are well known.
Mixing. A simpler, but not necessarily more economical, way of adding
energy is to add the hot gas to the cold mixture. This technique
requires care to achieve homogenous mixing.
6-15
-------
ENERGY CONSUMPTION
Reheating consumes energy in two ways: the heat transfer required to
raise the temperature of the gas and the fan energy required to force
the flue gas through the reheater.
When the flue gas is forced through a heat exchanger, the energy
requirement for a given temperature increment of reheat will be mini-
mized because all of the heat transferred goes directly to heating the
flue gas. On the other hand, the pressure drop through such a heat
exchanger may be significant and will require some fan energy. When
energy is conveyed to the flue gas by mixing it with a hotter gas,
the energy requirement for a given temperature increment is generally
larger than for the inline heat exchanger. This is because a larger
amount of gas must be heated (flue gas plus mixing gas). In general,
the pressure drop in a mixing chamber will be smaller than that in a
heat exchanger.
Gas mixing methods of reheat include indirect hot air reheat (with
air drawn from the outside, heated in an exchanger, and mixed with
the flue gas) and direct combustion reheat (clean fuel burned in air
and the combustion products mixed). They differ in their energy
requirements depending on the reheat increment. In the case of in-
direct hot air, the volume, rather than temperature, of hot air is
increased to raise the temperature of the scrubbed gas because a
constant temperature steam extraction source (or hot water) is avail-
able as the primary energy source. To achieve higher temperatures
with direct combustion reheat, more fuel is burned; but less energy
may be required because the combustion products are at higher tempera-
ture than for the heated air method.
The energy source for reheating is an important variable. Although
it may not affect the energy requirement, the electrical energy lost
6-16
-------
may well be different for different sources. For example, if waste
heat is used, by definition no electricity would be lost. On the
other hand, the penalty associated with steam use depends on the point
in the cycle at which the energy is extracted. If steam is extracted
at the highest temperature and pressure available (an unlikely choice),
the cost of that heat will be very high. If steam is extracted at some
intermediate point, the cost is lower. In determining the cost of
such heat, its value in terms of available electrical energy must be
considered. The boiler and turbine must be designed with the reheat
steam penalty in mind.
Bypass reheat has the advantage of low capital investment, negligible
operating cost, and simple and reliable operation. However, the maxi-
mum degree of reheat obtainable is limited by the overall S0_ removal
requirement versus the S0? removal capability of the FGD system.
The relative theoretical energy requirements for various degrees of
reheat have been calculated by Battelle (Ref. 6). Results of these
calculations are shown in Figure 6-9. The results are for a typical
coal-fired boiler. In actual practice, the energy consumption may
be slightly higher than that indicated because of mist eliminator
inefficiency (mist carryover and latent heat penalty).
It should be noted that for the same degree of reheat, indirect hot
air reheat has the highest energy requirement. However, for the same
amount of energy consumption, indirect hot air reheat may have equal
or better benefits because of its dilution effect on the plume.
6-17
-------
cc
O
00
OC
LJJ
UJ
U
QC
LJJ
Q.
CO
X
UJ
cc
SOURCE: REF.6
20
I
40 60
STACK GAS REHEAT,
80
100
Figure 6-9. Theoretical Reheat Energy Requirements for
Typical Coal-Fired Boiler
6-18
-------
PROBLEMS EXPERIENCED
This section examines some of the reheater problems experienced in oper-
ational flue gas desulfurization systems. Tables 6-1 and 6-2 indicate
the type of reheat in each generating station considered and Tables 6-3
through 6-6 highlight the problems within each system. It is of interest
to note that the following power plants operate without reheat:
Cane Run No. 4
Louisville Gas and Electric Company
Green River 1 and 2
Kentucky Utilities
Conesville No. 5
Columbus and Southern Ohio Electric Company
Duck Creek No. 1A
Central Illinois Light Company
Southwest 1
Springfield City Utilities »
Bypass reheat is employed by Martin Lake 1, Texas Utilities Company, and
additional FGD systems are under construction with provision for bypass
reheat. Operating experience to date is limited, but the selection of
duct lining material for the hot-cold gas interface area is a critical
design factor. The material must be capable of withstanding thermal
shock, the high temperature of the bypassed gas stream, and the corro-
sive effects of the scrubbed gas.
Table 6-1 classifies the heat exchanger type of reheat into three
subdivisions:
Direct steam
Intermediate hot air
Direct hot water
The direct steam reheater consists of a heat exchanger installed in
the duct containing the flue gas. Condensing steam inside the heat
exchanger heats the flue gas to the desired reheat temperature.
6-1-9'
-------
Table 6-1
REHEAT USING HEAT EXCHANGERS
Power Plant
La Cygne(,a)
Kansas City Power
and Light
Will County
Commonwealth
Edison
Cholla ,
Arizona Public
Service
Mohave Vertical
Module ,
Southern California
Edison
Mohave Horizontal
Module,
Southern California
Edison
Cols trip.
Montana Power
Valmont-5,
Public Service
Company of
Colorado
Hawthorn,
Kansas City Power
and Light
Lawrence ,
Kansas Power and
Light
Sherburne County,
Northern States
Power
Four Corners,
Arizona Public
Service
Reid Gardner,
Nevada Power
Type of
Reheat
Useful
Energy ,
Direct
Useful
Energy ,
Direct
Useful
Energy ,
Direct
Useful
Energy ,
Direct
Useful
Energy,
Inter-
mediate
Hot Air
Useful
Energy ,
Direct
Useful
Energy,
Direct
Useful
Energy,
Direct
Useful
Energy ,
Direct
Useful
Energy ,
Direct
Useful
Energy,
Inter-
mediate
Hot Air
Useful
Energy,
Inter-
mediate
Hot Air
Heating
Source
Steam @
140 psig &
690°F
Steam @
350 psig &
485°F
Steam @
250 psig &
405°F
Steam @
600 psig &
650°F
Steam @
600 psig S
650°F
Steam @
150 psig
360°F
Steam @
450 psig
900°F
Hot Water @
325°F
Hot Water @
250°F
Hot Water @
350°F
Steam @
600 psig &
650°F
Steam @
450 psig &
760°F
Tube
Style
Bare
Bare
Bare
Bare
Finned
Plate
Finned
Finned
Finned
Finned
Finned
Finned
Materials of
Construction
316LSS(b)
Bottom Section,
316LSS
-------
Table 6-2
REHEAT USING COMBUSTION GAS
Location of ,. Degree of
Power Plant the Combus- £pe Reheat
tion Chamber e AT°F
Eddystone 1A,
Philadelphia Electric
Phillips,
Duquesne Power
Elrama,
Duquesne Power
Bruce Mansfield,
Pennsylvania Power
St. Clair 6,
Detroit Edison
Paddy's Run 6,
Louisville Gas and Electric
*
D. H. Mitchell
Northern Indiana Public
Service
External'
Internal
Internal
Internal
External
Internal
Internal
Oil
Oil
Oil
Oil
Oil
Gas
Gas
100
30
30
NA
100
40
50
This reheat system has not yet been put in service.
6-2x
-------
Intermediate hot air for reheat is attained by passing outside air
through condensing steam in a heat exchanger and mixing with flue gas.
Direct hot water reheat is similar to the inline steam approach, except
that lower temperature differences are involved. The heat transfer
fluid is pressurized hot water.
Flue gas and intermediate air heat exchangers are normally tubular in
shape. Extended surface fins have been used to enhance heat transfer.
The following materials of construction have been used (see Table 6-1):
carbon steel, 304 stainless steel, 316 (low carbon) stainless steel,
Corten, Inconel 625, and Hastelloy G. The relative effectiveness of
these materials for reheater service is discussed in the following
section. Typically, the tubes are 1 inch or less in diameter. Plate
coils have also been used.
The degree of reheat for the installations listed in Tables 6-1 and 6-2
varies from 17 to 56 C (30 to 100 F). The amount of reheat required is
site specific, and there is no known correlation between generating
stations.
The problems associated with each method of reheat are distinct and
will be discussed as related to the following categories:
Direct steam and hot water reheat (Tables 6-3 and 6-4)
Intermediate heated air reheat (Table 6-5)
Gas and oil fired reheat (Table 6-6)
Direct reheaters have failed because of plugging, corrosion, and
vibration. Plugging occurs from entrainment. Once the scrubber
liquor deposits on the reheater, the dissolved and suspended solids
bake onto it. This deposit continues to grow, blocking the gas flow,
increasing the pressure drop, and helping to induce corrosion from
the localized high temperature and concentrated salts. The effective-
ness of the heat transfer surface is progressively reduced as the
deposit builds up.
622
-------
Table 6-3
DIRECT STEAM REHEAT OF SCRUBBED FLUE GASES
Power Plant
La Cygne,
Kansas City Power
and Light (7,8)
Will County,
Commonwealth Edison
(9,10,11)
Reheater Problem
The 304SS reheater
failed because of chlo-
ride stress corrosion
and acid condensate
corrosion.
The reheater plugged
with scrubber slurry
passed through the
mist eliminator.
Module A Reheater;
Reheater tubes failed
because of chloride
corrosion (1972-1974).
The reheater also
plugged with solids
passed through the
mist eliminator.
Module E Reheater:
Reheater tubes failed
because of vibration
during high gas flow
conditions (1972).
The reheater failed
because of chloride
corrosion (1973).
Note: Module B was
shut down April, 1973
and not restarted until
May, 1975.
Solution Attempted
Replaced the 304SS reheater
tubes with 316LSS tubes.
Hot air (supplied from the
boiler's air heater) was
injected below the reheater
to prevent acid condensation.
Increased the reheater flue
gas temperature from 147°F
to 175°F. The utility ex-
pects that 316LSS will with-
stand corrosion at 175°F.
Modified and corrected the
operating procedures relat-
ing to the mist eliminator
and reheater (details not
available) .
Modified the steam soot
blower.
One module is cleaned each
night (there are seven
modules) .
No corrective action taken
at the time of tube failure.
Module A reheater was put
back in service by cannibal-
izing the Module B reheater.
In addition, the mist
eliminator wash method was
improved by constant under-
spray and an intermittent
overspray.
The mist eliminator wash-
water was changed from pond
reclaim water to service
water.
Steam was injected during
scrubber outages to slow
down the tube deterioration.
A second stage demister was
installed above the original
demister to prevent wash-
water carryover to the
reheater.
The reheater was rebraced
and a baffle installed to
reduce the vibrations.
The Module B reheater was
rebuilt based on the expe-
rience gained on Module A.
316SS and carbon steel tubes
replaced the 304SS and
Corten tubes.
Solution Follow-Up
Not available
Intermittent heavy over
spray and continuous under-
spray have kept the mist-
eliminator relatively clean.
The soot blower modifica-
tion aided in keeping the
reheater clean.
Not available
The corrective actions re-
lieved but did not completely
eliminate corrosion.
Wyoming low sulfur coal
was burned during 1974; there
were very few problems.
Once the station started
burning high sulfur coal in
mid-1975, the mist eliminator
and reheater problems
reappeared and the module
was shut down for the balance
of 1975.
During mid-1975 the reheater
and mist eliminator were
replaced .
Not available
Carbon steel reheater tubes
failed.
6-23
-------
Table 6-3 (Cont'd)
Power Plant
Reheater Problem
Solution Attempted
Solution Follow-Up
Cholla,
Arizona Public
Service (12,13,14)
Mohave Vertical
Module,
Southern California
Edison (15)
Reheater tubes failed
because of vibration
(<1,000 hours opera-
tion) . The tube sup-
ports had worked loose
permitting the tubes
to vibrate.
The carbon steel tube
support frames
corroded.
Reheater plugging
occurred in the same
manner as Module A.
Module A Reheater;
Harmonic vibrations
with deflections as
much as 0.040 inches
occurred in the re-
heaters. The vibration
was attributed to im-
proper flue gas
distribution.
Little information is
available on Module A
reheater corrosion.
Most corrosion prob-
lems occurred in the
Module B reheater.
Module B Reheater:
Harmonic vibrations
also occurred in the
Module B reheater.
Both chlorides and sul-
furous acid corroded
the reheater. Sul-
furous acid formed on
the duct walls and
dripped onto the re-
heater tubes. Chlo-
ride attack occurred
when entrained water
containing chlorides
impinged on the tubes.
The reheater housing
also corroded.
There was no problem
with the vertical mod-
ule reheater other than
a few seams failing on
the seamless rated
tubing.
No reheater plugging
occurred during the
testing.
The tube supports were re-
designed and the carbon
steel tube bundles repaired.
No information available.
The same solutions were
followed for both Modules
A & B.
Baffles were installed.
Not available.
Baffles were installed.
To prevent sulfurous attack
the ducts above the re-
heater were insulated.
In addition a trough was
built around the duct to
catch any run off before
it reached the reheater.
No solution postulated for
chloride corrosion.
None
No reported reheater plug-
ging even when the mist-
eliminator scaled and plugged
while burning high sulfur
coal in mid-1975.
The baffles reduced but did
not eliminate the vibration,
but it was no longer con-
sidered a major problem.
The baffles reduced but did
not eliminate vibration.
It is the intent that new
reheat bundles would be
of Inconel 625 rather than
the original 316LSS.
Upon completion of the
testing, the tubes were in-
spected. Pits were detected
in the "E-Brite" 26-Cr 1 Mo
tubes.
6-24
-------
Table 6-3 (Cont'd)
Power Plane
Colstrip 1 and 2,
Montana Power ( I*)
Valmont 5,
Public Service
Company of
Colorado (14,16)
Reheater Problem
No problems have been
reported with the
inconel 625 and
Hastelloy G plate type
reheaters.
Some loose scale formed
on the reheaters but
did not cause any oper-
ating problems.
No reported reheater plug-
ging or corrosion during
short-terra SO, removal
tests. The scrubbers are
normally used for particu-
late removal only.
Solution Attempted
Solution Follow-Up
6-25
-------
Table 6-4
DIRECT HOT WATER REHEAT
Power Plant
Hawthorn,
Kansas City Power
and Light
(10,14,1?)
Lawrence,
Kansas Power and
Light
(6,10,14,18,19)
Sherburne,
Northern States
Utilities (14)
Reheater Problem
Hawthorn No. 3 Reheater
No major reheater prob-
lem has been reported.
Hawthorn No. 4 Reheater
No reheater corrosion
problem reported. Re-
heater plugging has
been a problem par-
ticularly in the B
side of the scrubber.
Lawrence No. 4 Reheater
This reheater plugged
frequently during the
early life of this
system.
Carbon steel reheater
completely failed
(1976).
Lawrence No. 5 Reheater
No problems reported on
this reheater. The
corrections made on
No. 4 were incorporated
into the Lawrence No . 5
scrubber system.
There has been multiple
failure of the carbon
steel finned tubes.
Reheater plugging is a
common event .
Solution Attempted
A section of the reheater
was removed to facilitate
cleaning and maintenance.
To alleviate the plugging
problem the following
changes were made:
a. Soot blowers were in-
stalled beneath the
reheater.
b. The mist eliminator was
redesigned.
c. Vanes were installed
under the marble bed to
improve gas distribution.
This scrubber will be com-
pletely razed and replaced
by a venturi rod scrubber
followed by a spray tower
(to be completed in 1977).
No information available
The scrubbers are shut down
on a three-day cycle for
cleaning.
Solution Follow-Up
It is normal practice to
shut down the scrubber
every three days for clean-
ing of the mist eliminator.
At that time the reheater
may be cleaned.
The soot blower pressure
flattened the copper fins.
The copper finned tubes
were replaced with carbon
steel finned tubes.
Mist eliminator plugging
continued to be a recurring
problem.
6-26
-------
Table 6-5
STEAM HEATED AIR REHEAT
Power Plant
Reheater Problem
Solution Attempted
Solution Follow-Up
Mohave Horizontal
Module,
Southern California
Edison (20)
Four-Corners,
Arizona Public
Service (21)
Reid Gardner,
Nevada Power
(14,22)
There were no reported
problems with this re-
heat system.
Horizontal Scrubber
Reheater:
The carbon steel tubes
developed pinholes.
Metallurgical examina-
tion revealed that the
pinholes developed from
inside the tube as a
result of stagnant con-
densate and air that
leaked Lntu thu reheat
system.
It is not known when
the pinholes developed
but the system had been
idle for a year prior
to testing. The pin-
holes probably devel-
oped during this idle
period.
Reheater leaks have
been reported but
their severity has
not been indicated.
The pinholes were brazed so
that the system testing
could continue. A nitrogen
blanket system was installed
in the reheater loop to ex-
clude air from the inside of
the tube.
The scrubber system was
moved to the Four-Corners
Generating Station for ad-
ditional testing.
The horizontal scrubber has
been shut down indefinitely.
There is no feedback as to
the nitrogen blanket
performance.
No information available
6-27
-------
Table 6-6
GAS AND OIL FIRED REHEAT
Power Plant
Reheater Problem
Solution Attempted
Solution Follow-Up
Eddystone 1A,
Philadelphia Electric
(14,23,24)
Phillips,
Duquesne Light
(6,14,25,26,27)
Elrama,
Duquesne Light
(14,27)
Bruce Mansfield
No. 1,
Pennsylvania Power
(28)
St. Clair 6,
Detroit Edison
(6,14)
Paddy's Run 6,
Louisville Gas and
Electric (14)
D. H. Mitchell,
Northern Indiana
Public Service
(30,31)
The combustion chamber
double brick refractory
lining failed. The
failure occurred be-
cause of torch place-
ment and problems with
the ultra violet flame
scanners. Also a por-
tion of the duct was
insulated which limited
heat dissipation and
caused subsequent fail-
ure of the refractory
and steel sheet.
Problems occurred with
the oil pumps, burners
and temperature control.
Corrosion has been re-
ported in the combus-
tion chamber.
Problems similar to
Phillips Power
Station's.
the
Combustion problems
have occurred because
of the flue gas mixing
with the combustion
air (i.e., problems in
maintaining a flame in
the combustion chamber.
Reheater vibration is
also a problem.
The only problem re-
ported is the failure
of a thermal controller
resulting in liner
damage.
In addition, the fire-
bricks in the flue gas
mixing zone fell off
because of vibration.
No reported problems
Reheat system has
not been put in
service.
No information available.
No information available.
(No priority has been
placed on operating the
reheater because of the oil
shortage.)
No information available.
These problems are to be
resolved by the equipment
supplier.
A new acid proof
stack liner has been
installed.
This reheat system has hardly
been operated. The scrubber
functions with a wet stack.
No information available.
No information available.
6-28
-------
Corrosion originating in this manner has led to serious failures.
The scrubber liquor usually contains dissolved chlorides that concen-
trate within the deposits and attack the metal. Stainless steel,
particularly type 304, is vulnerable to stress corrosion in the pres-
ence of chloride. Gaseous S02 also slowly attacks carbon steel and
Corten above the acid dewpoint and rapidly below. Sulfurous acid from
this source will condense on cold reheater tubes (when operated without
steam or hot water) and on cold duct surfaces during startup or low
load periods. If the reheater is in a vertical duct, acid condensate
formed on the duct will run down the duct onto the reheater tubes.
Reheaters have failed from vibration fatigue. If the natural fre-
quency of the reheater matches that of the fan, reheater tubes may
wear against their support baffles and fail.
The major problem with hot air injection systems, to date, has been
condensate corrosion. Typically, the failure has been with carbon
steel, a particularly vulnerable material.
With oil-fired systems the problems appear in the course of attempting
to maintain the flame within the main flue gas stream. This saves
space, but typically does not work well. Refractory failures have
occurred in the combustion chamber from flame impingement, attack from
condensation during downtime, or other reasons. Brick linings have
also failed from vibration. Most refractories are weak in tensile
strength, and both rapid heating and vibration tend to cause failure.
Burning oil in the presence of flue gas causes complications in light-
ing off the burner, flame-outs during operation, and formation of soot.
Burning natural gas in the flue gas stream is reported to have been
done successfully at the Louisville Gas and Electric Paddy's Run Station.
6-29
-------
DESIGN, OPERATION, AND MAINTENANCE OF REHEATERS
The previous section outlined typical problems that have occurred in
operating installations with reheaters. Avoiding these problems is
best done by making use of operating experience.
No particular method of reheat can be endorsed for general use. A
well designed, operated, and maintained reheat system should be able
to function with a minimum of problems. Selection .of a system is
increasingly a matter of sound design.
Design, operation, and maintenance of reheaters fall into the three
categories already identified:
e Direct steam and hot water
Intermediate hot air
Gas and oil combustion gas
Direct Steam and Hot Water Reheaters
Direct steam and hot water reheaters have experienced three major
problems: plugging, corrosion, and vibration.
Reheater Plugging. As described in the previous section, reheater
plugging is the result of scrubber liquor containing suspended and
dissolved solids passing through the mist eliminator and impinging on
the reheater tubes or plates. The purpose of the mist eliminator is
to prevent the scrubber liquor from leaving the scrubber. If the
mist eliminator allows scrubber liquor to pass, then it requires
improved design, operation, or maintenance. Proper mist eliminator
design is essential to prevent plugging. The following design param-
eters are critical:
Velocity through the mist eliminator
Material of construction
6-30
-------
Blade spacing (if any)
Number of stages
Distance between the upper stage of the scrubber
and the mist eliminator, or location within the
scrubber
The Colstrip and Mohave stations are reported to have been relatively
successful with direct steam reheater operation by correct integration
of the above factors into the mist eliminator design.
Most direct reheaters incorporate a steam or air soot blower to clean
any plugs that form. The quantity of steam or air used and the fre-
quency of cleaning (most stations clean at least once every 8 hours)
vary from station to station. Soot blowers are a standard feature
for maintaining clean reheat surfaces; operating cycles are best de-
rived from experience. <
The best method to alleviate plugging is to minimize entrainment.
Corrosion. Direct steam and hot water reheaters have suffered
repeated construction material failure because of corrosion. Oper-
ating experience suggests that certain materials should not be used,
such as carbon steel, 304SS, 316SS, and Corten. More exotic materials
such as Inconel 625 and Hastelloy G have been used successfully at
Colstrip. The use of these materials would approximately double the
cost of a 316LSS reheater.
So far, corrosion and pitting have been attributed to periodic acid
conditions and to chlorides. Effective design of the mist eliminator
will minimize these.
6-31
-------
Condensation can be avoided by keeping the reheat bundles hot and
by insulating ductwork downstream of the reheater. Provisions may
be desirable in future designs to keep the reheater warm during pro-
longed outages.
Vibration. Design against vibration is a structural matter depending
on the span of the tubing and its section modulus. A frequency analy-
sis is indispensable for proper design.
Operation and Maintenance. A well designed scrubber system has ample
lighting and access doors near the mist eliminator and reheater for
inspection and maintenance. Warning instruments are of great value.
The soot blower cleaning cycle must be designed with enough latitude
to meet actual service requirements.
Intermediate Hot Air Reheaters
Intermediate hot air reheaters have had fewer problems than inline types.
The reheaters are outside the flue gas duct so that slurry never reaches
the reheater surface. In one case corrosion of tubes has been experienced
(see Table 6-5).
Flue gas leakage into dead-ended cold reheaters can be avoided by
installing shutoff dampers between the in-service duct and the idle
reheater and by purging the idle zone with air or gas. In addition,
low pressure steam can be bled into the tubes to keep them hot. With
such precautions, even carbon steel may demonstrate acceptable service
life for some installations.
The temperature of the hot air before mixing is higher than 200 C
(400 F) and can destroy the usual coatings used to protect ducts and
scrubber walls. Hot air must be prevented from entering the system
without cold gas flow. This requires suitable interlocks.
6-32
-------
The same precaution should be exercised as with inline steam coils
in preventing tube vibration in the air heater.
Because the reheat equipment is outside the flue gas duct, hot air in-
jection and maintenance can be somewhat simpler than for the direct system.
Oil- and Gas-Fired Reheaters
The major problems in oil-fired reheaters have been those of arrange- ,
ment installing the combustion chamber too close to or inside the
flue gas duct. Oil-fired reheat systems should be designed with an
external combustion chamber so that complete combustion takes place
before the hot gas mixes with the flue gas, and the provision should be
made for off-line light-off; e.g., short stack.
A flame temperature of 1,600 C (3,000 F) can warp metals, destroy coat-
ings, and crack refractories. Combustion chambers must be amply sized,
and interlocks must be provided so that the reheater cannot be started
unless there is flue gas in the duct and the scrubber liquor is circu-
lating. Careful specification of refractory and subcomponents is
essential.
There have been no significant problems reported with the gas-fired
reheat system at Paddy's Run.
ALTERNATIVES TO REHEAT
A number of alternatives to reheating are available. Whether or not
they are feasible must be considered on a case by case basis. A major
factor in their applicability is whether a scrubber is being installed
as a retrofit or is being designed for a new plant.
Tall Stacks
To the extent that excessive ground concentration of pollutants is a
problem (due to reduced plume buoyancy from scrubbing), one alternative
6-33
-------
to reheating is to build a taller stack. A taller stack could be more
economical than reheating even though it involves a high capital cost.
There is, by comparison, no energy cost. However, under certain cir-
cumstances the stack height required for a particular location might
not be feasible. Meteorological modeling is a useful tool for deter-
mining the validity of such an alternative, but most dispersion models
have not been developed for wet plumes.
Corrosion Protection
To limit corrosion, one may either select materials that are inherently
resistant to corrosion or use coatings to cover materials subject to
corrosion.
Corrosion resistant materials include some costly alloys, such as the
high-chromium stainless steels. Nonmetallic materials resistant to
wet corrosion include fiberglass or fiberglass reinforced polyesters
(FRP) for use as linings for stacks and critical sections of ductwork.
However, there is an upper temperature limit on such materials in the
range of 100 to 150°C (200 to 300°F), well below that of stainless steel.
A number of noncorrosive coatings and linings include materials that
can be sprayed on, painted on, or prefabricated and bonded in place.
Polyesters, silicate-based inorganics, epoxies, and other plastics are
available. Rubber linings have long been used. Each has advantages
and disadvantages based on the material performance and cost for the
specific application.
A key problem is that of quality control of application of the lining.
The surface preparation and bonding must be timed for periods of warmth
and low humidity. Once applied, bonding adequacy is very difficult to
determine. If bond failure occurs and corrosion is discovered, repairs
can be costly. Further progress in the areas of lining bond quality
measurement and corrosion detection is desirable.
6-34
-------
Upstream Fan
If the purpose of reheat is to protect a downstream fan, an obvious
alternative is to place a fan upstream of the scrubber. This is only
feasible with an upstream collector or precipitator to remove erosive
particulate matter.
Mitigation of Plume Appearance
There is very little, other than reheat, that is effective in mitigat-
ing the vapor plume. However, plume appearance and, in particular,
the length of a visible plume, are strong functions of atmospheric
conditions.
Figures 6-10 and 6-11 show the calculated behavior of plumes for a
Southwestern United States 2,000 MW coal-fired power plant with sulfur
dioxide scrubbers (Ref, 32). These figures show the frequency of plumes
of various lengths. For both figures, Curve A shows the frequency dis-
tribution of plumes of a given length in the absence of reheat. It
may be noted, for example, that only 11 percent of the time are plumes
longer than 1 kilometer. Thus, if visibility is considered a problem
only for plumes longer than a few kilometers, reheat may be unnecessary
except for a small fraction of the time.
These same figures also show the effect of reheat to a given tempera-
ture on the frequency distribution of visible plumes. Figure 6-10
refers to direct reheat, meaning in this case any form of reheat that
does not add mass to the flue gas, while Figure 6-11 refers to indirect
reheat where it is assumed that air is first heated, then mixed with
the flue gas. These curves show that a substantial increase in temper-
ature is required to affect the frequency distribution significantly.
For example, to reduce the frequency of plumes as long as 1 kilometer
from 11 percent to 2 percent, it is necessary to apply almost 95 C
(170°F) of reheat. It can also be seen from these figures that for a
given amount of reheat, the indirect method is -more effective in
6-35
-------
reducing or mitigating visible water vapor plumes. However, it must be
kept in mind that the indirect application requires a higher energy in-
put for the same amount of temperature difference. It turns out that
the same effect on visibility requires approximately the same energy
input.
The frequency distribution curves do indicate, however, that if the
purpose of reheat is a cosmetic one (reduction of the frequency of
visible plumes), a logical technique would be to use'variable reheat.
That is, reheat could be limited to those periods of atmospheric con-
ditions during which a plume of objectionable length would otherwise
occur.
Mist Elimination
As indicated previously, absence of reheat may be accompanied by rain
or stack ice formation. Reheat, however, is costly and not necessarily
the best answer to such problems. A preferred alternative would be to
use more efficient mist eliminators to remove droplets of large size.
Heat transfer calculations can show that reheat consumes very large
amounts of energy before reducing significantly the size of large droplets.
One form of liquid emission that would not be affected by better mist
elimination results from condensate forming on the inside of the stack
and brought out of the stack by the rising flue gas. Such an effect
can be reduced or eliminated by the combination of insulation on the
stack to reduce condensation likelihood, plus the addition of liquid
catchers on the upper portion of the stack.
It can also be shown that the formation of large droplets in the atmos-
phere by agglomeration and growth is highly unlikely in the absence of
large droplet emission from the stack. Such formation should not con-
tribute significantly either to stack icing or rain.
6-36
-------
cr.
UJ
4.0
3.0 -
LLJ
D
o.
LU
m 2.0
LL
O
X
LU
1.0
0
0.01
0.1
C B A
FLUE GAS
TEMPERATURE
A 130 (NO REHEAT)
B 210
C 300
2.0
10 20 40 80
FREQUENCY OF LONGER PLUMES ,
98
2.5
LU
1.5 5
CO
1.0
e?
z
LU
0.5
99.9 99.99
Figure 6-10. Frequency of Occurrence of Visible Plumes with Direct Reheat
-------
CO
CO
FLUE GAS
TEMPERATURE
B
130 (NO REHEAT)
210
300 NOT VISIBLE
-- 2.0
2.5
1.0
LU
1-B
Li.
O
I
CJ
z
LU
0.5
0.01 0.1
10 20 40 80
FREQUENCY OF LONGER PLUMES , %
98
99.9 99.99
Figure 6-11. Frequency of Occurrence of Visible Plumes with Indirect Reheat
-------
Opacity Measurement Technology
It has been suggested that one value of reheat lies in the fact that
it simplifies the measurement of particle emissions without inter-
ference from liquid droplets or aerosols. While that is true, it
would probably be possible to measure particle concentration even in
the presence of condensed moisture. Techniques have been devised
that employ an optical sensing device in a reheated bleed stream,
for example. If all other arguments for reheating were eliminated,
using it solely to measure opacity would constitute the most expen-
sive opacity measurement of all.
REHEAT SUMMARY
This section has shown that while reheat of flue gas may have signifi-
cant benefits, it is clearly expensive. In some cases, particularly
with low sulfur, reheat can be more expensive than the alkali absorb-
ing costs themselves. In addition, it has been shown that most of
the benefits can be realized by alternative methods. Reheating should
not be considered as a necessity, but as one of a number of approaches
for consideration in optimizing sulfur dioxide absorption systems.
6-39
-------
REFERENCES FOR SECTION 6
1. Slack, A.V., and G.A. Hollinden, Sulfur Dioxide Removal from
Waste-Gases, Noyes Data Corporation, Park Ridge, New Jersey,
1975.
2- Bechtel Corporation, EPA Alkali Scrubbing Test Facility: Advanced
Program Third Progress Report, EPA Report, EPA-600/7-77-105,
September 1977.
3. Lisle, E.S., and J.D. Sensenbaugh, "The Determination of Sulfur
Trioxide and Acid Dew Point in Flue Gases," Combustion, January
1965, pp 12-15.
4. Ferrel, J.D., and E.W. Stenby, "Interface Design Problems Between
Coal Fired Boilers and S02 Scrubbing Systems," The Second Pacific
Chemical Engineering Congress, Denver, Colorado, August 28-31, 1977.
5. Slack, A.V., "Design Considerations in Lime-Limestone Scrubbing,"
The Second Pacific Chemical Engineering Congress, Denver, Colorado,
August 28-31, 1977.
6. Battelle Columbus Laboratories, Stack Gas Reheat For Wet Flue Gas
Desulfurization Systems, EPRI Report, FP-361, February 1977.
7. Isaacs, Gerald A., "Survey of Flue Gas Desulfurization, La Cygne
Station, Kansas City Power and Light Company and Kansas Gas and
Electric Co.," PB244-401 NTIS, V.S. Department of Commerce,
5285 Port Royal Road, Springfield, VA 22151, July 1975.
8. McDaniel, C.F., "Wet Scrubber Operating Experience at La Cygne
Station, No.l," EPA Flue Gas Desulfurization Symposium, Atlanta,
Georgia, November 1974.
9. Isaacs, Gerald A., "Survey of Flue Gas Desulfurization, Will County
Station, Commonwealth Edison Company," PB246-851 NTIS, U.S. Depart-
ment of Commerce, 5285 Port Royal Road, Springfield, VA 22151,
October 1975.
10. Devitt, T. W. and Zada F., "Status of Flue Gas Desulfurization
Systems in the United States," EPA Flue Gas Desulfurization Sym-
posium, Atlanta, Georgia, November 1974.
11. Stober, W. G., "Operational Status and Performance of the Common-
wealth Edison Company, Will County, Limestone Scrubber," EPA
Flue Gas Desulfurization Symposium, New Orleans, March 1976.
6-40
-------
12. Isaacs, Gerald A., "Survey of Flue Gas Desulfurization, Cholla
Power Generating Station, Arizona Public Service Company,"
PB244-141, NTIS, U.S. Department of Commerce, 5285 Port Royal
Road, Springfield, VA 22151, June 1975.
13. Mundth, L. K., "Operational Status and Performance of the Arizona
Public Service Company Flue Gas Desulfurization System at the
Cholla Station," EPA Flue Gas Desulfurization Symposium,
Atlanta, Georgia on November 1974.
14. Pedco Environmental Specialists, Inc., Summary Report Flue Gas
Desulfurization Systems, November-December 1976, 11499 Chester
Road, Cincinnati, Ohio 45246.
15. Bechtel Power Corporation, Navajo/Mohave 170 MW Vertical Test Module,
Inspection after Test Block 9, Norwalk, California, August 1975
Page 46.
16. Ness, Harvey M., Sondreal, Everett A., and Tufte, Philip H., "Status
of Flue Gas Desulfurization using Alkaline Fly Ash from Western
Coals," EPA Flue Gas Desulfurization Symposium, New Orleans, March
1976.
17. Isaacs, Gerald A., "Survey of Flue Gas Desulfurization Systems
Hawthorne Station, Kansas City Power and Light Company," PB246-
629 NTIS, U.S. Department of Commerce, 5285 Port Royal Road, Spring-
field, VA 22151, October 1975.
18. Isaacs, Gerald A., "Survey of Flue Gas Desulfurization Systems,
Lawrence Power Station, Kansas Power and Light Company," PB246-849
NTIS, U.S. Department of Commerce, 5285 Port Royal Road, Spring-
field, Va 22151.
19. Miller, D. M., "Recent Scrubber Experience at the Lawrence Energy
Center, the Kansas Power and Light Company," Symposium on Flue
Gas Desulfurization, New Orleans, March-1976.
20. Weir, A., Papay, L.T., Jones, D.G., Johnson, J.M. , Martin, W.C.,
"Results of the 170 MW Test Modules Program, Mohave Generating
Station, Southern California Edison Company," EPA Flue Gas Desul-
furization Symposium, New Orleans, March 1976.
21. Bechtel Power Corporation, Norwalk, California, inhouse information.
22. Gerstle, Richard W., "Survey of Flue Gas Desulfurization, Reid
Gardner Station, Nevada Power Company," PB-246-852 NTIS, U.S.
Department of Commerce, 5285 Port Royal Road, Springfield, Va
22151, October 1975.
6-41
-------
23. Gille, James, "Magnesium Oxide Scrubbing at Philadelphia Electric1s
Eddystone Station," EPA Flue Gas Desulfurization Symposium, New
Orleans, March 1976.
24. Isaacs, Gerald A., "Survey of Flue Gas Desulfurization Systems, '.
Eddystone Station, Philadelphia Electric Company," PB247-085 NTIS,
U.S. Department of Commerce, 5285 Port Royal Road, Springfield,
Va, 22151, September 1975.
25. Isaacs, Gerald A., "Survey of Flue Gas Desulfurization Systems,
Phillips Power Station, Duquesne Light Co.," PB246-285 NTIS, U.S.
Department of Commerce, 5285 Port Royal Road, Springfield Va,
July 1975.
26. Pernick Jr., S.L. and Knight, R.G., "Duquesne Light Company, Phil-
lips Power Station, Lime Scrubbing Facility," EPA Flue Gas Desul-
furization Symposium, Atlanta, Georgia, November 1974.
27. Knight, R.G. and Pernick S.L., "Duquesne Light Company Elrama and
Phillips Power Stations Lime Scrubbing Facilities," EPA Flue Gas
Desulfurization Symposium, New Orleans, March 1976.
28. Private Communication, Bechtel Corporation, Research and Engineering,
San Francisco, California.
29. Isaacs, Gerald A., "Survey of Flue Gas Desulfurization System,
Paddy's Run Station, Louisville Gas and Electric," PB246-136 NTIS,
U.S. Department of Commerce, 5285 Port Royal Road, Springfield, Va.,
August 1975.
30. Mann, E.L., "Power Plant Flue Gas Desulfurization by the Wellman-
Lord S02 Process, Part 1, Dean Mitchell Station," EPA Flue Gas
Desulfurization Symposium, Atlanta, Georgia, November 1974.
31. Mann, E.L., "Status of Demonstration of the Wellman-Lord/Allied
F.G.D. System at NIPSCO's, D.H. Mitchell Generating Station,"
EPA Flue Gas Desulfurization Symposium, New Orleans, March 1976.
32. Bechtel Power Corporation, Norwalk, California (calculations
developed for Nevada Power, Harry Allen Station, Environmental
Impact Statement).
6-42
-------
Appendix A
DATA FROM EPA ALKALI SCRUBBING
TEST FACILITY
These data were obtained at the EPA Alkali Scrubbing Test Facility
located near Paducah-, Kentucky. The test facility is integrated into
the flue gas ductwork of a 150-MW coal-fired boiler at the TVA Shawnee
Power Station. Two parallel wet scrubber systems are in operation:
a venturi/spray tower system and a Turbulent Contact Absorber (TCA) sys-
tem. Each system can treat approximately 10 MW equivalent (33,000 Nm3/hr,
or 30,000 acfm @ 300°F) of flue gas containing 1,500 to 4,500 ppm of SO .
Figures A-l and A-2 (drawn with major dimensions to scale) show the two
n
scrubber systems. The cross-sectional area of the spray tower is 6 m
2
(50 ft ) in both the scrubbing section and in the mist eliminator. The
2 2
cross-sectional area of the TCA scrubber is 3.0 m (32 ft ) in the scrub-
2 2
bing section and 4.6 m (49 ft ) in the mist eliminator section.
Figures A-3 through A-27 show the results of short-term factorial tests
6 to 8 hours in length and long-term tests about one week in length.*
The short-term tests were aimed at determining S02 removal efficiencies
without regard necessarily for possible long-term scaling conditions
that might be associated with the process conditions chosen. Therefore,
these particular data are not to be interpreted as evidence of long-
term reliable operation. However, during other portions of the Shawnee
program adequate demonstrations have been made that these FGD systems
can be operated for extended periods without problems with respect to
scaling.
*Bechtel Corporation, EPA Alkali Scrubbing Test Facility: Advanced
Program Third Progress Report, EPA Report EPA-600/7-77-105, September 1977.
A-l
-------
For the spray tower and TCA data, the S0? removal curves in the figures
have been produced by mathematical models which are presented in
Appendix B.
A-2
-------
YENTURI SCRUBBER AND SPRAY TOWER
Chevron Mist
Eliminator
SPRAY TOWER
INLET SLURRY
GAS OUT
MIST ELIMINATOR
WASH WATER
MIST ELIMINATOR
WASH LIQUOR
Adjustable Plug
VENTURI SCRUBBER
EFFLUENT SLURRY
Figure A-l. Schematic of Venturi Scrubber and Spray Tower
A-3
-------
TCA SCRUBBER
GAS OUT
tLimiNAlUK J
SH WATER *h A A A
XX&X&XX&XX
I
Chevron Mist / \
Eliminator
Datnininn Rnr nrMc f
neiaining por-grios \
GAS IN h
'
Y Y / MIST ELIMINATOR
1 1 H4 WASH LIQUOR
M^ luirv fi unnvr
A A A
- -o~0- -
° O
o o ^ o
v o o
00 0
o.p_o o. o
/o o
°00 ° °
\° o o ° o
QS^Q Q Q Q
O
O o
OO o
N /
1 '
/ Mobile Packing Spheres
^s^
5'
1 i
APPROX. SCALE
EFFLUENT SLURRY
Figure A-2. Schematic of Three-Bed TCA
A-4
-------
60
50 4-
u
UJ
UJ
cc
(M
O
CO
I-
z
UJ on
O 20
CC
UJ
a.
10
SO2 INLET CONCENTRATION = 2,650 - 2,950 ppm
SCRUBBER INLET LIQUOR pH = 5.7- 5.8
VENTURI A P = 9 IN. HjO
GAS FLOW RATE,
acfm @ 330 °F
35,000
27.500
20,000
SYMBOL
O
A
D
10 20 30 40
VENTURI LIQUID-TO-GAS RATIO, gal/Mcf
50
Figure A-3. Effect of Liquid-to-Gas Ratio and Gas Flow Rate
on S02 Removal Venturi Scrubber with Limestone
A-5
-------
70
SO2 INLET CONCENTRATION = 2,650 - 2,950 ppm
SCRUBBER INLET LIQUOR pH = 5.7-5.8
60
VENTURI AP = 9 IN. H2O
SLURRY FLOW RATE, gal/min. SYMBOL
600 O
300 A
u
uu
u
50 --
40
ID
CC
co 30
H
LLI
O
CC
20 -
10 - -
1
35,000
20,000
27,500
GAS FLOW RATE, acfm @ 330 °F
Figure A-4. Effect of Slurry and Gas Flow Rates on SC>2
Removal Venturi Scrubber with Limestone
A-6
-------
60
50
40
O
O
iZ
u_
in
O
ui
DC
CM
I-
8 20
cc
LU
a.
30
10
5.0
SO2 INLET CONCENTRATION = 2,440-2,880 ppm
LIQUID - TO - GAS RATIO = 27gal/Mcf
VENTURI AP = 9in. H2O
GAS FLOW RATE = 27,500 acfm @ 330° F
5.2 5.4 5.6
SCRUBBER INLET LIQUOR pH
5.8
6.0
Figure A-5. Effect of Scrubber Inlet Liquor pH on S02
Removal Venturi Scrubber with Limestone
A-7
-------
60
50
u
z
LU
o 40 +
H
LL
LU
LU
QC
CO
Z
LU
U
oc
LU
O.
30
20
10
O (5.8)
(5.6)
(5.7)
SO2 INLET CONCENTRATION = 2,550-2,800 ppm
LIQUID - TO - GAS RATIO = 27 gal/Mcf
GAS FLOW RATE = 27,500 acfm @ 330° F
SCRUBBER INLET LIQUOR pH IN PARENTHESES
7 8 9 10 11
VENTURI AP, in. H2O
12
13
Figure A-6. Effect of Venturi AP on S02 Removal Venturi
Scrubber with Limestone
A-8
-------
100
o
z
LU
o
Ul
O"
CO
Z
LU
tt
LU
O.
90 +
80 4-
70 +
LU
CC 60 4-
CM
50 4-
40 +
30
1 ' ' I T
S02 INLET CONCENTRATION = 2,140 - 3,630 ppm
SCRUBBER INLET LIQUOR pH = 5 7 - 5 3
VENTURI AP = 9in. H20
SLURRY FLOW RATE = 600 gal/min
GAS FLOW RATE,
ACFM@330°F
20,000
27,500
35,000
SYMBOL
* EFFECTIVE Mg++ =
- CI~/2.92
CI-/2.92
4-
2,000 4,000 6,000 8,000 10,000
EFFECTIVE LIQUOR Mg++ CONCENTRATION, *ppm
12,000
Figure A-7. Effect of Liquor Magnesium-Ion Concentration
and Gas Flow Rate on SC>2 Removal Venturi
Scrubber with Limestone
A-9
-------
100
90
>. 80
O
UJ
O
70 -
§
O
5
UJ
CC
CM 60
UJ
U
cc
40 -
30
9'300Ppm
SO2 INLET CONCENTRATION = 2.140 - 3,630 ppm
SCRUBBER INLET LIQUOR pH 5.7 5.8
VENTURI AP = 9 in. H2O
SLURRY FLOW RATE = 600 gal/min.
* EFFECTIVE Mg++ = Mg++- CI~/2.92
= O FOR Mg"l"+ < CI-/2.92
20,000 27,500
GAS FLOW RATE, acfm @ 330 °F
35,000
Figure A-8. Effect of Gas Flow Rate and Liquor Magnesium-Ion
Concentration on S02 Removal Venturi Scrubber
with Limestone
A-10
-------
60
O
UJ
O
u.
u.
UJ
O
CM
8
UI
O
cc
HI
Q.
Ill
SO2 INLET CONCENTRATION = 2,580 - 2,860 ppm
SCRUBBER LIQUOR INLET pH = 8.0-8.1
VENTURI AP = 9 IN. H2O
50 4. GAS FLOW RATE, SYMBOL
acfm @ 330 °F
35,000
27,500
20,000
30 -I-
10 +
H 1 , H
10 20 30 40
VENTURI LIQUID-TO-GAS RATIO, gal/Mcf
50
Figure A-9. Effect of Liquid-to-Gas Ratio and Gas Flow Rate
on SC>2 Removal Venturi Scrubber with Lime
A-ll
-------
70
60 --
UJ
O
LL
LL
UJ
O
CO
111
g
LU
D.
50
40 -
30 -
20 -
10 --
SO2 INLET CONCENTRATION = 2,580 - 2,860 ppm
SCRUBBER INLET LIQUOR pH = 8.0-8.1
VENTURI AP = 9in. H2O
1
35,000
20,000 27,500
GAS FLOW RATE, acfm @ 330° F
Figure A-10. Effect of Slurry and Gas Flow Rates on S02
Removal Venturi Scrubber with Lime
A-12
-------
60
50
O
z
UJ 40
O
LU
_l
<
30
UJ
CC
CM
20
cc
LU
a.
10
0
6.0
O
SO2 INLET CONCENTRATION = 2,580-2,860 ppm
LIQUID - TO - GAS RATIO = 27 gal/Mcf
VENTURI AP = 9in. H2O
GAS FLOW RATE = 27,500 acfm @ 330° F
7.0 8.0 9.0
SCRUBBER INLET LIQUOR pH
10.0
Figure A-ll. Effect of Scrubber Inlet Liquor pH on SCL
Removal Venturi Scrubber with Lime
A-13
-------
60
50
uj 40
U
uL
u.
UJ
30
UJ
cc
CM
o
CO
cc
01
Q.
20
10
SO2 INLET CONCENTRATION = 2,580-2,770 ppm
LIQUID - TO - GAS RATIO = 27 gal/Mcf
GAS FLOW RATE = 27,500 acfm @ 330° F
SCRUBBER INLET LIQUOR pH = 8.0-8.1
5 6
7
8
9
10
11
12
13
VENTURI AP, in. H2O
Figure A-12. Effect of Venturi AP on S02 Removal Venturi
Scrubber with Lime
A-14
-------
100
90
o
UJ
o
U4
o
5
UJ
CC
CM
O
111
o
cc
UJ
Q.
80
70
60
50 +
40 f
30
20
SCRUBBER INLETpH
FACTORIAL TESTS
O pH = 5.7-5.9
D pH = 5.4-5.6
A pH = 5.1-5.3
SCRUBBER GAS VELOCITY = 7.3 ft/sec
EFFECTIVE LIQUOR Mg++ CONCENTRATION = 0 ppm
INLET SO2 CONCENTRATION = 2,500-3,000 ppm
LIQUOR Cl~ CONCENTRATION = 12,000-16,000 ppm
H 1 1
30
40 50 60
LIQUID - TO - GAS RATIO, gal/Mcf
70
80
Figure A-13. Liquid-to-Gas Ratio and Scrubber Inlet pH Versus
Predicted (Equation B-2) and Measured S02
Removal Spray Tower with Limestone
A-15
-------
100
90
o.
z
UJ
I 30
LL
UJ
o
UJ
CC
CM
8
I-
LLt
O
CC
UJ
Q.
70 -
60
50
40
SLURRY FLOW RATE,gal/min-ft2
O 30 FACTORIAL TESTS
22.5 LONG - TERM TESTS
D 22.5 FACTORIAL TESTS
A 15 FACTORIAL TESTS
SCRUBBER INLET pH = 5.7-5.9
EFFECTIVE LIQUOR Mg++ CONCENTRATION = 0 ppm
2,000-3,000 ppm
CONCENTRATION = 3,500-16,000 ppm
INLET SO2 CONCENTRATION
LIQUOR cr
789
SPRAY TOWER GAS VELOCITY, ft/sec
10
11
Figure A-14.
Gas Velocity and Slurry Flow Rate Versus
Predicted (Equation B-2) and Measured S02
Removal Spray Tower with Limestone
A-16
-------
100
90
O 80
z
UJ
O
70
LU
OC
M
8
UJ
O
QC
UJ
Q.
60
50
40
30
5.0
LIQUID-TO-GAS RATIO-
FACTORIAL TESTS
O 68 gal/Mcf
D 51 gal/Mcf
A 34 gal/Mcf
SCRUBBER GAS VELOCITY = 7.3 ft/sec
EFFECTIVE LIQUOR Mg++ CONCENTRATION = 0 ppm
INLET SO2 CONCENTRATION = 2,500-3,000 ppm
LIQUOR C1~ CONCENTRATION = 12,000-16,000 ppm
I 1 1
5.2
5.4 5.6 5.8
SCRUBBER INLETpH
»
6.0
6.2
Figure A-15.
Scrubber Inlet pH and Liquid-to-Gas Ratio Versus
Predicted (Equation B-2) and Measured S02
Removal Spray Tower with Limestone
A-17
-------
100 -
90
O
UJ
O
UL
u_ 80 +
UJ
I
UJ
cc
70
CM
8
I-
Z
UJ
o
DC 60
UJ
Q.
50 -
40
EFFECTIVE LIQUOR Mg++
-FACTORIAL TESTS
O 8,000-10,000 ppm
D 3,000-5,000 ppm
^ 500-1,500 ppm
A 0 ppm
CONCENTRATION
SCRUBBER GAS VELOCITY = 7.3 ft/sec
SCRUBBER INLET pH = 5.4-5.6
INLET SO2 CONCENTRATION = 2,500-3,000 ppm
LIQUOR Cr CONCENTRATION = 8,000-16.000 ppm
4-
4-
20
30
40 50 60
LIQUID - TO - GAS RATIO, gal/Mcf
70
80
Figure A-16.
Liquid-to-Gas Ratio and Effective Magnesium Versus
Predicted (Equation B-2) and Measured 862 Removal
Spray Tower with Limestone
A-18
-------
100 -
EFFECTIVE LIQUOR Mg++ CONCENTRATION
-FACTORIAL TESTS
O 8,000-10,000 ppm
g 3,000-5,000 ppm
V 500-1,500 ppm
A 0 ppm
90
o
z
UJ
2 80
LL
LL
UJ
O
CC
CM
I-
LLJ
O
CC
70
50
-
40
SCRUBBER GAS VELOCITY = 7.3 ft/sec
LIQUID - TO - RATIO = 51 gal/Mcf
INLET SO2 CONCENTRATION = 2,500-3,000 ppm
LIQUOR Cl~ CONCENTRATION = 12,000-16,000 ppm
1 1
5.0
5.2
5.4 5.6
SCRUBBER INLET pH
I
5.8
6.0
Figure A-17.
Scrubber Inlet pH and Effective Magnesium Versus
Predicted (Equation B-2) and Measured S02
Removal Spray Tower with Limestone
A-19
-------
100
90 --
SCRUBBER INLET pH-
FACTORIAL TESTS
O pH = 5.7-5.9
D pH = 5.4-5.6
A pH = 5.1-5.3
SCRUBBER GAS VELOCITY = 7.3 ft/sec
LIQUID - TO - GAS RATIO = 51 gal/Mcf
INLET SO2 CONCENTRATION = 2,500-3,000 ppm
LIQUOR Cl~ CONCENTRATION = 12,000-16,000
40
I 1 I 1 1
2,000 4,000 6,000 8,000 10,000
EFFECTIVE LIQUOR MAGNESIUM CONCENTRATION, ppm
Figure A-18. Effective Magnesium and Scrubber Inlet pH Versus
Predicted (Equation B-2) and Measured SCL
Removal Spray Tower with Limestone
12,000
A-20
-------
100
90
O
z
UJ
O
LI 80
LU
UJ
$
O
5 70
UJ
QC
-------
100
90
o
z
UJ
g »
u_
UJ
I
111
IT
to
H-
LU
O
DC
UJ
a.
70 -
60 -
50 -
40
SLURRY FLOW RATE FOR
FACTORIAL TESTS
O 30 gal/min - ft2
A 22.5
D 15
SCRUBBER INLET pH = 7.9 - 8.1
SO2 INLET CONCENTRATION = 2,500 - 3,500 ppm
EFFECTIVE LIQUOR Mg"1"1" CONCENTRATION = 0 ppm
LIQUOR Cl~ CONCENTRATION = 8,000 - 13,000 ppm
_, 1 1
789
SPRAY TOWER GAS VELOCITY, ft/sec
-t-
10
11
Figure A-20.
Gas Velocity and Slurry Flow Rate Versus
Predicted (Equation B-l) and Measured
S02 Removal Spray Tower with Lime
A-22
-------
100
90
LIQUID-TO-GAS RATIO
FOR FACTORIAL TESTS
O 68 gal/mcf
A 51
D 34
ii! 80
o
uZ
u.
UJ
|70
UJ
oc
P«J
O
t/J
ui
O
DC
UJ
Q.
60
50
40
. SPRAY TOWER GAS VE LOCITY = 7.5 gal/Mcf
SO2 INLET CONCENTRATION = 2,500 - 3,000 ppm
EFFECTIVE LIQUOR Mg++CONCENTRATION
= 0 ppm
LIQUOR Cl~ CONCENTRATION = 8,000 -
13,000 ppm
I 1 ,
7 8
SCRUBBER INLET pH
10
Figure A-21
Scrubber Inlet pH and Liquid-to-Gas Ratio
Versus Predicted (Equation B-l) and Measured
SOo Removal Spray Tower with Lime
A-23
-------
100
90 +
SCRUBBER INLET pH
A 7.9 - 8.1 FACTORIAL TESTS
A 7.9 - 8.1 LONG - TERM TESTS
6.9 - 7.1 LONG - TERM TESTS
6.0 LONG - TERM TEST
50 +
40
SPRAY TOWER GAS VELOCITY = 7.3 - 9.3 ft/sec
LIQUID - TO - GAS RATIO = 50 gal/Mcf
SO2 INLET CONCENTRATION = 1,800 - 3,000 ppm
LIQUOR Cl~ CONCENTRATION = 8,000 - 13,000 ppm
500 1.000 1,500 2,000
EFFECTIVE LIQUOR Mg++ CONCENTRATION, ppm
2,500
Figure A-22.
Effective Magnesium and Scrubber Inlet pH
Versus Predicted (Equation B-l) and Measured
S02 Removal Spray Tower with Lime
A-24
-------
100
90
O
i! so
i!
o
u.
Ill
>
o
UJ
QC
CM
70
60
111
U
cc
HI
Q.
50
40
TCA GAS VELOCITY FOR
FACTORIAL TESTS
12.5 ft/sec
10.4
SCRUBBER INLET pH = 7.9 - 8.1
SO2 INLET CONCENTRATION = 2,200 - 2,800 ppm
TOTAL HEIGHT OF SPHERES = 15.0 in.
EFFECTIVE LIQUOR Mg1"1" CONCENTRATION = 0 ppm
4-
4-
20
30
40 50 60
LIQUID - TO - GAS RATIO, gal/Mcf
70
80
Figure A-23.
Liquid-to-Gas Ratio and Scrubber Gas Velocity
Versus Predicted (Equation B-3) and Measured
SO Removal TCA with Lime
A-25
-------
100
90
80
o
z
UJ
O
LL
70 --
O
UJ
CC 60
CM
O 50
UJ
o.
40
30
SLURRY FLOW RATE FOR
FACTORIAL TESTS
O 37.5 gal/min - ft2
A 28.1
D 18.8
° SLURRY FLOW RATE n = 37.5 gal/min - ft2
O
O
8-
28.1 gal/min - ft2
SCRUBBER INLET pH = 7.9 - 8.1
TOTAL HEIGHT OF SPHERES = 15.0 in.
SO2 INLET CONCENTRATION = 2,200 - 2,800 ppm
EFFECTIVE LIQUOR Mg*"1" CONCENTRATION = 0 ppm
4-
10 11
TCA GAS VELOCITY, ft/sec
12
13
Figure A-24.
TCA Gas Velocity and Slurry Flow Rate Versus
Predicted (Equation B-3) and Measured S02
Removal TCA with Lime
A-2 6
-------
100
90
u
jg 80
u
LL
HI
O
UJ
cc
CM
O
V)
UJ
U
CC
UJ
O.
70
60 ..
50 -
40
T
LIQUID - TO - GAS RATIO FOR
FACTORIAL TESTS
O
A
n
60 gal/Mcf
45 gal/Mcf
30 gal/Mcf
SCRUBBER GAS VELOCITY = 10.4 ft/sec
INLET SO2 CONCENTRATION = 2,200 -
2,800 ppm
TOTAL HEIGHT OF SPHERES = 15 .0 in.
EFFECTIVE LIQUOR Mg++
CONCENTRATION = 0 ppm
7 8
SCRUBBER INLETpH
10
Figure A-25. Scrubber Inlet pH and Liquid-to-Gas Ratio Versus
Predicted (Equation B-3) and Measured
Removal TCA with Lime
SO,
A-27
-------
100
90 -
40
SLURRY FLOW RATE FOR
FACTORIAL TESTS
O 37.5 gal/min - ft2
A 28.1
D 18.8
D
TCA GAS VELOCITY = 10.4 ft/sec
SCRUBBER INLET pH = 7.9 - 8.1
INLET SO2 CONCENTRATION = 2,200
- 2,800 ppm
EFFECTIVE LIQUOR Mg++
CONCENTRATION = 0 ppm
10 15
TOTAL HEIGHT OF SPHERES, in.
20
25
Figure A-26.
Total Height of Spheres and Slurry Flow Rate
Versus Predicted (Equation B-3) and Measured
S02 Removal TCA with Lime
A-28
-------
100
90
o
2
LLJ
O 80
\L
LL
LU
o
LJJ
GC
CM
O
CO
I-
z
111
o
cc
UJ
Q.
70 -
60
50 -
40
SCRUBBER INLET pH FOR
LONG - TERM TESTS
7.8-8.1
A 6.9-7.1
SCRUBBER GAS VELOCITY = 12.5 ft/sec
LIQUID - TO - GAS RATIO = 37 gal/Mcf
INLET SO2 CONCENTRATION = 3,000 - 3,500 ppm
TOTAL HEIGHT OF SPHERES =12-15 in.
4-
1,000 2,000
EFFECTIVE LIQUOR MgH
3,000 4,000
CONCENTRATION, ppm
5,000
Figure A-27.
Effective Magnesium and Scrubber Inlet pH Versus
Predicted (Equation B-3) and Measured S02
Removal TCA with Lime
A-29
-------
Appendix B
MATHEMATICAL MODELS AND NOMOGRAPHS FOR S02 REMOVAL BY
LIMESTONE AND LIME WET SCRUBBING
A semitheoretical model has been developed and fitted to the Shawnee
test data for predicting SC>2 removal by lime or limestone wet scrubbing
from operating variables in a spray tower or a TCA. This model accounts
for all major variables investigated at the Shawnee Test Facility except
for possible effects of liquor sulfate (gypsum) saturation or sulfite
oxidation on SC>2 removal at high magnesium-ion concentrations. Such
effects are relatively insignificant at low magnesium-ion concentrations
The fitted equations are as follows:
« Spray tower equation for S02 removal by lime slurry
Fraction S02 Removal = 1 - exp |- 0.0020 (L/G) exp J0.29 pH.
+ 2.8 x 10~4 (Mg)e + 4.7 x 10 5 Cll \ (B-l)
Spray tower equation for S0? removal by limestone slurry
/ _5 0.92 0.19
Fraction SO Removal = 1 - exp <- 9.8 x 10 (L/G) v
exp p + 1.35 x 10 (Mg) - 1.7 x 10 ** (S02)
+ 1.45 x 10~5C1)J > (B-2)
B-l
-------
TCA equation for SO removal by lime slurry
1. 12 0.65
/ 1.12
Fraction SO Removal = 1 - exp <- 0.0010 (L/G) v
exp JO. 18 pH± + 1.5 x 10~tf(Mg)(
- 2.2 x 10 (S02)i + 0.0039v [ -~^ + Nr]|}(B-3)
TCA equation for S0_ removal by limestone slurry
. -i* 0.81 0.36
Fraction SO Removal = 1 - exp { - 2.05 x 10 (L/G) v
|"4.3 x 10 3v(-jp^ + NG j + 0.81 pH±
+ 7.9 x 10 5 (Mg)e - 1.7 x 10~\S02)i
+ 1.3 x 10 5 Cl IV (B-4)
B-2
-------
where:
Cl = measured total dissolved chloride concentration, ppm
dg = diameter of TCA sphere, 1.5 inches at Shawnee
htot = total height of spheres in TCA, inches (zero for spray tower)
L/G = liquid-to-gas ratio in scrubher (125°F, humidified gas),
gal/Mcf
(Mg)g = effective magnesium-ion concentration, ppm
= [ppm Mg - (ppm Cl~/2.92)] for Mg"^1" > Cl~/2.92
= 0 for Kg4"1" < Cl~/2.92
where 2.92 = ratio by weight of Cl~ to Hg^ in MgCl-
N = number of grids or screens in TCA, 4 at Shawnee (zero
for spray tower)
pH. = scrubber inlet liquor pH
(SO.) . = inlet gas SCL concentration, ppm
v = gas velocity in scrubber (125 F, humidified gas), ft/sec
These equations are verified only for the range of conditions encountered
at the Shawnee Test Facility. Extrapolation significantly beyond this
range is not recommended. The ranges of the important operating
variables are given below:
Equation B-l
L/G = 30 to 100 gal/Mcf
pHi = 6.0 to 9.0
(Mg)e = 0. to 2,500 ppm
Cl = 0 to 12,000 ppm
v = 5.3 to 9.3 ft/sec
(S02)i = 2,000 to 4,000 ppm
B-3
-------
Equation B-2
L/G = 30 to 70 gal/Mcf
v = 5.3 to 9.3 ft/sec
pHi = 5.2 to 5.9
(Mg)e = 0 to 10,000 ppm
(S02)i = 1,500 to 4,500 ppm
Cl = 4,000 to 16,000 ppm
Equation B-3
L/G = 25 to 75 gal/Mcf
v = 8.5 to 12.5 ft/sec
htot = ° to 22'5 incnes
ds =1.5 inches
NG =4
pHi = 6.0 to 9.0
(Mg)e = 0 to 4,000 ppm
(S02)i = 2,000 to 4,000 ppm
Cl = 0 to 8,000 ppm
Equation B-4
L/G = 25 to 75 gal/Mcf
v = 8.5 to 12.5 ft/sec
pHi = 5.1 to 5.9
(Mg)e = 0 to 10,000 ppm
(S02)i = 1,500 to 4,500 ppm
Cl = 1,500 to 20,000 ppm
ds =1.5 inches
htot = 0 to 22.5 inches
NG =4
Figures B-l through B-5 are nomographs for predicting, from operating
variables, S0_ removal by limestone or lime wet scrubbing with a spray
tower or a TCA. Figure B-l can be used to predict the effect of a
change in SCL inlet concentration on SO- removal. Nomographs for lime
and limestone scrubbing with a spray tower, Figures B-2 and B-3,
respectively, are derived from Equations B-l and B-2. Figures B-4 and
B-5 for lime and limestone scrubbing with a TCA, respectively, are
derived from Equations B-3 and B-4. The nomographs cover the same
ranges of operation as the corresponding equations
B-4
-------
In practice, predicting S02 removal by nomographs is typically within
about one-half percent of the corresponding equation prediction. Speed
of calculation using one of these nomographs is about the same as for
solving the corresponding equation by electronic calculator.
The inset in Figure B-2 illustrates the use of these nomographs.
First, the liquid-to-gas ratio is located on the bottom curve.
Then, a vertical line is drawn to intersect with the next lowest hori-
zontal line, which represents the lowest level of the next variable -
scrubber inlet liquor pH in this case. A line is then drawn upward
from the horizontal line and parallel to the nearest pH curve, until
the y-axis level corresponding to the specified pH, 8.0 in this case,
is reached. Again, a vertical line is drawn to the horizontal line
for the next operating variable chloride-ion concentration. This
procedure is continued until the top of the nomograph is reached. The
predicted SO- removal is then the x-axis value at this point. In the
example case, the nomograph SCL removal prediction by Figure B-2 was
85 percent; Equation B-l correspondingly predicted 86 percent.
If SO removal is specified in a design, and prediction of another
operating variable is required, for example scrubber inlet liquor pH,
this prediction can be made from the nomograph by working upwards from
the bottom of the nomograph and downwards from the top. The two
resultant intersecting lines, one parallel to the nomograph pH curves
from the bottom up and the other vertical from the top of the pH scale,
meet at the y-axis value of the operating pH.
B-5
-------
100
86 -
O 80 -
CO
H
S
U
ff!
H
70 +
65
60 -
55 -
1,000 1.500 2,000 2,500 3,000 3,500 4,000
SOZ INLET .CONG. , ppm
Figure B-l. Predicted Effect of Inlet S02 Concentration on S(>2 Removal
Figure B-l can be used to predict the effect of a change in SC>2 inlet
concentration on 862 removal. When the S02 removal, (SC^R)!, at a
particular inlet S02 concentration, (SO^i)!, is known, the removal,
(S02R)2, at a different inlet S02 concentration, (S02i)2, is obtained
by first locating the point [(S02i)l, (S02R)l] and then paralleling
the nearest curve until (S02i)2 is reached. The corresponding value
on the y axis is then (S02R)2.
B-6
-------
PREDICTED PERCENT SO2 REMOVAL
30
30
40
50 60 70 80
PREDICTED PERCENT SO2 REMOVAL
Fraction SO2 Removal = 1 - exp |-0. 0020 (L/C) exp [ 0.29 pB{
+ 2.8 x 10'4 (Mg)e H- 4. 7 x lO'5 Cl]l
90
Figure B-2.
Nomograph for Percent S02 Removal
by Lime Scrubbing in a Spray Tower
B-7
-------
PREDICTED PERCENT SO2 REMOVAL
40 50 60 70 80
90
100
PREDICTED PERCENT SO, REMOVAL
Fraction SO2 Removal ] exp j -9.8 x lO-^I./C)
' 92 °'19
exp
!-4
1. 7 v JO"* y. I 1.41? x 10
HI
TpHj
+ 1.35 x 10
"4
Figure B-3. Nomograph for Percent S02 Removal by
Limestone Scrubbing in a Spray Tower
B-8
-------
PREDICTED PERCENT S02 REMOVAL
100
20
30
40
50 60 70 80
PREDICTED PERCENT SO2 REMOVAL
90
Fraction SO2 Removal - 1 exp 1-0. 0010 (L/G) v'' exp [0. 18 pH;
' 1. S *. 10-4 (Mg)e 2.2 x ID"4 y. ^ 0. 0039 v(^£ + NG)||
Figure B-4. Nomograph for Percent SC>2 Removal by
Lime Scrubbing in a TCA
B-9
-------
20
PREDICTED PERCENT S02 REMOVAL
40 50 60 70 80
90
100
30
40 50 60 70 80
PREDICTED PERCENT SO2 REMOVAL
90
100
Fraction S02 1 exp j - Z. 05 x 10'4 (L/G)°' 81 v°' 3f> exp
Removal «
+ 0.81 pH; +7. 9 x ID"5 (Mg)e 1. 7 * Kr4y. + 1.3 x 10'5 Cll [
Figure B-5. Nomograph for Percent S02 Removal by
Limestone Scrubbing in a TCA
B-10
-------
Appendix C
CONVERSION TABLE
To Convert From
scfm (60°P)
cfm
°F
ft
ft/hr
ft/sec
ft2
>~\
ft /tons per day
gal/Mcf
gpm
gpm/ft2
gr/scf
in.
in. H20
Ib
Ib-moles
Ib-moles/hr
Ib-moles/hr ft2
Ib-moles/min
To
nm3/hr (0°C)
m3/hr
°C
m
m/hr
m/sec
m2
n
m /metric tons per day
1m3
1/min
r\
1/min/nr
gm/m3
cm
mm Hg
gm
gm-moles
gm-moles/min
fj
gm-moles/min/m
gm-moles/ sec
Multiply By
1.61
1.70
(°F-32)/1.8
0.305
0.305
0.305
0.0929
0.102
0.134
3.79
40.8
2.29
2.54
1.87
454
454
7.56
81.4
7.56
C-l
-------
TECHNICAL REPORT DATA
(Please read laurjcnons on ilie reverse be tare completing)
1. REPORT NO.
EPA-600/7-78-030b
4. TITLE AND SUBTITLE JT lU6 GaS I>
Design and Operating Consi
Volume n. Technical Repo
2.
asulfurization Systems:
derations
rt.
7. AUTHORtSl
C. C. Leivo
9. PERFORMING ORGANIZATION NAME Al>.
Bechtel Corporation
50 Beale Street
San Francisco, California
ID ADDRESS
94119
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
3. RECIPIENT'S ACCESSION- NO.
5. REPORT DATE
March 1978
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO.
10. PROGRAM ELEMENT NO.
EHE624
11. CONTRACT/GRANT NO.
68-02-2616, Task 2
13. TYPE OF REPORT AND.PERIOD COVERED
Task Final; 4-12/77
14. SPONSORING AGENCY CODE
EPA/600/13
is. SUPPLEMENTARY NOTES EPA project officers are J.E.Williams (IERL-RTP, 919/541-2483)
andK.R.Durkee (OAQPS/ESED, 919/541-5301).
16. ABSTRACT rj-,^ repOrt describes flue gas desulfurization (FGD) systems and the design
and operating parameters that are monitored to ensure proper operation. It explains
how these parameters are varied to accommodate changing boiler loads and fuel char-
acteristics , and describes the control of parameters to prevent such problems as
scale buildup. It describes effects of designing and operating FGD systems for 90% or
greater SO2 removal efficiencies, based on current testing program data. It describes
effects of coal characteristics on FGD performance, along with operating and design
techniques used to compensate for coal property variations. It describes the purpose,
need, and methods for exhaust gas reheat, downstream of FGD systems. It discusses
alternatives to exhaust gas reheat.
17.
a. DESCRIPTORS
KEY WORDS AND DOCUMENT ANALYSIS
b.lOENTIFIERS/OPEN ENDED TERMS
Air Pollution Scrubbers Air Pollution Control
Flue Gases Calcium Oxides Stationary Sources
Desulfurization Limestone Alkali Scrubbing
Coal Sulfur Dioxide Particulate
Reheating Dust Venturi/Spray Towers
Alkalies Scale (Corrosion) Mist Eliminators
13. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (Tins Report)
Unclassified
2O. SECURITY CLASS (This page)
Unclassified
c. COSATI Field/Group
13B
21B 07B
07A,07D 08G
21D
13A 11G
11F
21. NO. OF PAGES
216
22. PRICE
EPA Form 2220-1 (9-73)
C-2
------- |