CD A U.S. Environmental Protection Agency Industrial Environmental Research EPA-600/7-78~032b
^•' •» Office of Research and Development Laboratory -i/\"»o
Research Triangle Park, North Carolina 27711 MafCn 1978
FLUE GAS DESULFURIZATION
SYSTEM CAPABILITIES FOR
COAL-FIRED STEAM GENERATORS
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 endorsemertt or recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
-------
ABSTRACT
This study presents an assessment of flue gas desul-
furization (FGD) systems with respect to both SO- removal
potential and system operability. This assessment was made
by reviewing the performance of operating systems and the
features of new systems designed to alleviate previous
problems. Five major FGD processes are described in detail.
The impact of key design parameters on removal efficiency
and process operability was investigated from a theoretical
point of view and by using data from operating systems. In
addition, the operating experience of major FGD installa-
tions was reviewed. Operating problems and their solutions
were analyzed with respect to their impact on process opera-
tion. Research data were used to evaluate the status of
process development and the potential for sustained system
operation and high removal efficiencies.
This study shows that the major systems, comprising
lime and limestone slurry, Wellman-Lord, magnesium oxide
scrubbing, and double alkali processes, are capable of re-
moving SO2 with high efficiencies when applied to both high-
and low-sulfur coal combustion facilities. Such results
have been obtained in pilot, prototype, and full-scale sys-
tems. Though sustained operation at high SO removal effi-
ciencies has not been widely achieved, a basis for design of
such systems has been developed. When required, new systems
are being designed for 90-percent removal efficiency and an
availability of 90-percent or greater.
An analysis of the effect of using spare scrubber
modules is also presented and shows that large improvements
in system availability can be achieved by installing a s
module.
11
-------
CONTENTS
Page
Abstract ii
Figures v
Tables xi
Acknowledgment xv
Metric Conversion Factors xvi
1.0 INTRODUCTION 1-1
1.1 History of FGD Systems 1-1
1.2 Application of FGD Systems 1-3
1.3 Report Contents 1-4
2.0 OVERVIEW OF ALTERNATIVE S0? EMISSION CONTROL 2-1
SYSTEMS
2.1 Low-Sulfur Coal 2-1
2.2 Physical Coal Cleaning 2-11
2.3 Emerging Technologies 2-17
3.0 FLUE GAS DESULFURIZATION SYSTEMS 3-1
3.1 Lime Slurry Flue Gas Desulfurization Systems 3-2
3.2 Limestone Slurry Flue Gas Desulfurization 3-73
Systems
3.3 Double Alkali Flue Gas Desulfurization 3-111
Systems
3.4 Magnesium Oxide Systems 3-194
3.5 The Wellman-Lord Process 3-261
3.6 Other FGD Systems 3-318
4.0 PERFORMANCE AND OPERABILITY OF FGD SYSTEMS 4-1
4.1 Operating Problems and Solutions 4-2
ill
-------
CONTENTS (Continued)
Page
4.2 System Operability and Availability 4-11
4.3 System Efficiency 4-27
APPENDIX A DOMESTIC LIME SLURRY FGD SCRUBBING A-l
SYSTEMS
APPENDIX B DOMESTIC LIMESTONE SLURRY FGD SCRUBBING B-l
SYSTEMS
IV
-------
FIGURES
Figure Page
2-1 Plot of the Maximum Permissible Sulfur Content 2-3
Versus Btu Content of Coal Required to Meet
Current NSPS
2-2 Recoverable Coal Reserve by Sulfur Content 2-6
2-3 Effect of Cleaning Variable on Coal SO- Emission 2-14
Potential
3-1 Flow Diagram/Material Balance - Lime Slurry 3-4
Scrubbing Systems for a 500-MW Boiler
3-2 Vapor Liquid Equilibrium Diagram SO- - 10 Percent 3-6
CaO Slurry
3-3 Process Flow Diagram - Green River No. 1, 2, and 3 3-13
3-4 Scrubber System Operability - Green River No. 1, 2 3-18
and 3
3-5 Simplified Process Diagram - Bruce Mansfield No. 1 3-22
3-6 Horizontal Test Module Mohave Plant 3-28
3-7 Effect of Circulating Liquor Flow Rate on SO- 3-30
Removal at Constant Gas Flow 212 m3/s (450,000 scfm)
Mohave Plant
3-8 Water Balances - Mohave Plant 3-32
3-9 Mist Eliminators Mohave Plant 3-36
3-10 Reheater Comparison for Equivalent Reduction in 3-37
Fog Formation - Mohave Plant
3-11 Liquid-to-gas Ratio and Scrubber Inlet pH Versus 3-43
Predicted and Measured SO,., Removal, Spray Tower
with Lime, Shawnee Plant
3-12 Liquid-to-gas Ratio and Scrubber Gas Velocity 3-45
Versus Predicted and Measured SO Removal, TCA
with Lime, Shawnee Plant
3-13 S02 Removal Versus L/G Ratio, 170-MW Horizontal 3-46
Module, Mohave Plant
3-14 Effect of pH on SCU Vapor Pressure over Buffered 3-48
Solutions
v
-------
FIGURES (Continued)
Figure
3-15 Scrubber Inlet pH and Liquid-to-Gas Ratio Versus 3-49
Predicted and Measured S02 Removal, TCA with Lime
Shawnee Plant
3-16 Distribution of Aqueous Sulfite Species as a 3-51
Function of pH
3-17 Scrubber Inlet pH and Liquid to Gas Ratio Versus 3-52
Predicted and Measured SO2 Removal, Spray Tower
with Lime, Shawnee Plant
3-18 Effect of Magnesium on SO2 Removal Efficiency 3-53
3-19 Dissolved Alkalinity Generated by Addition of MgO 3-55
3-20 170-MW Horizontal SC>2 Removal Versus Number of 3-56
Stages, Mohave Plant
3-21 Gas Velocity and Slurry Flow Rate Versus Predicted 3-58
and Measured SO2 Removal, Spray Tower with Lime,
Shawnee Plant
3-22 Effect of Inlet S02 Concentration on SC>2 Absorption 3-59
Efficiency for Fixed Design Condition
3-23 Scrubber System Operability Green River No. 1, 2, 3-61
and 3
3-24 Scrubber Operability, Cane Run No. 4 3-62
3-25 Scrubber Operability - Paddy's Run No. 6 3-64
3-26 Flow Diagram/Material Balance - 500-MW Limestone 3-76
Scrubbing System
3-27 Flow Diagram of One of the Eight FGD Modules - La 3-81
Cygne No. 1
3-28 La Cygne Availability History 3-28
3-29 Availability History Sherburne No. 1 and No. 2 3-94
3-30 L/G Ratio and Scrubber Inlet pH Versus Predicted 3-100
and Measured S02 Removal - TCA with Limestone
Shawnee Plant
VI
-------
FIGURES (Continued)
Figure Page
3-31 L/G Ratio Versus Actual Percent S02 Removal 170-MW 3-101
Test Module with Limestone - Mohave Plant
3-32 Scrubber Inlet AV Versus pH for Three L/G Ratios - 3-102
TCA Unit with Limestone - Shawnee Plant
3-33 L/G Ratio Versus Percent SO2 Removal at Various 3-104
Magnesium Ion Concentrations TCA with Limestone -
Shawnee Plant
3-34 Scrubber Inlet pH Versus Percent S02 Removal at 3-105
Various Magnesium Ion Concentrations TCA with
Limestone - Shawnee Plant
3-35 Height of Spheres Versus S02 Removal Efficiency, 3-106
TCA with Limestone - Shawnee Facility
3-36 FMC Double Alkali Pilot Plant: Schematic and 3-112
Simplified Process Flow Diagram
3-37 Schematic of a Double Alkali System with Lime or 3-119
Limestone Regeneration, Concentration Alkali, and
H2SO. Sulfate Removal
3-38 Schematic of a Double Alkali System with Lime 3-121
Regeneration, Dilute Alkali, and Carbonate
Softening
3-39 Material Balance for 500 MW Double Alkali ,FGD 3-122
System with Lime Regeneration.
3-40 Parma Double Alkali Scrubbing System: Schematic 3-130
and Simplified Process Flow Diagram
3-41 Joliet Double Alkali Scrubbing System: Schematic 3-142
and Simplified Process Flow Diagram
1-42 Pottstown Double Alkali Demonstration Plant: 3-148
Schematic and Simplified Process Flow Diagram
3-43 Envirotech Gadsby Double Alkali Pilot Plant: 3-153
Schematic and Simplified Process Flow Diagram
3-44 Dougle Alkali System at Newton Station 3-154
VII
-------
Figure
FIGURES (Continued)
Paqe
3-45 S09 Removal Versus L/G Ratio for the Enviro- 3-158
tech/Gadsby Pilot Plant with a Single Stage
Polysphere Absorber
3-46 SO9 Removal Versus Scrubber Effluent pH for 3-159
the Envirotech/Gadsby Pilot Plant at an L/G
of 2.5 1/m (19 Gal/lOOOacf) and using a two-
stage Absorber
3-47 Scholz Double Alkali Prototype Scrubber: 3-161
Schematic and Process Flow Diagram
3-48 Outlet SO,, Concentration Versus Scrubber 3-165
Bleed pH.
3-49 Kureha-Kawasaki Double Alkali Process at Tohoku 3-172
Electric's Shinsendai Station
3-50 Showa Denko Sodium-Limestone Process 3-178
3-51 Concentrated Active Alkali System: Sulfur 3-182
Dioxide Vapor Pressure vs. pH @ 54°C (130°F)
3-52 Magnesium Oxide Slurry FGD System 3-195
3-53 Mystic No. 6 FGD System and Essex Chemical 3-204
Plant Regeneration Facility
3-54 Effect of Inlet S02 Concentration and Venturi 3-207
Pressure Drop on SO- Removal for the Mystic
Venturi Absorber
3-55 Effect of Pressure Drop on S0~ Removal for 3-208
the Mystic Venturi Absorber
3-56 Dickerson No. 3 Regenerative FGD System: 3-212
General Process Diagram
3-57 Eddystone Unit 1A S02 Removal System 3-222
3-58 S02 Scrubber System at Eddystone Plant 3-223
3-59 SO2 Scrubbing System - Mg SO- Recovery 3-225
van
-------
FIGURES (Continued)
Figure Page
3-60 MgO Regeneration Plant 3-226
3-61 Flowchart - Onahama Reverberatory Furnace 3-231
SO2 Control System
3-62 The Affect of pH on S02 Scrubbing Efficiency 3-237
3-63 Comparison of PEPCO and Boston Edison Dryer 3-242
Operation. (For ranges of dryer gas outlet
temperature)
3-64 Typical Wellman-Lord S02 Recovery Process 3-262
3-65 Flow Diagram of the Wellman-Lord SO2 Recovery 3-268
Process and a Sulfur Removal Unit
3-66 Schematic of the Wellman-Lord S02 Recovery 3-279
Process at NIPSCO
3-67 Inlet and Outlet SO Concentrations During Run 3-285
No. 1
3-68 Inlet and Outlet SO Concentrations During Run 3-285
No. 2
3-69 Inlet and Outlet SO,, Concentrations During Run 3-286
No. 3
3-70 Flowsheet of Wellman-Lord Process at Japan 3-293
Rubber, Chiba, Japan
3-71 Flowsheet of the Wellman-Lord (MKK) Process at 3-296
Chubu Electric, Nagoya, Japan
3-72 Layout of Wellman-Lord FGD Plant at Chubu 3-297
Electric, Nagoya, Japan
3-73 Flowsheet - Chiyoda Thoroughbred 101 Process 3-321
3-74 Flowsheet - Citrate Process 3-323
3-75 Flowsheet - Westvaco Activated Carbon Process 3-328
ix
-------
FIGURES (Continued)
Figure Page
3-76 Flowsheet - BF/FW Dry Absorption Process 3-332
3-77 Flowsheet - Consol Scrubber 3-336
3-78 Flowsheet - Consol Commercial Regenerator 3-338
3-79 Flowsheet - Aqueous Carbonate Process 3-342
3-80 Flowsheet - Schell/UOP Copper Oxide Absorption 3-348
Process
3-81 Flowsheet of Lime/Alkaline Flyash Scrubbing 3-352
4-1 Average Plant FGD Availability/Operability 4-13
Versus Plant Start-Up Date
4-2 Average Availability for Selected FGD Systems 4-17
x
-------
TABLES
Table Page
2-1 Summary of Total Recoverable Coal Reserves by 2-5
Region and Emission Potential - Continental U.S.
3-1 Assumptions Used in the Material Balance Calcu- 3-5
lations for a 500-MW Boiler
3-2 Lime Based FGD Systems in the United States 3-10
3-3 Power Plant and FGD System Design Data 3-12
3-4 Green River Power Station Operational Data FGD 3-17
Unit
3-5 Power Plant and FGD System Design Data 3-21
Bruce Mansfield No. 1 - Pennsylvania Power Co.
3-6 Power Plant and FGD System Design Data 3-27
Mohave Test Plant - Southern California Edison
3-7 Unavailability History - Mohave 170 MW Test 3-34
Modules Program
3-8 Operational Data - Mohave Horizontal FGD Unit 3-35
3-9 Performance of Large SO.., Scrubbers Using Lime 3-38
Slurry on Boilers in Japan
3-10 Assumptions Used in the Material Balance Calcu- 3-75
lations for a 500-MW Boiler
3-11 Major Domestic FGD Installations - Limestone 3-79
Slurry
3-12 Power Plant and FGD System Design/Operating 3-80
Data, La Cygne No. 1
3-13 La Cygne Station Stack Sampling Test - Seven 3-88
Modules Operating
3-14 La Cygne Station Sampling Test - Seven Modules 3-90
Operating - Maximum Boiler Load
3-15 Sherburne County Generating Plant Unit 1 - 3-95
Performance Data
3-16 Performance of SO2 Scrubbers Using Limestone 3-97
Slurry on Boilers in Japan
XI
-------
TABLES (Continued)
Table .pa
-------
TABLES (Continued)
Table Page
3-30 Sodium-Based Double Alkali Process Installations 3-171
in Japan
3-31 MgO FGD System Material Balance 3-198
3-32 Particulate Removal - Test Results 3-203
Boston Edison - Mystic Station No. 6
3-33 SO Removal - Test Results 3-203
Boston Edison - Mystic Station No. 6
3-34 Operability of MgO System at Mystic No. 6 3-206
3-35 SO Emissions Test Results MgO FGD System - 3-215
Dickerson
3-36 Particulate Emission Tests Results for MgO System 3-216
at Dickerson
3-37 Removal Efficiency for Selected Particle Size 3-217
Ranges MgO System - Dickerson
3-38 Operability Data for Dickerson No. 3 3-218
3-39 MgO FGD Systems in Japan 3-229
3-40 MgO Plant Operational Specifications - Onahama 3-233
Smelter
3-41 Major Components of MgO Plant - Onahama Smelter 3-234
3-42 Summary of MgO Scrubbing System Operating Parameters 3-239
3-43 Material Balance for a Wellman-Lord FGD System 3-269
Serving a 500-MW Boiler
3-44 Operational Wellman-Lord FGD Systems in the U.S. 3-275
3-45 Wellman-Lord FGD Systems Under Construction in the 3-277
United States
3-46 Design Parameters for Wellman-Lord FGD Instal- 3-278
lations at San Juan Station of Public Service
Company of New Mexico
3-47 NIPSCO Operability Data 3-284
Xlll
-------
TABLES (Continued)
Table
3-48 Problems Encountered at the NIPSCO Wellman-Lord 3-290
Installation
3-49 Operational Wellman-Lord FGD Systems in Japan 3-292
4-1 Identification of Plants in Figure 4-1 4-14
4-2 FGD System Availability Projections Based Upon 4-19
No Spare Modules
4-3 FGD System Availability Projections Based Upon 4-21
Using One and Two Spare Modules (X + 1 and X + 2)
at Full Boiler Load
4-4 Probabilities and Outages for Specified Number of 4-25
Scrubber Modules Operating, Assuming Module Avail-
ability of 90 Percent - 1000-MW Boiler
4-5 Probabilities and Outages for Specified Number of 4-25
Scrubber Modules Operating, Assuming Module Avail-
ability of 80 Percent - 1000-MW Boiler
4-6 Summary of Availability Guarantees Offered by 4-26
Manufacturers
4-7 Plants Reproting 90 Percent or Greater SO2 Removal 4-28
4-8 Guarantees Offered by Manufacturers for SO,, Removal 4-29
xiv
-------
ACKNOWLEDGMENT
This report was prepared under the direction of Messrs.
Timothy W. Devitt and Richard W. Gerstle. Principal authors
were Messrs. Russell P. Klier, Larry L. Gibbs, J. Scott
Hartman, and Bernie Laseke.
Co-Project Officers for the Environmental Protection
Agency were Messrs. Kenneth R. Durkee and John Williams.
The authors appreciate the contributions made by Messrs.
Durkee and Williams and their co-workers.
XV
-------
METRIC CONVERSION FACTORS
To convert
English units
Multiply
by
To obtain
SI units
British thermal unit (Btu)
Cubic foot (ft3)
Degrees fahrenheit
Foot
Gallon (U.S. Liquid)
Gallon (U.S. Liquid)
Horsepower (hp)
Inch
Inch
Inches of water
Pound
Ton, short
1054
0.0283
5/9 (°F-32)
0.3048
0.0038
3.7854
746.0
0.0254
2.54
248.8
0.4536
0.9072
Joule (j)
Cubic meter (m )
Degrees Celsius (C)
Meter (m)
Cubic meter (m )
Liter (1)
Watt (w)
Meter (m)
Centimeter (cm)
Pascal (pa)
Kilogram (kg)
Metric ton (kkg)
xvi
-------
1.0 INTRODUCTION
The United States Environmental Protection Agency (EPA)
has undertaken a program to review the current New Source
Performance Standards (NSPS) limiting sulfur dioxide (S0~)
emissions from large coal-fired steam generators. The
current emission regulation was promulgated in 1971, and EPA
was recently mandated to review this regulation in light of
new technology. EPA's Office of Air Quality Planning and
Standards has prime responsibility for this review, which
will form a basis for determining whether the current NSPS
should be revised. To develop background information in
support of this review, EPA has contracted with PEDCo En-
vironmental, Inc., to evaluate the capabilities of flue gas
desulfurization (FGD) techniques.
1.1 HISTORY OF FGD SYSTEMS
The concept of scrubbing flue gases from coal-fired
utility boilers and other processes is not new. In 1926,
the 125-MW coal-fired Battersea Power Station in London,
England, was equipped with a spray-packed tower and a tail-
end alkaline wash section. The process was more than 90-
percent efficient in the removal of SO- and particulate from
the combustion gas of coal with a sulfur content of 0.9
percent. The alkaline waters of the River Thames provided
most of the absorptive alkali; about 20 percent was made up
from lime addition. Manganese sulfate was added to the
scrubber effluent to enhance the oxidation of calcium sul-
1-1
-------
fite to sulfate. The solution from the scrubber passed to a
settler and into an oxidizing tank before being discharged
back into the Thames,
Later air and. water pollution control programs, sup-
ported by Imperial Chemical Industries, Ltd., British Power
Authority consultants, and the Howden Construction Co., led
to the development of a closed-loop, lime-based system for
removal of sulfur dioxide. This was installed at the
Swansea power plant in 1935 and at the Fulham power plant in
1937. They operated successfully until the early period of
World War II when they were shut down because the vapor
plumes provided possible aerial guidance to enemy aircraft.
The next large-scale uses of lime/limestone scrubbing
were in Russia and Japan. A scrubber installation has been
operating in Russia since 1964, scrubbing sulfur dioxide
from about 850 m /s (1.8 million acfm) of waste gas from an
iron ore sintering plant. The first lime scrubber in Japan
was on a large sulfuric acid plant in 1966.
In the United States, the Tennessee Valley Authority
(TVA) conducted small-scale and limited pilot plant studies
in the 1950's. The first major pilot plant work appears to
have been that of Universal Oil Products (at a Wisconsin
utility installation) beginning in 1965. A limestone slurry
circulating through a mobile-bed scrubber treating 0.94 m /s
(2000 acfm) gave good SO,., removal.
In 1966, Combustion Engineering tested a system in-
volving injection of limestone, followed by scrubbing, in a
pd Lot unit [1.4 m /s (3000 acfm)] at a Detroit Edison power
plant. At a stoichiometric limestone-to-SO ratio of 1.1 to
1, the SO removal was 98 percent. On the basis of this
pilot plant work, the company offered the process to the
1-2
-------
utility industry, and five systems were installed. One
installation was at the Union Electric Co. in St. Louis,
Missouri (140 MW). Kansas Power and Light Co. installed one
on a 125-MW boiler in Lawrence in 1968; another on a 400-MW
unit at the same plant started up in 1971. Kansas City
Power and Light has used the process on two boilers, one at
ICO MW, and the other at 140 MW.
Because of major problems associated with dry limestone
injection, these systems proved inadequate. The Union
Electric installation has been abandoned, and the Kansas
Power and Light systems are being replaced by a system in
which a limestone slurry is introduced into the scrubber.
The Kansas City Power and Light installations are being
converted to lime scrubbing. Problems associated with
limestone injection include plugging (especially of the
boiler tubes), low sulfur dioxide absorption, and reduced
particulate collection in the electrostatic precipitators.
1.2 APPLICATION OF FGD SYSTEMS
As of August 1977, there were 141 FGD systems with an
equivalent electrical generating capacity of 54,840 MW
planned or in operation in this country. Of these systems,
29 were operational (8,914 MW); 28 were under construction
(11,810 MW); and 68 systems were planned (32,628 MW). An
additional 16 installations (8,592 MW) are considering the
use of FGD systems as well as other control strategies such
as low sulfur content coal. Approximately 12 to 15 boilers
(6000 MW) are planning to use FGD systems, but have not
publicly released their plans.
1-3
-------
1.3 REPORT CONTENTS
This study presents a brief overview of alternative S02
control techniques in Section 2 and their potential applica-
tion for meeting alternative NSPS levels. Section 3 pro-
vides an evaluation of the current capabilities of the lime,
limestone, double alkali, magnesium oxide, and Wellman Lord
FGD processes and a brief review of several other FGD
systems. Section 4 presents an overall analysis of FGD
system performance and operability based on the information
presented in Section 3.
Information for this study was obtained from the liter-
ature and from contacts with users and suppliers of FGD
systems.
1-4
-------
2.0 OVERVIEW OF ALTERNATIVE SO,, EMISSION CONTROL SYSTEMS
To comply with current NSPS, operators of coal-fired
steam generators with heat inputs greater than 260 GJ per
hour (250 million Btu per hour) must limit S0~ emissions to
0.52 g S02/MJ (1.2 Ib SO2/106 Btu) of heat input. Emissions
can be controlled to the required level by one of the fol-
lowing methods: burning naturally occurring low-sulfur coal,
lowering the sulfur content of the coal prior to combustion
(coal cleaning), removing S0? from flue gases after combus-
tion, or using fuel treatment processes and combustion
systems that remove S07 prior to, or during combustion (coal
gasification, coal liquefaction, and fluidized-bed combus-
tion) . Selection of a control technology for a new combus-
tion source is determined by the degree of control required,
process availability, and costs.
The following sections concern methods of controlling
SO- emissions other than by flue gas desulfurization (FGD).
These methods are discussed in terms of their possible
application in new, coal-fired boilers used by utilities to
meet the current NSPS level, as well as alternative levels,
of 90 percent reduction of S09 emissions and 0.25 g SO9/MJ
£.
(0.5 Ib S02/10 Btu) heat input.*
2.1 LOW-SULFUR COAL
Reserves of coal in the United States suffice to meet
projected, long-term demands for energy production, but only
about 9 percent of the recoverable reserves are of sufficient
* Alternative levels provided by EPA,
2-1
-------
quality to enable utilities to meet the current NSPS without
fuel processing or emission controls (see Table 2-1). If a
standard as stringent as 0.25 g S02/MJ (0.5 Ib S02/10 Btu)
were promulgated, less than 1 percent of the recoverable
coal could be burned without some kind of flue gas treat-
ment .
2.1.1 Definition
Coals can be categorized according to sulfur content in
relation to S09 emissions, since approximately 90 percent of
^-
the sulfur is emitted as S02 when coal is burned. No
definition of low-sulfur coal is universally accepted,
however. Coal can be burned in compliance with the current
Federal NSPS for boilers if it emits a maximum of 0.52 g
S09/MJ (1.2 Ib SO /10 Btu) gross heat value. Figure 2-1
^ £
relates fuel sulfur content and heat value required to meet
the current NSPS SO emission limit.
Reported data on coal sulfur content represent nominal,
long-term (annual) averages, and do not take into account
short-term fluctuations. If a revised NSPS were issued
based on a shorter averaging time, such as 24 hours, the
required sulfur content of the coal would have to be much
lower to satisfy the same emission limit. Accurate sta-
tistical data on sulfur variations in coal are required in
order to predict the level of coal sulfur content needed to
meet a standard that is based on averaging times of less
than one year.
2.1.2 Coal Reserves
Data on low-sulfur coal reserves are not exact and are
subject to estimation and interpretation. Estimated amounts
of reserve base relate to coal found in seams that can be
economically mined by conventional techniques. Not all
2-2
-------
CO
15,000
14,000
1?,000
12,000
=> 11,000
i 10,000
i—i
-------
this coal can be recovered, however. Estimates must be
reduced to allow for pillars for roof support, mining losses,
overburden, etc. Traditionally, about 50 percent of the
reserve base has been used to calculate recoverable reserves
for deep mines, and 80 to 90 percent for strip mines.
Recoverable coal reserves in the continental United
States (Table 2-1) are derived from U.S. Bureau of Mines
estimates and expressed on a basis of potential SO- emis-
sions, rather than on sulfur content alone. This method
takes into account both sulfur content and heating value and
is a more useful way of estimating emissions from a parti-
cular coal. As mentioned earlier, only about 9 percent or
22.9 billion Mg (25.2 billion tons) of total coal reserves
can achieve the current NSPS level of 0.52 g/MJ (1.2 lb/106
Btu) [0.7% sulfur content at 26,300 J/g (11,300 Btu/lb)
of coal] without prior treatment. An estimate by the
University of Illinois Center for Advanced Computation,
however, based on older Bureau of Mines data, shows a total
recoverable reserve of only 15 billion Mg (16.5 billion
tons) of coal capable of meeting this standard.
Figure 2-2 displays the data in Table 2-1 graphically
and shows that the amount of complying coal would decrease
significantly if the NSPS were lowered. Its reduction to
0.25 g/MJ (0.5 lb/10 Btu), for example, would mean that
less than 1 percent of the recoverable coal reserve could
meet the standard without pretreatment or flue gas desulfuri-
zation. Available data on coal sulfur content below 0.43
g/MJ (1 lb/10 Btu) are sparse and the amount of coal with
very low sulfur levels can be estimated only crudely. Cur-
rent data, however, indicate that reserves are very limited.
2-4
-------
Table 2-1. SUMMARY OF TOTAL RECOVERABLE COAL RESERVES BY
REGION AND EMISSION POTENTIAL - CONTINENTAL U.S.
(millions of short tons)a
Coal Type/Region
Bituminous
East
Midwest
West
Eubbit uminous
East
Midwest
West
Lignite
East
Midwest
West
Anthracite
East
Midwest
Hebt
Bituminous
Subb) turn) nous
Li gn i te
Anthraci te
4 of Total
Cum'i J at i vp t
Potential SOn Emission Rate', lb/10n Bfu
• 1.2
17, 366
925
4,167
22,456
0
0
0
0
2,766
0
9
2,775
22.45B
0
0
2,775
25,233
6. 77
fl. 77
1 . 201 - ] . b
7, 708
1,571
7,103
16,382
19,219
(Qua! ity c
0
0
0
3,567
0
0
3,567
16, 38?
] 9, 219
0
3, 567
3 5 , 1 fc 0
1 3. 6J
22. 38
i . C -- i
9, 302
2,017
43
11, 362
14 ,870
: 01 - 3
12,721
2,189
331
15,241
4 ,754
}f Rescjives unknown}
0
17,072
17,072
83
0
0
83
11,302
14,870
17,072
81
4 3 , 3 il 7
] ', . OR
37. 4(,
3,435
3,172
6,607
99
17
0
116
15,241
4,754
6, 607
in,
26,716
9. 28
46. 74
>3
57,896
290, 811
1 ,022
149 , 729
0
0
1 ,756
1,756
742
17
0
759
149,729
0
1 ,756
759
152, 244
52. 90
99.64
Total
104 ,993
97,513
12,666
215,172
0
0
38,843
1,026
3,435
22,000
26,461
7,257
34
9
7,300
215,172
38,843
26,46)
7 , 300
28^ , 776
Multiply by 0.907 to convert to Mg
NOTE: Calculated from data in References 2, 3 and 4 by
using the average heat value for each coal bed and
assuming 95% conversion of sulfur to SO-.
2-5
-------
UJ _J
> ID
C£ 00
UJ
OO Q
-<
O Q
O 2:
CQ oo
o oo
O UJ
UJ I
ct:
o
f—
50
40
30
20
10
5.0
2.0
1.0
0.5
0.1
g S02/MJ
0.25 0.5 0.75
1.0
,
la i
'
f
I
I
0.5 1.0
2.0
3.0
POTENTIAL EMISSION, LB. S02/10 Btu
Figure 2-2. Recoverable coal reserve by sulfur content.
2-6
-------
2.1.3 Other Factors Affecting Use of Low-Sulfur Coal
Other factors may determine whether coal that meets the
current NSPS or alternative levels will be used. They
include:
0 Availability of mining and transportation equip-
ment.
0 Environmental impact from mining activity -
0 New regulations affecting reclamation, safety, and
utilization of regionally available coal.
2.1.3.1 Coal Mining and Transportation Equipment - The
manufacture of mining equipment (especially large draglines
for strip mining) depends largely on the willingness of
mining companies to place firm orders well in advance of
delivery, which in turn hinges on whether utilities will
commit themselves to long-term contracts. The production of
large draglines, however, has been increasing about 25
percent per year and this trend should continue for the next
few years. The equipment problem, therefore, may be af-
fected more by a shortage of mining industry capital, than
by insufficient manufacturing capacity. A possible shortage
of hopper cars is related both to shortages of steel and
castings, and to the unwillingness of railroads to invest
money without strong projections on the demand for coal
shipments. In the past, however, hopper car builders have
been able to respond quickly to increases in demand, and it
appears unlikely that hopper car delivery time will exceed
that required to develop the coal mines. Also, utilities
may buy hopper cars and enter into lease-back arrangements
with the railroads.
2.1.3.2 Environmental Effects of Mining - Environmental
effects of coal production result primarily from mining and
2-7
-------
transportation. Although specific effects depend on the
location of the mine, some general tendencies are evident.
In strip mining, for example, topsoil and overburden
are removed to reach the coal seam. The topsoil is either
stored for later use, or used immediately to revegetate
spoil piles or other disturbed areas. The ecological bal-
ance is always upset by mining operations. Even if revege-
tation efforts are successful, the previous balance of flora
and fauna is never quite restored. In addition, removal of
ground cover causes erosion and fugitive dust.
The environmental effects of underground mining include
acidic mine drainage, unsightly piles of waste materials,
land subsidence, and dust caused by coal handling and trans-
portation.
Expansion of western coal production areas is hampered
severely by lack of water, which is needed for revegetation,
pipeline slurry transportation (if used), dust control on
haul roads and in underground mines, coal cleaning opera-
o
tions, and for drinking. Water shortage will undoubtedly
be the major problem in the development of some coal fields.
2.1.3.3 Regulatory Factors Affecting Coal Mine Develop-
ment - New State and Federal regulations requiring better
reclamation of strip-mined areas and more stringent safety
restrictions also affect further development of certain
operating mines and construction of new ones.
The August 1977 amendments to the Clean Air Act (Sec.
125) will also affect the development of coal mines, because
large coal-fired boilers may be prohibited from burning coal
Q
that is not mined locally or regionally.
2-8
-------
2.1.4 Conclusions
Limited amounts of naturally occurring coal are avail-
able to meet the current NSPS. Available data indicate that
approximately 9 percent, or about 22.9 billion Mg (25.2
billion tons), of recoverable coal in the continental United
States can be burned in compliance with the current NSPS SO,,
emission limit. If this limit is reduced to 0.25 g SO~/MJ
r ^-
(0.5 Ib S02/10 Btu), only a small portion (less than 1%) of
the recoverable reserves will be able to meet the S0~
limit.
2-9
-------
REFERENCES FOR SECTION 2.1
1. Thomson, R.D. and H.F. York. The Reserve Base of U.S.
Coals by Sulfur Content. Part 1. U.S. Department of
Interior, Bureau of Mines Information Circular 8680.
Wash. D.C. 1975, p. 8 and 9.
2. Ibid, p. 33-486.
3. Strippable Reserves of Bituminous Coal and Lignite in
the U.S. Department of Interior, Bureau of Mines
Information Circular 8531, 1971, p. 70-121.
4. Western Coal. Coal Age - Mid-April, 1973.
5. Op. cit. Ref. 1.
6. Rieber, M., et al. The Coal Future: Economic and
Technological Analysis of Initiatives and Innovations
to Secure Fuel Supply Independence. University of
Illinois. PB-247 678. Prepared for National Science
Foundation. May 1975. p. 111-12.
7. Asbury, J.G., and K.W. Costello. Price and Avail-
ability of Western Coal in the Midwestern Electric
Utility Market, 1974-1982. Argonne National Laboratory
Report ANLES-38. Argonne, Illinois. October 1974.
8. Evaluation of Low-Sulfur Western Coal Characteristics,
Utilization, and Combustion Experience. EPA-650/2-75-
046. U.S. Environmental Protection Agency, Office of
Research and Development. Washington, D.C. May 1975.
9. Congressional Record - House, August 3, 1977 p. H8519.
2-10
-------
2.2 PHYSICAL COAL CLEANING
Although coal beneficiation has been practiced for many
years, especially in the metallurgical industry, application
of this technique to meet S02 regulations, or to reduce the
efficiency required of an FGD system, is a relatively new
concept. Currently no coal cleaning plants are known to be
producing coal that will meet current NSPS for boilers; but
one plant designed for this (the Homer City Plant of Penn-
sylvania Electric Co.) is in the final stage of construc-
tion. It seems likely that coal cleaning techniques similar
to those used in the metallurgical industry will find at
least limited application as a control measure for complying
with current S0~ emission limits. However, as discussed in
this section, any reduction in allowable SO- emissions will
curtail this control option for new boilers.
The following subsections briefly describe several coal
cleaning processes and assess the potential benefits and
limitations of this control technology.
2.2.1 Coal Cleaning Processes
Coal is a complex mixture of organic and inorganic
materials. Its physical and chemical properties are not
homogeneous. Various materials, including sulfur compounds,
are chemically bound within the coal structure or discretely
dispersed through it. Although large inclusions can be
liberated and removed by physical methods such as crushing,
screening, grinding, and flotation, finely distributed
mineral inclusions or chemically bound inorganic species
usually do not respond to these methods.
The cleanability (lowering of sulfur and ash content)
of coal depends on its size, the original sulfur content,
2-11
-------
the pyritic fraction (FeS2), and the size distribution.
A given coal can be cleaned to various sulfur levels by
adjusting the size distribution and the effective separation
density in the cleaning process. As the sulfur content of
the coal is lowered, however, its heating value decreases
because of carbon losses.
All physical coal cleaning processes include two basic
operations: crushing, to a size that allows liberation of
unwanted materials; and separation of these impurities,
by exploiting the physical differences between them and the
coal particles. Although some processes take advantage of
differences in surface properties or magnetic properties,
most utilize differences in density- The density of coal is
about 1.3 g/ml, whereas the densities of most inorganic im-
purities found in it are much higher. The density of lime-
stone for example, is 2.7 g/ml, and that of pyrite is 5.0
g/ml.
The washability characteristics of a given coal reserve
can be determined by extensive core drilling. The samples
are tested by float-sink techniques and analyzed. Because
of the expense involved, these analyses usually yield only a
limited amount of detailed washability data. When a coal
bed is already being mined, washability characteristics are
determined by actual production samples.
Four physical cleaning methods are now being used on a
commercial scale: the dense-medium, hydraulic, froth flota-
tion, and pneumatic processes. They are described briefly
below.
In the dense-medium process, a liquid of higher density
than coal causes clean coal to float and the heavier im-
purities to sink. The principal dense-medium processes in-
2-12
-------
corporate either vessels that rely on gravity, or cyclones
that use centrifugal force to effect separation.
The principal hydraulic coal cleaning processes involve
the use of jigs, launders, hydrocyclones, or wet concen-
trating tables. All these processes use water and movement
or agitation to separate coal fractions.
Froth flotation exploits selective surface properties
of certain coal solids. When a coal slurry is mixed with
finely dispersed air bubbles, the bubbles preferentially
attach to coal particles, rather than to ash particles,
and bring them to the surface to be removed as a concen-
trate.
Most pneumatic cleaning methods use a pulsating upward
air current to classify the coal into stratified layers.
Ash and pyritic materials are withdrawn from pockets or
wells at the bottom of the bed. Cleaned coal is taken from
the top of the stratified coal column.
2.2.2 Potential for Meeting More Restrictive SO.-, Emission
Limitations
Data collected by the Bureau of Mines during wash-
ability studies on 455 U.S. coal samples indicate that an
additional 10 percent of these samples could meet an emis-
sion level of 0.52 g/106 J (1.2 Ib S02/106 Btu) if they were
physically cleaned by methods that allow recovery of 90
2 3
percent of the coal's heat value. ' Insufficient samples
were taken to show the variability in coal sulfur content,
or to determine average sulfur contents over a given period
of time. More extreme cleaning methods (such as finer
grinding) would yield lower sulfur contents, but at the
cost of lower heating value. Figure 2-3, which summarizes
the Bureau of Mines study, indicates that the promulgation
of a lower SO9 emission standard would decrease to a small
2-13
-------
g S02/MJ
456
RAW COAL
1-1/2.INCHES TOP
SIZE 90% BTU
RECOVERY
14 MESH TOP
SIZE 50% BTU
RECOVERY
"0 2 4 6 8 10 12 14 16 18 20 22 24
COAL POLLUTANT (OR SULFUR EMISSION) POTENTIAL
LB S00/106 BTU
Figure 2 3. Effect of cleaning variable on
coal S02 emission potential.
2-14
-------
percentage the amount of U.S. coal that can be burned, even
after cleaning, without some additional form of emission
control.
Another important fact determined from these studies is
that the final sulfur levels to which coals can be cleaned
vary from one coal region to another, from one coal bed to
another within the same region, and, to a lesser extent,
from one location to another within the same mine. Gen-
eralities are, however, difficult to make, and each coal
seam should be assessed based on data for that area. These
differences in cleaning potential result from variations in
organic and pyritic sulfur levels and in the morphology of
the coal-pyrite matrix, thus making it difficult to achieve
a specified short-term average sulfur level in coal.
REFERENCES FOR SECTION 2.2
1. Leonard, J.W., and David R. Mitchell. Coal Prepara-
tion. Third Edition. American Institute of Mining,
Metallurgical, and Petroleum Engineers, Inc., New
York. 1968.
2. Kilgroe, J.D. Implementation of Coal Cleaning for SO-
Emission Control. In: EPA Program Conference Report -
Fuel Cleaning Program: Coal. EPA-600/7-76-024.
October 1976. pg. 57-65.
3. Cavallaro, J.A., et al. Sulfur Reduction Potential of
U.S. Coals : A Revised Report of Investigation, EPA
600/2-76-091. Bureau of Mines RI 8118, Washington,
D.C. April 1976.
2-15
-------
2.3 EMERGING TECHNOLOGIES
A number of systems are being developed for firing coal
and reducing S0~ emissions. Those deemed applicable to
reduce SO,., from fuel combustion are fluid bed combustion
(FBC) systems; systems for generation of low- or medium-Btu
fuel gas from coal; and systems for production of liquid
fuel from coal. Table 2-2 shows the technologies and ap-
plications that have recently received the most attention.
123
They have been the subject of several recent reviews. ' '
The high cost of synthetic fuels, possible environ-
mental impacts from production and combustion of liquid
fuels, and competition for synthetic fuels for other uses
(as transportation and synthesis gas for the chemical
industry, for example) will probably combine to limit
severely the use of synthetic fuels in large boilers. At
present, only FBC and coal gasification appear to have near-
term (5-10 years) possibilities for commercial application
to boilers. It is difficult, however, to rule out any of
the energy systems in Table 2-2 with certainty. The ability
of both fluid bed combustion and coal conversion systems to
meet emission standards may determine their rate of develop-
ment.
A number of studies are being conducted for EPA to sum-
marize the status of these technologies and their ability to
achieve decreased S02 emissions. Because of the complex
nature of these emerging technologies, they are not dis-
cussed further in this report.
2-16
-------
Table 2-2. PROCESSES FOR UTILIZING COAL
TO REDUCE SO EMISSIONS
Basic technologies
Atmospheric pressure fluid
bed boilers
Application _
Utility or industrial boilers
Pressurized fluid bed
boilers
Combined gas cycles for genera-
tion of electricity
Gasification of coal
(low and medium Btu gas)
Production of fuel for combined
cycles for generation of elec-
tricity
Production of fuel for boilers
of conventional design
Liquefaction of coal
Production of fuel for combined
gas cycles for generation of
electricity
Production of fuel for boilers
of conventional design
2-17
-------
REFERENCES FOR SECTION 2.3
1. Fennelly, Paul F., et al. Preliminary Environmental
Assessment of Coal-Fired Combustion Systems. Environ-
mental Protection Agency, Research Triangle Park, N.C.
Contract No. EPA 600/7-77-054. May 1977. p. 9.
2. lammartino, Nicholas R. Fluidized-Bed Combustion: A
Better Way to Burn Coal? Chemical Engineering, June
1976. pp. 69-71.
3. Hirschfeld, F. What's Holding Up Coal Gasification?
Mech. Eng. August 1977. pp. 32-37.
2-18
-------
3.0 FLUE GAS DESULFURIZATION SYSTEMS
Flue gas desulfurization (FGD) is the process of remov-
ing sulfur oxides, primarily SO~ , from combustion gases. In
this process, flue gases are contacted with an absorbent
medium in either an absorber or a scrubber vessel. Typi-
cally, SO,., reacts with the absorbent or dissolves in the
solution to produce a slurry that contains dissolved or
solidified sulfur compounds. A number of dry adsorption
systems are also under active development.
FGD processes are grouped into two basic categories,
regenerable or nonregenerable, based on whether the sulfur
compounds are separated from the absorbent as a by-product,
or discarded along with the absorbent as waste. Nonregener-
able processes produce a sludge that requires disposal in an
environmentally sound manner. Regenerable processes include
additional steps to convert the sulfur into by-products such
as liquid sulfur dioxide, sulfuric acid, or elemental
sulfur.
Six basic types of FGD systems are used in this country
on large, coal-fired boilers. Of these, the lime scrubbing,
limestone scrubbing, sodium carbonate scrubbing, and dual
alkali systems come under the nonregenerable category. The
magnesium oxide scrubbing system and Wellman-Lord system are
regenerable. The following sections deal with these sys-
tems. Specific installations are described in detail, as
well as their efficiency and operability, the problems in-
volved, and their solutions.
3-1
-------
3.1 LIME SLURRY FLUE GAS DESULFURIZATION SYSTEMS
Lime slurry scrubbing is a wet scrubbing process that
uses lime slurry to remove SO2 from flue gas. Dry calcium
oxide (CaO) is fed to the system and reacted with water to
form a slurry. This slurry is contacted with the flue gas
and absorbs S02. The S02 reacts with the slurry to form
calcium sulfite, which is removed from the system as a
sludge.
3.1.1 Process Description
This section presents a process description for lime-
based FGD scrubbing systems and is divided under four sub-
headings :
0 Process Flow/Material Balance
0 Process Chemistry
0 Waste Disposal
0 Water Balance
3.1.1.1 Process Flow/Material Balance - The flow diagrams
for operating lime scrubbing FGD systems appear in the
appendix of this report. A model plant shown in Figure 3-1
shows the flue gas flowing from the boiler into ESP's, where
99 percent of the particulate is removed. The flue gas then
flows through an induced draft fan into the absorber module
(stream 1). Lime is fed from a feed bin to a slaker, where
it is slurried with fresh water, diluted with recycled water
and fed into the reaction tank (stream 13). A bleed stream
(15) from the recirculating slurry goes to a thickener, the
underflow from which is pumped to the sludge pond. Pond
water and overflow from the thickener are returned to the
system (streams 7 and 8). Makeup water enters as mist
eliminator wash and as fresh water feed to the slaker. In
actual practice, fresh water would also be used at the pump
seals, but it is not shown in this case. Clean gas from the
3-2
-------
MAKEUP
WATER
REHEATtR \~
ft
REHEAT
LIME
STORAGE
WATER
Descc ipt- 1 on
SOS
Bate. 103
Ib/hr
SCFM lo'
Temp. , °F
Fly .ash
10 Ib/hr
SO;, 103
1 b/hr
H20. 10
Ib/hr
HI, io3
Ib/hr
3
CO?, 10
Ib/hr
i
0 10
Ib/hr
1
Gas to
5570
1110
320
0. 5
25.5
201
3970
1030
)<3
2
Flue GJS
5840
1110
125
0. 5
2.55
490
)970
1 030
142
3
1360
299
400
12.4
1050
320
4
Has to liquid
"20 Rate, 10J
Ib/hr
1410 CPU
115 Solide, «
Sped f ic
0 . 5 qrav ity
Temp. , °F
2. 55
CaSO • 1/2
H 0 I0]
502 IB/hr
C«SO
5020 2H 8 10
Ib/hr
1030 HO IO3
Ib/hr
Deacr ] pt ion
• olid
streams
R«te, 103
Ib/hr
5
69.6
1 38
a
1 .00
60
69.6
Limp
to
8 laker
23.4
6
285
565
0
1 . 00
60
285
Fly a
45.5
P
2
4
.
2
sh
7
ond
0. 2
0.1
0
. 00
60
0.2
Gi 1 t
0. 234
8
300
595
0. 5
1 .00
125
300
9
320
635
0
1.00
120
320
10
Return
258
512
0
i . 00
1 20
258
11
Return
62.4
1 24
0
1 . 00
120
62. 4
12
92.8
156
25 0
1.183
1 )0
69.6
1 1
Slurry
to
155
277
15 0
..111
120
1 12
14
Scrubber
39500
72000
12.0
1 .089
125
3340
568
148DO
15
St^n t
slurry to
429
782
12.0
i. 089
125
41.7
6.17
378
16
Spent
slurry
129
199
40.0
1 .286
125
41.7
6.17
77 . 5
CO
I
LO
Figure 3-1. Flow diagram/material balance - lime slurry scrubbing systems
for a 500-MW boiler.
-------
scrubber absorber is reheated, using heated air, and dis-
charged to the stack (stream 4). The assumptions needed to
make this material balance are presented in Table 3-1.
The material balance for this system appears in Figure 3-1.
3.1.1.2 Process Chemistry - Lime Slurry Preparation - The
first step of the process is preparation of the lime slurry.
Lime (CaO) is reacted with fresh water in a slaker to pro-
duce a slurry of calcium hydroxide (Ca(OH)_) and water,
1
according to the following reaction: CaO + HO -*• Ca(OH)9 +
£ £
heat. This material is diluted with recycled water to about
15- to 20-percent solids.
SO., Absorption Mechanism
Droplets of slurry pass countercurrently (usually) to
the flue gas containing S02. The S02 passes from the gas
into the droplets through an interface driven by the con-
centration gradient. The vapor-liquid equilibrium curve
for S02 and a 10-percent lime solution is shown in Figure
3-2.
SO,, Reactions
The S02 molecules that are absorbed into the liquid
droplet must be removed from the solution if the absorption
process is to continue. Dissolved S02 molecules react with
the calcium species and water according to the following
reactions:
2OH~
Ca(OH)2(s)
20H~ + CO
co" + co2
Ca++ + C0~
CaC00 , . «-
3 (s)
S02 (g) -+
+--* Ca(OH)2 (aq) +
- CO" + H20
+ H?0 -^-> 2HCO
" -J
•^"> CaCO
+ CaC03 (aq) ^^ Ca
S02 (aq) + H20 *-->
- Ca++ +
++ + CO'
H+ HSQ-
HSO~ «--»• H+
3-4
-------
Table 3-1. ASSUMPTIONS USED IN THE MATERIAL BALANCE
CALCULATIONS FOR A 500-MW BOILER
SO- removal in absorber
Coal sulfur content
Coal heat value
Lime CaO content
Stoichiometry
Sludge oxidation (CaSO /CaSO )
Fly ash/bottom ash
Particulate removal in ESP
Excess air to air heater
Excess air to boiler
L/G
90 percent
3.5 percent
27,900 kJ/kg
(12,000 Btu/lb)
95 percent
1.1 percent
90/10
80/20
99 percent
40 percent
10 percent
5.3 1/m3
(40 gal/1000 scf)
3-5
-------
16 x 10
-4
12 x 10
-4
tM
O
2 8 x 10"
i
4 x 10'
EQUILIBRIUM
•LINE
0 2 x 10~5 4 x 10"5 6 x 10~5 8 x 10"5 10 x 10"5
HOLE FRACTION S02 IN LIQUID
Figure 3-2. Vapor liquid equilibrium diagram
S0» - 10 percent CaO slurry.
3-6
-------
Ca++
++
Ca
CaS03
+
H +
CaS03
+ s°r
+ so-'
+
+ H f-
__
HC03 f~>-
•1/2 H2
<--> CaS03
+ 1/2 H2<
++
•* Ca +
H2C03 ^
0 + 3/2
(aq)
0 «-+
HS03
-co2
H20 +
CaSO
(g)
1/2
3-1/2
+ H20
°2 ~
H2°(s)
CaS04
All the above reactions take place to some extent in every
lime/limestone scrubbing system; they depend on many factors,
including ionic concentrations, pH, temperature, and reten-
tion time in the reaction tank. The net effect of the
reactions is the removal of S02 from the flue gas, depletion
of calcium hydroxide in the slurry, and the generation of
calcium sulfite in the slurry. The calcium ions used in the
reactions are replenished by dissociation of the calcium
hydroxide. This is a primary difference between the lime
and limestone slurry scrubbing systems and the so-called
clear solution scrubbers (such as double alkali scrubbers),
where the predominant reactive alkali (sodium) is more
soluble.
In order to complete the S0_ removal reactions listed
above, a certain amount of retention time is needed for the
slurry to achieve chemical equilibrium before it is recir-
culated to the scrubber. To do this, a reaction tank is
usually used with a retention time of between 5 and 20
minutes.
3.1.1.3 Waste Disposal - The calcium sulfite and calcium
sulfate that are formed in the absorber and the absorber
reaction tank must be removed from the system. This is
usually accomplished by bleeding a stream from the recircu-
lating slurry to a thickener or clarifier. The calcium
sulfite and sulfate settle out and are removed from the
3-7
-------
underflow of the clarifier in a slurry of about 30-percent
solids.4 The overflow, a clear liquid of less than 1-
percent solids, is brought back to the system. The sludge
from the underflow is disposed of by several means.
One method of disposal is to pump the sludge to a pond.
In this case the sludge is pumped from the thickener under-
flow (or occasionally directly from the reaction tank) to a
large settling pond, where the solids settle. Clear water
is recycled from the pond, as necessary, to keep the pond
from overflowing. Return of all this clear liquor from the
pond is called a "closed-loop" operation.
Another method of disposal is to dewater the sludge in
a vacuum filter, chemically fixate it and dispose of it in a
landfill. In this case there are commercially available
processes to stabilize the sludge, such as the IOCS system.
When vacuum filtration is used, the filtrate is returned to
the system to create closed-loop operation.
3.1.1.4 Water Balance - To minimize liquid discharge,
closed-loop operation is desirable, and maintaining a water
balance in the scrubbing system is very important. The
quantity of water added can not exceed the amount lost in
the flue gas and in the sludge. The moisture content of the
flue gas is increased by cooling the incoming flue gases
from about 149°C (300°F) to their adiabatic saturation
temperature of about 53°C (128°F); this is achieved by
evaporating the water in the scrubbing system.7 Water is
lost in the sludge both as water of hydration and as free
water in the sludge. Water is added to the system as
slaker feed water, pump seal water, and mist eliminator wash
water.
3.1.2 Domestic Lime Scrubbing Units
The first major domestic lime FGD system, the Phillips
Power Station of Duquesne Light Co., started up in 1972.
3-1
-------
Since that time, 11 other major stations have installed lime
scrubber systems (Table 3-2). A description of each of
these is given in the Appendix A.
It is important to note that each of these facilities
that has been tested is operating at or above the efficiency
P
required to meet SO- emission legislation.
A more thorough review of three units, Green River,
Bruce Mansfield, and Mohave, is presented in the following
pages. Respectively, these units represent high FGD system
availability, a large system with high SO,., removal effi-
ciency, and a unit with very high SO,, removal efficiencies
on a low-sulfur coal application.
3.1.2.1 Green River Facility Description - The Green River
Station of Kentucky Utilities is on the Green River in
central Kentucky, approximately five miles north of Central
City. The plant contains four steam turbine generating
units having a total gross generating capacity of 242 MW.
Boilers 1, 2, and 3, the boilers with the FGD units, supply
steam for two steam turbine generators with a combined
generating capacity of 64 MW. These two electrical gener-
ating units are used for peak loads and normally operate on
a 5-day week, with one or more of the boilers often at
reduced capacity.
All three boilers are dry-bottomed and burn pulverized
coal. They were manufactured by Babcock and Wilcox and put
into service in 1949 and 1950. There are no plans to retire
them.10
The utility purchases a high-sulfur coal, which is used
in conjunction with the FGD system. This coal comes from
the Drake Mine in Muhlenberg County, Kentucky, and is
shipped to the plant via barge. A typical analysis of it
gives the following values: heating value, 25,100 kJ/kg
3-9
-------
Table 3-2. LIME BASED FGD SYSTEMS IN THE UNITED STATES1
Station
Bruce Mansfield
Cane Run
Colstrip
Conesville
Elrama
Four Corners
Green River
Mohave
Paddy ' s Run
Phillips
Rickenbacker
Shawnee
Company
Pennsylvania Power Co.
Louisville Gas & Electric
Montana Power Co.
Columbus & Southern Ohio
Electric Co.
Duquesne Light Co.
Arizona Public Service
Kentucky Utilities
So. California Edison
Louisville Gas & Electric
Duquesne Light Co.
USAF
TVA/EPA
Size,
(MW)
835
178
360
400
510
160
64
160
65
410
20
10
Start-up
date
April 1976
August 1976
December 1975
February 1977
October 1975
February 1976
September 1975
November 1973
April 1973
April 1972
March 1976
April 1972
Status
Operational
Operational
Operational
Operational
Operational
Test facility
shutdown
Operational
Test facility
shutdown
Operational
Operational
Operational
Test facility
New or
retrofit
N
R
N
N
R
R
R
R
R
R
R
R
Coal,
% sulfur
4.7
3.5-4
0.8
4.5-4.9
1-2.8
0.7
3.8-4
0.8
3.5-4
1-2.8
3.6
0.8-5.0
Design
SOj removal
efficiency, %
92
85
60
90
85
N.A.
80
N.A.
80
83
90
N.A.
LO
I
-------
(10,800 Btu/lb); sulfur content, 3.7 percent; ash content,
13.4 percent; total moisture, 12.1 percent.
Boilers 1, 2, and 3 are fitted with mechanical col-
lectors upstream of the FGD system. Design particulate
removal efficiency is 85 percent. The FGD system was de-
signed and installed by American Air Filter (AAF) and con-
sists of one scrubber module to handle a maximum flue gas
capacity of 170 m3/s (360,000 acfm) at 149°C (300°F). Table
12 13
3-3 presents pertinent plant design and operating data. '
The process is conveniently described in terms of two
basic operations: the flue gas scrubbing system, and the
lime slurry/recycle system. Figure 3-3 is a schematic flow
diagram of the process.
Flue Gas Scrubbing System
The flue gas from each boiler coupled into the scrub-
bing system initally passes through a series of mechanical
collectors (Western Precipitation, multicyclone, 9-in.
diameter, cast iron construction) where primary particulate
removal takes place. The flue gas is then drawn from the
existing breeching through a guillotine-type isolation
damper and associated ductwork to the scrubber fan. The
incorporation of the guillotine-type isolation dampers
allows selective bypassing of the flue gas around the
scrubbing system to an existing stack.
Before entering the scrubbing system, the flue gas
passes through a booster fan with a power rating of 1118 kW
(1500 Hp) and a pressure increase of 0.46 m (18 in.) H^O.
From the outlet of the scrubber booster fan, the gas flows
through a variable-throat flooded elbow venturi scrubber.
The venturi scrubber was installed to control particulate
emissions from the mechanical collectors and to quench the
hot gas. Quenching lowers the temperature of the inlet gas
3-11
-------
Table 3-3. POWER PLANT AND FGD SYSTEM DESIGN DATA
16
Green River - Kentucky Utilities
Boiler data
Coal data
FGD system
data
Generating capacity, MW
Year placed in service
Boiler manufacturer
Heat value
Ash content
Sulfur content
S02 removal efficiency
Particulate removal
efficiency
Start-up date
Flue gas rate
Flue gas temperature
Stack height
FGD vendor
64
1949
Babcock & Wilcox
25,100 kJ/kg
(10,797 Btu/lb)
13 to 14 percent
3.8 percent
80 percent design
99.7 percent
9/75
170 m3/s
(360,000 acfm)
149°C (300°F)
24 m (78 ft)
American Air
Filter
3-12
-------
OJ
I
ELECTRICAL
GENERATING
UNIT NO. 2
MAKEUP HATER
WKEUP WATER
Figure 3-3. Process flow diagram - Green River No. 1, 2, and 3.
15
-------
from 163°C (325°F) to approximately 47°C (116°F) within the
scrubber module. This substantially decreases the volume of
gas to be scrubbed and provides additional temperature
protection for the balls in the mobile bed contactor.
The venturi has a throat of 2.54 m (100 in.) diameter
operating with a 2.29-m (94-in.) plug. Pressure drop
through the venturi is maintained at 0.18 m (7 in.) H20 by a
Limitorque operator on the plug. Liquid flow is maintained
at 83 1/s (1360 gpm) through the top of the scrubber. The
scrubber shell is constructed of mild steel and lined with
acid-proof Precrete. The venturi throat is constructed of
stainless steel. After the venturi, the gas passes through
a flooded elbow and flows upward through the mobile bed con-
tactor at a rate of 136 m3/s (288,200 acfm) at 47°C (116°F).
Sulfur dioxide in the flue gas reacts with the lime slurry
at this stage. The SO2 absorber consists of a mobile bed
contactor section and a demister section. The mobile bed
stage contains approximately 175,000 to 190,000 solid poly-
ethylene balls of 3.2 cm (1.25 in.) diameter. It is con-
strained to a maximum thickness of 0.61 m (2 ft), or 0.41 m
(16 in.) at rest. The slurry is fed at a rate of 595 1/s
(9750 gpm) and applied both to the bed and to the upward
rising flue gas by 125 five-cm (2-in.) overhead nozzles and
sixty-five 3.2-cm (1.25-in.) nozzles spraying upward into
the bed. The contactor bed is compartmentalized into 10
individual sections. Underbed dampers are used to meet
flue gas turndown requirements. Pressure drop through the
contactor bed is 10 cm (4 in.) HO.
Following passage through the bed, the gases continue
upward to a single stage, spin vane demister. The demister
is continuously washed by spray nozzles at a rate of 3.1 1/s
3-14
-------
18
(50 gpm). Pressure drop is 2 in. HO.
3
The scrubbed flue gas (140 m /s @ 47°C or 296,300 acfm
@ 116°F) is then discharged to the atmosphere through a wet
stack. This stack is constructed of carbon steel and lined
with Carboline.
Lime Slurry/Recycle System
The scrubbing slurry feed and recycle system consists
of a partitioned concrete reactant tank, recycle pumps to
supply the scrubber and absorber module, a lime slurry
slaking and feed system, a bleed system for scrubbing waste
discharge to a settling pond, and a return water system for
recycling the water from the settling pond back to the
process.
The reactant tank, constructed of acid-proof concrete,
is 22 m (72 ft) long; 7.3 m (24 ft) wide; and 7.3 m (24 ft)
high. Two partitions form three individual compartments,
each agitated and connected by underflow openings. The
total liquid capacity of each compartment is 378.5 m
(100,000 gal.). The total retention time in the reactant
19
tank system is 21 minutes--? minutes per compartment.
From the reactant tanks, recycle pumps feed both the
venturi particulate scrubber and the mobile bed contactor.
These pumps (two operational, one spare) are rated at 360
1/s (5900 gpm) each. All pumps and agitators are rubber-
i • ^20
lined.
Reaction products and collected particulate matter are
pumped to an impervious clay-lined pond on the plant site
approximately 805 m (1/2 mile) from the scrubbing module.
The pond has a capacity of 182,630 m (148 acre ft) and a
depth of 6.1 m (20 ft). River water is used as makeup and
is introduced into the reactant tank, lime slaking tank, and
demister, as well as the pump seals. Total fresh water
21 22
makeup supplied to the system is 4.6 1/s (75 gpm). '
3-15
-------
Operating History
Commercial operation of the scrubbing system commenced
in the fall of 1975. Before commercial service, the system
was put through an extensive four-phase prestart-up testing
program, which included mechanical and electrical trials
operation on air and water, verification of mechanical
reliability, and hot flue gas operation.
The operation on flue gas began September 13, 1975. It
was conducted on a 50-percent load basis with 1.6- to 2.0-
percent sulfur coal. During this phase, several minor
problems were uncovered and corrected. The major areas of
concern included the pH sensors and SO,, analyzers. In
addition, some problems were encountered with plugged spray
nozzles.
Closed-loop, full capacity operation began in March
1976 with the initiation of a 6-month, vendor qualification
23
operation period. To date, the performance of the system
has been good. Mechanical reliability has been excellent.
The index values for average system availability, oper-
ability, reliability, and utilization (including start-up)
for 1976 are presented in Table 3-4. The operability is
plotted through May 1977 (Figure 3-4).
Based on continuous monitoring data recorded by AAF,
SC>2 removal efficiency has been well above the design
guarantee value, at about 90 percent.
Problems and Solutions
Some chemistry-related problems have been encountered.
Specifically, a hard coat of scale developed in the lower
sections of the absorber internals. This was believed to
have been caused by a high scrubbing solution pH combined
with high oxygen concentrations in the flue gas. Chemically,
this happened as follows: As the pH of the scrubbing solu-
3-16
-------
CO
I
Table 3-4. GREEN RIVER POWER STATION OPERATIONAL DATA FGD UNIT
FOR 1976 AND 1977 (SEPTEMBER)
25,26
Year
1976
1977
Month
January
February
March
April
May
June
July
August
September
October
November
December
January
February
March
April
May
June
July
August
September
Parameters
A
Hours
in
period
744
696
744
720
744
720
744
744
720
744
720
744
8784
744
672
744
720
744
720
744
744
740
B
Hours FGD
system
available
312.00
486.17
721.72
648.00
606.18
720.00
665.85
722.45
617. 20
744.00
720.00
539.31
7502.88
698.29
242. 80
0
288.00
735.65
720.00
744.00
744 .00
740.00
C
Hours FGD
called
upon
456.00
499. 38
408.66
552.00
455.88
596.43
583.53
744.00
571.20
698.55
704.25
591.48
6861. 36
744.00
266.12
0
166.82
526.55
34.38
0
0
0
D
Hours FGD
system
operated
64.00
210.75
386. 38
552.00
455.88
588.85
574.43
722.45
571. 20
698.55
704.25
517.20
6045. 94
698.26
242.80
0
164.00
513.27
34. 38
0
0
0
E
Hours
boilers
operated
571.55
499.38
457.53
552.00
455.88
596.43
583.53
744 . 00
571.20
698. 55
704. 25
535. 52
6969.82
744.00
266.12
0
166.82
526.55
34. 38
0
0
0
B/A
%
avail-
ability
41.9
69.9
97.0
90.0
81.4
100. 0
89.5
97.1
85.7
100.0
100.0
72.5
85.4
93.9
36.1
0
40.0
98.9
100.0
100.0
100.0
100.0
D/C
%
Reli-
ability
14.0
42.2
94 . 5
100.0
100.0
98.7
98.4
97.1
100.0
100.0
100.0
87.4
88 1
93. 9
91.2
0
98.3
97.5
100.0
0
0
0
D/E
%
oper-
ability
11.2
42.2
84 .4
100.0
100.0
98 .7
98.4
97.1
100.0
100.0
100.0
96.6
86.7
93. 9
91.2
0
98.2
97.5
100.0
0
0
0
D/A
%
utiliza-
tion
8.6
30.3
51.9
76.7
61.2
81.8
77.2
97.1
79. 3
93.9
97.8
69.5
68.8
93. 9
36.1
0
22.8
69. 0
4.8
0
0
0
-------
LO
I
J_
SHUT DOWN FOR
STACK REPAIR
JAN FEE MAR APR MAY
JUN JUL
1976
AUG SEP OCT
MONTHS
NOV DEC JAN FEB MAR APR MAY JUN
1977
Figure 3-4. Scrubber system operability - Green River No. 1, 2 and 3.
27
-------
tion became more alkaline, in the range from 9 to 10,
calcium sulfite tended to precipitate. In the presence of
high oxygen concentration, the sulfite was chemically
oxidized to sulfate, resulting in the formation of a hard
CaSO, scale. This problem was corrected by manually removing
the scale, lowering the oxygen content of the flue gas by
minimizing air leakage, and relocating the pH sensors to a
turbulent part of the absorber reaction tank. Following
completion of these modifications, scale formation has not
been a serious problem; and although some scale film has
been periodically noted, this has tended to disappear with
2 8
continued operation and periodic cleaning.
Acid fall-out from the plume has resulted in the need
to compensate a number of employees for damage to the finish
on their automobiles. A pH as low as 2 has been found in
29
moisture drained from the stack.
The system was shut down for 65 days from February to
April 1977, to cover one half of the inner stack wall for
its entire height with 9.5-mm (3/8-in.) steel plate. The
stack was then lined with 19-mm (3/4-in.) precrete over a
wire mesh. The original lining of Carboline had failed
around nearly half the circumference.
The utility is now replacing the radial vane demister
with a chevron type to reduce the acid mist loading. To
avoid further damage to automobiles and the possibility of
damage to an adjacent substation, AAF has been authorized to
design and install a stack gas heating system. The proposal
is to use extraction steam from another unit; this with heat
air to be injected into the stack and thus raise the stack
temperature by some 10°C (50°F) . 30'" 31
3.1.2.2 Bruce Mansfield Facility Description - The Bruce
Mansfield plant is a 2700-MW, three-unit, coal-fired facil-
3-19
-------
ity on the Ohio River in the Borough of Shippingport,
Pennsylvania. It was built by Pennsylvania Power Co., which
is acting on its own behalf and as an agent for the other
participating companies of the CAPCO consortium, namely, The
Cleveland Electric Illuminating Co.; Duquesne Light Co.;
Ohio Edison Co.; and Toledo Edison Co. All three boilers
will burn 4.7-percent-sulfur coal and require FGD to meet
emission standards. Unit No. 1 was started up in April
1976. Unit No. 2 is currently undergoing start-up and Unit
No. 3 is scheduled for start-up in 1980. The remainder of
this section discusses the Bruce Mansfield No. 1 facility.
Bruce Mansfield No. 1 is a coal-fired, once-through,
supercritical steam generator which fires 302 Mg/hr (333
tons/hr) of coal and generates approximately 2.9 Mg/hr (6.5
106 Ib/hr) of steam at 26.1 MPa (3785 psig) and 541°C
(1005°F). Emission control equipment required for this unit
is designed to meet state emission regulations of 25.85
S02/108 J (0.6 Ib S02/106 Btu) and 0.04 g/m3 (0.0175 gr/scf)
of particulate, when burning 27,700 kJ/kg (11,900 Btu/lb)
coal having average ash and sulfur contents of 12.5 and 4.7
percent respectively. Additional design related information
32
is presented in Table 3-5.
The emission control system is a venturi wet scrubber
system for removal of sulfur dioxide and particulates. The
system was designed and manufactured by Chemico Air Pollu-
tion Control Co. and uses the Dravo Corporation's thiosorbic
lime as the scrubbing absorbent. A flow diagram is shown in
Figure 3-5.
Each scrubbing train consists of a variable-throat
venturi, a 6711-kW (9000-hp) I.D. fan, and a fixed-throat
venturi absorber vessel. There are six scrubber trains, all
of which are required to handle the flue gas flow from the
3-20
-------
Table 3-5. POWER PLANT AND FGD SYSTEM DESIGN DATA
Bruce Mansfield No. 1 - Pennsylvania Power Co.
33
Boiler data
Coal data
FGD system
Generating capacity
Year placed in service
Boiler manufacturer
Heat value
Ash content
Sulfur content
SO2 removal efficiency
Particulate removal
efficiency
Start-up date
Flue gas rate
Flue gas temperature
Stack height
FGD vendor
825 MW (net)
1976
Foster-Wheeler
Corp.
27,700 kJ/kg
(11,900 Btu/lb)
12.5 percent
4.5 to 5.0 percent
92 percent
99.8 percent
4/76
1580 m3/s
(3,350,000 acfm)
196°C (385°F)
290 m (950 ft)
Chemico
3-21
-------
CLEAN GAS
FLUE
PLANT
BOILER
to
SCRUBBING
TRAIN
SPENT SLURRY
AND FLY ASH
TACK
THICKENED SLUDGE
WIN
SLURRY
row
TO SCRUBBERS
FOR MAKEUP
WATER
SUCTION
BOOSTER PUMP
SUPERNATANT
I \ 1 t J
\ > "7
1 RESERVOIR !
BACK-UP LINE
£ .BACK-UP UNf j
PIPELINES TO RESERVOIR
Figure 3-5. Simplified process diagram - Bruce Mansfield No. I.34
-------
boiler at full load. The scrubbers/absorbers are arranged
in two groups of three trains. The scrubbed flue gas is
ducted through a common reheat chamber (one for each three
train group) and then exhausted through two separate flues
in the 290-m (950-ft) stack.35
The adjustable-throat venturi scrubbing module removes
nearly all of the fly ash contained in the flue gas; the
absorber module removes most of the remaining fly ash.
Sulfur dioxide is absorbed in both the scrubber and absorber
by droplets of lime slurry, containing 2- to 6-percent
magnesium oxide. The unit is guaranteed to remove 92 per-
cent of the S0? and 99.8 percent of the particulate matter
o /r
from the flue gas.
The scrubber-recycle bleed is combined with a fly ash
slurry from the boiler and discharged to a thickener of 61 m
(200 ft) diameter. Sludge from the thickener underflow is
pumped to a waste disposal system, where it is mixed with a
stabilizing agent (Calcilox) and then pumped approximately
11 km (7 miles) to the Little Blue Run Ravine for landfill.
The initial shakedown and debugging phase of operation
38
began for part of the system in December 1975. Full com-
39
mercial operation began in April 1976.
Performance History
This facility reported 98-percent operability during
the first 7 months after commercial certification (June 1,
1976). Operating problems were solved without causing
boiler downtime. However, during the exceptionally cold
winter months of January and February 1977, the boiler lost
11 percent and 24 percent of generation respectively. On
March 21, 1977 the system was taken down for a 10-week
turbine overhaul. During this time, repairs also began on
the chimney flue for A, B, and C modules. The polyester
flakeglass lining had failed and was being replaced.
3-23
-------
Previously a spray application had been used, but the new
lining is being troweled on. The other chimney flue, for
the D, E, and F modules, also needs repair. Roughly one
year will be required to complete the work and the boiler
40
will be held to half load for that time. During the
boiler and chimney outages, Pennsylvania Power will be
making other system changes, as described in the following
paragraphs.
Performance Test - Pennsylvania Department of Environ-
mental Resources tested the unit July 19 and 20, 1977.
Sulfur dioxide emissions were 19 g/10 J (0.44 Ib S02/10
Btu) and 54 g/108 J (1.26 lb/106 Btu)., respectively. This
represents 89- and 78-percent removal. The allowable
emission rate is 26 g/108 J (0.6 lb/106 Btu). The details
and causes for such a wide variation in results have not
been published.
Operating Problems and Solutions
Mist Eliminator - The major problem is with the four-pass,
chevron-type, horizontal mist eliminator. The company has
experienced mist carry-over during scrubber operation. The
mist eliminator is designed for liquid carry-over of 2.3
3
g/m (1 gr/scf), but plant personnel estimate an actual
carry-over of about 6.9 g/m3 (3 gr/scf). The eliminator was
designed to operate at gas velocities of 2.4 to 3 m/s (8 to
10 ft/s). Pennsylvania Power would like to install a new
vertical mist eliminator in the ducting downstream of the
scrubber vessels; because of duct diameter and space restric-
tions, however, such a design would necessitate flow veloc-
ities as high as 15.2 m/s (50 ft/s). A plastic vertical
mist eliminator was installed, but collapsed as a result of
structural failure. The problem of plugging was also
apparent, but has been solved by relocating the lime feed to
the scrubber at a lower position.42'43
3-24
-------
Closed-Loop - The system is not closed-loop, since water is
being retained at the sludge disposal site and not recycled
to the process. Makeup water from the disposal pond is not
needed, because "fresh" water is added to the system as fan
sprays and for lime slaking. The company thinks it can
operate closed-loop; it has not done so because of its
preoccupation with the system's mechanical problems. In the
event it does not operate closed-loop, however, the plant
has a permit to discharge water to the Ohio river. Dis-
charge requirements for water are: total suspended solids
of 60 mg/1 or less; aluminum content of 10 mg/1 or less; and
44 45
a pH between 6 and 9. '
I.D. Fans - Pieces of scale were carried through the system;
the rubber lining on the fan scrolls was chipped away,
resulting in exposure and corrosion of the carbon steel.
These fan scrolls are now being replaced with ones made of
Inconel. Pitting has also occurred on the carbon steel
hubs, the result of seal leakage. The fan shafts are made
46
of carbon steel clad with Carpenter 20 alloy.
Reheater - The reheaters are not used because of duct
vibrations caused by resonant conditions. They are now
undergoing design changes to correct the problem. Without
reheat, plume buoyancy appears sufficient, but atmospheric
conditions can occasionally cause condensation and precipi-
tation of moisture from the plume. Winds are such that the
liquid fallout occurs over the nearby town of Shippingport.
This fallout is a clear liquid, but on drying leaves a film
on automobiles, windows, and other surface. The pH of the
47
condensed liquid is about neutral.
Scrubber - Some plugging of the venturi nozzles with scale
has occurred. This problem was eliminated by installation
of "baskets" in the main header to catch the scale particles
3-25
-------
upstream of the nozzles. The baskets can be cleaned without
shutting down the system. Maintenance of the scrubber
system is not performed on a routine basis, but only when
needed. Scheduled maintenance occurs every 12 months.
pH - pH had to be manually sampled and controlled because of
difficulties with the automatic system. This system is
49 50
being redesigned by plant personnel. '
3.1.2.3 Mohave Generating Station Facility Description - As
part of their compliance program, the participants of the
Navajo/Mohave Power Project funded a prototype scrubber
demonstration project at the Mohave Generating Station.
Participants in the project are:
Salt River Project Agricultural Improvement and Power
District
Arizona Public Service Company
Department of Water and Power of the City of Los
Angeles
Nevada Power Company
Tucson Gas and Electric Company
Bureau of Reclamation of the United States Department
of the Interior
Southern California Edison Company
Design of the 170-MW demonstration facility was based
on the results of an extensive pilot plant program in which
different scrubber configurations were tested, (Table 3-6).
Two types of absorber system were built for the demonstra-
tion tests. One was a horizontal cross-flow scrubber
(Figure 3-6) and the other a vertical counter-current flow
system. The horizontal module was designed to treat 212
m /s (450,000 scfm) of flue gas. In cross-section the unit
was 4.6 m (15 ft) high by 8.5 m (28 ft) wide. It was 14.6 m
(48 ft) long. A 1305-kW (1,750-hp) booster fan supplied
flue gas to the scrubber from the discharge of the ESP.
3-26
-------
Table 3-6. POWER PLANT AND FGD SYSTEM DESIGN DATA
Mohave Test Plant - Southern California Edison
53
Boiler data
Coal data
FGD system
Generating capacity
Year placed in service
Boiler manufacturer
Heat value
Ash content
Sulfur content
SO,, removal efficiency
Particulate removal
efficiency
Start-up date
Flue gas rate
Flue gas temperature
Stack height
FGD vendor
790 MW
1971
Combustion
Engineering
26,800 kJ/kg
(11,500 Btu/lb)
10 percent
0.8 percent
95 percent
93 percent
11/73
212 m3/s
(450,000 scfm)
149°C (300°F)
152 m (500 ft)
Southern California
Edison
Stearns Roger
3-27
-------
u>
I
M
CO
FOUR SCRUBBER STAGES
HOT AIR
INJECTION
BOOSTER FAN
MIST ELIMINATOR
Figure 3-6. Horizontal test module Mohave plant.54
-------
The demisters were in the discharge side of the absorber and
reheat was provided by externally heated ambient air. The
flue gas was discharged through a 152-m (500-ft) stack. The
module had four stages with two slurry pumps and one spare
for each stage. Flow was 549 1/s (9000 gpm) through 36
nozzles per stage. The slurry was sprayed perpendicularly
tc the gas flow (i.e. cross-current) and was recycled from
52
stage to stage countercurrent to gas flow.
This horizontal module was operated from January 16,
1974, until February 9, 1975, and was tested more exten-
sively than the vertical module. The dewatering system
consisted of a thickener tank and a sludge disposal pond.
Water from the pond was returned to the system.
The vertical module, which consisted of either a TCA or
packed grid (PPA), was also rated at an equivalent of 170
MW, but this system was operated mainly on limestone (see
Section 3.2),55
Performance Results - Sulfur dioxide removal efficiency was
excellent for all three types of absorbers. Although the
S09 inlet concentration was only 200 ppm, all three types of
^
absorber were capable of removing 95 percent of the inlet
S02. Outlet S02 loadings ranged from 1 to 10 ppm range.
The S0« removal efficiency was strongly dependent on
L/G for all three modules described in Section 3.1.4 and
shown in Figure 3-7. Note that the L/G shown for the
horizontal module represents the ratio in each stage.
Sulfur dioxide removal efficiency was directly related
to the number of contact stages employed. These data are
shown in Section 3.1.4.5 for the horizontal module. The
first stage removed 50 percent of the S0~ . Two-stage removal
3-29
-------
CIRCULATING LIQUOR FLOW RATE PER STAGE
(1/s)
0.5
0.75
1.0
1.25
1.5
1.75
100
99
98
97
2
CO
o
00
96
LU
95
94
93
92
91
90
HORIZONTAL
4 STAGES
LIME
10 15 20
CIRCULATING LIQUOR FLOW RATE PER STAGE
(1000 GPM)
25
30
Figure 3-7. Effect of circulating liquor flow rate
on SO- removal at constant gas flow 212 m /s
(450,000 scfm) Mohave plant.57
3-30
-------
efficiency was 82 percent. Successive stages removed more
SO,,, up to about 99 percent at five stages. These tests
3
were run at 212 m /s (450,000 scfm) at an inlet SO,, concen-
3
tration of 220 ppm. The L/G was 2.7 1/m (20 gpm/1000 scfm)
per stage. Both configurations of the vertical module
operating on limestone showed a similar relationship.
An attempt was made to maximize use of cooling tower
blowdown as makeup to the scrubber system, although it
contained 12,000 ppm total dissolved solids. The horizontal
module used 75 percent cooling water blowdown without
formation of scale in the scrubber. The tests on the TCA
and packed grid (PPA) absorbers were inconclusive, although
the PPA showed no signs of scale with 40-percent cooling
tower blowdown. Details of the water balances are shown in
Figure 3-8.
Water was lost from the horizontal system in the
f it 58
following ways:
9.03 1/s (148 gpm) as vapor in the flue gas
0.24 1/s (4 gpm) evaporation from the pond
0.24 1/s _(_4 gpm) with the sludge
9.51 1/s (156 gpm) Total
7.14 1/s (117 gpm) of makeup was supplied by cooling tower
blowdown and 2.4 1/s (39 gpm) as fresh water.
Water was lost from the vertical module in the follow-
ing ways :
9.03 1/s (148 gpm) as vapor in the flue gas
0.40 1/s (6.5 gpm) as evaporation from the pond
0.34 1/s (5.5 gpm) with the sludge
9~7T7 1/s (160 gpm) Total
Makeup was supplied from two sources: 5.9 1/s (96 gpm)
fresh water and 3.9 1/s (64 gpm) cooling tower blowdown.
3-31
-------
9.03 1/S (148 GPM)(eq.)
HORIZONTAL SCRUBBER
PLUS THICKENER
H20 VAPOR OUT STACK
MAKEUP M-
LIME SLURRY
185,000 PPM TDS
2.4 1/S (39 GPM)SERVICE H9«
7.1 1/S (117 GPM)COOLING ^
TOWER BLOWDOWN
12,000 PPM TDS
9.5 1/S (156GPMX255K FRESH H,0)
RETURN FROM POND
1.1 1/S (18 GPM)H20
TO POND
1.6 1/S (26 GPM)
0.24 1/S (4 GPM)EVAPORATION
0.18 1/S (3 GPMJFREE WATER IN SLUDGE
0.06 1/S (1 GPMJMATER OF HYBRATION
0.48 1/S (8 GPM)= 5.1% OF MAKEUP
VERTICAL PPA SCRUBBER
PLUS THICKENER
9.03 1/S (148 GPM)(eq.)
H00.VAPOR OUT STACK
MAKEUP H20-
LIMESTONE SLURRY
70,000 PPM TDS
5.9 1/S (96 GPM)SERVICE H00
3.9 1/S (64 GPM)COOLING i
TOWER BLOWDOWN
12,000 PPM TDS
9.8 1/S (160 GPM)(60% FRESH H20)
RETURN FROM FILTER
1.31 1/S H?0
(21.5 GPM) *
TO FILTER
2.04 1/S (33.5 GPM)H90
FILTER
0.40 1/S (6.5 GPM)EVAPORATION
0.27 1/S (4.5 GPM)FREE WATER IN FILTER CAKE
0.06 1/S (1 GPM) WATER OF HYDRATION
0.73 1/S (12 GPM) 7.5% OF MAKE-UP
Figure 3-8. Water balances - Mohave plant.
59
3-32
-------
Operating History - Table 3-7 shows the performance history
of the system. Since this was a test facility, several
design changes were made during the period and this work
contributed to the unavailability. It should be emphasized
that the problems with the horizontal mist eliminator gas
flow distribution, presented in Table 3-8, once solved,
would not be a problem to future modules of this type. As a
result, operability data are not very meaningful.
Several equipment-related test programs were conducted.
For-instance, two types of mist eliminators were tried, one
for the horizontal module and one for the vertical module
(Figure 3-9). It was concluded that a horizontal-type mist
eliminator with intermittent wash on the front and back
sides was preferred to the two section vertical-type mist
6")
eliminator with the described wash configuration. Also,
two types of reheat systems were investigated (Figure 3-10).
The same amount of heat input was used for both the direct-
and the indirect-type reheater. The indirect reheat did not
show corrosion attack because the steam eliminators were
only in contact with ambient air; the direct-type did show
corrosion attack.
3.1.3 Description of Foreign Installations
About half the Japanese FGD installations use lime or
64
limestone for the absorbent in the circulating slurry.
Table 3-9 gives performances of S0~ scrubbers on large
boilers in Japan using lime slurry. There are also many
lime scrubbing systems on smaller boilers and on new combus-
tion processes. Of the 13 boilers, 10 have SO_ removal
efficiencies in excess of 90 percent, some as high as 98
percent. In all except one case, the boilers are firing oil
of low- to medium-sulfur content. One boiler is firing
medium-sulfur coal.
3-33
-------
Table 3-7. UNAVAILABILITY HISTORY - MOHAVE 170 MW TEST MODULES PROGRAM
60
LO
I
OJ
JANUARY 16,1974 TO FEBRUARY 9,1975
HORIZONTAL
1.
2.
3.
4.
5
6.
7
8.
9.
MODIFY AND REPAIR PLASTIC DEMISTER BLADES
CORRECT BOOSTER FAN BALANCE PROBLEMS
REPAIR CRITICAL PUMPS
REPLACE WORN-OUT SPRAY NOZZLES
MODIFY INLET GAS FLOW DISTRIBUTION
REPAIR HOPPER LEAKS
REMOVE HARDHAT FROM THICKENER
MODIFY SLAKING WATER TO PREVENT SCALE
CONDUCT INSPECTIONS FOR LONG RUNS
TOTAL
TOTAL CALENDAR TIME
HOURS
503
317
256
238
162
135
82
45
19
1757
9328
NOVEMBER 2,1974 TO JULY 2,1975
VERTICAL
1.
2.
3.
4.
5.
6.
7.
8.
REPAIR GRIDS AND REDISTRIBUTE TCA BALLS
CLEAN SCALE FROM SCRUBBER INTERNALS
REPAIR OR REPLACE PLUGGED NOZZLES
REPAIR LEAKS IN TRAP-OUT TRAY
REPAIR /REALIGN PPA PACKING
CORRECT BOOSTER FAN TRIP PROBLEMS
CONDUCT INSPECTIONS FOR LONG RUNS
REMOVE HARDHAT FROM THICKENER
TOTAL
TOTAL CALENDAR TIME
HOURS
710
344
153
120
85
72
55
46
1585
5813
PERCENT UNAVAILABILITY = 18.7
PERCENT UNAVAILABILITY = 27.2
-------
Table 3-8. OPERATIONAL DATA - MOHAVE HORIZONTAL FGD UNIT
61
JANUARY 16, 1974 TO FEBRUARY 9, 1975
Months
1/74
2/74
3/74
4/75
5/74
6/74
7/74
8/74
9/75
10/75
11/74
12/74
1/75
2/75
Hours
in
period
360
672
744
720
- 744
720
744
744
720
744
720
744
744
208
Avail-
able
319
402
592
716
695
552
468
744
720
506
333
739
688
116
FGD hours
called
upon
358
481
590
563
620
695
596
615
575
.540
624
303
591
208
Opera-
ting
317
393
501
559
571
552
468
615
375
398
249
298
515
116
Boiler
hours
358
481
682
616
704
695
596
615
647
540
708
572
636
208
Parameter
1
Avail-
ability
89
60
80
99
93
77
63
100
100
68
46
99
90
56
2
Relia-
bility
89
82
85
99
92
79
79
100
100
74
40
98
87
56
3
Opera-
ability
89
82
73
91
81
79
79
100
58
74
4
Utiliza-
tion
88
58
67
78
77
77
63
83
52
53
35 35
52 40
J
81 ' 69
56
56
U)
I
OJ
U1
-------
HORIZONTAL MODULE TYPE
FRONT AND BACK SIDE
WASHED AT 8-HR. INTERVALS
VERTICAL MODULE TYPE
CONTINUOUS DOWN WASH ON
TOP SIDE OF FIRST SECTION
NO WASH ON SECOND SECTION
OJ
I
(_0
CTi
GAS
FROM
SCRUBBER
COLLECTED
DROPLETS
FALL BY
GRAVITY BACK]
TO SCRUBBER
COLLECTED DROPLETS FLOW
BY GRAVITY TO SUMP
GAS FROM SCRUBBER
NOTE: GAS FLOW FROM SCRUBBER TENDS TO
PUSH SMALLER DROPLETS BACK INTO
MIST ELIMINATOR BLADES
Figure 3-9. Mist eliminators Mohave plant.
66
-------
INDIRECT REHEAT
259 mj/s 0 52°C
(550,000 ACFM)
(P 125°F)
0.1 kg H20/kg GAS
(0.10 L6 H20/LB GAS)
(SATURATED)
56.4 mj/s P 24°C
(120,000 ACFM P 75°F)
0.01 kg H20/kg AIR
(0.01 LB H20/LB AIR)
334 nr/s P 77°C
(710,000 ACFM P 170°F)
0.08 kg H20/kg GAS
(0.08 LB H20/LB GAS)
18,100 kg/hr
(40,000 Ib/hr) STEAM
CARBON STEEL COILS (FINNED)
AMBIENT AIR FAN
DIRECT REHEAT
278 m3/s 0 88°C
(610,000 ACFM P 190°F)
0.10 kg H20/kg GAS
(0.10 LB H20/LB GAS)
18,100 kg/hr
(40,000 Ib/hr) STEAM
E - BRIGHT
'HIGH ALLOY COILS
(SMOOTH)
259 m3/s P 52°C (550,000 ACFM P 125°F)
0.10 kg H90/kg GAS (0.10 Ib H20/lb GAS)
(SATURATED)
Figure 3-10. Reheater comparison for equivalent reduction
in fog formation - Mohave plant.
3-37
-------
Table 3-9. PERFORMANCE OF LARGE SO2 SCRUBBERS USING
LIME SLURRY ON BOILERS IN JAPAN68
Plant site
Amagasaki
Omuta
Kainan
Hachinohe
Karita
Amagasaki
Niigata
Ehime
Mizushima
Owase
Iwakuni
Confidential
Confidential
Start-up
date
1972
1972
1974
1974
1974
1975
1975
1975
1975
1976
1976
1976
1976
Plant
capacity,
MW
32
164
150
122
175
125
170
86
192
2 x 375
64
125
384
Fuel
Oil
Coal
Oil
Oil
Oil
Oil
Oil
Oil
Oil
Oil
Oil
Oil
Oil
S02
concentration ,
ppm
Inlet
700
2000
550
850
800
500
700
1700
1050
1500
1400
500
550
Outlet
70
200
60
85
75
50
70
60
40
35
50
65
65
Removal
efficiency,
%
90
90
89
90
91
90
90
96
96
98
96
87
88
3-38
-------
All of the Japanese systems have attained an operability*
of 95 percent or better. The Omuta Plant, firing medium-
sulfur coal, for example, has been trouble-free since start-
up in 1972. Another example is the Karita plant, treating
flue gas from an oil-fired boiler. It went into operation
in November 1974 -and operated at 100-percent availability
through 1976 (last report date). Such high operability is
normal on Japanese systems. Several reasons are given as to
why Japanese operability is so much better than that in the
United States:70
0 Japanese emission legislation and enforcement is
much stricter than in the United States. As a
result, systems are installed with no bypass and
must operate or the boiler cannot operate. This
puts increased emphasis on installing reliable FGD
systems.
0 The Japanese systems are permitted to operate
open-loop and to have a purge stream to remove
soluble ions from the system.
0 The Japanese systems tend to operate at a lower
stoichiometric ratio, which tends to reduce the
solid loading and also the plugging in the ab-
sorber.
0 Japanese systems tend to be better designed, with
presaturators, vertical mist eliminators, and low-
sulfur oil reheat systems. Most systems in the
United States have not installed these items for
reasons of cost.
Oil-fired boilers tend to operate at slightly lower
excess air rates and produce less fly ash compared with
coal-fired boilers. However with high-efficiency fly ash
removal preceeding an FGD system, and design for the greater
excess air levels, lime systems should work equally well on
coal fuel.
*
Operability here equals an FGD plant's service time divided
by the scheduled operating hours of the process.
3-39
-------
Technology Advancements - Mist Eliminators - Because of the
strict particulate emission regulations, the Japanese
systems must collect all the mist after scrubbing. Much
more efficient mist elimination systems have been used; they
consist of vertical configurations with high gas velocities
6.1 m/s (20 fps). This high gas velocity usually requires
special eliminator vane design to prevent reentrainment.
Reheat - Japanese systems are usually installed with exter-
nal oil-fired reheat. This system is usually more expensive
to operate than in-line steam coils used in the United
72
States, but has proved much more reliable.
Scaling - Most Japanese units operate on low-sulfur fuel.
This may account for the fact that most of the systems do
not have major scale problems. When close to the stoichio-
metric amount (0.95 to 1.05) of lime is used, scaling is
reduced when compared with a stoichiometric amount of 1.1 to
1.3. Most Japanese systems use 0.95 to 1 of the stoichio-
metric amount of lime required. Another factor may be the
Japanese systems oxidize the recirculating slurry to sul-
fates and maintain smooth scrubber internals to reduce
buildup; these factors, however, have not been well-docu-
mented.
3.1.4 Engineering Design Parameters
Introduction - Sulfur dioxide removal efficiencies in
excess of 95 percent have been achieved using lime slurry
scrubbing on boilers in Japan. Although full-scale lime
scrubbing units in this country have not been designed for
the same high removal efficiencies (because the S0_ regula-
tions are not as strict as in Japan), test facilities in the
United States have demonstrated these same high removal
rates. These data are given in this section. Lime systems
3-40
-------
have successfully operated on both high- and low-sulfur coal
applications. Attainment of these high removal efficiencies
and successful operation is dependent on careful design and
operation. The importance of key design and operating
parameters is discussed in this section.
3.1.4.1 S02 Removal Potential of the System - Test facil-
ities in the United States have produced data that indicate
SO,., outlet concentrations from FGD scrubbing systems can be
as low as 1 to 5 ppm with removal efficiencies approaching
100 percent. Results of some of these tests are summarized
below:
0 At the Mohave plant of South California Edison, a
170-MW test facility, SO2 outlet concentrations
were less than 5 ppm with SC>2 inlet concentrations
of around 220 ppm. Net S02 removal efficiency was
about 98 percent. These tests were conducted over
a one-year period on low-sulfur coal.74
0 Recent tests at the Paddy's Run plant of Louis-
ville Gas and Electric have shown SC>2 removal
efficiencies in excess of 99 percent on 3-percent-
sulfur coal. This extremely high removal effi-
ciency was due to the addition of magnesium oxide
to the scrubbing slurry.75 Tests run for a 4-week
period in May and June 1976 resulted in 98-percent
SO2 removal.'6
0 The TVA Shawnee plant reported S02 removal effi-
ciencies of 95 to 99 percent for tests from
September 21, 1974, to November 11, 1974. The S02
inlet concentration ranged from 1800 to 3800 ppm
during this time.77
Full-scale facilities in Japan have shown similar high
performance, as documented in Section 3.1.3 of this report.
At these facilities, outlet concentrations as low as 35 ppm
with SO2 removal efficiencies in excess of 95 percent have
been attained.
3-41
-------
In summary,- current technology is capable of control-
ling SO- emissions to overall efficiencies in excess of 98
percent on both low- and high-sulfur coal applications, as
shown at the test facilities described above.
The ability of an SO- scrubbing system to remove high
levels of SO.-, is dependent on many design factors. The
operating conditions and results at the above plants vary
widely, but several key design parameters have proved im-
portant to high SO- removal efficiency. These key design
variables include:
0 L/G ratio - the mass transfer potential of the
system is directly related to the ratio of liquid
volume to gas volume. Sulfur dioxide removal
efficiency is greatly increased by increasing L/G.
0 pH - SC>2 removal efficiency is directly related to
slurry pH.
0 Magnesium ion concentration - recent tests of lime
slurry systems have shown increases in SC>2 removal
efficiency when magnesium is added to the slurry.
0 Gas velocity - the liquid-gas contact time is
directly related to the gas velocity through the
absorber and thus the SC>2 removal efficiency is
also directly effected by the gas velocity-
0 S02 inlet concentration - if all other parameters
remain constant, increased SC>2 loading in the
inlet gas will reduce the overall system effi-
ciency.
3.1.4.2 L/G Ratio - The L/G ratio is the flow of slurry
recirculating through the absorber in 1/s (gpm), divided by
the flow of flue gas in m3 (1000 ft3/min).
The theoretical effect of L/G on SC>2 removal efficiency
is shown in Figure 3-11 for a spray tower. This figure
shows SC>2 removal efficiencies predicted by Bechtel's
mathematical model for the Shawnee plant, based on actual
test data.
3-42
-------
LIQUID TO GAS RATIO, 1/rrT
100
10
90
80
D.
o:
o
10
CO
00
o
I/O
en
UJ
-------
The SO- inlet (2500 to 3000 ppm) is for a high-sulfur
eastern coal; gas velocity is 2.3 m/s (7.5 ft/s); Mg ion
concentration is 0 ppm and chloride content varied between
8000 and 13000 ppm. The increased S02 removal exhibited by
higher L/G is explained as follows:
0 The increased flow creates better conditions for
gas/liquid mass transfer at the interface because
of increased turbulence and better gas distribu-
tion, which thereby increases the mass transfer
driving force.
0 The increased liquid rate creates a larger mass
transfer area. This is multiplied by the greater
driving force.
Figure 3-12, illustrates the effects of L/G on SO-
removal in a TCA unit. In this example, both the predicted
and measured data indicate the strong dependence of SO-
removal on the L/G ratio. This test was run on high-sulfur
coal at the Shawnee plant. The magnesium concentration was
0 ppm. The scrubber inlet pH was 7.9 to 8.1; and total
height of spheres was 0.38 m (15.0 in.). The results are
expressed for three scrubber gas velocities: 3.8 m/s (12.5
ft/s); 3.2 m/s (10.4 ft/s); and 2.5 m/s (8.3 ft/s). In a
TCA, increased L/G does not significantly increase mass
transfer, since the mass contact area is determined largely
by the height of spheres. In this case, the increased S02
removal demonstrated must be attributed to the better
gas/liquid interface conditions.
Figure 3-13 shows the SO removal efficiency as a func-
tion of L/G in the horizontal test module at the 170-MW
Mohave facility. The L/G shown is per stage in a four-stage
system, so the total L/G is actually four times the indicated
vaiue. This facility fired low-sulfur coal (0.38 percent).
The extremely high S02 removal efficiencies shown (98%)
3-44
-------
LIQUID TO GAS RATIO, 1/nT
100
90
80
I 70
oo
CO
CM
o
60
O
a:
50
40
TCA GAS VELOCITY FOR
FACTORIAL TESTS
O 3.8 m/S (12.5 ft/sec)
A 3.2 m/S (10.4)
D 2.5 m/S (8.3)
10
20
D
SCRUBBER INLET pH =7.9-8.1
S09 INLET CONCENTRATION = 2,200 - 2,800 ppm
TOTAL HEIGHT OF SPHERES - 0.38 m (15.0 in)
EFFECTIVE LIQUOR Mg++CONCENTRATION = 0 ppm
I
30 40 50 60
LIQUID TO GAS RATIO, gal/mcf
70
80
Figure 3-12. Liquid-to-gas ratio and scrubber gas
velocity versus predicted and measured SO^ removal,
79
TCA with lime, Shawnee plant.
3-45
-------
L/G
100
95
90
85
80
CM
o
70
60
50
40
30
20
10
0
0
2
T
0.38% SULFUR COAL
10 15 20
L/G (gpm/1000 SCFM)
25
30
Figure 3-13. S0~ removal versus L/G ratio,
8 0
170-MW horizontal module, Mohave plant.
3-46
-------
yielded S09 outlet concentrations less than 10 ppm. Three
^ -3
curves are shown for 235, 216, and 165 m /s (500,000,
460,000 and 350,000 scfm) . As expected, SC>2 removal effi-
ciencies were lower at the higher gas flow rates because of
the shorter residence time. All three gas flow curves
illustrate increased SO,, removal efficiency at increased
L/G. Since this is a spray tower, the increased S0~ removal
efficiency can be attributed to both increased contact area
(which is directly proportional to liquid rate) and better
gas/liquid interface characteristics.
3.1.4.3 pH - Johnstone, in research in 1935, found that the
equilibrium vapor pressure of S02 over lime slurry was
inversely proportional to the slurry pH (Figure 3-14). This
phenomenon is caused by the higher alkalinity, which results
in lower SO,, equilibrium vapor pressure at higher pH.
The net effect of the pH on SO,, removal efficiency for
the TCA at the 10-MW Shawnee test facility is shown in
Figure 3-15. The trend predicted by the Bechtel Model is
indicated by the solid lines; the actual data points are
also shown. Although the pH effect is not as graphic as the
one shown primarily for L/G, the trend is obvious. Higher
pH results in higher SO- removal efficiencies. When used in
o
conjunction with a moderate L/G 8.04 1/m (60 gal/1000 ft3),
high pH can achieve 90-percent removal (top curve in Figure
3-15). There are, however, limitations. Higher pH requires
addition of excess lime. The excess lime, as well as raising
the pH, will travel through the scrubber system as excess
solids, and tend to increase scrubber scale formation. The
higher SO~ removal efficiency at higher pH is the result of
increased alkalinity of the circulating slurry. As shown
by the Johnstone graph, this tends to drive the SO,.,
into the circulating slurry. The same relationship of
3-47
-------
10
-2
5-10
-3
2-10
-3
10
S02 TOTAL
atm-kg H90
mol
2-10
-4
5-10
-5
4.0
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
4.5
5.0
pH
5.5
O
O
O
6.0
Figure 3-14. Effect of pH on SO2 vapor pressure over
81
buffered solutions.
3-48
-------
TOO
90-
80
C\J
° 70
CSL
LU
Cl_
60
50
40
1 I
LIQUID TO GAS RATIO FOR
FACTORIAL TESTS
3 (60 gal/1000 ft3)
(45 gal/1000 ft3)
D 4.0 1/nr (30 gal/1000 ftj)
O 8.0 1/m
A 6.0 1/m
SCRUBBER GAS VELOCITY =3.17 m/S(10.4 ft/sec'
INLET S09 CONCENTRATION = 2,200 - 2,800 ppm
TOTAL HEIGHT OF SPHERES =38.1 mm (15.0 in)
EFFECTIVE LIQUOR Mg++CONCENTRATION = 0 ppm
7 8
SCRUBBER INLET pH
10
Figure 3-15. Scrubber inlet pH and liquid-to-gas ratio
versus predicted and measured SO« removal, TCA
8 2
with lime, Shawnee plant.
3-49
-------
higher S09 removal efficiencies at higher pH is expected for
£t
all types of absorber modules. Figure 3-16 shows the
distribution of aqueous sulfite species as a function of pH.
The relationships are derived from the disassociation con-
stants of H SO It is apparent from these graphs that the
predominant sulfite species is a function of pH.
The predicted effect of pH on S02 removal efficiency
for the TVA Shawnee spray tower is shown in Figure 3-17.
The actual data do not agree closely with the predicted
values, but the trend is apparent, particularly for the L/G
of 6.8 1/m3 (51 gal/1000 ft3) curve, higher pH gives better
S0~ removal efficiencies.
3.1.4.4 Magnesium Ion Concentration - Addition of magnesium
ions to the recirculating slurry has demonstrated superior
S00 removal. Figure 3-18 shows the effect of magnesium ion
3
concentration on S02 removal efficiency at the 0.564-m /s
(1200-cfm) pilot plant run by Combustion Engineering Power
Systems for medium- and high-S02 inlet loadings.
Full-scale units have also reported increased S0?
removal with magnesium ion addition. For example, SO,.,
removal efficiency at the Paddy's Run facility was measured
with and without magnesium addition. Efficiency averaged 83
percent without magnesium; with magnesium, however, the
p cr
following results were obtained:
Date
7/6/77
7/8/77
Inlet S02
concentration, ppm
2150
2230
Outlet S02
concentration, ppm
1 to 5
1 to 5
Percentage
removal
99.7 to 99.9
99. 7 to 99.9
In addition, the Phillips Station conducted tests to demon-
strate the effect of increased magnesium ion concentration.
At this station the overall SO,., removal efficiency increased
3-50
-------
U)
I
U1
0.2
Figure 3-16. Distribution of aqueous sulfite species as a function of pH.
-------
TOO
90
80
CM
° 70
i—
•z.
UJ
60
50
40
LIQUID TO GAS RATIO
FOR FACTORIAL TESTS
O 9.1 1/m (68 gal/1000 ftj)
r\ f (™IT/ ^ f n \
A 6.81/m0 (51)
D 4.6 1/m3 (34)
T>
•\
\
-------
o
95
90
85
80
OJ
o
oo
75
70
I \
0 1000 2000 3000 4000 5000
MAGNESIUM CONCENTRATION, ppm
1 INLET S09 CONCENTRATION
2 INLET SO^ CONCENTRATION
2000 ppm
3000 ppm
Figure 3-18. Effect of magnesium on
n "J
S02 removal efficiency.
3-53
-------
from around 50-percent S02 removal to about 83-percent SO,,
removal when lime containing up to 10 percent magnesium ion
, 86
was used.
The reason for this phenomenon is that the magnesium
ionic species are much more soluble than the calcium ionic
species. Thus, the magnesium ions are in solution and ready
to react with the bisulfite and sulfite ionic species. This
increases the rate at which SO- can be removed from the
system. Figure 3-19 illustrates this increased alkalinity
as the magnesium ion concentration is increased.
3.1.4.5 Number of Contact Stages - When lime slurry con-
tacts S02~laden flue gases, S02 is removed until the solu-
tion approaches equilibrium. As this happens, the solution
has less potential to remove SO-, and the SO- removal rate
is therefore lower. One method used to maintain high S0_
removal is to contact the S02-laden flue gas with fresh
slurry in several stages. The effect of increased stages at
the Mohave plant is shown in Figure 3-20. This case is for
a horizontal cross-flow scrubber; additional stages are
added to the end of the absorber. This can also be accom-
plished in the TCA scrubber by additional trays, or in a
spray tower with additional spray modules. The figure shown
is for low-sulfur coal with L/G of 2.7 1/m3 (20 gal/1000
ft ) per stage. As a result, the total L/G is the number of
stages multiplied by 2.6 1/m3 (20 gal/1000 ft3).
3.1.4.6 Gas Velocity - The length of contact time between
the slurry and the flue gas will influence the amount of S02
removal. The longer the slurry is in contact with the flue
gas, the higher the overall S02 removal will be. This was
demonstrated in the test facilities at Mohave shown earlier
in Figure 3-13 for the horizontal module. The three flue
gas rates are directly proportional to velocity, and the
3-54
-------
0.20
0.10
0.05
to
£ 0.02
^-i
I—I
2
< 0.01
Q
UJ
>
O
0.005
0.002
0.001
I
TTTl I
Cl" = 0.10 MOLES/£
CaSO^ sat. = 1.0
Prn = 0.10 atm
LU2
T = 50°C
CaS04 sat. =0.3
_CaC03 sat
0.02 0.05 0.10 0.2 0.5 1.0
Mg CONCENTRATION (MOLES/K)
Figure 3-19. Dissolved alkalinity generated
88
by addition of MgO.
3-55
-------
100
80
60
CXJ
O
oo
O
.
UJ
Q-
40
20
CONDITIQNS
GAS FLOW =212 m3/S (450,000 SCFM)
L/G =2.7 1/ni (20 GPM/1000 SCFM)
INLET S0? = 220 ppm
_L
2 3
NUMBER OF STAGES
Figure 3-20. 170-MW horizontal S02 removal versus
number of stages, Mohave plant.
89
3-56
-------
data show that the higher gas flow rates yield a lower SO,,
removal efficiency.
Figure 3-21, for the spray tower at the 10-MW Shawnee
test facility, also shows the flue gas flow velocity inversely
related to the S02 removal efficiency at a given liquid
rate. The solid lines indicate the predicted values;
individual data points are also shown. For a packed or
turbulent type absorber, gas velocity is not so critical
since increased agitation and mixing partially offset the
decreased L/G ratio.
3.1.4.7 SO2 Inlet Concentration - Higher SO- inlet loadings
mean that less SO- is removed on a percentage basis at a
fixed operating condition. For a given set of operating
conditions, a lower SO- removal efficiency will be realized
if the S02 inlet concentration is increased. This is illus-
trated in Figure 3-22, which is again based on Shawnee test
results. These data show the theoretical and actual SO-
removal efficiencies for both spray tower and TCA absorber,
at an L/G of 11.4 1/m3 (85 gal/1000 ft3). Since all other
SO- removal parameters remain fixed in this example, the SO_
removal efficiency decreased as a function of inlet SO-. By
varying the operating conditions, for example by increasing
the L/G ratio, a desired efficiency can be maintained.
3.1.4.8 Overall Effects - All the variables discussed above
affect the system. To determine the overall effect, each
would have to be analyzed in conjunction with the remaining
factors. Bechtel has developed a model to define these
effects for the 10-MW test facility at Shawnee. This model
is useful for selecting the parameter needed to achieve high
94
SO- collection efficiencies.
3-57
-------
100
90
80
OL
CM
O
00
70
60
50
40
D
SLURRY FLOW RAVE FOR
FACTORIAL TESTS I
O 6.2 1/sm ?(30 gal/min
A 4.66 1/sni (22.5)
D 3.11 1/snT (15)
ft2)
CD
SCRUBBER INLET pH = 7.9 - 8.1
S09 INLET CONCENTRATION = 2,500 - 3,500 ppm
EFFECTIVE LIQUOR Mg CONCENTRATION = 0 ppm
LIQUOR CI'CONCENTRATION = 8,000 - 13,000 ppm
I
789
SPRAY TOWER GAS VELOCITY, ft/s
10
11
Figure 3-21. Gas velocity and slurry flow rate
versus predicted and measured S02 removal,
spray tower with lime, Shawnee plant.
90
3-58
-------
TOO
>-
z
LU
Q.
OIL
O
CM
o
on
O
o:
95
90
85
80
75
70
O
i 1 I 1
O EPA PILOT TCA
SPHERE HEIGHT = 178 mm (7 in. I/BED
LIQUID TO GAS RATIO =11.4 1/nT (85 gal/1000
TCA GAS VELOCITY = 2.29 m/s (7.5 ft/sec)
OTVA PILOT SPRAY TOWER
-I
LIQUID TO GAS RATIO = 11.41/m3 (85 gal/1000 ft3)
O
O
O
O
1,000
2,000 3,000
INLET S02 CONC. ppm
4,000
5,000
Figure 3-22. Effect of inlet S02 concentration on S02
91
absorption efficiency for fixed design condition.
3-59
-------
3.1.5 Operability
The Phillips lime scrubber, the first in the United
States, has had poor operability, but technology has bene-
fited from the experience gained from this and other units.
The Elrama lime scrubber, which is similar in design to the
Phillips system, was put in service approximately two years
later. Although only a percent of the boiler flue gas is
currently being treated, the operability of this five-module
92
plant has been good, with only limited forced outage time.
Of the new generation, for instance, the Green River facil-
ity of Kentucky Utilities, which started up in January 1976,
quickly rose to 100-percent operability and ran very well
through February 1977, when it was shut down for stack
repair. It started up again in April 1977 and once again
has good operability (Figure 3-23). This information is
detailed in Section 3.1.2.2.
The Bruce Mansfield No. 1 unit of Pennsylvania Power
(see Section 3.1.2.2) showed good operability during its
initial commercial operating period (98 percent availability
from June 1 to December 31, 1976). During the subsequent
months, a number of major operating problems arose, and
power production was reduced pending repairs and modifica-
tions. Although the major FGD-related problems have been or
are being solved, widespread stack liner failures have
caused half-load operations, which will probably continue
until early 1978. The availability of the half of the plant
still in service has been very good (95 to 99 percent during
June and July).
The Cane Run unit of Louisville Gas and Electric,
as shown in Figure 3-24, initially showed good start-up
operability, but was shut down in 1977 when river freez-
ing prevented lime delivery. During initial operation,
3-60
-------
U>
I
01
SHUT DOWN FOR
STACK REPAIR
AN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC JAN FEB MAR APR MAY
1976
MOUTHS
1977
Figure 3-23. Scrubber system operability Green River No. 1, 2, and 3.
95
-------
o
o;
100
90'
80
70
j 60
I
' 50
40
30
20
10
SCRUBBER SHUT DOWN1
BECAUSE WEATHER
CONDITIONS CUT OFF
LIME SUPPLY
,UG SEP
AFTER RESTART
SHUTDOWN FOR
MODIFICATIONS
(APR-JUL)
_L
OCT
1976
NOV DEC JAN FEB
MONTHS
MAR APR MAY JUN
1977
Figure 3-24. Scrubber operability,
Qfi
Cane Run No. 4.
3-62
-------
several design errors were detected and the unit was shut
down early in 1977 for adjustments; the mist eliminators
were replaced with Chevron-type units; L/G was increased; an
oil-fired reheater was added; and the stack lining replaced.
The Paddy's Run unit of Louisville Gas and Electric is
an older unit and the boiler is used for peak loads only.
As a result, the scrubber runs infrequently, but it does
have good operability during its working periods (Figure
3-25) .
The Colstrip No. 1 and No. 2 units of Montana Power
have been operating since September 1975 and May 1976
respectively. Both are equipped with SO2 and particulate
scrubbing systems. The alkalinity of the fly ash and
supplemental lime are the principal reagents. Systems
availability for both units has been good and has steadily
improved. For Unit No. 1, the annual average total avail-
ability values for 1976 and 1977 (August) were 86 and 89
percent respectively. Unit No. 2 values were 96 and 85
percent for 1976 (May startup) and 1977 (August) respec-
tively.
3.1.5.1 Effect of Increased SO^ Removal on Operability
Numerous corrosion and erosion problems have attended the
operation of lime scrubbing systems, as indicated in Section
3.1.2. As design techniques become more refined, however,
corrosion problems are being solved with improved construc-
tion materials. For example, the recirculation pumps at the
Phillips station were designed with Carpenter 20-type wetted
parts. It was soon apparent that this material is not
suited to lime slurry scrubbing. Subsequent installations
have incorporated rubber-lined or hardened iron parts and
98
have experienced few difficulties. Corrosion problems do
continue in some components. Linings on vessels and stacks,
3-63
-------
U)
I
90
80
70
(SYSTEM
*-1 DOWN
40
30
20
10
ISYSTEM
DOWN
SYSTEM
DOWN
ISYSTEM
DOWN
JAN FEB MAR APR MAY
JUN JUL
1976
AUG SEP OCT NOV DEC JAN FEB
MONTHS
MAR APR
1977
MAY
Figure 3—25. Scrubber operabllity — Paddy's Run No. 6.
97
-------
and in-line reheat coils have failed repeatedly. In some
cases, however, unlined carbon steel reheaters have per-
formed well. Although the higher reactivity of lime makes
it difficult for localized volumes of acidic solution to
accumulate, some degree of corrosive attach has occurred in
dampers, turning vanes, and expansion joints. The solution
to the problem is the replacement of corroded parts with
materials capable of withstanding acidic corrosion, such as
316 L SS or Monel.
Stacks associated with lime scrubbers have shown a high
degree of corrosion (at Cane Run and Green River for example)
apparently because of the lack of reheat in these lime
scrubber systems, or because of improper lining and mist
elimination. Operators of these systems are planning to
99
incorporate reheat. Increased SO- removal should not
amplify the corrosion problem; in fact, it should reduce
the problem of producing less corrosive flue gas.
Scaling and plugging problems are still of major
consequence in this country. Approaches to solving these
problems include regular shutdown and cleaning of each
absorber, addition of magnesium ion to the slurry, and
better pH and solids control. The magnesium ion dissolved
in the slurry increased alkalinity and the rate of chemical
reactions. Both these techniques show some merit, although
neither one alone provides the total answer to scaling and
plugging problems.
At present, the operational systems generally use 1.1
to 1.3 times the amount of lime theoretically required. If
continued, this strategy may increase chemical and mechanical
problems caused directly by plugging and scaling, which in
turn are due to the heavier load of solids in the system.
3-65
-------
New technology that might over-come these problems is
described in Section 3.1.2.4.
Scale and plugging also occur in the mist eliminator
section of the system. Washing techniques have been devel-
oped that appear to ease the problem.
3.1.6 Vendor Guarantees
For systems currently going on-line, the vendors are
guaranteeing higher SO- removal efficiencies. For example,
the Conesville Station has an 87.6 percent vendor guarantee;
for the Cane Run Station it is 85 percent; and at Bruce
Mansfield 92 percent. A brief vendor survey (see Section
4.2) was conducted for this review; all vendors surveyed
would guarantee 90-percent or better SO- removal efficiency
on their systems. In addition, vendors are now guaranteeing
operability periods. The Green River and Conesville Gener-
ating Stations for example, had six-month operability
guarantees. This means that the vendor will guarantee a
certain operability, usually 90 percent for a given length
of time.
3-66
-------
REFERENCES FOR SECTION 3.1
1. Kunesh, Charles J. The Calcination and Slaking of
Quicklime. 37th Annual Meeting, International Water
Conference, Pittsburgh, Pennsylvania, October 1975. p.
2.
2. UOP Air Correction Division. Air Pollution Control
Systems Qualifications.
3. PEDCo Environmental, Inc. Evaluation of the Flue Gas
Desulfurization System at the Phillips Power Station of
Duquesne Light Co. U.S. Environmental Protection
Agency, October 1976. p. 5-29.
4. U.O.P. Conesville Material Balance. Drawing No. H-60-
0102, 3/6/75.
5. Ibid.
6. Radian Corporation. Evaluation of Lime/Limestone
Sludge Disposal Options. U.S. Environmental Protection
Agency, 1973.
7. Slack, A.V., and G.A. Hollinden. Sulfur Dioxide
Removal From Waste Gases Noyes Data Corporation, 1975.
p. 62.
8. PEDCo Environmental. Summary Report - Flue Gas Desul-
furization Systems. U.S. Environmental Protection
Agency, June-July 1977.
9- Op. cit. No. 8.
10. Op. cit. No. 8, p. 132.
11. Op. cit. No. 8, p. 132.
12. Berst, A.H., and J. Reisinger. Start-up of American
Air Filter's Sulfur Dioxide Removal System at the
Kentucky Utilities Company's Green River Station.
Proceedings: Symposium on Flue Gas Desulfurization,
New Orleans. March 1976.
3-67
-------
13. Op. cit. No. 8, p. 132-133.
14. PEDCo Environmental. Trip Notes, March 10, 1977; March
3, 1976; June 30, 1976.
15. Op. cit. No. 12, Figure 1.
16. Op. cit. No. 8, p. 131.
17. Op. cit. No. 14.
18. Op. cit. No. 14.
19. Op. cit. No. 14.
20. Op. cit. No. 14.
21. Op. cit. No. 12.
22. Op. cit. No. 8.
23. Op. cit. No. 12.
24. Op. cit. No. 14.
25. Beard, J.B. Letter to Mr. D.R. Goodwin, July 21, 1977.
26. PEDCo Environmental. Summary Report - Flue Gas De-
sulfurization Systems. U.S. Environmental Protection
Agency, August-September 1977. p. 141-146.
27. Op. cit. No. 8, p. 135-138.
28. Op. cit. No. 14.
29. Op. cit. No. 14.
30. EPRI, Trip Report from November 23, 1976. p. 5-7.
31. Op. cit. No. 8, p. 135-138.
32. Op. cit. No. 8, p. 190.
33. Op. cit. No. 8, p. 190.
34. Op. cit. No. 8, p. B-22.
35. Op. cit. No. 8, p. 190.
3-68
-------
36. Op. cit. No. 8, p. 139.
37. Op. cit. No. 8, p. 190.
38. Richards, W.E., and C.G. Winters. Designing Large
Central Stations to Meet Environmental Standards.
Generation Planbook, 1976. p. 27-34.
39. Op. cit. No. 8, p. 189-91.
40. Op. cit. No. 8, p. 194-195.
41. Op. cit. No. 8, p. 199.
42. Durkee, K.R. Survey of Pennsylvania Power's Bruce
Mansfield Power Generating and Flue Gas Desulfurization
System. July 14, 1977. p. 2-6.
43. Op. cit. No. 8, p. 195.
44. Op. cit. No. 42, p. 4.
45. Op. cit. No. 8, p. 190.
46. Op. cit. No. 42, p. 4-5.
47. Op. cit. No. 42, p. 2.
48. Op. cit. No. 42, p. 5.
49. Op. cit. No. 42, p. 3.
50. Op. cit. No. 8, p. 189-91.
51. Weir, A. J., J.M. Johnson, D.G. Jones, C.T. Spencer.
the Horizontal Cross Flow Scrubber. Proceedings: Flue
Gas Desulfurization Symposium, Atlanta, 1974. U.S.
Environmental Protection Agency. p. 360.
52. Weir, A.J., L.T. Papay, D.G. Jones, J.M. Johnson, W.C.
Martin. Results of the 170 MW-Test Modules Program
Mohave Generating Station, Southern California Edison
Co. Presented at: Symposium on Flue Gas Desulfurization,
New Orleans, March 1976. p. 2.
53. Op. cit. No. 8, p. 328, 329.
3-69
-------
54. Op. cit. No. 8, p. B-38.
55. .Op. cit. No. 52, p. 1-8.
56. Op. cit. No. 52, p. 9.
57. Op. cit. No. 52, p. 10.
58. Op. cit. No. 52, p. 6.
59. Op. cit. No. 52, p. 16.
60. Op. cit. No. 52, p. 23.
61. PEDCo in-house files.
62. Op. cit. No. 52, p. 20.
63. Op. cit. No. 52, p. 18.
64. PEDCo Environmental. Ando, J. S02 Abatement for
Stationary Sources in Japan. U.S. Environmental
Protection Agency, 1977. p. 3-2.
65. Ibid, p. 3-4.
66. Op. cit. No. 52, p. 21.
67. Op. cit. No. 52, p. 19.
68. Op. cit. No. 64, p. 3-13.
69. Op. cit. No. 64, p. 3-11.
70. Slack, A.V- Private communication.
71. Slack, A.V. Technology for Power Plant Emission
Control. Survey of Developments in Japan, May-June
1977. p. 25.
72. Ando, J. S02 Abatement for Stationary Sources in
Japan, 1976.
73. Op. cit. No. 64, p. 3-11 to 3-14.
74. Op. cit. No. 52, p. 12.
3-70
-------
75. VanNess, R. Private communication, July 15, 1977.
76. Op. cit. No. 8, p. 147.
77. Head, H. EPA Alkali Scrubbing Test Facility Advanced
Program, Second Progress Report. U.S. Environmental
Protection Agency, September 1976. p. E-4.
78. Bechtel Corporation, EPA Alkali Scrubbing Test Facil-
ity, TVA Shawnee Power Plant, Paducah, Ky. Monthly
Progress Report for Period December 1, 1976 to December
31, 1976. P. 2-7.
79. Ibid, p. 2-19.
80. Op. cit. No. 51, p. 360.
81. Rochelle, G.T. Process Symthesis and Innovation in
Flue Gas Desulfurization, PhD Thesis, University of
California, Berkley, 1976. p. 382.
82. Op. cit. No. 78, p. 2-20.
83. Draemel, D. Part 1. A View of the Process Chemistry
of Identifiable and Attractive Schemes. U.S. Environ-
mental Protection Agency, May 1973. p. 6.
84. Op. cit. No. 78, p. 2-8.
85. Op. cit. No. 75.
86. Op. cit. No. 3, p. 4-13.
87. Frabotta, D. and P.C. Rader. Lime/Limestone Air
Quality Systems: Effect of Magnesium on System Per-
formance. ASME, December 1976. p. 4.
88. Op. cit. No. 81, p. 101.
89. Op. cit. No. 52, p. 11.
90. Op. cit. No. 78, p. 2-6.
91. Op. cit. No. 78, p. 4-4.
3-71
-------
92. Nelson, R.L. and R.A. O'Hara. Operating Experiences at
the Phillips and Elrama Flue Gas Desulfurization
Facilities. Proceedings from the Second Pacific
Chemical Engineering Congress, AIChE, Denver, Colorado.
August 1977. p. 310-311.
93. Berube, D. and C. Grimm. Status of the Performance of
the Montana Power Company's Flue Gas Desulfurization
System. Proceedings from the Fourth Symposium on Flue
Gas Desulfurization, Hollywood, Florida, November
8-11, 1977.
94. Bechtel Corporation. EPA Alkali Scrubbing Test Facil-
ity. TVA Shawnee Power Plant, Paducah, Kentucky. U.S.
Environmental Protection Agecny, January 1977. Monthly
Report. p. 3-2.
95. Op. cit. No. 8, p. 134-138.
96. Op. cit. No. 8, p. 141-143.
97. Op. cit. No. 8, p. 146-149.
98. Op. cit. No. 92.
99. PEDCo in-house files.
100. Op. cit. No. 8, p. 56.
101. Op. cit. No. 8, p. 139.
102. Op. cit. No. 8, p. 189.
3-72
-------
3.2 LIMESTONE SLURRY FLUE GAS DESULFURIZATION SYSTEMS
The limestone slurry scrubbing process is similar to
the lime slurry process described in Section 3.1. Since the
basic chemistry and SO,., removal mechanisms are similar to
the lime-based systems, technical details that were explained
in Section 3.1 will not be repeated here. Instead, this
section will highlight any differences in the two processes
and discuss the performance results of existing limestone
slurry FGD systems.
3.2.1 Process Description
Development of this process began much earlier than the
lime scrubbing systems. More than 40 years ago, calcium
carbonate (CaCO-.) was commercially used to remove SO,, from
waste gases. The process was applied to flue gases from
power plants in England. Four power plants were equipped
with scrubbers before 1940. Although these units were
inefficient in comparison with those being built today, two
were still in use in 1970.
Except for small-scale studies by TVA in the 1950's,
little interest was reported in limestone FGD until after
1960, when research began almost simultaneously in Japan,
the Soviet Union, and the United States. Work in the
Soviet Union was directed primarily toward abatement of SO,,
pollution from the metallurgical industry; several systems
are now in operation. The Japanese now have 12 major
(larger than 100 MW) limestone-based FGD systems on utility
2
and industrial boilers.
In the United States, the Universal Oil Products
Company tested a mobile-bed limestone scrubber on a power
plant in Wisconsin in 1965. The next year, Combustion
Engineering followed with a pilot plant on a Detroit Edison
power station. An acceleration of effort began shortly
3-73
-------
before 1970. Zurn Industries, in cooperation with the
National Air Pollution Control Administration (now EPA),
constructed a pilot limestone FGD system that was tested on
an oil-fired generating station at Key West, Florida, and
later on the coal-fired Shawnee station of the TVA at
Paducah, Kentucky. Also under EPA contract Bechtel built
and operates a large pilot plant at the Shawnee station and
conducts tests of both lime and limestone scrubbing.
During the EPA program, TVA modified one of the scrubbers
for a test series with other types of absorber internals.
Babcock and Wilcox performed bench-scale and pilot-plant
testing of limestone FGD at their Alliance, Ohio, research
and development center. Southern California Edison, in a
joint venture with five other western Power utilities,
tested limestone as a scrubbing reagent at the Mohave
generating station in Nevada. Equipment was later moved to
Four Corners, New Mexico, for additional testing. Combus-
tion Engineering, in cooperation with Northern States Power
Company, tested a prototype unit at the utility's Black Dog
station. Detroit Edison and Peabody Engineered Systems
conducted a pilot-plant program at the St. Clair power plant
from 1971 to 1973. A full-scale unit was built at St. Clair
in 1974 and was tested until 1976.3
A significant amount of theoretical work was conducted
in support of the plant testing and equipment development.
Most of the published findings are from studies funded under
EPA contracts.
The study of calcium- and magnesium-based FGD systems
has been one of the most intensive and broadly-based in-
dustrial research programs ever conducted in this country.
A massive body of data has been accumulated in public
3-74
-------
record, and still more is held as proprietary information by
various companies.
3.2.1.1 Process Flow/Material Balance - The flow diagram
for a model plant is shown in Figure 3-26. The assumptions
used to make this balance are presented in Table 3-10. This
flow diagram is identical to the one presented for lime
slurry scrubbing in Section 3.1.1, except for the limestone
slurry preparation equipment.
Table 3-10. ASSUMPTIONS USED IN THE MATERIAL BALANCE
CALCULATIONS FOR A 500-MW BOILER
SO removal in absorber
Coal sulfur content
Coal heat value
Limestone CaCO_ content
Stoichiometry
Sludge oxidation (CaSO_/CaSO.)
Fly ash/bottom ash
P articulate removal in ESP
Excess air to air heater
Excess air to boiler
L/G
90 percent
3. 5 percent
28 kJ/g
(12,000 Btu/lb)
90 percent
1.3
90/10
80/20
99 percent
40 percent
10 percent
5.36 1/m3
(40 gal/1000 scf)
.2 Process Chemistry - Limestone Slurry Preparation
The first step in the limestone scrubbing process, limestone
slurry preparation, is the only step that is significantly
different from the lime scrubbing process. In this step,
the limestone is taken from lump size [2.5 cm to 7.6 cm (1
in. to 3 in.)], pulverized to -200 to -300 mesh in a ball
4
mill and slurried with water. It is necessary to grind to
3-75
-------
to
r
C".
MAKE-UP
WATER-2
JREHEATER \—
fj
REHEAT
GAS FROM BOILER
13
STACK
EMISSION RATE - 0.525 lb/10 Btu S02
0.103 lb/106 Btu PARTICULATE
I
LIMESTONE
STORAGE
14
STORAGE
S.R.=1.3, 90S LIMESTONE
WATER
FEED BALLMILL & SLURRY
TANK PREPARATION
rn;8L-r 1 pt ion -
Ut». 10' Ib/hr
SCFK 103
T«»p.. *F
Fly *ah,
101 Ib/hr
SO,.
10' Ib/hr
MjO.
10 3 Ib/hr
»2.
101 Ib/hr
CO,,
103 Ib/hr
»2,
10J Ib/hr
1
CdU to
5570
1121
320
0.5
25.5
20.1
2970
1030
343
2
Flue
275
549
0.5
1.00
125
0.91
0.14
273.6
9
370.7
740
O.lt
1.00
70
0.91
0.14
369.}
10
148
231
35
1.2«
70
0
0
96.2
11
32037
56175
15
1.1
125
3134
46t
J7231
12
Sp«nt slurrv
to th«
434.5
719
15
1.1
125
41. t
4.2
3t9.3
1)
Sp«nt
•lurry
159.5
240
40.0
1.325
125
41. t
• .2
»5.7
Description -
•olid •tr«*M
LUt«, 10* Ib/hr
14
Li Ma ton*
5X.I
IS
Ply uh
45.4
Figure 3-26. Flow diagram/material balance - 500-MW limestone scrubbing system.
-------
this mesh, because limestone slurry is less reactive than
lime slurry under similar circumstance. The limestone
slurry reaction requires the dissolution of CaCO_; this is a
slow reaction. The rate of reaction is directly related to
the particle surface area and therefore to the particle
size .
SO-, Reactions
^ .
The reactions are the same as the reactions detailed in
the lime slurry scrubbing section, except the reactant
species is CaCO^, not Ca(OH) . The reactions are outlined
5
below.
CaC03 (s) - CaC03 (ag) ~ Ca++ + CO3"
co3 + co2 +
S°2 (g) ~ S°2
- +
HS03 «-•> H +
++ —
Ca + S03 +
+
H + HC03 ^+
H2° ""*
(aq)
„ _
S03
O T-T /"%
2 H20
H2CO3
2HC03
+
+ H20 ^^ H
•*->- CaS03-2H
(aq) ~ C°2
_,
+ HS03
i°/ \
2 (s)
, s + H00
(g) 2
-2H20 + 1/2 02 -«--»- CaSO4"2H O, .
HS03~ ++ H+ + S03
Ca++ + S0~~ + 1/2 HO •«--> CaSO -1/2 HO,
J ^ -j ^ ( S
H+ + -
«2H20 + 1/2 02 •*--> CaSO4-2H 0
3.2.2 Domestic Units
The first large domestic limestone scrubbing units
injected limestone into the boiler and often used a tail-end
scrubber to complete S0? removal. The first system was the
Meremac No. 2 unit of Union Electric Corporation, which
started up in 1968; the second was the Lawrence No. 4 unit
3-77
-------
of Kansas Power & Light, which started up in 1971. The
Meremac unit was eventually shut down and the scrubber at
Lawrence was converted to a conventional wet scrubbing
system.
Recently, all major limestone units have employed tail-
end scrubbing. A list of these domestic units appears in
Table 3-11. The major facilities are described in Appendix
B. It is important to note that each of these limestone
scrubbing facilities that has undergone performance testing
has met or exceeded design specifications. Design specifi-
cations usually call for 60- to 80-percent S02 removal
efficiency, which is sufficient to meet local codes. Three
units, La Cygne No. 1 and Sherburne No. 1 and No. 2, are
described in detail in this section.
3.2.2.1 La Cygne No. 1 - Kansas City Power & Light Co. -
The LaCygne Power Station is about 88 km (55 miles) south of
Kansas City, in Linn County, Kansas. The electric power
generating facilities of La Cygne No. 1 consist of a coal-
fired, base-loaded boiler with a capacity of 2,800,000 kg
steam/hr (6,200,000 Ib steam/hr) with an associated 820-MW
(net) steam turbine and electric generator. The plant also
has three oil-fired boilers, used primarily for start-up of
the large unit. The power generating facilities went into
service May 31, 1973.
The boiler at La Cygne No. 1 was designed by B&W and is
a wet-bottom, cyclone-fired unit. The pollution control
equipment consists of eight scrubbing modules, also built by
B&W as an integral part of the power generating facility.
It is not possible to bypass the boiler's flue gas around
the FGD system. The La Cygne Power Station uses about 24 MW
from its gross generating capacity of 870 MW to operate the
the emission control system. Another 30 MW are required to
operate auxiliary station equipment.
3-78
-------
Table 3-11. MAJOR DOMESTIC FGD INSTALLATIONS - LIMESTONE SLURRY'
OJ
I
Station/Unit
Power company
Cholla No. 1
Arizona Public Service
Duck Creek No. 1A
Central Illinois Light
La Cygne No. 1
Kansas City Power & Light
Lawrence No. 4
Kansas Power & Light
Lawrence No . 5
Kansas Power & Light
Martin Lake No. 1
Texas Utilities
Sherburne No. 1
Northern States Power Co.
Sherburne No. 2
Northern States Power Co.
Southwest No. 1
Springfield City Utilities
Widows Creek No. 8
Tennessee Valley Authority
Will County No. 1
Commonwealth Edison
Winyah No. 2
South Carolina Public Service
Size
(MW)
115
400
820
125
400
793
710
680
200
550
167
280
Start-up
date
10/73
8/78b
2/73
12/68
11/71
10/77
3/76
9/77
4/77
5/77
2/72
7/77
New or
retrofit
Retrofit
New
New
Retrofit
New
New
New
New
New
Retrofit
Retrofit
New
Coal
percentage
sulfur
0.4-1
2.5-3
5.0
0.5
0.5
1.0
0.8
0.8
3.5
3.7
4.0
1.0
Design
SO 2 removal
efficiency, %
92a
75
76
76
65
60
50
50
80
80
82
70
a The A-side train includes the packed absorber with limestones slurry circulated through the module.
S02 removal for the A-side is 92 percent. The B-side train does not include packing or
limestone slurry and removes approximately 25 percent of the S02- Total system removal
efficiency is 58.5 percent.
One module operated from September 1976 to April 1977.
-------
The coal ranges in gross heating value between 19,000
and 23,800 kJ/kg (8200 and 10,200 Btu/lb). Ash and sulfur
range from 20 to 30 percent and 5 to 6 percent respectively.
Table 3-12 presents pertinent data on plant design and
operation.
Table 3-12. POWER PLANT AND FGD SYSTEM DESIGN/
9 10
OPERATING DATA, LA CYGNE NO. 1 '
8
Maximum generating capacity
Boiler manufacturer
Year placed in service
Maximum coal consumption
Maximum heat input
Unit heat rate
Stack height above grade
Flue gas rate - maximum
Flue gas temperature
Particulate
Removal efficiency (actual)
SO2
Removal efficiency (actual)
No. of FGD modules
Process vendor
820 MW (net)
B&W
1973
366 metric ton/hr
(404 ton/hr)
8,105 106 kJ/hr
(7,676 106 Btu/hr)
9880 kJ/kWh
(9,360 Btu/kWh)
213 m (700 ft)
1,297 m3/s
(2,760,000 acfm)
141°C (285°F)
97 to 99 percent
70 to 83 percent
8
B&W
Flue Gas Scrubbing System
The FGD system consists of eight identical scrubbing
modules (Figure 3-27) with a venturi scrubber for particulate
emission control and an absorber tower for SO,, emission
control. Each module treats about one-eighth of the total
flue gas from the coal-fired boiler. As the hot flue gas
3-80
-------
FLUE GAS FROM
AIR HEATER STEAM
394,300 acfm
AT 285°F
HOT AIR FROM (MOT USED IN
AIR HEATER "D" MODULE)
DEMISTER
gpm (INTERMITTENT)
HYDROCLONE
800 gpm
Figure 3-27. Flow diagram of one of the eight
FGD modules - La Cygne No. 1.
3-81
-------
enters the venturi, it is subjected to jets of limestone
slurry injected through nozzles on the walls of the vessel.
The liquid-gas stream flows downward through the venturi
throat restriction, where the gas contacts the atomized
liquid droplets. The scrubbing efficiency is regulated by
adjusting the venturi throat gap. As the gas leaves the
venturi and enters the disengagement chamber, its velocity
decreases from about 40 m/s (130 ft/s) (at the throat) to
about 4.6 m/s (15 ft/s). This reduction in velocity separates
the limestone slurry droplets from the quenched gas. The
slurry drains into the recirculation tank. Gas enters the
SO- absorption tower at the base and moves upward through
two sieve trays in series. As the gas passes through the
34.9-mm-diameter (1-3/8-in.) holes of the sieve trays, it
contacts a shower of limestone slurry, which is sprayed into
the path of the rising gas. The scrubbed gas then passes
through a third sieve tray, which collects slurry carryover
and reduces the load on the demister. The gas then passes
through a chevron demister 25-cm (10-in.) high, where the
remaining fine droplets are collected. The flue gas is then
heated before it enters a plenum common to all modules and
is discharged to the stack through induced-draft fans.
The venturi and the absorber tower of each module share
a common limestone slurry recirculation tank, in which the
pH is maintained between 5.5 and 6.0. The pH is monitored
by means of a cell located in the slurry feed to the venturi
nozzles.
A hydrocyclone has been installed in the slurry circuit
of each module to remove large particles, to prevent plugging
of nozzles and strainers, and to reduce erosion in pumps,
pipes, and nozzles.
3-82
-------
The liquid level in the recirculation tank is main-
tained by pumping excess liquor to the sludge disposal pond.
This plant does not use facilities for sludge treatment or
fixation.
Limestone is ground on site. A 54,000-metric ton
(60,000-ton) supply of limestone rocks is kept near the coal
storage area. The limestone is transported intermittently
to the mill by the coal conveyor system. Two wet ball
mills, each rated at 98 metric tons (108 ton)/hr, are housed
in a building that also contains two limestone holding
tanks.
Current Design Parameters
The FGD installation was designed with venturi and
turbulent contact absorber (TCA) towers for fly ash and SO
removal. Many major modifications in the system have since
altered some of the original design parameters; the following
description represents current operating conditions.
The system was originally designed with seven modules
and space was provided for an eighth if needed. Shortly
after start-up, it was decided to install the eighth module
to allow higher boiler load, and to improve performance and
availability.
The venturi scrubber operation was originally designed
for an L/G ratio of 2.4 1/m3 (18 gal/1000 acf) of gas.
Since installation of the hydroclones, however, the recir-
culation pump head has increased so that the L/G is only
about 1.6 1/m3 (12 gal/1000 acf) of gas. The SO absorber
3
tower is designed for an L/G of about 3.5 1/m (26.5 gal/1000
acf) .
Gas velocity through the venturi throat is about 150
ft/s with the throat wide open. Gas velocity through the
demister section of the tower is 2.6 m/s (8.4 ft/s).
3-83
-------
The tower demisters are washed by underspray and over-
spray manifolds using pond water. Each demister is washed
continuously with 8.5 1/s (140 gpm) of underspray water.
The overspray operates intermittently at 130 1/s (2100 gpm)
for 1 minute during each 8-hour period.
The reheater tube bundles were originally made of 304
stainless steel. The original design reheater exit tempera-
ture was 64°C (147°F). Supplemental direct heating with hot
air injection had been practiced on six of the original
seven modules for gas temperature elevation purposes prior
to passage through the reheaters. Additional reheat capa-
city has been installed in each module to eliminate the need
for hot gas bypass.
System Performance
The FGD installation on the La Cygne Boiler No. 1 was
plagued with numerous start-up problems. Some of these
problems, such as vibrations of the induced-draft fans and
their sensitivity to imbalance, appeared even before the
boiler was fired and were not due to FGD system operation.
An account of the major problems follows.
Scrubbing Modules The process piping network is all
rubber-lined to protect the carbon steel base from the
abrasive slurry. In general, the rubber liners have per-
formed well, with only a few reported incidents of wear in
the area of the spent slurry valves. This wear was attrib-
uted to the throttling action of the valve to modulate the
flow of slurry. The problem was solved by operating the
valve only in a completely open or completely closed posi-
tion .
The venturi pumps on the circulation tanks are also
rubber lined; this lining has been damaged many times,
primarily because of plugging of the strainer at the suction
end of the pump. As the flow ceases or is drastically
3-84
-------
reduced because of plugging, the pump cavitates and the
liner is sucked into the path of the impeller and shredded.
Since the suction strainers plugged frequently, and they
were located inside the recirculation tank, the tank had
to be drained so the strainers could be cleaned. In order
to extend the life of the limestone slurry spray nozzles and
to reduce wear and erosion in the slurry recirculation loop,
a hydroclone was installed in the recirculation line of
each module. Thus it was possible to remove the strainer
and correct the pump problems.
In the past, the demister trays and the gas reheater
tubes have plugged severely- The demister trays consist of
Z-shaped fiberglass boards. Droplets of slurry are carried
over with the flue gas and are deposited on the trays. As
the slurry builds up, the gas flow is restricted and its
velocity through the demister increases. This leads to
solids carry-over and deposition on the reheater tubes. The
slurry carry-over also reaches the induced-draft fan and is
deposited on its blades.
These interrelated problems of carry-over to the de-
mister, the reheater, and the induced-draft fan have ne-
cessitated many modifications and corrections of operating
procedures. Currently, intermittent heavy overspray and
continuous underspray have kept the demisters relatively
clean. Steam soot blower modifications have been reasonably
successful in the proper maintenance of reheaters, and fan
washing has been greatly reduced.
The original reheater tubes (304 SS) began to fail
prematurely because of attack by acid condensate. They are
being replaced with 316L SS bundles.
3-85
-------
Induced-Draft Fans One problem was severe vibration in
the fan housing. The fans were initially sensitive to
imbalance, and it was found that the operating speed was
close to the critical speed. The vibrations caused cracks
in the inlet cones, requiring additional stiffeners to
strengthen the housing. The thick center plate was scal-
loped to hold down the weight, and the critical frequency
was moved farther away from the operating speed to reduce
the tendency to vibrate. The leading edge of each blade was
covered with a stainless steel clip to prevent erosion.
Particulate carry-over still requires that the fans be
washed on an intermittent basis, but the cleaning frequency
is being steadily reduced.
The high running temperature of the thrust collars on
the fan bearings presented another problem. Temperatures
reached the 82°C (180°F) alarm set point within a short time
after the fans were started. Temperature increases were
controlled by cutting oil grooves in the thrust collar and
installing forced-lubrication systems on all the fans.
These modifications caused the thrust collar temperature to
drop into a range from 60°C (140°F) to 71°C (160°F).
Performance History Despite the problems at start-up,
the availability of the system improved steadily. This
system is now one of the most reliable large domestic
utility FGD systems. Figure 3-28 summarizes availability
data. As shown, availability for 1976 averaged 91 percent;
the first half of 1977 averaged about 93 percent.13
This system was designed for 76-percent SO~ removal
efficiency. Actual S02 efficiency has been 80.18 percent
with the seven modules operating on 720 MW (Table 3-13) .
Under maximum load, the removal efficiency averaged 76.2
3-86
-------
U)
I
00
100
90
5 80
70
60
1975
1976 1977
YEAR
1978
Figure 3-28. La Cygne availability history.
-------
Table 3-13. LA CYGNE STATION STACK SAMPLING TEST - SEVEN MODULES OPERATING
14
Date :
Time:
Load Range:
Ambient Temp:
Average S02
Removal:
5/15/75
11:30 PM to 8:00 PM
700 to 720 MW Continuous
22°C (72°F)
80 percent
Gas flow indicated
Venturi slurry flow
Venturi Ap in. H_0
cm H20
Reheater Ap in. E^O
cm H20
Absorber Dem. Ap
in. HO
cm HO
S02 efficiency, %
Inlet, ppm
Outlet, ppm
A
320K
4000
7
17.8
4
10.2
~
-
76
4506
1068
B
280K
4000
6
15.2
3
7.6
4
10.2
81
4297
834
C
300K
4000
7
17.8
3
7.6
—
-
81
4663
892
D
340K
4000
7
17.8
8
20.3
8
20.3
82
4276
776
E
320K
4000
8
20.3
6
15.2
8
20.3
77
4982
1121
F
300K
4000
7
17.8
11
27.9
8
20.3
83
4156
704
G
300K
4000
8
20.3
8
20.3
8
20.3
81
834
917
LO
I
00
00
-------
percent (Table 3-14). Efficiencies under both conditions
should improve now that eight modules are operating.
3.2.2.2 Northern States Power Co. - Sherburne Station No. 1
and No. 2 - The Sherburne County Generating plant of the
Northern States Power Company is adjacent to the Mississippi
River in Sherburne County, near Becker, Minnesota. Each of
the two units has a net capability of 700 MW. The boilers
are controlled circulation, single-reheat, balanced draft
units manufactured by Combustion Engineering. Two addi-
tional units are also being planned to boost the total plant
output to 3100 MW by 1983.
The coal presently in use is subbituminous western coal.
It has 28-percent moisture, 9-percent ash, 0.8-percent
sulfur and a high heating value of 19,000 kJ/kg (8300
Btu/lb),16
Flue Gas Scrubbing System
The two Combustion Engineering limestone scrubber
systems each use twelve scrubber modules, eleven of which
are required for full-load operation. The flow through each
of the twelve scrubber modules is approximately 94 m /s at
154°C (200,000 acfm at 310°F). The approximate inlet
loading is 6.8 g/dry m (3.0 gr/dscf) particulate and 700
ppm S02. The approximate outlet loading is 0.09 g/dry m3
(0.04 gr/dscf) particulate and 350 ppm SC>2, yielding an SO2
collection efficiency of about 50 percent.
The scrubber system has a rod-type venturi throat, a
marble bed absorber, a two-stage fiberglass Chevron shaped
demister and a tube type reheater.
Limestone is fed into the Allis-Chalmers ball mill,
where it is ground to 200 mesh, slurried to 60-percent
solids and stored in an unbaffled tank. The tank is agitated
3-89
-------
Table 3-14. LA CYGNE STATION SAMPLING TEST - SEVEN MODULES OPERATING
.15
Date:
Time:
Power:
Outside Temp:
SO2 Removal
Average:
MAXIMUM BOILER LOAD'
March 24, 1975
6:10 PM to 10:25 PM
Min. - 800 MW (e) Gross
-3.3°C (26°F)
76.2%
- Peak - 830 MW (e)
m /s
Gas flows (cfm)
Venturi pump flow gpm
1/s
Venturi Ap in. W.C.
cm W.C.
Reheater Ap (X2) in. W.C.
(X2) cm W.C.
Abs-Dem. Ap in. W.C.
cm W.C.
SO~ efficiency, %
Inlet, ppm
Outlet, ppm
A
179
380,000
3000
183
8.0
20.3
4.0
10.2
-
-
75.1
5700
1419
B
169
360,000
4000
244
7.5
19.1
3.5
8.9
5.5
14.0
79.1
5138
1075
C
188
400,000
5000
305
-
-
3.0
7.6
-
-
72.2
5516
1533
D
193
410,000+
4000
244
11.0
27.9
3.5
8.9
8.5
21.6
72.0
4995
1220
E
188
400,000
5000
305
10.0
25.4
1.0
2.54
8.5
21.6
N/A
5700
N/A
F
183
390,000
5000
305
11.0
27.9
5.0
12.7
9.0
22.9
82.9
5017
857
G
160
340,000
5000
305
9.0
22.9
12.5
31.7
9.0
22.9
75.7
5120
1243
CO
I
O
-------
with an axial-type agitator that has air sparged through
nozzles in the agitator blades. Finally, the slurry is
diluted and recirculated through limestone slurry control
valves with a conventional recirculation loop. The lime-
stone is a low-magnesium, high-grade material with 95-
percent CaCO., content.
The I.D. fans are not considered part of the scrubber
and are controlled by the boiler plant. Each fan is about
12 ft in diameter, powered with a 4,500 kW (6,000 hp)
variable speed drive. Four fans are used on each generating
unit. Fans are carbon steel and manufactured by Arsene and
Co. The fans are capable of overcoming a 109-cm (43-in.)
H20 differential pressure across the scrubber. In opera-
tion, the pressure drop is slightly more than 51 cm (20 in.)
H20.18
The marble beds have been removed from two scrubber
modules, converting them into spray towers for test pur-
poses. In all units, strainers on the discharge of the
Worthington recirculation pumps have been bypassed. Instead,
screens have been added to the interior of the absorbers.
The screens are provided with water sprays to be used inter-
mittently to backflush the screens. The system is not
entirely successful and the nozzles plug because of pre-
cipitate formation (and subsequent dislodgement) behind the
screens. Also, settling occurs in the suction chamber when
the scrubber is not in operation. To avoid this, air
19
spargers are being added for use during shutdowns.
In the scrubber there are 54 spray nozzles (9 headers,
6 nozzles per header). The nozzles occasionally plug, owing
to inefficient strainers and to erosion. Metal nozzles
failed in a very short time. Ceramic nozzles are failing
because of erosion in the tangential cone-forming chamber.
3-91
-------
The venturi scrubber was a retrofit after completion of
initial design, since particulate removal was inadequate
without it. High velocity through the venturi has resulted
in a large amount of wear. The walls of the venturi are now
lined with a 9.5-mm (3/8 in.), 316L stainless steel wear
-, . 21,22
plate.
Reheaters are three-pass in-line type, using hot water
at 138°C (280°F) directly from boiler feed-water deaerators;
they operate with no major problems. A reheat of 4.4°C
(40°F) is added to the flue gas. Steam-operated soot blowers
are part of the design, to prevent buildup of solids in the
23
reheater and at the gas/liquid interface in the venturi.
The plant has two thickeners, one for each operating
unit. They are Dorr-Oliver units 48.8 m (160 ft) in diameter.
Last winter, the rake drive on one of the thickeners stopped
as a result of an overload. This situation was not detected
for a considerable time and the entire thickener filled with
sludge. Both units were then piped to one thickener. Since
operations were not hampered with only one thickener operat-
ing, they now operate only one at a time. Thickener design
was originally for 10-percent inlet solids, 25-percent
outlet solids, and zero solids in the overflow. As the
equipment is now operated, part of the recycled water is
used to sluice down an ash hopper in the boiler economizers,
and average thickener inlet is less than 5-percent solids.
Thickener bottoms are about 12 percent, and overflow is
about 1 percent. No flocculants are used.
Overflow from the thickeners is mixed with recycle
water from the ponds in a water makeup tank and then pumped
to the scrubbers and to the economizer ash hopper sluice
with 120-1/s (2,000-gpm) Worthington pumps. The Worthington
3-92
-------
recirculation pumps are made of Ni-hard, but they are wearing
excessively. Almost all impellers have been changed out at
least once. For testing purposes, impellers made of 28-
25
percent chrome and rubber-lined steel are on order.
Piping used for recirculated slurry is rubber-lined
except at the headers, where fiberglass is used. The system
has many reducers and sharp bends, and wear of the rubber
2 6
lining at these points is a problem.
Control of pH is manual, with operation monitored by
Universal Interlok pH instruments using a flow chamber
through which a slipstream from the recirculation pump
27
flows. Scrubber blowdown to the thickener is automatic,
using Texas Nuclear sensors and Leeds and Northrup con-
*. 1 1 28
trollers.
Recycle from the pond is controlle'd by maintaining the
makeup tank level. Bubble-tube controllers are used, but
are unsatisfactory and variable capacitance sensors may be
installed to replace them. In the thickener, pH rises to 8
because of continued reaction with limestone and with
quicklime in the fly ash.
Northern States Power and Combustion Engineering are
sponsoring a long-term optimization program for the scrubber
system.
Performance
Availability for Unit No. 1 averaged 85 percent for the
4 months of operation after start-up. For the past 12
months, availability has been in excess of 90 percent. Unit
No. 2 has shown even better start-up performance, with oper-
abilities averaging about 93 percent for the first 7 months.
These data are shown in Figure 3-29.
Table 3-15 shows pertinent operating data for the first
8 months of operation of No. 1 unit. The SC>2 removal
3-93
-------
OJ
I
100
t 80
60
40
20
i i i i i i i i i r
START-UP
NO. 1 UNIT
START-UP
NO. 2 UNIT
A NO. 1 UNIT
• NO. 2 UNIT
l I i I 1 i 1
I i
J I
MAY JUN JUL AUG SEPT OCT NOV DEC JAN FEB MAR APR MAY JUN JUL
1976 1977
YEAR
Figure 3-29. Availability history
Sherburne No. 1 and No. 2.
AUG SCP
-------
Table 3-15. SHERBURNE COUNTY GENERATING PLANT
UNIT 1 - PERFORMANCE DATA
1. Unit Data (based on May 1 through December 31, 1976 data]
Electrical output
Overall capacity factor
On-line duration
3,068,130 total MW-hr
60 percent
5,176 hr
II. Scrubber System Data (Averages)
Particulate concentration:
Inlet
Inlet
Outlet
Outlet
Removal efficiency
Sulfur dioxide concentration:
Inlet
Inlet
Outlet
Outlet
Removal efficiency
4.6 to 9.2 g/dry m
(2 to 4 gr/dscf)
1.7 to 3.4 g/kJ
(4 to 8 Ib/MM Btu) 3
0.080 to 0.10 g/dry m
(0.035 to 0.044 gr/dscf)
0.032 to 0.036 g/kJ
(0.075 to 0.085 Ib/MM Btu)
98 to 99 percent
400 to 800 ppm
0.730 to 0.859 g/kJ
(1.7 to 2.0 Ib/MM Btu)
200 to 400 ppm
0.370 to 0.41 g/kJ
(0.85 to 0.95 Ib/MM Btu)
50 to 55 percent
3-95
-------
efficiency was 50 to 55 percent, which was sufficient to
32
meet local regulations.
3.2.3 Foreign Units
Table 3-16 lists 17 major limestone slurry scrubbers on
boilers in Japan. The Japanese have led FGD technology for
several years and it is evident from this table that S02
removal efficiencies in the 90- to 95-percent range are
common in Japanese limestone scrubbing units. Many of these
units are oil-fired and none reports reliability problems.
The reasons for the Japanese success are detailed in Section
3.I.3.33
3.2.4 Engineering Design Parameters.
3.2.4.1 S00 Removal Potential of the System - The limita-
^/ __
tions of the system are identical to the limitations pre-
sented for lime scrubbing in Section 3.1 of this report and
indicate a theoretical removal efficiency of 100 percent.
This section presents the limestone system results as demon-
strated at the following actual scrubber installations:
0 At the Mohave 170-MW facility of the Southern
California Edison Co., SOo removal efficiencies in
excess of 95 percent were reported for a Turbulent
Contact Absorber using limestone slurry for a low-
sulfur coal application.34
0 Reports on the packed-bed module at the 115-MW
Cholla facility show 92-percent removal of sulfur
dioxide using limestone slurry scrubbing.35
The 10-MW test facility at the TVA Shawnee re-
peatedly reports SC>2 removal efficiencies in
excess of 90 percent. During one run in February
1976, efficiency reached 96 percent in the TCA
unit on a high-sulfur coal application.36
0 The initial operating performance of the Widows
Creek No. 8 550-MW scrubber facility indicates
that SC>2 removal efficiencies substantially higher
than the designed 75 percent value have been mea-
sured. These values were measured during manually
controlled periods of operation.37
3-96
-------
Table 3-16. PERFORMANCE OF SO SCRUBBERS USING
LIMESTONE SLURRY ON BOILERS IN JAPAN
Plant site
Yokosuka
Niigata
Kashima
Karatsu
Karatsu
Ainoura
Ainoura
Sakata
Sakata
Shimonoseki
Tamashima
Omuta
Takasago
Takasago
Isogo
Fuji
Yokkaichi
Start-up
date
1974
1976
1976
1976
1976
1976
1976
1976
1976
1976
1975
1975
1975
1976
1976
1974
1974
Plant
capacity
MW
133
140
143
243
190
243
243
367
367
400
487
184
280
280
2 x 300
20
83
Fuel
Oil
Oil
Oil
Oil
Oil
Oil
Oil
Oil
Oil
Oil
Oil
Oil
Coal
Coal
Coal
NA
NA
S02
concentration,
ppm
Inlet
250
550
1000
550
550
880
880
950
950
1600
1500
1500
1500
1500
500
1340
1600
Outlet
40
55
100
70
70
110
110
50
50
50
60
150
150
150
70
20
150
Removal
efficiency
t
84
9C
90
87
87
88
88
95
95
97
96
90
90
90
86
99
91
3-97
-------
0 During the St. Clair No. 6 limestone scrubbing
demonstration and test program conducted by
Detroit Edison and Peabody, SC>2 removal values
of 90 and 91 percent (90 percent design) were
measured on high (3.0 percent) and low (0.4 per-
cent) sulfur coals.38
0 At the 167-MW Will County No. 1 facility of the
Commonwealth Edison Co. , SC>2 removal efficiencies
as high as 86.8 percent were measured during a
high-sulfur coal (4.0 percent) burn program which
yielded inlet SC>2 loadings averaging 3573 ppm.39
0 SC>2 removal efficiencies in excess of 90 percent
have been measured in tests conducted by Combus-
tion Engineering and Kansas Power and Light at
Lawrence, Unit No. 4, in 1977. The testing was
conducted on the recently installed rod scrubber
and spray tower absorber system which replaced
the original limestone injection and scrubbing
system (1968 start-up). Actual efficiencies in
the 95.5 to 97.5 percent range were measured for
low-sulfur (0.55 percent) coal.40
In summary, current full-scale and test facilities in
the United States have demonstrated SO,., removal efficiencies
in excess of 90 percent on both low- and high-sulfur coal.
In Japan, results of 95-percent removal from low-sulfur fuel
oil are reported.
The ability of an S0~ scrubbing system to remove high
levels of S02 is dependent on many design factors. The
operating conditions and results at the above plants vary
widely, but several key design parameters have proved im-
portant to high SO- removal efficiency- These parameters
have been discussed for the lime slurry system in Section
3.1 of this report. The following key parameters will be
outlined in this section for the limestone system:
0 L/G ratio - the mass transfer potential of the
system is directly related to the ratio of liquid
volume to gas volume. S02 removal efficiency is
greatly increased by increasing L/G.
3-9!
-------
0 pH - S02 removal efficiency is directly related to
slurry pH.
0 Magnesium ion concentration - limestone slurry
systems have shown increases in SC>2 removal effi-
ciency when magnesium is added to the slurry.
0 Contact surface area - In all mass transfer opera-
tions, the rate of transfer is directly related to
the transfer area. In a TCA limestone scrubber
system, the transfer area (and the mass transfer
rate) is directly related to the height of the bed
of spheres.
3.2.4.2 L/G Ratio - The L/G ratio is the flow of slurry
recirculating through the absorber divided by the flow of
flue gas. The affect of L/G on S02 removal efficiency is
shown in Figure 3-30 for the TCA unit at Shawnee.
The S02 inlet (2400 to 2900 ppm) is for a high-sulfur
eastern coal; gas velocity is 3.17 m/s (10.4 ft/s); Mg ion
concentration is 0 ppm and chloride content varied between
3000 and 7000 ppm. In this example, both the predicted
(solid lines) and measured (data points) values indicate the
strong dependence of SO_ removal on the L/G ratio. In a
^
TCA, increased L/G does not significantly increase mass
transfer, since the mass contact area is determined largely
by the height of spheres, and, in this case, the increased
S02 removal must be attributed to the better gas/liquid
interface conditions.
Figure 3-31 shows a similar effect for a TCA absorber
at the 170-MW Mohave test facility, where a 99-percent SO2
removal efficiency was realized on a low-sulfur (0.38%)
western coal.
3.2.4.3 pH - The net effect of slurry pH on S02 removal
efficiency is illustrated in Figure 3-32 for the TCA at the
10-MW Shawnee test facility- The trend predicted by the
Bechtel model in the solid lines and the actual data points
3-99
-------
L/G l/m°
6 7
8
10
TOO
90
80
70
CM
O
1/1
o
OL
60
50
40
30
20
SCRUBBER INLET pH
• pH=-5.8 LONG-TERM TEST
O pH=5.7-5.9 FACTORIAL TESTS
D pH=5.4-5.6
A pH=5.1-5.3
FACTORIAL TESTS
FACTORIAL TESTS
SCRUBBER GAS VELOCITY=3.17 m/s (10.4 ft/sec^
TOTAL HEIGHT OF SPHERES=38.1 cm (15.0 in.)
EFFECTIVE LIQUOR Mg++ CONCENTRA,TION=0 ppm
INLET SO? CONCENTRATION=2,400-2,900 ppm
LIQUOR CICONCENTRATION=3,000-7,000 ppm
I i ' '
30
40 50 60
L/G, gal/mcf
70
80
Figure 3-30. L/G ratio and scrubber inlet pH versus
predicted and measured SO removal - TCA with limestone
Shawnee plant.
41
3-100
-------
L/G
0 1
100
99
98
97
i 96
LU
ce
CVJ
o
l/l
£ 95
94
93
92
91
90
TCA 4 STAGES
A-3 STAGES
PACKED GRIDS
COAL-0.38 % SULFUR
GAS RATE- 210 Nm3/S (450,000 SCFM)
10
20
30 40
L/G GAL/MCF
50
60
70
Figure 3-31. L/G ratio versus actual percent SC
^_
170-MW test module with limestone - Mohave plant.
removal
42
3-101
-------
100
90 -
o
to
o
£
a.
80
70
60
50
40
30
4.9
T
T
LIQUID - TO - GAS RATIO
FACTORIAL TESTS
'o 8 l/ir? (60 gal/mcf)
D 6 a/nr (45 gal/mcf)
A 4 Jl/m3 (30 gal/mcf)
SCRUBBER GAS VELOCITY=10.4 ft/sec
TOTAL HEIGHT OF SPHERES=15.0 in.
EFFECTIVE LIQUOR Mgtf CONCENTRATION=0 ppm
INLET S02 CONCENTRATION=2,300-2,700 ppm
LIQUOR CI" CONCENTRATION,000-7,000 ppm
5.1
5.3 5.5
SCRUBBER INLET pH
5.7
5.9
6.1
Figure 3-32. Scrubber inlet AV versus PH for three
L/G ratios - TCA unit with limestone -
Shawnee plant.
43
3-102
-------
are shown. Higher pH results yield higher SO removal effi-
ciencies because of the increased alkalinity of the cir-
culating slurry. Although high pH yields higher S09 re-
moval efficiencies, there are some limitations, since high
pH also increases scrubber scale formation.
3.2.4.4 Magnesium Ion Concentration - Addition of magnesium
ions to the recirculating slurry has demonstrated superior
S02 removal. The reason for this phenomenon is that the
magnesium ionic species are much more soluble than the
calcium ionic species. Thus, the magnesium ions are in
solution and ready to react with the bisulfite and sulfite
ionic species. This high dissolution rate increases the
rate at which S02 can be removed from the system. The
effect of magnesium ion concentration on alkalinity is shown
in Figure 3-19.
Actual data demonstrating the effect of magnesium ion
concentration on S0~ removal efficiency are shown in Figures
3-33 and 3-34. These data are for the 10-MW TCA test module
at Shawnee. The lower curve on both graphs shows a magne-
sium concentration of zero. The middle and upper curves
show 4700 and 8200 ppm magnesium respectively. It is ap-
parent from these graphs that SO- removal efficiencies in
3
excess of 90 percent are possible at moderate [3 1/m , (60
gal/1000 acf)] L/G ratios and moderate (5.7) pH values when
the recirculating slurry contains magnesium ions.
3.2.4.5 Mass Transfer Area - The rate of mass transfer (and
S07 removal) is directly related to the mass transfer area.
In the limestone slurry TCA system at Shawnee, the transfer
area (and SO2 removal efficiency) is directly related to the
height of the bed of spheres (Figure 3-35). Pressure drop
also increases with bed height.
3-103
-------
100
90
80
CM
O
CO
70
60
50
40
L/6yl/m°
6 7
I
_tt
I
EFFECTIVE LIQUOR Mg" CONCENTRATION
FACTORIAL TESTS
O 7,000-10,000 ppm
D 3,500-5,500 ppm
A 0-500 ppm
10
20
30
SCRUBBER GAS VELOCITY*3.17 m/s (10.4 ft/see^
TOTAL HEIGHT OF SPHERES=38.1 cm (15.0 1n.)
SCRUBBER INLET pH=5.4-5.6
INLET S0£ CONCENTRATION=2,200-2,800 ppm
LIQUOR CI° CONCENTRATION'S,000-16,000 ppm
40 50 60
L/6 RATIO, GAL/MCF
70
80
'j.gure j-33, L/G ratio versus percent SO removal at
various magnesium ion concentrations TCA
with limestone - Shawnee plant.
3-104
-------
TOO
90
c£
| 80
LU
Oi
CM
O
70
60
50
40
T
tt
EFFECTIVE LIQUOR Mg ' CONCENTRATION
FACTORIAL TESTS
07,000-10,000 ppm
D 3,500-5,500 ppm
A 0-600 ppm
5.0
SCRUBBER GAS VELOCITY=3.17 m/s,(10.4 ft/sec)
LIQUID - TO - GAS RATIO= 6 */mJ (45 gal/mcf)
TOTAL HEIGHT OF SPHERES= 38.1 cm. (15.0 in.)
INLET S02 CONCENTRATION=2,300-2,700 ppm
LIQUOR CI" CONCENTRATION-12,000-16,000 ppm
i i j
5.2
5.4 5.6
SCRUBBER INLET pH
5.8
6.0
F.i.gure 3-34. Scrubber inlet PH versus percent SO,
4
removal at various magnesium ion concentrations
45
TCA with limestone - Shawnee plant.
3-105
-------
HEIGHT OF SPHERES, CENTIMETERS
6 8 10 12 14 16 18
20 22
100
T
T
T
T
T
24
—r
SLURRY FLOW RATE
FACTORIAL TES"tS 2
O 0.22 Jl/s nC(38 gal/min ft 1
D 0.16 Jl/s rC (28 gal/min ftp
A o.ll Jl/s nr (19 gal/min ft )
SCRUBBER GAS VELOCITY=3.17 m/s (10.4 ft/sec)
SCRUBBER INLET pH=5.8
EFFECTIVE LIQUOR Mg+t CONCENTRATIONS ppm
INLET S02 CONCENTRATION=2,300-2,700 ppm
LIQUOR CI" CONCENTRATION-4,000-9,000 ppm
10
HEIGHT OF SPHERES, INCHES
Figure 3-35. Height of spheres versus SO, removal
efficiency, TCA with limestone - Shawnee facility.
3-106
-------
3.2.5 Operability
Although many limestone scrubbing systems have operated
poorly to date, several other large-scale commercial lime-
stone scrubbing facilities have demonstrated excellent
operability and availability. For example, Figure 3-29
shows availability for the Sherburne No. 1 unit for its 15
months of operation, including start-up. The FGD system has
demonstrated monthly system availabilities in excess of 90
percent since shortly after start-up. Also shown is the
excellent start-up history for Unit No. 2. Figure 3-28
shows the availability history for La Cygne. This system
has steadily improved performance since start-up in 1974 to
an average availability in excess of 90 percent for 1976 to
1977.
A combination of good design and maintenance have
achieved these high availabilities.
3-107
-------
REFERENCES FOR SECTION 3.2
1. Slack, A.V. and G.A. Hollinden. Sulfur Dioxide Removal
from Waste Gases, Noyes Data Corp., 1975. p. 49-50.
2. PEDCo Environmental. SO2 Abatement for Stationary
Sources in Japan. U.S. Environmental Protection
Agency. September 1977. pp. 4-1 to 4-6.
3. Op. cit. No. 1, p. 50.
4. TVA, Detailed Cost Estimates for Advanced Effluent
Desulfurization Processes, Environmental Protection
Agency, January 1975. p. 36.
5. OOP Air Correction Division. Air Pollution Control
System Qualifications, Sulfur Dioxide Absorption
Systems.
6. PEDCo Environmental. Summary Report - Flue Gas Desul-
furization Systems, June-July 1977. pp. 121-122, p.
335.
7. Ibid.
8. PEDCo Environmental Inc. Survey of Flue Gas Desul-
furization Systems, La Cygne Station. U.S. Environ-
mental Protection Agency. June-July 1975. p. viii.
9. Ibid, p. 2-3.
10. Op. cit. No. 6, p. 111.
11. Op. cit. No. 8, p. 3-1 to 3-4.
12. Op. cit. No. 8, p. 3-4 to 3-7.
13. Op. cit. No 6, p. 114 to 119-
3-108
-------
14. McDaniel, C.D. La Cygne Station Unit No. 1 Wet Scrub-
ber Operating Experience, EPA-FEA Coal Blending and
Utilization Conference, Des Moines, Iowa, June 16-17,
1976. Exhibit H.
15. Ibid. Exhibit F.
16. Op. cit. No. 6, p. 178 to 180.
17. Kruger, R.J., J.A. Noer. Sherburne County Wet Scrubber
System Experience. Sulfur Removal System Conference.
January 6-7, 1977. p. 4 to 5, Table 1.
18. PEDCo Environmental. Sherburne No. 1 and No. 2, Trip
Report, May 19, 1977.
19. Ibid.
20. Ibid.
21. Ibid.
22. Ibid.
23. Ibid.
24. Ibid.
25. Ibid.
26. Ibid.
27. Ibid.
28. Ibid.
29. Ibid.
30. Ibid.
31. PEDCo Environmental, Inc. Summary Report - Flue Gas
Desulfurization Systems, August-September 1977. pp.
194-197, and 201.
32. Kruger, R.J., M.F. Dinville. Northern States Power
Co., Sherburne County Generating Plant Limestone
Scrubber Experience. Utility Conference Wet Scrubbing,
February 1977, p. 28.
3-109
-------
33. PEDCo Environmental, Inc., J. Ando. SO? Abatement for
Stationary Sources in Japan, September 1977, p. 3-6 to
3-13.
34. Weir, A.J., L.T. Papay, P.G. Jones, J.M. Johnson, W.C.
Marlin. Results of the 170-MW Test Modules Program,
Mohave Generating Station. Southern California Edison
Company. Presented at EPA-FGD Symposium. March 1976.
p. 12.
35. Arizona Public Service. Response letter to Don R.
Goodwin, Environmental Protection Agency. July 20,
1977. p. 3.
36. Bechtel Corporation, EPA Alkali Scrubbing Test Facil-
ity: Advanced Program, Second Progress Report, EPA.
September 1976. p. H-ll.
37. Private communication, B. Laseke, PEDCo Environmental
and TVA, August 1977.
38. Op. cit. No. 31, pp. 304-308.
39. Stober, W.G. Operational Status and Performance of the
Commonwealth Edison Will County Limestone Scrubber,
U.S. Environmental Protection Agency Flue Gas Desul-
furization Symposium, New Orleans, LA, March 8-11,
1976.
40. Green, K. and Martin, J. Conversion of the Lawrence
Flue Gas Desulfurization System, U.S. Environmental
Protection Agency, Fourth Flue Gas Desulfurization
Symposium, November 8-11, 1977.
41. Bechtel Corporation, Flue Gas Desulfurization Implica-
tions of S02 Removal Requirement, Coal Properties and
Reheat. Environmental Protection Agency. July 1977.
p. 4-7.
42. Op. cit. No. 34, p. 10.
43. Op. cit. No. 41, p. 4-11.
44. Op. cit. No. 41, p. 4-15.
45. Op. cit. No. 41, p. 4-16.
46. Op. cit. No. 41, p. 4-13.
3-110
-------
3.3 DOUBLE ALKALI FLUE GAS DESULFURIZATION SYSTEMS
Double alkali scrubbing is an indirect lime/limestone
process that cleans flue gas with fewer of the plugging and
scaling problems found in direct lime/limestone systems.
The indirect technique involves absorption of SO^ by an
alkaline solution in the scrubbing device, and treatment of
the scrubber effluent with lime or limestone in a separate
reactor outside the scrubber loop. Calcium sulfites and
sulfates are precipitated and discarded, whereas the scrub-
bing liquor is regenerated and returned to the loop. Figure
3-36 presents a simplified process flow diagram of a typical
double alkali scrubbing system.
3.3.1 Process Description
Prior to the removal of sulfur oxides, particulate
matter is removed from the flue gas via an upstream ESP,
mechanical collector(s), and/or a venturi scrubber. This is
done to minimize fly ash erosion of the scrubber internals,
and to prevent any appreciable oxidation in the scrubber
circuit due to catalytic species in the fly ash.
Sulfur dioxide removal in the scrubber is effected by
the reaction of the gas with a circulating alkaline solution
sprayed countercurrent to the gas flow. The alkaline re-
actant employed may be a compound of either potassium,
sodium, or ammonia. In the United States, the scrubbing
liquor is generally a sodium salt solution.
Soda ash (sodium carbonate) is added to water to form
the basic absorbent for the removal of sulfur oxides.
Sulfur dioxide and the small amount of S03 are absorbed into
the scrubbing liquor and react with the sodium carbonate
(Na~CO~) and sodium hydroxide (NaOH) species to form sodium
sulfite (Na2SO ) , sodium sulfate (Na2SC>4), and sodium bisul-
fite (NaHSO_). Some of the S0_ reacts with sodium sulfite
3-111
-------
CO
I
3500 cfm
FLUE GAS
RECIRCULATION
PUMP
REGENERATION
CIRCUIT PUMP
PUMP
Figure 3-36.
FMC double alkali pilot plant: schematic and simplified process
flow diagram.
-------
in the presence of water to form sodium bisulfite. An
excess of sodium hydroxide, equivalent to about 10 percent
of that required to react with all the sulfur oxides in the
flue gas, is maintained in the scrubbing liquor. The excess
sodium hydroxide reacts with the sodium bisulfite to form
more sodium sulfite, which in turn may also be oxidized to
form sodium sulfates.
The absorption of sulfur oxides is represented by the
2 3
following reactions : '
Na2C03 + 2S02 + H20 -> 2NaHSO3 + CO^ (1)
NaOH + SO2 -> NaHSO (2)
2NaOH + S02 -»• Na2SC>3 + H20 (3)
2NaOH + SO3 + Na2SO. + H20 (4)
(5)
(6)
4NaHS03 + 02 + 2Na2SC>4 + 2S02 + 2H2O (7)
After reaction in the absorbing vessel, spent scrubbing
liquor is bled to a reactor tank. Here, inactive sodium
bisulfite species are reacted with lime or limestone. In
the resultant reactions, sodium bisulfite is converted to
sodium sulfite and calcium sulfite, and calcium sulfate
solids are precipitated. The reactions vary, depending on
whether lime or limestone is used.
Regeneration with limestone (calcium carbonate) takes
place over several stages. Sodium bisulfate reacts with the
limestone to form sodium bicarbonate (NaHCO,) and calcium
sulfite (CaSO.,) . The solution is acidic because sodium
bisulfite is the major reactant. In acidic solution, sodium
bicarbonate will break down to sodium hydroxide and CO-.
The sodium hydroxide then reacts with sodium bisulfite to
produce sodium sulfite. The series of reactions occurring
in the reactor tank when limestone is used as the additive
4 5
is as follows: '
3-113
-------
NaHS03 + CaC03 + 1/2H20 -»• NaHC03 + CaSO3'l/2H2O4- (8)
NaHC03 -»• NaOH + C02 (9)
NaOH + NaHS03 -»• Na2SC>3 + H20 (10)
When lime (calcium oxide) is used as the regeneration
additive, it must first be slaked with water to form calcium
hydroxide (Ca(OH)7). Calcium hydroxide reacts with sodium
bisulfite to form calcium sulfite and sodium sulfite .
Calcium hydroxide also reacts with sodium sulfite and sodium
sulfate to form calcium sulfite, calcium sulfate and sodium
hydroxide. The series of reactions is as follows: '
NaHS03 + Ca(OH)2 •* NaOH + CaS03 - 1/2H2(H + 1/2H2 2NaOH + CaSO-j- 1/2H2CH (12)
Ca(OH2) + 2H20 £ + 2NaOH + CaSO4 • 2H2CH (13)
2NaHS03 + Ca(OH)2 -> Na2S03 + CaS03' 1/2H2O4- + 3/2H20 (14)
In regeneration of the scrubbing liquor, hold times in
the reactor tank and utilization of limestone and lime depend
on solution concentration, temperature, agitation level, and
reactant stoichiometry. In general, however, the use of
limestone for regeneration requires a hold time of an hour or
Q
more to obtain 75 to 85 percent utilization. On the other
hand, regeneration with lime requires a hold time of only 10
9
minutes for utilization of 90 percent or more. The dis-
advantages of using lime instead of limestone include the
necessity of slaking, higher chemical costs, and better pH
control requirements .
Following bisulfite removal, a side stream is taken to
neutralize the sulfite and sulfate, and to regenerate the
original scrubbing liquor. Two distinct methods exist for
sulf ite/sulfate precipitation.
In most systems, where a dilute concentration (<0.10M)
of alkali is used in the scrubbing liquor, a lime regenera-
tion method is used. Lime is added to the scrubbing liquor,
3-114
-------
and reacts with the sodium sulfite and sodium sulfate in
solution to form calcium sulfite and calcium sulfate. These
reactions are limited by the equilibrium hydroxide ion
concentration and the sulfite ion concentration, which is
more active than the sulfate species when forming sodium
hydroxide. The reactions are as follows:
Ca(OH)2 + 1/2H2O -»• 2NaOH + CaS03 • 1/2H204- (15)
Na2S04 + Ca(OH)2 + 2H20.15 M) . In this process, a side stream from the
scrubber discharge is treated with sulfuric acid to precipi-
tate sulfates and regenerate the scrubbing liquor. At pres-
ent, this regenerating method has only been used in Japan.
For either regeneration method, regenerated liquor is
returned to the scrubbing loop after two additional steps.
First, makeup sodium must be added to the solution. This is
necessary because a small amount of the scrubbing liquor is
lost in the waste solids. Second, any excess calcium ion
concentration must be reduced in order to prevent scaling in
the scrubber. Both calcium sulfate and calcium sulfite are
produced during regeneration, as shown in reactions 15 and
16. Since the solubility product of calcium sulfate is
approximately two orders of magnitude greater than that of
3-115
-------
calcium sulfite, calcium sulfate is more likely to remain in
solution and to be carried in the liquor from the clarifier.
When a gypsum phase is present in the solids at equilibrium
(saturated solution), the regeneration system produces high
calcium ion concentrations. Therefore, if conditions develop
that cause the scrubbing liquor to become supersaturated
with calcium sulfate, a high potential for scaling in the
scrubber will result.
Any chemical reaction which removes excess minerals
from solution in order to reduce scaling potential is known
as a "softening" reaction. There are several methods of
softening the reactant solution before it returns to the
absorber. One of these, known as "sulfite softening", is
effected by returning the regenerated liquor to the bisul-
fite reaction vessel. The high calcium ion concentration is
reduced as the calcium combines with residual sulfite in the
reactor to form the highly insoluble CaSO,, • 1/2H?0, which is
removed. Sodium carbonate or sodium hydroxide may be added
to the liquor to replace lost sodium.
The second method of reducing calcium ion concentrations
is known as "carbonate softening". This process involves
the use of C0~ and/or Na~COv
If limestone regeneration is used, lost sodium is
replaced by adding sodium hydroxide or sodium carbonate to
the solution; however, only as much sodium hydroxide or
sodium carbonate is added to the liquor as sodium has been
lost in the waste discharge. The excess calcium might be
removed by contacting the solution with a stream of cleaned
boiler flue gas containing carbon dioxide, although this
has not yet been demonstrated.
The carbon dioxide reacts with the calcium ions as
shown in the following equation.
3-116
-------
Ca++ + C02 + H20 -> CaC03 + 2H+ (17)
The calcium ion concentration is now reduced to a level
below the saturation value for CaSO . , and the liquor can be
returned safely to the scrubbing loop. The precipitated
calcium carbonate settles out for ultimate disposal.
The same procedure for replenishing lost sodium and
reducing calcium ion concentrations in the liquor may be
used in a lime regeneration system. When lime is used, it
is better to add sodium carbonate to the liquor from the
clarifier; this not only replaces lost sodium, but also
"softens" the solution in the following single chemical
reaction.
NaC0 + Ca++ -> CaCO + 2Na+ (18)
Sodium is replaced in the scrubbing liquor by the addi-
tional sodium hydroxide, and excess calcium is removed as
precipitated calcium carbonate. As in the limestone system,
ideally only as much sodium is added to the liquor as has
been lost in the waste discharge. The amount of sodium
carbonate added, however, may be insufficient to soften the
solution satisfactorily, and an excess over stoichiometric
requirement may be added.
The various alternative steps described above can be
combined in a number of ways in double alkali FGD systems.
Three basic modes of operation exist for sodium/calcium
scrubbing, as defined below:
1. Lime regeneration, concentrated active alkali with
side stream sulfate treatment (removal).
2. Limestone regeneration, concentrated active alkali
with side stream sulfate treatment (removal) .
3. Lime regeneration, dilute active alkali, with
carbonate softening.
3-117
-------
These modes, their advantages and disadvantages, will be
discussed very briefly in the following sections. Many
variations on these three basic schemes can be made by
changing certain parameters associated with the systems.
1. Lime Regeneration, Concentrated Active Alkali, and
Side Stream Sulfate Treatmentllf12
The first mode of operation involves lime regeneration,
use of concentrated alkali solutions, and side stream treat-
ment for sulfate removal and control. Lime is used for the
bisulfite neutralization of the concentrated (>0.10 M) alkali
scrubbing solutions. A side stream is then taken from the
bisulfite reactor, where sulfuric acid is used to precipi-
tate sulfates and regenerate the scrubbing liquor. Sulfite
softening is again employed to remove residual calcium ions
(Figure 3-37).
This flow scheme has the disadvantage of being prone to
sulfate regeneration complications. Its advantages, however,
include relatively short hold times necessary for the bi-
sulfite/sulfite regeneration with lime, as well as high
reactant utilizations; an advanced state of development;
and high S02 removal efficiencies with low flows of the
concentrated alkali scrubbing liquors. This mode of opera-
tion appears to be promising, but its successful application
on a universal scale depends on development of an effective
treatment for sulfate removal and control.
2. Limestone Regeneration, Concentrated Active Alkali,
and Side Stream Sulfate Removal-1-^5114
This mode differs from the first only in its use of
limestone instead of lime for the bisulfite removal step.
Like the previous mode, it too is prone to sulfate regenera-
tion complications. It has the further disadvantage of
requiring long hold times for bisulfite/sulfite regeneration.
Its advantages, however, appear to be high S0? removal
3-118
-------
CLEANED
FLUE
GAS
VENTURI |
SCRUBBER /
CENTRIFUGE
OR FILTER
WASTE PRODUCT
(CaSO])
SULFURIC
ACID
Na2S04 SULFATE
+ CONVERSION
TANK
WASTE
PRODUCT
(GYPSUM)
CENTRIFUGE
OR FILTER
EFFLUENT
HOLD TANK
Figure 3-37 . Schematic of a double alkali system with lime or limestone
regeneration, concentrated alkali, and E^SO. sulfate removal.
-------
efficiencies with low liquor flow rates, and high reactant
utilizations. Successful large-scale application of this
flow scheme also depends on effective treatment for sulfate
removal and control.
3. Lime_Regeneration, Dilute Active Alkali, and Carbonate
Sof tenTng-fS, 16
The third mode of operation is lime regeneration using
dilute alkali with carbonate softening for calcium ion
(scaling) control. Spent scrubbing liquor is treated with
lime to neutralize bisulfite and to react with sulfite and
sulfate. This process uses carbonate softening (CO,, and
Na^CQ,.) to lower the calcium ion concentration. A typical
flow sheet is given in Figure 3-38. Among its disadvantages
are the possibility of inadequate sulfate regeneration;
scaling potential; and the need for large volumes of rela-
tively dilute scrubbing liquors. Advantages are found in
its ability to deal with high oxidation levels,, relatively
simple equipment requirements, and advanced state of develop-
ment.
Of the three process schemes described above, the most
promising for utility applications appears to be the first,
involving use of concentrated alkali with lime regeneration
and side stream sulfate treatment. The advantage lies
mainly in its higher relative SO,, removal capabilities and
17
lower capital costs.
A typical material balance for this operational scheme
for a 500-MW coal-fired plant is presented in Figure 3-39.
In all of the operational schemes presented above, a
calcium sulfite/sulfate sludge is generated that must be
disposed safely.
3-120
-------
VENTURI |
SCRUBBER
CLEANED
FLUE
GAS
OJ
i
CENTRIFUGE
OR FILTER
Figure 3-38. Schematic of a double alkali system with lime regeneration,
dilute alkali, and carbonate softening.
-------
M
Figure 3-39. Material balance for 500 MW doioble alkali
FGD system with lime regeneration.
-------
Stream no.
Description
Rate, 1000 kg/h
NmVs
mVs, Liquid
Particulates, kg/h
Temperature, °C
Specific gravity
Viscosity, pascal-second
Undissolved solids, %
pH
Sulfur as SO2, kg/h
1
Coal
to
boiler
170
11900
2
Air
to air
heater
2049
464
43.3
3
Air
to
boiler
1848
419
279
4
Gaa
to
economizer
1998
445
15300
477
11900
5
Gas
to air
heater
1998
445
15300
374
11900
6
Gas
to
ESP
2199
4 nn
15300
154
11900
7
Gas
to
fan
2186
490
2600
11900
8
Fan
outlet
gas
2186
490
2600
11900
9
Gas
bypass
77.3
17
91
440
10
Gas to
part .
3crbr .
2109
473
2510
11480
11
Ga> to
SO2 ab-
sorber
2196.6
503
12.2
51 . 7
10330
12
Gas
to
reheater
2180.8
501.4
12.2
51.7
770
Stream no.
Description
Rate, 1000 kg/h
NmVs
H\3/s, Liquid
Particulates , kg/h
Temperature, °C
Specific gravity
Viscosity, pascal-second
•Jndissolved solids, %
PH
Sulfur as SO2, kg/h
13
Reheater
discharge
gas
2180.8
501.4
103.2
78.9
770
14
Gas
to
stack
2258.1
518.4
103.2
78.9
1190
15
Soda ash
to
feed tank
0.790
16
Lime
to
slaker
11.46
17
Chemical to
f ixa t ion
tank
4. 58
18
Fresh
makeup
water
62
0.0172
1.0
-0-
19
Fregh
makeup
water
50.76
0.041
1.0
-0-
20
Makeup
water to
ven turi
44.16
0.0123
1.0
-0-
21
Part, scrbr.
slurry to
hold tank
4515
1.218
225750
1.03
5
103700
22
Recycle
from
hold tank
4465
1.204
223250
1.03
5
102520
Figure 3-39 (continued), Material balance for 500 MW double alkali
FGD system with lime regeneration.
-------
Stream no.
Description
Rate. 1000 kg/h
NroVs
nH/s, Liquid
Particulates, kg/h
Temperature, *c
Specific gravity
Viscosity, pascal-second
Undissolved solids, «
pK
Sulfur as 803, kg/h
23
Pond
recycle
water
45. 44
0.0126
1.0
-0-
24
88. 24
0.0245
1.0
-0-
25
4553.24
1. 228
223250
1.03
5
102520
26
Scrubber
to
part, scrbr.
4597.6
1. 240
223250
1.03
5
J02520
27
Demiater
wash
water
fi.6
0.00183
1 .0
28 ,
Demister
wash to
hold tank
6.6
0.00183
1 0
-0-
29
Spent
slurry to
hold tank
4900.2
1. 361
30
Recycle
from hold
tank
4906.8
1.363
1 . 0
31
Purge to
reactor
tank
110.2
0.0306
1 . 0
32
4796.6
1.332
1 0
OJ
I
to
Stream no.
Description
Rate. 1000 kg/h
NmVs
m^/B, Liquid
Particulates, kg/h
Temperature, "C
Specific gravity
Viscosity, pascal-second
Undissolved solids, %
pH
Sulfur ss SOj, kg/h
33
Soda
solution
92.4
0.0257
1 .0
-0-
12.3
34
Scrubbing
slurry to
absorber
4889
1.3581
1.0
35
Clarifier
feed
183.3
0.0467
1.09
15
36
Clarifier
overflow
91.6
0.0254
1.0
-0-
37
Clarifier
under f low
91.7
0.0212
1.20
30
38
Filter
cake
45.8
0.00848
1.5
60
39
Filtrate
45.9
0.01275
1.0
-0-
40
Mixing
water
3120
0.00087
1.0
-0-
41
42.8
0.0119
1.0
-0-
42
Slurry
to
reactor
73.13
0.0185
1.1
20
12.1
Stream no.
Description
Rate. 1000 kg/h
NmVs
mVs, Liquid
Particulates, kg/h
Temperature, °C
Specific gravity
Viscosity, pascal-second
Undissolved solids, %
pH
Sulfur as SO-, kg/h
43
Slaker
dump
0. 340
44
Particulate
slurry
product
50
0.0135
2500
0.03
45
Fly ash
disposal
12.7
12700
46
Sludge
to disposal
pond
, v 53.5
0.0099
1.5
60
Motea;
Calculations based on:
a. 130% stoichiometric lime: 95% CaO and 5% grit.
b. 60* of the grit is removed.
c. 90S SO2 removal.
d. 90* solid in ash pond.
e. CaSO.^1/2 «2O/CaSO4 • 2HjO = 40/60 by weight.
Figure 3-39 (continued). Material balance for 500 MW double alkali
FGD system with lime regeneration.
-------
3.3.2 Double Alkali Scrubber Units in the United States
Several successful bench-scale, pilot plant and proto-
type double alkali FGD systems have been tested on boiler
flue gas applications in the United States. The success of
these programs has resulted in commitments by three separate
electric utility companies to install full-scale double
alkali FGD systems on coal-fired boilers. As yet no full-
scale system is operating on a utility boiler in the United
States; but several systems are working on coal-fired indus-
trial boilers and one pilot plant system and one prototype
have been tested on utility units.
Industrial applications on coal-fired boilers are as
follows:
Company: General Motors, Inc.
Plant: Chevrolet
Location: Parma, Ohio
Stream treated: Off-gas from coal-fired boilers
System size: 124 m3/s (262,000 acfm) (32 MW)
SO2 inlet: 800 to 1300 ppm (1.5 to 3.0% S coal)
Start-up date: March 1974
Company: Caterpillar Tractor Co.
Plant: Joliet Plant
Location: Joliet, Illinois
Stream treated: Off-gas from coal-fired boilers
System size: 48.8 m3/s (103,500 acfm) (18 MW)
S02 inlet: 2300 ppm (4% sulfur coal)
Start-up date: September 1974
Company: Firestone Tire and Rubber Co.
Plant: Pottstown Plant
Location: Pottstown, Pennsylvania
Stream treated: Off-gas from a power boiler
System size: 6.6 m3/s (14,000 acfm)
S02 inlet: 1,000
Start-up date: January 1975
3-125
-------
Company: Caterpillar Tractor Co.
Plant: Mossville Plant
Location: Mossville, Illinois
Stream treated: Off-gas from 4 coal-fired boilers
System size: 113 m3/s (240,000 acfm) (57 MW)
Fuel properties: Coal, 3.2 percent sulfur average
Start-up date: October 1975
One prototype and one pilot plant double alkali system
have been operated on utility coal-fired boilers:
Utility: Utah Power and Light Co.
Unit: Gadsby Station, Unit No. 3
Location: Gadsby, Utah
Unit size: 1.2 m3/s (2500 acfm) (-0.6 MW)
Fuel properties: Coal, 0.4 percent sulfur average
Start-up date: 1971
Note: Terminated 1973
Utility: Gulf Power Co.
Unit: Scholz, Unit No. 1
Location: Chattahoochee, Florida
Unit size: 35 m3/s (75,000 acfm) (20 MW)
Fuel properties: Coal, 3 to 5 percent sulfur
Start-up date: February, 1975
Note: Terminated July 1976
As a result of the success of pilot and prototype systems,
three full-scale double alkali systems are scheduled for
operation soon on new coal-fired utility boilers. Tables
3-17, 3-18, and 3-19 present design information on these
units.
The following sections present detailed information on
each of the industrial and prototype units listed above.
3.3.2.1 General Motors - Parma Plant
22
System Description - Figure 3-40 is a process flow
diagram for Parma. Flue gases from four spreader stoker
boilers rated at a total of 32 MW are routed to four tray-
type absorbers. The gases enter through a prequench-section
at the bottom of each scrubber, and then flow in counter-
current to an aqueous sodium hydroxide sulfite/bisulfite
solution. The scrubber is a four-tray, impingement-type
3-126
-------
Table 3-17. DESIGN PARAMETERS FOR THE CANE RUN NO. 6
DOUBLE ALKALI SCRUBBER PLANT,18
LOUISVILLE GAS & ELECTRIC CO., LOUISVILLE, KY
FGD unit rating, MW
Fuel, characteristics
FGD system supplier
Process
Active alkali mode
Application
Start-up date
FGD modules
Module design
System performance
guarantees:
SO,., removal efficiency
Particulate emissions
Sodium consumption
Lime consumption
Power consumption
Filter cake properties
System availability
277
Coal: 26,700 J/g (11,500
Btu/lb) 3.5 to 4.0 percent
sulfur, 11.5 percent ash
CEA/ADL
Double alkali, sodium-calcium
Concentrated
New
2/79
Two
Two-stage tray tower absorber
200 ppm in the scrubber dis-
charge of 95 percent S02
removal if the sulfur content
of the coal is 5 percent or
greater
No additional loading over
inlet concentration
0.045 moles of soda ash makeup
per mole of S02 removed when
maximum coal chloride level is
0.06 percent. 0.05 moles of
soda ash additional for each
mole of chloride in the coal
above the 0.06-percent level
Maximum 1.05 moles/mole S02
removed
1.1 percent of peak operating
rate (300 MW)
55 percent insoluble solids
90 percent for a one-year
operating period
3-127
-------
Table 3-18. DESIGN PARAMETERS FOR THE A. B. BROWN NO. 1
19
DOUBLE ALKALI SCRUBBER PLANT,
SOUTHERN INDIANA GAS & ELECTRIC CO., WEST FRANKLIN, INDIANA
FGD unit rating, MW
Fuel, characteristics
FGD system supplier
Process
Application
Start-up date
FGD modules
Module design
SO2 removal efficiency,
(% design)
System performance
guarantees:
Availability
SO,, removal
Chemical consumption
Power consumption
Mechanical components
Sludge disposal
Projected unit capacity
factor
250
Coal: 26,700 J/g (11,500
Btu/lb) 3.53 percent S, 8.76
percent ash, 11.35 percent H.,0
FMC
Double alkali, sodium-calcium
New
4/79
Two
Two-stage disc contractor,
cascading liquor absorber
85a
95 percent for 60-day consecu-
tive test run
85 percent for 4.5 percent
sulfur coal
Yes (soda ash and lime)
Yesb
One year for mechanical failures
Hauled to off site landfill
90 percent for the first year
of operation; 70-percent
average for 35-year life span
Based on 4.5 percent maximum sulfur content in coal.
Specifics of guarantees not known.
3-128
-------
Table 3-19. DESIGN PARAMETERS FOR THE NEWTON NO. 1
DOUBLE ALKALI SCRUBBER PLANT,20'21 CENTRAL ILLINOIS
PUBLIC SERVICE CO., NEWTON, ILLINOIS
FGD unit rating, MW
Fuel, characteristics
FGD system supplier
Process
Active alkali mode
Application
Start-up date
FGD system design:
Modules
Module design
Regeneration
Dewatering
Filter cake disposal
SO,, removal efficiency
HC1 removal efficiency
Availability
Soda ash consumption Gg/yr,
(TPY)
Quicklime Gg/yr. (TPY)
Power consumption
575
Coal: 25,300 J/g (10,900
Btu/lb) 4-percent sulfur, 0.2-
percent chloride
Buell-Envirotech
Double alkali: sodium-calcium
Concentrated
New
11/79
Four
High velocity cocurrent spray
tower, countercurrent two-
tray-stage mobile ball bed
absorber (series arrangement)
Three process-causticizers and
and two thickeners
Three horizontal extraction
filters
Fly ash stabilization/on-site
landfill
95 percent, less than 200 ppm
90 percent
100 percent for 70-percent
load factor for 30-year life
span
3.4 (3800) (- 0.05 moles/mole
SO,., removed)
104 (115,000) (- 1.10 moles/
mole SO removed)
3 percent of peak operating
rate (550 MW)
3-129
-------
NaOH
u>
I
NaOM
FILTRATE
PUMP
SLURRY
TANK
SODA ASH
FEED PUMP
SURGE
TANK
SCRUBBER
-'FEED PUMP
Figure 3-40. Parma double alkali scrubbing system: schematic and
simplified process flow diagram.
-------
additional reaction and solids separation. Liquid effluent
is then pumped to the second clarifier, where it is softened
by addition of NaC03 only. Solution from this tank is re-
cycled to the scrubber recirculation loop. Underflow from
both clarifiers is pumped to the slurry storage tank for
batch processing through the vacuum filters. Filtrate is
returned to the primary clarifier for recovery of NaOH.
Pertinent design data are given in Table 3-20.
Distinguishing Features - This installation is charac-
terized by:
0 Operation in the dilute mode (0.10 normal or less)
with lime regeneration and carbonate softening.
0 Simultaneous particulate and SO_ control.
0 Firing with coal of 1.5 to 3.0 percent sulfur.
Operating History - The system has performed well with
regard to SO,., removal. Results of a 1-week test in 1974
indicate S0~ removal efficiencies in the 94- to 99-percent
range, with relatively low inlet S09 levels (600 ppm to
23
1200 ppm) and high excess air rates. A test was conducted
by A. D. Little and General Motors (GM) from August 19, 1974
to May 14, 1976. It consisted of three, 1-month intensive
test periods and 18 months of lower-level tests. Removal
of S09 reflects the variations in operating modes employed
by GM during the period, but removal efficiencies were at 90
24
percent for the viable operating modes. Operation during
April and May 1976 was excellent and A. D. Little recommended
continued operation in the mode used during this period.
The operability (hours the FGD system was operated/boiler
operating hours in a period expressed as a percentage) of
the Parma system from August 1974 through April 1977 is
presented in Table 3-21. The system's best period of
operation was May through August, 1976, when operability
3-131
-------
Table 3-20. GM PARMA. DOUBLE ALKALI SCRUBBING SYSTEM DESIGN,
24
OPERATING, AND PERFORMANCE CHARACTERISTICS
Application
Fuel, characteristics
System design
Process/mode
Pressure drop
Status
Start-up date
Inlet gas conditions:
Flow rate
so2
°2
Particulate
SO9 outlet concentration
£*
SO,-, removal efficiency
Scrubbing solution
characteristics:
PH
Total sodium
Active sodium
Calcium ion
Soda ash makeup
Lime utilization
Filter cake production rate
Filter cake solids
Filter cake disposal
Four coal-fired stoker boilers
Coal: 25,600 to 31,400 J/g
(11,000 to 13,500 Btu/lb) 1.5
to 3.0 percent sulfur
Four three-stage multiventuri
flexitray scrubber modules
Dilute active alkali
25 to 33 cm H2O (10 to 13 in.
H20)
Operational
3/74
30.9 m /s @ 27°C (65,500 acfm
@ 80°F)
800 to 1300 ppm
N/A3
0.7 g/m3 (0.3 gr/scf)(dry)
20 to 130 ppm
90 to 99 percent
5.5 to 7.5
0.58 to 0.96 M
0.087 to 0.13 M
305 to 490 ppm
>o.r
mole/mole SO,, removal
1.32 to 1.90 mole/mole S in
cake
40 to 55 percent
Offsite landfill
N/A = not available.
Nonsteady-state operations resulted in makeup rates of
0.028 to 0.05 moles Na /mole S0_ removed.
3-132
-------
unit with feed and recycle streams added at the top. The
absorption trays (Koch) have movable bubble caps for ad-
justment to variations in gas flow. Pressure drop through
the absorber is designed at 19cm (7.5 in) H^O and the maxi-
mum L/G ratio is 2.7 1/m3 (20 gal/1000 CF). The liquid
stream is composed of about 20 percent fresh feed and 80
percent recycle.
In the initial mode of operation, a side stream from
the scrubber recirculation loop was constantly fed to a
mix tank, where CaCO, slurry from the softener clarifier
was added for neutralization. Effluent from this tank was
pumped to another back-mix reactor, where lime slurry was
added for regeneration of caustic. The regenerated solution,
with a high concentration of fly ash and calcium precipitates,
flowed to two reactor clarifiers in series. In the first
clarifier, the solution went through additional reaction and
solids separation. Liquid effluent was pumped to the second
clarifier where it was softened by the addition of Na~CO.,
and CO,.,. Solution from this tank was recycled to the scrub-
bers as fresh feed; underflow from the primary clarifier was
pumped to a slurry storage tank for batch processing through
a vacuum filter; and cake from the filters was hauled to a
landfill for disposal. Filtrate was returned to the primary
back-mix reactor for recovery of NaOH.
Several changes were made in this operational mode that
resulted in a more successful operation. In the modified
system, a side stream from the scrubber recirculation loop
is constantly fed to the back-mix reactor, where lime slurry
is fed for regeneration of caustic. Effluent from this tank
is fed to another back-mix reactor for further reaction.
The regenerated solution flows to two reactor clarifiers in
series. In the first clarifier, the solution goes through
3-133
-------
Table 3-21. PARMA DOUBLE ALKALI SCRUBBING SYSTEM PERFORMANCE
HISTORY: OPERATION TIME, OPERABILITY, AND COMMENTS
OJ
I
M
OJ
Month
Aug. 74
Sep. 74
Oct. 74
Nov. 74
Dec. 74
Jan. 75
Feb. 75
Mar. 75
Apr. 75
May 75
Jun. 75
Operation hours
Boiler
1234
555
1190
1355
1245
1495
1825
1595
1105
1610
1190
FGD system
1234
475
1015
100
150
360
505
455
280
440
220
Operability %
1234
86
85
7
12
24
28
29
25
27
18
Comments
Units 1 and 3 were required for service. No. 3 scrubber
module was unavailable for approximately 25 hours.
All units were in service during the month. Total avail-
able time was approximately 40 hours.
Units 1, 2, and 4 were in service during the month. No. 1
scrubber module was the only module in service, operating
for approximately 100 hours.
All units were required for service. No. 1 scrubber
module was the only module in service, operating for
approximately 150 hours.
Units 1, 2, and 4 were required for service. No. 1
module was the only module in service, operating for
approximately 360 hours. Total scrubber system unavail-
able time was only 70 hours.
All units were required for service. Total scrubber
system unavailable time was approximately 245 hours.
All units and scrubbers were in service. Total scrubber
unavailable time amounted to approximately 230 hours.
All units were required for service. Scrubber No. 4
was not put in service, although it was available for
most of the month.
All units were in service. Scrubbers 1 and 4 operated
during the period. Total system unavailable time was
approximately 40 hours.
All units were required for service. Scrubber modules
1, 3, and 4 were operated during the month. Approxi-
mately 90 hours of scrubber unavailability time were
recorded during the month.
Recent operation has been intermittent. Tests have been
conducted for analysis of particulate loading at the
boiler outlet. The system was restarted in May, but was
shut down because of a plugged chemical feed line to
clarifier. After restart, around June 10, the system
was shut down because of similar plugging in another part
of the line. Unit ran for about 2 weeks in June.
-------
Table 3-21 (continued). PARMA DOUBLE ALKALI SCRUBBING SYSTEM
PERFORMANCE HISTORY: OPERATION TIME, OPERABILITY, AND COMMENTS
u>
I
Ul
Month
Jul. 75
Aug. 75
Sep. 75
Oct. 75
Nov. 75
Dec. 75
Jan. 76
Feb. 76
Mar. 76
Apr. 76
May 76
Jun. 76
Operation hours
Boiler
1234
2331
2135
1379
1084
1149
924
FGD system
1234
1848
1250
240
847
1042
816
Operability %
1234
79
59
7
78
91
88
87
Comments
FGD system was down in July and August for replacement
of gravity flow lines by an open flume.
FGD system restarted September 8, and operated during
this period at an availability factor of 80 percent.
(Note: The figures given for September and October
represent the operation hours from September 8 to
November 9 . )
The FGD system operated during the report period
(November-December), except for a scheduled holiday
shutdown from December 23 to January 4, 1976, Because
of problems with solids and solids carryover, GM con-
ducted steam tests during December to determine whether
solids carryover is due to a high solids recirculation
rate. The FGD system was down most of December because
of these tests, which caused the low operability factor.
GM has found solids concentration in the clarifier too
high for efficient system operation. A. D. Little is
scheduled to test the system in April as part of an EPA
evaluation program.
The low operability index during March resulted from
extensive modifications performed on the system by GM.
During this period, the scrubber blowdown and scrubber
flow indication system was revised.
From April 19 to the end of the month the system was
100 percent operable. During this period, all system
modifications were completed and all major problem areas
corrected.
No major problems were encountered during the May-June
reporting period. One system upset occurred during the
period because of an operator error. Boiler hours were
low in June because of low process demand. The scrubbing
system characterization study was concluded on May 14,
1976.
-------
Table 3-21 (continued). PARMA DOUBLE ALKALI SCRUBBING SYSTEM
PERFORMANCE HISTORY: OPERATION TIME, OPERABILITY, AND COMMENTS
to
I
CO
en
Month
Jul. 76
Aug. 76
Sep. 76
Oct. 76
Operation hours
Boiler
1234
599
715
809
1174
FGD system
1234
599
715
223
734
Operability %
1234
100
100
40
63
Comments
The boiler demand was very light during the report months
(July-August), requiring only the boiler's being continu-
ously on line. The system was down for one week because
of the annual plant inventory. Modifications to the
operation of the gas scrubbing system included:
0 Elimination of polymeric addition.
0 Elimination of sludge blanket for filtration.
0 High-percentage sulfate concentration in the filter cake.
0 Incorporation of a pH controller in the chemical mix
tank regulating the lime feed into the system. pH con-
trol should reduce the current stoichiometric require-
ments from the 1.25 to 1.50 range down to 1.10 to 1.20.
Two boilers and their corresponding scrubber modules were
in service during the month. Operation time was logged
primarily in the initial part of the month. A number of
problems were encountered, forcing a shutdown throughout
the remainder of the month. These included:
Solids deposition and plugging of the chemical tanks
and reactors, requiring shutdown and cleanout.
Agitator problems.
Boiler problems.
Replacement of demister pads.
The system was returned to service the second week in
October. Two boilers and their corresponding scrubbers
were put back in the gas stream. Problems encountered
stemmed primarily from a mechanical collector malfunction,
resulting in excessive dust loadings in the modules and
eventually causing widespread plugging of the lower bubble-
cape trays. Shutdown and cleanout of the modules plus
repairs to the mechanical collector were required. System
restarted November 8. Major modifications on the double
alkali system during the period included the installation
of a Great Lakes pH monitor and control unit for the regu-
lation of lime feed to the regeneration reactor-clarif ier
unit. This unit employs a digital readout and is scheduled
for operation in November. Running a seeding line back to
the reactor-clarif ier unit resulted in superior crystal
growth by the calcium sulf ite-sulfate salts.
-------
Table 3-21 (continued). PARMA DOUBLE ALKALI SCRUBBING SYSTEM
PERFORMANCE HISTORY: OPERATION TIME, OPERABILITY, AND COMMENTS
Month
Nov. 76
Dec. 76
Jan. 77
Feb. 77
Mar. 77
Apr. 77
Operation hours
Boiler
1234
1321
1559
1852
1462
1476
1300
FGD system
1234
787
349
608
892
1278
1158
Operability %
1234
60
22
33
61
87
89
Comments
The FGD figures reported for the month of December are
somewhat distorted, because only UAW personnel can operate
the system; therefore, the scrubbers could not be operated
during the Christmas holiday. During the latter part of
November and early December, the scrubbers were taken
out of the flue gas path, because of continued solids
buildup in the mix tank. In addition, some piping arrange-
ments were modified and the gear reducer in the mix tank
had to be replaced.
Lost an agitator on the No. 2 chemical mix tank. Low
operability factors for the winter months due primarily
to severe weather conditions and plant curtailments.
March and April operations were very satisfactory. SO2
removal has been measured well above the 90-percent level.
Particulate removal has been below par because of solids
carryover.
u>
I
Based on reference 23 and personnel communication between PEDCo and G.M.
personnel.
-------
averaged 94 percent. The GM Parma plant has several unique
characteristics that affect operability- Each boiler is
equipped with its own separate scrubbing module with no
provision for crossflow between modules. The plant is not
needed during the summer months, because operations are
shutdown during automobile model changeover changes and
because there is no need for heating. United Automobile
Workers personnel operate the scrubber plant, which precludes
operation of the scrubbers when they are not on the site.
The GM plant is a developmental system, and as such is
subject to modifications. Many of the low operability
periods were due to mechanical outages, or outages for
modifications to accomodate and test new modes of operation.
25
Operational Problems/Solutions - Major problems
encountered at the Parma plant were:
1) Scaling in the scrubber loop
2) Periodic excess entrainment
3) Excessive lime consumption
4) Excessive soda ash consumption
5) Low solids content in sludge
Solutions to these problems are presented below:
1) Scaling in the scrubber loop - During the test
program at the Parma plant, scaling occurred in the scrubber
loop under two sets of circumstances. In the first instance,
calcium carbonate scale formed on the top tray of the
scrubber, where fresh scrubber feed mixed with a recycle
stream. The high pH levels on this tray and the lack of
mixing resulted in the absorption of CO- from the flue gas
and formation of CaCO^. General Motors attempted to elimi-
nate the problem by introducing fresh scrubber feed to the
recycle tank for better pH control. In this scheme, calcium
3-13!
-------
sulfite scale formed on the scubber internals. This was
caused by the high pH in the recycle tank, high calcium ion
levels, and low sodium ion strength. The company reduced
this scaling by installing a line to mix regenerated liquor
with scrubber recycle liquor at the entrance to the top tray.
The major difference between this and the first mode of
operation is that in the new mode the recycle stream and the
fresh regenerated slurry feed are mixed entirely outside the
scrubber in a pipe. This assures good mixing and eliminates
localized high pH.
2) Periodic excess entrainment - Although entrainment
of scrubber effluent is within design specifications, at
times it can contribute significantly to particulate load-
ing. This was the result of poor gas distribution to the
demister, rendering a portion of it ineffective. Although
no remedial measures were attempted by GM, it was discovered
that replacement of the original stainless steel wire mesh
entrainment separators with polypropylene units virtually
eliminated visual evidence of entrainment problems.
3) Excessive lime consumption - Lime consumption has
been significantly higher than what was expected in theory.
It has been improved, however, from an initial stoichiometry
value of 1.92 moles Ca per mole S02 removed to a value of
1.32 moles Ca per mole SO- removed. Problems occurred
because of difficulties with the automatic lime feed system
which forced adoption of manual control, and resulted in the
low sodium level in the system. The scrubber liquor exhibited
an Na2S04 concentration of 0.33 M, at which the solubility of
[OH"] is less than 0.1 M. The lime feed rate was controlled
on the basis of a target hydroxide level of 0.1 M[OH ], re-
sulting in overfeeding of lime. Lime consumption was im-
proved by adjusting the feed rate to give 0.08 to 0.09 M
3-139
-------
[OH~] in chemical mix tank No. 2, compared with the previous
value of 0.1 M, and by installing a flow orifice to measure
the scrubber feed rate accurately -
4) Excessive soda ash consumption - In the early
stages of operation, solubles loss was reported in excess of
0.1 moles Na2 per mole SC>2 removed. This was caused by
failure to wash the filter cake, by high liquor content in
the filter cake, and by high sodium in the liquor. Washing
the filter cake has eliminated this problem to the point
that soda ash makeup is now determined by the minimum amount
required for the softening step. This amount has not yet
been determined, but with softening requirements correspond-
ing to Ca reduction of 150 ppm (to 650 ppm), the stoichio-
metry is 0.05 moles Na,., per mole SO? removed.
5) Low solids content in sludge - Excessive lime
consumption and low oxidation rates resulted in a solids
content in the filter cake of 39 to 46 percent. This was
improved to 56 percent in the latest test runs using cake
washing and the maintenance of low excess lime feed.
Summary - Operation of the Parma facility has shown the
viability of the double alkali process for removal of sulfur
dioxide. Removals of 90 percent or better have been ob-
tained on coals ranging from 1.5 to 3.3 percent sulfur
[25,600 to 31,400 J/g (11,000 to 13,500 Btu/lb)].26 Several
different operating modes have been investigated and signi-
ficant improvements have been obtained in both process and
mechanical performance. Although it has yet to be proved
over an extended test period, it is believed that in the
latest operational mode the system is capable of long-term
reliability-27
3-140
-------
3.3.2.2 Caterpillar Tractor - Joliet Plant
28 29 30
System Description ' ' - Figure 3-41 is a simpli-
fied process flow diagram for the Joliet double alkali FGD
system. Flue gases from two coal-fired stoker boilers are
routed to two Zurn Dustraxtor scrubbers. Here the gases are
continuously contacted with a dilute aqueous sodium hydrox-
ide/sulfite/bisulfite solution. The pressure drop through
the scrubber varies from 25 to 30 cm H90 (10 to 12 in. H»0),
o "
with a fresh feed L/G ratio of 8 1/m (60 gal/1000 acf).
A side stream of solution is withdrawn from the scrubber to
an external mix tank. Here, slaked quick lime is added for
neutralization and regeneration of active alkali. The
solution is drawn to a second mix tank, where the reaction
continues. Insoluble calcium salts precipitate as the Ca(OH)
reacts with the sodium salts. Underflow from the second mix
tank is pumped to a vacuum filter and the calcium salts are
dewatered to about 65 percent solids. This slurry is then
disposed of in an offsite landfill. The vacuum filter is
washed with fresh water to recover as much of the soluble
sodium as possible. Filtrate from the vacuum filter, which
contains a high concentration of calcium ions, is recycled
back to a clarifier. The clarifier also receives the re-
generated solution overflow from the thickener. Sodium
carbonate is added to the clarifier to "soften" the solution
by reacting with the calcium ion to form calcium carbonate
precipitate. Underflow from the clarifier is pumped to the
vacuum filter for dewatering of the calcium carbonate before
disposal. Overflow from the clarifier, regenerated active
alkali solution, is recycled to the scrubbers. Pertinent
design and operating data for the Joliet plant are presented
in Table 3-22.
3-141
-------
FRESH WATER
NJ
BOILER
NO. 2
50,000 LB/HR
STEAM
BOILER
NO. 3
100,000 LB/HR
STEAM
DRY CHEMICAL
TRANSFER
CaO
Na2C03
>- I—
cC X
S 3E
o:
Q_
LU
s: ^
CD t—
O **
0 s:
to
^^
CJ 1—
< x
— ^
WASTE PRODUCT
(10) PICK-UP
CONTAINERS
Figure 3-41.
Joliet double alkali scrubbing system: schematic and
simplified process flow diagram.
-------
Table 3-22. JOLIET DOUBLE ALKALI SCRUBBING SYSTEM
DESIGN, OPERATING, AND PERFORMANCE CHARACTERISTICS31
Application
Fuel, design characteristics
System design
Status
Start-up date
Design inlet gas conditions:
Flow rate
so2
SO- outlet concentration
Actual SO- removal efficiency
Soda ash consumption
Lime consumption
Water makeup
Power consumption
WAste disposal
Filter cake, percentage of
solids
Total operation time
System availability
Two coal-fired stoker boilers
Coal: 27,600 J/g (11,869
Btu/lb) 4 percent sulfur
Two dustraxtor scrubber modules
Operational
9/74
48.8 m3/s @ 182°C (103,500
acfm @ 360°F)
2325 ppm
115 to 350 ppm
85 to 95 percent
35.9 g/s (285 Ib/hr)
131 g/s (1,040 Ib/hr)
0.038 1/s (36 gph)
6.85 kWh/hr
655 g/s (5,200 Ib/hr)
50 to 60 percent
8600 hr
90 to 100 percent
3-143
-------
Distinguishing Features - This installation is charac-
terized by:
0 Operation in the dilute mode with lime regenera-
tion and carbonate softening.
0 Simultaneous particulate and SC>2 removal in
multitube impingement/entrainment scrubber.
0 Firing coal of about 4.0-percent sulfur content.
Operating History - The system has achieved excellent
SO,, removal efficiencies of between 85 and 95 percent under
32
various operating conditions. Sulfur dioxide inlet con-
centrations are high, about 2300 ppm. The system was de-
signed to attain an emission level of 0.86 g SO9/MJ (1.9 Ib
c *•
S0?/10 Btu) (75% S0« removal), but has consistently per-
formed much better than design.
The operability of the FGD system has been improving
steadily. Process availability for the period October 24,
1975, through June 1976 has been 100 percent. Most prob-
lems at the Joliet plant are mechanical; the majority are
solved while still on-stream or during scheduled shutdowns
resulting. As a consequence, there have been few forced
outages.
Operational Problems/Solutions ' - Problems en-
countered at the Joliet plant included:
1) Unreliability of the chemical feed system
2) Inoperability of the SO~ monitors
3) Excessive water consumption
4) Feed water pump impeller cavitation
5) Poor vacuum filter belt life
Reasons for these problems and their solutions are presented
below:
3-144
-------
1) Unreliability of the chemical feed system - Two
problems arose with the lime and soda ash supply systems.
The first was the failure of electronic components in the
feed control system, resulting in inaccurate feed rates.
This was alleviated by replacement of the faulty electronic
components. The second problem was the formation of lumps
inside the storage silos. They were caused by moisture
condensation on the inside walls and leakage of rain water
into the silos. The lumps also caused inaccurate feed
rates. The problem was solved by the installation of a
dry air purge system on each silo.
2) Inoperability of SO,., monitors - The system employs
several monitors, in the inlet duct and in the stack, to
measure SO_ concentrations. They experienced problems
caused by pluggage of the sample probes. The difficulties
were removed by increasing the pressure of the blow-back air
used to clean the probes, and by the installation of shields
around the probes.
3 Excessive water consumption - During early opera-
tion, the scrubber did not achieve closed-loop operation
with respect to water, because of the addition of excess
water through the process pump seals; from vacuum filter
wash sprays; and as a result of the low water evaporation
rates in the scrubber when it was under low boiler loads.
Closed-loop operation was obtained by installation of rota-
meters on the water supply lines of the process pump seals,
and by using clarified scrubber feed solution for the vacuum
filter belt wash. A seal water recirculation system was
also installed on the vacuum pumps.
4) Feed water pump impeller cavitation - Insufficient
suction head in the surge tank caused cavitation, which in
turn led to impeller damage. This was solved by increasing
3-145
-------
the surge tank height to raise the suction head at the feed
water pumps.
5) Poor vacuum filter belt life - Originally the life
of the multifilament polypropylene filter belt was poor,
because of the accumulation of solids in the interstices of
the cloth. The pressure of the belt wash sprays was in-
creased; and this resulted in extended belt life. Other
materials and weaves are being tested to see if belt life
can be increased even further.
Summary - The Joliet facility has demonstrated the
viability of a double alkali FGD system for control of SO™
on coal-fired boilers. Sulfur dioxide removals of 85 to 95
percent have been demonstrated on high-sulfur coal together
with very good operability. The problems encountered, which
were mechanical and generally associated with ancillary
hardware, have been effectively solved.
o/r 07
3.3.2.3 Caterpillar Tractor - Mossville Plant ' - The
double alkali FGD system at the Mossville Engine Plant of
Caterpillar Tractor Co. is the largest installed to date by
FMC Corporation. It went into commercial service in October
1975. The system employs a dual-throat, venturi-type scrub-
ber to remove SO- and particulates. The scrubber handles
flue gas from four coal-fired stoker boilers at the rate of
113 m /s (240,000 acfm). The boilers are rated at 58 kg/s
(460,000 Ibs/hr) of steam (57 MW equivalent).
The system operates in the concentrated mode (>0.15M
active alkali) using lime for regeneration of the sodium
sulfite scrubbing liquor. Precipitated calcium solids are
dewatered in a vacuum filter to about 60 percent solids
before disposal in a landfill.
FMC reports very high system availabilities (90%) and
90-percent S02 removal efficiencies. The problems experi-
3-146
-------
enced at Mossville are related to wear of control valves,
filter cloth, and conveyor belts. FMC feels that satisfac-
tory solutions have been reached for all these problems.
3.3.2.4 Firestone - Pottstown Plant
System Description ' - Figure 3-42 is a simplified
process flow diagram for the Pottstown plant. A slipstream
of flue gas equivalent to about 3 MW is routed to the scrub-
bing system from a pulverized-coal boiler (formerly burning
2-percent sulfur oil, but being converted to burn 3.0-
percent-sulfur coal). The flue gas is contacted with the
scrubbing liquor in a dual-throat venturi scrubber. It
then moves to a cyclone separator, which collects the scrub-
bing liquor and discharges it to a recirculation tank.
Cleaned flue gas then passes out the top of the separator,
through a mist eliminator, and on to the stack. A side
stream is drawn off the recycle tank and fed to a reaction
tank. Here, slaked lime is added to regenerate the sodium
sulfite, bringing about precipitation of calcium solids.
The solution is then pumped to a thickener, where the cal-
cium solids are settled and the thickened slurry passed to a
rotary drum filter. The slurry is dewatered to a cake of
about 55 percent solids and disposed of in a landfill.
Supernatant from the thickener and the filter is pumped back
to the recycle tank. At this stage, sodium carbonate is
added to make up for sodium lost in the filter cake. Per-
tinent design and operating data for the Pottstown Plant are
given in Table 3-23.
Distinguishing Features - This installation is charac-
terized by:
0 Operation in the concentrated mode (3 M active
alkali) with lime regeneration and carbonate
softening.
3-147
-------
OJ
i
CO
BOILER
FLUE GAS
14,000 cfm
130°F
TO EXHAUST
STACK
HYDRATED
LIME
DUAL
THROAT
VENTURI
CYCLONE
SEPARATOR
SODIUM
CARBONATE
SOLUTION
THICKENER
FILTER CAKE
AND FLY ASH
Figure 3-42. Pottstown double alkali demonstration plant: schematic and
simplified process flow diagram.4^
-------
Table 3-23. FIRESTONE POTTSTOWN DOUBLE ALKALI SCRUBBING
SYSTEM: DESIGN, OPERATING, AND PERFORMANCE CHARACTERISTICS
Application
Fuel, characteristics
System design
Pressure drop
Status
Start-up date
Inlet gas conditions:
Flow rate
S02
S0~ outlet concentration
Actual SO2 removal
efficiency44
Scrubbing solution
characteristics:
PH
Total sodium
Active sodium
Soda ash makeup of solution
Lime utilization
Filter cake production rate
Filter cake solids
Filter cake disposal
Actual system availability
45
Power boiler slipstream (3 MW
equivalent)
Oil, 2 percent sulfur; or
coal, 3 percent sulfur
Dual-throat venturi scrubber/
cyclonic separator
25 to 76 cm H90 (10 to 30 in.
H20)
Operational
1/75
6.6 m /s (14,000 acfm)
1,000 ppm
100 ppm
90 percent
6.5
3.0 M
0.2 to 0.3 M
0.019 m3/h (5 gph)
100 percent
23 g/s (180 Ib/hr)
55 percent
Landfill
94 percent
3-149
-------
0 Simultaneous particulate and S0_ removal.
0 Operation on 2-percent-sulfur oil or 3-percent-
sulfur coal.
Operating History - The system has exhibited excellent
S00 removal efficiencies of 90 percent on high-sulfur oil,
41
but no data are available for its performance on coal. it
has also achieved a very high availability: 99 percent for
42 43
the first 12 months of operation. ' Most downtime
periods were due to mechanical component failure or to
maintenance, and not to unwanted chemical changes or side
reactions. No scaling problems have been experienced.
Table 3-24 presents a summary of the causes for system
downtime during 1976.
Table 3-24. AVAILABILITY OF THE FIRESTONE POTTSTOWN
DOUBLE ALKALI SCRUBBING SYSTEM: 1976 OPERATING HISTORY
Item responsible Percentage of Percentage of operating
for outage total downtime period (year)
Thickener plugging 20.7 1.35
Pumps 16.1 1.04
Cake conveyor 16.0 1.04
Fan 15.4 1.00
Lime feeder 13.4 0.87
Spray nozzles 12.4 0.80
Instrumentation 4.1 0.27
Control valves 1.9 0.13
Total 100.0 6.50
Total availability for period =93.5 percent
3-150
-------
An ongoing test program is being conducted at the Pottstown
facility. Its goals include the determination of corrosion
characteristics of the various components; optimization of
filter and thickener systems; and performance of coal-firing
versus oil-firing.
Operation Problems/Solutions - Major problem areas
encountered at the Pottstown Plant included:
1) Thickener plugging
2) Failure of recirculating pump
3) Fan outage
Reasons for these problems and their solutions are given
below:
1) Thickener plugging - the single largest outage was
caused by plugging of the thickener underflow pipe
after a plastic beaker had been accidently dropped
into the unit. No mechanical or chemical upsets
occurred in the unit during the first year.
2) Failure of recirculation pump - recycle pump fail-
ure, due to separation of the pump lining from the
pump housing, was caused by improper maintenance
of the pump seal. A proper maintenance program
has eliminated this problem.
3) Fan outage - the fan failed because bearings
overheated, which in turn was caused by misalign-
ment of the fan shaft. Proper alignment has
alleviated this problem.
Summary - Operation of the Pottstown facility has
demonstrated the high efficiency and reliability of a system
operating in the concentrated mode. This pilot plant is a
scale-up of a smaller pilot plant facility; it has confirmed
virtually all the process data and relationships developed
4 7
in that program. Although the Pottstown facility has
demonstrated S02 removal efficiencies of 90 percent and an
availability of about 94 percent on oil-firing, future
3-151
-------
testing will provide an indication of performance on coal-
firing.
3.3.2.5 Utah Power & Light Co. - Gadsby Station
48 49
System Description ' - Figure 3-43 is a simplified
process flow diagram for the double alkali pilot plant
installed by Envirotech on Unit 3 at the Gadsby Station.
A simplified process flow diagram of a full scale
double alkali system now under construction at the 575-MW
Newton Station of Central Illinois Public Service Co., is
provided in Figure 3-44. At the Gadsby Station, a slipstream
of 1.2 actual m3/s (2500 acfm) from the 100-MW coal-fired
unit was routed to the pilot plant. The flue gases are
quenched in a venturi quench section and fly ash, chlorides,
and particulate matter are removed. The flue gas then flows
countercurrent to an aqueous solution of sodium hydroxide/-
sulfite/bisulfite in a mobile-bed scrubber with two trays of
polysphere balls. Entrained liquid in the flue gas is
removed in a mist eliminator before discharge of clean flue
gas to the stack. The spent scrubbing solution is collected
in a recycle tank, where atmospheric air is bubbled through
the solution to oxidize the sodium sulfite to sodium sul-
fate. A side stream from the recycle tank is pumped to two
reaction tanks and slaked lime is added for regeneration of
sodium hydroxide. The solution then passes to a reactor-
clarifier to settle the precipitated calcium sulfate solids.
Underflow from the reactor-clarifier is sent to a thickener,
in which calcium solids are dewatered.
The thickened slurry (40% solids) is then pumped to a
rotary vacuum filter so the slurry can be further dewatered
to a calcium sulfate cake (70-80% solids). The cake is
washed to recover any soluble sodium still present in it,
and the wash solution is pumped back to the reactor-clarifier.
3-152
-------
OJ
I
I-'
Ul
u>
CLEAN
GAS
RAW H20
FLUE GAS
2500 cfm
[LIME STACK pLANT
t| p USE
COOLING
TOWER
BLOW DOWN
PROCESS
NEEDS
Figure 3-43. Envirotech Gadsby double alkali pilot plant: schematic and
simplified process flow diagram.
-------
BASH HATER
EIMCO PROCESS-CAUSTICIZERS
ASH PUMP
FILTER CAKE
CONDITIONING
FACILITY
TO LANDFILL
Figure 3—44. Double aXkali system at Newton Station.
-------
Supernatant from the reactor-clarifier and the thick-
ener is pumped to a second reactor-clarifier. Soda ash
(sodium carbonate) is added to this reactor-clarifier as a
softener. Excess calcium ions are precipitated as calcium
carbonate, which is then pumped to the first reaction tank.
Overflow from the reactor-clarifier is pumped to the recycle
tank.
This installation is a pilot plant, and as such has
been tested in a number of different configurations. In
addition to the polysphere scrubber, a venturi scrubber has
been tested. Two types of mist eliminators have been used,
a spin vane in series with a cyclone, and the Euroform
Chevron type. Most of the testing was done in a dilute
mode, but some was also performed in the concentrated mode.
Pertinent design and operating data are given in Table
3-25.
Distinguishing Features - This installation is charac-
terized by:
0 Operation in both the dilute mode and the concen-
trated mode with lime regeneration and carbonate
softening.
0 Simultaneous particulate and S02 removal.
0 Firing with 0.4-percent-sulfur coal.
Operating History .- The scrubbing system has performed
well with respect to SO,., removal. Various modes of opera-
tion were tested using two types of absorbers. With the
polysphere absorber, SO- removals of 90 percent were achieved,
51
giving outlet concentrations of 15 to 40 ppm SO . With
the venturi absorber, efficiencies ranged from 80 to 85
percent SO- removal.
With the exception of the first three-month operating
period, during which some gypsum scaling problems were
encountered, dilute mode operations were conducted for
3-155
-------
Table 3-25. ENVIROTECH/GADSBY PILOT PLANT TEST PROGRAM:
DILUTE MODE PERFORMANCE AND RESULTS50
Application
Fuel, characteristics
Gas inlet conditions:
Flow rate
S02
°2
Particulate
S02 outlet
Particulate outlet
S02 removal efficiency
Particulate removal
efficiency
L/G ratio
Pressure drop
Recycle liquor composition:
PH
Active sodium
Calcium ion
Sulfate ion
Chloride ion
Lime consumption, percentage
stoichiometric
Reactor residence time
Filter cake composition:
CaSO.-2H20
CaCO
H20
Solubles
Ash
Pulverized coal power boiler
Coal: 0.4 percent sulfur
1.2 m3/s @ 10.4°C (2500 acfm @
220°F)
250 to 1500a ppm
5.0 percent
0.5 to 0.7 g/m3 (0.2 to 0.3
gr/scf)
15 to 40 ppm
0.02 to 0.05 g/m3 (0.01 to
0.02 gr/scf)
90 percent
93 to 95 percent
2.7 1/m3 (20 gal/1000 acf)
10 to 25 cm H7O (4 to 10 in.
H20)
5.8 to 7.7
0.10 M
500 ppm
40,000 to 45,000 ppm
1000 ppm
110 to 115 percent
120 min
42 to 57 percent
10 to 15 percent
30 to 40 percent
2 percent
1 to 2 percent
During later test runs S02 was added to the flue gas to
simulate higher sulfur conditions.
3-156
-------
52
almost two years without any major problems. No operating
problems causing shutdown were experienced between October
1972 and August 1974. It should be noted that for conven-
ience the system was shut down on weekends, but no drainage
of solution or cleaning of equipment took place during these
shutdowns.
The majority of the testing with the polysphere scrub-
ber was conducted at a total pressure drop of 10 cm HO (4
in. H_O) using two trays of balls. Figure 3-45 shows the
performance of a single polysphere tray; SO- removal is
-> *•
plotted against L/G ratio in 1/m (gal/1000 acf) at three
different pressure drops, 10, 18, and 25 cm (4, 7, and 10
in.) H-0 and at a pH from 5.8 to 7.1. Figure 3-46 shows a
plot of S0~ removal versus scrubber effluent pH at an L/G
3
ratio of 2.5 1/m (19 gal/1000 acf) at a total pressure drop
of 10 cm (4 in.) H,,0 across two polysphere trays. It can be
seen that, generally, two trays operating at a pressure drop
of 10 cm (4 in.) H~O are equivalent to one tray operating at
a pressure drop of 18 cm (7 in.) H90 with L/G ratios of 2.5
3
1/m (19 gal/1000 acf) in each case.
The scrubber/regeneration liquors had a chloride con-
tent of about 1000 ppm at steady state. Soluble chloride
usually concentrates in wet scrubbing systems, owing to the
conversion of coal chloride to HC1, which appears in the
flue gas and is absorbed by scrubber liquors. No obvious
detrimental effects were noted for chloride levels in the
range experienced in this double alkali application.
Operational Problems/Solutions - The major problem at
the Gadsby facility was gypsum scaling in the scrubber,
caused by excess calcium concentrations in the recycle
slurry. In the original operating flow scheme only one
reactor-clarifier was used. The scaling problem was solved,
therefore, by the addition of a second reactor-clarifier, in
3-157
-------
100
£/ACTUAL m°
234
95
90
CVJ
o
GO
85
80
75
10
20 30
L/G - GAL/MACF
40
Figure 3-45. SO- removal versus L/G ratio for the
Envirotech/Gadsby pilot plant with
a single stage polysphere absorber.
53
3-158
-------
100
90
Q
UJ
w 80
i
UJ
CC
o
70
60
o
o
o
o
o
o
4
8
SCRUBBER EFFLUENT pH
Figure 3-46. S0? removal versus scrubber effluent pH for
the Envirotech/Gadsby pilot plant at an L/G of 2.5 1/rrr
(19 gal/lOOOacf) and using a two-stage absorber.
3-159
-------
which the carbonate softening step is performed, and by the
addition of an organic coagulant to aid solids separation.
Summary - The Gadsby pilot plant has demonstrated that
a soluble alkali system operating in the dilute mode can
work reliably, removing 90 percent SO-, on flue gases from
utility boilers that burn low-sulfur coal. Good lime
utilization was achieved (110 to 115% stoichiometric re-
quirement) and waste solids exhibited good settling and
dewatering characteristics, thus enabling an end-product of
60 to 70 percent solids.
3.3.2.6 Gulf Power - Scholz Station
System Description^? _ Figure 3-47 presents a simpli-
fied schematic of the CEA/A.D. Little double alkali FGD
system as installed on the Scholz Steam Plant.
Fifty percent of the flue gas from a 40-MW pulverized
coal-fired boiler is routed to a high-efficiency ESP. The
efficiency of the ESP is variable, so that the impact of
various levels of particulate matter on the process can be
studied. The flue gases then pass to a variable-throat
venturi scrubber and an absorption tower arranged in series.
The tower can be operated as a spray tower, a tray tower, or
as a de-entrainment separator. The venturi can be used for
SC>2 absorption or for particulate collection, depending on
whether water or absorbent is circulated through it. The
pressure drop across the venturi can be varied from 13 to 25
cm (5 to 10 in.) H20. After the venturi, saturated flue gas
enters the bottom of the absorber, where it is contacted
countercurrently with the circulating absorption liquor.
Cleaned flue gas is reheated with an externally fired oil
burner before discharge to the stack. Spent solution is
collected in the internal recycle tank at the bottom of the
3-160
-------
CO
I
h-
01
SCRUBBED GAS
AIR
BOILER
FLUE GAS
75,000 acfm
276°F
SODA ASH
f SILO
SOLIDS
Figure 3-47. Scholz double alkali prototype scrubber: schematic and
process flow diagram.
-------
absorber. A bleed from the collected liquor (pH 5 to 7) is
sent to the venturi recirculation loop for additional S02
absorption. A continuous bleed stream (pH 4.8 to 6.0) is
drawn from the venturi recirculation loop and pumped to a
reactor system. Here, hydrated lime is added for regenera-
tion of active alkali and precipitation of calcium sulfite
and sulfate. Sufficient lime is added to produce a pH of
11.0 to 12.5 in the reactor underflow. The underflow, about
5-percent solids by weight, is pumped to a thickener, where
the slurry is concentrated to about 30-percent solids and
pumped to a vacuum filter. The solids cake on the filter is
washed to recover soluble sodium. The filter cake, at 45 to
60 percent solids, is trucked to a one-acre disposal area.
Liquid filtrate and wash liquor from the filters are
returned to the thickener. Sodium carbonate is fed to the
thickener to make up any sodium losses. This is not intended
as a softening step, since the Ca concentration is usually
less than 100 ppm; but some CaCO_ precipitation does occur.
Clear liquor overflow from the thickener is fed to a holding
tank for recycle to the absorber. The combined fresh feed
and recycle streams have a pH from 6 to 11. Pertinent
design data are presented in Table 3-26.
Distinguishing Features - This installation is charac-
terized by:
a Operation in the concentrated mode using lime for
regeneration.
0 Firing with coal of 0.9- to 3.1-percent sulfur.
0 Capability of several venturi/absorber configura-
tions.
3-162
-------
Table 3-26. CEA/ADL GULF POWER DEMONSTRATION UNIT:
c p
ORIGINAL DESIGN PARAMETERS
Identification Value
Boiler Unit No. 1
Boiler capacity
Nominal
Maximum
Fuel
Fuel characteristics
HHV
Sulfur
Ash
Absorber inlet conditions
Capacity
Generating capacity
Flow rate (standard)
Flow rate (actual)
Temperature
Sulfur dioxide loading
Particulate loading
Oxygen
S00 removal efficiency
Maximum S02 removal rate
Maximum particulate loading
Power consumption, percentage of
unit capacity
Without venturi scrubber
With venturi scrubber
40 MW
47.5 MW
Pulverized coal
28,400 J/g (12,200 Btu/lb)
3 percent
14 percent
20 MW
25.5 m3/s @ 21°C
(54,000 scfm <§ 70°F)
35.4 m3/s @ 136°C
(75,000 acfm @ 276°F)
136°C (276°F)
1800 to 3800 ppm (dry)
0.05 g/m3 (0.02 gr/scf) dry
6.5 percent (max.)
90 percent
694 kg/hr (1530 Ib/hr)
0.05 g/m3 (0.02 gr/scf) dry
1.0 to 1.5 percent
2.5 to 3.0 percent
3-163
-------
Operating History - The system started up February 3,
1975, and operated continuously through July 18, 1975,
when it was shut down for repairs and modifications. The
second period of operation was from September 16, 1975,
through January 2, 1976.
The system exhibited excellent S02 removal capabilities,
90 percent and greater. Using the combined venturi/
absorber configuration at a venturi liquor pH above 5.2,
outlet S00 concentrations below 50 ppm were achieved; this
59
corresponds to greater than 95-percent removal. Raising
the pH of the venturi liquor above 6.0 resulted in S00
59
removal efficiencies greater than 98 percent. Fxgure
3-48 presents the relationship of scrubber bleed liquor pH
to SO,., outlet concentration.
Table 3-27 summarizes the operation of the Scholz
facility with respect to various performance parameters.
The operability of the Scholz plant for the period
February 1975 through June 1976 is presented in Table 3-28.
The Scholz plant was designed to demonstrate the viability
of the double alkali process technology for application on
utility coal-fired boilers. As such, this prototype plant
had less spare equipment than would be normal in full-
scale applications. The operability of the system has
been steadily improved; during the last four months of
operation, it was 94 percent.
Operational Problems/Solutions - Problems occurring at
the Scholz plant have been mainly mechanical and equipment-
related. These are summarized in Table 3-29, together
with their solutions.
Summary - Operation of the Scholz prototype plant has
demonstrated the viability of 90-percent S02 removal at high
3-164
-------
500
400 _
300
00
i
H"
CTi
U1
Outlet S02
(ppm)
Active
Na+, (M)
Operational Configuration
. Venturi + 2 Trays 13-18 (5-7)
o Venturi + 2 Trays 20-28 (8-11
ft Venturi (No Feed to Trays) 20-28 (8-11
0.25-0.4
0.15-0.3
0.15-0.3
Inlet SO9 = 1050 - 1250 ppm
200 _
100 -
5.5 6.0
Scrubber Bleed Liquor pH
Figure 3-48. Outlet SO,., concentration versus scrubber bleed pH
63
-------
Table 3-27. CEA/ADL DOUBLE ALKALI PROTOTYPE SCRUBBER
SYSTEM OPERATION RELATIVE TO MAJOR PERFORMANCE PARAMETERS
60
SO- removal
Particulate removal
Oxidation
Sodium makeup
Lime utilization
Power consumption
Filter cake properties
The system generally operated at
S02 removal efficiencies of about
95 percent and demonstrated the
capability for removal of over 99
percent.
As was intended in the design, there
was no increase in loading.
Maximum oxidation rates of 25 to 30
percent in closed-loop operation
were experienced without serious
process upset.
Under these process conditions,
approximately 0.03 mole Na2C03 was
required per mole of SO,, removed
as makeup for soluble solids losses
in the washed cake. About an equal
amount of sodium makeup was un-
accounted for, probably because of
leakage from the system and errors
in soda ash makeup estimates.
Lime utilization typically ranged
from 95 percent to 100 percent.
The overall lime stoichiometry was
0.95 to 1 mole Ca(OH)2/mole of
S02 removed from the flue gas.
This was 2.5 to 3 percent of the
power generated by the utility
boiler with the system operating at
full fan and at full venturi and
absorber liquid recirculation
capacity. Correcting for the excess
fan and pump capacity, the power
consumed by the equipment actually
required in the application (Tray
tower at an L/G of 0.7 to 1.3 1/m3
(5 to 10 gal/1,000 ft3) is roughly
1 to 1.5 percent of the power gen-
erated at the design conditions.
Waste Solids Properties - The
system produced a washed filter
cake generally containing at least
50 percent insoluble solids. The
waste material, on the average,
contained 3 to 5 percent soluble
solids (dry cake basis).
3-166
-------
Table 3-28. CEA/ADL DOUBLE ALKALI PROTOTYPE SCRUBBER
PERFORMANCE HISTORY: OPERATION AND VIABILITY
PARAMETERS
Period
Total
period,
hr
Boiler
operation,
hr
Feb.
Mar.
Apr.
May
Jun.
Jul.
Aug.
Sep.
Oct.
Nov.
Dec.
Jan.
Feb.
Mar.
Apr.
May
Jun.
75
75
75
75
75
75
75
75
75
75
75
76
76
76
76
75
76
672
744
720
744
720
744
744
720
744
720
744
744
696
744
720
744
720
459
507
604
598
720
683
744
577
559
620
732
0
0
480
642
735
656
FGD
operation,
hr
454
485
336
375
720
221
0
254
559
560
732
0
0
445
616
651
641
FGD"
operability,
percent
98.9
95.7
55.6
63.2
100.0
32.4
0
44.0
100.0
90.3
100.0
0
0
92.7
95.9
88.6
97.7
FGD
utilization,
percent
67.6
65.2
45.2
50.4
100.0
29.7
0
35.3
75.1
77.8
98.4
0
0
59.8
85.6
87.5
89.0
a
FGD operability: The number of hours the FGD system was in
operation divided by the number of hours the boiler was in
operation, expressed as a percentage.
FGD utilization: The number of hours the FGD system was in
operation divided by the number of hours in the period, expressed
as a percentage.
3-167
-------
Table 3-29. SUMMARY OF THE PROBLEMS AND SOLUTIONS IN OPERATION
OF CEA/ADL DOUBLE ALKALI PROTOTYPE SCRUBBER
Unit operation
area
Scrubber-absorber
module
Regeneration
reactor
Thickener
Vacuum filter
Sodium carbonate
addit.-'.on
Problem
Failure of rubber linings
in bleed control valves
Plugging of dry lime feed
chute to first stage of
multistage reactor
Solids deposition in
first stage of multistage
reactor
Broken agitator blade in
second stage of multi-
stage reactor
Erosion and plugging of
flow-through pH meters
Plugging of thickener
underflow lines
Deterioration of
thickener
Erosion and cracking of
fiberglass scraper and
rocker arm
Erosion of plastic bridge
valve
Holes in filter cloth
Loss of vacuum due to
cracks in trunnion tubes
Insufficient agitation
Plugging in dry feeder
gate
Insufficient reliability
of sodium makeup feed
solution
Solution
Replaced with stainless steel
valves
Installation of feed vibrator
Design and installation of a
new first-stage unit
Part replacement
Replaced with immersion type
meter with sonic cleaner
Installation of flexible lines
and addition of back-flushing
procedures
Patch/replace lining section
Replaced with stainless steel
parts
Part replacement
Repaired holes
Patched cracks
New agitation system
Replaced feed control unit
Revised monitoring system
and employed manual techniques
3-16!
-------
Table 3-29 (Cont'd). SUMMARY OF THE PROBLEMS AND SOLUTIONS IN OPERATION
OF CEA/ADL DOUBLE ALKALI PROTOTYPE SCRUBBER
Unit operation
area
Roheater
Fans
Pumps
Tanks
Piping
Stack
Problem
Deterioration of refrac-
tory in burner chamber
Fly ash buildup on fan
blade
Deterioration of thrust
bearing
Leakage of scrubber re-
cycle pumps through pump
seals and seals and
piping
Broken shaft on reactor
pump
Failure to reactor pump
isolation valve
Freezing of pump seal
water rotameters
Cracks and pinholes in
the lining of the ab-
sorber recycle tank
Plugging of pH meter on
recycle tanks
Liner failure in the
thickener hold tank
Insufficient reliability
of level transmitter in
the thickener hold tank
Circuitry failure in
heat tracing
Failure of stack lining
Solution
Replaced refractory material
Clean and rebalance fan
Part replacement
Replaced packing and flange
gaskets
Part replacement
Overhauled
Replaced parts and modified
operating procedures
Repaired
Repaired and relocated
sampling lines
Replaced
Replaced
Replaced
Repair/replace
3-169
-------
availabilities on high- and low-sulfur coal applications„
The success of the prototype facility has enabled a full-
scale system to be installed on a utility boiler (Louisville
Gas and Electric's Cane Run 6).
3.3.3 Description of Japanese Double Alkali FGD Installa-
tions
There are 19 plants in Japan on which sodium-based
double alkali systems are operating. The plants differ from
U.S. applications in that by-product gypsum is produced
rather than disposable sludge. Table 3-30 lists the systems
and their applications. Of these installations, 13 are
industrial-size boilers, five are utility boilers, and one
is a sulfuric acid plant. All boilers are fired with oil,
but the principle of operation and equipment used is virtu-
ally identical to operation on coal-fired units.
Following is a brief description of the systems, as
offered by the suppliers, and a summary of system perform-
ance in Japan.
3.3.3.1 Kawasaki/Kureha - Kureha Chemical Industry Co.,
Ltd., and Kawasaki Heavy Industries, Ltd., have jointly
developed a commercial double alkali process utilizing
concentrated sodium solution and limestone regeneration.
The system is currently installed on five oil-fired utility
boilers.
System Description - A typical schematic diagram of the
system as installed at Tohoku Electric is shown in Figure
3-49. The system consists of an SO,, absorption section, a
limestone regeneration section, a sulfate removal section,
and a gypsum production section.
There is a precooler section at the entrance to the
scrubber, in which the hot incoming gas from the boiler is
3-170
-------
Table 3-30. SODIUM-BASED DOUBLE ALKALI PROCESS INSTA
S~ Q
IN JAPAN
Process developer
Showa Denko
Showa Denko-Ebara
Absorbent,
precipitant
Na2S03,
OJ
I
Kureha-Kawasaki
Tsukishima
Na SO,,
CaCO-
CaCO.
CaCO-
CaO
User
Showa Denko
Kanegafuchi Chem.
Showa Pet. Chem.
Nippon Mining
Yokahama Rubber
Misshin Oil
Poly Plastics
Aj inomoto
Kyowa Pet. Chem.
Japan Food
Yokohama Rubber
Asia Oil
Tohoku Electric
Shikoku Electric
Shikoku Electric
Myushu Electric
Tohoku Electric
Kinuura Utility
Daishowa Paper
Plant site
Chiba
Takasago
Kawaski
Saganoseki
Hiratsuka
Isogo
Fuji
Yokkaichi
Yokkaichi
Yokkaichi
Mie
Yokohama
Shinsendai
Sakaide
Anan
Buzen
Akita
Nagoya
Fuji
Capacity,
1,000 Nm3/hr
500
300
200
120
105
100
212
82
150
100
100
243
420
1,260
1, 260
730
1,050
185
264
Sourci
Industr:
Industr.
Industr:
H2so4 P:
Industr:
Industr:
Industr:
Industr:
Industr:
Industr:
Industr:
Industr:
Utility
Utility
Utility
Utility
Utility
Industr:
Industr:
-------
JNaOH WATER |
CaC03 VENT
u>
i
H2S04
ABSORBER
VENT
REGENERATION
REACTOR
SYSTEM
I
ABSORBER
S02
STRIPPER
FRESH
WATER
AIR
GYPSUM
CENTRIFUGE
Figure 3-49. Kureha-Kawasaki double alkali process at
Tohoku Electric's Shinsendai station.
-------
adiabatically cooled with recycle liquor. The scrubber is
of the grid-packed type, consisting of two separate stages
of packing; gas flows upward through the packing. Scrubber
recycle liquor and effluent from the top stage are fed to
the top of the bottom stage which is operated at an L/G
3 3
ratio of about 1.6 1/m (12 gal/1000 ft ). Regenerated
liquor alone is fed to the top stage, thus maximizing the pH
in this stage and achieving good countercurrent mass trans-
fer; this results in high S02 removal efficiency. Clean
flue gas leaving the absorber is reheated to 141°C (285°F)
by the direct firing of oil.
The regeneration reaction consists of neutralization of
NaHSO- with finely ground limestone to regenerate sodium
hydroxide. The limestone is ground to 10 pm in a wet,
vertical tower mill. Scrubber effluent liquor bleed taken
from the bottom-stage recycle stream is filtered to remove
oil soot before it is pumped to the regeneration reactors to
be reacted with limestone. The reactor system consists of
five closed reactors in series, operated at about 90°C to
promote the reaction between limestone and NaHSO-. in the
scrubber bleed. The solid product produced in the reactors
is separated from the mother liquor and washed in a vacuum
filter. The liquor is recycled to the scrubber system, and
the solids, mostly CaSO *l/2 HO. are fed to the oxidizer
system and to the sulfate removal section.
This system is designed to remove sulfate formed in the
scrubber as gypsum through reaction with sulfuric acid and
calcium sulfite. The process is fed with a bleed stream of
scrubber effluent, a portion of the calcium sulfite cake
(washed) from the filter and sulfuric acid. In order to
maximize sulfuric acid efficiency, a small absorber and
3-173
-------
stripper are incorporated in the sulfate treatment loop.
The incoming scrubber bleed is acidified in the absorber
with S02; this SC>2 is supplied by the stripper, which con-
verts NaHSO, to Na2S03, stripping off S02 by means of thermal
decomposition according to the following reaction:
2NaHS03 + Na2S03 + S02 + H20 (19)
Feed liquor to the stripper is produced by the acidic mother
liquor from the sulfuric acid reactor in which the sulfate
removal occurs. Stripped liquor is recycled to the regen-
eration reactor system.
Feed to the sulfuric acid reactor is thus already at a
low pH, and therefore requires less sulfuric acid addition
to achieve a given conversion of Na^SO. to gypsum.
Effluent slurry from the sulfuric acid reaction is
centrifuged to separate the gypsum produced from acidic
mother liquor. The separated solids are transferred to the
oxidizer system, together with the calcium sulfite solids
produced in the regeneration reaction. The centrifugate, as
previously stated, is recycled to the stripper to allow
recovery of acid values.
The oxidizer developed by Kawaski/Kureha is considered
to be proprietary. Oxidation air is supplied at 5 psig.
Repulped calcium sulfite slurry from regeneration, and
gypsum from the sulf ate-removal step are pumped to the
oxidizer. Sulfuric acid is added to adjust pH in the re-
action to the right conditions for the oxidation reaction.
Calcium sulfite is oxidized to gypsum by reaction with
atmospheric oxygen according to the overall reaction:
CaS03-l/2H20 + 1/202 + 3/2H24"2H20 (20)
The reaction mechanism may be acidic dissolution of CaSO.-,0
1/2H20, followed by oxidation of the bisulfite to sulfate
3-174
-------
which regenerates the acid value to dissolve additional
CaSO_•1/2H20. Gypsum product is separated from oxidizer
effluent and washed in centrifuges. Spent off-gas from the
oxidizer is recycled through the scrubber to remove S07 that
was released by the acidic oxidation reaction.
Typical conditions and performance data for the Tohoku
Electric FGD system are given below:
Flue Gas
Source: oil fired boiler
Fuel: 1.2 to 1.5 percent sulfur oil
Oxygen content: 3 percent
Flow rate: 420,000 m /hr (247,000 scfm)
S02 inlet concentration: 600 to 800 ppm
Particulate inlet: 2 to 10 mg/m3 (0.0009 to 0.0044 gr/scf)
Reheat: direct fire with 1.2 percent sulfur oil to 140°C
(285°F)
Absorber Systern
Absorber type: grid-packed,
SC>2 removal: 98 percent
SC>2 exit concentration: <10
L/G ratio: 1.6 1/m3 (
two packing stages
ppm
Scrubber pressure drop;
Scrubber liquor pH:
200
6.9 to
6.3 to
Recycle liquor composition:
Oxidation rate: 5 percent
12 gal/1000 ft3, bottom stage
mm H20 (7.9 in. H2O
7„6 inlet to top stage
6.5 bottom stage recycle
5 to 10 percent Na2S04
5 to 8 percent NaHS04
6 to 9 percent Na2SO3
Regeneration
Limestone stoichiometry: 100 percent of S02 inlet
Sodium makeup rate: 0.018 moles Na/mole S02 absorbed
Reactor system residence time: 3 hr (total, five reactors]
Reactor percentage solids: 5 percent (essentially all
CaS03«l/2H20)
Reactor filter cake solids: 40 percent
Regenerated liquor: 6,9 to 7.6 pH
20 percent total dissolved solids
<10 ppm Ca
Limestone particle size: lOym
3-175
-------
Solids Production (Gypsum)
Water content: 5 percent
Sodium content: <300 ppm on dry bases
Particle size: 30 x 200um
The system is said to be a closed-loop system, in that
there is no intentional water purge; however, based on the
rate of water addition to the system, and a comparison of
the rate of sodium makeup with the sodium content of the
product, it appears that there is some sodium loss from the
system other than the amount lost in the product gypsum.
The overall system is fairly complex when compared with some
of the systems under development in the United States. The
reasons for the additional complexity appear to be related
to the following: a) the system produces a relatively pure,
salable product, b) it is designed to achieve extremely high
S0~ and particulate removal, and c) limestone rather than
lime is used as the source of calcium. Apparently, con-
siderable effort has been devoted to minimizing the amount
of reagent used; i.e., sulfuric acid, limestone, and sodium.
The gypsum product has been used in cement manufacture;
however, it also meets the more stringent specifications for
wallboard production.
The very low oxidation (5%) in the absorption system is
attributed by Kawasaki/Kureha to "proprietary know-how."
3.3=3.2 Showa Denko KK/Ebara - Showa Denko KK and Ebara
Manufacturing Co., Ltd., Japan, jointly developed a con-
centrated double alkali process using limestone as the
calcium source, and producing a salable gypsum by-product.
The system is currently operating on 11 industrial oil-fired
boilers and a sulfuric acid plant.
3-176
-------
Figure 3-50 is a schematic representation of the Showa
Denko/Ebara process. Basically it is similar to the Kawasaki/
Kureha system. The full-scale unit (150 MW) uses four
inverted "vertical cone" scrubbers (similar to the Zurn
type). Liquor is entrained at the bottom of a conical draft
tube (truncated apex) by the entering flue gas and the
mixture passes up the tube. At the top of the tube (conical
base) there is a disengaging section, where liquor separates
and falls back to the bottom of the scrubber.
In the reactor system, limestone is reacted with spent
scrubber effluent to neutralize the NaHSO producing Na~SO
for recycle to the absorber, and CaSO •'L/2H~O, which is sent
to an air oxidizer to be converted to gypsum by air oxida-
tion (equation 28). A slipstream of spent scrubber liquor
is taken to a sulfate treatment section, which converts
Na_SO. to gypsum and NaHSO-. by reacting it with sulfuric
acid and calcium sulfite. The gypsum, which contains some
unconverted calcium sulfite, is combined with the calcium
sulfite produced in the limestone neutralization reaction
and pumped to the air oxidizer system. Sulfuric acid is
added to adjust the pH in the oxidizer system so that
optimum conditions can be achieved for calcium sulfite
oxidation.
The system operates in an open loop, since the amount
of water required to wash the gypsum product to below 300-
ppm sodium content is greater than water losses through
evaporation, water of hydration, and moisture in the final
product. The sodium content specification for wallboard
manufacture is less than 300 ppm. The corresponding specifi-
cation for gypsum used in cement manufacture is 900 ppm.
General conditions and performance data for the full-
scale system are as follows:
3-177
-------
CO
WASTE
GAS--
LiMESTOME
Figure 3-50. Showa Denko sodium-limestone process.
67
-------
Flue Gas
Source: oil-fired 150-MW boiler
Fuel: 2.5 to 3.0 percent sulfur
Particulate inlet concentration: 1.8 g/m3 (0.8 gr/scf)
SC>2 inlet concentration: 1200 to 1500 ppm
Reheat: direct-fired with 1 percent sulfur oil to
141 to 149°C (285 to 300°F)
Absorber System
Absorber type: four vertical cone scrubbers
SC>2 exit concentration: 60 to 100 ppm after reheat
with 1 percent sulfur oil
Particulate exit concentration: 0.057 to 0.80 g/m3
. (0.025 to 0.035"gr/scf)
L/G ratio: 1 1/m (7.5 gal/1000 ft3)
Scrubber liquor pH: 6.3 exit
Oxidation rate: <10 percent
Regeneration Chemical Usage
Limestone utilization: 96 percent
Sulfuric acid usage: 0.2 moles H-SO^mole SC>2 absorbed
Sodium usage: 0.10 to 0.12 moles of NaOH/mole SO,, absorbed
Oxidizer
Suspended solids content: 15 percent
Gypsum By-Product
Na content: <300 ppm
Water content: 7 to 8 percent
3.3.3.3 Performance of Double Alk.ali Systems in Japan -
The processes offered by Kureha-Kawasaki and Showa-Denko are
very effective for removal of SO At the Chiba Plant, the
Showa-Denko process has demonstrated 95-percent SO removal
68
on a flue gas stream containing 1200 to 1500 ppm S02-
At the Shinsendai plant, the Kureha-Kawasaki process has
demonstrated S09 removals of over 97 percent on a flue gas
stream containing 600 to 800 ppm.69 NO detailed information
3-179
-------
is available on reliability, but it is reported that it
approaches 100 percent.
3.3.4 Engineering Design Parameters
Sulfur dioxide efficiencies of over 90 percent can be
achieved with sodium-based double alkali FGD system on coal-
fired boilers firing high- or low-sulfur coal. The key
operating parameters may vary depending on the mode of
operation chosen (i.e. dilute vs. concentrated). In the
United States, only two basic scrubbing strategies have been
investigated: dilute active sodium with lime regeneration,
and concentrated active sodium with lime regeneration. Of
these two, the concentrated mode has received the major
emphasis in research and development on coal-fired boiler
flue gas applications and is the mode selected for full-
scale operation at the three utility boiler applications
described in Section 3.3.2.
Because of the high solubility of sodium alkalis and
their sulfur-containing reaction products, problems asso-
ciated with plugging, scaling, and limited removal capa-
bility are minimized. In regard to SO^-removal capability
and scale formation, the scrubbing solution pH range pro-
vides the key to optimal system efficiency. A pH control
value of 6.5 in the scrubbing effluent allows the achieve-
ment of maximum S0~ removal efficiency without the occur-
rence of chemical-related problems. If the scrubbing solu-
tion pH value exceeds 7, CO~ absorption would become sig-
nificant and could lead to calcium carbonate scaling. A
scrubbing solution pH level below 6 is likewise avoided,
because the vapor pressure of sulfur dioxide increases
dramatically for concentrated active alkali systems and can
lead to equilibrium-limited scrubbing conditions where
3-180
-------
outlet concentrations at or below 200 ppm are desired. This
chemical phenomenon is demonstrated in Figure 3-51, where
the sulfur dioxide vapor pressure is plotted as a function
of pH for a solution temperature of 54°C (130°F), a typical
saturation temperature for a boiler flue gas stream. For a
concentrated active alkali system, the vapor pressure of S0~
as a function of pH is significant. The SO., vapor pressure
at a pH of 6 is approximately 100 ppm and increases rapidly
as the pH drops.
For these reasons, concentrated systems tend to operate
in the pH range of 6 to 7, and preferably at a set point of
6.5, where the sulfite-bisulfite system is highly buffered.
The highly buffered system is preferred, because the scrub-
bing solution is able to adapt to rapid changes in flue gas
inlet conditions. This is in contrast to scrubbing systems
using a pH of 6 or less; these absorbents can be very
sensitive to rapid changes in flue gas inlet concentrations.
Thus, based upon the low S0~ back pressure of the
reacted species in solution at the optimum pH control value
of 6.5, S0~ concentration levels in the scrubber discharge
can be maintained as low as 10 to 20 ppm.
The principal area of concern in regard to the chemical
characteristics of double alkali processes is their ten-
dency to form inactive species through reaction with oxygen
or chlorides. Oxidation in a double alkali system refers to
the conversion of active sulfite and bisulfite species to
inactive sulfate. This can occur either in the gas (S02 +
1/2 02 -* 803) or aqueous phase (S0~ + 1/2 O2 -> S0~; HSO~ +
1/2 02 -> SOJ + H+) .
The formation of inactive sulfate in the scrubbing
solution can have a deleterious effect on process chemistry
3-181
-------
ex
Q.
oo
LU
QC
O
Q.
X
O
n
QC
oo
300
200
100
90
80
70
60
50
40
30
20
10
9
8
7
6
LABORATORY DATA
(SOLUTION IONIC
STRENGTH)^ 5.5
(TOTAL OXIDIZABLE
SULFUR)= 0.6
5.5
6.5
PH
Figure 3-51. Concentrated active alkali system:
sulfur dioxide vapor pressure vs. pH @ 54°C
(130°F).71
3-182
-------
because of a) its inability to react with S02, b) a reduced
ability to be regenerated into an active alkali form by
reaction with lime, and c) the potential to form gypsum
scale under certain conditions. Oxidation may occur in any
segment of the scrubber plant; it is generally considered to
be a function of the rate of dissolution of oxygen; scrub-
bing solution pH; trace contaminants (catalytic effects of
collected fly ash; lime and soda ash impurities); solution
ionic strength; absorber design; S0_ concentration of the
flue gas; O concentration of the flue gas; and flue gas
temperature. Evidence indicates, however, that approximately
90 percent of total system oxidation occurs in the scrubber
circuit. Sulfates along with sulfites are removed from
the spent scrubbing solution by reaction with lime in the
regeneration/precipitation reactor. In concentrated mode
systems, the unsaturated calcium sulfate mode of operation
is maintained by coprecipitation of calcium sulfite and
sulfate from slurry liquid that is saturated in calcium
sulfite and unsaturated in calcium sulfate and favored by
low sulfite-to-sulfate oxidation tendencies. Calcium sul-
fate can be coprecipitated with the principal sulfite pre-
cipitate up to a sulfate/sulfite mole ratio of approximately
0.20. Thus, unsaturated conditions can be maintained only
at relatively low sulfite-oxidation rates, achievable in
high SO-/0,., ratio emission sources. The minimum inlet gas
conditions that must be met in order to operate a concen-
trated mode system reliably and economically, are about 400
to 500 ppm S02 and 5 percent 0^. For flue gas applications
that equal or fall below these inlet gas requirements (such
as predominate in some industrial boilers, bark boilers, and
paper mills), dilute mode double alkali systems are tech-
3-183
-------
nically feasible, owing to their ability to withstand high
oxidation rates (up to 100 percent oxidation).
Inactive sodium chloride may also be formed in the
system as a result of absorption of hydrogen chloride by
the scrubbing liquor and the subsequent formation of hydro-
chloric acid (HC1) and NaCl. Where filter cake washing is
used to recover soluble sodium, higher NaCl concentrations
can build up in the liquor recycle and reduce SO,, removal
efficiency, brought about by loss of active sodium. Purge
streams or prior CHI removal must be practiced to minimize
this effect.
The double alkali system must be designed to limit
dissolution of oxygen, so that oxidation of sialfite/bisul-
fite to sulfate can be minimized. The chloride level in the
scrubber circuit must also be controlled by purging, in
order to hold down sodium losses via soluble sodium chloride;
to limit corrosion attack on scrubber internals; and to
permit operation in a mode that is sulfate-unsaturated.
Sophisticated solids dewatering is also necessary to pre-
vent soluble sodium losses, thereby limiting potential
environmental problems associated with waste disposal and
permitting closed-loop operation. Finally, instrumentation
for pH control of the absorber and regeneration systems must
be reliable to insure adequate SO2 removal in a scale-free
environment.
As stated previously, 90 percent of the oxidation of
sulfites to sulfates occurs in the scrubber circuit. Since
oxidation is a strong function of the rate of oxygen absorp-
tion, the number of contact stages in the scrubber module
should be limited to two (e.g., the FMC, CEA/ADL and En-
virotech full-scale electric utility applications all
3-184
-------
incorporate two-stage modules). In addition, the regen-
eration reactor must also employ an optimum residence time
to minimize the oxidation occurring in the reactor.
The dry removal of fly ash in a high-efficiency ESP
upstream from the scrubber plant also helps to limit sulfite
oxidation caused by catalyzing metal compounds.
The presence of chloride in the coal at or above 0.04
percent by weight can result in increased scaling in the
scrubber loop, attributable to sodium sulfate insolubility,
which itself is caused by high sodium chloride concentrations
(1.5 to 2.0 moles). In addition, all hydrogen chloride
absorbed in the scrubber circuit ties up the sodium as
dissolved sodium chloride, which will be completely lost as
inactive sodium. This increases the potential for pollu-
tion problems in surrounding ground and surface waters,
and also results in a higher sodium consumption. These
considerations, plus the possibility of chloride corrosion,
may force the inclusion of a separate water loop scrubber
before the SO., absorber, in order to remove most of the
chlorides from the flue gas.
Solids separation by means of a rotary drum filter,
belt filter, extraction filter, or centrifuge, in addition
to one or more thickeners, is necessary in order to operate
a closed-loop system, to minimize potential water pollution
problems, and to reduce soda ash consumption.
As stated previously, pH control in the S02 absorption
step determines the amount of SC>2 removed and precludes the
formation of carbonate scale. Also, pH control in the
regeneration reactor determines the composition of the
regenerated solution and precipitated solids. A pH of 8.5
is the titrametric end-point for sodium bisulfite. It is
3-185
-------
possible to precipitate additional sulfur and to regenerate
active sodium by controlling at more alkaline pH levels (11
through 12„5). The pH is more difficult to control in the
more alkaline regions, however, because of poorer lime
system responsiveness to solution pH values, resulting in
higher stoichiometric lime requirements, lower lime utiliza-
tion values, poorer filter cake quality, and a potential for
calcium ion supersaturation in the scrubber circuit. The pH
control loops must be designed and situated to respond
reliably and quickly if optimum system operations are to be
ensured.
In summary, a double alkali system designed to effect
90-percent SC>2 removal or greater on a high- or low-sulfur
coal-fired utility boiler would have the following design
parameters:
0 Operation in the concentrated alkali mode (e.g.
>0.15M active alkali concentration in the scrubber
feed).
Use of a venturi scrubber with a separate water
loop for control of particulates and chlorides for
high-chloride coal applications (i.e. >0.04% Cl by
weight in the coal).
0 Use of a two-stage tray or packed tower absorber
with an L/G ratio of from 1.3 to 2.7 1/m3 (10 to
20 gal/1000 acfm), a scrubbing liquor pH of 6.0 to
7.0, and a pressure drop of 15 to 30 cm (6 to 12
in.) H20.
0 Regeneration of the active sodium through addition
of lime or limestone to a reactor tank having a
residence time of 10 min for lime and more than 30
min for limestone. Utilization rates of 90
percent or greater should be obtained for lime and
75 to 85 percent for limestone.
0 Sodium makeup by addition of sodium carbonate with
its associated softening effect. Sodium makeup
should be equivalent to 0.05 moles of sodium car-
bonate per mole of S02 removed.
3-186
-------
0 The solids removal section should consist of
clarifiers and vacuum filters to ensure high
solids content in the final waste cake. Water
washing of the waste cake should be used to re-
cover soluble sodium present in the cake.
0 Operation should be in a closed-loop mode with
respect to the water balance by use of recycle
water.
3.3.5 Process Operability
Although no full-scale double alkali FGD systems are in
operation on coal-fired boilers, it is possible to predict
operability of the systems, based on experience with smaller
units on coal-fired boilers. As shown in Section 3.3.2, the
operability of double alkali FGD systems on coal-fired
utility boilers and prototype utility installations has been
improved steadily; it is now 90 percent and above. Most
operability problems were due to design-related equipment
shortcomings in these prototype installations. It should be
pointed out that most installations did not have spare
equipment; this would, however, be included in full-scale
utility systems.
Equipment-related problems were solved at each of the
double alkali installations and the experience gained will
benefit later installations. The vendors of double alkali
systems have developed confidence in their reliability: as
evidenced by guarantees of 90-percent availability for the
first year of operation and 100 percent for the life of the
plant (based on a boiler operating rate of 70 percent for
some of the new, full-scale utility applications). The
systems are all guaranteed to achieve 85- to 95-percent SC>2
removal efficiency on high-sulfur coal applications. No new
low-sulfur coal applications are planned, but similar
guarantees would be expected for such systems.
3-187
-------
Corrosion, erosion, and scaling problems have not been
important factors at double alkali FGD installations. Full-
scale versions of these systems are not expected to experi-
ence these problems either.
The double alkali system has demonstrated the ability
to perform well under fluctuating SO,., inlet concentrations.
At the Scholz plant, the design inlet SO,., concentration was
1800 ppm. At inlet concentrations varying from 800 to 1700
77
ppm, removal efficiencies were above 90 percent.
Sludge produced in the double alkali FGD process must
be disposed of safely. FMC Corporation has studied the
properties of the sludge produced by their double alkali
process. The study has revealed that the waste cake was
highly impermeable; the only potential problem was the
7 8
leaching of soluble salts contained in the cake. This can
be overcome by proper landfill management and/or the use of
a liner in the disposal area. The waste cake exhibits good
mechanical properties, making it unnecessary to use fixation
to obtain mechanical stability.
3-188
-------
REFERENCES FOR SECTION 3.3
1. Draemel, D. An EPA Overview of Sodium-based Double
Alkali Processes. Part 1. A view of the Process
Chemistry of Identifiable and Attractive Schemes, U.S.
Environmental Protection Agency, Research Triangle
Park, NC, May, 1973. p. 5.
2. Ibid. pp. 5 and 8.
3. Kaplan, N. An Overview of Double Alkali Processes for
Flue Gas Desulfurization. For presentation at the EPA
Symposium on Flue Gas Desulfurization. Atlanta, Ga.
November 4-7, 1974. pp. 453-454.
4. Op. cit. No. 1. p. 7.
5. LaMantia, C.R. and R.R. Lunt, Arthur D. Little, Inc.,
and I. S. Shah, Combustion Equipment Associates, Inc.,
Dual Alkali Process for Sulfur Dioxide Removal, Chemical
Engineering Process, Vol. 70, No. 6, June, 1974. p.
67.
6. Kaplan, N. Introduction to Double Alkali Flue Gas
Desulfurization Technology- For presentation at the
EPA Symposium on Flue Gas Desulfurization. New Orleans,
La. March 8-11, 1976. p. 3.
7. Operating Experience with the Zurn Double Alkali Flue
Gas Desulfurization Process. Zurn Industries, Inc.
For presentation at the EPA Symposium on Flue Gas
Desulfurization. New Orleans, La. March 8-11, 1976.
p. 2.
8. Op. cit. No. 1. p. 7.
9. Ibid.
10. Ibid. p. 15.
11. Ibid. p. 17.
3-189
-------
12. Kaplan, N. An Overview of Sodium-based Double Alkali
Processes. Part II. Status of Technology and Descrip-
tion of Attractive Schemes, U.S. Environmental Protec-
tion Agency, for Presentation at the EPA Flue Gas
Desulfurization Symposium, New Orleans, La. May 14-17,
1973. pp. 11, 13, 14.
13. Op. cit. No. 1. p. 18.
14. Op. cit. No. 12. p. 11, 13, 14.
15. Ibid. pp. 14-16.
16. Op. cit. No. 1. p. 18.
17. Ibid. pp. 16-17.
18. Summary Report - Flue Gas Desulfurization System,
Prepared by PEDCo Environmental, Inc., for U.S. Environ-
mental Protection Agency. Contract No. 68-01-4147,
Task No. 3. May-June, 1977. pp. 20, 241.
19. Ibid. pp. 1, 30, 242.
20. Ibid. pp. 5, 15.
21. Holhut, W., and T. Nelson, Central Illinois Public
Service Co.; W. Edison, Buell Envirotech; and J.
Wilhelm, Eimco BSP Division Envirotech. Zero-Effluent
Throwaway S02 System Design for High-Chloride High-
Sulfur Coal. Presented to American Power Conference,
38th Annual Meeting. April 22, 1976. pp. 9-10.
22. Dingo, T., and E. Piasecki, General Motors Corp. An
Initial Operating Experience with a Dual-Alkali S02
Removal System. For presentation at the Environmental
Protection Agency Symposium on Flue Gas Desulfurization
Atlanta, Ga. November 4-7, 1974. pp. 523-525.
23. Interes, E., A. D. Little, Inc. Evaluation of the
General Motors Double Alkali S02 Control System. EP/i
600/7-77-005. U.S. Environmental Protection Agency,
Research Triangle Park, NC. January 1977. pp. 1-34.
24. Ibid.
25. Ibid. pp. 27-34.
3-190
-------
26. Ibid. p. 1.
27. Ibid. p. 5.
28. Op. cit. No. 7. pp. 1-4.
29- System Allows Burning of High-Sulfur Coal by Removing
SOo from Flue Gases. Chemical Processing, June 1976.
pp. 142-143.
30. Operating Experience with a Large Industrial Double
Alkali Flue Gas Desulfurization System. Zurn Industries,
Inc., Presented at the 38th Annual Meeting of the
American Power Conference. pp. 1-3.
31. Ibid. pp. 2,6.
32. Ibid. p. 2.
33. Ibid. pp. 2,3.
34. Ibid. pp. 2,3.
35. Op. cit. No. 7. pp. 2,3.
36. PEDCo In-House data obtained under U.S. Environmental
Protection Agency Contract No. 68-02-2601, Task No. 4.
To be published in "Summary Report on S02 Control
Systems for Nonutility Combustion and Process Sources."
1977.
37. Legatski, L. K., et. al. FMC Corporation. The FMC
Concentrated Double Alkali Process. Prepared for
presentation at the U.S. Environmental Protection
Agency Flue Gas Desulfurization Symposium, New Orleans,
La. March 8-11, 1976. p. 23.
38. Ibid. p. 19.
39- Zaharchuk, R. and L. K. Legatski. S02 Scrubber Passes
Test at Firestone. Pollution Engineering. April 1977.
pp. 50-53.
40. Ibid. p. 51.
41. Ibid.
42. Ibid. p. 52.
3-191
-------
43.
44.
45.
46.
47.
48.
49.
Ibid. pp
Ibid. p.
Op. cit.
Ibid. pp
Ibid. p.
Op. cit.
Kaplan, N
. 50, 51.
53.
No. 37, p. 19.
. 19-23.
19-
No. 3, pp. 476-479.
An EPA Overview of Sodium-based Double
Alkali Processes. Part II. Status of Technology and
Description of Attractive Schemes. For presentation at
the EPA Flue Gas Desulfurization Symposium. New
50.
51.
52.
53.
54.
55.
56.
57.
Orleans,
Op. cit.
Ibid. p.
Ibid. p.
Ibid. p.
Ibid. p.
Ibid. p.
Ibid.
LaMantia,
La. May, 1973. pp. 21-23.
No. 3, pp. 479, 480.
479.
483.
481.
482.
483.
C. R. , et. al. Operating Experience - CEA/ADL
Dual Alkali Prototype System at Gulf Power/Southern
Services,
Inc. For presentation at the EPA Flue Gas
Desulfurization Symposium, New Orleans, La. March,
58.
59.
60.
61.
62.
1976. pp
Ibid. p.
ibid. p.
Ibid. pp
Ibid. p.
Ando , J .
. 7-13.
9.
29-
. 28-38.
2.
SO? Abatement for Stationary Sources in
Japan. PEDCo Environmental, Inc. Cincinnati, Ohio.
June, 1976. Table 3-6.
3-192
-------
63. Op. cit. No. 3. pp. 501-505.
64. Ibid. p. 502.
65. Ibid. p. 504.
66. Ibid. pp. 506-508.
67. Ibid. p. 505.
68. Ando, J., and G. Isaacs. SC>2 Abatement for Stationary
Sources in Japan. EPA - 600/2-76-013a. January, 1976.
p. 4-4.
69. Ibid. p. 4-13.
70. Op. cit. No. 3. pp. 505, 508.
71. Op. cit. No. 41. p. 9.
72. Op. cit. No. 61. p. 31.
73. Ibid. p. 33.
74. Op. cit. No. 3. p. 467.
75. Bloss, H. E., et. al. The Buell Double Alkali S02
Control Process. For presentation at the EPA Symposium
on Flue Gas Desulfurization. New Orleans, La. March,
1976. p. 4.
76. Op. cit. No. 61, p. 19.
77. Ibid. pp. 28-30.
78. Op. cit. No. 41. pp. 12-13.
3-193
-------
3.4 MAGNESIUM OXIDE SYSTEMS
This scrubbing system uses a slurry of magnesium oxide
in water to absorb SO,,. The reaction product, magnesium
sulfite, is then dried, and the magnesium oxide is regen-
erated and recirculated. Sulfur dioxide is recovered
for conversion to sulfuric acid or reduction to sulfur.
3.4.1 Process Description
In this system, fly ash and other particulates are
removed from the flue gas stream prior to the scrubbing
operation, either by electrostatic precipitators (ESP's)
or by primary (wet) scrubbers. The fly ash is removed to
reduce solids loading in the absorber and to reduce impurities
in the regenerated magnesium oxide.
The flue gas then enters an absorber for SO removal as
X
shown in Figure 3-52. In this system, reactive magnesium
oxide is slaked (reacted) with water to form magnesium
hydroxide [Mg(OH),,] slurry, which becomes the absorbent for
the removal of SO . Sulfur dioxide and the small amount of
J\.
SO- in exhaust gases from the boiler react with the magne-
sium hydroxide slurry to form magnesium sulfite (MgSO.,) and
magnesium sulfate (MgSO.) respectively. Some of the SO..,
reacts with the magnesium sulfite in the presence of water
to form magnesium bisulfite [Mg(HSO.,)2]. An excess of
magnesium oxide (2.5 - 5%) is maintained in the slurry
beyond that required to react with all the SO in the flue
j\.
gas. The excess magnesium oxide or hydroxide reacts with
the magnesium bisulfite, forming more magnesium sulfite.
Additional magnesium sulfate is formed by the oxidation of
magnesium sulfite.
Sulfur oxides absorption is represented by the fol-
1 2
lowing reactions: '
3-194
-------
U>
I
Figure 3-52. Magnesium oxide slurry FGD system.
-------
Mg(OH)2 + 5H20 + S02 -> MgS03 • 6H20 4- (1)
Mg(OH)2 + 2H20 + S02 -> MgS03 • 3H20 4- (2)
Mg(OH)2 + 6H20 + S03 + MgS04 • 7H2O 4- (3)
MgS03 • 6H20 ->• Mg(HS03)2 + 5H2O (4)
MgS03 • 3H20 -> Mg(HS03)2 + 2H2O (5)
Mg(HS03)2 + MgO + HH20 -> 2 MgS03 • 6H20 4- (6)
Mg(HS03)2 + MgO + 5H20 -> 2MgS03 • 3H20 (7)
2 MgS07 + 09 -> 2MgSO. (8)
O ^ ^*
The aqueous slurry absorbent (scrubbing medium) con-
tains both hydrated crystals of MgO, MgSO3 , and MgS04, as
well as a solution that is saturated with each of these
components. A portion of the scrubber stream, theoretically
the equivalent (expressed as magnesium sulfite) of the
sulfur oxide content being introduced, is constantly di-
verted to a clarifier/thickener for concentration, and the
concentrated slurry is then fed to a continuous centrifuge.
Alternatively, the scrubber slurry stream may be fed di-
rectly to the continuous centrifuge. A "wet cake" con-
taining crystals of MgSO., • 6R~0, MgSO., • 3HpO, MgSO. •
7H20, and unreacted MgO is produced.
The liquor removed from the crystals may be returned to
the main recirculating slurry stream, or may be used to
slake/react with fresh or regenerated magnesium oxide (MgO) ,
which is then added to the recirculating slurry stream as
makeup.
The "wet cake" is conveyed to a direct-contact dryer
where free and chemically bound water are removed. Dryer
temperatures typically are in the range of 176-232°C (350-
450°F) .
The chemical reactions occur in the dryer as follows:
3-196
-------
MgS03 • 6H20 -> MgSO., • 3H 0 + 3H O + (9)
MgSO • 3H 0 + MgSO + 3H O t (10)
J ^ A J ^
MgSO., • 6H90 + MgSO., + 6H90 t (11)
A
MgS04 • 7H20 -> MgS04 + 7H20 t (12)
The dry magnesium sulfite and magnesium sulfate crystals are
calcined at 800 to 1000°C (1472-1832°F) in the presence of
coke or a reducing atmosphere to regenerate the MgO and
3 4
release SO9 . The calciner reactions are as follows: '
MgS03 ^ MgO + SO2 t (13)
2C + 02 -»• 2CO t (14)
MgS04 + CO -> MgO + C02 f + S02 t (15)
Regenerated magnesium oxide is moved to storage for
reuse in the scrubber slurry system.
The SO- from the calciner is usually used as the feed
stock for a sulfuric acid unit. However, depending on
demands for sulfuric acid, it could be sold as liquified
S09, liquified SO.,, or elemental sulfur; or used to enrich
concentrated (98%) sulfuric acid to various grades of oleum.
Table 3-31 represents a material balance for this
system, as applied to a 500-MW coal-fired plant.
3.4.2 Magnesium Oxide Scrubber Units - U.S.
Three full-size units have been operated in the United
States. All used fuel with 2 to 2.5 percent sulfur content.
They are as follows:
Utility name: Boston Edison
Unit name: Mystic Station, Unit No. 6
Unit location: Everett, Massachusetts
Unit size: 155 MW
Fuel properties: No. 6 fuel oil, 2.5% sulfur
Start-up date: April 1972
Note: Terminated June 1974
3-197
-------
Table 3-31. MgO FGD SYSTEM MATERIAL BALANCE
CO
Stream No.
Description
Rate, Ibs/hr
scfm
gpm
Particulates, Ibs/hr
Temperature, °F
Specific gravity
Stream No.
Description
Rate, Ibs/hr
scfm
gpm
Particulates, Ibs/hr
Temperature, °F
Specific gravity
Stream No.
Description
Rate, Ibs/hr
scfm
gpm
Particulates, Ibs/hr
Temperature, °F
Specific gravity
1
Coal to the
boiler
375M
6
Gas to the
ESP
4,848M
1.040M
39. 3M
310
11
Makeup
water
28. 4M
56.8
2
Combustion air
to air heater
4,518M
984M
110°F
7
Gas to the
SC>2 scrubber
4,848M
1,040M
393
310
12
Total MgO
feed
16. 9M
3
Combustion air
to boiler
4,075M
888M
535°F
8
Gas to the
reheater
5,156M
1,144M
127
13
Liquor to
slaker
129M
246
4
Gas to the
economizer
4,404H
94 3M
39. 3M
890°F
9
Gas to the
stack
5,156M
1.144M
174
14
MgO slurry
to SO2
absorber
146M
263
1.11
5
Gas to
air heater
4,404M
943M
39. 3M
705°F
10
S02 absorber
slurry to tank
14,217M
26,085
127
15
Slurry to
thickener
1,088M
1,996
-------
Table 3-31 (continued). MgO FGD SYSTEM MATERIAL BALANCE
U)
i
Stream No.
Description
Rate, Ibs/hr
scfm
gpm
Particulates , Ibs/hr
Temperature, °F
Specific gravity
Stream No.
Description
Rate, Ibs/hr
scfm
gpm
Particulates, Ibs/hr
Temperature, °F
Specific gravity
Stream No.
Description
Rate, Ibs/hr
scfm
gpm
Particulates, Ibs/hr
Temperature, °F
Specific gravity
16
Thickener
overflow
884M
1666
1.06
21
Makeup
water
23. 4M
46.8
26
Dryer gas
to absorber
90. 1M
23. 2M
159
17
Recycle
slurry
14 ,013M
25,745
22
Recycle to
absorber
432M
802
27
Dryer gas
recycle to dryer
45.8M
11. 8M
318
18
Thickener
underflow
240M
340
19
Liquor to
centrate tank
137M
260
j
1.2
23
Feed to
dryer
73M
28
Dryer product
to storage
33. 4M
400
1.06
24
Dryer gas
to cyclone
136M
35M
5960
400
29
0.1 to fire
dryer
3M
20
Wet cake from
centrifuge
67M
25
Cyclone dust
to dryer
5.5M
30
Air to
fire dryer
60. 5M
13.2M
-------
Table 3-31 (continued). MgO FGD SYSTEM MATERIAL BALANCE
Stream No.
Description
Rate, Ibs./hr
scf m
gpm
Par ti culates , Ibs/hr
Temperature, °F
Specific gravity
Stream No.
Desci j f.t ion
Rate, jbs/hr
scf m
gpm
Part i c u 1 a t e s , i L s ,/' h r
Temperature, °F
Specific gravity
31
MgSC>3 feed
to calciner
210
32
Coke feed
to calciner
40. 1M
I
36
Calciner qas
to cyclone
37
Calciner gas
33
Feed to
calciner
3.1M
34
Oil to
calciner
3.1M
38
Cyclone dust
to bag filter! to calciner
35
Air to
calciner
44 .7M
9.7M
i
39
Baq filter dust
to calciner
! 1 - - - -
69. 1M
12.6M
1910
61. 1M
12.6M
191
1719
1600 ; '
187
Table 3-31 Material Balance Notes
40
SO 2 to
acid plant
134M
26. 9M
4
1. Calculation* based on th« following i
a. 105 percent stoichiometric magnesia
b. 9,000 Btu/kWh for th* unit
c, 12,000 Btu/lb of coal
d, ?.5 percent «ulfur coal (dry bails)
«. 92 p«rc»nt of th« sulfur to th* coal avolvea
as SO]
f. 14 percent ash in tha coal (a« fired ba«i«)
g. 75 percent of the ash in the coal avolvea am
fly £»h
h. 99.0 parcenc reooval of pa^ticulat.a» by the
ESP
SO2
-------
Utility name: Potomac Electric and Power
Unit name: Dickerson No. 3
Unit location: Dickerson, Maryland
Unit size: 95 MW
Fuel properties: Coal, 2.0% sulfur
Start-up date: September 1973
Note: Terminated August 1975
Utility name: Philadelphia Electric
Unit name: Eddystone No. 1A
Unit location: Eddystone, Pennsylvania
Unit size: 120 MW
Fuel properties: Coal, 2.5% sulfur
Start-up date: September, 1975
Suspended operation in January 1976 pending relocation
of calciner after acid plant shut down. Restarted in
June 1977.
As discussed in the following sections dealing with
individual units, SO,., removal efficiency at all three loca-
tions has been over 90 percent. Both PEPCo-Dickerson and
PECo-Eddystone have demonstrated over 99 percent particulate
removal efficiency.
In general, however, all three units have experienced
mechanical, material-related and product-related corrosion
and handling problems. These have limited on-stream opera-
tion to between 27 and 80 percent of the total operating
time of the generating units. The on-stream time was deter-
mined after an initial "shakedown" operating period. Boston
Edison's Mystic Station Unit operated 80 percent or more of
the time for three of the last four months of the test
period.
3.4.2.1 Boston Edison - Mystic Station, Unit No. 6
System description - The flue gases from the boiler did
not need to pass through any particulate removal equipment
because oil was burned. To remove SO,,, flue gas was con-
tacted cocurrently with an aqueous magnesium oxide slurry in
3-201
-------
a single-stage, venturi scrubber. A side stream from the
scrubber slurry recirculating loop was constantly fed to a
continuous centrifuge, and the aqueous phase from the
centrifuge pumped back to the scrubber. The "wet cake" from
the centrifuge was conveyed to a countercurrent, direct-
fired dryer. The dry product was stored for later shipment
to an off-site sulfuric acid plant, at Essex Chemical,
Rumford, R.I., where the magnesium oxide was regenerated in
a rotary kiln, and the resultant S00-rich gas stream used
R f>
for sulfuric acid production as shown in Figure 3-53. '
Distinguishing Features - This unit was characterized
by:
0 Firing with No. 6 fuel oil during FGD operation.
0 Fly ash from oil combustion was not removed
before it reached the SO,, absorber.
0 The portion of the absorbing MgO slurry removed
from the recirculating loop was not thickened
prior to being fed into the centrifuge.
0 The dryer and the calciner were both direct fired
rotary kilns.
0 There was no flue gas reheat prior to discharge
through the stack.
Operating History - The average of the test runs for
particulate removal in the single-stage venturi SO- scrubber
was 57 percent as shown in Table 3-32. Resulting particu-
late emissions were low and averaged 36 kg/J (0.084 lbs/10
Btu).7
The scrubber did an excellent job of SO,., removal. Test
data indicate an average removal efficiency of over 91.6
percent. This was achieved at gas flow rates ranging from
12,036 m /min (425,000 acfm), which was the desiqn value, to
3-202
-------
Table 3-32. PARTICULATE REMOVAL - TEST RESULTS
BOSTON EDISON - MYSTIC STATION NO. 69
Test
NO.
1
2
3
4
Boiler
load,
MW
146
144
151
148
Inlet gas rate
m^/s
211.0
230.0
310.7
237.6
(acfm)
(446,953)
(486,991)
(658,207)
(503,233)
Particulates,
Ib/hr
In
380
232
399
151
Out
116
115
150
82
%
removal
69.5
50.4
62.4
45.7
Particulates
Kg/J
31.0
36.2
47.8
29. 3
(lb/106 Btu)
(0.072)
(0.084)
(0.111)
(0.068)
U)
I
M
O
U>
Table 3-33. SO2 REMOVAL - TEST RESULTS
BOSTON EDISON - MYSTIC STATION NO. 6
10
Test
No.
1
2
3
4
Boiler
load,
MW
146
144
151
148
Inlet gas rate,
mVs
211.0
230.0
310.7
237.6
(acfm)
(446,953)
(486,991
(658,207)
(503,233)
S02 in
ppm - vol.
926.1
1004.5
983.9
833. 3
SO2 out
ppm - vol.
71.1
89.0
63.3
86.6
% S02
removal
92.3
91. 1
93.6
89.6
SO2 out
Kg/J
59.0
93.9
95.9
114.7
(lb/10b Btu)
(0.125)
(0.199)
(0.201)
(0.243)
-------
SULFURIC ACID
CONTACT ACID
PLANT
981 GRADE
u>
I
t-o
o
PUMP
Figure 3-53. Mystic No. 6 FGD System and Essex Chemical Plant
Regeneration Facility.
-------
18,640 m /min (658,000 acfm), more than 54 percent in excess
of the design value (Table 3-33). Outlet SO- concentrations
averaged 82.7 kg/J CO. 192 lb/106 Btu).8
The operability of the No. 6 Unit at Mystic for the
entire test run is presented in Table 3-34. Operability is
defined as hours of FGD operation divided by hours of boiler
operation in a given period, expressed as a percentage. The
unit worked best during its last four months, when opera-
bility was around 80 percent. It would have been approxi-
mately 85 percent but for a two-week outage of the off-site
sulfuric acid plant. The FGD system had to shut down since
MgO could not be regenerated. From April 12 to May 10,
12
1974, the system achieved 100 percent operability.
A correlation of SO2 removal efficiency as a function
of inlet S02 concentration and pressure drop across the
scrubber based on actual test data is shown in Figures 3-54
and 3-55. The data show consistently high SO2 removal
efficiency for pressure drops over six inches of water.
Although these data were obtained at the Mystic Station, the
13 14
results should apply to other similar MgO systems. '
To maintain S02 outlet concentrations of less than 100
ppm, the pH of the recycle magnesium slurry stream leaving
the absorber should be held between 6.8 and 7.5.
If a regenerated magnesium oxide of maximum reactivity
is required, the calcining kiln temperature and the control
of the reducing atmosphere (through coke addition) are
critical. The bulk density (specific gravity) of magnesium
oxide is the best indicator of its activity. The desired
bulk density is from 272 to 384 k/m3 (17 to 24 lb/ft3).
Overburning the magnesium oxide produces unreactive peri-
clase with a density as high as 480 to 721 kg/m (30 to 45
3-205
-------
Table 3-34. OPEPABILITY OF MgO SYSTEM AT MYSTIC NO. 6
Month/yr
Apr. 72
to
May 73
Jun. 73
Jul. 73
Aug. 73
Sep. 73
Oct. 73
Nov. 73
Dec. 73
Jan. 74
Feb. 74
Mar. 74
Apr. 74
May 74
Jun. 74
Boiler
operation
(hr)
592
575
637
627
629
658
555
541
4C8
585
488
559
Test period completed
Scrubber
operation
(hr)
402
351
Boiler shutdow
243
377
162
86
152
138
353
471
280
288
Operability
%
17
68
61
T
38
60
26
13
28
25
87
81
57
80
Comments
The module operated intermittently because of mechanical diffi-
culties. Operation of the magnesium sulfite crystals dryer pre-
sented a major problem.
The longest period of continuous operation, 7.5 days, occurred
during June and July.
The boiler was down for the annual overhaul. System operability
decreased during the last quarter of the year because of heavy
erosion/corrosion in the liquor recirculation pumps and centrifuge.
Boiler-related problems caused frequent shutdowns in January and
February.
Two 7-day, continuous operation periods occurred during the month.
The 57-percent operability index value for the month was due to a
2-week outage of the acid plant rather than to FGD system failure.
Demonstration program was completed and FGD system was shut down.
EPA funding of MgSO.. calcination has expired. There are no
definite plans for restarting of this unit. Major problem areas
encountered during the operation of this prototype unit included
formation of trihydrate instead of hexahydrate sulfite crystal,
dust problems in the dryer, lack of stack gas reheat, which caused
condensation in the stack, louver damper problems, erosion of
pumps, piping, and centrifuge, and minor ancillary equipment
failures.
U)
I
t\J
o
o-i
-------
U)
i
NJ
O
100
90
80
-------
O
00
90
80
70
_
ef
>
o
UJ
oi
CM
60
50
40
5.08
INLET
INLET S02
INLET S02
1000 ppm
700 ppm
400 ppm
VENTURI ABSORBER
AT L/G = 40 gal/1000 acf
= 5.3 £/m3
10.16
15.24 20.32 25.40
30.48
35.56 40.64 45.72
PRESSURE DROP - GRAMS/SQUARE CENTIMETER
Figure 3-55. Effect of pressure clrop on SO2 removal for the Mystic venturi absorber.
II
-------
Ib/ft ).* Data from the test program show that formation of
magnesia with low bulk density and high reactivity is pro-
moted by lower temperatures (600-7QO°C; 1100-1300°F) and by
increasing the carbon (coke) content (1.5-2%) in the cal-
f * 19
ciner feed.
Operating problems and solutions - The major problem at
Mystic was the production of a high percentage of magnesium
sulfite trihydrate (MgSO., • 3H20) , rather than the expected
magnesium sulfite hexahydrate (MgSCX, • 6E~O) . The magnesium
sulfite dryer was designed for operation with the relatively
large hexahydrate crystal. Fine trihydrate crystals caused
excessive dusting, solids buildup, loss of drying ability,
20
and ultimately a number of major mechanical problems.
More than a year was required to accomplish the follow-
ing:21'22'23'24
1) Modify dryer operation (trial and error)
2) Modify dryer internal configuration
3) Install hammers to prevent solids buildup on dryer
walls
4) Change centrifuge wet cake feed-point
5) Install scalper and lump breaker to handle mag-
nesium sulfite lumps leaving the dryer
6) Install pneumatic system for conveying material
from the dryer off-gas cyclone dust collector to
the storage silo. Dust was initially mixed with
centrifuge wet cake; the mixture was "set up" in
screw conveying equipment.
7) Vent dryer off-gas from cyclone dust collector
back to the venturi scrubber inlet.
Erosion and corrosion of the caxbon steel slurry
recirculating pipes resulted in the adoption of rubber-lined
* A dead-burned, hardened form of magnesium oxide.
3-209
-------
pumps, valves, and piping throughout the SO,, scrubbing
25,26,27,28,29
system.
The centrifuge bowl plows, head plows, and conveyor
surfaces underwent rapid wear, causing several shutdowns.
Tests on stellite hard surfacing showed it could be used at
least a year without repair. Wash connections were added to
the centrifuge casing and an internal wash pipe was included
to reduce or eliminate solids deposition. '
Handling regenerated magnesium oxide presented prob-
lems, including: 1) formation of overburned magnesia
(periclase); 2) formation of oversized magnesia particles;
and 3) difficulty in slaking the magnesia. Closer control
of the coke (reducing atmosphere) and kiln temperature
eliminated the periclase problem. Installation of a "scalp-
er" for removal of oversized particles, and a pulverizer,
reduced the number of oversized magnesia particles. Fi-
nally, raising the temperature of the magnesium oxide slurry
tank to 82.2°C (180°F) increased the magnesia slaking
. 32,33
rate.
When low-sulfur oil was burned in the boiler during a
test, the solids level in the recycle slurry stream from the
centrifuge rose uncontrollably, forcing unit shutdown. This
was caused by high levels of both magnesium sulfate and
unreacted magnesia, resulting in increased viscosity and
hindered settling. Use of low-sulfur fuel had changed the
SC>2/0~ ratio in the flue gas, generating greater oxidation
and an increase in magnesium sulfate. Low-sulfur fuel was
34
therefore abandoned.
Summary - The operation showed that magnesium oxide
scrubbing for sulfur oxide removal is technically feasible.
The scrubber performed at or above the design efficiency
3-210
-------
(90%) at gas flow rates 50 percent over design. Slurry
solids separation was achieved and magnesium sulfate con-
centration controlled. No plugging and scaling occurred.
Despite improvements, however, mechanical or equipment-
related problems prevented the scrubber from meeting power
plant needs, since it could not be operated for long, un-
interrupted periods. Moreover, losses of magnesium oxide
(20 to 50% makeup with virgin MgO, opposed to a design value
of 5%) were unacceptable for much of the trial. In the last
four months of operation, however, after numerous spills and
leaks had been repaired and the calciner had been made to
operate properly, losses were cut to 3.6 to 5 percent of the
total magnesia feed. This first-generation facility
experienced significant mechanical and design problems, but
many were solved by the end of the test program.
3.4.2.2 Potomac Electric and Power - Dickerson Station,
Unit No. 3
System Description - This boiler burned 2-percent
sulfur coal. Flue gas from the boiler normally passed
through an ESP for fly ash removal before entering a two-
stage venturi scrubber/absorber (Figure 3-56). The MgO
system was tested both with and without the ESP, which can
be bypassed. '
Particulates were removed in the first stage of a
variable-throat venturi, operating with water. The gas was
saturated and cooled to 49°C (120°F) in this stage.
Sulfur dioxide absorption was performed in the second
stage by magnesium oxide scrubbing; slurry contacted the gas
in a fixed-throat venturi.
3
The scrubber was designed to treat 139 m /s (295,000
acfm), which is half the output of a 190-MW boiler. After
3-211
-------
OJ
I
t\J
UHSCRUBBED GAS FOR REHEAT
FLUE GAS BYPASS
1ST STAGE
ELECTROSTATIC
PRECIPITATOR
TO DRY ASH HANDLING SYSTEM
THICKENER i
11 1
,. RECYCLED
4 POND WATER
L,
ASK POND
O-s. TAN'
FRllU STEAM
MgO FROM
ACID PLANT
=SCRUBBER/ABSORBER
YW
MgS03 TO ACID PLANT
''i gure 3—56. Dickerson No. 3
FGD system: General process diagram.
-------
scrubbing, flue gas entered an induced-draft (wet) fan. It
then went to a mist eliminator, where it was mixed with
unscrubbed flue gas to raise the exit gas temperature to
70°C (170°F) before being vented through the stack.39'40
The liquid streams from the particulate and SO9 scrub-
^
bers were separate. The particulate scrubber stage had two
Alloy-20 liquid recycle pumps, each capable of moving the
total flow. A side stream was continuously diverted to two
thickeners of forty-foot diameter. Thickener overflow
returned to the scrubber, together with any makeup water
that might have been required. Thickener underflow was
41
diluted and pumped to an ash pond.
The SO,, scrubber stream passed through a combination of
three recycle pumps, each capable of pumping half the total
slurry flow. A side stream was diverted continuously to a
42
centrifuge. The centrate (liquid phase) was pumped either
to the magnesium oxide slaking operation or directly to the
S02 scrubber section for recycling. From the centrigure,
wet magnesium sulfite cake was screw conveyed to a cocur-
43 44
rent, direct-fired rotary dryer. ' Dried cake went by
bucket elevator to storage silos.
Dryer off-gases passed through a cyclone dust separator
on route to the inlet of the second stage (SO- removal) of
the scrubber. The magnesium sulfite collected in the
4 5 46
cyclone went to the storage silos. '
The dry magnesium sulfite was transported from the
storage silos to an off-site calcining/regenerating plant.
Magnesium oxide was regenerated in a rotary kiln calciner
47
and the SCU gas stream sent to a sulfuric acid plant. The
MgO regeneration unit and sulfuric acid plant at Rumford,
R.I., were used for both the Mystic and Dickerson trials.
3-213
-------
Unique Features of the Dickerson FGD Unit
0 The unit was the first coal-fired utility applica-
tion of the MgO FGD process.
0 The particulate removal venturi and the S02
absorber venturi were both in the same shell.
0 The magnesium sulfite dryer was an oil-fired
rotary kiln using cocurrent flow.
Operating History - Dickerson Unit No. 3 achieved an
average SO,-, removal efficiency of 88.9 percent during
performance testing (Table 3-35). If the one low result
48
were omitted, the average would be 90.4 percent.
Concentrations of S02 ranged from 779 to 1419 ppm at
the scrubber inlet. The unit was designed for 90-percent
efficiency when burning 3-percent sulfur coal (equivalent to
49 50
about 1850 ppm inlet S02). ' At S02 inlet concentrations
over 1000 ppm and at the design gas flow rate, removal
efficiency was consistently above 90 percent. The L/G ratio
was 5.350 1/m (40 gal/1000 acf), and the pressure drop was
20 to 22 cm (8 to 9 in.) of H2O. At lower pressure drop and
lower inlet S02 concentration, reduced efficiency (82.9 to
88.7% removal) was noted. As pressure drop and/or inlet S09
51
concentration increased, removal efficiency increased.
Initially, the S02 removal was below design values (70% vs.
90%). The pressure drop across the absorber throat was then
increased from 7.5 cm (3 in.) H22 removal efficiencies in excess of 90
percent were achieved by better liquid/gas contact.52'53
With regenerated MgO, SO removal efficiency averaged 92
percent during tests in September 1973. This was at rated
gas flow and S02 inlet concentrations from 771 ppm to 1419
ppm.
3-214
-------
Table 3-35. SOV EMISSIONS TEST RESULTS
.A.
MgO FGD SYSTEM - DICKERSONa
Test
series
5A
5B
6
7
8
AP,b
cm (in.) H~0
38.4 (15.1)
16.8 (6.6)
13.0 (5.1)
37.3 (14.7)
13.2 (5.2)
SO? , ppm
Inlet
779
1373
800
1418
1419
Outlet
78
157
137
88
156
SO2 removal ,
%
90
88.7
82.9
93.9
89.0
SO^j , ppm
Inlet
34.6
47.5
2.9
1.8
Outlet
3.56
3.31
0.64
0. 41
U)
I
Ul
Test results abstracted from York Research Corporation, Final Report, Y-8513,
January 31, 1975.
Second Stage (absorber) pressure drop.
-------
Table 3-36. PARTICULATE EMISSION TESTS RESULTS FOR MgO SYSTEM AT DICKERSON
Test3
series
5A
5B
6
7
8
,
Boi ler load
MW
183
96
185
183
176
E5,Pb
Yes
Yes
Yes
NO
No
APC
cm (in. ) H20
25.9 (10.2)
28.2 (11.1)
26.9 (10.6)
27.2 (10.7)
25.4 (10.0)
Flue gas flow,
103 m3/s (acfm)
141 (299)
76 (161)
72 (153)
133 (282)
60 (127)
Inlet
Outlet
Efficiency
Inlet
Outlet
Efficiency
Inlet
Outlet
Efficiency
Inlet
Outlet
Efficiency
Inlet
Outlet
Efficiency
Temp.
°C
117
52
116
47
116
58
11-8
48
111
(°F)
(242)
(126)
(241)
(116)
(240)
(136)
(245)
(118)
(232)
C09
r
12.2
11.9
12.3
11.9
11.6
\
6.5
6.7
6.4
6.5
7.1
Grams/nm
(Gr/scfd)
12% C02
0.104
(0.190)
0.002
(0.002)
99.0
0.055
(0.101)
0.003
(0.005)
95.0
0.065
(0.120)
0.004
(0.007)
94.2
1.991
(3.654)
0.008
(0.014)
99.6
1.785
(3.276)
0.009
(0.017)
99.5
Grams/am
(Gr/acf)
0.071
(0.131)
0.001
(0.002)
98.5
0.038
(0.070)
0.002
(0.004)
91.3
0.046
(0.085)
0.004
(0.007)
91.8
1.348
(2.474)
0.006
(0.011)
99.6
1.205
(2.212)
0.008
(0.014)
99.4
to
I
NJ
h-'
cn
Average of four tests per series.
Indicates whether the flue gas treated by the scrubber stage passed through the ESP.
c First stage (scrubber) pressure drop.
-------
Table 3-37. REMOVAL EFFICIENCY FOR SELECTED
PARTICLE SIZE RANGES
MgO SYSTEM - DICKERSON
Test
Series
5-1
5-2
5-3
5-4
Average
Removal efficiency, %
Above 5 ym
99.0
99.8
99.0
98.8
99.2
1 to 5 ym
97.5
99.2
99.9
97.5
98.4
Below 1 ym
94.6
99.0
98.6
97.0
97.3
3-217
-------
As shown in Table 3-36, particulate removal to below
0.02 grain per standard dry cubic foot (gr/scfd) could be
attained, at any flue gas volume up to design flow rate,
irrespective of whether the scrubber was accepting flue gas
before or after it went through the ESP. Inlet loading to
the scrubber ranged from 0.03 mg/m (0.06 gr/scfd) to 1.99
mg/m (3.66 gr/scfd). Scrubber efficiency was over 99
percent for particles over 5 ym; over 98 percent for par-
ticles between 1 and 5 ym; and over 97 percent for particles
below 1 ym (Table 3-37).55'56
The operability of this MgO FGD system, defined as
hours of operation divided by boiler operating hours, is
shown in Table 3-38.
The unit was a prototype trial installation, built to
obtain operating data. As such, equipment and materials
were used that would not have been used in a long-term
installation on a new plant. This operation fulfilled its
purpose by showing areas where improvement was needed. The
operability data in Table 3-38 show the downtime caused by
mechanical and material failures.
57 5fi
Table 3-38. OPERABILITY DATA FOR DICKERSON NO. 2 '
Time period
Operability, %
September 13, 1973 - January 14, 1974
December 9, 1973 - January 14, 1974
April 15, 1974 - May 1, 1974
August 1, 1974 - August 31, 1974
November 1, 1974 - November 30, 1974
November 15, 1974 - November 30, 1974
December 1, 1974 - December 31, 1974
Test Runs Completed August, 1975
27.4
58.3
40.5
43.5
44.6
67.9
57.9
3-21!
-------
Problems and Solutions - The major equipment items
(scrubber vessel, thickeners, centrifuge, and dryer) all
performed well. Major problems were encountered, however,
with construction materials for handling systems. Carbon
steel pipe and slurry pumps were found to be inadequate
against the corrosive/erosive properties of the slurry.
Individual pump suction and discharge lines were necessary
since leaks in pipes at or after a header could make in-
stalled spare pumps useless. The designs of the centrifuge
discharge hopper, weigh belt feeders, and the dry magnesium
59
sulfite bucket elevator were improved.
Proper feeding of magnesium oxide to the slurry system
became a problem, because of plugging in the mix tank and
suction lines to the feed pumps. Later, there were diffi-
culties with the proper slaking of regenerated MgO. The
final solution to both problems was the installation of a
steam-sparged, agitated premix tank/slaker to promote dis-
. .. 60,61,62
solution.
Various rods, bellows and hangers, etc. corroded in the
reaction vessel. It was found that the wrong materials had
been used, particularly 304S.S. instead of the specified
316S.S. The proper materials were used to replace the cor-
fi 3
roded parts and this problem was solved. Corrosion and
erosion were severe in the recirculating piping of both
first- and second-stage scrubber slurry- Fiberglass rein-
forced polyester (FRP) and epoxy were used to make repairs.
For long-term commercial use, however, rubber-lined piping
should be specified.
Because the MgO regenerating facilities at the Essex
Chemical sulfuric acid plant at Rumford, Rhode Island, were
unavailable, the unit exhausted its supply of virgin magne-
3-219
-------
slum oxide.* When the calciner became available, it was
discovered that magnesium sulfite from Dickerson Station was
predominantly hexahydrate, whereas that obtained from
Boston Edison's Mystic Station was mostly trihydrate. This
necessitated additional testing, to determine the proper
operating parameters (temperature and feed rate) for cal-
65,66
ciner operation.
The bucket elevator conveying the dried magnesium
sulfite to the storage silo tended to overload and trip.
The problem was traced to the discharge chute from the
centrifuge, where wet magnesium sulfite cake tended to hang
up and then break off in large chunks. Modifications to the
discharge of the centrifuge outlet hopper and installation
of larger buckets for the sulfite elevator helped overcome
... 67,68
this.
Problems affecting the centrifuge included buildup of
material or wear inside the centrifuge, and changes in the
physical/chemical form of the magnesium oxide. Use of
washout ports and hardened steel surfacing, similar to those
69
changes introduced at Mystic, should remedy this.
The dryer was designed for use with chain sections, to
allow better product drying and the reduction or elimination
of product buildup. The chains were never installed,
however.
As is apparent from this summary, the major problems
encountered at the Dickerson Plant were related to materials
handling. They are now avoidable.
3.4.2.3 Philadelphia Electric-Eddystone Station, Unit
No. 1A
System Description - This generating unit burns 2.5-
71
percent sulfur coal. The flue gas passes through a mech-
*
The regeneration facility was being used to process MgSO.,
from the Mystic Plant at this time.
3-220
-------
anical collector and an ESP before entering a system of
three parallel scrubbing trains with a wet particulate
scrubber, a reheater, and a scrubber booster (induced draft)
fan in each train. A flue gas bypass system was also in-
72
stalled. One of the three trains (Train C) has a SO,,
venturi scrubber in series with the wet particulate scrubber
(Figure 3-57).
The particulate scrubbers are made of 316L stainless
steel. The particulate and SO,, scrubbers have separate
surge tanks for slurry collection, pH control, and makeup.
The surge tanks and the SO scrubber are carbon steel with a
7374
polyurethane coating; the piping systems are rubber-lined. '
In the particulate scrubbing system shown in Figure
3-58, incoming flue gas first enters a venturi-type wet
scrubber. The scrubber slurry goes to a surge/storage tank;
the recirculating pump returns the slurry to the scrubber;
and a continuous side stream is diverted to the waste water
system.
The gas stream is cooled from 146°C to 53°C (295-
127°F), and humidified in the particulate scrubber. Hu-
midified flue gas stream then enters the S0_ venturi-type
^ ^
(Ventri Rod ) scrubber (two stages of scrubbing and two
stages of mist elimination), where the magnesium oxide/
hydrate slurry scrubs out S0?. The slurry then goes to the
SO- scrubber surge tank, afterwhich the recirculating pump
returns the slurry to the scrubber at a pH of 6.0. A
continuous side stream is removed and sent to the magnesium
sulfite recovery system; and makeup magnesium hydroxide
slurry is fed to the SO.-, scrubber surge tank. The L/G ratio
O ^
is 6689 liters per am /min (50 gal./lOOO acfm); the scrubber
operates at a pressure drop of 12 in. H00. Following SO.-,
3-221
-------
BYPASS
LO
1
!\J
KJ
FROM NO. 1
BOILER
BYPASS
BYPASS
1A PART.
SCRUB.
IB PART.
SCRUB.
1C PART.
SCRUB.
1
PUT.
I J
IDF*
REHEATER
l~~~
1 FUT
I SCL
i 2
L-_
~1
. i
1
1
. — 1
--^-REHEA
IDF
TO NO.l
STACK
BYF
>ASS
so2
SCRUB.
•^- REHEA1
IDF*
* INDUCED DRAFT FAN (IDF)
3—57. Eddy stone unit 1A SO - remova.3. system.
-------
STACK GAS
FROM I D
FANS
PARTICULATE
SCRUBBER
ro
OJ
HUMIDIFIED
GAS STREAM
S02
SCRUBBER
CLEANED GAS TO
REHEAT AND STACK
Mg (OH)2 SLURRY
PARTICULATE
SCRUBBER
SURGE TANK
S02 SCRUBBER
SURGE TANK
TO Mg S03
RECOVERY
TO WASTE WATER
TREATMENT SYSTEM
Figure 3-58. S00 scrubber system at Eddystone Plant.
X ~~
-------
removal, scrubbed gas then enters a direct-fired (oil)
-i (• T -y
reheater. The reheater eliminates condensation and
corrosion in the stack, as well as the stack plume.
The magnesium sulfite slurry enters a thickener (Figure
3-59). Thickener underflow goes to a centrifuge; overflow
is recycled to the SO,., scrubber surge tank, or to the magne-
sium oxide slaking tank. Centrifuge liquor is handled in a
. . . 78
similar manner.
The centrifuge produces a wet magnesium sulfite cake,
which is fed to a cocurrent dryer. The dry product is
pneumatically conveyed to storage prior to shipment to an
79
off-site regeneration plant (Figure 3-60).
Exhaust gas from the dryer passes through a dust col-
lector (collected dust is pneumatically conveyed to product
storage) , and on to the inlet side of the SO.., scrubber
(Figure 3-59).8°
The off-site regeneration process uses a fluidized bed
calciner for regeneration of magnesium oxide and liberation
of SO,., for sulfuric acid production. Magnesium sulfite is
fed to the calciner from a solids feed tank by a screw
conveyor. Hot air (954°C or 1750°F) and calcined magnesium
oxide leave the fluid bed calciner and pass through an air
preheater to a recovery cyclone. There the magnesium oxide
is recovered and conveyed to storage. The gas stream is
cleaned and scrubbed for removal of entrained MgO and any
other particulates, before being piped to the sulfuric acid
plant.81
Operating History - Although the Eddystone FGD system
operated only a short time, it removed more than 90 percent
of the SO_ when both trays of ventri rods in the SO,., ab-
sorber were used and when the L/G was 6700 &/100 am /min
3-224
-------
TO S02
SCRUBBER INLET
Mg S03 -
OJ
i
SLURRY
THICKENER
TANK
MOTHER
LIQUOR
TANK
RECYCLE
TO MgO
SLAKING
TANK
Mg S03
STORAGE
SILO
TO MgO
REGENERATION
PLANT
Figure 3-59. SO,, scrubbing system - Mg SO., recovery.
-------
750°F
SOLIDS
FEED
LO
I
NJ
NJ
FLUIDIZED
BED
REACTOR
(REFRACT-
ORY-
LINED)
\\\\\\\\\\\
FUEL OIL
AIR
PREHEATER
K
IOOO°F
AIR
COMBUSTION
AIR BLOWER
1200
°F
S02 RICH GAS
I64°F
CYCLONE\7
RECOVERY
SYSTEM
TO
RECOVERED
MgO STORAGE
1
1
i
1
BLOW DOWN
1
VENTUR!
WET
SCRUBBER
;MAKEUP
'WATER
Figure 3-60. MgO regeneration plant.
-------
(50 gal./lOO acfm). Eddystone must meet an SO emission
standard of 0.26 g/MJ (0.6 Ib S02/106 Btu). Published test
data are not available.
Limited operability data given below show the initial
8 ?
start-up difficulties. Sulfur dioxide absorption ceased
from January 1976 until June 1977 while the regeneration
facility was being moved.
Particulate Scrubbers Operability
1C: July 23, 1975 - January 31, 1976: 70 percent
IB: August 15, 1975 - January 31, 1976: 55 percent
1A: October 2, 1975 - January 31, 1976: 24 percent
S02 Scrubber
1C: October 2, 1975 - December 31, 1975: 33 percent
The S02 scrubber was shut down because Olin Chemical
closed the acid plant serving the MgO calciner. The
regeneration equipment was relocated to Essex Chemical,
Newark, N.J., and restarted in June 1977.
Problems and Solutions - The particulate scrubbers
experienced many problems. All three scrubber booster fans
developed high shaft vibration (IB fan failed in November
1974, 1C failed in July 1976, and 1A failed in September
1976). These failures were caused by excessive clearance of
the slip fit between the shaft and the wheel hub, resulting
in high vibration at operating temperatures. This construc-
tion error meant that all three fans had to be returned to
the manufacturer for shrink-fit repair. They were back on
line in March 1977.83'84
The recirculation pumps for the particulate scrubber
were too small. This problem, too, was resolved with the
manufacturer.
The particulate scrubber by-pass dampers are operated
by a pneumatic cylinder through a lever arm. The lever arms
3-227
-------
failed several times, so heavier ones with longer stroke
pneumatic cyclinders have been installed.
The particulate scrubber suffered extensive corrosion
caused by pH levels as low as 0.5, chlorides of over 200
ppm, and the "shock" of a "hot start" procedure. The
problem has been lessened by a) the addition of caustic to
maintain the system pH at 2.5; b) higher blowdown to control
the chlorides; and c) the introduction of a "cold start"
87
procedure.
Within 3 months the polyurethane coating in the 1C S0_
scrubber and particulate surge tanks began to blister and
peel. All the vessels were sandblasted and coated with
88
"flake glass" polyester coatings. This has held up well.
The reheaters on all three systems developed refractory
problems, caused by faulty flame/torch placement, the
failure of ultraviolet scanners, and inadequate heat dis-
89
sipation. These problems have all been resolved.
In the most recent test period (May 1977 to August
1977) the main problems were with ancillary equipment
(rotary valves, pumps, etc.). Vendors are working with
United Engineers and PECo personnel to resolve these prob-
lems. The major units (scrubbers, absorber, etc.) worked
properly.
3.4.3 Description of Japanese MgO Scrubber Installations
Japan has three industrial MgO FGD installations (Table
3-39), none of which operates solely on boiler flue gas.
Brief descriptions of them are given in this section.
Although these units do not operate on boiler flue gas,
their operating principle is identical to that of systems on
coal-fired power plants in the United States, and the equip-
ment used is very similar. The Japanese plants reportedly
have worked well; specific operability data have not been
published, however.
3-228
-------
Table 3-39. FgO FGD SYSTEMS IN JAPAN
.90,91
Process
developer
Onahama-Tsukishima
Chemico-Mitsui
Mitsui Mining
(shut down)
User
Onahama
Smelting
Idemitsu
Kosan
Mitsui
Mining
Plant site
Onahama
Chiba
Hibi
Gas flow
1000 nm3/hr
84
(28 MW)*
500
(162 MW)*
80
(25 MW) *
Gas
source
Copper
smelter
Claus &
boiler
H2S04
plant
Inlet
SO2
ppm
25,000
(Not
availab-
(Not
availab.
Installed
1972
1974
-e)
1971
Le)
By-product
H2S04
240 T/D
S
H2SO4
18 T/D
* Equivalent MW based on gas flow from oil-fired boilers.
CO
I
-------
3.4.3.1 Onahama - The Onahama Smelting operation cleans the
gas stream from two reverberatory copper smelting furnaces.
Half the gas stream is cleaned using a MgO FGD system, the
92
other half by a lime scrubber system. The MgO scrubber is
not currently operating, because of the reduced operation of
the smelter and the increased demand for gypsum (the by-
product of lime scrubbing). For several years, however, the
lime and MgO systems operated in parallel (Figure 3-61)).
The furnace gas, containing 2.0 to 3.0 percent (20,000
to 30,000 ppm) SO-/ enters waste heat boilers, passes
through an ESP and a water wash system, and is further
cooled by surface coolers. At this point, the cooled gas
stream is divided and routed to the MgO or the lime FGD
system. The MgO system uses two turbulent contact S02
absorbers (TCA) . The primary product is MgSO^ • 6E<0. The
two MgO absorbers operate alternately every two weeks.
These TCA absorbers originally were backwashed with sulfuric
acid for removal of magnesium sulfite scale; more recently,
however, only water has been used. The problem of wear on
the polyethylene balls has necessitated their replacement
every few months. The absorber off-gas passes through a
mist eliminator and an ESP, to be discharged from the stack
Q? Q4
at 60°C (140°F).yj'y^
The MgSO-, stream from the scrubber enters a liquid
cyclone; the resulting concentrated slurry is centrifuged
and the MgSO-, cake that is produced is moved by screw con-
veyor to a rotary steam dryer. Centrifuge filtrate is
returned to the scrubber loop. The MgSO^ crystals that are
—3
formed are normally about 200 ym (7.87 x 10 in.) in size.
In the drying operation, less than 3.5 kg of steam/kg SO,, is
consumed. Both dryer and centrifuge operations have been
3-230
-------
Quick Lime
Slaker
_£
Absorption
Tower
L
Oxidation
Tower
Thickener
Centrifuge
Stock Yard
Gypsum
Reverb Gas
Hot Cottrell
Washing Tower
Gas cooler
Wet Cottrell
75m Stack
Magnesium Hydroxide
( makeup )
Storage Tank
1
Absorption
Tower
J
Centrifuge
Dryer
Calciner
Slaker
J.
Sulfuric
Acid Plant
Sulfuric Acid
Figure 3-61. Flowchart - Onahama reverberatory
furnace S0_ control system.
3-231
-------
major sources of trouble in the United States; in Japan,
however, these areas have been trouble-free. '
Steam from the dryer is condensed and discharged as
liquid waste. Dryer off-gas passes through a baghouse and
scrubber, and then through an exhaust fan to the discharge
97
stack. Typically, 10 percent of the product from the
dryer is MgSO,. Coke is added to the dried product through
a variable-speed rotary valve. This aids reduction of the
MgSO. to MgSO., in a direct, oil-fired, rotary calciner
operating at 750° to 800°C (1382-1472°F) and consuming oil
at the rate of less than 0.13 kg/kg S0~.
Of the regenerated product, only 1 percent is MgSO..
Magnesium oxide agglomerates are ground prior to slaking.
The material in the slaker comprises 92 percent regenerated
material and only 8 percent virgin MgO.
Losses of MgO occur primarily from the calciner ex-
haust. On an average, the proportion of SO^ in the exit gas
stream from the calciner is 11 percent (10 percent to 13
percent normal range). Calciner gases, after washing,
become feed stock to the sulfuric acid plant. Regenerated
MgO is slaked and recycled to the absorber. Plant opera-
tional specifications and component data are presented in
qp qq
Tables 3-40 and 3-41. '*y
This FGD system demonstrated S02 removal efficiencies
from 99.5 percent to over 99.8 percent from a concentrated
(2 to 3 percent S0«) reverberatory furnace gas stream
[utility boilers fired with coal with a 3.5 percent sulfur
content would typically have a concentration of 0.2 percent
S02 (2000 ppm]. The design called for 99.5-percent removal
from a 3 percent inlet gas stream. As a safety factor, the
smelter must be operated 50 percent to 70 percent in excess
3-232
-------
Table 3-40. MgO PLANT OPERATIONAL SPECIFICATIONS
ONAHAMA SMELTER
Reverberatory gas:
Volume (dry)
SO.,
°2
C02
N2
H2
Temperature
Pressure drop (MgO Plant)
Calciner (rotary kiln) gas:
S02
Tail gas:
Volume
S02
Steam consumption at dryer
Oil consumption at calciner
Mg(OH)„ makeup
1500 nm /min (52,950 cfm)
2.4 to 2.6 percent
6.0 percent
12.0 percent
Balance
0.12 (kg/kg dry gas)
300°C, 572°F
8.0 (in. H20 (20.3 cm)
10 to 13 percent
270 nm3/min (9531 cfm)
<100 ppm (-20 ppm)
<3.5 kg/kg SO2
<0.13 kg/kg S02
<0.08 kg/kg S0
3-233
-------
Table 3-41. MAJOR COMPONENTS OF MgO PLANT -
ONAHAMA SMELTER
Absorption step:
TCA tower 4.0 m (13.2 ft) dia. x 26.5 m (87.2 ft)
length x 2 (1 backup)
Blower 2200 nm3/min (77,660 cfm) x 900 Kw x
2 (1 backup)
Separation and drying step:
Centrifuge Essherwise type 10
Dryer 900 m2 (9720 ft2) x 2 [3 m (9.8 ft)
diameter, 25m (82 ft) long]
Thermal decomposition step:
Rotary kiln 3.4 m (11.2 ft) dia. x 52 m (171 ft)
length
3-234
-------
of the 99.5 percent goal. Operation, therefore, is strictly
controlled to remove 99.7 percent of the sulfur input. The
design outlet SO,, emission was 100 ppm. Actual operation
has achieved less than 50 PPm.10°'101'102'103'104
3.4.3.2 Idemitsu Kosan - At Idemitsu Kosan's Chiba re-
finery, Mitsui Miike Machinery Co. constructed a MgO FGD
system to clean 500,000 nm /hr (294,167 cfm) of waste gas
from industrial boilers and Glaus furnaces. The unit went
on stream in November 1974 and has consistently achieved 95
to 97 percent SC> removal efficiency at an inlet SO- con-
centration of 2,500 ppm.105'106
The venturi particulate scrubber and venturi SC>
absorber are both in the same shell. The MgSO., slurry
discharge from the absorber is pH-adjusted and filtered.
The MgSO., cake is dried in a rotary dryer countercurrent to
the hot gas stream; the dried MgSO,, is calcined in an oil-
fired, rotary kiln. The MgSO., cake contains 10 to 15
percent MgSO.; coke is added in the calcination step to
reduce the sulfate to sulfite. Regenerated MgO is slaked
and then wet-milled to increase its reactivity- Off-gas
from the calciner goes through a cyclone, a wet venturi, and
a wet precipitator. This cleaned gas, containing 10 to
12 percent S02, is fed to the Glaus furnace, where H2S
recovered in the refinery is reacted with S02 to produce
i * 107
elemental sulfur.
The form of magnesium sulfite, trihydrate, or hexahy-
drate is an important key to plant operation. Usually
hexahydrate is preferred, because it grows in much larger
crystals than does trihydrate. Mitsui Miike has found that
mixing the two forms gives the best results. There has been
no problem with dryer operation. The major problem, how-
3-235
-------
ever, was with the the firebricks in the rotary kiln. They
became dislodged, possibly the result of cooling by rain.
To prevent corrosion, chloride concentration in the scrubber
liquor is maintained below 2,300 ppm. Normally wastewater
is retained because of the low inlet chloride concentra-
108,109
tion.
3.4.3.3 Mitsui Mining - In 1971, Mitsu Mining and Smelting
Co. built an MgO system at the Hibi Works. This unit
3
treated 80,000 nm /hr (47,067 cfm) of tail gas with 1500 ppm
to 2000 ppm SO^ from a sulfuric acid plant. A cross-flow
absorber was used, and MgSCu-GH^O was formed. Sulfite
crystals were separated by centrifuge from the solution,
which contained some magnesium sulfate. The sulfite was
dried in a rotary dryer and then calcined in an indirect-
fired calciner at 750°C. The concentrated SO,., was returned
to the acid plant. Magnesium sulfate solution was con-
centrated to produce MgSC>4 • 7H«0, which was sold for fer-
tilizer and other uses. HO
3.4.4 Engineering Design Parameters
Sulfur dioxide collection efficiencies in excess of 90
percent can be achieved with a MgO FGD system on coal fired
boilers. Key design and operating parameters for 90 percent
removal include:
0 Operation of venturi scrubbers at pressure drops
of 2.5kPa (10 in. H2O) or greater, with an L/G
ratio of 40 to 50; use of a TCA scrubber operating
at 2.0 kPa (8 in. H20) on smelter reverberatory
furnace off-gas.
0 Maintenance of slurry pH (measured in the absorber
discharge tank) in the range of 6.0 to 7.5 (Figure
3-62).
0 Slaking (dissolution) of MgO in water in a heated,
agitated tank before it is added as makeup to the
absorber.
3-236
-------
s
^_
o
r^
^_
LU
O
t— H
LJ_
u_
LU
CJ3
t— i
CO
CO
Di
o
oo
1 UU
99
98
97
96
95
94
93
92
91
90
89
flft
(JO
871
86
85
Q/l
' 1 1 ' 1
— —
..-•"•---•W"1"""®8*"11"' -
— s*^^ —
~ /
~ /
- $
f
— / —
f
~ t
- J
-1
t
-i
f
i
j_ _
— —
1,1:1
6 7
pH OF SCRUBBING SLURRY
Figure 3-62. The affect of pH on SO,., scrubbing efficiency.
Ill
3-237
-------
0 Avoidance of over-calcining or dead-burning of
regenerated MgO, to maintain high reactivity.
0 Prescrubbing of flue gas with water for residual
fly ash removal and cooling, which minimizes
formation of trihydrate of MgSC>3 and insures
better handling, drying, and calcining of the
product.
Table 3-42 summarizes operating parameters and SOp collec-
tion efficiencies for the three plants operated in the
United States.
In the past, MgO FGD systems installed by Chemico
(Mystic and Dickerson) and United Engineers (Eddystone) had
no overall performance guarantees; rather, the manufacturer
of each piece of equipment guaranteed that particular unit
against manufacturing defects, guarantees varied greatly.
Now, however, Chemico is willing to guarantee the
entire MgO FGD system mechanically, as well as to specify
that the unit will meet the applicable SO9 regulation 0.52
/- ^-> -I -I **L
g/MJ (1.2 lb/10 Btu, 90 percent removal, etc.).
3.4.4.1 Applicable Research Studies - Chemico sponsored
work at the New Jersey Institute of Technology (NJIT),
Newark, New Jersey, to determine conditions favoring the
formation of magnesium sulfite trihydrate and hexahydrate.
Trihydrate is more difficult to handle physically and has a
smaller, more granular crystal than hexahydrate. (The
primary product at Boston Edison's Mystic unit was the
trihydrate form; while at PEPCO Dickerson, hexahydrate
predominated.)
The NJIT study of S02~bearing gas (at 160°C, 320°F),
with MgO slurry at 49°C (120°F), 57°C (135°F), 65°C (149°F)
and 80°C (176°F), showed that at 49°C (120°F) and 57°C
(135°F) the major product was MgSO -6H20, while at 65°C
(149°F) and 80°C (176°F) it was the trihydrate MgS03'3H20.
3-238
-------
oo
I
Table 3-42. SUMMARY OF MgO SCRUBBING SYSTEM OPERATING PARAMETERS
Plant
Mystic
(Boston Edison)
Dickerson
(Potomac
Electric)
Eddystone
(Philadelphia
Electric)
ppm
1042
1092
1500
Inlet S02
lb/106 Btu
1.05
1.60
1.90
g/MJ
0.452
0.689
0.819
Outlet S02
ppm
89
121
Below 150
SO? removal
efficiency,
percentage
91.5
88.9
>90
AP
KPa
2.0
2.0
2.5
in. H20
8
8
10
-L
a/1000 mj
4400
5350
6700
/G
gal/1000 acfm
33
40
50
Slurry
pH
7.1
7.1
6.3
OJ
-------
The presence of magnesium sulfate in solution (up to 18%),
and the passage of excess air, had no significant influence
on the nature of the hydrate formed.
Chemico also conducted research at the MgO regeneration
facility in Rumford, Rhode Island, where the reactions of
S0~ at 149°C (300°F) with both virgin and regenerated MgO
were studied at 57°C (135°F), 60 to 63°C (140 to 145°F) and
66 °C (150°F). The pH was maintained at 6,8 to 7.2. Samples
were collected periodically and the filtrates were returned
to the reactor. The results of the study agreed with those
performed at the NJIT; i.e., hexahydrate predominated at the
114
lower temperatures.
At 57°C (135°F) the major product was the hexahydrate;
trihydrate was the predominant product at 66°C (150°F). The
study at 60 to 63°C (140 to 145°F) was interesting—it
showed the gradual change from hexahydrate (original prod-
uct) to trihydrate (final product). The complete trans-
formation from hexahydrate to trihydrate required 4 to 5
hours. The study at 60 to 63°C (140 to 145°F) showed also
that the transition of hexahydrate to trihydrate was accom-
panied by a reduction of crystal size (from 200 to 50 ym) .
Analysis of centrifuge cakes from PEPCO and Boston
Edison supported the contention that: a) the ratio of the
two hydrates at any temperature is controlled by the equi-
librium value at the temperature; b) hexahydrate persists
even above the transition temperature; and c) the rate of
transition is slow.
Further work was done to determine the drying mechanism
of magnesium sulfite trihydrate and hexahydrate, and to
establish the drying rate at various temperatures, so that
dryer operation could be improved. The tests were performed
3-240
-------
in an oven operating at 40.5 kPa (12 in. Hg) vacuum and at
80°C (176°F), 120°C (248°F), and 150°C (302°F). At 80°C and
120°C, reaction rates were very slow for hexahydrate and
nonexistent for trihydrate. At 150°C, hexahydrate dehy-
drated to trihydrate in two hours, and then gradually lost
water until it reached MgSO «1/2H~0 (apparent composition).
At 150°C, the trihydrate was stable for 48 hours and then
rapidly dehydrated to the apparent MgSO., • 1/2H-0. Additional
work at 166°C (330°F) showed dehydration of the hexahydrate
in 10 minutes, and of the trihydrate in 25 minutes. The
dehydration study indicated that water molecules are more
117
strongly bound in the trihydrate than in the hexahydrate.
A comparison of the dryer product temperature with the
percentage of dry solids (MgSO^), as a function of the dryer
gas outlet temperature, showed that significantly more heat
energy must be used to dry the trihydrate, the predominant
product at Boston Edison, than is needed to dry the hexahy-
118
drate, the predominant product at PEPCO (Figure 3-63).
The S0~-removal capability of regenerated MgO was also
investigated. The size range of MgO particles returned for
slaking was found to be very important in this respect.
Experimental data indicate that grinding calcined magnesia
119
increases its reactivity, but only within certain limits.
Laboratory data now indicate that a size range of between 75
-3 -3
and 150 ym (2.9 x 10 and 5.9 x 10 in.) is best for
recycled, regenerated magnesia. The greatest increase in
MgO activity occurs when particles of 295+ ym are reduced to
_-> _2
150 to 295 ym (5.9 x 10 to 1.2 x 10 in.) additional
improvement in S00 removal is obtained when size is reduced
-3 -3
to 75 to 150 ym (2.9 x 10 to 5.9 x 10 in.). Further
reduction has only a slight effect on activity increase.
3-241
-------
01
Q
M
.-I
o
wj
8
Q
00
I
NJ
*.
to
100
90
80
70
38
PEPCO DRYER
171'C
66
93
BOSTON EDISON DRYER
121
149
177
204
DRYER BED PRODUCT TEMPERATURE,*C
Figure 3-63. Comparison of PEPCO and Boston Edison Dryer Operation.
(For ranges of dryer gas outlet temperature)
-------
The properties of regenerated MgO are determined by cal-
cining conditions, as well as by MgS04 and fly ash levels.
At PEPCO, flue gas was saturated and cooled in the wet
particulate scrubber to 49° to 55°C conditions that favor
hexahydrate formation in the absorber, At Boston Edison,
flue gas entered the single-stage venturi scrubber at 138°
to 160°C (280 to 320°F); favoring the formation of trihy-
drate.
Further work just completed by Radian Corporation
determined the following:
1. The trihydrate is the thermodynamically stable
form at scrubbing process conditions. Precipita-
tion of the hexahydrate occurs as the result of
kinetic phenomena. Precipitation may take place
by two mechanisms: by nucleation, and by growth on
existing crystals. Both mechanisms were studied.
Nucleation and crystal growth rates of the hexa-
hydrate are much faster than those of the tri-
hydrate at conditions encountered in the magnesium
oxide wet scrubbing process.121
2. Operating experience with magnesium oxide scrub-
bing processes indicates that magnesium sulfite
hexahydrate forms under conditions where, theo-
retically, trihydrate is the thermodynamically
stable form. Formation of the metastable hexa-
hydrate is due to kinetic phenomena, which are
described in this work.
3. The system under investigation is complex and
several phenomena can occur simultaneously. These
include trihydrate nucleation, hexahydrate nuclea-
tion, trihydrate crystal growth, hexahydrate
crystal growth, and hexahydrate dissolution.123
The results obtained in scrubber-like systems were
consistent with the results for tests of the
individual phenomena.124
4. The following observations about the relative
rates of the individual processes are important
for design considerations. The rate of trihydrate
3-243
-------
nucleation at magnesium oxide operating tempera-
tures of 55°C (131°F) appears to be very slow
compared with the rate of hexahydrate micleations.
Although hexahydrate nucleation was observed,
evidence of trihydrate nucleation in scrubber-like
media was not obtained. In addition, the rate of
hexahydrate crystal growth was much faster than
the rate of trihydrate crystal growth.1"
5. The following sequence was observed in scrubber-
like systems and the times were characterized:
hexahydrate precipitation (tens of minutes),
simultaneous hexahydrate dissolution and trihy-
drate precipitation (hundreds of minutes), tri-
hydrate precipitation and attainment of equilib-
rium (thousands of minutes).I26
6. By specifying the composition and particle size of
seed crystals, the solution composition (driving
force), and the reactor volume (residence time),
it is possible to design a system to produce
either of these hydrates as the solid product.
7. Design tools used in other SC>2 slurry-scrubbing
processes, such as the particle balance concept,
are applicable to this system.128
8. Since nucleation rate is an important factor,
equipment that cause high-energy changes in solu-
tion will have an effect on precipitation phe-
nomena. Turbulence effects of nozzles, pumping,
agitator speed, and control valve throttling will
be important. 2^
9. Hold-tank and reactcr volumes, and slurry density
determine the solid residence times. Solid
residence time in the system will be an important
design criterion.130
10. Temperature and solution composition can be used
to control seed crystal composition.131
11. In solutions containing only magnesium sulfite in
water, the hexahydrate is the stable form at
temperatures up to the transition temperature of
41°C (106°F). At higher temperatures, the tri-
3-244
-------
hydrate is the form that precipitates from pure
solutions. The analysis showed that the tri-
hydrate is the thermodynamically stable species in
solution compositions and at temperatures char-
acteristic of the scrubbing process. Thus, the
formation of hexahydrate crystals is explained on
the basis of kinetic factors.132
3.4.5 Process Operability Considerations - Data on both
systems indicate 27 to 80 percent operability, which is
expressed as the ratio of scrubber operating time to boiler
operating time for MgO systems applied to coal-fired boilers
in the United States. When reviewing these operability
levels, a number of factors must be kept in mind, including:
0 Sulfur dioxide collection efficiency was fre-
quently over 90 percent during test periods.
0 Two units (Mystic and Dickerson) were trial in-
stallations, built to obtain operating data. As
such, various construction materials were used
that would not have been used in a long-term
installation on a new plant.
0 Handling and drying properties of the wet MgSO3
were not adequately considered in system designs.
This caused problems in the ancillary equipment
serving the scrubber system, not in the scrubber
itself.
0 The sulfuric acid plant which was to receive SC>2
from the Eddystone MgO regeneration facility was
shut down by its owner and a new site had to be
found.
0 The single regeneration facility at Fumsford,
Rhode Island, could not simultaneously process
material from Mystic and Dickerson.
0 Many of the problems at Eddystone were related to
particulate scrubbing and not to the SOo absorber
section.
0 Many problems encountered and solved during these
early programs would not be repeated in subsequent
designs.
3-245
-------
This type of FGD system also has the distinct advantages of
having no sludge disposal problem, and, thanks to its pro-
cess chemistry, only minimal scaling and plugging problems.
3.4.5.1 Operational Problems and Solutions - Operational
difficulties with the two trial systems have led to better
design. The problems were primarily mechanical, caused by
faulty equipment or by a lack of understanding of the nature
of magnesium oxide and its sulfur compounds.
Heated makeup mixing tanks or slakers equipped with
agitation and steam spargers are necessary if the production
of makeup magnesium slurry is to be reliable. At Mystic and
Dickerson, attempts to add dry magnesia without prior
slurrying failed. Installation of a heated slurry tank with
agitation solved this problem. ' ' When regenerated
mangesia was first used, it was found to be less reactive
than virgin material; additional heat was required to slake
it. ' Other methods of introducing and slurrying
magnesia were either unsuccessful, or less successful than
138
high-speed agitation and steam sparging. Heating the
magnesia solution above 82°C (180°F) in the slaker must be
avoided, since scaling will result in the tank and pipes.
Corrosion and erosion occurred in both the particulate
scrubbing and the SO» absorbing systems, although rubber-
lined pipes, pumps, and valves reduced the problem. Crit-
• i u u u • 4. n , j 139,140,141,142,143,144
ical pumps should have installed spares. '
The pH in the particulate scrubbing system had been reported
145
as low as 0.5. This, coupled with the abrasiveness of
the fly ash resulted in the erosion, corrosion and rapid
failure of carbon steel. In the sulfur dioxide scrubbing
system, the failure of carbon steel was due to the abra-
siveness of magnesium oxide and to periodic pH swings from
At the Eddystone Station.
3-246
-------
0.5 to 9.0, caused by fluctuations in the magnesium oxide
146
makeup slurry feed. Inside the scrubbers, 316L sLainless
steel, Inconel 625, flaked glass, and corrosion-resistant
147 148 149
polyester resin coatings gave the best service. ' '
At Mystic and Dickerson, flaked glass and polyester resin
provided excellent corrosion protection. There was no
problem with scaling or buildup inside the scrubbers. ' ' '
Both magnesium sulfite trihydrate (small, hard-to-
handle crystals) and hexahydrate (large, easy-to-handle
crystals) are produced in the scrubber. The systems should
be capable of handling the more troublesome magnesium sul-
*-4. 4- -u A *. 154,155,156
fite trihydrate. ' '
Trihydrate requires halt as much heat of vaporization
to dewater the crystal as does hexahydrate; the latter loses
water at 100° to 120°C (212 to 248°F), however, while tri-
hydrate loses it at 180° to 190°C (356 to 374°F). Drying
the hexahydrate may be accomplished with low pressure steam,
while drying the trihydrate would probably require oil-fired
heating.
Centrifuge bowl plows, head plows, and conveyors
should all have hard surfaces (stellite, possibly ceramics)
to reduce or eliminate the kind of wear problems that were
experienced at Mystic and Dickerson. Also, wash-out
connections capable of thoroughly washing the centrifuge
internals should be installed.158'159'160 A final check on
the stellite-hardened surfaces at Mystic indicated a life
expectancy of over a year. The need for the wash-out
connection was shown at both Mystic and Dickerson, where the
centrifuges had to be dismantled to break loose hardened
material. Seizing of the centrifuge internals ended the
Dickerson trial prematurely. Consequently, in keeping
3-247
-------
w ch tilt, coianon practice, spare centrifuge Siiovilc; be In-
n n . .. 162,163,164
eluded in any new unit.
Further, the centrifuge should be designed to handle a
wide range of concentrations and viscosities of solids,
while giving good separation. This is necessary because
higher-than-normal solids levels, the result of excessive
magnesium sulfite trihydrate concentrations arid higher-than-
,.,_,,_-, 165,166,167,161
normal viscosities, have caused unit shutdown.
The centrifuge discharge hopper must be designed to
prevent the product from building up. Downstream surges,
caused by sudden release of large masses'of wet magnesium
sulfite cake, have caused equipment overloads, particularly
, , . 169,170
in conveyors and elevators.
The magnesium sulfite dryer should be cocurrent and
have internal chain sections and external rappers to dis-
171
lodge built-up product. Excessive sulfite carryover from
the dryer at Mystic illustrated the dangers of not having
cocurrent flow.172'173'174
The bucket elevator for transporting dried magnesium
sulfite to storage should have built-in excess capacity, so
that a product surge will not cause it to overload. The
elevator proved to be a bottleneck during prototype opera-
tions, causing the whole system to operate at less than
capacity.175'176
The dryer off-gas system should include a cyclone
separator for removal of entrained magnesium sulfite; the
"cleaned." gas stream should then be vented, into the sulfur
dioxide scrubber for further magnesium sulfite removal.
Finally, sulfite collected in the cyclone separator should
be discharged through a rotary valve into a pneumatic
g systarn and t- to z'-.c..?qe.
-------
3.4.5.2 Technological Transfer from the Pulp and Paper
Industry - Magnesium-based pulping operations have been in
operation in the pulp and paper industry since the 1940's.
The magnesium process (known as Magnefite) is the most
amenable to regeneration of the four primary pulping systems
(magnesium, calcium, ammonium, and sodium). As a rule waste
gas, 1.0 to 1.5 percent of which is S0~, is precooled to
43°C (110°F) to 66°C (150°F) prior to SC>2 absorption. The
scrubbing slurry consists of Mg(OH)2 and MgSO~ in water.
The exit gas normally contains less than 250 ppm SO,,,
representing 97.5 to 98 percent S0~ removal; and overall
operability is high. The MgSO., calcining takes place in a
reducing atmosphere. The regenerated MgO is slaked at 83°C
(180°F) to 94°C (200°F) prior to addition as makeup to the
178
absorber slurry.
3.4.5.3 Product Disposal - The magnesium oxide process is
a regenerable process, and as such does not present the
landfill and disposal problems associated with the "throw-
away" sulfur dioxide scrubbing processes (lime, limestone,
or double alkali). The magnesium sulfite produced, however,
must be regenerated. Points that must be considered in the
operation of an MgO system include the proximity of the unit
to a regeneration facility (if off-site); the freight costs
involved in transporting magnesium sulfite to that facility;
and the market for the resulting sulfuric acid (although the
sulfur dioxide, sulfur trioxide, and sulfur are other pos-
sible products).
3.4.6 Pilot Plant and Prototype Potential
Babcock and Wilcox (B&W) carried out a test program for
the EPA, developing the scrubbing step of the MgO FGD pro-
cess. They used a TCA absorber, which was later adapted for
3-249
-------
use at the successful MgO FGD system at Onahama Smelting and
Refining.
The two Chemico MgO FGD systems (Oil-fired at Mystic;
Coal-fired at Dickerson) were prototype units, built to
demonstrate the potentials of the process and to determine
the major areas for improvement. The venturi was adopted
for use in these prototypes. The units were built on a low
budget and include little redundancy, to enable a high
percentage of on-stream time. Considering this, they
achieved their purpose.
3-250
-------
REFERENCES FOR SECTION 3.4
1. Capsule Report, Flue Gas Desulfurization and Sulfuric
Acid Production via Magnesia Scrubbing, Environmental
Protection Agency, Washington, B.C. EPA-625/2-75-007.
pp. 2 and 6.
2. Survey of Flue Gas Desulfurization System at the
Dickerson Station, Potomac Electric Power Company,
PEDCo Environmental, Inc., Cincinnati, Ohio 45246.
Contract No. 68-02-1321, Task No. 6, U.S. Environmental
Protection Agency, Research Triangle Park, N.C. July
1975. pp. 3-4.
3. Ibid., pp. 3-5 and 8.
4. Op. cit. No. 1, p. 7.
5. Ibid., pp. 4-8.
6. Koehler, G.R. and E.J. Dober. New England SO2 Control
Project Final Results, EPA-650/2-74-126b, Environmental
Protection Agency, Washington, D.C., December 1974.
pp. 675-677.'
7. Op. cit. No. 1, p. 10.
8. Ibid.
9. Ibid.
10. Ibid.
11. Ibid., p. 12.
12. Op cit. No. 6, p. 680.
13. Op. cit. No. 1, pp. 10-11.
14. Op. cit. No. 6, pp. 684-685, 703-704.
15. Op. cit. No. 1, p. 11.
16. Ibid. pp. 10-11.
17. Op. cit. No. 6, pp. 684-685, 703-704.
3-251
-------
18. Ibid. pp. 688-692 and 707.
19. Ibid.
20. Op. cit. No. 2, pp. 3-5 and 8.
21. Quig, R.H. Chemico Experience For SC>2 Emission Control
on Coal-Fired Boilers, Coal and the Environment Tech-
nical Conference, October 23, 1974. pp. IV - 7,8.
22. Op. cit. No. 1, pp. 11 and 12.
23. Op. cit. No. 6, p. 678.
24. Quigley, C.P. and J.A. Burns. Assessment of Prototype
Operation and Future Expansion Study - Magnesia Scrub-
bing, Mystic Generating Station, EPA-650/2-74-126b,
Environmental Protection Agency, Washington, D.C.,
December 1974. p. 714.
25. ibid., p. 715.
26. Op. cit. No. 21, p. IV - 9.
27. Op. cit. No. 6, p. 679.
28. Op. cit. No. 1, p. 12.
29. Rosenberg, H.S. and D.A. Ball, et al. Status of Stack
Gas Control Technology - Battelle, Columbus Laboratories,
Columbus, Ohio. Report for EPRI to Kurt Yeager.
August 1975. pp. 3 and 10.
30. Op. cit. No. 1, p. 12.
31. Op. cit. No. 6, pp. 679 and 686.
32. Op. cit. No. 1, p. 12.
33. Op. cit. No. 6, pp. 693.
34. Ibid., pp. 687 and 704.
35, Ibid., pp. 673, 697, and 699.
36. Ibid., pp. 693-695.
3-252
-------
37. Erdman, D.A. Mag-Ox Scrubbing at the Coal-Fired
Dickerson Station, Potomac Electric Power Company, EPA-
650/2-74-126b, Environmental Protection Agency, Wash-
ington, B.C., December 1974. p. 730.
38. Taylor, R.B., P.R. Gambarani, and D. Erdman. Summary
of Operations of the Chemico-Basic MgO FGD System at
the Pepco Dickerson Generating Station, EPA-600/2-76-
136b, Environmental Protection Agency, Research Tri-
angle Park, N.C. May 1976. p. 737.
39. Op. cit. No. 37.
40. Op. cit. No. 38.
41. Ibid.
42. Ibid., pp. 737-738.
43. Ibid., p. 738.
44. Koehler, G. Magnesia Scrubbing Applied to a Coal-Fired
Power Plant, Chemico Air Pollution Control Company, New
York, New York 10001. EPA-600/7-77-018, U.S. Environ-
mental Protection Agency, Research Triangle Park, N.C.
March 1977. pp. 32-33.
45. Op. cit. No. 37.
46. Op. cit. No. 44, pp. 33-34.
47. Ibid., pp. 34-40.
48. Ibid., p. 80.
49. Ibid.
50. Op. cit. No. 38, p. 740.
51. Op. cit. No. 44, pp. 78-81.
52. Ibid., p. 49-
53. Op. cit. No. 21, p. IV - 6.
54. Op. cit. No. 44, pp. 5-6.
55. Ibid., pp. 76-79 and 83.
3-253
-------
56. Op. cit. No. 21, p. IV-6.
57. Op. cit. No. 37, pp. 736-738.
58, Op. cit. No. 44, pp. 58-61.
59. Op. cit. No. 38, pp. 738-741.
60. Ibid., pp. 738-740.
61. Op. cit. No. 37, pp. 731-732.
62. Op. cit. No. 44, pp. 5-6.
63. Op. cit. No. 37, pp. 731-732.
64. Op. cit. No. 44, pp. 74-75.
65. Ibid., pp. 48-52.
66. Op. cit. No. 38, pp. 738-739.
67. Ibid.
68. Op. cit. No. 2, pp. 4-8.
69. Op. cit. No. 38, pp. 741-742.
70. Ibid., p. 742.
71. Anz, B.M., C.C. Thompson, Jr. and J.T. Pinkston.
Design and Installation of a Prototype Magnesia Scrub-
bing Installation, United Engineers and Constructors,
Inc., Philadelphia, Pa. May 15, 1973. p. 2.
72. Gille, J.A. Magnesium Oxide Scrubbing at Philadelphia
Electric's Eddystone Station, Philadelphia Electric
Company, Philadelphia, Pa. EPA-600/2-76-136b, U.S.
Environmental Protection Agency, Research Triangle
Park, N.C. May 1976. p. 750.
73. Ibid., pp. 750-752.
74. Survey of the Flue Gas Desulfurization System at the
Eddystone Station, Philadelphia Electric Company.
PEDCo Environmental, Inc., Cincinnati, Ohio. Contract
No. 68-02-1321, Task No. 6. U.S. Environmental Pro-
tection Agency, Research Triangle Park, N.C. July 16,
1975. p. A-8.
3-254
-------
75. Op. cit. No. 72, p. 736.
76. Op. cit. No. 74, pp. 3-5 and 6.
77. Op. cit. No. 71, pp. 4-7.
78. Op. cit. No. 72, p. 757.
79. Ibid.
80. Ibid.
81. Ibid.
82. Ibid., pp. 751-752.
83. Ibid.
84. Summary Report - Flue Gas Desulfurization, Prepared by
PEDCo Environmental, Inc., Cincinnati, Ohio, Contract
No. 68-01-4147, Task No. 3, U.S. Environmental Pro-
tection Agency, Research Triangle Park, N.C. May-June
1977. pp. 185-186.
85. Op. cit. No. 72, p. 752.
86. Ibid.
87. Ibid.
88. Ibid.
89. Ibid., p. 753
90. Ando, J., Status of Flue Gas Desulfurization Technology
in Japan, EPA-650/2-74-126b, U.S. Environmental Pro-
tection Agency, Research Triangle Park, N.C. Dec.
1974. p. 134.
91. Ando, J. S02 Abatement for Stationary Sources in
Japan. PEDCo Environmental, Inc., Cincinnati, Ohio,
June 1976. Table 3-7.
92. Itakura, K., H. Ikuda, and M. Goto. Double Expansion
of Onahama Smelter and Refinery, Paper A 74-11, The
Metallurgical Society of AIME, 1974. pp. 4-6, 9 and
10.
3-255
-------
93. Ando, J. and G.A. Isaacs, S02 Abatement for Stationary
Sources in Japan, EPA-600/2-76-013a, Office of Research
and Development, U.S. Environmental Protection Agency,
Cincinnati, Ohio. January 1976. pp. 5-11, 12 and 13.
94. Evaluation of the Status of Pollution Control and
Process Technology-Japanese Primary Nonferrous Metals
Industry, PEDCo Environmental, Inc., Cincinnati, Ohio,
Contract No. 68-02-1375, Task No. 36, U.S. Environ-
mental Protection Aqency, Research Triangle Park, N.C.
July 1977. pp. B-89-92.
95.
96.
97.
98.
99.
100.
101.
102.
103.
Ibid.
Op. cit.
Op. cit.
Ibid.
Op. cit.
Ibid.
Op. cit.
Op. cit.
Matsuda,
No. 90, p. 133.
No. 94.
No. 93, pp 5-13.
No. 94, p. B-90-
No. 92, pp. 3 and 10.
S. Trip Report-Nonfe
PEDCo Environmental, Inc., Cincinnati, Ohio. Contract
No. 68-02-1321, Task 38, U.S. Environmental Protection
Agency, Cincinnati, Ohio. February 1977. Onahama
Smelter Section.
104. Semrau, K. Controlling the Industrial Process Sources
of Sulfur Oxides, American Chemical Society, Advances
in Chemistry Series No. 139. April 4-5, 1974. p. 7.
105. Op. cit. No. 90.
106. Ibid., p. 133.
107. Ibid.
10,3. Ibid.
109. Op. cit. No. 91, Section 6.
3-256
-------
110. Op. cit. No. 93, pp. 5-15.
111. Op. cit. No. 71, Appendix p. 2.
112. Telephone Conversation with Mr. Steve Achtner, Utili-
ties Sales Manager, Chemico Air Pollution Control,
August 19, 1977.
113. Op. cit. No. 44, pp. 91-92.
114. Ibid.
115. Ibid.
116. Ibid.
117. Op. cit. No. 44, pp. 93-95.
118. Ibid., p. 114.
119. Ibid., p. 96.
120. Ibid.
121. Lowell, P.S., F.B. Meserole, and T.B. Parsons. Pre-
cipitation Chemistry of Magnesium Sulfite Hydrates in
Magnesium Oxide Scrubbing, Radian Corporation, Austin,
Texas, Contract No. 68-02-1319, Tasks 36 and 54, U.S.
Environmental Protection Agency, Research Triangle
Park, N.C. June 24, 1977. p. 1.
122. Ibid., p. 5.
123. Ibid.
124. Ibid., p. 6.
125. Ibid.
126. Ibid.
127. Ibid., p. 7.
128. Ibid.
129. Ibid.
130. Ibid.
3-257
-------
131. Ibid.
132. Ibid., p. 9.
133. Op. cit. No. 38, pp. 738-740
134. Op. cit. No. 37, pp. 731-732,
135. Op. cit. No. 44, pp. 5-6.
136. Op. cit. No. 37, pp. 731-732.
137. Op. cit. No. 1, p. 12.
138. Ibid.
139. Op. cit. No. 24, p. 715.
140. Op. cit. No. 21, p. IV-9.
141. Op. cit. No. 6, p. 679.
142. Op. cit. No. 1, p. 12.
143. Op. cit. No. 29.
144. Op. cit. No. 44, pp. 74-75.
145. Op. cit. No. 72, p. 752.
146. Op. cit. No. 24.
147. Op. cit. No. 1, p. 11.
148. Op. cit. No. 37, pp. 731-732.
149. Op. cit. No. 38, p. 741.
150. Op. cit. No. 1, p. 11.
151. Op. cit. No. 37, pp. 731-732,
152. Op. cit. No. 6, p. 697.
153. Op. cit. No. 24, p. 718.
154. Op. cit. No. 1, p. 11.
3-258
-------
155. Op. cit. No. 21.
156. Op. cit. No. 71, p. 7.
157. Op. cit. No. 6, pp. 679 and 686.
158. Op. cit. No. 1, p. 12.
159. Op. cit. No. 6, pp. 679 and 686.
160. Survey of the Flue Gas Desulfurization System at the
Dickerson Station, Potomac Electric Power Company,
PEDCo Environmental, Inc., Cincinnati, Ohio, Contract
No. 68-02-1321, Task 6, U.S. Environmental Protection
Agency, Research Triangle Park, N.C. June 1975. pp.
4-7 and 8.
161. Op. cit. No. 38, p. 740.
162. Op. cit. No. 21.
163. Op. cit. No. 24, p. 723.
164. Op. cit. No. 148.
165. Op. cit. No. 6, pp. 679 and 686.
166. Op. cit. No. 37, p. 733.
167. Op. cit. No. 38, p. 739.
168. Op. cit. No. 44, pp. 7-8.
169. Op. cit. No. 38, pp. 738-739.
170. Op. cit. No. 2, pp. 4-8.
171. Op. cit. No. 44, pp. 5-6.
172. Op. cit. No. 21.
173. Op. cit. No. 3, p. 742.
174. Op. cit. No. 44, p. 2.
175. Op. cit. No. 21, p. iv-9.
3-259
-------
176. Op. cit. No. 148.
177. Op. cit. No. 44, p. 2.
178. Slack, A.V. and G.A. Hoilinden. Sulfur Dioxide
Removal from Waste Gases. Second edition. Noyes Data
Corporation, Park Ridge, N.J. 1975. pp. 227-229.
3-260
-------
3.5 THE WELLMAN-LORD PROCESS
This scrubbing system uses an aqueous sodium sulfite
solution to absorb S02. Sodium bisulfite is formed, and the
S0~ is released in a forced-circulation evaporator-crystal-
lizer. The regenerated sodium sulfite is crystallized and
dissolved for recycle to the absorber. The water vapor in
the concentrated S02 stream is condensed, and the resulting
stream is recovered as liquid S02, sulfuric acid, or elemental
sulfur. This system is used on combustion and other S09
emission sources.
A typical Wellman-Lord system as applied to a combus-
tion process is shown in Figure 3-64.
1 2
3.5.1 Process Description '
Primary removal of fly ash and other particulates is
generally provided by an ESP prior to the flue gas stream
entering the SO,., absorber. A variable-throat venturi pre-
scrubber provides cooling of flue gases to 49 to 54°C (120
to 130°F), additional particulate removal, and primary
chloride and sulfur trioxide removal. The need for an ESP
may be avoided by use of a larger, more efficient venturi;
however, the process vendor, Davy Powergas, does not recom-
mend this because of the possibility of reduced mechanical
reliability. '
This recovery process is based on the chemistry of the
sodium sulfite/bisulfite system. After appropriate pre-
treatment, the flue gas containing S02 enters the absorber
where the S09 is contacted countercurrently in the absorber
by a sodium sulfite solution. As the sodium sulfite absorbs
and reacts chemically with the SC>2, forming the more-soluble
sodium bisulfite, the solution becomes less saturated. This
means that fouling and scaling are unlikely to occur in the
absorber.
3-261
-------
CLEAN
FLUE GAS
U)
I
COOL IMS
WATER
FLY ASH PURGE
TO POND
KEY:
1 FOflCED DRAFT FAN
2 ORIFICE CONTACTOR
3 ABSORBER (3 TO 5)
4 ABSORBER SURGE TANK
5 FILTER
6 EVAPORATOR-CRYSTALLIZES
7 DUMP-DISSOLVING TANK
8 CONDENSER
9 CHILLER CRYSTALLIZER
10 CENTRIFUGE
11 DRYER
12 STORAGE BIN
13 A8SQRBE8 FEIDER TANK
S02 VAPORS
TO SULFUR
RECOVERY
DRIED SULFATE
PRODUCT
Figure 3-64. Typical Wellman-Lord SO- Recovery Process
-------
Oxygen and sulfur trioxide in the flue gas react with
the sodium sulfite, forming the unreactive sodium sulfate/
bisulfate. The presence of the unreactive species requires
a purge from the absorber to maintain the level of reactive
sodium sulfite and to reduce the possibility of scaling
u-, 6,7
problems. '
The absorber tower has a series (typically three to
five) of valve and collector trays made of Hastalloy G
(bottom trays) or 316L stainless steel. Because of the
extremely low L/G ratio 0.8 to 1.35 1/m (6 to 10 gal/1000
acf) required for absorption, each set of trays requires
recirculation of the sodium sulfite absorbing solution. This
recirculation provides the necessary amount of liquid
Q
required to load the valve trays hydraulically . The sodium
bisulf ite-rich solution is discharged from the bottom of the
absorber into a surge tank. It is then pumped to the sulfur
recovery area for regeneration.
A mist eliminator removes entrained liquid droplets
from the cleaned flue gas stream. The gases then enter a
reheater to eliminate the steam plume, to provide buoyancy
to the exhaust gas, and to minimize corrosion of the stack
9
from condensate.
The chemical reactions in the S02 absorber are as
follows :
50?
react to form sodium bisulfite.
S02 + Na2S03 + H20 ->-
Oxidation - Some oxidation of sodium sulfite to sodium
50? Absorption - Sulfur dioxide and sodium sulfite
form sodium bisulfite.
2 NaHS03 (1)
sulfate occurs.
2Na9SO-. + O? ->• 2Na_SO, (2)
The sodium ion makeup reactions are as follows:
3-263
-------
Sodium carbonate (soda ash) or sodium hydroxide (caus-
tic) reacts with sodium bisulfite to regenerate the S02
absorbent, sodium sulfite,,
Na2C03 + 2NaHS03 + 2Na2S03 + H20 + CC^i (3)
NaOH 4- NaHS03 ->- Na2S03 + H20 (4)
The product solution from the absorber is pumped
through filters for removal of any fly ash or other par-
ticulates . A small side stream of the filtered solution is
pumped to the purge treatment area for sodium sulfate
removal. The bulk of the stream is pumped to the evaporator
for regeneration as sodium sulfite.
Thermal regeneration of the bisulf ite-rich absorbing
solution is accomplished in a forced circulation vacuum
evaporator. The sodium bisulfite in the feed solution is
indirectly heated by low pressure steam and converted to
sodium sulfite by driving off SO^ and water vapor. The
sodium salts remaining in the evaporator circulating liquid
now form a crystalline slurry, the major component of which
is sodium sulfite. The concentration of crystals in this
evaporator circulation slurry is kept at a predetermined
level to minimize heat transfer area fouling. The slurry is
withdrawn continuously to the dissolving tank, where the
condensate from the evaporator is used to dissolve the
crystals. The resulting solution is mixed with makeup
sodium solution and pumped back to the top tray of the
absorber.10'11'12'13'14
The overhead SO^-water vapor mixture is concentrated
through a series of partial condensers. Depending upon the
size of the chemical plant, which is based on the amount of
sulfur dioxide absorbed,, the first partial condenser for a
portion of the S02~water vapor may be the heat exchanger for
a second stage evaporator. As the size of the chemical
3-264
-------
plant increases somewhere beyond the production of 2268 kg
(5000 Ibs) of S02 per hour, operating economics favor a
double-effect evaporator. A double-effect evaporator may
reduce steam consumption by 40 to 45 percent. The condensers
following the evaporators may be air- or water-cooled,
depending on the process absolute pressure, product purity
required, and cooling medium available. The product SO2 , at
least in the case of a power plant, will be used to make
elemental sulfur or sulfuric acid; it is therefore advan-
tageous to have the S02 concentration as high as possible.
Normally, this product gas will be approximately 85-volume-
percent SCK, with the balance being water vapor. '
The chemical reactions in the regeneration plant
are:17'18'19
Decomposition of the sodium bisulfite to sodium sulfite
so
2NaHS03 o NaS0 + H0+ + SC>+ (5)
2.
A small amount of the sodium bisulfite may react to
form other species when heated as follows:
Sodium pyrosulfite
2NaHS00 A Na0S0Oc + H00+ (6)
J ->- 2 2 D 2
Sodium thiosulfate (Na^B^C).-. )
Reaction unclear A Na S?0_ (7)
Sodium dithionate (Na-S.O,)
/ 4 o
Reaction unclear A Na2S4
-------
Sulfuric Acid - Sulfur dioxide reacts with oxygen in the
presence of vanadium pentoxide catalyst to
form sulfur trioxide
2S02 + 02 + 2S03 (9)
The sulfur trioxide reacts with water to
form sulfuric acid.
S03 + H20 + H2S04 (10)
Sulfur - Methane (natural gas) reacts with sulfur dioxide
to form hydrogen sulfide.
2CH4 + 3SO2 + 2C02 + 2H20 + 2H2S + S (11)
The hydrogen sulfide reacts with more sulfur
dioxide, forming water vapor and sulfur.
2H2S + S02 + 2H2<3 + 3S (12)
The overall reaction is:
CH4 + 2S02 ^ C02 + 2H2° + 2S (13)
Other commercially available reductants that can
be used in place of methane (natural gas) are
carbon monoxide, hydrogen, higher hydrocarbons up
through propane, and the products of coal gasifica-
tion (low-, medium-, and high-Btu coal gases).
Liquid S0? - The sulfur dioxide vapors are dried with silica
gel to remove moisture. The vapors then are
compressed and condensed; the resulting liquid
is collected and stored in pressurized tanks.
The purge streams go to a chiller crystallizer. With
controlled crystallization, the sulfate precipitates in a
much greater proportion than the other sodium compounds.
Thus, the solid phase has a high sulfate concentration while
the liquid phase has a higher sulfite concentration. The
solid phase is removed in a centrifuge and the liquid is
recycled to the process. The sulfate can be dried for sale
3-266
-------
or disposal, or can be neutralized and discharged as an
innocuous effluent.23
Table 3-43 represents a material balance for the
Wellman-Lord system as applied to a 500-MW, 3.5 percent-
sulfur coal-fired plant. The flow streams refer to Figure
3-65.
3.5.2 Wellman-Lord FGD Systems in the United States
Seven Wellman-Lord systems are currently in operation
in the United States as listed in Table 3-44. Six of these
units are installed on sulfuric acid or Glaus sulfur re-
covery units. The gas flow rates on these are small, 51,000
to 133,000 nm3/hr (30,000 to 78,000 scfm), in comparison
with a new 500-MW coal-fired boiler with a gas flow rate of
1,700,000 nm3/hr (1,000,000 scfm). The inlet S02 concentra-
tions on the seven small units range from 2700 to 10,000
ppm, which is about one to five times the SC>2 concentration
expected with a 3.5-percent-sulfur coal-fired boiler. The
smaller units, however, operate on fairly clean, dry streams
with low oxygen concentrations in comparison with boiler
flue gases.25'26
The S02 removal efficiency of the six units is typically
90 percent or greater; removal efficiencies in excess of 97
percent are reported. The collected S02 is either recycled
to the sulfuric acid or Glaus sulfur unit. Little opera-
tional data are available; it is reported, however, that the
six units have absorber on-stream times of greater than 97
percent.27'28
The No. 11 unit at the D.H. Mitchell Generating Station
of Northern Indiana Public Service Company is currently the
only utility operation in the United States, and also it is
29
the only coal-fired application. It is described further
in the next Section. In addition, three Wellman-Lord systems
3-267
-------
I
NJ
00
Figure 3-65. Flow diagram of the Wellman-Lord SO recovery process
and a sulfur removal unit.
-------
Table 3-43. MATERIAL BALANCE FOR A WELLMAN-LORD FGD SYSTEM
SERVING A 500-MW BOILER
Notes: 1. M indicates 1,000
2. Calculations based upon the following:
a. 105 percent stoichiometric sodium
b. 9,000 Btu/kWh for the unit
c. 12,000 Btu/pound of coal
d. 3.5 percent sulfur coal (dry basis)
e. 92 percent of the sulfur in the coal evolves as S02
f. 14 percent ash in the coal (as fired basis)
g. 75 percent of the ash in the coal evolves as fly ash
h. 99.0 percent removal of particulates by the ESP
OJ
I
Stream No.
Description
Rate, Ib/hr
scfm
gpm
Particulates , Ib/hr
Temperature, °F
Specific gravity
Viscosity, cps
Undissolved solids, %
PH
1
Coal to the
boiler
375 M
2
Combustion
air to air
heater
4,518 M
984 M
110
3
Combustion
air to the
boiler
4,518 M
984 M
535
4
Gas to the
economizer
4, 518 M
984 M
39.3 M
890
-------
Table 3-43 (Continued).
OJ
i
r-o
Stream No.
Description
Rate, Ib/hr
scfm
gpm
Particulates, Ib/hr
Temperature, °F
Specific gravity
Viscosity, cps
Undissolved solids, %
PH
5
Gas to
the ESP
4,518 M
984 M
39.3 M
310
6
Gas to
scrubber
4,518 M
984 M
393
310
7
Gas to SC>2
absorber
4,518 M
984 M
127
8
Gas to
reheater
4,518 M
984 M
127
9
Gas to
stack
4,518 M
984 M
170
11
Steam to
gas reheater
15 M
470
Stream No.
Description
Rate, Ib/hr
scfm
gpm
Particulates, Ibs/hr
Temperature, °F
Specific "ravity
Viscosity, cps
Undissolved solids, %
PH
12
Make up
water to
scrubber
202 M
403
13
Recycle
slurry to
scrubber
9,764 M
19.0 M
1.03
5
14
Scrubber
slurry to
surge tank
9,796 M
19.0 M
127
3
15
Particulate
slurry
purge
224 M
410
1.09
15
16
Pond water
recycle
190 M
381
-------
OJ
I
NJ
Stream NO.
Description
Rate, Ibs/hr
scfm
gpm
Participates, Ib/hr
Temperature, °F
Specific gravity
Viscosity, cps
Undissolved solids, %
pH
17
Make up
water to
mix tank
10. 6 M
21.2
18
Soda ash
to mix
tank
2,650
20
Make up
to dissolv-
ing tank
13.3 M
22.2
100
1.20
21
Feed solu-
tion to SO2
absorber
349 M
564
22
Recycle to
third
stage
2,052 M
3260
Stream No.
Description
Rate, Ib/hr
scfm
gpm
Particulates, Ib/hr
Temperature, °F
Specific gravity
Viscosity, cps
Undissolved solids, %
pH
23
Recycle to
second
stage
2,432 M
3,828
24
Recycle to
first stage
and product
2,820 M
4,408
25
Liquor to
surge
tank
370 M
580
26
Liquor to
chiller
214 M
334
68
27
Solution
to
chiller
1,395 M
2,607
25
1.07
28
Cooling
water-
refrigeration
389 M
777
-------
Table 3-43 (Continucr1) .
I
to
Stream No .
Description
Rate. lb/hr
scfm
gpm
Particulates ,
lb/hr
Temperature, °F
Specific gravity
Viscosity, cps
Undissolved solids,
%
pH
29
Slurry
centrifuge
214 M
334
35
30
Centrate
to heater
205 M
320
98
31
Cake to
conveyor
9,380
32
Steam
to heater
20.4 M
470
34
Sulfate
to bin
3,712
35
Liquor
to
evaporator
156 M
244
36
Feed to
evapora-
tor
120 MM
137 M
Stream No.
Description
Rate, lb/hr
scfm
gpm
Particulates,
lb/hr
Temperature, °F
Specific gravity
Viscosity, cps
Undissolved solids.
pH
37
Slurry
from
evaporator
120 MM
137 M
38
Slurry
to
heater
120 MM
137 M
39
Feed
to
heater
120 MM
137 M
40
Steam
to
heater
230 M
220
41
Slurry
evaporator
recycle
120 MM
137 M
43
Vapor to
primary
condenser
216 M
70.8 M
44
Water to
condensers
11,136 M
22.2 M
-------
nabic 3--"3 (Continued).
U)
I
NJ
Stream NO.
Description
Rate, Ib/hr
gpm
Particulates, Ib/hr
Temperature, °F
Specific gravity
Viscosity, cps
Undissolved solids, %
PH
45
Steam to
stripper
7,336
220
46
Condensate
to
scrubber
10,326
20.6
47
Condenate
to dis-
solving
tank
191 M
382
48
Gas to
compressor
21.4 M
49
Condensate
738
50
Gas to
SC-2
reduction
20.6 M
250
Stream No.
Description
Rate, Ib/hr
scfm
gpm
Particulates, Ib/hr
Temperature, °F
Specific gravity
Viscosity, cps
Undissolved solids,
PH
51
CH4 to
S02
reduction
3,306
1,130
200
52
Gases to
reheater
23.9 M
3,410
53
Recycle
to
dryer
55.5 M
220
54
Steam
to
steam
plant
9. 5 M
250
55
Boiler feed
water to
condensers
13.7
116
230
57
Sulfur
to
storage
9,342
10.4
59
CH4 to
incinera-
tor
243
83
-------
Table 3-43 (Continued).
Stream NO.
Description
Rate, Ibs/hr
scfm
gpm
Particulates , Ib/hr
Temperature, °F
60
Air to
incinerator
8,813
1,930
70
62
Water to
compressor
148 M
296
63
Steam to
sulfur storage
3000
290
-------
Table 3-44. OPERATIONAL WELLMAN-LORD FGD SYSTEMS IN THE U.S.
24
Company/ location
Olin Chemical*
Paulsboro, New Jersey
Std. Oil of California
El Segundo, California
Allied Chemical
Calumet, Illinois
Olin Chemical
Curtis Bay, Maryland
Std, Oil of California
Richmond, California
Std. Oil of California
El Segundo, California
Northern Indiana Public
Service
Gary, Indiana
Completion
date
July 1970
September 1972
November 1972
May 1973
August 1974
January 1975
December 1976**
Feed gas
origin
Sulfuric acid
plant
Glaus plants
Sulfuric acid
plants
Sulfuric acid
plants
Claus plant
Claus plant
Coal-fired
boiler
(115 MW)
Gas flow.
1000 Nm3/hr
76
51
51
133
51
51
527
(scfm)
(45,000)
(30,240)
(29,850)
(78,046)
(30,000)
(30,000)
(310,000)
Design
S02 -concentration,
ppm
in 6,000
out 500
in 10,000
out 250
in 2,700
out 250
in 4,000
out 250
in 10,000
out 250
in 10,000
out 250
in 2,200
out 200
Disposition
of SO2
Recycle to acid plant
Recycle to Claus plants
Recycle to acid plants
Recycle to acid plants
Recycle to Claus plant
Recycle to Claus plant
Elemental sulfur
UJ
I
Ul
* Plant operation suspended as of January 1, 1976.
**First integrated operation of the plant was in December, 1976; however, the Wellman-Lord system
began operations in July 1976, before completion of the Allied sulfur recovery unit.
-------
are currently under construction, (Table 3-45). '
The largest Wellman-Lord installations are being con-
structed on four coal-fired boilers at Public Service
Company of New Mexico's San Juan Station, Fruitland, New
Mexico. The retrofit installations on Units 1 and 2 should
start-up in early 1978; however, an explosion in Boiler 2
has delayed start up of that unit until mid-1978. The in-
stallation on the new No. 3 boiler is expected to be com-
pleted in 1979. Table 3-46 provides further details on
these units. The coal that is burned has the following
specifications: sulfur - 1.3 percent; ash - 22.4 percent;
moisture - 14,8 percent? and heating value - 18,870 kJ/kg
(8100 Btu/lb). On all three units, the Wellman-Lord system
is preceded by an ESP. Because of the size of the units,
multiple absorber modules are being installed. Rectangular
absorbers are used with tile-lined concrete walls that are
resistant to corrosion from the circulating solution. The
systems are designed for 90-percent SO.., removal, or better
and elemental sulfur will be recovered from the SOp collected.
The prescrubber blowdown will be used to condition the fly
ash for hauling by truck and disposal by burial in an adjacent
coal mine area.34'35'36'37
A more recent Wellman-Lord project is an installation
for ARCO/Polymers at Monaca, Pennsylvania, where a single
scrubber will receive the flue gas from three coal-fired
39 40
boilers with a combined generating capacity of 100 MW.
3.5,2.1 Northern Indiana Public Service Co. (NIPSCQ) - D.H.
Mitchell Unit No. 11 System Description - The 115-MW unit at
F-LPSO"- is the first application of a Wellman-Lord system to
a ^.oal-fired utility boiler, (Figure 3-66) . Flue gases from
the boiler pass through an ESP for primary particulate
removal (about 98.5%-efficiency). The booster blower delivers
3-276
-------
Table 3-45. WELLMAN-LORD FGD SYSTEMS UNDER CONSTRUCTION IN THE
UNITED STATES32'33
Company/location
Public Service Co. of
New Mexico, San Juan
Station
Fruitland, New Mexico
Public Service Co. of
New Mexico, San Juan
Station
Fruitland, New Mexico
ARCO/Polymers
Monaea, Pennsylvania
Feed gas
origin
Coal-fired boilers
(700 MW total)
Coal-fired boilers
(184 MW total)
Coal-fired boilers
(100-MW)
Gas flow,
1000 NmVhr
3058
889
425
(scfm)
(1,800,000)
(523,000)
(250,000)
Design S02
concentration ,
ppm
in 850
out 85
in 850
out 85
in 2000
out 250
Disposition
of SO2
Elemental suflur
Elemental sulfur
Unknown
u>
I
-------
CO
I
to
^J
CO
Table 3-46. DESIGN PARAMETERS FOR WELLMAN-LORD FGD INSTALLATIONS
38
AT SAN JUAN STATION OF PUBLIC SERVICE COMPANY OF NEW MEXICO
Boiler rating, MW (Gross)
Fuel rate, kg/hr
(Ib/hr)
Gas flow, m /a scfm at 26.7°C
(80°F)
Inlet SO,, concentration, ppm
Minimum
Maximum
Average
Inlet particulate loading
kg/108 J
Ub/106 Btu)
g/m
(gr/scf dry)
No. of scrubber trains
m /s
(acfm/ train)
Heating rate, kJ/kWh
(Btu/kWh)
Without scrubber
With scrubber
Sulfur regenerated, kg/hr (ton/hr)
Design
Actual
Unit 1
350
181,440
(400,000)
416.2
(882,000)
490
1200
850
0.0030
(0.07)
0.121
(0.053)
4
207.2
(439,000)
10,223
(9,682)
10,680
(10,115)
2041 (2.25)
852 (0.94)
Unit 2
350
181,440
(400,000)
373.7
(792,000)
490
1200
850
0.0043
(0.10)
0.18
(0.079)
4
200.1
(424,000)
10,203
(9,663)
10,660
(10,096)
2041 (2.25)
852 (0.94)
Unit 3
550
303,458
(669,000)
603.7
(1,279,400)
490
1200
850
0.0039
(0.09)
0.13
(0.056)
5
240.8
(510,000)
10,233
(9,691)
11,215
(10,621)
2041 (2.25)
1660 (1.83)
Unit 4
550
303,458
(669,000)
603.7
(1,279,400)
490
1200
850
0.0039
(0.09)
0.13
(0.056)
5
240.8
(510,000)
10,233
(9,691)
11,215
(10,621)
2041 (2.25)
1660 (1.83)
Each boiler will have an additional spare train to provide reserve capacity and improve
reliability.
-------
COOLING \
WA TER
FLUE GAS
FROM ESP ON
UNIT NO. 11
BOOSTER BLOWER
ORIFICE CONTACTOR
ABSORBER
ABSORBER SURGE TANK
EVAPORATOR-CRYSTALL HER
6 DUMP-DISSOLVING TANK
7 CONDENSER
8 SO; COMPRESSOR
9 CHILLER CRYSTALL/ZER
10 CENTRIFUGE
11 DRYER
12 STORAGE BIN
13 ABSORBER FEED TANK
VAPORS
DRIED SULFATE
PRODUCT
Figure 3-66. Schematic of the Wellman-Lcrd SO -
46
recovery process at NIPSCO.
3-279
-------
the flue gas through a variable-throat venturi prescrubber
to the absorber. The flue gas is cooled and saturated in
the prescrubber by water/slurry recirculated from the bottom
of the prescrubber and back to the venturi sprays. Fly ash
captured by the scrubbing solution is purged continuously
from the system to the fly ash pond0 For short periods, the
scrubber has the capability of handling considerably greater
than design fly ash loading. Water from Lake Michigan is
used for the makeup of water lost through the purge and
41,42,43,44
evaporation processes.
The absorption of the SO- from the prescrubbed flue gas
takes place in a three-stage absorber. Each stage consists
45
of a valve tray and a collector tray.
A sodium sulfite solution absorbs and chemically reacts
with the S02 to form sodium bisulfite. A mist eliminator
removes entrained liquid droplets from the gas exiting via
the absorber stack. There is a direct-fired, natural gas
reheat system in the absorber stack so that cleaned gas can
be reheated to 82°C (180°F) , if necessary, for dispersion of
the steam plume. '
The product solution collected on the bottom collector
tray of the absorber overflows to the absorber surge tank.
From this tank, the solution is pumped through a filter to
ensure that no fly ash will enter the evaporator system. A
small sidestream of the filtered solution is sent to the
purge treatment area to remove the sodium sulfate. The
purge treatment equipment consists of four chilled-wall
crystallizers; a slurry of sodium sulfate crystals forms in
these crystallizers and is then removed in a centrifuge.
Th'2 bulk of the product solution is pumped to the evaporacor
for regeneration of the sodium sulfite.49'50
3-280
-------
The evaporation system consists of a forced-circulation
vacuum evaporator. The filtered solution is recirculated in
the evaporator, where low-pressure (345 kPa, 50 psig) steam
is used to evaporate the water from the sodium bisulfite
solution. When sufficient water has been removed, sodium
sulfite crystals form and precipitate. Sulfur dioxide is
removed with the overhead vapors. The slurry formed by the
sodium sulfite crystals is withdrawn continuously to the
dump/dissolving tank, where condensate from the evaporator
is used to dissolve the crystals in the solution that is
51 52
pumped back to the top stage of the absorber. '
Water vapor is removed from the SO- in water-cooled
condensers. The S02 is compressed by a liquid ring com-
pressor for introduction to the Allied Chemical S0~-reduction
facility- The gas stream is about 85 percent SO», the
53 54
remainder mostly water vapor. '
Sodium lost as sulfate in the purge treatment system is
replenished by the addition of sodium carbonate to the
absorber solution. Soda ash is brought to the plant in
trucks and transferred to the storage bin by a pneumatic
conveying system. It is metered to the slurry tanks by a
bin activator and belt feeder. The soda ash slurry is
pumped to the absorber feed tank by parallel centrifugal
pumps.55'56
The small sidestream of filtered solution from the ab-
sorber is pumped to four chilled-wall crystallizers, where
sodium sulfate crystals form. The crystallized slurry is
centrifuged to extract the sodium sulfate crystals, and
clear solution is returned to the evaporator feed system.
The sodium sulfate crystals are melted and fed to a steam-
heated dryer, whose discharge product is then stored in
a bin until loaded in trucks for shipment. Any gases that
3-281
-------
evolve from the purge treatment are chemically scrubbed and
57
vented to the atmosphere.
The compressed S02 is fed to the Allied Chemical S0_-
reduction plant, where it is reacted with natural gas. The
resulting elemental sulfur is condensed and stored in molten
form for shipment. The off-gases are burned in a tail gas
incinerator to convert any hydrogen sulfite to S09 and
58 59
returned to the absorber inlet. '
Distinguishing features - This unit is characterized by
the following:
0 Fired with 3.15 to 3.5-percent-sulfur coal.
0 Most of the fly ash is removed in the ESP (primary
particulate removal) with some additional fly ash
removal in the wet venturi prescrubber.61,62
0 The absorber has three valve trays with separate
recirculation pumps for each stage.^3
0 The S02 is recovered as elemental sulfur. '
Process guarantees - The contracts between NIPSCO and
EPA and the other principals, Davy Powergas and Allied
Chemical, contain penalty-assessable guarantees for the
following:
0 SO2 emission levels - the process, according to the
contracts, will be at least 90-percent efficient for S02
removal when firing coal with a sulfur content of up to 3.5
percent (approximately 2300 ppm by volume of SO- in the
stack gas). As an override, no more than 200 ppm by volume
of S02 will be present in the exit gas.66'67'68
0 Mechanical reliability - The process plants will be
mechanically sound.
0 Utilities consumption - The aggregate cost of steam,
electric power, and natural gas required for operation of
3-282
-------
the complex will not exceed a specified cost; i.e., $56.00
per hour based on: '
Electric power - $0.007/kWh
Steam - $0.50/453.6 kg (1000 Ib) at 3792 kPa gauge
(550 psig), 399°C (750°F)
Natural gas - $0.55/1056 million J (million Btu)
0 Chemical requirement - The quantity of makeup chemi-
cals (caustic soda or soda ash) will not exceed a fixed
daily amount. The average chemical makeup over a 12-day
operating period at an average of 92 MW shall be no greater
than 6.6 ton of Na2C03/day.72'73
0 Product quality - The sulfur product is to be at
least 99.5-percent pure sulfur, and of a quality suitable
for use in the manufacture of sulfuric acid by the contact
74,75
process.
Operating history - Initial start-up of the NIPSCO
Wellman-Lord absorber was on July 19, 1976. An extended
shutdown period began on November 28, 1976, when high-
pressure steam supply failures from the boiler and from
emergency back-up systems to the FGD plant resulted in
freeze damage to the FGD plant. The shutdown lasted until
early January 1977. During the period from July through
November 1976, the Unit 11 boiler operated for 121 full days
and 10 partial days, while the SO- removal system of the FGD
plant operated for 71 full days and 23 partial days and was
down for 38 days. The steam supply failures previously
mentioned were responsible for 28 of the 38 days that the
system was down.
During the three sustained operating periods (Table
3-47), the absorber demonstrated the capability of greater
SO- removal than specified in the performance criteria. The
3-283
-------
77
concentrations are shown in Figures 3-67, 3-68, and 3-69.
Table 3-47- NIPSCO OPERABILITY DATA78'79
Duration
Run number (days) Period
1 15 Sept. 25 through Oct. 9, 1976
2 11 Oct. 13 through Oct. 23, 1976
3 14 Nov. 15 through Nov. 28, 1976
Acceptance 16 1/2 Aug. 29 through Sept. 14, 1977
[12 days at 151.0 actual m3/s
(320,000 acfm) (92 MW equivalent)
and 83 hours at 18.40 actual m^/s
(390,000 acfm) (110 MW equivalent)
During runs 1, 2, and 3, the booster fan delivered the
flue gas at the fixed rate of 544,000 actual nm3/hr (320,000
acfm), the 92-MW design level, to attain the acceptance test
levels. The load on Unit 11 varied from 60 MW to 108 MW.
For safety reasons the multileaf stack damper was open
during the operating periods; this allowed additional flue
gas with lower SO,, concentrations from Unit 6 to be pulled
83
across the stack to the absorber.
A boiler-related mishap on Jan. 15, 1977 caused Unit 11
to be shut down for repairs until May, 1977. The Wellman-
Lord unit was placed in service again on June 13, 1977. The
acceptance trail began on August 29, 1977, and was success-
fully completed on September 14, 1977. The Wellman-Lord FGD
system was officially accepted from Davy Powergas by NIPSCO
on September 16, 1977.
The 1-year demonstration run commenced September 15,
1977. During the trial, the evaporator and the absorber
both operated without interruption, except for two periods
(19 hours total) when the blower booster fan was shut down
3-284
-------
J,Ui)U
2,000 -
I
1.000 -
200 ---
5 10
RUN DURA TION, days
15
tXlRAPOLA TED THROUGH DA Y5 3 AND 5 UECA (7i£
OF I NOPE R A 7 / VE INS TR UMEN1A TION.
Figure 3-67. Inlet and outlet SO,., concentrations
during run no. 1.
80
2,000
1,500
7,000
500
200
OUT
5 10
RUN DURATION, day*
Figure 3-68. Inlet and outlet SO,., concentrations
, 0 81
during run no. 2.
3-285
-------
3,000
2,000 -
1,000 -
5 10
RUN DURA TION, days
THE POOR SO 2 RECOVERIES DURING THIS PERIOD
RESULTED FROM POOR QUALITY SOLUTION CAUSED
BY MECHANIC A L PROBLEMS IN THE SODA ASH FEED
SYSTEM AND EVAPORA TION AREA, AND LOW FEED
RATES TO THE ABSORBER WHILE BALANCING TANK
INVENTORIES.
Figure 3-69. Inlet and outlet SO,., concentrations
during run no. 3.
82
3-286
-------
by erroneous signals from a NIPSCO-installed pressure switch.
Several minor boiler-related and FGD system problems occurred,
but they caused no SCU-reinoval difficulties.
The only other interruption of the integrated (Wellman-
Lor-i and Allied units) operation was an 8-hour outage of the
Allied sulfur-recovery unit, caused by excessive pressure
drop across the reactor. This caused no problem in SO
d£
removal, however, because of the storage capacity of the
absorber surge tanks.
The inlet flue gas contained as high as 2800 ppm S0_;
the normal inlet SO- concentration, however, ranged from
2100 to 2300 ppm. Flue gas volume exceeded that expected.
Outlet S02 concentrations normally ranged from 170 to 190
ppm, demonstrating the capability of SO~ removal in excess
of 90 percent. During one 4-hour period, SO,, removal
efficiency was 89 percent; and over the same period at the
end of the acceptance trial a removal efficiency in excess
of 90 percent was obtained.
For both the 12-day run at 544,000 actual m /hr (320,000
acfm) (92 MW equivalent) and the 83-hour run at 662,000
actual m /hr (390,000 acfm) (110 MW equivalent) the average
S02 removal efficiency was 91 percent. The average S02
emission was 0.002 kg/1000 million J (0.5 Ib/million Btu
input), which is 42 percent of the current new source per-
formance standard (NSPS) of 0.0055 kg S02/100 million J
input) (1.2 Ib S02/million Btu input).
The makeup chemical requirement guarantee was 5988 kg
(6.6 ton) of soda ash (Na2C03) per day. The actual usage
was 5625 kg (6.2 ton) of soda ash per day, 94 percent of the
guaranteed maximum.
The absorber was designed for a flue gas stream contain-
3-287
-------
ing 5.2 percent excess oxygen, resulting in a purge of
sodium sulfate of 6350 kg (7 ton)/day. Actual excess oxygen
during the acceptance trial was 7 to 7.5 percent, with a
maximum of 8.5 percent. The average sodium sulfate purge
was 4082 kg (4.5 ton) per day, 64 percent of that expected
at design conditions although the actual excess oxygen was
about 40 percent in excess of design.
Utilities consumption was guaranteed not to exceed $56
per hour (1972 base costs). Actual utilities consumption,
when multiplied by the cost factors in the contract, was $43
per hour (1972 base costs), 77 percent of the guaranteed
maximum.
The required 99.5-percent minimum purity for the
recovered sulfur was met.
Operation problems/solutions - During the early operat-
ing period (July to August 1976), the FGD plant experienced
two major problems: 1) insufficient temperature control of
the low-pressure steam to the evaporator and reduction area,
and 2) insufficient turndown capabilities in the absorber
operation. A shutdown was scheduled to relocate the steam
desuperheater and to modify the absorber valve trays.
During the shutdown, an inspection of the absorber internals
revealed corrosion on the bottom surface of the lower-stage
collector tray, which had been installed without rubber
lining, and a seal failure where the trays connect to the
absorber walls. The shutdown was then extended so the
corroded tray surface could be lined with rubber [4 mm (5/32
in.) of tip-top rubber] and the tray-to-wall sealant could
be replaced.84'85
On two occasions during the operations period from July
through November 1976 the absorber was required to remove
particulate matter under the most severe conditions.
3-288
-------
During a 3-day period, NIPSCO experienced problems
with the ESP on Unit 11. The ESP was completely
out of service for about three hours, and operated
with only partial removal for the rest of the 3-
day period. The absorber continued to operate
through the period, but at reduced solution flow
rates in order to control the fly ash buildup in
the system. Following a brief power plant shut-
down to repair the ESP, the SC>2 removal unit was
started and leveled out, reaching normal operating
conditions in only 1 day. 86
On another occasion the flue gas from Unit 11 con-
tained a high loading of unburned coal. The Well-
man- Lord S02 Recovery Process continued to operate,
removing the particulate coal, but at reduced S02~
removal capabilities because of the lower absorber
solution feed
The sodium sulfate concentration level in the feed
solutions to the absorber was to be maintained at 7 to 8
percent. Because of mechanical problems in the purge treat-
ment area, however, the sulfate level reached 13 percent.
Problems were experienced with the purge dryer, which
restrained the crystallizer operation and prevented the
discharge of the dried sodium sulfate. The drive train,
feed box, and distribution problems in the dryer were cor-
rected. During the subsequent operating period an icy
crystal buildup formed on the crystallizer walls, restrict-
ing sulfate crystal formation. Several piping and control
changes were made to the refrigeration system and to the
crystallizers, which lengthened the operating period and
shortened the time needed to descale the crystallizers
go QQ
increasing the overall purge treatment operating cycle. '
A review of the operating problems from July through
November 1976 and the corrective actions taken is given in
Qn
Table 3-48.
During the acceptance trial, the only major problem was
3-289
-------
Table 3-48. PROBLEMS ENCOUNTERED AT THE NIPSCO WELLMAN-LORD INSTALLATION
(July to November, 1976)
93
OJ
I
KJ
i-D
O
Problem
No absorber turndown. The absorber
was to operate between 46 MW and
110 MW without dumping the liquid
from valve trays.
The collector tray seals leaked at
the walls.
Absorber roof and stack joint leaks
occurred.
Temperature control of the low-pressure
steam was inadequate.
There was corrosion on the bottom
surface of the lower collector tray.
SC>2 analyzer sample probes became
plugged.
Piping changes were made to the stack
reheat system.
Low pressure occurred in emergency
steam supply piping to the FGD plant.
Solution
The valve trays were leveled to within
3.2 mm (1/8 in.) across a distance of 7.62 m
(25 ft). Some valve cap weights were changed
and some valves were replaced with a different
valve type.
The original seal material between the
metal tray and the tile wall of the absorber
failed. This material was removed and replaced
with packing and silicone caulking.
The gaskets between the top of the absorber
wall and cover and between the reheat venturi
and the stack were reinforced.
The desuperheater on the low pressure steam
line was relocated to give better steam
saturation and temperature controls.
The bottom surface of the collector tray
(exposed to flue gas) was sand-blasted
and lined with cured rubber.
Newly designed, traced, and air purged sample
probes have been installed, eliminating the
plugging.
The size of the ring header supplying natural
gas to the four burners was increased, and
new regulators were installed to maintain
steady gas pressure to all four burners.
NIPSCO removed the flow meter orifice in
the emergency steam line.
-------
the shut downs of the blower booster turbine because of an
erroneous signal from a pressure switch. A NIPSCO-installed
pressure switch, designed to protect the boiler and the
ductwork from excessive positive or negative pressure,
became plugged. When cooled rapidly by rain or gusting
wiiid, the gas inside the tubing contracted and gave a false
indication of negative pressure. This signal shut down the
blower booster turbine and opened the bypass damper to the
stack. The pressure switch was restored to proper operation
and the problem was eliminated.
3.5.3 Description of Foreign Installations
Seventeen Wellman-Lord installations are operating in
Japan (Table 3-49). As of mid-1976 all reported to be
achieving S02 removal efficiency greater than 90 percent
(some in excess of 98 percent) while maintaining an on-
stream time greater than 98 percent for the absorption area.
91 92
None of these units operates on coal-fired boilers. '
3.5.3.1 Japan Synthetic Rubber - The first Wellman-Lord
unit on an industrial boiler was installed on two oil-fired
(4.0 to 4.2 percent-sulfur fuel oil) steam boilers at the
Japan Synthetic Rubber plant at Chiba. Each boiler generates
84 metric tons/hr (185,000 Ib/hr) of steam. The inlet SC»2
concentration is 1500 to 2000 ppm; the outlet S02 concentra-
tion is less than 200 ppm. The flue gas rate is 211,000
m3/hr (124,000 scfm).94'95'96 A flowsheet of the unit is
given in Figure 3-70.
The flue gas is washed by the prescrubber installed in
97
the lower part of the three-stage absorption tower.
The greater part of solids and S03 formed in the boiler
is removed with the wash liquor. The wash liquor is fil-
tered, blended with the Na2S04 purge stream (discussed
later), and neutralized, before discharge from the system.
3-291
-------
Table 3-49. OPERATIONAL WELLMAN-LORD FGD SYSTEMS IN JAPAN
OJ
I
NJ
Company/ location
Japan Synthetic Rubber,
Chiba
Toa Nenryo,
Kawasaki
Chubu Electric Power,
Nagoya
Sumitomo Chemical,
Sodeguara
Japan Synthetic Rubber,
Yokkaichi
Kashima Oil,
Kashima
Toa. Nenryo,
Hatsushima
Toyo Rayon ,
Nagoya
Japan National Railway,
Kawasaki
Mitsubishi Chemical,
Mizushima
Kurashiki Rayon,
Okayama
Fuji Film,
Fujinomiya
Shin Daikyowa ,
Yokkaichi
Sumitomo Chemical ,
Niihama
Mitsubishi Chemical,
Mizushima
Mitsubishi Chemical,
Kurosaki
Tohoku Electric Power,
Niigata
Completion
date
August 1971
August 1971
May 1973
November
1973
December
1973
February
1974
October
1974
December
1974
September
1974
April 1975
July 1975
1974
December
1975
February
1976
August 1976
September
1976
March 1977
Feed gas origin
Oil-fired boilers
Claus plants
Oil-fired boiler
(220 MW)
Oil-fired boiler
Steam boiler
(2% S fuel oil)
Claus plants
Claus plant
Oil-fired boiler
Steam boiler
(200 MW-3% S
fuel oil)
Oil-fired boiler
Oil-fired boiler
Oil-fired boiler
Oil-fired boiler
Oil-fired boiler
Oil-fired boiler
Oil-fired boiler
Oil-fired boiler
Gas flow.
100 mVhr
200
67
620
360
450
30
17
330
700
628
410
150
400
155
628
530
380
(scfm)
(124,000)
(41,000)
(390,000)
(225,000)
(280,000)
(20,200)
(10,000)
(218,000)
(435,000)
(373,000)
(248,000)
(89,000)
(253,000)
(91,000)
(390,000)
(330,000)
(236,000)
SO2 concentration,
ppm
in 2,100
out 200
in 6,500
out 200
in 1,600
out 150
in 1,550
out 250
in 1,000
out 100
in 11,000
out 200
in 18,580
out 250
in 1,500
out 150
in 1,500
out 45
in 1,500
out 150
in 1,500
out 150
in 1,300
out 125
in 1,500
out 150
in 1,600
out 130
in 1,300
out 130
in 1,500
out 75
in 1,000
out 100
Disposition of SO2
Sulfuric acid
Recycle to Claus plants
Sulfuric acid
Sulfuric acid
Sulfuric acid
Recycle to Claus plants
Recycle to Claus plant
Sulfuric acid
Sulfuric acid
Sulfuric acid
Sulfuric acid
Liquid SO,
Sulfuric acid
Sulfuric acid
Sulfuric acid
Sulfuric acid
Sulfuric acid
-------
UJ
I
to
S^ Oil-Fired
Afterburner
Sieve Tray
Absorption Tower
Fly ash
Evaporator
Cooling water
To sulfuric
acid plant
.Wastewater treatment
Sulfuric acid plant
N-
99
Figure 3-70. Flowsheet of Wellman-Lord process at Japan Rubber, Chiba, Japan.
-------
The partially cleaned gas rises in the tower while con-
tacting a countercurrent flow of concentrated sodium sulfite
solution that removes more than 90 percent of the inlet SO-.
98
The following scrubber operating parameters are typical:
Gas velocity (superficial), 1.65 m/s (5.4 ft/s)
pH, inlet 7.3 to 7.4
outlet 5.5
Solution concentration,
(percent by weight)
Na9SO~, inlet 20
^ J outlet 7
NaHSO,, inlet 2
outlet 20
Na_SO., inlet 5
outlet >5
Mist droplets are removed by the eliminator and demister
combination. The eliminator consists of a two-stage vertical
section housing; the first stage is a conventional chevron-
type and the second has a Tellerette packing. The demister
is a horizontal section also using Tellerette packing. The
gas is discharged at about 138°C (280°F) after reheating in
an oil-fired afterburner, requiring 5 percent of the total
energy consumption. The spent absorber, primarily NaHSO.,,
is discharged from the absorption tower and is stored in a
surge tank before being pumped to a two-stage evaporator.
Gas leaving the evaporator has a SO,., concentration of
90 percent after the water vapor is condensed in a cooler.
The concentrated SQ9 gas stream is sent to an acid plant.
Sodium sulfite is crystallized in the evaporator, separated
by centrifuging, redissolved in the condensate from the
cooler, and recycled to the absorber.
To reduce buildup of Na-SO. in the system, a portion of
the mother liquor from the centrifuge is purged. This purge
liquor is added to that from the cooler. The bleed taken
off in the cooler is required to remove contaminants (sodium
3-294
-------
pyrosulfite, thiosulfate, dithionate, etc.) which would
build up in the system. The purge streams undergo wastewater
treatment. The final wastewater is an essentially clear,
concentrated solution of sodium, which is discharged to the
102
bay.
This Wellman-Lord unit is noteworthy for its success-
ful operating history. It started operation in Jxme 1971
and operated at 97-percent availability during its first
year and 100 percent during the second and third years.
More than 90 percent of the S02 is removed, and only four
hours of an operator's time are required per 8-hour
shift.103'104'105'106
3.5.3.2 Chubu Electric Power - The first large installation
of a Wellman-Lord unit at a Japanese utility plant came on
line in late spring of 1973 at the Nagoya Station of the
Chubu Electric Power Company in Japan. The plant has the
capacity of treating 663,000 m /hr (390,000 scfm) of flue
gas, containing 2100 ppm SO,.,, from a 220-MW boiler which
burns oil containing 3-percent sulfur. It was designed and
constructed by Mitsubishi Kakoki Kaisha, a licensee of Davy
Powergas. This is a peak-shaving power station. The FGD
system was contractually required to handle stack gas
fluctuations from 35 percent to 105 percent of design flow
in 22 minutes, while maintaining outlet SO- emissions at
less than 150 ppm. In May 1973 the outlet S02 was tested by
the Japanese government at 130 to 135 ppm, which represents
over 93-percent S02 removal. '
The process flowsheet and the plant layout are shown in
Figures 3-71 and 3-72.
The flue gas from the boiler is cleaned in an ESP and
precooled to 58°C (136°F) before entering the main part of
the Wellman-Lord absorber. The L/G ratio is 0.7 1/nm (5.2
3-295
-------
STACK
VACUUM PUMP
OJ
I
to
^EVAPORATOR CENTRIFUGES
I ISUBR.INsk.Jl Lc-
~ THUFB 2?"l _J"£l
ozcmE
OXIDATION
NEUTRALIZING
TANK
CKILUR
Figure 3-71. Flowsheet of the Wellman-Lord (MKK) process at Chubu Electric,
109
Nagoya, Japan.
-------
STACK
oo
I
to
HASTE WATER
TREATING
FACILITY
ACID STORA6E TANK AREA
«««YSICTI
FtEB TASK
*'»e'«.
?^%>>..
**£ I'-
Fiaure 3-72. Layout of Wellman-Lord FGD plant at Chuba Electric, Nagoya, Japan,
110
-------
Inlet
Outlet
PH
7.3-7.5
5.5
Na2S
20
7
gal/1000 scf), and the pressure drop across the scrubber and
mist eliminator is 500 mm (2 in.) H20.
The pH and composition of absorbing liquor are as
follows:
, % NaHSO-,, % Na0SO. ,
3 24
3 5
22 5
The gas is then passed through two mist eliminator
units, the first a chevron above the sieve trays and the
second a packed vessel (Tellerette packing) in the dust
after the scrubber. The gas is reheated by an afterburner
to 140°C (284°F) in winter and 110°C (230°F) in summer.113
The scrubber effluent, rich in sodium bisulfite, is
heated in a double-effect evaporator by steam to crystallize
114
sodium sulfite and to regenerate SO-.
Sodium sulfite is separated by centrifuge, dissolved in
water, and returned to the absorber. Sulfur dioxide gas
containing steam is cooled in a condenser to separate water
and sent to a sulfuric acid plant. The tail gas from the
acid plant is treated by a small auxiliary scrubber for S09
115
removal.
The Chuba unit is built almost entirely of stainless
steel.116
A portion of the sodium sulfite is oxidized in the
scrubber by oxygen in the flue gas, to form sodium sulfate.
A small amount of sodium thiosulfate forms in the evapora-
tor. Neither the sulfate nor the thiosulfate can absorb
S02, and must be removed from the liquor. To remove the
sulfate, a portion of the liquor from the evaporator feed
tank is sent to a crystallizer; it is cooled to 0°C (32°F)
to crystallize the sulfate, which is then separated from the
3-298
-------
liquor by a centrifuge. Liquor from the centrifuge is
returned to the feed tank. Sulfate crystals containing a
small amount of sodium sulfite and bisulfite are dissolved
in a purge stream and combined with a small amount of sodium
sulfite and bisulfite, purge containing thiosulfate contam-
inant. The mixed liquor is treated with sulfuric acid to
decompose the sulfite, bisulfite, and thiosulfate; it is
then aerated (the SO2 released is sent to the auxiliary
scrubber), oxidized with ozone for further decomposition of
the thiosulfate, settled to precipitate solids, neutralized
with sodium hydroxide, and then discharged into the sea
along with other wastewater from power plants. The chemical
oxygen demand (COD) is kept below 10 ppm, as required by
regulations.
When the unit was originally put on line, the operators
inadvertently switched from low-sulfur (0.7%) to high-sulfur
(4%) fuel oil; however, there was little effect on the
outlet SO2 concentration, and because of the automated
adjustments to the process equipment at no time did it
118
exceed 150 ppm.
Ample absorber liquor surge capacity allows the evapor-
ator and acid plant to continue operation at minimum load at
119
weekends when the boiler and main scrubber are shutdown.
The SO? gas removed in the system is cooled in a con-
denser to remove water, and then used for acid production in
a sulfuric acid plant with a capacity of 73 metric ton/day
120
(80 TPD) was built as part of this installation.
The plant had some operating problems at first, includ-
ing vibration of the sieve assembly, difficulty in controlling
the draft at the I.D. fan at low load, and high turbidity of
the by-product acid all these, however, were solved in a few
months.121
3-299
-------
The Nagoya plant has proved reliable and easy to
control. The system is highly automated; only two operators
are required for the entire operation, including start-up
and shutdown. Automation is necessary, owing to the widely
fluctuating load and the fact that the boiler is shut down
. , 122,123
every weekend.
3.5.4 Engineering Design Parameters
Sulfur dioxide removal efficiencies in excess of 90
percent have been achieved with all the operating Wellman-
Lord FGD systems. Removal efficiencies in excess of 98
percent have been reported at Japanese units; they have
exceeded 97 percent at sulfuric acid and Clauss operations
in the U.S.; and 90 percent at the acceptance trials completed
on September 14, 1977, at NIPSCO. 124 ' 125< 126 ' 127 ' 128
To reduce the pressure drop required across the pre-
scrubber, and to maintain good prescrubber mechanical reli-
ability, Davy Powergas recommends the use of an ESP for
primary particulate control upstream of the prescrubber and
absorber. The prescrubber provides additional particulate
removal and saturation of the flue gas to the adiabatic
1 "3ft
temperature (approximately 49°C to 52°C, 120°F to 130°F) . '
The prescrubber may be constructed of rubber-lined mild
steel with acid-proof brick lining. In Japan, Nippon Steel
45M stainless steel (equivalent to 329 SS) was used for the
prescrubber. The L/G in the prescrubber has varied as a
function of particle size and the pressure drop. The L/G at
NIPSCO ranges from 800 to 1070 1/1000 m3 (6 to 8 gal/1000
1-51 1 O O
acf) at 1 kPa (4 in. H20) pressure drop. 'i<5
According to Davy Powergas, the L/G ratio for each tray
of the absorber ranges from 200 to 535 1/1000 m (1.5 to 4.0
gal/1000 acfm) . Davy Powergas currently favors about 260
1/1000 m3 (2 gal/1000 acfm). Each absorber will have three
3-300
-------
(for high inlet S0~ concentration) to five (low inlet SO.,
concentration) valve and collector trays. The overall L/G
for the absorber in a new installation will range from 800
to 1350 1/1000 m3 (6 to 10 gal/1000 acfm).133
The pH of the sodium sulfite absorbing solution is
maintained at approximately 6.0 at the absorber solution
inlet (in new designs, in the top tray of the absorber).
The pH of the purge stream from the absorber ranges from 4.4
to 4.5.134
The system makeup feed of fresh, 20-percent sodium
carbonate solution is approximately 67 1/1000 m of flue gas
(0.5 gallon/1000 acfm) per tray. All the fresh solution is
fed onto the top tray, so that chemical equilibrium is not
established and no theoretical lower limit to the exit S00
135
level exists.
For the valve tray assemblies within the absorber, the
design superficial gas velocity is 2.7 to 3.1 m/s (9 to 10
ft/s). The pressure drop across each absorber tray is
approximately 0.87 kPa (3.5 in. H-0). For three to five
trays, the pressure drop across the absorber itself is 2.6
to 4.4 kPa (10.5 to 17.5 H20).136'137
The absorber trays are Koch Multiventuri Flexitrays
(MYFT). The valve trays may be constructed of 316 SS, 316 L
SS, or Hastalloy G. Normally, Hastalloy G would only be
used for the bottom tray, where the most acidic conditions
occur.138'139
Scaling in the SO,-, absorber has never been reported as
140 141
a problem in any Wellman-Lord installation. '
The current maximum-size booster fan/scrubber/absorber
module which Davy Powergas recommends will handle approximately
3 142 143
8500 m /min (300,000 scfm) of flue gas. '
3-301
-------
As the S02 inlet concentration increases, the number of
mass transfer stages required to maintain a desired efficiency
may decrease. This is due to the higher driving force of
the more concentrated S02 gas stream. For example, for 90-
percent removal, an inlet S02 concentration of 1000 ppm may
require as many as five trays or mass transfer stages (e.g.,
San Juan). For the same 90-percent removal, an inlet S02
concentration of 3000 ppm may require only three mass
transfer stages (e.g., D.H. Mitchell). Each mass transfer
stage will add 0.6m to 1.5m (2 to 5 ft) to the absorber
height. Each additional stage added to the absorber accounts
for an approximate 10-percent increase in capital
. 144,145,146
cost. '
The water content of the flue gas for a given dry bulb
temperature will have a direct effect on the number of
stages required to satisfy the SO2 removal requirements.
The absorber will operate 2°C to 7°C (4°F to 12°F) higher
than the adiabatic saturation temperature of the flue gas.
Since this operating temperature of the absorbing section
will affect S02 equilibrium, a 2-volume-percent change in
water content of the flue gas may require one additional
mass transfer unit, since the saturation temperature will be
, . , 147
higher.
The chloride level in the absorber is maintained in the
range from 200 to 300 ppm to avoid corrosion of the stainless
steel internals. The chloride level is controlled by the
148
purge rate from the absorber and by the prescrubber.
The key component of the sulfur recovery area is the
forced-circulatory, double-effect evaporator. This is
constructed primarily of 316 L SS. The majority of the
remaining chemical plant components (partial evaporators,
condensate stripper, storage tanks, and related pumps and
3-302
-------
piping) are constructed primarily of 316 and 316 L SS and,
149
where possible, of mild steel.
The chloride level in the mother liquor stream may be
in the 2000-ppm range. This level has been measured at
NIPSCO without any observed serious damage to the compo-
nents. Again, the chloride level is controlled by removal
in the sulfate purge stream.
Several side-products arise from thermal regeneration
of sodium bisulfite solution in the evaporator to sodium
sulfite. For simplified reporting these products are often
expressed as thiosulfates„ The thiosulfate level is normally
maintained from 0.2 to 0.3 percent by careful temperature
and pressure control in the chemical plant and in the
bleed/purge stream. If the thiosulfate were permitted to
build up beyond the control range, it could break down to
elemental sulfur. Over an extended period, the sulfur
buildup could foul the absorber and evaporator surfaces.
The percentage of solids in the evaporator/crystallizer
is a control variable for the sodium sulfate regeneration
process.
The evaporator is often designed for 45 to 50 percent
solids (sulfite crystals). Evaporators have been operated
for prolonged periods at approximately 60 percent. The
limiting factors are formation of crystals on the heat
transfer surfaces and the viscosity, which limits the ability
152
to recirculate the slurry.
The evaporator may be either a single-or a double-
effect type. If less than 2268 kg (5000 Ib) of S02 per hour
is being handled, a single effect evaporator is normally
chosen. The operating conditions for a single effect
evaporator are 55 to 69 kPa (8 to 10 psia) operating pres-
sure and 88° to 99°C (190° to 210°F). A double-effect
3-303
-------
evaporator is normally chosen above 2268 kg (5000 Ib) of
S09/hr. The operating conditions are 27 to 41 kPa (4 to 6
psia) operating pressure and 71° to 79° (160° to 175°F).153'154
The storage or surge tanks in the chemical regeneration
plant are often designed with the storage capacity for 8 to
24 hours of full process stream flow. This enables the
regeneration plant to be shut down for maintenance without
interrupting SO- removal in the absorber section. The
necessity for extensive installed spare equipment is there-
fore reduced without sacrificing control reliability. ' '
The sulfate purge from the Wellman-Lord system can
amount to 5 to 10 percent of total sulfur removal. ' '
The size of the purge stream has increased with the in-
creasing size of new Wellman-Lord installations. Attempts
to prevent solution oxidation by the addition of antioxidants
were made, but were abandoned because of high cost and the
fact that purge would still be required. Selective removal
of the oxidized portion of the solution (sodium sulfate) by
chilled-wall crystallizers has been more successful, result-
ing in a five-to six-fold decrease in the purge stream and
the chemical makeup requirements. o^J-Dj
As higher S02 removal efficiencies are required, more
space is required for the chemical plant. One good aspect
of the Wellman-Lord process is that the absorption and
chemical plant areas can be separated as far apart as 800 m.
(0.5 mile) without any major capital or operating cost
increase.
3.5.5 Process Operability
In the United States, the six operating Wellman-Lord
units installed on sulfuric acid or Glaus sulfur recovery
units are reported to have an on-stream time of greater than
97 percent for the absorption area. In Japan in mid-1976
3-304
-------
the approximately thirteen Wellman-Lord units on oil-fired
boilers and Claus sulfur recovery units maintained an on-
stream time greater than 98 percent for the absorption area.
During this time S02 removal efficiencies of these units
were reported to exceed 90 percent, with some units report-
edly experiencing 97-to 98-percent S02 removal.165'166'167'168'159
The Wellman-Lord process has had S02 outlet concentra-
tions guaranteed in the range from 200 ppm (NIPSCO) to 70
ppm (San Juan 1 and 2). ' The corresponding removal
efficiencies would be 91 and 90 percent respectively.
Design outlet S02 concentrations within this range have been
met and have been below design in a number of cases. Most
recently, the S02 removal guarantee, the chemical usage
guarantee, and the utility usage guarantee have been met
during the acceptance trial at NIPSCO (8/29/77 through
9/14/77),172
Davy Powergas, the process vendor, when acting as
primary contractor, is willing to guarantee the minimum
percentage SO2 removal (e.g., 90-percent S02 removal)
and/or the maximum SO2 outlet concentration (e.g., 200 ppm
at NIPSCO), the chemical usage, the utility usage, the
particulate removal, and other items that the individual
173
client may request, depending upon the project parameters.
The only corrosion problem has been in the prescrubber/
absorber area. Use of rubber-coated 316 L SS or Hastalloy G
(44 percent Ni, 22 percent Cr, 20 percent Fe, 6.5 percent
Mo, 2.1 percent Cb + Ta, 2 percent Cu, and 0.05 percent C
max.) for the lowermost valve and collector tray protects
against the acidic conditions encountered there. At NIPSCO,
the rubber coating of the collector tray caused no unscheduled
. ., 174,175,176
shutdown. ' '
3-305
-------
No scaling has been reported in any of the 24 opera-
tional Wellman-Lord absorber units in the United States or
177,178
Japan.
The ability of several Wellman-Lord systems to adjust
to inlet flue gas composition variations is as follows:
0 At the automated Wellman-Lord unit at Chubu
Electric Power, Nagoya, Japan, shortly after the
unit went on line the operators inadvertently
switched from low-sulfur (0*7 percent) to high-
sulfur (4 percent) fuel oil. The outlet SC>2
concentration had been 130 to 135 ppm; after the
changeover, automated adjustments to the process
equipment kept the SC>2 emission below 150 ppm. 179
0 During a 3-day period, NIPSCO experienced problems
with the ESP on Unit No. 11. The ESP was com-
pletely out of service for about 3 hours, and
operated with only partial removal for the rest of
the 3-day period. The absorber continued to
operate through the period, but at reduced solu-
tion flow rates to control the fly ash buildup in
the system. Following a brief power plant shut-
down to repair the ESP, the SO2 removal unit was
restarted and leveled out, reaching normal oper-
ating conditions in only 1 day.180
0 On another occasion, the flue gas from Unit No. 11
contained a high loading of unburned coal. The
Wellman-Lord SO2 Recovery Process continued to
operate, removing the particulate coal, but
penalizing the SC>2 removal capabilities because of
the lower absorber solution feed rate.181
The Wellman-Lord units may produce sulfur,- liquid
sulfur dioxide, and/or sulfuric acid as salable products.
The market for sulfur and/or sulfuric acid must be considered
in the decision as to the form in which the SO- will be re-
covered. Often the product of choice is sulfuric acid.
The by-product of the Wellman-Lord process is sodium
sulfate, which is often dried for sale or for disposal. The
possible market for sodium sulfate or possible disposal
3-306
-------
problems must be investigated with regard to the location of
the individual Wellman-Lord unit.182'183'184'185
3.5.6 Pilot Plant and Prototype Potential
All inactive sodium salts must be purged from the
scrubber solution and their sodium ions replaced by active
sodium ions from soda ash or caustic soda. Research is
being conducted into methods for eliminating the purge
stream by reactivating the sodium ion. A number of pro-
cesses have been investigated in both laboratory and full-
scale applications, but few details have been disclosed.
Davy Powergas hopes to close the loop, by reactivating as
much of the inactive sodium ion as is economically feasible
and by generating sodium carbonate. In one process, the
purge stream would be incinerated to liberate the SO- and
recover the sodium salts, a concept similar to the recovery
boiler used in the pulp and paper industry. Two processes
are under development to reduce the sodium sulfate chemi-
cally, one of which uses natural gas and the other coal as
the reductant. Pilot plant testing has been successful. A
method employing high-temperature separation of the sulfate
has also been successfully tested on bench scale. ' '
The shortage of natural gas could be a major limitation
on the use of the Wellman-Lord process to recover elemental
sulfur. A 500-MW plant would require about 23 m /min (800
scfm) of natural gas, depending upon the sulfur content of
the coal and the efficiency of the recovery plant. High-
sulfur residual oils, coal gasification products, and
alternative hydrocarbon fuels such as propane have been
found to be useful as reducing compounds.
3-307
-------
REFERENCES FOR SECTION 3.5
1. EPA Technology Transfer Capsule Report - First Progress
Report: Wellman-Lord S02 Recovery Process-Flue Gas
Desulfurization Plant (Draft). EPA-625/2-77-011, U.S.
Environmental Protection Agency, Research Triangle
Park, North Carolina. May 1976. pp. 2-4.
2. Pedroso, R. An Update of the Wellman-Lord Flue Gas
Desulfurazation Process. In: Proceedings - Symposium
on Flue Gas Desulfurization - New Orleans, March 1977.
EPA-600/2-76-136a, U.S. Environmental Protection Agency,
Washington, D.C., May 1976. pp. 720-722.
3. Laseke, B.A., Davy Powergas Visit, Internal Memo, PEDCo
Environmental, Inc., Cincinnati, Ohio, June 14, 1977.
p. 2.
4. Christian, D., and H. Simmons. Power Plant Flue Gas
Desulfurization by the Wellman-Lord S02 Process. In:
Proceedings of the 12th Air Pollution and Industrial
Hygiene Conference on Air Quality Management in the
Electric Power Industry (Hal B.H. Cooper, Ed.). Center
for Energy Studies, The University of Texas at Austin,
January 28-30, 1976. p. 461.
5. Op. cit. No. 2, p. 720.
6. Ibid., p. 724.
7. Op. cit. No. 4, p. 466.
8. Hartman, J.S., Wellman-Lord FGD System conversation
with Davy Powergas. Internal Memo, PEDCo Environmental,
Inc., Cincinnati, Ohio. September 19, 1977.
9. Op. cit. No. 1. p. 2.
10. Ibid., pp. 2-3.
11. Op. cit. No. 2, p. 721.
12. Op. cit. No. 4, p. 462.
13. Bailey, E. Continuing Progress for Wellman-Lord S02
Process. In: See No. 48 EPA-650/2-74-126-b, U.S.
Environmental Protection Agency, Washington, D.C.,
December 1974. p. 757.
3-308
-------
14. EPA public hearing and conference on the Status of
Compliance with Sulfur Oxide Emissions Regulations by
Power Plants, Arlington, Va., Oct. 18-Nov. 2, 1973,
Transcript pp. 2434-2435.
15. Op. cit. No. 13, pp. 747-748.
16. Slack, A.V. and G.A. Hollinden. Sulfur Dioxide Removal
from Waste Gases, Second Edition, Noyes Data Corpora-
tion, Park Ridge, New Jersey. p. 209-
17. Op. cit. No. 1, p. 3.
18. Ando, J., and G. Isaacs. S02 Abatement for Stationary
Sources in Japan. EPA-600/2-76-013a, U.S. Environmental
Protection Agency, Washington, D.C., January 1976. pp.
5-7 and 5-8.
19. Ando, J. S02 Abatement for Stationary Sources in
Japan. PEDCo Environmental, Inc., Cincinnati, Ohio,
June 1976, Section 2.
20. Op. cit. No. 3, p. 3.
21. Op. cit. No. 4, pp. 463-464.
22. Op. cit. No. 2, pp. 721-722.
23. Op. cit. No. 13, p. 759.
24. Ibid.
25. Ibid.
26. Op. cit. No. 2, p. 728.
27. Ibid., p. 722.
28. Op. cit. No. 3, p. 1.
29. Op. cit. No. 14, p. 2445.
30. Op. cit. No. 13, p. 759.
31. Op. cit. No. 2, p. 728.
32. Op. cit. No. 13, p. 759.
33. Op. cit. No. 2, p. 728.
3-309
-------
34. Information provided by Public Service Company of New
Mexico in response to Edison Electric Institute Flue
Gas Desulfurization Questionaire. March 21, 1975. p.
2.
35. Op. cit. No. 4, pp. 487 and 491.
36. Op. cit. No. 13, pp. 745-748.
37. Op. cit. No. 34, 61 pages.
38. Ibid.
39. Op. cit. No. 2, p. 723.
40. Op. cit. No. 4, p. 492.
41. Op. cit. No. 1, p. 2.
42. Op. cit. No. 2, p. 722.
43. Mann, E.L. and R.C. Christman, Status of the Wellman-
Lord/Allied FGD system at NIPSCO's D.H. Mitchell Gen-
erating Station. From Proceedings - Symposium on Flue
Gas Desulfurization - New Orleans, March 1976. EPA-
600/ 2-76-136b. U.S. Environmental Protection Agency,
Washington, B.C., May 1976. pp. 704 and 705.
44. Op. cit. No. 14, p. 2445.
45. Op. cit. No. 43.
46. Op. cit. No. 1, p. 3.
47. Ibid., p. 2.
48. Mann, E.L. The Dean H. Mitchell Station (Northern
Indiana Public Service Company). Wellman-Lord (Davy
Powergas) - Allied Chemical Corporation SO2 Emission
Control Facility. In: Proceedings - Symposium on Flue
Gas Desulfurization - Atlanta, November 1974. EPA-
650/2-74-126-b, U.S. Environmental Protection Agency,
Washington, D.C., December 1974. p. 742.
49. Op. cit. No. 1, p. 2.
50. Op. cit. No. 48.
51. Op. cit. No. 1, pp. 2-3.
3-310
-------
52. Op. cit. No. 48.
53. Op. cit. No. 1, p. 3.
54. Op. cit. No. 48.
55. Ibid.
56. Op. cit. No. 1, p. 4.
57. Ibid.
58. Op. cit. No. 48.
59. Op. cit. No. l, p. 4.
60. Ibid., p. 5.
61. Ibid., p. 2.
62. Op. cit. No. 43.
63. Op. cit. No. 1, p. 2.
64. Op. cit. No. 48.
65. Op. cit. No. 1, p. 4.
66. Op. cit. No. 48, pp. 742-743.
67. LaKatos, S.F., A.W. Michner and W.D. Hunter, Jr.
Status of Demonstration Wellman-Lord/Allied Chemical
FGD Systems NIPSCO D.H. Mitchell Generating Station.
From Proceedings - Symposium on Flue Gas Desulfurization
New Orleans, March 1976. EPA-600/2-76-136b. U.S.
Environmental Protection Agency, Washington, D.C., May
1976. p. 713.
68. Op. cit. No. 1, pp. 5-6.
69- Op. cit. No. 48, p. 743.
70. Op. cit. No. 67.
71. Op. cit. No. 48, p. 743.
72. Ibid.
73. Ibid.
3-311
-------
74. Op. cit. No. 67.
75. Op. cit. No. 48, p. 743.
76. Summary Report - Flue Gas Desulfurization. Prepared by
PEDCo Environmental, Inc., Cincinnati, Ohio. U.S.
Environmental Protection Agency Contract No. 68-01-
4146, Task No. 3, June-July 1977. p. 177.
77. Op. cit. No. 1, p. 5.
78. Op. cit. No. 4, p. 492.
79. Op. cit. No. 1, p. 5.
80. Ibid.
81. Ibid.
82. Ibid., p. 6.
83. Ibid.
84. Op. cit. No. 76.
85. Op. cit. No. 1, p. 9.
86. Ibid.
87. Ibid.
88. Op. cit. No. 76.
89. Op. cit. No. 1, p. 9-
90. Ibid., p. 10.
91. Op. cit. No. 13, p. 759.
92. Op. cit. No. 2, p. 723.
93. Op. cit. No. 1, p. 10.
94. Op. cit. No. 14, pp. 2442-2443.
95. Op. cit. No. 4, p. 483.
96. Osborne, W. Jeff and C.B. Earl. Recent Experience of
the Wellman-Lord Sulfur Dioxide Recovery Process.
American Chemical Society, Advances in Chemistry Series
No. 139, April 4-5, 1974, p. 160.
3-312
-------
97. Op. cit. No. 16, p. 207.
98. Ibid.
99. Ibid., p. 208.
100. Ibid.
101. Ibid., p. 209.
102. Ibid.
103. Op. cit. No. 14, pp. 2442-2443.
104. Op. cit. No. 4, p. 483.
105. Op. cit. No. 96.
106. Op. cit. No. 13, p. 750.
107. Ibid.
108. Op. cit. No. 14, pp. 2436, 2443-2444
109. Op. cit. No. 18, pp. 5-4.
110. Ibid., p. 5-5.
111. Ibid., pp. 5-3 and 8.
112. Ibid., p. 5-7.
113. Ibid.
114. Ibid.
115. Ibid.
116. Ibid., p. 5-5.
117. Ibid., pp. 5-7 and 5-8.
118. Op. cit. No. 14, p. 2444.
119. Op. cit. No. 13, p. 750.
120. Op. cit. No. 18, p. 5-10.
121. Ibid., p. 5-8.
3-313
-------
122. Op. cit. No. 13, p. 750.
123. Op. Cit. No. 18, pp. 5-11.
124. Op. cit. No. 13, p. 759.
125. Op. cit. No. 2, p. 722.
126. Op. cit. No. 23, p. 1.
127. Hartman, J.S. Acceptance Trial at NIPSCO. Internal
memo, PEDCo Environmental, Inc., Cincinnati, Ohio,
September 19, 1977.
128. Op. cit. No. 2, p. 723.
129. Op. cit. No. 3.
130. Op. cit. No. 13, p. 756.
131. Op. cit. No. 8.
132. Hartman, J.S. Design Specifications of the Wellman-
Lord Process. Internal memo, PEDCo Environmental,
Inc., Cincinnati, Ohio, September 20, 1977.
133. Op. cit. No. 8.
134. Op. cit. No. 3, p. 3.
135. Hartman, J.S. Telephone conversation with Davy Powergas,
Internal memo, PEDCo, Environmental, Inc., Cincinnati,
Ohio, September 13, 1977.
136. Op. cit. No. 4.
137. Hartman, J.S. Davy Powergas telephone call. Internal
memo, PEDCo Environmental, Inc., Cincinnati, Ohio,
September 7, 1977.
138. Op. cit. No. 4.
139. Op. cit. No. 135.
140. Op. cit. No. 2, p. 720.
141. Op. cit. No. 14, pp. 2433, 2433a, and 2434.
142. Op. cit. No. 3, p. 4.
3-314
-------
143. Op. cit. No. 2, p. 726.
144. Op. cit. No. 3, p. 5.
145. Op. cit. No. 13, p. 752.
146. Op. cit. No. 4, p. 465.
147. Op. cit. No. 13, p. 753.
148. Op. cit. No. 3, p. 3.
149. Op. cit. No. 3.
150. Op. cit. No. 3, p. 3.
151. Ibid.
152. Op. cit. No. 132.
153. Op. cit. No. 13.
154. Op. cit. No. 132.
155. Op. cit. No. 2, p. 721.
156. Op. cit. No. 13, p. 747.
157. Op. cit. No. 14, pp. 2437-2438.
158. Ibid., p. 2448.
159. Op. cit. No. 48.
160. Op. cit. No. 16, p. 210.
161. Zada, F. Status of Flue Gas Desulfurization Processes
PEDCo Environmental, Inc., Cincinnati, Ohio, January
13, 1975. p. 113.
162. Op. cit. No. 14, p. 2448.
163. Op. cit. No. 2, pp. 724-725.
164. Op. cit. No. 14, p. 2437.
165. Op. cit. No. 13, p. 759.
3-315
-------
166. Op. cit. No. 2, p. 722.
167. Op. cit. No. 3, p. 1.
168. Op. cit. No. 127.
169. Op. cit. No. 2, p. 723.
170. Op. cit. No. 34.
171. Information provided by Northern Indiana Public Service
Company in response to Edison Electric Institute Flue
Gas Desulfurization Questionaire, March 27, 1975. p.
2.
172. Op. cit. No. 127.
173. Hartman, J.S. Davy Powergas Guarantees. Internal
memo, PEDCo Environmental, Inc., Cincinnati, Ohio,
August 29, 1977. pp. 1-3.
174. Op. cit. No. 4.
175. Op. cit. No. 135.
176. Perry, J.H. and C.H. Chilton, (Eds.). Chemical Engi-
meers1 Handbook, Fifth Edition, McGraw Hill Book Co.,
New York, N.Y. pp. 23-44 and 55.
177. Op. cit. No. 2, p. 720.
178. Op. cit. No. 14, pp. 2433, 2433a
179. Ibid., p. 2444.
180. Op. cit. No. 1, p. 9-
181. Ibid.
182. Op. cit. No. 2, p. 721.
183. Op. cit. No. 4, p. 462.
184. Op. cit. No. 13, p. 759.
185. Op. cit. No. 53. pp. 159-160.
3-316
-------
186. Op. cit. No. 2, pp. 724-725.
187. Op. cit. No. 13, p. 749.
188. Op. cit. No. 14, p. 2449-
3-317
-------
3.6 OTHER FGD SYSTEMS
This section describes other FGD systems that are not
widely used in the United States.
1234
3.6.1 Chiyoda Thoroughbred 101 Process ' ' '
In the Chiyoda process, SO- is absorbed in weak sul-
furic acid and then oxidized by air to sulfuric acid. Part
of the acid is neutralized with limestone to produce gypsum;
the remainder is circulated as SO- absorbent. Thirteen
utility oil-fired boilers in Japan, up to 350 MW in capa-
city, use the process with greater than 90 percent SO,,
c /- £
removal. ' A 23-MW prototype unit has been tested on a
coal-fired boiler at the Scholz Station of Gulf Power Company.
No full-scale systems have yet been installed on any U.S.
plants.
The Chiyoda process is a variation of the double alkali
process, which uses a clear solution in the scrubber. The
process has four basic operations: prescrubbing, absorbing,
cyrstallizing, and by-product/waste handling (Figure 3-73).
Following removal of most of the particulate matter by means
of an ESP, flue gas is quenched and cooled in a venturi-type
prescrubber, where fly ash and chlorides are removed.
Solids in the prescrubber liquid are removed in a thickener
and the liquid is recycled. Particle-free gas proceeds to a
fixed-bed scrubber, where SO- is absorbed in a counter-
current flow of weak sulfuric acid solution, containing a
ferric ion catalyst, to produce sulfurous acid. Flue gas
from the scrubber passes through a mist eliminator and
reheater, and is then discharged. Continuous withdrawal of
the absorbent maintains scrubber acid content at about 2.0
percent by weight. In a subsequent oxidation tower, air is
bubbled through the scrubber effluent, catalytically oxi-
dizing the S02 (by the ferric ion) to sulfuric acid. Part
of this liquor is neutralized with lime or limestone to
3-318
-------
produce gypsum; the rest is recycled to the absorber.
Chiyoda has developed a special gypsum crystallizer that
produces large crystals. Smaller gypsum crystals from the
overflow at the crystallizer and of the centrifuge overflow
are settled out in a clarifier and used as seed crystals.
Clarifier overflow is returned as mother liquor to the
absorber. A portion of this stream is purged to reduce the
buildup of soluble matter, such as chlorides. The basic
chemistry of the Chiyoda process is as follows:
H0 + S0 •> HS0
-> H2S04 (2)
H2S04 + CaC03 ->- CaS04 + H20 + C02 (3)
Based on Japanese operating experience with oil-fired
boilers, the Chiyoda 101 process has demonstrated a number
of advantages: relative simplicity, a high degree of reli-
ability, and the production of a high-quality gypsum by-
product. In addition, scaling is prevented in the absorber/
oxidizer or mist eliminator, since the scrubber liquid is
not saturated with calcium sulfate. Sulfur dioxide con-
centration in the tail gas can be 50 ppm or less, repre-
Q
senting 95 percent S02 removal. The main drawback of the
process is the high cost associated with its high liquid/gas
ratio. The weak acid solution cannot absorb and retain much
S02 per unit volume, and an L/G of 40 1/m (300 gal/1000
acf) for a 1000-ppm SO stream is required for 90-percent
-, 9
removal.
Chiyoda has developed a variation of this process for
their simultaneous NO -SO removal method known as Process
3C £f
102. Ozone, added to the entering gas stream, partially
oxidizes NO after prescrubbing and cooling. Both SO,, and
^C £~
3-319
-------
NO are then absorbed in the FGD unit, with part of the NO
x x
being converted to nitric acid and the remainder being
released, largely as nitrogen and nitrous oxide, with the
flue gas. Sixty-percent NO removal has been reported in a
-A.
Japanese pilot plant. No data on S02 removal capabilities
of Process 102 were found.
The first test of the Chiyoda 101 process on a coal-
fired boiler was conducted under the direction of Southern
Services, Inc., on Gulf Power Company's 23-MW Scholz plant
at Sneads, Florida. The system began operation in February
1975. Total availability for the remainder of 1975 was 60
percent. The system broke down for 6 weeks in August and
September, after which it operated at greater than 97-
percent availability to the end of the year. Typical S02
removal was 82 to 87 percent with relatively low-sulfur coal
12
(1.7 or 2.2 percent). The gas rate varied between 16.5
and 21.2 m /sec (35,000 and 45,000 scfm). Operation of the
system was terminated in March 1977.
The major operating problems occurred when both centri-
fuges became unbalanced because of a high, nonuniform wear
on the screw conveyors. The hard facing on the screws wore
much faster than was suggested by the Japanese experience,
possibly because a different facing technique was used.
Replacement of the screws alleviated the problem. Although
there were other, more common equipment problems, no scaling
occurred in the absorber, oxidizer, or mist eliminator.
Liquid waste streams from the plant were combined, neutral-
ized with limestone, and discharged to a settling pond,
whose overflow went into the plant ash pond and eventually
to a river.
3.6.2 Citrate Process13'14'15•16'17
The citrate process for flue gas desulfurization is a
buffered aqueous absorption process that uses sodium citrate,
3-320
-------
Reheater Absorber Oxidizer
I
U>
K)
Cleaned gas
Crystallizer
Limestone
Prescrubber \ Filter
Sludge
Centrifuge
Purge
to treatment
Figure 3-73. Flowsheet - Chiyoda Thoroughbred 101 process.
-------
or sometimes phosphate, as the buffering agent. The S02 is
absorbed in solution, then reacted with gaseous H2S to
precipitate sulfur and regenerate the solution for recy-
cling. Elemental sulfur is precipitated directly from the
scrubber solution. The process was developed mainly by the
U.S. Bureau of Mines; significant research was also con-
ducted by Pfizer-Peabody-McKee and Stauffer. The different
versions vary in such factors as absorber configuration,
absorber liquor pumping rates, type of buffer, operating pH,
and methods of H,,S production and sulfur separation. A
partially integrated system has been demonstrated.
The chemistry of the citrate process is very complex
and not yet fully understood. By buffering the absorption
solution at a pH of 3.0 to 3.7 with various organic acids,
principally citric acid, side reactions are controlled and
sulfur is produced without building up concentrations of
undesirable side products. Citrate suppresses side reac-
tions , yielding sulfur as the main product. Figure 3-74
illustrates the Bureau of Mines process; it is representa-
tive of various configurations. The S02-bearing gas is
first cleaned of particulate matter and sulfuric acid mist,
then cooled to saturation temperature. It is absorbed in an
absorption tower by a solution of sodium citrate, citric
acid, and sodium thiosulfate, to produce sodium sulfite.
Liquor from the absorber goes to a regeneration reactor,
where it is reduced with H^S, both to produce elemental
sulfur and to regenerate the solution. Sulfates are purged
from the liquor by cooling and crystallizing Glauber's salt
(Na2S04-10H2O), which is removed by filtration and then
washed. The sulfur is separated from the solution by oil
flotation and melting, and the regenerated solution is
recycled to the absorber. The Bureau of Mines process uses
3-322
-------
STACK GAS
REHEAT
FLUE GAS
FROM
PARTICIPATE
SEPARATOR
GO
I
GO
K>
CO
ABSORBER
OFF-GAS
TO BOILER
4 , REGENERATION
REACTOR
LOADED
LIQUOR
SULFATE PURGE STREAM
SULFATE
CRYSTALLI2ER
SULFUR
FLOATATION
UNIT
REGENERATED LIQUOR
REGENERATED LIQUOR
REGENERATED
LIQUOR
SULFUR SLURRY
SULFUR
MELTING
STEAM
SULFATE
BYPRODUCT
TO DISPOSAL
MOLTEN
SULFUR
SULFUR
TO ri2S
GENERATOR
PRODUCT
SULFUR
Figure 3-74. Flowsheet - Citrate process.
18
-------
kerosene for this flotation, whereas the Pfizer and Stauffer
processes use other proprietary methods to separate sulfur
from the product slurry. Flue gas from the absorber is
reheated and discharged. The H2S used in the process can be
produced by a variety of methods, including reacting two-
thirds of the product sulfur with natural gas and steam.
The simplified chemistry of the citrate process is as fol-
lows :
S02 + H20 £ HS03~ + H+ (1)
HSO~ + Na+ £ NaHSO3 (2)
Cit= or P04= + 3H+ £ H3Cit or H3PC>4 (3)
S02 + 2H2S -* 3S + 2H20 (4)
The major advantage of the citrate process is its
direct precipitation of sulfur from the scrubber solution.
Removal of 99 percent of SO- has been demonstrated, but
19 20
there is no effect on NO in the gas stream. ' Citrate
5C
is used as the buffering agent because of its good chemical
stability, low vapor pressure, adequate pH buffering capacity,
and the purity and physical character of the precipitated
sulfur. The Stauffer process employs phosphate: it is
cheaper and lasts longer than citrate. The resultant anhy-
drous Na-SO. is said to be a valuable by-product. The
citrate process is also efficient in terms of fuel con-
sumption —the incremental fuel requirements for a 500-MW
installation are estimated to be as follows: stack gas
reheat, 23.1 x 10 J/sec (39 MM Btu/hour); heat for anhy-
drous Na7SO by-product, 11.4 x 106 J/sec (39 MM Btu/hour).
Z t ,.
Tc tal incremental fuel consumption is 7.19 x 10 J/kWh (."82
Btu/kWh).
There are certain drawbacks to the citrate process.
Production of lS from methane and sulfur can be a problem,
3-324
-------
as a result of the shortage of natural gas (methane).
Further, the relatively low pH in the scrubber can result in
low efficiency, necessitating a large regeneration system,
if high bisulfite formation is attempted. Since the ab-
sorber is also large, retrofitting existing boilers with
citrate FGD systems may sometimes prove difficult. The
separation of coal fly ash, which is necessary to avoid
contamination of the product sulfur, further increases
costs. There is also a potential problem with hydrocarbon
losses from kerosene flotation.
Variations of the citrate process have been tested in
three pilot plant applications: a Bureau of Mines, 0.5-MW
plant on a slipstream from the Bunker Hill lead sintering
furnace in Idaho; the Pfizer 1-MW plant on a coal-fired
industrial boiler at Terre Haute, Indiana; and the Stauffer
plant on a low-sulfur, oil-fired boiler at Northeast Utili-
ties' Norwalk Harbor station in Connecticut, which used
phosphate instead of citrate. All three plants showed good
results. Kinetic models exist for scaling to larger sizes,
but have not been demonstrated.
The Bureau of Mines citrate plant operated for 4500
hours between February 1974 and November 1975, producing
more than 50 net tons of sulfur from a feed gas averaging
0.5 percent S02. The tail gas from the sinter plant was
first cleaned in a baghouse, cooled in a packed wet scrub-
ber, and then cleaned of acid mist and particulate in a wet
ESP. To better simulate conditions in a coal-fired boiler,
the gas steam was diluted to below 0.5 percent SC>2 (typi-
cally 0.07 to 0.13 percent). Kerosene or another hydro-
carbon oil was added for the sulfur flotation-separation
step; it was found that the use of air alone for flotation
produced a wet sulfur product of about 8 percent solids. A
glycolate scrubbing solution was satisfactorily substituted
3-325
-------
for citrate, at about half the cost, for about 450 hours of
the pilot plant operation. Because of the irregular opera-
tion of the sinter plant, the longest continuous run was 265
hours. The sulfur product at the Bunker Hill pilot plant
was of greater than 99.5 percent purity, with an ash content
of about 0.03 percent. Carbon content was 0.2 to 0.3 per-
cent with kerosene flotation, and less than 0.04 percent
21 3
with air flotation. With a gas flow of 0.47 m /sec (1000
scfm), SO0 removal exceeded 99 percent and feed gas ranged
22
from 70 to 550 ppm S09. Removal efficiency dropped to 95
323
percent for a 130-ppm S0~ gas at 0.61 m /sec (1300 scfm).
A number of problems arose at the Bunker Hill plant.
Incomplete regeneration of the absorbent solution was found
to be caused by insufficient retention time of the H~S gas
in the single-surface precipitation reactor. This was
solved by increasing contact time through the use of two
reactors, addition of sodium thiosulfate to inhibit oxida-
tion, and adding sodium carbonate to neutralize sulfuric
acid in the solution. An early problem with cloudy recycle
solution from kerosene flotation was caused by delayed
precipitation of colloidal sulfur. It was corrected by
increased contact time and by bleeding about 5 percent of
the absorption solution to the reactor effluent tank, where
it reacted with the excess absorbed H^S. There were also
some plugging problems, with sulfur in the pipelines between
the reactors, for example.
A demonstration plant of the Bureau of Mines Citrate'
Process is currently under construction at a coal-fired
power plant at the St. Joe Zinc Co. in western Pennsylvania.
3.6.3 Westvaco Activated Carbon Process ' '
The Westvaco process is a dry process featuring in situ
oxidation of S02 while it is still in the gas stream.
3-326
-------
Multistage fluidized beds are used for both adsorption of
SO- and for regeneration of the carbon by hydrogen reducing-
£•
gas, simultaneously with the production of elemental sulfur.
The major drawback is that it is a relatively energy-inten-
sive process with high hydrogen consumption. Reaction heat
is supplied by low-Btu gas from a gasifier or fuel oil* The
fluidized-bed approach permits the handling of relatively
large gas volumes. More than 90-percent removal has been
27
demonstrated from a flue gas containing 2000 ppm SO-.
In the Westvaco process, the catalytic and sorptive
character of carbon is exploited to convert S0« to sulfuric
acid within carbon granules. Carbon promotes the necessary
reactions, but does not directly take part in them. Figure
3-75 is a flow diagram of the process. The flue gas initially
contacts dry granular activated carbon at stack gas tempera-
ture. Shallow-fluidized beds, the number of which are de-
pendent on the boiler size, are used for 90-percent SO^
removal. The S0~ is removed from the gas through catalyzed
oxidation to SO_, followed by hydrolysis to sulfuric acid,
which remains sorbed in the carbon granules. In a second
fluidized-bed reactor, the carbon granules are reduced with
hydrogen sulfide at 150°C (300°F), producing elemental
sulfur and water vapor. Hydrogen sulfide is regenerated and
elemental sulfur removed by a thermal stripping process near
540°C (1000°F) in a third fluidized-bed reactor with an
external hydrogen source. The basic process chemistry is as
follows:
S00 + 1/200 + H~0 -> H9SO. (1)
^ Z Z Z *±
H2S04 + 3H2S -> 4S + 4H2 3H2S + S (3)
Because it is a dry process, the Westvaco process does
not involve the critical control of chemistry necessary in
3-327
-------
I
U)
to
CO
TO BOILERS
SULFUR GENERATOR
ACID CONVERTER
FLUE GAS So2
QLOWER ADSORBER
WATER
SPRAY
STEAM HEAT
EXCHANGER
^CARBON
PREriEATER
CARBON
MAKE'OF
SULFUR
CONDENSER
GENERATOR/
SULFUR
STRIPPER
SULFUR
FILTER
INERT GAS AND STEAM
WATER SPRAY
INERT GAS AND STEAM
Figure 3—75. Flowsheet - Westvaco activated carbon process.
28
-------
wet reactions. Both the separation of water from by-prod-
ucts and the reheating of stack gas are eliminated. In
addition, there is no sulfate problem and readily available
coal may be gasified for use as the reductant. Nor is high-
quality hydrogen essential; dilute gas from an air-blown
gasifier can be used if the tars are removed and carbon
monoxide is reduced to a low level. Removal of 99 percent
of the SO.., is possible with increases of only about 20
29
percent in adsorber size and pressure drop. Estimated
energy requirements for a 500-MW unit are: electric power,
9
13 MW; fuel oil, 79 x 10 J/hr (75 MM Btu/hr); steam, 68 x
108 J/hr (6.4 MM Btu/hr); and reducing gas, 18 x 1010 J/hr
(175 MM Btu/hr). Total incremental fuel requirements are
10.6 x 10 J/kWh (1000 Btu/kWh) equivalent to an energy
requirement of 50 MW.
Several potential design and operating problems are
involved in scaling-up the Westvaco process to commercial
size. Compared with most wet scrubbers, a fluidized-bed
unit operates at a higher pressure and at a lower gas veloc-
ity: which means it must have a larger diameter and a
higher power consumption. A 250-MW unit would require 2,323
2
square meters (25,000 ft ), a size large enough to create
problems for many retrofit installations. A hard, reactive
carbon absorbent is necessary to minimize raw material costs
and particulate emissions. Carbon fines from adsorbent
attrition in the highly agitated fluidized-bed reactors
present a potential particulate emission problem in the
exhaust, although control should be possible by moderate
pressure-drop cyclones. Removal of NO has not been demon-
^X
strated, although similar processes such as those of Bergbau-
Forschung and Reinluft have reported 40- to 50-percent
removal.
3-329
-------
The Westvaco process has been demonstrated on a slip-
stream of 0.16 m /sec (330 cfm) from an oil-fired industrial
boiler. The system featured three units, a sorber, an acid
converter, and a sulfur stripper/hydrogen sulfide generator.
The granulated carbon, nominal 12 x 40 mesh size with a
density of 641 kg/m (.40 Ib/ft )_ , flowed by gravity between
the units. Counter-current gas flows were used in the
reactors. The sorbing unit was cooled by direct water
spray, and the converter and stripper/generator were heated
by electrical resistance. Dust collection was performed by
cyclones and bag filters. The system operated for 350
hours, demonstrating better than 90-percent sulfur removal
from the inlet concentration of 2000 ppm SC>2. A maximum
desulfurization of 97 percent was obtained. ^ There was
only one period of downtime, caused by sulfur plugging, and
there were no corrosion or dew point problems. The granu-
lated carbon was recycled 21 times with only slight chemical
and mechanical carbon loss. Some initial carbon attrition
occurred, but it was reduced by using a combination of
larger particle sizes with improved hardness. There was no
evidence of undersired side reactions or trace contaminant
buildup. The carbon had no effect on the 150 ppm of NO in
jC
the flue gas; it appeared to suppress adsorption S0~ up to
an NO concentration of about 150 ppm. The by-product
2£
sulfur was of commercial grade, 99.7 percent pure, and
32
contained 380 ppm ash and 2500 ppm carbon. The carbon
33
could be readily filtered, yielding a 99.9 percent pure
sulfur. A preliminary, 15-MW prototype design has been
prepared for 90-percent SO- removal from a coal-fired boiler
34
using coal with a 3.5-percent sulfur content.
3-330
-------
3.6.4 Bergbau-Forschung/Foster Wheeler Dry Adsorption
D 35,36,37
Process ' '
The Bergbau-Forsclvung/Foster Wheeler (BF/FW) process
involves the adsorption of dry SO~ by a moving bed of acti-
vated char. The char is then thermally regenerated, and ;he
S02~rich gas is reduced to elemental sulfur in a proprietary
process using crushed anthracite coal. The system also
removes some particulate matter and NO from the gas stream.
.?c
Among the unique features of the process are the louvered,
moving-bed adsorber and the use of hot, inert sand to pro-
vide heat for char regeneration. Pilot plant studies have
•5 I
indicated 95-percent S0?, and 40- to 60-percent NO removal. '
^ -X
There are three basic process steps in the BF/FW proc-
ess: adsorption, regeneration, and reduction. The first
two were developed by Bergbau-Forschung and the latter by
Foster Wheeler. Figure 3-76 is a simplified flow diagram of
the process. In the two-stage adsorber, the flue gas passes
horizontally in cross flow through vertical columns of
activated char. Char in the beds is continuously recycled,
moving downward in mass flow and acting as a catalyst for
the adsorption of SO~, SO.,, NO , oxygen, and water vapor.
^ O X
Particulate matter is also collected on the char pellet
surfaces, which, because of their size and physical arrange-
ment, act as impingement filters. In the second step, the
activated char containing sulfuric acid is reduced by being
heated to 650°C (1200°F) in an inert atmosphere. Hot sand
that is not used in the reaction is used as the heat trans-
fer media. The sand/char mix flows downward through the
regenerator to be separated by a vibrating screen deck. The
char is spray-cooled to 104°C (220°F) and returned to the
adsorber; the sand is recycled to an oil-fired, fluidized-
bed heater. Particulate matter is physically separated from
3-331
-------
(TO BOILER)
_fi E S.Q X_I AI L_G AS_
STACK CAS
DILUT1J1M.
AIR
FLUE GAS
FROM
STACK
L
' 2 STAGE
ADSORBER
1.0. FANS
CRUSHED
ANTHRACIT E
COAL
S02-RICH GAS |
1' C-QflVEYOR- ¥
NT
ASH
PARTICULATE DISPOSAL
COLLECTION
EQUIPMENT
SERVICE WATERfr
N |
1
ERATOR
SAf
i FLL
I
y- -~*^
SAND HEATER
ASH DISPOSAL
SAND
HEATER
1
FUEL
STEAM
BOILER
FEED WATER
I
! \ REACTOR/ I^ONDENSER)
, / -^
SULFUR TO
STORAGE
CHAR COOLER
Figure 3-76. Flowsheet — BF/FW dry adsorption process!
39
-------
the saturated char before regeneration. The off gas from the
regenerator contains SO (25 to 40 percent by weight), CO- ,
water, and nitrogen. In the third step, Foster Wheeler's
RESOX process, the gas stream is introduced in counterflow
to a mass flow of crushed anthracite coal, thus reducing the
SOp. Molten sulfur is then recovered in an inclined shell-
and-tube condenser, removed to an insulated tank and stored
ar a liquid. The nitrogen and C02 in the gas stream do not
take part in the reaction. Tail gas from the condenser is
recycled to the boiler to oxidize the remaining sulfur
values to S02. The chemistry of the BF/FW process is as
follows :
+ H0 + HS0 (1)
H2S04 + 1/2C + H20 + 1/2C02 + SO2 (2)
2NO + C -> C02 + N2 (3)
S02 + C •> S + C02 (4)
As with other dry processes, BF/FW does not require
critical chemical control of wet reactions, separation of
water from by-products, or stack gas reheating. In addi-
tion, the design temperature for the adsorber, 96° to 149 °C
(205° to 300 °F) , is about normal for boiler flue gas after
an air preheater. In comparison with other FGD systems,
BF/FW is intermediate in terms of energy consumption, re-
quiring about 9.5 x 10 5 JAWh (900 Btu/kWh} for a 500-MW
plant. Other estimated inputs for a plant of this size are:
activated char, 1360 kg/hr (3000 Ib/hr) ; electric power, 8
MW; anthracite coal for the RESOX process, 17 x 10 J/hr
(160 MM Btu/hr) ; the oil-fired sand heater, 22 x 10 J/hr
(210 MM Btu/hr); and total incremental fuel consumption, 9.5
x 10 J/kWh (900 BtuAWh) . Total energy needs are equivalent
to approximately 10 percent of the power plant output «,
3-333
-------
Several potential problems are involved in scaling-up
the BF/FW process. Dilution of S02 by other gases is a
handicap in making sulfur. There is a need for reliable
solids-handling equipment that can operate up to 1500°F (the
temperature to which the sand is preheated before it is
mixed with the char in the regenerator). The materials-
handling subsystems for the adsorber and regenerator must
either be located near the stack or make use of expensive
conveying systems. In some cases this makes retrofit in-
stallations difficult. A shortage of activated char could
be a problem. High-efficiency ESP's are required to protect
the cleaning section from fly ash that has been separated
from the char out of the adsorber. An additional waste
stream is formed by fly ash, unreacted coal, and tars re-
moved from the bottom of the reduction vessel.
Elements of the BF/FW system have been demonstrated
both in Germany and in the United States. The adsorber and
regenerator were piloted by Bergbau-Forschung on a flue gas
slipstream from the coal-fired boiler at the Kellerman Power
Station in Liinen, Germany. A modified Glaus plant was used
to convert the S02-rich gas to elemental sulfur. About
6,000 hours of continuous operation indicated high desul-
furization and low pressure drop. Removal of up to 95
percent SO , 40 to 60 percent NO , and 95 percent inlet
40 x
particulate was demonstrated. The only effluents were
sulfur, cooling water, dry flyash, and small quantities of
char fines. The fully integrated BF/FW system has been
tested at the Gulf Power Scholz Steam Plant in Florida,
where pulverized coal is burned in the boiler. Pilot tests
have been conducted on simulated flue gas streams containing
41
various S02 concentrations. A high-efficiency ESP
removed 99.7 percent of the particulate matter in the gas
3-334
-------
stream, which contained 900 to 2150 ppm S09. Pressure
^
drops in the adsorber were lower than expected, but S09
removal was 96 percent or better. Poor char distribution in
the adsorber caused imbalances in the bed level and reduc-
tions in removal efficiency, and there were also problems in
the char feed, char/sand separation, hot sand conveying, and
char cooling sections of the regenerator. Two brief runs
made with the RESOX reduction system in 1975 indicated that
the reaction was controllable. Modifications to the system
are planned and testing will continue.
3.6.5 Consol Process43'44'45
The Consol process, developed by Consolidation Coal
Company, is an aqueous potassium formate process involving
SO- absorption followed by reduction of the product in
solution. Unlike other reduction methods, such as citrate
and Shell/UOP, which use sodium compounds for absorption,
the Consol technique employs potassium salts. The major
product of absorption is potassium thiosulfate (K^S^O-),
which is reduced in solution with CO at elevated temperature
and pressure to produce H2S. The H S can be used for pro-
duction of elemental sulfur by standard methods. The scrub-
ber has been tested on a 10-MW plant, whereas the regenera-
tion step has been operated only on the laboratory scale.
The Consol process is one of three that the EPA has selected
for a demonstration project, to be co-sponsored with a large
utility.
The chemistry of the Consol process is relatively
complex, and many side reactions can occur. Flue gas is
first quenched and humidified in a venturi scrubber, which
also removes S0_, HC1, and residual particulate matter
(Figure 3-77). The scrubber is a packed tower with a recir-
culation circuit through a reaction drum. The feed comprises
3-335
-------
U>
I
U)
u>
FLUE GAS
MAKE-UP
WATER
TO FLY ASH
VENTURP
POND 8r-
NEUTRALIZATION
KOH
or
FEED
A TO STACK GAS
.REHEATER
/
SCRUBBER
REACTION
DRUM
FILTER
CAKE
TO
REGENERATOR
SCRUBBER FEED
FROM REGENERATOR
•^-
Figure 3-77 Flowsheet - Consol scrubber.
-------
flue gas, spent scrubber product, and makeup potash. Vari-
ous scrubbing compounds are used, including KHS , K S ,
/— A.
KHCO , KCO3, and KOOCH . The key reaction in the scrubber is
reduction of potassium sulfite to thiosulfate by the follow-
ing process:
KHS + 2KHS03 -> 3/2 K S20 + 3/2 H20 (1)
The polysulfides are relatively minor components and are
consumed in the reaction drum. The KOOCH is a relatively
minor component and is essentially inert; it reacts with the
circulating scrub solution in the reaction drum. At a pH of
7, a residence time of less than 1.5 minutes is sufficient
to consume the K9S and KHS in the reaction drum completely.
^- X
The KHS must be completely consumed to prevent production of
H_S gas in the scrubber. Sulfur dioxide is removed from the
flue gas by reaction with potassium carbonates in the solu-
tion, as follows:
K2C03 + H20 + 2S02 -> 2KHS03 + C02 (2)
The scrubbing process is controlled by adjusting the feed
rate to produce an effluent pH of 7, and by maintaining the
proper proportions of KHS and K S to K-CO and KHCO
^ x £ j j
The basic reaction in the Consol regenerator is the
noncatalytic reduction of the thiosulfate scrubbing solution
with CO-rich gas:
3K2S203 + 12CO + 7H20 -> 2KHS + 4KHC03 + 4H2S + 8C02 (3)
Figure 3-78 is a flow diagram of the regeneration process.
Optimum reaction temperature is 232°C (450°F) at a pressure
of 3.4 to 6.9 MPa (500 to 1000 psig) . The use of a catalyst
to maintain the simplicity of the system was tested and
rejected. The acceptability ratio is controlled by the
proportion of CO,, in the reducing gas; the ratio of KHS to
KHC03 in the reduction product is controlled by the follow-
ing equation:
3-337
-------
CO
I
CO
CO
CO
SPENT SOLUTION
FROM SCRUBBER
CO_DEPLETED
r~GAS~~~'
REGENERATOR
450° F
750 PSIG
I I
SELECTIVE
H2S
REMOVAL
RECYCLE GAS
COMPENSATE w I
^1
30 -RICH SYNTHESIS GAS ^j^
m |
PARTIAL
OXIDATION
2500° F
850 PSIG
HI- SULFUR
^FUEL OIL
REGENERATED
^SOL'N. OUT™"
.H^i
CO 2
•H
, AIR
A L CLAUS
TAIL
CONDENSAT1
GAS
r
AIR
INCINERATOR
_HOT_
TO STACK
REHEAT
K2S04
OUT
SULFUR
OUT
Figure 3-78. Flowsheet - Consol commercial regenerator.^'
-------
KHS + CO, + H,0 -»• KHCO- + H0S (4}
^ £ -3 £
Final adjustment of the acceptability ratio to about 0.95 is
carried out by decomposition in a flash drum, simultaneously
releasing CO,,:
KHS + KHC03 •* K2C03 + H2S (5)
2 KHC03 -> K2C03 + C02 + H2° (65
Off-gas from the regenerator contains unreacted CO, C09 ,
£
H2S, and hydrogen. After cooling and steam condensation,
the H2S is selectively removed by scrubbing with a polar
organic solvent in one of several available proprietary
processes. The absorbed H_S-CO9 gas released during regen-
eration is combined with H,S-CO, gas from the flash drum and
sent to a Glaus plant for elemental sulfur production. The
Claus plant tail gas is incinerated with part of the scrub-
bed regenerator off-gas to produce hot combustion gas, which
is used for direct reheating of the scrubbed stack gas.
Carbon monoxide is produced from the regenerator off-gas by
a reverse water-gas shift in an oxygen-fired, partial-oxida-
tion unit.
Because of the high solubility of potassium thiosulfate
in water, the Consol scrubber can be operated with highly
concentrated salt solution. Sulfate problems are relatively
minor, because the low sulfite concentration maintained
inhibits oxidation to sulfate. Any K2S04 produced is
readily marketable.
The Consol scrubber has been demonstrated in a 0.47
m /sec (1000-acfm) pilot plant at the Consolidation Coal
Research Division in Library, Pennsylvania, and in a 14.2
m /sec (30,000-acfm) pilot plant at the Philadelphia Elec-
tric Cromby Station. The Library unit received feed-gas
from an oil-fired boiler; the Cromby installation was on a
slipstream from a coal-fired boiler. The regenerator has
3-339
-------
also been operated at Library/ including a closed-loop
integrated operation with the scrubber. Sulfur dioxide
absorption at Cromby varied from 88 to 99 percent, with the
higher rates achieved by improved liquid distribution,
smaller packing, and a higher gas rate. Sulfate formation
averaged about 1 percent of the S02 absorbed. There was no
appreciable NO absorption. In a 10-day run of the closed-
JX
loop, integrated operation, sulfur absorption averaged about
99 percent, with no accumulation of foreign material in the
product.
48 49 50 51
3.6.6 Aqueous Carbonate Process ' '
In the Aqueous Carbonate Process, developed by Atomics
International, SO- in the flue gas is absorbed by a very
concentrated sodium carbonate solution in a spray dryer.
The finely-divided sodium sulfite powder that results is
reduced with coke to regenerate the carbonate and to produce
hydrogen slufide. This can be used for the production of
elemental sulfur by standard methods. The process has been
designed for the removal of 90 to 95 percent of SO? from the
flue gases of coal-fired power plants. The carbonate
solution is said to provide effective S00 removal over wide
52
concentration ranges. In addition to its use of a spray
dryer for producing a granular salt mixture suitable for
regeneration, the process is also unique in that there is
complete reduction of the sodium salts in a molten pool.
There is no need for reheat, since the flue gas is not
saturated during contact with the solution. Individual
components of the system have been pilot tested, and the
process is also one of the three chosen for the EPA demon-
stration project.
There are six major process steps in the Atomics Inter-
national system: scrubbing, product collection, reduction,
3-340
-------
quenching and filtration, carbonation, and sulfur production
(Figure 3-79). The flue gas is precleaned to 0.22 g/m (0.1
gr/scf) flyash in cyclones or in an ESP- The spray dryer is
similar to those used in other applications for more than 50
years. Sodium carbonate solution is atomized by high-speed
centrifugal atomizers, and the fine mist absorbs S02 in
crossflow. The thermal energy of the water gas vaporizes
water without excessive saturation or gas cooling. Solid
sodium sulfite and sulfate granules are formed and removed
from the exit gas in cyclones or in an ESP. Because the
waste gas remains above the dew point, it can be vented
directly and without reheat. The reduction unit is a refrac-
tory-lined, steel vessel which contains molten sodium car-
bonate and sodium sulfide at 930° to 1040°C (1700° to 1900°F).
Dry feed from the spray dryer melts and mixes with the
pool, and the sulfite and sulfate are reduced by petroleum
coke to sulfide. Off-gas from the reducers is used as a source
of CO™ for later carbonation steps and as a source of pro-
cess heat. The molten mixture flows to a quench tank, where
it is dissolved. The solution is then cooled and filtered
to remove carbon and flyash. The quench and filter units
are similar to those used in the pulp and paper industry.
Quenching can be eliminated by cooling the reduced melt,
allowing it to solidify, and breaking it into chunks. The
solution is precarbonated with pure CO- to produce sodium
hydrosulfide. Final carbonation products are rich in I^S
and are combined and sent to a Glaus plant. The slurry from
the bicarbonator/crystallizer is decomposed to produce
sodium carbonate solution for the spray dryer and pure CC>2
for the precarbonator. The principal chemical reactions
occurring in the Aqueous Carbonate Process are as follows:
3-341
-------
EXISTING STACK
EXISTING
POWER
PLANT
I
U)
NEW ID
FAN
ENTRAINED
SOLIDS
REDUCER OFF-GAS
CO
O
Z
o
PRODUCT
COLLECTION
CO
CO
,
olo o
CO IC/3 O
CM' CM CSJ
re I eg co
2
SULFUR
PRODUCTION
so
2
14 —
ril
t-liQ_Diru
HaO
1 j
r
CARBONATION
Na2S-RICH
•$
SOLUTION
H2O
i
QUENCH AND
FILTRATION
*
'Si r
°; ^
REDUCTION
AIR
4-COKE
GAS
ELEMENTAL
SULFUR
ASH AND
COKE
Figure 3-79. Flowsheet - Aqueous Carbonate Process.
53
-------
SO2 + Na2CO3 •> Na2S03 + CO2 (1)
Na2S03 + 3/2 C -> Na2S + 3/2 C02 (2)
Na2S04 + 2C + Na2S + 2C02 (3)
Na2S + C02 + H20 -> Na2C03 + H2S (4)
In addition to the fact that reheat is not necessary,
there are several other advantages to the Aqueous Carbon? ce
Porcess. Petroleum coke is not only an economical reducing
agent; it reduces all sulfates, and thus avoids water pol-
lution problems. Coal would be a desirable reducing agent,
but it produces problems with flyash, chlorides, and trace
contaminants. Although the Aqueous Carbonate process has not
been tested on a full-scale application, similarly-sized
dryers are being used successfully on other processes. A
marketable sulfur by-product is produced, and no corrosion
problems have been reported. In addition to SC>2 removal,
there are also significant reductions in 863 , particulate
and halogen gases. Removal of NOX was found to be less than
5 percent, based on normal pilot plant operating conditions.
The only waste streams consist of a flyash filter cake dis-
charge and the chloride purge stream discharge contained in
the spent cooling water. The process is a relatively efficient
energy user. Based on a design application to a 500-MW plant,
total incremental fuel consumption is 72 x 10 J/hr (680
Btu/kWh) , Other requirements for a plant of this size
include: reducing coke, 21 x 1010 J/hr (201 MM Btu/hr) ;
fuel oil, 32 x 109 J/hr (32 MM Btu/hr); electricity, 10 MW;
and sodium carbonate, 385 Kg/hr (850 Ib/hr) .
The relatively large size of the Aqueous Carbonate
System, especially of the spray dryer, may cause problems in
retrofit installations. A particulate collection system is
necessary to recover sodium sulfite from the dryer off -gas,
3-343
-------
and a secondary particulate collection system will be
needed. Other potential problems include relatively high
heat requirements in the reduction step, control of the
molten salt reactor, material heat and corrosion stress, and
the possibility of a quench-melt explosion.
The individual process steps have been separately
tested. Spray dryers were pilot tested in a unit with a
diameter of 1.5 m (5 ft) on a slipstream from a coal-fired
power plant at Southern California Edison's Mohave Station,
and in a unit with a diameter of 2.1 tn (7 ft) at Bowen
Engineering's North Branch, New Jersey, laboratory. Over
100 tests were conducted at Mohave, with the dryer and
cyclone operating up to 1.9 m /sec (4000 scfm) . Over 90
percent S00 removal was achieved on a flue gas containing
54
from 200 to 4000 ppm S02. The cyclone collectors removed
from 89 to 99 percent of the sodium sulfite and sulfate
product. At the Bowen laboratory, it was found that SO..,
removal efficiency could be maintained by adjusting the
solution feed rate in direct proportion to the feed gas
rate. The regenerator has been tested in the Atomics Inter-
national laboratory. Integrated tests are still needed to
establish the viability of this process.
3.6.7 Aqueous Sodium Carbonate Absorption ' '
A throwaway aqueous FGD system has been developed by
Combustion Equipment Associates. Trona, a hydrous acid
sodium carbonate (Na9CO °NaHCO -2H 0) , is powdered, slurried,
£ -3 -3 <£,
and used as the absorbent in venturi-type scrubbers. Various
alternate forms of sodium carbonate have also been employed
in tho system, which has been successfully tested in both
pi^ot and commercial scale on a coal-fired power plant.
Many of the design details of the trona scrubbing system
are the property of Combustion Equipment Associates, but the
3-344
-------
commercial installation at Nevada Power's Reid Gardner
Station in Moapa, Nevada, can be described. A separate
modular system, consisting of two parallel venturi scrubbers
followed by a single-stage perforated-plate washing tower,
has been installed on each of two 120-MW boilers burning
pulverized coal. The design throughput of each scrubber is
223 m3/ser (473,000 acfm) at 177°F (350°F). Seventy-five
percent of the flyash is first removed from the hot flue gas
i.i a multicyclone; the gas is then split and sent to the
twin scrubbers, where it is quenched with scrubber liquor
59
sprayed from tangential nozzles. The pressure drop across
the scrubbers is 3.7 kPa (15 in. 1^0). The gas then goes to
a cylindrical separation tower, in which finer droplets are
removed by centrifugal force. It is bubbled through a sieve
tray and passed through a horizontal, single-pass, radial-
vane-type mist eliminator. Prior to discharge, the gas is
reheated by direct mixing with hot air that has been indi-
rectly heated with steam in a heat exchanger. The scrubber
liquor is prepared by mixing trona with water from the ash
pond to dissolve the sodium carbonate. Insolubles, mainly
sand, are settled from the slurry in a clarifier. The
scrubber liquor enters a recirculation circuit, in which the
blowdown from an intial tank is mixed with the alkaline
clarifier underflow and neutralized in a tank to pH 7.
Spent liquor from this tank, about 1 percent of the recycle
stream, is pumped to an ash settling pond and the overflow
to a larger evaporation pond. The sieve tray is flooded
with water from the ash pond, which is also the source of
the water mixed with the recycle liquor stream before it is
returned to the venturi. The chemistry of the trona system
is based on the following reactions.
Na2C03 + S02 + Na2S03 + C02 (1)
Na2SO3 + SO2 + H20 -> 2NaHSO3 (2)
3-345
-------
An open-l°°P soluble system such as the aqueous sodium
carbonate process minimizes operating difficulties that are
due to plugging and scaling. In addition, there is no
problem with absorbent volatilization. Disposal of soluble
sodium compounds, however, may create water pollution problems.
The trona system was piloted in 1972 with a unit of 3.3
to 4.2 m /sec (7000 to 9000 acfm) capacity at the Reid
Gardner Station. The full-sized units began operation in
April 1974. The sulfur content of the coal is between 0.4
and 0.7 percent, averaging 0.55 percent. No major operating
problems have been reported, and the scrubbers have achieved
greater than 90 percent SO., removal efficiency (326 to 577
fin
ppm inlet and 1 to 26 ppm outlet). The evaporation pond
is lined with clay. Because of the low relative humidity in
the area, this is thought to be appropriate for the disposal
of spent liquor. A shortage of trona, however, prevents
widespread use of the aqueous sodium carbonate system.
3.6.8 Shell/UOP Copper Oxide Adsorption61'62'63
The Shell/Universal Oil Products process is a dry metal
oxide system, using copper oxide (CuO) on an activated sor-
bent. The copper oxide is regenerated, and the SO- released
for further processing to elemental sulfur or liquid S0_. A
dry, selected adsorbent is used so that the complications of
a wet system can be avoided. Copper oxide was chosen
because of its high reactivity with SO , and because it can
be regenerated without thermal energy. A unique feature of
the process is its use of parallel passage, fixed-bed reactors.
The Shell/UOP process has been installed on a commercial-
sized, oil-fired boiler in Japan. It, too, is OPS of. the
three processes chosen by EPA for its demonstration project;
a demonstration unit of this type has been built on a slip-
stream from a coal-fired utility boiler in Florida.
3-346
-------
The Shell/OOP process features simultaneous adsorption/
regeneration, followed by SO,., recovery and concentration
(Figure 3-80) . A fixed, packed bed is used for adsorption,
since moving beds would involve greater attrition of the
copper oxide acceptor, increased particulate separation, and
relatively complex solids-handling operations. The gas
stream flows through open channels beside, and in contact
with, the copper oxide. This results in a low pressure drop
and reduced particulate plugging. As the acceptor becomes
loaded with SO,.,, collection efficiency decreases and copper
oxide is regenerated in-situ at the same temperature 400°C
(750°F). Reducing agents such as hydrogen, carbon monoxide,
or light hydrocarbons are used for regeneration; hydrogen is
preferred because of its high reactivity with copper sulfate.
The rate of S02 release can be controlled by the rate of
hydrogen addition. Hydrogen is catalytically oxidized by
copper oxide to water; the S02, free of oxygen and particu-
late matter, can be used for the production of liquid SO,., or
elemental sulfur. The use of two or more reactors in parallel,
with some in service while others are being regenerated,
permits continuous S02 removal. There are a number of ways
to concentrate and prepare S0~-rich regeneration gases for
further processing, including water adsorption/stripping,
solvent absorption, compression and condensation of water
vapor, and complete liquefaction. In addition, UOP has
developed other less energy-intensive procedures. The basic
chemistry of the Shell/UOP process is as follows:
CuO + 1/202 + S02 -»- CuS04 (1)
CuS04 + 2H2 -> Cu + S02 + H20 (2)
CuO + H2 + Cu + H20 (3)
3-347
-------
REDUCING GAS
QUENCH
TO AIR HEATER
ABSORBER
STRIPPER
BOILERS
TO
REDUCTION
REACTOR
U)
I
LO
oo FLUE GAS
BOTTOMS
X C ri A N G E P
WASTE HEAT
BOILER
OO
AIR -COOLED
EXCHANGER
RICH H20
STORAGE
REDUCTION REACTOR
S02 TO
REDUCTION
REACTOR
REDUCING GAS
WASTE HEAT
BOILER
2 STAGE CLAUS PLANT
SULFUR
TAIL GAS
TO INCINERATION
THEN TO FLUE GAS
Figure 3—80. Flowsheet - Shell/UOP copper oxide adsorption process.
64
-------
UOP guarantees 90-percent SO.-, removal with the copper
cc ^
oxide adsorption process. in addition, the process provides
almost complete SCX, removal. The only waste stream is
wastewater, containing about 50 ppm S0~ at 455 to 680 1pm
(120 to 180 gpm) for a 500-MW plant. Removal of 60 to 70
percent of NO has been demonstrated by the addition of
-X
ammonia.
Integration of a complete copper oxide system into a
power plant will be difficult, because a cyclic fixed-bed
adsorber and a variable flue gas stream cannot be compatibly
operated in conjunction with relatively inflexible hydrogen
and sulfur production units. An inlet dust load of less
than 0.22 g/m (0.1 gr/scf) is required. The fairly large
space requirements, and the necessity of reheating the exit
gas from 150° to 400°C (300° to 750°F), could create diffi-
culties in some retrofit installations. In addition to the
large amount of hydrogen reducing gas needed, energy demands
are high, because of the quantity of water circulated and
the steam necessary to heat the stripper reboiler. Total
incremental fuel requirements for a 500-MW plant are 17 x
10 J/kWh (1620 Btu/kWh); other requirements include:
ammonia, (2110 Ib/hr); reducing agent, 34 x 1010 J/hr (322
MM Btu/hr); steam, 30 x 1010 J/hr (288 MM Btu/hr); fuel oil,
g
52 x 10 (50 MM Btu/hr); electricity, 6 MW.
The Shell/UOP acceptor and regenerator were first demon-
strated in 1967 on a slipstream of 0.19 to 0.28 m /sec (400
to 600 scf/min) from a process heater using high-sulfur fuel
oil at Shell's Pernis refinery near Rotterdam. Ninety-per-
cent S0~ removal was demonstrated over 20,000 operating
66
hours, with estimated acceptor life of 1.5 years. A com-
mercial unit was installed in 1973 at a Showa Yokkaichi Sekiyu
(SYS) oil-fired, 40-MW boiler (125,000 Nm3/hr - 2500 ppm SO ).
3-349
-------
Only the acceptor and water-gathering systems were installed,
because there were existing hydrogen and Glaus plants.
This system had 68 unit-cells, identical in design but
larger than the single cell at Pernis. Ninety-percent S02
removal was shown, with the SO2 co-processed both in the
Glaus plant with normal refinery acid gas, and in an absorber/
39
stripper. The longest continuous run was 2 months. Simul-
taneous removal of 60 to 70 percent NO was achieved by the
jC
addition of ammonia, with a maximum ammonia carryover in the
exit gas of 2 ppm. Hydrogen consumption was 0.2 percent by
/- r-i
weight of sulfur recovered. The stripped water contained
20 ppm sulfur, 75 percent of which was present as sulfate.
A demonstration unit of the acceptor and regenerator has
also been installed on a slipstream from a coal-fired boiler
at Tampa Electric's Big Bend Station. It has five modules
identical to those at the SYS refinery, and is designed
for 90-percent desulfurization of a gas sgream with 1260 to
1400 ppm S02. Operating results have not been reported.
The EPA-utility joint demonstration plant will be the first
completely integrated unit on a coal-fired boiler.
3.6.9 Alkaline Flyash Scrubbing Process69'70
In the alkaline flyash scrubbing process, the inherent
alkalinity of the flyash from certain coals is used to
react with SO2. Lime is added to the scrubbing liquor for
pH control. The alkaline species in the scrubbing liquor
combine with the SO2 to form insoluble sulfates and sulfites,
which are then disposed of. The system has been successfully
tested in four pilot plant studies, and is currently operated
on two utility boilers. It will be employed on cix utility
boilers in the future.
Lime/alkaline flyash scrubbing is a throwaway process,
very similar to lime systems, except that the major source
3-350
-------
of the active alkali is flyash from coal. It is used pri-
marily in the West, where the flyash of some coals is high
in sodium oxide, calcium oxide, and magnesium oxide. The
alkali content of Western coal ash varies from under 10 to
over 50 percent, with the lower-rank coals (i.e., lignite)
having the highest alkali content. Figure 3-81 is a simpli-
fied flow diagram for a lime/alkaline flyash scrubbing system.
Flue gases enter the scrubber at approximately 150°C (300°F).
The scrubber applied to large utility boilers is a two-stage
device with a venturi and a spray tower in series. In the
venturi, most of the flyash is collected, whereas the spray
tower absorbs S02. Slurry from both stages is collected and
held in the bottom of the scrubber for reuse. The slurry is
usually agitated to hold the flyash in suspension. The
cleaned flue gas passes through a mist eliminator and re-
heater before discharge to the stack.
A purge stream is constantly withdrawn from the recycle
slurry in order to maintain approximately 12 percent solids
in the recycle stream. The purge stream is usually clari-
fied or thickened before disposal in a pond. Decanted water
from the clarifier/thickener is recycled to the absorbent
loop.
Makeup water is also added to the absorbent loop, to
compensate for water losses by evaporation and by retention
in sludge. The scrubber recycle is maintained at a constant
pH of about 5.0 by the addition of lime or limestone slurry
to the recycle stream. This is done to minimize scaling and
to promote ionization of the active alkali species in the
flyash.
Based on operation experience, the lime/alkaline flyash
system is considered attractive. Tests on Colstrip Units 1
and 2 revealed that an emission level of 0.19 kg/10 J (0.44
3-351
-------
I
CO
01
WASH TRAY POND
FLYASH POND
Figure 3-81. Flowsheet of lime/alkaline flyash scrubbing.71
-------
Ib SO9/10 Btu) or less could be attained using coal with
72
sulfur contents of 0.56 to 0.94 percent. No measurements
were made of inlet SO,., loadings to the scrubber; however,
assuming a 90-percent oxidation of the coal sulfur to SO-
in order to estimate the inlet SO,, loadings, an efficiency
of 70 to 90 percent was achieved. The major limitation
to the use of alkaline flyash scrubbing is the need for high
active alkali concentrations in the ash in relation to the
sulfur content of the fuel. Application would therefore be
limited in the United States to low-sulfur, high-alkaline
Western coals.
3-353
-------
REFERENCES FOR SECTION 3.6
1. Daken, R.B., et al. Interim Report on Chiyoda Thorough-
bred 101 Application plant at Gulf Power's Scholz
Plant. In: Proceedings of the Symposium on Flue Gas
Desulfurization. New Orleans, March 1976. EPA-600/2-
76-136a. U.S. Environmental Protection Agency.
Washington, B.C. May 1976. pp. 762-766.
2. Slack, A.V., and G.A. Hollinden. Sulfur Dioxide Re-
moval from Waste Gases. Second Edition. Noyes Data
Corp., Park Ridge, New Jersey, 1975. pp. 124-126.
3. Noguchi, M. Status Report on Chiyoda Thoroughbred 101
Process. From EPA-650/2-74-126-b. p. 838.
4. Slack, A.V. Second Generation Processes for Flue Gas
Desulfurization Introduction and Overview. In:
Proceedings of the Symposium on Flue Gas Desulfuriza-
tion, Atlanta, November 1974. EPA-650/2-74-126b. U.S.
Environmental Protection Agency. Washington, D.C.
December 1974. p. 1034.
5. Op. cit. No. 1. p. 762.
6. Op, cit. No 3. pp. 844, 847.
7. Op. cit. No. 2. p. 125.
8. Op. cit. No. 3. p. 839.
9. Ibid.
10. Op. cit. No 1. pp. 767-769.
11. Ibid.
12. Ibid. p. 764.
13. Op. cit. No. 2. pp. 215-217.
3-354
-------
14. Op. cit. No. 4. p. 1Q42.
15. Nissen, W.I., et al. Citrate Process for Flue Gas
Desulfurization. A Status Report. In: EPA-600/2-76-
136a, pp. 841-864.
1.6 Ottmers, D.M., Jr., and J.C. Diekerman. Status of
Second Generation Regenerative Flue Gas Desulfurization
Processes. In: Proceedings of the 12th Air Pollution
and Industrial Hygiene Conference on Air Quality
Management in the Electric Power Industry CHal B.H.
Cooper, Jr., Ed.). Center for Energy Studies, The
University of Texas at Austin. January 28-30, 1976.
pp. 420-422.
17. McKinney, W.A., et al. Pilot Plant Testing of the
Citrate Process for S02 Emission Control. From EPA-
650/2-74-126-b, Reference 4. pp. 1019-1066.
18. Op. cit. No. 16. p. 435.
19. Op. cit. No. 15. p. 843.
20. Op. cit. No. 17. p. 1059.
21. Ibid. p. 1061.
22. Ibid. p. 1059.
23. Ibid.
24. Op. cit. No. 4. p. 1040.
25. Op. cit. No. 16. pp. 406-408.
26. Ball, F.J., et al. Westvaco Activated Carbon Process
for SOX Recovery as Elemental Sulfur. In: EPA-650/2-
74-126-b, pp. 1151-1191.
27. Ibid. p. 1165.
28. Op. cit. No. 16. p. 430.
29. Op. cit. No. 26. p. 1172.
30. Ibid. p. 1165.
31. Ibid.
3-355
-------
32. Ibid. p. 1172.
33. Ibid. p. 1167.
34. Ibid. p. 1184.
35. Op cit. No. 4. p. 1040.
36. Op cit. No. 16. pp. 412-414
37 Strum, J.J., et al. BF Dry Adsorption System, Parts I
and II. In: EPA-600/2-76-136a, pp. 877-911.
38. Op. cit. No. 16. p. 413.
40. Op. cit. No. 37. p. 902.
41. Ibid. p. 883.
42. Ibid.
43. Op cit. No. 2.
44. Op cit. No. 4. p. 0142.
45. Struck, R.T., et al. The consol FGD Process. In: EPA-
600/2-76-136a, pp. 913-929.
46. Ibid.
47. Ibid.
48. Op cit. No. 2.
49. Op. cit. No. 4. p. 1041.
50. Op cit. No. 16. pp. 414-417.
51. Gehri, B.C., and R.D. Oldenkamp. Status and Economics
of the Atomics International Aqueous Carbonate Flue Gas
Desulfurization Process. In: EPA-600/2-76-136a, pp.
787-816.
52. Ibid. p. 798.
53. Ibid. p. 789
54. Ibid. p. 801.
3-356
-------
55. Ibid. p. 798.
56. Op. cit. No. 2.
57. Gerstle, R.W., and G.A. Isaacs. Survey of Floie Gas
Des-ulfurization Systems, Reid Gardner Station, Nevada
Power Co. EPA-650/2-75-057-j. U.S. Environmental
Protection Agency, Research Triangle Park, North
Carolina. October 1975. p. viii.
58. Performance Guarantee and Compliance Tests for CEA
Scrubber Units 1 and 2, located at Reid Gardner Gener-
ator Station, Nevade Power Company. York Research
Corporation. Stamford, Connecticut. September 1975.
59. Op. cit. No. 32. p. 3-1.
60. Ibid.
61. Op. cit. No. 16. pp 409-412.
62. Pohlenz, J.B. The Shell Flue Gas Desulfurization
Process. In: EPA-650/2-74-126-b, pp. 807-
835.
63. Vicari, F.A., and J.B. Pohlenz. Energy Requirements
for Shell FGD Process. In: EPA-600/2-76-136a, pp.
817-832.
64. Op. cit. No. 16. p. 431.
65. Ibid. p. 410.
66. Op. cit. No. 62. p. 816.
67. Ibid. p. 819.
68. Ibid. p. 821.
69. New, H.M., et al. Status of Flue Gas Desulfurization
Using Alkalia Flyash from Western Coals. In: EPA-
600/2-76-136a, pp. 269-324.
70. Gumm, C., et al. Particulate and S02 Removal at the
Colstrip Station of the Montana Power Co. In: Proceed-
ings of The Second Pacific Chemical Engineering Confer-
ence, Denver, August 1977-
71. Ibid. p. 324.
3-357
-------
-------
4.0 PERFORMANCE AND OPERABILITY OF FGD SYSTEMS
This section summarizes the current status of FGD
technology with respect to S0« removal efficiency and system
operability. All of the documentation in this section is
presented by reference to specific full scale or test facili-
ties. Information for these plants is presented in detail
in Section 3 and the appendicies of this report. This
information was obtained from the extensive research that
has been reported for pilot and demonstration scale units in
this country and also included data from full scale facili-
ties. The design and operating experience gained with first
generation FGD systems has resulted, and will continue to
result, in improved design and operation of subsequent
installations. FGD systems which are currently being engi-
neered will incorporate the latest design features and even
better performance can be expected as newer plants are
completed.
New plants that are designed with current knowledge of
process kinetics and chemistry, equipment capabilities,
experience with materials of construction, and adequate
redundancy can achieve removal efficiencies and operabil-
ities of, or in excess of, 90 percent.
In this section of our report we have summarized the
operating experience and problems which FGD systems have
encountered; the solutions (both engineering and opera-
tional) , which have resulted; and the operability, avail-
ability and control efficiencies that can be reasonably
expected in FGD technology.
4-1
-------
1 2
4.1 OPERATING PROBLEMS AND SOLUTIONS '
There have been and still are problems associated with
FGD systems; however, the majority of these problems have
basically been solved and the methods of reducing the
severity of the remaining items are much better understood.
While these problems can be minimized, it must be recognized
that large complex chemical processes like an FGD system
need constant attention and maintenance, and will never be
run in a completely troublefree manner.
To date, the problems encountered with FGD systems and
the severity of these problems varied both with system type
and within units of the same system. The more common prob-
lems encountered are listed below, followed by a more
detailed discription of each problem and how it is being
solved.
° Formation of scale in the absorber and associated
equipment in lime and limestone systems leading to
plugging and reduced capacity
0 Plugging of mist eliminators, lines, and some
types of absorbers
0 Failure of ancillary equipment such as pumps,
piping, pH sensing equipment, reheaters, centri-
fuges, fans and duct and stack linings
° Inadequate absorbent make-up preparation
0 Handling and disposal of sludge in nonregenerable
systems
4.1.1 Scaling and Plugging
In lime and limestone systems, scaling has been a
particular problem and has reduced operability in a number
of systems. Both a soft sulfite scale and a hard sulfate
scale may form in the absorber, mist eliminator, and an-
cillary tanks, pumps, and pipes. Much effort and investi-
4-2
-------
gation has shown that these problems are closely related to
the system chemistry and therefore, specific process control
techniques have produced significant improvements:
° Operation at subsaturation levels for calcium sulfate
and sulfite.
Full scale and test facilities in this country have
effectively reduced saturation and scaling by addition of
magnesium to the circulating slurry. The TVA Shawnee
facility, the Phillips facility, and the Paddy's Run facil-
ity demonstrated that the addition of magnesium to the lime
and limestone slurry eliminated scrubber scale formation.
Other large stations, the Bruce Mansfield and Conesville use
lime containing magnesium oxide to prevent scaling.
By maintaining high liquid to gas (L/G) ratios, the
proportion of unreacted lime or limestone remains high
relative to the absorbed SO-. There is thus less chance of
creating a supersaturated solution of sulfites or sulfates.
The higher L/G ratio also improves overall SO_ collection
efficiencies as described in Section 3.1. The actual L/G
will vary with the type of absorber, and values in excess of
10.8 1/m (80 gal/acf) have been used in spray towers.
Increased reaction tank holding time will also decrease
saturation by allowing further reaction between the absorbed
S02 and the lime or limestone slurry. Slurry residence time
at the Green River facility is greater than twenty minutes
and scale formation is not a major problem. This plant has
an excellent availability history.
0 pH Control.
Work at the EPA-Shawnee test facility has shown that
the most important parameter in controlling scale formation
is solution pH. Operational experience has shown that lower
pH values accelerate scale formation. This is documented
4-3
-------
at full scale facilities such as LaCygne. The measurement
of pH has also received considerable attention. More rugged
and dependable sensors are being used; they are located in
the slurry stream where they are subject to less breakage,
are more accessible, and where they yield data which is more
reliable and responsive for pH control. The Bruce Mansfield
facility has just completed a renovation program to incor-
porate these design features into their pH system.
0 Co-precipitation of sulfate.
Minimizing the oxygen content in the flue gas, by re-
ducing any air in-leakage, favors co-precipitation of sul-
fate with the sulfite crystal. Therefore, air exposure is
reduced by covering open reaction tanks, clarifiers, etc.
Plugging caused by deposition of solids on equipment
surfaces has sometimes restricted the passage of liquids or
gas in FGD systems. It is usually easily removed by flush-
ing with water or steam. Plugging in pipes can be prevented
through designs which avoid low flow velocities. Careful
control of raw material particle size and screening of the
slurry also decreases plugging problems, especially in spray
nozzles, pipes and pumps. Since this problem is caused by
the deposition of solids from the recirculating slurry,
reduction of the overall amount of solids will reduce the
plugging. The minimum stoichiometry that will effect the
required S02 removal efficiency should be used. This has
been demonstrated in this country at Shawnee and LaCygne.
The Japanese facilities use stoichiometries in the 0.95 to
1.05 range, have no plugging difficulties and have SO,,
&
removal efficiencies in excess of 95 percent.
In summary, the scaling/plugging problem can be con-
trolled or eliminated by carefull scrubber chemistry con-
trol. This has been demonstrated at operating facilities in
this country and Japan.
4-4
-------
4.1.2 Erosion and Corrosion
Problems with ancillary equipment plagued many of the
first FGD systems because of poor design, faulty component
manufacture, or improper installation. Many of these prob-
lems were due to corrosion and erosion.
Erosion in venturi pre-scrubbers has resulted from high
fly-ash loadings. Likewise, pre-scrubbers remove the bulk
of any chlorides and sulfur trioxide in the gas stream; both
of these components are highly corrosive. Corrosion occurs
more frequently in areas subject to wet saturated flue gas
as opposed to areas subject to alkaline slurry streams.
There are so many factors involved in FGD operation
which effect corrosion rates, that generalizations regarding
corrosion resistant materials is difficult. A sufficient
amount of data has been accumulated, however, to provide
general guidelines for the construction of critical elements
in FGD systems as summarized below:
(a) Some systems are incorporating such alloys as
Hastelloy C-276, Hastelloy G, Inconel 625, Incoloy
825, 317L stainless steel, 904L stainless steel
and Jessop JS700 in wet/dry high temperature, high
chloride environments, such as in presaturators.
The LaCygne Station has found that these materials
give excellent reheat service. The Bruce Mansfield
station has had good results with Hastelloy wetted
parts in the fan.
(b) Synthetic and natural rubber coatings predominate
in recycle tanks, pumps, and lines. These mate-
rials have been reported to give superior erosion
resistance once application problems have been
overcome. For instance rubber lined pumps have
been used successfully at the following facilities:
Green River, LaCygne, Bruce Mansfield, and Cones-
ville.
(c) For liners in the absorbers, exhaust ducts and
stacks, a number of materials such as resins,
ceramics, polyesters, polyvinyls, polyurethanes,
4-5
-------
carboline, and Gunite, have been used with varying
degrees of success. Although successful applica-
tions have been reported, widespread failures of
the liners have been attributed to the undepend-
ability and inexperience of lining applicators,
instability of the materials at high temperatures,
inconvenience of repair, and cost-related factors.
These problems are especially evident on higher
sulfur coals. Extensive effort in continuing by
FGD suppliers to fully solve this problem.
In summary, the corrosion and erosion problems have
been solved by employing more expensive materials of con-
struction. It is apparent that the technology exists to
control these problems, but it is more expensive than
materials that are normally used.
Increased SO2 removal efficiency should have no addi-
tional effect on corrosion or erosion problem in systems
which are designed properly. In fact, since higher S0?
removal efficiencies will reduce the SO,., content in exit
gases the downstream equipment corrosion may be reduced.
4.1.3 Equipment Design Improvements^
Solutions utilized to reduce problems with ancillary
equipment include:
0 Recirculation Pumps - Slurry recirculation pumps provide
the driving force for the liquid circuit in FGD systems. In
their design, special attention must be given to an accurate
service description (solution pH, specific gravity, solids
content, gas entrainment, flow rates, and head). A number
of general trends are evident, and summarized below:
(a) New systems must incorporate spare pumps. Spare
capacity from 50 percent (one spare for every two
operational) to 100 percent (one spare/one opera-
tional) is useful to avoid downtime.
This type of spare equipment is found at many new
large stations including Bruce Mansfield and
Conesville.
4-6
-------
(b) Natural and synthetic molded rubber lining should
be specified for wetted parts in the pumps.
(c) Flush-water wash systems are needed to purge the
pumps of solids, which tend to settle out during
periods of inactivity.
0 Mist Elimination - Chevron and baffle-type mist elimina-
tors have been and are currently being used in virtually
every FGD system in the United States. The popularity of
these collectors is due primarily to design simplicity, high
collection efficiency (for moderate to large size drops),
low pressure drop, wide-open construction, and low cost.
Within these two preferred types of mist eliminators, a
number of specific design and construction innovations have
been implemented:
(a) Chevron designs (continuous vane construction) are
predominant over baffle designs (discontinuous
slat construction).
(b) Fiberglass-reinforced plastic is now used at
nearly all facilities.
(c) The horizontal configuration (vertical gas flow)
is also used in almost all installations for cost
reasons.
(d) Two-stage designs predominate over single-stage
designs, because they yielded higher elimination
efficiencies.
(e) Bulk entrainment separators, perforated plates,
impingement plates and other precollection devices
are becoming integral parts of mist elimination
systems. These reduce plugging and improve
separation. The Conesville facility employs this
as well as LaCygne, Lawrence, Cane Run No. 5, and
Coal Creek.
(f) Mist eliminator wash systems that employ inter-
mittent, high-velocity sprays predominate over-
continuous wash systems. These produce a hydraulic
washing effect.
(g) Operation at high alkali utilization.
4-7
-------
Mist elimination in streams containing slurry
droplets is not new technology. The pulp and
paper and sulfuric acid industries have been using
it for many years. The problems that are encoun-
tered today in the FGD industry are similar to
those that have been experienced in other indus-
tries for some time. New design and operating
conditions are continually reducing the problems.
For instance, Combustion Engineering offers a bulk
entrainment separator system that has worked
effectively at the Lawrence and Sherburne stations.
0 Reheat - Virtually all the FGD systems coming on-line and
planned for future operation incorporate some type of stack
gas reheat system. These systems heat the flue gas to avoid
condensation with subsequent corrosion to downstream equip-
ment, ductwork and stack and to suppress plume visibility as
well as enhance plume rise and pollutant dispersion. To
date, a number of "wet stack" FGD systems (no reheat) have
been installed and have encountered corrosion problems. The
trend in reheat systems is towards heating of ambient air
and mixing with the flue gas and mixing of untreated flue
gas with scrubbed gas.* In-line reheat systems have been
subject to corrosion and solids deposition, the latter often
occurring because of inefficient upstream mist elimination.
Several existing facilities have found that they could
control reheater problems by fabricating the reheat bundles
from more expensive materials. Both Will County and Sher-
burne have found this to be true. At Sherburne and Lawrence,
carbon steel fin-heaters have been used with hot water as
the heating mechanism. Reheat bundle life expectancy is 5
to 7 years, which is considered acceptable for this service.
Since these tubes are carbon steel, corrosion failure is not
catastrophic but gradual. This permits planned shutdown and
maintenance. Both of these units have excellent mist elimi-
K Blending with untreated flue gas is, of course, limited when
high efficiency is desired. The amount of gas which can be
bypassed for reheat purposes depends upon the amount of SO.-,
resulting from the coal, the applicable regulations, the
particular emission levels.
4-8
-------
nation systems (multiple-stage FRP mist eliminators which
incorporate bulk entrainment separators), which contribute
to the reheat success.
Fans ~ Fans installed immediately after an FGD system (wet
fans) have experienced corrosion, chloride attack, and
solids deposition problems. Deposition problems have
caused fan imbalance resulting in excessive bearing wear and
damage to the fan. Three systems have reported this prob-
lem; Phillips Bruce Mansfield, and St. Clair. The problems
associated with fans installed upstream of the FGD system
(dry fans) include operation at higher temperatures (over
150°C) resulting in higher gas velocities and abrasion by
fly ash. Dry fan problems are more easily solved, and the
tendency is toward fans upstream of the FGD system. Where
necessary, however, the use of various steel alloys have
made wet fans a viable alternative, as at Phillips and Bruce
Mansfield.
0 Summary - It has been demonstrated that each of the major
equipment problems encountered has been solved. Usually
this requires more expensive design or operating techniques.
Most often, problems in new systems are not encountered
because the technology has failed, but because the proper
equipment is more expensive than that installed.
4.1.4 Feed Preparation
The preparation of feed slurries has presented problems
in lime, limestone, and magnesium oxide systems. Slaking,
prior to fresh feed storage tanks, is necessary for the lime
and magnesium oxide systems. Fine grinding has reduced
problems in the use of limestone. The Japanese have had
excellent results with limestone ground to less than 300-
mesh. This has increased SO,, removal and reduced plugging.
4.1.5 Solids Separation and Sludge Disposal
The disposal of sludge is a major consideration in
selecting a nonregenerable FGD system. Land availability,
4-9
-------
sludge handling, and waste water contamination are potential
problems with these systems. Increased SC>2 removal will
tend to increase this problem since slightly more sludge
will be generated. The impact on solid waste and environ-
mentally acceptable disposal techniques is being discussed
in detail in a separate study by the Aerospace Corp. for
EPA.
Solutions to sludge disposal have been directed toward
reducing its water content, reducing its volume, and in-
creasing its compressive strength. A number of dewatering
processes have been applied, and clarification, centrifuga-
tion, and vacuum filtration techniques predominate. In
addition to these techniques, a number of plants are ex-
perimenting with or employing forced oxidation of sulfite to
sulfate to enhance solids settling properties, to decrease
sludge disposal land requirements, and to maximize the
quality of recycled water. Two full-scale systems using
forced oxidation are Northern States, Sherburne No. 1 and
No. 2. Two additional full-scale systems using forced
oxidation on a short-term experimental basis are Arizona
Public Service, Cholla No. 1; and Commonwealth Edison, Will
County No. 1. Systems planning to use forced oxidation for
future operations are Northern States Power, Sherburne No. 3
and No. 4. Research is continuing in this area.
The two predominant methods of sludge disposal are on-
site ponding and landfilling with physically conditioned or
chemically treated waste. A number of specific disposal
methods are available in each of these categories. The
selection of a specific method depends upon a number of
site-specific factors such as space, distance to site, etc.
Currently, feasible alternates and trends include:
(a) Based upon systems now in service or planned for
future operations, approximately 75 percent of the
FGD operating capacity will be using on-site
4-10
-------
ponding for ultimate disposal of sludge. The
remaining 25 percent capacity will employ physical
and chemical treatment methods for landfill dis-
posal.
(b) Two major suppliers now offer proprietary methods
for chemically fixing sludge. The Dravo Corpora-
tion and IU Conversion Systems (IOCS) both market
commercial stabilization processes. Dravo and
IUCS supply systems that treat FGD waste from
full-scale FGD systems. These systems are being
used at the Phillips, Elrama and Conesville
stations.
(c) A number of utilities operate FGD systems that
include fly ash and lime stabilization of the
thickener underflow, or vacuum filtration of
solids prior to offsite landfill.
(d) A number of major FGD system suppliers now offer
sludge treatment processes in addition to emission
control systems. Research-Cottrell and Combustion
Engineering, for example, now provide sludge
treatment systems as separate packages from the
FGD systems.
Water from the disposal pond is frequently recycled to
the FGD system. Operation with no waste water disposal
(closed loop) has proven difficult in some cases due to the
need to purge dissolved solids from the system and the extra
water needed at times to wash demisters. The entire water
disposal problem is being investigated for EPA in a study
by the Radian Corp.
4.2 SYSTEM OPERABILITY* AND AVAILABILITY**
4.2.1 Operability Trends
The design and operation of newer FGD systems benefits
from the experience of older units. The observed trend of
improved operability should therefore continue to increase.
Based upon limited, utility-provided data. Figure 4-1 shows
the cumulative FGD operability percentage as a function of
* Hours FGD system operated/hours boiler operated.
** Hours FGD system was available/hours in that period.
4-11
-------
FGD system start-up date. In this presentation, the indi-
vidual unit availability or operability, from start-up to
July 1977, is plotted as a function of plant start-up date.
Operability is used in all cases except when all the flue
gas must pass through the FGD unit (no by-pass). In that
case availability is a more valid measure of operation. As
will be noted, the trend line, which is a least squares
regression fit, reflects greatly improved performance by
newer FGD systems. Table 4-1 identifies the plants plotted
in Figure 4-1.
Several examples of the improved availability and
operability of FGD systems are noted below. Performance
histories of these individual systems were presented in
Section 3 and in the Appendices to this report.
Lime Systems
0 Louisville Gas and Electric, Cane Run Unit No. 4,
Louisville, Kentucky (178 MW)
The unit began operation in August 1976. With the
exception of a shutdown caused by the lack of lime
(frozen rivers prevented barge deliveries), and
process modifications, the unit has averaged in
excess of 95 percent operability through July
1977.
0 Louisville Gas and Electric, Paddy's Run Unit No.
6, Louisville, Kentucky (65 MW)
The unit began operation in April 1973. Start-
up problems, modifications, and an extended boiler
shutdown kept operability low until October 1974.
Operability from October 1974, to August 1977,
has been in the 95- to 100-percent range, al-
though the boiler operates primarily as a peak
loading station.
0 Kentucky Utilities, Green River Units No. 1, 2, &
3, Central City, Kentucky (64 MW)
4-12
-------
I
1—•
U)
lOOi
£< 90
00 O
Si
CO
<:
80
Q.
I
< 70
z 60
UJ
it so
i 40
i
<* 30
— O
CO
3
A
15
NOTE: CORRELATION COEFFICIENT FOR THIS LEAST
SQUARES LINEAR PLOT INDICATES A
STATISTICAL CERTAINTY OF 99 PERCENT.
STANDARD DEVIATION = 18
A OPERABILITY
® AVAILABILITY
(§) RELIABILITY
1972
1973
1974
PLANT START-UP DATE
1975
1976
Figure 4-1. Average plant FGD availability/operability
versus plant start-up date.
(See Table 4-1 for plant identification)
-------
Table 4-1. IDENTIFICATION OF PLANTS IN FIGURE 4-1
1. Will County No. 1
Commonwealth Edison
2. Mystic No. 6
Boston Edison
3. Hawthorn No. 4
Kansas City Power and Light
4. Hawthorn No. 3
Kansas City Power and Light
5. LaCygne No. 1
Kansas City Power and Light
6. Paddys Run No, 6
Louisville Gas and Electric
7_ Choila No. 1
Arizona Electric Power Co-Op
8. Reid Gardner No. 1
Nevada Power
9. Reid Gardner No. 2
Nevada Power
10. Scholz No. IB and 2B
Gulf Power Co.
11. Scholz No. 1A
Gulf Power Co.
12. Green River No. 1 and 2
Kentucky Utilities
13. Sherburne County Station No. 1
Northern States Power Co.
14. Bruce Mansfield No. 1
Pennsylvania Power
15. Reid Gardner No. 3
Nevada Power
16. Cane Run No. 4
Louisville Gas and Electric
4-14
-------
This FGD unit began operation in December 1975.
After the initial operating period (Dec. 1975-Feb.
1976) , the operability of the unit was 97 percent
through June 1977.
Limestone Systems
0 Kansas City Power and Light, La Cygne Unit No. 1,
La Cygne, Kansas (.820 MW)
Operation of this FGD system began in January 1974,
Since all the flue gas must be treated by the FGD
system, availability is a more meaningful param-
eter. From February 1976, through July 1977, unit
availability averaged 93 percent.
0 Northern States Power, Sherburne Nos. 1 & 2,
Sherburne Minnesota (700 MW each)
After start-up, No. 1 unit averaged 92-percent
availability- No. 2 unit has averaged 95-percent
since start-up.
Magnesium-Oxide Systems
0 The first full-scale FGD unit (Philadelphia
Electric, Eddystone No. 1) designed as a con-
tinuing operation was recently restarted after
eighteen months of downtime for the forced re-
location of the magnesium sulfite regeneration
equipment. Two earlier installations (Boston
Edison's Mystic Station and Potomac Electric"s,
Dickerson Station) were short-term, prototype
units to explore the process potential and opera-
ting problem areas. Only data for the Mystic
Station are shown in Figure 4-1.
In summary, the availability of full scale scrubbing
facilities has increased steadily to where current systems
are demonstrating long-term availabilities in excess of 90
percent, in most cases without the use of spare modules.
4.2.2 Availability
The ability of an FGD system to operate is as important
as its efficiency in reducing S02 emissions. This ability
to operate has been defined a number of ways including:
4-15
-------
0 Availability Factor - Hours the FGD system was
available for operating (whether operated or
not)/hours in period, expressed as a percentage.
0 Reliability Factor - Hours the FGD system oper-
ated/hours FGD system was called upon to operate,
expressed as a percentage.
0 Operability Factor - Hours the FGD system was
operated/boiler operating hours in period, ex-
pressed as a percentage. __
0 Utilization Factor - Hours FGD system operated/
hours in period, expressed as a percentage.
Availability and operability are commonly interchanged
in usage, and, if a boiler operates continually over a
specific time period, they yield the same numerical value.
Historically FGD system availability has varied depending on
a variety of factors, and, naturally, individual module
availabilities have varied even more so. Trends in overall
system availability were shown earlier in Figure 4-1. These
data demonstrated continuing and positive improvement in
availability as newer units (with improved design and com-
ponent features) came on-line. In addition, only one of
these stations (Sherburne) has an extra or redundant module
available for use in the event of an operating module mal-
function or failure.
4.2.2.1 Module Availability - The range in average monthly
module availabilities for a number of selected FGD systems,
which historically have shown high average availability, is
shown in Figure 4-2. The annual average availability for
the overall FGD system is also shown. Because of the varia-
tions in system design, maintenance, and operational capa-
bility, it is difficult to generalize about module avail-
ability. It is, however, evident that a 90-percent average
module availability can be achieved for both low- and high-
sulfur coal applications.
4-16
-------
LaCYGNE
(5.4% S)
90
80
70
60
50
40
?n
—
—
E9M
en
en
1 —
10
en
en
en
~
en
C--
en
10
en
en
OL
o
LJ_
UJ
<:
C£
UJ
•<
—J
en
oo
_i
Q
CD
— 1
_1
oo
UJ
Q
O
— 1
1
o;
o
u-
LU
CD
LU
>
_
— |
_
r--
en
-
—
-
UD
r — .
en
f~*
ANNUAL AVERAGE
AVAILABILITY FOR
ENTIRE FGD SYSTEM
RANGE IN INDIVIDUAL
MONTHLY MODULE
AVAILABILITY
CHOLLA
(0.6% S)
SHERBURNE
(0.8% S)
BRUCE
MANSFIELD
:».Q% s MAX)
Figure 4-2. Average availability for selected FGD systems
4-17
-------
Also shown on this figure are trends in improved avail-
ability of modules at the same plant over a period of years.
At Sherburne No, 1, the ability to achieve an average avail-
ability of 90 percent for all 12 modules (1 redundant)
increased from 91 to 93 percent over a one-year period (1976
to 1977). At La Cygne, a steady increase in average avail-
ability from 76 to 93 percent occurred over a three-year
period (1974 to 1977). The high initial availability of
Cholla has varied only slightly over 4 years and is cur-
rently over 90 percent. At the Bruce Mansfield Station,
availability was near 98 percent during the first 8 months
of operation in 1976. During 1977, availability remained
high until the boiler was shutdown in March. After re-
starting at 50 percent of capacity due to stack relining,
the availability of the modules remained high.
4.2.2.2 Redundancy - In order to establish a basis for
selection of redundancy requirements to be used in cost
analysis, a statistical analysis was made of overall FGD
system availabilities for various numbers of required
modules. Table 4-2 presents availability projections for
various sized FGD systems and various assumed single module
availabilities. For this case, it was assumed that the
availability of the system was equal to the probability that
all modules are available. No consideration was given to
boiler availability and it was assumed that the boiler
required the FGD system 100 percent of the time (no bypass).
The formula for such a determination is
4-18
-------
Table 4-2. FGD SYSTEM AVAILABILITY PROJECTIONS
BASED UPON NO SPARE MODULES
Boiler
Size,
MW
100 - 150
150 - 300
300 - 450
450 - 600
600 - 750
750 - 900
900 - 1050
1050 - 1200
Minimum
Number of
Modules at
Full Load
1
2
3
4
5
6
7
8
FGD System Availability Based
Upon Single Module Availabilities of:
70
70
49
34
24
16
12
8
6
75
75
56
42
32
24
18
13
10
80
80
64
51
41
37
30
24
19
85
85
72
61
52
44
38
32
27
90
90
81
73
66
59
53
48
43
95
95
90
86
82
77
73
70
66
4-19
-------
where
P = 'probability that n modules are available,
n = number of modules, and
A = availability of a single module.
Inspection of the projected values indicates that a
definite reduction in overall FGD system availability can be
anticipated as the number of modules increases. To increase
system availability, it is necessary to add spare (redun-
dant) modules to compensate for those put out of service.
FGD system availabilities for various size systems with
one or two spare modules are presented in Table 4-3. In the
case of one spare module, the system availability is assumed
to be equal to the probability that all modules are avail-
able (including the spare) plus the probability that the
minimum required number of modules is available. This is
given by the following formula:
-tr -v £\ i" Li \ _L"™ii/ A"*
>n—l
where
P>n-l = probability that (n-1) or greater
— modules are available.
This assumes that all modules have the same availability and
are independent of each other.
For the case with two redundant modules, the system
availability is assumed to be equal to the probability that
all modules (including both spares) are available p]us the
probability that all modules but one are available plus the
probability that the minimum number of required modules (two
modules out of service) are available. This is given by the
following equation:
4-20
-------
Table 4-3. FGD SYSTEM AVAILABILITY PROJECTIONS BASED UPON USING
ONE AND TWO SPARE MODULES (X + 1 and X + 2) AT FULL BOILER LOAD
Boiler
Size
MW
100 - 150
150 - 300
300 - 450
450 - 600
600 - 750
750 - 900
900 - 1050
1050 - 1200
Minimum
Number of
Modules Required
to Operate at
Full Load (X)
1
2
3
4
5
6
7
8
System Availability Based Upon Single Module Availabilities of:
One Spare Module (X + 1)
70
91
78
65
53
42
33
26
20
75
94
84
74
63
53
44
37
30
80
96
90
82
74
66
58
50
44
85
98
94
89
84
78
72
66
60
90
99
97
95
92
89
85
82
77
95
99 +
99
99
98
97
96
94
93
Two Spare Modules (X + 2)
70
97
92
84
74
65
55
46
38
75
98
95
90
83
76
68
60
53
80
99
97
94
90
85
80
74
68
85
99 +
99
97
95
93
89
86
82
90
99 +
99 +
99
98
97
96
95
93
95
99 +
99 +
99 +
99 +
99 +
99
99
99
I
to
-------
>_n-2 2
Based on this analysis, it is expected that if each
module is available 90 percent of the time, one spare module
will achieve an overall system availability of 90 percent up
to a boiler size of about 750 MW. If each module is only 85
percent available, a single spare module will achieve system
availability of 80 percent up to about 750 MW. With two
spare modules, overall system availability is increased
dramatically.
The previous analysis assumes that the particular FGD
system is required 100 percent of the time. However, in
practice coal-fired boilers have approximately an 80-percent
requirement for availability. This inherent down-time
provides the opportunity for conducting preventative main-
tenance on the FGD system coincident with boiler servicing
periods. This maintenance should increase the operability
of the FGD system; that is, its capability to operate when
the boiler is in operation. In addition to the above con-
sideration, there is a reserve capacity margin for any
public utility's entire system of generating plants and/or a
regional power system of which one specific plant is a part.
These reserve capacities further allow a particular boiler
to reduce load and service its FGD module(s) without inter-
rupting the power supply- Typical percent reserve margins
for 1977 range from 18.7 to 50.9 percent during summer peak
periods and from 21.7 to 93.6 for winter peak periods.4
Total U. S. values are 29.2 percent for summer peak demand
and 4'o.9 percent for winter peak demand.
4-22
-------
Additionally, during times of forced outage on an FGD
system, there usually are "peaking" units within the par-
ticular utility system which can be used to provide the
makeup power required to cover the resultant partial load
operation. In case there are no other generating units
available within a particular utility system to meet the
outage demand, power can be purchased from another company
within the "grid" system which provides a regional back-up
reserve power margin.
It can be readily seen that a valid analysis of FGD
system redundancy needs to take many factors on system
reliability into consideration. ' ' Such an analysis would
require specific data for the utility company under study
and its interconnection with other companies in the same
grid. This analysis cannot be done within the scope of this
effort, however an analytical approach is outlined as fol-
lows:
Assume a reserve margin of 29 percent based on the
results reported by FPC.^ Furthermore, assume that the
plant size is 1000 MW, and eight scrubber modules are
installed (one spare). With one scrubber module down,
the plant can operate at full load. If two modules are
down, the plant can serve 6/7 or 85.7 percent of the
full load, and if three are down, 5/7 or 71.4 percent
of the full load. The probabilities of these different
situations, assuming an individual module availability
of 90 percent, are given in Table 4-4.
Thus, this particular FGD system will provide full-load
service (no outage) at least 81 (43+38) percent of the
time and 85.7 (6/7) percent of full load for an addi-
tional 15 percent of the time. Assuming the 29 percent
reserve margin, this FGD system can provide more than
85 percent of the load for 96 (81+15) percent of the
time. The FGD system will serve at least 71.4 percent
of full load 99.5 percent of the time.
4-23
-------
Conclusion; Thus, there would be a deficiency, with
respect to serving at least 71 percent of the full
load, only 0.5 percent of the time.
Assuming a module availability of 80 percent, the
probabilites shown in Table 4-5 for corresponding
outages are obtained.
In this case, the FGD system would serve at least 71
(5/7) percent of the full load for 94.5 percent of the
time and would be deficient 5.5 percent of the time.
Conclusion; One spare module would be considered
adequate under the assumption that each module has 80-
percent availability and that there is a 29-percent
reserve margin.
4.2.2.3 Manufacturer's Availability Guarantees - Seven out
of twelve FGD system suppliers indicated a willingness to
guarantee overall availability of their systems. Ninety
percent was the typical "level-of-performance" guaranteed as
shown in Table 4-6. In most cases the availability is
negotiable and subject to guarantees from component sup-
pliers for varying amounts of time up to two years.
In addition to SC^ removal efficiencies and avail-
abilities, vendors are guaranteeing additional parameters.
For instance, the following items are included in the
guarantee at the Conesville Power Plant in Ohio:
0 lime usage rate, Ibs/hr
0 makeup water rate, Ib/hr
o
o
o
pressure drop, inches of water
particulate emissions, no increase due to scrubber
system availability during a 6-month demonstration
test, 90 percent overall.
4.2.3 Summary
The ability to operate at a high level of pvailability
has been demonstrated by a number of FGD systems on both
high- and low-sulfur coal. High availability is largely a
function of design and maintenance, and as more operating
4-24
-------
Table 4-4. PROBABILITIES AND OUTAGES FOR SPECIFIED
NUMBER OF SCRUBBER MODULES OPERATING, ASSUMING
MODULE AVAILABILITY OF 90 PERCENT - 1000-MW BOILER
Number of scrubber Corresponding Corresponding
modules available probability outage, MW
8 (0.90)8 = 0.430 = 43% n
7 8(0.90)7(0.10) = 0.383 = 38% 0
6 28(0.90)6(0.10)2 = 0.149 = 15% 143
5 56(0.90)5(0.10)3 = 0.033 = 3.3% 286
< 5 0.005 >_ 286
Total Probability 1.000
Table 4-5. PROBABILITIES AND OUTAGES FOR SPECIFIED
NUMBER OF SCRUBBER MODULES OPERATING, ASSUMING
MODULE AVAILABILITY of 80 PERCENT - 1000-MW BOILER
Number of scrubber Corresponding Corresponding
modules available probability outage, MW
8 (0.80)8 = 0.168 = 17% 0
7 8(0.80)7(0.20) = 0.336 = 34% 0
6 28(0.80)6(0.20)2 = 0.294 = 29% 143
5 56(0.80)5(0.20)3 = 0.147 = 15% 286
£ 4 = 0.055 = 5.5% >_ 429
Total Probability 1.000
4-25
-------
Table 4-6. SUMMARY OF AVAILABILITY GUARANTEES
OFFERED BY MANUFACTURERS3
Company
Guarantee offered
Yes (level!
No
A
B
C
D
F
G
H
I
J
K
L
Normally better than 90%
Typically 90% during performance
testing; sometimes up to 95%
Maximum of 90% based on boiler
hours
Yes (level of guarantee not dis-
closed)
Have guaranteed in excess of 90%
Normally 85 to 90% for 1 or 2 years
Maximum of 90% on a case-by-case
basis
X
X
X
X
X
b
Represents the responses of 12 manufacturers.
Cqmpany names are deliberately withheld.
4-26
-------
experience is gained, the methods for achieving higher
availabilities are being implemented. Installations of an
extra or redundant scrubber module can significantly increase
overall system availability since repairs can be made with
the boiler in operation.
4.3 SYSTEM EFFICIENCY
As presented in Chapter 3, all five of the demonstrated
FGD systems (lime, limestone, double alkali, magnesium
oxide, and Wellman-Lord) have achieved S02 removal effi-
ciencies in excess of 90 percent. Systems which have
reported 90 percent SC>2 removal or greater are identified in
Table 4-7 by utility, station, unit no., removal efficiency,
and nature of application.
In Japanese FGD installations, these same processes are
reported to have demonstrated SC>2 removal efficiencies well
in excess of 90 percent, many in excess of 95 percent.
However, only a few of these are on coal fired boilers.
Extensive pilot, prototype, and demonstration scale
data have been reported to document the complex relation-
ships between process parameters and SC>2 removal efficiency.
Among the test facilities are:
Shawnee TVA 10 MW Lime/Limestone
Mohave So. Calif. Edison 170 MW Lime/Limestone
Four Corners Arizona Public Ser. 160 MW Lime
Mystic Boston Edison 150 MW Mag. Oxide
St. Clair Detroit Edison 163 MW Limestone
Scholz Gulf Power 20 MW Double alkali
20 MW Activated carbon
The overall removal efficiency capability has progressed
to the point that almost all vendors will guarantee SC>2
removal efficiences in excess of 90 percent, as shown in
Table 4-8.
4-27
-------
Table 4-7. PLANTS REPORTING 90 PERCENT OR GREATER S02 REMOVAL
Utility Company
Arizona Public Service
Duquesne Light
Louisville Gas & Electric
Northern Indiana Public
Service
Philadelphia Electric
Tennessee Valley Authority
Boston Edison
Detroit Edison
General Motors
Gulf Power
Potomac Electric & Power
Southern California
Edison
U.S. Aii Force
Kentucky Utilities
Station
Cholla
Four Corners
Phillips
Cane Run
Paddy ' s Run
Mitchell
Eddystone
Shawnee
Mystic
St. Clair
Parma
Scholz
Scholz
Scholz
Dickerson
Mohave
Mohave
Rickenbacker
Green River
Unit
No. 1
No. 5
Nos. 1-6
No. 4
No. 6
No. 11
No. 1
No. 10
No. 6
No. 6
No. 1-4
No. 1
No. 1-2
No. 2
No. 3
No. 1
No. 2
No. 1-9
No. 1-3
Nature
Full-scale
Demonstration
Full-scale
Full-scale
Demonstration
Demonstration
Demonstration
Prototype
Demonstration
Demonstration
Full-scale*
Prototype
Prototype
Prototype
Demonstration
Demonstration
Demonstration
Full-scale**
Full-scale
% SO 2 Removal
92
95
90+
90
99.5
90
95-98
95-99
90
90-91
90+
95
95
90+
90
95
95
99
90+
I
NJ
OO
**
Industria"1
Military Base
-------
Table 4-8. GUARANTEES OFFERED BY MANUFACTURERS FOR SO2 REMOVAL
I
tsJ
Company
C
D
E
F
<90
Would normally
guarantee 80-85%
Level of SO- removal guaranteed
90%
Minimum guarantee given
Minimum guarantee given
This guarantee is normally
given
This guarantee is given
where SC>2 inlet concentra-
tion is 500-4,000 ppm
This guarantee is given
where low sulfur coal is
utilized
Minimum guarantee given
This guarantee is usually
given with coal having 3-4%
sulfur
This guarantee is normally
given with low or high
sulfur coal
Minimum guarantee given
>90%
Is willing to offer 95%
guarantee on case by case basis
For >90%, it is based on inlet
SO- concentration
Would guarantee 95% in all cases
Would guarantee^ up to 92%
Have guaranteed >90% in the past
Depending upon the process, they
would guarantee >90%
Have guaranteed up to 95% in the
past
Are prepared to offer better than
90% with low or high sulfur coal,
but would not guarantee less than
50 ppm SO2 concentration in exit
stream
In many cases they guarantee 95%
with high sulfur coal
May guarantee up to 95% on a case
by case basis
Company names are deliberately withheld.
-------
4.3.1 Lime and Limestone Systems
Many of the current operating lime and limestone scrub-
bing systems were designed for S02 collection efficiencies
under 90 percent to satisify an applicable regulation.
Often, an efficiency in the range of 60 to 70 percent was
sufficient to meet the specific emission limitations and
therefore, values of this magnitude were used for design
criteria. Design of newer systems which are required to
achieve high efficiency must take into account a number of
key design variables including:
inlet SO2 concentration rate
liquid to gas ratio
scrubber gas velocity
scrubber liquor inlet pH
type of absorber
magnesium content
type of alkali
Inlet SO? concentration and rate - Higher removal effi-
ciencies can be more easily achieved at lower S02 inlet
concentrations because the amount of SQ2 that must be ab-
sorbed per unit of scrubbing liquor to achieve a specified
outlet concentration is smaller. When the scrubbing liquor
contacts a high SC>2 concentration it quickly reacts with the
SO2, decreasing further absorption capability- Since the
dissolution rate of the alkali in the scrubber media is
slower than the reaction rate with SC>2, a "back pressure" or
resistance to further absorption is created which can inhibit
a scrubber efficiency. At low S02 concentrations the
alkali in the liquor can react with a greater percentage of
the S02 and affect a greater removal efficiency under a
given set of operating conditions.8'9
As the back pressure of S02 in solution approaches the
partial pressure of S02 in the flue gas, absorption de-
creases rapidly. However, by contacting the lowest S02
4-30
-------
bearing gas with the fresh slurry (countercurrent flow),
high efficiency can be maintained. At the Mohave Station
greater than 95 percent efficiency was achieved at an S0_
inlet concentration of 200 ppm with both lime and limestone
and with three different types of absorbers. The TVA
Shawnee facility reports 95 percent removal at high S0~
inlet concentrations.
Under typical plant operation, the inlet SO concentra-
tions will vary directly with changes in fuel sulfur con-
tent. To maintain a desired level of efficiency, variable
amounts of alkali must be brought into contact with the flue
gas. This can be accomplished rapidly by directly varying
the liquor flow rate based on the inlet S02 readings, which
is a "feed-forward" type of control system. That is, the
scrubbing system responds to a signal which indicates SO..,
concentrations before the scrubber.
Liquid to gas ratio - Higher efficiencies are realized at
higher L/G ratios for lime and limestone systems, as pre-
viously shown in Figures 3-11 to 3-13, and 3-30 and 3-31,
respectively. For a given absorber, increased L/G ratios
will yield higher efficiencies until flooding and poor gas
distribution occur. For new designs, absorbers which can
accomodate high L/G ratios can be selected and high effi-
ciency maintained. Higher L/G ratios also require larger
pumps, pipes, and slurry reaction tanks. Again, these can
be designed into the system and should cause no unusual
operating problems. Additionally, scaling potential will
actually decrease since the slurry contains less SO per
unit volume.
4-31
-------
The following L/G ratios have yielded efficiencies in
excess of 90 percent:
Spray Tower: Lime or limestone - greater than 10. 1/m
(80 gal/1000 acf)
Turbulent Contact Absorbers: Lime - 8 1/m (60 gal/
1000 acf)
Turbulent Contact Absorbers: Limestone - 10-7 1/m (80
gal/1000 acf)
Scrubber Gas Velocity - The effects of changes in flue gas
absorber velocity on SO- removal efficiency, when other
variables are kept constant, vary with the type of absorber.
For a spray tower, the efficiency decreases at a fixed L/G
ratio (Figure 3-21). This effect is much less noticeable on
packed and turbulent contact type absorbers where increased
agitation and mixing partially off-set the decreased L/G
ratio. Velocities in the range of 2.1 to 2.4 m/s (7 to 8
ft/s) and 2.7 to 3 m/s (9 to 10 ft/s) appear desirable for
spray tower and turbulent contact absorbers, respectively.
Scrubber liquor inlet pH - This parameter has a direct and
noticeable effect on S02 collection efficiency as shown
previously in Figures 3-17 and 3-32 for lime and limestone
slurries, respectively. Increased efficiency is achieved at
higher pH since more alkali is available and higher dis-
solution rates are achieved (as previously noted, operation
at very high pH, however, causes scaling problems).
Based on tests at EPA's Shawnee Test Facility, a pH of 5.8 and
8.0 should be used for limestone and lime slurry systems,
respectively. Operation at these pH levels requires lime
and limestone in excess of that required to chemically react
with the S02 (about 20 percent excess for lime and 40 per-
cent excess for limestone). Maintenance of the desired pH
by careful measurement and close control of reagent feed and
4-32
-------
mixing system will prevent the pH variations which reduce
efficiency (if too low) or cause scaling (if too high).
Type of absorber - A large variety of absorber designs have
been utilized to achieve S02 removal efficiencies as high as
99 percent. These include cross-flow horizontal spray
chambers (Weir), spray towers, packed-grid towers, and TCA
(mobile bed) absorbers. The venturi type has also been
used, however, it is more useful as a particulate removal
scrubber and not as efficient for S02 absorption. The final
selection and design of an absorber is usually based on
previous test data and on the required liquid and gas flow
rates. Spray towers (either horizontal or vertical) offer a
number of advantages including simple internal design which
decreases scaling potential, acceptance of high liquid flows
and decreased maintenance as reported at the Mohave and
Shawnee plants.
Magnesium Content - The addition of relatively small amounts
of soluble magnesium (less than 1 percent by weight) to the
scrubber liquor in the form of magnesium oxide, magnesium
sulfate, or dolomitic lime (in lime systems) can greatly in-
crease the SC>2 collection efficiency of the system. Magne-
sium compounds are much more soluble, compared to calcium,
and can react rapidly in the liquid phase with SO-. The
effect of magnesium addition was shown in Figures 3-18 and
3-34 for lime and limestone systems respectively, and was
tested at the Paddy's Run plant and Phillips plant (lime
based FGD systems) . These data show SC>2 collection effi-
ciencies in excess of 90 percent when the effective mag-
nesium concentration is over 0.5 percent. Any chloride
present in the liquor will react with the magnesium to form
inactive magnesium chloride. Additional magnesium must be
added, above that which will react with any chlorides, to
4-33
-------
yield an effective or available magnesium concentration.
Since soluble magnesium compounds can be lost in the sludge,
possible treatment of any water discharge and addition of
fresh magnesium to replenish losses is required.
Two proprietary absorbents have been developed, one by
Dravo (Thiosorbic lime: 2 to 6 percent magnesium oxide
lime) , the other by Pullman Kellogg (Catalytic limestone:
3 to 27 percent magnesium sulfate limestone). Three full-
scale, operational FGD systems (Columbus and Southern Ohio
Electric, Conesville No. 5; Pennsylvania Power, Bruce Mans-
field No. 1 and No. 2) are now using Dravo's thiosorbic lime
reagent. In addition, six more full-scale systems plan to
use thiosorbic lime in future FGD operations:
Big Rivers Rural Electric Power Cooperative, Reid No. 2
Columbus and Southern Ohio Electric, Conesville No. 6
Duquesne Light, Elrama Nos. 1-4
Duquesne Light and Phillips Nos. 1-6
Indianapolis Power and Light, Petersburg No. 3
Pennsylvania Power, Bruce Mansfield No. 3
Type of alkali - Both lime and limestone are effective
compounds for SO- removal. Lime is more reactive and lime
scrubbing systems are more efficient than limestone systems
under a given set of conditions. However, since the type of
alkali must be initially considered in the design phase,
either system can be engineered to yield high efficiencies.
A model has been developed by Bechtel, based on their
tests at the Shawnee pilot plant, that incorporates the
various factors that affect lime and limestone scrubbing
systems. These models are useful for selecting design
factors for conditions within the model's bounds.10
4-34
-------
4.3.2 Sodium Carbonate and Double Alkali FGD Systems
There are currently three sodium carbonate scrubbing
systems in operation. These systems do not recycle any
sodium carbonate but operate on a "once-through" basis (see
Section 3.6.7). Double alkali systems, which produce a
sludge and recycle the absorbent solution, are in operation
in a number of large industrial sized boilers, but not on
utility sized units (Section 3.3). Because of the more
reactive nature of sodium carbonate and its solubility in
water, the sodium carbonate scrubbing systems have in
general experienced high efficiencies (greater than 90
percent at Reid Gardner) and few operating problems in the
absorber area.
Based on the limited application of these systems, the
following guidelines should be followed to achieve high
efficiency:
(a) Utilization of a prescrubber with a separate water
recirculating system for control of particulates
and chlorides for high chloride coal (>0.04 per-
cent Cl by weight in the coal).
(b) The use of a two-stage tray or packed tower ab-
sorber with an L/G of 1.3 to 2.7 1/m3 (10 to 20
gal/1000 acf). Typically the absorber pressure
drop is 15 to 30 cm (6 to 12 inches) of water.
(c) The absorber scrubbing liquor pH being recycled to
the absorber should be in the 6.0 to 7.0 pH range.
(d) If lime regeneration is used in a double alkali
system, the reaction tank residence time should be
approximately 10 minutes. Lime utilization of
about 90 percent may be expected.
(e) If limestone regeneration is used, the reactor
tank residence time should be approximately 30
minutes. Limestone utilization of 75 to 85 per-
cent may be expected.
(f) Carbonate softening and sodium ion make-up can be
accomplished through addition of sodium carbonate
(soda ash).
4-35
-------
4.3.3 Magnesium Oxide FGD Systems
This system has received considerable attention in
pilot plant studies and is currently installed on one coal
fired power plant in this country and two industrial appli-
cations in Japan (Section 3.4). Due to the reactive nature
of magnesium oxide, efficiencies in excess of 90 percent
have been achieved with various types of absorbers. Based
on experience at the Mystic and Dickerson Stations, the
following conditions should be used as guidelines to achieve
high efficiency:
(a) High efficiencies (99%) particulate removal should
be accomplished by use of ESP(s) or venturi scrub-
ber (s) .
(b) A prescrubber should be used to remove any remain-
ing particulate and most of the chlorides and SO.,.
(c) Utilize venturi absorbers, typically operating at
a pressure drop of 25 cm (10 inches) of water or
greater, or Turbulent Contact Absorbers operating
at approximately 20 cm (8 inches) of water pres-
sure drop, at an L/G of 5.3 to 6.7 1/m3 (40 to 50
gal/1000 acf).
(d) The absorber superficial gas velocity should not
exceed approximately 3.0 m/sec (10 ft/sec) range.
(e) The slurry pH measured at the absorber discharge
should be maintained in the 6.0 to 7.5 range.
4.3.4 Wellman-Lord FGD Systems
Experimental data and full scale operating data on the
single Wellman-Lord FGD system on a coal-fired boiler in
this country have demonstrated levels of SO,, collection
efficiency in excess of 90 percent with 3.5 percent-sulfur
coal. The following conditions should be used as guidelines
to obtain high efficiency.
4-36
-------
(a) Install a prescrubber with a separate water recir-
culation system for final particulate control and
reduction of 803 and chlorides.
(b) Use a three to five tray absorber with an L/G of
1.0 to 1.3 1/m3 (6 to 10 gal/1000 acf). Typically
the absorber pressure drop is 27 to 44 cm (10.5 to
17.5 inches) of water.
(c) A superficial gas velocity in the range of 2.7 to
3.1 m/sec (9 to 10 ft/sec) should be used.
(d) Maintain the recycled sodium sulfite scrubbing
solution at a pH of 6.0 at the absorber inlet.
(e) The system makeup of fresh, 20-percent sodium
carbonate solution should be approximately 0.07
1/m3 (0.5 gallon/1000 acf) per tray.
(f) As the inlet S02 concentration decreases, the
number of trays required to obtain high SO-
removal should be increased.
4.3.5 Summary
The ability to achieve 90 percent S02 removal effici-
ency has been demonstrated by lime, limestone, double alkali,
Wellman Lord and magnesium oxide FGD systems. However, all
FGD systems have not demonstrated equal capability to achieve
high efficiency on both high- and low-sulfur coal. Thus,
for a particular application proper scrubber selection and
design is required to achieve high efficiency.
4-37
-------
REFERENCES FOR SECTION 4.0
1. Summary Report - Flue Gas Desulfurization Systems -
June-July 1977, PEDCo Environmental, Inc., Cincinnati,
Ohio. EPA Contract 68-01-4147, Task No. 3.
2. Personal Communication with FGD System Suppliers and
Users Performed in Support of Contract 68-01-4147, Task
No. 3. July, August 1977.
3. Report on Equipment Availability for the Ten-Year
Period - 1966-1975. EEI Publication No. 76-85.
December 1976.
4. Electric Power Supply and Demand 1977-1986 as Projected
by the Regional Electric Reliability Councils in Their
April 1, 1977 Responses to FPC Order 383-4 (Docket
R-362). Federal Power Commission, Bureau of Power
Staff Report.
5. Weiss, Joel R. A Reliability Model for Interconnected
Electric Power Systems. IEEE Transaction on Reliability,
UR R-24, No. 2. June 1975.
6. Billinton, Roy, et al. Power-System Reliability Cal-
culations. MIT Press, Cambridge, Massachusetts. 1973.
7= Billinton, Roy. Power System Reliability Evaluation.
Gordon and Breach, Science Publishers, New York. 1970.
8. Potts, J.M., et al. Removal of Sulfur Dioxide from
Stack Gases by Scrubbing with Limestone Slurry: Small-
Scale Studies at TVA. Second International Lime/Lime-
stone Wet Scrubbing Symposium, New Orleans, November
8-12, 1971.
9. 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.
10. Bechtel Corp. EPA Alkali Scrubbing Test Facility. TVA
Shawnee Power Plant, Paducah, Kentucky. U.S. Environ-
mental Protection Agency. January 1977 Monthly Report.
p. 3-2.
4-38
-------
APPENDIX A
DOMESTIC LIME SLURRY FGD SCRUBBING SYSTEMS
A-l
-------
This appendix summarizes data on the three lime scrub-
bing FGD systems presented in Section 3.1 and also describes
the other operating lime systems in this country.
The following FGD systems are described:
0 Bruce Mansfield No. 1
0 Louisville Gas and Electric - Cane Run No. 4
0 Montana Power Co. - Colstrip Nos. 1 and 2
0 Columbus and Southern Ohio Electric Co.
Conesville No. 5
" Duquesne Light Co. - Elrama Power Station
0 Arizona Public Service - Four Corners No. 5
0 Kentucky Utilities - Green River Boilers 1, 2,
and 3
0 Southern California Edison - Mohave
0 Louisville Gas and Electric Co. - Paddy's Run No.
6
0 Duquesne Light Co. - Phillips Power Station No. 6
0 Rickenbacker Air Force Base
0 Tennessee Valley Authority - Shawnee No. 10A and
10B
A-2
-------
PENNSYLVANIA POWER
BRUCE MANSFIELD NO. 1
BACKGROUND INFORMATION
The Bruce Mansfield plant is a 2700-MW, three-boiler,
coal-fired facility located on the Ohio River in Shipping-
port, Pennsylvania. This facility was built by Pennsylvania
Power Co., which is acting on its own behalf and as an agent
for the other participating companies, the Cleveland Elec-
tric Illuminating Co., Duquesne Light Co., Ohio Edison Co.,
and Toledo Edison Co.
Bruce Mansfield Nos. 1 and 2 are coal-fired, once-
through, supercritical steam generators, which fire 333
tons/hr of coal and generate approximately 6.5 MM Ib/hr
(each) of steam at 3785 psig and 1005°F. The emission
control equipment required for this unit is designed to meet
state emission regulations of 0.6 Ib sulfur dioxide/10 Btu
of heat input and 0.0175 gr/scf of particulate when burning
11,900 Btu/lb coal having average ash and sulfur contents of
12.5 and 4.7 percent, respectively. Additional design-
related information is presented in Table A-l.
A-3
-------
Table A-l. FGD SYSTEM DATA, BRUCE MANSFIELD NO. 1
Unit location
Unit rating
Fuel characteristics
FGD vendor
Process
New or retrofit
Start-up date
Efficiency, particulates
Efficiency, S02
Sludge disposal
Shippingport, Pennsylvania
825 MW
Coal with 4.7 percent
sulfur content (average)
Chemico
Lime scrubbing
New
4/76
99.8 percent (design)
92 percent (design)
Stabilized sludge is dis-
posed of in unlined
reservoir
POLLUTION CONTROL
The emission control system is a venturi wet-scrubber
system for removal of fly ash and sulfur dioxide. It was
manufactured by Chemico Air Pollution Control Co. and
utilizes Dravo Corporation's Thiosorbic limea as the scrub-
bing absorbent. The emission control systems for Boilers 1
and 2 are identical. The system planned for Boiler 3 in
1978 will utilize a Weir Horizontal Scrubber system.
Each scrubbing train consists of a variable-throat
venturi, a 9000-hp I.D. fan, and a fixed-throat venturi in
series. There are six scrubber trains per boiler. The
design called for five trains to handle the flue gas flow
from the boiler at full load, but due to higher flue gas
Dravo's patented lime which contains 6 to 12 percent
magnesium oxide.
A-4
-------
flows, all six are required. The scrubbed flue gas is re-
heated by a direct oil-fired burner and then exhausted
through a 950-ft stack.
The variable-throat venturi removes fly ash contained
in the flue gas and the absorber module removes the re-
maining fly ash. Sulfur dioxide is absorbed in both the
particulate scrubber and the absorber by the lime slurry,
which contains 2 to 6 percent magnesium oxide. The unit is
guaranteed to remove 92 percent of the sulfur dioxide and
99.8 percent of the particulate matter from the flue gas.
The scrubber discharge stream is combined with fly ash
slurry and discharged to a 200-foot-diameter thickener.
Sludge from the thickener underflow is pumped to a waste
disposal system, where it is mixed with a stabilizing agent
(Calcilox), and then pumped approximately 7 miles to the
Little Blue Run Ravine for landfill.
The initial shakedown and debugging phase of operation
began for part of the system in December 1975. Full com-
mercial operation commenced in May 1976.
PERFORMANCE HISTORY
This facility reported 100 percent operability during
the first seven months after start-up. Operating problems
were solved without causing boiler downtime. However,
during the cold winter months of January and February 1977,
the boiler lost 11 and 24 percent generating capability,
respectively, because of extreme weather conditions. On
March 21, 1977, the boiler was shut down for a 10-week
turbine overhaul. During this time, work began on replacing
the polyester flakeglass lining in the exit flue serving
three scrubber modules. Failure resulted from poor applica-
A-5
-------
tion of the original lining, which was accomplished by
spraying, and probably from the lack of flue gas reheat and
poor mist elimination. The new lining is being troweled on.
The other chimney flue, for the D, E, and F modules, also
requires repair. Roughly 1 year will be required to com-
plete this work, and the boiler will be held to half load
for that time. Also, during the reduced load operation,
Pennsylvania Power will be making other system changes as
2 3
described in the following paragraphs. '
Mist Eliminator - The company has experienced excessive mist
carry-over during scrubber operation. The mist eliminator
was designed for 1 gr/scf liquid carry-over but an actual
carry-over of about 3 gr/scf has been estimated by plant
personnel. It is also designed to operate at gas velocities
of 8 to 10 ft/sec. Pennsylvania Power would like to install
a vertical mist eliminator in the ducting downstream of the
scrubber vessels. Because of duct diameter and space re-
strictions, however, such a design would experience flow
velocities as high as 50 ft/sec. A vertical mist eliminator
was installed but collapsed as a result of structural fail-
ure caused by high flue gas velocity. The problem of
plugging was also apparent, but it has been solved by re-
locating the lime slurry feed to the scrubber to a lower
position.
Closed Loop - The system is not being operated closed-loop,
because water is being retained at the sludge disposal site
and not recycled to the process. Makeup water from the
disposal pond is not needed since "fresh" water is added to
the system in the fan sprays and during lime slaking. The
company thinks they can operate closed loop, but have not as
yet because they have been concentrating on the system's
mechanical problems. Also, in the event they do not operate
A-6
-------
closed loop, the plant has a permit to discharge water to
the Ohio river. Discharge requirements for water are
total suspended solids of 60 mg/1 or less; chloride con-
tent of 10 mg/1 or less; and a pH between 6 and 9.
I.D. Fans - Pieces of scale carried through the system have
caused the rubber lining of the fan scrolls to be chipped
away, resulting in exposure and corrosion of the carbon
steel. These fan scrolls are now being replaced with ones
manufactured of Inconel. Pitting has also occurred to the
carbon steel hubs because of seal leakage. The fan shafts
are fabricated of carbon steel clad with Carpenter 20.
Reheater - The reheaters are not used because of duct vibra-
tions caused by resonance. The reheaters are now undergoing
design changes to correct the resonance problem. Plume
buoyancy appears to be sufficient without reheat, but on
certain days atmospheric conditions cause condensation and
precipitation of moisture from the plume. Winds are such
that the liquid fallout occurs in Shippingport. The fallout
occurs as a clear liquid, but leaves a film upon drying.
The pH of the condensed liquid is about neutral.
Scrubber - Scaling of the system has not been a serious
problem. Scale has built up to about 1/2 to 1 inch over a
1-year period. Some plugging of the venturi nozzles with
scale has occurred. This problem was eliminated by instal-
lation of "baskets" in the main header to catch the scale
particles upstream of the nozzles. The baskets can be
cleaned without shutting the system down. Maintenance of
the scrubber system is not performed on a preventive basis
but is done on the basis of need. Scheduled maintenance
occurs every 12 months.
Performance Test - The Pennsylvania Department of Environ-
mental Resources tested the unit on July 19 and 20, 1977.
A-7
-------
S02 emissions were 0.44 Ib SO /106 Btu and 1.26 lb/10 Btu,
respectively. This represents 94 and 83 percent removal.
The allowable emission rate is 0.6 lb/10 Btu. Since these
are recent tests, only the results are available at this
time. The details and causes for such a wide variation in
24
results have not been published. EPA, OAQPS, is obtaining
test data from this facility by using continuous monitors.
These data will be presented in a separate EPA report.
A-8
-------
LOUISVILLE GAS AND ELECTRIC
CANE RUN NO. 4
BACKGROUND INFORMATION
The Cane Run Power Station, located in Louisville,
Kentucky, is operated by the Louisville Gas and Electric
Company. The plant has six electric power steam generating
units, providing a total steam turbine net generating capa-
city of 992 MW.
Unit No. 4 is a coal-fired boiler with a continuous net
generating capacity of 178 MW= The unit has a maximum power
generation capacity of 190 MW. The unit heat rate is 10,030
Btu/kWh. The boiler is currently burning coal with a gross
heating value of 11,500 Btu/lb and average sulfur and ash
contents of 3.5 to 4.0 percent and 11 to 12 percent, re-
spectively.
POLLUTION CONTROL
The emission control system for this unit consists of
an electrostatic precipitator (ESP) upstream of a wet scrub-
bing system. The ESP provides primary particulate control
while the wet scrubbing system provides additional particu-
late removal and primary sulfur dioxide control.
The FGD system consists of two identical parallel
scrubbing trains designed and installed by the American Air
Filter (AAF) Company. The wet scrubbing system utilizes a
slurry of carbide lime (a waste by-product obtained from a
nearby acetylene manufacturing plant). The hydrated lime
A-9
-------
contains 90 to 92 percent calcium hydroxide, 2.0 to 2.5
percent silica, 3 to 8 percent calcium carbonate, and 0.1
percent magnesium oxide. Design data are presented in Table
A-2.
Table A-2. FGD SYSTEM DATA, CANE RUN NO. 4
Unit location
Unit rating
Fuel characteristics
FGD vendor
Process
New or retrofit
Start-up date
Efficiency, particulates
Efficiency, S02
Sludge disposal
Louisville, Kentucky
178 MW
Coal 3.5 to 4.0 percent
sulfur
American Air Filter
Lime scrubbing
Retrofit
August 1976
99 percent (guaranteed)
85 percent (guaranteed)
Stabilized sludge is dis-
posed of in unlined pond
Each scrubbing train is equipped with a guillotine-type
bypass damper allowing bypassing of the gas around the
scrubbers. Each scrubbing train contains the following
major pieces of equipment: I.D. booster fan, quench sec-
tion, flooded elbow, mobile bed contactor, centrifugal
demister, and a three-section reactant tank system. The
waste disposal system consists of a 75-foot-diameter thick-
ener for liquid-solids separation and a fly ash pond for
ultimate disposal of the thickener underflow. The scrubbing
wastes are stabilized with fly ash, and water is recycled
A-10
-------
back to the process. The system does not use stack gas
reheat. A wet stack lined with Carbolyne obviates the
4
necessity of reheat.
PERFORMANCE HISTORY
Available data on operability of the FGD system at Cane
Run No. 4 are given in Figure 3-24. These data show a high
level (90%) of operation during the first four months of
operation, before the lime supply was curtailed. Operation
in early 1977 was very limited before the unit was shut down
for modification; which consisted of the addition of turning
vanes in the absorber and conversion to a chevron-type mist
eliminator.
There are no published SO,, efficiency data. In re-
sponse to a questionnaire from the EPA in June of 1977, the
utility reported preliminary SO- removal efficiencies of 80
5
percent. These data were taken before the absorber gas
distribution modifications were made. The EPA is currently
running more definitive efficiency tests, which will be
documented in a separate EPA report.
A-ll
-------
MONTANA POWER CO.
COLSTRIP NOS. 1 AND 2
BACKGROUND
The Colstrip Power Station located near Colstrip,
Montana, is operated by the Montana Power Company. The
station consists of two, relatively new pulverized-coal-
fired boilers. Each boiler is rated at 360 MW of power
generating capacity. The station is located at the mouth of
a mine and burns low-sulfur coal. Further details can be
found in Table A-3. Because of the relatively low sulfur
content of the coal (0.8%), a design efficiency of only 60
percent was utilized.
POLLUTION CONTROL
Both Unit No. 1 and Unit No. 2 are equipped with FGD
systems. The systems are Arthur D. Little/Combustion Equip-
ment Associates (ADL/CEA) lime/alkaline fly ash scrubbing
systems designed to remove both particulate matter and
sulfur dioxide from the flue gas. Design removal efficien-
cies are 99 percent for particulate matter and 60 percent
for sulfur dioxide, as shown in Table A-3. The efficiencies
are sufficient to meet existing applicable codes.
A-12
-------
Table A-3. FGD SYSTEM DATA, COLSTRIP NOS. 1 AND 2
Unit location Colstrip, Montana
Unit rating 360 MW each
Fuel characteristics Coal, 0.8 percent sulfur
FGD vendor ADL/LEA
Process Lime/alkaline fly ash
scrubbing
New or retrofit New
Start-up date July 1976
Efficiency, particulates 99 percent
Efficiency, SO 60 percent
Sludge disposal Unstabilized sludge is
disposed of in unlined
pond
Each boiler is equipped with three scrubber modules,
each module capable of handling 40 percent of the boiler
flue gas capacity (1,430,000 acfm at 291°F). Each scrubber
module incorporates a variable-throat venturi (design super-
ficial gas velocity - 200 ft/sec), a spray absorption sec-
tion (design superficial gas velocity - 11.5 ft/sec), an
internal recycle tank, a wash tray, and a chevron-type mist
eliminator. The venturi and spray sections operate on a
recirculating slurry loop containing 12 percent solids
(consisting primarily of calcium salts and fly ash). Total
module pressure drop is approximately 23.5 inches ^O.
Total module liquid-to-gas ratio is approximately 33 gal/
1000 acf. A wash tray is situated beneath the mist elimina-
A-13
-------
tor in order to dilute any entrainment reaching the com-
ponent, thus reducing the potential for plugging and scal-
ing. Slurry for this tray comes from a separate recycle
tank and recirculation loop. Reheat is supplied by steam in
heat exchanger tube bundles. The reheaters are also equipped
with soot blowers for cleaning.
The scrubber booster fans (induced draft, 43.5 inches
H?0) are located downstream of the scrubbers and operate on
dry, reheated flue gas. The fans were located downstream in
this scrubber design because of the high particulate loading
of the inlet flue gas.
The scrubbing wastes (unreacted fly ash, 9 parts
calcium sulfate/1 part calcium sulfite) are discharged to an
initial holding basin (520 acre-feet) adjacent to the scrub-
ber plant and then pumped (50.8 dry tons per hour/2 units)
3 miles to a permanent sludge disposal pond (2218 acre-feet,
clay-lined where necessary). Clear liquor is returned from
ponds for additional scrubbing service. Fresh makeup water
to the system is limited to approximately the amount of
water lost in the scrubber plant via evaporation and sludge.
PERFORMANCE HISTORY
The Unit No. 1 FGD system was started up in late 1975.
Following a shakedown and debugging period, the unit has
been operating at full commercial capacity, and, according
to the designers, approaching 100 percent availability. The
Unit No. 2 FGD system started up in July 1976. The utility
tested the units in 1976 and found both of them to be in
compliance with the regulation calling for 1.2 Ib SO2/10
Btu. Some test runs indicated S09 removal efficiencies in
c ^
excess of 90 percent.
A-14
-------
COLUMBUS AND SOUTHERN OHIO ELECTRIC CO.
CONESVILLE NO. 5
BACKGROUND
The Conesville Power Station is located on the Muskingum
River, near Conshocton, in northeast Ohio. The plant has a
current capacity of 1644 MW (design) with an additional 411
MW under construction. Units 1, 2, and 3 have a combined
capacity of 433 MW and share a common stack. Unit 4 is
rated at 800 MW, and Unit 5 is rated at 411 MW. Unit 6,
currently under construction, will also be rated at 411 MW.
Units 4, 5, and 6 each have a separate stack.
Boiler 5 is a dry-bottom, pulverized-coal-fired Combus-
tion Engineering unit, installed in 1976. The plant burns a
mixture of high-sulfur Ohio coals, with sulfur contents of
from 4.2 to 5.1 percent, ash contents of from 12 to 19
percent, and heating values of from 10,300 to 11,220 Btu/lb.
Forty percent of the coal is delivered by conveyor from a
nearby (7 miles distant) coal mine complex. The remainder
is trucked in from southeast Ohio.
POLLUTION CONTROL
The air pollution control system at Conesville No. 5
consists of a Research-Cottrell cold-side ESP, followed by
two Universal Oil Products (UOP) SO2 absorber modules in
parallel. The ESP is designed for 99.65 percent removal
efficiency and the Turbulent Contact Absorbers (TCA) are
designed for 89.6 percent S02 removal efficiency. The
A-15
-------
system is designed for an outlet SC>2 loading of 1.0 lb/
million Btu of heat input. Boiler I.D. fans are located
immediately downstream of the ESP-
Following the I.D. fans, the flue gas enters the two
parallel TCA scrubbing trains. Each absorber is capable of
handling 60 percent of the flue gas flow. A presaturator
section lowers the flue gas temperature from 286° to 125°F,
and provides some initial S02 removal. The gas then enters
the neoprene-lined, carbon steel absorber modules, where two
stages of 1.5-inch plastic balls provide a contacting sur-
face between the lime slurry and the flue gas. Following
each absorber module, the flue gas passes through a fiber-
glass entrainment separator and two horizontal banks of
chevron-type mist eliminators. The bottom of the trap-out
tray is washed intermittently and the lower mist eliminator
is washed continually with recycled pond water. The flue
gas from the parallel absorber trains then enters the 800-
ft, Ceilcote-lined stack. Following the boiler I.D. fan
is a bypass breeching around the entire scrubber loop.
Each module can be bypassed independently. No stack gas
reheat is currently being employed, although it is possible
that reheat will be added on at some future date.
Dravo thiosorbic lime from Maysville, Kentucky, is
utilized in the UOP scrubber modules at a stoichiometric
ratio of 1.1. The calcined, pelletized lime has a nominal
particle diameter of 1.75 inch, an MgO content of 3 to 8
percent and a CaO content of 90 to 95 percent. The lime
slaker discharges the 20 percent solids slurry into an
acitated lime slurry sump, where it is retained for a 5
minutes before being transferred to the lime slurry storage
tank, which handles the surge requirements of the absorption
A-16
-------
system. The transfer of slurry from the storage tank to the
TCA recycle tanks is accomplished by variable speed pumps,
which respond to changing SO2 concentrations and boiler load
conditions via a pH monitor. The scrubbing liquor contains
about 7 to 12 percent solids and is recirculated by four
pumps (one standby), each rated at 12,000 gpm. Scrubber
outlet pH is 5.8, and pH in the recycle tank is approxi-
mat?~Ly 6.8.
A bleed stream of spent reaction products is contin-
uously withdrawn from the recycle tank and pumped to the
thickener system. The thickener is 100 ft in diameter and
14 ft deep in the center. Here the reaction product slurry
is concentrated to an underflow composition of approximately
40 percent solids. This "underflow is cycled to the IU
Conversion Systems, Inc. (IUCS) fixation facilities where it
is further thickened, vacuum-filtered, and mixed with a
blend of dry fly ash and lime to form a 73 percent solid
substance (IUCS Poz-o-tec). The product is currently being
discharged to a 3500-acre-foot diked pond.
The wastewater pond receives ash sluice water, cooling
water blowdown, and water from the sludge treatment plant.
This system is not operating closed loop at the present
time. Design information is presented in Table A-4.
PERFORMANCE HISTORY
The unit was originally scheduled to startup in December
1976. Prior to start-up, module 5A was destroyed by a fire
in the neoprene lining. This module is being replaced and
start-up is now scheduled for November 1977. The other
module, No. 5B, began operation in January 1977. This was a
very cold month and initial start-up problems were com-
pounded by the severe weather. Frozen equipment and pipe-
A-17
-------
Table A-4. FGD SYSTEM DATA, CONESVILLE NO. 5
Unit location Conesville, Ohio
Unit rating 400 MW
Fuel characteristics Coal, 4.5 to 4.9 percent
sulfur
FGD vendor Universal Oil Products
Process Lime scrubbing
New or retrofit New
Start-up date February 1977
Efficiency, particulates 99.6 percent
Efficiency, SO- 89.7 percent
Sludge disposal Stabilized sludge is dis-
posed of in unlined
pond
lines were common. In the following months the facility
continued to experience difficulties, including ruptured
plastic piping, SO,, analyzer failure, poor velocity distri-
bution through the absorber, and carry-over of scrubber
liquor into the mist eliminator. During shutdown in August,
UOP replaced the pressurized plastic piping with stainless
steel. Problems have included corrosion in the presatura-
tor, failure of the rubber liner, and poor quality of the
lime supply. The lime supplier is installing mechanical
separators and metal detectors to improve the lime quality.
The utility is considering retrofitting the unit with a
reheat system.
No S0_ removal efficiency data are available at this
time.
A-18
-------
DUQUESNE LIGHT CO.
ELRAMA POWER STATION
BACKGROUND INFORMATION
The Elrama Power Station is owned and operated by the
Duquesne Light Company. The station's four pulverized-coal-
fired boilers have a generating capacity of 510 MW. De-
signed by Babcock and Wilcox, the boilers were placed into
service in 1952, 1953, 1954, and 1960. Additional plant
Q
details can be found in Table A-5.
POLLUTION CONTROL
The scrubber facility at the 510-MW Elrama Station is
almost identical to the system at the Phillips Station,
also owned by Duquesne Light. Five Chemico single-stage
venturi scrubbers have been installed after the mechanical
collectors and electrostatic precipitators.
It was planned that the knowledge gained from the test
program at the Phillips Station would be applied to Elrama
9
to enable compliance with emission regulations. The first
Elrama scrubber was placed in service on October 26, 1975.
It had been scheduled for an earlier start-up date, but
because of the severity and number of problems encountered
at Phillips, start-up was delayed until many of the problems
at Phillips were resolved and the modifications could be
incorporated at both stations.
Boiler No. 2 was initially connected to the FGD system
on October 26, 1975. This boiler has an equivalent capacity
See description later in this Appendix.
A-19
-------
Table A-5. FGD SYSTEM DATA, ELRAMA POWER STATION
Unit location
Unit rating
Fuel characteristics
FGD vendor
Process
New or retrofit
Start-up date
FGD status
Efficiency, particulates
Efficiency, S02
Sludge disposal
Elrama, Pennsylvania
510 MW
Coal, 1.0 to 2.8 percent
sulfur
Chemico
Lime scrubbing
Retrofit
October 1975
Operational
99 percent, design
83 percent design - 2
percent sulfur coal
Stabilized sludge is dis-
posed of in unlined
pond
of approximately 100 MW, and the emissions are handled by
one scrubber. To ensure reliability in the case of a scrub-
ber malfunction, two scrubbers are operated, each at partial
load.
The boiler operated continuously on the scrubber system
*
through January 1976. Two minor outages occurred during
this period, one caused by inoperative throat dampers and
the other by failure of a lime feeder belt.
Until additional construction is completed, two boilers
are the maximum that can be served by the scrubber system.
Further operability data have not been reported due to
recent litigation.
A-20
-------
To date the utility has encountered the usual minor start-up
problems in addition to some major problems. The major
problems entailed frozen pipes and hardware and design
problems associated with recirculation of the sludge within
the thickeners to attain 30 to 40 percent solids concentra-
tion.
The single-stage scrubber trains were not designed for
SO- removal, but through the use of hydrated calcium lime a
50 percent removal rate is being obtained in the interim
period. Tests have shown that single-stage scrubbers can
obtain at least 80 percent SO., removal by using a modified
lime containing 6 to 10 percent MgO, which will eventually
be used when all scrubber lime and sludge handling equipment
is completed in early 1978.
A-21
-------
ARIZONA PUBLIC SERVICE
FOUR CORNERS NO. 5
BACKGROUND INFORMATION
The Four Corners Power Station, located near Farmington,
New Mexico, is jointly owned by the Arizona Public Service
Company and the Southern California Edison Company. This
station contains five coal-fired boilers with an equivalent
electric generating capacity of 2233 MW.
Unit 5 is a coal-fired boiler with a net generating
capacity of 800 MW. This unit was placed in commercial
operation in July 1970. Approximately 20 percent of the
flue gas generated by this boiler was scrubbed during a test
program, to establish what sulfur dioxide and particulate
removal efficiencies could be obtained.
POLLUTION CONTROL
The FGD system, which was retrofitted on Unit 5 in
February 1976, is a Weir horizontal cross flow spray scrub-
ber that was formerly installed and operated in a test
program at the Mohave Generating Station of the Southern
California Edison Company. Following completion of tests at
Mohave, the module was dismantled, transported, and reas-
sembled at the Four Corners Generating Station for further
testing.
The horizontal module is a four-stage scrubbing chamber
48 feet long, 28 feet wide, and 15 feet high. The module
A-22
-------
itself is a hollow chamber consisting of four stages of 36
spray nozzles each. The slurry is cycled through the
scrubber in a countercurrent manner at a rate of 36,000
gpm/stage. Fresh lime slurry is sprayed across the flue gas
at the fourth stage, which is the discharge end of the
scrubbing chamber. The scrubbing solution is then collected
and recycled successively to the third, second, and first
stages of the scrubbing module. This allows for more com-
plete lime utilization and insures that fresh lime slurry
contacts gas with the lowest sulfur dioxide concentration.
The liquid recirculation rate can be adjusted over a wide
range. The module is designed to treat a maximum flue gas
capacity of 450,000 scfm (170 MW). The design gas velocity
through the module is 21.6 ft/sec. The cleaned gases then
pass through a demister and are reheated before being dis-
charged to the stack. The spent slurry is discharged from
the scrubbing system to a thickener and the underflow is
pumped to an on-site settling pond. Water is returned to
the process for further use.
The results of the test program, which was completed in
December 1976, have not yet been released.
A-23
-------
KENTUCKY UTILITIES
GREEN RIVER BOILERS 1, 2 AND 3
BACKGROUND INFORMATION
The Green River Station of Kentucky Utilities is
located on the Green River in central Kentucky, approxi-
mately 5 miles north of Central City. The local terrain
surrounding the power plant is sparsely populated and
heavily wooded.
The plant itself contains four steam turbine generating
units with a total gross generating capacity of 242 MW.
Boilers 1, 2, and 3 supply steam to generate 64 MW. These
boilers are used for peak loads and normally operate on a 5-
day week, with one or more of the boilers often at reduced
load. All three boilers are dry-bottom, pulverized-coal-
fired units manufactured by Babcock and Wilcox and installed
in 1949 and 1950. There are presently no plans to retire
these units. A coal with a 3.8 percent sulfur content is
burned.
POLLUTION CONTROL
Boilers 1, 2, and 3 are fitted with mechanical col-
lectors upstream of the FGD system. Design particulate
removal efficiency is 85 percent. The FGD system was de-
signed and installed by AAF and consists of a venturi scrub-
ber and a single mobile-bed contactor module to handle a
maximum flue gas capacity of 360,000 acfm at 300°F. Table
A-6 summarizes the FGD system data.
A-24
-------
Table A-6. FGD SYSTEM DATA,
GREEN RIVER NOS. 1, 2 AND 3
Unit location Central City, Kentucky
Unit rating 64 MW
Fuel characteristics Coal, 3.8 percent sulfur
FGD vendor American Air Filter
Process Lime scrubbing
New or retrofit Retrofit
Start-up date 9/75
Efficiency, particulates 99.7 percent (design)
(overall)
Efficiency, SO2 80 percent (guaranteed)
Water makeup 1.20 gpm/MW
Sludge disposal Sludge is disposed of in
unlined pond
The flue gases from each boiler pass through a series
of mechanical collectors where primary particulate removal
takes place. The flue gas is then drawn from the existing
breeching through a guillotine-type isolation damper and
associated ductwork to the scrubber fan. The guillotine-
type isolation dampers allow selective bypassing of the flue
gas around the scrubbing system to the existing stack. The
single-stage mobile-bed contactor consists of plastic balls
through which the flue gas passes countercurrent to the lime
slurry flow.
A-25
-------
The scrubbing slurry feed and recycle system consists
of a partitioned concrete reactant tank that includes
recycle pumps to supply the scrubber and absorber module, a
lime slurry slaking and feed system, and a system for dis-
charging scrubbing waste to a settling pond. Water from the
settling pond is recycled back to the scrubber.
PERFORMANCE HISTORY
This system has demonstrated good operability, as
shown in Figure 3-23. The problems encountered have been
primarily of a general mechanical nature such as rubber
peeling of the impellers. These problems were solved with-
out significant scrubber down time.
Some chemistry-related problems also occurred. Specif-
ically, a hard coat of scale developed in the lower sections
of the absorber internals. This was believed to have
occurred because of a high scrubbing solution pH combined
with high oxygen concentrations in the flue gas. This
problem was corrected by manually removing the scale, lower-
ing the oxygen content of the flue gas by minimizing air
leakage, and maintaining better pH control by modifying and
relocating the pH sensors in the absorber. Following the
completion of these modifications, scale formation has not
been a problem although some scale film has been noted peri-
odically, and has tended to disappear with continued opera-
tion .
In February 1977, the Carboline-lined stack was dis-
covered to be badly deteriorated. The lining had failed in
many areas, and the steel stack had corroded completely
through in some areas.
The stack was repaired first by welding a backup metal
plate to the portions of the stack where pitting occurred.
A-2 6
-------
Then the entire stack was lined with refractory coating
called Precrete G-8 by American Air Filter (AAF). The
utility reports that the lining failure was due to poor
application. Stack repairs were completed on March 1, 1977.
Kentucky Utilities is now planning to install reheat on the
system to prevent this problem from recurring. This stack
liner problem resulted in the first delayed outage caused by
the scrubber system.
A-27
-------
SOUTHERN CALIFORNIA EDISON
MOHAVE
BACKGROUND INFORMATION
Two prototype sulfur dioxide absorber modules were
installed in 1973 at the Mohave Generating Station of the
Southern California Edison Company. A vertical module,
rated at 170 MW, was installed to treat a 450,000-scfm
portion of the flue gas from Unit 1. A horizontal module,
also rated at 170 MW, was installed to treat a similar flue
gas portion from Unit 2. Units 1 and 2 are identical boilers,
Each has a maximum net continuous generating capacity of 790
MW, and each burns a western, low-sulfur coal with a heating
value of 11,500 Btu/lb and ash and sulfur contents of 10 and
0.5 to 0.8 percent, respectively.
POLLUTION CONTROL
The vertical module, a Universal Oil Products turbulent
contact absorber (TCA), was tested in two modes: first, as
a TCA unit, from November 2, 1974, to April 30, 1975, and
second, as a polygrid-packed absorber (PPA), from April 30,
1975 to July 2, 1975.
Start-up of the vertical module was initiated on sched-
ule, January 1, 1974. On January 24, 1974, a fire burned
most of the chlorobutyl lining, and the system was shut down
until October 1, 1974. Testing of this module started a
month later and was completed after 3131 hours of operation.
System design data are presented in Table A-7. The module
A-2 8
-------
is presently shut down in a cold stand-by condition at the
Mohave Generating Station. Refer to Section 3.1 of this
report for detailed test data.
Table A-7. FGD SYSTEM DATA, MOHAVE NO. 1A
Unit location Laughlin, Nevada
Unit rating 170 MW
Fuel characteristics Coal, 0.5 to 0.8 percent
sulfur
FGD vendor Universal Oil Products
Process Lime or limestone scrubbing
New or retrofit Retrofit
Start-up date November 1974
Efficiency, particulates 93 percent
S0? removal measured 95 percent
Water makeup 0.92 gpm/MW
Horizontal Module
The horizontal module, a four-stage horizontal cross-
flow spray scrubber, was operated by the utility in a short
series of start-up tests that ended on January 16, 1974,
when a formal test program was initiated to assess the
performance and reliability characteristics of the scrubbing
system. The test program was terminated on February 9,
1975, following 5927 hours of operation. System design data
are shown in Table A-8. The module was subsequently dis-
A-29
-------
mantled, transported, and reassembled at the Four Corners
Plant of Arizona Public Service, where additional testing
has been completed, but was not made available.
Test data are presented in Section 3.1 of this report,
Table A-8. FGD SYSTEM DATA, MOHAVE NO. 2A
Unit location
Unit rating
Fuel characteristics
FGD vendor
Process
New or retrofit
Start-up date
Efficiency, particulates
SC>2 removal measured
Water makeup
Sludge disposal
Laughlin, Nevada
170 MW
Coal, 0.5 to 0.8 percent
sulfur
Southern California
Edison/Stearns Roger
Lime or limestone scrubbing
Retrofit
November 1973
93 percent
95 percent
1.0 gpm/MW
Stabilized sludge is dis-
posed of in unlined
pond
A-30
-------
LOUISVILLE GAS AND ELECTRIC CO.
PADDY'S RUN NO. 6
BACKGROUND INFORMATION
The Paddy's Run Power Station of Louisville Gas and
Electric Company is located on the Ohio River in Rubbertown,
about 10 miles southwest of the center of Louisville,
Kentucky- Of the six units at Paddy's Run, only the boiler
on Unit 6 is retrofitted with an FGD system.
This boiler is a dry-bottom, pulverized-coal-fired unit
designed and installed by Foster-Wheeler in 1951. It
operates as a peaking unit and has a design rating equiv-
alent to 65 MW. The station as a whole operated at only an
approximate 5 percent load factor in 1974. Because of this
low load factor, the other boilers are not being fitted with
an FGD system. Data on plant operation and emissions
appear in Table A-9.
The coal now being burned has an average heating value
(as received) of 12,400 Btu/lb. Ash and sulfur contents are
14 and 3.5 to 4 percent, respectively.
POLLUTION CONTROL
A Research-Cottrell electrostatic precipitator (ESP) with
an efficiency of 99 percent provides primary control of
particulate emissions. Particulate loading at the outlet of
the ESP unit is approximately 0.05 gr/scf.
A-31
-------
Table A-9. FGD SYSTEM DATA, PADDY'S RUN NO. 6
Unit location Louisville, Kentucky
Unit rating 65 MW
Fuel characteristics Coal, 3.5 to 4.0 percent
sulfur
FGD vendor Combustion Engineering
Process Lime scrubbing
New or retrofit Retrofit
Start-up date April 1973
Efficiency, particulates 99 percent (design)
Efficiency, SO2 80 percent (design)
Water makeup Open-loop, 0.7 gpm/MW
Sludge disposal Unstabilized sludge is dis-
posed of in unlined pond
Atmospheric emissions of sulfur dioxide are limited to
1.2 Ib/MM Btu of heat input, and continuous monitoring
equipment shows that S02 emissions are within this limit.
The FGD system was designed by Combustion Engineering,
Inc. Start-up for the FGD plant occurred in April 1973.
The lime scrubbing system utilizes calcium hydroxide sludge,
which is generated as a waste by-product at a nearby acet-
ylene manufacturing plant. In this FGD process a slurried
mixture of calcium hydroxide and calcium sulfite in water
constitutes the scrubbing liquor.
The FGD system consists of two identical modules,
each sized to handle 175,000 acfm of flue gas at 350°F.
A-3 2
-------
Flue gas enters each scrubber module near the base and
contacts nonatomizing sprays which provide a constant supply
of slurry to the underside of the two stages of marble beds.
After emerging from the second marble bed, the clean flue
gas passes through a two-stage chevron mist eliminator,
where entrained water droplets are agglomerated and removed.
The flue gas then passes through a gas-fired reheater, a
booster fan, and out the stack.
Scale formation does not occur at this plant as long as
both of the following conditions are met: 1) fly ash must
not constitute more than about 6 percent of the slurry
solids, and 2) slurry pH in the reaction tank must be main-
tained between 8.0 and 9.5. Slurry pH is measured down-
stream from the reaction tank. Additive lime is pumped to
the reaction tank from the additive slurry tank. Scaling
potential does not seem to be influenced by the oxygen
content of the exhaust gas, which typically ranges between 6
and 9 percent.
The solids content of the slurry leaving the absorber
ranges between 9.5 and 10.5 percent. The ratio of sulfite
to sulfate in the recirculating slurry ranges between 40 and
50:1 on a weight basis. Liquid-to-gas ratio (L/G) ranges
between 15 and 18 gal/1000 ft3 of gas at 125°F per stage.
Fresh slurry additive contains 20 to 30 percent solids and
has a pH of 12.6. Lime is added to the thickener tank to
stabilize the sludge that is formed at a rate of about 100
pounds of lime per ton of dry sludge. The sludge is trucked
to a 10-acre landfill that ranges from 20 to 30 feet in
depth. Solids content of the effluent from the thickener
averages about 25 percent. This material is then dewatered
by vacuum filtration to form a stabilized sludge containing
A-33
-------
45 percent solids. Alternately, the thickener effluent can
be mixed with dry fly ash and lime to form stabilized sludge
with a 45 percent solids content.
No major start-up problems were encountered that were
associated with the scrubbers. Problems with other system
components included the following:
Thickener Size - The original thickener was too small
to handle the untreated slurry at full load. Preliminary
data had indicated that the slurry material entering the
thickener would contain more calcium sulfate than was
actually encountered. The slurry actually contained a high
ratio of calcium sulfite to calcium sulfate.
Carbide Lime Feed System - Some early problems were
experienced with plugging of the mesh strainer on the
additive tank. This problem was solved by installing a
Rietz mechanical disintegrator in the carbide feed line to
the additive tank.
Mist Eliminator Wash System - The original mist elim-
inator wash system consisted of rotating nonretractable
lances with 3/16-inch diameter nozzles every 6 inches. The
system would not efficiently clean the upper mist elimina-
tor. The nonretractable lances were replaced with retract-
able half-track lances with oscillating 1/2-inch nozzles.
The larger solid stream of water from these nozzles keeps
the mist eliminators clean.
Damper Leakage - The FGD system can be bypassed by
means of louvered dampers. Deposits on these dampers were
found to prevent them from closing completely, so leakage
through the dampers resulted in the formation of a visible
plume. By cleaning the dampers periodically the leakage was
i p
minimized and plume formation effectively suppressed.
A-34
-------
PERFORMANCE
In July 1977, FGD efficiency tests were run with the
addition of MgO. Results from those two 8-hour test runs
are as follows:
SC>2 concentration, ppm
Date Inlet Outlet Removal, %
July 8, 1977 2230 1 to 5 99.7-99.9
July 16, 1977 2150 1 to 5 99.7-99.9
Additional test data are presented in Section 3.1 of
this report.
Availability figures approaching 100 percent have been
achieved during recent operation, but the power plant, which
is primarily a summer peaking unit, is operated infrequently.
The capacity factor for Boiler 6 in 1976 was about 14.6
percent. The FGD unit was operated only 7 months in 1976.
A-35
-------
DUQUESNE LIGHT CO.
PHILLIPS POWER STATION NO. 6
BACKGROUND INFORMATION
The Phillips Power Station of the Duquesne Light
Company is located on the Ohio River in Allegheny County,
Pennsylvania, 20 miles northwest of Pittsburgh, Pennsyl-
vania. The area is heavily industrialized, and the plant
supplies electricity to residential, commercial, and indus-
trial users in Allegheny and Beaver Counties.
Six generating units at this station constitute a gross
continuous generating capacity of 408 MW. The net station
capacity is 373 MW when all four particulate scrubbers and
the single SO,, absorber are operating. All the boilers are
dry-bottom, pulverized-coal-fired units manufactured by
Foster-Wheeler. The first unit was installed in 1942 and
the sixth one in 1956. Coal at this station has an average
as-received gross heating value of 11,350 Btu/lb. Ash
content, on a dry basis, is 18.2 percent, and sulfur content
14
is 2.15 percent. All the generators are cycling base load
units. Unit 6 is the largest boiler, with a net capacity of
143 MW.
POLLUTION CONTROL
Primary particulate emission control at this plant is
provided by Research-Cottrell mechanical collectors followed
by a Research-Cottrell electrostatic precipitator on each
boiler. The collection efficiency for this system was
A-36
-------
originally estimated to be 90 percent; however, the addition
of the FGD system caused disruption of the gas flow dis-
tribution through the ESP, and the combined efficiency
through the mechanical and electrostatic system now appears
to be only about 80 percent.
Final particulate control is achieved by four parallel
Chemico venturi scrubbers with a design efficiency of 99
percent. These scrubbers also remove about 50 percent of
the S02 when lime is added to the water. FGD data are
presented in Table A-10.
Table A-10. FGD SYSTEM DATA, PHILLIPS POWER STATION NO. 6
Unit location
Unit rating
Fuel characteristics
FGD vendor
Process
New or retrofit
Start-up date
Efficiency, particulates
Efficiency, S02
Sludge disposal
South Height, Pennsylvania
143 MW
Coal, 1.0 to 2.8 percent
sulfur
Chemico
Lime scrubbing
Retrofit
July 1973
99 percent (design)
83 percent (design, 2
percent sulfur coal)
Stabilized sludge is dis-
posed of in unlined
pond
Emissions from one of the particulate scrubbers are
further controlled by a Chemico second-stage venturi ab-
sorber utilizing a lime solution. Sulfur dioxide removal
A-37
-------
efficiency for this two-stage train has averaged about 90
percent, based on performance guarantee tests. Sulfur
dioxide removal efficiency of the single-stage scrubber is
50 to 60 percent when the pH is maintained between 6 and 7-
By December 1977, through the use of a single-stage scrubber
and the firing of coal with 2 percent sulfur content, the
unit should be able to achieve 83 percent S02 removal when
thiosorbic lime (6 to 12% Mg) is used as the absorbent.
PERFORMANCE
The Phillips Station was the first lime FGD system in
the United States. The Phillips scrubber system was started
up in July 1973. Several problems developed in the form of
erosion of fans and the lime feed system, causing outages of
the scrubber system. After an extended outage, the scrubber
system was returned to service in March 1974. Scale de-
posits were removed, corroded parts were replaced, and "off-
spec" components were replaced with the originally specified
material. Not all of the operating problems have been
solved, and system availability has been lower than that of
many other FGD systems. Scale continues to be a problem,
and the utility is planning to use lime with magnesium oxide
added in an attempt to reduce the problem. Sometimes the FGD
system has been bypassed so that outages would not reduce
boiler output. Problem areas are discussed below:
Recycle Pumps - The original Carpenter 20 alloy impellers
and casings used in the Ingersoll-Rand scrubber pumps have
failed repeatedly.
Approximately 30 percent of the metal on both impellers
and casings is being eroded/corroded, thus limiting pump
service life to about 1000 to 3000 hours. An extensive pump
A-38
-------
reliability study is under way. Several materials of con-
struction are being tried, including rubber and hard iron.
Throat Dampers - Throat dampers have been a major problem
because of fly ash and hard scale deposits. The deposits
caused an increase in pressure drop across the throat from 6
inches to 12 inches or more. This increase in pressure drop
reduced the amount of gas that could be scrubbed. In an
attempt to clear the deposits, Duquesne Light maintained a
continuous flow of water through the 3/4-inch openings of
the damper housing. When this effort failed, air was forced
through an open vent on the sides of the damper housings.
The problem remains unsolved. Significant accumulations are
still being found in the throat damper. Manual cleaning
requires removal of the train from service and takes about
288 man-hours per scrubber.
A total of 60 throat damper arms (12 per vessel) were
specified as 316L SS, but were apparently fabricated from
304 SS. Thirty-two of these throat damper arms corroded
badly. The top and bottom throat damper scraper blades were
also found to be 304 SS; they were essentially destroyed by
corrosion. All of the damaged throat damper arms and damper
scraper blades were replaced with 316L SS components.
Wet Induced Draft Fans - Several problems have occurred in
operation of the induced draft fans. The fan shaft shrouds,
the 316 SS-clad fan dampers, and the stiffener bars have all
corroded. The fan spray sump pumps have also been wearing
out much faster than anticipated, partly because of a
chloride concentration of about 3500 to 4500 ppm and a com-
paratively low pH in the scrubber liquor. Several steps
were taken to alleviate the problems: the fan spray pumps
were redesigned; the affected fan shaft shroud parts were
A-39
-------
coated with 1/4-inch neoprene rubber; addition of lime to
the system was increased to achieve optimum pH; double
(reinforcing) plates were installed on each of the blades to
reduce the fan stresses causing metal deformation; and
deposits were manually removed from the fans. Despite all
of these steps, significant pitting is still taking place on
the back of the fan blades, primarily because of fly ash/
sludge deposits.
Lime Slurry Preparation - The lime slaker at the Phillips
plant was not designed for use with all six boilers con-
nected to the system. Discontinuous operation of the slaker
system has resulted in deviations from proper pH and has
caused deposition of solids.
A-40
-------
U.S. AIR FORCE
RICKENBACKER AIR FORCE BASE
BACKGROUND INFORMATION
The FGD system at the Rickenbacker Air Force Base
located near Columbus, Ohio, is a Research-Cottrell/Bahco
System which was started up in March 1976.
This first U.S. Bahco installation handles the flue gas
from seven stoker-coal-fired boilers. Three boilers were
first placed in service in 1952, and the remaining four in
1954. One is rated at 31 million Btu/hr and the other six
at 60 million Btu/hr each. The units fire a high-sulfur
(3.6 percent) coal at high combustion air rates (150 to 160
percent excess air). The plant burns approximately 40,000
tons of coal per year. Plant and FGD data are presented in
Table A-ll. As shown in the table, the FGD system is
designed to remove between 82 and 98 percent of the SO-.
POLLUTION CONTROL
The Rickenbacker emission control system consists of a
mechanical collector, Bahco scrubber tower, lime storage and
handling system, clarifier, booster fan and associated
ductwork, pumps, controls, and a sludge disposal pond.
Particulate and SO -laden gas taken from each of the
ri£
existing seven stacks is fed through a common header and
into a mechanical collector, where primary particulate
removal takes place. The mechanical collector has a design
removal efficiency of 70 percent and was installed primarily
A-41
-------
to reduce fan wear. The partly cleaned gas then flows to
the fan and into the Bahco tower, where additional particu-
late and primary S02 removal takes place.
Table A-11. FGD SYSTEM DATA, RICKENBACKER AIR FORCE BASE
Unit location
Unit rating
Fuel characteristics
FGD vendor
Process
New or retrofit
Efficiency, particulate
Efficiency, SO2
Sludge disposal
Columbus , Ohio
20 MW
Coal, 3.6 percent sulfur
Research-Cottrell
Research-Cottrell/Bahco
Lime Scrubbing
Retrofit
98 percent
90 percent (82 to 98% de-
pending on stoichiometric
lime feed rate)
Unstabilized
sludge to lined pond
The Bahco scrubbing vessel is a vertical tower consist-
ing of two converted venturi stages. Gas flows through the
tower, where it is contacted with a lime slurry solution.
After passing through the first scrubbing stage, entrained
moisture is removed by a cyclonic-type mist eliminator. The
gas then passes through a second scrubbing stage and a
cyclonic-type mist eliminator. The cleaned gas is dis-
charged to the atmosphere through a stack mounted directly
on the tower.
Pebble lime is used as the reagent in the scrubbing
system and is prepared by slaking and dissolution in a mix/
A-4 2
-------
hold tank. The scrubbing slurry is then pumped to the
absorber, where it is contacted with the flue gas in a
countercurrent fashion.
The spent scrubbing solution is discharged to the
thickener, where the waste solids settle out. Thickener
overflow is returned to the mix/hold tank. Thickener
underflow is discharged to a 5-acre Hypalon-lined disposal
pond. This pond is located approximately 400 feet from the
scrubbing tower. Clear water from the pond is returned to
process as slaking water.
Research-Cottrell has recently released data that
indicates the SO_ removal efficiency can approach 100 per-
18
cent if the stoichiometric ratio exceeds 1.0.
A-4 3
-------
TENNESSEE VALLEY AUTHORITY
SHAWNEE NO. IDA AND 10B
BACKGROUND INFORMATION
The Shawnee Steam Plant is located on the Ohio River
about 10 miles northwest of Paducah, Kentucky. The plant
has 10 separate pulverized-coal-fired boilers, each con-
nected to its own stack and generating turbine. The boilers,
which were manufactured by Babcock & Wilcox, are identical
in design, each rated at 175 MW capacity. The first unit
went into commercial operation in April 1953. By June 1957,
all 10 units were generating power.
POLLUTION CONTROL
Particulate emissions from each boiler are controlled
by a mechanical collector followed by an electrostatic
precipitator. An experimental FGD prototype plant, jointly
sponsored by EPA and TVA, is being operated on a portion of
the flue gases from Unit 10. Data from this system are
presented in Tables A-12 and A-13.
The two parallel, retrofitted FGD systems began opera-
tion in April 1972. Universal Oil Products manufactured one
of the FGD systems, a turbulent contact absorber. Chemical
Construction Company (Chemico) produced the other FGD
system, which consists of a venturi scrubber followed by a
spray tower. Each FGD system is capable of treating 30,000
acfm, the equivalent of approximately 10 MW of power plant
generating capacity. A third (experimental) FGD system,
A-4 4
-------
manufactured by Combustion Engineering, Inc., was operated
for a time at Shawnee, but testing of this system has been
concluded.
In June 1974, the EPA began a 3-year advanced test
program at the Shawnee facility. The major goals are as
follows: (1) continuation of long-term testing, with
emphasis placed upon the reliable operation of mist elimina-
tion systems; (2) investigation of advanced process and
equipment design variations for improving system reliability
and process economics; and (3) long-term reliability testing
on promising process and equipment design variations. The
testing program is scheduled to run through June 1978.
Because of the experimental nature of the FGD systems, no
19
system operability data are published. The system has
operated well, however, and yielded useful data. Operating
data and test results are presented in Section 3.1.
Many of the operational problems of FGD systems have
been solved at this facility. Some of the most important
findings are discussed below. Mist eliminator scaling was
controlled by using a top wash with fresh water every 4 to 8
hours when alkali utilization was in excess of 85 percent.
The benefits and techniques of forced oxidation were also
studied, because they are related to alkali utilization.
The effects of magnesium liquid-to-gas ratio, chloride, pH,
and stoichiometry on S02 removal were studies. Turndown
capabilities were demonstrated for the venturi/spray sys-
20
tern.
A-45
-------
Table A-12. FGD SYSTEM DATA SHAWNEE NO. IDA
Unit location
Unit rating
Fuel characteristics
FGD vendor
Process
New or retrofit
Start-up date
Removal Efficiency, S0_
Paducah, Kentucky
10 MW
Coal, 2.9 percent sulfur
Universal oil products
Lime/limestone scrubbing
Retrofit
April 1972
60 to 98 percent
Table A-13. FGD SYSTEM DATA SHAWNEE NO. 10B
Unit location
Unit rating
Fuel characteristics
FGD vendor
Process
New or retrofit
Start-up date
Removal Efficiency, SO,.,
Paducah, Kentucky
10 MW
Coal, 2.9 percent sulfur
Chemico
Lime/limestone scrubbing
Retrofit
April 1972
60 to 90 percent
A-46
-------
REFERENCES FOR APPENDIX A
1. PEDCo Environmental, Inc. Sumary of the Status of Flue
Gas Desulfurization Systems in the U.S. Pennsylvania
Department of Environmental Resources. January 21,
1977, and July 1977. p. 5-59.
2. PEDCo Environmenta, Inc. Summary Report - Flue Gas
Desulfurization Systems June-July 1977. Prepared for
Division of Stationary Source Enforcement and Indus-
trial Environmental Research Laboratory. U.S. En-
vironmental Protection Agency, Research Triangle Park,
North Carolina. p. 197.
3. Durkee, K., Trip Report. U.S. Environmental Protection
Agency- Office of Air Quality Planning Standards
Research Triangle Park, North Carolina. July 14, 1977.
4. Op. Cit. No. 1. p. 5-43.
5. Letter from Mr. R. VanNess to Mr. D. Goodwin, EPA, July
20, 1977.
6. Grimm, C., J. Z. Abrams, I.A. Rabin, and C. Lamantia.
Particulate and S02 Removal at the Colstrip Station of
the Montana Power Co. The Second PACHEC Conference.
Denver, Colorado. August 1977.
7. Op. Cit. No. 2. p. 56-58.
8. Op. Cit. No. 1. p. 5-31.
9. Op. Cit. No. 2. p. 87.
10. Op. Cit. No. 2. p. 278.
11. Op. Cit. No. 2. p. 133.
A-47
-------
12. PEDCo Environmental, Inc. Survey of the Flue Gas
Desulfurization System at the Paddy's Run Power Sta-
tion, Louisville Gas and Electric Company. Prepared
for Control Systems Laboratory. U.S. Environmental
Protection Agency- Research Triangle Park, North
Carolina. July 9, 1975.
13. Op. Cit. No. 2. p. 146.
14. PEDCo Environmental, Inc. Survey of the Flue Gas
Desulfurization System at the Phillips Power Station -
Duquesne Light Company. Prepared for Control Systems
Laboratories, U.S. Environmental Protection Agency -
Research Triangle Park, North Carolina. June 16, 1975.
15. PEDCo Environmental, Inc. Evaluation of the Flue Gas
Desulfurization System at the Phillips Power Station of
Duquesne Light Company. Prepared for U.S. Environ-
mental Protection Agency. Research Triangle Park,
North Carolina. October 1977.
16. Rasor, J.B. (Associate Base Engineer) A Lime Slurry
Scrubbing System in Operation. U.S. Air Force, Ricken-
backer Air Force Base, Ohio. Presented at 1977 AIPE
Convention. June 20-22. Atlanta, Georgia.
17, PEDCo Environmental, Inc. Survey Report on S02 Con-
trol Systems for Nonutility combustion and Process
Sources. Prepared for Industrial Environmental Re-
search Laboratory. U.S. Environmental Protection
Agency. Research Triangle Park, North Carolina. May
1977. p. 43.
18. Biedell, E.L. Particulate and Sulfur Dioxide Removal
System at Rickenbacker Air Force Base, Columbus, Ohio.
Research-Cottrell, Inc., Bound Brook, New Jersey.
Presented at The Second Pacific Chemical Engineering
Congress, Denver, Colorado. August 29-31, 1977.
19. Op. Cit. No. 2. p. 211.
20. Williams, J.E. Summary of Operation and Testing at the
Shawnee Prototype Lime/Limestone Test Facility, U.S.
EPA. April 1977.
A-48
-------
APPENDIX B
DOMESTIC LIMESTONE SLURRY FGD SCRUBBING SYSTEMS
B-l
-------
This appendix summarizes information on limestone
scrubbing FGD systems presented in Section 3.2 and also
describes other operating limestone scrubbing systems in
this country-
The following FGD systems are described:
0 Arizona Public Service - Cholla No. 1
0 Kansas City Power & Light - La Cygne No. 1
0 Kansas City Power & Light - Lawrence Nos. 4 and 5
0 Northern States Power Co.
0 Springfield Utilities
0 Widows Creek No. 8
0 South Carolina Public Service Co.
0 Will County No. 1
The EPA test facility at Shwnee uses both lime and lime-
stone. It is described in Sections 3.1 and 3.2 and in
Appendix A.
B-2
-------
ARIZONA PUBLIC SERVICE
CHOLLA NO. 1
BACKGROUND INFORMATION
The Cholla Power Generating Station is located near
Joseph City, in Navajo County, Arizona. At present Cholla
operates only Unit 1, a dry-bottom, pulverized-coal-fired
boiler with a net generating capacity of 115 MW. The
boiler was designed by Combustion Engineering, Inc. Two
250-MW boilers are under construction and two 350-MW boilers
are also being planned at the Cholla Power Station.
The coal now being burned has typical fuel values, as
received, of 10,290 Btu/lb, 10.4 percent ash and 0.4 to 1.0
percent sulfur. It is fired at a rate of 54 tons/hr at
capacity.
POLLUTION CONTROL
A Research-Cottrell multicyclone-type collector, oper-
ating with an efficiency of about 75 percent, provides
primary control of particulate emissions. Design particu-
late loading at the outlet of the multicyclones is approxi-
mately 2 gr/scf.
In 1973, the boiler was retrofitted with a flooded-disc
scrubber system for particulate and S02 control. Designed
by Research-Cottrell, Inc. (R-C), the wet limestone system
removes 99.7 percent of the particulate matter and 92
percent of the SO,,.* The flue gas desulfurization
system consists of two parallel scrubbing modules, each
*
Average SO2 removal effxciency for the SO2 and particulate
scrubber is 58.5 percent (92 percent for A-side and 25 per-
cent for B-side, 92+25/2 - 58.5).
B-3
-------
designed to accommodate 50 percent of the flue gas. Module
A of the system includes an adjustable flooded-disc scrub-
ber for particulate control, followed by a packed tower,
which utilizes a limestone slurry, for S02 removal. Module
B also incorporates a flooded-disc scrubber for particulate
control but does not currently utilize a second stage.
Sulfur dioxide removal efficiency for Module B is 25 percent,
whereas that for Module A is 92 percent. FGD data are
summarized in Table B-l.
Table B-l.
Unit location
Unit rating
Fuel characteristics
FGD SYSTEM DATA, CHOLLA NO. 1
Joseph City, Arizona
115 MW
FGD vendor
Process
New or retrofit
Start-up date
Efficiency, particulates
Efficiency, S02 (actual)
Water makeup
Sludge disposal
Coalt 0.44 to 1 percent
sulfur
Research-Cottrell
Limestone scrubbing
Retrofit
October 1973
99.7 percent
58.5 percent overall
92 percent in absorber
Open-loop, 1.04 gpm/MW
Unstabilized sludge is
disposed of in unlined
pond
PERFORMANCE HISTORY
Although the FGD system at Cholla has a good operating
history, operation has not been trouble-free. Reheater
B-4
-------
vibration was an early problem, caused by a difference in
the size of the main duct and the reheater transition duct.
Baffles were installed to eliminate this problem and to
prevent acid condensation on the reheater tubes, where
corrosion was occurring.
Solids buildup and line plugging occurred, particularly
when the system operated at low flow rates. To solve these
problems, the piping was modified to eliminate stagnant
pockets where solids could accumulate, and pumps were
flushed immediately after removal from service.
Overall operability has been inconsistent, with many
periods of 100 percent and a low period of 46 percent.
The overall S0? removal efficiency is 58.5 percent,
which is sufficient to meet the local regulations.
FUTURE PLANS
Arizona Public Service (APS) plans to install an FGD
system on Cholla Unit 2, which is presently under construc-
tion. APS signed a contract with R-C to purchase a partic-
ulate scrubber and an FGD system for that boiler similar to
the system on Unit 1. R-C is preparing preliminary engi-
neering designs and plans to use two modules to treat the
123
flue gases from that boiler.* ' '
Personal communication from Mr. Lyman K. Mundth, Arizona
Public Service Co. to Mr. Don R. Goodwin, U.S. Environ-
mental Protection Agency. July 20, 1977.
B-5
-------
KANSAS CITY POWER & LIGHT
LA CYGNE NO. 1
BACKGROUND INFORMATION
The La Cygne Power Station of Kansas City Power and
Light (KCPSL) is a new station located about 55 miles south
of Kansas City, in Linn County, Kansas.
The electric power generating facilities consist of one
Babcock & Wilcox 6,200,000 Ib steam/hr, coal-fired, base-
load boiler with associated 820-MW (net) steam turbine and
electric generator. The plant also has three oil-fired
boilers, used primarily for start-up of the large unit and
to supply steam to a separate 22-MW turbine generator.
The coal now being burned ranges in gross heating value
(as received) from 8200 to 10,200 Btu per pound. Ash and
sulfur range from 20 to 30 percent and from 5 to 6 percent,
respectively.
POLLUTION CONTROL
The FGD system consists of eight particulate and S0~
scrubbing modules, with on-site limestone grinding and
storage facilities. All flue gases are treated, and the
ductwork does not provide for the bypassing of flue gas
around the scrubbers. As the hot flue gas enters the
venturi, it is subjected to jets of limestone slurry in-
jected through nozzles on the walls of the vessel. The
liquid-gas stream flows downward through the venturi throat,
where the gas contacts the atomized liquid droplets. The
B-6
-------
scrubbing efficiency is regulated by adjusting the venturi
throat gap. As the gas exits from the venturi and enters
the disengagement chamber, its velocity decreases from about
130 ft/sec (at the throat) to about 15 ft/sec. This reduc-
tion in velocity separates the limestone slurry droplets
from the quenched gas. The slurry drains into the recir-
culation tank. The gas then enters the SO,., absorber tower
base and moves upward through two sieve trays in series. As
the gas passes through the 1.375-inch-diameter holes in the
sieve trays, it contacts limestone slurry that is sprayed
into the path of the rising gas. The scrubbed gas then
passes through a third sieve tray, which collects slurry
carryover and reduces the mist load on the demister. The
gas then passes through a 10-inch-high chevron demister,
where the remaining fine droplets coalesce and drip back
down through the gas stream into the recirculation tank.
The flue gas is then reheated from about 121° to 175°F.
Reheating is accomplished primarily by means of steam coils,
with additional heat provided by injecting hot air from the
boiler combustion air heater. Additional steam reheat
capacity is currently being installed to eliminate the need
for hot air bypass.
Additional FGD system data are presented in Table B-2.
PERFORMANCE HISTORY
The FGD system has encountered typical problems. Some
start-up problems, such as vibrations of the I.D. fans and
the sensitivity of these fans to imbalance, occurred even
before the boiler was fired and are not related to the FGD
system. As these fabrication problems were corrected, other
types of problems appeared. These included plugging of the
B-7
-------
4
Table B-2. FGD SYSTEM DATA, LA CYGNE NO. 1
Unit location La Cygne, Kansas
Unit rating 820 MW (net)
Fuel characteristics Coal, 5.0 percent sulfur
FGD vendor Babcock & Wilcox
Process Limestone scrubbing
New or retrofit New
Start-up date February 1973
Efficiency, particulates 98 percent
(actual)
Efficiency, S02 (actual) 80.14 percent
Water makeup Open-loop, 1.4 gpm/MW
Sludge disposal Unstabilized sludge is
disposed of in unlined
pond
demister and strainers, wearing of spray nozzles, and cor-
rosion of reheater tubes. Problems of this type are the
result of collected fly ash and flow characteristics of the
recirculated slurry. Several modifications have been made
to the original plant design in an effort to solve these
problems. These include hydroclones to separate fly ash,
increased reheat capacity, and installation of a spare
scrubber module.
Since there is no bypass, a considerable amount of
ef.'fort is expended toward maintaining the modules by nightly
cleaning of one module per night.
B-f
-------
This system has shown continued performance improvement
as shown in Figure 3-28. Availability for the last 18
months has been in excess of 90 percent.
The results of a stack sampling test conducted on May
5, 1975, with the unit continuously operating from 700 to
720 MW, indicated an average S02 removal efficiency of 80.14
percent, and removal of 98.2 percent of the particulates.
S09 removal efficiency is 80.14 percent, which is
A C £ *7
sufficient to meet local regulations. ' ' '
B-9
-------
KANSAS POWER & LIGHT COMPANY
LAWRENCE NO. 4 AND NO. 5
BACKGROUND INFORMATION
The Lawrence Power Station of Kansas Power and Light
Company is located in a lightly industrialized area on the
outskirts of Lawrence, Kansas. The plant operates two
Combustion Engineering steam boilers, which are equipped to
burn coal or natural gas supplemented with oil, or a com-
bination of these three fuels. Boiler 4, the older of the
two units, was first placed in service in 1959. It is
operated as a cyclic load boiler. The maximum electric
generating capacity of this unit varies with the type of
fuel being burned; output can be as high as 143 MW when
burning natural gas, whereas it decreases to 125 MW when
burning coal plus natural gas. The retrofitting of this
boiler with an FGD system in 1968 reduced the boiler capa-
city to 115 MW.
The second unit at the plant is Boiler 5. Its rated
capacity, when burning coal plus natural gas, is 400 MW.
The unit and the FGD system were placed in service in
November 1971. Boiler 5 is similar to Boiler 4, and it also
is classified as a cyclic load unit.
The company has now switched from high-sulfur Kansas
coal to Wyoming coal, which contains from 0.4 to 0.8 percent
sulfur and 10 percent ash. The coal has a gross heating
value of 10,000 Btu/lb. Tables B-3 and B-4 present a sum-
mary of the FGD systems.
B-10
-------
Table B-3. FGD SYSTEM DATA, LAWRENCE NO. 4
Unit location Lawrence, Kansas
Unit rating 125 MW
Fuel characteristics Coal, 0.4-0.8 percent sulfur
FGD vendor Combustion Engineering
Process Limestone scrubbing
New or retrofit Retrofit
Start-up date December 1968
Efficiency, particulates 99 percent
Efficiency, SO,, 75 percent
Sludge disposal Unstabilized sludge is
disposed of in unlined
pond
Table B-4. FGD SYSTEM DATA, LAWRENCE NO. 5
Unit location Lawrence, Kansas
Unit rating 400 MW
Fuel characteristics Coal, 0.4-0.8 percent
sulfur
FGD vendor Combustion Engineering
Process Limestone scrubbing
New or retrofit New
Start-up date November 1971
Efficiency, particulates 99 percent
Efficiency, SO2 65 percent
Sludge disposal Unstabilized sludge is
disposed of in unlined
pond
B-ll
-------
POLLUTION CONTROL
Unit 4 was originally a dry limestone injection system,
but it has been converted to a limestone flue gas scrubbing
system. Boiler 4 has two FGD modules, and Boiler 5 has
eight. All are identical in size, and each is designed to
handle approximately 150,000 scfm of flue gas. Originally,
each module consisted of a single stage of 3/4-inch glass
marbles. The bed was about 3 to 4 inches thick and was fitted
with overflow pots to collect the liquor drained from the
top of the bed. The scrubbing liquor was sprayed through
nozzles located below the bed. The utility is in the proc-
ess of replacing the marble bed absorbers with spray towers.
Boiler 4 was completed in 1977; Boiler 5 is scheduled for
completion in 1978.
Each of the two modules on Boiler 4 is connected
(through an I.D. fan) to a separate 120-ft stack, whereas the
gases from all eight modules on Boiler 5 are discharged
through a common 375-ft stack.
PERFORMANCE HISTORY
Availability figures have not been reported; however,
the availability is apparently adequate for this station,
since it has a low load factor and up to 50 percent of the
modules can be shut down for maintenance. No lost boiler
capacity has been reported on Unit 4 since January 1977,
8 9 TO
when the limestone scrubbing spray towers were started.
Measured SO- efficiency has not been reported.
B-12
-------
NORTHERN STATES POWER CO.
SHERBURNE COUNTY STATION NO. 1 AND NO. 2
BACKGROUND INFORMATION
The Sherburne County generating plant of the Northern
States Power Company is located adjacent to the Mississippi
River in Sherburne County, near Becker, Minnesota. Each of
the two units has a net capability of 700 MW. The boilers
are controlled-circulation, single-reheat, balanced-draft
units manufactured by Combustion Engineering. Two addi-
tional units are being planned to boost the total plant
output to 3000 MW by 1980.
The coal presently in use is subbituminous western
coal. It contains approximately 28 percent moisture, 9
percent ash, 0.8 percent sulfur and 8300 Btu/lb.
POLLUTION CONTROL
The scrubber system -utilizes wet limestone scrubbing
for both particulate and S0~ removal. Twelve scrubber
modules are provided for each unit with only 11 required for
full load operation. Design particulate and SO,, removal
efficiencies of these systems are 99 percent and 50 percent,
respectively, as shown in Tables B-5 and B-6.
The Combustion Engineering scrubber systems consist of
a rod-type venturi throat, a marble bed absorber, a two-stage
fiberglass Chevron demister, and a fine tube reheater.
B-13
-------
PERFORMANCE HISTORY
Problems encountered by Sherburne include some mechani-
cal problems, partly due to a last-minute decision to in-
clude a particulate scrubber. Mud deposits in and around
the venturi throat and erosion of the walls and rods in the
throat were a problem. These areas have been relined with
316L stainless steel. The components are still wearing, but
at a slower rate.
Other problems are loss of marbles with the sludge,
reheater leaks, failure of rubber lining in pipes, erosion
of pump internals, ineffectiveness of Zurn duplex strainer,
and limited capacity of limestone grinding mills. All of
these problems are being dealt with either by design modifi-
cation such as 316 stainless wear plates or by replacement
parts.
Availability for Unit 1 averaged 85 percent for the
first 4 months of operation after start-up. For the past 12
months, availability has been in excess of 90 percent. Unit
2 has shown even better start-up performance, with oper-
ability averaging about 95 percent for the first 4 months.
These data are shown in Figure 3-29.
Unit 1
1977 - The unit was down until June 6 for completion
of the first year's inspection. It came back
on line June 18, 1977.
FUTURE PLANS
Northern States Power and C-E are sponsoring a long-
term optimization program for the units.11'12'13
B-14
-------
Table B-5. FGD SYSTEM DATA,
Unit location
Unit rating
Fuel characteristics
FGD vendor
Process
New or retrofit
Start-up date
Efficiency, particulates
Efficiency, S0_
Water makeup
Sludge disposal
SHERBURNE COUNTY STATION NO. 1
Becker, Minnesota
710 MW
Coal, 0.8 percent sulfur
Combustion Engineering
Limestone scrubbing
New
March 1976
99 percent (guaranteed)
50 percent (guaranteed)
Open-loop, 1.13 gpm/MW
Unstabilized sludge is
disposed of in lined
pond
Table B-6. FGD SYSTEM DATA, SHERBURNE COUNTY STATION NO. 5
Unit location
Unit rating
Fuel characteristics
FGD vendor
Process
New or retrofit
Start-up date
Efficiency, particulates
Efficiency, SO2
Water makeup
Sludge disposal
Becker, Minnesota
680 MW
Coal, 0.8 percent sulfur
Combustion Engineering
Limestone scrubbing
New
April 1977
99 percent (guaranteed)
50 percent (guaranteed)
Open-loop, 1.13 gpm/MW
Unstabilized sludge is
disposed of in lined
pond
B-15
-------
SPRINGFIELD CITY UTILITIES
SOUTHWEST NO. 1
BACKGROUND INFORMATION
The Southwest Power Station of Springfield City Util-
ities is located in Springfield Missouri. The plant oper-
ates one coal-fired boiler (Unit 1) with a rating of 200 MW.
The boiler uses coal with a 3.5 percent sulfur content.
POLLUTION CONTROL
The two turbulent contact absorber modules at the South-
west Power Station are supplied by Universal Oil Products.
These limestone scrubbers are downstream of an ESP and are
designed for overall removal of 99.7 percent of the partic-
ulate emissions and 80 percent of the SO,, emissions. This
unit, which was put on-line in April 1977, has experienced
some start-up problems. Specifically, these have included
mist eliminator plugging, corrosion in the presaturator
chamber, corrosion of the inlet/outlet gas dampers, and
failure of the lining materials in both the scrubber and
stack. These problems are being solved by a number of modifi-
cations :
0 Modification of the scrubber outlet damper
0 Replacement of scrubber packing
0 Relining of the scrubber outlet ductwork
0 Insulating of the inlet dampers
B-16
-------
Sludge produced by the FGD system is deposited in an
unlined pond after stabilization. FGD system data are
summarized in Table B-7.
FUTURE PLANS
One module is being taken off line to implement modifi-
cations while the other remains on-line. If the modifica-
tions are successful, then the other module will also be
modified.
PERFORMANCE HISTORY
Figures on operability and actual S00 removal effi-
ciency are not yet available on this FGD system.
14,15,16
Table B-7. FGD SYSTEM DATA, SOUTHWEST POWER STATION NO. 1
Unit rating
Fuel characteristics
FGD vendor
Process
New or retrofit
Start-up date
FGD modules
Efficiency, particulates
Efficiency, SO,,
Sludge disposal
200 MW
Coal, 3.5 percent sulfur
Universal Oil Products
Limestone scrubbing
New
April 1977
Two
99.7 percent
80 percent
Stabilized sludge is
disposed of in unlined
pond
B-17
-------
WIDOWS CREEK NO. 8
BACKGROUND INFORMATION
The Widows Creek Power Station of Tennessee Valley
Authority (TVA) is located in Bridgeport, Alabama. Widows
Creek No. 8 is now the only coal-fired boiler at the plant
that is equipped with an FGD system; a scrubber is also
planned for Unit 7. Unit 8 has a capacity of 550 MW and
uses coal with a sulfur content of approximately 3.7 per-
cent.
POLLUTION CONTROL
The FGD system was designed and built by TVA. It con-
sists of four Polycon limestone scrubber modules downstream
of an ESP. Each scrubber is designed to handle 25 percent
of the total flue gas flow. The modules each consist of a
large venturi throat to compensate for poor particulate
collection of the installed ESP, a grid type absorber, a
Chevron vane type demister and a steam reheater. The over-
all particulate and S0_ removal efficiencies of the modules
are 99.5 percent and 80 percent, respectively.
The FGD system is still in the debugging stage and has
not yet achieved simultaneous running of all four modules.
Shakedown is due to be completed at the end of December.
Sludge produced by the FGD system is deposited in an
unstabilized form in an unlined pond. FGD system data are
presented in Table B-8.
B-18
-------
Data on SO2 collection efficiency and operability are
not yet available on this FGD system since it started
17 18 19
operation only recently as shown in Table B-8. ' '
Table B-8. FGD SYSTEM DATA, WIDOWS CREEK NO. 8
Unit rating
Fuel characteristics
FGD vendor
Process
New or retrofit
Start-up date
Efficiency, particulates
Efficiency, SO-
Sludge disposal
550 MW
Coal, 3.7 percent sulfur
Tennessee Valley Authority
Limestone scrubbing
Retrofit
May 1977
99.5 percent
80 percent
Unstabilized sludge is
disposed of in unlined
pond
B-19
-------
SOUTH CAROLINA PUBLIC SERVICE CO.
WINYAH NO. 2
BACKGROUND INFORMATION
The Winyah Power Station of South Carolina Public
Service Company is located in Georgetown, South Carolina.
The plant now operates two coal-fired boilers; however, only
Boiler 2 is equipped with an FGD system. Unit 2 has a
rating of 280 MW, which is 35 MW less than Unit 1. Both
boilers use a low-sulfur coal(approximately 1 percent sul-
fur) as fuel.
POLLUTION CONTROL
Two Babcock and Wilcox limestone scrubbing modules are
located downstream of an ESP at Winyah Unit 2. The two new
modules scrub only 50 percent of the total flue gas and
remove 69 or 70 percent of the SO,, in this stream. The
particulate removal efficiency of the system is 99.4 per-
cent .
Since the system started up in July 1977, the only
reported problems are those with electrical relays. Sludge
produced by the FGD system is deposited in an unstabilized,
unlined pond, which is estimated to have a 20-year capacity.
FGD system data are presented in Table B-9.
FUTURE PLANS
South Carolina Public Service intends to build two more
boilers at the Winyah site. These will be of the same size
as the Boiler 2 (280 MW).
B-20
-------
Table B-9. FGD SYSTEM DATA, WINYAH NO. 2
Unit rating 280 MW
Fuel characteristics Coal, 1 percent sulfur
FGD vendor Babcock & Wilcox
Process Limestone scrubbing
New or retrofit New
Start-up date July 1977
FGD modules Two
Efficiency, particulates 99.4 percent
Efficiency, SO- 70 percent
Sludge disposal Unstabilized sludge is
disposed of in unlined
pond
PERFORMANCE HISTORY
Figures on availability, reliability, and operability
are not yet available on the FGD system because of its very
...... 20,21,22,23
recent start-up.
B-21
-------
WILL COUNTY NO. 1
BACKGROUND INFORMATION
The Will County Power Station of Commonwealth Edison
Company is located near Romeoville, Illinois. The plant has
four electric generating units with a total rated capacity
of 1147 MW. Only Unit 1 has been retrofitted with an FGD
system.
Unit 1, a wet-bottom, pulverized-coal-fired boiler
designed by Babcock & Wilcox, was installed in 1955. Rated
at 167 Wi, the unit now burns a low-sulfur (0.4%) Western
coal and a high-sulfur (4.0%) coal from Illinois.
POLLUTION CONTROL
The boiler is fitted with an ESP manufactured by
Western Precipitation Division. The ESP has a 79 percent
actual particulate collection efficiency, and it is gen-
erally used when the FGD system is out of service. The
present particulate emission rate from the FGD system is
equivalent to 0.06 Ib/MM Btu.
In 1972, the boiler was retrofitted with an FGD system
for particulate and SO,, control. Designed by Babcock &
Wilcox, the wet limestone system removes 99.7 percent of the
particulate matter and 82 to 90 percent of the SO2. FGD
system data are presented in Table B-10.
B-22
-------
Table B-10. FGD SYSTEM DATA, WILL COUNTY NO. 1
Unit location
Unit rating
Fuel characteristics
FGD vendor
Process
New or retrofit
Start-up date
Efficiency, particulates
Efficiency, S02
Water makeup
Sludge disposal
Romeoville, Illinois
167 MW
Coal, 4.0 percent sulfur
(design)
Babcock & Wilcox
Limestone scrubbing
Retrofit
February 1972
98 percent
82 percent
Open-loop, 1.66 gpm/MW
Stabilized sludge is
disposed of in lined
pond
The two scrubber modules were designed for 385,000 acfm
at 355°F and handle 50 percent of the total boiler exhaust
gas flow. Flue gas passes through the existing ESP and
enters the venturi scrubber. Particulate removal efficiency
is maintained by regulating the pressure drop across the
adjustable venturi throat to maintain a pressure drop of
about 9 inches of water. Gas velocity through the venturi
is about 135 ft/sec.
The quenched flue gas flows upward through the S0?
absorber tower and passes through two perforated trays. The
trays are wetted with limestone slurry sprays located above
the trays. The trays provide an extended wetted surface for
B-23
-------
absorption of S02 by the circulated slurry. Superficial
velocity through the absorber is 12.2 ft/sec. Pressure drop
through the two trays is 6 inches of water. Total system
pressure drop is 25 inches of water.
The cleaned flue gas passes upward through a two-stage
Z-shaped demister and then enters the reheater unit, where
its temperature is raised from 128° to about 165°F. Reheat
is necessary to prevent condensation in the fans, ducts and
the existing brick-lined stack and to increase plume buoy-
ancy.
Shortly after start-up in 1972 and during the initial
shakedown, the plant encountered many problems. Demister
plugging was an early problem and the demisters were even-
tually replaced. Modifications to the reheater were ef-
fected to reduce a vibration problem. Erosion and corrosion
of equipment were also problems. Commonwealth Edison shut
down Module B to concentrate on solving the problems of
Module A, and Module B was not restarted until May 1975.
Module A was then taken out of service in June 1975 to
allow installation of new demisters and a new reheater. Be-
cause of delivery delays on the new reheater for Module A,
this module was not restarted until late March 1976. Addi-
tional scrubber outages were due to delays in receiving
replacement parts and a lack of sludge disposal pond capa-
city.
PERFORMANCE HISTORY
FGD system operability for this system is highly vari-
able, but has been generally poor. The FGD system was
designed for 76 percent S02 removal, and the utility re-
ports an actual removal efficiency of 82 percent.
B-24
-------
is:24'25'26
March 1976
summary of the reasons for recent poor operability
July 1976
Oct.-Nov. 1976
May 1977
Jan. 1977
April 1977
May 1977
- The B-side module remained out of
service until March 29 because of delays
in the installation of a repaired re-
heater. During March, high-sulfur
Illinois coal was fired in the boiler.
- The A-side remained out of service the
entire month because of repairs to a
damaged absorber pump and pond thickener
sludge overloading.
- Problems related to the A-side were due
primarily to repairs to a recycle pump
and delays in the shipment of replace-
ment parts.
- The inlet valve on the recycle pump
failed and the pond water return pump
failed.
- The spent slurry valve failure, which
occurred in December, kept the B-side
out of service until January 27.
- B-side outages occurred twice for ab-
sorber suction head repair and once to
wash the I.D. booster fan.
- Both modules were out of service for the
first 15 days of the month due to trans-
former failure on the pond-water return
pump. In addition, side A was out once
because of inlet valve failure on an
absorber pump.
B-25
-------
REFERENCES FOR APPENDIX B
1. PEDCo Environmental, Inc. Summary of the Status of
Flue Gas Desulfurization Systems in the United States.
Prepared for U.S. Environmental Protection Agency,
Division of Stationary Source Enforcement, Washington,
D.C. Cincinnati, Ohio. EPA Contract No. 68-01- 4147.
Task 25. July 1977. pp. 5-17 to 5-21.
2. PEDCo Environmental, Inc. Summary Report - Flue Gas
Desulfurization Systems. Prepared for U.S. Environ-
mental Protection Agency, Division of Stationary Source
Enforcement, Washington, D.C. Cincinnati, Ohio. EPA
Contract No. 68-01-4147. Task 3. June-July 1977. pp.
52-55.
3. Isaacs, G.A., and F.K. Zada. Survey of the Flue Gas
Desulfurization System at the Cholla Power Generating
Station. Arizona Public Service Company. Prepared for
U.S. Environmental Protection Agency- Control Systems
Laboratory. Research Triangle Park, North Carolina.
PEDCo Environmental, Inc., Cincinnati, Ohio. EPA
Contract No. 68-02-1321. Task 6. May 15, 1975.
4. Op. Cit. No. 1. pp. 5-39 to 5-42.
5. Op. Cit. No. 2. pp. Ill to 119.
6. McDaniel, C.F. La Cygne Station Unit No. 1 Wet Scrub-
ber Operating Experience. Kansas City Power & Light
Co. Prepared for presentation at EPRI-FEA Coal Blend-
ing and Utilization Conference. June 16-17, 1976.
7. Isaacs, G.A., and F.K. Zada. Survey of Flue Gas De-
sulfurization Systems. La Cygne Station, Kansas City
Power & Light. Prepared for U.S. Environmental Protec-
tion Agency. Control Systems Laboratory, Research
Triangle Park, North Carolina. PEDCo Environmental,
Inc. Cincinnati, Ohio. EPA Contract No. 68-02-1321.
Task 66. July 1975.
B-26
-------
8. Op. Cit. No. 1. p. 4-1.
9. Op. Cit. No. 2. pp. 121 to 130.
10. Isaacs, G.A., and F.K. Zada. Survey of Flue Gas Desul-
furization Systems. Lawrence Power Station, Kansas
Power & Light Co. Prepared for U.S. Environmental
Protection Agency- Office of Research and Development,
Washington, D.C. PEDCo Environmental, Inc. Cincinnati,
Ohio. September 1975.
11. Op. Cit. No. 1. pp. 5-54 to 5-58.
12. Op. Cit. No. 2. pp. 179 to 188.
13. Kruger, R.J., and M.F. Dinville. Northern States Power
Company. Sherburne County Generating Plant Limestone
Scrubber Experience. Presented at Utility Representa-
tive Conference on Wet Scrubbing. February 23-25,
1977.
14. Op. Cit. No. 1. p. 4-4.
15. Op. Cit. No. 2. pp. 208 to 209.
16. PEDCo in house files.
17. Op. Cit. No. 1. pp. 221 to 223.
18. Op. Cit. No. 2. pp. 221 to 223.
19. PEDCo in house files.
20. Op. Cit. No. 1. p. 4-5.
21. Op. Cit. No. 2. p. 207.
22. Electrical World Directory of Electric Utilities.
McGraw-Hill. New York. 1976. p 710.
23. PEDCo in house files.
24. Op. Cit. No. 1. pp. 5-17 to 5-27.
25. Op. Cit. No. 2. pp. 65 to 81.
B-27
-------
26. Isaacs, G.A., and F.K. Zada. Survey of Flue Gas Desul-
furization System at the Will County Station Common-
wealth Edison Company. Prepared for U.S. Environmental
Protection Agency. Control Systems Laboratory, Re-
search Triangle Park, North Carolina. PEDCo Environ-
mental, Inc. Cincinnati, Ohio. EPA Contract No.
68-02-1321. Task 6. October 1975.
B-28
-------
TECHNICAL REPORT DATA
(Please read ! 'tu _.c uons on the reverse before completing)
REPORT NO.
EPA-600/7-78-032b
2.
3. RECIPIENT'S ACCESSION NO,
Flue Gas Desulfurization System Capa-
dlities for Coal-fired Steam Generators
Volume EL Technical Report
5, REPORT DATE
March 1978
6. PERFORMING ORGANIZATION CODE
AUTHOR(S)
T.Devitt, R.Gerstle, L. Gibbs , S.Hartman, and
R.Klier
8. PERFORMING ORGANIZATION REPORT NO.
PERFORMING ORGANIZATION NAME AND ADOR6SS
PEDCo. Environmental, Inc.
1499 Chester Road
:incinnati, Ohio 45248
10. PROGRAM ELEMENT NO.
EHE624
11. CONTRACT/GRANT NO.
68-02-2603, Task 1
2. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development*
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Task Final; 4-12/77
14. SPONSORING AGENCY CODE
EPA/600/13
Vs.SUPPLEMENTARY NOTES (*) Cosponsored by EPA's Office of Air and Waste Management.
Project officers are J.E.Williams (IERL-RTP, 919/541-2483) and K. R. Durkee
(OAQPS/SSSD. 919/541-5301).
16.ABSTRACT
repOrj- discusses the availability of technology for reducing SO2 emis-
sions from coal-fired steam generators using flue gas desulfurization (FGD) systems,
Foreign and domestic lime, limestone, double alkali, magnesium slurry, and We 11-
man-Lord FGD systems are described, and the design parameters and operating
experiences are discussed. Steps that have been taken to achieve high system opera-
bility are discussed. Also, disposal of FGD system wastes is discussed briefly.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Air Pollution
Flue Gases
Desulfurization
Coal
Boilers
Wastes
Alkalies
Scrubbers
Calcium Oxides
Limestone
Sulfur Dioxide
Dust
Air Pollution Control
Stationary Sources
Alkali Scrubbing
Particulate
Venturi/Spray Towers
Mist Eliminators
13B
2 IB
07A,07D 07B
21D 08G
13A
11G
18. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
510
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
B-29
------- |