EPA-650/2-73-038
December 1973
ENVIRONMENTAL PROTECTION TECHNOLOGY SERIES
irljXlKijIWijKx-xjiwI^'JlSkiJlIf
:::;
-:
S
-------
EPA-650/2-73-038
PROCEEDINGS:
FLUE GAS DESULFURIZATION
SYMPOSIUM - 1973
May 14-17, 1973
Jung Hotel
New Orleans, Louisiana
E. L. Plyler, Chairman
M. A. Maxwell, Vice Chairman
Control Systems Laboratory
National Environmental Research Center
Research Triangle Park, North Carolina 27711
. ROAP No. 21ACY-30
Program Element No. 1AB013
NATIONAL ENVIRONMENTAL RESEARCH CENTER
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, N.C. 27711
December 1973
-------
PREFACE
A Flue Gas Desulfurization Symposium was held May 14-17, 1973, in
the Jung Hotel, New Orleans, Louisiana, under the sponsorship of the
Environmental Protection Agency, Office of Research and Development,
Control Systems Laboratory. The primary purpose of the symposium was
to present the current status of "throwaway" and "regenerable" flue
gas desulfurization processes as applied to controlling SO^ emissions
from full-scale facilities.
The symposium included sessions on technology and application of
first generation processes such as lime/limestone scrubbing, magnesia
scrubbing, catalytic oxidation and sodium sulfite scrubbing as well
as a session on second generation or advanced processes. In addition,
a panel discussion was held concerning the disposal and uses of by-
products from flue gas desulfurization processes.
Over 430 representatives of government and industry were in
attendance during the 4-day symposium.
These proceedings have been reviewed by the Environmental Protec-
tion Agency and approved for publication. Except for minor editing
for consistency of presentation, the contents of this report are as
received from the authors. Approval does not signify that the contents
necessarily reflect the views and policies of the Agency, nor does men-
tion of trade names or commercial products constitute endorsement or
recommendation for use. All papers presented (as well as transcriptions
of the panel discussions and the symposium summary) are included in
these proceedings.
Copies of this report are available free of charge to Federal
employees, current contractors and grantees, and nonprofit organiza-
tions -- as supplies permit -- from the Air Pollution Technical Infor-
mation Center, Environmental Protection Agency, Research Triangle Park,
North Carolina 27711.
Publication No. EPA-650/2-73-038
iii
-------
CONTENTS
TITLE PAGE
OPENING SESSION
Keynote Address - THE HEALTH RATIONALE FOR STRICT CONTROL
OF SULFUR OXIDE EMISSIONS
V. A. Newill, Environmental Protection Agency,
Washington, D. C 1
TECHNOLOGICAL ALTERNATIVES TO FLUE GAS DESULFURIZATION
S. J. Gage, Council on Environmental Quality,
Washington, D. C 13
STATUS OF FLUE GAS DESULFURIZATION TECHNOLOGY
R. E. Harrington, Environmental Protection Agency,
Washington, D. C 57
STATUS OF JAPANESE FLUE GAS DESULFURIZATION TECHNOLOGY
J. Ando, Chuo University, Tokyo, Japan 69
ECONOMICS OF FLUE GAS DESULFURIZATION
G. T. Rochelle, Environmental Protection Agency,
Research Triangle Park, North Carolina 103
THROWAWAY PROCESSES - PART I
Session Chairman - F. T. Princiotta, Environmental
Protection Agency, Research Triangle Park,
North Carolina
STATUS OF TECHNOLOGY OF COMMERCIALLY OFFERED LIME AND
LIMESTONE FLUE GAS DESULFURIZATION SYSTEMS
I. Raben, Bechtel Corporation, San Francisco,
California 133
WASTE PRODUCTS FROM THROWAWAY FLUE GAS CLEANING PROCESSES -
ECOLOGICALLY SOUND TREATMENT AND DISPOSAL
J. W. Jones, R. D. Stern, and F. T. Princiotta,
Environmental Protection Agency, Research Triangle
Park, North Carolina 187
TEST RESULTS FROM THE EPA LIME/LIMESTONE SCRUBBING TEST
FACILITY
M. Epstein, Bechtel Corporation, San Francisco,
California 235
-------
TITLE PAGE
OPERABILITY AND RELIABILITY OF THE EPA LIME/LIMESTONE
SCRUBBING TEST FACILITY
H. W. Elder and P. E. Stone, Tennessee Valley Authority,
Muscle Shoals, Alabama
L. Sybert, Bechtel Corporation, San Francisco,
California
J. E. Williams, Environmental Protection Agency,
Research Triangle Park, North Carolina 333
SCRUBBING EXPERIMENTS AT THE MOHAVE GENERATING STATION
A. Weir, Jr., and L. T. Papay, Southern California
Edison, Rosemead, California 357
THROWAWAY PROCESSES - PART II
Session Chairman - F. T. Princiotta, Environmental
Protection Agency, Research Triangle Park,
North Carolina
A REVIEW OF BABCOCK AND WILCOX AIR POLLUTION CONTROL
SYSTEMS FOR UTILITY BOILERS
J. Stewart, Babcock and Wilcox Co., Barberton, Ohio . . 393
ONE YEAR'S PERFORMANCE AND OPERABILITY OF THE CHEMICO/
MITSUI CARBIDE SLUDGE (LIME) ADDITIVE SO? SCRUBBING SYSTEM
J. Sakanishi, Mitsui Aluminum Co., Omuta, Japan
R. H. Quig, Chemical Construction Corporation,
New York, New York 419
BRIEF PANEL DISCUSSION - SIGNIFICANCE OF OPERATION TO DATE
OF 156 MW CHEMICO/MITSUI LIME SCRUBBING SYSTEM
P. Wechselblatt, Chemical Construction Corporation,
New York, New York
J. Craig, Southern Services, Birmingham, Alabama
H. W. Elder, Tennessee Valley Authority, Muscle Shoals,
Alabama
F. T. Princiotta, Environmental Protection Agency,
Research Triangle Park, North Carolina 451
THE TVA WIDOWS CREEK LIMESTONE SCRUBBING FACILITY
PART I - FULL SCALE FACILITY
B. G. McKinney, Tennessee Valley Authority,
Chattanooga, Tennessee
A. F. Little, Tennessee Valley Authority,
Muscle Shoals, Alabama
J. A. Hudson, Tennessee Valley Authority,
Knoxville, Tennessee 475
VI
-------
TITLE PAGE
PART II - PILOT PLANT AND PROTOTYPE OPERATING EXPERIENCE
J. J. Schultz, Tennessee Valley Authority,
Muscle Shoals, Alabama
T. M. Kelso, Tennessee Valley Authority,
Muscle Shoals, Alabama
J. L. Graham, Tennessee Valley Authority,
Muscle Shoals, Alabama
J. K. Metcalfe, Tennessee Valley Authority,
Muscle Shoals, Alabama
N. D. Moore, Tennessee Valley Authority,
Chattanooga, Tennessee 495
STATUS OF C.E.'S AIR QUALITY CONTROL SYSTEMS
M. R. Gogineni, J. R. Martin, P. G. Maurin,
Combustion Engineering, Inc., Windsor, Connecticut . . 539
REGENERABLE PROCESSES - PART I
Session Chairman - G. G. McGlamery, Tennessee Valley
Authority, Muscle Shoals, Alabama
MAGNESIA SCRUBBING
G. G. McGlamery, Tennessee Valley Authority,
Muscle Shoals, Alabama 553
OPERATIONAL PERFORMANCE OF THE CHEMICO BASIC MAGNESIUM
OXIDE SYSTEM AT THE BOSTON EDISON COMPANY
PART I - G. Koehler, Chemical Construction Corporation,
New York, New York 579
Part II- C, P, Quigley, Boston Edison Company,
Boston, Massachusetts 605
DESIGN AND INSTALLATION OF A PROTOTYPE MAGNESIA SCRUBBING
INSTALLATION
B. M. Anz, C. C. Thompson, and J. T. Pinkston, United
Engineers and Constructors, Philadelphia, Pennsylvania. 619
REGENERABLE PROCESSES - PART II
Session Chairman - N. Plaks, Environmental Protection
Agency, Research Triangle Park, North Carolina
APPLICATION OF THE WELLMAN-LORD S02 RECOVERY PROCESS TO
STACK GAS DESULFURIZATION
R. T. Schneider and C. B. Earl, Davy Powergas, Inc.,
Lakeland, Florida 641
vii
-------
TITLE PAGE
APPLICATION OF S02 REDUCTION IN STACK GAS DESULFURIZATION
SYSTEMS
W. D. Hunter, Jr., Allied Chemical Corporation,
Morristown, New Jersey 657
THE CAT-OX PROJECT AT ILLINOIS POWER
W. E. Miller, Illinois Power Company, Decatur,
Illinois 673
MITRE TEST SUPPORT FOR THE CAT-OX DEMONSTRATION PROGRAM
G. Erskine and E. Jamgochian, Mitre Corporation,
McLean, Virginia 687
DISPOSAL AND USE OF BY-PRODUCTS FROM FLUE GAS DESULFURIZATION
PROCESSES: INTRODUCTION AND OVERVIEW
Panel Chairman - A. V. Slack, Tennessee Valley Authority,
Muscle Shoals, Alabama
J. M. Potts, Tennessee Valley Authority, Muscle Shoals,
Alabama 747
STUDY OF DISPOSAL AND UTILIZATION OF BY-PRODUCTS FROM
THROWAWAY DESULFURIZATION PROCESSES
J. Rossoff, R. C. Rossi, and J. Meltzer, Aerospace
Corporation, El Segundo, California 77S
EXPERIENCE IN THE DISPOSAL AND UTILIZATION OF SLUDGE FROM
LIME/LIMESTONE SCRUBBING PROCESSES
W. C. Taylor, Combustion Engineering, Inc.,
Windsor, Connecticut 799
FIXATION AND DISPOSAL OF FLUE GAS WASTE PRODUCTS: TECHNICAL
AND ECONOMIC ASSESSMENT
L. J. Minnick, IU Conversion Systems, Inc.,
Plymouth Meeting, Pennsylvania 835
UTILIZING AND DISPOSING OF SULFUR PRODUCTS FROM FLUE GAS
DESULFURIZATION PROCESSES IN JAPAN
J. Ando, Chuo University, Tokyo, Japan 875
LONG RANGE MARKET PROJECTIONS FOR BY-PRODUCTS OF REGENERABLE
FLUE GAS DESULFURIZATION PROCESSES
M. H. Farmer, Esso Research and Engineering Company,
Linden, New Jersey 891
NEW USES FOR SULFUR - THEIR STATUS AND PROSPECTS
H. L. Fike and J. S. Platou, The Sulphur Institute,
Washington, D. C 921
PANEL DISCUSSION: DISPOSAL AND USE OF BY-PRODUCTS
FROM FLUE GAS DESULFURIZATION PROCESSES 931
viii
-------
TITLE PAGE
ADVANCED PROCESSES
Session Chairman - J. H. Abbott, Environmental
Protection Agency, Research Triangle Park,
North Carolina
REMOVAL OF SO- FROM STACK GASES BY SCRUBBING WITH AMMONIACAL
SOLUTIONS: PILOT SCALE STUDIES AT TVA
G. A. Hollinden and N. D. Moore, Tennessee Valley
Authority, Chattanooga, Tennessee
P. C. Williamson, Tennessee Valley Authority,
Muscle Shoals, Alabama
D. A. Denny, Environmental Protection Agency,
Research Triangle Park, North Carolina 961
AN EPA OVERVIEW OF SODIUM BASED DOUBLE ALKALI PROCESSES
PART I - A VIEW OF THE PROCESS CHEMISTRY OF IDENTIFIABLE
AND ATTRACTIVE SCHEMES
D. C. Draemel, Environmental Protection Agency,
Research Triangle Park, North Carolina 997
PART II - STATUS OF TECHNOLOGY AND DESCRIPTION OF
ATTRACTIVE SCHEMES
N. Kaplan, Environmental Protection Agency,
Research Triangle Park, North Carolina 1019
STONE AND WEBSTER/IONICS SO REMOVAL AND RECOVERY PROCESS
N. L. Foskett and E. G. Lowrance, Stone and Webster
Engineering Corporation, Boston, Massachusetts
W. A. McRae, Ionics, Inc., Watertown, Massachusetts . . 1061
FOSTER WHEELER/BERBAU-FORSCHUNG DRY ADSORPTION SYSTEM FOR
FLUE GAS CLEANUP
W. F. Bischoff, Foster Wheeler, Livingston, New Jersey. 1081
THE ATOMICS INTERNATIONAL MOLTEN CARBONATE PROCESS FOR SO
REMOVAL FROM STACK GASES
W. V. Botts and R. D. Oldenkamp, Atomics International,
Canoga Park, California 1101
SYMPOSIUM SUMMARY
F. T. Princiotta, Environmental Protection Agency,
Research Triangle Park, North Carolina 1133
ix
-------
HEALTH RATIONALE
FOR STRICT CONTROL OF SULFUR OXIDE EMISSIONS
by
Vaun A. Newill, M. D. , S. M. Hyg,
Jean D. French, Dr. P. H.
Office of the Administrator
Environmental Protection Agency
Washington, D.C.
-------
.Man has been subject to air pollution since his primordial ancestor
lit the first fire. It was not, however, until people became crowded
together in cities that pollution was more than a family problem from
the hearth. Today with the phenomenal growth of both the population
and use of power in the technologically advanced countries, pollution
has reached such magnitude that it threatens the health and well being
of the population on a global scale.
Even though there had been occasional speculation and complaints
about the health effects of air pollution since the days of Edward I,
it was not until the occurrence of certain air pollution episodes in
the Valley of the Meuse, in London and in Donora, Pennsylvania that
people finally came to realize that it was a threat to human health.
These episodes made it obvious that under certain conditions air
pollution could kill.
Despite the drama of the episodes, the greatest impact of air
pollutants on human health results from day to day exposure under
unexceptional conditions.
Environmental pollutants can effect the health of individuals
or communities over a broad range of biological responses as shown
in Figure 1. At any point in time more severe effects such as death
will be manifested in relatively small proportions of the population.
Mortality studies have shown that elevated levels of S02 contribute
to approximately 1 percent of excess deaths.
-------
Illness from sulfur oxides may be either short-lived (acute) or
relatively permanent and irreversible (chronic).
Bronchitis is an example of a long-term or chronic effect.
Investigators in Japan, Britain, and the United States have demonstrated
excess bronchitis morbidity in areas with moderately elevated 862
exposure. A significant and consistent excess bronchitis morbidity
3
occurred over an exposure range of 100 to 350 yg/m S02 with associated
total suspended particulate levels of 66 to 365 yg/m3. The following
slides show results from several studies conducted by the Environmental
Protection Agency in different areas of the United States.
Table 1 demonstrates that nonsmokers as well as smokers from high
exposure communities of four areas consistently reported more chronic
bronchitis than their smoking counterparts in low exposure neighborhoods.
As shown in the last row of Table 2, the contribution of air pollu-
tion towards the prevalence of chronic bronchitis is one-fifth to equal
that of cigarette smoking among males in the four areas.
A remarkably consistent excess of acute respiratory disease ranging
from 14 to 64 percent was found in children exposed to S02 levels of
50 to 275 yg/m3. These findings pertained largely to children who had
lived in the polluted areas for three years or longer suggesting a
chronic alteration in their defense mechanisms.
Figure 2 shows excess croup frequency among children from high
pollution neighborhoods in the Salt Lake Basin who had lived in that
neighborhood for three or more years. Similar findings in children
in Idaho-Montana are presented in Figure 3. Aggravation of symptoms
-------
in diseased subjects such as asthmatics and cardie-pulmonary subjects
have been repeatedly demonstrated at daily S02 levels well below the
indicated threshold for excess mortality. However, some studies have
demonstrated illness aggravation to be most strongly related to
suspended sulfates. Figure 4 reveals that at the same total suspended
particulate level, asthma attack rates were higher on days with high
sulfate concentrations. These data are from a study conducted by EPA
in the Salt Lake Basin.
As shown in Figure 5, when symptom aggravation in heart and lung
patients in New York were computed against exposure within a defined
temperature, daily levels of suspended sulfates more so than SC>2 or
total suspended particulates were closely associated with these attacks
of illness. The threshold for the adverse effect occurred at sulfate
concentrations of 9.2 ug/m , a level below the average daily sulfate
concentrations of most northeastern cities.
These findings confirm results from animal studies in which metallic
sulfates were shown to exert adverse biological effects at concentrations
below those of SOn. These results are physiologically coherent since
S0£ alone tends to be absorbed high in the respiratory tract while
sulfates can be delivered deep into the lung. It is important for
environmental scientists to determine the relationship between $©2 and
suspended sulfates. Although St^ serves as a precursor to the formation
of sulfates. the relationship is nonlinear; there is reason to believe
that S02 reacts in a complex manner with particulate matter. Certain
-------
metal particles, water vapor and oxidizing chemicals in the atmosphere
may promote the formation of more potent respiratory irritant sulfate
aerosols.
The section of the biological response spectrum entitled "functional
changes preceding disease" can be exemplified by pulmonary function
tests. A diminution in the pulmonary function test indicates some
resistance in air flow to the lungs and serves as a sentinel that a
disease process may ensue. Studies have shown lung function of school
children living in areas of moderate S0« and TSP pollution was impaired
compared with children living in low exposure communities. In a study
conducted in Japan, lung function of school children responded rapidly
to seasonal changes in air pollution. Function became worse in polluted
seasons and during less polluted months returned to levels of children
in low exposure areas. These data suggest that the lung function may
be reversible and improved air quality will enhance lung function in
children.
Experimental laboratory studies on humans and on lower animals
exposed to artificially produced S02 and suspended sulfates have shown
subtle physiologic changes, the significance of which is still obscure.
Russian investigators have demonstrated neurobehavioral effects of
sulfur oxides including changes in the electroencephalogram. Histo-
pathological studies in animals showed marked histopathological changes
in nasal epithelium at concentrations of S02 which had no direct effect
deeper in the lungs.
-------
There are no existing data on the relationship between sulfur
oxides and pollutant burdens.
Table 3 summarizes the levels of S02 and suspended sulfates
associated with biological responses over the entire spectrum. As
shown in the table, adverse health consequences ranging from functional
changes preceding disease to death itself have been associated with SC^
q
exposures in the range of 80 to 120 ug/nr for one or more days. Even
in communities where national primary standards for S02 have been
achieved, daily sulfate levels in the range of 7 to 14 ug/tn3 have been
associated with aggravation of symptoms in particularly vulnerable
population groups such as asthmatics and cardiopulmonary subjects.
Although pressure is currently being exerted to relax our existing
national air pollution standards because of the energy crisis, we must
bear in mind that such action could be costly in terms of safeguarding
the nation's health. We must pursue a prudent course whereby the
energy needs of the country are met in a manner which poses a minimum
threat to the public health.
-------
ACKNOWLEDGEMENT
The authors gratefully acknowledge the Human Studies Laboratory
of the U. S. Environmental Protection Agency for providing from their
research much of the data used in the figures and tables.
-------
Table 1. RATIOS OF CHRONIC RESPIRATORY DISEASE AREA PREVALENCE RATIOS
(MALES) BY GEOGRAPHIC AREA AND BY CATEGORIES
OF AIR POLLUTION AND SMOKING
Exposure
CRD prevalence ratios (males)
Utah
Idaho-
Montana
New
York
Chicago
Low
Nonsmoker
Smoker
High
Nonsmoker
Smoker
(3.0)'
6.5
2.3
8.9
1.0
(1.25)
13.6
2.8
14.7
1.0
(4.6)
3.0
3.4
4.7
Smoker 8^9 [ 14.7 j 417__
aBase period prevalence rate per 100 people in parentheses
1.0
(4.0)
3.8
1.3
4.5
Table 2. RATIOS OF CHRONIC RESPIRATORY DISEASE AREA PREVALENCE RATIOS
(MALES) BY POLLUTION AND SMOKING SOURCES OF RISK
Exposure
Pollution3
Smoking
Pollution
Smoking
CRD relative prevalence (males)
Utah
2.27
6.53
1
3
Idaho-
Montana
2.78
13.64
1
5
New
York
3.45
3.02
1
1
Chicago
1.33
3.80
1
3
Nonsmokers of high exposure/nonsmokers of low exposure area.
'Smokers of low exposure area/nonsmokers of same area.
-------
Table 3. EFFECTS OF SULFUR OXIDES ON HEALTH
BY LEVEL OF RESPONSE
Level of
resoonse
SQ2,
(ppm)
500-1000 (0.20-0.40)3
80-275 (0.03-0.11)a
90-120 (0.035-0.45)1
500-1000 (0.20-0.40)c
No data
Sulfates,
jjg/m3
Death
Illness (acute and
chronic)
Functional changes
preceding disease
Changes of uncertain
significance
Pollutant burdens
24 hour average.
bn
Annual average.
Experimental studies with artificial S02 or other sulfur oxides,
No data
7 14a
9-lla
25QC
No data
ADVERSE
HEALTH
EFFECTS
PHYSIOLOGIC \ T
HANGES \
PHYSIOLOGIC CHANGES OF
UNCERTAIN SIGNIFICANCE
POLLUTANT BURDENS
\
PROPORTION OF POPULATION AFFECTED •
Figure 1. Spectrum of biological response to pollutant exposure.1
-------
30
25
£ 20
0
c/o
K
UJ
Q_
O
o
S 15
UJ
ca-
ll.
•=2
0
cc
o
10
5
COMMUNITY EXPOSURE
•
15.7
LOW
INTERf
i MT ror
m 1 tnl
HIGH
11.7
MEDIATE
MEDIATE
11.7
v:::::v::::::;
9.4
•
II
•
•Xv.vXv.
X'X'I'X'X'I
mm
16.9
14.5
16.5
:::X::;X::j:v
26.4
1 ^
I
'XvIvivX
.••••"•"•"•*» «V
1
^R%%
•x-x-x-x-:
x'$Sx$§
"**••**• * i
>3 (n-7763)
RESIDENCE DURATION, years
Figure 2. Salt Lake Basin: incidence rates for croup in children by residence
duration and sulfur oxides exposure.^
10
-------
12
COMMUNITY EXPOSURE
POOLED LOW
POOLED HIGH
o
J
•
8.1
6.2
11.2
<3(n=1468) RESIDENCE DURATION, yea.
Figure 3. Idaho-Montana: Incidence rates for croup in children by residence
duration and sulfur oxides exposure.2
11
-------
25
£ 20
£E
UJ
D_
«ar
OS
£
15
10
WITH HIGH SULFATES
WITHOUT HIGH SULFATES
(26)
I
(71)
I
I
I
i
50
150
100
TSP CONCENTRATION,
Figure 4. Salt Lake Basin: asthma attack rates by total suspended particulate
level and by category of sulfate content.2 (Minimum temperature >51 °F.)
200
oc.
CD
o
t
a.
uu
oc.
UJ
CD
•f
o
K.
10 15 20
DAILY SS CONCENTRATIONS, jig/n>3
Figure 5. New York: Percentage of cardiopulmonary subjects reporting sympton
aggravation by daily suspended sulfate (SS) concentrations.2
25
12
-------
TECHNOLOGICAL ALTERNATIVES
TO FLUE GAS DESULFURIZATION
by
Stephen J . Gage
Council on Environmental Quality
Washington, D.C.
-------
TECHNOLOGICAL ALTERNATIVES
TO FLUE GAS DESULFURIZATION
Stephen J. Gage
Council on Environmental Quality
Washington, D. C»
Introduction
The single roost important factor today in controlling
air pollution from stationary industrial sources is the
reduction in emissions of sulfur oxides. The omnipresent
sulfur in domestic coals and in both domestic and imported
oils poses a serious challenge to the efforts of Federal,
state, and local environmental agencies to significantly
diminish the release of over 30 million tons of sulfur
oxides into the atmosphere each year.
Natural gas is, of course, the cleanest fuel but it is
already in short supply and, even with accelerated
domestic development and importation of LNG, demand will
undoubtedly outstrip supply. Also new supplies are likely
to be bid away from industrial and utility uses for
residential and commercial uses. Domestic oil production
has leveled out and it is expected that much of our future
oil needs will be met by imports of foreign, mostly Mid-
Eastern, oils. But these oils generally contain sulfur in
concentrations which will, in many cases, preclude use of
the residuum without some measure of hydrodesulfurization.
Desulfurization of imported oil certainly represents one
strategy for meeting the.need for low sulfur fuels. But
such a solution would spawn a set of new problems. There
are the fundamental problems of national and economic
security arising from reliance on large imports of a vital
fuel. Since new refineries equipped for desulfurization of
residual fuel would be required, siting and environmental
problems of supertanker ports and refineries emerge. If
such facilities are located offshore, there is further
aggravation of the balance-of-payments problem. However,
importation of crude oil for desulfurization in this country
and of desulfurized residual fuel will be important in
meeting our clean fuel needs. The technology for
14
-------
desulfurization of oil is reasonably well established so it
will not be covered in this analysis of technological options.
Coal, our most abundant fossil fuel resource, presents
the serious environmental challenge. As the traditional fuel
used in utility and industrial power plants in large sections
of the nation, particularly in the industrial Midwest, coal
is the source of a significant fraction of the sulfur oxides
emitted in those areas. Nearly all of the coal in the
Central coal region has sulfur content exceeding 3 percent
and much of the coal in the Appalachian coal region has
sulfur content in the 2 to 3 percent range. In the eastern
half of the country, only Central Appalachia* has significant
deposits of low sulfur coal. And, unfortunately, the terrain
in this area is very mountainous so the a11-too-common
ravages of surface mining—landslides, slumps, massive
erosion, and acid mine drainage—are even more severe.
Most of the coals in the vast sub-bituminous and lignite
fields of the West have lower sulfur content but their use
to solve air pollution problems in the Midwest appears to be
limited to the western fringe of the industrial region
(Minneapolis, Chicago, Kansas City) because of the high cost
of transportation from Montana and Wyoming. Although Western
coals are already being used locally, limited water avail-
ability and problems with re-establishing vegetation on
mined lands are potential constraints.
Technology which removes the sulfur from the coal before
combustion or from the gases during or after combustion is
therefore essential if air quality is to be improved and if
coal production is not to be displaced by a flood of imported
oil. Maintenance of a viable coal industry is also essential
if part of our clean fuel needs are to be met in the 1980's
with synthetic fuels produced from coal.
*Southern West Virginia, western Virginia, eastern Kentucky,and
northeastern Tennessee
15
-------
One technology for reducing sulfur oxide emissions from
coal-burning plants is the subject of this symposium—flue
gas desulfurization or stack gas cleaning. The past year has
seen both significant progress in the development of commercial
stack gas control systems and a rising chorus of concerns
about the application of such systems. Questions about the
reliability, costs, and associated environmental effects of
stack gas control systems have generated a more fundamental
question: Because of the current and potential problems with
stack gas cleaning devices, and the promise of precombustion
coal treatment processes, shouldn't application of sulfur
oxide abatement technologies be delayed until the new coal
cleaning processes become available? Even if the premises to
this question are conceded, the answer is not a simple yes
or no; rather the answer may be both no, in the near-term,
and yes, in the intermediate-term.
This paper will attempt to place into perspective the
current status of stack gas cleaning systems relative to the
status of coal cleaning processes. It will review the status
of the coal precombustion cleaning processes, estimates of
their costs, and forecasts of their application. The focus
will be on liquefaction and gasification processes which
will convert high sulfur coals into usable clean fuels in
either liquid or gaseous form.
Status of Coal Liquefaction Processes
One approach to precombustion cleaning of coal is to
liquefy the coal and to remove the ash and sulfur from the
liquid phase. There has been considerable experimental work
over the past decade to produce a synthetic crude oil which,
in turn, could be used in the manufacture of gasoline.
Conversion of coal to a liquid fuel requires the addition of
hydrogen, but considerably less hydrogen is required to
produce a power plant fuel than is required to produce feed-
stock for a gasoline plant. In fact, a liquid product is not
required for power plant application, although a liquid
phase is required for the hydrogenation and separation of ash
and sulfur.
Solvent Refined Coal
One of the most advanced process for producing an
ashless, low-sulfur power plant fuel is known as solvent
refining [1,2,3]. In this process, pulverized coal is first
16
-------
mixed with a heavy aromatic solvent (derived from coal) and
then passed through a reactor under a hydrogen atmosphere at
about 1000 psi and 800°F. Small quantities of hydrocarbon
gases, hydrogen sulfide, and light liquids are formed along
with a heavy organic material called Solvent Refined Coal
(SRC). The gases are separated and treated; the solution
is filtered to remove the ash and some undissolved coal; the
solvent is recovered through flash evaporation. A schematic
diagram of the SRC process is shown in Figure 1.
The SRC is a black pitch-like substance with a melting
point of about 350°F containing about 0.1 percent ash. Its
heating value is about 16,000 Btu's per pound regardless
of the quality of the coal feedstock. Most of the
inorganic sulfur is removed in this process and as high as
60-75 percent of the organic sulfur. For instance, it
appears possible to convert Illinois coal with 3.5 percent
sulfur to SRC with 0.6 percent sulfur. With the higher heat
content, SRC with 0.6 percent sulfur would release about 0.4
pounds of sulfur per million Btu.
A 50 ton/day pilot plant is now being constructed near
Tacoma, Washington by Pittsburgh and Midway Coal Mining
Company, one of the developers of the SRC process [4,5,6,].
'This project is being sponsored by the Office of Coal Research,
Scheduled for completion in early 1974, this plant should
Provide the engineering data necessary to build a commercial
Plant. A smaller pilot plant, 6 ton/day, is being constructed
under the sponsorship of the Southern Company and the Edison
Electric Institute and is scheduled for completion in late
1973. This plant, which will use a solvent refining process
developed by Consolidated Coal Company, will be built near
Wilsonville, Alabama [4,5].
H-Coal Process
The H-Coal process, developed by Hydrocarbon Research,
Inc. (HRI), introduces a coal slurry into a ebullating bed
reactor where, in the presence of a catalyst (cobalt
molybdate), hydrogenation occurs [2,3,5]. The upward flow
of the feed—which consists of crushed coal mixed with
recycle oil and hydrogen—maintains the catalyst in a state
of rapid motion and permits the continuous passage of
Unconverted coal and ash from the reactor. The H-Coal
reactor operates at 2700 psi and 850°F. The problems of
introducing feed into the pressurized reactor vessel and
separating the unconverted coal and ash from the synthetic
17
-------
Mining
Fuel Gas
Hydrogen (30-40 Ibs/ton Coal)
Coal
Prep'n
Vent Gas
Power
Vent
gas
Solv.
Ret.
Energy.
1
RESIDUE
DRYING^
j
FURNACE
Stack
gas
cont. '
en
1
ACID GAS
REMOVAL
H2S
SULFUR
PLANT
Ash
Sulfur
Solvent Recycle
Chem. Prod.
Light Oil
Gas Oil
De-ashed
Fuel, Lig.
or Solid
FIGURE 1. Solvent Refined Coal (SRC) Process
-------
oil await final solution. HRI has stated that the processing
conditions can be varied to obtain 1.5 to 1.0 percent sulfur
without solids removal, 0.5 percent sulfur with ash removal
and increased conversion severity, and 0.25 percent sulfur with
secondary coal conversion [3]. The H-Coal process has been
tested in a 2.5 ton/day pilot plant with high conversion and
good catalyst recovery and without serious difficulties [1,5].
HRI has stated interest in building a 250 ton/day demonstration
plant, but the financing for such a plant has not been
secured [5]. A schematic diagram is shown in Figure 2.
COED Process
Another process which produces a liquid product from coal
is the COED (Coal Oil Energy Development) process [2,3,4],
Developed by Food Machinery Corporation (FMC) under sponsor-
ship of the Office of Coal Research, this multiproduct process
uses fluidized-bed pyrolysis. A synthetic crude oil, a gas
stream, and a char product are produced. Powdered coal is
subjected to increased temperatures in four successive beds.
Volatile liquid products are withdrawn at every stage and not
subjected to the higher temperature where further cracking
would occur. The liquid fuels yield is lower than for coal
hydrogenation. Further, the sulfur content of the product
char may be too high to meet air quality standards in many
areas, without further desulfurization. The COED process can
t>e modified somewhat to yield different product mixes, i.e.,
the hydrogen-rich gas stream can be used to convert more of
the char to fuel oil. Pilot plant experience has indicated
yield of about 1.2 barrels of fuel oil and 9000 standard
cubic feet of gas per ton of coal. FMC has been operating a
36 ton/day pilot plant near Princeton, New Jersey, for
several years to obtain process information on various types
of coal [4]. No plans have been announced for a scale-up
to a demonstration plant.
.Coal Hvdrodesulfurization
Another promising coal liquefaction process being
developed by the Bureau of Mines (BOM) promotes
hydrodesulfurization of the coal slurry by turbulent flow
with hydrogen through a fixed bed of catalyst [1,3,7,8].
The coal, suspended in coal - derived product oil, is pumped
with hydrogen through a packed bed of pelletized cobalt
ttolybdate alumina. Turbulent flow through the bed prevents
plugging (as the coal passes through its stickly phase prior
to becoming liquid), keeps the catalyst surface clean for
19
-------
Vent Gas
Hydrogen
Mining
Preparation
©HSLUE
PREHEATER
OVERHEAD
SEPARATOR
H-COAL
REACTOR
2700 psi
850 F
CCTOolybdateCat
Cat.I F I
Liq.L, ,_!
Disengaging
EBULLATED
CATALYS3J
Note: P=Pump
Recycle Slurry Oil
Hot Oil
Recycle
_». Light Liq. Hydrocarbon
Atm. Bottoms Product
Overhead Product
>*• Vac^ jgttoma Slurry
fetin. Flashed Slurry
i-Char-Oil
FIGURE 2. H-Coal Process
-------
good contact, and promotes transport of the hydrogen into
the slurry. The reactor operates at 2000 psi and 850°F.
The goal of this design is to do just enough hydrogenation
to remove the sulfur with the liquefaction of the coal
essentially being a side effect, therefore use of expensive
hydrogen is minimized. Sulfur is removed as H2S which can be
easily converted in elemental sulfur. Experiments have
been conducted with a 0.1 ton/day reactor on 5 different
coals. For example, a Kentucky coal having 4.6 percent
sulfur and 16.0 percent ash was converted into a synthetic
fuel oil having only 0.19 percent sulfur and 1 percent ash.
BOM is now scaling up the size of the operation and is de-
signing a 5 ton/day pilot plant.
Status of Coal Gasification Processes
One promising approach to precombustion cleaning of
coal is gasification. Because of growing shortages in natural
gas supply, interest is running very high in gasification
processes which can produce a pipeline quality gas, i.e., a
synthetic gas composed primarily of methane with heat content
approaching 1000 Btu per standard cubic feet. More recently,
interest in producing a low-Btu content gas as a clean-
burning fuel for power plants has risen sharply. Because
high-Btu synthetic gas would be too expensive, e.g., $1.00-
1.50 per million Btu, for power plant use, this review will
concentrate on the direct use of low-Btu synthetic gas in
power plants. However, in order to assess the status of the
several gasification technologies, it will be necessary to
Review high-Btu gasification research.
For the sake of simplicity, coal gasification to low-
Btu gas can be divided into two categories: present
generation gasifiers (PGG) and advanced gasifiers (AG).
Present generation gasifiers include coal gasification
processes—Lurgi, Koppers-Totzek and Wellman Galusha—that
are now commercially available. In this country, industrial
interest is focused on the Lurgi process. Plans for
construction of full-scale commercial coal gasification
plants for production of pipeline quality gas have been
announced by two companies. In addition, one utility has.
initiated construction of a combined-cycle (gas turbine-
steam turbine) demonstration plant using a Lurgi gasifier to
produce the gas.
21
-------
As mentioned above, research on the production of
pipeline quality gas from coal has been receiving con-
siderable governmental and industrial support. A number of
advanced gasification processes are under development and a
cooperative government-industry effort, comprised primarily
of the Interior Department and the American Gas Association,
is currently evaluating four processes. Pilot plants are in
operation for two of the processes and under construction for
another two.
With utility interest in low-Btu gas for combined-cycle
plants growing, several study efforts have been initiated to
evaluate the possibility of modifying the advanced gasifiers
now under development to produce low-Btu gas. Since low-Btu
gas is produced in the initial stages of all of the advanced
gasifiers (prior to methanation of the hydrogen and carbon
monoxide), the proposed modifications appear to have merit.
Present Generation Gasifiers (PGG)
Coal gasification processes involve reacting coal,
steam, and oxygen under heat and pressure in a specially
designed reactor to form a synthesis gas of hydrogen, carbon
monoxide, carbon dioxide, sulfur componds and varying amounts
of methane [1,2,3,9,10].
The Koppers-Totzek and Wellman Galusha processes operate
at atmospheric pressure, produce no methane directly, and
require a rather high oxygen consumption [1*4]. The Lurgi
gasifier is the only high pressure gasifier (ca.300-450 psi)
commercially available [1,10].
The Lurgi Gasification Process. In the Lurgi process,
the synthesis reaction takes place in a water-jacketed vessel
[2,11]. The crushed coal (with the fines briquetted with
process tar) is fed through a coal lock chamber downward
into the gasifier. Steam and oxygen are introduced at the
bottom of the gasifier, heated by the high temperature ash at
the bottom, and flow upward through the reaction zone. The
crude gas and unreacted steam are quenched upon leaving the
gasifier to cool and remove dust, tars, ammonia, and
phenols. A cross-sectional view of the Lurgi gasifier is
shown in Figure 3.
22
-------
FIGURE 3. Lurgi Gasification Process
23
-------
For the production of pipeline quality gas, relatively
pure (>98 percent) oxygen must be supplied to the gasifier,
since any nitrogen dilutes the pipeline gas [11]. The crude
gas leaving the gasifier at 700 to 900°F will, in this case,
have a heat content, above 400 Btu per cubic foot. For
pipeline quality gas, the crude gas is subjected, in turn, to
a catalytic shift conversion process to establish the H2~to-
CO ratio required for the subsequent methanation step, a
physical absorption process to remove sulfur compounds and
carbon dioxide, and a methane synthesis process which
catalytically combines carbon monoxide and hydrogen to form
methane and water. The resulting gas, when compressed and
dried, has a heat content of 950 to 970 Btu per cubic foot
[11] .
On the other hand, for the production of low-Btu gas,
air may be fed to the gasifier, yielding a gas with a heat
content of less than 200 Btu per cubic foot [1,2,10,12,13].
Before using in either a conventional gas-fired boiler or in
a combined-cycle power plant, the sulfur componds (largely
H2S with some COS) must be removed. Although there are a
variety of available gas cleaning processes (Lurgi's pro-
prietary Rectisol process which uses low temperature methanol,
Stretford process, potassium carbonate scrubbing, etc.), all
require cooling the gas to 100 to 200°F [1,2,13]. This
requirement reduces the overall thermodynamic efficiency of
the gasification-power plant system. This represents a
serious fuel and cost penalty in the design of a combined-
cycle plant.
At the present state of technology, sulfur compounds
must be removed from the gas before it is introduced into the
gas turbine because of the sensitivity of the turbine blades
to sulfur exposure. However, the loss of efficiency in
cooling and reheating the gas for scrubbing significantly
degrades the attractive theoretical efficiency of the
combined-cycle. Research on processes to remove the sulfur
compounds in a high temperature environment is underway.
Development of a high temperature gas cleaning process is
probably essential for widespread application of both PGG and
AG systems.
-------
Commerical Applications of Lurgi Gasifiers. Abroad,
the Lurgi gasifiers have found commercial use in India,
Australia, West Germany, South Africa, USSR, and UK [9,11].
Recently, Lurgi gasifiers have been installed to provide
synthetic gas for a 165 MWe combined-cycle power plant in
Luenen, Germany [5,10].
Two full-scale commercial coal gasification plants have
recently been announced [11,14]. Both are to be located near
Burnham, New Mexico, and are to use strip-mined coal from the
Navaho Indian Reservation. One will use 8.5 and the other
9.7 million tons per year of coal. Each plant will employ
approximately 30 Lurgi gasifiers to produce 250 million cubic
feet of 970 Btu per cubic foot gas. One plant is planned by
El Paso Natural Gas and the other by the combination of
Transwestern Coal Gasification Company, Pacific Coal
Gasification Company, and Western Gasification Company. The
Plants are scheduled to go on line in the 1976-1978 period.
Commonwealth Edison is assessing the feasibility of
using Lurgi technology to gasify Illinois coal for direct
firing or in a combined cycle at its Powerton Station [4,12],
Called the "Clean Power Fuel Demonstration Plant," this
project seeks to demonstrate Lurgi technology with American
coals for power plant operation. A schematic of the
gasification plant for the Powerton Station is shown in
Figure 4. The project, sponsored in part by the Edison
Electric Institute, will involve 3 Lurgi gasifiers and is
slated to be in operation in late 1975.
j^dyanced Gasifiers (AG)
As indicated above, a number of advanced gasifiers
are under development. Four of the processes are now at the
Pilot plant stage while a number of other promising systems
are still in the experimental stage. Because of the
Possibility that some of these advanced gasifiers can be
adapted for production of low-Btu utility gas, they will be
briefly reviewed below.
Advanced Gasifier Pilot Plant Studies. The AG processes
which are at the pilot plant stage are indicated in Table 1,
along with the developmental and sponsoring agencies
and the pilot plant status [1,2,3,4,9].
25
-------
FIGURE 4. Commonwealth Edison's Clean Power Fuel
Demonstration Plant
EXPANDER /COMPRESSOR /
GENERATOR
DESULFURIZATION
SYSTEM
WATER
EXISTING WWERTON *4
120 MW STEAM TURBINE
-13-
-------
All of the four employ fluidized bed reactors with the
synthesis gas glowing through a coal bed. All processes use
the shift reaction, purification, and methanation steps
described above for production of pipeline quality gas with
the Lurgi gasifiers. Schematic diagrams of the four pilot
plant projects are presented in Figure 5.
TABLE 1
Summary of Advanced Coal Gasification Processes
in Pilot Plant Stage
Process
1• HYGAS
(Electrothermal)
2. CSG (CO2 Acceptor)
3• BI-GAS
4• Synthane
Development
Agency
IGT
Consol
BCR
BOM
Sponsoring
Agency
OCR-AGA
OCR-AGA
OCR-AGA
BOM
Pilot Plant
Status
In Operation
In Operation
Under Con-
struction
Under Con-
struction
IGT: Institute for Gas Technology
BCR: Bituminous Coal Research
Consol:Consolidated Coal Company
BOM: Bureau of Mines, Department of Interior
OCR: Office of Coal Research, Department of Interior
AGA: American Gas Association
Methane is produced in the AG by (a) devolatilization
°f the coal and (b) reaction of freshly devolatilized coal
with H2 and CO present in synthesis gas and steam introduced
into the reactor [2]. The four AG systems now in the pilot
Plant stage have a number of similarities but basically
Differ on the mode of production and composition of the
synthesis gas supplied to the reaction of H2 and CO with
freshly devolatilized coal.
27
-------
Hygas-electrothermal process
Fuel
iteom ond powei <;eneiolion
High BID goi
OGJ
Bigas process
low got
Stjam
Hijh-BIUiat
Carbon dioxide acceptor process
Synlheiii
]«
Dry lignilt
Steam
Synthane process
Hoi corfconoti
IttUui
FIGURE 5. Schematic Diagrams of Advanced Gasification
Pilot Plants
28
-------
Three of the processes—HYGAS, BI-GAS, and Synthane—
are fairly similar, in that ground coal is reacted with
steam and oxygen to produce a synthesis gas [1,2]. If the
coal is of the caking variety, it is typically pretreated
with oxygen to render it noncaking. The coal is then
introduced into the reactor vessel which operates around
1000 psi, the higher pressure shifting the equilibrium mixture
to higher concentrations of methane.
In the HYGAS-Electrothermal process, the hydrogen-rich
synthesis gas is produced by electrically-heated
gasification of char. In the BI-GAS process, oxygen is fed
to a high-temperature stage with recycled char and steam to
provide energy for production of the synthesis gas.
Similarly, in the Synthane process, synthesis gas is produced
by oxygen and steam reaction with char.
The C02-Acceptor process uses quite a different method
for supplying heat to the gasifier to generate the
synthesis gas. Process heat is carried to the reactor by
calcined dolomite (MgO-CaO) which releases both sensible and
chemical energy as it absorbs C02. The calcined dolomite is
regenerated with char in a separate fluidized-bed reactor.
This allows elimination of the electrothermal gasifier or
oxygen plant but reduces the gasification temperature so that
only the most reactive coals such as lignites can be used and
increases the complexity of the system.
Other Advanced Gasification Processes. In addition to
the four processes just described, there are a number of
other processes for which data has been obtained in
experimental units [1,2,9]. The Institute of Gas
Technology is developing two other process, both of which
use the same gasifier as employed in the Hygas-Electrothermal
process. In the HYGAS-Oxygen process (similar to the
Texaco process), oxygen is supplied to a separate char
gasifier, the synthesis gas from which is purified by the
removal of CC>2, the addition of steam, and catalytic shifting
before being fed into the hydrogasifier. In IGT's Steam-
Iron process, synthesis gas produced in an air-blown char
gasifier is used to reduce iron ore Fe304 to Fe + FeO. The
latter is used to decompose steam to hydrogen, reforming
Fe304 which is recycled. The steam is fed to a hydrogasifier.
29
-------
The Bureau of Mines is developing a second gasification
process called the Hydrane process which offers the unique
possibility of directly converting raw coal to methane by
reacting the coal with hydrogen [1,15]. While the economic
advantages of such an approach are well known, the tendency
of most American coals to agglomerate has led to a
pretreatment stage in which the coal feed stream is partially
oxidized. Such pretreatment also reduces the coal's
reactivity for methane production, thus requiring the indirect
steps of steam-oxygen gasification, water-gas shift reaction,
and methanation. In the Hydrane process, almost all (up to
95 percent) of the methane is produced by direct hydro-
gasification of raw coal in a "falling dilute phase" reactor
[15]. The remainder of the pipeline methane is produced by
methanation of carbon monoxide which constitutes 3 to 4
percent of the synthetic gas. The hydrogen for the hydro-
gasification is produced by gasification of the residual char.
The Hydrane process has been tested at bench-scale in a 3-
inch diameter reactor. The process offers the promise of both
high thermal conversion efficiency and low gas production costs,
Two other processes, both of which gasify coal with
steam and oxygen in molten baths, deserve mention. The
Kellogg process uses a molten sodium carbonate bath and the
ATGAS process uses a molten iron bath with a limestone flux
[1,2,4,16]. In the Kellogg process, the sodium carbonate
strongly catalyzes the steam-coal reaction, permitting
complete gasification at temperatures low enough to allow
methane production in the gasifier [1,2,16], The
gasification temperatures can be lowered from 1900°F to
1700°F. The molten carbonate also supplies heat to the
reaction, disperses the coal and steam uniformly throughout
the reactor, permitting the direct gasification of
agglomerating coal, and reacts with the sulfur to release
H2S in the product gas.
In the ATGAS process, the coal is dissolved in the molten
iron bath, releasing organic and inorganic sulfur
constituents [1,4]. Because of the high affinity of sulfur
for iron, iron sulfides are formed which migrate to a slag
floating on the molten iron bath. The iron sulfides react
with the limestone flux, releasing the iron and fixing the
sulfur as calcium sulfate which must be discarded. The
carbon which is dissolved in the molten iron is reacted with
steam and air to produce a sulfur-free synthesis gas of CO,
H2/ and nitrogen oxides. Little methane is produced directly
30
-------
so the indirect steps of shift reaction and methanation are
required to increase the heat content of the gas. It may be
possible to use the synthesis gas directly in a combined cycle
plant. The process operates at atmospheric pressures and
agglomerating coals can be gasified without devolatilization.
Outstanding problems with this process include lance con-
struction materials, refractory lining, and control of
particulate emissions. Experimental design data has been
obtained on a gasifier of about 2MWe capacity [4].
Adaptation of AG Processes for Low-Btu Gas Production.
The possibility of adapting AG processes for the production
of low-Btu utility gas is being investigated in a number of
laboratories. Early efforts have already verified the
potential for such adaptation as well as have identified some
of the unresolved problems [13,17]. Because of the loss of
efficiency associated with cooling the synthesis gas for sulfur
removal (as described above), probably the major unresolved
problem is the yet undeveloped technology for hot cleanup of
tar, dust, and sulfur from the synthesis gas.
It should be noted that, while the Office of Coal
Research has been given the responsibility for developing low-
Btu gasification processes as well as those for pipeline gas,
the trend in both OCR and industry funding is not supportive
°f adaptation of the largely developed AG processes. Rather,
OCR and industry appear to be pursuing new AG processes
specifically designed for low-Btu gasification.
Design of New AG Processes for Low-Btu Gas Production.
within the past two years, several new AG systems have been
Proposed for production of low-Btu utility gas production.
The major difference between these designs and adaptation of
existing high-Btu.systems is that the new designs are
optimized for production of low-Btu gas. Although several
e£forts are underway, probably the most advanced is the team
e£fort headed by Westinghouse and partially funded by OCR.
^e team also includes Public Service Indiana, Amax Coal
Company, and Bechtel [18,19], This effort would be oriented
toward developing a total combined-cycle system using a
fluidized combustor to produce a low-Btu (150 to 200 Btu
Per cubic foot)fuel gas. Lime would be added to the de-
volatilization stage where it reacts with the sulfur, producing
calcium sulfide particles. The synthesis gas emerging from
the reactor still has to be cleaned of particulates but the
9as is relatively free of sulfur compounds. The cleaned hot
9as can then be combusted to drive first a gas turbine and
31
-------
then give up the remainder of its heat to steam boiler.
Plans call for a pilot plant to be constructed in the
Pittsburgh area by mid-1974 to be followed by a commercial
plant to be constructed at a public service Indiana plant
near Terre Haute by late 1977.
Other industry teams are also working with the Office
of Coal Research to define design characteristics for low-Btu
AG and combined cycle systems [4]. One team is composed of
Combustion Engineering and Consolidated Edison. Another team
is made up of Pittsburgh and Midway, Poster-Wheeler, and
Pratt and Whitney.
Outstanding problems for all of these processes include
advanced gasifier design, hot gas clean-up, and advanced turbine
materials technology.
There are a number of other coal gasification processes
in various stages of R&D which have not been mentioned in
this review. Also the classification of the processes reviewed
has been somewhat arbitrary, e.g., the ATGAS process could
have placed in the last category considered since it appears
to have greater potential for production of utility fuel than
for pipeline gas. The omissions and commissions reflect only
the need for brevity and the judgment of the author.
Status of Advanced Combustors
There is another class of emerging technology which
should be mentioned because it bridges the gap between con-
ventional coal-fired boilers and coal gasification process.
This technology involves the combustion of coal in the
presence of a sulfur acceptor such as limestone. Because
of the necessity to continuously remove the ash and the
reacted and unreacted limestone, it is necessary to use a
moving or fluidized-bed combustor. The first generation
combustors operate at atmospheric pressure. In addition to
reducing sulfur emissions, such combustors may, because of
lower flame temperatures than in pulverized coal boiler, re-
duce nitrogen oxide formation and may offer cost savings.
One systems which should be mentioned is that developed
by Pope, Evans, and Robbins [2]. This fluidi«ed-bed boiler
has operated successfully in the pilot plant stage and
is ready for scaling to a demonstration plant. The Office
32
-------
of Coal Research has recently granted Pope, Evans, and
Robbins a contract for the design, construction and
operation of a 30 MWe fluidized-bed boiler. In England, the
National Coal Board is developing an atmospheric, fluidized-
bed combustor. Partial funding for these projects has been
provided by the Environmental Protection Agency and the
Office of Coal Research.
These systems do involve changes in boiler design and
manufacture and boiler operation which, although much less
extensive than with coal gasification, have represented a
barrier to acceptance by manufacturers and utilities. The
general opinion seems to be that, if significant departures
from conventional boiler design are to be made, then it is
logical to go the extra step to gasification with potential
for much improved thermal efficiency and decreased environ-
mental problems or to liquefaction with potential for
decreased environmental problems. Similar reasoning seems
to apply in the case of stack gas cleaning which is viewed
t>y many utility managers as a troublesome, short-lived
solution.
Comparison of Alternative Technologies
The alternative technologies described above embody
different physical processes which are introduced into the
utility fuel cycle in different ways, e.g., liquefaction may
produce a new fuel to substitute directly for coal while
gasification may require an integrated gasifier-gas turbine-
steam boiler system. Further, the various technologies are
at considerably different stages in the R&D process. In a
number of. cases, engineering studies of systems incorporating
the technologies have not been conducted. These factors
niake it quite difficult to compare the alternative
technologies on a common basis. Insofar as possible, the
technologies will be compared on the basis of (1) applic-
ability for power plant use; (2) capital and operating
costs; and (3) associated environmental factors. Because
of the limited utility interest in advanced combustors, the
comparison will be limited to coal liquefaction, present
generation gasifiers, and advanced gasifiers.
Applicability for Power Plant Use.
The major difference between the coal liquefaction and
gasification processes for utility boilers is that liquefield coal
products have high energy content (16,000 Btu per pound) and are
in a physical form convenient for both transportation and storage
33
-------
while the low heat content and physical form of low-Btu gas
from coal is such that it can be neither transported (by
pipeline) nor stored.
Coal Liquefaction. In terms of physical suitability,
liquefied coal products (LCP's) could,with minor boiler
modifications, be used in nearly all utility and industrial
boilers. For instance, SRC could be used in either the
pelletized solid form or, if heated, could be handled as a
viscous fuel oil. Coal liquefaction would probably be
accomplished at a large centralized, possibly mine-mouth,plant
and could probably thereby achieve economies of scale. The
ash and sulfur disposal problems at the processing plant could
be quite significant and may severely limit the number of
potential plant sites. The sulfur would probably be converted
to elemental sulfur which could be more easily stored than
many sulfur compounds.
LCP1s,largely ash-free, still contain some sulfur. SRC
with 0.6 to 0.7 percent sulfur could be used in nearly all areas
of the country under the New Source Performance Standards and
in most areas under standards established by State
Implementation Plans or by state and local regulations. But
the sulfur content in SRC just slips under the requirements
in many areas and is too high in some such as New Jersey and
New York. If allowable sulfur concentrations are reduced in
the future, SRC would probably not be adequate. By
comparison, the sulfur emissions with SRC are as much as twice
those apparently achievable with stack gas cleaning. Other LCP
processes, such as the Bureau of Mines turbulent catalytic
hydrodesulfurization, have demonstrated the ability to
reduce the sulfur content by one-half or even two-thirds
below that achievable with SRC. Supposedly such low con-
centrations can be achieved with H-Coal although the matter is
somewhat confused at this point [5 ].
Other factors which may affect the applicability of LCP's
are supply and cost. Much of the particulate and sulfur
emissions in urban and industrial areas comes from industrial
rather than utility boilers. Many of these boilers, because
of space limitations and highly unfavorable economics, cannot
be retrofitted with SGC or gasification systems. LCP's
represent, therefore, the only technological alternative to
low sulfur oil or natural gas. As LCP's become commercially
available, in limited supply at first, industrial boilers
may pay the premium price for the clean fuel Il8l• At the
same time, utilities may elect to use LCP's for their older,
34
-------
less efficient, lower load capacity boilers which cannot
be retrofitted with SGC systems and, under regulatory and
market pressures, retrofit their newer boilers with SGC
units or use low sulfur coal. LCP's may provide the option for
utilities to continue to use their older and smaller units
with reduced air pollution to meet peak loads—this could be
extremely important if siting problems, new plant startup
difficulties, and regulatory delays persist.
Present Generation Gasifiers. If the synthesis gas is
only to be combusted in a modified conventional boiler, then
PGG systems can be used with both new and existing plants.
However, to offset the increased cost of the Lurgi (or other)
gasifiers, a combined-cycle must be used to increase the
thermal efficiency. Commonwealth Edison plans to install a
gas turbine as part of the gasification demonstration
project at its existing Powerton Station. It appears then
that the application of PGG system will be restricted
essentially to new plants. There may be some limited
application of PGG systems for conventional boilers until hot
sulfur and ash clean-up processes and improved turbine blades are
developed but PGG systems will generally be applied on new systems,
Advanced Gasifiers. For the reasons just given, AG
systems will be applied only to new power plants. All AG
systems, either as adaptations of high-Btu content gasifiers
or as new designs involve the integration of a gasifier with
a combined-cycle unit.
.Comparative Costs of Alternative Technologies
A most important comparison, and the one most fraught
with uncertainties, is that of costs, both capital and
operating. Cost analyses of emerging technologies are
hazardous enough, but estimates of capital and operating
costs for processes only at the level of bench-top experiments
approach speculation. Also because of recent cost escalations
it is very important to compare costs calculated in the same
period and preferably as recently as possible. Cost analyses
as recent as 1970 and 1971 may be quite out-of-date. Some
recent cost data will be presented to indicate, at least,
relative economic attractiveness.
35
-------
One recent comparison based on conservative assumptions
casts an equally jaundiced eye at all technological alter-
natives [5]. This sobering comparison prepared by Chem Systems
Inc. should bracket the maximum costs that may be anticipated.
These costs estimates are given in Table 2.
Table 2
Cost Comparison of Technological Alternatives
for Sulfur Control
Technological Total Annualized Total Capital ,
Alternative Energy Costs,=* Investment Costs,—'
jzf/106 Btu $/KWe
Stack Gas
Scrubbing* 80-85 104
Low-Btu Gas 90-95 132
Liquefied Coal
Product 85-90 (?)
^•Interior coal priced at $7/ton (or 30^/106 Btu) and a load
factor of 0.70
2Includes contingency of 15% of capital expenses.
*The costs of stack gas scrubbing quoted by Chem Systems
were based on an analysis by the Federal Power Commission.
The system analyzed was a 1000 MWe boiler using high sulfur
Interior coal. The system economics were penalized by
inclusion of an electrostatic precipitator, coal storage and
preparation facilities and ash, dust, slag systems — the entire
waste disposal system — with the wet lime/limestone scrubber,
giving $70.5 per KWe. $17 million was included for utilities
and $13.5 million was included for contingency, yielding
$103.5 per KWe for total investment. The individual costs
were broken down as follows:
Coal 30>z< /!06Btu
Operating Costs 14^
Waste Disposal lljzf /106Btu
Capital Charges 3ljz? /106Btu
/106Btu
Chem System indicates that the "relatively high estimated
costs may have been unfavorably influenced by Commonwealth
Edison's recent experience on the Will County stack gas unit" [5]
36
-------
The estimates of low-Btu gas costs were developed by
the Federal Power Commission Coal Gasification Task Force
for a 1000 MWe plant using a combined cycle. The onsite
process costs were $71.8 per KWe and the total plant invest-
ment including contingency was $132 per KWe. The total energy
costs (in terms of gas) were 92ff per million Btu.
The economics of LCP's are somewhat more complicated in
that the cost of the product is quite sensitive to size of
plant. To illustrate the trends for SRC, the cost of SRC as
a function of coal feed to the plant is shown in Figure 6.
For a plant producing 7,700 tons of SRC per day* required
to fuel a 1000 MWe generator, the energy cost (in terms of
SRC) were 88^ per million Btu. As envisioned, one LCP plant,
probably located at the mine-mouth, would produce enough
fuel for several power plants. It is suggested that central
plants processing 40,000 to 50,000 of coal per day may eventu-
al ly be developed.
There are a number of other cost estimates for the
various technologies, however, it is indeed fortuitous when
the costs have been computed on a common basis. Three recent
publications gave cost estimates of electrical power production
from low-Btu gas, although not in directly comparable terms.
Lurgi indicates that, for a 330 MWe plant using 30jzf per million
Btu coal, the energy production costs would be about 10 mills
per Kw-hr and the capital costs would be about $200 per KWe [19]
IGT has estimated that, for a 1000 MWe plant using 24jz< per
million Btu coal, the capital costs would be $200 per KWe and,
assuming a 90 percent load factor, the total energy production
costs would be 6.9 mills per Kw-hr [20]. Finally, ATC has
estimated that, for a 1000 MWe plant using 30^ per million
Btu coal, the capital investment is $163 per KWe and, assuming
a 90 percent load factor, the energy costs would be 6.8 mills
Kw-hr [21] .
As mentioned above, cost analyses of LCP plants are
considerably more difficult because the effect of plant
size on the economics. Several recent publications have
estimated the costs of SRC and H-Coal, although again the
costs are not on a comparable basis. On the basis of an
analysis on a SRC plant processing 10,000 tons of coal per
day, Pittsburgh and Midway suggests that, as a rule of
*In the Pittsburgh and Midway SRC process, this output
would require 13,300 tons of Interior coal.
37
-------
S.R.C. SELLING PRICE
/MM BTU
i.oo- -
.90
.80
10,000
20,000
30,000
40,000
50,000
COAL FEED - TONS/DAY
FIGURE 6. SRC Economics as a Function of Plant
Capacity and Coal Price
38
-------
thumb, SRC will sell at a price equal to the cost of raw
coal plus 35jz? per million Btu [ 6 ] . For 3Q& per million
Btu coal, this gives an energy cost of only 65^ per million
Btu, considerably below the adjusted Chem Systems estimates.
Analysis by Chem Systems indicated for an SRC processing
44,000 tons of 35j£ per million Btu coal per day would yield a
fuel oil (produced by blending enough of the lighter hydro-
carbons formed in the SRC process with the SRC to make
127,000 barrels per day of a pumpable product) which costs about
7Q& per million Btu. The total plant investment would be
$209 million which,reflected in incremental capital costs to
electrical plants, would be about $63 per KWe [5]. For the
H-Coal process, two feed coals—both priced at 35jt* per million
Btu—were analyzed. Processing Illinois seam coal with
3.4 percent sulfur at the rate of 29,000 tons per day resulted
in an energy cost of 83.5^ per million Btu while a Pittsburgh
seam coal with 4.2 percent sulfur at 25,000 tons per day
resulted in 88.8/z? per million Btu. Both plants would
produce about 70,000 barrels per day of a pipeline product.
The incremental capital investments for the two plants would
be $78 per KWe and $90 per KWe, respectively. It should be
noted, however, that the interest rates used for the LCP
plants are those for normal refinery financing, rather than
utility financing, as usually used for stack gas scrubbing
and for low-Btu gas production.
Although not really germane to these comparisions, an
excellent common base cost comparison for pipeline coal
gasification was recently published [22]. To the extent
that some of the PGG and AG processes may be adapted for
low-Btu gas production, this is an informative comparison.
The economics of five processes are compared in Table 3.
The study also computed costs for coal at 35^ per million
Btu and for 9 and 15 percent gross return on rate base.
This study did not directly take into account the differences
between lignite and bituminous coals, although some of the
processes are operable only with specific grades of coal,
i.e., the CSG process must use lignite.
Associated Environmental Factors
While it is too early to characterize all of the
environmental effects associated with the alternative
technologies, it is possible to describe them in general and
to give a limited amount of specific information. Most of
the environmental factors, except for those associated with
mining, are most easily discussed by considering each
technology generically.
39
-------
Table 3
Comparison of Pipeline Gasification Process Costs [22]
Total Capital 20-year Average 2/
Investment,— Gas Selling PriceT
Process $106 jrf / 106 Btu
CSG (CO2-Acceptor) 112 73.5
BI-GAS 179 88.9
HYGAS (Electrothermal), 165 89.3
Kellogg Molten Salt 167 87.4
Lurgi 297 119.6
Plant capacity 250 million cubic foot per day; 20-year
plant life; 7.5 percent interest; taxes and insurance at
3% of total fixed investment
•j
Coal price 25jzf per million Btu; 12 percent gross return
on rate base; load factor 95 percent; by-product credit:
sulfur $5 per ton, ammonia $20 per ton, phenol $80 per ton,
char $0.15 per million Btu.
-------
Probably the major environmental abuse, one for which
we have not discovered a technological fix, is that due to
strip mining for coal extraction. If we increasingly rely on
strip mining for more of our coal production and do encounter
increased need for coal to feed coal liquefaction and
gasification plants, much improved surface mining regulation
and enforcement and real innovation in mining and reclamation
technology will be required. Because of the losses associated
with the conversion techniques, the land disturbed by mining
will correspondingly increase unless offsetting improvements
are made in the power generation cycles (such as combined-
cycles) .
Coal Liquefaction. The land impacts are those associated
with clearing and preparing the site for the plant, including
coal storage, handling and preparation facilities, ash and
sulfur disposal facilities, product storage facilities, and
transportation facilities. If the plant is located at the
mine-mouth, coal transportation would be minimized but LCP's
would have to be transported by train or pipeline to the power
plant. For many LCP plants,siting of both the plant and the
transportation corridor may be a significant problem. Disposal
of the ash, and possibly of the sulfur, may represent another
serious problem area. If the plant is located at the mine-
mouth, then disposal in the mined area might be feasible,
otherwise the solid wastes generated in a large LCP plant would
represent a massive landfill operation. A plant processing
40,000 tons of coal per day would generate 5,000 tons of ash
(dry weight) -per day using Eastern and Interior coals and up
to 10,000 tons per day using Western coals .
The air and water impacts of coal liquefaction plants
are not well characterized at this point. Most of the sulfur
removed from the coal is converted to H2S and then converted
to elemental sulfur by various processes. Light hydrocarbons
are generated in the liquefaction processes and some of these
will be vented to the atmosphere. Emission standards for these
and for H2S (not converted to sulfur) will have to established
for these large sources. A major variation in the air
emissions may depend on whether the char or other mineral
residue is fired for process heat or sold for combustion
elsewhere. Particulates and sulfur oxides from such combustion
will be difficult sources to control. Water clean-up does
not appear to represent a major problem, at least at this time.
-------
From an environmental point of view, LCP plants do hold
considerable promise for shifting much of the environmental
burden from the dispersed power plants to a centralized
location. This is quite attractive to the utilities.
Emissions to air and water may be more effectively controlled
at a larger central plant, however, this has not yet been
demonstrated. The concentration factor will undoubtedly
exacerbate other impacts such as those on land.
Coal Gasification. The environmental impacts of gasifi-
cation plants are better known, primarily because environ-
mental assessments have been completed for both commercial
high-Btu gasification plant proposals. The Federal Power
Commission's Natural Gas Survey [1] estimated the following
potential pollutants would be generated in a 250 million
cubic foot per day plant, assuming use of Interior coal with
3.7 percent sulfur and 10 percent ash:
Tons per day
Sulfur (Mainly as hydrogen sulphide) 300-400
Ammonia 100-150
Phenols 10-70
Benzene 50-30
Oil and Tars trace to 400
Ash (based on coal with 10% ash) 1500
Control processes are known for many of these potential
pollutants and are being incorporated into the engineering
design of gasification plants.
As an example of the type of control planned for a high-
Btu gasification plant, the sulfur balance diagram for the
proposed Transwestern gasification plant is shown in
Figure 7. The plant will use 25,600 tons per day of low
sulfur coal for the gasifiers and process heat boilers. The
sulfur input is 233 tons per day and only 7.5 tons are to be
emitted, over 60 percent of which are from the coal-fired
boiler. The maximum 24-hour average SOX concentration expected
is about 0.03 ppm. Concentrations of NOX, K^S, COS, and
particulates are a fraction of the allowable New Mexico
standards [11].
-------
-tr
U)
S* 3.0 (SULFUa LEAU GA5)
ANALYSIS, WT%
MOISTURE IZ.-4
ASH 25.6
FIXED CAR&Oti 33.6
VOL. H&TTEO. 28-2
1OO.O
=0.336
L8S/UMBTU
SULFUR WT%
HHV,BTU/LB 63 IO
LHV, BTU/LB 7B6O
5-0.0077
LB5/MMBTU
50^=0.0154-
LSSfMMBTU
BOILES
COAL
123
14.1
40.Z
334
too.o
0.87
987O
942O
0.28
H.49Q
16, 500
Total Sulfur Emissions
7.46 Tons/Day
NOTE'
ALL QUANTITIES ABE in SHOUT rows
EXCEPT SHOWU OTHEBWISE.
FIGURE 7.
Sulfur Process Diagram for Transwestern
Coal Gasification Plant
-------
Water impacts of coal gasification plants of two types:
pollution potential and consumption. For instance, in the
Transwestern proposal, all the gas liquor streams are
treated to remove the phenols, ammonia, and I^S and all
waste water streams treated to remove suspended oils and
solids and residual organics. Water consumption, minimized
in the Transwestern proposal by recycling, is 5100 gallons
per minute (or 8,260 acre-ft per year) Twenty percent of
this is consumed directly in the gasifier and 50 percent is
evaporated in cooling towers and settling ponds. Without
recycle, the consumption could have been 8,000 to 10,000
gallons per minute. In water-scarce areas, such water
consumption for both PGG and AG systems may become an
important factor.
Forecasts of Application
Forecasting application of emerging technologies is a
difficult matter. However, because of their relevance to policy
development for air pollution control, even rough forecasts
are infinitely better than none at all. This section will
review the major features of several recent forecasts of the
application of technological options for clean fuel.
CEQ Forecast
A forecast of the application of all of the major
technological alternatives for the use of high sulfur coals
for steam-electric plants has been developed by the Council
on Environmental Quality [18] . This forecast include four
options: staclc gas cleaning (SGC), present generation
gasifiers (PGG), advanced gasifier (AG), and coal
liquefaction (CL). The predictions assume that the forecasts
for SGC installations in the "realistic" scenario of the
SOCTAP report are valid from 1974 to 1977. In the period
1978 to 1980, it is assumed that only new coal-fired plants are
equipped with SGC technology; beyond 1980 it is assumed that
no plants are outfitted with SGC systems. These conservative
assumptions would mean that 84,000 MWe, of which 61,000 MWe
represent new plants, would have SGC devices by the end of
1980. Limited application of PGG processes would begin in
1975 and application of AG system would begin in 1980.
Application of CL plants would begin in 1979 and would expand
rapidly. The allocation of these processes to fuel new or
existing plants are predicated on many factors which have
been discussed above. The results of these assumptions are
presented in Table 4 and are shown graphically in Figure 8.
44
-------
TABLE 4.
PROJECTIONS OF UTILIZATION OF CLEAN FUEL OPTIONS
WITH LIMITED APPLICATION OF SGC SYSTEMS
Ul
Tech- Type of
nology Install,
new
SGC retro
new
PGG retro
new
AG retro
new
CL retro
Annual Capacity
Addition, KWe
Cumulative Capacity,
KWe
Coal Usable,
105 tons
Total Steam-Electric
Coal Use. 106tons
Percentage with Control
Coal-Fired Steam Electric Capacity with
74 75 76 77 78
900 3.800 8,200 12,000 12,000
1,600 3.600 5,600 12.000
1.000 1.000 1,000 1,000
2,500 8,400 14,800 25,000" 13,000
2,500 11.000 26.000 51,000 64,000
6.4 28 66 129 162
384
7.3%
Clean Fuel Technological
79 80 81
12,000 12,000
1.000 2,000
10,000
1,000 5,000 8,000
13,000 18,000 17.000
77,000 95,000 112.000
195 240 283
460
52%
Options.
82
4.000
12,000
8,000
24,000
136.000
344
MWe
83 84 85
8,000 12.000 12,000
10,000 10,000 10,000
8.UOO 10,000 10.00C
26,000 32,000 32,00n
162.000 194,000 226,000
410 491 572
613
93%
-------
300 -
Elec.
20O -
Capacity,
1OOO MWe
100 -
FIGURE 8. CEQ Projections of Utilization
of Clean Fuel Alternative Technologies
YEARS
-------
This forecast shows dramatically that the only viable
technological alternative for reducing sulfur oxide emissions
until the early 1980's is stack gas cleaning. Even with rapid
expansion in LCP availability, the first priority for use of
LCP's will be for existing plants. Competition from
industrial boilers also could limit availability of LCP's.
In this forecast, equal amounts of coal cleaning would be
achieved by SGC and CL in 1985, since AG and CL supply all
clean fuel needs after 1980.
Hittman Forecast
The application of stack gas cleaning and solvent refined
coal for power plants has been forecast by Hittman
Associates [4]. Hittman emphasized only coal liquefaction
because,in their opinion, SRC is the most advanced clean
fuel option and has the greatest likelihood of success.
Hittman assumed three cases: Base, Higher SGC, and Sooner SRC.
In the Base case, it is assumed that commercial application
of SGC begins in 1976 and the maximum supply capability is
limited to installation on 20,000 MWe per year. SRC is
commercially available in 1981 and is also supply limited for
the first decade. The Higher SGC case differs from the
Base case only in that the maximum supply capability is
assumed to be 40,000 MWe per year. The Sooner SRC case
differs from the Base case only in that SRC is commercially
available in 1979. The projected application of SGC and
SRC technologies for these three cases is shown in Figure 9.
Some features of this forecasts are well-illustrated in
Figure 9. SGC application is supply constrained until the
1983-1986 period; after then SGC application is demand
constrained. SRC application is supply constrained and
since, from the utilities' point of view, is the desirable
approach to coal cleaning, SRC substitutes directly for
SGC to the extent possible. Utilization of SGC continues,
in this forecast, not only after 1980 but even after the
1983-1986 period. This application continues, although only
on new plants, because Hittman feels that SGC will offer
the least expensive alternative in many cases. The effect
of the technological alternatives in the three cases on the
reduction of sulfur oxide emissions is shown in Figure 10.
-------
16 i—
14
£H
O1
12
Ho
J~
O '
10
O
U
u
2
w
fn Q
OW
BASE CASE - SGC
HIGHER SGC-SGC
SOONERSRC-SGC
SOONER SRC-SRC
1975
BASK C A SI-]- SRC
HIGHER SGC-SRC
1980
,-^T' ' -J—I «—i
1885
YEAR
1
1U90
-t. i t L
1E95
2000
FIGURE 9. Hittman Projections of Utilization of Clean Fuel
Alternative Technologies
-------
32
28
u
<;
AH
I .1.
J L
J L
FIGURE 10. Reduction in Sulfur Oxide Emissions Associated
with Hittman Projections
-------
MITRE Forecast
Application of SGC and SRC processes as well as synthetic
crude oil production and gasification were also projected by
MITRE [23]. While the projection given in Figure 11 covers total
coal utilization, the SGC and SRC components would probably
be used mostly by the electric utilities. MITRE projects an
optimistic expansion of coal utilization which triples over
the next 15 years. Interestingly, MITRE assumes that 250
million tons of coal could be burned in 1980 with SGC which is
roughly 100,000 MWe. The MITRE projection for SRC and
synthetic crude production accounts for 150 million tons of
coal which is much larger than that projected by any other source,
Chem Systems Projections
Chem Systems has made predictions when the CL and PGG/AG
systems might become available [5 ]. Chem Systems believes that
CL processes will become available by 1978 with construction
and initial operation of both SRC and H-Coal plants. Chem
Systems predicts that PGG plants will become commercial in
the U.S. by 1976-1977 and application will start before the
end of the decade. AG power cycles will be 3-4 years behind
the PGG systems. Chem Systems suggests the following
scenario for PGG system application:
1976 Commonwealth Edison 3-gasifier system
successfully in operation.
1978 Installation of gasifiers on two 300-500 MW
retrofit generators
1980 Gasifier systems in operation on three new
1000 MW generators and on five additional
300-500 MW retrofit installations.
For this optimistic scenario, Chem Systems predicts a total
of 5000-6000 MW of generator capacity on low Btu coal gas by
1980. Thus, low Btu gas could account for 2-3% of installed
coal-burning generator capacity by 1980, or more significantly
4-5% of installed coal-burning utility generators feeding
high sulfur Interior coals. These predictions agree
reasonably well with the application of PGG and AG systems
forecast by CEQ.
50
-------
FIGURE 11. MITRE Projection of U.S. Coal Production and Utilization
1800-1
1600
1400
COAL
PRODUCTION,
TREATMENT,
AMD USE
(MILLION TONS
PER YEAR)
100 SOLVENT REFINED
UNCONTROLLED
HfGH SULFUR
400-
200-
1825 TOTA L U.S. COAL PR OO UCTI ON
280 CONVERTED TO GAS
330 CONVERTED TO OIL
325 HIGH SULFUR COAL BURNED WITH
STACK GAS CLEANING
225 SOLVENT REFINED
ZERO HIGH SULFUR COAL BURNED
WITHOUT POLLUTION CONTROLS
NATURALLY
66P LOW SULFUR
1970
1975
1980
1985
[SOURCE: MITRE ANALYSIS]
-------
Summary
In summary, this review of coal liquefaction, gasification,
and advanced combustion processes indicates that technological
alternatives to stack gas cleaning are just beginning to
enter the pilot plant stage. Years of development remain
before any of these technologies will see widespread commercial
application. Low-Btu gas production for utility boilers using
Lurgi gasifiers and solvent refining of coal appear to be the
first technologies that will be applied. Two coal liquefaction
processes—H-Coal and COED—and high-Btu gas production appear
to lag somewhat behind the PGG and SRC technologies. Many other
CL and AG processes with promising characteristics are only at
the bench-top experimental stage.
All of the processes are expensive and it appears that
electric power production costs will increase significantly
as these new technologies are introduced. Although it is
nearly impossible to generalize, it appears that stack gas
cleaning will be significantly less expensive than the
technological alternatives for many new and at least some existing
power plants. The relative economics of the various
technologies, of course, must be evaluated for the conditions
of a given plant. Among the various alternatives, SRC appears
to provide the least expensive fuel which should meet air
quality standards in most areas. SRC may also enjoy economics
of scale not available to PGG, AG and some CL processes. If,
and there are a number of if's, AG combined-cycle systems can
be successfully developed, they may compete with CL processes.
Environmentally, CL processes take the burden off the
utilities to control air emissions and manage solid wastes
and shift it to the processing plant. But the same ash and
sulfur has to be disposed of at the processing plant. PGG
and AG processes remove the sulfur from the synthetic gas
stream and may, if hot clean-up processes are developed, do this
without loss of thermal efficiency. But disposal of the ash
and sulfur remains at the power plant. Air and water emissions
from CL, PGG, and AG processes are not well characterized yet but
it appears that some new problems may arise through the
chemical processing of coal. Developmental work on advanced
clean-up processes remains underway.
52
-------
Concluding with the forecasts of application, a common
element in all the forecasts is the 5 to 8 year lag between
the commercial availability of SGC and that of the alternative
clean fuel technologies. To the degree that sulfur oxide
emissions must be controlled to meet primary air quality
standards, it is quite clear that SGC is now and will continue
to be for some years the only viable alternative to fuel
switching. A delay in the application of SGC to plants
requiring control until another technological alternative is
available places too much burden on the air quality and too
much reliance on yet-unproved technologies which, if
successfully developed, many cost significantly more.
53
-------
References
1. Department of Interior, "Environmental Statement
for the Proposed Prototype Oil Shale Leasing Program,"
Volume II, unpublished [March, 1973].
2. Hottel, H.C., and J.B. Howard, Kew Energy Teclmolog-y—
Some Facts and Assessments, MIT Press, Cambridge [1971].
3. Mills, G.N., and H. Perry, "Fossil Fuel > Power +
Pollution," Chemtech 53-63 [Jan. 1973].
4. Hittman Associates, Inc., "Assessment of S02 Control
Alternatives and Implementation Patterns For the
Electric Utility Industry," Contract for Office of
Science and Technology [March 1973].
5. Chem Systems, Inc., "Desulfurization Strategy Options
Including Fossil Fuel Supply Alternatives For the
Electric Power Industry," Contract for Office of Science
and Technology [March 1973].
6. "Big SRC Pilot Unit Under Construction." Oil and Gas
Journal [Nov. 6, 1972].
7. Akhtar, S., S. Friedman, and P.M. Yavorsky, "Process for
Hydrodesulfurization of Coal in a Turbulent Flow Fixed-
Bed Reactor," American Institute of Chemical Engineers
national Meeting, Dallas [Feb. 20-23, 1972],
8. Yavorsky, P.M., S. Akhtar, and S. Friedman, "Process
Developments: Fixed-Bed Catalysis of Coal to Fuel Oil,"
American Institute of Chemical Engineers Annual
Meeting, New York [Nov. 26-30, 1972].
9. Mills, G. A., "Gas from Coal: Fuel of the Future,"
Environmental Science and Technology 5:12, 1178-1183
[Dec. 1971].
10. Squires, A.M., "Capturing Sulfur During Combustion,"
Technology Review, 52-59 [Dec. 1971].
11. Transwestern Coal Gasification Company et al, "Detailed
Environmental Analysis Concerning a Proposed Coal
Gasification Plant and the Expansion of a Strip Mine
Operation Near Burnham, New Mexico," submitted to Federal
Power Commission [Feb. 1, 19731.
54
-------
12. Agosta, J., et al., "Status of Low Btu Gas as a
Strategy For Power Station Emission Control,"
American Institute of Chemical Engineers Annual
Meeting, New York [Nov.1972].
13. Matthews, C.W., "A Design Basis for Utility Gas
from Coal," Third International Conference on
Fluidized Bed Combustion, Hueston Woods State Park,
Ohio [Oct. 30-Nov.l, 1972].
14. El Paso Natural Gas Company, "Report on Environmental
Factors Burham Coal Gasification Project," submitted
to Federal Power Commission [Nov.1972].
15. Feldman, H.F., J.A. Mima, and P. M. Yavorsky, "Pressurized
Hydrogasification of Raw Coal in a Dilute-Phase
Reactor," American Chemical Society, Dallas
[April 8-13, 1973].
16. Cover, A.E., W.C. Schreiner, and G. T. Skaperdas,
"The Kellogg Coal Gasification Process: Single Vessel
Operation," Synthetic Gas Production Symposium,
New York [Nov. 1972].
17. Bituminous Coal Research, Inc., "Economics of
Generating Clean Fuel Gas from Coal Using an Air-
Blown 2-Stage Gasifier," Office of Coal Research
R&D Report No. 20 [Aug. 5, 1971].
18. Council on Environmental Quality, "Technological
Alternatives for Using Domestic Coal Resources,"
Appendix A, The Potential for Energy Conservation;
Substitution for Scarce Fuels, Office of Emergency
Preparedness [Jan. 1973] .
55
-------
19. American Lurgi Corporation, "Clean Fuel From Gas,"
Lurgi Quick Information, New York, Publication
No. 01007/10.7.
20. Matthews, C.W., "Coal Gasification for Electrical
Power," Institute of Gas Technology, presented at
American Power Conference, Chicago [April 18-20, 1972]
21. Advanced Technology Corporation, "SO2 Free 2-Stage
Coal Combustion Process," prepared for Environmental
Protection Agency [Aug. 1972].
22. Mehta, D.C., and B.L. Crynes, "How Coal Gasification
Common Base Costs Compare," Oil and Gas Journal
[Feb. 5, 1973].
56
-------
STATUS OF FLUE GAS DESULFURIZATION TECHNOLOGY
by
R. E. Harrington, Director
Air Pollution Control Division
Office of Research and Development
Environmental Protection Agency
Washington, D.C.
57
-------
THE STATUS OF DESULFURIZATION TECHNOLOGY
Implicit in our efforts to assess the status of sulfur oxide
control technology is the assumption that there is either a well-defined
measure of the problem that the technology is designed to address or
that there is a well-defined market for the technology, or both. This
paper deals largely with the factors that must be considered in the
development of a yardstick for measuring the status of sulfur oxide
control technology. ,
In this conference, several papers will be given which will
specifically address the technical, economic and regulatory factors
relating to specific S02 control processes. It is the purpose of
this paper to help provide a yardstick against which you can better
judge for yourself the status of each process and the overall status
of the technology.
To help put the sulfur oxide problem in perspective in the United
States, let us first compare the mass emissions of the five principal
pollutants. This is done in Figure 1. Using 1970 mass emission data,
the pie-chart on the left shows the total mass emission of the five
major ambient air quality pollutants: SOX, NOX, Particulate, Hydro-
carbons, and Carbon Monoxide. For comparison, the pie-chart on the
right replots the 1970 mass emission data but excludes emissions from
transportation sources. The chart on the right, therefore, shows the
distribution of these five pollutants which are emitted from stationary
and some natural sources. Comparing the mass emissions, we see that
stationary sources in the United States account for most of the sulfur
58
-------
oxide and particulate emissions. They account for roughly half of the
nitrogen oxide and hydrocarbons and about 1/5 of the carbon monoxide.
Mass emissions, however, do not provide an adequate index of the
impact of a source on the ambient air quality. Other major factors
influencing the impact of pollutants on ambient air quality include:
1. The geographical point of discharge
2. The height of the source above the receptors
3. Topography surrounding the discharge point
4. Proximity of source to other sources
5. Other factors affecting distribution such as
a. Wind Direction
b. Wind Speed
c. Ambient Temperature
To provide a measure of the relative impact of sources on ambient
air quality, we have developed and applied a dispersion model which
takes into account most of the factors mentioned above. This model
assumes a rectangular dispersion of pollutants downstream from the
source and considers the relative location of sources, height of dis-
charge, wind direction and velocity. Using the model, it is possible
to predict the ambient exposure of a receptor in the impacted area.
Using this model, we have calculated the average impact of various
kinds of emitting sources. Figure 2 shows the relationship between the
mass emissions and their impact on ambient air quality. In this figure
we have classified sources into four major categories:
59
-------
PP--Power Plants
IC--Industrial Combustion
IP--Industrial Processes
AS-Area Sources
Measured in terms of tons of pollutants emitted, the fossil fuel
fired power plant accounts for over 50 percent of S02 emissions,
industrial combustion accounts for 24 percent of the S02 emissions,
10 percent are from industrial processes (such as sulfuric acid plants,
pulp and paper plants, refineries and smelters) and 12 percent from
area sources (residential and commercial).
When the factors affecting ambient air quality other than mass
emissions are applied (location, height, proximity to other sources,
etc.), the impact of the sources are as reflected in the bar-graph on
the right of Figure 2. Industrial combustion is seen to be by far the
most important single source accounting for about 40 percent of the
ambient air quality impact. Power plants and area sources each account
for about 25 percent of the ambient air quality impact while industrial
processes continues to be the smallest impactor accounting for only
about 14 percent.
The data depicted in Figure 2 are a composite of 8 U. S. industrial
cities. The data represent calculated annual averages which have been
verified against measured annual averages. The verification or corre-
lation with measured annual averages is quite good.
Combustion Sources
Since combustion sources represent the principal source of sulfur
dioxide impacting ambient air quality, it is useful to look more
60
-------
closely at this problem. Industrial combustion sources are characterized
by being a large number of individual sources. They span a relatively
large size range and relatively large range of different types of
combustion from small gas burning systems to large residual oil and
coal burning units. In general, they are further characterized as
being too small to permit the economic use of desulfurization processes
developed for larger power plants.
One obvious possible solution to this segment of the problem is to
use low sulfur fuel thereby obviating the sulfur oxide emissions.
Figure 3 shows the relationship between the demand for fossil fuels
by the utility industry in the United States over the next 16 years.
Also shown is the source of supply of clean fuels supplemented by flue
gas cleaning control technology required to achieve Ambient Air Quality
Standards.
The principal observations that can be made from this figure are
as follows:
1. Naturally occurring low sulfur fuel located in areas where
it can be reasonably and economically used is inadequate.
2. Cleaned fuels will not make a significant impact on the
supply pattern until the latter half of this decade.
If these projections are correct, they lead to several important
conclusions. First, clean fuels is not the answer—at least not for
the next 10 years. Second, the shifting of the limited supply of low
sulfur fuels from utilities to industrial combustion will help in
certain cases, but it is not likely to be a broadly applicable solution.
61
-------
A key factor limiting the fuel switching approach will be the present
ownership and the existence of extant contract arrangements surrounding
these low sulfur fuels. Much of the low sulfur coals are captive to
industrial processes and will probably not be dedicated to electric
power generation. Many utilities have long-term contracts for low
sulfur fuel and will understandably not be willing to give up their
solution to the air pollution problem to make their low sulfur fuels
available to other fuel users. Third, flue gas cleaning will play a
principal role in the near future control of sulfur oxides. Flue gas
cleaning will undoubtedly be the work-horse of the industry for at
least the next ten years.
What Level of Control is Needed
So far, we have considered the average pollution problem—the
average source category and the average solution. By considering only
the average pollution levels in industrial cities, we can conclude that
average reductions in pollutants of 10, 30, or 50 percent can result in
meeting Ambient Air Quality Standards.
How meaningful, however, are the averages? For the individual
source that must control its emissions to meet regulations, is it
possible to control to the average level of reduction needed by the
geographical area in which the source is located and meet his emission
control obligations? Let's consider this question further.
62
-------
In order to get an estimate of the level of control that a given
individual source would have to achieve, the dispersion model described
above was again exercised. The objective was to estimate the control
capability that a specific control system would have to be able to
achieve when installed on a single source to permit that source to
meet applicable regulations. This, after all, is the real question
that must be answered by the regulatory agency, the polluter and the
research and development community charged with developing control
technology to meet the regulatory goals.
Three cities, New York, Philadelphia, and the Niagara Frontier
were selected largely because of the availability of data. In each
city, the ten most seriously impacted receptor points based on ambient
air quality annual averages were used.
For simplicity, the assumption was made that the contributions of
electric power generating plants would be eliminated (that is 100
percent control of power plants could be achieved—this permits the most
conservative control estimate for other sources). The level of control
that would be required by other impacting sources to achieve ambient
air quality standards were then calculated. In each case, it was found
that greater than 90 percent control would be required for many of •
principal contributing sources. The calculations did not take into
consideration further industrial growth in the areas studied.
While it might be argued that these three cities are not repre-
sentative of all U. S. industrial cities, it seems almost certain that
similar situations will be found in most other industrial cities
63
-------
in which the impact of one or more sources will be such as to contribute
to "hot spots" within that area.
The results of this study serve to establish a reliable estimate
of the control capability required of control processes and equipment
to deal with specific source problems. As a limit, flue gas desulfuri-
zation processes and other methods of control must be capable of at
least 90 percent removal.
This is not to say that all processes must be capable of 90 percent
S02 control. In many cases, lower levels of control will be adequate to
meet standards. In marginal cases, processes such as the dry limestone
injection process which appears to be capable of only 25 percent SO^
control may be adequate to permit compliance. Each case must be evalu-
ated in terms of its own unique needs and economics.
Summary
What js^ the status of sulfur oxide control technology? Many factors
enter into the answer. This paper deals with a few of these factors
that relate to the current definition of the problem and the current
perspective of the need for flue gas desulfurization. In summary,
these are as follows:
1. Low sulfur fuel will be a major tool in achieving ambient
air quality standards over the next few years.
2. Low sulfur fuel, however, falls far short of being a
sufficient solution for SO control.
-------
3. Fuel substitution can and will be used to a limited extent.
Major problems exist of a proprietary, legal and logistical
nature that will preclude our approaching the full theoretical
potential of fuel substitution.
4. Fuel cleaning technology is at least ten years away from
making a major impact on the increased availability of
clean fuel.
5. Flue gas cleaning technology is the principal tool available
to the Nation for SOz control for at least the next 10 years.
6. Industrial combustion is the No. 1 problem source of S02 in
the United States needing solution.
7. Considerable effort must be made to extend the application
of processes which have been designed for utility application
to industrial combustion control. In addition, increased
attention must be directed toward developing processes uniquely
applicable to small industrial combustion, industrial processes
and area sources.
65
-------
TOTAL EMISSIONS,
267 X 106 tons
PARTI CU LATE
23X70*10*I26*}*
STATIONARY SOURCE EMISSIONS,
122 X 106 tons
PARTICIPATE
25 X 106 tons
20%
Figure 1. Emission relationships, 1970.
66
-------
POWER
PLANTS, 53%
INDUSTRIAL
COMBUSTION, 24%
INDUSTRIAL
PROCESSES, 10%
AREA
SOURCES, 12%
TOTAL MASS
EMISSIONS
POWER
PLANTS, 22%
INDUSTRIAL
COMBUSTION, 40%
INDUSTRIAL
PROCESSES, 14%
AREA
SOURCES, 24%
DISTRIBUTION IN
AMBIENT AIR
Figure 2. Distribution of SOX emissions in ambient air.
67
-------
FUEL FOR
FLUE-GAS
CLEANING
NATURAL
LOW-SULFUR
FUEL
cc.
UJ
1990
Figure 3. Power demand and utility fossil fuels required to achieve ambient air quality standards.
68
-------
STATUS OF JAPANESE FLUE GAS DESULFURIZATION TECHNOLOGY
by
Jumpei Ando
Faculty of Science and Engineering
Chuo University
Kasuga, Bunkyo-ku, Tokyo, Japan
69
-------
Status of Japanese Flue Gas Desulfurization Technology
Jumpei Ando
Per-capita energy consumption in Japan is now about one-third that in
the U.S.A. but consumption per acre of level land is eight times the
U.S. figure resulting in serious environmental problems. Stringent
regulation of SO,, emissions has prompted the development of various
desulfurization processes. More than 100 commercial and prototype
plants for waste-gas desulfurization are on stream now. The present
paper will chiefly describe seven of the processes which seem to be of
more interest for possible application in the U.S.A.
1 Outline of SO- control activity
Eecently, about 70$ of energy supply in Japan has depended upon imported
crude oil. Heavy oil, which is a residue from atmospheric distillation
of crude and contains 2-4$ sulfur, has been a major fuel for power
generation as well as for other industrial activities. Hydrodesulfurization
of heavy oil has been carried out since 196?• In 1971i nearly 200 million
barrels of heavy oil, about one-fourth of the total quantity of heavy oil
consumed in Japan, was subjected to hydrodesulfurization giving 287,000 tons
of elemental sulfur as by-product. Still nearly 3 nillion tons of sulfur
in heavy oil burned produced nearly 6 million tons of S0?, constituting
the chief source of SO,, emissions. Today's stringent controls on SO,,
emissions have renderea the desulfurized heavy oil, with its 1$ or higher
sulfur content, unsatisfactory for large power plants.
Under such situation, flue-gas desulfurization has assumed greater
importance, which fact has led to construction of more than 100 commercial
and prototype plants. Most of the plants built so far are of relatively
small capacity, designed to treat waste gas from industrial boilers,
Chemical and smelting plants, etc. Some of the larger ones are listed
in Table 1. Major plants under construction or being designed are listed
in Table 2. Nine maj*or power companies, supplying more than JCffi of the
total electric power needs in Japan, were originally interested in dry
processes but have recently decided to build many plants using wet processes
because wet processes with reheating of the treated gas have proved to
"be less expensive than dry processes. With some of the wet processes,
moreover, trouble-free continuous operation has been demonstrated and
wastewater has been eliminated. The total capacity of the desulfurization
plants of the major power companies will increase from 375MW in 1972 to
2,700MW in 1974 and to 4,800MW in 1976.
Most of the desulfurization plants have produced salable by-products such
as sodium sulfite, sulfuric acid, and gypsum, because Japan is subject
to limitations in domestic supply of sulfur and its compounds as well as
in land space available for disposal of useless by-products. As the
desulfurization projects are making very rapid progress, however, it is
likely that a considerable oversupply will occur in future necessitating
abandonment of a substantial portion of the by-products.
70
-------
Table 1 Major waste-gas desulfurization plants in operation
Process developer
Kureha Chemical
Showa Denko
Wellman-MKK
Vellman-SCEC
Mitsubishi-JECCO
Mitsubishi-JECCO
Chiyoda
Chiyoda
Nippon Kokan
Onahama Smelting
Chemico-Mitsui
Sumitomo S.M.
Mitsubishi H.I.
Hitachi Ltd.
Absorbent
NaOH
NaOH
NaOH
NaOH
Ca(OH)2
Ca(OH)2
H-SO., CaCO.-
H-SO,, CaCO-j
24 3
NH , Ca(OH)2
MgO
Ca(OH)2
Carbon
MnOx, HE,
Carbon, CaCO,
Product
Na2SO
Na2SO
H2S04
so2, s
GypBum
Gypsum
Gypsum
Gypsum
Gypsum
H S04
CaSO.,
H2S°4
I "MTT A O/~\
V liil t I f^ &\J .
Gypsum
User
Kureha Chemical
Ajinomoto
Japan Synth. Rub.
Toa Nenryo
Kansai Electric
Onahama Smelting
Fuji Kosan
Mitsubishi Rayon
Nippon Kokan
Onahama Smelting
Mitsui Aluminum
Kansai Electric
Chubu Electric
Tokyo Electric
Plant site
Nishiki
Kawasaki
Chiba
Negishi
Amagasaki
Onahama
Hainan
Otake
Keihin
Onahama
Omuta
Sakai
Yokkaichi
Kashima
Unit capacity
(l.OOOscfm)
176a, 176a
159a
118a
Tj
35
59a
54°
93a
53a
.a
88d
53°
226e
iooa
193a
250a
Date of
completion
1968
1971
1971
1971
1972
1972
1972
1973
1972
1972
1972
1971
1972
1972
a:" Oil-fired boiler
d: Sintering plant
b: Glaus furnace 01 Smelting furnace
e: Coal-fired boiler
-------
Table 2
Process developer
Vellman-MKK
Wellmari-MKX
Wellman-SCEC
Showa Denko
Shell
Mitsubishi-JECCO
Mitsubishi-JECCO
Mitsubishi-JECCO
Kureha-Kavasaki
Rureha-Kawasaki
Babcock-Hitachl
Chiyoda
Chiyoda
Absorbent
NaOH
NaOH
HaOH
NaOH, CaCO,
CuO
Ca(OH)2
Ca(OH)2
CaCO,
NaOH, CaC03
NaOH, CaCO'
CaCO
,
3
, CaCO,
Major flue-gas desulfurization plants under construction
or being designed (oil-fired boilers)
Product
H2S04
Gypsum
so2, s
Gypsum
Gypsum
Gypsum
Gypsum
Gypsum
Gypsum
Gypsum
Gypsum
User
Chubu Electric
Japan Synth. Rub.
Sumitomo C.C.
Showa Denko
Showa Y.S.
Tohoku Electric
Kansai Electric
Tokyo Electric
Shikoku Electric
Tohoku Electric
Chugoku Electric
Eokuriku Electric
Mitsubishi P.O.
Plant site
Nishinagoya
Yokkaichi
Chiba
Chiba
Yokkaichi
Hachinoe
Kainan
Yokosuka
Shintokufchiffl
Sendai
Mizushima
Shinminato
Yokkaichi
Capacity
(MW)
220
160
125
200
42
125
150, 130
150
a 150
130
100
250
240
Date of
completion
1973
1974
1973
1973
1973
1974
1974
1974
1974
1974
1974
1974
1974
-------
2 MItsubishi-JECCO lime (limestone)-gypsum process1*2)
Developer Mitsubishi Heavy Industries
5-1t 2-chome, Marunouchi, Chiyoda-ku, Tokyo
Japan Engineering Consulting Co.
1-4» Ogawamachi, Kanda, Chiyoda-ku, Tokyo
Rrocess description Waste gas is first washed with water for dust removal
and cooling to about 140°]?. As the water becomes acidic and dissolves
metallic components of dust, it is neutralized with milk of lime to
precipitate metallic ions, which are filtered off together with the dust.
The filtrate is used for slaking of lime. The cooled gas is then sent to
an absorbing step. At three plants built recently two plastic-grid packed
absorbers in series, which are put together in one tower, are used as
shown in Figure 1. (For new -plants which are being
designed a one-absorber system will be used.} Milk of lime is fed to the
No. 2 absorber. The gas is introduced into the tfo. 1 absorber and then into
the ITo. 2 absorber. The slurry discharged from the No. 2 absorber is
a mixture of calcium sulfite and unreacted lime with a small amount of
gypsum. The slurry is then led to the Ho. 1 absorber, where the remaining
lime is reacted to fom calcium sulfite, a portion of the sulfite is
converted to bisulfite. The pH of the slurry discharged from the Ho. 1
absorber is 4-4»5- The concentration of the slurries in the absorbers
Is about 15$. A relatively large liquid/gas ratio (20-50 gal/1,OOOscf)
is used to prevent scaling.
The pH of the slurry is then adjusted to about 4 to promote oxidation in
the following step. If required, a small amount of sulfuric acid,
normally less than one ton per 100 tons of inlet SO-, is added to the
slurry for the adjustment. The slurry is then sent to an oxidizing tower
where the sulfite and bisulfite are converted to gypsum by air oxidation
using rotary atomizers invented by Japan Engineering Consulting Co.(jECCO)
at a pressure of 50-57psig and a temperature of 120-180°F. The atomizer
is quite effective in producing fine bubbles and is free from scaling,
erosion and corrosion. The gas leaving the oxidizer contains some SO-,
and is returned to the absorber. The gypsum is centrifuged. All of
the liquor and wash water are used for the gas washing and cooling step.
The gypsum grows into large crystals; its moisture content after ce'ntrifu-
gation is only 8-10$. The gypsum thus obtained is of high purity and good
quality, which make it suitable for use in cement and gypsum board. The
gas from the No. 2 absorber is passed through a demister, reheated, and
led to a stack. Wash water of the mist eliminator is also used in the
system. Normally no wastewater is emitted from the system. More than
of SOg is recovered.
73
-------
of d'-velc^-ient Four lants are in operation, and five others
TJser
Nippon Kokan
Kansai Electric
Onahama Smelting
Tomakomai Chemical
Kawasaki Steel
Tokyo Electric
Kansai Electric
Tohoku Electric
Kansai Electric
PI -.r.t site
Koyasu
Amagasaki
Onahama
Tomakomai
Chiba
Yokosuka
Hainan
Hachinoe
Hainan
Capacity
fscfn)
37,000
59,000
54,000
35,000
71,000
235,000
235,000
224,000
221,000
Number of
A.b s orb ont ab z o rb o r z
Ca(OH)2
Ca(OH)2
Ca(OE)2
Ca(OH)2
Ca(OH)2
CaCO
Ca(OE)2
Ca(OH)2
Ca(OH)2
2
2
2
2
1
1
1
1
1
are being constructed or designed as shown ir. the following table:
Year of
completion
1964
1972
1972
1972
1973
1974
1974
1974
1974
Status _of techr.olcp' Based on extensive studies with a pilot plant,
Mitsubishi has succeeded in scale prevention. Scaling can be prevented
by the use of a suitable material, shape, and arrangement of the grid
in the absorber, by the adjustment of the slurry concentration and pH
as well as of the liquid/gas ratio, by the addition of gypsum crystal
seed and thorough mixing of lime and the circulating slurry.
The Amagasaki plant has been in continuous operation since its start in
April 1972 except for the period of shutdown of the power plant. The
desulfurization plant treats a fraction of flue gas 86,000scfm from a
1JJ6MW boiler'containing about 700ppn S0? to recover about JOJj of the SO-.
The gas velocity in the absorber is about 11 feet/sec. The pressure
drop in the whole system including the cooler, absorbers and demister is
6 in.HpO. More than "95/a of calcium sulfite is oxidized into gypsum in
the absorbers due to the low S0« concentration; the oxidizing tower is
almost unnecessary. The amount of water added to the system is maintained
equal to that removed from the 'system by evaporation in the cooler, by
hydration of gypsum, etc. Ho water is wasted from the plant.
The Onahama plant treats 54,000scfm of gas from a copper smelter containing
20,000-25,000ppm S02. More than 99.5$ of the SO- is recovered with a
stoichiometric amount of lime by feeding milk of lime mainly to the No. 2
and partly to the No. 1 absorber. The SO^ content of the outlet gas is
less than 50ppm. The plant came on-stream at the end of October 1972
and has been in continuous operation without trouble except for a period
of scheduled shutdown for inspection at the end of November, ifo scaling
was observed at the inspection. The gas supplied from tjtie smelter is a
-------
wet gas at 155°F and results in less evaporation of water in the system.
Therefore, the amount of water fed to the desulfurizaticn plant slightly
exceeds that by evaporation, hydration, etc. A small amount of water
is wasted after being treated for pollution control. About 450 tons/day
gypsum is produced; three oxidizing towers are provided for the oxidation
of calcium sulfite into gypsum; little oxidation occurs in the absorbers
due to the high concentration of SO-.
A one-absorber system will be used for the new plants to save invest-
ment cost. To ensure high SO,, recovery excessive amounts of the
absorbents, about 105;' of stoichiometric for line and about llO^j for
limestone, will be used. For pH adjustment prior to oxidation, a
considerable amount of sulfuric acid will be required to convert the
excessive absorbent to gypsum. Other facilities and treatments are the
same as in the two-absorber system.
In the limestone scrubbing plant in Yokosuka, seawater will be used for
cooling flue gas from an oil-fired boiler. The seawater is gradually
concentrated in the gas-cooler and therefore should be wasted after
"being duly treated. Salable gypsum of high purity and good quality will
"be by-produced using limestone as absorbent.
Economics Investment cost for the Amagasaki plant (59»OQOscfm) was
$1.46 million including various equipment for automatic control and for
tests, while that for Tomakomai Chemical (31,OOOscfm) was $0.32 million.
The cost for larger plants (224,000-235,OOOscfm) is estimated to be
$2.6-2.9 million in battery limits. The desulfurization cost for a
plant to treat 100,000-150,OOOscfm gas from an oil-fired boiler is
estimated at 50.67-0.88/bl of oil containing 2.5-3.0^ sulfur, including
depreciation and credit on gypsum at $6/t.
Advantages High recovery of SO- is attained and good-quality gypsum is
obtained using either lime or limestone without scaling problems. The
jfotary atomizer is quite effective for oxidation, involving no operational
prgblem. No water is wasted when gas at temperatures above 250°F is
treated. Seawater can be used for cooling, although in this case the
used seawater should be discharged.
Disadvantage Lime is more expensive than limestone. Although limestone
is used for the absorbent, an appreciable amount of sulfuric acid is
required to produce gypsum of high purity and good quality. Oversupply
of eypsum might occur within several years if too many gypsum-producing
desulfurization plants are built.
75
-------
Washing
and
cooling
Cooling water
Lime
0>
Stack Cooling
_ water
Demlster Blower I
--Q-Ll
pK adjusting
<-l| JlCentrlfuge
tentrl-
,, *
Soot
Lime
Milk of lime
NeuSrallzer
L
Thick-
ener
Liquor tnak
Figure 1 Flow sheet of Mitsubishi-JECCO lime-gypsum process
Reed
crystal
tank
-------
125)
3 Chemico-Mitsui lime process ' '
Developer Chemico, U.S.A.
Mitsui Aluminum Co. (Miike Power Station), Orauta
Constructor Mitsui Miike Machinery Co. Ltd.
1-1, 2-chome, Muromachi, Kihonbashi, Chuo-ku, Tokyo
Process description Chemico scrubbers (two-stage venturi) are used
for SC>2 and dust removal from flue gas from coal-fired boiler (156MW).
Carbide sludge (primary calcium sulfite) is used as absorbent.
Figure 2 presents a description of the total scrubber system. The
flue gas (302,OOOscfm) after passing through an electrostatic precipitator
contains 0.3 grain/scf of dust and 1,300 to 2,200ppni of S02 at 300°?.
About 75/S of the gas is handled by the scrubber. Two scruobers were
installed but one of them has been used with the other as a back-up. The
gas flows down through the first venturi section, up through the mist
eliminator section, then passes through the second venturi and mist eliminates
sections, is reheated and exhausted to a stack along with the unscrubbed
fraction of the gas. Milk of lime is mixed with the discharge from the
second venturi; the mixed slurry is partly recycled to the second venturi
and partly fed into a delay tank. The slurry in the delay tank is sent
to the first-stage venturi. The discharge from the first venturi,
consisting mainly of calcium sulfite with small amounts of calcium sulfate,
unreacted calcium hydroxide and-fly ash, is sent partly to a delay tank
and partly to a disposal pond about a nile from the plant. The decanting
or settling of the solids takes place there and the supernatant from the
pond is recycled to the scrubbing system to prepare milk of lime and also
to wash mist eliminators. The outlet gas frcia the scrubber contains
Otl grain/scf of dust and 200 to 300ppm of SO-.
Status of technology The scrubber is 33 feet in diameter and 66 feet high,
is constructed of stainless steel, and is lined with glass flake reinforced
polyester material. The following operation conditions have been'used:
L/G (venturi + spray) 1st stage, 46 to 59 gal/1,OOOscf
2nd stage, 42 to 55 gal/1,OOOscf
Stoichiometry 100 to 105$ aa pure Ca(OH)2
Percentage solids in slurry 3-5$
Total pressure drop 16 in.E-O
Prior to the completion of the plant, extensive pilot plant tests were
carried out by Mitsui Aluminum Co. leading to the establishment of
operation know-how for scale prevention. Precise control of pH to a
certain narrow range is important for the scale prevention.
State of development Operation of the commercial plant started in March
1972. After 8 months of satisfactory continuous operation, the plant
vas subjected to a scheduled shutdown for inspection, wliich revealed
essentially no scaling. Operation was resumed soon and has since been
77
-------
carried out smoothly. The waste disposal pond has a capacity for
holding solids discharged over a period of several years. However, to
eliminate any possible pollution with eulfite ion in wastewater which could
"be emitted when the facilities are cleaned in a periodical -shutdown of
the plant for inspection, some means of converting calcium sulfite to gypsum
will be adopted within 1973* Tests have shown that the gypsum can be used
for gypsum board and cement although it contains a snail amount of fly
ash.
Economics The commercial plant cost 33«3 million including two scrubbers.
The desulfurization cost is a litule less than 1 mil/Vah. About 70j£ of
the desulfurization cost is accounted fox by depreciation and interest.
Advantages Stable operation without scale formation is achieved, removing
both SO- and dust satisfactorily. Carbide sludge, a cheap source of line,
is used. The Chemico scrubber is suited for treating a large amount of-
gas.
Disadvantage Large amounts of slurry and water must be recycled because
of the use of high L/G and low slurry concentration in order to ensure
scale prevention.
78
-------
Boiler
flue
VD
Stack
--CD
Wet carbide pit
N j
«D« Fan
n
•*-
A
Reheat D
furnace
eoiste
r
f
Makeup
i slurry
Makeup slurry
feed pump
let stage
recycle*
"" "" ""* — — — pump "~ ""
Q~-
Aah pond liquor t
return pump
^r
Dry carbide pit
slurry
,. —. Makeup slurry
,•__» Bleed slurry
____ Return liquor
>LL
\'=r^~^r^/
Waste disposal pond
Note: Only one scrubber systea presently In operation*
• Indicates a closed valve during 2/72 to 9/72.
M indicates a closed damper during 2/72 to 9/72.
Figure 2 Mttsul-Chemico lime process
-------
Chiyoda dilute sulfuric acid process
(Thoroughbred 101 process)
Developer Chiyoda Cheaical Engineering &. Construction
1580 Tsurumi-cho, Tsurumi-ku, Yokohama
Outline, of the process Flue gas is washed with dilute sulfuric acid
which contains an iron catalyst and is saturated with oxygen. SO-
is absorbed and converted to sulfuric acid. Part of the acid is
reacted with linestone to produce gypsum. The rest is diluted with
gypsum wash water and returned to the absorber.
Description A flow sheet is shown in Figure 3,
Flue gas is first
treated by a prescrubber to eliminate dust and to cool the gas to
l*fO°F. The cooled gas is led into a packed tower absorber containing
1 inch Telleretts. Dilute sulfuric acid (2 to 5$ H-SO.) which contains
ferric ion as a catalyst and is nearly saturated with oxygen, is fed to
the packed tower. About 90$ of S02 is absorbed, and partly oxidized
into sulfuric acid.
The product acid is led to the oxidising tower into which air is bubbled
from the bottom to complete the oxidation. Most of the acid at 120-150*F
nearly saturated with oxygen is returned to the absorber. Part of the
acid is treated with powdered limestone (minus 200 mesh) to produce
gypsum. A special type of crystallizer has been developed to obtain good
crystalline gypsum 100 to 300 microns in size. The gypsum is centrifuged
from the mother liquor and washed with water. The product gypsum is of
good quality and salable.
The mother liquor and wash water are sent to the scrubber. The amount
of input water—wash water and the water to the prescrubber—is kept equal
to the amount of water lost by evaporation in the scrubbers and by
hydration of gypsum. No wastewater is emitted.
State of development The operation of a pilot plant (6,000scfm) has led
to the construction of the following commercial plants:
Commercial plants by Chiyoda process
User
Plant site
Mizushima
Kainan
Nippon Mining
Fuji Kbsan
Mitsubishi Rayon Otake
Tohoku Oil Sendai
Daicel Ltd. Aboshi
Source of gas
Glaus furnace
Oil-fired boiler
Oil-fired boiler
Glaus furnace
Oil-fired boiler
Size, scfm
20,600
94,100
52,900
8,200
59,000
Hokuriku Electric Shinminato Oil-fired boiler 442,000
Mitsubishi Yokkaichi Oil-fired boiler 413,000
Petrochem.
Completion
October 1972
October 1972
December 1973
January 1973
October 1973
June 1974
December 1974
80
-------
Status of technology The iron catalyst is less reactive at. low
temperature but is as reactive as manganese catalyst at operation
temperatures above 120eF (Figure ^0. I* *s no* poisoned by
impurities in the gas, even when flue gas from a coal-fired boiler
is used. Catalyst loss is very small (Figure 5). The towers of
the commercial plants are provided with rubber or FEP linings.
Stainless steel is also usable; the ferric catalyst works also as a
corrosion inhibitor.
A large L/G ratio is required to attain high S0« recovery as shown in
in Figure 6 ; large pumps and fairly large absorber and oxidizer are
required as shown in the following table.
Size of towers (feet)
Ab so rb_er_ Oxidizer
Ca-pacity fscfm) Diameter Height Dianeter Eeight
117,600 29.7 49-5 13.4 62.7
294,100 49-5 49-5 21.1 62.7
A double-cylinder type reactor (Figure 7) including an oxidizing
section in the center and a scrubbing section in the outer part has
been developed recently instead of using two towers. The absorbing
liquor goes down the scrubbing section, then goes up in the oxidizing
section and overflows to the scrubbing section. The reactor enables
some savings to be made in floor space and investment cost.
Advantages The process is simple and the plant is easy to operate.
Even in the event that the gypsum-producing system has to be stopped
for a day or two for repairs, the absorbing system can be operated
continuously. The concentration of sulfuric acid increases by 1 or
Zfo in this case but S0« recovery is not decreased. Catalyst is cheap
and is not poisoned by impurities in the gas. Salable gypsum of good*
quality is obtained from limestone without scaling problems.
Disadvantage Large pumps and a fairly large scrubber and oxidizer are
required. A large L/G is required when SO- concentration of inlet gas
Is high.
81
-------
Cost estimation* ($1 = ?308)
(inlet gas S02 l,000ppm, dust 0.08 grain/scf)
(Outlet gas SO. lOOppin, dust 0.04 grain/scf)
Capacity
Plant cost (3) (A)
Fixed cost (S/year) (B)
Direct cost (S/year)
Limestone (0.250/lb)
Electricity (0.70/kUh)
Water (80/1,000 gal.)
Fuel oil ($3.18/bl)
Steam (0.120/lb)
Catalyst (60/lb)
Labor (312,000/year/capita)
Kaintenance
Subtotal (C)
Net running cost (3 + C = D)
Overhead (s) (12JS of C)
Sunning cost (D 4- E)
Desulfurization cost f'($/bl)
without by-product $/MWhi
credit v, N '
800IW
(1.51 x lO^scfn1)
17.2 x 106
3.10 x io6
430,000
789,000
25,400
1,035,400
39,700
11,000
144,000
344,000
2,818,500
5,918,500
338,200
6,256,700
0.645
•) 0.992
500MW
(0.95 x lO^scfn)
12.5 x IO6
2.25 x IO6
272,000
529,000
17,000
648,700
25,100
8,000
96,000
250,000
1,845,800
4,095,800
221,500
4,317,300
0.712
1.095
250M&T
(0.475 x 106scfn)
7.0 x 10°
1.26 x IO6
135,600
308,000
8,400
325,600
12,600
4,000
96,000
140,000
1,030,200
2,290,200
123,600
2,413,800
0.797
1.225
* Estimation made by Chlyoda in February 1973
82
-------
I
Water dJU Repeater
Prescrubber
CO
Dust
eliminator
Waste gas
Limestone Water
Dilute
""1
4
Air
O
Water
Absorber Oxidlzer
Centrifuge
V
V/ater
^^.u—^V^W ^ -C*^
L-o
Crystallizer
Gypsum
Figure 3 Flow sheet of Chiyoda Thoroughbred 101 process
-------
-P
03
C
O
r4
•P
03
•o
«H
K
O
80
60
40
20
Figure
Fe
•H-f-
60 80 100 120 140
Temperature (°F)
Temperature and oxidation ratio
with catalysts
1.0
ra
*>
0>
>»
0.6
1,000 2,000 3,000
S(>2 content (ppm)
Figure 5 Catalyst requirement
-------
?
o
§
CM
O
03
100
80
60
20 .
-X Physical-chemical absorption
(Chiyoda)
Physical absorption
Flue gas containing l,300ppm S02
H 1 1 1 1 1 1 1 1 H
200 400 600 800 1,000
Liquid/gas ratio ( gal./lfOOOscf )
Figure 6 Liquid/gas ratio and S0£ removal
T Gas outlet
Liquor
distri-
buter
Liquor outlet
Liquor
outlet
Liquor
inlet
Figure 7 Double-cylinder type reactor
85
-------
1 2)
5 Weliman-KSZ sodium process *
Developer Well man-Lord, U.S.A.
Mitsubishi Chemical Machinery (MK£)
6-2, 2-chorne, Marunouchi, Chiyoda-ku, Tokyo
Process description Flue gas is first washed by a prescrubber installed
in the lower part of the sieve tray absorbtion tower (pigure 8).
The partially cleaned gas rises in the tower while contacting a
countercurrent flow of a concentrated sodium sulfite solution which
eliminates nore than 9C$ of the inlet S09 to fora sodium bisulfite.
Hist droplets are removed by the elininator and deoister combination
in the upper part of the tower and the gas is discharged at about 2650?
after reheating in the oil-fired after-burner. A sodium bisulfite
solution is discharged from the absorption tower and is stored in a
surge tank before it is sumped to an evaporator. In the evaporator
the sodium sulfite solution- is heated with steam and decomposed into
S0« gas and sodium sulfite.
2 Na2SO, + S02 -
The S00 concentration in the gas leaving the evaporator is 90$ after
the water vapor is condensed in a cooler. In this evaporator, sodium
sulfite is gradually concentrated and crystallized. The crystalline
sodium sulfite is centrifugally separated from the nother liquor,
dissolved in a condensate from the cooler, and the solution is recycled
to "be used as absorbent. The recovered S02 is used for sulfuric acid
production. Tail gas from the acid plant is led into the absorption tower.
Sodium sulfite is gradually oxidized into sulfate by the oxygen in the
flue gas. To keep ITa^SO. concentration to a minimum and optimum figure,
a. small portion of the mcfther liquor is purged; this purge liquor is
used for gas cleaning in the prescrubber. The bleed is taken off this
prescrubber circuit for the purpose of removing contaminants, which
otherwise would build up in the system. The major contaminants are
sodium sulfate and sodium polythionate. The purge otreaa is subjected
to wastewater treatment that involves the following: (l) addition of
HpSO. to convert NaSO- and ITaHSO, to Ka-SO,; SO. evolved is sent to
tfie sulfuric acid plant; (2) alkali is added to^form the hydroxide of
soluble metal ions (vanadium, nickel, iron); this precipitates then as
hydroxides; (j) removal of solids by filtering; (4) neutralization by
adding H2SO^. The final wastewater is a largely clear concentrated
solution of sodium sulfate. This effluent eventually is sent to the bay.
State of development A commercial plant designed to treat 118,OCOscfm
of flue gas front oil-fired toilers has been in operation since June
1971 at Chlba plant, Japan Synthetic Rubber Co. A larger plant with a
capacity of treating 365tOOOscfm of flue gas from oil-fired boilers is
under construction at Nishinagoya Station, Chubu Electric Power Co. to
start operation in July 1973. It has been recently decided to build
& 237»000sbfa plant (flue gas from oil-fired boiler) for Japan Synthetic
Rubber Co. at Yokkaichi.
Status of technology In Chiba plant, Japan Synthetic Rubber, flue gag
from two oil-fired boilers (1JO tona/hr each) is treated by two 16-ft
86
-------
square sieve tray absorption towers at a rate of 90,000scfa per absorber.
The plant came on-streara in June 1971 a^-cL has been operated for more than
8,000 hours in a year. The S00 concentration in inlet gas normally
ranges from 1,000 to 2,OOOppm and that in outlet gas from 100 to 200ppm.
The major problem associated with the process is the necessity to bleed
a waste stream from the absorber licuor circuit to avoid build-up of
contaminants, primarily sodium sulfate. The following tabulation
summarizes the present composition of absorber and waste streams:
Absorber feed Absorber Cut Vastewater to Wastewater after
treatment treatment
Ka2SO- 16-19$ by wt
Ka2S2°5a °
Fa2S04 5-7$
Suspended
solids
pH
COD
Plow rate
2-4$ 0
14-17$ 4-6$
5-7$ 3-6$
1-2$
(10, 000-20, OOOppm)
5-5-5
20, OOOppm
•» MM
0
0
7-16$
2-10ppm
7+0. 1
200ppia
1-1.5 tons/or
r
Clear sodium sulfate solution is emitted from Chiba plant. In new plants
to be built in future, the sodium sulfate solution will be evaporated to
produce solid sodiun sulfate as a by-produc'i; or treated with lime to
precipitate gyp.sua and to recover a sodium hydroxide solution, thus
eliminating the wastewater.
Economics The investment cost of the Chiba plant was $2,600,000 including
the cost for sulfuric acid plant. The plant consumes approximately
755,000bbl/year of fuel and produces 13,200 tons/year of aulfuric acid.
The requirements of the desulfurization plant are shown belcw.
Make-up caustic soda 3-4 Ib/bl oil for 86-93$ recovery of SO,,
After-burner fuel 10.5 Ib/bl oil to heat to 266°F
Steam 175 Ib/bl oil
Cooling water 6.5 tons/hi oil
The investment cost for Nishinagoya plant, Chubu Electric (365,000ccfn)
is $5,200,000 including the sulfuric acid plant with a capacity of 90
tons/day.
Advantages Stable reliable operation. High recovery of S02- The
sulfuric-acid plant is much smaller than usual because the concentrated
SO- recovered is used.
Pi sadvantage The treatment of sodium sulfate formed by the oxidation is
not simple*
87
-------
----- M
H
Boiler
S02 gas -to I
Stack
gas to
Plant
^2^0h plant j, /\ Absorber
1 ] cL j>
Evaporator
Coolin
water
NaQH
Soot
en
CO
Wastewater treatment
section
Absorbing section Recovery section
Figure 8 Wellman-MKK process
-------
6 Kureha sodium-limestone process 1«2)
Developer Kureha Chemical Industry Co.
1-8, Horidonecho t Nihonbashi, Chuo-ku, Tokyo
State of development Kureha first developed a sodium scrubbing process
to produce solid sotiiun s\ilfite to be sold to paper aills. In addition to two
176,000scfm plants operated by Kureha since 19°9» two plants have been
licensed, one to Mitsui Toatsu Chemical (ll2,000scfd) which began
operation in September 1971, and the other to Konan Utility (I23,000scfm)
which started operation in late 1972. Since the demand for sodiun
sulfite is limited, Kureha has recently developed a sodium- calcium
double-alkali process. Tests with a small pilot plant led to the
construction of a larger pilot plant (3fCOOscfm) which has been in
operation since July 1972. The larger pilot plant program is & joint
effort with Kawasaki Heavy Industries. Two commercial plants will be
completed in 1974 to treat flue gas from oil-fired boilers at power
companies, one at Shinsendai station, Tohoku Electric Power (l50MW),and
the other at Shintokushima station, Shikoku Electric Power (150MW).
Process description A flow sheet of the sodium-calcium process is shown
In Figure 9. The scrubbing system consists of a venturi scrubber
vhere water is used to remove particulates and to cool the gas followed
"by a rubber-lined, grid-packed scrubber where SO- is absorbed in a sodiua
sulfite solution. The water from the dust scrubBer is discharged at
a pH of about 2.5. The pH of the liquor from the absorber is controlled
to 6.0-6.5. With an inlet concentration of l,500ppa SO-, 9&fo removal is
achieved; the liquid gas ratio is 7gal/l ,000ft* of gas. The feed to the
absorber contains 20-25^ sodiua sulfite and has a pH of 7-Q» the calcium
content was reported to be about J>Qppm. The scrubber discharge contains
about 10$ sodium sulfite, lOye sodium bisulfite, and 2-5$ sodium sulfate.
Limestone pulverized in a wet aill equipped with a cyclone classifier
Is fed continuously along with scrubber liquor into an atmospheric
pressure vessel where sodium bisulfite reacts with limestone to form
calcium and sodium sulfites.
2KaHSO, + CaCOj— ^ CaSOj'l/fflgO + NagSOj + 1/2H20 + COg
The reaction temperature is somewhat higher than the scrubber temperature
vhich IB about J40°F; residence tine for conversion is about 2 hours.
The slurry from the decomposer is passed through a centrifuge where the
ealciua sulfite crystals are separated from sodium sulfite liquor, which
Is then returned to the scrubber*
89
-------
The calcium sulfite is reacted with air at atmospheric pressure in an
oxidizer developed by Kuraha. Gypsum is removed from the oxidizer
discharge stream by a centrifuge, The product is suitable for use in
wallboard and ceaent.
Oxidation of sulfite in the scrubbing and decomposition steps results
in the formation of sodium sulfate which cannot be regenerated by
reaction with limestone. I;i order to control the eulfate level, a
sidestream fron the scrubber discharge is nixed with calcium sulfite
crystals and sulfuric acid ia added. The net effect is to convert the
sodium sulfate to calciua sulf&te and produce sodium bisulfite for recycle.
Gypsun is separated by a centrifuge and added to the oxidizer loop.
B02 + CaS03-l/2E20
Status of pilot Plant operation The pilot plant (j.OOOscfm) has been
operated continuously since its coaple^ion in July 1972 except for the
scheduled shutdown for inspection in September and December. Almost
no scaling was observed. Operation of the centrifuges has given
effective separation of solid and liquid phases. Both calcium sulfite
and gypsum discharged from the centrifuge are dump solids which can be
transported by solid-handling equipment, if desired. The crystals of
gypsum grow to around 100 microns. The flue gas contains about 6?&
oxygen. About ?# of the recovered S02 is oxidized to form sodiun
Btilfate. The sulfuric acid requirement for the decomposition of the
snlfate is 125$ of the theoretical amount. Therefore, about 8.756 of
the product gypsum is derived from sulfuric acid.
Economics For the production of a ton of gypsum, 1,200 Ib of limestone,
18 Ib of caustic soda (100#) for make-up, 100 Ib of sulfuric acid (9W,
340kWh of electric power, 1,460 Ib of steam, and 18 tons of water are
required. Investment cost is uncertain, but the investment requirement
would be split among the process steps as follows:
Absorption 30#, Decomposition 50#, Sulfate conversion
Oxidation
Advantages High recovery of S02 is achieved with- limestone. No scaling.
Sodium eulfate is decomposed to recover sodium bisulfite and gypsum.
Good quality of salable gypsum is obtained. Both gypsua and calcium
eulfite discharged from the centrifuge hare less moisture and are easy
to handle.
Disadvantages The process is less simple than the lime-gypsum process.
STSTff* considerable amount of sulfuric acid it disadvantageous
for plants where the product calcium sulfite or gypsum must be discarded.
90
-------
Fuel
Air
X
Make-up
NaOH
Limestone
NaHS03
HpSOjj.
^
J Gypsum
(1) Reheater (2) Absorber (3) Mill and classifier
Decomposition tank (5) Centrifuge (6) Sulfate conversion tank
(8) Oxidizer (9) Centrifuge
(?) Air compressor
Figure 9 Plow sheet of Kureha sodium-limestone process
-------
7 Showa Denko aodiun-limeatone process
PeTeloper Showa Denko K.K.
34» Shiba Miyamotocho, Minato-ku, Tokyo
Ebara Manufacturing Co. Ltd.
11-1, Asahiaaehi, Haneda, Ota-ku, Tokyo
State of development Showa Denko, Jointly with Ebara, recently constructed
commercial plants for SO- recovery by sodium scrubbing to produce sodium
sulfite for paper mills. The plants include Kawasaki plant (88,000scfm)
of Showa Denko, Kawasaki plant (I39*000scfm) of Ajinonoto, and Yokohama
plant (I42,000scfm) of Asia Oil for flue gas from oil-fired boilers, and
Sodegaura plant (47,000scfa) of Nippon Phosphoric Acid to treat tail gas
from a sulfuric acid plant. As demand for sodium sulfite is limited,
Showa Denko and Ebara have started Joint tests on sodium-calcium process
to by-produce salable gypsum. A pilot plant (5»900scfm) has been in
operation at Kawasaki plant of Showa Denko since 1971* A commercial
plant to treat 34°»°00scfa of flue gas is being constructed at Chiba plant,
Showa Denko, to start operation in June 1973-
Process description The Showa Denko-Zbara process features the use of
a vertical-cone type absorber as shown in Figures 10 and 12.
A liquid (sodium sulfite solution) is charged from the bottom, blown up
by the gas to absorb SOO, and flows back to the liquor inlet by gravity.
Very good contact between gas and liquid particles is attained ensuring
95-98$ desulfurization at a liquid/gas ratio of 7-14 gal./l,000scf
(Figure 11). Pressure drop ranges from 8 to 15 in.H?0. A flow sheet
of the sodium-calcium process is shown in Figure 12. Flue gas from
an .oil-fired boiler containing l,500ppm SO- and about 1 grain/of dust is
led directly into the scrubber; 95$ of the SO^ and about 60?b of the dust
is removed by a sodium sulfite solution. Most of the liquor discharged
from the scrubber is recycled to the scrubber. A portion of the liquor
la led to a reactor and treated with pulverized limestone.
SHaHSO, + CaCO, = Na2SO, + CaSO, + H20 + C02
The calcium sulfite is separated from the sodium sulfate solution; the
solution is returned to the scrubber. Calcium sulfite is oxidized in
an oxidizer to form gypsum. As sodium sulfate gradually forma in the
solution and tends to accumulate, a portion of the liquor discharged
from the scrubber is sent to a sulfate conversion step to maintain the
sulfate concentration at a certain level. In the conversion step, the
sulfate is treated with calcium sulfite and sulfuric acid to produce
gypsum and sodium bisulfite.
Na2S04 + 2CaS05 + H2S04 + 2H20 = 2(CaS04«2H20> + 2NaHSO,
The bisulfite solution is led to the reactor.
92
-------
Status of technology The pilot plant has 'oeen operated for more than
one year without serious trouble. Both line and limestone have been
used for comparison. Limestone reacts slowly with sodium bisulfite
requiring a few hours. By using line the reaction proceeds rapidly.
However, limestone will be used in the commercial plant because it is
anich cheaper than lime, and moreover, larger crystals of gypsum is
obtained with limestone. An oxidizer developed by Showa Denko and Ebara
will be used in the commercial plant. The oxidation proceeds a little
«ore slowly but the crystals of gypsum grow larger than with a rotary
atoaizer developed by JECCO.
Advantages The scrubber is very effective for deeulfurization. High
recovery of S02 ia achieved consulting limestone.
Disadvantages The process is less simple than the lime-gypsum process.
The use of a considerable amount of sulfuric acid is a demerit for
plants whose by-product gypsum or calcium sulfite must be discarded.
.Economics The estimated de sulfur 12 at ion cost for a 120,000scfm plant
for 9556 removal of SO. is shown below.
Plant cost $1.7 million
7,000 hours operation in a year
By-product gypsum 192,000 tons/year
Requirements (lb -oer bl oil)
KaOH
0.56
E2S°4
5.0
CaO
35.8
Water
560
Steam
0.38
Electricity
lO(kWh)
Yariable cost $O.J31/bl oil
Fixed coat $0.687/bl oil
Desulfurization coat $1.018/bl oil {without credit for gypstua)
$0.936/bl oil (with credit for gypsum)
93
-------
Gas
1
10 feet^.
<£>
4=
Liquid
Gas T
Gas-liquid
separator
Absorbing
section
I
28 feet
I
I
I
Liquid
Figure 10 Vertical-cone type
absorber for 60,000acfm
150
3 100
bO
0)
H
-P
O
a
CM
O
Ki
50 ..
0
L/G = 14
0 1,000 2,000 3,000 4,000
SOg ppm of inlet gas
Figure H Relation between
liquid/gas ratio (gal./l,000scf)
and SO^ removal (pH 6,5}
-------
1C
Ul
Gas mixer
Sulfate treatment
Mist
eliminator
i Oxldlzer
Centrifuge
Gypsum
Figure 12 Showa Denko sodium-limestone process
-------
2 4)
8 NEK ammonia process and ammonia-lime process
Developer Nippon Kokan Kabushiki Eaisha
1-1-3, Otemachi, Chiyodaku, Tokyo
Process description Waste gas at 250°P from an iron ore sintering plant,
containing 400 to l,000ppm SO. is first led into an electrostatic
precipitator and then cooled to 140 °F in a cooler with water spray
(Figure 13). The gas is then led into a screen type scrubber
(jinkoshi type scrubber ) for the absorption of S02 by a liquor
containing ammonium sulfite. In the scrubber 16 mesh screens of stainless
steel are placed with some inclination in five stages. On three of the
screens placed at the middle of the scrubber, the ammonium sulfite
solution flows slowly forming a liquor film which readily absorbs S0«.
On the other two screens placed at the upper part of the scrubber,
water flows slowly forming a water film to decrease plume formed by the
reaction of S0? and ammonia* About 95$ of S0« is removed when the pH
of the circulating liquor is about 6. Virtually no ammonia is lost when
the pH of the liquor is 6 or below.
The outlet liquor containing ammonium bisulfite is sent to an ammonia
absorber* Coke oven gas containing a small amount of ammonia is introduced
into the absorber. The liquor is sprayed to absorb ammonia and to form
an ammonium sulfite solution. A large portion of the solution is returned
to the scrubber to absorb SO-. The rest of the solution is sent to an
oxidizer where the sulfite is oxidized into sulfate by air bubbles
produced by rotary atomizers. The ammonium sulfate solution is evaporated
to produce crystal ammonium sulfate.
Nippon Kbkan has recently developed an ammonia-lime double alkali process
(Figure 1*0. The SO- absorbing part is the same as in the ammonia
process except that no coke oven gas is used. The liquor from the scrubber
contains about (NSL) SO, 7-5$, NH,ES03 7»5$ and (NH.)2SO. 15$. A portion
of the liquor is sent td a reactor end Is reacted with milK of lime (10$
concentration) under normal pressure at 210°F. The ammonia released here
Is sent to the ammonia absorber to be absorbed by the liquor from the
scrubber. Calcium sulfate and sulfite are precipitated in the reactor.
The slurry from the reactor is acidified with sulfuric acid to adjust
the pH to 4 to promote oxidation. The slurry is then led into an oxidizer
equipped with rotary atomizers to convert calcium sulfite to gypsum,
which is then centrifuged. Salable gypsum with a good quality is obtained.
The gas from the oxidizer contains S02 and is sent to the scrubber.
State of development After tests with a pilot plant to treat 17»000scfm
of waste gas from the iron ore sintering plant, a prototype plant to treat
88,000scfm gas to produce ammonium sulfate by reaction with coke oven gas
was courpleted in early 1972 at the Keihin works of Nippon Kbkan.
96
-------
Additional units for the ammonia-lime double alkali process were
completed in November 1972. The construction and the operation of
the plants have been carried out as a research project by Japan Iron &
Steel Federation.
Status of technology The Jinkoshi type scrubber capable of treating
88,000scfm gas is 56 feet high with a cross section of 14 feet x 23 feet.
The following conditions are used for S0« absorption:
L/G 12gal/l,000scf Gas velocity 7 to 13ft/sec.
Pressure drop 10 to 12 inches H«0
Inlet SOg 400 to l.COOppm Outlet S02 15 to 50ppm
The outlet gas is at 140°F and is not preheated; heavy plume is observed
from the stack. Tests have indicated that the plume becomes slight when
the gas is reheated to 180°F and nearly invisible at 240°F.
When coke oven gas is used as the source of ammonia, H«S in the gas is
absorbed to form thiosulfate.
2LS + 2SO," -f 2HSO," —^>3S20, + 3H20
The thiosulfate is not oxidized into sulfate in the oxidizer. It is
decomposed by addition of sulfuric acid to the liquor discharged from
the oxidizer.
By the decomposition SO. is released which is sent to the scrubber.
Elemental sulfur -formed by the reaction is removed by filtration.
In the ammonia-lime process, the oxidation of calcium sulfite into sulfate
is hindered under the presence of thiosulfate. Therefore, it is better
not to use coke oven gas in this case. To make up a small amount of ammonia
(about jfo} lost from the system, ammonium sulfate is added to the reactor
to react with lime and to generate ammonia. The pH in the reactor is
maintained at about 11 to ensure gypsum crystal growth in the reactor.
To promote the oxidation of calcium sulfite present in the slurry from
the reactor, the slurry is acidified to pH 4 by addition of sulfuric acid
and then led into the oxidizer. Gypsum grows into big crystals (100
to 300 microns) and can easily Tbe centrifuged to a low moisture content
(about 10$). The liquor from the centrifuge which is acidic is neutralized,
clarified and returned to the system. No wastewater is emitted.
97
-------
_Adv£r Adages. -rhe 3cx-&c-a type ssrabb^r it; effective for SO,. renaval with
B relatively lev pressure drop. A lar,^- niuount cf g^s5 up to about
5GG,GQGscf2i can. be treated in cr.e s crabber. 2y the acnonia process,
both SO,. In. waste gs.3 and. ar^o.iia in. coke over, gas are utilised to
produce salable ammonium sulfats. 3y the air.onia-Iiiae process, salable
gyp sun of good quality is obtained with no scaling problem. No wastewater
iB emitted.
coke oven, gas is used, hydrogen sulfide in thfc gas
tates additional facilities. The screen in the scrubber is
to corrosion under inadequate operating condition.
orL A cost estimation for the anmariiE-line process TO treat
ilue gas froa. an oil-fired boiler is shown in the following table in
cc-parison with -;hat for the liae-gyps-on prccesa developed by Nippon
Kokan which is sinilar to the Mitsubishi- JECCO lime gypsum process.
Cost: eatisation (31 = ¥308)
(SO in inlet gas l^COpprc, in outlet gas 70ppm)
Agmsr-i a- 1 J_m e p r o c e s_s Liae-sypsua process
^moxint of gas treated (scfm) 235,000 88? .000 235_T_OOp 882.000
Investment cost ($1,000; 2,110 4,545 2f27J 5|644
Fixed cost (SljOOO/year)
Depreciation 382 822 409 1,055
Labor (? persons) 45 45 45 45
Hepair 63 136 68 175
Insurance 25 J 7
Hana^ement 95 553 84 295
Variable cost (Si f 000/year)
Electric power (€11 mil/teWh) 168 634 1$1 822
Steam I©Sl.5/ton) 96 363
Industrial water («15 nil/ton) 4 14 3 11
Seawater(03 nil/ton) 1 2
Quick line {©313. I/ton) 168 688
Slaked lime t«S15. I/ton) 2?6 951
Sulfuric acid (^328. 3/ton) 45 180
Ammonia (f=>S87.6/ton) 37 148
Fuel C
-------
Coke-oven gas
Gas to stack
; Scrubber i p—» j-
*taJL^^ I — * 1
, I • ^^w^» - f^MvMH
10
Gas
i
A
Cooler
-***
„.*'
„*••
L
1 1
3
w
^
„(
,S
i
-K!
aterl i
'
"i ¥ '
Wi
1
1
1
Cooler i
ID '
>
•
-»
»— -p.
»
L
Te
Purified
coke-oven gas
Ammonia absorber f
i
Absor-
ber
~U
o
Filter
Tank
Evaporator
Air Ammonia
Ammonium
sulfate
Figure 13 Nlppn Kokan ammonium sulfate process
-------
Recovered water
Gas to stack
Gas to prescrubber
| NH3 absorber
Prescrubber
o
o
Thi ckener
Sludge
Reactor pH
adjusting
tank
Neutra-
lizer
S02 recovery
section
V
NH-j recovery |
section j
NHo regeneration
section
Gypsum
production
section
Water
recovery
section
Figure
Nippon Kokan ammonia-lime process
-------
References
1) H. ¥. Elder, P. T. Princiotta, G. A. Hollinden, and S. J. Gage,
Sulfur Oxide Control Technology, Visits in Japan—August 1972,
Interagency Technical Comittee, U.S.A., Oct. 1972
2) J. Ando, Recent Developments in Desulfurization of Fuel Oil and
Waste Gas in Japan (1973). prepared for U.S. Environmental
Protection Agency through Processes Research, Inc., April 1973
(in English)
3) Haiendatsuryu no Subete (All About Waste-gas Desulfurization), Jukogyo
Shimbunsha, Nov. 1972
4) M. Yokoi, Ryusan to Kogyo (Sulfuric Acid and Industry), Vol. 26,
No. 1, 1973
5) J. Sakanishi, ibid.
101
-------
ECONOMICS OF FLUE GAS DESULFURIZATION
by
Gary T. Rochelle
Control Systems Laboratory
National Environmental Research Center
Environmental Protection Agency
Research Triangle Park, N. C.
103
-------
Summary and Conclusions
This paper summarizes the results and conclusions of an analysis of the
costs of flue gas desulfurization for fossil fuel boiler plants. Cost details
of the desulfurization processes have been evaluated and, presented as a con-
sistent set of data from which it is possible to estimate capital and annualized
costs for any given plant conditions. The following conclusions were derived
from that data.
1) Flue gas desulfurization can be applied to 75% of existing fossil fuel
utility capacity at an annualized cost of 1.5 to 3.0 mills/kwh, which
is less costly than substitute low-sulfur fuels, typically costing 3.0
to 6.0 mills/kwh.
2) Regenerable processes are generally less costly than throwaway processes
at waste disposal costs above $3/ton wet sludge.
3) Newer processes such as citrate and double alkali may be up to 20% less
costly than currently available processes such as Wellman-Lord and lime-
stone scrubbing.
4) Annualized costs for a given process can easily vary from 2 to 6
mills/kwh with typical variations of plant sizes and load factors.
5) Flue gas desulfurization can be significantly cheaper to use with
medium sulfur coals if only a portion of the flue gas needs to be
treated in order to meet emission standards.
6) Extrapolation of this costing technique to industrial boiler conditions
indicates annualized costs of 60 to 130 C/MMBTU witli tlirowaway processes,
7) This cost data base consistently predicts actual first-of-a-kind project
capital costs within 10 to 20%.
104
-------
Introduction
A thorough evaluation of technology for control of SCL from power plants
must include consideration of numerous processes and approaches as well as
several criteria of usefulness. The scope of such an evaluation is illustrated
in Figure 1. The technology to be considered includes not only flue gas clean-
ing and low sulfur fuels for short term solutions, but also longer term
alternatives such as gasification, fluidized bed combustion, or solvent refining
of coal. This technology must be evaluated on the basis of the following
criteria:
1) environmental impact
2) stage of development
3) economics
This paper evaluates flue gas cleaning processes on the basis of economics* Its
scope is indicated by the dotted line in Figure 1.
For purposes of evaluation the flue gas cleaning processes considered here
can be divided into three groups:
1) Available throwaway - Limestone scrubbing and lime scrubbing
2) Available regenerable - Wellman-Lord, MgO scrubbing, and Cat-Ox
3) New/developing
a) Throwaway - Double alkali
b) Regenerable - Citrate and Stone & Webster/Ionics
These processes were selected as being representative; they do not include all
potentially important systems.
Each of the three groups appears to be identified with one of the evaluation
criteria listed above. Throwaway processes have serious problems with environ-
mental impact of waste disposal. Regenerable processes are often cited as
105
-------
LOW SULFUR
FUELS
GASIFICATION
FLUID-BED
COMBUSTION
SOLVENT
REFINING
OTHER
APPROACHES
TECHNOLOGY
FLUE GAS
CLEANING
THROWAWAY
NEW PROCESSES
REGENERABLE
ENVIRONMENTAL
IMPACT
DEVELOPMENT
STATUS
ECONOMICS
Figure 1. Evaluation of 502 control technology.
-------
having higher costs. New processes have not yet reached a stage of development
adequate for commercial applications. To provide a quantitative evaluation for
process selection, these different criteria must be quantified for all of the
process groups. This paper concentrates upon quantifying the economics of flue
gas cleaning processes, so that they can be compared with each other and with
other approaches to S02 control. Flue gas cleaning will be evaluated on the
basis of all three criteria -in-«£ paper to be presented at the APCA annual meet-
ing in June 1973.
In a paper presented in Fall 1972, we summarized and analyzed the cost
data for five wet scrubbing processes: limestone scrubbing, lime scrubbing,
Wellman-Lord, MgO scrubbing, and Stone & Webster/Ionics. That data base has
been expanded to include the double alkali and citrate processes as well as
preliminary information on Cat-Ox. The capital cost data for Wellman-Lord and
Stone & Webster/Ionics has been modified to reflect the higher costs of purge
handling and S02 reduction being estimated for the NIPSCO Wellman-Lord demonstra-
tion. Corrections and supplements to the 1972 paper are being prepared and
will be available shortly. This paper summarizes the results and conclusions
of the revised cost analysis.
The major sources of cost data are given in Table 1. The primary sources are
important, detailed estimates. The secondary sources include project, scoping,
and open literature references of lesser detail or importance.
The Cost Model
On the basis of the aggregated cost data, CSL has developed a series of
equations and graphs that allow cost estimates to be generated as a function
of process type and conditions of specific installations. For ease of repre-
sentation, the scrubbing processes are divided into two process areas:
107
-------
Table 1. SOURCES OF COST DATA
Primary
(2)
Catalytic, Inc. — Limestone scrubbing
(3)
TVA — MgO conceptual study; includes limestone
(4)
Davy-Power Gas — Demonstration project costs and proprietary data
Allied Chemical — Proprietary data on sulfur production
Secondary
M. W. Kellogg — Cost study of 12 processes
Stone & Webster/Ionics — Costs of electrolytic regeneration
(8)
TVA — Widow's Creek project estimate (limestone)
(9)
Monsanto/Illinois Power — Cat-Ox project costs
Chemico/Boston Edison — MgO project costs
Bechtel — Limestone scrubbing costs
108
-------
scrubbing and alkali handling. Scrubbing costs are dependent on the flue gas
rate but not on the amount of sulfur removed. Alkali handling costs vary with
sulfur rate but are independent of flue gas rate.
The structure of economics is illustrated in Figure 2. Capital costs
are made up of four segments: process direct, other direct, contractor
indirect, and user indirect. The process direct costs include labor and
material for process equipment. For scrubbing, process direct costs are repre-
sented as a function of process type, flue gas rate, and retrofit conditions.
Alkali handling process direct costs are a function of process type and the
amount of sulfur removed per hour. The other direct costs and indirect costs
ate estimated as a percentage of process direct costs. In this analysis total
capital costs are computed as 170% of process direct costs, unless otherwise
noted. Capital costs can be expressed as absolute dollars or $/kw of generating
capacity.
Annualized costs include contributions from utilities and raw materials,
labor, maintenance and capital charges. Utilities and raw materials cost per
year are functions of gas rate or sulfur rate, process type, and load factor.
Labor is assumed to be $225,000/year regardless of plant size. Maintenance
cost is estimated as 4.5% of capital cost per year. Capital charges at 17.5% of
capital cost are meant to include depreciation, return on investment, taxes, and
insurance. Annualized costs can be expressed as absolute dollars per year or
as cost per unit throughput, such as mills/kwh, C/MMBTU, or $/ton sulfur.
Generally CSL has used mills/kwh. Variations of annualized cost with load
factor depend on the units used. At lower load absolute costs per year will
decrease, but relative costs per unit throughput will Increase.
109
-------
ANNUALIZED
COSTS
UTILITIES
AND RAW
MATERIALS
LABOR
MAINTENANCE
CAPITAL
CHARGES
CAPITAL
COSTS
PROCESS
DIRECT
OTHER
DIRECT
CONTRACTOR
INDIRECT
USER
INDIRECT
Figure 2. Structure of economics.
-------
Estimated Versus Actual Costs
Using this cost data base and estimating techniques CSL has evaluated
the costs of six actual construction projects. The results of these evalua-
tions are given in Table 2. Generally all but 10 to 20% of the actual or
contractor estimated investment costs were predicted from process configurations
and site conditions. This consistently good prediction of costs was maintained
over a wide range of retrofit conditions, plant size, fuel type, and process
type. Therefore this technique of cost prediction should be accurate in pre-
dicting variations in process costs and can be relied upon for process comparisons.
However, the estimated costs are consistently about 15% lower than the
actual project costs. This error may result from insufficient detail on actual
contractor estimates or from real costs for first-of-a-kind installations. When
cost detail is lacking, assumptions must be made for the extent of inflation and
other direct and indirect costs. These assumptions will probably be more con-
servative than the assumptions of a contractor trying to make a profit. The
costs presented here are meant to be accurate for widespread commercial applica-
tions and do not include first-of-a-kind costs. Nevertheless, we recognize
that the bias in our estimated capital costs may be partially real, At the
worst however, the 15% lower capital cost projects would result in annualized
costs that are 5 to 10% low.
In order to better illustrate project cost estimating, two of the cases
will be discussed in greater detail. Calculations for the Will County limestone
scrubbing facility are summarized in Table 3. The power plant is designed
for a new generation of 163 Mw, but its gross capacity is 177 Mw. We have estimated
the design flue gas rate at 360,000 SCFM or 180 Mw gas equivalent. The actual
111
-------
Table 2. ESTIMATED CAPITAL COSTS OF DEMONSTRATION PROJECTS ($MM)
Explained casts
Actual costs
Unexplained costs,
% of actual costs
Plant characteristics
K)
Will
County
11.3
13.3
15%
Limestone
Retrofit
3% S Coal
177 Mw
Widow's
Creek
27.6
35.0
21%
Limestone
Retrofit
4% S Coal
550 Mw
Sherburne
County
31.0
36.1
14%
Limestone
New
1% S Coal
1360 Mw
LaCygne
29.5
32.5
9%
Limestone
New
5% S Coal
820 Mw
Mystic
5.6
6.2
10%
MgO
Retrofit
2% S Oil
155 Mw
Mitchell
7.7
9.6
20%
Wellman-Lord
Retrofit
3% S Coal
115 Mw
-------
Table 3. ESTIMATED COSTS OF WILL COUNTY LIMESTONE DEMONSTRATION
(Basis: 180 Mw (360,000 SCFM), 15 tons limestone/hour, difficult retrofit.)
Process area Cost, $MM
Scrubber, two modules 8.6
Alkali handling, without pond, grinding
for 24 tons limestone/hr 2^7
Total 11.3
Actual Costs (includes considerable overtime) 13.3
Table 4. ESTIMATED COSTS OF THE NIPSCO WELLMAN-LORD DEMONSTRATION
Process area Comments Estimated3 Actual Unexplaine<
by model Estimate
Scrubber
Evaporator
Purge handling
Sulfur plant
Boiler
Total
Single shell with particulate scrubbing
Very small, almost pilot scale
Earlier paper
Very small, almost pilot scale
Earlier paper
Not usually required
3.00
1.65
0.65
(0.30)
1.50
(0.95)
0.90
7.7
3.0
1.8
1.3
-
2.6
-
0.90
9.6
—
+ 0.15
+ 0.65
(+ 1.0)
+ 1.10
(+ 1.8)
_
1.9
a
Includes actual indirect cost rates with predicted direct costs.
113
-------
sulfur design rate is not disclosed, but the limestone rate is 15 tons/hr,
equivalent to about 3 tons/hr of sulfur. The scrubbers were installed under
severe space limitations and are therefore rated as difficult retrofits.
Normally, the total gas flow could possibly go through a single large scrubber,
but two smaller modules have been used. Furthermore, the limestone grinding
facilities were designed for almost double capacity (24 tons/hr). The actual
costs are about $2 MM more than the estimated costs, but include a large amount
of overtime labor not usually accounted for by the estimating procedure. A
similar facility designed for 180 Mw, 3 tons S/hr with easy retrofit conditions
and using only one scrubber and matching sized limestone grinding should cost
only $7.0 MM compared to the estimated cost of $11.3 MM for this facility.
The estimated costs for the Mitchell Station Wellman-Lord process are
given in Table 4. The costs are compared on an area by area basis with the
actual contractor estimates. Costs for the scrubber and evaporator agree very
well. Costs for purge handling and the sulfur plant reported in the earlier
CSL paper were not well founded and have actually been modified for this paper
to reflect the much higher costs estimated for the Mitchell Station. Even as
modified, the contractor costs for these two areas are almost a factor of two
greater than our estimated costs. However, both the purge handling and sulfur
plant are very small, first-of-a-kind, almost pilot plant facilities. Since
the estimating technique is intended to be most representative of larger,
widespread applications, the lack of accuracy under these extreme conditions
is not unexpected. The Mitchell estimate also includes provision for a
package boiler which would normally not be required. Another factor of great
importance is inflation. The Mitchell estimate includes actual costs through
1975. Our time basis is late 1972, so as much as 15% should be added to our
114
-------
estimated costs to account for inflation. This would put estimated costs
within 5% of actual project estimates.
Comparative Process Costs
In order to provide a point of process comparison, capital and annualized
costs for a single, base case power plant are given in Table 5. The range of
annualized costs between processes is quite small, varying from 1.95 to 2.75
mills/kwh. At the conditions selected, lime and limestone scrubbing are somewhat
less expensive than the regenerable processes. The new processes, double alkali
and citrate, appear to offer savings up to 20% in annualized costs over available
throwaway and regenerable processes. The costs for the Cat-Ox process have been
scaled from the actual demonstration project costs and are therefore somewhat
preliminary. Nevertheless, the estimated annualized cost of Cat-Ox is within
5 to 10% of the other regenerable systems, and actual costs could easily turn
out to be competitive. At conditions different from the base case, the relative
cost ranking can change. For example, higher costs of waste disposal would make
throwaway processes more expensive.
Regenerable Versus Throwaway - The comparative costs of regenerable and
throwaway processes are typified by the annualized costs of Wellmah-Lord and
limestone scrubbing as compared in Table 6. In the scrubbing area limestone
is more expensive because of the higher capital expense of using limestone
slurry with larger liquid recirculation, more complicated demisting, and
erosion resistant materials. On the other hand the annualized costs of alkali
handling appear to be greater for Wellman-Lord, because of contributions from
large capital expense and utility requirements. Therefore the major cost com-
ponents balance out, and overall annualized costs differ by only 10Z.
115
-------
Table 5. COMPARATIVE PROCESS COSTS
(Basis: 500 Mw, 3.5% S coal, retrofit, 60% load, waste at $3/ton wet
sludge, sulfur credit at $15/ton, particulate removal included.)
Process
Comparative Process Costs
Capital, Annualized,
$/kw mills/kwh
Throwaway
Double alkali
Lime scrubbing
Limestone scrubbing
24
35
36
1.95
2.40
2.45
Regenerable
Citrate
MgO (to S)
Wellman-Lord (to S)
Stone & Webster/Ionics (to S)
Cat-Ox
39
49
50
50
55
1.95
2.40
2.65
2.70
2.75
a
Capital costs do not include disposal facilities, usually $5-15/kw.
116
-------
Table 6. ANNUALIZED COSTS OF LIMESTONE SCRUBBING AND WELLMAN-LORD
Scrubbing
Operating utilities and labor
Capital charges and maintenance
Subtotal
Annualized costs (mills/kwh)
Limestone Wellman-Lord
scrubbing
0.21
1.33
1.54
0.21
1.09
1.29
Alkali handling
Operating utilities and raw materials
Labor
Capital charges and maintenance
Waste disposal costs and sulfur credit
Subtotal
0.23
0.06
Q
0.18a
0.45
0.45
0.06
0.99
- 0.15
0.92
1.35
Total annualized cost
2.46
2.64
Capital charges are not included for a disposal pond. Its costs are
represented by waste disposal costs.
117
-------
The assumed sulfur credit for Wellman-Lord is $15/ton, but at that level
it only amounts to 0.15 mills/kwh. If acid were produced rather than sulfur,
the costs would be reduced about 0.15 mills/kwh in addition to the byproduct
credit which could amount to as much as 0.30 mills/kwh. Therefore, acid pro-
duction could be as much as 0.30 mills/kwh less costly than sulfur production.
The key assumption affecting the relative costs of regenerable and throwaway
processes is the cost of waste disposal. The assumed cost in Table 6 is $3/ton
wet sludge ($1.50/ton dry solids) and corresponds to the typical costs of a
disposal pond. Figure 3 illustrates the variation of the annualized costs of
lime and limestone scrubbing with waste disposal costs. The middle range of
regenerable process costs is also shown. The range of disposal costs at which
regenerable processes are competitive varies from $1 to 3/ton wet sludge. In
practice, costs for disposing of sludge via dewatering and hauling can run as
high as $7/ton sludge. Therefore, it is certainly not accurate to say that
throwaway processes are less expensive than regenerable systems, though there
will certainly be localities of low-cost waste disposal where throwaway systems
will be 5 to 20% less expensive.
New Processes - The double alkali and citrate processes are typical of the
cost benefit to be achieved by new systems. The capital cost of double alkali
is compared with lime scrubbing in Table 7. The double alkali scrubbing system
is much cheaper because it can achieve S09 and particulate removal in a very
simple single stage venturi scrubber with little concern for entrainment and
erosion problems. Of course the double alkali process requires more equipment
for alkali handling, but the overall capital cost is still 15 to 30% less
expensive than lime scrubbing. Utility and raw materials costs are the same
118
-------
3.0
CO
o
o
2.5
LIMESTONE SCRUBBING
2.0
.IME SCRUBBING
12345
COST OF WASTE DISPOSAL, $/ton WET SLUDGE
Figure 3. Process costs versus waste disposal costs.
-------
Table 7. CAPITAL COSTS OF DOUBLE-ALKALI AND LIME SCRUBBING
(Basis: 500 Mw, 3.5% S coal, typical retrofit, includes particulate scrubber.)
Process area
Capital Costs, $/kw
Double alkali/lime Lime scrubbing
Scrubbing
Alkali handling (not including pond)
Total
18.1
6.8
24.9
32.0 (25.9)
2.7
34.7 (28.6)'
"without particulate scrubber.
Table 8. ALKALI HANDLING COSTS OF CITRATE AND WELLMAN-LORD
(Basis: 500 Mw, 3.5% S coal, 60% load.)
Annua1ized costs, mills/kwh
Cost factor
Operating utilities and raw materials
Labor
Capital charges and maintenance
Sulfur credit
Total (alkali handling)
Citrate
0.20
0.06
0.57
-0.15
0.68
Wellman-Lord
0.45
0.06
0.99
-0,15
1.35
120
-------
for both processes, so the difference in annualized costs is a result of capital
cost differences.
The scrubbing costs of citrate and Wellman-Lord are essentially the same
since both processes use clear solution scrubbing. The annualized costs of
alkali handling are compared in Table 8. The citrate process regenerates
scrubbing solution with H2S to sulfur, so no steam is required as with Wellman-
Lord. Furthermore, the regeneration reactor is quite a bit simpler than the
Wellman-Lord evaporator. The citrate process requires a unit to produce H_S
from sulfur, but it should be cost equivalent to the Wellman-Lord unit for
production of sulfur from S02. Therefore, the capital costs, and hence main-
tenance and capital charges, are 40% less for citrate. The overall costs for
alkali handling are about 50% lower for citrate.
The reduced costs for new processes may not totally materialize. The cost
estimates are necessarily based on preliminary information. Additional processing
equipment may be identified as the processes become better developed.
Cost Variations with Source Parameters
As just' explained, the total variation of annualized costs with process type
at the base conditions is limited to a factor of about 1.4 from double alkali to
Cat-Ox. However variation of costs with source parameters such as plant size,
fuel sulfur content, and load factor is much more pronounced. The variation
of total capital costs (including waste disposal) with plant size is illustrated
in Figure 4. Costs for the available systems, Wellman-Lord and limestone, vary
a factor of 1.65 from $65-75/kw at 100 Mw to $45/kw at 1000 Mv. Citrate and
double alkali capital costs are 20 to 50% lover.
121
-------
BASIS: 3.5°, S COAL RETROFIT, 1 k* - 2 scfm
WELLMAN-LORD
NJ
ro
DOUBLE^ALKALI LIME
200 300
PLANT SIZE, MW
Figure 4. Total capital costs.
-------
The variation of annualized costs with plant size and load factor is
illustrated in Figure 5. Over a range of 100 to 1000 Mw, 40 to 70% load
factor, the annualized costs for a typical retrofit plant with 3.5% S coal
vary a factor of 3 from 2 to 6 tnills/kwh. The sludge waste disposal cost was
treated in Figure 5 as a materials operating cost rather than capital cost.
Therefore the limestone process is relatively less costly than Wellman-Lord
at smaller plant sizes and lower load where capital costs have a greater
impact than materials costs.
Figure 6 illustrates the variation of annualized costs of limestone
scrubbing with sulfur content. In most estimates it is assumed that all of
the gas will pass through the scrubber, in which case the cost varies a factor
of 1.4 in going from 1 to 5% S coal. The other curves in Figure 6 represent
costs if some flue gas can be bypassed around the scrubber while maintaining
a sulfur emission standard. Clearly the cost to treat flue gas from combustion
of 1% S coal will be zero if the sulfur emission standard requires 1% S. Actual
costs for 3.5% S coal with a 2% S standard can be as much as 50% less than
costs for treatment of all the flue gas. With a 1% S standard, the cost of
using 2% coal with flue gas scrubbing would only be 1.25 mills/kwh (12.Sc/MMBTU)
Therefore, substitution of low sulfur fuel is economical only if its incremental
cost is less than 12.5C/MMBTU. With a 1% S standard, annualized costs vary a
factor of 2 from 1.25 mills/kwh at 2% S to 2.5 mills/kwh at 5% S.
Actual Distribution of Annualized Costs
It is somewhat academic to discuss process costs over a range of source
conditions as discussed earlier. In real life,costs must be considered as applied
to actual plants. Figure 7 represents a distribution of calculated annualized
123
-------
K>
jr
6.0
5.0
4.0
8 3.0
UJ
M
2.0
LIMESTONE
SCRUBBING
LIMESTONE SCRUBBING
1.0
BASIS: 3.5*, S COAL, RETROFIT, WASTE AT S3, TON
WET SLUDGE, SULFUR CREDIT AT S15/TON S
I
I
I
I I
50
100 200 500
PLANT SIZE, MW
Figure 5. Annualized costs versus plant size and load factor.
1000
2000
-------
3.0
Ul
2.5
I 2.8
*i
o
o
" ^^
i
5
.v
1.0
0.5
BASIS: LIMESTONE SCRUBBING,
500 MW, 60% LOAD, RETROFIT
90% REMOVAL
I
1.0 2.0 3.0
COAL SULFUR CONTENT, wt %
Figure 6. Annualized costs versus sulfur content.
4.0
5.0
-------
6.0
5,0
4.0
o 3.0
a
UJ
ISI
2.0
1.0
I
I
I
i
I
10 20 30 40 50 60 70
COAL-FIRED CAPACITY THAT CAN BE RETROFITTED AT COST OR LESS, percent
80
90
Figure 7. Distribution of annualized costs over utility population, limestone scrubbing.
-------
costs over a random sample of coal-fired utility plants including 25% of the
plants in the 1969 FPC form 67 survey. Costs of limestone scrubbing for each
plant were calculated as a function of plant size, load factor, and fuel
sulfur content. Retrofit factors were estimated on the basis of unit age and
size. Each system was assumed to treat all of the flue gas for SO and
X
particulate removal. Waste disposal was represented as an operating cost at
$37ton sludge. In addition a random variation of plus or minus 15% (over two
standard deviations) was incorporated into the costs to account for variations
in waste disposal costs and retrofit factors that could not be correlated
with plant parameters.
As shown in Figure 7, 75% of the coal-fired utility capacity in the U. S,
could be retrofitted with limestone scrubbing at a cost of 3.0 mills/kwh or
less, 50% at a cost of 2.0 mills/kwh or less. Thus representative costs for
widespread application of flue gas cleaning would be 1.5 to 3.0 mills/kwh.
It is also apparent from Figure 9 that 25% of the capacity would cost more
than 3.0 raills/kwh (about 30C/MMBTU) to retrofit with flue gas scrubbing.
This capacity would probably be most economically controlled by the use of
low-sulfur fuels, if available at costs less than the cost of flue gas
scrubbing. It would appear from Figure 9 that 15% of the capacity would cost
more than 6 mills/kwh (60C/MMBTU). This capacity would certainly consider
clean fuels or another more economic approach to sulfur emission abatement.
Since substitution of low-sulfur fuels is expected to add an incremental
cost of 30 to 60C/MMBTU, flue gas cleaning would be the better economic
choice for 75 to 85% of current coal-fired capacity.
As shown in Table 9, costs for typical new plants would be about 1.85
mills/kwh (18.5C/MMBTU) for coal or 1.3 mlllc/lcwh (13C/MMBTU) for residual
127
-------
Table 9. COST OF LIMESTONE SCRUBBING FOR NEW PLANTS
(Basis: 500 Mw, 70% load factor, no particulate scrubbing, 70% indirect
costs, and 1972 dollars.)
*»
Annual!zed, Capital,
Coal,
Oil,
Fuel
3.5% S
2.5% S
mllls/kwh
1.85
1,31
$/kw
25
20
Waste disposal included in operating cost, but not in capital cost.
Disposal ponds would cost $5-10/kw.
128
-------
26
oil. Thus the use of substitute low-sulfur coals or even desulfurized oil
at incremental costs over 20C/MMBTU would not be economically favored.
Therefore most new fossil-fired plants will probably use flue gas cleaning
until a cheaper alternative (such as fluidized bed combustion) can be
developed.
Costs with Industrial Boilers
Generally the costs presented thus far have related primarily to applica-
tions on utility boilers. Coal-fired industrial boilers present a different
picture. A typical coal-fired industrial boiler plant has three boilers with
a total capacity of 500 MMBTU/hr (50 Mw). The average load factor is probably
about 50%. Higher excess air (50% compared to 20% for utilities) is used
with the typical stoker-fired boilers, so larger scrubbers must be used.
Capital expense is a major factor that can have more than twice as great an
impact on industrial boiler annualized costs. Industrial facilities are built
primarily with equity, not 50% debt/50% equity as with utilities; therefore
taxes are substantially higher. Furthermore, non-regulated industry necessarily
requires higher return on investment, because of the risk involved. The com-
bined effect of equity financing and higher investment returns is that capital
charges for depreciation, return on investment, and taxes amount to 33Z of the
capital cost per year, not 17.5% as with utilities. Also, the relative impact
of capital costs is increased because of the small scale of operation.
Figure 8 presents the annualized costs of lime scrubbing and double alkali
as applied to coal-fired industrial boilers. The capital costs were extra-
polated from utility costs. Actual capital costs may be lower if shop-fabricated
129
-------
180
160
140
"120
CO
o
S
o
Nj 100
_J
I
80
60
1 T
40
30
BASIS: 2 Ib S 106 Btu REMOVED, 50*. LOAD,
38'o CAPITAL AND MAINTENANCE CHARGES
I
I
I
I
I I
50
400
100 200
BOILER PLANT SIZE, 1Q6 Btu/hr FUEL
Figure 8. Annualized costs of throwaway processes on industrial boilers.
600
800 1000
-------
systems are practical. Labor and overhead costs were assumed to be constant
at $40,000/year. Waste disposal costs were assumed to be $5/ton of wet sludge.
The annualized costs on a boiler plant of 500 MMBTU/hr are 60 to 80C/MMBTU.
On a plant of 100 MMBTU/hr the costs are 100 to 130C/MMBTU. Such prohibitive
costs indicate that substitute low-sulfur fuels will probably be used if they
are available.
References
1. Burchard, J.K., Rochelle, G.T. e£ al^. "Some General Economic Considerations
of Flue Gas Scrubbing for Utilities," Sulfur in Utility Fuels: The Growing
Dilemma, pp. 91-124, McGraw-Hill, Inc. (1972).
2. Catalytic, A Process Cost Estimate for Limestone Slurry Scrubbing of Flue Gas,
NTIS No. PB219-016 (1973).
3. Tennessee Valley Authority, Sulfur Oxide Removal from Power Plant Stack Gas:
Magnesia Scrubbing-Regeneration. Contract No. TV-29233A (draft, 1973).
4. Davy - Power Gas, Project Manual for S00 Demonstration Unit, Contract No.
EPA-68-02-0621 (draft, 197TT. - ~ ~~~~
5. Allied Chemical, proprietary data, 1972.
6. M.W. Kellogg, Evaluation of S02 Control Processes, NTIS No. PB204-711 (1971).
7. Humphries, J. J. e£ jrl., An SO Removal and Recovery Process, Chemical
Engineering Progress 67 (5): 65, May, 1971.
8. Tennessee Valley Authority, private communication.
9. Control Systems Laboratory, EPA, "Briefing Document: Cat-Ox Demonstration"
(1972).
10. Quigley, C.P., "Magnesium Oxide Scrubbing System at Boston Edison Company's
Mystic Station," Sulfur in Utility Fuels; The Growing Dilemma, pp 287-290,
McGraw-Hill (1972). ~
11. Sherwin, R.M. £t al., "Economics of Limestone Wet Scrubbing Systems," Proceed-
ings £f Second International Lime/Limestone Wet-Scrubbing Symposium. APTD-
1161, pp. 745-64 (1972).
131
-------
Notes
1. The mention of company names or products is not to be considered
as endorsement or recommendation for use by the U. S. Environmental
Protection Agency.
2. EPA policy is to express all measurements in Agency documents in
metric units. When implementing this practice will result in undue
cost or difficulty in clarity, NERC/RTP is providing conversion
factors for the particular non-metric units used in the document.
For this report these factors are:
British
1 SCFM (60°F)
1 short ton
1 MMBTU
Metric
1.61 NM3/hour (0°C)
.91 metric tons
252,000 kilocalories
132
-------
STATUS OF TECHNOLOGY OF COMMERCIALLY
OFFERED LIME AND LIMESTONE FLUE
GAS DESULFURIZATION SYSTEMS
by
I. A. Raben
Bechtel Corporation
San Francisco, California
133
-------
SUMMARY
The status of technology for commercially offered lime/limestone
SO_ removal processes has been evaluated. Suppliers have been con-
L*
tacted to determine the present status of design criteria. It is im-
portant to note that most commercial system suppliers will guarantee
80- and 90-percent SO removals using limestone and lime scrubbing,
u
respectively.
Between 1970 and 1973, liquid-to-gas ratio has increased from 20-30
to 40-80 gallons per thousand cubic feet, and stoichiometry has de-
creased from 1.7 to as low as 1. 2. The new values for both parameters
tend to make designs more reliable and free of scaling.
There are 21 full-size units being designed or constructed. The num-
ber of SO. removal systems (using alkali addition into scrubber) opera-
ting for any period of time is quite limited. The Will County Station of
Commonwealth Edison is the only domestic coal-fired unit with scrubber
addition that has operated for a significant period. The facility has
shown high SO2 removal but has been plagued with mechanical problems.
Much has been learned from its operation to advance commercial
technology.
Capital and operating costs have been estimated for new units. The
capital cost of a 500-megawatt unit with the use of limestone is esti-
mated at $45 per kilowatt; a typical operating cost with limestone is
134
-------
estimated at 23-25£ per million Btu. It is not possible to generalize
capital costs for retrofit units because of the special requirements for
each plant. New SO- removal systems using lime will cost $38 per
kilowatt with typical annual operating costs of 23£ per million Btu.
The two areas requiring further study are demonstration of long-term
reliability and sludge disposal management. Long-term reliability
will be evaluated and hopefully demonstrated when many of the 21 units
come on stream in 1973 and 1974. To develop ecologically sound sludge
disposal methods, techniques must be studied in much greater detail.
135
-------
ACKNOWLEDGMENT
The author wishes to express his appreciation for contributions
from the engineers of Bechtel's Scientific Development Air Qual-
ity Group. The cooperation of the SO- removal system suppliers
is also gratefully acknowledged.
136
-------
Section 1
INTRODUCTION
When the first EPA Symposium on Lime/Lime stone Scrubbing was
held in November 1971 in New Orleans, Louisiana, there were only
three wet scrubbing systems operating. These units consisted of
alkali injection into the boiler furnace followed by wet scrubbing.
Today there are over 20 commercial systems being designed, con-
structed, or operated. We have gained much knowledge in the past
18 months, but we still have more to learn.
The lime/limestone sulfur-dioxide removal process is the most ad-
vanced system at this time. The primary reasons for this are:
The process is more fully characterized than other
processes. Approximately 20 pilot plant test pro-
grams have been conducted over the last 3 years.
In addition, the EPA Test Facility has been operat-
ing for over a year, and test results will be reported
at this meeting.
The process has relatively low capital and operating
costs.
The process can achieve high sulfur-dioxide removal
efficiency.
137
-------
However, these systems are not without problems, which include con-
trol of scaling and erosion/corrosion and solid waste disposal.
This paper describes the commercially offered lime/limestone pro-
cesses for SO^ removal and indicates the type of alkali, capital and
operating costs, types of scrubbers, and design criteria. It presents
the process chemistry as it relates to system reliability and equip-
ment selection, and it discusses the commercial systems as to boiler
size, scrubber type, vendor, and status. The paper also evaluates
solid waste disposal.
138
-------
Section 2
CHEMISTRY OF LIME-LIMESTONE SCRUBBING OF
The chemistry of lime-limestone scrubbing is complicated because of
the large number of species present in the system at equilibrium.
2. 1 RAW MATERIALS AND CONSTITUENTS
The three raw materials — gas, fly ash, and limestone — each contain
several constituents that affect the chemical composition of the system.
For the power plant, the gas supplies SO , SO3> CO , NO, and NO ;
the ash contributes Na, K, Ca, Cl, Fe, Si, and others. Limestone
gives Ca, Mg, and other constituents in minor proportions — Na and K.
2. 2 REACTIONS
The main reactions in the scrubbers are assumed to be:
• Absorption of SO
• Hydrolysis to form H SO acid
• Reaction of sulfite ion from H_SO with calcium ion from
CaC03 or Ca(OH>2 3
These reactions are affected in several ways by other constituents in
system. Detailed studies of the system chemistry have been
139
-------
carried out by TVA (Ref. 1) and the Radian Corporation (Ref. 2).
TVA studied the effect of supersaturation, ionic strength, and sul-
fite oxidation. Radian developed a computer program using 41 spe-
cies and 28 equations to predict equilibrium compositions in the
scrubber circuit.
The main equations in the scrubber can be written (Ref. 3):
• H SO j*.HSO" + H+
- + =
• HSO J± H + SO
• CaCO3(s)^CaCO3(aq)4±Ca+"1" + CO3 + H+«± CaHCO*
• CaHCO* £ Ca++ + HCO~
• Ca++ + SO3 + 0. 5 H2O ;2CaS03 • 0. 5
• CaSO, + 1/2 O^^i
3 2
If the system is assumed to be one of sulfurous acid formation fol-
lowed by the reaction of acid with lime or limestone, then the fol-
lowing effects may influence the overall kinetic rate:
1. Diffusion of SO_ to and through the gas film at the
liquid surface
2. Dissolution of SO^
3. Hydration of SO- to H2SO3> H*. and HSO"
4. Dissociation of HSO, to form SO_
5. Diffusion of H-SO. and ions through the liquid film at
the droplet surface and into the droplet interior
6k Hydration of CaO to Ca(OH>2 when CaO is used
140
-------
7. Dissolution of Ca(OH) or CaCO.
3
8. Reaction of Ca(OH) or CaCO, with H to give C&++
++ =
9. Reaction of Ca with SO, to precipitate CaSO,
•3 3
Available data indicate reactions in steps 3, 4, 8, and 9 are rapid.
The controlling mechanisms are therefore either gas diffusion, liquid
diffusion, CaO hydration, or dissolution of CaCO, or Ca(OH) . For
•5 Ci
the most used design case — introduction of CaCO into the scrubber —
gas phase mass transfer of SO and CaCO^ dissolution are the con-
trolling steps. This case was studied by Boll (Ref. 4) in a three-stage
scrubber, and he found that these two steps were most critical.
2. 3 INFLUENCE OF PROCESS CHEMISTRY ON SO REMOVAL
DESIGN L
In gas absorbers, one of the major criteria establishing the size of the
equipment is the rate of mass transfer from the gas to the scrubbing
medium. For lime or limestone scrubbing, gas-phase absorption or
chemical reaction rate, or both, may be rate-controlling resistances.
For optimum SO mass transfer, it is necessary to maximize the sur-
L*
face area of the calcium-contributing reactant and to minimize the re-
sistance of the gas-to-liquid interface — at minimum cost. Both hy-
drated lime, Ca(OH)2> and precipitated CaCO_ [resulting from the
reaction of Ca(OH)2 with CO. in the flue gas] have very high specific
surface area compared to ground limestone. Hence, if limestone is
used, it is important to supply a large surface area per volume of
SO, per unit of time in the scrubber to keep the solution supplied
with calcium ions.
141
-------
Resistance to gas phase SO mass transfer can be minimized by the
proper choice of scrubber type and high gas velocity. Maximizing
reactant surface area can be accomplished by:
• Large liquid holdup of slurry in the scrubber
• High slurry recirculation rate
• High solids content in slurry
• Increasing feed stoichiometry
• Reducing the particle size of feed
Liquid holdup depends mainly on scrubber type, in descending order as
follows: tray-type, packed-bed, floating-bed, spray tower, and ven-
turi. Slurry recirculation rates in the order of 20 to 80 gallons per
thousand cubic feet of saturated gas would be required, depending on
the type of scrubber used. Solids content of 5 to 15 weight-percent
are normally used in the slurry. Limestone feed stoichiometry may
vary from 1. 2 to 1.5 (stoichiometry being defined as the molal ratio
of alkali feed to SO. absorbed).
Reducing the particle size of limestone is an obvious way of increasing
surface area. Good results have been obtained at a particle size of
80 to 90 percent through 200 mesh. * It should also be noted that if
!U. S. Mesh No. Typical Median Range
90% through basis, microns
100 27-66
200 14-35
300 7-18
142
-------
particles are too fine, solids settling may become a problem. In
addition, the mill capacity will decrease and power consumption in-
crease for finer size reduction at a given limestone hardness. *
High slurry recirculation rate and high solids content in the slurry will
not only maximize reactant surface area, but will also minimize scal-
ing potential, if sufficient delay time (approximately 10 minutes) is
provided in the recirculating tank.
2.4 WATER BALANCE AND AMBIENT QUALITY
CONSIDERATIONS
It is necessary that wet scrubbing systems be operated in closed loop
to minimize effects on adjacent water basins. In closed loop opera-
tion the amount of make-up water used will be equal to the amount of
water evaporated to saturate the flue gas plus the amount entrained
or combined with solid waste produced. The remaining liquor (i.e.,
pond or thickener overflow) is recycled to the process. From a
process standpoint, closed loop operation requires recycling saturated
solutions. Process conditions must be carefully controlled to prevent
scaling.
* Lime stone hardness "Work Index" — 6-10.
143
-------
Section 3
COMMERCIAL SYSTEM DESIGN
3.1 PROCESS APPROACH
The development of the lime- lime stone scrubbing system has taken
three process routes:
• Introduction of limestone into the scrubber circuit
(Figure 1)
• Introduction of limestone into the boiler to produce
CaO, followed by scrubbing of the flue gas (Figure Z)
• Introduction of lime into the scrubber circuit
(Figure 3)
The most used process route is introduction of limestone into the
scrubber. This approach has the advantage of minimum effect on
the power plant; it can achieve high SO_ removal with minimum
scaling and plugging. The disadvantage of this process is that
limestone is less reactive than lime; to offset this limitation, a
higher stoichiometric ratio of limestone to SO must be used, more
Lt
slurry must be recirculated {higher liquid-to-gas ratio), and a counter-
current scrubber with several stages is required.
The second process approach of introducing limestone into the boiler
furnace produces a calcined limestone. Calcium oxide (CaO) enters
144
-------
GAS TO STACK
STACK
GAS
CfS03+CiS04
TO WASTE
Figure 1. Scrubber Addition of Limestone — SO2 Removal
I • GAS TO STACK
IOH.EI
C«OGAS
SCIUIIH
PUMP
TANK
••
SITTllft
J
~1
CsCOj*-
TO WASTE
Figure 2. Boiler Injection of Limestone Followed by Wet Scrubbing
STACK GAS-
•*- 6AS TO STACK
SCKUIIEI
ۥ0
PUMP
TANK
~l
CiSOs4CiS04
TO WASTE
Figure 3. Scrubber Addition of Lime - SO, Removal
145
-------
the scrubber with the flue gas. Problems with this process include
boiler fouling, inactivating the lime by everburning, and increased
scaling in the scrubber at the dry-wet interface. This approach is
no longer offered by the SO, removal system supplier who originally
Ct
offered it, unless requested by the customer.
Scrubbing efficiency can be increased in the third process approach
by using lime in the scrubber circuit. This process has two dis-
advantages — the higher cost of lime over limestone and the greater
potential for scaling under certain process configurations.
3.2 SCRUBBER DESIGN
In designing a scrubber system, the following main design criteria
must be considered:
• Sulfur content of fuel
• Ash content of fuel (Is fly ash removed separately?)
• Percent SO removal required
£
• Scrubber type
• System turndown
• Scrubber size and spare capacity
• Mist separators
• Gas reheat
• Waste disposal
• System power losses
• Materials of construction
146
-------
3.2.1 Sulfur and Ash Content of Fuel
The sulfur content of the fuel burned in the power plant determines
the required SO removal to meet the new performance standards
(1. 2 Ib SO /million Btu for coal and 0. 8 Ib SO /million Btu for
Lt w
oil). Three-percent sulfur coal requires approximately 80 percent
removal of SO from flue gas, whereas 0. 8 percent sulfur coal
Lt
(depending on heat value) requires 40 percent removal. Pilot plant
data have shown that SO removal from low sulfur flue gas is less
L*
difficult for the same liquid-to-gas ratios (40-80) because the amount
of SO absorbed per pass is significantly less. This can be seen by
comparing inlet SO values of 2400 ppm (high sulfur coal) to 600 ppm
Ct
(low sulfur coal. The SO absorption per pass is 1920 ppm for the
£»
high sulfur case and only 240 ppm for the low sulfur case. Control
of scaling and supersaturation in the scrubbing systems is more
predictable and thereby provides greater system reliability. In
addition, the lime or limestone requirements are significantly less
for low sulfur fuels. Waste disposal sites also require less land.
If the fly ash is removed from the flue gas before SO removal is
accomplished, the scrubbing system design can be different. This
is particularly true as to scrubber selection. A spray column could
be used to obtain 80 to 90 percent SO removal with lower gas pressure
Lt
drop. Waste solids disposal is also influenced by this design con-
sideration.
If SO2 and fly ash are removed simultaneously, a venturi followed by
an afterabsorber, two stages of venturi, or turbulent contact absorber
147
-------
(multistage) offer good possible choices for high removal. Waste
solids requirements are higher, and process chemistry becomes
more complicated.
3.2.2 Scrubber Type
The major criterion for scrubber selection is its capability to remove
both sulfur dioxide and particulates with high efficiency (SO removal
greater than 80 percent and particulate removal greater than 99 per-
cent). Other factors considered are ability to handle slurries without
plugging, cost, control, and pressure drop.
Additional variables are:
• Liquid-to-gas ratio
• pH control
• Stoichiometric ratio of alkali to SO
C*
• Percent of solids in slurry
• Limestone particle size
• Number of stages
The scrubber types that have been tested to date include the following:
Venturi. The venturi scrubber (Figure 4) has been used
when both particulate (fly ash) and sulfur dioxide must
be removed. The venturi has good capability to remove
fly ash down to 0. 02 gr/SCF with pressure drops of 10
to 15 inches HO and liquid-to-gas ratios of 10 to 15 gpm
per 1000 cu ft gas for typical dust loadings and particle
size distributions from power plant stack gases. The
148
-------
TWO-STAGE VENTURI
(CHEMICO)
UMIIOROUE DRIVC
FOR PIUMI K»
GASINUT
KCYCU
GASOUtllT
ONE-STAGE VENTURI
(CHEMICOI
"PIUMIIOS"
INUTS
PIUMI 101 SHAFT
MUiriPU
rANCENTIAl UMTS'
CAS OUTUT
HOT CASINUT
SPRAY NOZZLE IARR&
IFOR PIUMB BOH
Zntf. STAG!
•.MIST SEPARATOR
SPRAY
•DENTAL KHK."
"PLUMB 101"
MIST SEPARATORS
VENTURI UNIT WITH AUTOMATIC THROAT AREA CONTROL
-------
venturi may contain an adjustable throat area that
permits control of pressure drop over a wide range
of flow conditions. The venturi scrubber is limited
in SO removal to 40 to 50 percent per stage with
lime-lime stone due to the short liquid residence time.
It therefore requires two stages of Venturis or an
afterabsorber to achieve 85 to 95 percent SO removal
from high sulfur fuels at total liquid-to-gas ratio of 80.
Turbulent Contact Absorber (TCA). The TCA (Figure 5)
is a countercurrent multistage scrubber consisting
of screens that both support and restrain the plastic
spheres. The spheres move in a turbulent fashion
providing good gas-liquid contact and scale removal.
The number of stages generally are between two and four
for high SO removal with liquid-to-gae ratio of 40 to
50. The pressure drop per stage is approximately
2 to 2. 5 inches HO.
Lf
Marble-Bed Absorber. The marble-bed absorber
(Figure 6) utilizes a 4-inch bed of packing of glass
spheres (marbles) that are in slight vibratory motion.
A turbulent layer of liquid and gas above the glass
spheres increases mass transfer and particulate
removal. This scrubber has been used mainly in
the process where the limestone is added to the furnace
for calcining, and the flue gases are scrubbed to
remove SO . Pressure drop is generally 4 to 6 inches
HO. Liquid-to-gas ratios of 25 to 30 have been used.
Packed-Bed Absorber. The packed-bed absorber must
use open packing to prevent plugging. Packed towers
have been tested in a pilot plant with high SO removal
and no significant scaling with low sulfur coal. Pressure
drops are low — 0. 4 inches HO per foot of packing.
Scale control is extremely important in this type of scrubber,
requiring liquid-to-gas ratios of 30 to 60.
Spray Colurnn. The spray column (Figure 7) is a counter-
current type scrubber that has a low pressure drop. The
spray tower requires high liquid-to-gas ratios (L/G=80)
and several stages of sprays to achieve high SO removal.
Lt
150
-------
RECIRCULATION
OUTLET NOZZLE
Figure 5. Turbulent Contact Absorber for SO.
and Fly Ash Collection (UOP)
DEMISTER
DEMISTER SPRAY
DRAIN POT
TURBULENT LAYER
MARBLE BED
UNDER BED SPRAY
GAS OUTLET
GAS INLET
Figure 6. Marble-Bed Absorber
151
-------
GAS OUTLET
HOT GAS
INLET
DEMISTER
DEMISTER
WASH SPRAY
SPRAY
SPRAY
Figure 7. Spray Coltimn
152
-------
It has been tested by TVA, Peabody Engineering/Detroit
Edison, and Ontario Hydro, The spray column is being
tested at the EPA Test Facility, Paducah, Kentucky.
Results also will be reported in a separate paper today
(Ret. 5).
• Tray Column. The tray column offers high liquid hold-up
and high SO removal at relatively low pressure drop.
The main disadvantage is the potential for scaling. High
liquid-to-gas ratios (L/G = 40) are required. In addition,
undersprays are used to wash off soft scale.
• Cross Flow Absorber. The cross flow absorber has a
short gas path with the scrubber installed in a horizontal
position. It has a low pressure drop and has been tested
with packing or sprays. It requires a higher L/G ratio
to obtain high SO2 removal.
• Screen or Grid Scrubber. The screen or grid scrubber
has been recently tested by TVA. It contains five to
ten screens (7/8-inch openings). A liquid-to-gas ratio
of 50 and a stoichiometric ratio of 1. 5 give SO removal
of 75 to 80 percent. Low pressure drops were observed.
3.2.3 Scrubber Size, Turndown, and Spare Capacity
To keep the SO removal system investment as low as possible,
2 3
scrubber sizes of 500, 000 ft /min have been developed. This is
required because the flue gas from a power plant is high — 3, 000
ACFM/MW. An additional design consideration is scrubber turn-
down, due to changes in boiler load. Some scrubbers can be turned
down to 50 percent of design and others require compartmentalization.
In order to insure greater system reliability, spare capacity should
be installed. Thus, a 500-MW unit could be designed with four
modules (no spare) or five modules to insure backup capacity and
system reliability. An economic analysis of a five module 500-MW
unit will be presented later in this paper.
153
-------
3.2.4 Mist Separators
Mist separators come in various designs, but are typically of open
construction and low-flow resistance, similar to cooling tower packing.
The techniques under development to control mist eliminator plugging
employ such principles as:
VELOCITY
REDUCTION IN
TCA
VS,
Mini
VELOCITY
REDUCTION IN
VENTURI
Reduction in upward
gas velocity to 8 feet
per second or less, at
the same time attempt-
ing not to disturb the
uniformity of gas
distribution
WASH
ROW
RATE
n n n
TIME
vs.
MASH
FLOW
RATE
nnnnn
TIME
More frequent back-
washing, "which may
upset water balance
in closed operations
IU
vs.
A A A
Y/////////////S
Y YY
More efficient back-
washing, for example
employment of a tra-
veling retractable
washer similar to a
conventional soot
blower, in place of
stationary sprays, or
washing from both
sides
154
-------
vs.
THREE - PASS
SINGLE - PASS
Simplification of the
mist separator design
itself with possible
sacrifice of entrain-
ment removal
efficiency
tttt
WATER.
VS.
MM
rywvwyyx:
n I
Installation of an inter-
mediate wash tray to
lower the slurry solids
entering the separator
DRAIN FLOW
j fttt
VS.
•ROW
1
DRAIN
Reorientation of the
separator to "cross-
flow" conditions with
the gas flowing hori-
zontally and the sepa-
rators draining verti-
cally at 90 rather
than countercurrent
at 180°
vs.
'.'.'i
DRAIN
Wet precipitation
with a variety of
configurations
155
-------
It seems likely that a satisfactory answer to the problem of mist
separator plugging will soon be found, and that it will consist of
one or more of these measures.
3.2.5 Stack Gas Reheat
As the gas passes through a wet absorber, it will be humidified and
cooled to the adiabatic saturation temperature (normally 120 to 130 F).
Some water droplets may be present since the eliminator does not
remove all of the mist formed in the scrubber. If the scrubber exit
gas is not reheated, it will lose some of its buoyancy as a result
of cooling. Thus, the effective stack height and the plume dispersion
will be reduced. This should not be serious if a high degree of SO
removal is achieved. However, due to the low efficiency of nitrogen
oxide removal, the ground level concentration of NO may become
a serious problem, especially for existing plants where the stack
height is already fixed. For new plants, this effect must be taken
into consideration in stack design.
Humidification of the stack gas is also objectionable because condensa-
tion may occur and cause formation of a visible plume. This acidic
plume may also create a corrosive environment around the stack area,
although the stack itself is corrosion resistant.
The obvious solution is to reheat the scrubber exit gas. There is general
agreement that the saturated and cooled gas should be reheated, but
there is no agreement on the degree of reheating necessary or desirable.
The consensus seems to be that it is not necessary to reheat it to the
156
-------
original temperature (too expensive), but that it should be heated at
least to a high enough temperature to minimize the water-vapor plume
under most atmospheric conditions. This, of course, depends on the
local meteorological conditions, but 50 F is the most commonly used
reheat design figure at present.
The methods of reheating the stack gas may be divided into direct
heating and indirect heating.
Direct Heating. The major advantage of direct heating is operational
reliability, because there is no heat transfer surface on which fouling
can occur. Other advantages are low investment, flexibility in degree
of reheat in some cases, low pressure drop, and low maintenance.
There are several methods by which direct heating can be accomplished.
Direct-Fired Reh^f^i. Either natural gas or low- sulfur
fuel oil may be used, depending on availability of fuel
and cost. Natural gas is the most expensive fuel, but it
is convenient and clean. Oil would be less expensive,
but receiving, storing, and handling would be additional
problems. Complete combustion of oil in contact with
wet gas may also be a problem. The addition of sulfur
dioxide and ash to the treated gas should be considered.
Flue Gas Bypassing. Bypassing the scrubber with
part of the gas stream — followed by mixing this gas with
the scrubber exit gas — requires minimum investment
and adds essentially no operating cost. However, this
is possible only when the overall SO2 removal require-
ment is low enough to allow such bypass, i. e. , 50 to 60-
percent removal or less. This may be possible for low-
sulfur coal-fired boilers when the scrubber is preceded
by a precipitator, but it is impossible for high-sulfur
coal application.
157
-------
Hot Air Injection. If the air preheater is designed to
produce excess hot air over the amount required for normal
boiler consumption, this hot air may be used to mix with the
scrubber exit gas to obtain the degree of reheat required.
In doing so, the heat efficiency of the boiler will, of course,
be lowered somewhat. Alternately, air can be preheated
in a separate exchanger (with steam for instance), then
mixed with the scrubber exit gas. This alternative gives
the clean, nonfouling service expected, but it consumes more
steam, because of sensible heat losses in the air, compared
to direct-steam-to-stack-gas exchange. The added cost might
be justified on the basis of reliability.
Indirect Heating, Indirect heating of the stack gas requires an exchanger
to transfer heat from the heating medium to the gas. The advantage of
this method is usually low operating cost, especially when heat from the
inlet gas is transferred to the scrubber exit gas (recuperative heat ex-
change). The main disadvantages of this method are a higher gas pres-
sure drop, higher investment, and possible fouling on heat transfer sur-
faces. If this method is chosen, finned tubes should be avoided, and de-
vices to keep the exchanger surface clean (such as a soot blower) may
be required. The following are alternative indirect reheat methods:
Reheat with Steam. The exit gas may be heated with
steam from the turbine cycle in a heat exchanger at the
scrubber outlet. This method would require additional
coal firing in the boiler to generate the extra steam and
modification of the turbine to allow higher than normal
extraction rates. In a new plant, a system could be de-
signed to provide the steam required. The additional
coal firing would still be more economical than direct-
gas or oil-fired reheater.
Recuperative Reheat System. A heat exchanger may be
used for direct transfer of heat from the scrubber inlet
gas to the exit gas. With this method, heat that would
158
-------
be wasted is recovered. The disadvantages are the
large heat exchanger required (because of low-transfer
coefficient and low-temperature differential), high pres-
sure drop, and possibility of fouling. Corrosion would
also be a problem when the inlet gas is cooled to below
the acid dewpoint. A cyclic-liquid heat exchange sys-
tem with heat transfer from the inlet gas to water and
from the water to the scrubber exit gas would permit
use of smaller exchangers than those required for gas-
gas exchange, and the smaller surface would reduce
pressure drop. Other disadvantages remain, however.
The choice of a reheating method is not entirely an economic
consideration; the system reliability must be considered as well.
3.2.6 System Power Losses
Wet scrubbing systems have significant energy requirements in
terms of pumping and fan losses. When particulatc removal is not
required, the energy requirement is reduced substantially. Figures
8 and 9 show the order of magnitude to expect for fan and pumping
losses from these systems. To these must be added an allowance
for discharge pumps, small auxiliaries, lighting, and instrumentation.
Certain assumptions are incorporated in these figures, and the user
is cautioned to consider results obtained from them as preliminary esti-
mates only. The assumptions made in their construction are as follows:
• Compressibility effects for the fan are negligible.
• Fan efficiency is 60 percent.
• Motor efficiencies are 93 percent.
159
-------
u
§
«/>
O
at
O
13
1
5 —
4 —
3—
1—
' VENTURI FOLLOWED BY SPRAY TOWER
'TCA
' SPRAY TOWER
I
10
30
I
40
50
I
60
PRESSURE DROP,INCHES OF WATER
Figure 8. Station Electrical Loss as a Function
of Draft Requirements
Pump efficiencies are 85 percent.
Scrubbing liquor solids are 10 percent,
Liquid flow is turbulent. The sum of potential and friction
losses is 80 velocity-heads. (This is typical of 80-
foot elevation, plus pipe and fitting losses, with an
allowance for control-throttling losses. )
There is single-stage absorption.
Flue gas generation was assumed to be 3000 actual
cubic feet per minute per megawatt.
160
-------
5—1
£ 4—
o
O£
O
u.
o
I/)
(/)
2
o
z
i
1
20
1
40
1
60
1
80
1 1
100
1
120
1 1
140
1 1
160
l/G RATIO,GAl/1000 ACFM (INLET)
Figure 9. Station Electrical Loss as a Function of
L/G Ratio and Nozzle Pressure
3. 2. 7 Materials of Construction
Any part of a wet scrubbing system that is in contact with wet SO gas
&
or acidic scrubbing liquor should be constructed of acid-resistant mate-
rial. If solids are also present in the system, abrasion-resistant mate-
rial should be used. The material specified should be able to withstand
the highest temperatures encountered during normal operation and up-
set conditions.
161
-------
Stainless steel 316L is sometimes used for scrubber construction.
However, if chloride is present, stress corrosion will be a problem.
Chloride stress corrosion is frequently characterized by so massive
a failure that the affected metallic part must be totally replaced. A
common form of prevention is to provide an upgraded alloy (e. g. , Alloy
20, Incoloy, or Inconel), but an alloy demands a premium cost consistent
with its higher nickel content.
Elastomer-lined (soft rubber or neoprene) carbon steel can be effective
under abrasive conditions and temperatures up to about 175 F. The
temperature must be well controlled to protect the lining. Field rubber
lining costs roughly four dollars per square foot. A further advantage
of these linings is the lower maintenance costs in descaling. The
scale-substrate bond is more readily broken than in the case of
stainless steel or alloy metals.
Rubber-linings applied in the shop have relatively strong vulcanized
bonds because they are autoclave cured. Field linings employ chemical
curing agents rather than heat, and their quality is not always compara-
ble. Ambient temperatures must not fall below 70 F during the curing
period. Vessel components with short-axis dimensions exceeding 10
to 12 feet cannot be shipped by rail, hence, they must be field-lined.
A serious problem area inside scrubbers is the zone where hot gases
first contact the liquor medium, because temperature control is some-
times difficult. An acid-proof gunite is now being offered by Pennwalt
for protection of these areas. The gunite is bonded directly to the steel
with thickness from 0. 5 to 1. 5 inches. The gunite will withstand tem-
162
-------
peratures up to 750 F,and the applied cost is $4 to $10 per square foot.
Gunite can also be used for duct and stack lining. In using gunite, the
added weights and stresses must be considered.
Another lining material of promise is glass-flake reinforced-polyester
resin, such as the Ceilcote 100 and 200 series and Carboglass 1601.
This material is trowel-applied at 40 to 80 mils thickness on carbon
steel and can withstand temperatures up to 200 F under conditions of
continuous operation. The applied cost is $3 to $4 per square foot. Its
abrasion resistance improves when well wetted, and it can be field-
patched with relative ease, compared to elastomeric linings.
3. 3 WASTE DISPOSAL
The nature of calcium sulfite and calcium sulfate solids leaving the sys-
tem has created a problem that remains incompletely defined. Judging
from the experience at Mitsui Aluminum and at Will County Station
No. 1, there seems to be some indication that the solids from limestone
systems may not settle as completely in the pond as those from lime
systems.
The limestone solids at Will County have been observed to settle only
to 35 percent. Under such conditions they have no bearing strength,
but behave like quicksand. This not only increases the acreage require-
ments, but it also creates a possible hazard to stray animals and
wildlife.
163
-------
One possible solution is to install supplemental mechanical dewatering
equipment. Another is to use lime in place of limestone, although at
greater cost. Also, the use of coagulants to improve settling is a possi-
bility. Furthermore, it is not clear whether the lack of consolidation
originates from the use of calcium carbonate (as opposed to calcium hy-
droxide), or whether it can be primarily attributed to excessive size
reduction of the limestone feed. Until this is established, limestone
systems should have built-in turndown on their size reduction circuits.
The determining factor in size reduction is the classifier circuit, com-
prised of liquid cyclones. These return the oversized material to
the ball mill, but reject the material of requisite fineness so that it
passes on to the process.
It has been found very difficult to reliably simulate the properties of
these solids on a laboratory scale. The reasons for this appear re-
lated to ionic strength — i. e. , the slow buildup of minor soluble salts
in the circulation slurry that alters a solubility relationship of its
principal constituents. Invariably laboratory slurries turn out to be
faster settling than encountered at full scale. Further detailed studies
based on pilot operation are required,
The waste sludge produced in the lime/limestone scrubbing process
requires special study for proper management. The sludge consists
primarily of CaSO3 • 1/2H2O, CaSO4 • 2H2O, CaCO3 or Ca(OH), and
fly ash. The quantities are significant, 500,000 tons/yr, 50-percent
solids for a 500 MW power plant burning 3-percent sulfur coal. The
principal method for most of the 20 full-size installations is disposal
in a pond located on the power plant site. However, there are many
plants that have insufficient land and therefore must condition the
solids for disposal some distance from the plant.
164
-------
Suitable areas that are ecologically sound must be found. Studies
should be carried out to determine the feasibility of disposing of
these solids in the mines where the coal was produced. The unit
trains that transport the coal to the power plant could logically carry
the sludge back to the mines. The sludge must be dewatered so that
it can be transferred easily to and from the trains. Studies shoxild
bo funded to evaluate this concept. Other sites that should be studied
are old limestone quarries and salt domes. These should be accept-
able from an ecology point of view. The large disposal sites might
be operated by state agency and a fee be charged for disposal.
Another important disposal technique used where land is not available
at the plant site involves dewatering the sludge and transporting it by
truck or barge to a suitable landfill. Further studies are necessary
to better understand this technique.
Chemical fixation processes are also being developed. These pro-
cesses generally involve pozzolanic (cementitious) chemical reactions
requiring the presence of lime. Such reactions lead to the formation
of dry solids with potential landfill and reclamation applications. This
technique requires additional study to define optimum conditions.
Lirnitr.-d research on the use of fly ash and limestone-modified fly ash
has boon performed. Additional research is required to explorr llu>
utilization of lirmj/limestone sludge to perhaps 25 percent of tin- quan-
tity produced. In Japan, the sulfite is usually oxidi/.ed to sulfate and
the j/ypsum produced is used for construction. This potential markcl
should be further explored in the UnitedStates.
165
-------
The Coal Research Bureau of West Virginia University has been per-
forming research under partial support from the Environmental Pro-
tection Agency to develop and evaluate utilization of such a solid-waste
fly-ash mixture. As a result of this research, several possible areas
of utilization have emerged. These include production of autoclaved
materials such as calcium-silicate brick or block, aerated or foamed
concrete, and cement materials.
Based on the above comments, it is evident that further study is criti-
cally needed to define a sound plan of solid waste management.
166
-------
Section 4
STATUS OF FULL-SIZE INSTALLATIONS
Although process development for SO removal using lime /limestone
scrubbing continues at an intensive rate, the U.S. industry has decided
to install full-size units to meet air quality regulations and to accele-
rate technology. To date, there are 21 commercial-size installations
using lime or limestone scrubbing (see Table 1). There are also four
scrubber systems for particulate removal only (not listed in Table 1)
that have contributed to the state of the art. Table 1 gives the opera-
tional status and design criteria of 21 installations; 11 units were retro-
fits and 10 installations are planned for new power stations. The larg-
est facilities are the Mansfield plant of Ohio Edison (two 900 MW -
3 percent S coal), and the Navajo plant of Salt River Project (three
750 MW - 0. 5 to 0. 8 percent S coal). These units are designed for alkali
addition into the scrubber circuits.
Other points that should be emphasized are:
• Units Operating — 8
— Alkali furnace injection — 3
— Alkali into scrubber circuit — 5
Total MW installed - 1900
167
-------
• Units being engineered and constructed— 12
— Alkali furnace injection— 0
— Alkali into scrubber circuit— 12
- Total MW to be installed - 9600
In 1968 and 1969 at Union Electric's Meramec Station and Kansas Power
& Light's Lawrence Station, the first full-size units were installed.
Both employed finely pulverized limestone injected into the boiler fur-
nace followed by wet scrubbing utilizing a marble bed scrubber.
Operating experience revealed a number of problems that included
scaling and corrosion. A typical flow diagram for this process ap-
proach is shown in Figure 10. The system at Meramec has been aban-
doned, and the Lawrence Unit has been substantially modified. A
second unit at Lawrence Station was installed and started up in
December 1971. Reported SO removal is 70 percent at a stoichio-
(L
metry of 0. 7.
The first full-size limestone scrubbing system in this country utilizing
alkali addition into the scrubber liquor circuit was installed as a retro-
fit unit at the Will County Station of Commonwealth Edison. The unit
involved is a 180-MW cyclone boiler with a two-stage venturi/absorber
to remove fly ash, followed by SO . More has probably been learned
from this installation about practical operating problems with lime-
stone during its one year of operation than from any other pilot plant to
date. The major problems experienced are maintaining clean de-
misters, system reliability due to mechanical problems, and disposal
of waste solids. SO removal efficiencies reported were 75 to 85 per-
£•
cent. A typical flow diagram for this process is shown in Figure 11.
Table 2 presents preliminary test data for 1972 (Ref. 7).
168
-------
en
vo
Table 1. OPERATING AND PLANNED FULL-SIZE LIME/LIMESTONE DESULFURIZATION
FACILITIES IN THE UNITED STATES
Utility Company/Plant
1. Union. Electric Co.
9. Kansas City Power & Light/
La Cygne Station
10. Arizona Public Service Co. /
Cholla Station
11. Duquesne Light Co. (Pittsburgh)/
Phillips Station
12. Detroit Edison Co. f
St. Clair Station No. 6
13. Ohio Edison/Mansfield Station
14. Tennessee Valley Authority/
Widow's Creek Station No. S
15. The Montana Po-wer Co. /
Colstrip Units 1 & 2
16. Northern States Power Co.
(Minnesota)/
Sherburne County Station
No. 1 & Z
17. Northern Indiana Public Service
Co./
Kankakee 14
18. Mohave-Navajo/
Mohave Module (Vertical*)
19. Mohave-Navajo/
Mohave Module (Horizontal?)
20. Salt River Project/
Navajo Station
21. Southern California Edison/
Mohave Station (Horiz/Vert)
Absorbent
CaO
CaO
CaO
CaCO,
3
CaCO
CaO
CaO
Ca
-------
S1ACK
1.0, FAN
- STACK (.AS RtHIAItfi
DRAI!>i POTS
RFCYCU
MAKE UP 1VAUH
TO ASH
DISPOSAL POkD
Figure 10.
Particulate and SC>2 Removal System Employing
Marble Bed with Limestone Calcining in the Boiler
(Source: Combustion Engineering)
170
-------
Table 2. WILL COUNTY UNIT 1 WET SCRUBBER TEST DATA ("A" SCRUBBER, MAY 1972}
Test Number
Date
Load, MW
Gas Flow, 103 CFM
Scrubber System, Pressure
Difference, Inches HO
Dust Inlet, gr/SCFD
Dust Outlet, gr/SCFD
SO_ Inlet, ppm
SO Outlet, ppm
SO Removal Efficiency, %
Absorber Slurry Density, %
Absorber pH
1
5-18
113
335
24.5
—
.0232
1145
67
94
3.4
6.5
2
5-18
113
335
29
.0944
.0079
1140
75
93
5.2
6.3
3
5-19
114
335
21
. 1440
.0073
890
294
67
5.5
7.4
4
5-19
115
340
25
.1470
.0298
930
35
96
5.2
6.3
5
5-20
111
335
24
. 1105
.0261
1130
285
75
2.5
5.7
6
5-20
112
320
25.5
. 1790
.0255
1000
118
88
4.3
5.8
7
5-21
113
315
22.5
—
—
640
18
97
5.0
7.2
8
5-21
115
310
22.0
-
-
910
45
95
-
5.7
9
5-22
110
315
23.2
.3060
.0205
1000
223
81
2.9
5.9
10
5-22
111
335
23.0
.2580
.0334
545
180
67
2.2
5.4
11
5-23
205
16.0
—
—
1200
45
96
-
6. 1
12
5-23
58
215
18.0
-
-
1150
50
96
1.5
6.1
-------
""'k^-il
fU e
LIMESTONE
MAKE-UP | | IUHKU
WATER
I.D. IOOSTE«
y, '*"
STEAM
RIHEATER
'TO SETTUH6
POND
Figure 11
g
Will County Station Unit No. 1 (Limestone Added
to Scrubber - Venturi plus TCA Absorber)
172
-------
The first scrubbing system utiliaing lime introduced into the scrubber
circuit has been installed at the Phillips Station of Duquesne Power
and Light. It is presently going through startup and should supply
valuable information when in full operation later in 1973, This unit
is reported to be designed based on the Mitsui Aluminum unit in Japan,
which uses a calcium hydroxide scrubber (Ref. 8). To date the Japanese
system is the most successful operating unit based on a throwaway pro-
cesa. It has been operating since March 29, 1972 without any significant
downtime. SO- removal efficiencies have been reported to be 80 to 90
percent. A typical flow diagram for this process is shown in Figure 12.
The Hawthorne Units 3 and 4 of Kansas City Power & Light are pres-
ently in the first stages of operation. One uses alkali injection into
the boiler furnace and the second has been modified to inject lime
into the scrubber.
The Paddyrs Run Station of Louisville Gas & Electric is the first plant
to test carbide sludge [Ca(OH)2]. Test results based on a relatively
short period of operation indicate high removal of SO
£t
Kansas City Power fc Light's LaCygne plant is presently in startup
and should be in full operation during the next 60 days.
When most of the 20 units are operating, significant progress toward
commercial demonstration will be realized.
173
-------
I.D. FAN
FLUE GAS
BOILER
I
RE
FUII
DEMISTERS
—1 ; n
Ui: !
MAKEUP SLURRY
TANK
r: :z:._
j MAKEUP SLURRY
FEED PUMF
BLEED SLURRY
TRANSFER PUMP
~i
DRY CARBIDE PIT
RECYCLE SLURRY
MAKEUP SLURRY
BLEED SLURRY
_ RETURN LIQUOR
ASH POND LIQUOR
RETURN PUMP
\
WASTE DISPOSAL POND
Figure 12. Chemico /Mitsui Flue Gas Desulfurization
System (Lime Added to Scrubber)
174
-------
Section 5
LIME/LIMESTONE COMMERCIAL OPERATING CONDITIONS
OFFERED BY SO2 REMOVAL SYSTEM SUPPLIERS
Suppliers for SO_ removal systems are growing steadily. A review
£
of the market reveals seven suppliers with full-scale experience.
There are five other companies that have just recently entered the
market.
The seven companies with full-scale experience were contacted to ob-
tain design information concerning SO removal guarantees related to
L*
percent sulfur fuel, scrubber type, liquid-to-gas ratio, stoichio-
metry, percent solids in slurry, alkali type, hold tank residence time,
and any qualifications concerning system design. The results of this
study are presented in Table 3.
It is interesting to note most companies will supply systems using lime
or limestone as the alkali and provide SO removal guarantees that
£*
vary from 70 to 90 percent, respectively, or as required to meet per-
formance standards. Important operating parameters such as liquid-to-
gas ratio and stoichiometry have changed in design range. Liquid-to-
gas ratio has increased from 20 to as high as 100, the average being 60.
Stoichiometry has decreased from 1.75 to as low as 1.0 based on SO
absorbed, the average being 1.2. There appears little restriction on
the percent sulfur in the fuel as it relates to guarantees or size of units.
175
-------
Mc,dui<; designs, as largo as ISO megawatts each, for units up to 800
megawatts, are provided by U.S. suppliers. All process designs are
based on alkali addition to the scrubber circuit.
When schedule was discussed, most suppliers stated a minimum time
schedule of 24 months from purchase order to startup. A preferred
time period would be 30 months, depending on backlog. The time
required for a bidder to respond to a functional specification ranges
from 4 to 10 weeks, with 8 weeks being typical. An additional period
of 8 weeks is required to review proposals, select supplier, and con-
duct preaward negotiations. Based on the above requirements, the
total time period is approximately 36 months. In addition, pilot
plant testing to evaluate limestone reactivity is desirable.
176
-------
Table 3.
OPERATING CONDITIONS OFFERED BY S02 REMOVAL SYSTEM SUPPLIERS
Supplier
Chemical
Construction Co.
Combustion
Engineering
Babcock & Wilcox
Peabody
Engineering
UOP - Air Correction
Combustion
Equipment Assoc.
Research Cottrell
Scrubber
Venturi 2 stage
spray c olumn
1 or 2 marble beds1*1
(bottom spray only)
-Low pressure drop
quencher plu« tray
absorber
Venturi plus
spray column
Turbulent Contact
Absorber (3 stages)
Venturi (if particulate
included) (2 stage)
spray column
(S02 only)
Multicontact
absorber
Banco systenv'
Alkali
Lime
Limestone
Lime
Limestone
Carbide sludf
Lime
Limestone
Carbide sludj
Lime
Limestone
Lime
Limestone
Lime
Limestone
Limestone
Lime
Liquid to Gas
Ratio, gal/MCF
40-80
80
25-30
25-30
e 25-30
40-50
40-50
e 40-50
(b)
(c)
40
40
60
80
ffl
90lfl
Full Size
Yes
Yea
Yes
Yes
Yes
Yes
Yes
a. Furnace injection of akali not offered
unless requested.
b. L/C = 20 for venturi, 50-60 for spray
column with lime.
c. L/C * 80 for »pray columnwth
limestone.
d. L/C (1 st stage) = 50-60; L/C (2nd
stage) = 15-30.
t. Research Cottrell has no restriction on system design or guarantee
for units using fuel with 2 percent or less sulfur. For fuel with
greater than 2 percent sulfur, they require a 100-MW demonstra-
tion unit before building a full-size (i.e., 800 MW( installation.
f, Bahco is limited to 100,000 CFM, or 35 MW, and offers lime-
stone systems.
-------
Section 6
CAPITAL AND OPERATING COSTS OF FULL-SIZE
SO REMOVAL SYSTEMS
LJ
In order to develop capital and operating costs, Bechtel cost estimates
were used as a basis. The basis for design of lime/Limestone scrub-
bing systems was a "typical" new 520-megawatt boiler burning high
sulfur coal (3 percent S). SO removal required is approximately 80
£r
percent. Particulate removal proceeded the SO removal. A spare
LJ
train was provided for system reliability, and a pond is available for
disposal of the waste solids.
It has generally been the practice to cite capital costs on a "per kw"
basis, whereas a "per CFM" cost basis is more meaningful across the
range of combustion conditions encountered. Economics of scale pro-
vide unit savings for scrubbers up to 500,000 CFM. Above this size,
duplication of scrubber trains is necessary, and cost savings cease
to apply.
A common reason for high "per kw" cost is the installation of a complete
scrubber train with isolation dampening so as to make on-stream main-
tenance. For a five-train system, this is equivalent to 25 percent of
installed cost. This practice is not always necessary, being depen-
dent on utility load balancing as well as on reliability and proven
178
-------
maintenance record for the particular system employed. Other rea-
sons for "high" costs per kw are due to high contingencies for system
modifications and high escalation rates.
It should be noted that no costs were shown for real estate deeded over
to sludge ponding. Instead, a flat charge of $2 per ton of wet sludge
(40-percent solids) was allowed. A slight advantage was given to lime
sludges in that it was assumed the solids would settle to 50-percent
solids. In addition, it should be emphasized that disposal costs are
highly site-sensitive. If the sludge must be treated for disposal off
site, the charge per ton could double to $4. 00 and thus increase opera-
ting costs by 15 percent.
Capital costs have been divided into direct costs and indirect costs.
Direct costs include yard facilities, raw material handling, scrubber
system, and sludge handling. Indirect costs include field costs, engi-
neering, home office, escalation, and fee. Estimated capital cost for
a "typical" new 520-megawatt unit using limestone scrubbing is
$23, 400, 000 or $45 per kw. The total installed cost is $ 1 0. 70 per
ACFM. The estimated cost per kwh is 2.26 mills, based on limestone
at $5. 00 per ton delivered (stoichiometry at 1. 5 and a sludge charge
of $2. 00 per ton). If limestone can be purchased at $3. 00 per ton, the
annual cost would be 2. 15 mills/kwh. Based on 10,000 Btu/lb coal,
the cost per million Btu is estimated to be 23£, If sludge disposal cost is
$4.00 per ton, the annual operating cost would be 26£ per million Btu.
Tables 4 and 5 summarize these costs and give the basis for the estimate.
The estimated capital cost for a "new" 520-megawatt lime scrubbing
system is $19,600,000 or $38/kw. This is equivalent to $9/ACFM.
179
-------
The estimated cost per kwh is 2. 16 mills based on lime at $20 per ton
delivered. Sludge disposal costs are reduced due to less sludge pro-
duced because of lower stoichiometry and better settling of solids to
50 percent by weight. Annual cost for SO removal is estimated to be
per million Btu. Tables 6 and 7 summarize these costs
Retrofit costs for SO removal are greatly affected by specific plant
C*
conditions. Therefore, no effort has been made to present costs for
retrofit units. More detailed information on cost considerations can
be found in a paper presented by Burchard (Ref. 9).
Table 4
CAPITAL COST - LIMESTONE SCRUBBING
(a)
Cost Item
Yard facilities
Raw material handling
Scrubber system
Sludge handling (on- site)
Direct costs, subtotal
Field costs, engineering
and home office, escala-
tion, contingency, and fee
Total installed
Thousands of
Dollars
648
2, 347
7, 427
755
11, 177
12,253
23,430
Dollars/kw
45.05
Dollars/ AC FM
Installed
10.70
a.
520-MW gross output; 3. 1-percent S coal; 81-percent SO2 removal;
particulate matter removed by others; 80-percent load factor; with
ponding; spare train
180
-------
Table 5
ANNUAL OPERATING COST - LIMESTONE SCRUBBING
(a)
Coat Item
Basic alkali, tons
Process -water, gal x 10
Power, demand, kw
Power, energy, kwh x 10
Steam, demand, annual Btu x 10
Steam, energy, Btu x 10
Operating labor, 3/shift
Operating labor overhead at 75%
Maintenance at 5% of direct coat
Maintenance overhead at 30%
Capital recovery, 15 yr.
7-1/2% f. 1315 multiplier)
Insurance and property taxes
at 1.9%
Sludge disposal, tons
Total Cost
Total
Requirement
205,000
461,000
7,940
53,400
800,000
615,000
Z5.200
659,000
Range of
Unit Cost
Experience
$ 3.00-9.00
$ 0.02-0. 15
$15.00-40. 00
$ 0.90-5.00
$ 0.08-0.25
$ 0.10-0.55
-
$ 0.75-5.00
Unit Cost,
This Study
$ 5.00
$ 0.03
$35.00
$ 5.00
$ 0.20
$ 0.55
$ 6.25
$2.oo(b)
Annual
Cost,
$Thousand
1,025
9
278
267
160
338
153
118
557
167
3, 081
444
1,318
7,920
Cost
per kwh,
mills
0.293
0.003
0.078
0.076
0.046
0.097
0.045
0.034
0.159
0.048
0.881
0.127
0,377
2.26
a. 520-MWgroes output; 3. 1-percent S coal; 81-percent SO removal; particulate removal by others;
80-percent load factor; with ponding; spare train
b. Waste solids are assumed to be 40 percent by weight dry solids. At $4 per ton for sludge disposal,
annual cost increases to $9,238,000 or 2.64 mills per kwh
181
-------
Table 6
CAPITAL COST - LIME SCRUBBING
(a)
Cost Item
Yard facilities
Raw material handling
Scrubber system
Sludge handling (on- site)
Direct costs, subtotal
Field costs, engineering and
home office, escalation,
contingency, and fee
Total installed
Thousands of
Dollars
555
592
7.427
755
9,329
10, 262
19,591
Dollar s/kw
37.68
Dollars/ AC FM
Installed
8.96
a.
520-MW gross output; 3. 1-percent S coal; 81-percent SC>2 removal;
particulate removal by others; 80-percent load factor; with ponding;
spare train
182
-------
Table 7
ANNUAL OPERATING COST - LIME SCRUBBING
(a)
Cost Item
Basic alkali, tons
Process water, gal x 10
Power, demand, kw
Power, energy, kwh x 10
Steam, demand, annual Btu x 10
6
Steam, energy, Btu x 10
Operating labor, 3 /shift
Operating labor overhead @ 75%
Maintenance @ 5% of direct cost
Maintenance overhead @ 30%
Capital recovery, 15 yr, 7-1/2%
(. 1315 multiplier)
Insurance and property taxes
at 1.9%
Sludge disposal, tons
Total Cost
Total
Requirement
94,000
1,754,000
1Z, 109
81,500
800,000
615,000
25,200
-
-
-
—
-
416,000
Range of
Unit Cost
Experience
$15-25
$.02-. 15
$15-40
$.90-5.00
$.08-. 25
$.10-. 55
-
-
-
-
—
—
$.75-5.00
Unit Cost
This Study
$20.00
$00. 02
$35.00
$ 5.00
$00.20
$00.55
$ 6.25
-
-
-
—
—
$2.00(b>
Annual
Cost,
$ Thousand
1,880
35
424
407
160
338
158
118
466
140
2,576
372
832
7.906
Cost
per kwh,
mills
. 537
.010
. 121
. 116
.046
.097
.045
. 034
. 133
.040
.736
. 106
.238
2.260
a, 520-MW gross output: 3. 1-percent S coal; 81-percent SO, removal; particulate removal by others;
80-percent load factor; with ponding; spare train
b. Waste solids are assumed to be 50-percent by weight dry solids. At $4 per ton for sludge disposal,
annual cost increases to $8,738, 000 or 2.5 mills per kwh
183
-------
Section 7
CONCLUSIONS
1. Commercial technology has advanced to the point where
system suppliers offer guarantees of 80 to 90 percent
removal for limestone and lime, respectively.
2. Process design has become more conservative with higher
liquid-to-gas and lower stoichiometric ratios. Both should
give greater system reliability.
3. Twenty-one full-size lime/limestone scrubbing units for
SO removal are being engineered, constructed, or operated.
4. Capital cost for new 500-megawatt SO^-removal limestone
units burning 3 percent sulfur coal is $45 per kw. Operating
cost is 23 to 25 cents per million Btu (including capital).
5. Capital cost for lime scrubbing units under the same con-
ditions as for the previous conclusion is $38 per kw.
Operating cost is 23 cents per million Btu.
6. Two areas requiring further study and evaluation are
reliability and solid waste disposal.
184
-------
Section 8
REFERENCES
1. TVA Special Reports on Pilot Plant Operations, Aug 1971,
Nov 1971
2. P. Lowell et al, "A Theoretical Description of the Limestone
Scrubbing Process," (Jun 1971) Vol. 1. Report No.
PB 193-029, Clearinghouse for Technical Information,
Va. , 2251
3. A.V. Slack et al, "Sulfur Oxide Removal From Waste Gases,"
Journal of Air Pollution Control Association, Mar 1972
4. R.A. Boll, "Mathematic Model for SO Absorption by Lime-
£
stone Slurry," Limestone Scrubbing Symposium, Perdido
Bay, Pensacola, Fla. , Mar 1970
5. M. Epstein, "Test Results from EPA Lime/Lime stone Scrub-
bing Test Facility, " Flue Gas Desulfurization Symposium,
May 1973, New Orleans
6. J. Jones et al, "Waste Product From Throwaway Flue Gas
Cleaning Processes," Flue Gas Desulfurization Symposium,
May 1973, New Orleans
7. D. C. Gifford, "Will County, Unit 1 Limestone Wet Scrubber,"
Proceedings Electrical World Conference - Sulfur in Utility
Fuels, Chicago, 111., Qct 1972
185
-------
8. Interagency Report on Japanese Sulfur Oxide Technology,
Aug 1972
9. J. Burchard et al, "General Economic Considerations
of Flue Gas Scrubbing for Utilities," Proceedings Elec-
tric World Conference— Sulfur in Utility Fuels, Chicago,
111., Oct 1972
186
-------
WASTE PRODUCTS FROM THROWAWAY
FLUE GAS CLEANING PROCESSES -
ECOLOGICALLY SOUND TREATMENT AND DISPOSAL
by
Julian W. Jones
Richard D. Stern
Development Engineering Branch
Control Systems Laboratory
Office of Research and Monitoring
U.S. Environmental Protection Agency
Research Triangle Park, N. C.
187
-------
ACKNOWLEDGEMENTS
In presenting this paper, the authors are indebted to numerous
organizations, including Chicago FlyAsh Co., Commonwealth Edison Co.,
Dravo Corporation, EPA National Environmental Research Centers in
Cincinnati, Ohio, and Corvallis, Oregon, EPA Office of Solid Waste
Management Programs, International Utilities Conversion Systems, and
the Tennessee Valley Authority. The authors are also indebted to
several individuals, but especially to Mr. Frank Princiotta for his
helpful comments and direction, and to Mrs. Charlotte Bercegeay, for
her patience in typing the many "revised editions."
-------
NOTES
1. The mention of company names or products is not to be considered
as endorsement or recommendation for use by the U. S. Environmental
Protection Agency.
2. EPA policy is to express all measurements in Agency documents
in metric units. When implementing this practice results in
difficulty in clarity, NERC-RTP provides conversion factors for
the particular non-metric units used in the document. For this
paper these factors are:
British
1 acre
1 acre-ft
5/9 (°F-32)
1 ft
1 ft2
1 ft3
1 ft3/short ton
lb/in.2
1 mile2
1 mile2-ft
1 ton (short)
Metric
4047 meters'
1233.6192 meters3
0.3048 meter
2
0.0929 meters
0.0283 meters
0.0312 cubic meters per
metric ton
70.31 grams per centimeter2
2
2.59 kilometers
0.7894 square kilometer-meter
0.9072 metric tons
189
-------
ABSTRACT
Extensive application of lime/limestone throwaway processes will
create large quantities of wet sludge. The sludge solids consist
primarily of flyash or oil soot plus sulfonated and unsulfonated calcium
salts and contain many impurities and trace elements associated with
the raw materials involved. In addition, neutralization of unmarketable
abatement sulfuric acid can potentially create large quantities of
similar sludge. Based on current and projected utilization of lime/
limestone throwaway processes, the sludge problem is quantified.
Results of a preliminary assessment of current technical approaches
and techniques for treatment and disposal of sludge are discussed,
including potential for ground and surface water degradation. An EPA
program to determine ecologically and economically acceptable methods
for disposal of lime/limestone sludge is described. The program includes
the physical and chemical characterization of sludge materials from:
representative sorbent/fuel combinations; an evaluation of potential
water pollution or other problems associated with disposal of both
untreated and treated (subject to dewatering, oxidation, chemical
fixation, etc.) sludge; and a study of the economics of various
disposal/treatment combinations. Preliminary data and analyses are
presented.
190
-------
1.0 INTRODUCTION
A major problem inherent in any flue gas desulfurization system
is the necessity of disposing of or utilizing large quantities of a
sulfur product. The sulfur compounds produced are either saleable
(sulfuric acid, sulfur dioxide, or elemental sulfur) or throwaway
(calcium-sulfur compound sludge materials). The technology of
processes producing abatement sulfuric acid is currently more advanced
than for those producing the other saleable products. To date, however,
most utilities have favored utilization of lime or limestone scrubbing
throwaway processes.
Although research and development is currently underway in a
number of processes with saleable sulfur products, most will not be
commercially available until the late 1973-1975 period. Allowing 2
years for retrofit to an existing plant and 5 years for a new installa-
tion including a control system, the earliest application of these
processes could be late 1975 and 1978, respectively. It is significant
to note that of 23 full-size desulfurization control processes for
power plants in design, construction, or operation, 17 are lime/limestone
scrubbing throwaway systems. This majority, undoubtedly based on a
higher degree of confidence in understanding, process characterization,
and current commercial availability, and the timing indicated above,
suggests that throwaway processes will comprise a very significant
percentage of flue gas cleaning installations in the near term at least
to 1980. (It should be noted that the double alkali process—soluble
alkali scrubbing, lime/limestone regeneration of soluble alkali—is
191
-------
currently regarded by many to be an attractive alternative to lime/
limestone slurry scrubbing. However, this process results in a very
similar throwaway product.) The majority of the remaining installa-
tions will very likely be comprised of processes producing sulfuric
acid since these processes appear to be further advanced than those
producing other saleable sulfur products.
Extensive application of the throwaway processes will create
large quantities of wet sludge with solids consisting primarily of
flyash or oil soot plus CaS03 • 1/2 11,0, CaS04 • 2 IL,0, CaC03, and
possibly some Ca(OH)9 from lime scrubbing processes, and CaS03 • 1/2 H20>
CaS04 • 2 H-,0, and CaCO, from limestone scrubbing processes. In addition,
neutralization of unmarketable abatement sulfuric acid can potentially
create large quantities of sludge consisting primarily of CaS04 • 2 l^O.
Typically, the water in equilibrium with these sludges will not only
contain varying amounts of these materials as dissolved solids in the
3,000-15,000 ppm range (primarily Ca**, S04=, Mg++, and S03=), but also
many impurities and trace elements found in the applicable raw materials
such as the alkali sorbent, the fuel material, and the process water.
For a number of years the Environmental Protection Agency
has been engaged in research and developnent concerned with
utilization of lime/limestone-modified pulverized flyash resulting
from lime/limestone based desulfurization processes. Acknowledging the
potential solid waste and water pollution problems associated with these
processes, the prime objective was that an air pollution problem should
not be transferred to these other areas. A study of the utilization of
192
-------
' indicated that only about 7-10 percent of flyash is currently
utilized and that under current technology and economic conditions,
utilization would probably not exceed about 25 percent in the near terra.
Similar limiting conditions are expected to apply to sludge materials.
Additionally, analysis of dry-collected limestone-modified flyash
indicated the presence of potentially toxic elements.
On the basis of these results, the rapid emergence of lime/limestone
wet scrubbing as the dominant desulfurization process, and a lack of
knowledge regarding potential heavy metal and toxic element involvement
in the scrubber chemistry, EPA reoriented its activities toward ecologi-
cally sound and "safe" treatment/disposal of the waste products from
these throwaway processes. Techniques for treatment and disposition
may be applicable to sludges produced by neutralization processes.
This paper quantifies the problem of lime/limestone throwaway
process generated sludge, examines and assesses potential treatment
and disposal alternatives, identifies problem areas, and presents the
EPA program, including results to date.
193
-------
2.0 QUANTIFICATION OF THE PROBLEM
Typical quantities of throwaway sulfur products, potential saleable
products, and coal ash production for a typical 1000 Mw coal-fired boiler
are presented in Table 1. From this table rough comparisons between the
production rate and storage volumes* for typical throwaway sulfur products
may be compared to typical saleable products and to coal ash, the normal
waste product for a coal-fueled plant. The following are important
observations which can be made:
1. The production rates (dry basis) of throwaway sulfur product
are approximately 45-80 percent larger (dry) than those of ash normally
produced; this leads to a total (sludge plus ash) throwaway requirement
about 2.5-3.0 times the normal coal ash disposal rate. The ranges
reflect lime and limestone sludges, respectively.
2. Large storage volumes are required for ultimate disposition of
the sludges. For example, for a 1000 Mw coal-fired unit over a 20-year
lifetime, about 860-1100 acres** (1.3-1.7 square miles) of disposal
land would be required for a wet sludge (50 percent solids) ponded to
a 10-foot depth. The ash alone would require about 250 acres (0.4
square miles) ponded to the same depth. Thus, the volume required for
the throwaway sulfur product is approximately 275 percent greater than
*It should be noted that rough estimates for solids packing specific
volume were used to calculate potential storage volumes required for
20 years of boiler operation; depending on the process, lime sludges
can be allowed to settle in a storage pond to a maximum of about 50
percent solids slurry, or can be dewatered to about 70 percent or
more solids by various techniques.
**Factors for converting these and other English units to their metric
equivalents may be found on page iii.
194
-------
Coal Ash
Table 1. TYPICAL QUANTITIES OF ASH AND SULFUR
PRODUCTS FROM A 1000 MW COAL-FIRED BOILER CONTROLLED WITH
A LIME OR LIMESTONE SCRUBBING FLUE GAS DESULFURIZATION SYSTEM
Yearly Production, 1000 MIV
(tons/yr)
Assumed Packing Volume
(ft3/ton of dry material)
Approximate Volume
Required for Storage
1000 MW for 20 Years
(acre-feet)
288,000 (dry)
19
(wet; 20% water)
2510
to
Ul
Lime Sludge (dry)
73% CaS03-l/2H70
11% Ca(OH)2
11% CaS04-2H20
5% CaC03
Lime Sludge (wet;50% solids)
Limestone Sludge (dry)
58% CaS03-l/2H20
9% CaS04-2H20
33% CaC03
304,000
47,000
45,000
21,000
417,000
834,000
305,000
45,000
170,000
520,000
22 4210
45 8600
22 5225
Limestone Sludge (wet;
50% solids)
1,040,000
45
(assumed same as for wet
lime sludge)
10,800
Sulfuric Acid (95% cone)
Sulfur (dry)
243,000
75,600
17.6
23 (80% packing)
1960
800
Assumptions: Coal - 3.5 percent S content, 12 percent ash content.
Coal Rate - 2,400,000 tons/yr for 1000 Mw on stream for 6400 hours/yr (0.75 Ib/Kwh).
Lime Sludge - Based on performance of Chemico Ca(OH)2 scrubbing unit in Japan (Mitsui Aluminum Co.)
at 1.28 Ca(OH)2/S02 mole ratio and 85 percent utilization.
Limestone Sludge - Based on preliminary EPA/Bechtel/TVA data from the TVA Shawnee Power Plant at
1.65 CaC03/S02 mole ratio and 85 percent utilization.
Removals - 90 percent of S07 in flue gas converted to sulfur ^r—luct, 100 percent of ash collected.
-------
for ash normally produced and the total volume requirement is about
4 times that for ash alone. The ranges reflect lime and limestone
sludges, respectively.
3. Potentially large quantities of sulfuric acid can be produced
by certain flue gas desulfurization processes such as Catalytic
Oxidation, Wellman-Lord, and magnesium oxide scrubbing. Approximately
24 million tons per year of concentrated sulfuric acid can be produced
per 100,000 Mw of flue gas desulfurization capability. This is close
to the total U. S. sulfuric acid production rate, which was 29.3 million
tons in 1971. Additionally, the use of a sulfuric acid plant and
neutralization as a control strategy in the smelting industry
has the potential for producing large quantities of sludge. Assuming
stoichiomctric quantities, for each ton of sulfuric acid neutralized,
approximately 1.75 tons of dry sludge would result.
4. Elemental sulfur appears the most attractive product in terms
of production rates and potential storage volume. Only about 7b,000
tons per year of sulfur would be produced per 1000 Mw of flue gas
desulfurization; this leads to a storage area of about 80 acres CO.12
square miles) over a 20-year lifetime, assuming a 10-foot depth.
The numbers of Table 1 become significant when one considers that
in the foreseeable future, probably through 1980, U. S. electric
utilities will likely continue the current pattern of ordering wet
scrubbing systems; a majority of these orders will probably be for wet
lime/limestone processes producing a throwaway sludge. Forecasts based
on a government interagency Sulfur Oxide Control Technology Assessment
Panel (SOCTAP) report^ indicate that for coal:
196
-------
1. Over 20,000 Mw of generating capacity could be equipped with
S02 scrubbing systems by 1975, but the equipped capacity is estimated
to be closer to 10,000 Mw.
2. By 1977, the equipped capacity may be 48,000-80,000 Mw, but
the lower end of the range is considered a more realistic estimate;
i.e., approximately 50,000 Mw.
3. By 1980, at least 75 percent of the coal-fired capacity could
be equipped with stack gas scrubbers. This is equivalent to about
144,000 Mw. Assuming 75 percent of the total stack gas scrubbing
installations utilize a lime/limestone throwaway process results in
108,000 Mw controlled in this way.
Known commitments for lime/limestone scrubbing systems indicate
control of approximately 1900 Mw in 1972, 6700 Mw by 1975, and 9000 Mw
by 1977.
The above data were used to prepare Figure 1 which presents the
growth of coal-fired generating capacity and a comparison with known
committed and forecast utilization of lime/limestone scrubbing processes
for flue gas cleaning. Although a function of plant size, sulfur
content, and other factors, the total waste sludge generated to 1980
was roughly estimated using the data from Table 1. The results are
shown in Figure 2. In addition, an estimate of limestone sludge production
rates and disposal area requirements as a function of plant size and coal
sulfur content was also determined. These are shown in Figure 3.
197
-------
CAPACITY, KNOWN CONTROL
COMMITMENTS
1972 1973 1974 1975 1976 1977 1978 1979 1980
YEAR
Figuie 1. Comparison of estimated electric utility coal generation capacity with current
and forecast control by lime/limestone scrubbing.
193
-------
-------
1.4
1.2
^S"»
c:
,f i.o
UJ
0.8
a
I 0.6
0.4
0.2
T
SULFUR CONTENT OF COAL, \ 4.5
2.5
2.0
1.5
OJ
1.0
-------
The magnitude of the 1980 estimates of Figure 2 appears significant
from the following comparisons:
1. Strip and Surface Mining - The estimated 1800 square mile-feet*
required to be committed for sludge disposal (50 percent solids) by
1980 is:
a. About 89 percent of the total land disturbed by strip
and surface mining of coal as of January 1965.
b. About 3.5 times the square mile-feet which will be
utilized in strip and contour mining of coal from
1973 to 1980. The strip and contour mining utilization
Ls estimated at 73.5 square-mile feet per year based on
1967 Bureau of Mines reports. ' This rate is
undoubtedly lower than the present and forecast rate,
but the comparative order of magnitude is obvious.
2. Coal and Ash Production - The estimated 112 million tons/year
production rate of wet sludge by 1980 is:
a. Approximately 25 percent of the weight of all the coal
(450 million tons) estimated to be required for electric
f41
power generation in 1980.
b. Approximately 175 percent of the weight of all the wet
ash (approximately 20 percent moisture) from all the
coal required for power generation in 1980. An ash
content of 12 percent was assumed for the coal.
*Square miles X ft(depth); e.g., 180 mi2 X 10 ft = 1800 mi2-ft,
201
-------
3.0 PRELIMINARY ASSESSMENT OF TREATMENT/DISPOSAL TECHNOLOGY
Currently, all of the existing or planned lime/limestone scrubbing
facilities for power plants in the United States employ the two major
ash disposal techniques, ponding and landfill, for lime/limestone waste
sludge disposal. Disposal of the sludge has been chosen because of
unknown technical and economic factors and the lack of a readily available
market for large-scale utilization of the waste material, although
research and development has shown at least limited use is feasible in
some structural applications. This section discusses the results of a
preliminary assessment of treatment/disposal technology made in late
1972, to assist in formulation of the F!PA program.
3.1 Ponding
Since ponds are widely used by electric utilities for the disposal
of ash, it is not surprising that ponding is considered to be a prime
means of sludge disposal. In fact, ponds are employed for disposal of
sludge in about 60 percent of existing or planned U. S. power plant
locations of lime/limestone scrubbing facilities. In addition, small
ponds are used for partial dewatering and temporary storage of the
sludge in almost all locations where other methods are being employed.
In general, ponds are used where a large area of inexpensive land is
available near the power plant. If sufficient land is available, the
pond is designed to eventually store wot sludge material over the
lifetime of the power plant.
The production rates (tons/year, dry basis) of sulfur product
waste material are approximately 50 percent larger than the ash normally
202
-------
produced. This leads to a total (sludge plus ash) waste disposal
requirement about 2.5 times the normal coal ash disposal rate. In
addition, since some sludge materials retain up to 60 percent moisture
even after an extended period of settling time, pond volumes several
times the size of normal ash ponds are required for sludge disposal.
This moisture-retaining tendency, or tendency for the solids not
to compact well upon settling, has been attributed to the thin platelet-
like crystal structure of calcium sulfite (CaSOg • 1/2 H-0), which
usually represents the majority of the sulfur in the sludge. The calcium
sulfate (CaSC>4 • 2 FUO) crystals, on the other hand, are "blocky," and
tend to compact well. This situation has created interest in means of
oxidizing the sulfite in the pond to sulfate. Oxidation of the sulfite
could reduce the pond volume requirements, as well as make the pond
area more easily reclaimable, rather than the possibility of indefinitely
remaining a wet "bog." Oxidation can also reduce the potential oxygen
demand of the sulfite in solution. However, oxygen consumption by the
sulfite does not appear to be a major problem because of the slowness
of the reaction under normal pond conditions.
Oxidation of the sulfite could lead to other problems, however.
Over an ambient temperature range, the calcium sulfate is on the order
of 50 to 100 times more soluble than the sulfite, so the pond water
pollution potential would be greatly increased. This solubility could
cause problems with reclaimed pond areas through sub-surface dissolution.
An alternative to oxidation might be dewatering using filtration,
centrifugation, etc. prior to ponding. Solids with approximately
203
-------
30 percent moisture have been generated by filtration in limestone
scrubbing pilot tests at EPA. Unfortunately, it is not currently
known whether the dewatered material will "set-up" permanently; i.e.,
will harden or remain dewatered after exposure to rainfall. (See
Section 5.2.2.)
With either a lime or limestone scrubbing system, ponding creates
the potential for water pollution. Lime sludge solids consist primarily
of CaS03 • 1/2 H90, CaS04 • 2 H-,0, CaCCL, and possibly some Ca(OH)2,
and flyash; limestone sludge solids consist primarily of the same com-
pounds, excluding Ca(OH) . Both sludges contain liquor which is, in
all likelihood, saturated in calcium sulfitc and sulfate. In fact,
pilot and full-scale operating data have shown that it may be super-
saturated in calcium sulfate. Magnesium, a normal impurity in lime
and limestone, is also contained in the sludge. Because of the high
solubility of magnesium sulfite and sulfate, magnesium is a serious
potential pollutant in the liquor associated with the sludge. It is
anticipated that in most applications, a relatively low limit may have
to be put on the magnesium contained in the lime or limestone, possibly
rendering dolomitic limestones unsuitable.
A variety of additional chemicals from sources other than lime
and limestone are also present in the sludge. The chlorides in the
coal, for example, are given off in the combustion process as HC1. The
HC1 is removed in the scrubber and, through internal recycle of scrubber
liquor, the concentration of soluble chlorides builds up in the liquor.
The concentration levels reached are dependent on: the quantity of
chlorine in the coal, the solubilities of the chlorides formed, and the
204
-------
quantity of liquor associated with the waste sludge. The dissolved
solids level in the scrubber liquor is greatly affected by this chloride
build up, which presents another potential water pollution problem.
It is generally recognized that flyash contains trace quantities
of potentially toxic elements, although usually in extremely insoluble
form. For example, limited analysis of dry-collected limestone-modified
flyash from EPA's dry limestone injection program indicated 50-200 ppm
As, about 200 ppm each of Ba and V, and 200-500 ppm Pb. It is also
recognized that trace quantities of toxic elements present in coal and
oil are volatilized in the combustion process and are potentially
removed from the flue gas in a wet scrubbing process, subsequently
ending up in the waste sludge. In addition, lime and limestone can
contain trace impurities which will be present to some extent in the
process streams. Any trace chemicals in the makeup water will also be
present. Because of the internal recycle of scrubber slurry, and the
recycle of water from the pond back to the scrubber circuit, it is
possible that trace chemicals from all of the above sources will build
up in concentration with time. Similar to the chloride problem, the
concentration levels reached are dependent on: quantities in the above
sources, their solubilities, and the quantity of liquor associated with
the sludge. Potential exists, therefore, for trace chemical contamina-
tion of both surface and groundwaters unless proper precautions are
taken.
To prevent unintentional seepage of liquor through the dikes and
floor of the pond into groundwater, it may be necessary to use a
205
-------
sealant material. The pond diking should also he carefully designed
to avoid overflow and run-off of sludge liquor into surface water.
In addition, the system must be operated in a closed- or nearly-closed-
loop mode of operation. This means that all liquor entering the pond
is recycled to the scrubber circuit; no sludge liquor is released to
any watercourse.
In the majority of full-scale installations currently in operation,
attempts have been made to operate in a closed-loop mode. However, in
most of these facilities, seepage, run-off, and other mechanisms could
be postulated which would allow liquor to be released into surface or
groundwatcr, at least periodically. It should be noted that ash ponds
which have been used for many years have similar potential water pollution
problems, although there are little available data related to water con-
tamination by ash liquors. However, soil amendment studies using flyash
have shown trace element ingestion by plant life, suggesting the
possibility of leaching from flyash ponds.
In summary, although ponding is currently widely used, the following
potential problems arc evident:
1. Volumes several times the size of normal ash ponds,
because of high moisture content in solids containing a
significant percentage of calcium sulfite.
2. Potential sulfite oxygen demand.
3. Potential dissolution of calcium sulfate, with resultant
water pollution potential.
206
-------
4. Potential reslurry of dewatered solids when exposed
to rainfall.
5. Impurities and trace chemicals in the sludge, with
inherent water pollution potential.
6. Seepage, run-off, and/or overflow of liquor from the pond.
3.2 Landfill
Although landfill operations are included in less than 40 percent
of existing or planned lime/limestone scrubber locations, ash disposal
has involved techniques other than ponding in these locations.
(Actually, in most cases ponding ultimately results in a landfill-type
operation; to reclaim the land when the pond is filled, the "pond" must
be covered with earth, similar to a landfill.)
Since the sludge material contains large quantities of liquor and
therefore is difficult to handle, dewatering of the material for trans-
portation is likely although not absolutely necessary. Clarification
alone will probably result in not more than 50 percent solids, so
further dewatering by either filtering or centrifuging will likely be
necessary. The filtered or centrifuged material can then be transported
to a suitable landfill site. As mentioned in the previous sub-section,
it is not currently known whether the dewatered material will remain
dewatered after exposure to rainfall. (See Section 5.2.2.) If it does
not, a fixation or stabilization technique must be applied.
207
-------
Chemical fixation processes are currently being developed by many
companies, including IU Conversion Systems, Dravo, Chicago FlyAsh,
Chemfix, and others. Such processes generally involve pozzolanic*
chemical reactions between flyash and lime. In most lime processes,
the hydrated lime in the scrubber and the slurry recycle tank is
almost completely converted to calcium carbonate. Therefore, although
the percentage of unreacted calcium carbonate in a lime process sludge
is less than that in sludge from a limestone process, the sludge
components are identical. This means that lime must be added to the
sludge to generate the pozzolanic process, possibly along with an
accelerating agent to decrease the curing time. Reaction of the lime
and flyash leads to the formation of a relatively dry solid in which
the sludge components are physically, and possibly chemically, bound
up. This is desirable because it decreases the potential dissolution
of these components and thus the possibility of groundwater contamina-
tion through leaching or nearby surface water contamination through
run-off.
Binding up the sludge components through a chemical fixation process
is currently being applied at the limestone scrubbing facility at
Commonwealth Edison's Will County Station. It is also planned for the
lime scrubbing facility at Duquesne Light's Phillips Station, to be
*Pozzolan: Generic name for cementitious material which is based on
ordinal use of volcanic rock or ash containing silica, alumina, lime,
P^;,n?-a rf ?r , construction in ancient aqueducts. Named after
Pozzuoli, Italy, where it was first found.
208
-------
started up in mid-1973. It is not currently known whether fixation
techniques will be applied to additional projects, although it is
known that the techniques are being considered for a large new coal-
fired facility.
In summary, problems similar to those of ponding exist with landfill
operations, for example:
1. Potential reslurry of dewatered solids when exposed to
rainfall.
2. Dissolution of calcium compounds and trace chemicals in
the sludge material, with resultant water pollution
potential.
Fixation processes may alleviate the problems associated with landfill,
but these require further evaluation.
3.3 Utilization
As development activity in limestone scrubbing has increased during
the past few years, investigations in the utilization of flyash and
limestone waste sludge material have been undertaken. Private companies
involved in this area have included Combustion Engineering, IU Conversion
Systems (formerly G&WH Corson), and others. In Japan, oxidation of the
sulfur products to gypsum for construction use is underway. EPA has
sponsored work at West Virginia University's Coal Research Bureau(7) and
at Aerospace Corporation.
The approach taken in the EPA programs was to determine current and
potential utilization of flyash, assess applicability of limestone waste
sludge material and, based on its properties, assess its potential for
209
-------
new applications. However, it was found that in 1970, only about
7 percent of all the flyash produced by coal-burning power plants was
actually utilized - primarily for concrete, structural fill, lightweight
aggregate, raw material for cement, filler for bituminous products, and
road base material. It was estimated that the maximum practical
potential usage under current technology and associated market conditions
would only be about 25 percent. Thus, approximately 75 percent of the
flyash produced in the United States was not considered marketable in
the near term because of: expected variation in the composition of the
flyash due to coal composition changes during operation, unfamiliarity
with this material, and the general inability to readily compete economi-
cally with conventionally used materials. Because of similar limitations
there is not expected to he widespread utilization of limestone sludge
material in the United States, at least not in the near future.
It can he concluded, therefore, that although utilization is
desirable in the long run, near-terra treatment/disposal solutions to
the sludge problem must occupy a higher priority for EPA research and
development. However, promising utilization approaches which result in
materials which can be disposed of as "waste," to be utilized at some
time in the future, are considered applicable to the current problem.
3,4 Conclusions
This preliminary assessment of treatment/disposal technology,
made in late 1972, indicated a considerable number of potential problem
areas, all described earlier in this paper. Some efforts are being
expended by private industry, but these seem to be directed toward
210
-------
solving the problems of a specific application, and are in many
instances quite limited in scope. Because of the wide variety of
scrubber systems, sorbent/fuel combinations, power plant land use
considerations, etc., there is quite an apparent need for a program
with a broad, national scope which will encompass as many situations
as possible. This broad-based program would take into account currently
available information, generate new information, and disseminate findings
to all electric utilities, private industry, and the general public. It
also would have the advantage of reducing the cost of specific application-
oriented development by, among other things, avoiding duplication of
effort.
EPA has initiated a program with the objective of meeting as many
of these requirements as practicable. Due to fiscal and other constraints,
the program, although broad in scope, includes neither the full range of
sorbent/fuel combinations, nor their associated treatment disposal tech-
niques. In addition, demonstration of the most promising treatment/
disposal technology is yet to be undertaken. However, the EPA program
described in Section 4.0 is considered a major first step.
211
-------
4.0 THE EPA PROGRAM
An EPA contract, "Wet Collected Limestone-Modified Fly Ash
Characterization and Evaluation of Potentially Toxic Hazards" was
formalized late in 1972 with the Aerospace Corporation. The contract
provides for a detailed characterization of wet collected limestone-
modified flyash and an evaluation of the potential toxic hazards posed
in processes that may be performed in subsequent handling, disposal,
or utilization of the sludge. The need for this study was based on
the results of previous programs conducted by LPA which indicated the
following:
1. Analyses performed on dry collected limestone-modified
flyash identified the presence of heavy metal trace
elements that could pose a toxic hazard, depending upon
their chemical/physical state.
2. The wet collected limestone-modified ash may contain even
greater quantities of the heavy metal elements as a con-
sequence of the wet scrubbing of the ash and flue gases,
and poses a potential hazard in its disposal or utilization.
3. Water-soluble elements pose an additional health hazard in
the disposal of waste waters because of the large quantities
of water required for the process scrubbing.
However, when procurement for the current contract was initiated,
potential utility sources to obtain representative sample types were
limited; disposal was essentially limited to ponding; and commercial
acceptability of throwaway processes and the coresponding quantity of
material requiring disposal was unknown. These factors led to a
212
-------
program of limited scope with prime ecological emphasis on toxicity.
Since that time, additional utility sources using different sorbent/
fuel combinations and applying other treatment/disposal techniques
have become operative. These additional combinations with their
different elemental compositions and treatment/disposal techniques
needed to be taken into consideration. Although toxicity is still
important, toxic element concentrations are expected to represent only
a very small percentage of the total quantities to be disposed of.
Based on the above and the current and projected magnitude of the
sludge problem, EPA now plans an expanded program to allow a more
complete assessment of ecological acceptability, technical state of the
art, and economics for the various treatment/disposal techniques.
The objective of the expanded program is to determine ecologically
and economically acceptable methods for treatment/disposal of lime/
limestone sludge and to provide pertinent input for the establishment
of realistic water and solid waste regulations and standards. Sample
materials, representative of as many situations of lime/limestone wet
scrubbing process applications as practicable, will be obtained. In
addition, test, operational and economics data from a wide variety of
sources will be taken into consideration.
The basic elements of the program are as follows:
1. An inventory of sludge components, including chemical
analysis of various types of sludge and the raw materials
from which they are formed (lime or limestone, coal or oil,
213
-------
process water). Sorbent/fuel combinations to be studied
are limestone/Eastern and Western coals, lime/Eastern coal,
and double alkali/Eastern or Western coal.
2. An evaluation of the potential water pollution and solid
waste problems associated with disposal of the sludge,
including consideration of existing or proposed water
effluent, water quality and solid waste standards or guide-
lines. The information would also assist in the evaluation
of potential treatment/disposal techniques described below.
3. An evaluation of treatment/disposal techniques with
emphasis on ponding and "fixed" and "unfixed" landfill
(and related land use applications). Physical analyses
and tests of various sludges will be conducted, including
determination of the effects of dewatering, oxidation,
chemical fixation, aging, etc., on stability, compactibility,
leachability of solubles, potential pond seepage, potential
run-off problems, and other disposal considerations. The
economics of various treatment/disposal combinations will
also be studied.
4. A recommendation of the best available technology for sludge
treatment/disposal based on the evaluation described above.
Pertinent input for the establishment of realistic water
and solid waste regulations and standards may also be provided.
214
-------
The approach taken by the EPA in undertaking this program is to
utilize EPA funded or co-funded scrubber sources, the cooperation of
several utilities who have installed scrubber systems and (as much
as possible) the knowledge and technical expertise of EPA water and
solid waste organizations. EPA personnel conducted a survey of
potential electric utility participants for the planned program.
Factors in the study included facility availability, additive/fuel
combination, facility size, treatment techniques (dewatering, oxidation,
chemical fixation, etc.). and final sludge disposition (ponding,
landfill, other). Table 2 presents the results in order of facility
availability to support the EPA program.
Upon completion of the survey, contacts were made with several of
the utilities to determine their willingness to provide sludge and raw
material samples, as well as information concerning their sludge treat-
ment and disposal activities. In addition, discussions were held with
several scrubber system companies, waste treatment/utilization/disposal
companies, and EPA water and solid waste personnel regarding review and
integration of activities and information.
Aerospace Corporation efforts have been limited to the scope of
the current EPA contract. However, these efforts have been conducted
and coordinated to the maximum extent possible with EPA efforts toward
meeting the objectives of the planned program expansion. Surveys of
current treatment, disposal, and utilization technology, as well as
chemical analyses and physical testing of samples from the TVA Shawnee
facility, have been conducted. Progress of Aerospace and EPA efforts
are discussed in Section 5.0.
215
-------
Table 2. POTENTIAL SLUDGE TREATMENT/DISPOSAL PROGRAM UTILITY PARTICIPANTS
(X = Current; P = Possible Additions)
Facility
(Availability
Status)
TVA-Shawnee
(Current]
City of Key
West-Stock
Island
(Current!
Commonwealth
Edison Co. -Will
County
(Current)
Southern
California
Edison-Mohave
Kansas City
Power 5 Light-
Havv'thorn
(Current)
Kansas Power §
Light-
Lawrence
(Current)
Louisville Gas
& Electric-
Paddy's Run
(Current)
Sorbeut
Fuel
Limestone now, .^^^
lime later .^^^
*s^^ Eastern
^^^ coai
Limestone .->**
j^""^^
(coral marl) ^^
^*^^ Residual
^^^ oil
Limestone ^^^^
.s^ Eastern
.^ coal
Limestone ^^'
5 lime ^*-^^
^^-^^'^ Western
^^^ coal
Boiler *^
injected s/~^
limestone s^
S^ Coal
.S (possible E&W
^ blend)
Boiler ,^"
injected ^^^^
1 imes tone ^^~^
.^*^ Eastern
^^"^^ coal
Carbide ^^**
sludge ^^"'^
(Ca(OHK}^^
^s- — Eastern
<-^"^ coal
Scale
Proto-
type
Full
Full
Pilot
Full
Full
Full
Dewatering Technique
Ciuri-
f icr
X
X
X
X
Filter
X
V
\
X
Centri-
fuge
X
P
Dryer
P
Pond
X
X
(Well
points)
X
(Well
points)
X
Final Disposition
Ponding
X
X
X
X
Landfill
X
(Unfixed)
X
(Sealed)
(Fixed)
X
Other.
-------
Table 2 (Continued). POTENTIAL SLUDGE TREATMENT/DISPOSAL PROGRAM UTILITY PARTICIPANTS
CX = Current; P = Possible Additions)
Facility
(Availability
Status)
Sorbent
Fuel
Scale
Clari-
fier
Dewatering Technique
Filter
Centri-
fuge
Dryer
Pond
Final Disposition
Ponding
Landfill
Other
Northern
States Power-
Black Dog
(Current)
Limestone
Western
coal
Pilot
Kansas City
Power §
Light-
LaCygne
(Current)
Limestone
Eastern
coal
Full
Arizona
Public
Service-
Choi la
(Approx
mid-1973)
Limestone
Full
Western
coal
X
(Solar
evap)
Duquesne
Light-
Phillips
(Approx
Bid-1973)
Lime
Full
Eastern
coal
X
Curing)
X
(Fixed)
Detroit
Edison-
St. Clair
(Late 1973)
Limestone
Full
Eastern
coal
(Unfixed)
-------
5.0 PROGRAM PROGRESS
5.1 Industry and Intra-EPA Coordination
The information highlights discussed here are based on an initial
survey conducted as part of the EPA program via informal meetings
and/or recent publications. The information generally falls into the
categories of general, technical, and economic.
5.1.1 General - Lea.chate pollution of groundwater or water
courses appears to be the primary concern for sludge disposal. Results
of various studies have shown a cause for this concern. For example,
greenhouse studies have shown that application of selected samples of
flyash to soil increases the availability of boron, molybdenum, zinc,
phosphorus, and potassium to plants by supplying soluble forms of these
( Q~\
elements and/or modifying the soil pH. } These results add credibility
to the implication that these elements and others can leach from flyash
even though the ash consists primarily of glassy silicates considered
relatively insoluble in aqueous environments. While leaching in this
particular application is undoubtedly a function of surface area
exposed to permeating moisture, it provides additional evidence that
water quality may be sacrificed by unsuitable disposal of flyash/
sulfur sludges. Further, at least one state has preliminarily disapproved
disposal of a particular sulfur sludge on the basis that potentially toxic
materials in the sludge were in excess of potable water standards.
Coordination with EPA water personnel indicated that because of a
lack of detailed information, there are presently no Federal water
218
-------
effluent, water quality, or groundwater standards* which take into
account power plants with on-site desulfurization systems. As a
result, State and local regulatory agencies are applying available
water standards dealing with receiving stream quality and potable
water. The Federal Water Pollution Control Act Amendments of 1972
require the setting of standards applicable to effluents and ground-
water, and also establish a goal of zero pollution discharge by 1985.
Based on the above, a study on recycle, reuse, or treatment of
effluent waters related to flue gas lime/limestone wet scrubbing
processes is under consideration by EPA water personnel. This study
would be closely coordinated with the program discussed in this paper
and represents additional broadening of the overall EPA program. The
study, directed toward determining the implications of open-loop or
partially open-loop systems, would involve analyzing effluent liquor
downstream of the scrubber system. In addition, the study would
include the potential for treatment of the water after scrubbing and/or
dewatering. Noxious effects as well as toxic hazards would be con-
sidered, with technology and economics determined for all cases.
Coordination with EPA solid waste personnel also indicated that
there are no specific Federal standards* related to power plant sludge
disposal. However, considering the current lack of information concerning
sludge properties, disposal of sludge in sanitary landfills is not
expected to be permitted.
*Current and proposed State and Federal guidelines, standards, etc.,
are being compiled for correlation with EPA program results.
219
-------
For these and other reasons, there is considerable interest in
treatment of the sludge to combine primary and trace elements into
new crystalline phases and also to reduce the leaching rate. One
such treatment technique under development results in the production
of synthetic aggregate, which would be used immediately or stored for
subsequent use. This lime/flyash/sulfur sludge aggregate would be
used in load-bearing applications in place of natural aggregate with
either conventional cement binder material or with lime/flyash/sulfur
sludge binder material. Synthetic aggregate production was advanced
as a means of raw material (limestone aggregate) conservation as well
as ecologically sound disposition of sludge. This would free limestone
supplies for scrubbing and other uses. Without this approach, local
limestone availability in sufficient quantity to supply sorbent materials
for large scale wet scrubbing installations has been questioned.
On-site flyash availability in sufficient quantity for some treat-
ment formulations may be questionable. In other words, even with
all the plant flyash available, some utilities may need additional
flyash or a substitute, depending on final disposition of the treated
sludge. For utilities which currently market their flyash, the problem
is even greater. In fact, one such utility is currently looking for a
flyash substitute for their sludge disposal process. Although optimum
formulations are not presently known, applications such as structural
or non-structural landfill, base course materials, and synthetic
aggregate are expected to require different mix proportions.
220
-------
Additional information related to sludge disposal has been generated
through EPA efforts in mine drainage pollution control and solid waste
disposal. EPA mine drainage activities have resulted in numerous
reports dealing with sludge produced by neutralization of acid mine
drainage. This type of sludge was one of those tested at Transpo '72.
(See Section 5.1.2, below.) The reports cover areas such as in-situ
sludge precipitation, sludge supernatant treatment, thickening and
dewatering, use of latex as a soil sealant, and technical and economic
feasibility of bulk transport.
Although primarily concerned with municipal sludge, many EPA solid
waste activities relate to the CPA sludge program under discussion.
Examples include the following:
1. Methods of removing pollutants from leachate water.
2. Determination of organic and metal characteristics of
leachate from selected landfills.
3. Evaluation of landfill liners.
4. Development of mathematical models to determine effects
of landfill leachate on groundwaters.
5. Leachate pollutant attenuation in soils.
6. Moisture movement in landfill cover material.
The EPA reports and other results of continuing coordination will
be taken into account under the current program.
5.1.2 Technical - A wide divergence in physical properties
between untreated sludges has been reported. Settling rates varied
considerably and thickener design parameters have shown large spreads
221
-------
fg-j
from source to source and even within the same source.v J Although
the large variations observed in sludge behavior are currently not
completely explained, it is known that sulfite/sulfate ratio, percent
flyash, and other measures of chemical composition have a definite
effect. However, the operating conditions which result in the difference
in sludge properties (or its chemical composition] are not as well
defined.
As mentioned earlier, there is considerable interest in treatment
of sludges in pozzolanic fixation processes using lime. The production
of the sulfur sludge pozzolan is based on:
1. The reaction of lime with soluble sulfates originating in
either the flyash or the sludge water to form calcium sulfate
and tie up water.
2. The reaction of lime, sulfate, iron oxide, and/or alumina
to form complex crystalline sulfoferrites or sulfoaluminates
such as ettringite (A1203 • 3 CaS04 - 31 H20).
3. The reaction of lime with the glassy silica of the flyash
resulting in the well-known pozzolanic reaction proceeding
slowly to form the calcium silicate, tobermorite.
These reactions reportedly result in significant changes in physical
and chemical properties of the sludge. A reduction in permeability, an impor-
tant property directly related to ecological disposal, has been reported.
This is based on falling head permeability tests indicating a decrease
by about 2 orders of magnitude between raw sludge and treated sludge
222
-------
(10,11)
after a week. It has also been reported that the trace element
concentration in leachate from a sulfur sludge pozzolan produced from
acid mine waste sludge was also reduced by about 2 orders of magnitude.
Considering these results, the availability of soluble contaminants
to groundwaters could be reduced on the order of 10,000 times.
It is postulated that the expansive nature of ettringite crystalli-
zation during hardening seals the mass and prevents shrinkage cracks,
thereby reducing permeability. Thus, dimensional stability appears to
be another important property of the sulfur sludge pozzolan especially
when compared to untreated sludge.
There is also some evidence that inclusion of sulfur sludges
improves the properties of conventional (limc/flyash/aggregate) pozzolans.
The mix containing sulfur sludge is reported to develop superior early
strength; greater final strengths appear possible. For example, com-
pressive strengths of up to 350, 750, and 1100 psi have been reported
for sulfur sludge pozzolan samples cured at 100°F for 2, 7, and 28 days,
respectively. *- J With unspecified formulations and cure conditions,
unconfined compressive strengths of 800 and 1600 psi after 14 days and
1550 and 2700 psi after 28 days were also reported/11-' It is postulated
that these properties are also the result of the ettringite crystalli-
zation.
(12)
Transpo '72 results/ on the other hand, indicated paving
material compressive strengths of about 100 and 225 psi for an acid
mine drainage sludge pozzolan and an S02 scrubber sludge pozzolan,
respectively, after 28 days at 73°F. However, this is a single test
223
-------
result and time did not permit optimization of the formulation. In
addition, it is reported that inadequate compaction, sealing and
curing time due to inclement weather and other factors, may have com-
promised performance. Cube compressive strengths of synthetic
aggregate made from acid mine drainage sludge and SC^ scrubber sludge
and cured at 73°F were approximately 3500 and 3300 psi, respectively,
after 28 days; and approximately 6500 and 6000 psi, respectively,
after 90 days. Synthetic aggregate used directly as a crushed aggre-
gate in conventional binder material and as an aggregate in the
sludge/flyash/lime paving mixture was reported to perform adequately,
with no evidence of failure.
5.1.3 Economics - The cost of local (within 20 miles) disposal
of wet (50 percent solids) lime/limestone sludge, including an additive
for pozzolanic fixation, has been estimated by one source to be as low
as $2.50 and by another to be as high as $10/ton. ' However, the
bases for these estimates are not completely known and may be signifi-
cantly different from each other. The cost of the additive (primarily
lime) has been estimated at from less than $18 up to $20/ton; a 3
percent by weight addition for sludge treatment would result in a cost
of $0.50 to $0.60/ton of wet (50 percent solids) sludge.
The cost of aggregate formation from the sludge has been estimated
from about $5 to $8/ton. This compares to a cost of from $1.50 to $8/ton
for naturally occurring aggregate; this cost depends on local supply,
transportation, etc. A typical average cost is reportedly about $2.50/ton,
224
-------
The choice between disposal by landfill and synthetic aggregate
production depends on many factors, such as natural aggregate availa-
bility and costs, transportation costs, land availability, and plant
size. Because of the many variations in these factors from one power
plant to another, it is difficult to select a single overriding con-
sideration. However, it has been estimated that the break-even point
for synthetic aggregate production (versus disposal cost) is approxi-
mately 200,000 tons/year wet sludge production. Below this sludge
production rate, the economics reportedly favor direct disposal. The
bases for this break-even point is unknown. However, the plant size
for 200,000 tons/year of wet sludge can be estimated from Figure 3,
Section 2.0.
5.2 EPA/Aerospace Corporation Contract
Initial results of analytical and physical property tests on
limestone scrubbing sludge and process materials from the TVA Shawnee
power plant burning Eastern coal are discussed below. It should be
noted that these results are of a very preliminary nature and are based
on a single sample from one source. As such, these results may or may
not be representative of those from other samples from the same source
or from different sources.
5.2.1 Chemical Analysis - Emission spectrographic analysis
of clarifier underflow liquor and solids (separated by centrifugation)
and of the limestone indicated:
1. Liquor - No toxic components; dissolved solids content
representative of equilibrium calcium sulfate concentration.
225
-------
2. Solids - Presence of several heavy metals, including
lead, gallium, chromium and titanium. In this specific
case, based on the above results of analysis of liquor
which had been in equilibrium with the solids, these
elements do not appear to present a toxic hazard. However,
the specific compounds and their solubilities under other
conditions are presently unknown.
3. Limestone - Presence of strontium in addition to carbon,
sodium, magnesium, aluminum, silicon, potassium and
calcium. All other elements found are constituents of
tramp clay minerals in the limestone. No other mineral
phases were found. The limestone showed few major impurities
that could cause potentially toxic effects.
Ion microprobc mass analyzer (IMMA) results have shown good agree-
ment with the emission spectrographic analysis. More precise analyses
of solids by spark source mass spectrometry (SSMS) and liquors by atomic
absorption (AA) are in progress. Available physiological concentration
effects data and toxic and hazardous element standards are being com-
piled for correlation with analyses of untreated and treated sludge
materials and run-off, overflow, and leachate liquors.
Tests to determine whether potentially toxic components may
sublime from a pond surface subjected to drying by solar heating indi-
cated that this was not a problem up to 650°F.
5.2.2 Physical Properties Testing - Testing of unconditioned
(raw) sludge from the clarifier underflow produced the following
preliminary results:
226
-------
1. Drainability and Water Retention - In a falling head
column, a naturally settled sludge of approximately
50 percent solids showed steady state drainage of 7.5 X 10
cm/sec. After air drying to less than 15-20 percent
moisture, the column was found to be virtually impervious
until several days later. At that time, initial drainage
was noted, but at a rate about half that indicated above.
At intermediate column dryness, initial rate and delay in
drainage were intermediate. In each case the final drainage
rate was approximately equivalent to the steady state rate
with the 50 percent solids sludge. In the experiments with
sludge dried above 50 percent solids, the column appeared to
regain sufficient water to return to the initial water content
(about 50 percent) and permeability. The effects of extended
time, and other dewatcring and compaction techniques on this
sludge behavior are not currently known.
2. Corrosion - Sample specimens of 1100 aluminum and 1010 mild
steel have been exposed to sludge solids and supernatant
liquor. The aluminum specimens have shown no degradation.
The effect on steel specimens differs between that portion
suspended within the solids and that immersed within the
liquor. After 1 month, the area in contact with the solids
appeared dull black,and a weight loss of about 1 percent
was measured. After 2 months, the area within the solids
was heavily encrusted with corrosion products and bound
227
-------
sludge particles and a net weight gain of more than
10 percent was measured. The area in the liquor dis-
played rust (Fe(OH)3) which was non-adhering and
easily rinsed from the specimen.
3. Viscosity - The viscosity of sludge samples with solids
contents varying from 50 to 60 percent was measured with
a viscometer which uses a rotating cylindrical cup immersed
in the fluid being measured. Peak viscosities of 120 poises
for 60 percent solids and 20 poises for 50 percent solids
were measured. Thixotropic behavior was exhibited by all
samples.
4. Bulk Density - Bulk density was determined for sludge with
water content from zero to 100 percent. Normally settled
and dried sludge bulk density was 1.2 g/cm3 (75 Ib/ft3).
With increasing water content, the bulk density increased to
a maximum of about 1.7 g/cm3 (106 lb/ft3) for a water content
of about 30 percent.
5. Shrinkage - Sludge cast into a known volume and allowed to
air dry exhibited linear shrinkage of about 4 percent.
Shrinkage of this magnitude could produce cracks in dried
ponds, providing paths for subsequent leaching.
5.3 Concluding Remarks
The authors were concerned with presenting results obtained so
early in the EPA program. However, the overriding consideration in
presenting these results was to transfer information in this
important area as soon as possible.
228
-------
Additional information and details pertaining to industrial
activities and EPA program progress related to sludge treatment
and disposal can be found in other papers presented at this symposium,
229
-------
6.0 PRELIMINARY CONCLUSIONS
In addition to the obvious conclusion that much extensive
research and development remains to be done, the following preliminary
conclusions, based on initial progress, are drawn.
6.1 Untreated Sludge
1. Based on reported sludge variability and divergence of
properties between and within sources, there is a
definite need to thoroughly characterize sludge materials
before a disposal/utilization scheme can be determined.
2. Sublimation of potentially toxic components docs not
appear to be a problem at upper extremes in ambient
temperature.
3. Air-dried, naturally settled sludge exhibits a tendency to
regain moisture back to its equilibrium moisture content and
initial permeability. Effects of other treatments and extended
time arc currently unknown.
4. Mild steel does not appear to be a satisfactory material for
use in storage or handling of untreated sludge.
5. Sludge pumpability appears to be significantly reduced as
solids content increases from 50 to 60 percent solids.
i
6. Air-dried sludge exhibits a tendency to shrink enough to
produce cracks which could serve as paths for leaching to
occur.
6.2 Treated Sludge
1. The treatment of sludge in a pozzolanic fixation process
.appears to significantly reduce the permeability of the
230
-------
material; the leachability of trace elements from within
the material is not clearly established.
2. The extent to which lime, flyash, and sulfur sludge
proportions influence permeability, leachability, and
strength development is not clearly established.
3, The inclusion of sulfur sludge in conventional (lime/
flyash/aggregate) pozzolans, for load-bearing applications,
produces a material which appears to develop superior early
strength and (possibly) greater final strength.
4. Pozzolanic mixtures of lime, flyash, aggregate and sulfur
sludge may be useful for road basecourse construction;
they do not appear useful for wearing surfaces.
5. Synthetic aggregate produced from lime, flyash, and sulfur
sludge may be suitable for basecourse construction either
directly as crushed aggregate with conventional binder
materials or as aggregate with a lime/flyash/sulfur sludge
binder material. Substitution of synthetic aggregate for
all cr part of the natural aggregate required in a construc-
tion mixture could significantly increase the use of the
waste sludge input.
6. Optimum formulations for the various disposal/utilization
techniques are not clearly established.
6.3 General
The reader is reminded of the nature of the results on which
these preliminary conclusions are based. Much more data and information
must be obtained and interpreted before firm conclusions can be reached.
231
-------
The attainment of these firm conclusions in the shortest possible
time will require a high degree of cooperation between industry and
governmental agencies to assess, develop, and apply the sludge treat-
ment, disposal, and utilization technologies.
232
-------
References
1. The Aerospace Corporation, "Technical and Economic Factors
Associated with Flyash Utilization," NTIS No. PB 209-280, final
report on EPA Contract F04701-70-C-0059.
2. Brackett, C. E., "Production and Utilization of Ash in the
United States," presented at the Third International Ash Utilization
Symposium, Pittsburgh, Pennsylvania, March 13-14, 1973.
3. Sulfur Oxide Control Technology Assessment Panel (SOCTAP), "Final
Report on Projected Utilization of Stack Gas Cleaning Systems by
Steam-Electric Plants," submitted to the Federal Interagency
Committee, Evaluation of State Air Implementation Plans, April 15, 1973.
4. Princiotta, F. T., Kaplan, N., "Control of Sulfur Oxide Pollutants
from Power Plants," presented at EASCON '72, Washington, D. C.,
October 18, 1972.
5. U. S. Department of the Interior, "Surface Mining and Our Environment -
A Special Report to the Nation," 1967.
6. Young, W. T., "Thickness of Bituminous Coal and Lignite Seams Mined
in 1965," Bureau of Mines 1C 8345, August 1967.
7. Condry, L. Z., Muter, R. B., and Lawrence, W. F., "Potential
Utilization of Solid Waste from Lime/Limestone Wet Scrubbing of
Flue Gases," presented at the Second International Lime/Limestone
Wet Scrubbing Symposium, New Orleans, Louisiana, November 8-12, 1971.
8. Martens, D. C., Plank, C. 0., "Basic Soil Benefits from Ash
Utilization," presented at the Third International Ash Utilization
Symposium, Pittsburgh, Pennsylvania, March 13-14, 1973.
233
-------
9. Selmeczi, J. G., Knight, R. G., "Properties of Power Plant Waste
Sludges," presented at Third International Ash Utilization
Symposium, Pittsburgh, Pennsylvania, March 13-14, 1973.
10. Minnick, L. John, "Multiple By-Product Utilization," presented
at the Third International Ash Utilization Symposium, Pittsburgh,
Pennsylvania, March 13-14, 1973.
11. Bitler, J. A., Minnick, L. John, "Lime-Sulfur Dioxide Scrubbing
System and Technology for Utilization of Underflow Sludge,"
Industrial Waste, March/April 1973.
12. Brink, R. I!., "Use of Waste Sult'ate on Transpo '72 Parking Lot,"
presented at Third International Ash Utilization Symposium,
Pittsburgh, Pennsylvania, March 13-14, 1975.
13. Gifford, D. C., "Will County Unit 1 Limestone K'et Scrubber,"
presented at AIChE 65th Annual Meeting, New York City, N. Y.,
November 28, 1972.
234
-------
TEST RESULTS FROM THE EPA
LIME/LIMESTONE SCRUBBING TEST FACILITY
by
M. Epstein, C. C. Leivo,
C. H. Rowland, S. C. Wang
Bechtel Corporation
San Francisco, California
235
-------
ACKNOWLEDGEMENT
The authors wish to acknowledge the valuable contribution of
the Bechtel, TVA and EPA on-site personnel at the Shawnee
Test Facility.
236
-------
Section 1
INTRODUCTION
In June 1968, a three phase program was initiated whose aim was the
testing of a large, versatile prototype system to fully characterize
wet lime and limestone scrubbing for removal of sulfur dioxide and
particulates from boiler flue gas. The Office of Research and Mon-
itoring (OR&M) of the Environmental Protection Agency (EPA) is
sponsoring this program, with Bechtel Corporation of San Francisco
as the major contractor and test director, and the Tennessee Valley
Authority (TVA) as the constructor and facility operator.
Phase I of the test program consisted of preliminary engineering,
equipment evaluation and site selection. Phase II involved the detailed
design and construction of the facility and the development of the test
plan and mathematical models for predicting system performance.
Phase III, the testing portion of the program, began in March 1972.
The test facility consists of three parallel scrubber systems each
capable of treating approximately 30, 000 acfm (10 Mw equivalent)
of flue gas, which are integrated into the flue gas ductwork of an
existing coal-fired boiler at the TVA Shawnee Power Station, Padu-
cah, Kentucky.
237
-------
This paper will cover, primarily, the test results for wet limestone
scrubbing from September 1972 to April 1973 at the test facility.
The operability and reliability of the facility during the limestone
testing will be covered in a second paper at this symposium (Ref. 1).
The results of air-water and sodium carbonate testing from May 1972
to August 1972 have been presented in Ref. 2.
238
-------
Section 2
TEST PROGRAM OBJECTIVES
The overall objectives of this program are to evaluate the perfor-
mance, reliability and economics of closed-loop limestone and lime
wet scrubbing processes. The following are specific goals of the
program:
• Investigate and solve operating and design problems,
such as scaling, demister plugging, corrosion and
erosion (see Ref. 1).
• Generate test data to characterize scrubber and sys-
tem performances as a function of the important
process variables.
• Study various sludge disposal methods.
• Develop mathematical models to allow economic
scale-up of attractive operating configurations to full-
size scrubber facilities and to estimate capital and
operating costs for the scaled-up system designs.
• Determine operating conditions for optimum SO2
and particulate removal, consistent with operating
cost considerations.
• Perform long-term reliability testing.
239
-------
Section 3
TEST FACILITY
The test facility has been described in detail in Ref. 3. It consists of
three parallel scrubber systems, each with its own slurry handling
system. Scrubbers are of prototype size, each capable of treating
approximately 30,000 acfm of flue gas from the TVA Shawnee coal-fired
boiler No. 10. Therefore, each circuit is handling the equivalent
of approximately 10 Mw of power plant generation capacity. The
equipment selected was sized for minimum cost consistent with the
ability to extrapolate results to commercial scale. The 30,000 acfm
scrubber train was judged to meet these requirements. Boiler No. 10
burns a high-sulfur bituminous coal leading to SO2 concentrations of
2300-3300 ppm and inlet grain loadings of about 2 to 6 grains/scf in
the flue gas.
Based on their ability to handle slurries without plugging or excessive
pressure drop and on their reasonable cost and maintenance, the fol-
lowing scrubbers were selected for testing:
(1) Venturi followed by a spray-tower (after
absorber)
(2) Turbulent contact absorber (TCA),
(3) Marble-bed absorber (Hydro-Filter)
240
-------
Figures 1, 2, and 3 depict, approximately to scale, the three scrub-
ber types along with the demisters (and demister wash sprays) se-
lected for de-entraining liquor in the gas streams. Note that, for
the TCA scrubber (Figure 2), a Koch wash tray has been installed
between the uppermost stage and the demister, in order to provide
a liquid barrier to prevent the bulk of the entrained slurry from im-
pinging directly onto the demister.
Typical system configurations for the venturi, TCA and Hydro-Filter
systems are shown in Figures 4, 5, and 6, respectively. Such pro-
cess details as flue gas slurry saturation sprays and demister or
Koch tray wash sprays are not shown.
241
-------
FIGURE I
SCHEMATIC OF VENTURI SCRUBBER
AND AFTER-SCRUBBER
CHEVRON DEMISTER
AFTER-SCRUBBER
INLET LIQUOR
THROAT
ADJUSTABLE PLUG
VENTURI SCRUBBER
GAS OUT
DEMISTER WASH
INLET LIQUOR
EFFLUENT LIQUOR
APPROX. SCALE
EFFLUENT LIQUOR
242
-------
FIGURE Z
SCHEMATIC OF THREE-STAGE TCA SCRUBBER
WITHOUT TRAP-OUT TRAY
GAS OUT
CHEVRON DEMISTER
INLET KOCH TRAY
LIQUOR
KOCH TRAY
EFFLUENT KOCH
TRAY WASH LIQUOR
STEAM WASH
RETAINING GRIDS
IN
0 o
°°o°o
o o ow
o
o o °o
00
INLET LIQUOR
MOBILE PACKING SPHERES
I-
5'
H
APPROX. SCALE
EFFLUENT LIQUOR
243
-------
FIGURE 3
SCHEMATIC OF HYDRO-FILTER SCRUBBER
DEMISTER WASH
INLET LIQUOR
INLET LIQUOR
GAS IN
GAS OUT
CHEVRON DEMISTERS
TURBULENT LAYER
GLASS SPHERES
EFFLUENT LIQUOR
y
i \
APPROX. SCALE
EFFLUENT LIQUOR
244
-------
N>
J=
w*
FLUE
VEWW1
SCRUBBER
®
FIGURE 4
EPA TEST FACILITY
TYPICAL PMKfSS FLOW DIAGRAM
FOR VENTURI SYSTEM
ft
i
»}»»
A r^ A
/* f\ f\
f+ ^ ^
WWW
\ /
"AFTER
SCRUBBER
1.0. FAN
SETTLING PONO
O Gas Composition
® Paniculate Cwnposition & Loading
© Slurrv or Soliffs Composition
— — to Stream
— Liquor Stream
-------
FIGURE 5
EPA TEST FACILITY
TYPICAL PROCESS FLOW DIAGRAM
FOR TCA SYSTEM
KJ
JT
en
SEHLINGPOND
O Gas Composition
® Particulate Composition & Loading
(•) Slurry or Solids Composition
Gas Stream
— Liquor Stream
-------
KJ
FIGURE 6
EFA TEST FACILITY
TYPICAL PIOCESS FlOW DIACfAM
FOt HTMO-FILT1I SfSTEl
SETTLING POND
O G»sComp«iUon
® Particulatc Cocnpositttm 4 Loading
© Slurry or Solids Composition
—> _ Gas Stream
__ Liquor Sirnm
-------
Section 4
TEST PROGRAM
The following contains a brief description of the test program. A
more detailed description has been presented in Refs. 2 and 3. In
Table 1, a description of the reports which are presently scheduled
for general distribution is presented,
4. 1 TEST PERIODS AND TEST PROGRAM SCHEDULE
The following sequential test blocks have been defined for the test
program.
(1) Air-Water Testing
(2) Sodium Carbonate Testing
(3) Limestone Wet Scrubbing Testing
(4) Lime Wet Scrubbing Testing
The test program schedule is presented in Figure 7. As can be
seen, the air-water and sodium carbonate tests have been completed.
As of mid April 1973, limestone wet scrubbing short-term factorial
tests were approximately 90% complete and longer-term reliability
verification tests were approximately 25% complete.
248
-------
Table 1
TOPICAL AND FINAL REPORT DESCRIPTION
K>
£
Report Title
1. Air-water, Sodium
Carbonate and Open-
Loop Limestone Test
Results
2. Limestone Wet-Scrubbing
Test Results
3. Lime Wet-Scrubbing
Test Results/Limestone
Reliability Test Results
Status
4. Final Report
Information to be Included
Summary of operational problems and resolutions,
planned and actual test designs, results of air-water
and NazCOj testing, utilization of data for model de-
velopment, results of open-loop limestone testing
with interpretation of data.
Summary of operating problems and resolutions
associated with reliability verification testing,
planned and actual test design, results of closed-
loop factorial tests, interpretation of data, status of
process model development and selection of param-
eters for limestone long-term reliability testing.
Summary of operational problems and resolutions
associated with lime reliability verification testing,
planned and actual test designs, results of factorial
lime testing, status of process model development,
interpretation of data and status of limestone relia-
bility testing.
Summary of total test program with particular em-
phasis on lime and limestone reliability test results,
mathematical models, scale-up design and economic
studies.
Estimated General
Publication^ Date
June, 1973
September, 1973
February, 1974
July, 1974
-------
FIGURE 7
PHASE III
SHAWNEE TEST SCHEDULE
TEST PROGRAM FUNCTIONS
SYSTEM CHECK-OUT
AIR-WATER & SODIUM CARBONATE TESTING
LIMESTONE WET-SCRUBBING TESTING:
Short-Term Factorial Tests
Reliability Verification Tests
Short -Term Factorial Tests
Reliability Tests
LIME WET-SCRUBBING TESTING:
Short -Term Factorial Tests
Reliability Verification Tests
Reliability Tests
ENGINEERING & COST ESTIMATE STUDIES
1972
MAMJ JASOND
123456789 10
•^•MMM
OPEN-LOOP TESTING-*
BOILER OUTAGE
SYSTEM MODIFICATK
1973
JFMAMJ JASOND
11 12 13 14 15 16 17 18 19 20 21 22
i 1 j — *-CLOSED-LOOP TESTING
_J |
i j
/ -
DNS
«••••
1974
J F M A M J
23 24 25 26 27 28
••
KJ
U1
O
-------
4.1.1 Air-Water Testing
These experiments, which use air to simulate flue gas and water to
simulate alkali slurry, are designed to determine pressure drop model
coefficients and observe fluid hydrodynamics (e.g. Hydro-Filter tur-
bulent layer) for all three scrubbers in clean systems.
4.1.2 Sodium Carbonate Testing
Two series of sodium carbonate tests have been designed. The first,
or high concentration series, utilizes concentrated (~1 wt. % sodium
ion) water solutions of sodium carbonate to absorb SO_ from flue gas
c*
and from a synthetic flue gas composed of air and SO_. These tests
L*
are designed to determine uncertain model coefficients for the case
where gas-side mass transfer is rate controlling. The second, or low
concentration series, uses dilute (~0. 1Z5 wt. % sodium ion) sodium
carbonate solutions to absorb SO from flue gas and synthetic flue gas.
£
For this series, gas-side mass transfer is not rate controlling and
liquid-side mass transfer uncertain coefficients can be calculated
using relationships for gas-side coefficients developed from the high
concentration tests. These runs also help ascertain the absorption
capability of liquors associated with some variations of the Double
Alkali scrubbing process (see Ref. 4) over a range of operating
conditions.
251
-------
4.1.3 Limestone Wet-Scrubbing Testing
The primary objectives of these test sequences for closed liquor
'i;
loop operation, are:
(1) To characterize, as completely as practicable, the
effect of important independent variables on partic-
ulate removal and SO;? removal.
(2) To identify and resolve operating problems, such as
scaling and demister plugging.
(3) To identify areas or regions for reliable operation
of the three scrubber systems, consistent with reason-
able SC>2 removal, and to choose attractive (economic)
operating configurations from within these regions.
(4) To determine long-term operating reliability with
attractive configurations for one or more of the scrub-
ber systems and to develop more definitive process
economics data and scale-up factors.
In order to accomplish the first objective, it will be necessary to
make a large number of "short-term factorial tests" for each scrubber
system. Based on tests performed to date, it appears that meaning-
ful data on SO and particulate removal can be obtained during tests
LJ
lasting from 4 hours to one day.
In order to accomplish the second and third objectives, a relatively
small number of longer-term (2+ week) "reliability verification tests"
will be made on each scrubber system. These longer-term tests
will also be useful to:
u
Closed vs. open liquor loop operation will be discussed subsequently.
252
-------
Obtain more reliable material balances.
Quantify any variations in SO and particulate removal
and system slurry compositions with time.
The fourth objective will be accomplished by running "reliability tests, "
lasting from 4 to 1 0 months, on attractive operating configurations for
one or more of the scrubber systems. During these tests, the systems
will be carefully monitored for potential long-term reliability problems,
such as erosion and corrosion of system components. The ability to
effectively operate such systems under varying load and SO inlet
conditions will also be studied during this test period.
Early during the short-term factorial test period (see Figure 7 ) it
became apparent that it was not feasible to operate the test facility
o..
*i*
in a totally closed liquor loop, without facility modifications. A
closed-loop test is a test wherein the raw water input to the system
is nearly equal to the water normally exiting the system from the
humidified flue gas and the waste sludge transferred to the pond. In
an open loop system, raw water input is significantly greater than
the water outflow in the exit gas and sludge. Therefore, process
liquor must be discharged from the system to maintain an overall
These modifications were completed during a five-week boiler-outage
in February and March, 1973 (see Figure 7 ). The major modifica-
tions included: eliminating pump sealing water on the Allen-Sherman-
Hoff pumps by changing from Hydroseals to Centriseals; humidifying
the hot inlet flue gases with slurry instead of with raw water; and
washing demisters with process liquor diluted with raw water instead
of with raw water only. Other major modifications to the systems during
the boiler outage, not necessarily effecting the water balance, are dis-
cussed in Ref. 1.
253
-------
water balance. In a commercial system such discharge may not be
acceptable due to potential water pollution problems. Also, during
open-loop operation reliability may be unintentionally enhanced since
the additional raw water added tends to desaturate liquors returning
to the scrubber, thereby tending to reduce scaling and plugging. It
is expected, however, that important SO removal data has been ob-
tained during the short-term open-loop factorial testing, since other
pilot-scale testing has indicated that SO removal is relatively in-
dependent of liquor composition for limestone scrubbing systems.
To date, therefore, the order of limestone testing has been (see
Figure 7 ):
(1) Open-loop short-term factorial testing.
(2) Closed-loop long-term reliability verification testing.
4.1.4 Lime Wet Scrubbing Testing
This test series, which involves introduction of hydrated lime (calcium
hydroxide) directly in the scrubber circuit, will resemble the lime-
stone wet scrubbing test program. The major difference will be the
absence of any open-loop tests. Again, tests will be divided into
three general categories (see Figure 7 ): short-term factorial tests,
longer-term reliability verification tests and long-term reliability
tests.
254
-------
4. 2. LIMESTONE WET SCRUBBING TEST DESIGNS
4.2.1 Short-Term Factorial Testing
The test sequences for the short-term factorial experiments are all
full or partial factorial designs (with centerpoint replicates included)
based upon the chosen independent variables, their levels and the
restraints of time as outlined in Figure 7. The choice of the inde-
pendent variables and their levels was based upon pilot plant test
results, the restraints of the systems and results from mathemati-
cal models which relate the dependent and independent variables.
Table 2 shows the independent variables and levels which are to be
investigated during limestone factorial testing for the three scrubber
systems. These variables and levels are always tentative, since the
experimental test program is continuously reviewed and up-dated, as
information is generated. Intermediate variable levels for center-
point replicate runs (e.g., stoichiometric ratio =1.5) have not been
included in Table 2.
The results, to date, of short-term (open-loop) factorial testing for
the three scrubber systems are presented in Section 5.
4.2.2 Reliability Verification Testing
Assuming that each reliability verification test will last about 2 weeks,
and assuming 1/3 downtime for each system ( for inspections, clean-
ings, etc.), only about six tests can be made for each scrubber system,
255
-------
TABLE 2
INDEPENDENT VARIABLES AND LEVELS FOR LIMESTONE FACTORIAL EXPERIMENTS
Venturi System
Variable
Scrubber configuration
Gag Flow Rate
Liquor Flow Rate to
Venturi and to Spray Tower*
Stoichiometric Ratio ~
Percent Solids
Recirculated
Effluent Hold Tank
Residence Time
Venturi Pressure Drop
Number of Spray Headers
in Spray Tower
Levels
1 Venturi alone
2 Spray tower alone
3 Venturi and Spray tower
1 IS.OOOacfm
Z 30,000
1 300 gpm
2 600
1 1.00 mole/mole
2 1.75
1 6%
2 IS
1 2. 3 min
2 4.6
3 40
* 60
1 6 in. HO
29
* 12
1 2 headers
2 4
TCA System
Variable
Number of Stages
Gas Flow Rate
Liquor Flow Rate
Stoichiometric Ratio
Percent Solids
Recirculated
Effluent Hold Tank
Residence Time
Levels
1 1 stage
Z 3 stages
1 IS.OOOacfm
2 30,000
1 600 gpm
2 1200
1 1.00 mole /mole
2 1.75
1 67.
2 15
1 4. 6 min
2 30
Hydro-Filter System
Variable
Marble- Bed Height
Gas Flow Rate
Total Liquor Flow
Rate
Stoichiometric Ratio
Percent Solids
Recirculated
Levels
1 ? inches
Z 5
1 20, 000 acfm
2 30,000
1 400 gpm
2 BOO
1 1.00 mole/male
2 1.75
1 6%
2 15
10
Wl
* After Mid-May 1973, the maximum liquor flow to the spray tower will be 1200 gpm.
** Stoichiometric ratio is defined as moles CaCO^/mole SO inlet.
-------
given the restraints of time outlined in Figure 7. Obviously, not all
variables which are assumed to affect system reliability can be com-
prehensively studied within the six- run limitation.*
Earlier open-loop factorial runs were all made at high stoichiometries
(greater than 1.75 moles CaCO,/mole SO_ abosrbed) and had scrubber
•J Lt ••- ™ -•"-•--— •
inlet liquor pH's ranging from 6. 0 to 6. 3. It is planned that a majority
of the reliability verification tests will be made at reduced scrubber
inlet liquor pH's (5. 7-5. 9), reduced stoichiometries (1. 1- 1. 5 moles
CaCO /mole SG>2 absorbed) and, of course, slightly reduced SO-,
removals. These "reduced-pH" runs should result in improved over-
all system reliability, because of larger oxidation rates at the lower
pH's and, of course, increased limestone utilizations (less CaCO3
added/mole SC^ absorbed).
The higher oxidation rates at reduced pH's will result in improved clari-
fier operation and in a larger percentage of "seed" CaSO4 crystals within
the process slurry**. An increase in the percentage of CaSO4 crystals
will probably result in reliable operation at lower percent solids recir-
culated and/or in smaller effluent hold tank sizes (residence times),
which is economically desirable. The increase in limestone utilization
results, of course, in a reduction in waste mass solids handling and in
limestone requirements. These improvements may be offset, however,
by the need for larger scrubbers in order to obtain the specified removals.
*
These tests will be supplemented with reliability verification tests
with the EPA pilot TCA scrubbers at Research Triangle Park,
N.C. (see Section 4. 4),
Oxidation of sulfite to sulfate in the E. H. T. can also be increased
by air- sparging (Ref.._5).
257
-------
On Tables 3, 4, and 5, the proposed limestone reliability verification
test plans for the venturi, TCA, and Hydro-Filter systems are shown.
For the venturi system, the effect of percent solids and gas rate on
reliability, with two different demister types, will be tested. For the
TCA system, the effect on reliability of percent solids, residence time,
air sparging, and gas rates are tested. For the Hydro-Filter system,
the effect on reliability of percent solids, gas rate, and scrubber inlet
pH (at 10% solids) are tested. Solids separation tests for the clarifier,
filter, and centrifuge will be made on the three systems throughout the
entire test period.
The runs listed in Tables 3, 4, and 5 have been listed in the expected
order in which they will be made. An attempt has been made to start
the testing on each system at conditions which are expected to give
high probability for reliable operation (e.g. high L/G, high effluent
residence time).
As of late April 19*73, the initial runs on Tables 3, 4, and 5 for the
three scrubber systems have been in progress for about 20 days each
(see Section 7).
4.3 ANALYTICAL SCHEDULE
Samples of slurry, flue gas and limestone are taken during each test
run for chemical analyses, particulate size sampling and limestone
reactivity tests. Batch samples of coal are taken periodically for
chemical analysis. Locations of sample points are shown on Fig-
ures 4, 5, and 6. In addition, clarifier settling tests, filter leaf
tests, and filter and centrifuge "operational tests" are also being
258
-------
Table 3
PROPOSED TEST PLAN FOR RELIABILITY VERIFICATION
TESTS WITH VENTURI SYSTEM
Venturi and Spray Tower
Venturi pressure drop = 9 in. H~O
Liquor rate to venturi = 600 gpm
Residence time =20 min
Stoichiometric ratio = 1. 3-1. 5 moles CaCO3/moles SO2 absorbed (expected)
Inlet liquor pH = 5. 7-5. 9
Percent Expected
Run Solids Spray Tower Gas Rate, Demister Percent
No. Recirculated Liquor Rate, gpm acfm Type SO2 Removal
1
2
3
4
5
6
15
8
15
8
(a)
(a)
600
600
600
600
1200
1200
20,000
20,000
20,000
20,000
20,000
30,000
A
A
B
B
(a)
(a)
60-70(b)
60-70
60-70
60-70
70-80
65-75
(a) To be determined from runs 1-4.
(b) Expected SO2 removal at high-pH (6. 0-6. 3) is from 70-80%.
259
-------
Table 4
PROPOSED TEST PLAN FOR RELIABILITY VERIFICATION
TESTS WITH TCA SYSTEM
Three stages of TCA spheres, 5" per stage
Liquor rate =1200 gpm
Stoichiometric ratio = 1.3-1.5 moles CaCO3/moles SO2 absorbed (expected)
Inlet liquor pH = 5. 7 - 5. 9
Expected SO2 removal = 80 - 90%. For high stoichiometry (> 1. 5) and high
pH (6. 0 - 6. 3) expected SO2 removal is from 90 - 95%.
Run
No.
1
2
3
4
5
6
7(b)
Percent Solids
Recirculated
15
8
(a)
15
8
15
15
Residence
Time, min.
20 (EHT)
20 (EHT)
20 (EHT)
5-6 (EHT)
5-6 (EHT)
5-6 (EHT)
4. 6 (RT)
Air-
Sparging
No
No
Yes
Yes
Yes
Yes
No
Gas
Rate, acfm
20, 000
20,000
20, 000
20, 000
20,000
30, 000
20,000
(a) To be determined from results of runs 1 and 2.
(b) Will be run if time permits.
Note; "EHT" stands for effluent hold tank and "RT" stands for
recirculation tank.
260
-------
Table 5
PROPOSED TEST PLAN FOR RELIABILITY
VERIFICATION TESTS WITH
HYDRO-FILTER SYSTEM
Height of marble-bed = 3-1/2 inches
Liquor rate = 800 gpm (total)
Residence time = 30 min
Stoichiometric ratio = adjusted to give desired pH or SO2 removal
(expected between 1.3 and 1. 6 moles CaCO3/moles SO2 absorbed)
Run Percent Solids Approximate Gas Expected
No. Recirculated pH to Scrubber Rate, acfm SOg Removal, %
1 10 5.7 20,000 70
2 10 6.2 20,000 80
3 15 5,7 20,000 70
4 6 5.7 20,000 70
5 15 5.7 30,000 70
6 6 5.7 30,000 70
261
-------
conducted periodically. A typical analytical schedule for a limestone
\vet-scrubbing test and a summary of the analytical methods for deter-
mining important species in slurry, coal and alkali are presented in
Ref. 2.
4.4 EPA PILOT PLANT SUPPORT
In support of the Shawnee prototype testing activities, two smaller
scrubbing systems (300 acfm each), which are capable of operating
over a wide range of operating conditions, have been installed at
EPA's Research Triangle Park, N.C. facility. The small pilot scale
units are capable of simulating the TCA scrubber circuit and have
already generated large quantities of closed-loop data on certain TCA
configurations.
262
-------
Section 5
SHORT-TERM FACTORIAL LIMESTONE TEST RESULTS
5. 1 SO REMOVAL RESULTS
In this section, the significant SO removal results from the open-
loop limestone factorial test sequences are presented graphically.
The complete listing of all pertinent data, in tabular form, for all
of the open-loop runs made prior to the boiler outage (see Figure 7),
will be presented in the first interim report (see Table 1).
The SO removals during open-loop testing have all been corrected
for the dilution effect of water vapor and reheater gas pickup by the
flue gas. The SO removals have also been corrected for DuPont
SO- analyzer calibration errors associated with unstable composi-
tion of SO- calibration gas (from September 15 to October 13, 1972)
and the deterioration of the DuPont analyzer optical filters (from
June 23 to December 1, 1972)."
After mid-November, 1972, when a 60 wt. % limestone slurry addition
##
system was installed for the three scrubber systems, some problems
developed with the calibrations of the limestone additive magnetic
flowmeters at the reduced slurry flow rates. All of the stoichiometric
*These correction factors were furnished by the DuPont Co. The
corrections were small for SO2 removals greater than 70%.
**Previously, a 1* wt. % limestone addition system was in use.
263
-------
ratios after mid-November were higher than their "nominal" (see Table 2)
values and in excess of 1. 5 moles CaCO3 added/mole SO2 inlet. For these
high values of stoichiometric ratio, and at the prevailing high-inlet
SO_ concentrations (2300-3300 ppm), the SO. removals were not
Ci t*
significantly affected by changes in stoichiometric ratio (see Section
6. 1 and 6. 2).
Prior to system modifications made during the five week boiler outage,
reasonable material balance closures for calcium and sulfur could
only be obtained with the TCA system. During this period of opera-
tion, the venturi and Hydro-Filter systems still had the clarifiers
and process water hold tanks included in the main slurry loops, while
the TCA system had been modified so that the main slurry stream
circulated between the hold tank and scrubber, with a bleed stream
from the main slurry stream routed to the solids separation area
*
(clarifier, vacuum filter or centrifuge) . The poor material balances
for the venturi and Hydro-Filter systems were attributable to solids
build-ups (or depletions) within the clarifiers. For the TCA scrubber,
the calcium and sulfur leaving the system could be obtained from the
measured flow rate and solids analysis of the "bleed stream" to the
solids separation area, and the clarifier could be excluded from the
material balance enclosure. Typical system configurations for the
open-loop limestone tests have been presented in Ref. 2, along with
the material balance results for six TCA runs. The closures were
all within the limits of the estimated experimental accuracies during
those periods of time.
*
During the five-week boiler outage, the flow configurations for the
venturi and Hydro-Filter were converted to ones similar to that
of the TCA system. These modified configurations have been shown
in Figures 4, 5 and 6.
264
-------
Although satisfactory material balance closures were not obtained
during the open-loop factorial testing for the venturi and Hydro-Filter
systems, confidence in the generated data is based on the following:
• "Wet" chemical analyses for SO- in the inlet and
exit gas streams repeatedly corroborated
DuPont SO analyzer measurements.
• Sulfur removals in longer-term reliability veri-
fication runs, with excellent material balance
closures for calcium and sulfur (see Section 7),
have been in close agreement with open-loop
replicate'- runs for the venturi, Hydro-Filter and
TCA systems.
It is recognized that SO removal is affected by SO. inlet gas con-
centration and scrubber inlet liquor temperature (Ref. 6). Car«,
therefore, has been exercised in segregating these non-controlled
independent variables in the presentation of the data.
5. 1. 1 Venturi System
In Figure 8, the effect of gas and liquor flow rates on SO, removal
&
for the venturi scrubber with 9 inches of pressure drop is shown.
Only data taken after December 1, 1972 has been plotted, since there
is some evidence that SO- removal corrections (for the deteriora-
Lt
tion of the DuPont optical filters) were not accurate at low SO,
tt
removals.
j_
Replicate runs are made with identical values for all controlled
independent variables.
265
-------
FIGURE 8
PRELIMINARY RESULTS FOR SCL REMOVAL IN THE
CHEMICO VENTURI WITH A NINE INCH PRESSURE DROP.
60
50 ••
40 .
i
Ul
otf
CM
O 30 +
Ul
u
20 -f
LIQUOR FLOW RATE = 600 gpm
LIQUOR FLOW RATE = 300 gpm
STOICHIOMETRIC RATIO > 2 moles CaCOa/mole SO2 absorbed
10 -f SOj INLET CONCENTRATION = 2200-3000 pom
PERCENT SOLIDS = 5-8%
HOLD TANK RESIDENCE TIME = 33-70 min.
SCRUBBER OUTLET LIQUOR TEMP.= 107-120 °F
15,000 20,000 25,000 30,000
GAS FLOW RATE, acfm @ 330°F
266
-------
Figure 9 illustrates the effect of gas and liquor flow rates on SO re-
moval in the four-header spray tower. The outlet liquor temperatures
(shown in the figure) varied considerably from run to run. The effect
of outlet liquor temperature on SO removal was generally consistent,
with higher SO removal at lower temperature. A curve representing
c*
a median liquor temperature of approximately 100 F has been drawn
for each liquor flow rate.
Figure 10 is a cross-plot of Figure 9, showing the effect of liquid-to-
gas ratio and gas velocity on SO- removal at a scrubber outlet liquor
Ct
temperature of about 100 F. The SO_ removals are outside the range
£*
•tf
of interest for commercially acceptable gas velocities (>7 ft/sec). '
The results from Figure 10 appear to agree reasonably well with the
spray tower data taken by the Hydro-Electric Power Commission of
Ontario (Ref. 7), after correcting for the effects of inlet gas SO»
concentration (see Section 5. 1.2).
;'»;'*
A spray tower limestone depletion run 463-1A was made to deter-
mine the effect of stoichiometric ratio and inlet scrubber liquor pH
on SO_ removal. Results from this run are presented in Table «> and
£
in Figure 11. SO^ removals for this run were low because, at that
L*
time, the liquid-to-gas ratio could not be maintained greater than
A modification to increase the maximum liquor rate from 600 to
1200 gpm for the spray tower is scheduled for completion by end
of May, 1973. This will allow for a liquid-to-gas ratio of 53 gal/
mcf at a gas velocity of 7. 5 ft/ sec, which should produce an SO2
removal of about 65%.
##
A limestone depletion run is a run in which no limestone make-up
is added during the test period.
267
-------
FIGURE 9
EFFECT OF GAS AND LIQUOR FLOW RATES
ON S02 REMOVAL IN THE FOUR-HEADER SPRAY TOWER
70 -•
1(78-92)
60 ••
$50
£
(71-103)
320 gpm
2-3% SOLIDS
30 ••
O LIQUOR RATE = 450 gpm
^ LIQUOR RATE = 300 gpm
STOICHIOMETRIC RATIO = 1.5-3.0
PERCENT SOLIDS = 5-9%
RESIDENCE TIME = 40-106 mfn.
SO2 INLET CONCENTRATION
= 2,500-3,400 ppm
(98-111)
(102-114)
(109-114)
(79-116)
(97-113)
LIMESTONE
DEPLETION
RUN 463-1A
HIGH STOICH.
RATIO
(113-120)
(111-118)
NUMBERS IN PARENTHESES REPRESENT
LIQUOR OUTLET TEMPERATURES( °F).
J80-101)
(116-123)0-—
20
10,000
20,000
GASRATE,acfm@330°F
30,000
268
-------
FIGURE 10
EFFECT OF LIQUID-TO-GAS RATIO AND GAS VELOCITY
ON S02 REMOVAL IN THE FOUR-HEADER SPRAY TOWER
70 ••
60 ••
40 -•
30 ••
20
SO2 INLET CONCENTRATION = 2,500-3,400 ppm
STOICHIOMETRIC RATIO = 1.5-3.0
PERCENT SOLIDS = 5-9%
HOLD TANK RESIDENCE TIME = 40-106 min.
SCRUBBER OUTLET LIQUOR TEMP.» 100 °F
3.7 ft/sec
2.5 ft/sec
10
20
30 40 50
LIQUID-TO-GAS RATIO, gal/mcf
60
70
269
-------
TABLE 6
SPRAY TOWKH LIMES CONE
IUIN WITH FOUIl UF.ADKRS
RUN NO. 4l.f-IA
f
PATE
1/21/71
IN
TIME (5"
noiw w
MOO 7.
1200
lino
1400
1500 6.
li.OO
1700
1806
LET .Sl.UHKV
-IRS
, ,i f) 1 1
. •
70 7. 30
n i.. =>o
(.. 10
6.25
6.20
1900 6.40 6.30
1/24/73
2000
2100
2200
2300 . <:.
2400
0100
0200
0300 7.
0400
0500
0600
(>.2O
(,.20
u.OO
9=, f.. 10
6. 10
1'., 2 »
6.30
93 6, 10
6. 30
6. 10
6.20
O700 8.92 6.15
0800
0900
1000
6. 10
(,.25
!,. JO
1100 9.00 6.10
1200
1300
i,. 50
6. 10
mif.T ^
<:Sc. REM°
ppm
1125 71.
1094 70.
1094 60.
106) 58.
303! 53.
1031 49,
1011 48.
1000 47.
2969 47.
3000 46.
1125 43.
3250 44.
3187 44.
1063 44.
1063 44.
3063 44.
3000 41.
3063 42.
3061 41.
3125 42.
3125 42.
3156 42.
3156 41.
3031 42.
2875 41.
2813 41.
2813 42.
SrOIClHOMETRJC RATIO,
VAL. 10 PAL MOI.RSC.^O,
MOI.K SOj ARSOHBRD
0
0 1H
0 19
0 13
0 10
0 H. 6
0 7.4
0 6.5
0 S. H
» S.I
5 4.H
5 4. S
0 4. 1
5 3.'*
5 l.i.
0 1.4
5 1.2
5 1. 1
0 2.9
0 2.H
0 2.7
5 2.6
0 2. 5
0 2.4
0 2.3
5 2.2
5 2. 1
1400-V
• j System down due to high fan vibration.
210oJ
2200
6. 15
2300 5.95 6.20
1/Z5/7J
2400
0100
0200
6.20
6. 10
t.05
0300 6.60 6.00
0400
0500
0600
6, 10
6.00
6.00
0700 7.95 6.20
0800
0900
1000
6.20
6,00
6.0;
1100 6.62 6.10
1200
1100
1400
1500
6. 10
6.00
2812 51.
2812 47.
2750 43.
2812 44.
2562 41,
5 2. 1
0 2. 1
5 2,0
5 2.0
5 .0
2812 47.4 .9
2750 46.0 .8
2812 47.0 .8
2750 44.5 ,8
2750 44.
2813 45.
5 .7
5 .7
2813 43.0 .6
2813 43.
2500 44.
5 .6
5 .6
2438 43.0 .0
1375 43.0 .5
2375 43.0 .5
5.90
1600 5. B5 5.90
1700
180(1
5.95
5.70
1900 7.50 5.80
2000
2100
2200
5.70
5. 65
5. 60
2300 6.22 5.70
1/26/73
2400
0100
02 00
5.70
5. bO
5.60
0300 6, 7H 5. 50
0400
0500
0400
5.40
S.45
5.35
0700 6.99 5.45
osao
0900
1000
5.40
5.20
5.20
1100 h. 10 S.40
1200
5. 30
2375 40.0 .5
2312 36.0 .5
2000 36.0 .4
2312 36.0 .4
218K 35.0 .4
1812 32.5 .4
1750 30.0 ,4
2000 33.5 .4
1918 31.5 .4
1875 32.5 .3
1688 11.5 .3
1781 31.5 .3
1 906 11.5 . ,
1906 10.0 .3
1813 29.5 .1
1906 29.0 .}
2000 2H.O .)
2344 2f,.0 .3
2344 24.0 .2
2500 23,0 .2
2500 21.0 .2
2500 21.0 .2
270
-------
FIGURE 11
EFFECTOFINLETLIQUORpH
ON S02 REMOVAL IN THE FOUR-HEADER SPRAY TOWER
(LIMESTONE DEPLETION RUN NO. 463-1A)
60
50 +
§ 40 ••
O
*-
Ul
y 30 ••
20 ••
GAS RATE = 20,000 ocfm @ 330 °F
LIQUOR RATE = 450 gpm
SO2 INLET CONCENTRATION =1,750-3,200 ppm
PERCENT SOLIDS = 6-9%
SCRUBBER OUTLET LIQUOR TEMP. = 97-113 °F
HOLD TANK RESIDENCE TIME = 56 min.
HIGH STOICH. RATIO
(GREATER THAN APPROX.
1.4 MOLES CaCO«/MOLE
SO ABSORBED)
10 ' I I I I I 1 | I I 1 I I I I I I I I |
5.0 5.5 6.0 6.5 7.0
SCRUBBER INLET LIQUOR pH
271
-------
approximately 30 gal/mcf.
j-
In Table 6, the stoichiometric ratio has been calculated from the
estimation of the original Ib-moles of CaCO in the system and of
the SO absorbed. A comparison between the stoichiometric ratios
calculated in this manner with those obtained from the solids analysis
could not be made, unfortunately, because of uncertain solids analy-
tical results during this period.
As can be seen from Table 6, the SO removal and inlet liquor pH
L*
remained at steady values of 40-44% and 6. 0-6. 3, respectively, for
a long period of time (about 35 hours) before the removal and pH
began to drop. The stoichiometric ratios for this period of time were
greater than 1.4 moles CaCO- added /mole SO absorbed. The SO
3 £• {*
removal for this high stoichiometry region has been included as a
data point in Figure 9.
Figure 11 shows the effect of inlet liquor pH on SO removal for Run
463-1A as the limestone in the system was depleted. Similar effects
of pH upon SO. removal have been reported in Refs. 8 and 9.
The stoichiometric ratio (moles CaCOj/moles SO2 absorbed) of
the scrubber inlet liquor changes with time as the SO? is absorbed
(i. e., one mole of CaSOx is formed and one mole of CO2 is
evolved for every mole of SO, absorbed).
272
-------
5.1.2 TCA System
i't
The results of the EPA TCA limestone runs are summarized in Fig-
ures 12 through 15.
Figure 12 shows the effect of height of spheres (5 and 10 inches;/stage)
and gas rate on SO_ removal in the TCA with six grids and three stages,
b
The effect of spheres vs. no spheres in the six-grid TCA on SO~ re-
moval is illustrated in Figure 13.
The liquor and gas rate effects on SO removal in the four-grid three-
stage TCA are presented in Figure 14. Figure 15 is a cross-plot of
Figure 14, showing the effect of liquid-to-gas ratio and gas velocity
on SO removal.
tL
The variation in SO removal for 5 inches of spheres per stage in the
six-grid, three-stage TCA (shown as open circles in Figures 12 and
13) is attributed, mainly, to differing average values of SO- inlet
concentrations. In Table 7, the SO removals and operating conditions
£
for these runs have been compared..
R.H. Borgwardt (Ref. 9) of EPA has reported that, for his pilot scale
TCA (see Section 4.4), the percent SO- removal is inversely pro-
iL
portional to the one-tenth power of inlet SO? concentration. The
From November 4, 1972 to January 15, 1973, TVA has conducted
a special series of tests with the TCA scrubber to provide process
and equipment scale-up and design information for the 550 Mw coal-
fired TVA Widows Creek Unit 8 retrofit limestone scrubbing system.
The results shown on Figures 12 through 15 do not include the TVA
tests.
273
-------
FIGURE 12
EFFECT OF HEIGHT OF SPHERES AND GAS RATE
ON S02 REMOVAL IN THE SIX-GRID THREE-STAGE TCA
100
95 -
A (7.1-7.9 In.H.O)
90 •• J-
m^
Q 85 4
0(4.4-5.1 ln,H90)
80 -
75 ••
70 ••
O"
.J
(8.2-9.9 In. H20)
(5.9-7.7 in.H20)
1(12-15 !n.H2O)
LIQUOR RATE = 1 , 170-1 ,220 gpm
SO2 INLET CONCENTRATION = 2,400-3,300 ppm
STOICHIOMETRIC RATIO = 1.2-2.0
PERCENT SOLIDS = 6.5-10.5%
HOLD TANK RESIDENCE TIME = 4.6 min.
SCRUBBER OUTLET LIQUOR TEMP. = 111-125 °F
HEIGHT OF SPHERES
5 INCHES/STAGE
10 INCHES/STAGE
65
NUMBERS IN PARENTHESES REPRESENT TOTAL
PRESSURE DROPS (EXCLUDING DEMISTER).
1 1 1 1 1 1 1 1 1 1 1 1 1 1
15,000
20,000 25,000
GAS RATE, ocfm@280°F
30,000
274
-------
100
95 -•
90 -•
85 ••
I
cTso
LU
u
75 -•
70 --
65 --
60
FIGURE 13
EFFECT OF SPHERES VS NO SPHERES AND
GAS RATE ON S02 REMOVAL IN THE SIX-GRID TCA
1 1 1 1 1
/ 102-1 18 °F
T /
i
X
(5.5in.H2O)
(6.2fn.H2O)
L) (2.0in.H20)
M (2.5in.H2O)
|(3.6?n.H0O)-
LIQUOR RATE =1,190-1,210 gpm .
SO2 INLET CONCENTRATION = 1,700-2,950 ppm
STOICHIOMETRIC RATIO ^ l .2-2.0
PERCENT SOLIDS = 7.5-11.5%
HOLD TANK RESIDENCE TIME = 4.6 min.
SCRUBBER OUTLET LIQUOR TEMP. = 110-120 °F (EXCEPT AS NOTED)
HEIGHT OF SPHERES
O 5 INCHES/STAGE (3 STAGES)
NO SPHERES
NUMBERS IN PARENTHESES REPRESENT TOTAL PRESSURE DROPS
(EXCLUDING DEMISTER & KOCH TRAY).
•H 1 1 1 1 1-
1 1 1 1 1
20,000
GAS FLOW RATE,acfm @ 280 °F
275
15,000
25,000
-------
FIGURE 14
EFFECT OF LIQUOR AND GAS RATE
ON S02 REMOVAL IN THE FOUR-GRID THREE-STAGE TCA
100
(3.6 In.
95 ••
90 -•
35 . .LIQUOR RATE=900 gpm'
80 ••
I/I
1
£ 70
65 -
60 -•
55 -•
50
T I
LIQUOR RATE=1200 gpm
(5.5 ln.H20) n(7.
(3.8ln.H20)_
LIQUOR RATE=600 gpm
(4.4in.H2O)
SO2 INLET CONCENTRATION = 1,800-2,500 ppm
STOICHIOMETRIC RATIO = 1.5-3.0
PERCENT SOLIDS = 6-11%
HOLD TANK RESIDENCE TIME = 18-35 mtn.
SCRUBBER OUTLET LIQUOR TEMP. = 111-123 °F
HEIGHT OF SPHERES - 5 INCHES/STAGE
«
NUMBERS IN PARENTHESES REPRESENT TOTAL PRESSURE
DROPS (EXCLUDING DEMISTER & KOCH TRAY).
H 1 1 1 1 1 1 1 1 1 1 1 1 1
15,000
20,000
GAS RATE,acfm @ 280 °F
25,000
276
-------
100
95 -•
90 -
85 -•
0^75 +
= 70-1-
65 -•
60 -
55 ••
50
FIGURE 15
EFFECT OF LIQUID -TO-GAS RATIO AND GAS VELOCITY
ON S02 REMOVAL IN THE FOUR-GRID THREE-STAGE TCA
9.8 ft/sec
sec
SO2 INLET CONCENTRATION - 1,800-2,500 ppm _
STOICHIOMETRIC RATIO » 1.5-3.0
PERCENT SOLIDS = 6-11%
HOLD TANK RESIDENCE TIME = 18-35 mln.
SCRUBBER OUTLET LIQUOR TEMP. - 111-123 °F
HEIGHT OF SPHERES = 5 INCHES/STAGE
20 30 40 50 60 70
LIQUID-TO-GAS RATIO, gal/mcf
-4-
80
90
277
-------
Table 7
EFFECT OF INLET SC>2 CONCENTRATION ON
REMOVAL IN A SIX-GRID THREE-STAGE TCA
Run No.
SO2 Removal, %
Inlet SOg Cone. , ppm
Gas Rate, acfm @ 280°F
Liquor Rate, gpm
Stoichiometric Ratio
Scrubber Outlet Liquor
Temperature, °F
Percent Solids Recirc.
Hold Tank Res. Time, min.
Ht. of Spheres/Stage, in.
Pressure Drop, in. ^O
409-2A
&
414- 2A
90+3
2800-3250
20, 100
1, 190
1.4-1.6
112-122
7-11
4.6
5
5.9-7.7
416-2A
95+1
1750-2200
20,000
1,195
>1.5
111-118
8-9
4.6
5
5. 8-6. 6
410-2A
b
411-2A
87+3
2500-3150
15, 100
1, 180
1.2-2.0
111-120
7-8. 5
4.6
5
4.4-5. 1
415-2A
95+2*
2250-2750
15,250
1, 200
>1.5
102-118*
7-11
4.6
5
5. 0-6. 0
*High removal may also be due to lower outlet liquor temperature.
278
-------
difference of 5% in SO removal between runs 409-2A and 416-2A,
for average SO_ concentration differences of 3000 and 2000 ppm,
respectively, is in agreement with EPA's pilot results. The 8%
difference in the SO removal between runs 410-2A and 415-2A is
attributable both to the differences in inlet SO_ concentrations and
the differences in the scrubber outlet liquor temperatures.
5.1.3 Hydro-Filter System
Figure 16 summarizes the effect of gas and liquor flow rates on SO_
L*
removal in the Hydro-Filter with five inches of marble bed height.
Figure 17 is a cross-plot of Figure 16, showing the effect of liquid-
to-gas ratio and gas velocity.
The extended dash-lines showing the ranges of SO_ removal in Figures
16 and 17 indicate where the ranges of SO_ removal would have been
£4
if corrections had not been made (for the deterioration of the DuPont
analyzer optical filters) for these runs. As mentioned previously,
there is some doubt about the accuracy of these corrections at low
SO removals.
5.2 ANALYTICAL RESULTS
A comparison between measured and predicted liquid and solids analy-
tical data for the venturi and TCA systems during open-loop testing is
presented in Section 6.2. Analytical data for the closed-loop testing
is presented in Section 7. 1.
279
-------
FIGURE 16
EFFECT OF GAS AND LIQUOR FLOW RATES ON
S02REMOVAL IN THE HYDRO-FILTER WITH FIVE INCHES OF MARBLES
2 80 +
04
8
I-
£ 60 ••
40 ••
20
TOTAL LIQUOR RATE = 400 gpm
TOTAL LIQUOR RATE = 600 gpm
TOTAL LIQUOR RATE = 800 gpm
SO2 INLET CONCENTRATION = 2,400-3,200 ppm
STOICHIOMETRIC RATIO = 1.1-1.9
PERCENT SOLIDS = 5-7%
HOLD TANK RESIDENCE TIME = 50 mtn.
100 + SCRUBBER OUTLET LIQUOR TEMP. = 115-125 °F
-11 1n.H0O)
*
(10-12 in.H2O)
ri T
U(B-9?n.H20) i
U(9-10Fn.H2O)
NUMBERS IN PARENTHESES REPRESENT HYDRO-FILTER PRESSURE
DROPS (EXCLUDING DEMISTER) IN A SCALE-FREE BED.
•I 1 1 1 1 1 1 1 1 1 1 » 1 h—
20,000 25,000 30,000
GAS FLOW RATE,acfm @ 330 °F
280
-------
FIGURE 17
EFFECT OF LIQIUD-TO-GAS RATIO AND GAS VELOCITY ON
S02 REMOVAL IN THE HYDRO-FILTER WITH FIVE INCHES OF MARBLES
100
80 ••
0 60
LUI
CM
-40 +
20 ••
10
7.7 ft/sec
SO2 INLET CONCENTRATION = 2,400-3,200 ppm
STOICHIOMETRIC RATIO = 1.1-1.9
PERCENT SOLIDS = 5-7%
HOLD TANK RESIDENCE TIME = 50 mm.
SCRUBBER OUTLET LIQUOR TEMP. = 115-125 °F
20
30 40 50
LIQUID-TO-GAS RATIO, gal/mcf
60
70
281
-------
5.2.1 Liquid Data
*
Table 8 shows the average scrubber inlet liquor compositions for the
open-loop factorial runs. During the period of open-loop testing,
there did not appear to be a continual build-up of magnesium, sodium
or chloride ions within the liquor. The large concentrations of chlor-
ide ions are attributable to chlorides present in the coal which were
converted to HC1 and absorbed from the flue gas in the scrubber. A.
Saleem of Ontario Hydro (Ref. 11) has reported similar chloride
concentrations during limestone wet scrubbing tests with flue gas
from a coal-fired boiler.
Table 8 shows that the venturi and Hydro-Filter systems had lower
overall dissolved solids than the TCA system. This was expected,
since the quantity of input raw water for these systems was greater
than for the TCA system (the TCA system was more "closed-loop").
Actually, the concentration of sulfate within the TCA liquor is close
to the predicted "saturation" level for sulfates.
The liquid analytical data are tested for consistency by inputting
the measured compositions and pH's into a modified Radian
Equilibrium Computer Program (Ref. 10), which then predicts
the ionic imbalance. For the data shown in Table 8, the ionic
imbalances were all less than 10%.
282
-------
Table 8
AVERAGE LIQUOR COMPOSITIONS AT THE SHAWNEE TEST
FACILITY DURING OCTOBER, 1972
C!** •pii'KI") f* i*
System
Species Concentration, mg/1 (ppm)
S
-------
5.2.2 Solids Data
Analyses of the Fredonia Valley limestone used at the Shawnee facil-
ity showed an average composition of 90% CaCO3, 5% MgCO3 and 5%
inerts. Dry sieve analyses showed approximately 90% of the ground
limestone passing through 325 mesh. A MikroPul* particle size analysis
showed approximately 7% of the ground limestone less than 3 microns,
30% less than 6 microns and 85% less than 20 microns.
The coal burned in boiler No. 10 during these limestone tests is Old
Ben 24 and contains approximately 18% ash, 10% total moisture, 3.2%
sulfur and 0. 3% chloride. The analyses of ash from boiler No. 10
showed about 50% SiO2, 18% Al2O3, 16% Fe2O3, 7% CaO, 1. 3% MgO,
1. 3% SO3, 2. 3% K2O, 1% Na2O and 3. 2% ignition loss.
The composition of solids in the slurry is determined by the moles
CaCC>3 added per mole SC>2 absorbed, the overall percent oxidation of
sulfite to sulfate within the system and the percent of fly ash. The mole
percent oxidations averaged approximately 30% during the open-loop
factorial testing and the fly ash comprised from 30 to 50 wt. % of the
solids for the three scrubber systems.
5. 3 PARTICULATE REMOVAL RESULTS
Particulate removal results for the three scrubbers are presented in
Tables 9, 10, and 11. Only those data which were taken at close-to-
isokinetic sampling conditions have been included in the tables. All
A division of United States Filter Corporation
284
-------
Table 9
PARTICULATE REMOVAL IN VENTURI AND SPRAY TOWER SCRUBBER
DURING OPEN-LOOP TESTS
K)
00
U1
Run No.
415-1A
414-1D
414-1D
414-1C
417-1A
414 -IE
418-1C
*
453-1B
*
454- IB
: #
456-1 A
Date
11-09-72
11-12-72
11-14-72
11-15-72
12-22-72
12-25-72
12-27-72
12-31-72
1-04-73
1-05-73
Gas
Rate,
acfm @ 330°F
30,000
30,000
29,900
29,900
30,000
30,000
14, 900
14,900
14,900
14,900
Liquor Rate,
gpm
Vehturi 1 Spray Tower
305 0
305 0
305 0
305 0
605 0
300 0
600 0
12 460
12 450
12 450
Pressure Drop,
in. HO
Venturi
9.0
9.0
9.0
6.4
9.5
12.0
12. 5
2.5
0.75
0. 70
Spray Tower
1. 1
1. 0
1.0
1.0
1.0
1. 0
0.2
0. 25
0. 25
0. 25
Grain Loading,
g/scf
Inlet
4. 38
2. 1
3. 32
3.40
3. 38
4. 17
6.39
2.6
4. 62
3. 38
Outlet
0.012
0. 010
0. 013
0.02
0. 012
0.009
0. 114
0. 004**
0.07
0.056
Percent
Removal
99. 7
99. 5
99. 6
99.4
99.6
99.8
98.2
99.8
98. 5
98. 3
Spray tower only.
"Data point questionable.
-------
Table 10
PARTICULATE REMOVAL IN TCA SCRUBBER WITH NO SPHERES
DURING OPEN-LOOP TESTS
Run No.
Date
Gas
Rate,
acfm @ 330°F
Liquor
Rate,
gpm
Pressure
Drop, in. H2O
Grain Loading, g/scf
Inlet
Outlet
Percent
Removal
KJ
00
WC-5
WC-5A
WC-5A
WC-11
WC-12
12-21-72
1-06-73
1-09-73
1-12-73
1-14-73
19,200
19,300
19,300
19,400
19,300
730
730
730
745
375
1. 5
1. 5
1.5
1.5
1.0
*
1. 70
4.16
1. 32
3.29
3.65
0. 004
0. 029
0.019
0. 017
0. 022
99.8
99. 3
98. 6
99. 5
99. 4
These listed values are the assumed pressure drops across the scrubber. Increases in total pressure
drop for these runs were most likely due to pluggage of the inlet gas duct during testing.
-------
Table 11
PARTICULATE REMOVAL IN HYDRO-FILTER SCRUBBER DURING OPEN-LOOP TESTS
Run No.
427-3A
427-3A
426-3B
427-3C
427-3B
428-3A
428-3A
428-3A
438-3A
440-3A
440-3A
Date
11-13-72
11-16-72
11-28-72
12-02-72
12-24-72
12-28-72
12-29-72
12-30-72
1-07-73
1-11-73
1-13-73
Gas
Rate,
acfm @ 330°F
20,000
20,000
20,000
20,000
20,000
20,000
20,000
20,000
19,900
12,500
12,500
Liquor
Rate,
gpm
810
810
810
800
805
810
810
810
400
600
600
Pressure
Drop, in. H^O
12. 0
12.0
10.0
12. 5
11. 0
11.5
11. 5
11. 5
7.0
6.8
6.8.
Grain Loading, g/scf
Inlet
2.6
3. 32
4.43
4. 24
2.19
3. 78
4. 12
3.63
4.20
3. 82
3.59
Outlet
0.030
0. 035
0.032
0.033
0.027
0. 025
0.016
0. 035
0.020
0. 042
0. 066
Percent
Removal
98. 8
98. 9
99. 3
99.2
98, 8
99. 3
99.6
99.0
99.5
98.9
98.2
hJ
00
-------
of the outlet particulate data have been corrected for soot-
contamination from the gas rcheaters. The soot amounted to less
than 30% of the total mass of the outlet particulates.
During the open-loop factorial testing there were solids accumulations
*.tf
'i1*
(depositions) in the demisters for much of the test period. These
solids accumulations may explain some of the very low measured out-
let grain loadings in Tables 9, 10, and 11, especially for the TCA
(multi-grid tower) at 1. 5 in. HO of pressure drop.
Ci
During open-loop testing, the demisters were all washed from above
with raw water (see Figures 1, 2, and 3 in Ref. 2). During the boiler
outage, provisions were made for washing the demisters from below,
with a mixture of clarified liquor and raw water, and for the installa-
tion of a Koch tray in the TCA scrubber (see Figure 2).
288
-------
Section 6
MODELING OF OPEN-LOOP LIMESTONE DATA
Results of the open-loop factorial limestone test program are pre-
sented in Section 6. 1 as equations relating sulfur removal to mea-
sured parameters for the three scrubber systems. These "linear"
equations, which were produced by a statistical analysis of the data,
are the simplest available that can adequately depict the data within
the range of variables measured.
In Section 6.2, a theoretical approach is employed, for the spray
tower, TCA and Hydro-Filter, to again relate SO removal to the
c*
measured parameters. A general closed-form equation is developed,
which is not incompatible with boundary constraints, and which should
permit reasonable extrapolations. Those variables which were
found to be significant in Section 6. I were introduced into the general
closed-form equations.
In Section 6. 3, complex mathematical models are discussed for com-
puting SO2 removal and slurry compositions for the scrubber systems.
These models are, generally, systems of partial differential equa-
tions, which are solved with numerical computer methods.
Other models for predicting pressure drops, particulate removal and
gas and liquid.side mass transfer coefficients, will be presented in
the first interim report (Table 1).
289
-------
6. 1 STATISTICAL MODELS FOR SO2 REMOVAL
Presented in this section are the results of a statistical analysis of
the data from the open-loop limestone factorial runs. The objective
of the linear equations presented in this section is to identify the im-
portant independent variables affecting SO2 removal.
Variables may not appear in the linear equations for a number of
reasons. Some of these are:
The variable did not significantly affect
removal over the range tested, i.e. the vari-
able was not statistically significant in improv-
ing the fit of the equations over the range tested.
The effect of the variable was masked by a simul-
taneous variation of more significant variables.
The variable was substantially constant for the data
set being analyzed.
The variable was "non-controlled" (e.g., inlet gas
SO2 concentration, liquor temperature) and may
not have varied in a manner condusive to determina-
tion of its effect on SO removal.
6. 1. 1 Venturi Scrubber
A stepwise regression analysis of the 10 venturi runs (with no slurry
in after-scrubber) made in December, 1972, resulted in the following
equation:
290
-------
% Removal = 9.4 + 0. 04 L + 0. 9 Ap H)
The range of variables covered by these runs, and therefore the reg-
ion of application, is:
Gas Flow Rate: 15, 000-30, 000 acfm
Liquor Flow Rate (L): 300-600 gpm
Pressure Drop (Ap): 6—12 in. I^O
Inlet SO2 Concentration: 2, 400-2, 800 ppm
Stoichiometric Ratio: 1.0—3.0** moles CaCC>3/mole SO2 inlet
Percent Solids Recirculated: 6-7%
Hold Tank Residence Time: 33-70 min.
Scrubber Outlet Liquor Temperature: 112—117 F
Equation 1 accounts for 89% of the total variation of the data. Gas
rate, Stoichiometric ratio, hold tank residence time, percent solids,
inlet SO concentration and liquor temperature were not statistically
ti»
significant (over the above ranges) in improving the fit of the equation.
6.1.2 Spray Tower
The following equation was fit to data from the 15 open-loop test runs
made with the four-header spray tower with no liquid to the venturi
(see Figures 9 and 10):
The effect of pressure drop on SO2 removal was only observed be-
low 9 in. H2O. Changes in pressure drop above this value (e.g.,
9-12 in. H2O) had little effect on SO2 removal.
High stoichiometries O1.75 moles CaCOs/mole SO2 inlet), after
mid-November 1972, were the result of calibration problems with
the limestone additive flowmeters (see Section 5. 1).
291
-------
% Removal = 16 + 0.9 L/C (2)
The range of variables covered by these runs, and therefore the reg-
ion of application, is:
Gas Flow Rate: 10, 000-30, 000 acfm
Gas Velocity: 2.5-7.5 ft/sec
Liquid-to-gas Ratio (L/C): 13-61 gal/mcf
Inlet SO2 Concentration: 2, 700—3, 300 ppm
Stoichiometric Ratio: 1. 0-3. 0 moles CaCO3/moles SO2 inlet
Percent Solids Recirculated: 2-8%
Hold Tank Residence Time: 40-106 min.
Scrubber Outlet Liquor Temperature: 98—128°F
Equation 2 accounts for 95% of the total variation of the data. Inlet
SO- concentration, Stoichiometric ratio, percent solids recirculated,
hold tank residence time and liquor temperature were not statistically
significant (over the above ranges) in improving the fit of the equation.
At constant L/C, a gas velocity of 7. 5 ft/sec gave slightly more
removal (~3%) than 5 ft/sec. This velocity effect was not observed
below about 5 ft/sec. Percent solids were actually between 6—8%
for most runs, with only one run at 2%.
6. 1. 3 TCA Scrubber
The following equation was fit to the 31 EPA TCA runs (see Figures
12 through 15):
% Removal = 47 + 0. 034 L + 1.4 PSR 4- 0. 5 Hp - 0. 006 ppm (3)
292
-------
The equation accounts for 85% of the total variation of the data. The
range of variables covered by these runs, and therefore the region of
application, is:
Gas Flow Rate: 15,000-27, 500 acfrn
Gas Velocity (V): 6-11 ft/sec
Liquor Flow Rate (L): 600-1,200 gpm
Percent Solids Recirculated (PSR): 6-14%
Inlet SO2 Concentration (ppm): 1, 800-3, 200 ppm
Stoichiometric Ratio: 1. 0—3. 0 moles CaCO3/mole SO2 inlet
Hold Tank Residence Time: 3. 8-35 min.
Scrubber Outlet Liquor Temperature: 105—122°F
Number of Grids: 4, 6
Total Height of Packing (Hp): 0-30 in.
Percent removal decreases with increasing inlet SO_ concentration
£*
(~6% per 1000 ppm). Gas velocity, Stoichiometric ratio, hold tank
residence time, scrubber outlet liquor temperature, and the number
of grids were not statistically significant (over the above ranges) in
improving the fit of the equation. Although not an independent varia-
ble, the pressure drop in the scrubber was also examined and was
found to not have a significant effect on removal. For example,
runs giving 92% removal have been made at pressure drops of 4, 6
and 9 in. H2O.
The 17 EPA six-grid TCA runs were analyzed as a group. The follow-
ing equation was fit to these runs (see Figures 12 and 13):
% Removal = 67 + 0. 02 L + 1. 0 V + 0. 44 Hp - 0, 006 ppm (4)
The equation accounts for 78% of the total variation of the data. The
analysis is restricted to the previously mentioned range of variables,
293
-------
with the exception of percent solids recirculated, which only varied
from 7—10%. Note the gas velocity term for this group, which was
not statistically significant for the entire set of runs (see Eq. 3).
Again, stoichiometric ratio, hold tank residence time, and scrubber
outlet liquor temperature did not significantly affect SO removal
over the above ranges.
The 14 EPA four-grid three-stage TCA runs were also analyzed as
a group. The following equation was fit to these runs (see Figures
14 and 15):
% Removal = 53 + 0. 04 L + 1.4 PSR - 0.007 ppm (5)
The equation accounts for 96% of the total variation of the data. These
test runs were made with 5 inches of spheres per stage, for a total of
15 inches. The range of variables is otherwise the same as that for
the 31 test run group. Gas velocity, stoichiometric ratio, hold tank
residence time, and scrubber outlet liquor temperature did not sig-
nificantly affect SO_ removal over the above ranges.
£+
The following equation was fit to the 16 runs made without spheres in
the five-grid TCA tower (TVA tests):
% Removal = 90 + 0.034 L-0.46 T_ (6)
Lj
The range of variables covered by these runs, and therefore the reg-
ion of application, is:
294
-------
Gas Flow Rate: 19, 000-30, 000 acfm
Gas Velocity: 7.5-12 ft/sec
Liquor Flow Rate (L): 375—1, 070 gpm
Inlet SO2 Concentration: 2, 200—3, 200 ppm
Stoichiometric Ratio: 1. 0—3. 0 moles CaCC^/moles SC>2 inlet
Percent Solids Recirculated: 14%
Hold Tank Residence Time: 5—15 min.
Scrubber Inlet Liquor Temperature (T ): 63—110 F
Scrubber Outlet Liquor Temperature: ^9—115 F
Pressure Drop: 1—7 in. HO
Equation 6 accounts for 65% of the total variation of the data. Stoi-
chiometric ratio, inlet SO- concentration and hold tank residence
time were not statistically significant (over the above ranges) in im-
proving the fit of the equation.
6.1.4 Hydro-Filter Scrubber
*
A stepwise regression analysis of 27 Hydro-Filter runs resulted in
the following equation (see Figures 16 and 17):
% Removal = 17. 9 + 0.1 L - 2. 0 V (7)
The range of variables covered by these runs, and therefore the reg-
ion of application, is:
Eleven runs made during October, 1972, were excluded from the
analysis due to doubtful low values of SO2 removal obtained dur-
ing this period. Recent closed-loop data has validated this
exclusion.
295
-------
Gas Flow Rate: 12,000-30,000 acfm
Gas Velocity (V): 3-8 ft. /sec.
Liquor Flow Rate (L): 200-800 gpm
Inlet SOZ Concentration: 2, 000-3, 500 ppm
Stoichiometric Ratio: 1.5—3. 0 moles CaCO3/mole SO2 inlet
Percent Solids Recirculated: 6—12%
Hold Tank Residence Time: 50 min.
Scrubber Outlet Liquor Temperature: 85—125 F
Height of Marbles: 3-5 in.
Equation 7 accounts for 94% of the total variation of the data. Inlet
SO- concentration, Stoichiometric ratio, percent solids, liquor tern.
perature, and height of marbles were not statistically significant
(over the above ranges) in improving the fit of the equation.
6. 2 CLOSED-FORM CORRELATIONS FOR PREDICTING SO-
REMOVAL
Analysis of the closed-loop limestone data, using the Radian Equilib-
rium Computer Program (Ref. 10), has shown that the equilibrium
Jf
mole fraction of SO_ over the bulk liquid,^- , is negligible with res-
pect to the SO- mole fraction within the gas, for the spray tower, TCA
* *
and Hydro-Filter Scrubbers . For this condition, the following rela-
tionship represents SO- absorption:
Due to low liquor residence times, the amount of limestone dis-
solved within the venturi scrubber is relatively small. Hence, y-
can be significant. Also, U/f can be significant for the spray tower,
TCA and Hydro-Filter scrubbers operating at low stoichiometries
(<1.5 moles CaCOs/ mole SO2 absorbed) and, consequently, low
inlet liquor pH's «6.0).
296
-------
where
- gas -side mass transfer coefficient
= gas -liquid interfacial area per scrubber volume
= axial distance
-gas rate per cross. sectional area
- Henry's law constant
- liquid- side mass transfer coefficient
Also, scrubber computer models utilizing previously fitted gas-side
mass transfer coefficients (see Ref. 2) have shown that liquid-side
resistance controls (i.e. , -f?t/tfZ
-------
T~ = liquor temperature, F.
The liquid-side coefficient is expected, therefore, to be a function of
gas and liquor flow rates, scrubber configuration, amount of dissolved
reactant, inte
temperature.
-i-
reactant, interfacial concentration of dissolved SO? (H SO ) and
£* C* j
The form of Eq. 9 has been fitted (by multiple regression) to the open-
loop limestone data for the spray tower, TCA and Hydro-Filter scrub-
bers. The significance of the independent variables in the fitted equa-
tions was demonstrated by the statistical analysis (see previous
section).
The open-loop limestone data was all obtained at relatively high stoi-
chiometries (greater than 1.5 moles CaCO /moles SO absorbed) and
consequently, high scrubber inlet liquor pH's (from 6. 0 to 6. 3). Within
this regime of operation, stoichiometric ratio showed no apparent ef-
fect upon SO- removal (see previous section). The effect of stoichio-
£+
metric ratio for the scrubber systems will, hopefully, be obtained
during the closed-loop testing now in progress. Other variables
which showed negligible effects upon SO- removal during the open-
loop testing, such as percent solids recirculated, may also have more
significant effects at reduced stoichiometries during closed-loop
testing.
* \j
As the concentration of SO, in the gas increases, XA increases and
^L decreases. This is an explanation for the empirical fact that
as the SOg inlet gas concentration increases, for limestone wet-
scrubbing systems, the SO£ removal decreases.
298
-------
The effect of inlet gas SC>2 concentration (a non- controlled independent
variable) upon SC>2 removal has only been included in the fitted equa-
tion for the four-grid three-stage TCA, although it is presumed that
a similar effect exists for the other scrubbers. Also, the effect of
inlet scrubber liquor temperature (a non-controlled independent vari-
able), which was determined to be significant from the TVA TCA runs
(see Eq. 6), has not presently been included in the closed-form equa-
tions. The effects of inlet gas concentration and temperature will be
included in the final forms of all the correlations, once the analyses
of other pilot data and the Shawnee closed-loop data has been completed.
6.2.1 Spray Tower
The following equation was fit to 15 open-loop limestone test runs
made with the 4-header spray-tower (see Figures 9 and 10 and
Eq. 2):
where:
= liquid to gas ratio in scrubber, gal/mcf
Equation 1 1 accounts for 94% of the variation in the data (correlation
coefficient of 0. 97) with a standard error of estimate of 3. 7% SO,
&
removal.
299
-------
6.2.2 TCA
The following equation was fit to 11 open-loop limestone test runs
made with the four-grid, three- stage TCA scrubber (see Figures
*
14 and 15 and Eq. 5):
»'-a*/ , \a^1
L ("K* J
where
L = liquor flow rate per cross -sectional area, gpm/ft
_ concentrati°n *n inlet gas, mole fraction
Equation 12 accounts for 99% of the variation in the data with a stan-
dard error of estimate of 1. 3% SO- removal. As previously men-
fit
tioned, it is assumed that the measured effect of ffSO*. for the TCA
scrubber will be similar for the other systems.
6.2.3 Hydro-Filter
The following equation was fit to 27 open-loop limestone test runs
made with the Hydro-Filter scrubber (see Figures 16 and 17 and Eq.
7):**
Two runs at relatively high weight percent solids and one "lime-
stone depletion" run were eliminated from this analysis.
*#
Eleven runs made during October, 1972, were excluded from the
analysis due to anomalously low values of SO£ removal obtained
during this period. Recent closed-loop data has affirmed this
exclusion.
300
-------
Equation 13 accounts for 95% of the variation in the data with a stan-
dard error of estimate of 4. 1% SO, removal.
Lt
6. 3 COMPUTER MODELS FOR PREDICTING SO, REMOVAL
AND SLURRY COMPOSITIONS
6. 3. 1 Scrubber Models
In Ref. 12, mathematical models were presented for predicting SO-
removal in the venturi, TCA and Hydro-Filter scrubbers. The models
are, generally, sets of partial differential equations which describe
SO. absorption into the process liquor (in accordance with the two-
film theory), 'reaction between the absorbed SO, (H SO ) and the
species in the liquor and the dissolution of solids (e.g., CaCO.) with-
in the process liquor. The assumption has been made, for these
systems, that the liquor is at all times in equilibrium with an inter -
facial vapor pressure of 0. 1 atrn of CO,, i. e., the rate of absorp-
tion of CO, from the flue gas is large. The thermodynamic equilibria
for the models are obtained from the Radian Computer Program
(Ref. 10).
Computer models have been written for the three scrubber systems,
which numerically solve the systems of differential equations. It has
been planned to fit the gas and liquid-side mass transfer coefficients
to the high and low-concentration soda-ash data (see Section 4. 1.2)
and then fit the solids dissolution rate constants to the limestone data.
The fitting of the gas.side coefficients for all three scrubbers has
been presented in Ref. 2. To date, only the liquid-side coefficient
301
-------
for the venturi scrubber has been fit to the low concentration soda-ash
data. The results of this fit will be presented in the first interim re-
port (Table 1), along with the correlations for the gas-side coefficients.
As discussed in the previous section, the equilibrium mole fraction
of SO? over the bulk liquid is essentially zero for the spray tower,
TCA and Hydro-Filter scrubbers, for the open-loop (high-stoichio-
metry) data. For this regime, therefore, the models describing SO_
absorption for the three scrubbers can be greatly simplified (see
Eq. 8). For the venturi scrubber, however, the residence time of
the liquor is low (<0.1 sec.), the dissolution of limestone within
the scrubber is small and, consequently, the equilibrium mole frac-
tion of SO, over the bulk liquid not zero everywhere, for the ranges
£•
of variables tested. Results from the venturi computer model, using
the previously fitted gas and liquid-side mass transfer coefficients,
have shown that an assumption of zero dissolution of solids will give
a reasonable fit to the open-loop limestone data. These results will
be presented in detail in the first interim report (Table 1).
6.3.2 Simulation Model
The simulation model is a computer model which determines the slur-
ry compositions of the waste streams and the scrubber inlet and out-
let streams for the three scrubber systems. The major assumption
in the model is equilibrium between the liquid and solids in the slurry
This implies that the kinetic rate of dissolution of limestone within
the scrubbers is high.
302
-------
leaving the effluent hold tank, at a specified equilibrium partial pres-
sure of CO . The equilibrium relationships between the liquid and
solid species are obtained from the Radian Equilibrium Program.
The simulation model takes as imput all of the independent variables
(e.g. , stoichiometric ratio), the percent sulfite oxidation, the percent
ash in the solids and the concentration of chloride and magnesium in
the process liquor. If a scrubber model (either simplified closed-
form or computer model) is used, the simulation model will (itera-
tively) predict SO- removal, as well as the slurry compositions. If
L4
no scrubber model is used, then SO removal must be input into the
simulation program, along with the independent variables.
Two results from the simulation model will be presented here. The
first simulation, for venturi runs 419-1A and 421-1A, is shown pic-
torially in Figure 18 and incorporates the venturi scrubber computer
model (which assumes zero dissolution of solids). The predicted
removal of 45% is close to the average measured values of 42 ± 5%
(see Figure 8 and Eq. 1). The second simulation, for TCA run
412-2A, is shown pictorially in Figure 19, and does not incorporate
a scrubber model (the measured removal of 96% was input to the
The specified CO2 partial pressure was chosen to match the mea-
sured E.H. T. outlet liquor pH's and compositions. Predictions
with the Radian program indicate relatively constant CO? equilib-
rium partial pressures from 0. 05 to 0. 1 atm.
*** *£
"" Ultimately, models will be developed for predicting sulfite oxi-
dation, ash in the solids and the concentrations of chloride and
magnesium in the liquor.
303
-------
FIGURE IS
BECHTEL LIMESTONE / LIME WET- SCRUBBING SIMULATION PROGRAM
JOjttMOVAL- 45%
SIMULATION OF VENTUtl SOtUlK* SYSTEM KUNS
419-lA/ttt-IA ON DECEMKR 23-25, 1*72.
KAL SCtUUH MOOCl USED IN SIMULATION.
Scnibbw Inlet Uquor lot. • 600 PJM ( 1 4,9
Go. lot.- I5,000ocf«.( 1.600t>-w>le/W>@ 3»°F
Inlet SO 2 Concentration . 2, 900 par*
Stolchloinerrlc Ratio « 1 .5
L/C- Mool/mcF
Percent Solid* Reclrcuwled - 6 %
Percent Soli* (Xtetwraed - 14 %
Percent Flyori) In Soil* - 40%
Percent Sulflfe Oxidation -20%
Concentration of Chloridd * 14 pnt-moli/IOOO liten
Scrubov T«np«aiuni < II7°F
MAX* ASSUMPTIONS:
o Effluent Held Tonfc Slurry in evitltbrtim
wlnSO.05ott.COj.
e G»SO3 & CaSO4 oancenrrationi In liquor
drmm leaving E.H.T. or» 1 J. 1 time* rtvt
evulllbrtum lalwrotian levelt. nwpectiwly.
e Seiubber Sluny (In contact with flue-gal)
In equilibrium with O.I arm CO., i.e.
rot* of frontier of CC>2 it l«nj*.
• Zero ditnlution of iclidt In icnibbor
fall diuolution of »li-«ole/nt C02^^
X
V r-
SPt
Total
Total
Total
Toto
Total
Toto
H2O
6.0 Ib-fwU/nr CoCO
0 Ib-mole/hr MgCC
^^-»H«3.4
,
Flow Rate, Ib-mole/hr
5reC>" UquW | Solid
Totol Co 9.7 109.3
Total MO 0 0
Ton! COj 0.6 78.0
TotolSOj 2.2 27.1
Total SO. 6.6 4.2
Total 0 4.3 0
MjO 16,900
1
„.„ Flow Rote. *t-«ole/nr
Clt- Liquid 1 Solid
Co 10.0 109.1
MQ 0 0
CO2 0.9 78.0
SO2 0.7 26.9
SO3 6.6 4.2
Cl 4.3 0
16,900
3 1 >• 1 . 4lb-mol.Ar CO j
EFFLUENT
WASTE DISPOSAL Wmre Stream
\
«dES ^%
Total Co 6.0
Total Mg 0
Total CO2 4.3
Total SOj 1.4
Total SO] 0.3
Total Cl 0.1
H2O 370
1 Mot«-uo Water
140pm
-------
SOj REMOVAL - 96 %
SIMULATION Of TO SCtUUER SYSTEM RUN
4J2-2A ON OCTOBEt 29-31, I972t
REAL SCRUBBER MODEL NOT USED IN SIMULATION
(SO. REMOVAL INFUT TO PROGRAM.)
o
VI
0.4fc-«oMirSO2
300 Ib-mle/hr CO,""* "I
9.0tt>-i»VWS02 J
300 fe-molo/W C02 ~
FIGURE »
BECHTa LIMESTONE , LIME WET-SCRUBBING SIMULATION PROGRAM
Scrubber Inlet Liquor Rote »1. ISOgpm (32,700 Ib-mole/M
Ga Rota = 27,SOOaefm ( 3.000 Ib-mole/W) @ 290°F
Inlet SO2 Concentration = 3,000p(xn
Stoichiometric Katie - 1.25
L/G = 55 gol/mef
Percent Solidt Reciroulated - 8.5%
P*rc««> Soli* Oivctwx^w) = 18%
Percent Flyah in Solldi « 30%
Percent Sulfite Oxidation = 30 %
Concentration of Chloride! * 42.5gm-fiiole/10QO liten
SaubUr Tempemlure = 120°F
I
MAJOR ASSUMPTIONS-.
e Effluent Hold Tank Slurry in equilibrium
wtthO.OSatmCO,.
e CbSOj & CoSO4 eonc*n*raHoni In liqwx
itream leaving E.M.T. are I & I fimei the
•avrllbrtum tatuntlon leveli, reipecflvely.
e Scrubber Slurry On contact with flue-gal)
in equilibrium •!* 0.1 atm CO i.e.
rate of tranler of CO2 li lanje.
pH-e.3
SPECIES
Flow Rare, Ib-molt/V
Total Co
Total Mg
Total CO 2
Total SO2
Total SO3
Total Cl
H20
Liquid | Solid
25.4
0
1.6
1.2
tl.S
25.0
32,700
321.«
0
72.2
177.9
71.4
0
11.3 lb-mol.Ar CoCO j <
0 *
Ib-mole/hr CO j
Nor Computed
pH = Not Computed
SPECIES
Total Co
Total Mg
Total CO2
Total SO2
Total SO3
Total Cl
HjO
Flow Rote,tt>-mole/Kr
Liquid | Solid
Not Computed
EFFLUENT
HOLD TANK
Waita Stream
1990"
SPECIES
Total Co
Total Mg
Total CO2
Total SO2
Total SOj
Total O
H2O
Flow RoM,
Ib-moleAr
11.3
0
2.6
6.0
2.6
0.4
4tt>
Make-up Water
-------
program). The agreement between the predicted and measured scrub-
ber inlet slurry compositions for this TCA simulation is shown in
Table 12.
306
-------
Table 12
COMPARISON OF MEASURED AND PREDICTED SLURRY
COMPOSITIONS AT SCRUBBER INLET FOR TCA RUN 412-2A
Gas Rate = 27,500 acfm
Liquor Rate = 1, 170 gpm
L/C = 5? gal/mcf
Pressure Drop = 14 in.
Three stages, 5 inches/stage
Species
pH
SO/
CO/
so/
Ca + +
Mg+ +
cr
Species Concentrations, gm mole/1
Liquid
Measured 1
5.9
1.8
1.2
24
35
5.5
43
Predicted
6.3
2. 1
2.7
19
43
0*
43*
000 liters
Solid
Measured
—
21
0
220
86
51
2
0
1
1 Predicted
—
300
120
120
550
0
—
:': Input to computer model.
307
-------
Section 7
RELIABILITY VERIFICATION TEST RESULTS
As mentioned previously, the objects of the reliability verification
tests are to (1) identify areas or regions for reliable operation con-
sistent with reasonable SO removal, (2) choose attractive operating
configurations from within these regions, (3) obtain more reliable
material balances, and (4) quantify any variations in SO and parti-
L*
culate removal and system slurry compositions with time.
As discussed in Section 4. 2. 2, the initial tests are to be run at re-
duced scrubber inlet liquor pH's (5.7—5.9), to increase system relia-
bility and limestone utilization. A modest reduction in SO_ removal
Ct
(from high pH performance) is the price of the increased reliability
and limestone utilization.
Presently, initial runs are in progress on all three scrubbers (see
run Nos. 1 on Tables 3, 4 and 5). The Hydro-Filter (Run 501-3A)
was sUrted.up on March 14, the TCA (Run 501-2A) on March 22 and
the venturi-spray tower (Run 501-1A) on April 9, 1973.
The performance data for these three reliability verification runs
will be presented in Section 7. 1 and the results of material balances
for sulfur and calcium (which were satisfactory) in Section 7. 2.
308
-------
7. 1 PERFORMANCE DATA
Data for the first 400-500 hours of operation on the initial runs is
summarized in Figures 20, 21 and 22. The upper section of each fig-
ure shows the operating periods (blank space indicates shut-down),
and such critical variables as SC>2 removal, liquor pH and stoichio-
metric ratio. The middle plot gives some analyses of solids in the
scrubber inlet liquor. The lower plot gives concentrations of some
dissolved species in the scrubber inlet liquor.
Also shown on Figures 20, 21 and 22, are the depletion (line-out)
periods for the tests. Fresh limestone slurries (no CaSO "seeding")
were introduced into the effluent hold tanks and SO- absorption was
ft
used to reduce the slurry pH until the desired level of SO- removal
was attained. This level was approximately 10% below that attain-
able in high-pH open-loop operation (see Section 4.2.2).
Before beginning limestone addition, the systems were inspected to
be sure that they were free of scaling or erosion that might have
occurred in the high-pH period of the line-out. Periodic (approxi-
mately weekly) inspection shut-downs are scheduled in order to moni-
tor scaling and erosion in sensitive areas of the systems.
Operability and reliability of the three scrubber systems during the
initial runs is discussed in another paper presented at this symposium
(Ref. 1).
309
-------
Ol£
•HOMUHtl-IA
jpnif
nW.\
r*»
, i 8
10 'I .? !L g fr
8 » > 7, §
n
. . -+>
3 O
1
ft)
ro
o
o
-a
a>
3
LQ
s-
sr
o
n>
-5
3
Ol
O
3' '- " I 1 go
n t— is*
fi §
^-» 3 3 IX*
? I" a.P
5 3
i
• TOfAL WHtHVIO lOHt» «- MAGNESIUM («• ** )
O CAtCWM fU ** ) fl iOOK»*CN«*l
cmMMia-l
O o«iON»n ico,, • I
CALEMOAI MY
-------
Ll£
z ™ -g -if
? i § s
o - -< x
.0' .o'
g S
I
txi
Q — • -i
I I
o 5 8
-s1 < a
1 s i
3 3'
8
•7
i
tn
o
-n
_>.
Ǥ
O
~
i
m
Q-
n>
O
-s
r+
C
-s
-5
01
O
T —• O"
i5 °- o- 3_
CQ n 5 «
ffl O "*
^ 3 J on
7^ £^- — /^i
^-ro =* M
°- § "° *
P i " f
n - N> -
SO r*0
I O
° K)
I ,
-------
ZL£
; I S O -g S
* -, < ii ? g"
«? 5" 8 2* *
1 * I O
=•53 o
n c c "
g" B: f 3;
* j? * *
D I i- O
5 0. "* M
I <; JF
O-
01
o
-s
c
CJ1
o
t\3
O I II O
O —• I
1
z
s
ff
2/40
z Q »
2s
5 ^ 10
n - 0
-X A
\ / \
\ / x — ^--
ill .
e y o 8
1 ^ »
ill
Clli CONC(
INSCIUMf
SOtlOS,
5 B
\
1/73 t V73 I J^ I I/SI
TEST tlMt, kun
^4 I V17 I 3XII
O.L1NOAI UV
* TOTAL OtISQLVCD SO4.HM * UAGNiSHJM 4M« )
O CAIOUM Ki ** t ^ IOWUM 0* * 1
D SUUATI 80 'I • JUIFITI (SOj- )
A OlOtHK (Cl- 1 O CAOOHAH fCOj - I
CT
? »
,.i
ion 5
8
-------
T •-
3C 2
<
Q TO Q
fi • A
S s' 3"
• _ o>
-S i' y
r * f
2 "
S S
©
3
9
-s
PO
o
o
rt
CD
o.
o
-a
n>
2
a.
Cli
01
o
-I
o
-s
cn
o
ro
3=>
?£ if
O 5 5"
o _ —
n
S
-
cj
S t
-n° 3
^
TOIM 611501VID HXIW * «»O«SIU*I
-------
:i ™ p C C O
3 x 8 £| 1
JL H # " i f
3T 30 O O
fill
28 i ^ -"
3 «
« -I °°
f j 3
I .
• "
" s
s
3
1 r
©
a i« «
|>
i
TOTAl, EXCLUDING KMISTM t KCWTH r*A*
-DCMISTEI & ICOCM Iwv
(£1
CD
ro
o r
M* -tl
O C
s~ o
-» M
•\
^OUTIEI fL>ll
O
o
3
O.
i^sl -»
o 8 '„ §
§ - - i
rll
o
T3
CU
f+
—i.
3
CL
fa
a>
-s
3
CJ1
I
r-o
HIT TIME, ham
4/13 I 4/l<
CAUNDAI OAY
• TOTAL DISSOIVIO SOLICH • MAONI1HIM (Ma " 1
O CALCIUM {fa** ) 6 1QOHJM [N= ' )
D SOLFATIBO^-) • SUlFlnOO,')
CWOtlOE Kl - )
O CAUONATE (CO3 p I
'•""i
&
. t"0.
4/13 li 4/IJ 1 4/14 I «/13 I '''^
CALENQU OAV
-------
SLE
*5 J?1 r
r y |- a f «
f §" "^ > 1 sS
f S « 5. 8 I
' =* - f «
« 3 f a |
'= i S I ©
® ^ W
S. IS
ro
ro
ro
O
•a
(T>
-s
o
-s
Q.
O
ff
-5
-S
c
S
?f 51 ?
n c c »
§" £ f ^
i e I *
o I s- o
? 1 i
- 2
P I
HiT TIMi, k..
3/19 I V30
CALErNDAI DAY
a.
00
00
* TOIAI. OIHOLVtO ICXHO
O CALCIUW fCa ** )
CHICHIW (Ct - )
SODIUM [Na " )
O OUKDNATE (CO., • 1
1,000 O
» «
3/U 1 V\7 1 J/18
A~.
'
i/n i J/23 i
CALENDAR DAY
-------
„
1 * f i f §* if
« -1 ^ " * ? ° •
y *> g « ? 3
i 8 Q. « S » j "
x °- *^ — " &j a" i * «
3- 8 *« 1 8 §
T =1 ^ f o
<* 1 ^ ? ?
= i* f® , »
r f8 js-
3 o | ^
-n r
-J j t
3 ii -
ro " "
no
D r: }C p '••
--- s' i a s g j | ,..
Q r s: f j- Ml,
O o~
3 » 9 1 ° 6 •". ,„
r-+ • — ^ 3 <" - ? ^ '
n g- 5- o
:± £T £ ° w
C S i. I' ?
n> 2 ^ i I
Q. "~* ii "" • • /
^~* s v[ 1 M Isi"
g-g-o- azaJO
"^ n ^ ~~ ' ^ 3 5 °
o> 3tK^ § s ' n
(-h ' 0 S ° ? -'
N 3 "n S • ' to
,3 f 2" ill •
^ J > I ^ = •
Q.
& rt. '
I- ' '
Q5 1 1 '
-« If8 .
Q. s «
3 lh »
4 Jig.
in sss
— " 5- »
(D "M
-s
.-s
14,000
tn
o ta,M
i
s 0
3 > '•"
i. I '•""
^
oo 1 ••""
ca :
. I i.«
3
1 •
3
a MO
£
1
i *°°
•e
»
IOC
o
9L£
fND (UN KI-M ! ! ItGlN RUN Ml -31
^^^^ SYSIU* DWINtD TO tINt-OUT IMSKCTIOM
V\
\ A >*
v V/~. \ /^/
^\ \ \ i/
L V 1
/— IO?M. (XCLUDINO 01 Mint
sc — ^ ^ n^
A^,(0
y~ Of MIITEt
,-. INLEKUn ^"CXl,
K *~i~~^ '~~~$
i — DCWNCOMH OUTLET ILAII 1
^— Will OVIMLOW CXJTLET (LAI) ^
-
C JdO 3W 300 HO MO 140 180 400 4M t*0 460 *
Ttil TIME, hwn
cAiiMOAi a*v
^ - ^ -
# "*
"
/ I I*
^^" "^J
!
-^ ^^" *^^
v
" \ '_
i — TOTAV SUlFUi [SOjl ^J .f^
^— CALCIUM (C*O) y— iUUHE t5OJ / n
3- ^.j-i^^ f d J^ r-t*^~* — •»--.
OUOnO)BOnoMOHOMO*00*?04«9«40««
nsr TIM*. hM,
I J/73 1 J/I4 1 VW 1 Wt 1 V» 1 V30 1 J/ll 1 4/1 1 */5 1 4/1
CAUNMI OAV
* TOTAL Dt»OLVlD VTHIOS * MAGNCSHJM (*»»*")
O CAlCW* (C« ** ) n WO*** (Si. » )
O SULFAH &Qt ' 1 • iULFin tJOj" 1
A CHLOtIM (O - I O OMONAn {CO} * )
•
"
-
*]
A
O *'
0 0"
D-
. •-
^
0 *
i °, i , , , , i i , i . . i n
-1 o
» "
-B - |
-73 „
«S
-:«
1 1?
JlO z
j* "J 2
j! -°
i(j
1: ??
!:! Hi
J"^ ?
S H
.« 1 1 1
D
Z
„ i i
i i
D ! £
^
:E i
0 I I
' I ?
•
5 r- ^ ^
1 y i
, 1*1
0 ~"
I J
5 S n
XM1
" 'H
» "6
9
14,000
11,006
W'°" s
Ii
0
4,flOO 5
Z
is
I, MO c
5
0 7
s
*» B
» ?
i
400 ^
JOO
»0
T»
0
1/3) 1 1/J* 1
-------
L\.£
" —- * ~ t
V 70 O & Q a
2. a o. eg s "
11?Ii S
•5 2 0.1 8 §
r
0?
g
tw
Id
I
O
o
3
s i 1 s
f 5: £ 3-
O
T3
0>
pu
3
10
o.
Q)
C+
O>
O
o
-s
c
00
CO
D3
O •* — •
~ «.OCD
S 1,000
* TOTAL DIUOIVCO KM.IOS * MAONHH** PT* BOj' )
A CMOUDt CO - ) O CAttOUTE (CO, * 1
4/11 I VI*
-------
An overall summary of the initial run data appears in Table 13, which
presents average values for some significant parameters (from
Figures 20, 21 and 22). The listed pH values may be in error.
A summary of the liquid analytical data is presented in Table 14.
Most dissolved species appear to have approached steady state con-
centrations during the period of operation. However, magnesium
ion (Mg ) concentration exhibited a steady
and TCA systems (see Figures 20 and 21),
ion (Mg ) concentration exhibited a steady increase in the venturi
It is of interest to compare the liquid analytical data for the closed-
loop and open-loop runs (see Tables 8 and 14). For the TCA system,
sulfate concentrations for both periods of operation were close to the
"saturation" levels. The chloride concentrations, however, differed
significantly. For the venturi and Hydro-Filter systems, the sulfate
concentrations during the initial closed-loop runs was well above that
measured in the open-loop tests. As expected, in all three systems,
the level of total dissolved solids during the closed-loop runs was
far greater than that obtained during open-loop testing.
Lack of confidence in the pH meters led to a decision to control SO,
removal in the initial tests rather than to directly control pH at the
desired 5. 7 to 5.9 region. Results of open-loop limestone depletion
runs were used to estimate SO removals consistent with the desired
* * *
pH. In general, these SO? removal levels were selected at about
*
There have been some problems with pH meters at the facility.
##
See Figure 11 and Table 6.
318
-------
Table 13
AVERAGE CONDITIONS FOR INITIAL
RELIABILITY VERIFICATION RUNS
Parameters
Operating Time, hrs
Gas Velocity, ft/sec
L/C, gal/mcf
Pressure Drop, in. H.O
Percent Solids Recirc.
Percent SO_ Removal
Stoichiometric Ratio
Limestone Utilization
Inlet Liquor pH
Percent Oxidation
Dissolved Solids, pprn
Hydro -Filter
Run 501. 3A
520
5
53
9
11
65-70
1.25
80%
5,8
30
8000
TCA
Run 501-2A
550
7.8
80
6
15
80-85
1.20
83%
5.8
20-30
7500
Venturi
Run 501. 1A
410
5
40*»
10.5**
15
70.75
1.5
67%
5.8-5.9
15
7000
L/G's of 40 for spray tower and 40 for venturi.
9 inches across venturi and 1. 5 inches across spray tower.
319
-------
Table 14
AVERAGE LIQUOR COMPOSITIONS FOR INITIAL
RELIABILITY VERIFICATION RUNS
Scrubber
System
Liquor
so3= co/
Venturi 200 200
TCA 150 150
Hydro-
Filter 300 150
Species
so/
1500
1800
1800
Concentrations, mg/1 (ppm)
c."
2000
2000
2000
Mg+*
250
300
200
Na+
50
50
50
Cl"
3000
3000
3500
Total
7200
7400
8000
This species increased gradually throughout the time period. The
values shown on this table are the maximum values, at the end of
the plotted periods.
320
-------
10% below that attainable in open-loop operations at a pH above 6. 0
(high pH). Control of SO removal was established by varying the
rate of limestone addition.
7. 1. 1 Venturi Run 501-1A (see Figure 20)
Open-loop testing at high-pH indicated approximate SO, removals to
*
be 42% in the venturi and 57% in the spray tower, which is equiva-
lent to an overall removal, for the combined system, of 75%. Thus,
in order to achieve the desired low-pH operation , a target of 65%
removal is indicated.
From April 14 to April 21, removal was controlled at about 74%
with an average stoichiometric ration of 1.5 (moles CaCO, per mole
SO- absorbed) and an average oxidation of 15%. This implied higher
pH than desired, although the scrubber inlet liquor pH (which is sub-
ject to question) was apparently 5. 8 — 6. 0 during this period.
From April 22 to April 27, SO, removal was controlled at about 70%.
In this period the stoichiometric ratio rose from 1. 3 to 1.8. This
increase in stoichiometry, while maintaining the same SO, removal,
^
This removal was estimated from Eq. 11 and "corrected" for the
change in inlet SO concentrations (see Eq. 12 and Section 5. 1. 2).
&
** The high-pH removal at the selected venturi run conditions was
originally estimated at 85% for the closed-loop operation. The
early part of Run 501-1A is thus at a pH and SO, removal some-
what higher than desired.
321
-------
may indicate "degradation" in the system (e.g. , drop in limestone
reactivity, erosion of spray nozzles). Oxidation and apparent inlet
liquor pH remained at 15% and 5.7 — 6.0 respectively.
Toward the end of the operating period depicted on Figure 20, the
SO_ removal dropped below 70% and was restored to a value slightly
above 70%. On April 28 (low removal) the stoichiometric ratio was
about 1. 5 and then rose to above 2. 0 by April 30. The apparent
scrubber inlet liquor pH remained at 5. 9 — 6. 0 in this period.
7.1.2 TCA Run 501-2A (see Figure 21)
The estimated high-pH value of SO. removal for the TCA system,
,tf
operating at the test conditions of Run 501-2A, is about 95%' (see Eqs.
5 and 12). Therefore, the controlled SO, removal for Run 501-2A
£
was chosen at 85%. After an initial operating period (from March 24
to March 31) in which there were relatively large fluctuations in SO,
removal, a relatively steady period of about 5 days ensued (from April
1 to April 6), where the SO, removal varied from 80 to 85% and the
£»
scrubber inlet liquor pH from about 5. 7 to 5. 9. The stoichiometric
ratio during this period was about 1.15 (which corresponds to a
limestone utilization of about 87%) and the percent oxidation was
about 30%.
*
A removal of about 96% was obtained in the line-out (depletion)
period for Run 501-2A. This confirms the estimate from the open-
loop data.
322
-------
Towards the latter part of the plotted operating periods (from April
12 to April 15 and from April 18 to April 20), while the SO removal
L^
was still controlled between 80 and 85%, there appeared to be an in-
crease in the stoichiometry to an average value of 1. 4 (limestone
utilization of 71%), which again could indicate "degradation" in the
system (e.g. , drop in limestone reactivity, pluggage of spray noz-
zles). The percent oxidation during these periods dropped slightly
to about an average of 20% and the inlet pH ranged between 5. 7 and
6.0.
The system was shut down a number of times due to solids pluggage
of the inlet duct in the vicinity of the humidification section, and re-
sulting increase in total system pressure drop (see Ref. 1).
7.1.3 Hydro-Filter Run 501-3A fe 3B (see Figure 22)
The predicted high-pH value of SO_ removal for the Hydro-Filter
system, operating under the test conditions of Hun 501-3A, is about
80% (see Eqs. 7 and 13). Therefore, the controlled SO7 removal
C»
target for Run 501-3A was 70%. During most of the operating period
for Run 501-3A, the SO^ removal was controlled between 65 and 70%,
the average stoichiometric ratio was about 1. 3 and the average per-
cent oxidation and inlet liquor pH were about 30% and 5. 8, respec-
tively. As mentioned previously, the measured pH's, during this
operational period, were in doubt.
#
During a brief period of high stoichiometric ratio (about 1.5) and
inlet liquor pH (about 6. 1) at about 110 hrs. in Figure 22, the SO2
removal rose to about 80%, This substantiates the predicted high
pH removal.
323
-------
After the system was drained to remove debris (marbles) on March
28 and March 29, another depletion (or line-out) period was conducted
for Run 501-3B. From April 3 to April 13, the SO- removal was held
between 65 and 70%, and the average stoichiometric ratio was about
1.4. The stoichiometry, during this period of the run, appeared to
gradually increase, form an initial average value of about 1. 3 (from
April 3 to April 6) to a final average of about 1. 5. The percent oxi-
dation remained relatively steady during this period (at about 30%)
as well as the inlet liquor pH (at about 5.8). The increase in stoichio-
metry, for the same SO_ removal, could again indicate some "degra-
dation" within the system.
7.2 MATERIAL BALANCES
As mentioned previously (see Section 5. 1), during open-loop testing,
good material balances for calcium and sulfur could only be obtained
with the TCA. The poor material balances for the venturi and Hydro-
Filter systems were attributable to solids build-ups (or depletions)
in the clarifiers, which could not be excluded from the material bal-
*
ance enclosures. During the five-week boiler outage, the venturi
and Hydro-Filter flow configurations were modified to ones similar
to that of the TCA system (see Figures 4, 5 and 6). It was expected,
therefore, that good material balances for calcium and sulfur would
be obtained on all three scrubber systems, based on the measured
*
For the TCA, the main slurry stream circulated between the
hold tank and scrubber, with a "bleed stream" from the main
slurry stream routed to the solids separation area.
324
-------
flow rate and solids compositions of the bleed streams to the solids
separation area and the measured limestone addition rates and SO.
removals.
7.2.1 Venturi Run No. 501-1A
Table 15 gives the results of a material balance for calcium and sul-
fur for venturi Run 501-1A, during a continuous 142 hour operating
period from April 14 to April 19, 1973 (see Figure 20).
The results of the balance showed that the measured sulfur discharged
(4.38 Ib-moles/hr) is 3. 1% less than the measured SCX absorbed
(4.52 Ib-moles/hr) and that the measured calcium added (6. 10 Ib-
moles/hr) is 6.0% less than the measured calcium discharged (6.49
Ib-moles/hr). Both closures are satisfactory, in spite of some dif-
ficulties experienced in measuring the limestone feed rate during the
*
initial reliability tests.
The ionic balances for the solids analyses, from which the calcium
and sulfur discharge rates were calculated, averaged less than +3%
(more cations than anions) for the bleed stream shown in Table 15.
Note that for both sulfur and calcium the measured inlet and outlet
rates do not necessarily balance during each individual computational
period in Table 15. This is due to the unsteady conditions which
prevail (e.g. changing percent solids) and the resultant accumulation
This measurement problem will be alleviated after May 4, when
replacement flowmeter elements (for smaller flow ranges) are in-
stalled in all three limestone feed system magnetic flowmeters.
325
-------
FABLE 15
MATERIAL BALANCES FOR VENTUKI RUN NO. 501-1A
Date
Time
Length
of
Period.
hours
Gi«
Flow
Rate.
acfm
@ 330 OF
Inlet
S02
Cone. ,
ppm
S02
Removal.
%
Bleed Stream
Flow
Rate.
gpm
4/H 0200-0800 6 19.750 3.000 73 12.5
4/14 0800-2400 16 19.850 2.950 72 14.8
4/15 0000-2400 24 19,750 1.050 74 It. 6
4JJ6 0000-0800 8 19.90(1 1.150 76. 16.6
4/16-17 OBOO-OSOO 24 20. 000 3.05.0 73 IS. E
4/17 0800-2400 16 19.900 2.850 75 15.5
4/18 0000-2400 24 ZO.OOO 2.900 73 15. S
4/19 OOOD-2400 24 20,000 2.950 72 15.8
Totals 142
Solids
In
Bleed.
•urt. %
13.7
14.0
14.6
14.9
15, S
It. 0
15.7
16.0
SOX in
Solids
»sSO3,
«.!-•%
29-3
24.9
24.0
27.4
24.4
32.4
29.3
25.3
Ca in
Solids
as CaO.
wt.%
_
*'
Flow
Rate,
gpm
26-0 1.2J
27.6 1.47
27.7 1.53
29.0 1.25
27.3 1.45
29. S 1.41
27.5 1.40
26.6 1.59
Solid s
Content,
wt. 1,
Sulfur Balance
SQt
Absorbed,
Ib-moles
56.7 Z7
59. 3 70
57.4 111
53, 0 40
59,! HI
59. S 71
58.1 106
57.5 106
642
SOX in
Solids
Ditch. ,
Ib-xnolea
Calcium
Ca in
L-S
Feed,
Ib- moles
21 31
56 101
96 150
37 41
9-9 149
&9 98
iia uo
106 156
<>22 &66
Balance
Ca in
Solid I
Ditch. ,
Ib-molet
26
89
158
54
155
U6
159
160
922
CJ
KJ
01
Average rates. Ib-moles/hr:
SO2 absorbed = 642/142 ~- 4.52
SOX discharged = 622/142 = 4. 38
Ca added = 866^142 = 6. 10
C» discharged = 922/142 = 6.49
Average stoichiometric ratio, moles Ca added/mole SOj absorbed:
Based on limes, one added and SOj absorbed - S66/642 = I. 35
Based on solids analysis = 922/622 = 1.48
-------
(or depletion) of the species in. the system. However, over a long
period of time (e. g. ^150 hours) the accumulation term becomes
negligible as compared to the total input or output for the entire
computational period.
The average stoichiometric ratio in Table 15 of 1.48 moles Ca/mole
SO9 absorbed based on solids analysis is probably more accurate
£t
than the value of 1. 35 based on the measured limestone addition rate
and SO, absorption, because of uncertainties in the limestone slurry
C*
feed rate.
7.2.2 TCA Run No. 501-2A
Table 16 gives the results of material balance calculations for TCA
Run 501-2A, covering a period of 150 hrs. uninterrupted operation
from March 30 to April 6, 1973 (see Figure 21).
The results of the balance showed that the sulfur discharged {4, 34
Ib-moles/hr) ie 7% less than the SO. absorbed (4.67 Ib-moles/hr),
while the calcium added (4.45 Ib-moles/hr) is 11% less than that
discharged (4.99 Ib-moles Air). The closures are considered to be
quite acceptable.
jn Table 16, the sulfur input in each individual computational period
is generally greater than ftie output, and the reverse is true for
calcium. The ionic imbalances for the solids analyses during these
periods were mostly positive (more cations than anions) and averaged
abovit +5%. In other words, the reported sulfur content in the bleed
solids might have been too low, or the calcium content too high, or
327
-------
TABLE 16
MATERIAL BALANCES FOR TCA RUN NO. S01-ZA
3/30-31
3/31
4/1
4/Z
4^3
•4/4
4/5-6
Total*
2300-0800
0800-2400
0000-2400
0000-2400
0000-2400
0000-2400
OOOD-0500
Length
of
Period.
hours
9
Ih
24
14
24
24
2<>
150
Gas
Flo*
Rate.
acfm
@3oo°r
20,200
20,000
20,000
20,000
20, 000
20,000
20. 000
Inlet
S02
Cone. ,
ppm
2410
2460
2520
2540
2700
2680
2580
S02
Removal,
1,
84
87
S2
85
82
82
84
Bleed Stream
Flow
Rate.
ipm
14. 1
!+. 5
1 J. 7
13,9
14.9
14. 0
13.7
Solids
in
Bleed,
wt. %
14. 3
14. 1
15.3
15.0
15.5
14.2
14.6
SO, in
Solids
as SOj.
wt. %
27.8
24.3
28.9
27.8
33.2
34.6
31. 5
Ca in
Solids
as CaO,
wt. %
25.2
22. 6
Z4.4
24.0
it,. 6
24.2
2!. 5
Limestone Feed
Flo*'
Rate.
gpm
1. 03
1. 10
1. 17
1.25
1. 07
0. 96
0. «
Solids
Content,
wt. 7,
S8.4
i9.2
S3. 5
S6.3
fco.o
59.6
60. 3
Sulfur Balance "x
S02
Absorbed.
Ib-moles
43
74
107
112
115
114
13*
701
SOX in
Solids
Disch. ,
Ib-moles
35
54
100
96
127
113
126
451
Calcium
Ca in
L-S
Feed.
Ib-moles
3"!
75
103
119
112
100
119
fc(,7
Balance
Ca in
Solids
Disch. .
Ib-moles
45
72
121
118
141.
113
134
749
Average rates. Ib-moles/hr;
SOz absorbed - 701/150 = 4.67
SOX discharged - 651/150 = 4.34
Ca added - f.C7/i^O -i.45
Ca discharged 749/150 = 4.99
Average stoiehiometric ratio, molea Ca added/mole SO2 absorbed;
Based on limestone added and $02 absorbed
Based on solids analysis
667/701 = 0.95
749/651 - 1.15
-------
both. If this factor is taken into account, either or both of the sul-
fur and calcium balances would be better than those reported.
Again, due to uncertainties in limestone addition measurement, the
average stoichiometric ratio of 1. 15 moles Ca/mole SO absorbed
based on the solids analysis is a more reliable number than the value
of 0. 95 based on the measured limestone addition rate and SO-
Cf
absorption.
7.2.3 Hydro-Filter Run No. 501-3A
Table 17 gives the results of material balance calculations for Hydro-
Filter Run No. 501-3A, covering a period of 150 operating hours from
March 16 to March 22, 1973 (see Figure 22).
For sulfur, the average discharge rate (4. 11 Ib-moles/hr) is only 3%
less than the SO,, absorption (input) rate (4.24 Ib-moles/hr). For
calcium, the rate of addition {4.49 Ib-moles/hr) is 13% less than
the discharge rate (5. 16 Ib-moles/hr). The balance is satisfactory,
considering the uncertainties in the limestone slurry addition rate
during the period.
The ionic imbalances for the solids analyses, from which the calcium
and sulfur discharge rates were calculated, averaged less than +2%
(more cations than anions) for the bleed stream shown in Table 17.
Again, the average stoichiometric ratio in Table 17 of 1.26 moles Ca/
mole SO, absorbed based on solids analysis is probably more accurate
than the value of 1. 06 based on the measured limestone addition rate
and SO. absorption.
329
-------
TABLE 17
MATERIAL BALANCES FOR HYDRO'-FILTER RUN NO. 501-3A
Date
3/16
3/17
3/17
3/17
3/16
3/18
3/18
3/19
3/19
3/19
3/20
3/20
J/ZO
J/21
3/2'l
3/Z1
3/22
3/22
3/22
Total a
Time
1600-2400
0000-0800
0800-1600
1600-2400
0000-0900
1100-1600
1600-2400
0000-0800
0800-1600
1600-2400
0000-0800
0800-1600
1600-2400
0000-0800
0800-1600
1600-2400
0000-0800
0800-1600
1600.2400
Length
of
Period,
hours
8
8
8
8
9
5*
8
8
8
8
8
8
8
8
8
8
8
8
8
150
Gas
Flow
Rate,
acfm
@310°F
20,400
20.000
20,300
20.500
20,300
20,300
20,200
20.700
20,350
20.300
20,350
ZO, 600
20,200
20,250
20.150
20, 150
20,150
20,450
20,100
Inlet
S02
Cone. ,
ppm
2950
3050
3100
3150
3320
3050
2800
2900
2800
2750
2800
3020
2950
Z850
2850
2730
2750
2900
2870
S02
Removal,
1,
66
70
72
68
66
70
73
68
77
71
67
68
69
70
64
63
69
67
65
Bleed Stream
Flow
Rate,
gpm
24.7
22.5
26.3
25.4
22. 3
22.6
23.7
21.6
21. 6
25.0
22.1
19.9
19.0
19.2
17.3
16.5
18.2
19.6
19.6
Solids
in
Bleed,
wt. %
9.9
9.2
9.0
9.0
9.8
10.8
10.4
10.2
10.7
10.7
10.9
11.5
11.7
11.3
10.9
10.7
10.9
10.8
10.2
SOX in
Solids
as S03,
wt. %
25.4
21.6
23.0
24.4
27.9
30.0
30.4
31.6
26.0
28.7
28.7
30.3
33.5
26.4
27.5
27. 1
27.5
28. 3
27. 3
Ca in
Solids
as CaO,
wt. %
23. 7
22.8
23. 7
25.8
28.7
26.2
25. 6
24.9
25.6
24,3
23.3
24.5
24.9
22.7
23.3
22.5
22.8
22.5
23.7
Limestone Feed
Flow
Rate,
gpm
0.98
1. 12
1.42
1.60
1.75
1.68*
1. 47
1.93
1.87
1.35
1.52
1.45
1. 19
0.80
0.84
0.97
0.98
0.96
1.00
Solids
Content,
wt. %
35.0
57.6
59.8
50.6
49.9
51.3
53. 1
33.5
55.6
45.4
41.7
43.9
57.0
62.8
60.7
61.3
59.9
60.7
62. 1
Sulfur Balance
so2
Absorbed,
Ib- moles
33. 1
35.6
37.7
36.6
41.7
22.6
34. 4
34.0
36.5
31.0
31.8
35.2
34.2
33.6
30.6
28.9
32.2
33. 1
31.5
f,3f.
SOX in
Solids
Disch. ,
Ib-moles
33. !
23.7
28.8
29.5
36.5
24.5
40.0
37. 1
32.1
41. 1
37. 1
37.3
40.2
30.8
27.8
25.6
29.3
32. 1
29. 1
fiUi
Calcium
Ca in
L-S
Feed,
Ib-moles
16.0
36.8
49.4
43.2
52. 1
40. 5
42.5
30.0
58.2
31.4
31,3
32. 1
38.4
30.1
30. 0
35. Z
34. 1
34. 3
37.0
673
Balan.c
Ca in
.Solids
Disth. ,
Ib-molcs
44. 1
35.7
42.4
44.6
53.6
30.6
48. 1
41.8
45.2
49.7
43.0
43. 1
42.6
37.8
33,7
30.4
34.6
36.4
36.1
774
OJ
to
O
ivera^e rates, Ib-moles/hr
SO2 absorbed =
SOX discharged -
Ca added
Ca discharged =
636/150 =
616/150 =
673/150 =
774/150 =
4.24
4. 11
4.49
5. 16
Average stoichiometric ratio, moles Ca added/mole SOg absorbed :
Based on limestone added and SO^ absorbed = 673/636
Based on solids analysis = 774/M6
1.06
1.26
' The scrubber was shut down on 3/18, 0900-1100 hours (no flue gas addition). However, the limestone addition was continued during this 2-hour period* and the length of limestone
addition was 7 hours instead of 5 hours.
-------
Section 8
REFERENCES
1. H. W. Elder, et al. , "Operability and Reliability of the EPA
Lime/Limestone Scrubbing Test Facility, " presented at Flue
Gas Desulfurization Symposium, New Orleans, Louisiana,
May 14-17, 1973
2. F. T. Princiotta and M. Epstein, "Operating Experience with a
Prototype Lime-Limestone Scrubbing Test Facility, " presented
at the Sixty-Fifth Annual Meeting of the A. I. Ch. E. , New York City,
November 26-30, 1972
3. M. Epstein, et al. , "Test Program for the EPA Alkali Scrubbing
Test Facility at the Shawnee Power Plant, " presented at Second
International Lime/Limestone Wet Scrubbing Symposium, New
Orleans, Louisiana, November 8-12, 1971
4. F. T. Princiotta and N. Kaplan, "Control of Sulfur Oxide Pollution
from Power Plants, " presented at EASCON, Washington, D. C. ,
October 18, 1972
5. R. H. Borgwardt. Limestone Scrubbing at EPA Pilot Plant, Progress
Report No. 3. EPA Report, October 1972
6. J. M. Potts, et al. , "Removal of Sulfur Dioxide from Stack Gases
by Scrubbing with Limestone Slurry: Small Scale Studies at TVA, "
presented at Second International Lime /Limestone Wet Scrubbing
Symposium, New Orleans, Louisiana, November 8-12, 1971
7. A. Saleem, et al. , "Sulphur Dioxide Removal by Limestone Slurry
in a Spray Tower, ibid.
8. A. D. Little, Inc., Evaluation of Problems Related to Scaling in
Limestone Wet Scrubbing, EPA Report, April 1973
9. R. H. Borgwardt, Limestone Scrubbing at EPA Pilot Plant, Progress
Report No. 6, EPA Report, January 1973
10. Radian Corporation, A Theoretical Description of the Limestone-
Injection Wet Scrubbing Process, NAPCA Report, June 9, 1970
jl. A. Saleem, J. Air Pollution Control Assoc., Vol. 22, No. 3,
March 1972
331
-------
12. M. Eptstein, et al., "Mathematical Models for Pressure Drop,
Particulate Removal and SOg Removal in Venturi, TCA and
Hydro-Filter Scrubbers, " presented at Second International
Lime/Lime s tone Wet Scrubbing Symposium, New Orleans,
Louisiana, November 8-12, 1971
332
-------
OPERABILITY AND RELIABILITY
OF EPA LIME/LIMESTONE SCRUBBING
TEST FACILITY
by
H. W. Elder
Steering Committee Member
Tennessee Valley Authority
L. Sybert
Test Program Director
Bechtel Corporation
J. E. Williams
Onsite Technical Representative
Environmental Protection Agency
P. E. Stone
Test Facility Supervisor
Tennessee Valley Authority
333
-------
OPERABILITY AND RELIABILITY OF THE EPA LIME/LIMESTONE SCRUBBING TEST FACILITY
ABSTRACT
Results of research and development programs to study lime/
limestone wet scrubbing have shown that effective sulfur dioxide removal
is feasible. One of the major questions remaining is the long-term reli-
ability of the process when applied under actual conditions on a power
generating unit. The objective of the EPA-Bechtel-TVA test program at
Shawnee is to define a system which operates both effectively and reliably.
The test facility design considerations to permit investigations of opera-
bility are covered in this presentation along with the experience during
about 1 year of operation. Preliminary conclusions which might affect
selection of components for full-scale systems are presented.
Introduction
As part of an extensive research and development program by the
Environmental Protection Agency (EPA), an experimental test facility was
designed and constructed to study and evaluate the feasibility and economics
of closed-loop lime/limestone wet-scrubbing processes and develop mathe-
matical models to allow effective and economic scale-up of practical
operating configurations to full-size scrubber facilities. Interest in
alkali scrubbing has grown because of its basic simplicity and comparatively
low capital cost. The drawbacks of alkali scrubbing are the need to dis-
pose of large quantities of solid wastes and the tendency of accumulation
of solids to plug the system.
The emphasis on development of technology for sulfur dioxide
control during a relatively short period of time resulted in pilot-plant
studies of limestone and lime scrubbing by several organizations. These
programs led to the conclusion that lime/limestone scrubbing is based on
feasible technology. However, the process alternatives are not well de-
fined and the long-term reliability of the configurations tested to date is
still questionable. One of the main objectives of the prototype-scale test
program is to provide a better understanding of the technology and to study
the factors which affect reliability. The program is funded and directed
by EPA; Bechtel Corporation prepared the detailed design of the test fa-
cility, and developed, the test program; and TVA constructed and operates
the test facility.
334
-------
Three parallel scrubbing systems (10-raw each) were installed
at TVA's Shavnee station near Paducah, Kentucky. The system was described
at the limestone symposium in New Orleans during October 1971. The facility
has been operated now for approximately 1 year. The process results are
covered in a paper presented by Dr. M. Epstein (Bechtel Corporation) at
this meeting. The intent of this paper is to review the operability and
reliability of the test equipment during the early phases of the program.
The design considerations in providing methods for evaluation of reliability
are reviewed followed by a description of operating experience.
Design Considerations
The test facility is not sufficiently large to have a significant
effect on stack gas effluent quality at the station where it is installed—
its only function is to provide information on lime/limestone scrubbing
technology. Therefore, it was designed to be a versatile, relatively sophis-
ticated system with emphasis on data generation and collection.
Unit Size
Most of the test facilities used to study sulfur oxide and particu-
late matter removal are small ones capable of treating only up to a few
thousand cftn of flue gas. Because of the differences in both the performance
characteristics of the equipment and scrubbing efficiencies of small-scale
test facilities and commercial-size units, and the associated uncertainty
and difficulty involved in using small-scale data in full-scale unit design,
the three scrubbers at the Shawnee test facility are of prototype size,
capable of treating up to 30,000 acfm of flue gas (320°F, 1^.3 psia).
The size selected will allow extrapolation by a factor of 10 (an
acceptable scale-up) to commercial-size units. The systea is designed so
that the three scrubbers can be operated simultaneously in order to acquire
information on different types in the shortest period. In retrospect,
because of the multitude of problems associated even with routine operation,
maintenance, and evaluation of process data, it might have been more prudent
to concentrate on a single, versatile, system.
Flexibility
The selection of the scrubbers was based on the evaluation of
the performance characteristics of the various designs in terms of their
overall effectiveness as S02 absorbers and particulate collectors under the
operating conditions expected for the process in commercial applications.
The selected types of scrubbers are:
335
-------
1. Variable throat venturi scrubber followed by an after-scrubber
absorption section (spray tower or Pall-ring packed bed).
2. Turbulent contact absorbers (TCA).
3. Hydro-Filter marble-bed absorber.
It is beyond the scope of this paper to discuss the types of scrubbers
considered and their advantages and disadvantages.
The following describes some of the flexibilities that have been
designed into the test facility:
• The three scrubbers, including the flue gas inlet and exhaust
facilities, slurry handling equipment, and clarifiers are in-
stalled in parallel. Only the alkali addition system and the
final solids dewatering and disposal facilities are common.
• Scrubber internals can be changed. For example, the packed-bed
section of the venturi after-scrubber can be replaced by a
four-header spray tower (the four headers can be operated in
any combination). The TCA can be operated as a one-, two-, or
three-bed unit, with a variety of liquor flow piping arrangement.
• Arrangements are being made to provide parallel limestone and
hydrate addition systems.
• The scrubbers can accommodate different types of mist eliminators.
• Each scrubber system has its own oil-fired reheater for inde-
pendent evaluation of exhaust gas characteristics with up to
125°F (above scrubber exhaust temperature) reheat capability.
• A heat exchanger was provided to evaluate the effect of cooling
of the slurry feed on scrubber performance (only one scrubber
at a time).
• Various solid disposal configurations can be evaluated. They
include: clarifier/pond, clarifier/centrifuge/pond, and
clarifier/rotary vacuum filter/pond.
• To evaluate the degree of supersaturation in the circulating
slurry system and the associated possible scaling problems, the
residence time of the circulating slurry solutions in the scrub-
ber effluent hold tanks can be varied between k and 60 minutes
(not necessarily in all three scrubber systems).
336
-------
To determine uncertain model coefficients for gas side con-
trolling mass transfer and obtain design information for
sodium carbonate aqueous scrubbing, a common S02 additive
system was provided. This system can also be used to compen-
sate for rapid and significant drop in S02 concentration in
the flue gas feed.
The ducts, piping, and pumps in each scrubber system are
designed to allow the required variation of the turndown ratio
and the ratio of liquid to gas flow rates.
Variable speed pumps have been used for flow control wherever
possible in order to avoid potential solids buildup and/or
rapid erosion of throttling mechanisms (rubber pinch valves
or metal control valves).
Instrumentation
The control of the test facility operation is carried out from a
central control room. Important process control variables are continuously
recorded or indicated on graphic panelboards by remote conventional
(pressure, temperature level, and water, liquor, and air flows) and special-
ized instrumentation (radiation-type densitometers, magnetic-type flow meters).
The SOg concentration of both the flue gas feed and scrubbed gas streams to
and from the scrubbers can be automatically analyzed by UV spectrophotometer;
also C02, 1^0, 02, and H2 can be continuously monitored by gas chromatographs.
The analytical results of these instruments are also remotely recorded in the
central control room.
The number of local and board-mounted instruments is about 1500.
Some of the more important control loops are listed below:
• The flue gas flow to each scrubber is controlled by a flow-
indicating control loop which senses the flow via a venturi flow
element and sets the position of the damper at the inlet of the
induced-draft fan.
• The S02 injection system to the flue gas feed is on a control
loop set by the Du Pont UV analyzer using a valve in the S02
line as the final control element.
• The reheater outlet temperature is controlled by a temperature-
indicating control loop which senses the exit temperature and
adjusts the fuel oil and combustion air rates to the burners.
• The reheater and the induced fan are interlocked such that the
fan has to be operational in order to fire the reheater.
337
-------
• The internals of the venturi after-scrubber and the TCA and
Hydro-Filter scrubbers are lined with neoprene rubber. To
protect the rubber lining, a temperature sensor is provided
at the bottom of each scrubber which trips the ID fan and
reheater systems and admits water to the emergency sprays in
the scrubber gas inlet duct. (The temperature trip is set at
190°F.)
• The induced draft fans are protected by automatic shutdown
control loops in case of high bearing vibration and temperature.
• Each scrubber is protected by a vacuum relief system which pro-
tects the induced draft fan against high vacuum (-3^ in. water gage).
• The amount of,limestone and water to be added to the limestone
slurry tank is controlled by a cascade of a level indicator con-
troller, to a weight of limestone, to volume of water ratio
controller. This system maintains the desired limestone slurry
concentration at any set level in the mixing tank. The limestone
slurry is then pumped by remotely controlled pumps to feed the
scrubber systems separately.
• The scrubbing slurry to each system is tied to a flow-indicating
control loop. The loop uses a magnetic flow meter as a sensing
device and a variable speed pump as the final element to maintain
the desired flow.
• The levels in the effluent hold tanks are controlled by the amount
of slurry bled to disposal. The levels are sensed by diaphragm-
type elements.
• The clarifier underflow is set by a density-indicating control
loop. The underflow density is sensed by a gamma radiation type
element and the flow rate is set by a signal to the variable speed
pump.
• Alarms for high and low values of different variables are also
provided where necessary.
An electronic digital data acquisition system is utilized to record
automatically and continuously over 150 selected operating data points. In
addition, about 125 channels are provided for alarm condition printouts. This
system is hard wired for data output in engineering units directly on magnetic
tape which is transmitted from the field to Bechtel Corporation in San Francisco
for data evaluation along with manually recorded pertinent information.
An X-ray Fluorescence Spectrometer is used to determine the following
species in liquid samples:
Ca**, S, 1C*", Cl"
338
-------
Solid samples can also be analyzed after dissolution in an appropriate
solvent (e.g. HCl). All functions of the X-ray unit (such as sample
presentation, goniometer angle, slit, collimator) are controlled by a
Nova 1200 minicomputer. Sample identification is fed to the system via
an ASR 35 teletype. All the data are reduced by the computer and printed
out on a line printer in addition to being recorded on a magnetic tape.
Provision for Inspection and Cleaning
An equipment inspection program for the test facility is scheduled
to observe the performance and condition of the equipment after periodic
intervals of operation. The results of the inspection are correlated to
process conditions (temperature, pH, chemical composition, particle size,
distribution of limestone slurry feed) and the time period for which the
equipment was exposed to the above conditions. Trends based on quantitative
data and observations are being analyzed and evaluated, and corrective actions
will be taken to resolve the problems.
Two general categories are covered by the program:
• Evaluation of the corrosive tendencies of the various process
fluids using corrosion coupons installed in preselected lo-
cations in the process equipment.
• Observation and evaluation of the equipment performance with
regard to erosion, solids deposition, scaling, etc.
Corrosion test racks are installed in the flue gas inle*b ducts,
scrubber exhaust ducts downstream of the reheaters, at different locations
in each scrubber, in the effluent hold tanks, recirculating tanks, and clari-
fiers; twenty coupons are mounted on each test rack.1 The selection of the
coupons and the location of the racks were based on experience and anticipated
locations of corrosive conditions.
The equipment performance checks are made periodically at various
points in the system for:
Erosion
Deposition of solids and slurry sludge
Corrosion (this is separate from the evaluation
of the corrosion coupons)
Formation of chemical scale
General conditions
1 Carbon steel (A-283), Cor-Ten B (A-588), l&l S.S., 3l6L S.S., 18-18-2 S.S.,
Alloy 20 Cb-3, lHO S.S. , W»6 S.S. , E-Brite 26-1 S.S. , Hastelloy B, Hastelloy
C-276, Incoloy 800, Incoloy 825, Inconel 625, Monel 400, 70-30 Cu-Ni,
Aluminum 3003, Flakeline 200 (Coated M.S. ), fiberglass-reinforced plastic
(Bondstrand), Transite. In addition, specimens of fiberglass-reinforced
Furan, resin, natural rubber, neoprene and butyl rubber, and five stressed
alloy specimens were installed at selected locations.
339
-------
The location of the inspection points and the frequency of inspection are
based on the severity of the process conditions (temperature, pH, etc.),
hydraulic characteristics of the system, equipment vendor recommendations,
and experience by others. Wherever possible, photographic records of the
condition of the equipment are established and used in the evaluation of
the system.
The observation windows provided on the venturi, TCA, and Hydro-
Filter scrubbers have only limited use in limestone service because the slurry
interferes with visibility.
Access doors on major equipment and Victualic couplings on the
slurry piping system permit quick cleaning and disassembly.
Materials of Construction
The selection of material for construction was based on anticipated
process conditions and design and economic considerations. For example:
Rubber lining of the scrubbers was selected for erosion and corrosion protec-
tion; the stainless steel construction of the venturi scrubber was necessitated
by high throat velocities.
The following list gives an overall picture of the materials of
construction:
• The inlet flue gas ducts are of carbon steel and are insulated.
• The scrubber exhaust gas ducts are of 31&L stainless steel.
• The induced-draft fan internals are of 3l6l> stainless steel.
• The venturi scrubber internals are of 316L stainless steel.
The after scrubber shell is of carbon steel coated with I/1*-inch
neoprene lining.
• The TCA and the Hydro-Filter scrubber internals are of carbon
steel lined with neoprene.
• The chevron demisters are of stainless steel.
• The direct-fired reheater shells are of carbon steel lined
with refractory.
• The tanks are of carbon steel with internals lined with Flake-
line (polyester glass) or with neoprene rubber. The tank
agitators are rubber coated.
340
-------
• The limestone slurry tank is of 3l6*L stainless steel.
• The carbon steel clarifiers are lined with Slakeline and
the rakes are stainless steel.
• All of the slurry piping above 2-1/2-inch diameter is of
carbon steel with internal neoprene rubber lining. The lines
are not heat traced or insulated.
• All of the slurry piping smaller than 2-1/2-inch diameter
is of stainless steel. The lines are not heat traced or
insulated.
• The water lines are of carbon steel and are heat traced and
insulated.
• The slurry pumps have rubber-lined casings and impellers.
• The slurry piping is valved with rubber-lined plug valves or
rubber pinch valves.
• The entire SOg injection system is of 3l6L stainless steel.
• The instrument probes (temperature for example) in the
scrubber and duct internals are of Jl6L stainless steel.
• The solid bowl centrifuge and the rotary drum vacuum filter
are constructed of J16L stainless steel.
• Slurry piping to the pH and conductivity meters is made of FVC.
Operating
Firm conclusions regarding the effect of operating variables on
reliability can be made only after relatively long-term tests. Operation
to date has been mainly restricted to short duration testing of performance
variables. However, observations of trends and short-term effects are perti-
nent. Also, the experience with instrumentation, solids handling, and
materials evaluation should be of interest.
Effect of Operating Variables on Reliability
Although the effects of all variables may not be evident from the
experience to date, the factors are listed below together with applicable
observations from the test program and from supporting studies.
341
-------
Open Versus Closed-Loop Operation. One of the objectives of
development work on lime and limestone scrubbing is to establish conditions
for operation without liquid discharge. The criteria for quality of liquid
effluent streams are not well developed; therefore, zero discharge may not
be required but certainly is a safe level. One of the main factors which
affects reliability of lime/limestone scrubbing technology is the low solu-
bility of the calcium absorbents and reaction products. If the scrubbing
slurry were discarded after a single pass through the scrubber, a minimum
amount of precipitation would occur. When the slurry is recycled, dissolved
solids build up in the liquid phase and more precipitation occurs in the
scrubber. The precipitation of calcium sulfite and calcium sulfate leads
to scaling in the scrubber if provisions are not made for controlling the
precipitation.
The test facility design included pumps with water seals for bear-
ing protection, water quench sprays for gas cooling, water sprays for mist
eliminator wash, and dilute slurry feed. The water required for operation
exceeded the makeup requirement so that the systems operated for about 6
months with partially open liquor loops during limestone scrubbing tests.
This was not considered to be a serious problem because the results of
factorial testing of performance variables are not likely to be significantly
influenced by the water balance. However, not a great deal was learned about
the effect of scaling potential on reliability during this period. Essen-.
tially no scaling occurred.
During February and early March of this year, the systems were
modified to permit closed liquor loop operation. The absorbent feed system
was changed to allow feeding of slurry with up to 60$ solids. The water
seals on the pumps were converted to mechanical seals supplemented with air
purge. The quench spray system was modified to use slurry and the mist elim-
inator wash system was converted to use clarified water. Required revisions
to flow measurement and control devices were made.
Since the modifications were completed in mid-March the systems
have been operated with liquid discharge approaching the quantity of moisture
that would reside with settled sludge. No serious scaling has been observed
during the relatively brief period of operation. One serious encounter with
scale during a special test is discussed in the section below related to
stoichiometry.
Gas Velocity. The effect of gas velocity on entrainment is a
major reliability factor. The need for high solids concentration in the
recirculated slurry aggravates mist recovery because the extent of entrain-
raent influences the amount of solids that impacts on eliminator surfaces
and must be removed by washing.
342
-------
Most of the tests in the Hydro-Filter scrubber have been carried
out at about 5 ft/sec; a few were made at about T-5 ft/sec. Because of the
few comparative data points the correlation between velocity and pressure
drop increase in the mist eliminator is not clear. However, the rate of
solids accumulation was higher at the higher velocity. During the tests at
the lower velocity, manual cleaning of the mist eliminator was required
about,once per month. Only limited wash with fresh water was used during
this period. An improved wash configuration has been installed to reduce
the cleaning requirement.
The TCA scrubber has been operated during most of the test periods
at three different velocity levels, 5.9, 7.8, and 9.8 ft/sec; a few tests
were run at 11 ft/sec. This system has been operated less than the Hydro-
Filter scrubber and no conclusions can be reached regarding effect of gas
rate on carryover.
During a special test to simulate the TVA pilot-plant configura-
tion, slurry entrainment was excessive at 12 ft/sec and the gas rate had to
be reduced to 8 ft/sec for the run. The higher value had been acceptable in
the pilot-plant (l-mw size) work; this indicates that gas-liquor distribution
may not have been as good in the larger scrubber.
Most of the testing in the venturi-spray tower system has been with
slurry fed only to the venturi; in this mode of operation, the spray tower
serves as a large disengaging chamber and carryover to the mist eliminator
was not a problem. When slurry was introduced in the spray tower, the gas
velocities were generally low, 2.5 to 5.0 ft/sec, although a few tests were
run at the maximum velocity of 7.5 ft/sec. Mist eliminator deposits were
minor in all tests.
Until the recent modifications were made, the mist eliminators
were washed intermittently with limited amounts of fresh water on the top side
only. Piping was changed so that wash liquor can be applied to both sides
which should improve the flushing action: a mixture of makeup water and clari-
fied recycle liquor (about half and half) is now being used. There has been
no indication of significant scaling in the demister since the changes were
made. However, use of recycle liquor for wash has caused sulfate scaling in
pilot-plant studies and this could be a potential problem during extended
tests.
Liquid Rate. The primary effect of liquor rate on operability is
to control scaling by providing sufficient volume to accommodate the "make"
of reaction products. Of equal importance is the flushing action in the
scrubber to avoid silting of suspended solids. The latter requirement is
met incidentally because the trend in development of lime/limestone scrubbing
technology has been toward higher and higher liquor rates. At the test fa-
cility, operation has generally been at maximum liquor rate—800 gal/min
(L/G of 27-5*0 in the Hydro-Filter, 1200 gal/min (L/G of Uo-8o) for the TCA,
and ^50 gal/min (L/G of 20-^0) for the spray tower. The spray tower system '
is being modified to increase the liquor rate to 1200 gal/min.
343
-------
Except for minor deposits in the Hydro-Filter bed, silting has not
occurred. Scaling has not been a problem because of the open-loop operation
during most of the program to date. The minor amounts of scale formed during
the brief period of closed-loop operation is encouraging.
Solids Concentration. The suspended solids concentration in the
recirculated slurry is important to the process because sufficient unreacted
absorbent to provide the required dissolved calcium is needed and recycled
calcium salts provide seed sites for precipitation; the effect of fly ash has
not been established. On the other hand, erosion increases as the suspended
solid concentration goes up; therefore, an optimum level should be established
and, although this has not been done, it remains one of the objectives of
the program. The composition of recirculated solids is probably also important,
but the degree of control is questionable. If the sulfate content could be
increased, the total solids requirement might be reduced. Tests of methods
to promote oxidation are planned.
Percent solids recirculated has been varied from 5 to 12; a few
tests were made with 15$ solids. The most serious effect of suspended solids
has been erosion of spray nozzles. The Hydro-Filter nozzles are constructed
from stainless steel and have polyurethane internal liners and swirl vanes.
The liners failed after about 1800 hours of use and have been replaced with
an improved design. Open-type spiral nozzles constructed from stainless steel
were chosen for the spray tower and these are also showing signs of wear; the
same type except with stellite tips have been ordered for replacements. The
large (300 gpm) open-type nozzles used in the TCA scrubber have given good
performance.
Because of frequent changes and short periods of operation at each
solids level, the effect of solids concentration on erosion rates was not
determined. This information will be obtained during the longer term tests.
Stoichiometry. Much of the pilot-plant work on limestone scrubbing
has been done with a limestone feed rate equivalent to 1.2-1.5 times the
amount required to remove all the sulfur dioxide from the inlet gas. During
a special test carried out in the TCA system to simulate the TVA pilot plant,
the intended feed stoichiometry was 1.5. Problems with flow control resulted
in an actual stoichiometry of about 2.7. The scrubber walls and grids were
heavily scaled during the 3-week, closed-loop test. It has since been deter-
mined in pilot-plant studies that the excess limestone was the primary cause
of scaling.
EPA is conducting a test program in a small pilot plant (500 cfm)
to support the Shawnee program. This program, under the direction of R. H.
Borgwardt, has been highly productive. In one series of tests the pilot
plant was configured to simulate the TVA grid-packed tower and several poten-
tial causes of scaling were studied including limestone type, limestone
particle size, inlet gas temperature, presence of fly ash, and stoichiometry.
344
-------
The factors were systematically tested and a-11 except stoichiometry had no
effect on scaling in the ranges studied. With the same stone ground to the
same particle size, scale formed rapidly (less than 40 hours) at a stoichi-
ometry of 2.5; the scale was similar to that found at Shawnee (primarily
CaSOa). The scale growth rate was estimated to be 60 mg/m2. At a stoichi-
ometry of 1.25* no scale formed during Uoo hours of operation even when the
hot gas (285°F) entered the scrubber without a precooling step. The effect
of high stoichiometry was also confirmed in the TVA pilot plant.
From the results of the EPA pilot-plant studies, it appears that
calcium sulfite scaling occurs when the precipitation of calcium sulfite in
the tower exceeds a critical rate that is influenced by pH (rate of dissolu-
tion of limestone). It was estimated that the critical pH for the system
tested is 6.2 - 0.1. When this value is exceeded, scaling is likely to
occur. The effect of pH (stoichiometry) on the dissolved calcium sulfite
in the scrubber discharge is shown below.
Average Liquid-Phase Composition of Scrubber Slurrya
at 1.25 and 2.|? Stoichiometry
(L/G = 37.4, no fly ash)
Stoichiometry
Scrubber effluent 1.25x 2.5*
so3, ng/i 3,?4o 1,580
S0a 774 424
C02 240 292
Ca 771* 670
pH 5.8 6.2
a From "Limestone Scrubbing of S02 at EPA Pilot
Plant," Progress Report 8, March 1973, p. H*
The difference in sulfite level represents the increased precipitation (and
scaling) at the higher stoichiometry.
The test program planned for Shawnee specified a stoichiometry
range of 1.25 to 1.75» Actual values were normally higher than 1.5 as a
result of problems with the feed system. Based on the results of the pilot-
plant tests, the stoichiometry will be lowered in order to reduce the risk
of scale formation. Minimum levels will be established by desired sulfur
dioxide removal efficiency. A potential problem of operation at relatively
low pH is loss of limestone reactivity, probably because of calcium sulfite
precipitation on the particles. This phenomenon has been observed in the
TVA pilot-plant program.
345
-------
Limestone Particle Size, The effect of stoichiometry on pH points
out the importance of any factors which will influence dissolution rate of
the absorbent. Limestone particle size is one of these factors and although
no systematic study has "been carried out in the test program, differences
in sulfur dioxide removal have been attributed to particle size of the stone.
An empirical test to relate particle size, feed stoichiometry, and unreacted
stone in recycled slurry to dissolution rate is needed for application of
results to different situations.
Reheat. The scrubbed gas is saturated when it leaves the scrubber
and entrainment plus any condensation which might result from cooling would
be detrimental to downstream equipment. Therefore, some level of reheat
is required. Reheat for each system is provided by combustion of oil in
direct-fired burners installed in a combustion chamber in the exhaust duct.
Operation of these units has been unacceptable. The cooling effect of the
cool, oxygen-deficient, saturated scrubber exhaust plus poor atomization of
oil as a result of operation of a single nozzle over a wide range of flow
rates have caused incomplete combustion. In addition to interference of
soot with particulate measurements, the poor combustion resulted in deposits
of unburned oil and soot in the ductwork above the reheater. These accumula-
tions of combustibles ignited on two occasions and could have caused injury
to personnel or damage to equipment. Fortunately the fires were extinguished
without serious effects. After the second incident, the local plant manage-
ment advised that the test facility should not operate until the hazard was
eliminated.
Both parts of the problem were investigated simultaneously during
a period while the test facility was not operating because of a scheduled
boiler outage. Stainless steel sleeves (4o-in diameter by U-ft high) were
installed to isolate the combustion mixture from the scrubber exhaust until
combustion of the oil could be essentially completedj the effective combus-
tion volume is about 50 ft3. Mechanical atomizing nozzles were installed
instead of the turbulent mixing type provided initially with the combustion
system. The types now in use are designed for a narrow range of oil flow
rates and will have to be changed when the reheat requirement changes. How-
ever, nozzle replacement is a simple job.
The changes have been effective. Essentially no soot is visible
in the stack gas and particulate samples have shown no evidence of carbon
from the reheaters. Plans for installation of an external combustion system
have been deferred.
346
-------
Instrumentation
S0g Analyzers. Experience with the six Du Pont Model ^00 UV S02
analyzers at the test facility has generally "been good. Most problems have
been associated with the sample handling system rather than with the instru-
ments. Initially the sample handling system was particularly vulnerable to
condensation, dust, oil, corrosion, or combinations of these factors. This
led primarily to line leakage, line plugging, plugging of the filters, or
coating of the lens. All of these effects resulted in erroneous S02 read-
ings. It was also found during early attempts to calibrate using standard
reference gas, that the certified values are given for a specified tempera-
ture. Use of the reference gas at other than the specified temperatures can
also lead to erroneous results.
In order to eliminate the problem areas, the sample handling system
was modified in November 19T2 as follows:
1. All heat sinks and sharp bends in the sample lines were eliminated.
A new 5/8-inch diameter Dekeron sample line was installed to replace
the original 1/Wnch stainless steel line. Heat tracing was in-
stalled the full length of the sample line.
2. Stainless steel shields furnished by Du Pont were installed around
the probe filters. The original ceramic probe filters were also
replaced by type Jl6 stainless steel probe filters recently developed
by Du Pont.
3. An automatic zero and air blow-back system was installed on the S02
analyzers in the inlet gas duct as had originally "been installed on
those in the outlet gas ducts.
k. All stainless steel lines and fittings were replaced with Dekeron
or teflon wherever possible.
5. Calibration methods were changed to use a stainless steel wire mesh
reference filter rather than use of standard reference gas bottles.
One additional problem associated with the instrument rather than
the sample handling system was also encountered. The interference filter in
the optic section of all six analyzers was found to deteriorate with time.
All of these filters, which screen out all except the desired light wave
lengths, were replaced. The failure and subsequent deterioration was attri-
buted by Du Pont to have been caused by exposure to freezing conditions prior
to installation. The freezing resulted in minute cracks which then deterio-
rated with time as theorized by Du Pont.
Following the modifications made to the sample handling system
and the replacement of the interference filters,'operation for the past 5 to 6
months has been essentially trouble-free with only a minimum of preventive
maintenance time.
347
-------
Magnetic Flow Meters. Operating experience with the Poxboro mag-
netic flow meters at the test facility has generally "been good. The main
problem has been in obtaining accurate flow measurements at very low flow
rates with meters designed to measure flow over too wide a range. This is
a problem inherent in the nature of a test facility where testing over a
wide range of variables is desired. To assure accuracy, Foxboro recommends
a minimum linear velocity of 5 ft/sec through the flow element. Periodic
cleaning of the electrodes and calibration checks are also required to
compensate for drifting of the meters. Routine cleaning and maintenance
checks are now made on a monthly basis.
Control Valves. Operating experience with control valves at the
test facility is similar to experience with flow meters. Operation has been
generally good when control valves are used in reasonable design flow ranges.
However, when excessive throttling is required to obtain very low flow rates,
severe erosion in a relatively short period of time results from the increased
velocity through the throttled valve. This has been observed with both
stainless steel plug valves and rubber pinch valves used as control valves.
pH Meters. Operating experience with pH meters at the test fa-
cility has thus far been limited to in-line flow-type meters. No significant
scaling of electrodes has been noted to date with limestone. However, fre-
quent calibration checks using a buffer solution are required to maintain the
desired accuracy. The current frequency in use at the test facility is twice
per week on a routine basis or more frequent if required. Because of the
desirability to be able to control pH to within t 0.1 pH unit, future test
program plans include evaluation of another type of pH meter*
Density Meters. Operating experience with density meters at the
test facility includes both the Ohmart radiation-type meter and the bubble-
type meter. Both systems require further study and modification to achieve
adequate reliability in their respective control service.
Data Acquisition System. Operating experience with the EMC data
acquisition system at the test facility has been improved during the past 6
months. Early in the test program much difficulty was experienced in recover-
ing the data from the tapes recorded automatically onsite. Extraneous
characters were being recorded on the tapes and were interfering with the
data recovery. Changes were made to reduce the industrial noise of the system
and a special computer program was also written to help recover the data from
the tapes. The system is also sensitive to dust, and particularly to coal
dust. Since the system is neither enclosed nor located within a pressurized
area, periodic cleaning of the recorder on a weekly basis was initiated.
Subsequent to these changes, operation of the data acquisition system has
been very good.
348
-------
Based on the operating experience at the test facility, two areas
appear to be critical* Some of the field effect transistor (FET) circuits
appear to be extremely sensitive and should probably be redesigned if time
permits. Compatibility of the tape recorder with the equipment used to
recover the data from the tapes also appears to be very sensitive to dust,
noise, alignment, etc. However, neither problem area appears to be insur-
mountable with proper care or design of the equipment.
X-Ray Computer System. Operating experience with the Siemens
X-ray computer system at the test facility has been similar to that on the
data acquisition system. Some mistakes made earlier in the program on the
data acquisition system were avoided on the X-ray computer system. Both the
X-ray unit and the computer were enclosed in a pressurized air conditioned
room. This has minimized the problems of recovering the data from the tapes.
Also when problems do occur in removing the data from the tapes, the special
computer program prepared for use with the data acquisition system is
available.
Operation of the X-ray unit has been very satisfactory. Most prob-
lems have been associated with either the interface between the X-ray and the
computer or in the peripheral hardware equipment. Once these problems were
corrected and a weekly cleaning schedule was established, operation of all
associated equipment has improved significantly.
Waste Solids Handling
One of the most important considerations in use of lime/limestone
scrubbing technology is disposal of waste solids. A section of this meeting
is scheduled for discussion of the work related to handling and disposal of
sludge. The test facility is equipped to study alternate methods for sepa-
ration of solid and liquid phases in the scrubber discharge stream.
Pond. A three-section settling pond was constructed in an area
that had previously been an ash storage pond. The dikes are made from fly
ash and are covered with local clay. A small starter pond was used during
the early open-loop tests and has been filled and retired. The pond arrange-
ment which will be used for the remainder of the closed-loop program is
discharge of sludge into a large settling area and return of the supernate
through a smaller "polishing" pond. The slurry to the pond can come directly
from the scrubber circuit, from the thickener underflow, or from the filter
and centrifuge as reslurried cake. The ponds are being equipped with instru-
mentation for evaluation of seepage. During the early stage of use before a
significant area of the pond is covered with settled solids, seepage has been
excessive. Small-scale permeability tests are being made in an effort to
determine the ability of settled sludge to seal the clay lining. An im-
pervious membrane may be required to reduce seepage to an acceptable level.
349
-------
Clarifier. Each scrubbing system is equipped with a separate
clarifier; the venturi and Hydro-Filter systems have 20-ft diameter units and
the TCA system, because of a higher recirculation rate, has a JO-ft diameter
thickener. The performance of the larger unit has been satisfactory but
the solids carryover in the overflow has been a problem with the smaller
units. The settling characteristics of the reaction products (particularly
calcium sulfite) are poor and insufficient time is available during the
compression or hindered settling period to produce a clear overflow. Vari-
ation in the feed rate may contribute to poor performance. It might be
necessary to operate with a turbid overflow. This probably would not be
serious if the solids in the overflow concentration could be controlled.
The concentration of solids in the underflow approaches the ex-
pected final settled density of sludge (approximately
The program for improved clarification and thickening includes:
• Maintain the feed flow to the clarifiers as steady as possible.
• Optimize the oxidation of calcium sulfite to calcium sulfate.
• Investigate the effect of the particle-size distribution of the
ground limestone on both the SQ^ absorption efficiency and
settling characteristics and adjust limestone grinding accordingly.
• Maintain the percent solids in the clarifier underflow under
density control and let the sludge level vary in the tank. The
set point of the density controller could then be adjusted
(i.e., the underflow rate increased) to prevent sludge overflow.
• Operate the clarifier in series with the vacuum filter or centri-
fuge to attain minimum solids concentration in the recycle liquor
and maximum final solids concentration in the solids disposal
stream.
• Use of flocculents.
Filter. Initial tests of a rotary vacuum filter to separate the
solid and liquid phases in the clarifier underflow were not successful. The
cake is thixotropic and, although it appeared dry and firm under vacuum,
internal water was released and the cake became fluid when the vacuum was
released. The wet, sticky cake would not separate from the filter cloth.
Further dewatering was restricted by formation of cracks in the cake which
allowed bypassing of air.
350
-------
The vendor has recommended installation of compression rolls and
air blast discharge to reduce the final moisture to approximately ^0$.
These changes are being made*
Centrifuge, Short-term, exploratory tests of the centrifuge were
carried out in late April. Both clarifier underflow and a sidestream from
the scrubber recirculation loop were tested. The results of these cursory
tests are shown below.
Summary of Centrifuge Tests
Machine Centrifuge Peed Wt $ Wt # Wt %
Test speed, feed rates, solids moisture solids in
series rpm source gpm in feed in cake centrate
I 2000 HP clarifier 11-22 15-22a ^3-^7 0.2-0.6
bottoms
II 2000 HP clarifier 10-22
bottoms
0.5-0.5
III 2000 HP clarifier 10-22 19-29* 39-^2 0.1-0.5
bottoms
IV 2500 HP clarifier 9-33 19-2?a 36-^0 0.1-1.1
bottoms
V 2500 Scrubber bleed 11-35 10-1^ 37-^1 0.1-0.6
(clarifier
bypassed)
a Increase the values by about 3 for pump seal water correction.
* Test Series II was a replicate of Test Series I.
It appears that the centrifuge is effective in reducing the
solids content well below the level attained by settling and that the cen-
trate clarity is satisfactory. The centrifuge will be incorporated into
the loop of one of the scrubbing systems for long-term tests.
351
-------
Materials Evaluation
Selection of construction materials for components of scrubbing
systems is an important economic consideration. One objective of the test
program is to evaluate corrosion/erosion rates for alternate materials.
The evaluation involves comparison of several different materials in
installed components as well as exposure of test coupons at appropriate
locations.
Equipment Inspection. A thorough inspection of al.1 system com-
ponents was conducted during the extended boiler outage. The systems had
been operated during limestone scrubbing tests that totaled about 1800 hours
for each train.
The mild steel gas ducts between the boiler and scrubber inlet
transitions had localized deposits of loose fly ash. The surfaces had a
thin coating of iron oxide scale except at uninsulated flanged connections
where moderate pitting had occurred; these flanges are now insulated.
The rubber lining in the scrubbers was in excellent condition;
no erosion or deterioration was noted. The rubber linings in pumps, piping,
and process water tanks were also in excellent condition. Slight wear was
noted on some of the rubber-coated agitator blades.
The only damage noted in the Flakeline lining in effluent hold
tanks and clarifier tanks were several hairline cracks which did not appear
to penetrate the entire thickness. The cracks were most prevalent at the
junctions between the baffles and tank walls.
The most severe corrosion was found on type J>l6 stainless steel
surfaces particularly on mist eliminator blades in the TCA system. The
general attack was in the form of pitting and some pits were as large as
1/16-inch diameter and 50 to 55 mils deep.
The only significant erosion of equipment noted was on pump sleeves
and at intersections of wires of support grids in the TCA scrubber. Weight
loss of mobile bed packing material has also been detected. The polypropylene
(and polyethylene) spheres in the TCA scrubber have worn noticeably. Some
are so thin that they have lost mechanical strength and the walls have col-
lapsed; a random sample of these showed about 60$ weight loss. The bulk of
the spheres were still intact but had an average weight loss of 20$. The
glass marbles in the Hydro-Filter have lost about 6% of their initial weight.
352
-------
Test Coupons* Test coupons of several different materials of con-
struction were exposed for 70 days or more in various slurry and gas
environments. Stressed and welded metallic materials were also tested.
Corrosion of Hastelloy C-276 was negligible to 5 mils per year;
this alloy showed no evidence of localized attack in any test location.
Next in resistance were alloys Inconel 625, Incoloy 825, Carpenter 20Cb-3,
and Type Jl^L stainless steel with corrosion rates for each material rang-
ing from negligible to 5* 7, I1*, and 15 mils per year, respectively. These
alloys had very few minute pits and/or crevice corrosion. One specimen of
Type 3l6 stainless steel was grooved and the weld of another specimen was
attacked. Type J516L is the fifth alloy in resistance and the least expensive
of this group.
Three nonferrous alloys, Cupro-Nickel 70-30, Monel UOO, and
Hastelloy B each had minimum rates of < 1 mil and maximum rates of ^9, 57,
and 100 mils per year, respectively, with 1 or 2 specimens pitted. In three
tests of Monel and in one test of Cupro-Nickel 70-30, the welds were inferior
to the parent metal.
A group of five alloys, Type kk6 stainless steel, E-Brite 26-1,
Incoloy 800, USS lS-lB-2, and Type 301* stainless steel, had rates that ranged
from negligible to a "greater than" (>) value which indicates that the speci-
men was completely destroyed at one or more test locations. The values for
failures ranged from > l4o mils per year for Type kk6 to > 200 for both
USS 18-18-2 and Type 30UL stainless steels. These five alloys were highly
susceptible to localized corrosion.
Another group of alloys, Type 1*10 stainless steel, aluminum 3003,
mild steel A-283, and Cor-Ten B, had minimum rates of < 1 mil per year and
maximum rate of >250 for Type UlO to > lUoo for mild steel and Cor-Ten B.
Pitting and crevice corrosion occurred on the four alloys.
In general the stressed specimens (five alloys only) were not
corroded at rates higher than their counterpart disk-type specimens.
Specimens of Bondstrand ^000, Flakeline 200, and Transite were
tested at 21 locations. Bondstrand showed good resistance in 12 tests and
poor in 9 tests. The evaluations for Flakeline and Transite were: good,
2 and l^j fair, l4 and 2; and poor, 5 and 5, respectively. Only 6 specimens
each of Qua-Corr plastic and of butyl, natural, and neoprene rubbers were
tested. The results were 5 good and 1 poor for Qua-Corr and 6 good for each
of the rubbers.
353
-------
Recommendations Based on Experience to Date
Some of the preliminary conclusions discussed in this section may
change as more information is obtained during long-term tests. However,
several large-scale installations are in the design stage and some of the
trends which have been noted during operation of the test facility are worth
considering.
Procea s jgentrol
During the open-loop testing, runs at a given set of conditions
were normally short (2 days or less). Process control during these tests was
accomplished by setting the limestone feed rate at the value required for the
desired stoichiometry based on S02 concentration in the inlet gas; slurry pH
and removal efficiency were, therefore, dependent variables. This approach
met the objectives of the test program.
When closed-loop testing was begun in March, the program emphasis
was placed on long-term reliability tests. Because of the effect of pH on
scale formation in the tests discussed under stoichiometry, it was decided to
operate with pH control; the limestone feed rate is manually adjusted to main-
tain a nearly constant value of pH. System response has been good in the
relatively low pH range (5,8-6.0) studied. It appears that if reliable pH
meter operation can be established, automatic control of limestone feed rate
to maintain pH at values below 6.0 would be feasible.
Gas Cooling
If a heat sensitive lining is used in the scrubber for erosion/
corrosion protection, gas cooling ahead of the scrubber is required. Efforts
to control solids deposits at the point of liquor or slurry addition have been
only marginally effective when sprays are used. The venturi scrubber configu-
ration provides a clean separation between the wet and dry areas and solids
have not accumulated at the transition point.
When gas cooling by humidification ahead of the absorber is required,
a venturi scrubber may be the most effective device for addition of the liquor.
Mist Eliminator
Most of the testing to date has been with excessive fresh water
addition for washing the mist eliminators. This operation is not consistent
with the long-term reliability evaluation. The systems recently have been
354
-------
modified to permit use of recycled water for wash of the mist eliminators.
The liquor has varied from a ratio 1 part fresh water to 3 parts clarified
liquor to a half and half mixture. In the Hydro-Filter and venturi scrubbers
the full underside is washed intermittently at a rate of about 1 gpro/ft2 on
a cycle that has averaged 1 minute on and 3 minutes off. In the TCA, the
wash is added on a valve tray beneath the mist eliminator and entrained liquor
from the tray flushes the underside of the eliminator.
No significant scaling has been detected although some solids de-
posits have occurred. It appears that precipitation on mist eliminator
surfaces can be controlled by dilution of clarified liquor with available
makeup water.
Reheaters
Use of direct-fired, in-line, oil burners for reheat can lead to
incomplete combustion and accumulation of oil-saturated soot. The problem
apparently has been solved at the test facility but turndown is complicated
by the need for nozzle substitution.
Use of an external combustion system and admixture of completed
burned combustion products with the scrubber exhaust gas would be a more
satisfactory system for direct-fired reheat.
Clarifiers
The small (20-ft diameter) Clarifiers appear to be underdesigned
to provide both a dense underflow (kQ% solids) and a clear overflow. The
settling rates and settled density of the sludge are poorer than the design
basis and preliminary results indicated the volume of the Clarifiers should
be increased; the additional volume required has not been established.
The Clarifiers appear to be underdesigned to provide both a dense
underflow (kofi solids) and a clear overflow. The settling rates and settled
density of the sludge are poorer than the design basis and preliminary re-
sults indicated the volume of the Clarifiers should be doubled.
SOo Analyzer
The continuous, on-line S0a analyzers (Du Pont Model 400 UV) are
accurate and reliable provided that high-quality gas samples are delivered
to the units. A nonplugging probe is needed and heated, well-insulated sample
lines (Dekeron) are required to prevent condensation and plugging. Precise
calibration techniques are necessary.
Materials of Construction
Rubber lining of scrubbers, tanks, pumps, and piping has been
entirely satisfactory. Use of less expensive polyester-glass lining may be
acceptable, but further exposure is required before a firm conclusion can be
reached* Erosion of mobile-bed spheres has been excessive; improved materials
355
-------
are being tested. Type Jl^L stainless steel has been badly pitted in
partially wetted areas, particularly the surfaces of mist eliminators.
Nonmetallic materials will be tested in these areas. Test coupons of
highly resistant alloys (HastelUoy, Carpenter 20) have shown no signs of
attack, but these are relatively expensive.
356
-------
SCRUBBING EXPERIMENTS
AT THE MOHAVE GENERATING STATION
by
Alexander Weir, Jr., Principal Scientist for Air Quality
Lawrence T. Papay, Director of Research and Development
Southern California Edison Company
2244 Walnut Grove Avenue
Rosemead, California
357
-------
PREFACE
The pilot testing program was set up to make comparative tests
of different scrubber types and reagents to determine SO^ and
particulate removal capabilities. Reliability testing of mech-
anical equipment and hardware was not a part of the program.
The performance results in the succeeding discussions therefore
do not imply that reliable operation of these levels can be
attained. The follow-on module program described will hopefully
give a better measure of the total performance. It should be
pointed out though that with the press of time a large step is
being taken in proceeding from the pilot to the full size module,
viz., good engineering practice would dictate an intermediate
prototype first.
358
-------
This paper presents work performed at the Mohave Generating
Station during 1971 and 1972. Eight pilot plant scrubbers and
four different reagents - soda ash, limestone, lime and ammonia -
were studied. We intend to present some of the highlights of
this work as well as the conclusions which allowed us to select
two of the scrubber types for construction of two 450,000 scfm
scrubbers. The work reported here was performed by a number of
contractors and paid for by an even larger number of organizations
Figure 1 (attached) presents some of the organizations which
participated in this work.
Figure 2 presents a further description of the scrubbers.
The scrubbers were all essentially cylindrical in shape and,
as may be seen, most of the scrubbers tested were between two
and three feet in diameter, with the exceptions being the six
foot diameter spray drier and the 19-inch diameter TCA. The
scrubber lengths ranged from eight feet to 27 feet. However,
although the scrubbers were somewhat comparable in physical size,
their maximum gas handling capacity varied considerably ranging
from 1100 scfm in the WPS to 3500 scfm in the SCE scrubber. It
should be noted that the flow rates listed were not necessarily
the optimum flow rate for SC^ or particulate removal but the
maximum (except for the SCE scrubber) under which the scrubber
would operate. The capacity of a scrubber, rated in scfm per
megawatt varies from station to station, depending on the type
of fuel, the excess air used in combustion, the amount of leak-
age through air preheaters and other factors. At Mohave we have
determined that 2800 scfm is approximately equal to one megawatt
downstream of the electrostatic precipitators resulting in tested
scrubber capacities from 0.4 to \\ megawatts. Actually, the
capacity of the scrubbers is a function of the actual linear
velocity through the scrubbers. While the temperature of the
gas entering the scrubber varies considerably, the exit gas
temperature remains constant, ranging from 120° to 130° F regard-
less of the outside air temperature. The gas at the exit is
saturated with water. By using the exit gas conditions, the
cross-sectional area of the scrubber, and the maximum scfm
capacity of the scrubber we are able to calculate the "super-
ficial" (i.e. disregarding scrubber packing) linear velocity
through the scrubber and thus determine a relative "size factor"
for extrapolation to larger scrubbers. In Figure 2 the scrubbers
are listed in order of increasing linear velocity.
Finally, the last column in Figure 2 indicates the reagents
tested with the scrubbers. Most of the experiments with AIS
scrubber were performed with soda ash but a few experiments
359
-------
Figure 1
MOHAVE SCRUBBER PROGRAM
SCRUBBER
MANUFACTURED BY TESTING PERFORMED BY WORK SPONSORED BY
U)
CT>
O
WPS HYDRO PRECIPITROL
CVS CHEMICO VENTURI
US PEABODY-LURGI IMPINGE-
MENT SCRUBBER
DESEVERSKY ELECTRO- SCE-TRUESDAIL LABS
NATOM CORP.
CHEMICAL CONSTRUC-
TION COMPANY
PEABODY ENGINEERING
CO.
PPA POLYPROPYLENE ABSORBER HEIL WITH FLUOR
PACKING
TCA TURBULENT CONTACT
ABSORBER
SCE EDISON SCRUBBER
AIS ATOMICS INTERNATIONAL
AQUEOUS CARBONATE
SCRUBBER
RHS ROTATING HORIZONTAL
SCRUBBER
UNIVERSAL OIL
PRODUCTS
STEARNS-ROGER, INC.
BOWEN ENGINEERING
CO.
HAZEN RESEARCH, INC.
SCE-BECHTEL CORP.
SCE-BECHTEL CORP.
SCE-BECHTEL CORP.
SCE-BECHTEL
SCE-TRUESDAIL LABS
SCE-ATOMICS INT'L.
SCE, U.S. LIME.
HAZEN RESEARCH AND
NLA
SCE
NAVAJO AND MOHAVE
PARTICIPANTS
NAVAJO AND MOHAVE
PARTICIPANTS
NAVAJO AND MOHAVE
PARTICIPANTS
NAVAJO AND MOHAVE
PARTICIPANTS
SCE
16 UTILITIES WHO ARE
MEMBERS OF WEST
ASSOCIATES
NATIONAL LIME ASS'N
AND SCE
-------
Figure 2
MOHAVE PILOT PLANT SCRUBBER TEST CONDITIONS
U>
SCRUBBER
AIS CONVENTIONAL
SPRAYDRIER
FOLLOWED BY CYCLONE
SEPARATOR
RHS HORIZONTAL LIME KILN
CONTAINING BALLS OR
CHAINS
CVS SINGLE STAGE
VENTURI WITH FIXED
ANNULAR THROAT
FOLLOWED BY CENTRI-
FUGAL SEPARATING
CHAMBER
WPS VERTICAL ABSORBER
FOLLOWED BY WET ELEC-
TROSTATIC PRECIPITATOR
LIS VARIABLE THROAT
VENTURI FOLLOWED BY
THREE-STAGE IMPINGE-
MENT TRAY VERTICAL
ABSORBER
PPA VERTICAL ABSORBER
PACKED WITH 7 1/2 FT.
OF PLASTIC "EGG-CRATE"
PACKING
TCA VERTICAL TURBULENT
CONTACT ABSORBER
CONTAINING THREE
STAGES OF "PING PONG
BALLS"
SCE FOUR-STAGE
HORIZONTAL CROSS-
FLOW SCUBBER
HEIGHT
OR
(LENGTH)
12
10
8
15 1/2
14
12 1/2
27
DIAMETER
(INCHES)
I.D.
72
36
26 3/4
27 1/4
30
24
19
MAX. GAS
FLOW RATE
TESTED
(SCFM 60°F)
1375
1300
1037
1100
1332
1850
1230
LINEAR
VELOCITY
AT EXIT
0.9
3.4
5.0
5.1
5.1
11.0
11.7
15
22
3500
24.9
TEST
REAGENTS
SODA ASH
(LIME)
LIME
SODA ASH
LIME
LIMESTONE
AMMONIA
SODA ASH
LIME
LIMESTONE
SODA ASH
LIME
LIMESTONE
SODA ASH
LIME
LIMESTONE
, LIME v
(LIMESTONE)
-------
were performed with lime. On the other hand, most of the
experiments with the SCE scrubber were performed with lime,
with only a few experiments being performed with limestone.
Ammonia, as gaseous ammonia and as ammonium hydroxide, was used
only with the WPS.
Some of the variables studied in addition to the scrubber and
reagent type included reagent composition (including pH and
percent solids) and, in the case of lime and limestone slurries,
the effect of soluble sodium salts such as occur in cooling tower
blowdown water. The gas flow rate was an important parameter.
As previously mentioned, there appeared to be an optimum gas flow
rate for each of the scrubbers tested. Some of the scrubbers
were operable over wider ranges than the other due to their con-
struction. For example, with the TCA scrubber at too low a gas
velocity the ping pong balls stayed immobile on the lower support
screen while at a higher velocity they congregate at the top of
the stage, i.e., the bottom of the next higher stage. The liquid
flow rate and the L/G ratio (GPM/1000 scfm) were also important
variables.
The number of stages was varied with the TCA scrubber and the SCE
scrubber and we were quite pleased to find that our theoretical
predictions on the effect of staging were confirmed experimentally,
The gas pressure drop was measured under a variety of conditions
and this, of course, determined the fan requirements for larger
size scrubbers. We operated with both FD and ID fans and per-
haps we can summarize our findings in this area by indicating
that both of our 450,000 scfm test modules will have FD fans even
though the power required because of the higher gas temperature
will be greater.
We also studied sludge separation with centrifuges, rotary filters
and thickeners, the degree of oxidation from sulfite to sulfate
and the properties of the slurry and the sludge itself.
It is not possible in the time allotted to present much more than
a cursory overview of the effect of the different variables or
to itemize all of the operating problems encountered with the
various scrubbers. Primarily, the latter resulted from plugging
in scrubbers, demisters, reheaters, centrifuges, and piping.
Effect of Inlet SC>2 Concentration and L/G
The Mohave Generating Station is located some 90 miles south of
Las Vegas and burns low sulfur coal transported by a 285-mile
slurry pipeline from the Black Mesa Mine. The average sulfur
362
-------
content of the coal burned over the last two years was about 0.3870
sulfur which results in less than 200 ppm of S02 in the exhaust
gas.
After we initiated our experiments, it became evident that the
S02 inlet concentration varied with the station operating condi-
tions. In order to have consistent data we had to maintain the
inlet S0« concentration at a constant value. Accordingly, we
arbitrarily spiked the inlet gas to 400 ppm of SOo. This corres-
ponds to about 0.83% sulfur coal, the highest sulfur content coal
predicted for Mohave. In the two 450,000 scfm scrubbers that are
currently under construction we will have to burn about five tons
per day of sulfur per each scrubber in order to increase the con-
centration to 400 ppm for test purposes.
Most of the data presented in the remainder of this paper is
based on an inlet gas concentration of 400 ppm of S02-
We did perform some experiments, however, with five scrubbers on
the effects of inlet S02 concentration. The data from one of
these scrubbers is presented in Figure 3.
Prior to performing these experiments, our calculations, based
on a rather simplistic theory, indicated that we should obtain
straight, parallel lines at lower SOo concentrations, the degree
to which they can be considered parallel is debatable. The shape
of the curve near the origin varies with different scrubbers,
being convex in some cases rather than concave as shown here.
Figure 3 does show, however, the effect of L/G on S02 exit con-
centration, i.e. increasing L/G increases the S02 removal.
However, the CVS scrubber exhibited a minimum in the curve.
That is, after a certain L/G was reached an increase in L/G
resulted in an increase in S02 exit concentrations.
Often considerations of SCH removal specify scrubber performance
in terms of percent sulfur removal without regard to the inlet
S02 concentration. When the data in Figure 3 is presented in
terms of percent SOo removed, as is shown in Figure 4, it is
obvious that the inlet S02 concentration is an important factor
in characterizing scrubber performance. While this data is for
one scrubber at one gas flow rate, we have observed similar
trends at other flow rates with four other scrubbers. Apparently
this is due to the solubility of the reagent, the gas flow rate
and the gas-liquid contacting mechanism.
In summary, Figure 4 indicates that for a given L/G ratio and gas
flow rate, a decrease in the inlet S02 concentration results in
an increase in the S02 removal efficiency, until some maximum
363
-------
Figure 3
EFFECT OF L/G ON OUTLET SO2
CONCENTRATION AT DIFFERENT INLET
CONCENTRATIONS SCE SCRUBBER
-2500 SCFM
1OOO
1000 PPM
INLET
1600 PPM
/ INLET
600 PPM
INLET
200 PPM
INLET
400 PPM
INLET
L/G
364
GPM
1OOO SCFM
-------
FIGURE 4
EFFECT OF INLET SC>2 CONCENTRATION
ON PERCENT SO2 REMOVAL
100
SCE SCRUBBER
2500 SCFM
200
400
600
1000
1600
INLET SO2 CONCENTRATION (PPM)
365
-------
removed efficiency (around 400 ppm of 862) is reached. At lower
values of SOo inlet concentration, it becomes increasingly dif-
ficult to achieve high percentage 862 removal. Thus blanket
edicts to remove a certain percentage of the sulfur in the fuel
pose a problem for the scrubber designer as well as the utility
user.
Reagent Comparison
Figure 5 presents a comparison of three different reagents in the
same scrubber, the packed grid tower (PPA). As may be seen, with
this scrubber soda ash is a much more efficient scrubbing reagent
than limestone while lime is intermediate between the two. Also
shown on Figure 5 are a few data points obtained with ammonia.
Similar data for soda ash, lime and limestone obtained with the
TCA scrubber are presented in Figure 6 along with the same ammonia
scrubbing data obtained with the WPS. The data of Figures 5 and 6
indicate that the combination of ammonia used with the WPS scrubber
is superior in S02 removal to soda ash, lime, or limestone used
in either the PPA or TCA scrubbers. However, the experiments with
ammonia were discontinued because of plugging difficulties in the
scrubbing section of the WPS when operating on boiler flue gas.
In addition, significant amounts of unreacted ammonia passed
through the scrubber which creates an additional air pollution
problem by substituting NHg for S02-
The problems associated with NH^: its cost, the solubility of
the sulfite-sulfate in water, disposal, the lack of a regeneration
process, and the possibility of substituting ammonia for S02,
contributed to the decision not to perform any further tests with
NH3, even though it appears to be the most reactive reagent.
While Figure 5 indicated that soda ash was the most effective
scrubbing reagent at any given L/G ratio with the PPA, the data
of Figure 6, with the TCA scrubber operating at its optimum gas
flow rate of 1020 scfra, present a different story. With this
scrubber, the data presented indicate that lime is a superior
reagent to soda ash over the operable L/G range, and that lime-
stone is superior to both lime and soda ash at L/Gs greater than
58. A similar anomaly also exists with the LIS scrubber as shown
in Figure 7. With this scrubber, while soda ash is superior to
lime, limestone is superior to soda ash at L/G ratios above 47.
The superiority of limestone to lime in this scrubber was ex-
plained by the plugging with lime slurry on the underside of the
impingement trays. The reason for the superiority of limestone
to soda ash is difficult to explain. You will note, however,
that the S02 levels in the region of interest in Figure 7 are
below 6 ppm S02. The same type of instruments (Dynasciences)
366
-------
Figure 5
REAGANT COMPARISON
400
T
T
T
T
T
a.
o.
<
cc
LLJ
O
O
o
CM
o
CO
K-
III
200
Q 100
80
40
20
10
8
LIMESTONE
PPA-139O SCFM
LIME
PPA-139O SCFM
©
AMMONIA
WPA-1OOOSCFM
SODA ASH
PPA-139O SCFM
I
10
20
30
L/G
40
GPM
50
1000 SCFM
60
70
367
-------
Figure 6
REAGANT COMPARION
6OO
4OO
2OO
Q.
Q_
<
DC
LU
o
LIMESTONE
TCA 1O2O SCFM
1OO
80
6O
O 4O
LU
_l
O
2O
1O
SODA ASH
TCA-1O2O SCFM
- AMMONIA
0 0 WPS-
1OOO SCFM
LIME
TCA-IO2OSCFM
I
I
I
I
1
10
2O 30
L/G
4O 5O
GPM
6O
7O 8O
1000 SCFM
368
-------
600
400
Figure 7
REAGANT COMPARISON
LIS-980SCFM
200
Q.
Q.
O
UJ
O
O
CM
100-
L/G
1000 SCFM
369
-------
were used in testing the PPA, TCA, LIS, CVS, SCE and AIS scrubbers
and in all four instruments the presence of NOX influenced the
wet electrochemical cello Specifically NOX interference in the
instrument resulted in readings of SOo 8-10 ppm higher than actual.
This correction had to be subtracted from the instrument reading.
It is possible that the superiority of limestone to soda ash in
these two scrubbers is possibly due to measurement error. However,
the superiority of lime over soda ash in the TCA scrubber is more
difficult to explain and neither our test contractor nor their
consultants was able to explain this anomaly.
With the few experiments with lime performed in the AIS scrubber,
soda ash was more reactive, while in the SCE scrubber lime was
a more reactive reagent than limestone. Comparative data with
the three reagents utilizing the Chemico Venturi scrubber is
presented in Figure 8. With this scrubber the exit SOo levels
are high enough (168 ppm with limestone, 120 ppm with lime, and
55 ppm with soda ash, based on 400 ppm inlet concentration) to
eliminate possible instrument measuring errors. This data indi-
cates, as expected, that soda ash is superior to lime which in
turn is superior to limestone. In this scrubber, operating at
its optimum gas flow rate of 950 scfm, the S02 removal efficiency
(based on 400 ppm inlet S02 concentration) at the optimum L/G
ratio of 43, was about 86% with soda ash compared to 58% with
limestone.
Because of the solubility of sodium salts, a sodium system offers
many advantages due to reduced probability of scrubber plugging.
However, the cost, without regeneration, is high. Approximately
$500,000 dollars was spent trying to regenerate the sodium scrub-
bing solution on a pilot plant scale without success. Laboratory
experiments indicated that we could regenerate a sulfite-bisulfite
solution with lime, but not with limestone. However, on the pilot
plant scale using actual stack gases, we found considerable oxida-
tion of the sulfite to sulfate and we were not able to regenerate
the sulfate. In addition, the regeneration pilot plant design
was such that the calcium and sodium systems become intermingled
and we had plugging of scrubbers and centrifuges with the calcium
salts. We hope that the other papers in this symposium on double
alkali processes will be able to report more success with their
experiments than we are able to report.
Scrubber Comparison
Comparative data on five scrubbers operating with soda ash is
presented in Figure 9. This indicates that the AIS scrubber
requires a much lower L/G ratio to reach an exit gas concentra-
tion of 40 ppm (90% removal) than the TCA, PPA, LIS or CVS scrubber
370
-------
500
400
Figure 8
REAGANT COMPARISON
CVS-950 SCFM
10
L/G
1000 SCFM
371
-------
Figure 9
SCRUBBER COMPARISON-SODA ASH
CVS
88O SCFM
AIS
SCFM
/1300
TCA
1O1O SCFM
PPA
144O SCFM
- 9OO SCFM
1O 2O
30
L/G
4O 5O
GPM
10OO SCFM
6O 70 8O
372
-------
as well as indicating that the CVS scrubber, as tested, was unable
to achieve an exit S(>2 level of 40 ppm even with soda ash.
Similar data with limestone is presented in Figure 10. As indi-
cated, the LIS scrubber (a Lurgi Venturi followed by a three-
tray tower) was able to achieve the lowest S02 exit concentra-
tions (around 7 ppm 802) while the CVS scrubber minimum SO? con-
centration (ca. 170 ppm) was considerably above the desired 40
ppm.
Data on six scrubbers with lime is presented in Figure 11. This
data indicates that the RHS scrubber requires the lowest L/G ratio
to achieve the desired 40 ppm S02 outlet concentration. The
mechanism of operation of this scrubber makes it difficult to
really determine the real ratio of liquid to gas because of the
"random" method of contacting. The L/G presented in Figure 11
for the RHS is based on the lime feed rate, the actual amount
of liquid contacted per unit of gas flow is undoubtedly higher.
The L/G ratio used for the SCE scrubber was the actual amount
of liquid contacted with the gas, but since this scrubber is a
four-stage device, this amount of liquid is contacted with the
gas four different times. Thus, the L/G as plotted is correct
for comparison of mass transfer performance, but the amount of
liquid pumped should be multiplied by four to compute the horse-
power required for liquid pumping to be on a comparable basis
with the other scrubbers tested. The total horsepower required
is composed not only of the liquid pumping horsepower (and in the
case of the RHS, the power required to rotate the scrubber) but
also that of mixers, thickeners, instruments, etc. The major
contributor in most of the scrubbers tested, the fan power re-
quired, is a function of the pressure drop through the system.
Reagent Utilization
The composition of the scrubbing reagent influences the degree
of reagent utilization as well as scrubber performance. At Mohave,
cooling tower blowdown water is to be used as makeup water for
the scrubbers, therefore experiments with added soluble sodium
salts were performed. This was found to increase the scrubber
performance with limestone slurries. Calculations indicated
that with a 5% NaCl - 5% Na2S04 solution saturated with CaC03,
the sorption capacity on a once-through basis was almost five
times greater than with a solution saturated with CaC03 alone.
These calculations also indicated that the contribution of sodium
sulfate is more effective than that of sodium chloride. Experi-
ments with sodium salts added to a soluble lime slurry did not
seem to improve the performance of the SCE scrubber.
373
-------
Figure 10
SCRUBBER COMPARISON - LIMESTON
4OO
CL
Q.
-------
Figure 11
SCRUBBER COMPARISON - LIME
4OO
28O
2OO
a.
o.
O 12°
z
UJ
o
§
o
ON
CO
UJ
80
6O
40
32
28
2O
16
12
CVS
1OOO SCFM
II
II
-RHS
975 SCFM
I
\ \
0
PPA
139O SCFM
TCA
1O7O SCFM
SCE
3OOO SCFM
- 1
LIS
© 131O SCFM
10
2O 3O
L/G
375
40
GPM
SO
6O
7O
1OOO SCFM
-------
Increasing the suspended solids (i.e. calcium sulfate, calcium
sulfite and unreacted calcium carbonate) concentration from 5%
to 15% and reducing the calcium carbonate concentration from
1.5 to 0.5% would increase the limestone utilization rate since
less unreacted limestone would be lost with the filter cake.
Thus, in the first case
limestone utilization - 1 - 1.5 - 70%
While in the second case
limestone utilization - 1 - 0.5 = 96.6%
Experimental values in the first case were around 66%, while
in the second case were around 95% as indicated in Figure 12.
As indicated, experimental determination of the degree of reagent
utilization is difficult. In this four-day test, five different
methods were used, four of them based on time-consuming wet
chemical analyses of the cake and the slurry, the fifth based
on the S02 removal from the gas. As may be seen, this latter
method indicated a limestone utilization rate of 117%.
It should be pointed out that, while increasing the solids content
in a limestone scrubbing slurry increases the degree of reagent
utilization, it also increases erosion of nozzles and pumps as
well as increasing the probability of plugging. It was for this
reason that our 450,000 scfm Vertical Module is being designed
to use a 5% rather than a 157o solids slurry.
Experiments with the SCE scrubber using a soluble lime solution
were made to determine the degree of reagent utilization. In one
75 -hour test, this was determined to be 93.3% while in a second
159-hour test, the degree of reagent utilization was 92.8%.
Particulate Removal
An objective of the Mohave Test Program equal in importance to S02
removal has been that of particulate removal downstream of the
98% efficient electrostatic precipi tator s . The dust loading down-
stream of these precipi ta tor s , based on two years of operating
experience and the manufacturer's guaranteed performance with all
sections in service, is 0.05 grains /scf. Optimum precipitator
performance has typically met the manufacturer's expected per-
formance with all sections in service of a dust loading of 0.03
grains /scf. Most of the data presented in Figure 13 was based
on inlet grain loadings ranging from 0.01 to 0.02 gr/scf. How-
ever, with some precipitator sections out of service, due to
discharge wire failures, the precipitator manufacturer would
376
-------
FIGURE 12
DETERMINATION OF LIMESTONE UTILIZATION
BY DIFFERENT METHODS
10
ttt>
-------
expect grain loadings as high as 0.095 gr/scf and we have occasion-
ally recorded grain loadings this high. Figure 13 presents a
composite of the particulate removal data obtained downstream of
these precipitators with five different scrubbers. There are
many variables which affect particulate removal performance, for
example increasing the gas flow rate (and hence the pressure drop)
seems to increase the particulate removal performance on the CVC,
PPA, TCA and LIS scrubbers and in the SCE scrubber at gas flow
rates above 1750 scfm. Increasing the liquid flow rate on the SCE
scrubber is beneficial, at 2500 scfm, the particulate removal
efficiency is increased from 75% to 90% by increasing the liquid
flow rate from 50 to 66 GPM. At a given liquid and gas flow rate,
we have obtained good data correlations (with TCA, PPA, CVS, and
SCE scrubbers) on a log-log plot of % particulate removal versus
inlet grain loading. A straight line results with increased
particulate removal at higher inlet grain loadings.
This increased efficiency at higher inlet grain loadings is un-
doubtedly due to the fact that at higher dust loadings the stack
gas is composed of larger diameter particles due to precipitator
sections being out of service.
Particle sizes in the stack gas at Mohave are normally very small.
About 90 cumulative wt % of the particles are less than 4 microns
in diameter with 70 wt % less than 1 micron in diameter, 40 wt %
less than 0.5 micron in diameter and about 15 wt % less than 0.3
micron in diameter. With the SCE scrubber, our test data indi-
cated about 75% removal of 0.3 micron diameter particles, 87%
removal of 0.5 micron particles, 96% removal of 1.0 micron dia-
meter particles, and 97% removal of 1.5 micron diameter particles.
Pressure Drop
One important characteristic of a scrubber which contributes a
great deal to both the capital cost and the operating cost is the
pressure drop through the scrubber. The pressure drop is, of
course, primarily a function of the mechanical design of the
scrubber. Plugging of the scrubber or demister and operation
near the flooding point (particularly with the TCA scrubber and
the LIS scrubber; caused extremely high pressure drops. Partial
plugging occurred with all of the scrubbers tested, particularly
with hydrated lime slurries, except with the SCE scrubber. This
scrubber also had the lowest pressure drop (ranging from
-------
FIGURE 13
MOHAVE PILOT PLANTS
PARTICULATE
REMOVAL DATA
20
500
1000 1500 2000
GAS FLOW RATE-SCFM
379
2500
3000
-------
It was found that the contribution of skin friction and wake
contribution of the stage separations was about 7 (% £ v2) while
that of the liquid spray was about 0.4 (% fv2)(L/G) so that pres-
sure drop could be correlated by the following equation:
P - [7 + 0.4 (L/C)] (% e v2)
Comparison of pressure drop through the various scrubbers is
presented in Figure 14.
Operating Problems
Insufficient time prevents a detailed discussion of all of the
operating problems encountered in the pilot plant program. As
previously mentioned, tests with the WPS were discontinued pri-
marily due to plugging, although unreacted ammonia passing through
the device also was a contributing factor in this decision. There
was an accumulation of solid material at the wet-dry interface
with the RHS device, the lime utilization rate was high, and
particulate removal performance was marginal.
The planned test program of the AIS was completed but the mode
of operation was found to be different than that originally
proposed. Initially it was planned to operate the spray drier
with a thick slurry feed so that the spray droplets would be
"BB sized11 particles when dried. It was postulated that this size
particle could easily be collected in the downstream cyclone
separator. However, when the spray drier was operated in the
slurry (as contrasted to solution) mode (about 30 wt % Na2COo)
it was found that the soda ash utilization rate was about 25%,
making the reagent cost prohibitive. When the spray drier was
operated with lower concentration solutions, the droplets formed,
and hence the dried particles, were too small to be collected
efficiently in the mechanical separator In many of the experi-
ments, the dust loading leaving the system was greater than the
inlet dust loading.
Bechtel's test report on the LIS device stated "The Lurgi im-
pingement scrubber scaled up so badly several times in the lime
slurry tests that the desired gas flow rate could not be maintained."
Various parts had to be wire brushed, scraped, or acid cleaned.
The vanes above the venturi also had to be acid cleaned and the
vane demister plugged. In the limestone tests, appreciable build-
up occurred on the underside of the impingement trays as well as
a considerable accumulation of solids in the bottom of the liquid
downcomer trays. Flooding difficulties caused cancellation of
the tests at the higher liquid and gas rates.
380
-------
oo
Figure 14
SCRUBBER PRESSURE DROP
AP GAS FLOW RATE L/G
SCRUBBER INCHES OF H?0 SCFM GPM/100Q SCFM
SCE 10 3000 17
PPA 2.5 1390 54
TCA 6.2 1020 67
CVS 9.0 950 36
US 11.5 1300 35
-------
The PPA liquid inlet spray header plugged several times as did
the coarse screen and strainer intended to protect this header.
A thick scale was also deposited on the egg crate packing. With
regard to the demister, the Bechtel test report stated "The PPA
absorber demister was completely unsatisfactory for hydrated lime
slurry operation." After a ten-day test with limestone, it was
found that a hole had been eroded in the fiberglass wall from the
force of the slurry spray.
During initial limestone tests with the CVS there was an apprecia-
ble buildup of solids (about one inch thick) in the duct from the
venturi to the demister. After a ten-day run, the thickness of
these solids was between 2 and 3 inches. A slight buildup of
solids occurred in the separator. With the lime tests, only a
relatively thin scale was deposited in the throat, but thick
(1% inch) scale was deposited on the vessel walls and the line
to the entrainment separator. This scale also plugged the pressure
measurement lines, thus explaining some of the erratic BP measure-
ments obtained. The major problem with this scrubber was its
relatively low S02 removal efficiency.
The major difficulty with the TCA scrubber concerned erosion of
the "ping pong" balls. Initial measurements after about 200 hours
of operation indicated a wear rate of about 0.5 percent per day.
For the ten-day test with limestone, new balls were used and the
loss in weight over 10.8 days was 4.457o loss (0.246 grams/ball)
or 0.43% per day loss. The problem with ball erosion, in addi-
tion to the cost of ball replacement every three or four months,
is that when the balls wear through, they fill with slurry and
the balls remain immobile. Eventually, the slurry solidifies
and if enough balls are immobile the pressure drop increases
greatly. This spring, balls constructed of a new flexible soft
material instead of the previous hard material were tested.
However, in several days of testing so many of these balls split
in half that they were all replaced with the old style balls.
There were other difficulties with the TCA scrubber; spray nozzle
erosion, demister plugging, and some buildup of solids in low
velocity areas.
As previously indicated, there were no difficulties with the SCE
scrubber itself. There were problems with pump seals, as there
were with all of the scrubbers, but the design of this scrubber
is such that one pump can be taken out of service and repaired
without degradation of the exit SC^ concentration.
As previously indicated, difficulty was experienced with I.D.
fans due to uneven buildup of material on the rotors. More
problems were experienced with centrifuges than rotary filters.
382
-------
Reagent Cost
The annual cost of the reagents is a significant contributor to
operating costs. Since each mole of S02 required to be removed
requires one mole of lime (CaO) or limestone (CaCOo) or soda ash
(Na^COo), the theoretical minimum requirements (100% purity of
reagents, 100% utilization) to remove 32 Ibs of sulfur or 64 Ibs
of S02 are 56 Ibs of CaO or 100 Ibs of CaC03 or 106 Ibs of Na2C03«
Based on supplier quotations for delivery to the Mohave Generating
Station, 90 miles south of Las Vegas, Nevada, and including a
delivery charge of $5/ton, the information presented in Figure 15
was prepared. The important fact to note is that if limestone is
used in the "non-optimal utilization" system, lime costs on an
annual basis are lower. On the other hand, if an "optimal utiliza-
tion" limestone system is used, maintenance costs could be higher
due to the higher solids concentration in the slurry system.
Furthermore, this system has a greater potential for plugging and
particulate carryover.
Reagent and Scrubber Selection for Full Scale Testing
It should be emphasized at this point that the evaluation made
by Southern California Edison of various scrubbers and reagents
applies only to the Mohave Generating Station. Different condi-
tions at other generating stations might lead to conclusions dif-
ferent than the ones reached here.
The criteria on reagent cost and operation have been summarized
on a comparative basis in Figure 16. The numerical ratings
range from 0 to 3, with 3 representing outstanding performance,
2 average, 1 poor and 0 assigned where the minimum required per-
formance was not demonstrated.
Figure 16 compares potential scrubbing reagents in terms of
performance, cost and control factors. Based on the pilot plant
data, the best reagent system is soluble lime. Except for the
potential plugging problems with certain types of scrubbers, this
reagent has exhibited overall good characteristics. Without
regeneration, the cost and waste disposal problems associated
with soda ash are prohibitive. Limestone, the other possible
reagent choice, suffers in regard to removal capability and
system control. The addition of sodium and/or solids to lime-
stone salts improves performance but increases the waste disposal
problem. Ammonia is not recommended at all due to the potential
presence of ammonia in the stack gas.
Based on the experimental data, operating problems and the SO2
and particulate removal needs at Mohave and Navajo Projects, the
383
-------
FIGURE 15
% UTILIZATION
REAGENT COSTS
LIMESTONE +
LIMESTONE (NA & SOLIDS) LIME SODA ASH
66
95
98
99
REAGENT REQUIRED
(LB/LB OF SO2)
1.56/1
1.56/1
1.88/1 1.66/1
u>
00
REAGENT COST
($/TON TO MOHAVE )
12.50
12.50
21
50
ANNUAL COST
$1,725,378 $1,198,684 $1,090,500 $4,896,160
-------
Figure 16
COMPARISON OF SCRUBBING REAGENTS
CRITERIA
5O2 REMOVAL
UTILIZATION
U)
s EROSION & WEAR
MATERIAL COST
DISPOSAL
LIQUOR CONTROL
PLUGGING & SCALING
TOTAL
SODA
ASH
3
3
3
1
0
2
3
15
SOLUBLE
LIME
2.5
3
3
3
3
3
2
19.5
0.5%
LIME
2.5
2
2.5
3
3
2
1
16
LIME-
STONE
1
2
1
2
1
1
2
12
LIME-
STONE
+ SODIUM
2
2
1
2
1
1
2
11
LIME-
STONE
+ SODIUM
+ SOLIDS
2
3
0
3
1
1
1
11
AMMONIA
3
1
3
1
0
2
2
12
-------
following conclusions were reached:
1. It is essential that a full scale prototype scrubbing
system be constructed and operated reliably before
selection and installation of equipment to treat all
of the stack gas at either Mohave or Navajo.
2. Construction of two different test module systems would
give a higher probability that a workable system for the
entire plant could be developed in a reasonable time
frame.
3. The scrubber/reagent combinations which seemed to offer
the greatest chance of success were the TCA scrubber
using limestone and the SCE scrubber using lime.
4. Flexibility should be included in the systems design
including provisions to incorporate packing other than
ping pong balls in the TCA scrubber and the ability
to test this scrubber with lime and the SCE scrubber
with limestone.
Although we believe the Mohave Generating Station today is one
of the "cleanest" coal burning stations in the United States,
with S02 emission levels ranging from 200 to 250 ppm and dust
loadings from 0.01 to 0.03 grains/scf, Clark County, Nevada
District Board of Health believed that an improvement was neces-
sary. Clark County adopted a regulation limiting S0« emissions
to 0.15 Ib S02/million Btu after the plant was in operation.
In January of this year, the Clark County authorities granted
variances for Mohave which require compliance with this S02 emis-
sion limit. This is intended to provide the time required to
carry out the required prototype module test program. This program
includes installation and testing of two different scrubber module
prototypes each having a capacity equal to one-fifth the stack
gas flow for each of the two Mohave boilers. Based on module
test results and operating experience, the type and specific
design for the full scale scrubber systems would then be selected.
In accordance with this compliance plan and variance, we started
construction February 15, 1973, of two 450,000 scfm test modules
at Mohave. The time allowed for design and construction of these
modules is extremely short, particularly since we believe that a
450,000 scfm scrubber is larger than any operating today in the
United States. We are required to start checkout of the Horizontal
Module by November 1, 1973 and commence operation by December 1,
1973 with checkout of the Vertical Module to start January 1,
1974 and operations commencing March 1, 1974. However, we are
proceeding with this "forced draft" schedule in order to have
386
-------
time for an abbreviated test program before selecting the final
module design. At the conclusion of our test period, we will
construct at least ten more modules of similar size at Mohave.
These test modules are based on the results of our pilot plant
tests but represent a considerable increase in size. Normally,
good chemical engineering practice would be to extrapolate pilot
plant data by a factor of 10 when installing a new chemical pro-
cess. However, with the Horizontal Module, which is based on the
SCE scrubber, we also have a new unconventional piece of equipment,
the scrubber itself. Considering the difficulties encountered
with other large scale scrubbing systems, it required considerable
management fortitude to start construction of a new device 150
times larger than was tested on a pilot plant scale. The relative
scale up is shown in Figure 17 which is an artist's rendering of
the full size scrubber compared to the initial pilot plant.
The Vertical Module is based on the TCA scrubber. Construction of
a 450,000 scfm TCA scrubber represents a size factor increase of
450 over the 1,000 scfm pilot plant scrubber data. While TCA
scrubbers have been constructed in larger sizes at other locations,
operating difficulties informally reported to us by other utilities
did nothing to allay our fears. Therefore, provisions are being
incorporated in the design of this scrubber to allow us to replace
the balls and screens, if necessary, with the PPA type egg crate
packing or other packing material or possibly convert the device
to a simple spray tower. An artist's rendering of the Vertical
Module is presented in Figure 18, compared to the initial pilot
plant size. Unfortunately, the man in Figure 18 is smaller than
the man in Figure 17 so that a visual size comparison of the two
scrubbers is not available.
Some comparative information being incorporated in the test modules
design is presented in Figure 19. It is perhaps of interest that
an 18 MW transformer is being installed to provide power for these
two 160 MW modules. However, the 6000 hp F.D. fan motor for the
Vertical Module will obtain power directly from a 13.8 kV source.
Acknowledgement
As indicated previously, testing of the TCA, PPA, US and CVS
scrubbers was paid for by the Navajo Project Participants (Salt
River Project - Operating Agent, Arizona Public Service Company,
City of Los Angeles Department of Water and Power, Nevada Power
Company, Tucson Gas and Electric Company and the United States
Bureau of Reclamation) and the Mohave Project Participants
(Southern California Edison Company - Operating Agent, City of
Los Angles Deoartraent of Water and Power, Salt River Project,
aSd Nevada Power Smpany). Southern California Edison served
as Program Manlger for these tests while Bechtel Corporation
served as the test contractor.
387
-------
SIZE COMPARISON BETWEEN PILOT
AND FULL-SCALE HORIZONTAL SCRUBBERS
oo
oo
MliT FACILITY
FULL-SIZE HORIZONTAL TEST MODULE
Figure 17
-------
CO
ID
SIZE COMPARISON BETWEEN THE VERTICAL
TEST MODULE, PILOT-PLANT AND EXISTING FACILITIES
VERTICAL PILOT
iODULE FACILITY
MOHAVE
GENERATING STATION
-------
u>
t£
o
SO2 REMOVAL WITH LIME SLURRY
SO2 REMOVAL W/LIMESTONE SLURRY
THICKENER TANK
SCRUBBER DIMENSIONS
GAS VELOCITY IN SCRUBBER AT 128° F
GAS REHEAT TEMPERATURE RISE
SYSTEM PRESSURE DROP
NOMINAL
MAXIMUM
FAN HORSEPOWER
NOMINAL
MAXIMUM
SLURRY PUMPS
NOMINAL
MAXIMUM
LIQUID FLOW RATE
NOMINAL
MAXIMUM
US RATIO IGPM/1000 SCFM)
NOMINAL
MAXIMUM
ELECTRIC POWER
NOMINAL
MAXIMUM
FIGURE 19
450.000 SCFM MODULES
MOHAVE GENERATING STATION
HORIZONTAL MODULE
400 PPM TO 40 PPM
250 PPM TO 40 PPM
60 FT. DIA. x 16 FT. HIGH
15 FT x 30 FT. x 60 FT LONG
21.6 FT/SEC.
80° F
6 INCHES H20
10 INCHES HO
1200 HP
1750 HP
8 - 300 HP
12-300 HP
12.000GPM FOR EACH OF 4 STAGES
24,000 GPM FOR EACH OF 4 STAGES
20
40
3.0MW
5.0MW
VERTICAL MODULE
250 PPM TO 40 PPM
400 PPM TO 40 PPM
50 FT DIA. BY 8 FT. HIGH
18 FT x 40 FT. x 90 FT HIGH
12.6 FT/SEC.
80° F
31 INCHES H2O
40 INCHES H2O
4850 HP
6000 HP
4 - 500 HP
6 - 500 HP
37,350 GPM
37.350 GPM
83
83
8.8 MW
10 MW
-------
Testing of the AIS scrubber was paid for by the following members
of WEST Associates: Southern California Edison Company, Arizona
Public Service Company, City of Los Angeles Department of Water
and Power, City of Colorado Springs Department of Public Utilities,
Colorado-Ute Electric Association, Inc., El Paso Electric Company,
Idaho Power Company, Montana Power Company, Nevada Power Company,
Pacific Power and Light Company, Public Service Company of Colorado,
Public Service Company of New Mexico, Salt River Agricultural
Improvement and Power District, San Diego Gas and Electric Company,
Tucson Gas and Electric Company, and Utah Power and Light Company.
Southern California Edison Company served as Program Manager for
these tests while the Atomics International Division of North
American Rockwell served as the test contractor.
Funding for the RHS scrubber tests was primarily provided by the
National Lime Association although Southern California Edison
Company provided some financial support. Both organizations pro-
vided test personnel.
With regard to the WPS, EPA provided advice on instrumentation
and test procedures, but the funding was provided solely by
Southern California Edison. The test program was conducted by
Southern California Edison with assistance from Truesdail Labora-
tories.
Funding for the SCE scrubber test program described here was pro-
vided solely by Southern California Edison Company. The test
program was conducted by Southern California Edison and Truesdail
Laboratories.
Funding for the two 450,000 scfm test modules is being provided
by the Navajo Project Participants (represented by Mr. Tom Morong,
Chief Engineer and Assistant General Manager of the Salt River
Project) and the Mohave Project Participants (represented by Mr.
Jack B. Moore, Vice President-Advanced Engineering of the Southern
California Edison Company). Program Management of this Test
Modules Program is the responsibility of Southern California
Edison Company. Stearns-Roger, Inc. is responsible for the
design and procurement of equipment for the Horizontal Module,
with the Bechtel Power Corporation serving as the contractor at
the site. The Bechtel Power Corporation is also responsible for
the overall design and installation of the Vertical Module with
Universal Oil Products providing the scrubber and process design
under subcontract to Bechtel Power Corporation,
Finally, we feel somewhat embarassed about taking credit for the
work of so many individuals and would like to publicly acknowledge
that the individuals listed below made many identifiable contri-
butions to the experiments reported here.
Atomics International Division of North American Rockwell
Dr. Dennis Gehri
Mr. Donald Gylfe
391
-------
Bechtel Power Corporation
Mr. Paul"Grimm
Mr. Robert Keen
Dr. Wen Kuo
Mr. Fred Miller
Mr. Angelo Sassi
Dr. J. L. Shapiro
National Lime Association
Mr. Clifford Lewis
Radian Corporation
Dr0 Phil Lowell
Dr. Delbert Otmers
Southern California Edison Company
Mr. S. T. Carlisle
Mr. E. J. Fletcher
Mr. John M. Johnson
Dr. Dale G. Jones
Dr. E. A. Manker
Mr. W. Carl Martin
Mr. Richard B. Rolfe
Steams-Roger, Inc.
Dr. Robert M. Christiansen
Mr. Keith S. Campbell
Mr, John D. Ferrell
Mr. Dave Naulty
Dr. J. Louis York
Truesdail Laboratories
Mr. Harold Decker
Mr. Harold A. Kerry
Dr. Olgart Klejnot
Dr. Marty Prieto
Mr. Eli San Jose
U. S. Lime Division of Flintkote Corporation
Mr. Dan Walker
392
-------
A REVIEW OF BABCOCK & WILCOX
AIR POLLUTION CONTROL SYSTEMS
FOR UTILITY BOILERS
by
J. F. Stewart
Fossil Power Generation Division
Power Generation Group
Barberton, Ohio
393
-------
A REVIEW OF BABCOCK & WILCOX AIR POLLUTION CONTROL
SYSTEMS FOR UTILITY BOILERS
J. F. Stewart, Fossil Power Generation Division
Power Generation Group, Barberton, Ohio
Presented to
Environmental Protection Agency
Flue Gas Desulfurization Symposium
New Orleans, Louisiana
May 14-17,1973
INTRODUCTION
The Federal Air Quality Act of 1967 was set up in
an attempt to deal with air pollution problems in
this country on a regional basis. Over 90 air quality
control regions were originally designated to be the
basis for regional administration and enforcement
of the Act. There presently exist a total of 247 air
quality control regions through out the nation. The
Act also called for the Department of Health,
Education, and Welfare (HEW) to develop and
publish air pollution criteria that indicate the
extent to which pollutants are detrimental to the
health and property of people and how emissions
can be limited and controlled.
In an effort to speed up progress toward cleaner
air, Congress passed the Clean Air Amendments of
1970 that set up the Environmental Protection
Agency (EPA), directed the EPA to set national
primary and secondary ground level air quality
standards, and set a timetable for their
implementation. The initial set of standards was
issued April 30, 1971, and covered six common
pollutants, including particulate matter, sulfur
dioxide and nitrogen oxides. Table I lists ground
level standards that would primarily affect utility
boilers. Primary standards are intended to protect
the public health with an adequate margin of
safety, and the secondary standards are intended to
protect the public welfare, and consider such
factors as soiling, corrosion and vegetation damage.
Each state was responsible, by January 31, 1972,
to develop and submit an implementation plan for
approval by the EPA. The state plans are required
to include emission control limits designed to
achieve and maintain these national ambient air
quality standards. The States have until July, 1975
to achieve air quality equal to or better than the
national primary standards. Secondary standards
are to be met 27 months later, depending on the
availability of adequate control technology, land
use and transportation control.
The ammended Clean Air Act of 1970 also has
resulted in the establishment of emission standards
for new stationary sources. Limits for particulate
matter, sulfur dioxide (SO2), and nitrogen oxides
(NOx) related to fossil fuel-fired steam generators
of more than 250 million BTU/hour fuel imput are
shown on Table II.
This paper will deal with the research and
development programs, and demonstration
scrubber systems that Babcock & Wilcox has been
developing during recent years in an attempt to
provide the electric utilities with solutions to their
present and future air pollution control problems.
PARTICULATE MATTER
Particulate emissions from most coal-fired and
some oil-fired utility boilers have been controlled
for many years with mechanical collectors,
electrostatic precipitators or a combination of
both. In order to meet the stringent particulate
collection efficiency required by the EPA for new
sources, the engineer today is faced with a difficult
problem. Further, many states are adopting
particulate emission limits for existing units which
are more stringent that the EPA standards for new
sources. For example, the state of New Mexico has
recently passed an Air Quality regulation that will
limit particulate stack emission levels to .05
Ib/MKB input by 1975.
394
-------
Engineers today can choose from four basic types
of collection devices when selecting participate
removal equipment. These devices include:
1. Cyclone-type mechanical collectors and
classifying hoppers.
2. Electrostatic precipitators.
3. Wet impingement-type scrubbers.
4. Bag filter houses.
1. Mechanical Collectors
Mechanical collectors and classifying
hopper devices are typically low collection
efficiency components when handling
flyash from a coal-fired boiler. Collection
efficiencies for these devices will range
from 20 to 75%, considerably below the
requirements being set by the EPA and
state authorities.
2. Electrostatic Precipitators
The electrostatic precipitator has become
the principal gas cleaning device for boilers
where fine particles cannot be collected in
a mechanical device. Properly designed and
arranged, precipitators are able to perform
at high collection efficiencies over a wide
range of particle sizes. However, as these
systems age, electrodes corrode and break,
hoppers bridge, deposits form on
insulators, and more frequent flashover
occurs with a resulting increase in the
emission rate. To compensate for these
malfunctions, and they most certainly do
occur, the designer must allow for some
redundancy and conservatism in the
precipitator design so that design efficiency
can be maintained without reducing boiler
load or incurring an outage. It is
conceivable that a precipitator designed for
92 to 96% collection efficiency can
continue to perform near design efficiency
with some degree of malfunction; however,
the margin of safety becomes very slim
when a unit is designed for greater than
99% collection efficiency. These are the
minimal efficiencies that are going to be
required in order to meet the 0.1 Ib/MKB
set forth by the EPA. A typical 10,000
Btu/lb fuel with 20% ash will require the
precipitator to perform at an average
efficiency of 99.6% in order to meet this
requirement. It would appear that the
prudent designer may have to provide spare
precipitator sections which could be
isolated from the operating precipitator
modules with the boiler in service to permit
routine maintenance, hopper cleaning, etc.
if the required emission level is to be met
on a day in day out basis. One alternate to
providing a modular precipitator design
would be for the engineer to select
conservative precipitator designs that are
less subject to discharge electrode failure.
Precipitator efficiency is controlled by
many factors which include dust size and
loading, gas temperature, sulfur oxides
concentrations, moisture content, ash
chemical composition, treatment time and
gas velocity. The shift of many power plant
operators to low sulfur western fuels is
currently having a marked affect on the
operation of their existing precipitators,
which in many cases, were designed for
high sulfur eastern fuels. The immediate
result can be a marked increase in stack
emission due to many factors not all of
which are entirely related to the sulfur
content of the fuel. In some cases, the
higher moisture content of the western fuel
can have a compensating effect on the
reduction of the sulfur of the fuel with
respect to precipitator performance.
Many operators are redesigning old
precipitators, adding on new sections, or
installing completely new precipitators to
perform on low sulfur fuels based on little
operating experience or resistivity data
applicable to their specific gas conditions.
Also, little is known today to what extent
coal composition can be varied without
affecting precipitator efficiency. It would
appear that considerable effort is required
to establish what effect parameters such as
coal and ash composition chemistry, mode
of burning, flue gas moisture, SO2 and SO3
concentration, temperature, and velocity
have on precipitator performance. This will
be especially important in designing high
efficiency precipitators for low sulfur fuels
that are expected to have a wide variation
in coal composition over the plant life. It is
expected that these plants will be supplying
a large portion of our future fossil energy
requirements.
395
-------
3. Wet Impingement-Type Scrubbers
Wet impingement-type scrubbers have been
solving a multitude of varied problems for
the chemical, pulp and paper and steel
industries for many years. Within the past
four years, a number of utilities have begun
to install prototype and demonstration
scrubbers for the purpose of removing
particulate and/or SO2 from boiler flue gas.
Particulate removal in these scrubbers is
accomplished in a number of ways. Some
designs rely on quenching the flue gases to
the adiabatic saturation temperature with
wetting and/or agglomeration of the
particles in a low velocity duct. Their
removal from the gas stream is
accomplished by gravitational forces and
en trainmen t separators.
Other devices impact the quenched gases
on a wetted packing, such as marbles, balls,
or bubble caps, and remove particulate by
the process of inertia! impaction. Particles
are removed from the gas stream in this
process because the particles are unable to
follow the gas stream around the packing,
resulting in the particle impacting against
the packing. Collection efficiencies for
devices of this type are dependent on the
particle size distribution of the dust
entering the scrubber. Most scrubbers have
good collection efficiencies on large
particles greater than one micron; however,
the collection efficiency for submicron
particles can decrease rapidly unless the
particles are accelerated sufficiently to
cause impaction on the packing surface.
The performance of wetted packing can be
affected significantly unless the gas and
liquor distribution remains uniform over
the bed surface.
Removal of submicron particles can be
accomplished in a wet scrubber if the
particles are accelerated sufficiently and
then permitted to collide with or on a
droplet surface. This can be accomplished
in a high energy venturi scrubber. The
smaller the size of the particle to be
removed, the higher the velocity and
energy required. Most of the energy losses
in a venturi result from accelerating the
scrubbing liquid. In a venturi scrubber, the
probability of a particle colliding with a
water droplet is greatly increased by
maximizing the number of water droplets
in the throat area. This can be
accomplished to a degree by first atomizing
the liquid; however, more complete
atomization of the liquid droplet can be
produced by the shearing action of the gas
stream. The accelerated particles impact on
the fine liquid droplets which subsequently
collide with each other and agglomerate.
The gas stream is then decelerated and the
water droplets with their captured particles
are removed from the gas stream by gravity
or inertia! separation.
Venturi scrubbers have been used for many
years to scrub fine fumes such as the salt
cake generated in Kraft recovery boilers
where 40 to 50% of the particles are less
than one micron in size. More recently,
high energy venturi scrubbers have been
employed to scrub the iron oxide fume
emitted from Open Hearth and Basic
Oxygen Furnaces. In this application,
where 90% of the particles are less than one
micron, the energy requirements amount to
a 50 to 60 in. wg gas side pressure loss to
obtain virtually a clear stack.
What energy requirements are required in
the case of boiler fly-ash? It is first
necessary to define the particle size
distribution of the fly-ash to be collected.
The classical method for obtaining particle
size distribution for fly-ash has been to
obtain a fly-ash sample according to ASME
PTC-27 and determine specific gravity and
particle size distribution by Banco Analyses
(ASME PTC-28). Figure 1 is a plot showing
typical fly-ash particle size distribution for
large pulverized coal-fired or cyclone-fired
boilers. Banco data is usually not reported
below a particle diameter of 2 microns
because the smallest size fraction
determined with this analysis is 1.7 to 2
microns. It can readily be seen from Figure
1 that a significant percentage of the
fly-ash (6 to 7%) exists in the fraction
below two microns, which may or may not
follow the same distribution slope as the
larger material fraction. The fraction below
396
-------
2 microns is the most difficult to collect,
regardless of the type of collection device
employed.
A venturi scrubber functions much like a
sieve; that is, it has a cut-off point, which is
a function of gas velocity and recycle
liquor rate. These two parameters
determine the venturi pressure loss which
for a given collection efficiency must be
made to vary inversely with dust particle
size.
The theoretical venturi grade efficiency
curves shown in Figure 2 illustrate how the
cut-off point for a venturi can be shifted
with energy level. These are calculated
curves and assume ideal conditions that do
not occur in an actual operating venturi.
The collection efficiencies indicated are
significantly greater than those expected
for an operating unit and can be effected
by venturi design. These curves illustrate
how the particle size cut-off point in a
venturi can be shifted with changes in
venturi energy level.
It therefore is quite apparent that
submicron particle size distribution for a
fly-ash must be determined before the
required venturi energy levels for a given
collection efficiency can be calculated.
Submicron fractions have in the past been
examined using light scattering devices,
ultracentrifuge techniques, transmission
electron microscope, and the scanning
electron microscope. These methods,
however, suffer from the same
shortcoming, i.e., not reproducing the
particle size distribution as it exists in the
gas stream. It was expected that dust
loading and particle size distribution can
vary considerably, according to application;
therefore, some tool was needed that
would measure the actual dust loading and
particle size distribution as exist in the gas
stream of an intended installation. Such a
technique would eliminate the need for
expensive and time consuming field pilot
plant tests, which only indirectly give
particle size distribution and are subject to
error in extrapolation to commercial sizes.
To perform these measurements, B&W
utilizes a commercial cascade impactor
modified to include a cyclone separator in
series with from one to seven impactor
stages.1 These components are assembled
in a probe that can be inserted in a duct for
dust sampling. An iso-kinetic gas sample is.
drawn through the sample probe, which*
acts like seven Venturis in series. Figure 3
shows the sampling train employed.
Each impactor stage has an orifice and
collection cup as illustrated in Figure 4.
The orifice diameter and the distance
between the orifice and cup determine the
particulate collection characteristics of the
stage. Some typical particle size
distributions for fly-ash determined with
this device are shown in Figure 5 along
with a typical Bahco analysis for fly-ash.
Using theoretical analysis coupled with
experimental results, the Research and
Development Center of Babcock & Wilcox
at Alliance has developed a mathematical
model to predict venturi scrubber
performance. With this model,1 we are able
to construct theoretical grade efficiency
curves, shown typically in Figure 2 for
various venturi design configurations, and
apply particle size distribution data
determined with the cascade impactor to
predict theoretical venturi performance for
different energy levels.
Wet scrubbers may not necessarily be the
final answer to every dust collection
problem especially in water-scarce areas or
where visible vapor plumes are
objectionable. Large quantities of water are
evaporated in cooling the flue gas stream to
its adiabatic saturation temperature. This
water quantity can be as much as 900 gpm
for an 850 MW scrubber unit. Another loss
is the dilution water required to remove the
ash as a slurry from the system. Some of
the dilution water can be recovered with
suitable thickening and dewatering
equipment; however, some degree of
blowdown will be required for these
systems due to dissolved solids buildup.
The extent of the blowdown will depend
on the chemistry of the fly-ash and the
make-up water supply to the scrubber.
397
-------
Bag Filter Houses
In applications where extremely low
discharge emissions are required, the energy
requirements for a wet scrubber may
become excessive when compared to that
required to obtain the same performance
with a bag filter house.
Highly efficient filter houses are commonly
used to collect dusts from cement plants,
fertilizer plants, metallurgical furnaces, and
other applications where collection of
submicron material is required. The
successful application of a full-scale filter
bag house to a 320 MW oil-fired boiler for
the control of visible stack emissions was
demonstrated in 1968.3 A number of pilot
filter houses have been installed in recent
years on pulverized coal-fired boilers to
determine pressure drop requirements, air
to cloth ratios, and operating performance
as well as determining the economics of
applying these systems to coal-fired units.
The performance of a large full-scale filter
house applied to a coal-fired unit has not
yet been demonstrated. Application of
filter houses to coal-fired units in the
future will probably be warranted for
special situations where extremely high
collection efficiencies are required and
after taking into consideration site
location, fuel source and water availability.
When the engineer today starts to consider
the type of particulate collection
equipment that will best fit his present and
future needs, there will be factors which in
the past did not enter into his decision
making. Preference may be given by some
to wet collection systems since many
designs may be augmented in the future
with various basic materials, such as
limestone for removal of sulfur dioxide.
Others may have a ready market for
dry-collected fly-ash and will find a
combination precipitator for removal of
the bulk of the ash and wet scrubbers for
removal of submicron dusts and/or SO2
removal the best overall solution. Some will
find that application of a wet scrubber to
their particular fuel will result in serious
scaling of the slurry lines due to the
chemistry of the ash. When more
experience is gained, it may be possible to
predict what ashes are likely to cause
scaling problems in scrubbers,
SULFUR DIOXIDE REMOVAL SYSTEMS
No processes have received more attention by the
different engineering disciplines in recent years
than those that are being developed for removal of
sulfur dioxide from boiler flue gas. In the mid to
late sixties, the primary incentive for their
development was the attractive price for sulfur
which peaked out at $40 per long ton in 1968. The
price for sulfur in recent years has steadily dropped
to levels that have forced the closing of a number
of Frasch process mines in the Gulf Coast region.
Recent prices for Canadian sulfur, most of which
comes as a by-product of natural gas production,
have been quoted as low as $9.50 per long ton
delivered to some midwest markets. There will
always be a price available for recovered sulfur and
sulfur products, but it would appear that this price
will be adversely influenced in the future as more
crude oil stocks are desulverized and as sulfur is
recovered from flue gas sources.
NON RECOVERY SYSTEMS
The development of flue gas desulfurization
systems by B&W has been concentrated on both
recovery and non-recovery systems. A number of
bases can be employed in a scrubber system for
removal of SO: as a waste product. These bases
include lime, high calcium limestone, sodium
carbonate, sodium hydroxide and ammonia. A
development program was initiated by B&W five
years ago to determine the performance for various
basic materials in a number of wet scrubber
devices. Out of these investigations emerged the
B&W limestone wet scrubbing system. Bases such
as sodium carbonate, sodium hydroxide, ammonia
and lime were found to give better SO2 removal
performance than limestone, but these systems
have high raw material costs and many of the bases
result in sulfur products that have a high solubility
in water and would be difficult and expensive to
dispose of.
The chemical costs of lime could be reduced by
injecting pulverized limestone into the boiler
furnace to accomplish its decomposition to
calcium oxide. There are several disadvantages to
398
-------
this method of operation:
1. Injected pulverized stone increases the dust
loading and duty on the scrubber system.
2. Universal application of the injection
system was not deemed possible due to
possible pluggage in reheater and
economizer sections and slagging
conditions that could occur due to boiler
design or the type of fuel utilized.
3. An injection system could not be applied
to oil-fired units due to the tendency for
limestone to deposit on the furnace walls
or convection surfaces resulting in serious
changes to the furnace, superheater, and
reheater heat absorption.
4. The problems associated with circulating a
lime slurry are well documented in calcium
base pulping. The problem of maintaining
system chemistry to prevent scaling of
piping and other hardware in a lime system
was considered more critical than the
control required for the limestone system.
Concurrent with determining that limestone was an
acceptable base for an SOj removal system, a
program was initiated for evaluating and
determining the scrubber that had the potential for
at least 80% sulfur dioxide removal.
This degree of SO2 removal could not be
accomplished in a single or double stage venturi
with limestone. A counter current tray absorber
did look promising, however, when considering
both performance and probable operating
problems.
Again, the engineers at B&W's Research and
Development Center at Alliance attacked the
absorber problem with a fundamental approach.
They felt it would be extremely dangerous to
. scale-up pilot plant test results to a 125 MW size
absorber unless the absorption mechanism with
limestone was understood. A mathematical model4
was developed to determine what effect the
significant variables have on scrubber performance.
The model was later confirmed in the laboratory
pilot plant shown in Figure 6. This model considers
the normal operating variables such as flue gas
flow, recycle liquor rates, slurry concentration,
reaction rates and diffusion constants for the
chemical species involved. In addition, the
comparative reactivity of various limestones was
determined so the prediction of SOa absorption
could be adjusted accordingly.
Other factors that influenced the decision to
proceed on development of limestone wet
scrubbing as a first generation system for SO3
removal were the low cost and high abundance of
high calcium limestones in most areas of the
United States. The reaction products from this
system, calcium sulfite and calcium sulfate, have
low water solubilities which reduce the potential
for this system to create a water pollution problem
from disposal of spent react ants.
In order to evaluate the suitability of various
limestones for use in the limestone system, several
methods for measuring limestone reactivity were
developed. One method involves chemical titration
of a slurry sample prepared from a pulverized
sample of the limestone. The quantity of titrant is
plotted as a function of time, while simultaneously
taking into account the change occurring in stone
fineness during the titration. This result is
compared with the titration rate for the standard
stone sample utilized for the pilot plant and model
test work. This test is used primarily for screening
purposes to determine those materials that should
be further screened in the small pilot plant.
This method of laboratory pilot plant testing of
stones for use in limestone systems provides for
excellent control over all test conditions. In
addition, the testing methods are not subject to the
many uncontrolled variables that occur when
conducting tests with costly field pilot plants.
Confirming tests of limestone performance have
been run under closed cycle conditions which are
very close to those the stone will experience at the
final installation. Closed cycle testing has been
conducted in a larger laboratory pilot plant that
includes a furnace that can burn 500 Ib/hr of
pulverized coal, a steam generating bank to cool
the combusion gases, a particulate venturi
scrubber, a B&W counter current tray absorber,
steam coil reheater and ID fan.
The slurry portion of the system includes a
limestone preparation and recirculation system, a
thickener and a vacuum belt filter. It is possible to
operate this system with maximum recovery of
water to determine the effects of dissolved solids
buildup on scaling and SO} absorption for various
limestones and fuels. This pilot plant is
instrumented with controls, sampling equipment
and is capable of continuous round-the-clock
operation. Figure 7 is a schematic showing the
closed cycle pilot plant which is located at our
Research Center.
399
-------
Waste Disposal
Spent slurry and fly-ash disposal requirements for a
coal-fired boiler with a limestone wet scrubbing
system will be about double that normally handled
for the boiler alone. This will be a severe burden at
many locations and could require this material be
dewatered and hauled away for disposal. All the
problems associated with sludge disposal are not
fully known. The possibility of utilizing the waste
stream form this process as a useful or valuable
product is considered highly remote in this
country.
B&W's efforts at waste disposal with this process
have been directed toward conversion of the waste
stream to a form that will facilitate its disposal,
minimize its effects on the environment and
reduce, for the customer, the quantity of sludge
for disposal. Research efforts currently in progress
have not progressed sufficiently to permit a
meaningful report at this time.
Demonstration Systems
The first full scale demonstration of the B&W
limestone wet scrubbing system is located at the
Will County Station of Commonwealth Company.
This system is retrofitted on a 163 MW net B&W
radiant cyclone - fired boiler that went in service in
1955. Two venturi • absorber scrubber modules
treat the entire flue gas flow leaving the unit,
770,000 ACFM.
The expected performance for this system
corresponds to an outlet dust emission of 0.0124
grains per DSCF and a sulfur dioxide removal
efficiency 81.6 percent. The system and its major
components are shown schematically in figure 8.
Each scrubber module consists of a variable throat
area venturi that removes the particulate and
provides an initial stage of sulfur dioxide removal,
followed by a B&W countercurrent tray absorber.
Venturi and absorber sprays both drain into
separate recirculation tanks that provide for delay
time to complete chemical reactions prior to
recirculating the slurry back to the module. Three
100 percent capacity venturi recirculation pumps,
each rated at 7300 gpm, and four-60 percent
capacity absorber recirculation pumps, each rated
5250 gpm were provided for the two modules.
Flue gases leaving each absorber pass through a
bare tube steam coil reheater and then to a 2250
hp ID booster fan. Both booster fans discharge into
the existing ID fan inlets.
The major control functions for the system,
limestone feed rate, venturi spray liquor rate,
venturi AP, slurry solids concentration and
limestone milling system operation are all
controlled automatically from the scrubber control
panel located in the existing boiler control room,
Figure 9 shows the instrumentation and control
diagram for the scrubber portion of the system at
Will County.
Limestone is received by river barge and stacked
with the existing coal handling equipment at Will
County. Stone is transferred by conveyor to two
storage silos. Each silo discharges on to a
gravemitric feeder that supplies one of two 100
percent capacity wet ball mill and classification
systems. Each mill has a capacity to grind 12 tons
per hour of limestone to a fineness of 95 percent
minus 325 mesh. Twenty percent solids feed
limestone slurry leaving the milling system is stored
in a slurry storage tank from where it is transferred
to the scrubber modules.
Spent slurry from the scrubber is pumped to a 65
foot diameter thickener. Clarified recycle water
discharges to a 5.5 acre pond and returned to the
cycle with the reclaim pumping system. Thickened
slurry underflow is pumped to a loading station
where fly-ash and other dry additives will be
blended with the thickener underflow, to modify
the sludge sufficiently to produce a stable land fill
material.
Detail engineering for the Will County project
commenced in September, 1970. One scrubber
module was placed in service on February 23,1972
with the second module going in service on April 7,
1972.
A second demonstration limestone wet scrubbing
system is being supplied for a new electric power
generating station located near La Cygne, Kansas.
The 820 MW net station is a joint project of Kansas
City Power & Light Company and Kansas Gas and
Electric Company. The 6,500,000 PPH Universal
Pressure Cyclone boiler will be fueled by a low
grade bituminous coal obtained from nearby
surface mines.
The air quality system, shown schematically in
Figure 10, consists of seven venturi-absorber
scrubber modules designed to handle the entire
flue gas flow of 2,370,000 ACFM. The system is
designed for 98.75% particulate removal which
corresponds to an outlet particulate emission of
0.10 lb. per million BTU fuel input. Sulfur dioxide
removal for the system is designed for 80%
efficiency.
400
-------
Each scrubber module consists of a variable throat
area venturi followed by a B&W countercurrent
tray absorber. Each module has a 100% capacity
venturi recirculation pump, each rated at 7750
gpm and supplied with 350 hp drives. A total of
seven, 100% capacity, 10,300 gpm absorber
recirculation pumps, were supplied, each with 400
hp drives.
Flue gases leaving the seven modules, are first
reheated 25 F with a bare-tube steam coil reheater
and discharged by six induced draft fans, 7000 hp
each, through a 700 ft. stack.
Limestone consumed by the scrubber system
comes from local quarries and is delivered by truck
and conveyed with the plants coal handling
equipment to two limestone storage silos. Two full
size 110-ton per hour wet ball mills grind the
limestone to a fineness of 95% minus 200 mesh.
Pulverized limestone slurry is then stored in two
200,000 gallon capacity slurry storage tanks.
Spent slurry, containing fly ash and
calcium-sulfur-reaction products, is pumped to a
160 acre settling pond, with clarified recycle water
returned to the system with two-100% capacity
pond return pumps.
This generating unit is currently in the start phase
with a scheduled commercial operating date of
May 1,1973.
SULFUR RECOVERY SYSTEMS
It is too early to determine the total magnitude of
sludge disposal problems associated with the
nonrecovery sulfur removal systems. Some
operators may find no economic means for
disposal of waste products and will direct their
attention to processes that minimize this problem.
One sulfur recovery process that can be applied as
a retrofit to existing units is a wet MgO system
B&W has been developing for the past six years.
B&W MgO System
Scrubbing flue gas with MgO and recovering the
sulfur values is not a new system. For over 20
years, many calcium sulfite pulping processes have
been converted to an advanced pulping-recovery
process developed and patented jointly by Howard
Smith Paper Mills Ltd., Weyerhaeuser Company,
and Babcock & Wilcox. Over twenty installations
of this type both in the United States and abroad
have been installed at both new and existing pulp
mills that utilize this process to recover sulfur
dioxide from the flue gas leaving chemical recovery
boilers. Figure 11 shows a flow schematic for a
typical Magnesium Bi-Sulfite Pulping and Recovery
process. The scrubbing liquor, a mixture -of
magnesium sulfite and bisulfite, is utilized as* a
cooking liquor during the pulping process. Weak
liquor from the digester is concentrated to 50 to
55% solids in a multiple effect evaporator and
burned in a B&W recovery furnace. Dissolved
lignins from the pulping process supply the fuel
that maintains the combustion process. Magnesium
sulfites and sulfates are thermally decomposed to
sulfur dioxide and magnesium oxide. Magnesium
oxide is removed from the flue gas stream with
mechanical collectors, washed to remove soluble
impurities, and slaked to magnesium hydroxide in
hot water. The slaked magnesium hydroxide is
then added to the sulfite-bisulfite scrubbing
solution to remove the sulfur values from the flue
gas stream in a venturi or contact tower before
exhausted the gases to the atmosphere.
The application of this process to utility boiler
stack gas application results in a number of process
changes. Most of these are associated with the
regeneration portion of the cycle due to the
thermal requirements for drying and
decomposition which must be supplied from fossil
fuels, since the recovered magnesium salts have no
heating values. A schematic of this process is
shown in Figure 12.
The first step in the process involves quenching of
the hot flue gases and removal of particulate from
the flue gas stream. Particulate removal can also be
accomplished in a high efficiency electrostatic
precipitator. Sulfur dioxide removal is
accomplished in an absorber of the same design
utilized for limestone scrubbing. Magnesium sulfite
hexahydrate slurry is removed as a blowdown from
the process, concentrated and passed to a
dewatering step where partial dehydration of the
crystals takes place. Surface moisture and
additional waters of hydration are removed in a
drier. The dried crystals are decomposed in an oil
or gas-fired furnace that discharges a sulfur-rich gas
suitable as a feed to a contact acid plant or to an
elemental sulfur conversion plant, depending on
the desired product. Magnesium oxide formed
during the decomposition step is recovered with
mechanical collectors, slaked and recycled to the
scrubber.
401
-------
B&W-Esso Dry Sorbent Flue Gas Desulfurization Process
SUMMARY
One of the more promising dry flue gas
desulfurization processes is the system being
developed jointly by B&W and Esso Research and
Engineering Company (ERE) with the support of
17 electric utility companies in the United States
and Canada, Dry sorbent systems are attractive for
many reasons. They avoid plume problems created
by wet scrubbing and do not have some of the
water disposal problems associated with wet
scrubbers. Dry systems which operate in the
temperature range of 600 to 700 P places the
sulfur dioxide removal equipment ahead of the air
heater in a boiler cycle. This should result in
reduced maintenance and improved performance
for the air heater.
In August, 1967, B&W and Esso Research and
Engineering jointly began the study of a dry
sorbent flue gas desulfurization process. These
studies showed that the development of a good
sorbent material that would absorb and desorb
sulfur dioxide in the temperature range of 600 to
700 F was feasible. Utility support was solicited
and obtained in 1969, and a three-phase research
and development program established. The final
phase of this program will involve the design,
installation, operation and testing of a 150 MW
demonstration system. The B&W-Esso process
utilizes an electrostatic precipitator for controlling
particulate emissions and a dry sorbent for
controlling sulfur dioxide and nitrogren oxide
emissions.
It provides advantages over existing systems in
that the process reduces nitrogen oxide and sulfur
oxide emissions simultaneously and also minimizes
the problem of disposing of spent material.
Product gas from the process is suitable for the
production of saleable sulfur or sulfuric acid.
Figure 13 represents an 800 MW coal-fired utility
boiler with the B&W-Esso system.
The programs described here are part of a
continuing research and development effort that
will provide the power industry with some of the
answers to their air pollution control problems.
Coal remains as our most abundant fossil fuel, with
proven reserves estimated at 830 billion tons, most
of which cannot be utilized for future power
production without some degree of sulfur removal.
REFERENCES
1. Downs, W., Strom, S.S., New Particle Size
Measuring Probe • Application to Aerosol
Collector and Emissions Evaluations, ASME
Paper No. 71 - WA/PTC-7., 1971.
2. Boll, R. H., "Particle Collection and Pressure
Drop in Venturi Scrubbers", I&EC
Fundamentals, Vol 12, pp 40-49, February
1973
3. Bagwell, F.A., Cox, L.F., Pirsh, E.A., Design
and Operating Experience With a Filterhouse
Installed On an Oil-Fired Boiler, Air Pollution
Control Association, St. Paul, Minnesota, 1968.
4. Boll, R.H., A Mathematical Model of SO2
Absorption by Limestone Slurry, First
International Lime/Limestone Wet Scrubbing
Symposium, Pensacola, Florida, March 16,
1970.
5. Downs, W., Kubasco, A.J., Magnesia Base Wet
Scrubbing of Pulverized Coal Generated Flue
Gas • Pilot Demonstration, Project Sponsored
by NAPCA, Contract CPA-22-69-162,
September 28, 1970.
102
-------
TABLE I
National Primary and Secondary
Ambient Air Quality Standards
Pollutant
Ground Level Concentration
Not To Exceed
Primary
Secondary
Particulate Matter (a)
Sulfur Oxides (b)
Nitrogen Oxides {b)
75
80
100
60
60
100
(a) Annual geometric mean
(b) Annual arithmetic mean
403
-------
TABLE 2
EPA Standards of Performance for New
Fossil Fuel-Fired Steam Generators
Fuel
Emission Limits Not To Exceed
Parti culate
Ib/MKB
Coal . 1 (a)
Oil {a)
Gas (a)
SO2 NOX
Ib/MKB Ib/MKB
1.2 .7
.8 .3
.2
(a) Visible emissions will be limited to not greater than 20%
opacity.
404
-------
Particle diameter,
DP (microns)
100
10
I
01
5 10 50
Cumulative mass, % less than DP
90
99
Figure 1
Typical fly ash particle size distribution
for large PC and Cyclone fired boilers
Banco-Analysis
405
-------
Theoretical
grade efficiency,
100
10
Press, drop
in wg
6
10
20
40
.01
.1
Particle diameter, DP, (microns)
Figure 2
Theoretical venturi grade efficiency curves
i.o
406
-------
Temperature
10
TO
o
Cascade
impactor
0)
w ra
'o *-•
Drying
Column
cu
-*—'
03
E
o
4-J
o
Vacuum
pump
Figure 3
Cascade impactor sample train
407
-------
o
1 1
1
1 1
inch
Enlarged section A-A
Section A-A
Collection
cup
Spring
Jet spindle
Gasket
Figure 4
Cascade impactor
408
-------
Particle diameter,
DP (microns)
100
10
0.1
Boiler A
Boiler B
Boiler C
0.1 1 10 50 90 99
Cumulative mass, % less than DP
Figure 5
Cumulative particle size distribution
409
-------
Figure 6
Laboratory pilot plant
410
-------
Figure 7
Limestone wet scrubbing pilot plant
-------
Limestone
bunker
To sluGge
treatment plant
Figure 8
Will County limestone wet scrubbing system
-------
Venturi -.Absorber Modules
< Slurry
Di
867 GPM
11.612GPM
Figure 9
WILL COUNTY INSTRUMENTATION AND CONTROL DIAGRAM
-------
Figure in
La Cygne Limestone wet scrubbing system
-------
CWS
STEAM FOR
PROCESS & POWER
RECOVERY FURNACE
MECH. DUST
COLLECTOR
I I
DIRECT-CONTACT
EVAPORATOR
MAKE-UP
(M9(OH)2
COOKING
ACID
STORAGE
MULTIPLE-EFFECT
EVAPORATORS
SULFUR BURNER -
GAS COOLER
STRONG
RED
LIQUOR
STORAGE
WEAK
RED
LIQUOR
STORAGE
Figure 11
Flow diagram of magnesium base pulping and recovery
-------
SO2 GAS TO
ACID PLANT
ABSORBER
RECIRC.
PUMP
VENTURI
RECIRC.
PUMP
F. O. FAN
TO SETTLING POND
MAKE UP WATER &
ADDATIVE FOR pH CONTROL
RECYCLE &
MAKE-UP WATER
Figure 12
Magnesia scrubbing system
BOILER
H2O MAKE-UP
-------
Electrostatic
precipitator
Stack
Primary air heater
(tubular)
Air in
Tempering air duct
Primary air outlet
Figure 13
800 MW B&W-ESSO system
-------
ONE YEAR'S PERFORMANCE AND OPERABILITY
OF THE CHEMICO/MITSUI CARBIDE SLUDGE
(LIME) ADDITIVE SO2 SCRUBBING SYSTEM
at
OHMUTA #1
(156 MW-Coal Fired)
by
Jun Sakanishi, General Manager
Miike Power Station
Mitsui Aluminum Co.
Ohmuta City, Japan
Robert H. Quig, P. E., Vice President
Pollution Control Company
Chemical Construction Corporation
One Penn Plaza
New York, New York
419
-------
FOREWORD
It is a great honor and pleasure to introduce and participate at this sym-
posium with Jun Sakanishi, General Manager of the Miike Power Station
(referred to by Chemico as Ohmuta No. 1, which is part of the Mitsui
Aluminum Company Ltd., Japan. Sakanishi San has provided the super-
visory leadership necessary to assure the success of this full-scale
multi-stage SO2 Scrubbing System which was retrofitted onto his 156 MW
coal-fired boiler.
Sakanishi San's participation has spanned the total history of the project
ranging from the basic design and pilot testing phases, through engineer-
ing, construction, and start-up; to over one year's operation. It was
Sakanishi San who trained his existing power plant operators to thoroughly
understand all functions of the system in terms of what had to be done
and, most important -- why it had to be done. The interaction of design
engineering, comprehensive training and skillful operation has paid off--
the first commercial-sized, successfully operating, highly efficient,
totally reliable SO Removal Scrubbing System.
While it is Chemico's pleasure to assume credit for the basic design en-
gineering and the associated process know-how, major acknowledgments
must be extended to the outstanding engineering and construction staffs
of the Mitsui Miike and Mitsui Aluminum companies who completed the
project in the unbelievably short period of nine months --within budget.
Similar work in the U.S.A., taking into consideration problems of work
practices and productivity in the construction area as well as the many
factors which can adversely affect engineering and procurement schedules,
would require a much longer period of time.
The System has been totally reliable and available for all power generation
requirements of the Mitsui Aluminum Plant --a facility which is totally
dependent upon a low cost, reliable power supply. Much credit for this
reliability must be attributed to the exceptional quality of Japanese equip-
ment manufacturers. The supply of conventional system components such
as fans, pumps, controls, etc., and the associated equipment servicing
by the Japanese, has proved to be everything which they claimed and more.
This is something which American vendor counterparts should note that
they need to seriously improve upon.
The success of this installation is indicative of the effectiveness which
can be achieved, if required, on American power plants because of the
following reasons:
420
-------
(a) The inlet SO2 concentrations at the Ohmuta Scrubbing System
are quite similar to most of the SC>2 concentration ranges
measured by Chemico during different pilot testing operations
at 17 different power plant sites throughout the U.S.A. over
the past five years. Only Western coal-fired operations have
been different in that these inlet 803 concentrations are gen-
erally lower than that experienced at Ohmuta.
(b) Carbide sludge (calcium hydroxide) used at Ohmuta and vari-
ous grades of lime which are expected to be proposed for most
applications in the U.S.A., have sufficiently common proper-
ties for SO2 scrubbing purposes to correlate their mutual
successes as additives. Simultaneous SO2 absorption pilot
testing, using both types of materials under similar conditions,
has demonstrated these correlations. This is not the case for
considerations involving limestone.
(c) The Ohmuta Plant is essentially a base loaded operation but
has been subjected to sufficient load swings and 803 concen-
tration variations to demonstrate that the scrubbing system
can readily handle various turndown conditions.
Since American applications envision the use of 150 MW
modular trains, which is the approximate size of Ohmuta,
there are no further scale-up factors involved.
(d) Fly ash removal at Ohmuta is accomplished by depending upon
a relatively efficient, previously existing precipitator. The
first stage scrubber/absorption section also acts as a polish-
ing function to remove residual ash in the flue gas and has
functioned to provide a system outlet loading as low as
0.001 grains/SCF.
The decision to incorporate a precipitator on American in-
stallations versus solely using a wet scrubber for dust col-
lection, will be a function of client considerations for site
conditions and the results of associated cost /benefit studies.
SO. absorption is common to both dust collection techniques.
£t
(e) The Ohmuta €3803/804 effluent bleed disposal system is
designed to operate, and has operated, as a closed recycle
loop between the scrubbers and the disposal ponds. The re-
cycle liquor has been totally saturated with sulfate for ex-
tended periods of time (months) without significant scale,
build-up or other deposition occurring in the scrubbers.
421
-------
This success of precluding build-up has been essentially
attributed to pH control throughout the liquid system com-
bined with strategic utilization of fresh make-up water which
is always required in any scrubber system to compensate
for conventional stack evaporative losses.
Some "spokesmen" have attempted to speculate that the pond
is purposely "blown down" or "dumped" into an adjacent bay.
This is not so. There were periods during the year, however,
when extensive rainfall from typhoons created an emergency
pond overflow condition. This problem has been resolved by
increasing the pond area and building up of higher dike re-
tainer walls.
Pond management of this effluent material is achievable at
conditions which are no more or no less complex than that
experienced on existing plant sites in this country today.
It is clear that the removal of SC^ for power plant stacks by the use of
any method, is going to significantly increase steam /electric generation
production costs in this country. The rationale for or against this SO%
control has obviously become a matter of multi-opinions on many issues
ranging from questionable need for such SC>2 compliance in the first place,
cost benefits associated with the various methods, and who should pay for
what and how. These issues are subjects, the rhetoric of which, Chemico
defers to others. We do say, however, that the SC>2 can be chemically
absorbed from the flue gas at high efficiency in a reliable manner, and
have demonstrated as such at Ohmuta. Thus, Chemico respectfully sub-
mits to this symposium that the criteria, established by an adhoc panel
of the National Academy of Engineers concerning demonstrated SC>2 tech-
nology for over one year, has been achieved.
There is no technical reason why SC>2 removal via the demonstrated
Ohmuta scrubbing technique cannot be achieved in a manner which would
not adversely affect availability and reliability of the steam /electric
generation facilities.
Robert H. Quig, P.E.
Vice President
Chemical Construction Corporation
Air Pollution Control Company
New York, New York
April 26, 1973
RHQ:rg
422
-------
HEMICO
FLY ASH - SO2 SCRUBBING SYSTEM
MITSUI ALUMINUM CO
OHMUTA PLANT (156 MW)
SCHEMATIC PROCESS FLOWSHEET
NJ
OJ
1— _
1
EXISTING
FACILITIES
1—
*
RETRO- FITTED^
FACJLITIES
•
BOIlER
PRECIP i.D STACK BOOSTER
FAN ~~
DELAY
TANK
TWO STAGE
VENTURI SCRUBBER
ONE MILE - C.SOj /C. 50./FLY ASH
CARBIDE SLUDGE
TANK
E_XISTIN&^ ASH
POMD
-------
1. Background of pro.ject of SQg control plant
Mitsui Aluminum Co., Ltd. was established in 1968 with the
objectives of 1) to support the local coal industry whose economic
base was being weakened due to the drastic change in structure
of the fuel source supply and 2) to develop an Aluminum smelting
plant which had been planned by Mitsui Group for many years.
We believe that a basic requirement for the Aluminum Smelting
Industry is to secure a cheap and stable supply of electricity.
Miike Mining Station of Mitsui Mining Co., Ltd. located in Ohmuta
City, is enjoying the largest quantity of coal production (6 million t/y)
in Japan and produces high quality coal for the Steel Industry. In
the process of this production, however, a large quantity of slime
coal which is unsalable waste coal is also produced. Under the sit-
uation, the basic requirement for Aluminum Smelting Industry has
been resdlved by utilizing the low-grade slime coal.
Although the coal produced by Miike Mining Station (so-called "Miike
coal") is the best in quality in Japan, it has two unfavorable character-
istics. First, it is easy to cause clinker problems and second, the
coal has a high sulfur content. At the time of commencing the project
of the power plant (1967-1968), therefore, our primary efforts in boiler
design were "how to avoid clinker trouble" and this was successful.
As to the high sulfur content, nobody anticipated at that time the
severe pollution control codes which are prevailing today and there
were no SO^ standards to be applied by codes. Under the situation,
we thought that it might be acceptable enough to build a 130 meters
high stack as a measure of SO2 control which was also a recommendation
of the Japanese MITI (Ministry of International Trade and Industry).
The rapid progress of the Japanese economy has made it necessary to
quickly strengthen pollution controls. Asa result, power companies
and others have been obliged to change their fuel sources. The com-
bination use of low-sulfur oil and high stacks was one of the most pre-
vailing counter-measures at that time. In order to meet the initial
objectives of establishing the company in the first place, Mitsui
Aluminum Company was determined to continue the use of coal as a
primary fuel source and to solve the SO2 problem through the operation
of coal fired boilers withSC'2 control
The power plant was completed in March 1971. At the start of opera-
tions coal which was to have a heat value of 9,000 Btu/Lb and 2.1 %
of sulfur content as had been originally planned, was changed to coal
-------
with a heat value of 10, 300 Btu/Lb and 1.7% of sulfur content. One
year later, in March 1972, the SO,, control system was completed
and placed in service. Since then, the power plant with the SO2
scrubbing system has been operated smoothly without any major
trouble, satisfying not only the national government control code,
but also the agreements with the local municipal administration.
Attached TABLE I shows the history of the power plant and the SO2
control plant at Mitsui Aluminum Co.
The work which Chemico had done to date best suited our criteria.
Mitsui Miike Machinery executed a license agreement with Chemico
and then we began to design the Ohmuta system.
Reasons for selection of Chemico/Mitsui Miike process are as
follow s:
First: - We believe that Chemico is one of the most experienced
companies in the field of chemical process engineering
and construction in the world. They are especially well
experienced in applications of the large-sized venturi
scrubber which is a critical part of a totally engineered
scrubbing system. Accordingly, easy size-up of the
venturi absorption tower and reliable process analysis
could be expected with Chemico.
Second: - The process including the venturi-absorption tower is
relatively very simple when compared to other SO2 control
technologies. This advantage means relative easier
operation and maintenance of the plant which requires a
minimum training of operators.
Third : - The use of venturi vessel scrubber designs allowed that
continuous long-period operation and high efficiency
would be assured. In addition, the Chemico type SOg
control plant allowed for a shorter construction schedule
which could meet our target schedule of completion of
the plant construction by March, 1972 to satisfy the time
schedule of MITI's air pollution code.
Fourth: - The estimate of construction cost was comparatively low,
relative to other technologies.
Fifth : - Mitsui Miike Machinery has had extensive experience
125
-------
in the supply of chemical machinery. Its factory is
located near the site so that the sufficient arrangement
could be expected for our various technical requirements.
I wish to note that, in order to assure the success in operation of the
first commercial SO2 control plant applied to a large-scale thermal
power station based on the wet Calcium base in cooperation with
Chemico/Mitsui Miike, we installed a pilot plant witti a gas flow of
1/200 of the commercial plant at the site of the power plant when we
placed an order for the commercial plant. The operation of this pilot
plant in technical cooperation with Chemico provided much knowledge and
experience in construction and operation of the commercial plant.
426
-------
2. Brief history of selection of Chemico/Mitsui Miike process.
In Japan, the dry type SO, control processes had been under develop-
ment, sponsored by MITI and power companies. Dry type SO2 control
processes were omitted from our selection because there still seemed
to be many difficulties to be resolved in both the technical and econo-
mical aspects. In addition, these dry processes were considered in-
applicable to our coal fired power plant which exhausts so much fly
ash.
Since we had to achieve reliable operation of the SC>2 control plant within
the limited time of construction, the following criteria were established
for process selection after our various investigation and examinations.
First: - Economical requirement.
Relatively low capital and operating costs are critical in order
to maintain power generation cost as low as possible. To
meet this requirement the wet type scrubbing process using
a Calcium basis is more advantageous.
Second:- Simplicity of the process.
Only relative simple developments or improvements to the
process could be tolerated in order to achieve reliable oper-
ation easily. This requirement is very important since it
will result in lower maintanance cost and easier operation of
the plant.
Third: - Requirement for the absorbent.
There is a plant of The Electro Chemical Industrial Co., Ltd.
(Denki Kagaku.Kogyo Co., Ltd.) located in Ohmuta City
where a lot of wet carbide sludge was disposed and dry carbide
sludge is also being produced. As these absorbents were
available at low prices, we planned to adopt a Calcium additive
SO2 scrubber system even if we had a disadvantage of a possible
problem of build up of scale - a major concern of all at the time.
Fourth: - By-product from waste disposal.
By-products such as sulfuric acid, elemental sulfur and sodium
sulfite were not preferable under the market conditions at
that time and the local conditions at the site. In case of Calcium
base being adopted, either throw-away system or Gypsum
process without the secondary pollution is suitable to our
requirements.
127
-------
Fifth:- Requirement for space available.
Since very small space was available, the system which had
the least number of scrubber-absorption towers required
was one of our basic requirements. Therefore, we required
that the process designer should be well experienced.
Sixth:- High efficiency.
It was imperative that high removal efficiencies of dust and
SO_ at low stoichiometric requirement should be achieved.
Seventh:- Assurance of stable and continuous long period operation.
It was also required that the process should not be too much sen-
sitive in operation, but rather flexible to meet the change in
load factor of boiler operation and the fluctuation of sulfur
content in coal.
428
-------
3. Outline of SO 2 control plant and basic specification.
The outline of this Power Plant and the SO- control plant is shown
in the attached TABLE II.
First:- History of SO2 scrubbing with many different types of
calcium additive processes has shown that build-up of
scale is one of the most difficult problems. However, we
made the decision to employ this process despite the opinion
mentioned above being widely accepted.
The basic resolution was to establish optimum operating con-
ditions which would minimize or eliminate scale by first using
a pilot plant. Furthermore, we planned to install spare cap-
acity for respective parts of the system including an extra
scrubber vessel. This stand-by scrubber unit and other
spare equipment was found unnecessary later when the internal
inspection of the scrubber system was carried out in November,
1972 after the system had been in service for six months.
Second:- The project of by-product Gypsum production was postponed
in order to first achieve successful performance of SO- control
project. We selected the throw-away system which enabled us to
utilize the existing ash pond for disposal of slurry in a form of
Calcium Sulfite (approx. 80 wt.%) and Calcium Sulfate, (Now
we are going to start Gypsum production.)
Third:- By the request of The Electro Chemical Industrial Co.,
Ltd. who supplied carbide sludge to the SO2 control plant
where carbide sludge is used as absorbent, we installed a
carbide sludge receiving system to handle either wet carbide
or dry carbide whichever supplied.
Fourth:- In order to avoid problems in boiler operation due to rapid
pressure change in the duct when an unexpected trouble with a
booster fan occured, we developed a specially designed duct
which can handle backward flow of gas without any damper.
Fifth:- In determining the gas flow, we settled K value with a
figure of 10 anticipating that it would be strengthened to 5 in the
future and we decided to install two units of scrubbers having a
capacity of 75% each of total gas flow. Two reasons for the
reserve capacity were a possible build-up of scale and the anti-
cipated strengthening of K value in future when two scrubbers will
handle total gas flow. The operation of one scrubber is enough
to meet the present local code.
429
-------
We proceed with the project with the basic specifications mentioned
now and we succeeded in completion of construction of the SO,
control plant within a short period of nine months in a limitea area
of approx. 2, 000 square meters (21, 500 square feet). In the mean-
time, the preventive measures against the disposal of waste water
was also taken. Flow sheet and arrangement are shown in the at-
tached drawings 1 and 2 respectively.
430
-------
4. Outline of operation status.
On March 29, 1972, the operation of the SO2 control plant was started
and an excellent record of continuous operation for 202 days until
October 17, 1972 was established.
Further continuous operation was assured and no necessity of major
modification was confirmed when the internal inspection of the plant
was carried out for the first time during the annual maintenance shut-
down of the power plant from October 17 to November 10, 1972. The
plant has been operated further quite satisfactorily since the plant
operation was re-started on November 11, 1972 except a shut-down of
the power plant for 4 days in January this year due to mechanical
trouble in boiler.
Following is outline of operation status.
First:- Continuous operation for a long period.
Fortunately, the operation was carried out without shut-down
although there were several troubles, all of which were just
minor and resolved without difficulty while the system remained
in service.
Please refer to TABLE III for details.
Second;- Internal inspection during the annual maintenance shut-down
of the power plant. Every detail of internal parts was inspected
twice, namely during the annual maintenance shut-down of the
power plant in October/November last year and at the time of
boiler shut-down in January this year, and no major defect was
found.
Please refer to TABLE IV.
Third:- One of the major items for daily maintenance work. PH
values is being controlled strictly by means of sampling of
recycle liquor at each 1st and 2nd stage every hour, measur-
ing pHvalue with a portable pH meter and then adjusting the
quantity of make-up slurry at each 1st and 2nd stage using control
valves so as to maintain pH value at a pH with the tolerance of
plus and minus 0.2 within the range of preferable pH value.
Fourth:- Basic requirement for operation.
Since it was the first commercial SO2 control plant in the
world and any failure should be avoided, the operation was
carried out at lower side of pH value within the range of
preferable value where build-up of scale was not noticed
during pilot operations. The SO2 removal efficiency was
431
-------
sacrificed to some extent at the beginning. We are now trying
to raise up the pH value gradually maintaining the smooth
operation of the plant.
Fifth;- Ash disposal pond construction.
Please see the attached drawing 3, Ash Pond Layout. Originally
the ash pond was constructed for ash disposal, however, it
is used now for settlement of waste slurry from the throw-away
system. Supernatant is returned to the plant for re-use. The
ash pond occupies approx. a million square feet and is located
at an area of approx. a mile away from the power plant. The
wall of south-western side is lined with a polyvinyl film to seal
the leak water in order to avoid a secondary pollution problem.
Perhaps because of the heavy rains and mine water which flow
into the pond, there is no accumulation of sulfuric acid, and
we have had no trouble at all maintaining the plant operation
during this one year. The returned water is re-used for spray
water, level control and make-up of Carbide slurry.
Fortunately, operation is being carried out quite satisfactorily. The
requirements for a large quantity of recycle water had been regarded
as one of the disadvantages of Chemico process at the time of process
evaluation. Actually, however, this disadvantage was proved later
to be one of the biggest advantages because of low solid concentration,
high SCL removal efficiency with lower pH value resulting in prevention
from build-up of scale. The data on operation experience are shown
in the attached TABLE V.
432
-------
E conomic Features
Plant cost is about one billion Japanese Yen which is about 3 million
dollars and this amount works out at about 6, 500 Japanese Yen which
is about 25 dollars per KW excluding Gypsum plant.
Actual operating costs greatly underran the originally planned budget
of 0.3 Japanese Yen/KWH which is about 0,1 cent/KWh for the past
year of plant operation. We can definitely say that we have been able
to achieve the goals in our economic evaluation which had been planned
as one of the most critical requirements at the time of commencing
our project. Our analysis indicates that firstly, 70% of the operation
cost is spent for interest, which means that the lower plant cost will
greatly contribute to the lower operation and secondly, the low cost
of absorbent by use of waste Carbide sludge shows clearly an advantage
of wet type Calcium basis SC<2 control process. In addition, the mea-
sures taken to protect the plant from the possible build-up of scale
have been also very useful to reduce the maintenance cost.
The use of the scrubber system has allowed the increased use of lower
cost slime coal in our boiler operation. We have increased the quantity
of slime coal from 30% (on dry basis) to more than 80%. This lower
cost coal contributed in avoiding an increase of the power generation
cost by offsetting the increased cost due to construction and operation
of the SO2 control plant.
The ash disposal pond is presently being filled with the waste slurry
consisting of mainly Calcium Sulfite. We are planning to start Gypsum
production within this year so that we can expect further reduction of
the operating cost by sales of this by-product in place of waste dis-
posal. There is a difinite market for gypsum in Japan.
Conclusion
It is almost one year since the plant was completed. In spite of being
one of the first large-sized commercial plants in the world, it has
shown an excellent result beyond our expectation with great satisfaction.
Moreover, to our surprise, so many people concerned not only in
Japan, but also in the United States and many overseas countries are
very interested in our success.
It is our intention that to further try to attain higher performance and
more economical operation of the plant. I believe that our strong
intent to maintain coal firing power station brought us the great success
in the wet calcium base SOg plant.
433
-------
TABLE I
"Brief history of projects of Miike Power Station
and SO3 Control Plant"
Year
1967
1968
1969
1970
1971
1972
Miike Power Station
Project was started.
(Nov.) MITI approved project.
Construction was started.
Control Plant
(Mar.) Test operation was
completed and the power plant
operation was started.
(Fuel coal: 10,250 Btu/Lb
S: 1.7%)
(Feb. ) K value: 29.5
Investigation was started.
Approach to domestic and
foreign process owners
was started.
Research test was carried
out at Mitsui Engyo Co. , Ltd.
(Mitsui Salt Ind. Co., Ltd.)
using IHI-TCA test plant.
(Feb.) K value: 17.5
(Nov.) First survey team
was sent to overseas
countries.
(April) Second survey team
was sent.
(June) Purchase order was
placed to Chemico/Mitsui
Miike Machinery Co., Ltd.
(July) Test using a pilot
plant was started.
(1,500 SCFM)
(Jan.) K value: 11.7
(Mar. 29) Operation of SO2
control plant was started.
(Apr.) SO2 control plant
passed acceptance test by MITI.
-------
Year
1972
Miike Power Station
SO2 Control Plant
(June) Installation of an additional (May) Performance test
pulverized coal dryer was completed, was carried out.
(Fuel Coal: 9, 550 Btu/Lb
S: 1.8%)
(Oct.) Power plant was shut down
due to the second annual maintenance
and inspection.
1973
(Jan. ) Power plant was shut down
due to boiler tube leakage (for
4 days).
(Feb.) (Fuel Coal:
9,900 Btu/Lb
S: 2.2%)
(Sept. ) Plant performance
was checked under various
operating conditions.
(Oct.) Inspection was made at
the first time to check
internals of equipment of SO2
control plant during the
power plant shut down.
(Oct.) Test operation of
Gypsum plant was completed.
(Jan.) K value: 9.34
(Jan. ) Inspection was made
again during shut down of the
power plant.
435
-------
TABLE II
"Outline of Power Plant and
SO» Control Plant"
Electric power plant
(1) Generator:
(2) Tubine:
(3) Boiler:
(4) Combustion system:
(5) Coal treatment:
(6) Electrostatci Precipitator;
(7) Stack:
SO,, control plant
'"""""Si """" ~ ~ L J_J- J-TLLJ- ^™^^^
(1) Process:
(2)
Capacity:
(3) Performance guarantee:
(4) Equipment details:
174, 000 KVA, 15, 000 V, 60 HZ
156, 250 KW 3, 600 rpm
490 T/Hr, Manufactured by
F.W. - I.H.I., Single drum natural
circulation type.
Pulverized coal firing. Front
burners arrangement type.
Two slime coal dryers, having
a capacity of 24 T/Hr and 45
T/Hr. each.
Dust removal efficiency: 98.7%
Outlet dust loading: 0.25 Gr/SCFD
425 ft high, concrete outer shell.
Chemico/Mitsui Miike Machinery
process, wet type and Calcium base.
2 units of gas flow of 241, OOOSCFM
each.
One unit only is operated for the time
being to handle 75% (241, 000 SCFM)
of total gas flow (319, 000 SCFM)
Two units will be operated simultan-
eously when required to handle to
total gas flow in future.
SO 2 removal efficiency: 90% or mere
Dust " " : 90% or more
Stoichiometric requirement of
absorbent 120% or less
Continuous reliable operation.
Boost-up Fans : 1, 000 KW x 2 units
Recycle Pumps: approx. 200 KW x
6 units.
436
-------
Scrubbers : 241, 000 SCFM x
2 units.
Reheat Furnace : 1 unit incl. Fan etc,
SO _ Analyzers : 2 units.
(5) Disposal system: Throw-away system.
Waste slurry is transferred to ash
pond and supernatant is returned
from ash pond to SO- control plant.
(6) Gypsum plant: Under planning.
437
-------
TABLE III
"Report on minor troubles occured in
operation of SO Control Plant"
1.
Modification for improvement
of carbide sludge receiving
system:
2.
Improvement of ash pond water
return pump:
3.
4.
Improvement of flushing
schedule of washing spray at
mist eliminator:
Change of make-up point of
Carbide slurry:
1) Protector was installed to
protect Agitator from damage
caused by various foreign
matters contained in wet
carbide sludge.
2) Dry carbide sludge contains
approx. 3% of carbon particles
and other foreign matters.
Cyclone separator was installed
to avoid possible troubles of
piling-up in a slurry tank and
errosion caused by carbon
particles and foreign matters.
Wooden chips flowed from
timber yard into ash pond by
heavy rain of typhoon in June,
1972 and plugged suction side
of ash pond water return pump.
Use of substituted water caused
increase of pressure drop of
gas at 1st stage mist eliminator
due to insufficient pressure of
water for mist eliminator
washing and may have caused
build-up of scale. Improvement
was made and strainer of
pump was installed.
Flusing schedule was improved.
Feed point at 2nd recycle line
was improved to avoid build-up
of scale.
438
-------
5. Improvement of connecting
tubes for langential nozzles of
recycle liquor of scrubber:
6. Improvement of rubber lining
arrangement after butterfly
valve:
1) Scrubbing liquor ran out due
to loose fitting of chemical
tubes. The tubes were changed
to rubber tubes.
1) Rubber lining of pipe inside was
damaged at a part after butterfly.
valve due to turbulence. The
rouble was resolved by instal-
lation of orifice.
439
-------
TABLE IV
"Status of equipment and internal
parts of SO2 control Plant inspected
during annual maintenance shut-down
of the power plant"
Carbide Sludge
receiving system:
Scrubber:
Recycle liquor piping;
(rubber lining)
Mist eliminator:
5. Piping to disposal pond:
6. Recycle water pipeline:
Piling-up of carbon particles was
found in slurry tank.
No defect was found at glass flakes
lining. No build-up of scale was
also found.
No defect was found in general.
Thin deposit of scale (1/25" to 1/12"
thick) was found but the plant was
operated without trouble.
Slight damage of rubber lining was
found at a part of piping after
butterfly valve.
Partial build-up of scale was found
at the parts of:
1) area affected by trouble with
ash pond water return pump at
the time of typhoon in June, 1972.
2) outer area water coverage of
spraying water was not sufficient-
No trouble is anticipated in future
operation because of minor modifi-
cation in mist eliminator sprays.
No erosion, corrosion and build-up
of scale were found.
Muddy scale was found, but removed
completely without trouble.
440
-------
7. Pumps:
8. Ducts:
9. Delay tank:
No trouble was found.
Slight damage was found at expansion
.pint.
Piling up of carbon particles contained
in slurry was found.
Exhaust gas containing a very small
quantity of 803 was connected with
a stack thru, vent gas piping,
441
-------
TABLE V
"Operation Data"
Most of figures were
obtained in operation
during January/February
1973.
1. Characteristics of coal.
Heat value: 9.800 Btu/Lb
Moisture: 6%
Ash content: 2.91%
Volatile matter: 35.44%
Fixed carbon: 36.86%
Moisture content: 1.59%
Sulfur: 2.2%
Ash fusibility: S.P. 2330<>F
2. Characteristics of carbide sludge:
Moisture Consumption Purity
Wet carbide: 55% approx. 80 T/D 85% as Ca(OH>2
Dry 6% " 30 T/D 90%
Total: - 110 T/D (wet base)
Ca(OH)2
wt.%
j:84.3
87.8
CaSOg
1/2H00
1.8
3.9
CaSO4
2H2O
1.0
0.3
CaCOg
6.9
4.9
Si02
1.7
2.7
A1203
0.7
0.7
Fe2
wt
°3
.%
1.1
0.
7
200 mesh
thru.
%
99.72
87.12
Dry
3. Main operation data (as of October, 1972)
Output: 156, 000 KW
Coal: 9,700 Btu/Lb S:1.9%
Gas load to handle: 75%
442
-------
SO 2 concentration Boiler outlet 1, 900 ppm
Scrubber outlet 180 ppm
Stack 610 ppm
Gas temperature Boiler outlet 315°F
Scrubber outlet 130°F
Stack 212QF
Reheated gas 172op
K value : 7.84
Ca(OH)24 Stoichiometric requirement:
105 to 110%
4. Return liquor from ash pond.
pH 8.2
ppm Mg++70 ppm Cl~ 728 ppm NO2-0.28ppm
— 1,012 " Fe++0.45 " Mn++-
SS 5.2 " COD 6.4 "
5. Slurry analysis.
Ca(OH)2 CaSOq CaSO4 CaCQ3 SiO2 A^Oa Fe2O
Make-up 87.3 0.5 0.7 2.3 1.9 0.24 0.8
Bleed 6.0 69.5 12.2 1.02 5.8 2.6 0.3
-------
TABLE VI
Explanation of "K value"
which is used as the SC^
emission standard in Japan
K value
Constant ranging from 5.26 to 29. 5 according to the area stipulated
by the Air Pollution Control Law of Japan and is shown in the following
equation concerning SCX allowable emissions.
Q = K x 10-3 x He2
Where Q:
Quantity of SO,, allowable emissions from
the stack. *(NM3/HR)
* Cubic meters at 0°C and 1 atm.
He: Effective height of stack (Meter)
calculated by the following Bosanquet's formula.
He = Ho + 0.65 (Hm + Ht)
Ho: Actual height of stack (Meter)
Hm: Equivalent height of stack (Meter)
effected by discharged flue gas
momentum.
Ht: Equivalent height of stack (Meter)
effected by discharged flue gas
temperature.
Both Hm and Ht are calculated by the following formulas
Wm - 4-77 x /Q-1 vg
Hm ' 1*0.43 V_ V^
Vg
Ht = 6.37 gxQl - t (loge j2 + _2_ 2)
V3T! J
J = V (0.43 PI . 0 28 VgTl
g*t
-------
Where Qi: Quantity of wet flue gas discharged from
(M3/SEC)
Vg: Discharge velocity of the above gas (M/SEC)
V: Wind velocity (M/SEC)
TI: Temperature at which density of discharged
flue gas becomes equal to that of atmosphere (°K)
At: Difference between temperature of
discharged flue gas and T^ (°C)
G: Temperature gradient (°C /M)
g: Gravity constant 9.81 (M/SEC2)
T! and G are usually assumed to be 288 K (15 5C) and
0.0033 respectively.
In case of Ohmuta city, (where Miike Power Station is located.)
K = 9.34 usually employed
K = 5.26 employed in an emergency when
judged by the local government.
For the Ohmuta area, the K-value has been progressively decreased, leading
to more stringent emission limitations with time, such as
1968: K = 29.50
1970: K = 17.50
1972 K = 11.70
1973 K = 9.34
445
-------
•t
4S
01
1st st-ze
venturl
recycle
Vet carbide pit
Aflh pond llouor
return pucp
Dry carbide ptt
Recycle slurry
... — Makeup slurry
__.__ Bleed slurry
.„ Return liquor
Note: Only one scrubber systea presently In operation.
• indicates a closed valve during f/72 to r>/72.
M indicates a closed damper during f/72 to 9/72.
Waite disposal pond
Process Flow Sheet
-------
-tr
* «V
*«ey jf\<\^ V*
^f^y^\\^ *
-------
T=«b 1173
Ash
-------
-
t
#1 Ohmuta scrubber
-------
.-•
Ul
O
#2 Ohmuta pond system
-------
PANEL DISCUSSION:
SIGNIFICANCE OF OPERATION TO DATE
OF 156-MW CHEMICO/MITSUI LIME SCRUBBING SYSTEM
451
-------
PANELISTS
P. Wechselblatt - Chemical Construction Corporation
J. Craig - Southern Services
H.W. Elder - TVA, Muscle Shoals
F.T. Princiotta - EPA
Pond Saturation
P. Wechselblatt— In the questions that were raised perhaps the
one that everyone was most interested in was the degree of saturation
of the pond. And I think that in order to view this, it is necessary
to understand exactly what is happening in the system. A great amount
of liquid is being recycled (perhaps 90 gal/1000 cfm in total), and there
is a huge storage of liquid. The amount of liquid going to the pond and
coming back from the pond is very low. In fact, if the liquid coming back
from the pond were boiler feed water, it would change the degree of
supersaturation by only 4 percent in the scrubber. Therefore,
except for the fact that Bob Quig took a walk on it—and I am sorry there
was no picture of just a hard hat where he sank—the pond was completely
irrelevant to the degree of saturation or supersaturation in the scrubber.
And when we argue about the pond points, I would suggest that if you
looked at the points in the scrubber recycle loop you would find invariably
that the liquid is saturated or supersaturated; therefore, the only line in
question on the job is the return line from the pond to the scrubber. That
line, of course, could scale if you had supersaturation, but all other lines
in the system are operating in the supersaturated mode.
452
-------
Lime/Limestone Buffering
The second thing that I would comment on is the buffering action of
limestone and the buffering action of lime. There really is very little
black magic in this. The numbers are reproducible. Everyone has done a
great deal of work in the last 5 or 6 years. When I was at the startup
of the plant in Omuta and helped with the performance tests, we altered
the carbide sludge to the system because our guarantee to Mitsui Aluminum
was a more severe guarantee than their requirement from the State, from
the municipality. So it was necessary to run the system at times for
guarantee purposes for the period for the guaranteed tests at a stoichi-
ometry where we could obtain better than 90 percent S02 removal. Under
most of the yearly conditions, since the code requirement is not that
stringent, the plant operates between 80 and 85 percent removal with a
lower utilization of carbide sludge. The system was remarkably responsive
to alterations in the carbide sludge feed. For example, if you wanted
to raise the pH of the system from 7 to 8.5 (which might have a
stoichiometric change of perhaps 0.9 to 1.2), it was possible to watch
the pH change with the change in lime feed. The response time of the
system is excellent. Now you try to balance the response time of the
lime system against the buffering action of the limestone system. I
really do not want to get into a hassle about it, but this rapid
response really is a tremendous aid. Particularly when
putting the system on automatic control (which would necessitate sulfur
dioxide measurements, gas volume measurements, and things of that sort),
you want the system to respond very quickly. This system responds almost
instantaneously. And I might say that with a change in feed you can
453
-------
watch the S02 out of the stack increase as you change the hydroxide feed.
It is almost instantaneous. You can watch it take place.
Mist Eliminators
Another comment that was made this morning had to do with mist
eleminators. Chemico operates mist eliminators very successfully. We
have had problems on other jobs, but we have never had a problem with
mist eliminators. We operate them on a rotating cycle and at a higher
velocity than anyone talked about this morning. In fact, if you take a
look at the design, you will find that usually about 20 percent (15 to
20 percent) of the cross-sectional area is taken up by supporting
structure. Therefore, the mist eliminators in general operate at about
20 percent higher superficial velocity than anyone calculates when they
calculate a cross-sectional area of the vessel.
Japanese Versus American Results
I was impressed when I saw the plant in operation. I have seen
plants of various sorts in this country, and I would just like to say
this, and I feel very strongly about it: it is kind of a challenge. I
am disheartened when I see a system designed in this country that was
erected by the Japanese in 9 months. We started at the same time
in this country, doing the same design work, and the first unit we will
have operational in the United States will be in another 2 to 3 months—
almost a 30-month schedule compared with their 9-month schedule. They
take our design sheets and process flow sheets, which they pay for;
they build a plant; and they start operating a plant. They have no
particular problem, and if they can do it there, we can do it here. And
I think it is a challenge to the industry. There is no reason why the
Japanese can do it better than we can.
454
-------
H. W. Elder - Okay, the subject is of significance and I think we
will all have to agree that in Japan the significance is unquestionable.
It has been an excellent installation. The Japanese have done one
fantastic job of installing equipment and making it run. And it is a
real credit to the initiative and fortitude of the Japanese that they
had this large measure of success. We have a problem now in translating
that technology to application in this country, and this is in no way
taking a shot at what has been done in Japan.
Transfer of Technology
I think there are some questions that must be raised about transfer
of technology from one application to another. I hate to beat this horse
some more, but this closed-loop business is important, I think; and Pete
(Wechselblatt), I am not sure I agree with your total conclusions on that
point. Number 1, it is too bad we have to talk about closed-loop systems.
We need some guidelines on quality of water that can be discharged from
any plant, not only limestone scrubbing power plants, but any situation.
And that would give us some guidelines then as to how much blowdown we
could take and how much fresh water we can add. Therein lies the
important point, because if you can add no fresh water, I can practically
guarantee that you will have scaling. Based on TVA's pilot plant operation
and my conversations with experienced persons, I believe that
use of saturated and supersaturated solutions for washing mist
eliminators (or for points where dilution is required) is absolutely
essential if you are going to avoid scaling. So I think that you have
to use at least some makeup water to make the system run. So then it
is just a question of the amount. For your situation or for the Mitsui
455
-------
situation, particularly with the low solids concentration in the
recirculated stream, the sulfate concentration can be rather critical.
I think that the accepted (well, it is pretty well agreed) way to
control sulfate scaling is by recirculation of calcium sulfate to provide
precipitation sites. And the lower the solids concentration, of course,
the lower the sulfate concentration. So if it becomes saturated with
sulfate, I would expect scaling to occur. The low solids in the loop
coupled with the clarifier liquor returned from the pond with a high
sulfate concentration could be dangerous it seems to me. A couple of
other points of uncertainty, and again I am just raising questions, the
fact that the waste material settles in a fashion that is a little bit
unusual based on our experience is a little perplexing. I think that if
the sulfite/sulfate ratio in the reacted solids is similar to that which we
have experienced, you would expect it to settle very poorly. So this says
to me that maybe there is something different about the carbide sludge as
compared with freshly calcined lime—differences in particle size or
differences in surface characteristics, for example. But if the sulfite/
sulfate ratio is the same and if the particle size is the same as freshly
calcined lime, you would expect the reacted product to behave the same
way. As far as flyash removal from the system, I am not sure how important
this is. If I were personally putting in an order for a system for control
of S02 from a new power plant, I would certainly want to take the flyash
out as well as avoid the expenditure for a precipitator. What effect
the dust has in admixture with the calcium reaction products is really
not clear, but it is worth raising the question. It affects both the
chemistry of the liquor and, of course, it affects the erosion or
456
-------
mechanical considerations in the scrubber system itself. So there are
really four main points that I am raising relating to the transfer of
technology: the amount of fresh water and how it can be added; the
question of low solids; the effect of sulfate on precipitation; and the
effect of flyash in the recirculating slurry loop. As I say, I am not
throwing rocks. I am just raising questions.
J. Craig—I will keep my discussion pretty brief. I feel there will
be a number of questions from the audience and probably some discussion
on the panel. Apparently, the installation is meeting Mitsui's design
criteria and operating goals. Mr. Sakanishi (J. Sakanishi, Mitsui
Aluminum Co.) and all the employees at his station are to be commended
on a well run and well maintained operation. Now, as to what I saw there:
I saw a station operating base-loaded, with apparently very little variation
in the flue gas flow rate. We observed operating logs for 8 days; based
upon that, there was probably plus or minus 2 MW variation from around the
156-MW capacity of the unit. In comparing that variation to our stations,
where we can go from probably 40 percent to 110 percent of the flue gas
flow rate in a matter of 24 hours, it leaves some question in my mind
whether you can just extrapolate Mitsui's operation to a typical plant in
our system. We also observed that sulfur content in the coal was pretty
constant at Mitsui; this is because of where they get their coal and how
it is processed. Again, if you look at our sulfur content, it could vary
as much as 2 to 4 percent sulfur in a relatively short period of time. I
believe both of these favorable operating conditions allow Mitsui to run a
chemical plant the way a chemical plant should be run—either at or near
steady state conditions. I am not fully convinced myself that we can run
457
-------
our typical units in this same kind of mode. I, too, walked on the pond,
but that is a little bit misleading—I am not a Sumo wrestler. The pond
has a crusted surface approximately 6 inches thick. It held me up, but
Mr. Sakanishi warned me not to go too far out. We penetrated the crust
very easily with a stick; below the crust was a very pasty material. I
might indicate that the top layer was thixotropic in nature. As a con-
sumer, we would like to buy a system that includes a solution to the prob-
lem of waste disposal. Mitsui apparently has solved their problem for a
short operating time; they are going to a gypsum final product. Again,
I am not sure that is the best solution for our needs. In addition,
there is sonic question regarding the water balance. Members of the group
walked around the pond and could not find any visible overflows from the
pond. That was on one given day and I cannot say that that was the way
it was for a year. We were not able to resolve the question of open-
loop versus closed-loop operation.
Areas of Commonality
F. T. Princiotta — First of all, I think we should discuss areas
of commonality and areas of difference, with perhaps the first statement
being that I do not think any one plant in the United States, for that mat-
ter, is completely representative of all other coal-fired plants; it is
an impossibility. We tried to do it at Shawnee, I might add, and it is
just impossible. Everybody has a different sulfur content in the coal,
a different ash content, and different sludge disposal logistics. So I
think it is unreasonable to expect one particular plant, particularly in
458
-------
a foreign country, to answer all questons regarding this technology.
However, I think anyone who has studied the system carefully has seen that
there are really amazing parallels to many U. S. applications, many
areas of commonality. In the first place, it is a lime-based American
system utilizing American-based technology. And, as everyone knows,
much potential exists for these systems in the United States due primarily
to the low cost of the input alkali. In my opinion (based on discussions
I have had with Combustion Engineering people, Chemico's pilot plant
experience, and the English experience 40 years ago), there is no reason
to believe, well there is no hard evidence, that there is any difference
in the characteristics of carbide sludge as opposed to calcium hydroxide.
The major difference that I see is that it is a pure form of calcium
hydroxide. Other areas of commonality include unit size, percentage of
sulfur in the coal, and use of a precipitator. The 156-Mw unit can be
easily scaled up without any increase in size of the basic module which
can handle this amount of flue gas. The system uses 2 percent sulfur
coal, whereas many coals in the United States may have higher sulfur con-
tent. Yet, if you averaged out the sulfur content of all the coals in
the United States, I suspect it would be close to 2 percent. Admittedly,
it would be nicer if that plant happened to have 3,4, or 5 percent sul-
fur coal; but it is relatively close to many U. S. coals. Regarding the
precipitator upstream of the scrubber, as you well know from several dis-
cussions so far, many utilities plan to employ a precipitator. The Navajo/
Mohave experiments will have a precipitator upstream of the scrubber. And
it is not necessarily any dramatic, in my opinion at least, cost advantage
459
-------
to go In with an all-scrubber system as opposed to a precipitator/
scrubber. I think Gary Rochelle (G. T. Rochelle, EPA) treated this in
one of his earlier papers. So again, it is not representative to all
cases, but a reasonable case.
Areas of Difference
In my mind, the major areas of difference that are really signi-
ficant were brought out by John Craig; they are in the areas of the sys-
tem being base-loaded with relatively constant coal. Again, of course,
there are many systems in the United States that approximate this condi-
tion but there are also many plants, particularly the older units, that
do not. They do have wide swing loads, as we heard for the Will County
unit. This is an area of difference and could affect the control of the
system. I think we should also mention that within several months,
Pete Wechselblatt referred to it informally, there will be a Duquesne
Light scrubber on the line. This will be very similar to the Japanese
Chemlco scrubber and will, I think, answer quite a few questions about
relating this technology to the U. S. situation. I believe that much
of the success did relate to the use of Japanese equipment and Japanese
construction personnel. I think there is no question (based on Shawnee
experience, and experience of chemical plants, refineries, and other chem-
ical operations) that the reliability and quality control of American
equipment leaves a lot to be desired. There Is no question about that
in anyone's mind, really. Regarding the blowdown situation, my own belief
is that It is not very relevent because the model proposed for sulfate
460
-------
scaling, which is what is at issue here, indicates that the degree of super-
saturation of calcium sulfate is the significant parameter. Now this has
not been proven, but it is the most widely held theory and I think that
Pete Wechselblatt is right in the calculation that there is only a minimum
of 4 percent difference in this degree of supersaturation in the scrubber
itself. So I am not worried about whether the pond leaches or not. Of
course, it presents a water pollution problem that is certainly serious
in the U. S. But, as far as operability is concerned, that may not be
significant. I think I should briefly mention the settling differences.
Based on the results presented yesterday in the sludge discussions, I
think it should be mentioned that probably what we are seeing here is
not any difference in sludge. What we are seeing is drying of the top
layer of sludge. Based on everything that I have seen and heard, I
would suspect that, in times of rain, this would probably end up reslurry-
ing again and forming the boggy type sludge materials that have plagued
other installations, like Will County. So in my mind, there is nothing
magical about the sludge. And I think we heard earlier that, given the
right weather contitions, people have walked on Will County Sludge.
Open Discussion
We are now open for discussions and anyone who wants to challenge
anyone else on the panel, go right ahead.
P. Wechselblatt -~ I am not picking points. I recall one fact which
has not been stated and that is that the sludge going to the pond contains
approximately 15 percent calcium sulfate and, therefore, 85 percent calcium
sulfite. And the significance of that is that calcium sulfate must be
461
-------
coming out of solution because 15 percent is too high a number for 863
collection. So there is calcium sulfate coming out of solution and this
is a demonstration that the solution is in fact over saturation.
H. W. Elder — Frank {Princiotta), I would like to question one
of your points. You say that the average sulfur in U. S. coals is about
2 percent, but the system does not operate on a national average sulfur
level, you know. It operates on what you have to burn. I think you will
have to agree that there is a significant difference in removing lOOOppm
and SOOOppm of flue gas. I think that the level certainly must influence
the liquid-to-gas ratios and other important design considerations. So
differences between the levels operated at Mitsui and those that we face
in the TVA system, for instance, are important differences.
F. T. Princiotta — Yes, and I agree. But 2 percent, in my opinion,
is not very far from many high sulfur coal problems, the average probably
being around 3 percent.
I. S. Shah (Combustion Equipment Associates, formerly with Chemico)
— I was part of the successful installation at Mitsui/Miike. Last year
we were talking about whether to use lime or limestone, and I think there
is only one system in the world that is operating successfully without mech-
anical problems of any kind and that happens to be Mitsui Aluminum. We
have a lot of installations with limestone, regardless of equipment prob-
lems or process problems, that are not still operating successfully. So
now here we see a successful operating unit. Instead of asking questions
about why it works, let us start asking how it works. And let us learn
something from that and make the technology better, rather than fighting
462
-------
with "My lime system works and my limestone system does not work." The
second question regards the saturation solubility. Everybody has talked
about solubility of sulfur/sulfite, in the presence of all the compo-
nents. Everybody is theorizing. It should be this because that is what
I have. TVA got some that they say is the sulfite/sulfate ratio. Maybe
they have a unique oxygen, unique coal, unique flyash. Maybe, then, no-
body else has it. We talked about the saturation solubility within the
scrubber just like Pete (Wechselblatt) said. It is 90; it must be almost
beyond supersaturation because the amount of bleed that you are taking out
from the scrubber loop is practically, I would say, within 5-10 percent
of the total recycled liquor. Now we talked about the ponds. The next
installation that is going to come on line and which is under design, I
think, is TVA's Widow's Creek.
Open - versus Closed-Loop Operation
Can I ask you a question, Bill (Elder)? Have you designed that pond
for no leaching, no overflow, or complete recycle loop? And if not, why?
H.W. Elder — Because we cannot control scaling if we operate in a
completely closed loop.
I.S. Shah — So you said open-loop is required for reliability of
operation.
H.W. Elder — Well, you will hear more about this during the Widow's
Creek discussion, but basically we are. . . .
I.S. Shah — I was asking a question. Is open-loop necessary for
reliable operation of the S02 systems or not? In my opinion it may be
necessary.
463
-------
H.W. Elder — I was trying to answer the question. The system does
have a horizontal mist eliminator and that mist eliminator is washed
with once-through water. The scrubber loop itself is closed-loop except,
of course, for the makeup required for evaporator of that that goes with
the solids. But the design of the mist eliminator loop itself is open-
loop. We would like to learn how to put that back into the scrubber sys-
tem. Based on some of the progress we have made at Shawnee, that may
happen. But right now, it is designed for the mist eliminator to have a
blowdown from it. You are right, we probably do not know exactly what
the sulfite/sulfate levels are. But we know what happens when it gets too
high: we have grown some mighty nice gypsum crystals in the mist elimina-
tor by trying to use saturated (or at least high sulfate level) liquor for
wash. So we know the effect even if we do not know the reasons.
Alex Weir, Jr. (Southern California Edison) — We have been able to
operate, in essence, a closed-loop system for about 3 months, and have
washed with lime slurry mist eliminators without any really serious prob-
lems.
Bob Sherwin (Bechtel Corporation) — There is one point that is not
clear to me. I would just like to ask what is being done with the precipi-
tator ash at the Mitsui aluminum plant. Does this go out to the lime pond?
F.T. Princiotta — I do not believe it does, but why not ask Pete (Wech-
selblatt)?
P. Wechselblatt — Well, I do not know where it goes. Mr. Sakanishi
would. But it does not go the lime pond because Mitsui is trying to
keep the calcium sulfite as pure as possible for gypsum.
464
-------
A. Weir — Okay, it does not get mixed.
P. Wechselblatt — No, it does not.
Joe Selemczi (Dravo Corporation) — I also had the pleasure of the
congenial hospitality of Mr. Sakanishi. There are two points I would like
to contribute to this discussion. One is that when we are talking about
sulfur content of coal, we have to look at the Btu value, which was somewhat
lower for Japanese coals than for the Eastern coals in the United States.
So that coal would calculate out to something like 2.5 to 2.7 percent sul-
fur based on Eastern coals. There is also a lot of varied thinking about
the sludge and why somebody can walk on the water of Gethsemane there.
Actually, what Mitsui has going into the pond is a lean slurry of calcium
sulfite sludge containing about 3 to 4 percent solids. You can compare
this with a horizontal, long settling tank. The heaviest fractions settle
out first and build up close to the discharge points. That is what you
can walk on; nobody walked out to the section of the pond that is farther
away from the discharge point. Now I can say this with confidence because
we have examined the sludge from Japan as well as many dozens of other
sludges. When you operate the thickener with 3-4-5 percent solids content
in the sludge, you are in the free settling zone. Correspondingly, your
discharge will behave entirely differently in your pond. When you have a
20 percent or higher solids content thickener underflow, settling 1s going
to be hindered. The discharge 1s not going to settle preferentially. You
are not going to have the larger particles settle out close to the discharg
point and the fine particles travel maybe a few hundred feet or maybe a few
thousand feet If you happen to have that kind of distance. These are the
465
-------
comments I wanted to make.
Ab Saleem (Peabody Engineering) — My comments are related to the L/G's
which are used and the S02 removal that can be obtained. I would like to
submit to the panelists and the audience: if you sprayed all that water
to an open shell with no internals you will get removal of 90 percent or
more. And to me the 300-mm system pressure drop is excessive and unnecessary.
F.T. Princiotta — Sounds like you may be selling a spray unit. I
might add that, although it was not brought out, there is some particulate
removal requirement in the venturi scrubbers. As I recall, it is approx-
imately 0.007 gr/scf of particulate outlet of that system. As far as I
know there is nothing in the United States that can match that right now.
So keep in mind pressure drop does buy us some particulate removal.
J. Ando (Chuo University) — Mr. Sakanishi wants me to explain some-
thing for him. There is a variation of the operating load. As I said,
the operating load is fairly uniform for the Mitsui aluminum plant. But
he has the experience that on a few occasions the operating load could
amount to nearly 50 percent of the usual operating load, but he could not
follow up the variation. It is not very often, so they operated very care-
fully and they have done it without any problem.
J. Craig — Let me go a little bit further then. It is not often
and for what time period did this occur?
J. Ando — (after conferring with Mr. Sakanishi) - Several times 1n
1 year.
466
-------
J. Craig — But not on a daily basis?
J. Ando — Just for one day he dropped, and then backed up.
J. Craig — For one day?
Question — I would like to ask Pete (Wechselblatt) a question. Pete,
how critical is water makeup at various locations in the scrubber and mist
eliminator? And if it is extremely critical, what would happen if that
loop were closed and how, when you talk about the pond's being supersatu-
rated, does it take a little bit of water to drop the supersaturation in
order for it not to scale?
f. Wechselblatt — Our newest jobs in this country are designed to
utilize fresh water makeup; that is, the liquid lost by evaporation. This
water is used preferentially for pump seals, for fan spray wash water
(if in fact there is a wet fan), and for mist eliminator wash water. Of
course there are differences. I am sure different equipment suppliers and
different designers have different rates at which they wash mist eliminators,
for example. But the rate set at Mitsui is such that the normal evapora-
tion loss of the system (and therefore that concomitant makeup) is satis-
factory to use all the fresh water makeup to wash the mist eliminator and
allow you to have approximately 50-75 gal/min left over for fan spray wash
water and 5 gal/min per pump seal. So in our view, the quantity of evap-
oration down to about 50 percent load is satisfactory to add the fresh
water to where we think it is most important.
Paul Cho (EPA Region V) — I would like to answer one of the questions
that John (Craig) brought up. That is, your worry about some of the plants
which are operated under a fluctuating load condition, I think that is
not a technical problem in itself. And secondly, people are concerned
467
-------
about open-loop or closed-loop. I think this is a little bit confusing
here without defining what percentage of makeup water we are talking
about: if operating loss itself were defined, you could see an absolute
closed-loop system. And I further suggest that a power plant in this
country has cooling water operations and there is some border proton water
available. I was wondering if anyone has looked into this area as makeup
water.
J. Craig — Well, the discussion really is the applicability of Mitsui
and what Mitsui demonstrates to the U.S. Regarding the ability to bypass
gas, are you saying you bypass a certain percentage of untreated flue gas
to the stack? If so, will the regulations in Japan allow you to do this?
P. Cho •— No, I think that the bypass I am talking about is that
instead of using one scrubber, probably you can use two. This is perfect
application in terms of your fluctuating load condition. If you have two
modules and bypass you can vary the stoichiometric ratio as well as the
amount of flue gas introduced into the scrubber.
J. Craig — Well, we are talking about probabilities versus what we
saw at Mitsui and the probability is an expensive gamble.
F.T. Princiotta — I think it is a good point though. To some extent,
John (Craig), would you not agree, that costs can help you buy reliability;
having extra scrubbers, I do not think one can argue, does improve reli-
ability.
H.W. Elder— I think the important point here on turndown though,
Frank (Princiotta), is that with variation in gas flow rate or variation
in sulfur level, the trick is to absorb a certain amount of sulfur in a
given volume of liquor and to precipitate part of it and accommodate some
as soluble solids until it gets out of the scrubber. And if the ratio of
468
-------
absorbed sulfur per unit volume changes for any reason, it can upset
this balance between solubility in the scrubber and, therefore, pre-
cipitation in the scrubber. And that is the key point, really.
A.V. Slack (TVA, Muscle Shoals) — There have been a great many
postulations here. I wish we had a little more in the way of hard data.
Pete (Wechselblatt), as I understand it, you consider that it does not
make any difference, as far as operation of the scrubber is concerned,
whether one operates with saturated liquor return or not. Is that a
correct interpretation of your remarks?
P. Wechselblatt — Yes.
A.V. Slack — Well, this does not agree at all with not only data
at TVA and EPA, but it does not agree with data in Europe. I have seen
a plant there in which the lime system is being operated closed-loop with
saturated return liquor, and there was scaling. And I cannot see much
difference in data, really, between your operation and the other. So I
think this is a postulation on your part. Now if you have pilot plant
data in which you have actually operated a saturated return liquor, I would
like to see it. I think, though, that at Mitsui the data that Gerry
McGlamery (TVA, Muscle Shoals) showed indicates or proves that the return
liquor is not saturated and until you do saturate it you cannot say that
you would not have trouble either in Japan or in this country under such
an operating mode.
P. Wechselblatt — In answer to that, I would point out the very small
quantity of liquid going to and from the pond. If the liquid coming back
from the pond is, for example, at 50 percent of saturation, this system
is a more closed system than if you operated a pilot plant with a thickener
469
-------
and underflow at 35 percent solids.
Pat Rapier (Burns & Rowe) — In continuance of the question that came
up regarding the 300-mm pressure drop as being excessive, I have some ex-
perience. It was many years ago, but it is directly related to this. If
you take out the mist eliminators; if you take out all the impingement
baffles in the collector, whatever it is; if you use, as a substitute,
a 316 stainless cyclonic collector; if you precede the collector with a
316 stainless fan and keep the tip speed of the fan blades at over 12,000
ft/min ( I believe that will correspond to a 4- to 6-inch pressure gain
through the fan); and then if you put the slurry in through the fan it-
self and let the fan do the contacting, you create an extremely high
collision rate between the particles in the air stream and with the water,
which indeed the fan itself can accelerate and atomize. We tried this
using a 30,000 ft^/min unit, which is much smaller than we are talking
about here, but the same type of operation could be done on the larger units.
It would scrub out about 100 percent of the dust particles that were available
down to the micron range, and down probably about as low as 0.25 micron.
The pH of the liq'uor was not regulated because in those days we were not
interested in trying to remove sulfur, but it was removing sulfur dioxide.
My recollection is that the pH of that liquor was coming out at about 5.5.
The circulation rate going in for 30,000 ft3/min scrubbing was 72 gal, a
large portion of which was recirculated. We were furnishing a moderate
amount of evaporation water for makeup* but the main portion was recircu-
lated.
P. Wechselblatt —• Commercially, a device such as that might be repre-
470
-------
sentative of a Typson disintegrator, which I believe is comercially avail-
able for particulate removal.
Del Ottmers (Radian) — I have a question to be directed to Mr Wech-
selblatt. As I understand it, you have a scrubber with a hold tank. You
are operating with two feeds, two scrubbers with a total L/G of 90. I
was interested in the supersaturation and how you control this. In that
regard, I wondered about the volume of the hold tank that you are using
and the size of the stream of slurry that you are taking to your pond.
P. Wechselblatt — Well, the size of the hold tank provides approx-
imately 30-min residence time, and the bleed from the system to the pond
I believe is 3 percent of the liquid rate. But again, it is the kind of
number I would rather give you after the meeting. It can be made available
to you; I just do not recall the number.
R.H. -Quig (Chemical) — Bill (Elder), I direct a question to you.
You and I have had a running commentary from time to time on this recycle
loop. I just want to put a couple of things in perspective. Do you essen-
tially feel that this problem of closed-loop saturation falls into two areas
of concern? One is that you are concerned that, if you have a saturated
return system, you will plug and scale the scrubber vessels regardless of
the different geometries or shapes or what they are - spray towers, ping
pong balls, venturi, whatever. Is that a major point of concern? The
second point is, and I agree with you for sure on this one, the concern
for overall pond management and the impact of water quality standards.
But could you first comment on your concern for plugging and scaling as
related to this problem of saturation. How fast do you plug and scale
your pilot units when you truly have saturated liquor 1n terms of sulfate
471
-------
coming back?
H.W. Elder — Well, you know, I can not say flatly that if it is
saturated it will scale. What I do say is that if you can not add any
fresh water in the system, it would certainly scale. Now it becomes
relative then, and the less saturated the recycle water the more fresh
water you can add. And that has to be in the right direction. Now the
other point, how long does the pilot plant operate before it scales up?
You know, we have scaled it up at 4 hours and we have used lime. And
we have run for 1000 hours with limestone without scaling it up.
R.H. Quig — Saturated?
H.W. Elder — Yes, well saturated to the extent that the purge from
the system is a slurry with 40 percent solids, 60 percent water. That
goes out and does not come back. That simulates what would happen if you
put it in a storage pond. But under those conditions, yes, we have oper-
ated without scaling for a continuous run of 1000 hours. We, I think, have
a pretty good understanding of the system, the chemistry, and the tech-
niques with limestone. I do not understand the lime system that well and
perhaps if I did I would be less concerned about some of these points that
I have raised. Again, it gets back to the reactivity of the absorbent.
I talked about that yesterday: the difference in stoichiometry and its
effect on scaling in the scrubber itself. Now what we are saying really
is that the reactivity changed. More calcium was available to react with
the $02 that was absorbed. Lime is a more reactive material than limestone.
Therefore, the amount of calcium available with the lime system is higher
than with limestone. I would expect the tendency to precipitate sulfite
472
-------
in the system 1s greater with lime. As I say, there are some questions
1n my own mind about the difference between lime and limestone chemistry
that sort of confound our conversation.
R.H. Quig — Well, yes, perhaps so. But I would like to refer to
this chart that you fellows prepared here. It has several points of data
you have grabbed, based on information from visitors to the Omuta system.
And that chart (which we do not really agree with, but just let us assume
at the moment that it is correct) shows a saturation condition of several
weeks. We did not plug and scale, and I know the unit was operating at
full load with the impact of the changing S02 concentrations due to the
different coals. Now your chart says in effect that the scrubber system
operated saturated for about 6.5 weeks, if we look at the chart, under a
saturated or supersaturated condition. And then your chart says that, for
the rest of the year, it operated at less than that. So I submit that if
your data are correct ( and it is not because you only have partial infor-
mation due to returning visitors) it operated for 6 weeks totally saturated.
I think that in itself has significance. We happen to know that it opera-
ted much longer than that. But giving you credit for you assumptions, 1t
still operated.
H.W. Elder — Well, of course, you know the obvious question. If 1t
operated a few weeks, why not a few months, or a few years? We would have
had a lot more confidence in the performance had it operated 6 months
rather than 6 weeks with saturated conditions.
R.H. Quig — Well, I would just like to conclude our position on that
issue: we have gone through ranges of saturation and we find no impact
there.
F.T. Princiotta — I think we a running out of time, and I am sure
473
-------
people are anxious for a break, so we would like to conclude discussions
on the panel. I would like to thank all the panel members.
474
-------
THE
TVA WIDOWS CREEK
LIMESTONE SCRUBBING FACILITY
PART I
FULL SCALE FACILITY
by
B . G . McKinney
Power Research Staff
Tennessee Valley Authority
Chattanooga, Tennessee
A. F. Little
Division of Chemical Development
Tennessee Valley Authority
Muscle Shoals, Alabama
J. A. Hudson
Division of Engineering Design
Tennessee Valley Authority
Knoxville, Tennessee
475
-------
THE
TVA WIDOWS CREEK
LIMESTONE SCRUBBING FACILITY
PART I
FULL SCALE FACILITY
BY
B. G. McKinney
Power Research Staff
Tennessee Valley Authority
Chattanooga, Tennessee
A. F. Little
Division of Chemical Development
Tennessee Valley Authority
Muscle Shoals, Alabama
J. A. Hudson
Division of Engineering Design
Tennessee Valley Authority
Knoocville, Tennessee
ABSTRACT
The design of a limestone slurry scrubbing system for TVA's Widows
Creek Unit 8 (550-MWj plant located in Northeast Alabama) is discussed. Basic
design premises on which the design is based are presented. Descriptions of
major components of the scrubbing facility are described.
Estimated capital and operating costs are contained in this paper.
The estimated capital cost of this installation is $^2,000,000 not including
a portion of the solids disposal costs. The estimated capital cost is $1*3,636,000
if the full initial scrubber effluent ponding costs are Included. The estimated
operating cost is approximately 2.9 mills per killowatt hour generated.
476
-------
THE
TVA WIDOWS CPEEK
LIMESTONE SCRUBBING FACILITY
PART I
FULL SCALE FACILITY
By
B. G. McKinney
Power Research Staff
Tennessee Valley Authority
Chattanooga, Tennessee
A. F. Little
Division of Chemical Development
Tennessee Valley Authority
Muscle Shoals, Alabama
J. A. Hudson
Division of Engineering Design
Tennessee Valley Authority
Knoxville, Tennessee
In mid-1970 TVA made a decision to install a full-scale
demonstration limestone scrubbing SOg removal system on generating Unit 8
at Widows Creek Steam Plant (in Northeast Alabama, near Chattanooga,
Tennessee). The primary objective is to work out design and operating
problems that affect both S02 removal efficiency and process reliability,
with emphasis on the latter.
Since late 1970 considerable small scale, pilot plant and
prototype developmental work has been done on the limestone scrubbing
process in an attemptto establish design parameters for the full scale
demonstration unit. The results of the developmental work are described
in a separate paper. A draft environmental statement covering the full-
scale demonstration S02 removal system and waste disposal pond was prepared
and issued on June 30, 1972. Following receipt and resolution of various
state and federal agencies' conanents, the final environmental statement
was issued January 15, 1973?
477
-------
Following finalization of the environmental statement, the project
authorization was submitted to and approved by TVA's Board of Directors in
February, 1973- Since engineering had commenced prior to the project author-
ization it enabled construction to begin in March, 1973- The engineering
for the project is about 50 percent complete. Equipment bids have been
received and contracts awarded on most of the long delivery equipment items.
Specifications and requisitions for other material and equipment are at
various stages of completion.
The mechanical completion date for the facility based on the present
project schedule is mid-1975• Pre-operational testing is scheduled to be con-
ducted in the third quarter of 1975.
An attempt is made in this paper to discuss the design of the
facility and the estimated capital and operating costs. The design areas
of the facility to be discussed are: basic design premises, limestone
handling and grinding system, scrubber system, and solids disposal system.
Figure 1 is a general plot plan of the Widows Creek Steam Plant
showing the general location of the scrubber facility, limestone storage,
handling and grinding facilities, and the solids disposal area with respect
to existing facilities.
Basic Design Premises
The following design premises were established for the scrubbing
facility during the planning stages of the project:
1, Coal analysis (as fired basis)
a. Ash content, 25$
b. Sulfur content, b.yf>
c. Moisture, 10$
d. Heating value, 10,000 Btu/lb
2. Capacity
a. Maximum power generation rate for Unit No. 8, 550-MW
b. Stack gas rate at capacity, 1,600,000 acfm at 280°F
(5,325,000 Ib/hr)
Based on a total of 33$ excess air including air heater leakage.
478
-------
NEW ASH DISPOSAL POND
EXISTING ASH DISPOSAL
AREA FOR UNITS T & 8
3OO KV SWYO
POND RECYCLE WATEfr
PUMPING STATION
POWERHOUSE
UNITS 7*9
POWERHOUSE
UNITS 1-6
RELOCATED
WIDOWS CREEK
CHANNEL
UNIT 8 SCRUBBING
SYSTEM SOLIDS
DISPOSAL POND
UNIT 8 SCRUBBING SYSTEM
RIVER WATER PUMPING STATION
LIMESTONE GRINDING AREA
LIMESTONE DEAD STORAGE AREA
GUNTERSVILLE RESERVOIR
FIGURE I
GENERAL PLOT PLAN
-------
3. Sulfur dioxide removal
a. Percent removal, 80
b. Inlet concentration, 3^40 ppm (wet basis); 37to ppm (dry basis)
c. Outlet concentration, 650 ppm (wet basis); 750 ppm (dry basis)
k. Particulate removal
a. Inlet particulate loading, 5.6 gr/scf (dry); 5.1 gr/scf (wet;
3.6 gr/acf (280°F)
b. Particulate level at scrubber exit,2 0.020 gr/acf (125°F saturated);
0.022 gr/scf (wet); 0.026 gr/scf (dry)
5. Stack gas reheat temperature, 175°F (50°F rise)
Limestone Handling and Grinding
A wet grinding system for limestone was chosen over dry grinding
because it is less expensive and does not produce a dust problem. Even if
dry grinding were used, the ground limestone probably would be slurried
before feeding to the scrubber.
A schematic drawing of the limestone handling and grinding system
is shown in Figure 2. The facility is designed for receiving limestone by
both rail and truck from the quarry. The limestone is conveyed from an
unloading hopper to either the live storage silo or the dead storage area.
Material will be reclaimed from the dead storage area as required to maintain
an adequate level in the live storage silo.
Limestone is conveyed from the live storage silo to a wet ball
mill where it is ground from the purchased size (3A by 0 in.) to the
desired size. The resulting slurry is pumped from the ball mill through
a classifier where the oversized particles are separated and recycled to
the ball mill. The product slurry (Uo# solids) from the classifier goes
to a surge tank from which it is pumped to the scrubbing area limestone
slurry storage tank.
Based on **• conventional ASME sampling technique.
480
-------
GROUND
LEVEL?
BELT
FEEDER
UNLOADING
a RECLAIM
HOPPER
TO FUTURE
SURGE HOPPER
FROM POND WATER
^RECYCLE PUMPS
CYCLONE
CLASSIFIERS,
00
TRANSFER
POINT
BELT
CONVEYORS
LIVE
STORAGE
SILO
'BELT
CONVEYOR
DEAD
STORAGE
BUCKET
ELEVATOR
CONVEYOR
MILL SLURRY SUMP
TANK a PUMPS
VIBRATING
FEEDERS
L.S. SLURRY SURGE
TANK a TRANSFER
PUMPS
TO L.S. SLURRY STORAGE
TANK (SCRUBBER AREA)
FIGURE 2
LIMESTONE HANDLING AND GRINDING SYSTEM
-------
The grinding system is designed to produce a ground particle
size of 90 percent minus 200 mesh at the design rate of 50 tons per hour
dry limestone. The 50 tons per hour limestone rate corresponds to a
limestone feed stoichiometry of 1.5 at full load and design conditions.
Pilot plant work indicates that little or no improvement in SO^ removal is
obtained by using limestone quantities above 1.5 of stoichiometric amounts.
The unloading and conveying facilities up to the live storage
silo are designed for a 900 tons per hour limestone rate. The live storage
silo has an effective capacity of 6400 tons. The unloading and storage
facilities up to and including the live storage silo are designed to
accomodate a future limestone scrubbing facility on Widows Creek No. 7
generating unit which has a maximum rated capacity of 575-MW. The dead
storage area is designed to accomodate 120,000 tons of limestone corres-
ponding to approximately 50 days capacity at maximum limestone requirements
for Units 7 and 8. The live storage silo will have about six days' capacity
initially and about three days if used to supply the maximum limestone
requirements for both Units 7 and 8. The design of the unloading and
conveying systems will permit the use of only day-shift personnel for the
limestone unloading and reclaim operations.
The belt conveyor and bucket elevator conveying limestone from
the live storage silo to the limestone surge hopper are designed for 200
tons per hour to accomodate both the Unit 8 scrubbing facility and a future
Unit 7 scrubbing facility. The limestone surge hopper has a 50 ton limestone
capacity and is designed only for Unit 8 as are all the limestone facilities
downstream of it.
A weigh feeder is used to feed limestone from the surge hopper to
the ball mill at controlled rates. The grinding system is a closed circuit
wet ball mill system with cyclone classifiers. The ball mill has a 11'-0"
inside diameter shell 20'-0" long and is equipped with a 1250 horsepower
motor. The ball mill and cyclone classifiers are equipped with rubber liners.
A rubber liner was selected for the ball mill rather than a steel liner
because it provides a longer wear life, less maintenance, and a lover noise
level.
The mill discharge slurry (about 65 percent solids) goes into the
primary compartment of a two compartment sump tank from which it Is pumped
to the cyclonic classifiers. Sufficient pond water is added to the primary
sump compartment BO that the overflow or product slurry from the classifiers is
a 1*0 percent solids slurry. The overflow from the cyclonic classifiers goes
to the secondary compartment of the mill slurry sump. The product slurry
overflows from the secondary compartment of the mill slurry sump tank into
a limestone product slurry surge tank which has about five minutes capacity.
482
-------
At reduced rates slurry underflows from the secondary compartment into the
primary compartment of the mill slurry sump tank for recirculation to the
classifiers to maintain a constant feed rate to the classifiers. Three 50
percent capacity cyclonic classifiers are provided with one being an
installed spare.
Underflow from the classifiers containing oversized particles is
fed back to the ball mill for regrinding. Sufficient pond water is also
added to the ball mill to maintain a 65 percent solids slurry in the mill
discharge.
The product slurry (lv&f> solids) is pumped from the product slurry
surge tank to a limestone slurry storage tank located adjacent to the
scrubber facility about 300 yards from the grinding facilities. The slurry
storage tank has a capacity of about 181,000 gallons corresponding to about
eight hours storage capacity based on the «M>vimny limestone slurry requirement.
Scrubber System Design
A simplified flow diagram of the scrubber system is depicted in
Figure 3> The scrubbing system consists of four identical trains each
handling 25 percent of the flue gas from Unit 8. The flow diagram shows
the major components typical for each train and those components common to
the four trains.
The flue gas after exiting the existing electrostatic precipitators
passes through a fan, venturi, liquid-gas separation chamber, absorber,
entrainment separator, and a reheater for each train into an existing stack.
Flue gas bypass ducts are provided around each scrubber train to prevent an
undue amount of boiler downtime because of scrubbing system malfunctions,
particularly during the initial shakedown operation of the system. A plenum
connecting the inlet ducts to the four fans is provided to permit operation
of up to T5# load with one of the scrubbing trains out of service.
Limestone slurry, pond water recycle, and make-up river water are
added to an absorber circulation tank. Slurry is pumped from the absorber
circulation tank into the absorber and drains back into the tank. The
recycle pond water is regulated to maintain the solids concentration level,
about 10# by weight, in the absorber circulation tank. The absorber circul-
ation tank overflows into a venturi circulation tank from which slurry is
pumped to the venturi and drains from the separation chamber back into the
circulation tank. The solids concentration in the venturi circulation tank
is dependent upon the inlet flue gas flyash loading but will normally run
about 15 percent.
483
-------
CO
•p
RE HEATER-^ r-
""""* I FROM STEAM / 1
[ B.cao tf t
\_ I
n TO ASH ,*_! 4-
UbrUoAL rTJNU
, , . — •• >,
•^
1 VENTURI \
-TO STACK PLENUM f /
v.
_ FROM ESP X \ '
^J •*• j^ J
^3— .
,-. FROM B.C a D TRAIN VENTURI
^ VENTURI CIRC. TANKS ^| n lAfcutt
^v^ TO SETTLING
'"^ POND \y-
ENTRAPMENT
SEPARATOR
r
TkT.itirA-/r
J \
1
{
r
i *•
i
i
\
* fc
TRAINS < )—
B.caD ^J
1
ABSORBER
L
i
ir^ '
tir
1
1
1
^
CIRC. ABSORBER CIRC.
PUMPS TANK a PUMPS
i
i ,
SLURRY PUMP
SEAL WATER
, HEADER
/ ^t ^.
™ I'D
rRAINS<*— 1
r
i
TO P-«
TRAINS/ f—
B.C8D J ^__
__i— ~s RIVFR
— «T>*Jl!l£i j
WATER PUMPS
FROM POND WATER A-N
RECYCLE PUMPS ^
FROM LS. SLURRY /->
TRANSFER PUMPS ^
1 i
L S. SLURRY
STORAGE TANK
EFFLUENT SLURRY
SURGE TANK 8 PUMPS
FEED PUMPS
FIGURE 3
SCRUBBER SYSTEM FLOW DIAGRAM
-------
The venturi circulation tanks on each of the four trains overflows
into a common effluent slurry surge tank from which the effluent is pumped
to a settling pond. Supernatant pond water is recycled to the absorber
circulation tanks and to the limestone grinding system.
Four flue gas fans are provided, one per train, to supply the
static pressure for the system losses. The design capacity for each of the
fans is 1»OO,000 acfm (280°F) with a test block capacity of U60,000 acfm
(280°F). The fans will be equipped with induction motors and variable speed
fluid drives to permit flue gas flow regulation. The static pressure will
be sufficient to convert the Unit 8 furnace from a forced draft furnace to
a balanced draft furnace. The fans are a double width, double inlet and
radial tip blade design with an A-2U2 steel casing and impeller with high
strength alloy wear plates. After investigating and evaluating the advantages
and disadvantages of locating the fans upstream of the Venturis versus
downstream of the reheaters, it was decided to locate them upstream of the
Venturis. It was concluded that the dry fly ash erosion problems would be
less severe than corrosion and solids build-up problems vith the fans located
downstream of the reheaters.
The venturi is used to cool, saturate, and remove fly ash from the
flue gas. It is designed to operate with a differential pressure of five
inches of water at full load conditions; however, higher differential pressures
may be used at reduced loads. The venturi is a rectangular throat design
equipped with a motor operated variable throat mechanism. The venturi
housing material is 3l6L stainless steel and the convergent section is lined
with castable silicon carbide (2 inches thick). The throat damper blades
are lined with fired alumina brick. The venturi throat is 23*-0" wide
(inside), and has an effective throat depth of 28" (2-lU" sections). The
expected SOg removal in the venturi is about 10 percent.
Constant speed elastomer lined centrifugal slurry pumps will be
used to pump slurry from the venturi circulation tank to the venturi. A
total of six venturi circulation pumps are provided, one operating pump per
train and a common spare for each two trains. The pumps are designed to supply
a liquid circulation rate to the Venturis corresponding to an L/G of 10 gal/
Mcf of saturated flue gas at the design flue gas rate. At reduced flue gas
rates the L/G will be higher.
The venturi circulation tanks have a capacity of ^0,000 gallons
and are l8'-0" in diameter and 25'-0" high. The retention time in the tanks
is about 12 minutes based on the venturi circulation rate. The tanks will be
provided with agitators to maintain solids suspension.
485
-------
The absorber is a grid type absorption tower. It is designed
for a superficial gas velocity of approximately 12 feet per second at the
design gas rate. Its dimensions are 30 feet vide, l6 feet deep, and $k
feet high. The absorber has five grids, 65-70 percent open area, spaced
either four or five feet apart. A final grid selection has not been made,
but preliminary plans are for them to be of the floor grating type and made
of 316 stainless steel (top and bottom grids) and fiberglass reinforced
polyester (intermediate grids). The expected SOg removal efficiency for the
multi-grid absorber is about 70 percent and the pressure drop is expected
to be about 2 inches of water with clean grids. The corresponding overall
expected S02 removal including the venturi is about 80 percent. The slurry
is distributed in the absorbers using 316 stainless steel distribution
headers equipped with low pressure drop (5 psi) spray nozzles. The use of
turning vanes in the venturi-absorber sump to give better gas distribution
in the absorber is being studied.
The absorber circulation pumps will be elastomer lined centrifugal
slurry pumps designed to supply a maximum slurry rate to the absorbers
corresponding to an L/G of 60 gal/Mcf at the design gas rate. A total of
10 pumps are provided, two operating pumps per train and a common spare for
each two trains. The pumps are equipped with variable speed drives for
varying the slurry rate to the absorbers.
The absorber circulation tanks have a capacity of 1^8,000 gallons
and are 33 feet in diameter and 25 feet high. The retention time in the
tanks based on the maximum absorber circulation rate is seven minutes.
Agitators are provided in the tanks to maintain solids suspension.
The entrainment separator is a chevron vane type (k pass) located
vertically in a horizontal shell. The vanes will be 316 stainless steel,
12 inches deep and spaced on 1^ inch centers. The superficial gas velocity
in the entrainment separator is approximately nine feet per second at the
design gas rate. The face of the eliminator blades are washed continuously
with once through river water at a rate of 1 gpm/ft. The wash water drains
from the entrainment separators to the ash disposal pond. The environmental
statement may be referred to for the water quality considerations for the
ash pond overflow which will consist of ash sluice water from all Widows
Creek generating units and the entrainment separator wash water from the
Unit 8 scrubbing facility.
486
-------
The venturi liquid-gas separation chambers, absorber shells,
venturi circulation tanks, and absorber circulation tanks will be
constructed of Corten A steel and coated or lined inside. Although a
final coating or lining selection has not been made, the primary con-
siderations are for a poXyurethane elastomer base coating (50 mils)
for the separation chamber and absorber shell, and flakeglass reinforced
polyester resin (60-8© mils) for the tanks.
Indirect tubular exchangers (one pass) are used for reheating
the flue gas from the saturation temperature of about 125°F to 175°P
(50°F reheat) to desaturate and provide buoyancy for the gas. Steam at
approximately 500 psig and 650°F from Unit 8 is used as the heating medium.
The material of construction to be used for the tubes has not been defined.
Corrosion specimen are being evaluated in pilot plant tests to determine
the material of construction for the reheater tubes.
Slurry piping four inches and larger in diameter will be soft
rubber-lined carbon steel. Smaller slurry piping will be FVC-coated
aluminum or 3l6L stainless steel. A decision has not yet been made on
the type of slurry block valves but they will likely be either an
elastomer lined plug type valve or a knife gate type valve.
Steam operated soot blowers will be provided at the locations
in which solids deposition is expected to occur based on research and
development experience. The locations where soot blowers are to be
provided include the inlet ducts to the venturi, the elbow between the
absorber and the entrainment separator, and the reheater section.
A perspective drawing of the scrubber area is given in Figure k
which shows the general arrangement of the equipment. The controls for
the system are located in the existing Unit 8 control room to minimize
operating personnel requirements. Adequate controls and instrumentation
are being provided to minimize operating personnel, to monitor and control
the system performance, and to provide the necessary data for studying system
variables.
Provisions have been made in the design of the scrubber system to
make the system as flexible as practical. The absorber is being designed
so that it can be operated as a multi-grid, turbulent contact or spray
tower absorber. The base case for design is the multi-grid absorber which
has been discussed above.
487
-------
RE HEATER
fNTRAlNMENT SEPARATOR
ABSORBER CIRCULATION
TANKS
-ABSORBER CIRCULlTIOtl
PUMPS
f UJE G*S DUCT
POWER HOUSE
EKISTING ESP'S
FIGURE 4
SCRUBBER AREA PERSPECTIVE VIEW
-------
Provisions axe made in the absorber design BO that it can readily
be converted to a mobile bed type absorber should it prove desirable in the
future to obtain improved S02 removal. The variable speed absorber circu-
lation pumps permit the use of lower L/G's required for mobile bed absorbers..
Also the flue gas fans are designed with sufficient static pressure head
to accomodate the higher pressure drops associated vith mobile bed scrubbers.
Another back-up which is being investigated is the changing of
the limestone scrubbing process to a double-alkali process. Investigations
have been conducted which indicate that the conversion to a double-alkali
process is feasible. Pilot plant tests are planned in late fiscal year
1973 to further define a double-alkali process for the Widows Creek Unit 8
scrubber facility. Hopefully pilot plant testing will show that a double-
alkali system is a possible and practical back-up process for the limestone
scrubbing process being installed.
Solids Disposal System Design
Ponding was chosen as the most feasible method of disposing of
the waste by-product solids from the Widows Creek scrubbing facility. The
by-product solids consisting of flyash, reaction products (hydrates of
calcium sulfite and sulfate) and unreacted limestone are pumped to the
pond as a 15-16 percent solids slurry. The supernatant liquor from the
pond is recycled to the limestone grinding area and the scrubber area.
Thickeners could be used to concentrate the -urge slurry and
reduce the pumping rates to and from the ponds. However, the savings in
pumping costs do not Justify the additional capital investment for
thickeners. Also there is doubt as to how effective thickeners would be
in concentrating the purge slurry because of the very poor settling
characteristics of the reactant products in the slurry.
The poor settling characteristics of the solids in the bleed
slurry to the pond result in large pond volume requirements for solids
disposal. Based on the TVA pilot plant data, the weight percent water in
the settled solids after settling for extensive periods (up to 2to days)
ranges from 57 to 66. For planning the solids disposal pond requirements
for the scrubbing facility 60 percent water in the settled solids was
assumed. This corresponds to a pond volume requirement of about 1.^ cubic
yards per thousand pounds of solids pumped to the disposal pond. Based
on the expected average coal analyses with regard to sulfur ezid ash content
the solids disposal rate in terms of settled solids corresponds to about
150 cubic yards per hour at full load conditions.
489
-------
The initial scrubber effluent solids disposal pond vill have an
effective volumetric capacity of U.5 MM cubic yards. The area of the pond
including dikes is about 100 acres and the dikes average about 30 feet in
height. The life of the pond is estimated to be about seven years based
on projected capacity factors. The pond dikes are designed so that they
can be raised 10 feet to increase the pond capacity to 5.8 MM cubic yards
(increase of 1.3 MM cubic yards). The total estimated pond volume required
over the remaining life of the planl^ based on the above settled solids
density data and the projected capacity factors, is 9«3 MM cubic yards and
thus an additional storage volume of approximately 3.5 MM cubic yards would
be required for which no provisions are made. Due to the developmental
nature of the project and the possibility of developing means of enhancing
the settling characteristics of the solids, it was not deemed justifiable
to provide for the total estimated pond requirements for the remaining life of
the Unit 8 generating unit, TVA is continuing research and development work
on means to increase the compaction of the settled solids to reduce the pond
volume requirements for solids disposal.
Estimated Capital Costs
In Table I, the estimated capital costs for the wet limestone
scrubbing facility at TVA's Widows Creek Unit 8 are summarized. These costs
are based on cost estimates prepared by TVA, which are based on detailed
layouts and firm prices for much of the major equipment. The estimate is
based on construction beginning in March, 1973, and with completion in
August, 1975. The estimated direct cost for the scrubber facility is
$22,36*0,000 not including a portion of the solids disposal costs. If the
additional cost of solids disposal is included, the direct capital cost is
$23,61*0,000. The total field construction costs are $31.400,000 and
$33,018,000 respectively. The total project costs are $1*2,000,000 and $^3>
636,000 respectively. The cost estimates do not include the cost of land
which is associated with the pond or the other parts of the scrubber facility.
Estimated Operating Cost
The estimated operating costs are summarized in Table II. The
estimated total operating cost is approximately 2.9 mills per kilowatt
hour generated. For an investor-owned utility, the capital charges portion
of the operating cost would be higher for the same total investment due to
differences in cost of money, income taxes, and so forth. The estimated
total operating cost for an investor-owned utility, based on the same para-
meters (capital investment, operating life, capacity factor, etc.) except for
capital charges, is approximately 3.5 mills per kilowatt hour generated based
on capital charges of 15 percent of the total investment per year.
490
-------
TABLE I WidovB Creek Unit 8
Limestone Wet Scrubber Facility
Capital Cost Estimate Summary
Item Estimated Cost, M$
Grading, landscaping, yard drainage, surfacing ITT
Roads, sidewalks, bridges 608
Power house modifications 35
Electrical equipment building 100
Ductwork 2000
Fans 960
Reheaters and soot blowers (includes steam & condensate piping) 1525
Railroad facilities 320
Ball mill building 200
Limestone grinding facilities 255
Limestone conveying facilities 919
Limestone storage facilities 562
Mobile equipment for limestone handling 155
Scrubber area foundations 350
Scrubber area steel structures 1110
Scrubbers 1280
Pumps T65
Tanks (including linings and agitators) 390
Entrainment separators 150
Piping 2^T1
Elevator 80
Painting 30
Instruments and controls 550
Electrical work 2821
Electrical transmission plant 1526
Cranes and hoists 92
Solids disposal area T6T
Construction facilities 2162
Total direct cost subtotal 22,360
Field general expense 26TO
Allowance for shakedown modifications 2000
Contingencies
Total field construction subtotal
491
-------
Table I, Continued
Item Estimated Coat, M$
Indirect Costs
Engineering design
Managers office - Office of Engineering Design & Construction
Office of Power (research, development, & coordination)
Initial limestone supply and preoperational testing
Employee compensation benefits
Administrative and general expenses
Interest during construction
Other
Total indirect cost subtotal 10,600
Total project cost excluding supplemental pond costs 42,000
Additional supplemental pond cost allocations
Direct costs
Field general expenses
Contingencies
Indirect costs
Subtotal 1,636
Total project cost including supplemental pond costs 43,636
Notes
1. The total project cost with additional pond cost allocations from
separate authorization contains a total of about $2.8 million
($2.2 million direct cost) for scrubber effluent solids disposal
pond. The pond will have an effective capacity of 4.5 million
cubic yards which will last about seven years at projected Unit 8
load factors based on present expected settling characteristics.
The pond dikes are designed so that they can be raised in elevation
to obtain an additional storage capacity of 1.3 million cubic yards
at an estimated cost of about $1.6 million ($1.2 million direct
costs).
2. The estimate includes about $2.1 million ($1.2 million direct
costs) for solids disposal over and above the pond costs.
3- The estimate includes about $7.0 million ($4.0 million direct
costs) for limestone handling, storage and grinding facilities.
492
-------
TABLE II
Estimated Annual Operating Cost
TVA Widows Creek Unit 8
Limestone Scrubbing System
Estimated Annual Cost,
Item
Direct costs
Raw material - limestone^/
Conversion costs . .
Operating labor & supervision.!/
Utilities
Analyses .
Maintenance^/
Subtotal conversion costs
Subtotal direct costs
Limestone ,
Processings
299
Scrubber
Area
821
105
986
30
1390
2511
3332
Solids
Disposal
27
5
10
172
172
Total
821
159
1033
50
2$
3803
Indirect costs
Capital charges^/
TOO
3170
770
3/
4
Overhead
Plant, 20# conversion cost 60
Administrative, 1056 operating labor 3
Subtotal indirect costs 763
Total annual operating cost
Operating cost, mills/kWh generated
Notes
I/ Based on capacity factor of 65 percent (3135 x 10° kWh/year generated).
2/ Limestone handling, grinding and storage facilities.
273,600 tons limestone at $3 per ton delivered.
Operating labor and supervision at $6 per hour.
Annual maintenance costs are based on 4.0, 6.0, and 3*0 percent of total
field construction of $5.6 million, $23.1 million, and $U.3 million,
respectively,for limestone, scrubber and solids disposal areas.
6/ Annual capital charges based on 10 percent of total investment (25-year
life) except capital charges on $2.8 million for pond costs in the solids
disposal area are based on 20 percent (7-year life).
493
-------
It should be noted the effect which capital investment has on
the operating costs. The capital investment is approximately 50$ of the
operating costs. The second largest contribution to the operating cost
is maintenance and the third largest contribution is utilities.
Also, the capacity factor (65 percent) used in this estimate
may be high for a unit approximately 10 years old. A decrease in the
capacity factor would increase the operating cost substantially.
494
-------
THE
TVA WIDOWS CREEK
LIMESTONE SCRUBBING FACILITY
PART II
PILOT-PLANT AND PROTOTYPE OPERATING EXPERIENCE
by
J. J. Schultz, T. M. Kelso, J. L. Graham, and J. K. Metcalfe
Division of Chemical Development
Tennessee Valley Authority
Muscle Shoals, Alabama
N. D. Moore
Power Research Staff
Tennessee Valley Authority
Chattanooga, Tennessee
495
-------
THE
TVA WIDOWS CREEK
LIMESTONE SCRUBBING FACILITY
PART II
PILOT-PLANT AMD PROTOTYPE OPERATING EXPERIENCE
By
J. J. Schultz, T. M. Kelso, J. L. Graham, and J. K. Metcalfe
Division of Chemical Development
Tennessee Valley Authority
Muscle Shoals, Alabama
And
N. D. Moore
Power Research Staff
Tennessee Valley Authority
Chattanooga, Tennessee
ABSTRACT
The Tennessee Valley Authority (TVA) has "been engaged in a
limestone - wet-scrubbing pilot-plant test program for the past 2 years.
The purpose of the program is to provide information for the design and
operation of a demonstration-scale (550 mw) scrubber system now under
construction at TVA's Widows Creek power station in northeast Alabama.
The pilot—plant testing program consisted of two phases. The
objective of the first phase (l year) was to select a scrubber type for
the Widows Creek project. Four different types of scrubber systems
were evaluated. As a result of this evaluation, a venturi (for particu-
late removal) followed by a multigrid scrubber was selected as a basis
for the design of the Widows Creek system.
The objective of the second phase was to closely duplicate
the Widows Creek design in the pilot plant and evaluate the following.
• Process equipment and construction materials.
• Mist eliminator system.
• Scrubber operation during simulated changes in the
operation of the boiler.
• The effect of the scrubber grid configuration and the
temperature of the scrubbing slurry on S02 removal
efficiency.
496
-------
The following conclusions can te drawn from the pilot—plant
studies.
• Limestone wet scrubbing provides an effective method for
removing S02 from boiler flue gas.
• An S02 removal efficiency of JCffr can "be expected using
the multigrid type of scrubber with a scrubbing liquor
recirculation rate (L/G) of 50 to 60 gallons per 1000
cubic feet of gas and a Ca:S02 mole ratio of 1.5. The
efficiency can he increased to a"bout 9df> if a two-stage
mobile—"bed scrubber is used.
• Techniques for adequate long-term disposal of the spent
solids from the scrubbing slurry (including fly ash)
have not "been developed. Initially, ponding of the spent
solids will "be used for the Widows Creek system.
• The long-term operating reliability of the limestone - wet-
scrubbing system will be largely determined by proper design
and selection of suitable construction materials to cope
with the erosive and plugging characteristics inherent in
slurry scrubbing.
497
-------
THE
TVA WIDOWS CREEK
LIMESTONE SCRUBBING FACILITY
PART II
PILOT-PLANT AND PROTOTYPE OPERATING EXPERIENCE
By
J. J. Schultz, T. M, Kelso, J. L. Graham, and J. K. Met calfe
Division of Chemical Development
Tennessee Valley Authority
Muscle Shoals, Alabama
And
N. D. Moore
Power Research Staff
Tennessee Valley Authority
Chattanooga, Tennessee
Introduction
TVA has been engaged in a limestone — wet—scrubbing pilot—plant
test program for the past 2 years. The program is designed to provide
information for the design and operation of the full-scale (550 mw) lime-
stone wet scrubber which is being installed on unit 8 at the Widows Creek
generating station. The information obtained from the pilots-plant program
includes the following.
• Selection of scrubber type.
• Selection of optimum operating conditions.
• Evaluation of scrubber performance during simulated changes
in boiler operation;
• Identification of suitable process equipment and construction
materials.
The program was performed in pilot plants located at TVA's
Colbert Power Plant near the fertilizer research and development
facilities at Muscle Shoals, Alabama.
The first pilot plant, a temporary unit located near the unit
No. 3 boiler, was operated for 1 year before it was dismantled to make
room for the installation of an'electrostatic precipitator. The evalua-
tion of four basic scrubber types was done in this pilot plant. The
498
-------
second one is a permanent unit located near the unit 5 boiler, designed
to closely duplicate the system planned for Widows Creek. It has been
operated for 1 year and is still being operated. Except for changes
in the scrubber type, the same general mode of operation was used in
T>oth pilot plants; a discussion of the work performed in both follows.
General Description of Process
Flue gas for the pilot plant is obtained from a pulverized
coal-fired boiler in which coal containing about 4$ sulfur and 15$ ash
is burned. The boiler is normally operated with 20$ excess air. The
flue gas contains about 6$ oxygen (including air heater leakage) and
2800 ppm S02. It can be withdrawn from the boiler outlet duct on
either side of an electrostatic precipitator; fly ash loadings are
about 4 gr/scf or 0.1 gr/scf. The temperature of the gas entering the
pilot plant is about 300° F.
In the pilot plant the flue gas is scrubbed with a limestone
slurry containing 15$ undissolved solids (combined limestone reaction
products and fly ash). This level of solids is optimum for maintaining
the stability of the slurry pH and S02 removal efficiency. Spent
scrubbing slurry is continuously bled to a series of settling tanks
where the solids settle. The settled solids contain from 55 to 65$
water "by weight. The supernatant liquor from the settling tanks is
returned to the scrubbing loop to maintain the solids content at 15$.
The liquor system is operated on a closed—loop basis; no liquor is
overflowed to waste. This mode of operation is planned for the
Widows Creek system. The only liquor lost from the system that must
be replaced with fresh makeup water is (l) the water required to
humidify the hot (300°F) flue gas and (2) the free and chemically
combined water in the settled solids. This loss is equivalent to about
1 gpm of makeup water in the 1-^aegawatt pilot plant. With this mode of
operation, 'the liquor phase of the slurry system contains about 10g/l
total dissolved solids and is supersaturated with calcium and sulfur
salts.
Pulverized limestone (normally 75$ minus 200 mesh) is mixed
with water (part of the makeup) to give a 60$ solids slurry. Slurry
of this concentration has excellent storage characteristics and requires
only mild agitation to maintain homogeneity. The fresh slurry is fed
to the scrubbing slurry retention tank at the required rate (usually
1. 5 moles of Ca per mole of S02 in the inlet flue gas).
499
-------
vtork. Performed in the First Pilot
Plant (Temporary Unit at Boiler No. 3)
The following types of scrubbers were tested in the unit 3
pilot plant.
• Ventri-Rod1 followed by a packed—bed Crossflow scrubber.
• Ventri—Rod/open spray tower.
• Mobile-bed scrubber (TCA)2.
• Ventri-Rod followed by a multigrid scrubber.
The objective of the evaluation was to determine which system
tfould give the "best combination of S02 removal efficiency and operational
reliability. Pilot—plant operating experience and data for the first
three scrubber types were presented last year at the Second International
Lime/Limestone Wet Scrubbing Symposium ("Scrubber-Type Comparison/1 by
T. M. Kelso, P. C. Williamson, and J. J. Schultz).
The present paper will cover the testing and development of
the Ventri—Rod multigrid scrubber system which ultimately led to its
selection for the commercial-size installation at Widows Creek.
All of the scrubbers tested gave acceptable S02 removal
efficiencies at approximately the same liquid to gas ratio (L/G,
gallons/1000 ft3). The TCA and multigrid scrubbers could be operated
at higher gas velocities than the Crossflow or spray tower (12 ft/a vs.
8 ft/s). These data are shown in Table I.
The Crossflow scrubber was abandoned for slurry scrubbing
because of rapid plugging of the packed bed. The Ventri—Rod spray
tower, although nonplugging, did not give acceptable S02 removal
efficiency at allowable pressure drop (less than 15 in H20) if the gas
velocity exceeded 5 ft/s. The lack of sufficient space for fitting to
the Widows Creek unit precluded the use of a' low-velocity scrubber.
The TCA and multigrid scrubbers were nonplugging and reliable.
The TCA scrubber (containing two or three 1—foot beds of spheres) had
excellent gas—liquid contacting characteristics and gave the best S02
removal efficiency (90$). The major disadvantage of the TCA scrubber
was erosion of the packing spheres (polypropylene). A 2.% weight loss
of the spheres occurred during 1000 hours of operation and failures
due to puncturing began to occur. Universal Oil Products (manufacturer
of the TCA scrubber) has a program underway to develop a more erosion-
resistant sphere material; thermoplastic rubber has shown promise.
A type of venturi scrubber manufactured by Environeering, Inc.
2 Turbulent Contact Absorber manufactured by Universal Oil Products
Company. 500
-------
TABLE I
in
O
Typical Operating Characteristics
Scrubber system
Ventri-Kod/Crossflow
Ventri— Rod/ spray tower
Three stage mobile bed (TCA)
Ventri-Rod/multigrid
of Scrubber Systems Tested
L/G,
gal/1000 ft3
55
65
55
50
in TVA Pilot
AP,
in HP0
10
15
10
8
Plant
Velocity,
ft/s
5
5
12
12
S02
removal
efficiency,
80
75
90
75
-------
The multigrid scrubber, while not as effective in removing
S02 as the TCA, has the advantage of having few internal parts. The
pressure drop across the multigrid scrubber (containing five 65$ open
grids spaced on 5—foot intervals) is only about 2 inches of HeO;
consequently the particulate (fly ash) removal efficiency is poor. To
remove the fly ash, a venturi (Ventri—Rod) was installed upstream of the
multigrid scrubber. The Ventri—Rod, operated at an L/G of 10 and a
pressure drop of 5 inches of H20, was effective in removing the fly
ash from the flue gas and also providing a sharp wet-dry junction in
the inlet gas duct.
Selection of Scrubber for Widows Creek
As a result of this evaluation, the venturi/multigrid scrubber
system was selected as the basis for the design of the Widows Creek
system. A Ventri—Rod type of venturi was used in the pilot plant; however,
a standard venturi with similar operating characteristics may be used in
the Widows Creek installation. The expected S02 removal efficiency using
the multigrid scrubber is about 75$ (TOO ppm S02 in the gas exhausted to
the atmosphere). If higher S02 removal efficiency is required, the
scrubber can be converted to a two- or three—stage TCA type. A new pilot
plant was "built to continue the evaluation of this system.
It should be noted that a venturi or some other type of scrubber
effective for particulate removal is necessary at Widows Creek because
the present electrostatic precipitator is not very efficient. The ash
collected in the precipitator is slurried to the ash disposal pond and
that collected in the scrubber goes to the scrubber solids disposal system.
Work Performed in the Second Pilot
Plant (Permanent Unit at Boiler No. 5)
A perspective view and flow diagram of the unit 5 pilot plant
are shown in Figures 1 and 2. The gas capacity is equivalent to 1
megawatt (2^00 acfm at 120°F), and the design closely duplicates that
planned for the Widows Creek installation. The objective of the new
pilot—plant test program was to evaluate the following:
• Mist eliminator design and operation.
• The effect of grid configuration and slurry temperature
on S02 removal efficiency.
• Scrubber performance during simulated changes in power
plant operation.
• Process equipment and construction materials.
502
-------
FIGURE 1
Limestone -- Wet—Scrubbing Pilot Plant at Colbert Power Plant (Permanent Unit)
503
-------
Ul
o
WATER
FLUE
GAS
I
I
SLURRY
DISTRIBUTOR
MULTIGRID
SCRUBBER
FORCED DRAFT FANS
VENTRI-ROD
ELEMENT
PULVERIZED
LIMESTONE
SLURRY FROM F-4
«
F-2
LIMESTONE SLURRY
FEED TANK
F-3
SLURRY
RETENTION
TANK
MIST ELIMINATOR
TO SEWER
MAKE-UP WATER
SCRUBBED
GAS
DIRECT
FIRED
REHEATER
F-4
SLURRY
RETENTION
TANK
(OPTIONAL)
SETTLING
TANKS
FIGURE 2
Flow Diagram; Limestone - "We-t>-Scrubbing Pilots-Plant (Permanent Unit)
-------
A discussion of these items follows.
Mist Eliminator Design; Removal of entrained scrubbing slurry,
and consequently particulate, from the gas leaving the scrubber was a
major problem identified during the unit 3 pilot-plant program. Commer-
cially available mist elimination systems installed in the vertical
scrubber housings were inadequate. The most severe problem with this
type of system was plugging of the mist eliminator element by accumulation
of solids from the entrained slurry. The plugging occurred "because there
was not a sufficient amount of fresh water available to continuously wash
the mist eliminator when the scrubber slurry system was operated on a
closed-loop basis. A total of only 1 gpm of fresh water was required
for makeup in the 1-megawatt pilot plant, whereas most vendors recommend
a continuous wash of about 2 gpm per square foot of mist eliminator face
area. This is equivalent to about 8 gpm in the 1-megawatt pilot plant,
or 8 times the required makeup rate. In some tests a blend of 3 parts
clarified liquor from the slurry settling tanks and 1 part fresh water
was used to wash the mist eliminator intermittently (twice an hour) at
the vendor's recommended rate of 2 gpm per ft2. This method was not
satisfactory because of scaling (calcium sulfite and sulfate) caused
by the supersaturated clarified liquor (Figure 3).
An effective mist elimination system was ultimately developed
and is now in use. It consists of a horizontal housing connected to the
outlet of the vertical scrubber tower. The superficial gas velocity
through the horizontal housing is 9 ft/s. A V-pass chevron element is
installed in a vertical position in the housing. The element (Type 316
stainless steel blades spaced on 1—1/2—in centers) is 12 inches deep and
has a face area of k ft2. It is washed continously with fresh water
which is discarded to the fly ash disposal pond separate from the lime-
stone slurry disposal system. Fly ash pond effluent water, used in a
few tests for washing, was as effective as fresh water. The particulate
loading of the gases exhausted to the atmosphere averaged 0.04 gr/scf.
Tests were made to determine the minimum effective wash rate
with this configuration. A wash rate of 0.25 gpm/ft2 (l gpm total) was
effective, but was marginal in regard to uniform coverage of the chevron
element. At this rate (1 gpm total) the wash water was drained into
the slurry system as makeup. Maximum effectiveness was obtained with a
wash rate of 1 gpm/ft2 (4 gpm total). Since this was approximately 4
times the makeup required for the slurry system, the excess water was
routed to the fly ash disposal pond rather than to the limestone slurry
disposal system. The composition of the discarded wash water was
monitored in the pilot plant at various wash rates to determine the
environmental impact caused by this mode of operation (Table II). Based
on these data a wash rate of 1 gpm/ft2 is planned for the Widows Creek
installation. The excess wash water not required for makeup will be
discarded to the ash disposal pond.
505
-------
-
-
,HARD SCALE
SOFT SOLIDS
FIGURE 3
Mist Eliminator After 270 Hours of Operation
(Washed Intermittently from Both Sides wifrh a Blend of Clarified Settling Tank Liquor and Fresh Water)
-------
TABLE n
Composition of Effluent Water from Once-Through Mist Eliminator Wash System*
Limestone - Wet-Scrubbing Pilot Plant
Composition, mg/1 at the following wash rates
en
o
••J
Wash rate, ^pm/ft2
Component
Total dissolved solids
Aluminum
Ammonia nitrogen
Arsenic
Barium
Beryllium
Cadmium
Calcium
Chloride
Chromium.
Copper
Cyanide
Iron
Lead
Magnesium
Manganese
Mercury
Nickel
Phosphate
Potassium
Selenium
Silver
Sodium
Sulfatf
Zinc
PH
1.0
1000
<0.2
0.21
0.002
<0.1
<0.01
0.0042
220
2k
<0.05
0.02
<0.01
5.8
0.033
6.5
0.16
<0.0002
<0.05
0.11
2.2
0.012
<0.01
8.1
TOO
0.0?
3.1
0.50
1900
<0.2
0.25
0.002
<0.1
<0.01
0.0013
440
hO
<0.05
0.02
<0.01
0.07
0.011
8.2
0.14
<0.0002
<0.05
0.03
3-0
0.024
<0.01
8.8
1000
0.02
7.2
0.25
2200
<0.2
0.34
0.010
<0,1
<0. 01
0.0031
430
120
<0.05
0.03
<0.01
5.5
0.016
18
0.37
<0. 0002
<0.05
0.03
2.6
—
<0.01
11
1200
0.14
2.7
0.25b
1900
0.5
1.9
0.001
<0.1
<0.01
0.0015
370
70
-------
Grid Configuration; The S02 removal efficiency obtained in the
rrrultigrid scrubber is dependent upon the gas-liquid contact imparted by
the grids. Tests were made to determine the effect on S02 removal effi-
ciency and pressure drop caused by varying the number and open area of
the grids. When the scrubber was operated with 5 wire mesh—type grids
(75% open area, 1/8-inch—diameter wire with 7/8—inch—square openings)
at an L/G of 50, the S02 removal efficiency and pressure drop averaged
70# and 1. 7 inches of H20, respectively. Only 2 to 3 percentage points
increase in S02 removal efficiency was obtained when the number of
grids was increased to 10. The pressure drop increased only slightly
(less than 0. 5 inch H20). The relationship of the number of grids to
S02 removal efficiency is shown in Figure k.
Another series of tests was made to determine the effect of
less grid open area and consequently more contact surface on the pressure
drop and S02 removal efficiency. Combinations of five and six stainless
steel wire mesh grids containing 40 and 6Cff> open area were tested. The
hOf> open grids were made of 0.075-inch-diameter wire with 1/8-inch-square
openings. The 6Of> open grids were made of 0.09*1—inch-diameter wire with
1/2— by 5/l6-inch openings. The wire used in these grids was smaller
and more closely spaced than that previously used to provide more surface
(target) area for gas and slurry contact. In some tests, additional
grid surface area was obtained (with respect to the downward flow of
slurry) "by installing a kCff> open grid on a 65-degree angle "between two
horizontal grids. A summary of the grid combinations tested and
operating data obtained are shown in Table III.
There was no significant improvement (compared with previous
tests with 75% open area grids) in the S02 removal efficiency or
pressure drop across the scrubber when the open area was reduced to
60$. The removal efficiency and pressure drop increased to 78$ and
3.7 inches of H20, respectively, when the bottom four 6of> open grids
were replaced with 40# open area grids. The S02 removal efficiency
was further increased to Q2f> by increasing the L/G from 50 to 60. The
addition of a sixth grid (^C# open, slanted between the bottom two
grids) did not cause any significant change. Grids with kQJ> open area,
although better for S02 removal, are not practical because of the
tendency of the grid to collect trash that would normally pass through
a grid having 6Cff> or more open area.
On the basis of this evaluation, five grids containing
approximately 65% open area will be used in the Widows Creek installa-
tion. The construction of the grids may be similar to floor grating
(5/8- by 3-3/l|-inch-rectangular openings) made of reinforced plastic or
Type 316 stainless steel. The pressure drop and S02 removal efficiency
obtained in the pilot plant with this type of grid construction was
essentially the same as that obtained with the standard wire mesh grids.
508
-------
fc
Ul
o
u.
CO jjr
^^^
fT -J
0 §
^1. '^^^
o
llJ 2
O UJ
< o:
z ^
o iii O
*° O ^
Ul ?
CL
Ul
CO
Ul
££
0
«t
3
2
1
n c
1 1 I 1 I '
rv/^
: xr :
>^ S02 REMOVAL EFFICIENCY
.S WAS 70% WITH FIVE GRIDS
S 1 1 1 1 f
67 8 9 IO
NUMBER OF GRIDS (75 7o OPEN AREA)
FIGURE 4
Effect of the Mumber of Grids on SO^ Removal Efficiency
Limestone - Wet-Scrubbing Pilot Plant
fl
-------
TABLE III
(Jl
o
Evaluation of Grids in the
Multigrid Scrubber*
Limestone — Wet-Scrubbing Pilot Plant
Grids
Distribution13
Test No.
1
2
2
k
5
Total
5
5
5
6
6
No.
5
1
^
1
1
k
1
l
h
1
Open area, $
60
60
to
60
60
to
to (Slanted)
60
to
to (Slanted)
L/G
50
50
60
50
60
SO g removal
APj in H^O efficiency, #
1-7 70
3-7 78
k.Q 82
4.1 77
4.U 84
* All tests made at 12 ft/s superficial gas velocity.
° Grid distribution listed in order starting from top of scrubber.
-------
Slurry Temperature; During the course of a year's operation
of the pilot plant, the S02 removal efficiency varied as much as 10$
(Figure 5). This change in efficiency is largely attributed to the
change in temperature of the scrubbing slurry (about 25°F, from 100°F
in winter to 125°F in summer) caused by ambient temperature fluctuations.
The lower vapor pressure of S02 over the scrubbing slurry at the lower
temperatures presumably gave better S02 removal during cool weather.
The feasibility of removing heat from the scrubbing slurry
to obtain satisfactory S02 removal efficiency was tested in the
pilot plant. A Teflon-tube heat exchanger manufactured by DuPont was
submerged in the multigrid slurry retention tank and river water (57°F)
passed through the tubes at a rate of 10 gpm. The heat exchanger was
effective in reducing the temperature of the slurry from the normal
120° to 125°F and maintaining it at 100° F. About 5056 of the heat load
for the cooling was accounted for in the cooling water discharged from
the heat exchanger; the remainder was assumed to be lost by convection.
An overall heat transfer coefficient (U) of 25 was obtained. This method
of increasing the S02 removal efficiency is not practical because of
the large heat transfer area and high water rate required, which would
be even higher in a larger unit because a smaller percentage of the
cooling would be accomplished by convection. Also, the temperature of
the cooling water obtained from the river in summer would be about 80°F,
reducing the AT significantly. An equivalent increase (L/G of about 20)
in the pumping rate of uncooled slurry to the scrubber would increase
the S02 removal efficiency about as much as cooling the slurry to 100° F.
Power Plant Simulation Tests; A series of tests was made in
the pilot plant to determine whether the limestone — wet-scrubbing process
could te operated and controlled over a range of simulated power plant
operation conditions. The flow rate and S02 concentration of the gas,
the recirculation rate of the scrubbing slurry, and the Ca:S02 mole
ratio were varied to simulate scheduled and unscheduled changes in boiler
operation. A diagram of the pilot-plant configuration used for this
series of* tests is shown in Figure 6.
Six tests were run consecutively, all on a closed-loop "basis.
The retention time and the solids content of the scrubbing slurry were
9 minutes (based on a pumping rate of 105 gpm to the multigrid scrubber)
and ll$, respectively. The S02 removal efficiency ranged from ¥$ to
82#; 44$ when the Ca:S02 mole ratio was reduced to 0.8, and Q2f> when
the gas velocity through the multigrid scrubber was reduced to 8 ft/s
and the Ca:S02 mole ratio and L/G were 1.5 and 77 > respectively.
Operation of the pilot plant was routine and no difficulty was
encountered in maintaining control of the process. The S02 removal
511
-------
IOO
90
o
z
y eo
o
ti-
ll.
UJ
in
—« «s
KJ UJ
(VI
O
o 70
60
50
10
L/G = 5O AT A GAS VELOCITY OF 8 FT. /SEC
JL
15
20
25
30
35
SLURRY RATE TO SCRUBBER,GPM/FT.2
OF SCRUBBER CROSS SECTION
FIGURE 5
Effect of Slurry Temperature on SOp Removal Efficiency
4C
-------
SCRUBBED
CAS
WATCH
MtST ELIMINATOR
SLUKRY
DISTRIBUTOR"
MUtTl«fll|> ~
SCRUBBER\
(BORIDSt 1--
Ul
«>
CO
FORCED DRAFT FANS
(FROM CYLINDERS)
VENTRI-ROO
ELEMENT
PULVERIZED
UME3TOHC
SLURRY FROM F-4
'TO SEWER
MAKE-UP MMTER
r-z
UMESTOHE SLURRY
FCfO TAHK
F-3 F-4 SETTLING
SLURRY RETENTION SLURRY RETENTION TANKS
TANK TANK
C9MIMUTC RETENTION TIME 1
FIGURE 6
DIRECT FIRED
REHEATER
Flow Diagram: Limestone — Wet-Scrubbing Pilot—Plant Configuration Used During Power Plant Simulation Tests
-------
efficiency fluctuated with changes in L/G and in the gas flow (simulating
a 15 and 35$ change in load). The pH of the scrubbing slurry and S02
removal efficiency declined rapidly when the flow of fresh limestone was
stopped to simulate a pump failure. However, operation returned to
normal within 4 hours after the flow of fresh limestone was restored. A
summary of the test conditions, S02 removal efficiency, and pH of the
scrubbing slurry for each test is shown in Figures 7 through 12.
Slurry Settling and Compaction; The slurry disposal pond
planned for the Widows Creek installation represents a significant
portion of the project cost. The initial size and estimated life
planned for the pond depend upon the settling and compaction characteristics
of the spent slurry. In the pilot-plant tests, initial settling occurred
rapidly (within 2k hours) to produce a clear liquor, but little further
settling took place even over extended periods of time. A summary of the
long—term settling data is shown in Table IV. The poor settling and
compaction are attributed to the presence of small flat calcium sulfite
crystals which normally constitute about 90$ of the calcium—sulfur
reaction products. Only about 10$ of the reacted sulfur forms gypsum,
which would be a more desirable form "because the crystals are large and
settle well.
Attempts were made to convert (oxidize) the calcium sulfite
to calcium sulfate (gypsum) by treating the spent slurry with excess
oxygen (10 times stoichiometric). These tests were unsuccessful. In
other tests oxidation of the slurry was promoted when the flue gas was
mixed with ambient air to give a 1:1 volume ratio "before the scrubber.
It would not "be practical to do this in a commercial—size unit because
of the large volume of air required, but the method was used to determine
the effect of increased oxidation on the settling and compaction charac-
teristics of the slurry. The flue gas—air blend contained iy/o oxygen
and about 1100 ppm S02. In one short test oxidation was increased to
the extent that 90$ of the reacted sulfur was in the form, of gypsum. No
quantitative settling data were obtained during this test, but compaction
of the solids appeared much improved over that of less oxidized slurry.
In a longer test (9 days) only 53$ of the reacted sulfur formed gypsum.
There was no significant improvement in the short— and long—term settling
characteristics of this slurry (53$ oxidized). These data are shown in
Figure 13 and Table V.
Optimum Operating Conditions: Various modes of operation
and test conditions were evaluated during the pilot—plant program to
determine the best combination for S02 removal and ease and reliability .
of operation. The following conditions are considered optimum and are
being used for the planning and design of the Widows Creek system.
• Pressure drop across the venturi (or Ventri—Rod)
for fly ash removal 5 in H20
• L/G to venturi (slurry) 10 gal/1000 ft3
514
-------
O >
S O
UJ Z
a: uj
Sfc
UJ
tn
in
80
75
I
a.
BOILER SIMULATION TEST NO. I
UNSCHEDULED SHORT TERM DECREASE OF LOAD
BASE CONDITIONS
GAS FLOW
r REDUCED 15%
-H*
BASE CONDITIONS
GAS FLOW. 2100ACFM AT I20°F
Ca/SO2 MOLE RATIO:
L/G ro
L/G TO MULTIGRID;50
L/G TO V-R;I2
L/G TO M-G:59
SLURRY TO SCRUBBER
F-3 SLURRY FROM SCRUBBER
23456
HOURS
FIGURE 7
The Effect of a 15% Reduction in the Gas Flow on the S0g
8
Removal Efficiency and pH of the Scrubbing Slurry
-------
o
'£
uj
BOILER SIMULATION TEST N0.2
UNSCHEDULED SHORT TERM INCREASE IN LOAD
75
-BASE CONDITIONS
GAS FLOW INCREASED 15%
BASE CONDITIONS -
TO
6
- GAS FLOW:2100 ACFM
—Co/S02'. 15
. L/G TO V-RMO
L/G TO M-G.'SO
Ca/S02:1.3
F-4 SLURRY TO SCRUBBER
F-3 SLURRY FROM SCRUBBER
I
23456
HOURS
FIGURE 8
The Effect of a iyf> Increase in the Gas Flow on the S0g Bemoval
Efficiency and pH of the Scrubbing Slurry
8
-------
o >
SE "
ij Z
-------
BOILER SIMULATION TEST N0.4
00
S02 REMOVAL
EFFICIENCY, %
Q.
75
70
6
5
SCHEDULED SHORT TERM DECREASE IN LOAD
(FLOWS WERE ADJUSTED TO MAINTAIN A Ca/SOo MOLE RATIO OF 1.5 AND
AN L/G OF 10 AND 50 TO THE V-R AND M-G)
^S**-r^ ,_ r— X'^ BASE ~"
GAS FLOW REDUCED 35% , CONDITIONS
^r *H ^1
^ BASE CONDITIONS
F-4 SLURRY TO SCRUBBER
— —
1 ii i i i 1 i 1
123456789 10
HOURS
FIGURE 10
The_Effe_ct__of a 35% Decrease in the Gas Flow on the S0g
Removal Efficiency and pH of the Scrubbing Slurry
-------
§".
O >
2 u
(O
en
to
-
u.
U.
UJ
80
70
60
50
40
BOILER SIMULATION TEST NO.5
UNSCHEDULED AND SCHEDULED INCREASES IN SO2 LOAD TO SCRUBBER
-BASE CONDITIONS
4000 PPM SO2 Ca/S02'-l.5
-7IOO PPM S02, Ca/S02: O.8
•BASE CONDITIONS
F-4 SLURRY TO SCRUBBER
F-3 SLURRY FROM SCRUBBER
BASE CONDITIONS
2100 ACFM
2800 PPM S02
Ca/S02:i.50
L/G TO V-R:IO
L/G TO M-G:50
8 12 16 20 24
HOURS
FIGURE 11
28
32
36
40
44
The Effect of Q02 Concentration and Ca:SOg Mole Ratio on the
Removal Efficiency and pH of the Scrubbing Slurry
-------
o >
2 £
UJ Z
0* UJ
U1
KJ
O
85
SO
75
7O
x
ex
BOILER SIMULATION TEST NO.6
SCHEDULED DECREASES IN LOAD (OVERNIGHT) USING A CONSTANT
PUMPING RATE TO SCRUBBER
GAS FLOW REDUCED 35%
L/G TO V-RM5
L/G TO M-G=77
BASE CONDITIONS
DAYTIME
GAS FLOW.-2IOOACFM
L/G TO V-R:IO
L/G TO M-G'50
OVERNIGHT
Ca/S02 MOLE RATIO: 1.5
F-4 SLURRY TO SCRUBBER
F-3 SLURRY FROM SCRUBBER
16
32
48
64
HOURS
SO
96
112
128
FIGURE 12
The Effect of Reducing the Gas Flow 35% and Maintaining a Constant Pumping Rate
to the Scrubber on the S03 Removal Efficiency and pH of the Scrubbing Slurry
-------
TABLE IV
Long-Term Settling and Compaction Data
Limestone - Wet-Scrubbing Pilot Plant
Percent
water in settled
Limestone Percent solids Volume percent after solids after
source d in slurry at gO days 60 days 160 days 240 days 300 day? Days
Test and grind start of test Solids Liquid Solids Liquid Solids Liquid Solids Liquid Solids Liquid |o55160 240300
lb A 13.7 33 67 32 68 - - 32 68 32 68 67 66 - 66 66
b
in
to
2
c
Ub
B
C
D
12.9
ik.O
15.0
31
-
29
69
-
U
30
2k
—
70
76
—
27
23
—
73
77
— —
27
23
29
73
77
71
66
-
60
66
57
60
63 -
57
— —
63
57
60
a
Includes fly ash removed from the flue gas.
Flue gas contained a fly asli loading of 2 to 5 grains per standard cubic foot.
c Flue gas contained a fly ash loading of 0.1 to 0.5 grain per standard cubic foot.
d A = The Stoneroan, Inc. (Tiftonia), Chattanooga, Tennessee, Jk% minus 200 mesh.
B = The Stoneman, Inc. (Tiftonia), Chattanooga, Tennessee, 8856 minus 325 mesh.
C = Longview Lime Company, Birmingham, Alabama, 75$ minus' 200 mesh.
D = Fredonia Valley Ojiarries, Inc., Fredonia, Kentucky, 795& minus 200 mesh.
-------
NJ
KJ
te iv/v/
uj
l-
u.
**
oc
o
•=>
3 75
||SO
CO tig.
i_ i^
ujw
o
£ 25
0.
Ld
O
> 0
1 I I
i I 1 1 1
TESTS MADE USING A 1.50 Co/S02 MOLE RATIO,
LONGVJEW LIMESTONE (75% -200 MESH), AND
15% UNDISSOLVED SOLIDS
—
o— ^^-ltl--^.^><
r"
IN THE SCRUBBING SLURRY
Jto>*
^j^^ -f^" ^^
^N> °^
O USING NORMAL FLUE GAS,
260O PPM S02 AND 5%
10% OXIDATION
_
• USING 5O/50 BLEND OF
AMBIENT AIR, 1120 PPM
53% OXIDATION
1 1 1
1 2 3
02
—
FLUE GAS AND
S02 AND 13% 02
: 1 I II
456789
DAYS OF PILOT PLANT OPERATION
FIGURE 13
Effect of Increased Oxidation on Settling of Solids from Scrubbing Slurry
-------
TABLE V
Effect of Oxidation on the Settling and Compaction Characteristics
of Spent Scrubbing Slurry — Limestone - Wet-Scrubbing Pilot Plant
Ul
NJ
u>
Test
1
(I0ff> oxidation)
2
(53% oxidation)
Percent solids
in spent slurry
at start of test
15.0
16.7
Percent water
in settled
Volume percent after solids after
kO days 120 days Days
Solids Liquid Solids Liquic- 40 120
23.5 76.5 23.1 76.9 57 57
27.0 73.0 26.0 74.0 53 53
NOTE: The percent oxidation is defined as the percent of the total reacted sulfur in the slurry existing as calcium
sulfate (gypsum).
-------
• Gas velocity through multigrid scrubber
(Superficial, based on outlet conditions)
• L/G to multigrid scrubber (slurry)
• Number and spacing of grids in multigrid
scrubber
• Open area of grids
• Solids content of scrubbing slurry (total
undissolved, including fly ash)
• Retention time in delay tank of slurry fed
to multigrid scrubber (based on pumping rate)
• Retention time in delay tank of slurry fed
to venturi
• Ca:S02 mole ratio (based on inlet S02
concentration)-
• Limestone particle size
• Solids content of fresh limestone slurry
• Mist eliminator wash rate (fresh water on a
once—through basis)
12 ft/s
50-60 gal/1000 ft3
5 on 5— ft centers
60-65*
1936
7 minutes
Not critical
1.5
-£00 mesh
gpm/ft2
Operation of Prototype Scrubber
Tests were made in one of the three EPA prototype (10 mw) scrubbers
(designed "by Bechtel Corporation) located at TW's Shawnee Power Plant near
Paducah, Kentucky. Over a period of 2 months, tests were made which were
designed to simulate previous operation in the 1-megawatt pilot plant at
Colbert and to provide scale-up and design information for the Widows Creek
installation. The TCA scrubber was used; the packing spheres were removed
and 5 grids were installed at approximately 1»—foot intervals. The slurry
system was modified to closely duplicate the closed—loop mode of operation
used in the pilot plant. A diagram of the prototype plant is shown in
Figure 14. Initial operation was suspended after kQ hours of operation
because the chevron mist eliminator, located in the vertical scrubber
housing, became plugged with solids from entrained scrubbing slurry. This
problem was anticipated based on pilot—plant experience.
The design or the prototype system precluded the installation
of a horizontal mist eliminator housing at the outlet of the scrubber
to duplicate the system developed in the pilot plant. Instead, a multiple
venturi tray device (FlexiTray, manufactured by Koch Engineering Co., Inc.)
was installed in the vertical scrubber tower between the slurry distributor
and mist eliminator. The FlexiTray was designed to provide an interface
between the closed—loop scrubbing slurry system and a once—through wash
524
-------
10
ui
DIRECT FJREO REHEATER
SCRUBBED GAS
CHEVRON MIST ELIMINATOR
FREOON1A
LIMESTONE
MULT 16 RID
SCRUBBER
(MODIFIED TCA)
UNDERFLOW TO
DISPOSAL POND
SLURRY
RETENTION
TANK
CLARIRED
LIQUOR
HOLD TANK
THICKENER
FIGURE
Flow Diagram - Modified TCA Prototype Scrubber at Shawnee Paver Plant
-------
system for the mist eliminator. Fresh water, fed across the top of tne
FlexiTray at a rate of about 60 gpm, was effective in keeping the top
of the tray and the mist eliminator clean; the water was discarded.
Blasting the "bottom of the tray with steam (150 psig) once per shift
for a"bout 1 minute was effective in preventing accumulation of solids
in this region. A diagram of the prototype scrubber and FlexiTray
installation is shown in Figure 15.
When the scrubber was operated at a gas velocity of 12 ft/s
(design rate for Widows Creek at full load), carryover of scrubbing slurry
into the open-loop FlexiTray wash system was excessive (30 gpm). Continuous
operation could not 'be maintained at the high gas flow rate because (l) the
maximum slurry purge rate to the disposal pond during closed—loop operation
is only about 1 gpm per megawatt (10 gpm total) and (2) slurry lost to
the FlexiTray wash system was routed to the fly ash disposal pond and could
not be returned to the closed-loop slurry system. Excessive carryover of
scrubbing slurry did not occur in the pilot plant when the scrubber was
operated at a gas velocity of 12 to 13 ft/s. These data are shown in
Figure 16. The reason for the excessive entrainment of slurry in the
prototype scrubber is not known but may "be caused by the smaller ratio
of scrubber wall area to gas volume as the size of the scrubber increases.
The relationship between velocity and entrainment for the larger Widows
Creek scru'bber is not predictable.
Carryover of scrubbing slurry into the FlexiTray wash system
was negligible when the gas velocity was reduced to 8 ft/s (two-thirds
of Widows Creek design). This velocity was selected for a long—term
test designed to compare the operating characteristics of the prototype
system with the Colbert pilot plant. Scrubbing slurry containing 15$
undissolved solids was fed to the multigrid scrubber at an L/G of 50.
The hot flue gas was humidified with slurry fed to the inlet duct at a
rate of JO to 50 gpm. The retention time of the scrubbing slurry in the
delay tank was 7 minutes based on the pumping rate to the multigrid.
The tap of the FlexiTray was irrigated with fresh water on a once—through
basis at a rate of 45 to 60 gpm. The 'bottom of the tray was blasted with
steam (150 psig) once per shift.
Crushed limestone (1-1/2—inch) was supplied by Predonia Valley
Quarries, Inc., Fredonia, Kentucky, and pulverized by TVA at the Shawnee
site. The limestone is relatively soft and easily pulverized to produce
a large fraction (95$) of minus 325-mesh material, of which about 50$ -is
minus 8 microns. The pulverized limestone was fed as a 60$ slurry into the
scrubber system at a rate intended to maintain a Ca:S02 mole ratio of
1. 5 based on the inlet S02 concentration; however, chemical analysis of
the solids from the spent scrubbing slurry taken near the end of the
test indicated a Ca:S02 mole ratio equivalent to 3.0. The magnetic
flowmeter used to measure the fresh limestone feed rate was calibrated
at the beginning and end of the test, and the indicated flow rate at the
end of the test was only about one—half of the actual rate.
526
-------
DIRECT FIRED REHEATER
ui
K>
FLEXITRAY
FLEXITRAY WASH
WATER TO SEWER
MULT 16 RID
SCRUBBER
FREOON.A --^ DISPOSAL POND
SLURRY
RETENTION
TANK
CLARIRED
LIQUOR
HOLD TANK
THICKENER
FIGURE 15
Flow Diagram — Modified TCA Prototype Scrubber with FlexiTray
-------
Q_
O
or
o
en
o
UJ
UJ
9.0
7.0
5.0
4.0
3.0
2.0
1.0
0.5
PILOT PLANT
I-MW
PROTOTYPE
PLANT
10-MW
01
0
5 10 15
VELOCITY, FT./SECOND
FIGURE 16
Effect of Gas Velocity on Entrainment
of Scrubbing Slurry
20
528
-------
The S02 removal efficiency gradually increased from about
70$ in the beginning of the test to 80$ at the end. The pressure drop
across the system (including the wet-dry junction, scrubber, FlexiTray,
and mist eliminator) increased from 3.5 to 6.0 inches of H20 during
the course of the test (19 days). The undissolved and dissolved solids
contents of the scrubbing slurry -were steady at about 14 and 0.5$>
respectively. The pH of the slurry fed to the scrubber ranged from
5.8 to 6.0. The trends in S02 removal efficiency, pH and solids content
of the slurry, and pressure drop are shown in Figure 17.
From 10 to 2056 (5-10 gpm) of the FlexiTray wash water fell
through the tray into the multigrid scrubber. The fall-through water
exceeded the closed—loop water "balance based on a water content of 50$
in the discarded spent slurry. The extra water (5-10 gpm) was removed
from the system with the dilute thickener underflow (about 25$ solids)
which was discarded to the disposal pond. No clarified liquor was
returned to the scrubbing system from the disposal pond.
Operation of the scrubbing system was routine and control of
the system was maintained without difficulty. However, the performance
of the thickener was marginal, yielding a more dilute underflow than
expected. The dilute underflow provided a convenient means of purging
the excess water from the system caused by leaking of the FlexiTray
wash water into the slurry system.
After the long-term test was completed, and before the plant
was shut down, a series of short>-term screening tests was made designed
to determine the effects of L/G, gas velocity, and Ca:S02 mole ratio on
the S02 removal efficiency. These data are shown in Table VT. In all
tests, the S02 removal efficiency increased with increased L/G and with
increased gas velocity (over the range of 8 to 10.4 ft/s). The effect
of the Ca:S02 mole ratio on the S02 removal efficiency could not be
identified. This is attributed to the high mole ratio due to the error
in the flowmeter. These data agree closely with data obtained in the
pilot plant while operating under similar conditions (except for the
high Ca:S02 mole ratio).
Operation of the prototype plant was discontinued after
completing the screening tests. The plant had Tseen operated continuously
for 27 days. Following the shutdown the plant was inspected. Heavy
deposits of scale (mostly calcium sulfite and sulfate) had accumulated
on the grids and walls of the scrubber. A discussion of the inspection
follows.
Wet-Dry Junction; The flue gas entered the scrubber through
a UO-inch-square horizontal duct. Scrubbing slurry (about 30-50 gpm)
was sprayed into the duct (90 degrees to the gas flow) h feet from the
entrance to the scrubber. The slurry spray was effective in reducing
the temperature of the flue gas from 300°F to about 125° to 150° F before
entering the rubber-lined scrubber, thus affording protection to the
liner. About 70$ of the cross-section axea of the duct was plugged with
an accumulation of fly ash and solids from the scrubbing slurry. The
accumulation extended about 2 feet upstream from the point of slurry
529
-------
U Ul
e 5
«u-
o u.
to w
90
80
TO
60
i 1 1 1 i
I 234 5 6 789 IO II 12 13 14 15 16 17 18 13
DAYS
FIGURE I?
Prototype Multigrid Scrubber—Trends in S0g Removal, pH and Solids Content of the
Scrubbing Slurry, and Pressure Drop Across System
-------
TABLE VI
Screening Tests, Prototype Limestone - Wet-Scrubbing Pilot Plant at Snavnee Stean Plant
Flue gas
(Maltigrid Scrubber
Veloc ity
Flov rate, through AP
Test Duration, acfm at scrubber, Ca:SOj> nole ratio sc:
So. hr 120 °F ft/s
Ul
LO
WC-4A
HC-5A
WC-6A
WC-7A
WC-BA
WC-9A
HC-10
WC-U
WC-12
8
4J2
8
3
8
4
4
24
64
a Combined pressure
14,720
14,720
14,720
14,720
19,170
19,170
19,170
14,720
14,720
drop across
8.0
8.0
8.0
8.0
10.4
10.4
10.4
8,0
8.0
wet-dry
Indicated
1.5
1-5
1-5
1-5
1.5
1.5
1.5
1.2
1.2
Junction and
o i o*.+<*4 >w m
Actual i:
3-0
3.0 1
3-0
3-0
3-0
3-0
3-0.
2.4
2.4
grid section
across
rubber, Ito
a Hj£a i
3.8
.4-3.5
3-0
2.7
5-6
6-3
6-5
4.4
3-9
of scrubber
vith Five Grids)
PlexiTrav
Once-through wash vater
tray, From tr&yc
SPP
60
60
60
60
60
60
60
60
60
.
gpn Solids, $>
48 0.10
56 0.10
48 0.10
46
85 3-6
85
86
50 0.7
45 1.1
Fall through, £ P
12 2.5
4 2.4
12 2.4
14 2.4
3.4
3-5
3-4
10 2.6
15 2-6
Scrubbing slurry
L/C
70
50
35
25
25
35
50
50
25
5-9
5-9
5-9
6.0
5-8
5.8
5-9
5-9
5-9
Temp,
•F
99
-
-
99
97
94
105
103
110
Solids,
14
14
14
15
14
14
14
13
14
soa
removal,
88
75
76
72
79
86
90
86
75
Conbined pressure drop across FlexiTray and chevron mist eliminator.
DOTE: Slurry retention time v&3 7.2 minutes, based on lU,720 aefn at 120*F and an L/G of 50.
-------
injection. The soot blower (operated once per shift with air at 90 psig)
was not effective in keeping this area clean. No accumulation of solids
occurred downstream from the point of slurry injection.
Grids: The scrubber contained five Type Jl6L stainless steel
grid assemblies spaced at approximately 4—foot intervals. The open area of
the wire mesh (1—1/8—inch—square openings with 0. 148-inch—diameter wire) was
75%f but the effective open area of the grid assembly was only about 50$
because of the grid support and retaining structures.
A coating of hard scale formed on all of the grids and about 40
to 50$ of the open area was obstructed. The scale deposits were identified
as predominately calcium sulfite, although small amounts of calcium sulfate
and calcium carbonate were present.
The bottom grid was scaled the most. The scale was about 1/8-inch
thick on the grid wires and retaining assemblies. Nodules of scale (about
3/4—inch diameter) had formed on the top and bottom sides of the grid where
the wires intersected. The scale formation was most severe in the center
section of the grid; an area of about 1 square foot was completely plugged.
The deposit on the intermediate grids was uniform and about 1/16-inch
thick. The top grid was partially plugged with a mixture of scale and soft
solids from the scrubbing slurry.
Scrubber Walls: The rubber-lined walls of the scrubber were coated
with hard scale ranging from 1/8— to 1—inch thick. Petrographic examination
indicated alternate layers of calcium sulfite and calcium sulfate of varying
thickness. Approximately equal portions of calcium sulfite and calcium
sulfate were present according to chemical analysis.
In addition to the general scaling, an accumulation of solids
resembling a stalactite occurred on one wall between the top two grids. The
formation, about 12 inches thick and J5 feet long, was a soft mixture of scale
and solids from the scrubbing slurry.
•
FlexiTray and Mist Eliminator; Solids from entrained scrubbing
slurry had accumulated on the bottom of the FlexiTray to a thickness of
2 to 4 inches, but the openings in the tray were not plugged.
The chevron mist eliminator was generally clean except for the
portion located directly above the FlexiTray overflow and drain trough.
The pressure drop across the FlexiTray and mist eliminator assembly
increased from 2.J to 2.6 inches during the period of operation (27 days).
Pilot—Plant Simulation of Prototype Operation
Scaling of the scrubber did not occur during prolonged operation
(8000 hours) of the Colbert pilot plant. The reason for the heavy accumu-
lation of scale in the prototype scrubber is. unknown, but it may have been
caused by the following:
• Inadequate humidification of the hot flue gas "before it
entered the scrubber.
532
-------
• Poor distribution of the gas and scrubbing liquor in the
scrubber.
• Pine particle size of the Fredonia limestone.
• High Ca:S02 mole ratio (1.5 to J. 0) caused "by inaccurate
metering of the fresh limestone feed.
A test was made in the pilot plant to determine if the
conditions used in the prototype plant would also cause scaling in the
pilot plant.
The pilot plant was arranged to closely duplicate the equipment
configuration and operating conditions used in the prototype. The velocity
of the gas through the scrubber was maintained at 8 ft/a. Scrubbing slurry
containing 15$ solids was fed to the scrubber at an L/G of 50. The reten-
tion time of the scrubbing slurry was 7 minutes based on the pumping rate
to the scrubber. Finely ground Fredonia limestone was obtained from the
prototype grinding and storage facility. The Ca:S02 mole ratio was main-
tained at 2.5 to duplicate the average limestone feed rate used during
the test in the prototype system. Makeup water was added to simulate the
partial open—loop operation of the prototype system caused by the dilute
thickener underflow.
Operation was routine. The S02 removal efficiency averaged
68$ compared to 75$ in the prototype. A comparison of the trends in
the S02 removal efficiency and pH and solids content of the scrubbing
slurry for the prototype and pilot plants is shown in Figure 18.
The test was terminated after 1J days of continuous operation.
The grids and walls of the scrubber were coated with heavy deposits of
scale. The extent and composition of the scale deposits in the pilot-
plant scrubber were very similar to the deposits found in the prototype
A comparison of the chemical composition of the scrubbing slurry and
scale deposits from the pilot-plant and prototype operations is shown
in Table VTI.
Based on these tests and similar tests conducted by EPA in
a small pilot plant at Research Triangle Park, North Carolina, it has
been concluded that the scaling of the prototype and pilot-plant scrubbers
was primarily caused by the combination of (l) high Ca:S02 mole ratio and
(2) the fine particle size of the soft Fredonia limestone.
Process Equipment and Construction Materials
One of the primary objectives of the pilot-plant program was
to evaluate process equipment and construction materials. This information
is required for the design of the Widows Creek system.
533
-------
PROTOTYPE (O-O-O)
en
U)
4=
gjf
Ui Ul
o u.
in u
IT
o:
CO
£D
O
(0
80
70
60
50
6.4
6.2
6.0
5.8
CO
UJ
(A
O
14 -
13-
12 -
FIGURE 18
Comparison of Trends in 50^ Removal Efficiency and pH and Solids Content
of the Scrubbing Slurry in the Prototype and Pilot Plant
-------
TABLE VII
Comparison of Chemical Analysis of Limestone - Wet-Scrubbing Slurry_
and Scale Deposits from the Prototype Multigrid Scrubber
at Shavnee Steam Plant and Pilot Plant at Colbert Steam
HLant
Scrubbing slurry
Filtrate, gm/1
Ul
U)
en
Plant
Prototype
Pilot plant
Duration,
hr
500
310
Limestone
sourcea
Fredonia
Fredonia
Ca
27.4
30-7
Cakej % by vt
Sulfite
MB s
2.0 3-8
0.9 4.2
Total
S Ca
5-5 0.8
4.9 1-1
Total
S
0.4
0-5
Total
dissolved
solids
5-4
7.1
Ca
23.2
21.5
Chenical scale, % by wt
Grid
Sulfite
Mg S
0.4 10.4
0.1 14.5
Walls
Total
S
16.4
17-1
Ca
25-5
19-7
Mg
0.1
0.4
Sulfite
S
9-3
8-9
Total
S
19.2
11.3
Limestone supplied by Fredonia Valley Quarries, Inc., Fredonia, Kentucky, and ground by 3VA (50# -8 microns) in prototype dry ball mill at the Shawnee
Steam Plant.
Scrubbing slurry contained 15$ undissolved solids by weight.
IJOTE: Both plants vere operated using a Ca:SOa mole ratio of 2.5-5-0 based on the S02 concentration of the inlet flue gas (normally 2800 ppn). The
SQS removal efficiency varied from 65 to 755&-
-------
In as many cases as possible, commercially available equip-
ment was used and observed over extended periods of time. Pumps,
valves, piping, and process instrumentation were evaluated in this
manner. Most of the construction materials and coatings were evaluated
in a special test device designed to simultaneously expose several
different materials to the same chemical and erosive environment.
A summary of the conclusions drawn from this evaluation
follows.
Pumps; Soft rubber—lined (natural or neoprenej centrifugal
pumps were normally used to circulate the scrubbing slurry. The rubber
housing liner and impeller coating showed very little wear after 9000
hours of operation. Severe erosion of the housing and impeller occurred
within 500 hours when an unlined Type JOU stainless steel pump was used.
Flow Control and Isolation Valves: Plug, pneumatic pinch, and
"butterfly valves were evaluated for controlling the flow of scrubbing
slurry. The plug and pinch valves were lined with soft rubber. The
body and disk of the "butterfly valve was coated with abrasions-resistant
polyethylene.
The 3/l6-inch-thick rubber coating on the "body and plug of the
plug valve eroded through during 1200 hours of operation. The pneumatic
pinch valve was not suitable "because of severe vibration caused "by pulsa-
tion of the rubber sleeve. The butterfly valve provided an effective
means of trimming the flow of slurry, although positive shutoff was not
possible. The polyethylene lining was in good condition after 1000 hours
of service.
Knife—type gate valves (unlined stainless steel construction)
were effective for isolation and on-off service.
Piping; Rubber hoses were used for most of the slurry piping
in the pilot plant. The hoses had soft rubber (neoprene) liners and
gave excellent service. The velocity of the slurry through the hoses
and piping normally did not exceed 8 ft/s. In addition to the rubber-
lined hose, special test sections of pipe of the following construction
were tested:
• Mild steel
• Type 3l6L stainless steel
• Glass-reinforced polyester
• Mild steel lined with polypropylene
• Mild steel lined with urethane
• Mild steel lined and coated with polyethylene
536
-------
Unlined mild steel pipe is not suitable because of severe
erosion and corrosion. Some types of urethane—lined steel pipe failed
because of blistering of the liner. All other materials tested gave
good service. Selection of any one of the types of pipe construction
tested in the pilot plant (except mild steel and some urethane linings)
for a commercial unit would depend largely upon cost.
Instrumentation; Several instruments were used for process
control. A discussion of the performance of the instruments follows.
• SOP Determination-—An ultraviolet analyzer (DuPont photometric)
was used to monitor the concentration of S02 in the flue gas
before and after the scrubber. This unit has given good service
and required only a minimum of maintenance. Leaky valves in
the automatic sample selection system caused the initial instal-
lation to be unreliable. Replacement of the automatic sample
selection valves with manual valves cured the problem.
• pH Measurement'—The pH of the slurry in the multigrid retention
tank was measured continuously with a system manufactured by
Universal Interloc, Inc., consisting of a waterproof sensor
and amplifier assembly submerged directly in the slurry reten-
tion tank. The pH measurement is indicated and recorded by
an instrument located in the control room. This system has
worked very well and agrees closely with laboratory pH measurements,
The unit has required no maintenance, -nere has been no indication
of electrode scaling or erosion over 6 months of continuous expo-
sure to the scrubbing slurry.
• Slurry Density—The solids content (density) of the slurry was
monitored continuously with a differential pressure (bubble
tube) device. Two tubes (l-inch pipe) are submerged directly
in the slurry retention tank at a fixed pressure differential.
A constant air purge is supplied to each tube. The differential
in pressure of the air purge is proportional to the density of
the slurry. This device has been very reliable and useful as
a trend indicator. The absolute solids content of the slurry
is determined gravimetrically in the pilot-plant laboratory.
• Measurement of Slurry Flow—The flow rates of the fresh
limestone slurry (6(# solids) and the scrubbing slurry
solids) are measured with magnetic flowmeters manufactured
by the Foxboro Company. The meters are lined with an abrasion-
resistant material (Adiprene—L) similar to soft rubber. The
readout system is all electronic, consisting of a field-mounted
signal converter and a recorder located in the control room.
The performance of these meters has been very good and only
a minimum of maintenance has been required over a period of
18 months.
537
-------
Metals and Coatings; Several metals and coatings were
evaluated for possible use in the Widows Creek installation. Types 316
and 3l6L stainless steel showed good resistance to corrosion and erosion
in the scrubber (grids) and piping system. Severe pitting and general
corrosion occurred when stainless steel was used as a tube material for
an indirect reheater in the scrubber exhaust gas duct. Reheat tubes made
of other materials (inconel 625, Incoloy 825, Carpenter 20 Cb-3, and
Hastelloy C-276) are currently being evaluated. Mild steel, Cor-Ten and
Type 304 stainless steel are generally not suitable for use in the slurry
or gas systems because of severe pitting and general corrosion.
Soft rubber (natural or neoprene) showed good resistance to
erosion; however, blistering and separation from the base metal (mild
steel) occurred frequently. Excellent bonding and resistance to erosion
was obtained with a urethane elastomer (Urecal 2003) manufactured "by the
Urecal Corporation.
Limestone wet scrubbing provides an effective method for
removing S02 from boiler flue gases. An S02 removal efficiency of 70$
can be expected with a multigrid type of scrubber at an L/G of 50 to 60.
The efficiency can be increased to about 9<$ if a two-stage mobile- bed
(TCA) scrubber is used.
Long— term operating reliability is largely determined by proper
design and selection of suitable construction materials to cope with the
erosive and plugging characteristics inherent in slurry scrubbing.
538
-------
STATUS OF C-E's AIR QUALITY
CONTROL SYSTEMS
by
M. R. Gogineni, Supervisor
Chemical Process Design and Development
J. R. Martin, Supervisor
Systems and Equipment Development
Kreisinger Development Laboratory
P. G. Maurin, Assistant to Manager
Air Quality Control Systems
C-E Combustion Division
Combustion Engineering, Inc.
Windsor, Connecticut
539
-------
STATUS OF C-E's AIR QUALITY CONTROL SYSTEMS
INTRODUCTION
The need for removing S02 from stack gases of oil
and coal-fired combustion equipment has been and is
being emphasized by the stringent emission limitations
established by the Environmental Protection Agency
and other government agencies. The EPA standards
are 1.2 lb/106 Btu for coal and 0.8 lb/106 Btu for oil.
More stringent requirements in certain districts have
been passed. An example is Clark County, Nevada
which limits S02 emissions for a 1,500,000 kw steam
generating station to 0.15 lb/106 Btu or ]^th of the
EPA requirement. In order to meet these require-
ments, a very high percentage of 862 produced by the
combustion of most oil and coal fuels must be removed
from stack gases.
There are several ways to classify processes for the
removal of S02 from stack gases: wet or dry, recovery
or non-recovery, and absorption, adsorption, or cata-
lytic oxidation/1' In reviewing these processes and
many others, Combustion Engineering decided that
wet lime/limestone scrubbing without recovery of
sulfur was most worthy of development.
The development of C-E's air quality control system
started in 1964 with the construction of a small pilot
facility in our laboratories. A second pilot application
on a Detroit Edison Co. unit in 1966 and 1967 followed.
Table I lists the full-scale installations'2' that have
been sold to date by C-E. In a move to further accele-
rate development, a large prototype unit of 12,500-cfm
capacity was constructed in our laboratory and began
operations in early 1970.<3)
This paper describes the C-E AQCS for the removal
of particulate matter and SOa from stack gases of
steam generators emphasizing C-E's experience with
the full-size field units.
SYSTEM DESCRIPTION
In the tail-end C-E Air Quality Control System
(Fig. 1), a slurry of pulverized limestone or slaked lime
enters directly into the reaction tank (located at the
bottom of the scrubber). Recirculation pumps convey
the scrubbing slurry from the reaction tank to underbed
spray nozzles. The incoming gas, laden with dust and
S02, contacts the sprayed slurry and continues to the
bed. The removal of SOz and particulate matter occurs
in the bed. The scrubbing cycle continues with the
reacted materials draining to the reaction tank which
is designed to provide for completion of chemical re-
actions and precipitation of solids. From the reaction
tank, pumps recirculate the scrubbing solution.
A bleed line provides the necessary solids removal to
a clarifier or pond. Here, solids settle and clarified
water is available for recirculation. The cleansed flue
gas passes through a mist eliminator for removal of the
remaining water and is then reheated for induced-draft
fan protection and plume control.
The furnace injection process (Fig. 2) involves the
injection of an additive which contains a high per-
centage of calcium, such as limestone, into the furnace
of a steam generating unit. The pulverized additive
calcines in the furnace and reacts with the combustion
gases, removing 20 to 30% of the sulfur oxides, in-
cluding all the 80s. The flue gas enters the scrubber
and the process proceeds as described above. Due to
additive distribution and boiler deposit problems, the
furnace injection system is not now being offered.
The basic design variations of tail-end systems result
from additive selection — limestone or limj — and are
additive preparation and system control. Although both
additives enter the system as a slurry, the limestone
(usually received in the — Yi to —2-inch size range) re-
Utility
Union Electric
Kansas Power & Light
Kansas Power & Light
Kansas City Power & Light
Kansas City Power & Light
Northern States Power
Northern States Power
Northern States Power
Louisville Gas & Electric
Southwest Public Service
C-E CONTRACTS FOR
Unit
Meramec 2
Lawrence 4
Lawrence 5
Hawthorn 3
Hawthorn 4
Sherburne 1
Sherburne 2
Black Dog
Paddy's Run 6
Nichols Station B
TABLE I
AIR QUALITY CONTROL SYSTEMS
Size, Mw
140
125
430
100
100
690
690
Prototype
ISOOOcfm
65
350
Fuel
Coal
Coal/Gas
Coal/Oil/Gas
Coal/Gas
Coal/Gas
Coal, Low Sulfur
Coal, Low Sulfur
Coal, Low Sulfur
Coal/Gas
Coal/Gas
System
Furnace Injection,
Limestone
Furnace Injection,
Limestone
Furnace Injection, •
Limestone
Furnace Injection,
Limestone
Converted Tail-End,
Limestone
Tail-End, Limestone
Tail-End, Limestone
Tail-End. Limestone
Tail-End, Lime
(Ca(OH), Sludge)
Tail-End, Lime
540
-------
mcir MS
10,
9*««*«T9»-
_m»viet
wjrr*
_jro OTH«
KHirS
s«ri~\
r
.A
...-•" h
,,**'
"- CJICJOK
* 1
a.
y
r^M
V
«V<
\-JTIM»««
— MIXCft
\-5MI«r PUMP
cr/ov nwr
*— «« now MCTE«
*» V*
ni*r
umtKrum nuf
SCDUCKHS
Fig. 1: C-f foil-end AOCS
RUCTION TANK
III
Fig. 2: C-E furnace injection AQCS
Quires size reduction and water addition. A wet milling
sYstem is used for this. Lime requires a slaking opera-
tion to form a calcium hydroxide (Ca(OH)2) slurry.
A direct supply of solid "Ca(OH)2 is slurried to facilitate
"lira-system transport and control.
For a CaO/Ca(OH)2 system, the additive enters the
faction tank as required to maintain the spray slurry
Pi I at a level high enough to maintain SO2 removal
efficiency, yet low enough to avoid CaS03 scaling. For
a calcium carbonate system, fuel sulfur content
Determines additive feed rate.
SYSTEM DESIGN"*
System Chemistry
The possible reactions taking place in the wet
'''ne/limestonc SOa removal system have been studied.
'•*ur conclusions relative to the controlling reactions are
''ttswl on consideration of theoretical equations in light
°r operating experience in both field and lalwratory
systems. Because there are considerable differences in
the operating conditions required to provide adequate
SOj removal in the absence of scale or deposit forma-
tion when utilizing lime or limestone as additives, the
chemistry of the systems are treated separately. The
essential reactions governing these systems are:
Calcium hydroxide or lime system reactions
CaO + H20>=tCa(OH)2 (1)
Ca(OH)2 + S02 & CaS03 + H20 (2)
CaSOs + SO2 4- H20 * Ca(HS03)2 (3)
Ca(HS03)2 + Ca(OH)2 * 2CaS03 + 2H20 (4)
2CaS03 + 02-»2CaSO< (5)
Calcium carbonate or limestone system reactions
CaCOs + C02 -r- H20 ^ Ca(HC03)2 (6;
2S02 + Ca(HC03)2 ^ Ca(HS03)2 + 2C02 (7)
CaS03 + S02 + H20 ^ Ca(HS03)2 (8)
2CaS03 + 02 ^ 2CaSO4 (9)
Ca(HS03)z + 2CaC03 n Ca(HC03)2 + 2CaS03 (10)
Following the initial steps of hydration (Eq. 1) and
formation of calcium sulfite (Eq. 2), removal of SOa in
the lime or calcium hydroxide system depends upon
the formation of calcium bisulfite by reaction of sus-
pended calcium sulfite with sulfur dioxide and water
(Eq. 3).
The control of sulfite scaling requires that a minimum
amount of free hydroxide ion be recirculaled to the
scrubber, therefore, fresh additive (lime or slaked limf)
is added in a reaction tank external to the scrubber
where calcium sulfite is formed (Eq. i). An amount of
calcium sulfile equivalent to the 862 removed (or lh«
541
-------
fresh Ca(OH)z added) is conveyed from the system to
a pond or vacuum filter and the remainder recycled to
continue the removal process.
The principal absorption reactions for the calcium
carbonate system are shown in Eqs. 6, 7 and 8. Sulfur
dioxide reacts with the relatively soluble bicarbonate
to form calcium bisulfite. In addition, solid calcium
sulfite recycled from the reaction tank reacts with 862
to form bisulfite.
The reactions in which sulfite is oxidized to sulfate
(Eqs. 5 and 9) and soluble bisulfite is converted to in-
soluble calcium sulfite (Eq. 10) account for the waste
products as well as the regeneration of the solid calcium
sulfite reactant that is recirculated to the scrubber. The
ratio of calcium sulfite to calcium sulfate found in the
air quality control system solid waste depends upon the
extent to which these reactions go to completion.
Scru66cr Design
The primary function of the scrubber is to transfer
SOz from the flue gas into the liquid. The SOz remains
in the liquid or is converted partially to solid sulfur
compounds in the scrubber. Hence, knowledge of vapor-
liquid mass transfer rates is important in scrubber
design. The marble bed scrubber, which has a turbulent
layer, acts like an absorption tray.
The vapor-liquid equilibrium line and efficiency are
needed for tray design. Typical operating and equilib-
rium lines are shown in Fig. 3. The operating line is the
material balance line and has a negative slope of L/G
(liquid to gas ratio).
x0 x
MOL£ FRACTION OF S02 IN UQUID, X
Fig. 3i Vapor-liquid matt trontfw
The inlet gas composition YI is known, and the SOz
concentration in the incoming liquid Xlt which is
usually zero, is also known. Hence, point 1, which
represents the inlet condition is known. The operating
line can be drawn through Xj Yi with a negative slope
of L/G.
The intersection of the operating and equilibrium
lines (X, Ye) represents the scrubber outlet gas and
liquid conditions for a theoretical stage which repre-
sents 100% overall tray efficiency. The overall tray
efficiency E is defined as:
E
Y.-Y,
Y,-Y.
(ID
and is usually less than 100%. It can be seen both from
Fig. 3 and Eq. 11 that the actual outlet liquid and gas
compositions (Xo, Y0) can be predicted using the tray
efficiency and the point of intersection of the operating
and the equilibrium lines.
Laboratory tests show that the tray efficiency of the
marble bed scrubber is 90 to 95%. This indicates that
the marble bed scrubber is a good liquid-gas contactor.
Once the tray efficiency is known the number of
marble beds needed to obtain a required SOz removal
can be determined from the operating and the equilib-
rium lines.
Knowing the total alkalinity of the liquid available
in the scrubber, which is a function of the additive
dissolution rate for slurry systems, the equilibrium line
is plotted using a computer program.<5> For a specific
L/G, the equilibrium line shown in Fig. 3 moves to
the left or right depending on the alkalinity of liquid
available in the scrubber and, hence, affects the SOj
removal efficiency. The SOz removal efficiency is
defined as:
Y,-
(12)
If the additive dissolution rate is high enough to
maximize the available alkalinity in the liquid in the
scrubber, the equilibrium line will move to the far
right to a point where Y. can become zero. This repre-
sents the most favorable condition for SOa transfer
from gas to liquid. Comparison of Eqs. 11 and 12
shows that the SOj removal efficiency will approach
the tray efficiency as Y. approaches zero.
Reaction Tanks
The function of the reaction tanks is to provide:
a. Dissolution of the additive in order to convert
the highly soluble bisulfite present in the liquid
leaving the scrubber to relatively insoluble sulfite.
b. Precipitation of calcium sulfate which is formed
in the system (scrubber and/or reaction tanks)
due to the oxidation of sulfite.
c. Precipitation of calcium sulfite.
Additive dissolution rates vary considerably with
the type, origin, preparation, and concentration of the
additive. At C-E, a prototype scrubber system, pilot
542
-------
plant scrubber system, continuous flow stirred tank
reactors, and batch reactors have been used to deter-
mine the dissolution rates for individual additives used.
The following rate expression for the precipitation
of calcium sulfate has been developed:'6'
R - -KZ(C-C.)2 (13)
The rate of desupersaturation, R, is proportional to
the gypsum seed crystal concentration, Z. The differ-
ence between the actual concentration, C, and the
equilibrium concentration, C« of SO^ or Ca++ is the
driving force terra. Although it is more accurate to
express the -driving force in terms of the activity and
the solubility products, the driving force in Eq. 13 is
given in concentrations for convenience, by including
the factor for converting activities into concentrations
in the rate constant K. Using the rate constants de-
termined by experimentation,'7' the rate expression
given by Eq. 13 is employed in the design of reaction
tanks to ensure that calcium sulfate supersaturation is
eliminated.
Laboratory studies are in progress to determine the
calcium sulfite precipitation kinetics and the oxidation
kinetics of sulfite to sulfate.
DEPOSIT AND SCALE FORMATION
Calcium sulfite and calcium sulfate scaling can be a
problem for lime/limestone wet scrubber systems.
Scaling occurs when the solutions are supersaturated
to a point where heterogeneous crystallization takes
place resulting from nucleation. The ratios of the
products of the activities (A) of Ca++ and SO4 or SO3
to their solubility product constants (KSp) as a measure
of the degree of supersaturation are:
~AQa-HAgQ- "1 <1 Subsaturation
^ I =1 Saturation
Sp(CaSOa)J >1 Supersaturation
SP(CaS04)
Laboratory experiments have shown that hetero-
geneous crystallization is not significant until the ratio
of the activity product to the solubility product con-
stant reaches about 1.5 for calcium sulfate and about
7 for calcium sulfite.
Heterogeneous crystallization is minimized by pro-
viding seed crystals for homogeneous crystallization
and by designing the reaction tanks so that the liquid
leaving them is close to saturation and not highly
supersaturated. This requires knowledge of precipita-
tion kinetics of calcium sulfate and sulfite.
Calcium Sulfite Deposition
Calcium sulfite (CaSOa • HH20) is formed in the
scrubber under those conditions that favor sulfite
formation. These conditions are apparent when one
considers the sulftte-bisulfite equilibrium and compares
the relative solubilities of the corresponding calcium
salts. As seen in Fig. 4, extremely soluble bisulfite in
MOLE FRACTION SULFUROUS ACIO-OISUl.FITE-SUI.riTe
pH
Fig. 4: Mole fraction sulfurou* ocid-bitulfite-tulfiit vs. pH
solution changes to relatively insoluble sulfite when the
solution pH shifts from 4 to 10. When SOz is absorbed,
the scrubber solution is usually between pH 4 and 6
and, therefore, the predominant species is bisulfite. If
the pH of the scrubber solution containing bisulfite is
suddenly raised either in localized areas or in a reaction
tank, crystallization of calcium sulfite will occur.
Experimental work with lime scrubbing has shown
that sulfite scaling occurs in the scrubber bed when
free hydroxide is introduced according to the above
explanation. By proper control of the pH of the spray
slurry (less than 10) entering the scrubber, calcium
sulfite scaling will not occur in the scrubber.*8*
In the calcium carbonate system, the buffering action
of the carbonate-bicarbonate couple (Eq. 6) maintains
a system pH between 5 and 7, thus sulfite scaling is
not encountered.
Calcium Sulfate Deposition
The solubility of calcium sulfate is only slightly in-
creased with increasing pH, and calcium sulfate scaling
is related to the tendency of this material to form
extensively stable supersaturated solutions. While
chemical theory predicts that a given ionizable species
will not remain in'solution when the solubility product
of its component ions has been exceeded, calcium
sulfate may be held in solution to an extent twice thot
predicted before crystallization of calcium sulfate will
occur (gypsum, CaSO« • 2HaO).
The significance of this phenomena to scrubber
operation is that SOa removal can be accomplished
while scrubbing with solutions containing more than
the theoretical calcium and sulfate ion concentrations,
but less than some experimentally determined level at
which precipitation will occur within the scrubber
proper.
Crystallization from supersaturated solutions car
occur by two processes, formation of new crystals or
nucleation and growth of existing crystals.
543
-------
The internal surfaces of the scrubber can provide
nucleation sites, thus resulting in scale formation. For
many crystal systems, growth will occur without sig-
nificant nucleation if sufficient seed crystals are pro-
vided. Work by other investigators*61 has shown that
supersaturated calcium sulfate solutions can be effec-
tively desupersaturated by circulation of 1 to 5%
gypsum seed crystals.
By employing this technique, operation free of
calcium sulfate scaling has been demonstrated in both
laboratory and field installations. This seeding tech-
nique is the key to "closed loop" operation in which the
only liquid leaving the system is by evaporation or
combined with the solid by-product of the scrubbing
system.
The potential for sulfate scaling is generally more
prevalent in the calcium carbonate system than in the
calcium hydroxide system. The calcium hydroxide
system reaches steady state below the value of calcium
sulfate saturation.
Control Techniques™
The controls for the C-E AQCS depend upon the
additive type. The limestone tail-end system can be
controlled by the limestone feed to the system. The
limestone feed is determined by the 862 in the flue gas
or sulfur in the fuel. The solids concentration in the
slurry is maintained to prevent calcium sulfate scaling.
The tail-end lime system is controlled by the pH
of the spray slurry and the total solids in the system.
Slurry Circulation
Slurries containing 2 to 10 wt. % solids are circulated
in the lime/limestone wet scrubber systems in order to
control scaling in the system and to improve SOa
removal and additive utilization. The circulation of the
high solids slurries in the system could cause serious
erosion problems in pumps, piping, nozzles, etc.
AQCS solids are different from other solids. Little
technology is available to design pumps, piping, and
nozzles that can handle AQCS solids, therefore, C-E
undertook extensive testing of these components.
Several pumps of different materials have been in-
stalled on C-E systems both in the field and in the
laboratory. The information obtained from the evalu-
ation of these pumps is being used in current system
Several different materials have been evaluated to
determine piping material that can withstand erosion
and corrosion. The piping is designed so that slurry
velocities in the pipes are high enough to prevent
settling.
Severe nozzle plugging problems have occurred when
commercially available nozzles were used. Therefore,
C-E has developed a nozzle that does not plug, erode,
or corrode in the scrubber environment.
Solid-Liquid Separation
Solid-liquid separators such as clarifier-thickeners
and vacuum filters separate the solids from the liquid.
The solids are disposed of and the liquid is returned to
the system. This equipment is presently being designed
by vendors to C-E's specifications.
Gas-Liquid Separation
The gas leaving the bed carries water droplets which
contain solid particles and dissolved solids. Mist elimi-
nators are used to remove water droplets from the gas.
Proper design of these sections is essential to prevent
both plugging with solids and re-en train men t of the
liquid collected on the surfaces.
Gas velocity and the distance between the bed and
the mist eliminators are important operational and
design variables. Extensive test work has been con-
ducted in the laboratory and in the field to evaluate
these variables. Results of these studies are being used
in the design of mist eliminators for C-E systems.
Reheating
The purpose of reheating the gas leaving the scrubber
bed is to protect the I.D. fan and to reduce plume
formation. The amount of reheat required depends
upon several factors such as atmospheric conditions
and stack height.
Additive Preparation and feeding
The limestone is ground to a small size (about 80%
thru 200 mesh). The slurry from the limestone mill
scalpers (about 60 to 70% by weight) is stored in the
limestone slurry storage tanks. The slurry is then trans-
ferred to tanks, where it is diluted with make-up water
to reduce the solids content to 10 to 15% by weight.
This dilute slurry is added to the reaction tanks for use
with the slurry spray to the scrubber bed.
For a lime system, slaking, storage, dilution, and
transport steps are involved. Use of Ca(OH)2 rather
than CaO eliminates the slaking operation.
The additive handling system is designed so that the
solids do not settle in pipes and tanks.
EPA CONTRACT
C-E has recently completed a research contract for
the Environmental Protection Agency (EPA) to op-
timize lime/limestone wet scrubbing processes for SOj
and participate removal in a marble bed scrubber.
Three types of tests were conducted on the 12,000-
CFM C-E marble bed scrubber. Soluble system tests
using once-through sodium carbonate scrubbing solu-
tions were conducted using a single marble bed to
determine the vapor-liquid mass transfer character-
istics. Limestone furnace injection tests were conducted
using boiler calcined limestone and flyash mixture to
determine the system performance and the solid-liquid
mass transfer rates. Limestone tail-end system tests
544
-------
were conducted using limestone slurry to determine
whether single marble bed results can be extrapolated
to predict the performance of a two-bed scrubber and
to determine the solid-liquid mass transfer rates.
The results of the once-through soluble system tests
using a Na2COs scrubbing solution show that:
1. The marble bed scrubber is a good liquid-gas
contacting device with an overall tray efficiency
of 90 to 95%.
2. The SOz removal in the marble bed scrubber is
limited by the vapor-liquid equilibrium.
3. Liquid to gas (L/G) ratio and the scrubber liquor
alkali composition strongly influence 862 removal.
The C-E scrubber at EPA's Shawnee test facility
was originally supplied with one marble bed. The
current commercial systems offered by C-E for high
sulfur fuel applications have scrubbers with two marble
beds. Therefore, C-E recommended that a second
•narble bed be installed in the scrubber so that a
meaningful evaluation of C-E's current commercial
offering could be accomplished. Due to the lack of
available funds and the scheduling problems, EPA
decided against installing the second bed. As a reason-
able alternative, both C-E and EPA agreed that tests
would be run on the C-E AQCS prototype to determine
whether the performance of a two-bed scrublx-r could
be predicted from the single marble bed perform;,.,,',-
of the EPA test facility at Shawnee.
A series of tests were run on the C-E \QCS prolutyp,.
using limestone slurry with one and two marble U-ds
Test results show that the performance (SO2 removal
efficiency and elimination of scaling) of the scruhbrr
with two marble beds can be predicted by exIrapolaliriL'
the single bed test results. The S02 removal clfinem-ics
of the lower and the upper beds appear to l«: the same
based on the SOo concentrations entering the respective
bed.
Kansas Power and Light — Lawrence 4
This 125-Mw coal and gas fired C-E boiler was
retrofitted with a furnace injection AQCS in 1968. The
AQCS consisted of two wet scrubbers and related
equipment (Fig. 5). During 1971, it was recognized Unit
this system could not continue to operate as an open
system (i.e., liquid blowdown). While liquid was never
discharged to a natural body of water, the sludjje
disposal pond constructed at the plant site in 1%8
never reached saturation because of the small quantity
of scrubber effluent flow into this pond during the throe
years of operation and because of intermittent operat ion
of the AQCS on Unit 4.
STACK
xESTONE SUPPLY
COM. SUPPLY
SETTLING POND
Fig. 5: Flow diagram of AQCS for Unit 4 at KP&L
545
-------
Utilizing the results of tests conducted at our
Kreisinger Development Laboratory during 1971, it
was decided that the system could be modified to
ensure scale-free closed-loop operation. In order to
verify the design modifications, a series of tests were
conducted in February and March of 1972 which
demonstrated scale-free operation for periods extending
to two weeks.
During these tests, all process streams were measured
for chemical constituents in both the solid and liquid
phases. In addition, the flue gas entering and leaving
the AQCS was measured for both gaseous and solid
constituents using EPA approved methods. Table II
gives a typical set of data obtained in this test series.
TABLE II
TYPICAL OPERATING DATA
KANSAS POWER & LIGHT 4
February, 1972
5.0 gr/DSCF
0.06 gr/DSCF
1800 ppm
550 ppm
9.2
Inlet dust
Outlet dust
Inlet SO,
Outlet SOj
PH
% Solids
CaO
SO3
SO4
Solid analysis
CaS04 • 2H>0
CaSO, •
Flyash
CaO
Underbid Slurry
7.3
960 ppm
250 ppm
2,150 ppm
20%
15%
60%
5%
Pot Effluent
7.3
1.250
1.300
2,450
Since these results agreed with the data obtained in our
laboratory with respect to seed crystal concentration
required to prevent CaSC>4 scaling, and SOa removal
was maintained at 70 percent for the single marble bed
scrubber, it was decided to complete the modification
in order to operate the system at the 7 to 10% solids
level at full load.
The modifications to Unit 4 were made during the
spring and summer of last year. Subsequent operation
of the system in the fall of last year revealed a problem
with maintaining mist eliminator performance while
operating with high solids (7 to 10%). An extensive test
program has been undertaken to solve this problem.
Kansas Power and Light — Lawrence 5
Unit 5 at the Lawrence Station of Kansas Power and
Light, a 430-Mw C-E coal, gas, or oil fired unit was
first placed in service in the spring of 1971. The AQCS
supplied for this unit was based on the same design as
Unit 4. The system consisted of six scrubbers, each
handling approximately 165,000 CFM of flue gas at
125 F. The additive, calcium carbonate, was pulverized
and injected into the boiler, calcined, and then conveyed
to the scrubber for use in the SO2 absorption process.
The AQCS was not placed in service until November
1971 since the unit was operated on gas until that time.
The system was operated for three weeks without any
serious problems until the sludge pond became satu-
rated with calcium sulfate, at which time scaling in'lhe
scrubber was observed. Since tests were already in
progress on Unit 4 to enable operation of the AQCS as
a closed system without scale formation, the AQCS on
Unit 5 was operated on an intermittent basis during
the winter season when the unit fired coal.
After the successful completion of the closed loop
modifications on Unit 4, plans were made to "update"
Unit 5. Figure 2 shows the modifications that were
made. These modifications included installation of a
reaction tank where crystallization, dissolution, and
precipitation reactions take place; mixing devices for
the reaction tank; fiberglass and/or rubber lined piping
for the high solids recycle system; new underbed slurry
pumps; improved controls that can withstand the wet
scrubbing environment; redesigned Blurry nozzles to
decrease chemical and mechanical wear problems; and
a new demist system. Further, it was decided from test
data obtained during the winter of 1971-72 that we
should reduce the gas throughput of the marble bed
scrubbers in order to minimize carryover from the bed
due to localized high velocities, and incorporate the
concept of spare scrubbing capacity by adding two
additional scrubbers to this unit.
The modifications were completed late last fall and
the AQCS was returned to service for the winter period
of coal operation.
The unit has operated approximately 1800 hours
(since January). The results duplicate closely the tests
run last year on Unit 4.
Presently, an extensive series of performance tests
are being conducted on this unit by C-E to determine
full-load capability of the system.
Kansas City Power and Light — Hawthorn it
At the Hawthorn Station of the Kansas City Power
and Light, a furnace injection system was retrofitted
on a 100-Mw low-load-factor unit. The system consists
of two scrubber units and related equipment, and
includes a water seal by-pass (Fig. 6). The system was
started up in September 1972.
Preliminary operation of the system indicated that a
serious maldistribution of additive existed between the
two scrubbers. After several weeks of testing, it was
also determined that the back-pass of this boiler was
not capable of handling the additional dust loading
created by the furnace injection of limestone.
Therefore, in late 1972, it was decided to modify the
system to a tail-end limestone, single marble bed
546
-------
MODIFICATIONS
LATE
*-TO POND
FROM POND
Fig. 6: Proeew flow schematic KP&L Hawthorn Station Units 3 and 4
system with the pulverized limestone being blown into
the scrubber inlet duct (Fig. 6). Modifications have
been completed, but the system has not been tested
to date.
A test series was run during March to determine the
Performance of this system on low sulfur, high alkali
Wyoming coal. The limestone feed \vas maintained at
Zero for these tests because there was significant amount
of alkali in the coal. Some coal samples indicate as
much as 200% stoic-biometry of alkali (if all were
available for reaction with the sulfur in the coal). The
results shown in Table III were obtained during this
TABLE III
OPERATING DATA
KANSAS CITY POWER & LIGHT
HAWTHORN 4, WYOMING COAL TEST
Inlet SOj: 700 - 300 ppm
Outlet SO]: 300 • 40 ppm
% Solids 3.0 - 6.0
Underbed slurry pH 5.7 • 7.1
UNDERBED SLURRY LIQUID
CHEMICAL ANALYSIS
CaO 850 ppm
SOi 150 ppm
S0< 220 • 2520 ppm
lcst program. The S02 removal varied from 60 to 90
-------
SLUDGE
Fig. 7: Proeew flow schematic IG4E Paddy'* Run No. 6
This system represents the culmination of a three-
year effort initiated in 1970. Pilot plant tests were
conducted at C-E's Kreiainger Development Labora-
tory in 1971 to determine if an AQCS could be designed
that would operate with the Ca(OH>2 sludge that
Louisville Gas & Electric had available. The Ca(OH)2
sludge is a waste product from the manufacture of
acetylene.
In June 1971, a 100-hour continuous test on the
12,000-CFM prototype scrubber system at KDL was
completed and, as a result, the system design was
finalized and fabrication of the system began. Later in
1971, it became apparent because of more stringent
water pollution regulations that the system would have
to be operated as a closed system. Therefore, additional
testing in the laboratory was done in February 1972.
The findings of these tests showed that the closed
system was accomplished by obtaining essentially zero
oxidation and providing sufficient seed crystals to
precipitate the calcium sulfite without forming scale.
This system is presently in the process of starting up.
At this time, there has been approximately 300 hours
of operation with one scrubber. The chemistry of the
system has undergone preliminary checkout and con-
firms the findings of our laboratory work.
The scrubbers have been operating exceptionally
well. S02 removal has been 85% or higher depending
on the L/G ratio and additive feed rate being main-
tained. It appears that the guarantee level of 80%
sulfur dioxide removal can be obtained with a total
L/G of 40.
The Ca(OH)2 sludge feed rate is controlled auto-
matically by the reaction tank pH. This control system
has been in service continuously since start-up without
a problem.
The vacuum filter system has also operated without
problems. The filter cake has varied from 30 to 50%
solids, the lower solids being quite muddy and the
high solids being handled easily.
The data in Table IV are for a 150-hour run con-
ducted during the last week in April.
Northern Stales Power — Sherburne County 1
The systems for Northern States Power, Sherburne
County Station have been designed to clean the flue
gas from two 690-Mw C-E coal-fired units. These units
548
-------
TABLE IV
LOUISVILLE GAS & ELECTRIC
OPERATING DATA
Inlet SOi 2400 • 2800 ppm
Outlet SOj 300 - 450 ppm
Undarbed Slurry
pH 9.2
CaO 850 ppm
SOT 20° PPm
SOJ 800 ppm
% Solids 9.5
Pot Effluent
5.2
1.100 ppm
1,300 ppm
850 ppm
9
Solids
CaS04-2HjO 1-2%
CaSOj • ViHjO 78%
Flyash 20%
will fire a low sulfur Western coal. Typical analysis of
this coal is given in Table V. The addition of limestone
to the system will he varied to maintain the sulfur
dioxide removal to satisfy air quality requirements and
balance the chemistry in the scrubbers. Since significant
quantities of alkali exist in the flyash of this Western
coal, the limestone addition will serve as a supplement.
Tests run at our facilities indicate that for certain
TABLE V
NORTHERN STATES ROW Eh
TYPICAL COAL ANALYSIS
Btu/lb
Moisture
Sulfur
Ash
MgO
CaO
SO.
COAL
ASH
8,130
23.5%
0.8%
9.0%
5.9%
21.9%
1.4%
sulfur concentrations in the coal, all the alkali required
can be supplied by the flyash.
Another interesting aspect of the design of these
systems is related to the appreciable quantity of mag-
nesium in the flyash of the coal to be fired. Tin; sulfur
salts of magnesium are extremely soluble, thcn-fori'.
the rate of dissolution of magnesium must he deter-
mined in order to predict the total dissolved solids in
the systems at steady state. The prediction of total
dissolved solids was required to determine possible
effects on the system chemistry (i.e., ionic strength of
CUSTOMER
DISPOSAL
Fig. 8: Proposed process flow schematic NSP Black Dog Unit
549
-------
tbe process solutions) and to provide information to the
Water Resources Commission of Minnesota. A develop-
ment program conducted at our laboratory predicted
a total dissolved solids level of 15,000 ppm, dictating
closed loop operation. Closed loop system operation
has been incorporated by providing adequate seed
crystals and a reaction tank external to the .scrubber
beds to carry out crystallization and precipitation.
Northern States Power — Black Dog Prototype
AB part of the project to develop the tail-end lime-
stone scrubbing systems for the Sherburne County
Station of Northern States Power, a prototype facility
has been constructed at NSP's Black Dog Station. The
purpose of this facility is to verify the design for
Sherburne County, test various system components,
and provide a training facility for NSP operating
personnel.
This facility is a scaled down duplicate of the Sher-
burne County design except that no additive grinding
equipment was incorporated. The scrubber module has
a capacity of 12,000 ACFM at 130 F. All other related
equipment is included in the facility as shown in Fig. 8.
Tbe Black Dog system began air/water testing
during the first two weeks of March 1973. On Marehll,
it was placed in service with a stream of flue gas being
diverted from the Black Dog unit precipitator inlet
duct. Since that date, approximately 1,000 hoars of
operation have been logged.
The test series at the Black Dog facility includes the
evaluation of the following parameters: (a) gas velocity,
(b) additive feed rate, (c) percent solids being circu-
lated, and (d) L/G ratio variation. The steady-state
concentration of soluble cations (i.e., magnesium) is
being studied because of their relatively high concen-
tration in Western coal.
The initial tests at Black Dog agree with our labora-
tory pilot plant data as shown in Table VI. Current
plans are to complete operation of this facility by late
August of this year.
TABLE VI
OPERATING DATA FROM
LABORATORY PILOT PLANT TEST
FOR NORTHERN STATES POWER
Inlet SOi BOO ppm
Outlet SOi 200 ppm
I-
V)
o
o
u>
100
80-
60-
0-
o
< 40-
O
u.
O 20-
3*
0-
—
—
< < * j
!.
,f, "
£ ?
\ *
1 "]>
1 *" f V
Ai.*.. >•
5-
S 5. *».
* i / t
v <
f > \
\Jf /^
> i ; j<
, <
< ^ " <
^ :
L/G
Underbed slurry pH
Pot effluent pH
Underbed slurry
Li Quid analysis
SOT
SOT
20
6.5-7.1
5.8 - 6.3
600 ppm
900 ppm
150 ppm
4.000 ppm
MILLS
PER
$/HR KWHR
-------
ECONOMICS
Table VII shows the assumptions for a 600-Mw unit
and Fig. 9 shows the estimated operating costs.
TABLE VII
ASSUMPTIONS FOR DETERMINING
COSTS FOR A TYPICAL 600-Mw UNIT
Heat rate
Load factor
Fuel ash
Sulfur
HHV
Ash drop out
To precipitator
Electrostatic precipitator
AQCS tail end
9500 Btu/Kwhr
70%
12%
Additive limestone
Electric power
Water
Disposal
Maintenance
Labor
Capital
1 1,700 Btu/lb
15%
85%
90% effective
SO, Emission Rate 1.2 lb/10« Btu
95% pure • 130% stolen.
$3.SO/ton delivered
$l,00/ton wet grinding
2.5% of 600 Mw
$.007/Kwhr
1.75 gpm/Mw
$.13/1000 gal.
$4/ton
50% dry solids
4.3% of capital cost
4 men/shift - 4 shifts - $15.000/yr/man
17.5% of $34/Kw
REFERENCES
1. MAUBIN, P. G. and JONAKIN, J., "Removing Sulfur
Oxides from Stacks," Chemical Engineering, April
27,1970.
2. JONAKIN, J. and MARTIN, J. R., "Applications of the
C-E Air Pollution Control System" Second Inter-
national Lime/Limestone Wet Scrubbing Sympo-
sium, New Orleans, La., November 8-12, 1971.
3. PLUMLEY, A. L. and GOGINEM, M, R., "Research
and Development in Wet Scrubber Systems," Second
International Lime/Limestone Wet Scrubbing Sym-
posium, New Orleans, La., November 8-12,1971.
4. GOGINENI, M. R., TAYLOR, W. C, PLUMLEY, A* L.,
and JONAKIN, J., "Wet Scrubbing of Sulfur Oxides
from Flue Gases," American Chemical Society Na-
tional Meeting, New York, New York, August 27-
September 1, 1972.
5, Radian Corporation, "A Theoretical Description of
the Limestone Injection — Wet Scrubbing Process,"
A Report to NAPCA, HEW contract No. CPA-22-
69-138, June 1970.
6. LESSINO, R., "The Development of a Process of Flue
Gas Washing Without Effluent" Journal of the
Society of Chemical Industry, Transactions and
Communication, 57, pp. 373-388, November 1938.
7. RADEH, P. C., "Bench Scale Studies of CaSO* De-
supersaiuration Kinetics," C-E Internal Reports.
8. MAUBIN, P. G.t "The Combustion Engineering Air
Pollution Control System,1" Instrument Society of
America, 15th Annual Power Division Symposium,
Dallas, Texas, May 22-24, 1972.
551
-------
MAGNESIA SCRUBBING
by
Gerald G. McGlamery
Design Branch
Tennessee Valley Authority
Muscle Shaols, Alabama
553
-------
MAGNESIA SCRUBBING _!/
Gerald G. McGlamery, Design Branch
Tennessee Valley Authority, Muscle Shoals, Alabama
Under the sponsorship of the Environmental Protection Agency (EPA),
the Division of Chemical Development of the Tennessee Valley Authority (TVA)
has just completed an intensive design and cost study of magnesia scrubbing-
regeneration processes for sulfur oxides removal from power plant stack gas.
This investigation, which has been carried out over the past tvo years, covers
scrubbing systems using both aqueous slurries and solutions.
When published shortly, the study report will be the fourth in a series
of conceptual design and cost studies which TVA has prepared for EPA since 1967*
The first two studies were concerned with the use of lime or limestone as
absorbents, which convert the gaseous sulfur oxides to solid compounds (calcium
sulfite and calcium sulfate) that are discarded. These were called "throw-
away" processes. The third study included processes using aqueous ammonia
solutions as the scrubbing medium and recovering the sulfur oxides as ammonium
sulfites which were converted to sulfate and used as an intermediate in the
production of fertilizer products. This was the first recovery system examined
in which materials could be produced for sale to offset, at least partially,
the cost of operation.
I/ Paper presented at EPA Flue Gas Desulfurization Symposium, New Orleans,
Louisiana, May 1^-17, 1973-
554
-------
Processes recovering the sulfur oxides in a useful form are
potentially superior to the throwaway type because they do not generate
solid waste disposal problems and also offer the possibility that sales
revenue will reduce the cost of sulfur oxide removal. It should be kept
in mind, however, that recovery processes are generally more complex, may
be more expensive to install, and require a definite commitment to sell
the products produced.
In recent years, numerous processes have been proposed for sulfur
oxide recovery and some of these are currently undergoing demonstration on
large scale (100 MW or larger). The purpose of the EPA-TVA conceptual design
series is to subject the more promising of these to a detailed study in which
the best design is developed from available data; capital and operating costs
are estimated on a uniform basis; a market survey is made to estimate sales
revenue; total cash flow is related to economic promise; and needed research
and development are identified.
Scrubbing with magnesium oxide slurry to form magnesium sulfite,
followed by decomposition to produce concentrated SOg (MgO is recycled) is
one of the more promising processes for sulfur dioxide removal. Slurries
of magnesia are good absorbents; however, the most outstanding assets of
the concept are (l) the ease of separation of the sulfite salts formed from
the scrubber liquor, (2) the ability to regenerate and recycle the absorbent,
magnesium oxide, (3) the avoidance of a solids disposal problem, and (k) the
capability of separating, both financially and operationally, the power unit
scrubbing system from the chemical manufacturing and marketing function. At
the same time, the process does require extra expense for (l) drying and
calcining the intermediate MgSOg and MgSOj^ formed, and (2) the apparent need
555
-------
for two scrubbing stages on coal-fired units to avoid mixing fly ash with the
undiBsolved absorbent. As with all aqueous scrubbing processes, stack gas
reheating, if required, would also add expense. The potential of the process,
however, is outstanding enough to merit demonstration on a 155-MW, oil-fired
power unit of Boston Edison. This system, jointly funded by EPA and a large
group of chemical companies and utilities, started up in early 1972.
In the regeneration of the absorbent, sulfur dioxide is released
at concentrations practical for conversion to sulfuric acid, liquified
sulfur dioxide, or elemental sulfur. With the limited market for liquified
sulfur dioxide and the higher cost of conversion to sulfur, the product
receiving primary attention in the report is sulfuric acid. Commercial
grades of acid including 98$ concentration and oleum are easily produced in
the process.
Around the world, development work on magnesia scrubbing for
power plant stack gas has followed at least three major technological routes.
The Russians, Japanese, and Americans have concentrated on the use of mag-
nesium sulflte-magnesium oxide slurries having a basic pH; whereas a German
company, Grillo, has researched the use of an absorbent activator, manganese
dioxide, with the scrubbing slurry. In addition, using technology associated
with sulfite pulping practice, at least one American company has also investigated
the use of magnesium sulfites in acidic solution so that simultaneous particulate
and 802 removal can be accomplished with a single scrubber in coal-fired unit
applications.
Each of these three scrubbing schemes are given detailed review in
the present study and are described as follows:
556
-------
Scheme A • magnesia slurry variation—Wet scrubbing with magnesium
oxide-magnesium sulfite-water slurry to absorb S0g and form undissolved
MgSOyfiHgO plus some MgSOg.SHgO. The MgSO^.fiBUO is thermally converted to
trihydrate and dried to form anhydrous MgSCU. This material along with any
sulfate formed by oxidation is calcined with coke to generate MgO for recycle
and S02 for production of HgSO^ by the contact process. A flow diagram of
Scheme A is shown in Figure 1.
Scheme B - MgO-MnOg slurry variation—Wet scrubbing with magnesium
oxide-magnesium sulfite slurry containing a scrubbing reaction activator,
manganese dioxide. The. sulfites, sulfates, and unreacted manganese dioxide
are dried and calcined to regenerate the absorbent and activator with the
S0g rich gas being processed to HpSO, .
Scheme C - clear liquor variation—Wet scrubbing of stack gas to
remove particulates and absorb SOg simultaneously with an acidic solution
of magnesium sulfites, followed by separation of Insoluble fly ash and
liquor and addition of MgO to the liquor to precipitate MgSOo.&gO. The
crystals of sulfite are then converted to trihydrate, dried, and calcined.
MgO Is recycled and SOg processed to acid.
The above variations can be applied to multiple power units more
economically than with Individual power plants by taking advantage of a
concept called "central processing", fy processing (calcining and acid
production) the dried sulfite material from several scrubbing operation*
in a single, large plant, a more efficient operation (higher annual operating
time) can be derived and economy of scale can be achieved. This concept is
not a technological variation, but deserves separate consideration (Scheme D)
to evaluate the economic merit of the idea which of course, can be used in •
other type absorbent processes as veil.
557
-------
Stack
gag —
Water
Particulate
Removal
System
tn
vi
CO
Thickener
I
Ash to pond
Recycled
Pond
Water
To Stack
H
Sulfur Dioxide
Absorber
Slurry
Devatering
System
Liquor
Water MgO
11
Makeup System
Dryer
Air
Fuel
MgSO,
Recycle MgO
SulfurIc Acid
Plant and/or
Liquified S02
and/or
Sulfur
Calciner
-Fuel
Figure 1 Flov Diagram - Magnesia Slurry Scrubbing-Begeneration
-------
Some earlier preliminary investigations (1970) indicated the
possible use of a variation in magnesia scrubbing for N0x control. Recent
vork (1972), however, indicates that no more than 10$ removal can be expected;
therefore, magnesia scrubbing should not be counted as a means of meeting new
Federal NO emission standards.
A
Study Assumptions
Recovery process economics depend on several factors including
power plant size, type of fuel burned, sulfur content of fuel, operating
factor, plant location, unit efficiency, and unit status (new vs. existing).
For detailed design and cost estimating purposes, it was necessary to assume
a combination of conditions as a base case for both oil- and coal-fired units.
In the economic evaluation, the effect of variation in the major parameters
was determined. The basic conditions assumed are as follows:
fewer unit size, MW 500
Sulfur content of coal, % 3.5
Sulfur content of oil, % 2.5
Ash content of coal, % 12
Heating value of coal, Btu/lb 12,000
Heating value of oil, Btu/lb 18,500
Boiler excess air and leakage, %
coal-fired unit 33
oil-fired unit 15
Degree of dust removal, % 99
Degree of S0p removal, $
slurry Schemes A,B,D 90
solution Scheme C 77
Boiler type Horizontal, frontal-fired
Plant location Midwest
Capactiy factor, % of nameplate
rating
first to 10th year 80
llth to 15th year 57
l6th to 20th year UO
21st to 30th year IT
Avg over life of unit 1*8.5
Air preheater exhaust temp, °F 310°F
Stack gas reheat temp, °F 175°F
Unit heat rate, Btu/kWh 9,000
Product storage, days 30
559
-------
Process Equipment
The scrubbers, ductwork, and fans are the most expensive items
in a magnesia scrubbing process because they must handle the full flow of
gas (over 1,000,000 tons/day for a 1000-MW boiler). The slurry or solution
processing, drying, calcining, and sulfuric acid units handle a lower through-
put of material depending on S content of fuel. Stagewise scrubbing (see
Figure 2) will be necessary when using slurry scrubbing on coal-fired units to
keep the majority of the fly ash from entering the drying and calcining opera-
tions. For particulate removal, a venturi type device using clarified, circu-
lated water is chosen although electrostatic precipitators and bag filters
could be used* Electrostatic precipitators have shown variations in outlet-
loadings due to operating characteristics and time, and bag filters are
more expensive.
In scrubbing sulfur oxides, slurry systems can utilize venturi,
mobile bed (plastic sphere type) or spray units. For solution scrubbing
service (Scheme C}, plate and packed scrubbers might be added to the list
of acceptable devices, but consideration must be given to residual fly ash
carryover causing plugging. In any case, corrosion and erosion protection
should be provided by linings such as rubber or polyester-fiberglass resins.
At this time, mist elimination performance in slurry scrubbing
service is of concern with a variety of designs and materials of construction
currently in use or under study. In this design, the chevron vane type device
constructed of a corrosion-erosion resistant material or coated is used.
560
-------
Figure 2
PERSPECTIVE VIEW OF A MAGNESIA
SCRUBBING SYSTEM FOR SO2 REMOVAL
ON A COAL-FIRED 500-MW POWER UNIT-
TWO-STAGE VENTUR1 CONCEPT
-------
Reheating can be accomplished by indirect steam heat exchange on
new units for which design provisions have been made in the steam cycle;
however, existing units are not likely to have excess steam available; there-
fore, direct fuel oil reheat is preferred. Neither of these methods is the
most economical choice available, but the reliability of the indirect liquid-gas
heat exchange method considered in previous studies has become suspect.
Solids separation in the magnesia slurry processing area probably
can be accomplished best by first thickening the 10$ solids slurry to kOfi
and then centrifuging to a cake containing less than 1556 free water. Since
good test data are not available, separation by filtration cannot be ruled
out; however, cakes containing less than 15% water may be more difficult to
obtain.
Although rotary type devices are being utilized for drying and
calcining in the Boston Edison demonstration project, discussions with
vendors have Indicated that possible greater efficiency and lower cost
could be obtained with fluid bed units. In the absence of test data, some
doubt remains; however, fluid bed systems appear to be the better choice.
The "dry" gas cleanup system for calciner off-gas and the sulfuric
acid plant utilize relatively well established technology; therefore, few
unforeseen problems should arise in these areas. If desirable, the magnesia
process could easily be added to existing acid units which currently burn
elemental sulfur.
562
-------
Investment Requirements
Summarized investment under various combinations of conditions are
given for Scheme A and limestone-vet scrubbing in Table 1. Depending on sulfur
in the fuel, fuel type, plant size, and status, the magnesia Scheme A Invest-
ment varies from k-2Tf> higher than limestone-vet scrubbing; however, salable
product is produced rather than a waste material. Scheme C has the lowest
investment requirement ($36.2/kW for base case) for coal-fired power unit
scrubbing systems; however, SOg removal for Scheme C may not be sufficient in
all cases to meet Federal emission standards for new units. In addition, the
data supporting clear liquor scrubbing are limited; therefore, the scheme should
not be considered as the leading process.
A detailed, direct investment breakdown of the 500-MW base case
covering Scheme A is shown in Table 2.
Operating Costs
A summary of average annual and unit operating costs at TOGO hrs/yr
operation under regulated economics are given in Table 3 for both magnesia
Scheme A and limestone scrubbing (low and high cost). A detailed operating
cost breakdown for the Scheme A base case (500-MW, 3.5$ S in coal, new unit)
is shown in Table h.
Evaluation Considerations
Evaluation of recovery processes brings in factors such as product
marketability and price, return on investment and taxes, and project financial
promise, all of which make the analysis more difficult than for throwaway
processes. It would be desirable, of course, that recovery methods show promise
of a net profit, but this is not essential because recovery should be preferable
to throwaway, even at a net loss, aa long as the loss is lower than for throw-
systems. The cost of limestone-wet scrubbing was used as the criterion for
563
-------
Table 1
Capital Requirements for Magnesia Scheme- A
And Limestone-Wet Scrubbing
Conditions Capital. $/kW of power generating capacity
Magnesia Scheme A Limestone-Vet
vet scrubbing scrubbing
Base case - coal-fired units
(500-MW, new power unit, 3«5$ S in
coal,reheat to 175°F) ^3.5 35.2
Exceptions to base case (coal-fired units)
Existing power unit ^9.3 39.9
2% sulfur 37.6 32.3
5% sulfur 1*8.5 37.8
200-Mtf 58.4 46.0
1000-Mtf 33-1 27.4
Base case - oil-fired units
(500-MVf, new power unit, 2.5# S in
oil, reheat to 175*F) 2k.9 21.4
Exceptions to base case (oil-fired units)
Existing unit 27.8 24.8
1% sulfur 19.8 19.0
4# sulfur 29.1 23.U
200-MW 33.4 28.5
1000-MW 18.8 16.6
561
-------
Table 2 Process Equipment and Installation
Analysis-Direct Cost for Scheme Aa (Thousands of Dollars)
Particulate
scrubbing
Equipment
Material 828
Labor 2*0
Piping & Insulation
Material 327
Labor 177
Ductwork, dampers, &
Insulation
Material 752
Labor Inc.
Concrete-foundations
Material 105
Labor Inc.
Structural
u, Material 135
% Labor 180
Electrical
Material 2l6
Labor Inc.
Instruments
Material 111
Labor 57
Faint
Material 66
Labor Inc.
S02
scrubbing
997
256
236
101
878
Inc.
115
Inc.
1*5
190
3*0
Inc.
135
75
60
Inc.
Slurry
processing Drying
*24
75
52
20
-
-
30
Inc.
26
31*
73
Inc.
(Additional
29
10
12
Inc.
*90
176
2
3
20
Inc.
38
Inc.
8
11
43
Inc.
Calcining
X
665b
215
2
3
44
Inc.
46
Inc.
1*
18
39
Inc.
HgO
slurrying
115
27
15
k
-
-
10
Inc.
3
4
32
Inc.
Instruments)
11
k
k
Inc.
32
11
5
Inc.
39
13
2
Inc.
New
HgSOij.
production
925
Inc.
410
Inc.
741
Inc.
188
Inc.
99
Inc.
207
Inc.
185
Inc.
66
Inc.
ngSO^ Optional by- Fuel oil
storage pass duct storage Total
163
Inc.
.
-
-
—
18
Inc.
3
k
14
Inc.
Inc.
Inc.
1
Inc.
122
22
Inc.
Inc.
454 Inc.
Inc. Inc.
21
Inc.
2
Inc.
10
Inc.
1
Inc.
Inc.
- Inc.
4729
1011
104*
308
2889
Inc.
571
Inc.
435
441
974
Inc.
_• _
5*3
170
216
Inc.
^l^^^BBI*^
Sub-total
Direct costs 319* 3528
785
810
26k
2821
203
178 13,331
ft Hev plant, coal-fired, 500-Mtf, 3.556 sulfur In coal, l,0to,000 scfln stack gas, 378 tpd B^SO^. Inc. = included.
b Includes most instrumentation.
-------
Table 3 Average Annual and Unit Operating Costa
in
o*
o»
Conditions
Magnesia Scheme A
Average
Annual
Cost $
Unit
Operating
i t
Cost $/ton coal
Limestone-Wet
Low Limestone Cost,
On-Site
Average
Annual
Cost $
Solids Disposal
Unit Operating
Cost $/ton
Coal
Scrubbing
High Limestone Cost.0
Off -Site
Average
Annual
Cost $
Solids Disposal
Unit Operating
Cost $/ton
Coal
Base case - Coal-fired units
(500-MW, new power
S In coal, reheat
unit, 3.5^
to 175°F)
7,01*8,900
Exceptions to base case (coal-fired
Existing Power unit 7,762,500
5.0# Sulfur 8,066,600
200 m
1000 MW
3,870,700
10,635, too
5-37
units)
5.79
6.15
7.21
4.19
5,376,300
5,927,900
5,894,000
2,869,200
8,230,900
4.10 7
4.42 8
4.49 8
5.35 3
3*24 12
,621,500
,253,700
,861,700
,633,1*00
,883,100
5.81
6.15
6.75
6.77
5.08
JL
Base case - Oil-fired units
(500-MW, new power unit, 2.5#
S in oil, reheat to 175°P)
4,159,800
$/bbl oil
0.83
3,343,600
$/bbl oil
0.66
JL.
4,112,500
$/bbl oil
0.82
Exceptions to base case (oil-fired units)
Existing Power Unit
lf.0# Sulfur
200 MW
1000 MW
7000 hrs/year operation
Limestone at $2.05/ton and variable on-site disposal costs for calcium solids, ranges from $2.85/ton to $1.33/ton
4,51*8,800
4,973,500
2,305,600
6,317,100
0.88
0.99
1.12
0.65
3,755,100
3>747,300
1,836,700
5,160,400
0.73
0.74
0.89
0-53
4,566,000
5,046,700
2,046,900
6,848,000
0.89
1.00
0.99
0.70
Limestone at $6.00/ton and $6.00/ton for disposal of calcium solids
-------
Table k t Regulated Company Economics—Total Venture Average Annual
Manufacturing Costs for 98% H2SO4 from Scrubbed Power Plant Stack Gasa
Scheme A-Magnesia Slurry Scrubbing-Regeneration
(500-mw new coal-fired power unit, 3.5% S
Annual quantity
Direct Costs
Delivered raw material
Lime
Magnesium oxide (98%)
Coke
Catalyst
Subtotal raw material
Conversion costs
Operating labor and
supervision
Utilities
Fuel oil
Steam
Heat credit
Process water
Electricity
Maintenance
134.1 tons
1,086 tons
763 tons
1,800 liters
39,200 man-hr
5,356,000 gal
440,000 M Ib
20,300 MM Btu
2,207,500 M gal
66,760,000 kwh
i/i fuel; 1 1 0,400 tons/yr 1 00% H2 SO4 )
Total annual
Unit cost, $ cost, $
16.00/ton
102.40/ ton
23.50/ton
1.51/liter
6.00/ man-hr
0.09/gal
0.55/M lbb
-0.40/MM Btu
0.03/M galb
0.006/kwhb
Labor and material, .06 x 21 ,732,000
Analyses
Subtotal conversion costs
Subtotal direct costs
Indirect Costs
Average capital charges at 14.9%
of initial fixed investment
Overhead
Plant, 20% of conversion costs
2,100
111,200
17,900
2,700
133,900
235,200
482,000
242,000
(8,100)
66,200
400,600
1,303,900
85,000
2,806,800
2,940,700
3,238,100
561,400
Cost/ton
of acid, $
.019
1.007
.162
.024
1.212
2.130
4.366
2.192
(.073)
.600
3.629
11.811
.770
25.425
26.637
29.331
5.085
Administrative, research, and service,
1 1% of conversion costs
Subtotal indirect costs
Total annual manufacturing costs
forH2S04
Cost/ton
of coal
burned, $
5.371
308,700
4,108,200
Total
annual
cost, $
7,048,900
2.797
37.213
Cost/ton
of acid t$
63.850
aBasis:
Remaining life of power plant, 30 yi.
Coal burned, 1,312,500 tons/yr-9,000 Btu/kwh.
Stack gas rehca t to 17 5° F.
Power unit on-strcam time, 7,000 hr/yr; acid plant on-stream time, 7,000 hr/yr.
Midwest plant location-1972 costs.
Capital investment, $21,732,000; working capital, $505,600.
''Cost of utility supplied from power plant at full value.
567
-------
comparison. Both high cost ($6/ton limestone, $6/ton solids disposal cost) and
low cost ($2.05/ton limestone, variable cost on-site solids disposal) limestone
systems were estimated using the same basis as used for the magnesia schemes.
The basis on which the recovery process is financed is a major
consideration in evaluating economic promise and acceptability. If a power
company finances the entire project, it can be assumed that the investment
would become part of the rate base on which the company is allowed to earn
what the regulatory authority regards as a reasonable return on investment.
If sulfur oxide removal, either by throwaway or recovery method, were to
increase operating cost, then the price of power to consumers presumably
could be raised to offset the extra cost. Under such basis, sulfur oxide
removal (even by recovery) could be considered as necessary for production of
power just as is the boiler operation, dust removal, or cooling water system,
and the costs, therefore, passed on to the consumer. It is true that rate
increases are often strongly contested and delayed, and that the full adjust-
ment may not always be allowed, so that the power company has the incentive
to avoid extra investment and expense. In general, however, the power com-
pany has a more or less assured profit. For this reason, there is little
risk and capital can be attracted at regulated rates of return.
Since power companies generally are not familiar with chemical
production and marketing, there would be some advantage if a chemical company
built and operated the recovery process for a fee and marketed the products.
For a private, nonregulated company to enter into such an activity, however,
the project would have to be promising enough to attract the necessary
capital from investors. It is difficult to say how much promise is needed
because this varies with the situation. Generally, it is considered that
the projected cash flow (depreciation plus profit after taxes) should pay out
568
-------
the original investment in less than five years or on another basis which takes
into account the time value of money, the interest rate of return after income
taxes should be about 15$. For the relatively high investment required by sulfur
oxide recovery processes, this is a major hurdle.
A characteristic of the central process concept is that features
of both regulated and nonregulated economics can be utilized to advantage.
If by cooperative arrangement, magnesia scrubbing and drying operations are
placed under power industry (regulated) economics, and regeneration and acid
manufacture are covered under nonregulated economics, the power company can
avoid any responsibility for acid production and marketing, and the chemical
company can reduce its capital responsibility to levels more likely to achieve
successful profitability. By charging the power company for regenerated MgO
and also by selling sulfuric acid, enough revenue might be obtained for a
chemical company to Justify manufacturing acid from magnesium sulflte rather
than by the more conventional purchase of elemental sulfur. The revenue
obtained for regenerated MgO would, of Bourse, depend on the resultant cost
to the power company for magnesia scrubbing as opposed to limestone-wet
scrubbing or other feasible alternatives.
Comparison and Profitability
An important consideration for comparison under regulated
economics, and in profitability analysis under nonregulated economics, is
net sales revenue for the sulfuric acid. A market review for the acid
resulted in the following conclusions:
1. The growth rate of sulfuric acid production is about U-6# per year
generally paralleling that of the phosphate fertilizer industry.
Capacity from magnesia scrubbing-regeneration will most likely enter
569
-------
the market at a moderate to slow pace.
2. The best end-use market appears to be the phosphate fertilizer industry
as an acidulant for phosphate rock. The product is used in many other
applications, however, any of which merit consideration.
3' The most promising locations for magnesia scrubbing-regeneration
systems appear to be on waterways serving the areas where sulfuric
acid is now heavily marketed. Areas on the Ohio and Mississippi
Rivers, and along the Gulf and East Coasts are prime spots.
k. Sales price will be based on competition in each individual area
plus flexibility of demand. In those areas where by-product acid
or low cost sulfur are available, competiton will be greatest.
5. Expected net sales revenue after shipping and sales expense are
deducted could average about $8.00/ton of 100$ acid for single-site
systems and $12.00/ton for large central processing units. In the
better locations, these net backs and maybe more should be attainable
through the 1970's.
6. Long term marketing contracts appear to be practical since the
likelihood of escalating sales revenue due to prolonged product
shortages is not expected.
Comparison under regulated economics--A comparison of Scheme A
with both high and low cost limestone scrubbing is given in Table 5- The
values shown are the cumulative net annual costs over the life of the power
unit; thus, the values represent the total bill including return on invest-
ment and income taxes for particulate and SOg control over the power plant
life.
570
-------
5. Cost of Magnesia Scheme A vs. Limestoixe-Wet Scrubbing under Regulated Economies
Conditions
Cumulative net annual costs, $ millions
Scheme A~ Limestone-Wet Scrubbing
__________ Low limestone process coat" High limestone process costc
Base Case - coal-fired units
(500-Mtf, new unit, 3.5$ S
in coal, ^8.556 avg. capacity
factor over 30 yrs, reheat to 175°F)
Exceptions to base case (coal-fired units)
Existing unit (27 yr life)
5% sulfur
200-MW
1000-MW
162.1
160.2
180.5
91.1
237-9
136.2
133.4
148.5
72.7
208.3
170.6
162.3
194.3
82.6
283.2
Base Case - oil-fired units
(50O-MW, new unit, 2.5# S in oil,
48.5# avg. capacity factor over 30
yrs, reheat to 175°F)
Exceptions to base case (oil-fired units)
Existing unit (27 yr life)
4# sulfur
200-MSf
1000-MW
95.9
8k. 2
93-7
111.0
54.4
ilfl.6
84.1
93-8
46.3
129.5
94.2
92.3
112.1
47-7
154.3
a Net sales revenue assumed at $8.00/ton of acid.
k Limestone cost - $2.05/ton; on-site pond disposal of solids.
Limestone cost - $6.00/ton; off-site solids disposal cost, $6.00/ton.
-------
The only magnesia case with costs lower than the low cost limestone
system is a 1000-MW, coal-fired unit using Scheme C, the least developed
variation. Units smaller than 3°0-MW would most likely use limestone scrubbing
if funding were under regulated economics. Because the incremental cost of
producing additional acid exceeds $8.00/ton (net sales revenue for acid),
increasing on-stream time and higher sulfur content of fuel do not improve
the magnesia process economics.
Profitability under non-regulated economics—Based on projected
revenue from acid sales alone, all magnesia cases examined have negative
interest rates of return and no payout. If additional revenue in the form
of a fee equivalent to the cost of limestone-wet scrubbing or other compet-
itive 502 control method is charged by the chemical company for sulfur oxide
abatement, profitability can be derived. Shown in Table 6 are payout periods
in years and interest rates of return in % for Scheme A assuming revenue from
both a fee and acid sales.
As would be expected, the results depend on the size of the fee
charged; for a fee equivalent to a high cost limestone scrubbing process,
desirable profitability could be achieved in some cases and for a smaller
fee equivalent to a low cost limestone process, low profitability would
result in all cases. Funding under this concept will probably be limited.
Profitability of cooperative central process ventures—With the
separation of investment and operating responsibility and the advantage of
economy of scale for large central acid complexes, cooperative ventures
(Scheme D) between power companies and chemical companies are the best
route to financial funding of magnesia systems. Given in Table 7 are the
572
-------
Ul
-»j
w
Conditions*
Base Case - coal-fired units
(500-MV, new unit, 3.536 S in coal
48.5# avg. capacity factor over
30 yrs, reheat to 175*F)
Exceptions to base case (coal fired)
Existing unit (27 yr life)
5% sulfur
200-MW
XOOO-Mf
Base Case - oil-fired units
(500-MW, new unit, 2.5* in oil,
48.5# avg. capacity factor over
30 yrs, reheat to 175°F)
Exceptions to base case (oil fired)
Existing unit (27 yr life)
Itjt sulfur
200-MW
1000-MW
for Pollution Abatement
Low
Payout
7.6
coal
iver
Ired)
7.7
7.7
8.3
7-1
7.2
1,
ver
red)
7.0
7-5
7.6
6.8
equivalent payment1*
Interest rate
, yrs of return, $
8.8
8.4
8.5
7.4
10.0
9.8
10.0
9.0
8.8
10.9
High equivi
Payout, yrs
5-6
5.7
5.4
6.7
4.8
6.1
6.0
5.8
7.1
5.3
Interest rate
of return, j>
14.9
13.6
15.7
11.0
18.1
13.0
12.7
14.0
9.9
15.8
Bet sales revenue assumed at $8.00/ton of acid.
b
Equivalent to limestone-wet scrubbing cost assuming low limestone price, on-site pond disposal of solids.
Equivalent to limestone-wet scrubbing cost assuming high limestone price, off-site disposal of solids.
-------
Table ?.
Profitability of Central Regeneration-Acid Manufacturing Unit Under
Cooperative Economics.* Magnesium Sulfite Supplied from Combinations
of New 200, 500, or 1000-MW Units Burning Coal with 3.556 Sulfur.
Regulated Magnesia Scrubbing Costs Equalized to High and Cow Projected
Limestone-Wet Scrubbing Process Costs.
Payout, years
Interest rate of return, %
Case
Units and size
200-MW equiv.
5 x 200-MW equiv.
10 x 200-MW equiv.
15 x 200-MW equiv.
Recycle MgO
at $25/ton
None
6.6
5.2
4.6
Recycle MgO
at $55/ton
8.4
3.4
2.7
Recycle MgO1
at $25/ton
Neg.
8.2
14.0
17.2
Recycle Mgu
at &55/ton
3-3
26.6
35-5
1*0.6
500-MW equiv.
2 x 500-MW equiv.
4 x 500-MW equiv.
6 x 500-MW equiv.
Recycle MgO1
at $15/ton
None
9-9
7.7
6.5
Recycle MgOc
at $55/ton
5-1
3-5
2.8
2.4
Recycle MgO
at $15/ton
Neg.
0.3
5.1
8.7
Recycle MgOc
at $55/ton,
14.4
25-3
34.1
39-7
1000-MW equiv.
2 x 1000-MW equiv.
3 x 1000-MW equiv.
Recycle MgO
at $10/ton
None
9.9
8.3
Recycle MgO
at $55/ton
3.6
2.9
2.5
Recycle MgO
at $10/ton
Neg.
0.1
3-5
Recycle MgOc
at $55/ton .
24.4
33.0
38.5
a Nonregulated portion of system with 10 yr life; acid revenue - $12/ton.
b
Equivalent to limestone-wet scrubbing costs assuming low limestone price, on-site
pond disposal of solids.
0 Equivalent to limestone-wet scrubbing costs assuming high limestone price, off-site
disposal of solids.
574
-------
payouts and interest rates of return for Scheme 0 systems assuming revenue
from both acid and recycle MgO sales. The price of recycle MgO must be such
that magnesia scrubbing cost does not exceed that of competitive limestone
scrubbing for the same power unit. For a 500-MW, coal-fired unit, only
about $15-$20/ton could be paid for recycle MgO before exceeding the low
cost limestone system; however, approximately $55/ton could be paid if
competition came from a high cost limestone system.
The results in Table 7 indicate that the smaller the power unit
supplying MgSO^ and the larger the acid complex, the better the profitability
which could be achieved. A 3000-MW equivalent acid plant supplied by fifteen
200-Mtf units would show excellent profit making potential - 17.2$ interest
rate of return with $25/ton for recycle MgO and $12/ton for sulfuric acid or
1»0.6f> return for $55/ton recycle MgO and $12/ton acid.
Conclusions
The more important conclusions derived from, this study can be
summarized as follows:
1. Sulfur dioxide absorption by magnesia slurry scrubbing is effective
and the major portions of the process as conceptualized utilize
proven technology.
2. Magnesia scrubbing, like limestone scrubbing, is not an effective
means of NOX removal from power plant stack gas.
3. Magnesia slurry scrubbing-regeneration has been tested in laboratory
and pilot plant stages and at least one large scale demonstration is
underway.
k. Although limited experience is available to guarantee performance,
equipment for commercial systems can be obtained at this time from
vendors and fabricators.
575
-------
5. For moat U. S. fossil -fueled power plants, achievable net sales revenue
for recovered 9&f> sulfuric acid will probably average only $8-12/ton
over the next decade or so; however, there will be applications where
better netbacks are obtainable. Competition will continue from other
sources of by-product sulfuric acid and virgin acid made from low cost
sulfur.
6. Primary economic factors are investment, product volume (depending on
power unit size and sulfur content of fuel), net sales revenue (from
all sources), competitive cost of alternatives and basis of financing.
Raw material, labor, shipping costs, on-stream time, and plant age
are significant, but not nearly as Important as the primary factors.
7* Under total regulated financing, magnesia systems can compete with
limestone scrubbing on larger (toO-MW or greater) power units. The
limestone process would be favored in rural areas (low cost limestone
and space for solids disposal) whereas the magnesia scrubbing-regeneration
process would appear more desirable in crowded metropolitan areas.
8. Total nonregulated industry financing and operation appear unlikely;
however, with a large fee for pollution abatement and large size units,
such funding can be considered.
9* A cooperative venture between several power companies and a chemical
company with each supplying capital for and operating their portion
of the process, appears to be a good way to fund a magnesia system.
It will be necessary for the regeneration-acid plant to charge a
service fee for MgO processed from MgSCU in order to obtain sufficient
revenue for desirable profitability.
576
-------
10. Thus far, Interest in the magnesia scrubbing-regeneration process has
centered more on replacement of sulfur as raw material in existing
sulfuric acid plants with existing markets rather than for added
capacity to meet increasing acid markets.
11. There are a limited number of locations that can support a central
process installation. The Midwest, along the Ohio and Mississippi
Rivers, and the Gulf and East Coasts are prime targets.
12. For short range shipping distances (O-50-miles), the cost of shipping
MgSOo and MgO between sites in a small part of the total process cost
and will not greatly influence process application; however, as
distances exceed 100 miles, shipping cost becomes much more significant.
Additional research and development of the process should be performed
primarily on the demonstration level to determine effect of process factors
such as contamination build-up over long periods of MgO recycle, corrosion-
erosion of construction materials, scaling difficulties, and adaptability
to power plant operation. Some work on oxidation, crystal growth,
effectiveness of additives such as manganese dioxide and the manufacture
of sulfur in the calciner should be performed on the bench or pilot levels.
577
-------
OPERATIONAL PERFORMANCE OF THE
CHEMICO BASIC MAGNESIUM OXIDE SYSTEM
AT THE BOSTON EDISON COMPANY
PART I
by
George R. Koehler
Chemical Construction Corporation
New York, New York
This paper is Part I of a two part paper and supplements
Part II of this paper being presented by Mr. C. P. Quigley
of Boston Edison Company.
579
-------
The work upon which this publication is based was
performed pursuant to Contract No. CPA 70-114
with the Environmental Protection Agency.
580
-------
OPERATIONAL PEFORMANCE OF THE
CHEMICO/BASIC MAGNESIUM OXIDE SYSTEM
AT THE BOSTON EDISON COMPANY
PART I
G. R. Koehler
INTRODUCTION
The necessity for control of sulfur oxides emissions to the atmosphere
has become increasingly apparent. The deleterious effect of this compound
on human health and the environment has been documented. The growing de-
mand for energy compounds the problem. Total SO2 emissions in 1965 were
25 million tons per year. It is anticipated that within the next decade, if no
control is imposed, these emissions will rise to 75 million tons per year
in the United States. Of this amount, two-thirds will originate from the pro-
ducts of combustion of fossil fuel fired power generating stations. Of the avail-
able technology, one of the most promising seems to be flue gas scrubbing at the
generating facility itself.
In its search for suitable technology to control sulfur oxides emissions
CHEMICO developed a process scheme which uses the well established techno-
logy of acid gas absorption by alkaline scrubbing media with precipitation of the
sulfite formed as a solid product. Further process steps include dewatering of
the solid product and eventual reclamation of the alkali with recovery of the SO2.
In June 1970, the United States Environmental Protection Agency and the
Boston Edison Company agreed to provide the funds for a large prototype Sulfur
Dioxide Recovery Plant using the Chemico/Basic MgO Sulfur Recovery Process.
These integrated facilities were comprised of an SO2 absorption plant at Boston
Edison's Mystic Station and a regeneration facility at Essex Chemical's Rumford,
B.I. Sulfuric Acid Plant. Construction at both plants was completed by April,
1972.
Capital funds for the absorption system were provided by the Boston Edison
Company while E.P. A. furnished funds for installation of the magnesium oxide
regeneration system and appropriate acid plant modifications as well as opera-
ting funds for the absorption/regeneration systems.
581
-------
The SO^ absorption plant is designed to remove 90% of the sulfur dioxide
formed when burning fuel oil containing two and one half percent sulfur. The
regeneration facility is designed to recover the alkali in a form suitable for re-
cycling to the absorption facility and send the recovered SC^to a sulfuric acid
plant for the production of 98% sulfuric acid.
Operations to date have demonstrated each of the process steps. SO re-
£
moval efficiencies of 90% or better have been consistently obtained despite the
use of a lower sulfur fuel oil than originally expected. Several hundred tons
of commercial grade sulfuric acid have been produced and marketed in the con-
ventional manner. Operations have, however, been intermittent due to the nec-
essity to make numerous relatively minor changes and adjustments to the system.
Yet to be demonstrated is the long term operability of the process and the ability
to use Magnesia which has been recycled many times. Present efforts are being
concentrated on entering into long term operations in order to ascertain system
availability and reliability.
Problems associated with process consideration are addressed in this Part
I. Mr. C. P. Quigley of Boston Edison will describe mechanical problems en-
countered, corrosion—errosion experience and related equipment malfunctions
in the second part of this paper.
GENERAL PROCESS DESCRIPTION
Chemico/Basic's Magnesium Oxide System for the recovery of sulfur dio-
xide from power plant flue gases is shown schematically in Figure 1.
This process, which utilizes the sulfur dioxide absorption characteristics
of an aqueous slurry containing magnesium oxide, magnesium sulfite and
magnesium sulfate is comprised of five primary operations:
Absorption System
Centrifuge System
Dryer System
Magnesium Oxide System
Calcination System
582
-------
Absorption System
The flue gas containing sulfur oxides enters a venturi absorber (Figure
2) and contacts the absorbing media which is an aqueous slurry magnesium
oxide, magnesium sulfite and magnesium sulfate. The process of SO2 re-
moval is explained by conventional mass transfer principles. The venturi
absorber can be shown by analogy to correspond to a co-current packed vessel.
In the scrubber, the liquid slurry is injected and flows on surfaces over
which passes an accelerating gas stream. The high velocity gas passing over
the liquid causes wave motion on those surfaces. The waves increase in ampli-
tude finally dispersing as fine droplets in the gas stream. Thus, the whole mass
of liquid is dispersed in the form of atomized droplets.
In the process described in this paper, the dispersed droplets have a median
size of 400 micron and the surface area available for mass transfer averages
13 ft 2per ft3 of gas. This absorption surface is dispersed thru and flows with
the gas stream eliminating the problems of plugging associated with conventional
packed towers (the surface area per unit volume is approximately equivalent to
dumped 3 inch rachig rings). Due to system dynamics this surface area relation
is relatively invarient over wide turn-down ratios in the venturi scrubber and
can be used as an approximation over the power plant boiler's operating range
of 40 to 155 MW.
Similarly, system efficiency can be predicted with a fair degree of accuracy
by substituting a Sherwood Number of 2 (SH =2) in conventional mass transfer
relations. Using this estimate a maximum efficiency of 96% can be predicted
for the system, deviations from that removal efficiency being caused by equili-
brium partial pressures of SO^ over the droplet surfaces greater than zero in
the dynamic system.
To date, the absorption system has operated satisfactorily. Sulfur dioxide
removal efficiencies of 90% or greater have been consistently obtained with
SO2 concentrations in the outlet flue gas being generally 90 ppm or less measured
by both a DuPont 460 process analyzer and wet chemical methods. No scaling
or pluggage in the venturi absorber has been encountered despite almost conti-
583
-------
nual operation of this equipment in a recycle mode during the duration of the
operations at the Mystic Station.
Centrifuge Systern^
A bleed from the absorption system (Figure 3) enters the centrifuge where
the solids in the slurry are separated and the centrate is returned to the absorp-
tion system. This bleed stream is controlled to maintain a constant solids con-
tent in the recycle slurry and to remove product magnesium sulfite and any un-
reacted magnesium oxide and precipitated magnesium sulfate. The system is
operated so that the absorbed SO is removed as an equivalent amount of magne-
sium sulfur compounds. A Bird solid bowl centrifuge is used to dewater the
solids preparatory to drying. Except for a few isolated incidents of shear pin
breakage, this item of equipment has operated satisfactorily, yielding 50% or
greater removal of the solids in the slurry bleed stream.
Dryer System
The wet centrifuge cake containing magnesium sulfite, magnesium oxide
and magnesium sulfate plus carbon and other solids removed in the venturi
absorber-centrifuge system is passed to a rotary counter-current dryer to re-
move both unbound water and water of crystallization.
The dryer with its associated feed conveyor and product conveying system
has been the most frequent cause of the operational difficulties which have caused
shut-down of the complete system. The problems encountered have ranged from
an initial complete carry-over of dryer product into the off-gas due to excessive
dryer gas velocity to an inability to convey the dryer product resulting from
auxilary equipment breakdown. In order of occurrence these problems and their
subsequent solutions have been:
1. Total carryover of dryer material in the exit gas stream.
The dryer was designed to process a material of much
greater crystal size than encountered in actual operation.
High draft resulted in a gas velocity which exceeded the
pneumatic transport velocity of the dried material. Dryer
action placed all of the dryer inventory in the air stream.
All lifter flights were modified or removed in order to re-
duce internal dusting in the dryer. Normal operating draft
584
-------
was reduced from a design of 1" to a range of 0.1" to
0.01" (this caused subsequent problems in draft con-
trol) .
Adherance of centrifuge cake to dryer walls causing
plugging of the dryer. This was caused by a change
in the consistency of the centrifuge cake from a dry,
fine sand-like material to a wet, fluid mud-like material.
Increasing exit gas temperatures to increase feed end
surface temperatures resulted in granulation of the feed
(this problem of adherance was a recurring one and was
associated with other properties of the centrifuge cake as
described in later sections).
Granulation of the dried material resulted in a nonuniform
product. Many of the granules were several inches in dia-
meter and jammed product materials handling equipment.
The cross screw conveyor and weigh belt at the top of the
product silo were bypassed and a lump crusher was installed
at the boot of the bucket elevator to reduce the size of any
large granules to 1" or less.
Disintegration of the hydrated crystal during drying caused
excessive production of "fines" (Figure 4). Heavy loading
of fines in the dryer off gas caused excessive carryover of
particulate matter from the cyclone emission control equip-
ment installed with the dryer. Dryer off-gas was ducted to
introduce it to the venturi absorber after initial clean up in
the cyclones. The venturi absorber, operating at a five inch
pressure drop across it's throat provides satisfactory parti-
culate emission control of the dust carryover from the dryer
Further complications were anticipated as the dryer off-gas
was expected to furnish 15°F of reheat to the saturated flue gas
after sulfur oxide removal. With the dryer gas now diverted to
the inlet of the absorber during normal operations, it now enters
the stack at saturation temperature. To date no raining or pre-
585
-------
cipitation from the stack has been noted despite exit
gas velocities of nearly 80 ft.per second. A careful
watch is being kept over this part of the operation and
initial design of a reheat system has been completed
should it be needed.
5. Dust accumulation on a baffle installed over the spill
back bin of the dryer built up to restrict the flow of gas
from the dryer. Initially, this material was manually
pushed from the shelf. Recently the feed screw conveyor
has been lengthened to introduce the centrifuge cake further
into the dryer and a dust collection system was installed to
transport dust accumulated in the spill back bin to the product
silo, allowing removal of the baffle.
6. Addition of cyclone underflow to the centrifuge cake fed to the
dryer caused setting of the centrifuge cake accompanied by
fouling of the screw and overloading of the dryer feed conveyor
motor. Cyclone underflow was diverted from the feed screw con-
veyor; recently a dust collection system was installed to remove
cyclone underflow to the product silo.
7. Operation with recycled MgO caused a reoccurrence
of dryer feed sticking on the dryer walls which eventually
pluged the dryer. It is felt that this was caused in part by a
higher percentage of unreacted MgO in the centrifuge cake due
to inadequate slaking of the recycled MgO produced in the early
reclamation attempts. Continued efforts to improve the quality
of the recycled MgO in the calcination facility, installation of
"knockers11 on the dryer feed end, provision for heating the
MgO slurry tank to improve the slaking of the recycled MgO
(Table 1) appears to have eliminated this problem.
586
-------
Magnesium Oxidejystem^
The anhydrous magnesium sulfite and magnesium sulfate produced in the
dryer is conveyed to a storage silo before transportation by truck to the re-
covery acid plant. The same transportation facilities are used to return re-
generated magnesia to the magnesium oxide silo at the power plant. No pro-
blems of pluggage of these silos has been encountered to date. The measured
rate of rehydration of the magnesium sulfite on standing in the atmosphere is
relatively low, Figure 5. The "lump" crusher installed upstream of the silo
allowed the passage of very large pieces (1-1/2" x 2IC). These pieces proved
difficult to handle in the pneumatic trucks and Rumford's conveying system.
A scalping screen was installed in the truck loading chute and oversize material
is conveyed to the load-out belt by additional conveying equipment.
Recycled alkali absorbent and make up magnesium oxide are fed with water
to an agitated tank where a slurry is prepared for introduction to the recycle
system manually on pH control. Numerous pluggage problems were traced to
tramp material entering at this point.including granules of magnesium sulfite
not completely removed from the truck in the transport operation. Initially,
system protection was provided by interposing a vibratory screen at the dis-
charge of the weigh belt feed. Eventually the necessity to slake the recycled
MgO at higher temperatures in order to increase its reactivity required the
addition of an MgO tank equipped with seal legs to prevent steaming in the weigh
house. This device now serves as a "tramp" materials trap.
Calcination System
The dry product transported from the power plant is received, weighed and
pneumatically conveyed to a storage silo. It is fed to a direct fired rotary cal-
ciner at a metered rate, and calcined to generate sulfur dioxide gas while re-
generating magnesium oxide. Coke can be added to provide a reducing atmos-
phere to reduce the residual magnesium sulfate to MgO and sulfur dioxide. The
hot flue gas containing sulfur dioxide and MgO dust enters a hot cyclone where
essentially all the MgO dust is returned to the calciner. The flue gas then
inters a venturi scrubber for final MgO dust cleaning. At the same time the
is cooled and adiabatically saturated.
587
-------
At the Essex Chemicals installation, the resultant product from the
recovery of the sulfur dioxide is 98% sulfuric acid. The saturated flue
gas is cooled in a direct contact cooler to meet the requirements of the
acid plant water balance. The cleaned, cooled flue gas then directly enters
the drying tower of an existing 50T/D acid plant.
The regenerated magnesia is cooled, conveyed to the magnesia storage
silo and recycled back to the power plant site for reuse.
As would be expected in a plant of this complexity numerous operating
problems occurred. These problems and 'their solutions were:
1. Excessive leakage at the seals of the rotary calciner
preventing the attainment of the required neutral or
reducing atmosphere. All seals had to be remachined
to provide minimum clearances. Careful attention was
required to see that all view and instrument ports are
closed during operation. Finally a new seal was designed
and installed to completely correct the problem.
2. Granular nature of dryer product caused material handling
problems which required the installation of heavy duty
belts on weigh feeders to prevent ripping of the belts.
3. Balancing the draft requirements of the calcination facility
and acid plant caused "puffing" at feed hood and weigh house.
Modification of the gas exit transition of the calciner to provide
smooth gas flow was one of the steps taken to reduce pressure
loss.
4. Gritty nature of calciner product caused unreactivity in reuse
at Boston Edison. Installation of pulverizing equipment to
provide a grind of 100% thru 100 mesh and 70% thru 325 mesh proved
satisfactory.
588
-------
5. Tramp materials in calciner product caused
breakage of pulverizer. Installation of magnetic
separator and vibrator screen at screw conveyor
discharge eliminated this.
6. Severe dusting in calciner obscured flame, tripping
flame safety controls and interrupting operations.
Installation of secondary "flame scanner" and exten-
sion tubes partly alleviated problem. Eventual employ-
ment of full time operator on the firing platform during
operation reduced the number of shutdowns.
7. Inability to use coke. Initial charge of coke received was
off specification, and contained 47% ash. Fear of contami-
nation of the calcined product and the acid plant precluded
its use. For several months the calcining facility was
operated without the introduction of additional carbon and
attempts were made to adjust the fuel ratio to compensate;
this resulted in a harder "burn" of the product magnesium
oxide. Finally a coke of less than 10% ash was substituted
and coupled with the improved operations detailed above
produced dramatic results. A comparison of the two cases
is given in Table 2.
DESCRIPTION OF INSTALLED FACILITIES
Boston Edison's Mystic Station located in Everett, Mass, is presently
comprised of six oil-fired power generating units having a total rate capacity
of 619 MW. All units are equipped with electrostatic precipitators which are
presently de-energized as the station burns fuel oil exclusively instead of
coal. An additional unit which will be nominally rated at 600 MW is now under
construction.
The Chemico/Basic Magnesium Oxide System is operated on Unit No. 6 of
this installation. This unit has a rated capacity of 150 MW and was placed in
commercial operation in May 1961. The boiler is a Combustion Engineering
controlled circulation unit with a rating of 935,000 pounds per hour continuous
589
-------
capacity at 2,150 psi max. The unit burns 9700 gallons per hour
of No. 6 fuel oil at full rated capacity. Design criteria for the scrubber
installation are given in Table 3.
PROCESS CHEMISTRY
Magnesium oxide is introduced into a mixing tank with water where con-
version to the hydroxide commences. This mixture is added to the recycle
stream of the venturi absorber where it is contacted with the entering flue
gas containing sulfur oxides. The reaction between these two components
produces principally magnesium sulfite, some of which is oxidized to magnesium
sulfate. Conventional representation of the chemical reactions is shown in
Figure 6.
A better representation of the postulated reaction mechanism shown in
Figure 7 which indicates that the formation of the magnesium sulfite proceeds
thru the reaction of magnesium ion in solution and sulfite ion in solution. The
product of this reaction is a salt of low solubility (MgSOQ solubility 15 gm/liter
o
versus MgSO^ solubility of nearly 1800gm/liter). As mentioned in the previous
section a bleed stream is sent to a centrifuge where the crystalline magnesium
sulfite is removed in a quantity sufficient to balance the incoming sulfur oxides,
while the centrate (a slurry containing magnesium sulfite nuclei) is returned to
the absorption unit.
Upon initiation of operations from a cold start, a major change in the con-
sistency of the centrifuge cake takes place after a period varying from 8 to 16
hours. The cake changes from a dry, fine sand-like consistency to a thin mud.
Quantitatively, during the initial period the centrifuge cake may contain as much
as 65% crystals of size 40 to 80 micron while in operations after the first several
hours the particle size of the centrifuge cake is less than 15 micron.
It was also determined that the state of hydration of the crystalline magnesium
sulfite varied between these two conditions. The larger particle size material
was associated with a higher percentage of MgSOg. 6H-O while the centrifuge
cake containing more of the finer crystals contained approximately equal amounts
of MgSO3. 6H2O and MgSO,. 3H2O. The distribution of this latter case is shown
in Table 4.
590
-------
Initially, the slurry had been held at 120°p for several days with no
flue gas scrubbing. The complete absence of MgSO3.6H2O in the slurry
confirms that at temperatures above U0°p only MgSO3-3H2O is stable,
and the hexahydrate formed has a metastable existance.
The proposed mechanism for this phenonema is shown in Figurte 7.
The existance of the (Mg. 6H2O) coordinate ion is known and has been
identified by infrared spectroscopy. Apparently this material dehydrates in
solution at elevated temperatures to form the trihydrate. A proposed alter-
nate path shows the formation of the coordinate ion 3.3H2O.
In the system as currently operated, the compounds crystallize as dis-
tinct entities. The magnesium sulfite hexahydrate as a rhombic crystal and the
trihydrate of the same salt crystallizes as a trigonal pyramid.
Operating difficulties have been associated with this transition of the centri-
fuge cake. However, adherance of dryer feed to the dryer walls has also
occurred while drying a centrifuge cake containing 40 to 80 micron size crystals.
Dryer heat load is not adversely;affected by the transition during start-up as the
MgSOg.3H2O contains half as much water of crystallization as the hexahydrate
and therefore an equivalent amount of unbound moisture can be removed in the
dryer without affecting process economics.
A comparison between the average stream analysis obtained during operations
with virgin magnesia and recycled regenerated MgO produced prior to sustained
coke introduction is given in Table 5. Recent operations incorporating all the
improvements outlined thus far have shown a return to a stream analysis
similar to that for virgin MgO; most importantly, a reduction in unreacted MgO
in the centrifuge cake to 1. 4% instead of the higher average value indicative of
hard burned magnesia.
FUTURE INSTALLATIONS
With limited operation of the plant and lack of data on the effect of repeated
recycle of the absorbing alkali it is difficult to determine what other modifica-
tions might be incorporated in a new plant. Based on the initial operating ex-
perience at Boston Edison's Mystic unit it appears that future installations
should have additional "lump breakers" and screening equipment located at
591
-------
at strategic points in the plant. These should be installed at the MgO tank in
the absorber recycle stream at the dryer product discharge and the magnesium
sulfite storage silo discharge.
The type of centrifuge cake encountered up to the present time could be
handled readily in a dryer of different design. A dryer incorporating internal
chains, external rappers, lowered gas velocity and provisions to reduce the
internal dust recycle would be suitable.
The calcining facility should incorporate pulverizing equipment of more ro-
bust design in order to ensure more continuous operation with a grind suitable
for reuse in the absorption system.
POTOMAC ELECTRIC POWER CO. - DICKERSON #3
The previously mentioned provisions have been incorporated into-the
pollution control facility currently being constructed for Potomac Electric Power
Co. This plant is a coal fired unit and the venturi absorption system has been
designed to handle the flue gas equivalent to 100 MW from the boiler (Figure 9).
Flue gas containing SO2 and fly ash passes into the first stage of a two stage
venturi scrubber where fly ash is removed using recirculated water as the scrub-
bing media. A bleed stream from the scrubber is thickened to concentrate the
fly ash as a slurry underflow which is pumped to a disposal area. Overflow from
the thickener is returned to the scrubber circuit for reuse. A bypass flue gas duct
has been installed around the precipitator to the scrubber absorber. This will
allow the introduction of clean flue gas containing little fly ash into the first stage.
Dampers have been provided in these ducts which will allow the scrubber absorber
to handle either flue gas partially cleaned in the precipitators or flue gas coming
directly from the air preheaters. Thus the operation will provide the means of
evaluating wide ranges of particulate emissions on the magnesia slurry system for
the design of new plants and retrofit systems.
Flue gas leaving the particulate removal system enters the second stage
of the two stage venturi where it is contacted with the aqueous slurry containing
magnesium oxide, magnesium sulfite and magnesium sulfate and from that point
follows the process description of an oil fired magnesium oxide system.
592
-------
In addition to the incorporation of additional size reduction equipment in
this operation the principal difference is in the design of the dryer for the
system. The dryer supplied is a co-current unit, equipped with flights of
internal chains, some external rappers, toothed lifter flights and a diameter
sufficient to reduce the gas velocity to one half that obtained at the Mystic
Station.
While it is anticipated that any additional advantages incorporated at
Boston Edison will be added to the Potomac Electric facility it appears at this
time that they will be minor ones.
593
-------
Test Coupon Location
FIGURE 1
-------
ANNULUS
SPRAYS
CONE
WASH
TANGENTIAL
NASH
CLEAN GAS
OUTLET
TO STACK
INTERMITTENT
MIST
ELIMINATOR L— X
SPRAYS "
--NORMAL LIQUOR LEVEL
PUMP
SUCTION
Figure 2. Sectional view of absorber at Boston Edison Company. Boston, Mass.
595
-------
KOUCEO- DRAFT FAN
CONVEYOR
tn
(0
0
ELEVATOR
CONVEYOR I
f EffiH FEEDER
FMi CAL.CIIIWG
SYSIE»
FROM DISCHARGE EXISTING MUCED4RAFT FAN
•OTHER
LIQUOR
PUB'S
CEXTRIFUGE STSTEN
DRYER SYSTEM M»j + IfSCM + «fO)
Figure 3. Process flow<9uoram toi magnesia slurry SOj recovery system
at Boston Edison Co., Boston. Mass.
-------
L JrystalE - plus'200 rsesh 'i-accion
IOOX
Aiii.yc'rons I.eSC. - plus 200 :nesh Tractiou dried
IOOX
Fiqure 4
597
-------
tn
(£>
CO
g
u
S.
of
8 10 12 14 16 18 20 22 24 26 28 30
Figure 5. Percent water absorbed by anhydrous MgSOs on exposure to air.
-------
CHEMISTRY OF MAGNESIA SLURRY S02
Ul
to
to
RECOVERY PROCESS
ABSORPTION
MAIN REACTIONS
MgO + SO + 3 HO •* MgSO • 3 H 0
— +• «j f,
MgO + S02 + 6 H20 -*
SIDE REACTIONS
S09 + H~(
H £
2 + MgO
-T + 7 H (
3 -* Mg(HS03)2
-> 2 MgS03 + H2
) -> MgSO. - 7
MgO +
+ 1/2 02 + 7 H20 -v MgS04
Figure 6
-------
POTOMAC ELECTRIC POWER CO. PROTOTYPE
PRECIPITATOR/SCRUBBER - ABSORBER
MgO ADDITIVE SYSTEM FOR S02 RECOVERY
SCHEMATIC PROCESS FLOW SHEET
BY-PASS
SCRUBBER/ABSORBER
STACK REHEAT
ALTERNATIVES
• FUEL BURNERS
• STEAM COtLS
ELECTRO-STATIC
PRECIPITATOR
TO DRY ASH HANDLING SYSTEM
DUST COLLECTOR
CENTRIFUGE
MgO I i
SILO ' *
RECYCLED
POND WATER
MOTHER
LIQUOR
TANK
SLURRY
TANK
TRANSFER
TANK
CRYSTAL
DRYER
MgO FROM ACID PLANT
MgSO3 TO ACID PLANT
-------
TABLE 1
Effect of slaking temperature on recycle MgO utilization
Slurry Tank
Temperature
58° F
108°
180°
MgO Hulk
Density
Ib/ft3
32.3 "
Unreacted MgO In
Centrifuge Cake
17*
1.
TABLE 2
Effect of coke addition (4- day continuous operation with coke
feed interrupted for 12 hours.)
With coke addition
Without coke addition
Product Bulk
Density
2^.7 lb/ft3
68.2 lb/ft3
SQ2 Rating
21.9
601
-------
Table 3
UNIT NO. 6 S00 RECOVERY SYSTEM
OPERATING CRITERIA
GAS FLOW, ACFM 425,000
GAS TEMPERATURE, °F 300
DRY GAS, LB/MIN 21,369
S WATER VAPOR, LB/MIN 1,620
S02, LB/MIN 63.2
FLY ASH, LB/MIN 0.91
S02, PPM (DRY BASIS) 1,410
FLY ASH, GR/SCFD (DRY BASIS) 0.0228
FUEL OIL, CONSUMPTION, GAL/HR 9,700
FUEL OIL SULFUR CONTENT, % 2.5
-------
Table 4 OBSERVATION OF CRYSTAL SPECIES DISTRIBUTION
Sample
% Solids
07202-05-AT08
07202-09-AT08 6.52
§ 07202-11-AT08
LO
07202-14-AT08
07212-02-AT08
0712-06-AT 08
+ 200 Mesh
{Vol. %)
4.39
6.52
8.03
8.82
6.25
6.3
0
0.65
1.45
1.05
0.65
0.55
Composition {%)
- 200 Mesh
•MgS03 MgS04
Tri Hex
100
50
51
56
43.3
49.5
50
34.
24.
40.
36.
6
3
9
5
4.
6.
4.
6.
4
4
0
2
20
12
16.
14.
18.
2
0
0
% Moisture
Centrifuge Dryer
Cake Product
8
24
13
-------
Table 5 STREAM ANALYSIS
S02 Abatement System, Mystic Station #6
Slurry
Solids
MgO
MgSO:,
MgSOjJ
Filtrate
Centrifuge Cake
Solids
MgO
MgSOo
MgSOif
Centra te
Solids
Dryer Product
Solids
MgO
Average Compositions
Fresh MgO Feed Recycle MgO Feed
10;8
6.2
55.51
85*0
1.36
57.85
Mr. 12:
87.2
2.6
8^,63
11.9
6.83
59.61
10.8
83.0.
58.03
2.59
. 67
95.66
Acid Insoluble
.17
7.68
0.55
-------
OPERATIONAL PERFORMANCE OF THE
CHEMICO MAGNESIUM OXIDE SYSTEM
AT THE BOSTON EDISON COMPANY
PART II
by
Christopher P. Quigley
Boston Edison Company
Boston, Massachusetts
This paper if Part II of a two part paper and supplements
Part I of this paper being presented by Mr. G. Koehler
of Chemical Construction Corporation.
605
-------
OPERATIONAL PERFORMANCE OF THE
CHEMICO MAGNESIUM OXIDE SYSTEM
AT THE BOSTON EDISON COMPANY
PART II
C. P. Quigley
INTRODUCTION
The magnesia scrubbing system on Boston Edison Company's Mystic
Unit #6 has now been in operation for over a year. From the utility company's
point of view we consider the results obtained to date to be a combination of
successes and failures.
Success has been achieved in that the process has been demonstrated
chemically. The scrubber can consistently remove high levels of SO2 from the
flue gas. The sulfites produced have been successfully calcined at the acid
plant. The SO2 liberated in the calciner has been used to make 98% commercial
grade acid. The regenerated magnesium oxide has been reused successfully in
the scrubber for short periods of time. Further, the scrubber itself has re-
mained free of the scaling and plugging problems that have so seriously hindered
the development of calcium scrubbing systems in this country.
In spite of these successes in demonstrating the chemistry of the
process on this large scale, the equipment and the process have failed to dem-
onstrate a reasonable availability commensurate with power plant equipment needs.
In this first year, total scrubber operating time has not exceeded 1200 hours.
Availability has been a very low 177. of the Unit #6 operating hours. The longest
sustained run has been four and one half days. On regenerated magnesium oxide,
the longest run has been limited to two days.
OPERATING EXPERIENCE
Process Chemistry
Of the problems encountered in the past year, the most serious have been
1) related to the crystal size produced in the scrubber, 2) the inability to date to
606
-------
line out the calciner operation so as to produce an acceptable absorbent in
terms of sulfa.ce content and magnesium oxide reactivity and 3) the scaling of
dryer internals currently occurring when scrubbing with recycled magnesium oxide.
These problems are addressed in detail in a supplement to this paper presented at
this symposium by Mr. George Koehler of the Chemical Construction Corporation.
Beyond these process problems, other problems have occurred although
they must be considered relatively minor and more readily correctable in nature.
The following outlines some of the experience with equipment components during
this initial running period.
Corrosion/Eroaion Experience
The equipment components and piping of the scrubbing system are fabricated
mainly of carbon steel. Therefore, considerable attention was paid to corrosion
control. The internals of the scrubber and the flue gas ductwork have been coctftd
with corrosion resistant materials. pH of the slurry is maintained normally above
7.0 with only a few short duration tests being conducted at about 6.0 during the
year. A few excursions to 4.5- 5.0 pH have occurred for very short periods.
The internals of the scrubber and the flue gas ductwork leaving the
scrubber have been coated with a spray-applied polyester lining. This lining has
stood up well. Erosion has been noted only at the leading edge of the venturi
annulus. This annulus is subject to direct impingement by slurry crystals on its
relatively sharp-edged top surface. Some metal wasteage has occurred where this
coating has been eroded. A cladding of the top surfaces of the annulus ring with
stainless steel should correct this problem.
The flue gas ductwork to the scrubber has been protected with a I*1 lining
of a corrosion Inhibiting gunlte. During this first year, no corrosion of this
ductwork has bean experienced although the uncoated fan casings of the new scrubber
booster fans were corroding severely in the same environment. Subsequent guniting
of these fan casings has eliminated this problem.
607
-------
Viton expansion joints were installed in the flue gas ductwork as a
corrosion protection measure and have stood up well to date.
Severe corrosion/erosion of cast iron pump impellers occurred in the
slurry recirculating pumps. These impellers required replacement within the
first six months of intermittent operation. New 316 stainless steel impellers
have been installed and show no corrosive attack at this time.
Ring header piping at the top of the absorber feeds slurry to several
tangential wash nozzles above the venturi throat. Pip.ing in this header has
experienced heavy wall thinning at points where slurry sharply changes flow
direction, such as at reducers and at tees. Several leaks have resulted requiring
maintenance welding. No metallurical evaluation has yet been made to determine the
cause but erosion due to slurry impingement is believed to be responsible.
Scaling and Plugging Experience
A small ring of scale has deposited at the interface between the enter*
ing flue gas and the slurry liquor at the top of the scrubber. This scale deposit
has been too small to affect the flows of slurry or flue gas and does not appear to
increase in size. No maintenance work for removal has been required. The remainder
of the scrubber, a rather open vessel and the demisters have shown no evidence of
scale or deposit buildups. The pressure drop across the demister has not increased
since initial operation.
Scale like deposits have been experienced in dead end legs of magnesium
oxide recirculating piping. This occurrence resulted when steam piping used for
freeze protection was activated this winter. It is believed that over heating of
this static slurry accelerated hydration and resulted in the buildup of hard scale*
like deposits plugging the line. Provisions for rodding at pipe ends and provisions
for bleed piping to prevent static slurry buildups have been installed.
Slurry settling to the bottom of inactive vertical discharge lines at the
slurry recirculating pumps resulted in plugging of these lines. The material de-
posited and built up at the closed pump discharge valves. Small bleed piping was
608
-------
installed above each valve to appoint of lower hydraulic pressure at Che common
suction header to the pumps. The problem has been alleviated somewhat by providing
a continuous flow path for the settling slurry. In addition, when the system is
shutdown, one of the recirculating pumps is operated continuously to move the
slurry in the recirculating lines.
The tendency of the centrifuged wet cake to scale has been the process
phenomena that has inhibited continuous operation of the scrubber system. Hard
scale deposits in the centrifuge, the ribbon conveyor feeding the dryer and the
dryer itself have plagued this installation.
Hardened scale deposits left in the centrifuge after shutdown have
caused repeated shear pin failures and centrifuge motor overloads on startups.
Provisions to backwash the centrifuge were installed and the centrifuge is washed
at each shutdown. The problem is still present but appears now to be manageable.
Both rubber housings and steel housings have been tried with relatively
poor results in an effort to cope with scale deposits in the dryer feed conveyor.
This conveyor continues to be a high maintenance item and a source of many system
shutdowns.
The scaling in the rotary dryer of course has been the greatest problem
of all and until resolved, a austained run of this plant cannot be achieved.
Winter Conditions
Winter operation pointed out deficencies unique to cold weather operation.
Freezing and failure of steam traps on steam tracing lines, freezing of control air
lines because of inadequate air drying equipment and discomfort for operator and
maintenance personnel required to work for long periods on equipment outdoors
had to be faced.
Miscellaneous Problems
Flue gas dampers are installed to isolate the scrubber from the gas
path to the stack. These dampers are driven by pneumatic power drives. Several
occasions of sticking dampers have been experienced and on three occasions,
609
-------
misoperation of these dampers resulted in an unsafe condition in the boiler
requiring emergency tripout of the unit. New more powerful damper drives are
about to be installed, in the meantime, operating procedures are in effect to
assure proper damper operation so as to prevent the cutoff of a flue gas path
to the stack by the failure of dampers to operate.
Spillage and Housekeeping
The conveying and handling of the magnesium sulfite and magnesium oxide
material, much of which is of dust size, the removal of scale buildups in the
dryer, the occasional spillage of -slurry through the scrubber overflow line, and
the bleeding of plugged piping have all contributed to a continuing housekeeping
problem and an excessive maintenance cost relative to cleanups.
Operation Staffing
Operation at Mystic has been staffed by two Edison operators on a 24
hour basis. Chemico has supplemented this force with one man coverage around the
clock for operational guidance and test purposes.
Operation at the calciner facility is covered by two Essex Chemical
operators and by one Chemico operator per shift.
Operating Costs
Edison operating costs for the first year of the test and demonstration
program at Mystic Station are now projected to be $525,000. This projection is
based on accrued costs for the first eight months and estimated costs for the
remainder of the year.
Of these costs, $235,000 are for Edison operating labor, $240,000 are
for Edison and invoice maintenance labor and materials and $50,000 are for metered
services to the scrubber system such as water, power and fuel.
These costs are heavily unbalanced because of the low operating hours
and the many problems encountered during this first year.
A substantial part of the operating labor and maintenance invoice costs
are related to cleanup work associated with the dryer scaling and material spillage
610
-------
problems. These costs would be substantially reduced if the scaling problem were
solved and the system was operable for reasonable periods of time.
Maintenance costs, although only 27, of plant investment, must be considered
excessive in view of the few operating hours. At this time, we do not have sufficient
operating experience to accurately project annual maintenance costs.
The cost for services would increase four fold for the fully operable plant
but must be considered high based on actual operating time. This is due to the use
of power and fuel through much of the plant standby period this past year.
Magnesium Oxide Requirements
MgO makeup requirements can not yet be ascertained. A sustained run of
both plants for at least a few months will be required to get meaningful data in
this area.
Operating Constraints On Boiler
There have been no constraints placed on the operation of this unit by
the addition of the scrubber system. Whether the scrubber is in service or not,
the unit can continue to operate. Through use of • bypass damper system, the
scrubber can be put into and taken out of service without influencing boiler
operation. When the scrubber is operating, high sulfur residual oil is burned
in the boiler. By simple valving, the boiler is changed over to a low sulfur oil
supply before the scrubber is taken off the line.
Modifications Dictated By Operating Experience
The following features in hindsight would be of value if available on
this prototype:
1. Tank for draining scrubber system, tank to be equipped to assure
slurry will stay in suspension.
2. Clean outs at pipe bends and dead ends so as to facilitate redding lines.
3. Provisions for continuous flow in slurry lines which are valved closed.
4. Careful design of piping to eliminate high velocity impingements of
slurry at tees, ells and reducers. Careful selection of piping mater-
ials to provide greater erosion resistance when required.
5. Strategic location of crushing systems to handle magnesium sulfite
slags.
611
-------
6. Strategic location of screening and/or crushing equipment to
handle contamination of .feeds in transport, etc. that have con-
tributed to plugging problems.
LARGE SCALE SCRUBBING STUDY
A recent estimate of the cost to install and operate scrubbers at
Mystic Station has been made (Figs. 1 & 2). Under this scheme, four units
totaling 1050 MW would be equipped vith magnesium oxide scrubbing systems and
would burn 2.5% S residual fuel oil. A separate estimate for calciner operating
costs was made (Fig. 3). The overall economics of scrubbing and calcining as
compared to the present requirement of burning 0.571 maximum sulfur content fuel
oil (Figs. 4 & 5) clearly shows a large economic advantage to scrubbing.
Scrubbing costs equate to $1.00 per barrel of fuel oil burned ($0.16 per
million Btu). Present cost differential between low and high sulfur oil is
twice this amount and savings in excess of $7,000,000 per year are potentially
available to our customers who now pay these costs through a fuel adjustment
clause.
OVERVIEW
At this writing, we are still attempting to get this process on-stream
for a continuous operating period. A continuous run is required to successfully
level out the calciner operation. A continuous run is required to develop cost
data for the process. A continuous run is required to demonstrate that the process
has the reliability necessary for adaptation to power plant equipment.
We are confident that the dryer scaling problem will eventually be
overcome. It should be susceptible to solution by equipment modification or by
equipment replacement. We are hopeful that a solution to this scaling problem
will then result in continuous operation of both plants so that refinement of
operation, in particular operation of the calciner, can then be accomplished.
Many of the problems now encountered would be eliminated as they relate directly
612
-------
to our inability to keep the equipment running.
Further, we are confident that this process represents some of the
most promising technology today and if successfully developed can assist many
utilities who are in a position to combine their operations with a chemical
operation for disposal of the S02 by-product.
613
-------
FIGURE 1
Scrubbing Systems
Unit
MW
4 150
5 150
6 150
7 600
Preliminary Estimate
Capital Investment Requirements for
MgO Scrubbing of Stack Gas SC>2 at
Mystic Station
Capital Investment
$ 2,000,000
2,000,000
2,000,000
5.000,000
MgSOj Ccntrifuging, Drying
MgO Storage
MgS03 Storage
Media Preparation
Initial M&O Charge
$11,000,000
3,300,000
400,000
700,000
100,000
300,000
Other Captlal Costs
Services to Battery Limits
Foundations
Site Work
Edison Direct Engineering
Total Capital (1972 $)
Less Existing Unit #6 Base Unit
Total Additional Captial
Escalation to 1976
Contingency
Total Additional Investment
$ 500,000
1,000,000
250,000
250,000
2,000,000
$17,800,000
(1,000,000)
15,800,000
4,200,000
1.000,000
$21,000,000
614
-------
FIGURE 2
Annual Owning and Operating Costs for
MgO Scrubbing Systems at
Mystic Station
Basts: Mystic Units M, f>5, #6, #7 = 1,050 MW
Fuel Oil Usage o 7,300,000 bbls/year @ 2.57. sulfur
S02 removal equivalent to 0.3"{, sulfur fuel oil
Cciptlal Investment « $21,000,000
Fixed Charges $3,100,000
Maintenance 800,000
Labor 200,000
Supervision 100,000
Utilities
Electric Power 800,000
Fuel Oil 400,000
Water 100,000
Mg,0 Make-Up 250,000
Total Annual Owning and Operating Cost $5,750,000
615
-------
FIGURE 3
Annual Operating Costs
Calcining Plant
Capital Investment: $4,500,000
Amortization
Interest
Maintenance
Insurance and Taxes
Labor + Supervision
+ Overhead
General Administration
Utilities
Electric Power
Voter (Process, Cooling)
Fuel Oil
Chemicals
Subtotal
Management Fee
Total Annual Operating Cost
$ 450,000
200,000
200,000
150,000
200,000
70,000
50,000
5,000
300,000
25.000
$1,650,000
250.000
$1,900,000
S16
-------
FIGURE 4
Overall Economics
MgO Processing of Flue Gas SO2
Annual Operating Costs
MgO Scrubbing $5,750,000
Calcining 1,900,000
Transportation 80,000
Gross Scrubbing Costs 7,730,000
Sulfur Credits 400.000
Net Scrubbing Costs $7,330,000
Cost/bbl fuel oil $1.00
FIGURE 5
Estimate of Annual Savings
Scrubbers at Mystic Station
Annual Fuel Useage (bbls.) 7.300.000
Annual Cost - LS oil ($4.70/bbl.) $ 34,300,000
Annual Cost - HS oil ($2.70/bbi.) -19.700.000
Gross Fuel Savings/yr. 14,600,000
Annual Scrubber Costs __- 7,300.000
Net Annual Scrubber Savings 7,300,000
617
-------
DESIGN AND INSTALLATION OF A PROTOTYPE
MAGNESIA SCRUBBING INSTALLATION
by
B . M. Anz, C . C. Thompson, Jr.,
and J. T. Pinkston
United Engineers & Constructors Inc.
1401 Arch Street
Philadelphia, Pennsylvania
619
-------
INTRODUCTION
The design and installation of this prototype magnesia scrubbing
installation represent an action of the Philadelphia Electric Company to
cope with the problem of stack gas emissions at its Eddystone Generating
Station. Although the topic of major interest here today is the removal
of S02 from flue gas, at Eddystone there was a problem of removal of particu-
lates as well. Because of this and other complications, Philadelphia
Electric chose to follow the procedure described in this paper rather than
to purchase any package system being offered for S0~ removal.
The design which will be described here is the result of engineering
work by Philadelphia Electric and their Architect-Engineer, United Engineers
and Constructors Inc. It had the following three main goals:
1. Build a prototype S02 removal system with a capacity equivalent
to a 120 megawatt unit.
2. Add equipment which would achieve reduction in the emission of
particulates.
3. Minimize the possibility that the operation of this prototype
S02 scrubber would affect the availability of Eddystone No. 1
to generate power.
This design was based on solid engineering principles which were applied to
known chemical process work involving magnesium oxide.
All of this makes it necessary to describe some of the character-
istics of Eddystone No. 1, the station to which this magnesia scrubbing
620
-------
system is being retrofitted. Eddystone No. 2, which is similar to Eddystone
No. 1 in general size and some other characteristics, will also have scrubbing
units retrofitted to it after the successful operation of this prototype.
Both stations burn coal averaging about 2.5% sulfur, 8 to 12% ash and 13,500
Btu/lb. heating value.
Eddystone No. 1, which went into service in November 1959, has a
nameplate rating of 325 Mw, but a generating capacity of 360 Mw. Steam con-
ditions are 5000 psi and 1150°F, with two reheats at 1050°F. This pressure
and temperature are the highest used in any power station in the U.S., and,
as far as we know, in the world. It is a matter of considerable difficulty
to start up or to shut down this station. Accordingly, it was essential that
special care be taken to prevent any upset condition in the scrubber from
reflecting on the generating unit itself. A lot of flexibility was engineered
into this scrubbing unit.
The space available for the scrubbing system was severely limited.
It was a factor the engineering team always had to keep in mind, and had an
effect on the number and sizes of the parallel scrubbing trains. When
complete, this overall installation will have essentially three scrubbing
trains for each 360 MW boiler unit through which the stack gases' from
Eddystone No. 1 and Eddystone No. 2 can be passed. Elaborate ducting will
permit various combinations of these scrubbing lines to be used. Although
particulate scrubbers, ducts and valving will be installed for all three
trains for Unit No. 1, only one complete scrubbing line, i.e., both particulate
and S02 scrubbing, equivalent to 120 MW will be installed in this,first phase.
621
-------
The Eddystone station is exactly in line with the main runways used
by jet aircraft at the Philadelphia Airport, and only a few miles away.
Obviously no unusual plume could be tolerated. Even if we could have convinced
the environmentalists that this steam is harmless, the safety consideration
would require that a plume be avoided.
The best approaches to papers for oral presentation with slides,
and papers in written form, are not entirely consistent. In order for slides
to be read at all by people in the back of the room and to be understood in
the short length of time they can be shown, they must be kept very simple.
As an aid to describing this process to you, several ultra-simple slides have
been prepared. They are included in the written text, because they may help
clarify some points, but mostly because the figures had to be prepared anyway.
Also, in the written there is provided a process flowsheet with a material
balance. It was not originally prepared for this type of use, and there really
is given much too much information for the space available. Still, with a low
power magnifying glass, you can read it.
Finally, let me repeat that this is not a stripped down plant.
In all aspects, the engineering reliability was placed ahead of cost. There
are several items of equipment which we might not put in a second generation
plant. There is more than liberable interstage storage.
We hope to put th±s prototype scrubber into operation during the
fall of 1973.
622
-------
PROCESS DESCRIPTION
Humidiflcation and Particulate Removal
Eddystone No. 1 is equipped with mechanical collectors and electro-
static precipitators upstream of the induced draft fans. The collection
efficiency is about 93%. This is no longer good enough and the plant design
was required to provide for particulate removal as well as SO- removal.
Obviously, one approach to additional particulate removal was the
installation of wet scrubbers. An early design question was whether to install
a separate wet scrubber from the S0« scrubbing system or to combine the func-
tions. Clearly the scrubbing device which we have in mind for S02 removal
would remove enough particulate matter to bring the station into compliance
with regulations.
The decision was reached to provide a separate venturi-type wet
scrubber using only water for the removal of particulates and for the humidi-
fication of the stack gas. There are several reasons for this decision. First,
it was expected that the humidification of the stack gas prior to bringing it
into contact with any chemical scrubbing medium might assist in avoiding
plugging problems. Second, it has been established that some components of
fly ash such as vanadium and iron compounds can catalyze the oxidation of
magnesium sulfite (MgSO,) to magnesium sulfate (MgSO.). The presence of some
sulfate in the scrubbing system is, of course, unavoidable because some of the
623
-------
S00 is oxidized to SO- before it reaches the scrubber. Nevertheless, it
4- -}
appeared desirable to hold down the quantity of this complicating sulfate as
much as possible. Third, there is no easy way to eliminate fly ash from the
circulating scrubbing slurry. A blowdown stream might be taken, but this
would cause a large loss of magnesium and would add to the solids disposal
problem. Finally, there is the matter of general system reliability. In the
present design, if difficulties in the operation of the S0_ scrubber are en-
countered, particulate removal can be continued by simply by-passing the SCL
scrubber.
A price was paid for this decision. Not only is there the capital
cost of the first scrubber in which only water is used as a scrubbing medium,
but there is a serious operating cost. The consumption of electrical power
for the entire scrubbing system was almost doubled by the use of this first
stage wet scrubber.
S02 Scrubbing Step
A fundamental feature of the process being installed at Eddystone
is that the chemical compound which reacts in the scrubber to bind the SO. is
not MgO, or Mg(OH)2, but MgSC>3. MgStX reacts with H.O and SO. to produce
Mg(HS03)2. As you know, MgSC>3 (more properly MgSO^.GHJ) in this system) is
not very soluble in H20. Accordingly, it is actually used as an approximately
10% slurry of crystals in a dilute solution of Mg(HSO,)_, with a pH of about 6,
624
-------
On the other hand, Mg(HSO,)_ is very soluble. Thus, the fundamental
chemical reaction which occurs in the scrubber tower results Ln the conversion
of a slightly soluble material to a very soluble material. This approach offers
a good possibility of avoiding the problem of plugging in the scrubber tower.
A price may be paid for this feature, because the equilibrium vapor
pressure of S02 over a slurry of MgSO« in a slightly acid dilute solution of
MgCHSO ) is higher than that for a slurry of MgO or Mg(OH>2 in H-O. Accord-
ingly, it is possible that removal of S02 from stack gases will not be quite
as effective in the Eddystone plant as if an alkaline slurry were used.
Precise equilibrium values in the concentration range of interest are not
available to us. From the curve relating scrubbing efficiency to the
acidity of the scrubbing slurry in a particular experimental unit under
consistent operating conditions, it can be seen that the difference between
S02 removal at a pH of 6 and at pH 8 is not very great.
The equipment in which S02 is removed from the stack gases is a
commercial device manufactured by Environeering, Inc. known as a Ventrl-Rod
unit. A simplified sketch of this contactor is shown. Essentially there are
two scrubbing stages. In each a slurry of magnesium sulfite is sprayed upward,
co-current with the stack gas, through openings between cylindrical rods. A
venturi-like effect similar to a fountain is obtained as the streams flow
between these rods, assisting in the contacting between the stack gas and the
slurry of magnesium sulfite.
625
-------
Mist is eliminated by the use of backwash sprays of water onto
louvers in the top section of the scrubber.
This scrubber has an overall height of fifty-three feet. It is
rectangular in shape, fourteen feet six inches by twenty-five feet.
Magnesium sulfite slurry is circulated to it at a rate of 13,400
gpm, or 133,800 Ib/min. The humidified gas rate is 268,000 actual cubic feet
per minute or 17,300 Ib/min. giving an L/G of 50. This high slurry-to-gas
ratio is designed to minimize concentration changes across the SO- scrubbing
system proper, and to insure a rather constant acidity of the scrubbing
medium.
Neutralization
In the scrubber surge tank the following reaction occurs:
Mg(HS03)2 + Mg(OH)2 + 10H20 - >2MgS03.6H20
Here we have the first formation of a slightly soluble material. By carrying
out this reaction in a large agitated tank - its capacity is 60,000 gallons -
having a hold up time of four minutes, any plugging problems should be held
to a minimum. The temperature in this tank where the neutralization occurs
is close to 129 F, a condition under which the hexahydrate is formed. At
higher temperatures the trihydrate can form. Its crystals tend to be smaller
than those of the hexahydrate, and might cause more handling problems.
626
-------
The pH of the slurry in this surge tank ts controlled at 6 by
regulating the rate of addition of Mg(OH) slurry.
MgSO- Recovery
The recovery of the MgSO,,.6HJ) formed in scrubber surge tank is
begun by diverting a drag stream from the main scrubber circulation loop.
It flows first to a thickener. Here, again, is a piece of equipment which
may not be absolutely necessary. The conclusion to include it was based on
the desirability for more surge capacity in the recovery portion of the plant
and the prospect for improved centrifuge operation.
The thickener has a diameter of forty feet, and a straight side
height of twelve feet. It receives 1975 Ib./min of slurry containing 165
Ib./min MgS03,6H20. The underflow from the thickener going to the centrifuge
contains about 25% MgS03.6H20.
In the stainless steel solid bowl centrifuge, a cake of
MgS03.6H20 crystals wet with a solution of MgSO, is recovered. These wet
crystals of MgS03.6H20 are discharged from the centrifuge, through a vertical
chute into a screw feeder which provides a seal and a continous flow of wet
solids, into a rotary kiln type dryer. Combustion gases from an oil burner,
tempered by a side stream of stack gas from the first induced draft fans,
flows co-currently with the crystals. Anhydrous MgSO» is discharged from
the dryer and conveyed to the storage silos.
Some MgSO^ will also, of course, be present.
627
-------
Regeneration of MgO
For this prototype facility, the MgSO_ will be trucked to a
sulfuric acid plant about twenty miles from Eddystone. There it will be
heated in an oil-fired fluidized bed reactor to decomposition. The MgO
formed goes overhead with the SO^ and the combustion gases, from which it
will be separated by cyclones. The stack gas is cleaned up by a combination
venturi and a packed scrubber tower before being admitted to the sulfuric
acid plant itself.
The selection of an optimum temperature in this fluidized bed
regenerator represents a compromise. On the one hand it is desired to reduce
the MgSO, to MgSO and to insure the substantially complete decomposition of
MgSO- to MgO and SO , both of which are favored by higher temperatures. On
the other hand, excessive temperatures will produce "dead burned" MgO which
is not very reactive chemically and is thus not really effective for further
S0? removal. The final operating temperature for this stage has not been
selected, but will probably lie in the range of 1650° - 1750°F. Fluidized
bed reactors are particularly well suited to precise temperature control. They
also allow for precise control of oxygen in the reactor which eliminates the
necessity of adding oxygen scavengers such as carbon to the bed so the
MgSO, can be decomposed.
The regenerated MgO will be trucked back to Eddystone for re-use.
For a prototype operation, this sort of thing can be tolerated. For a full
628
-------
scale operation at a major generating station, however, this trucking will
not be practical. The sulfuric acid plant should be located at the station
site.
Stack Gas Reheat
When the stack gas emerges from the S02 scrubber, its temperature
is about 130 F and it is substantially saturated with water vapor. To avoid
a billowy white plume from the stack as this mixture emerges and water con-
denses, some kind of reheat is necessary. In this retrofit unit at Eddystone,
the practical way to accomplish this was to install an oil burner mounted
directly in the flue gas duct leaving the S02 scrubber.
Close to one million Btu/hr. will be required for every one F of
reheat for the total output of Eddystone No. 1 (all three final scrubbing
lines). The fuel cost alone will be close to $5000 per year per °F of
reheat. Obviously, the amount of reheat used will be held to a minimum,
depending on local conditions.
629
-------
ECONOMICS
Capital Cost
The best estimate we can make now is that the total capital cost
for this version of a magnesium oxide scrubber is in the range of $55 - $65
per kilowatt. This is for a new unit. For a retrofit, you should add about
$15 per kilowatt.
Of this basic cost, $45 - $50/KW is for the scrubbing unit itself.
About $10 - $15 is for the facility to convert MgSO« to MgO for recycle and
to carry the liberated S0« on to concentrated sulfuric acid. Our thought is
that the sulfuric acid plant should be located adjacent to the power station.
The hauling of MgSO- and MgO to and from a distant regeneration facility and
sulfuric acid plant can become very expensive.
This operation of a sulfuric acid plant in conjunction with a power
station is not something that will bring universal happiness. To the power
station staff it will be an utter nuisance. Its presence may become an up-
setting factor in the local sulfuric acid market. This acid which is produced
has got to go somewhere, at whatever price can be obtained for .it. Moreover,
the acid plant which will ultimately be built at Eddystone will be distinctly
below the best economic size, since it will have a daily capacity of only
about 350 tons. At present, in the Delaware Valley, we do not see any over-
riding problem in this regard. This brings us on to the matter of operating
costs.
630
-------
Operating Cost
We estimate that the overall operating cost of the final facility
at Eddystone, including the scrubbing unit itself, the MgO regenerator and'-
the sulfuric acid plant, will lie between 1.3 and 1.4 rails per kilowatt hour.
The factors going into this estimate are as follows:
Fuel cost $0.70/MM Btu
Power cost 7 mils/KWH
Fixed charges 15% of capital/yr.
Maintenance 3% of capital/yr.
It was further assumed that there will be a loss of MgO of 5% per
cycle. This is probably a very conservative estimate, but in the absence of
operating experience, it appeared best to include it.
The operation of this scrubbing facility as described earlier, with
a wet scrubber ahead of the S02 scrubbing unit, will consume approximately 3%
of the power generated by the station. If the wet scrubber were eliminated,
power consumption would drop to about 1.75%.
631
-------
C.EACTIOAJ 1AJ 5£!2U&BEIZ
O-i
OJ
NJ
-------
PH
MA60E61UM SYSTEM
iwu
99
98
%
V
V
3
vi
5
kJ
V3
2
t£l
VJ
VP
95
^
^
92.
91
90
99
;:~
gk
66
64
^^-»--^-"^*""-t
X
/
if
!
i
1
i
l
1
i
i
iii
6 6 7 6
OF sazuaaiiaQ 6i.ora.izy
633
-------
•6LUQJ2V IAJI_feT
-------
I2EACTIOAJ 1AJ 5CG.UBBEIZ 6UIZG.E TAAJtf.
U>
Ul
-------
6-TAC.k. 6.A/S
FfZOM I.Q.
PAr2.TlCUL.ATE
U)
HUMIDIFIED
PATZ-flCDLATE.
To
6JATET2
CLE-AAJEC? GA-6 To
I2.EMEAT AUD 5TAO!
I
11
To
-------
-TO Mq CD
rze&Eju
PLAAJ-T
-------
1-760°
FEED
V
CPl
w
00
l\\\\\\\\\\\
IZEACTOIZ
LtfcJE.0)
FDEU
Atrz
PrZEU&ATErz.
1OQO
AlC
1
•ro
rZEC.Ol/ETlE.0
&LOUJ
j,
6J£T
fZEGEUEfZATlQJSJ PLAMT
-------
r—
— i 4
ha?-:
.£
©
i '
I i
©
JL.
fflol
s
/
1
V
r
i
i
i
*> V
"Jgd XCJ 1"»"l ^M^ XL
' I
..I/;© os . U
' N
i
i
^ri
f~
^ t^S '"K1^1 H-"^
— u
rw CUM rL
>
i
/
/
— =%*;
-"•""
^ r'1
t~-1
^p
Efe
34 11 ?* « Jo ii i? M !4
-------
CAPITAL COST
AJELJ 3OiUfcBlAJ£» UAJIT
10 - $ 1 5/H.LJ f2E6EAJEfZATlO/J U/UIT AlUP W5O PLA/oT
FOfZ JZETfZOFIT - ADD A&OUT
OPEflATlAJd <105T
Ol
Jt
o
01/EI2ALL
FACTO125 COAJ5lDEfZBD
FUEL. ^O."70/MM
POMEfZ -7 mil6/l^liJU
FI^ED ^UAfZ.GE.5 15% af £apif
-------
APPLICATION OF THE WELLMAN-LORD SO2
RECOVERY PROCESS TO STACK GAS DESULFURIZATION
by
Raymond T. Scheider
Christopher B. Earl
Davy Power gas, Inc.
Lakeland, Florida
641
-------
ABSTRACT
The Wellman-Lord S02 Recovery Process has been successfully
commercialized in the U.S.A. and Japan. Six full-scale installa-
tions have now built up an aggregate operating experience of
eight years. The fourteen emission sources controlled by the
process include sulfuric acid plants, claus plants and oil-fired
boilers.
Since the start-up of the original plant at Paulsboro, New
Jersey for Olin, the process has demonstrated long-term relia-
bility of operation and its ability to attain SO9 emission
levels as low or lower than required by the regulatory agencies.
This paper describes the basic process and reviews the cur-
rent status of this technology. The application to large fossil
fuel-fired steam generating plants is reviewed,and capital and
operating costs for a specific application are included.
642
-------
THE WELLMAN-LORD SCU RECOVERY PROCESS
INTRODUCTION
Six commercial installations of the Wellman-Lord SO2 Recovery Process
are operating successfully in the U.S.A. and Japan. These applications
include seven sulfuric acid plants, two oil-fired boilers and five claus
units (see Table 1} .
The SO2 emission level first envisioned with the use of the process was
500 ppmv since this was a statutory requirement in 1969. The operations
of the five plants have demonstrated the process's ability to maintain
emission levels at consistently less than 200 ppmv.
CHEMISTRY OF THE PROCESS
The process is based on a sodium sulfite/bisulf ite cycle. The reactions
that take place in the process can be abbreviated for simplicity as
follows:
ABSORPTION
S02 + Na2S03 + H2O - »- 2NaHS03
REGENERATION
2N3HSO - »- Na2SO | + S02 | + H O 4
Apart from the two major reactions above, sodium sulfate (Na^SO,^) ,
which is nonregenerable in the normal process, is formed in the
absorber as a result of solution contact with oxygen and sulfur trioxide
as follows:
Na2S04
and
The sodium sulfate so formed is controlled at a level of approximately
five percent by weight in the absorber feed stream by maintaining a
continuous purge from the system. This purge stream is discussed later.
additional source of sodium sulfate and thiosulfate is the autooxidation
or so-called disproportionation reaction which takes place in the regen-
eration section:
643
-------
TABLE I
DAVY POWERGAS INC.
Type Contract
E - Engineer ing
P - Procurement
C - Contraction
T - Turnkey
WELLMAN-LORD SOj RECOVERY PROCESS
Project
Number
3659
l<120
4126
412?
M37
4H
-------
6NaHSO >- 2Na2S04 + Na S~0o + 2SO + 3H-O
«3 £ *• -^ ^ ^
Laboratory research and commercial experience have guided the selection
of operating conditions which minimize all these sources of sodium
sulfate formation.
A make-up of caustic is required to replace that lost in the purge
stream. The caustic make-up solution reacts with the sodium bisulfite
in the absorber solution to form additional sodium sulfite:
NaOH + NaHS03 »- Na2S03 + H2O
Soda Ash (Na2CO3) can also be used as the makeup source of sodium.
The simple regeneration scheme of the patented Wellman-Lord process
relies on the favorable solubilities of the sodium system. The bisulfite
form has almost twice the solubility of sulfite at the temperatures con-
sidered for the process. Because of this it is possible to feed the
absorber with a saturated sulfite solution, or even a slurry, without
any fear of additional crystallization or scale forming despite consid-
erable evaporation of water. This is due to the fact that as S02
absorption proceeds, the composition of the solution is shifted in the
direction of increasing solubility.
There are several advantages to operating the absorber with highly con-
centrated solutions. The solubility of oxygen decreases rapidly as salt
concentration increases and so the mass transfer of oxygen into the
solution is slowed appreciably. This reduces oxidation of the sodium
salts to sodium sulfate. The steam requirements of the process are
directly related to the quantity of water pumped around the system.
Operating at or near saturation with respect to sulfite thus reduces
the overall steam consumption.
The same solubility effect is taken advantage of in a revers fashion
in the regeneration section. AS SO2 is stripped from the concentrated
solution,, the sulfite salt is formed and rapidly reaches its solubility
limit thus precipitating as crystals. in effect the low vapor
pressure component (sodium sulfite) in the solution being regenerated,
is removed from the equilibria balance. Thus the regeneration
operation can proceed with a constant high percentage of bisulfite
in solution permitting a considerable reduction in the stripping
steam requirements.
DESCRIPTION OF FLOW
The flow scheme for the basic process is shown in figure 1. The SC>2
rich gas is contacted counter-currently in the absorber by the
sodium sulfite solution and passes out the top stripped of SO2-
The solution leaving the bottom of the tower, now rich in bisulfite,
645
-------
FIG, I BASIC FLOW DIAGRAM
«<£•«
'%
ICONULNSER
77?k
PRODUCT
G
ABSORBER
SOLUTION
STORAGE
VAPORATOR
CRYSTALLIZE.R
DISSOLVING
TANK
WELLMAN-LORD SO., RECOVERY PROCESS
-------
is discharged to a surge tank and then pumped to the regeneration
section.
Low pressure steam is used to heat the evaporator and drive off tJ02
and water vapor. The sodium sulfite precipitates as it forms and
builds a dense slurry of crystals.
The overhead stream is subjected to a partial condensation to remove
the majority of the water vapor. The condensate recycles to the
dissolving tank and the product SC>2 is discharged from the process.
A stream of slurry is withdrawn from the evaporator and is redissolved
by the overhead condensate. The lean solution is pumped to a surge
tank prior to being fed back to the absorber.
GAS PRETREATMENT
Normally the flue-gas from a fossil fuel-fired boiler or a smelter
will require cooling and particulate removal prior to absorption
of the SC>2»
In the case of a claus plant, the tail gas must first be
incinerated under conditions which destroy the H2S, COS and
CS2»
This pretreatment step must be studied on a case-by-case basis
to ensure the selection of the optimum design in relation to the
overall facility.
The type of pretreatment ultimately chosen will depend on the
following parameters:
Gas Temperature
Particulate Content
Organic Sulfur Content
Sulfur Triodide Content
Acid Mist or Vapor Content
Humidity.
SO? ABSORBER
The absorber is a simple two or three stage contacting device.
DPG commercial experience with both tray and packed towers
permits selection of the design most suited to each specific
application.
647
-------
For most gases treated the absorber will require recirculation
around each individual stage, since the quantity of feed solution
is insufficient to adequately wet the tray or packing.
During operations the recirculation rate can be throttled until
the SC>2 emission increases up to the design point. This will
minimize oxidation since the absorption of oxygen is liquid-film
controlling and therefore proportional to liquid rate. The
absorption of S02 is gas film controlled and therefore relatively
insensitive to liquid rate.
An inherent advantage of the Wellman-Lord Process is that the
absorption system can be physically separated by a large distance
from the regeneration system. This also means that it is practical
to treat gases from more than one plant by installing separate
absorbers for each source of SO2» with all the absorbers being
supplied by a common regeneration system.
SOLUTION SURGE TANKS
The use of solution storage not only enables the process to
operate smoothly during frequent changes of gas flow or S02
concentration, but it also permits a complete shutdown of the
regeneration section to perform scheduled maintenance acitvities,
without any pause in S02 absorption.
This feature minimizes the amount of expensive spare equipment
required with no sacrifice in basic pollution control reliability.
CHEMICAL REGENERATION SYSTEM
The heart of the regeneration system is a conventional forced-
circulation evaporator/crystallizer. The design parameters of
this unit have been developed such that long term continuous
operation is assured. The evaporator can be designed to use very
low pressure exhaust steam which might otherwise be discharged
to the atmosphere.
In very large plants or in case of high cost steam it is
practical and economical to operate the regeneration system as
a double effect evaporator which will reduce steam consumption
by about 40-45%.
PRODUCT SULFUR DIOXIDE
The vapor leaving the evaporator is subjected to one or more
stages of partial condensation to remove water. Existing
plants are operating with both air and water-cooled condensers
648
-------
in this service. The final product SO2 can be delivered at
whatever quality is required for further processing. It is
suitable for conversion to high grade sulfuric acid, elemental
sulfur or liquid SO2•
Existing Wellman-Lord plant are recycling the gas back to
sulfuric acid and claus plants or sending it to be processed in
a satellite sulfuric acid plant.
PURGE TREATMENT
A small amount of the circulating solution is oxidized to a
non-regenerable form and a purge stream is removed to control
the build-up. This stream can be dried for sale or disposal, or
it can be treated to permit its discharge as an innucuous effluent,
Other process steps are available which recover the sodium values,
thus allowing the system to operate as a closed loop.
OPERATIONS
The first Wellman-Lord plants quickly proved the viability of the
process but were not without some start-up problems. Needless to
say, the lessons learned in these plants were quickly incorporated in
subsequent designs. An indication of this is that three of the first
five users already have second plants in operation or in the process of
being engineered.
STANDARD OIL OF CALIFORNIA
One of our recent projects for Standard oil of California started
up in September, 1972 and from the first day has operated continu-
ously in a most satisfying manner.
This plant included large surge tanks which if necessary would
permit a three day shut-down of the regeneration section, while
the absorption section continues to run. The reliability pro-
vided by this feature was vital to protect the integrity of this
large refinery located close to Los Angeles International Airport
and surrounded by a residential area. The regulations governing
this situation require full air pollution control 24 hours a day,
365 days per year.
A side benefit of this relatively cheap means of providing
reliability was that the plant could handle excursions of SO2
concentration up to three times the design level while still
attaining an SO2 emission less than design. The unit mentioned
aboye^has surpassed the performance guarantees, and has been
officially accepted by standard oil of California.
649
-------
JAPAN SYNTHETIC RUBBER COMPANY AT CHIBA, JAPAN
This installation is operating on the flue gas from two oil
fired boilers each with a capacity of 286,000 Ibs/steam per hour.
The fuel oil contains 4.0 - 4.2% sulfur and produces a flue gas
varying from 1500 to over 2000 ppm SO2. The plant has consis-
tently reduced the SO2 emission to below 200 ppm since its start
up in June of 1971. The plant currently averages an inlet SO2
concentration of 1500 ppm with an outlet emission 150 ppm.
During the period June 1971 - March 1973, the boilers operated a
total of 12,972 hours. During this same period of time the
Wellman-Lord process unit operated 12,533 hours for a 97% on-
stream factor. During the period May 9, 1972 - March 1, 1973,
the boilers operated a total of 6,978 hours, and the Wellman-
Lord process unit was in operation during all of these for a
100% on-stream factor.
In this plant each boiler is connected to a five meter square
by 21 meter high sieve tray absorption tower, and the recovery
section utilizes a double effect evaporator for steam economy.
The tail gas after the absorber is reheated to 130°C for plume
dispersal.
The product SO2 gas is processed in a slightly modified,
conventional contact sulfuric acid plant, and produces high
grade, merchant quality, 98% sulfuric acid which is sold in
normal markets.
APPLICATION TO POWER PLANTS
THE NIPSCO DEMONSTRATION PLANT
As a consequence of the successful operating units mentioned
above, Davy Powergas has been awarded a full scale demonstration1
plant project for Northern Indiana Public Service Company,
funded jointly by the U.S. Environmental Protection Agency and
NIPSCO. This facility will treat the total stack gas from a
115 mw coal fire boiler at the Dean H. Mitchell station in Gary,
Indiana. The end product of this plant will be elemental
sulfur. SO2 Removal from the flue gas will be performed by the
Wellman-Lord SO2 Recovery Process, which will deliver the
concentrated SO2 to a reduction facility that uses Allied chemical
Corporation's technology to produce high quality elemental sulfur
as the final product.
650
-------
DEMONSTRATION OBJECTIVES
The broad overall objective of this project is to demonstrate
the applicability of the above two commercially proven processes
to large scale coal fired boiler plants. In order to meet the
overall objectives of the project, specific contractual guarantees
are being developed in accordance with the following guidelines:
1. The process will accomplish a minimum of 90% removal sulfur
dioxide from the stack gas when firing coal up to 3.5% sulfur
content producing an emission of 200 ppm SC>2 or less. 3.5%
sulfur content of the coal is equivalent to approximately
2300 ppm by volume sulfur dioxide in the stack gas.
2. The aggregate cost of steam, electric power, water, and
natural gas will not exceed a specified amount.
3. The required quantity of make-up chemicals will not exceed
a specific amount.
4. No other water or air pollution sources will be created.
5. The sulfur produced will be of a quality suitable for the
manufacture of sulfuric acid by the contact process.
6. The plant will comply with all applicable federal, state and
local air and water quality regulations in effect at the time
detailed engineering is begun.
PHASING AND SCHEDULING OF THE PROJECT
Phase I of the project consists of preliminary engineering and
definitive cost estimates. This phase of the project is
essentially complete at this writing.
Phase II includes the completion of detailed engineering, pro-
curement, construction and start-up. This phase of the work is
scheduled to be completed during the latter half of 1974.
Phase ill, following acceptance tests at the conclusion of Phase
II, covers a one year period of operation by Allied Chemical and
will be funded entirely by NIPSCO.
TRW/EPA TEST PROGRAM
Concurrent with Phase ill of the project, a comprehensive emission
testing program will be conducted by EPA. To implement this test
651
-------
and evaluation program the EPA has retained the services of TRW
Inc. The primary objectives of the test and evaluation program
are:
1. A validation of the emission reduction achieved by the
WeiIman/Allied process.
2. A validation of the estimated installation and operating
costs of the demonstration unit.
3. A determination of the applicability of the Wellman/Allied
process to the general population of utility boilers. The
test program will include extensive measurements of those
parameters that characterize the sulfur removal and economic
performance of the Wellman and Allied process units.
In addition to specific test work, TRW will conduct a survey
for the purpose of identifying the potential users of the
Wellman/ Allied process. User boiler units having
applicability potential are selectable based upon considerations
such as expected useful life, capacity, average load factor,
low sulfur fuel availability and the marketability of sulfur
and sodium sulfate in the region where the boiler is located.
PURGE CRYSTALLIZATION SYSTEM
In order to meet the objectives stated above that no additional
water or air pollution sources would be created, a purge cry-
stallization section has been included in this project, in this
section of the plant the purge stream will be subjected to
fractional crystallization in order to concentrate the sodium
sulfate and sodium thiosulfate components of the purge. This
portion of the purge stream will be dried to produce a solid
crystalline byproduct of high sodium sulfate and sodium thiosulfate
content. The resulting sulfite/bisulite stream from which most
of the sulfate and thiosulfate has been removed will be recycled
to the S02 Recover Process.
ALLIED CHEMICAL CORPORATION SO? REDUCTION SYSTEM
This project is unique in that it will produce elemental sulfur
as a by-product. The Allied chemical Corp. Process is the
subject of the next paper and will not be covered here.
SITE CONSIDERATIONS AT THE D. H. MITCHELL STATION
The demonstration plant will be installed on unit 11 of NIPSCO's
652
-------
Dean H. Mitchell Station. This slide is a view of this station
with unit 11 at the right. Unit 11 is a Babcock and Wilcox
pulverized coal feed boiler capable of producing 821,000 Ibs/hr
of steam. The superheater outlet pressure is 1870 Ibs/sq in.
gauge at 1005°F. It is equipped with an electrostatic precipitator
installation rated at 98% collection efficiency. The coal
presently being fired in this unit is from Southern Illinois
with an as received average analysis of 3.16% sulfur, 10.5% ash,
and 11.8% moisture.
NIPSCO clearly has reason to be interested in a method for removal
of sulfur dioxide from the stack gas. Such a method, if effective,
reliable, and economically viable, permits NIPSCO to comply with
emission limits while continuing to fire coal from assured
relatively near by sources.
Stack gas treatment processes can be diveded into two classes,
those that produce a so-called throw-away product and those
that yield a product that can in theory at least be marketed.
The throw-away process was rejected by NIPSCO for ecological
reasons.
The Dean H. Mitchell Station is located on Lake Michigan in
the Northwest corner of Gary, Indiana. As this slide shows,
its immediate environs are highly industrial, but farther out
the plant community is residential and densely populated.
Considerations of land and water use made a throw away process
environmentally unacceptable at such a site.
This slide shows an artists rendering of the S02 recovery and
reduction installation that will be made at the Dean H. Mitchell
Station superimposed on a photograph of that station. The red
structures are those of the Wellman-Lord process, and the
Allied process is housed within the orange building.
CAPITAL AND OPERATING COSTS
Actual costs for power plant flue gas scrubbing and S02 recovery
will vary considerably depending on the specific application. Local
site conditions and the, difficulty of retrofit can cause costs to vary
over a wide range. It is very misleading to compare estimates that
are derived from different bases.
Capital and operating costs for applying the W-L process to two large
plants (500 and 1,000 MW) have been projected from the detailed
studies made of the 115 MW plant at the Dean H. Mitchell Station of
653
-------
Northern Indiana Public Service Company. The following table indicates
the basic parameters used in estimating these costs:
TABLE 2
500 MW 1000 MW
Absorber Design Flow
SCFM 3 @ 400,000 ea 6 d> 400>000 ea
Regeneration Plant
Lb/Hr SO2 Product 1 @ 18,000 2 @ 18,000 ea
Sulfur Plant
LTPD Sulfur Product 1 @ 92 1 @ 184
The studies of the 500 and 1,000 MW plants are based on 3.2% sulfur
coal and the flue gas flows are based on 20% excess air.
The regeneration plant and sulfur dioxide reduction plant sizing
correspond to an 80% load factor. Absorber operation at peak capacity
would be handled by solution inventory in large surge tanks.
The annual utility costs presented in Table 3 are based on a continuous
80% load factor for 330 days per year which is equivalent to an over-
all load factor of 72%.
654
-------
TABLE 3 - CAPITAL AND OPERATING COSTS
Capital Cost
Annual Costs
Chemicals
Natural Gas d>55C/MM BTU
Electricity @ 7 Mil/KWH
Steam © 50C/1000 Lb
Cooling Water @ 2C/1000 Gal
Labor © $5.80/hr
Supervision
Overhead
Maintenance © 4.5% of Capital
Laboratory
Miscellaneous Supplies
capital Charges © 14.5%
TOTAL
Sulfur Credit @ $18/LT
NET TOTAL
Clean-up Cost mils/KWH
C/MM BTU
500 MW
$18,700,000
$/Year
956,000
336,000
499,000
554,000
111,000
169,000
47,000
343,000
842,000
72,000
45,000
2,712,000
6,686,000
(546,000)
6,140,000
1.9
20
1000 MW
$33,200,000
$/Year
1,912,000
672,000
993,000
1,108,000
222,000
229,000
47,000
506,000
1,494,000
72,000
90,000
4,814,000
12,159,000
(1,092,000)
11,067,000
1.7
18
655
-------
APPLICATION OF 803 REDUCTION
IN STACK GAS DESULFURIZATION SYSTEMS
by
William D. Hunter, Jr.
Allied Chemical Corporation
Industrial Chemicals Division
Morristown, New Jersey
657
-------
Allied Chemical's experience in the design and operation of
a large, single train plant for reduction of SO to elemental
sulfur has made possible the application of new technology to
emission control in metallurgical operations and fossil fuel
combustion.
This Allied Chemical technology has been proven in two
years' successful commercial operation where it functioned as
the emission control system at a Palconbridge Nickel Mines, Ltd.
facility near Sudbury, Ontario, Canada. The quantity of sulfur
in the gas stream treated in the Allied unit at Falconbridge
(500 long tons per day) is equivalent to that contained in the
largest metallurgical sources, or in flue gas from 2000 megawatts
of electric generating capacity fired with 3% sulfur coal. As
a consequence, future applications of Allied1s technology will
involve scale down of equipment size with full assurance of
optimum plant performance.
Availability of a proven process capable of converting SO
to elemental sulfur offers freedom from the problems and cost 2
of disposal of the waste products of neutralization. Likewise,
costly storage and transport of by-product sulfuric acid can be
avoided at locations remote from acid markets.
The Allied Chemical S02 reduction process can be applied
directly to gas sources where the SO2 content is 4-5% or more
and where oxygen content is limited. Various metallurgical
emissions can, therefore, be processed directly in the Allied
system. Since oxygen reacts with the reducing agent in the
same manner as the SO2, it is desirable in optimizing the
system to achieve the lowest oxygen content possible in order
to minimize reducing agent consumption. The Allied process
has the versatility of being adaptable to feed gases containing
any SO2 concentration up to 100% (dry basis). Hence, in those
applications where the oxygen content is too high or SO2
concentration too low for direct application to the emission
source, the Allied SO- reduction process may be joined with one of
several regenerable flue gas desulfurization process which
recovers the SO2 from the flue gas in a concentrated, low
oxygen form.
658
-------
Adaptability of Allied's SC>2 reduction technology to a
feed gas containing 100% SO2 (dry basis) will be demonstrated
at the D. H. Mitchell station of Northern Indiana Public Service
Co. at Gary, Indiana. There, the process will be combined with
the Wellman-Lord SO recovery process to provide a flue gas
desulfurization system for a 115 megawatt, coal-fired boiler
in a project jointly funded by NIPSCO and the Environmental
protection Agency.
Engineering, procurement and construction of the entire
facility will be the responsibility af Davy Powergas Inc.
Allied is providing the S02 reduction process technology as
well as technical and start-up services under contract with
Davy Powergas. Then, under a separate agreement with NIPSCO,
Allied will operate the entire flue gas desulfurization system
and will market salable by-products.
The role that will be undertaken with NIPSCO is a familiar
one for the Allied organization which has for many years
provided services to oil refineries in the design, construction
and operation of satellite plants for the conversion of
hydrogen sulfide to elemental sulfur and the decomposition of
spent sulfuric acids, in addition Allied has performed services
to metallurgical industries in the building and operation of
facilities for the roasting of sulfide ores, and the production
and marketing of sulfuric acid. The development of SO, reduc-
tion technology enabled Allied to extend the range of its services
directed to environmental improvement.
The Falconbridge plant, which involved the fluidized bed
roasting of pyrrhotite ore, could only have been built if
provisions were made to comply with Canadian regulations
requiring the removal of at least 90% of the sulfur values
contained in the roaster gases before their release to the
atmosphere. Two alternatives were considered: the manufacture
of sulfuric acid or reduction of the SO2 to sulfur. There
is no substantial market for either in the immediate area, but
the recovery of elemental sulfur offered-the advantages of a
commercial product that is more easily stored than sulfuric
acid and has a much wider economic shipping range.
659
-------
The Falconbridge installation demonstrated the required
capability of removing at least 90% of the SO- from roaster
gases at a rated capacity of 500 long tons or sulfur per day
in the incoming gases. Reliability of performance was
established as well as flexibility in turndown to two-thirds
and one-third of design capacity with essentially constant
operating efficiency at all rates.
At Falconbridge, Allied engineered and constructed the S02
reduction plant. The facility was run by Allied supervisors
utilizing an operating and maintenance staff assigned Allied
by Falconbridge Nickel Mines.
Product sulfur was, for the most part, distributed to
Allied's sulfuric acid manufacturing locations serving the
U.S. merchant acid market. The Falconbridge product was used
interchangeably at these locations with sulfur from Frasch
producers and other by-product sulfur sources.
COMMERCIAL PLANT DESCRIPTION
A flow diagram of the Allied Chemical SO- reduction
process as it is applied to a sulfide ore roasting operation
like that at Falconbridge is shown in Figure I. The hot
SO2 gas from the roasters is passed through hot gas heat
exchangers (1) and (2) where part of the heat content of
the gases is utilized to reheat other process gas streams.
These will be described in more detail later. At this point
the roaster gas still contains fine dust particles as well as
gaseous contaminants which must be removed before the gas
reaches the reduction reactor. This gas purification is
accomplished in a two-stage aqueous scrubbing system consisting
of a two-leg gas cooling tower (3) and a packed condensing
tower (4). The bulk of the dust and other contaminants are
collected in the gas cooling tower while the gas is cooled and
saturated by contact with a recirculated weak sulfuric acid
solution. The demister pad at the tower outlet is continuously
sprayed with weak acid from the condensing tower. The under-
flow from the gas cooling tower is treated with lime to precipitate
dissolved metallic impurities removed from the gas, and
neutralize the acidity, before being delivered to a waste pond
where the solids are allowed to settle.
660
-------
The process gas is further cooled in the condensing tower
(4) by circulating weak acid which is cooled externally in
impervious graphite heat exchangers (5) . Entrained droplets
of acid mist are removed from the gas in electrostatic
precipitators (6) . Drips from the precipitators are returned
to the gas cooling tower.
The temperature of the clean gas is then raised above the
dew point of sulfuric acid by admixing with a reheated stream of
the same gas in mist tower (7) . This recycle gas stream is
heated by circulation through hot gas heat exchanger (2). The
process gas is drawn through the wet purification system, then
forced by a centrifugal blower (8) through the balance of the
plant. Natural gas, which serves as the reducing agent, is
introduced into the process gas stream at the blower discharge
and the mixture passed through hot gas heat exchanger (1) to
raise its temperature above the dew point of sulfur before
entering the reduction reactor system.
The principal function of the catalytic reduction system is
to achieve maximum utilization of the reductant while producing
both sulfur and H2S, so the H^S/SO ratio in the gas stream
leaving the system is essentially that required for the subsequent
Glaus reaction. Although the chemistry of the primary reaction
system is extremely complex and includes reactions involving
11 different elements and compounds, it may be summarized in
the following equations:
CH4 + 2 S02 - j. C02 + 2 H20 + S2
4 CH + 6 SO2 - ^ 4 CO + 4 HO + 4 H2S + S2
The preheated process and natural gas mixture enters the
catalytic reduction system through a four-way flow reversing
valve (9) and is further preheated as it flows upward through a
packed-bed heat regenerator (10) before entering the catalytic
reactor (11) . The temperature of the gases entering the reactor
is held constant by continuously by-passing a varying quantity
of cold process gas around the upflow heat regenerator. The
661
-------
heat that is generated in reactor (11) by the exothermic
reactions serves to sustain the overall heat in the system. Upon
leaving the reactor, the main gas flow passes downward through a
second heat regenerator (12), giving up its heat to the packing
in that vessel before leaving the catalytic reduction system
through flow reversing valve (9). A thermal balance is maintained
in the system by passing a minor flow of the hot gases from
reactor (11) around the downflow regenerator and the flow
reversing valve (9) and remixing with the main stream before
entering sulfur condenser (17)
The primary function of the heat regenerators (10) and (12) ,
then, is to remove heat from the gases leaving the catalytic
reactor (11), and to utilize this heat to raise the temperature
of the incoming gases to the point where the S0«-natural gas
reaction will initiate. The direction of flow through the two
regenerators is periodically reversed to interchange their functions
of heating and cooling the gases. This is accomplished by the
flow reversing valve (9) and four water-cooled butterfly valves
(13), (14), (15) and (16). The valve arrangement shown in the
flow diagram Figure I is specially designed to maintain the gas
flow through the catalytic reactor (11) in one direction only.
All five valves are operated from a central control system which
synchronizes their movement perfectly so that each flow
reversal is completed in less than one second.
The elemental sulfur that is formed in the primary reactor
system is condensed in a horizontal shell-and-tube steaming
condenser (17). This represents over 40% of the total recovered
sulfur. The process gas stream then enters the first stage
(18) of a two-stage Glaus reactor system where the following
exothermic reaction occurs:
2 H2S + S02 ^ 3/2 S2 + 2 H2O
After the first stage of Glaus conversion, the gas is cooled
and additional sulfur condensed by passage through a vertical
steaming condenser (19). Further conversion of H2S and SO2 to
sulfur takes place in the second stage Glaus reactor (20).
This sulfur is condensed in a third steaming unit (21). A
coalescer (22), containing a mesh pad, then removes entrained
liquid from the gas stream. Molten sulfur from the three con-
densers and the coalescer is collected in a sulfur holding pit
662
-------
(23) from which it is pumped to storage. Residual H_S in the
gas from the process is oxidized to SC<2 in the presence of
excess air in an incinerator (24) before being exhausted to the
atmosphere through a stack (25) .
APPLICATION IN COMBINED SYSTEMS
The versatility of the Allied Chemical SCU reduction
technology can best be illustrated by considering its application
as a component part of several stack gas treatment systems for
electric utility generating stations. As was mentioned earlier,
the Allied Chemical process can accommodate a wide range of SC>2
concentrations in the feed gas. A process flow diagram of the
SOn reduction unit as it is applied to flue gas desulfurization
systems which produce strong SO gas, say 40% or more (dry basis)
is shown in Figure II. Since tne gases from S02 concentrating
systems are, generally cleaned by some type of wet scrubbing
operation before delivery to the SO- reduction unit, and are
saturated with water vapor at the scrubbing temperature, it is
necessary to raise the temperature of the gases sufficiently
above the dew point that no condensation of weak acid occurs
in the main blower. After the preheated gases are compressed,
the reducing gas is added in the correct proportion and the gas
mixture then passed through a feed gas heater where its temperature
is raised above the dew point of the sulfur that is formed in the
primary reaction system. The purpose of this is to preclude
sulfur condensation in the four-way reversing valve mentioned
in the earlier process description. The source of this heat
will be discussed subsequently.
The functions of the reduction reactor and the heat re-
generators are essentially the same when treating gases having
high SC-2 concentrations as when treating roaster gases. The
exothermic reactions which occur in this primary reactor system
result in the gases leaving the system at a substantially higher
temperature. A portion of this heat is used to heat the feed
gas in a shell and tube heat exchanger. However, because of the
excess heat liberated in the reactor system, the exit gases are
too hot, and, as a result, are too corrosive to be passed directly
through this heat exchanger. Consequently, they are first tempered
by diverting a part through a gas cooler, and remixing this side
stream with the main stream before entry into the feed gas heater.
Some sulfur is condensed in the cooler and withdrawn to a sulfur
holding pit. Low pressure steam is generated on the shell side
of the cooler.
663
-------
The two-stage secondary reactor, or Glaus converter, system
is similar to that used for roaster gases. Residual H2S in the
exit gas from the Glaus system is oxidized to S02 in the presence
of free oxygen in an incinerator. Instead of being released to
the atmosphere, however, as in the roaster gas treating process,
the incinerator tail gas is recycled to the flue gas desulfuriza-
tion unit. The S02 emission to the atmosphere is thus reduced
to the minimum level attainable.
Still another adaptation of the Allied Chemical SC>2 reduction
process is represented in the process flow diagram, Figure III.
Here, the only significant difference is the quantity of heat
generated in the reactor-regenerator system, with an 8% SO2,
low oxygen gas, the temperature rise in the reduction reactor
system is inadequate for the exit gas from the reduction reactor
system to be used efficiently in the feed gas heater. This makes
it necessary to use the incinerator tail gas for this purpose,
as shown in the flow diagram. The remainder of the process is
the same as for the more concentrated SO2 gases. The tail gas
is again returned to the front end S02 gathering system.
For a given emission control problem, it will be apparent
that the fixed amount of sulfur delivered into the Allied SO_
reduction unit can be contained in greatly differing gas volumes
depending upon the choice of flue gas treating system. This
is illustrated in Table I.
Note that the quantity of sulfur available for recovery
in the case of a 600 MW boiler fired with oil of 4% sulfur
content is 130 long tons per day compared to the 500 long tons
per day in the roaster gases at Falconbridge when operating at
rated capacity. It will also be seen that the volume of gas in
which that 130 long tons per day of sulfur is contained may
range from as little as 2,500 SCFM to nearly 32,000 SCFM. As
an added note of perspective, this latter value is substantially
less than one-half of the roaster gas flow entering the S02
reduction unit at Falconbridge.
REDUCING AGENT REQUIREMENTS
The reducing agent is the largest single element of operating
cost associated with the SO., reduction process. The consumption
varies directly with the quantity of SG>2 to be reacted, as is
illustrated in Figures IV and V.
The natural gas requirements for the reduction of SC<2
recovered from coal burning electric utility stacks is shown in
664
-------
Figure IV. The sulfur values on the curves are based on the use
of 10.6 tons of coal (containing 10% moisture) per megawatt day.
Corresponding figures for electric power generation using 2%
and 4% sulfur content residual fuel oil are presented in Figure V.
Natural gas was the logical reducing agent for the sulfur
recovery facility at Falconbridge because of the proximity of a
major trans-Canada pipeline. One phase of Allied's continuing
effort in the development of advanced SO,, reduction technology
is focused on the substitution of other gaseous and liquid
hydrocarbons for the natural gas utilized in the present process.
CAPITAL COST COMPARISON
The installed cost of an SO2 reduction plant is primarily
a function of the gas volume processed and secondarily the quantity
of sulfur contained in that process gas stream. Figure VI shows
the comparative relationship of the estimated installed costs of
Allied Chemical SO2 plants applied to the three feed gas concen-
trations we have examined — 8%, 40%, and 100% S02* dry basis.
The effect of S02 concentration on capital cost is readily
apparent, as is the fact that SO2 reduction is strongly favored
by increasing system size.
Such a comparison does not, of course, reflect total flue
gas desulfurization system installed cost., In fact SO2 reduction
will be a minor fraction of the total installed cost in every case.
Allied Chemical's studies to date have indicated that location
factors have a greater influence on total installed cost than on the
cost of the S02 reduction component alone. As a consequence, it
does not automatically follow that the lowest capital cost S02
reduction unit will insure the lowest total system installed cost.
Operating factors, particularly energy requirements and costs,
can vary widely among different combinations of flue gas desulfuriza-
tions systems. The optimization of operating requirements can only
be undertaken for the combined SO2 gathering and S09 reduction systems,
Allied Chemical is prepared to work with organizations having
SO2 emission problems, making its commercially proven SO2 reduction
technology available for direct application or in combination
with SO2 concentration systems provided by others.
665
-------
ALLIED CHEMICAL S02 REDUCTION TECHNOLOGY TYPICAL ROASTER GAS APPLICATION
Fig. I
O>
REGEN
1
— JT
NATU
GA
MA
CO
rtEAT
ERAT(
S02 GAS
I HOT GAS HEAT , ,
1 EXCHANGERS 2
—3 L_ -JT- ' j L^
I'll
RAL J ^
5
EFFLUENT
t
i . . , „
--K ^1^ /^^ 6
3 ^ ^ x CONDENSING " frrf
^_____Z^:±:TI 4- TOWER
i — _, -^^-s- / \ luni-n .
^ -^TJIC «. •
. llflS •" , „- J Jl-H
in C0°L|NG -i ^^^
IU1 Tfiiurn .,/r i r"~7-N I
NEUTRALIZATION 1UiilJI ri^5^ ^| ~r
8 3 1^ f*~
N BLOWER «cmE ^P
__ J
LD GAS BY-PASS
r— M-— *-txi ,
14 —^5 |
n REDUCTION
11 REACTOR
13 16
1 o
¥ C/*
j • 2j
17 Sr in HEAT
|R 1Z ^ '" REGENERATOR
_ i—
<=> i
ac f
A J
ix>
T9
— n ,, fltIO UDOUR !
DRIP ACID
1
MIST TOWER 1 "
_t \ TWn-STAGE A|n
CLAUS REACTOR ftm
^20\ NATUR
STEAM *n CTC»U f7c
oltHm fc STEAM UAb
/^^=J -^ STEAM — nr hir— ' ^^
i — Lt5=~t~77 m «, cumin ^21
i ^^^S' / CIHCIID IS 21 lULriiK 11*01*0
fiurae^i SULFUR ',?, r.1 rnNnrjjcrD **
SULFUR T rnunruccD 1' i iunutnatK 22
i/UNULNdLn — H — _
REHEAT GAS 1
1 SULFUR SULFUR \ SULFUR SULFUR
1 -
\ T fc Tfl CTHDAPr
J f n •" olORAht
r — M
[ 23 1
SULFUR HOLDING PIT
x^
-^1
ELECTROSTATIC
.PRECIPITATORS
n
~^1
25
&_J»]
INCINER
-i ft CTi
P5
COALESCER
ATOF
CK
Allietl
rhemical
-------
ALLIED CHEMICAL SQ2 REDUCTION TECHNOLOGY 100% AND 40% SO? FEED (DRY BASIS)
Fig. II
SIEAM REDUCING
1 GAS ~] \ u=:=i i
tft nip . / \ • 3
olio bAo •• — 1 f-c j 2 .
PRFHFATpT T Su FEED GAS
PRtHtATER f MAIN HEATER
BLOWER
CTF1U — -" — ~~ ~~~~~~~~-.
STEAM
i ^
, T
^•fcj t^t>— ^ i
^-^K^^y J>7=CT
C|iir||R 1^7 ** iMii tTFAM *
oULrUK f J pnuutDTrnp T — *-oltABI
CONDENSER T CONVERTERS 8
cniniR ~^ cm FUR
10
outrun outrun
CONDENSER CONDENSER
°i SULFUR
. REDUCTION STEAM
REACTOR i
SYSTEM 1 Y^l^^^
REACTOR EXJT^J^]
GAS COOLER T |
1 1 1
(
STEAM i 1
***** FUEL *
n GAS ^ 13 — '
1 , COALESCER L_ INCINERATOR
SULFUR
9 -Tn ^TnRARF & — A|R . —
TAIL
-GAS
SULFUR HOLDING
PIT lc
12
STEAM
-------
ALLIED CHEMICAL SO, REDUCTION TECHNOLOGY 8% SO, FEED (DRY BASIS}
Fig.II!
TAIL GAS
SQgGAS
REDUCTION
REACTOR
SYSTEM
4
SULFUR
TO STORAGE
SULFUR HOLDING PIT
AUied
(jlemical
-------
_ 4
3
GO
1
0
NATURAL GAS REQUIRED FOR S02 REDUCTION
AT COAL BURNING ELECTRIC UTILITIES
100 200 300 400 500
ELECTRICAL POWER GENERATED-MEGAWATTS
Fig. IV
-------
^ 3.0
2 2.5
O
c
CO
- 2.0
1.5
00
oo
= 1.0
0.5
0
NATURAL GAS REQUIRED FOR S02 REDUCTION
AT ELECTRICAL UTILITIES BURNING FUEL OIL
100 200 300 400 500
ELECTRICAL POWER GENERATED-MEGAWATTS
600
-------
$ (MILLIONS)
10
8
6
5
4
5 3
Fig. VI
2
100
8% SO 2
40% SO2
100% S02
ALLIED CHEMICAL S02 REDUCTION
TOTAL INSTALLED COST
(Battery Limits Plant)
1 1 1 1 1 1
S02
1 1
COMPOSITIONS ARE DRY
1 1 1
BASIS
200 300 400
CAPACITY (MEGAWATTS)
600
-------
Table I
FEED GAS TO ALLIED CHEMICAL S02 REDUCTION UNIT
MEGAWATTS GENERATED
100 300 600
EQUIVALENT SULFUR IN GAS (Long Tons/Day]
21.7 65
130
% SO 2 IN GAS (Dry Basis)
100%
40%
VOLUME OF GAS (MSCFM)
0.4 1.3 2.5
1.0 3.1
6.3
5.3 15.8 31.6
Based on 4% sulfur in fuel oil assuming 90% recovery
of SO 2 in the associated flue gas desulfurization step.
-------
THE CAT-OX PROJECT AT ILLINOIS POWER
by
W. E. Miller, Director, Environmental Affairs
Illinois Power Company
Decatur, Illinois
673
-------
THE CAT-OX PROJECT AT ILLINOIS POWER
ABSTRACT
The Catalytic Oxidation method developed by Monsanto Enviro-Chem Systems, Inc.
for removing sulfur dioxide from flue gas of fossil fuel generating stations has
passed the pilot plant stage. The first commercially sized installation was
operated for a short period in September, 1972 on the 100 Mw Wood River #4 unit
of Illinois Power Company. The project is being financed jointly by the Control
Systems Laboratory of the Office of Research & Monitoring of Federal EPA and by
Illinois Power Company. This article describes why Illinois Power Company chose
Cat-Ox for a demonstration installation in an attempt to control sulfur oxide
emissions and also describes the operation of the Cat-Ox System.
INTRODUCTION
The removal of sulfur dioxide from the flue gas of fossil combustion units is a
major concern of public utilities since they rely heavily on fossil fuel for the
generation of electricity; therefore, much effort is being expended in close co-
operation with the chemical industry in the development of an acceptable method of
sulfur removal. At the present time a proven method of removing sulfur dioxide
from flue gas is not available although many demonstration installations of the
scrubber type are now being installed by utilities. The Cat-Ox System, the chemi-
cal conversion method described in this paper, has now been installed on a demon-
stration basis by Illinois Power Company. A demonstration is considered to be a
successful installation whose scale and length of operation are great enough to
provide full determination of feasibility for commercial application. It is hoped,
therefore, that the Cat-Ox System now installed on a demonstration basis will be
successful and can be offered for general use.
674
-------
ILLINOIS POWER COMPANY
Illinois Power Company is a combination electric and gas investor owned utility
serving territories in central and southern Illinois. To supply its electrical
needs, the five fossil-fueled generation stations shown in Figure 1 combine to
provide a capacity of 2200 Mw.
Baldwin #2, a 600 Mw coal-fired unit has been completed recently and Baldwin #3,
another 600 Mw coal unit, will be completed in 1975. With the addition of Baldwin
#3, the company capacity will be 3,400 Mw in 1975. Since coal is used for gener-
ating the major portion of this capacity, it is vitally important to Illinois
Power that reasonable methods of eliminating pollutants from coal either before
or after combustion be developed as rapidly as possible.
The Wood River Station of Illinois Power Company is located in Madison County
near East Alton, Illinois, which is in the St. Louis Major Metropolitan Area.
The Cat-Ox System is being installed, on Unit #4 at Wood River which has a nominal
rating of 100 Mw; the boiler is a Combustion Engineering reheat, natural circula-
tion, balanced draft, pulverized coal, tangentially fired, unit with steam condi-
tions of 1500 psig and 1005°F at the superheater outlet. The Cat-Ox System is
designed for a flue gas flow rate of l,120,000#/hour which is equivalent to a
load of about 110 Mw.
The unit burns approximately 275,000 tons of coal per year with a sulfur content
of 3.1%. Based on this, the Cat-Ox System theoretically should produce about
19,000 tons per year of equivalent 1007. sulfuric acid.
675
-------
CAPITAL AND OPERATING COSTS
The capital costs involved in the design and construction of the Cat-Ox System
were estimated at $7,300,000, including the costs of providing off-battery ser-
vices. Of this amount Illinois Power Company contracted to pay $3,800,000 and
the Office of Research and Monitoring of Federal EPA contracted to pay $3,500,000.
The battery limit services include supply facilities for natural gas and fuel oil
for reheat burners, cooling water for acid coolers and miscellaneous uses, electric
services for ID fans, electrostatic precipitators and pumps and facilities for ash
disposal. Final costs now appear to be $8,150,000, giving us an installed cost of
$74 per Kw based on the 110 megawatt gross capacity of the generating unit.
The Federal EPA will undertake an extensive test and evaluation program to fully
characterize the Cat-Ox System and to provide data on its emission reduction capa-
bilities as well as the operating costs for the system to form a basis for compari-
son with other installations. It is expected that the data will be made available
to all interested parties and will be used by EPA, regulatory agencies, and members
of the utility industry.
For a period of five years, Enviro-Chem will be responsible for the disposal of
the sulfuric acid produced by the system with net revenues from acid sales credit-
ed 757o to Illinois Power and 25% to Enviro-Chem. Operating and maintenance costs
are expected to run $600,000 per year after the initial de-bugging period; assuming
a price of $8 per ton of acid, the net operating and maintenance costs should be
about $500,000 per year, excluding fixed charges on capital costs.
676
-------
THE CAT-OX PROCESS
The operation of the Cat-Ox System consists basically of the following six
separate phases:
1. Fly Ash Collection
2. Conversion of S02 to 503
3. Heat recovery
4, Removal of sulfuric acid
5. Acid mist elimination
6. Acid storage and loading
These basic steps are shown diagrammatically in Figure 2 and are described below.
1. Fly Ash Collection
The existing mechanical collector remains in service on Unit #4
to remove most of the fly ash from the flue gas. A new Research-Cottrell electro-
static precipitator with a design efficiency of 99.6% has been installed to work
in series with the mechanical collector to remove essentially all the particulate
matter from the flue gas. After leaving the electrostatic precipitator, the cleaned
flue gas is heated and passes into the converter of the Cat-Ox System or, during
start-up or unusual operation, can be by-passed directly to the stack. The fly
ash collected by the precipitators is conveyed pneumatically to the existing ash
pit area. The electrostatic precipitator installation was completed in February
1972 and has been operating with Unit #4 since that time.
2. Conversion of S02 to SO-j
The temperature of the flue gas leaving the electrostatic precipi-
tator is 310°F and must be reheated to 850°F to allow a 90% conversion of S02 to
503. It was proposed that this be done by two in-line reheat burners using
natural gas or No. 2 fuel oil and by recovery of sensible heat from the treated
677
-------
flue gas. The reheat burners were to be designed to maintain the 850°F conversion
temperature regardless of boiler load. Modifications are now being made on this
re-heat system to permit operation totally on #2 oil. Following reheat to con-
version temperature, the flue gas enters the converter where the Cat-Ox \^} catalyst
(a vanadium pentoxide catalyst) reacts with the S02 to form 803. The converter is
designed so that the catalyst bed can be emptied onto a conveyor system for trans-
port to a screening process after which the cleaned, catalyst is conveyed back to
the converter. About 2.57, of the catalyst mass is lost during each cleaning process
which is anticipated to occur about four times per year. About forty-eight hours
is required for each catalyst cleaning.
3. Heat Recovery
The treated flue gas, now containing 803, passes to a Ljungstrom-
type heat exchanger where about 400°F sensible heat is recovered to heat the in-
coming untreated flue gas. As a result of heat recovery in this exchanger, the
overall need for fuel usage is to add 150°F of sensible heat in the Cat-Ox process.
The temperature of the gas is maintained well above the dew point. Normal flue gas
leakage in a regenerative heat exchanger of this type will allow about 5% of the
flue gas to by-pass the unit thereby reducing the overall efficiency of S02 removal
to approximately 85%.
4. Removal of Sulfuric Acid
The flue gas is further cooled in a packed-bed absorbing tower
which operates in conjunction with an external shell and tube heat exchanger.
During cooling, the H20 and 863 combine to form sulfurlc acid which is subsequent-
ly condensed. The tower brings a cool stream of sulfuric acid into direct contact
with the rising hot flue gas. Exit gas leaves the packed section at about 250°F
while hot acid is constantly being removed from the bottom of the tower and cooled
678
-------
in the external heat exchanger and sent to storage.
5. Acid Mist Elimination
Very fine mist particles of sulfuric acid are formed, in the gas
as it is cooled in the absorbing tower. These mist particles in the flue gas
are removed along with some entrained droplets of circulating acid from the
tower by the Brink vS/ mist eliminator system. The packed section of the absorb-
ing tower and. the mist eliminators are contained within one vessel. The flue gas
leaving the mist eliminator to enter the exit stack contains less SOo than the
amount normally emitted from the combustion process.
6. Acid Storage and Loading
The cooled acid amounting to 12 gallons per minute of 78% l^SO^ is
collected in two 400,000 gallon steel storage tanks. An acid loading pump and
tank car loading facilities are provided adjacent to the storage tanks. Tank
trucks may be loaded from this station if desired.
PROJECT TIMETABLE
The time schedule for the Cat-Ox System includes:
Design and Capital Cost Estimates - Initiated June 1970
Detailed Engineering & Equipment Procurement - Initiated November 1970
On Site Construction - Initiated January 1971
Electrostatic Precipitator - Placed in Operation February 1972
Entire Cat-Ox unit scheduled for operation mid-1973
Data Phase including one year operational testing by Federal EPA *
estimated to be completed October 1974.
DELAY IN TIMETABLE
The Research-Cottrell precipitator was completed on schedule and was placed in
operation in February, 1972. The Cat-Ox System was scheduled to begin operation
679
-------
in June, 1972 but, due to construction delays, initial start-up using gas as the
in-line heater fuel did not occur until September 4, 1972. Preliminary opera-
tions in September and October indicate that the unit is meeting the performance
criteria set for the system. Conversion of S02 to 303 reached 93% which is over
the 90% guaranteed. The Cat-Ox unit removed 89% of S02 admitted to it which is
over the 85% guaranteed. Acid of 78% concentration was produced when operating
the absorbing tower at design temperature. Acid mist exiting the system was below
0.3 mg/cu. ft. against a design criterion of 1.0 mg/cu. ft. The system handled the
full 1,120,000 Ib/hr. of flue gas when meeting these criteria.
Over 1,400 tons of acid were produced during this period and two tank cars of the
acid have been sold to a fertilizer manufacturer to assure its acceptability.
In October it became evident that the shortage of natural gas would preclude its
use as a reheat fuel for the Cat-Ox unit. The existing burner equipment in the
Cat-Ox had to therefore be modified to operate totally on #2 fuel oil. Design
modification and equipment changes are ia progress. As soon as they are completed,
the Cat-Ox System will again be brought on stream.
In a new generating unit, the necessity for reheat would not exist. The flue gas
would be taken from the boiler at a point near the economizer or super-heater at
a temperature of 850-900°F. A hot precipitator would then remove the particulate
matter and let the flue gas enter the converter at the required 850°F. For this
reason, based on the limited operation we have had to date, we believe the Cat-Ox
System might be a feasible system for installation in a new coal fired station.
Retrofit application will be proven or disproven by the Wood River #4 installation-
680
-------
GUARANTEE AND OPERATING PROGRAM
One or two 24-hour tests by Enviro-Chem will be conducted to assure that the
Cat-Ox System meets the following guarantees:
1. The system is capable of operating with a gas flow of
l,120,000#/hour entering the system at 310°F.
2. The system is capable of producing 60° Baume' (77.7% l^SO^)
sulfuric acid.
3. The exit gas emitted to the stack does not contain, on the average,
more than 1.0 milligrams of 100% sulfuric acid mist per actual cubic
foot of gas when the system is operated at rated capacity.
4. The conversion of S02 to S03 of the gas entering the converter
is at least 90% at rated capacity.
5. The system operates so that over 99% of the fly ash in the flue
gas leaving the boiler is removed when operating at rated capacity.
6, The system removes 85% of the S02 in the flue gas entering the system.
Following the 24-hour test period, Illinois Power Company will operate the Cat-Ox
System for a minimum of one year as mentioned previously. During this period, and
for a subsequent period of four years, if Illinois Power Company decides to continue
to operate the system, data will be obtained to evaluate the following items:
1. Operating characteristics and plant performance (relative to
S02 and fly ash removals and to H2SO^ recovery).
2. Maintenance procedures, requirements, and costs,
3. Total process operating costs.
Figures 3 and 4 show the Cat-Ox installation as it nears completion.
SUMMARY
In an attempt to advance the frontiers of knowledge in the science of removing
sulfur dioxide from flue gas, Illinois Power, with Battelle Institute, conducted
681
-------
a comprehensive survey of possible sulfur removal systems. The possibility of
using low sulfur Western coal was considered but it was found to be less efficient
than local coal, and precipitator efficiency declined with its use; in addition, it
cost three times as much. As a result of these studies, Illinois Power Company
decided that the Cat-Ox System, after both a pilot installation and prototype
installation, was most nearly ready for commercial demonstration. In 1970 the
capital costs being considered for Cat-Ox were much higher than those proposed for
other S02 control systems. However, as more actual experience has been gained
since 1970, it seems that the capital costs originally estimated by Enviro-Chem
for Cat-Ox were more realistic than those estimated for other systems, primarily
because of the decade of research and the advanced stage of development of the
Cat-Ox System.
The Cat-Ox System is expected to remove 85% of the sulfur dioxide from the flue
gas as 787o sulfuric acid and to remove essentially all of the fly ash. Sale of
the sulfuric acid could offset some of the operating costs. Illinois Power Company,
by removing the pollutants from the environment and conserving natural resources
by recovering a product which is presently being thrown away, is not solving One
pollution problem and creating another one. While some start-up difficulties have
been encountered with the Cat-Ox System, we are still hopeful it will prove to be
a feasible method and that it may solve the sulfur removal problem in a new coal
fired generating unit and. possibly even in a retrofit installation.
ACKNOWLEDGMENT
We hereby express grateful acknowledgment to the Control Systems Laboratory of
the Office of Research and Monitoring of the Federal Environmental Protection
Agency for their financial assistance and excellent cooperation in this major
project.
682
-------
.'> . qw«u_J:, .L^; r „_!
FIGURE 1 TERRITORY SERVED BY ILLINOIS POWER COMPANY
683
-------
THE REHEAT
CAT-OX
SYSTEM
FLUE GAS
FROM EXISTING
ID FAN
PRECIPITATOR
CONVERTER
REHEAT;*
BURNERn
-DAMPERS
t I
GAS HEAT
EXCHANGER
SULFURIC
ACID
STACK
CAT-OX
MIST
ELIMINATOR
ABSORBING
TOWER
ACID
COOLER
RECYCLE
STORAGE
FIGURE 2
-------
Monsanto
3
CO
tn
Monsanto
Envjrol
IChem
Systems Inc
REHEAT CAT-OX SYSTEM
-------
REHEAT CAT-OX SYSTEM
ILLINOIS POWER - WOOD RIVER UNIT NO. 4
686
-------
MITRE TEST SUPPORT
FOR THE GAT-OX DEMONSTRATION
PROGRAM
by
G. Erskine
E. Jamgochian
The Mitre Corporation
McLean, Virginia
687
-------
ACKNOWLEDGMENTS
The work described In this paper wae performed by the MITRE
Corporation under the sponsorship of Mr. G. S. Haselberger of the
Control Systems Laboratory of the Office of Research and Monitoring,
U. S. Environmental Protection Agency. Members of the MITRE team
who participated in the program include: Dr. J. S. Burton, Mr. G. Erskine,
Mr. E. M. Jamgochian, Mr. N. T. Miller, Mr. J. P. Morris, Mr. R. Reale,
and Mr. W. L. Wheaton. Contributions to the program were also made by
Mr. D. C. Simpkins of Consultants and Designers, Inc., and members of
the Midwest Research Institute under subcontract to the MITRE Corporation.
688
-------
ABSTRACT
This paper describes MITRE test support for the Cat-Ox Demonstration
Program. The overall scope of MITRE*s test program is outlined, and
specific task efforts accomplished to date are presented. A detailed
description is provided on the individual tests of the Baseline Measure-
ment Program, and select test results are provided concerning: grain
loading measurements, comparisons of continuous and manual measurements
of S02 with theoretical values, comparisons of continuous and manual
measurements of NO , sulfur balance measurements, and comparisons of
Jt
elemental concentrations of coal and ash throughout the steam generator
system. The paper further discusses current activity in preparation
for the One-Year Demonstration Test Program. In particular, the
measurement parameters and design of the continuous measurement instru-
mentation system for the evaluation of process performance are presented;
and the instrumentation and facilities for the measurement of gas con-
centrations, gas flow, and automatic recording of data are reviewed.
689
-------
MITRE TEST SUPPORT
FOR THE
CAT-OX DEMONSTRATION PROGRAM
Background
MITRE's contract with the Environmental Protection Agency was
initiated in April of 1971. Under this contract, MITRE is required to
provide technical support for the Cat-Ox Demonstration Program in terms
of: 1) program management assistance, 2) the development of test
plans, 3) the design, development, and installation of measurement
systems, 4) services related to the implementation of test programs,
and 5) evaluation of test results with application of the results to
the total industry.
As a first step, a number of program plan alternatives were
developed and evaluated. These plans, and the details of the overall
program plan which was selected, are described in our Management Plan
for Test Support for the Cat-Ox Demonstration Program (MITRE Technical
Report No. 6054 dated July 1971). A summary of the MITRE task efforts
defined for the program is provided in Figure 1.
Major Task Areas
The first of the task areas concerns the "Definition of Test
Requirements," in which MITRE examined the requirements of potential
users, reviewed the technical and economic capabilities of the Cat-Ox
process, and defined the "window" of test conditions at the steam
generator/Cat-Ox interface for the demonstration test (see Figure 2).
This task was completed in November of 1971 with the definition of
690
-------
TASK 1
TASK 2
TASK 3
cr>
vo
TASK 4
DEFINE TEST REQUIREMENTS
BASELINE MEASUREMENT PROGRAM
CAT-OX DEMONSTRATION
- TEST PREPARATION
- ONE-YEAR MEASUREMENT PROGRAM
EVALUATION OF RESULTS
(COMPLETED)
(COMPLETED)
TO BE COMPLETED AUGUST 1973
SEPTEMBER 1973 - SEPTEMBER 1974
SEPTEMBER 1974 - MARCH 1975
FIGURE 1
PROGRAM SUMMARY
-------
EXAMINE REQUIREMENTS OF POTENTIAL USERS
• SPECIFY TECHNICAL & ECONOMIC CAPABILITIES
OF CAT-OX PROCESS
o> • DEFINE "WINDOW" OF TEST CONDITIONS AT
£ STEAM GENERATOR/CAT-OX INTERFACE FOR
DEMONSTRATION
FIGURE 2
TASK 1 - DEFINITION OF TEST REQUIREMENTS
-------
interface flue gas conditions (temperature, gas flow rate, SO. concen-
tration, and grain loading) for 89 existing steam generators which
were categorized as potential users of the reheat Cat-Ox system.
Interface flue gas conditions were also determined for 75 steam genera-
tors to be constructed by 1976 which were categorized as potential
users of the integral design Cat-Ox system.
The second of our task areas concerns the "Baseline Measurement
Program." The specific objectives of the Baseline Measurement Program
as shown in Figure 3 are to 1) characterize baseline performance in
terms of operability, reliability, and emission levels prior to
installation of the Cat-Ox process; 2) to determine the relationship
between control settings and operating conditions for the steam
generator, and flue gas properties at the Cat-Ox steam generator
interface; 3) to test and calibrate measurement procedures and hardware
to be used in the one year demonstration test; and 4) to obtain
quantitative data supporting the establishment of realistic performance
standards for pollutants other than SO- and particulates.
The third task area concerns the performance of a "One-Year
Demonstration Test" wherein a measurement program is conducted which
will fully characterize Cat-Ox process emission control performance;
quantify the operating economics of the process; establish the
operability, reliability, and maintainability of the Cat-Ox process;
and determine the resulting effects on the steam generator with which
the process is integrated.
693
-------
• DETERMINE BASELINE EMISSION LEVELS, OPERATING EFFICIENCY, AND
RELIABILITY OF STEAM GENERATOR IN ACCORDANCE WITH A.S.M.E.
POWER TEST CODES
• DEFINE STEAM GENERATOR CONTROL SETTINGS AND OPERATING CONDITIONS
TO PRODUCE DESIRED FLUE GAS PROPERTIES AT THE CAT-OX INTERFACE
TEST AND CALIBRATE MEASUREMENT PROCEDURES AND HARDWARE TO BE USED
IN THE ONE-YEAR DEMONSTRATION
PROVIDE EMISSION INFORMATION TO EPA TO SUPPORT THE ESTABLISHMENT
OF REALISTIC EMISSIONS STANDARDS
FIGURE 3
BASELINE MEASUREMENT OBJECTIVES
-------
The fourth task area concerns the "Demonstration Evaluation" where
reduced test data are translated into quantified statements on the
technical and economic adequacy of the process.
Scope of Baseline Measurement Program
To meet the objectives of the Baseline Measurement task (Task 2),
a test program was initiated with six preliminary tests wherein
necessary background information was obtained on isokinetic sampling
techniques, rates of particulate loading in sampling equipment, and
effects of power plant ambient conditions on the measurements. This
information was then used to determine ranges of gas and particulate
concentrations to be encountered in the measurement program, to confirm
the sampling frequency and sampling positions utilized in the measure-
ment program, and to identify any changes in operating procedures or
modifications in test equipment necessary for the primary measurement
effort. These preliminary tests were conducted over a two-week period
beginning September 28, 1971 (see Figure 4).
The main measurement program was initiated on November 8, 1971
following the completion of the preliminary measurement effort and the
completion of all required facility modifications (ports, sheds, and
stack access platform), Each of the twenty-one tests of the main
measurement program was of approximately 10 hours duration. During
the 10 hour period of each test, the major steam generator parameters
(load factor, fuel type, soot blowing cycle, excess air, and burner
angle) were fixed at a predetermined operating level. The combinations
695
-------
en
(£>
6 PRELIMINARY TESTS BASED ON:
- LOAD FACTOR: 35%, 100%
- EXCESS AIR: MINIMUM, MAXIMUM
- SOOT BLOWERS: NONE, RETRACTABLE AND WALL BLOWERS
- TEMPERATURE AND VELOCITY TRAVERSE AT ECONOMIZER
- TEMPERATURE, VELOCITY, AND GRAIN LOADING TRAVERSE AT AIR HEATER
• 21 INDIVIDUAL TESTS BASED ON:
- LOAD FACTOR: 35%, 50%, 75%, 100%
- FUEL TYPE: 3 LBS S/M BTU, 1.5 LBS S/M BTU, 1.0 LBS S/M BTU
- SOOT BLOWING: NONE, RETRACTABLE BLOWER AND WALL BLOWERS
- EXCESS AIR: MINIMUM, NORMAL, MAXIMUM
- BURNER ANGLE: NORMAL
• 2 TESTS BASED ON:
- TEMPERATURE, VELOCITY, AND GAS CONCENTRATION TRAVERSALS AT STACK AND AT
AIR HEATER
- LOAD LEVEL; 100%, 50%
- FUEL TYPE: 3 LBS S/M BTU, 1.5 LBS S/M BTU
1 TEST BASED ON:
- NO GAS CONCENTRATIONS AT STACK AND ECONOMIZER
- BURNER ANGLE: MINIMUM THROUGH MAXIMUM
FIGURE 4
BASELINE MEASUREMENT PROGRAM
-------
of operating levels were selected so as to provide the maximum of
information in a minimum number of tests, varying the parameters on
a "one-at-a-time" basis.
Two supplementary gas traversal tests were also conducted to
determine the pattern of leakage at the air heater and the gas flow
pattern midway in the stack.
A supplementary test was also conducted in which all factors were
held constant except for burner angle, which was varied in steps from
the minimum to maximum position.
Baseline Measurement Parameters
For all of the tests, key steam generator operating parameters
were monitored, samples of coal and ash were obtained at various points
in the steam generator, gas samples were manually obtained, particulate
grain loadings were determined by manual sampling, and temperatures,
pressures, gas flows and gas concentrations were monitored by a MITRE
designed continuous measurement system.
These key parameters were measured as shown on Figure 5 at three
locations: 1) Location 1 - Prior to the economizer, was selected to
provide data at the relatively high temperatures corresponding to the
inlet of an integral Cat-Ox system designed for installation in new
steam generators, 2) Location 2 - between the upper and lower sections
of the air preheater - selected as representative of a lover temperature
condition prior to any existing or planned flue gas treatment system,
and 3) Mid stack - approximately 135 feet above the foundation of the
697
-------
LOCATION 1 PRIOR TO ECONOMIZER
LOCATION 3 MIDWAY IN STACK
TEMPERATURE
PRESSURE
FLOW
GAS COMPOSITION (SO , CO, C02, HC,
NO, N02, 02, H20^VAPOR)
01
V£>
00
LOCATION 2 IN AIR PREHEATER
LOCATION 3 MIDWAY IN STACK
TEMPERATURE
PRESSURE
FLOW
GRAIN LOADING
PARTICULATE COMPOSITION
PARTICLE SIZE DISTRIBUTION
ADSORBED MATERIALS
FIGURE 5
BASELINE MEASUREMENT PARAMETERS
-------
250 foot stack, selected to provide a condition representative of the
flue gas emitted to the atmosphere (and, in this instance, conditions
which will be seen at the reheat Cat-Ox/steam generator interface).
At all locations, the ports and sampling points were selected so as to
provide the proper number and distribution of measurement points in
accordance with ASTM Standard D 2928-71 ("Standard Method for Sampling
Stack for Particulate Matter").
Results of Baseline Measurement Program
A brief summary of the results of the Baseline Measurement Program
is provided in Figure 6. This paper summarizes a few of our major
findings, whereas the detailed results of our Baseline program are
presented in a report to be distributed by the Office of Research and
Monitoring ("Baseline Measurement Test Results for the Cat-Ox Demonstra-
tion Program," April 1973, Report No. EPA-R2-73-189).
Representative grain loading measurements are shown in Figure 7.
These measurements were taken at our position number 3, midway in the
stack. All measurements were taken using the EPA particulate sampling
train (with impingers) as described in the Federal Register, Vol. 36,
No. 247, December 23, 1971. The upper set of data shown in Figure 7
summarizes the results from a series of tests which were run with
coal from the Southwestern Mine of the Peabody Coal Company. Ranges
are shown in the data set for the sulfur content, ash content, and
the measured grain loading, representing the extremes of the measure-
ment. (As for example, six tests were run at the 100 mw power level
699
-------
COMPLETE CHARACTERIZATION OF POWER PLANT IN TERMS OF:
• TEMPERATURES, PRESSURES, AND GAS VOLUME FLOWS THROUGHOUT SYSTEM
• GAS CONCENTRATIONS THROUGHOUT SYSTEM AND EMITTED AT STACK
- S02> 02, C02, CO, HO, NO, N02, H20 VAPOR
• GRAIN LOADING AT AIR HEATER AND EMITTED AT STACK
• COAL FEED COMPOSITION: PROXIMATE, ULTIMATE, AND ELEMENTAL ANALYSES
• FLY ASH PARTICLE SIZE, COMPOSITION, ADSORBED MATERIAL
-j
o
0 • BOTTOM ASH COMPOSITION
COMPLETE SULFUR BALANCE
"MAPPING" OF STRATIFIED GAS FLOW AT AIR HEATER AND STACK
N0x EMISSIONS RELATED TO BURNER ANGLE
FIGURE 6
BASELINE MEASUREMENT RESULTS
-------
ASH CONTENT
LOAD
FACTOR
100 MW
75 MW
50 MW
100 MW
75 MW
50 MW
FUEL
TYPE
3.24-3.40%S
2.87-3.37%S
2.74-3.47%S
1.65%S
1.83%S
1.68%S
OF COAL
(AS RECEIVED)
9.9-10.7%
9.8-10.4%
9.6-10.8%
16.7%
13.6%
16.0%
GRAIN LOADING
(GRAINS/ACF)
.6-. 95
.5-. 57
.6-. 75
.78-. 90
.59
.76
EMISSION RATE
(LBS/HR)
2,100-3,000
1,000-1,200
650-1,000
2,500
1,300
1,100
FIGURE 7
BASELINE GRAIN LOADING MEASUREMENTS
(MIDWAY IN STACK)
-------
for which the fuel for each individual test ranged from 3.24% sulfur
to 3.40%) The lower set of data shown in Figure 7 summarizes the
findings from three tests each run at a single power level, using as
fuel a mixture of natural gas and coal from the Orient No. 3 mine in
Jefferson County, Illinois. The coal and gas were fired in combination
so as to simulate fuel with a sulfur content of 1 Ibs Sulfur/10 BTU.
The values shown in Figure 7 do, however, represent the actual measured
percentage sulfur and ash of the coal constituent of the mixture. The
data show that although the ash content of the coal in the lower set
of data was substantially higher than the upper set, grain loadings
and emission rates were comparable because of the diluting effect of
the natural gas fired with the coal.
Data from the same tests are shown in Figure 8, compared with
respect to measured SO,, concentration. Again, the lower data set
which shows sulfur content of the coal as 1.65%S, 1.83%S, and 1.68%S
is fuel equivalent to 1 Ibs S/10 BTU because of the blending with
sulfur-free natural gas. The theoretical values shown for S0_ con-
centration (and the other measured values shown for SCL concentration)
are based upon the actual measured excess air, they are not corrected
to any standard excess air condition. The manual SO,, values were
obtained using the procedures defined by EPA in the Federal Register
for determining sulfur dioxide emissions from stationary sources
(Federal Register, Vol. 36, No. 247 - December 23, 1971). The
continuous S0? values were obtained using MITRE's continuous measure-
ment instrumentation system, which included a Dupont 460 UV analyzer
702
-------
o
w
LOAD
FACTOR
100 MW
75 MW
50 MW
100 MW
75 MW
50 MW
FUEL
TYPE
3.24-3.40%S
2.87-3.37%S
2.74-3.47%S
1.65%S
1.83%S
1.68%S
MEASURED
EXCESS AIR
36-64%
27-52%
38-69%
41%
35%
34%
THEORETICAL
(PP&)
1760-1990
1780-2170
1760-2110
580
755
533
MANUAL
so,
(PPH)
1300-1850
1300-1430
360
425
445
385
CONTINUOUS
(PP&)
1755-2040
1740-2080
1780-1815
535
630
565
FIGURE 8
COMPARISONS OF CONTINUOUS & MANUAL
MEASUREMENTS OF S02 WITH THEORETICAL VALUES
(MIDWAY IN STACK)
-------
for SO. measurement. Both the upper and lower data sets shown in
Figure 8 are typical of the results of the total test effort where
good agreement was found between the continuous measurements and the
theoretical measurements, whereas the manual measurements were generally
lower than the theoretical values (averaged 76% of the theoretical
values).
Figure 9 shows a comparison between continuous and manual measure-
ments of NO for the same set of tests. The manual measurements were
j£
obtained using the standard EPA method for determination of nitrogen
oxide emissions from stationary sources as described in the previously
cited Federal Register. The continuous measurement results were
obtained using a Dupont 461-C UV instrument which measured both NO and
NO.. The results shown in the continuous measurement column are
primarily NO. No consistent patterns were found with respect to the
effect of test conditions on NO concentrations, except for the tests
performed with gas and coal fuel firing which produced NO levels
A
significantly lower than in the tests performed with the other fuel types.
Representative sulfur balance calculations are shown in Figure 10.
For each of the tests, the average coal consumption rate was determined
utilizing coal scale readings. The average sulfur content of the coal
(as determined by chemical analysis) was then used with the coal scale
readings to determine the rate of sulfur feed to the steam generator.
Average SO, mass flow readings from the continuous measurement system
were then used to determine the average sulfur flow from the stack.
The results are based upon measurement of SO,, concentrations in the
704
-------
o
VI
LOAD
FACTOR
100 MW
75 MW
50 MW
100 MW
75 MW
50 MW
FUEL
TYPE
3.24-3.40*8
2.87-3.37ZS
2.74-3.47ZS
1.65ZS
1.83ZS
1.68ZS
MEASURED
EXCESS AIR
36-64%
27-52Z
38-69Z
41Z
35Z
34Z
MANUAL
NO
(Ppfi)
320-650
270-450
325-405
265-365
308
235-260
CONTINUOUS
NO
(PPfi)
285
330-345
345
105
240
150
FIGURE 9
COMPARISONS OF CONTINUOUS & MANUAL MEASUREMENTS
OF NO (MIDWAY IN STACK)
-------
AVERAGE
SULFUR FEED IN
AVERAGE
SULFUR FLOW FROM
VJ
O
LOAD
FACTOR
100 MW
75 MW
50 MW
100 MW
75 MW
50 MW
FUEL
TYPE
3.24-3.40%S
2.87-3.37%S
2.74-3.47%S
1.65%S
1.83%S
1.68%S
COAL
(LBS/MIN)
35.8-46.0
31.7-36.0
21.5-29.2
10.4
12,8
7.3
STACK AS SO.
(LBS/MIN)
31.8-45.5
26.5-35.1
19.8-22.9
10.0
11.3
6.3
FIGURE 10
SULFUR BALANCE MEASUREMENTS
-------
stack and do not include measurement of the sulfur exhausted from the
stack as SO,, or the sulfur adsorbed on the surface of the ash through-
out the system. The results, however, show good agreement on the
sulfur balance leading to the conclusion that the total combined error
in SO2 and gas flow measurements was'low. For all except two tests,
the sulfur feed rate exceeded the sulfur flow rate measured in the
stack indicating that there were, in fact, small unmeasured losses of
sulfur.
Trace element concentrations were determined in four tests on the
coal feed, pulverizer rejects from the mills, bottom ash, and the fly
ash collected from the bottom of the air heater, the mechanical collector,
and in the duct at the air heater and at mid stack. Trace element con-
centrations were also determined for samples of fly ash collected in
the duct at the air heater and at mid stack for four additional tests,
as well as trace elemental analysis of pulverized coal for six additional
tests. In all cases, the analysis was performed for approximately 27
elements using atomic absorption. A sample comparison from these
analyses is provided in Figure 11 for thirteen of the elements of
special interest aa potentially hazardous materials. An exact material
balance of the elements cannot be performed because of the fact that
certain material flows such as bottom ash and ash from the mechanical
separator were not measured and can only be estimated. This type of
balance is also precluded by the fact that many elemental concentrations
are expressed as "less than" values. The results are, however, of
special value in estimating emission rates for a number of elements
707
-------
CO
ELEMENT
Ba
Be
Cd
Cr
Cu
Hg
Mn
Ni
Pb
Se
Sn
V
Zn
PULVERIZED
COAL
<.03
<.0002
.0006
.002
.002
<.0002
.009
.009
< 0.003
<.06
<.05
<.02
.11
PULVERIZER
REJECTS
<.03
<.0002
.0004
.002
.003
.00006
.007
.0009
< 0.003
<.06
<.05
<.02
.011
BOTTOM
ASH
_
.0004
<.005
.016
.007
.0003
.057
.013
.009
<.05
<.l
.04
.038
ASH FROM
MECHANICAL
SEPARATOR
.03
.0008
.002
.013
.007
.00004
.036
.040
< 0.003
<.07
<.05
<.02
,057
ASH FROM
AIR HEATER
DUCT
<.04
.001
.05
.010
<. 00002
.050
.050
0.020
<.05
<.05
.02
____
ASH FROM
MID STACI
<.04
.001
.002
.020
.080
0.020
<.05
<.05
.03
.090
FIGURE 11
COMPARISONS OF ELEMENTAL CONCENTRATIONS IN COAL
AND ASH THROUGHOUT SYSTEM
(WEIGHT PERCENT)
-------
not usually examined In emission testing programs. For this purpose,
the elemental concentrations mast be combined with the mass loading
measurements to provide the emissions data In useful form. An estimate
of the overall accuracy of the measurements can be noted by comparing
the values given for the ash from mid stack and with ash from the air
heater duct. The values from these two columns should be nearly
Identical, and as such represent essentially duplicate measurements.
Cat-Ox Demonstration Teat
Task 3 of MITRE's support program concerns the Cat-Ox Demonstration
Test. Preparation for the demonstration test Is currently underway.
This preparatory work Is scheduled to be completed In August of 1973,
after which MITRE will Initiate our one-year measurement program In
September of 1973.
The major areas of concern which MITRE will Investigate In the
one-year test program are summarized In Figures 12. and 13. A test
plan has been drafted which examines these areas In a systematic step-
by-step fashion. The prime area for Investigation concerns the overall
operating characteristics and performance of the Cat-Ox system relative
to the removal of S02 and fly ash and the production of sulfuric acid.
Other areas, Include the effectiveness of the catalyst as a function
of time, the effectiveness of the mist eliminator and any special
requirements for maintaining its effectiveness, the degree of seal
leakage encountered in the Ljungstrom rotary heat exchanger,
corrosion rates of the plant equipment, and the response of the Cat-Ox
709
-------
OPERATING CHARACTERISTICS AND PLANT PERFORMANCE (RELATIVE
TO S02 AND FLY ASH REMOVAL AND H2SO, RECOVERY)
• LONGEVITY OF THE CATALYST
• NECESSITY AND FREQUENCY OF MIST ELIMINATOR WASHING OPERATIONS
• DEGREE OF REGENERATIVE HEAT EXCHANGER SEAL LEAKAGE WITH TIME
FIGURE 12
CAT-OX DEMONSTRATION TEST - MAJOR AREAS OF CONCERN
-------
• CORROSION RATES OF PLANT EQUIPMENT (WITH SPECIAL EMPHASIS ON THE
ACID LOOP AND HEAT TRANSFER EQUIPMENT)
• RESPONSE OF PROCESS TO FUELS OF VARYING SULFUR CONTENT
• EFFECT OF CAT-OX SYSTEM FAILURE ON POWER PRODUCTION
• COMPONENT PRESSURE DROPS (AS A FUNCTION OF TIME)
• ABILITY OF POWER PLANT PERSONNEL TO OPERATE AND MAINTAIN SYSTEM
FIGURE 13
CAT-OX DEMONSTRATION TEST - MAJOR AREAS OF CONCERN
-------
system to fuels of various sulfur content. Also of interest in our
test program is the possible effect of Cat-Ox system failure on power
production, pressure drops across system components as a function of
time, and the ability of power plant personnel to operate and maintain
the Cat-Ox System.
Manual and Continuous Measurements
Both manual and continuous measurement instrumentation will be
used for evaluation of the Cat-Ox process performance. Figure 14
states the reasons for utilizing both methods. With regard to the
measurement of particulates, sulfur trioxide and sulfuric acid mist,
there is presently not available completely automatic continuous
measurement instrumentation, and, therefore, manual techniques must
be employed.
Continuous recording of data has several advantages. One advantage
is the ability to obtain immediate indication of results instead of
having to wait for several days until sample solutions are analyzed
in the laboratory. Also real-time availability of data provides an
early check on the performance of the measuring instrumentation so
that necessary repairs can be made quickly, thereby minimizing the
loss of data. The installation of continuous measuring instrumentation
is cost effective provided the measurement program for which it will
be used is particularly long, as it will be for Cat-Ox (one year), so
that the initial cost of equipment purchase and installation is offset
by reduced manpower requirements during the test itself. Furthermore,
712
-------
MANUAL SAMPLING METHODS
- PROVIDE BASIS FOR COMPARISON
- SOLE METHOD FOR DETERMINING CERTAIN POLLUTANTS
- METHOD FOR CALIBRATION
CONTINUOUS MONITORING SYSTEM
- CONTINUOUS RECORD OF EMISSIONS
- AUTOMATIC SAMPLING FROM TWELVE LOCATIONS IN PROCESS
- AUTOMATIC CALIBRATION ON PERIODIC BASIS
- IMMEDIATE DATA OUTPUT ON GAUGES, STRIP CHART, AND TELETYPE
- DATA STORAGE ON MAGNETIC TAPE FOR COMPUTER PROCESSING
FIGURE 14
CAT-OX DEMONSTRATION TEST - MEASUREMENT METHODS
-------
if one Is evaluating a complex process as is the case with regard to
Cat-Ox where simultaneous measurements are desirable from several
locations, continuous measurements provide a particular advantage
because of the manpower which would be required to man all the sites
if the data were to be obtained manually. Continuous measurement
instrumentation also permits test conditions to be changed more quickly,
and permits observation of process performance during transitions from
one test condition to another. Finally, once it has been decided to
utilize continuous recording instrumentation, there is a tremendous
advantage to storing the information on magnetic tape so that it can
be readily processed by computer. The process of transferring data
from strip charts to punched cards is expensive in time and manpower
when large quantities of data are involved.
Cat-Ox Process and Measurement Points
The continuous measurement instrumentation represents a major
part of the MITRE program so that the discussion which follows will
be devoted primarily to that area of the MITRE work. The process
itself is discussed by W. Miller in a paper entitled "The Cat-Ox
Project at Illinois Power" included in these same proceedings and,
therefore, will be reviewed in this paper to the extent required to
explain the interfaces to MITRE Instrumentation.
Figure 15 shows the flow diagram of the steam generator and
Cat-Ox process. In this Figure, the dashed line separates the
steam generator from the process. The Cat-Ox process, which is of
714
-------
•FLUE GAS
U1
ELECTROSTATIC
PRECIPITATOR
GAS HEAT
EXCHANGER
REHEAT
BURNER
SYSTEM "B"
(IN DUCT)
J
BY-PASS TO STACK
FIGURE 15
STEAM GENERATOR & CAT-OX PROCESS
-------
the reheat type, has been installed between the existing I.D. fans
and stack. The electrostatic precipitator is regarded as part of the
process because it was installed with the process, and is required by
the process to minimize contamination of the catalyst in the converter.
Flue gas from the boiler passes through the economizer and air
heater, and then enters the mechanical collectors where particulates
are initially removed. Particulates are further reduced to a very low
level by the electrostatic precipitator. During the course of this
flow, the gas temperature has dropped approximately 400°F and therefore
must be reheated prior to entering the converter. The reheating is
accomplished by burners "A" and "B" and the heat exchanger. The SO.
in the flue gas is then converted to SO., and combined with water vapor
in the absorbing tower to formulate sulfuric acid. The mist eliminate
removes sulfuric acid mist which may escape from the absorbing tower
with the flue gas. Because of the pressure drop throughout the process,
a second I.D. fan is required to'make up the losses and restore flow
to the stack. Acid from the absorbing tower and mist eliminator are
pumped, cooled, and finally stored in a large tank from which it can
be periodically removed.
The numbers shown in the flow diagram identify the locations at
which measurements will be made. Points 1', 2f, and 14 are locations
where measurements were made during the baseline measurement program.
Measurements will be repeated at these three locations to permit
716
-------
correlation with the baseline results and to reevaluate improvements
which have been made in the operational efficiency of the mechanical
collector. The other measurement points identified are pertinent to
the evaluation of the overall Cat-Ox process and major subsystems
of the process.
Purpose of Measurements
Figure 16 shows the measurement points which have been identified,
the parameters which will be measured at each point, and the purposes
for making particular measurements. The significant parameters which
will be measured by continuous recording instrumentation are gaseous
concentrations of SO-, HJ) vapor, 02, C02, NO , N02, and THC. Also
differential pressure, static pressure, and temperature will be
recorded continuously to obtain gaseous volume flow, system pressure
drops, and gas temperatures. In addition, manual measurements will
be made of gaseous S0~, H^SO, mist, and particulates.
These measurements will be utilized to determine subsystem per-
formance of the power plant and Cat-Ox process as follows:
• Efficiency of the mechanical collector
• Efficiency of the electrostatic precipitator
• Pressure drops across process subsystems and overall system
• Efficiency of the heat exchanger
• Percent leakage of heat exchanger
• Efficiency of converter
• Gas distribution in converter
717
-------
POINT LOCATION
1* INPUT ECONOMIZER
PARAMETERS
, 0, C0, AP, P, T
PURPOSE
CONFIRMATION BASELINE
TEST; NON-REHEAT PROCESS
CONDITIONS
AIR HEATER
MASS LOADING
EFF. MECHANICAL COLLECTOR
00
INPUT ESP
OUTPUT ESP
MASS LOADING ,AP, P, T
MASS LOADING ,AP, P, T
EFFICIENCY ESP; PRESSURE
DROP ESP/CAT-OX
EFFICIENCY ESP;PRESSURE
DROP ESP
MANUAL MEASUREMENT
FIGURE 16
MEASUREMENT POINTS ONE-YEAR DEMONSTRATION TEST
-------
POINT
4
LOCATION
PARAMETERS
PURPOSE
INPUT HEAT EXCHANGER S02, H20, 0,,, CO ,AP, P, T EFFICIENCY/PERCENT LEAKAGE
HEAT EXCHANGER
OUTPUT GAS HEAT
EXCHANGER (INPUT
CONVERTER)
S02, S03 „ H20, 02, C02,AP, EFFICIENCY/PERCENT LEAKAGE
P T HEAT EXCHANGER; EFFICIENCY
' CONVERTER
to
8
CONVERTER
OUTPUT CONVERTER
so3 , so2,
, co
S02,
3 , HO,
MASS LOADING , AP, P, T
, C02,
GAS DISTRIBUTION
EFFICIENCY CONVERTER VS.
FLOW RATE, T, AND TIME;
ASH ACCUMULATION; EFFICIENCY/
PERCENT LEAKAGE HEAT EXCHANGER
MANUAL MEASUREMENT
FIGURE 16 (CONTINUED)
MEASUREMENT POINTS ONE-YEAR DEMONSTRATION TEST
-------
POINT
LOCATION
10 INPUT ABSORBING TOWER
PARAMETERS
AP, P, T
PURPOSE
EFFICIENCY/PERCENT LEAKAGE
HEAT EXCHANGER; HjSO, IN
ABSORBING TOWER
11 OUTPUT MIST ELIMINATOR H9S°4 MIST > Ap» p» T
PERFORMANCE MIST ELIMINATOR
KJ
o
13 INPUT STACK
1A MIDWAY IN STACK
P, T
S02,
S04 MIST , C02, 02,
N02, NO . THC, H20, MASS
LOADING , AP, P, T
PRESSURE DROP ACROSS CAT-OX
EMISSIONS
MANUAL MEASUREMENT
FIGURE 16 (CONCLUDED)
MEASUREMENT POINTS ONE-YEAR DEMONSTRATION TEST
-------
• Ash accumulation in converter
• Formation of H.SO, in absorbing tower
• Performance of mist eliminator
• Emissions from the stack
The measurements performed at point 1' will be used to confirm the
baseline test conditions and to obtain data at a location in the steam
generator which is typical of that to which an integral type Cat-Ox
process would be connected.
Overall Continuous Measurement System
Figure 17 shows the overall continuous measurement system. The
system consists of four subsystems: the time-shared gas measurement
subsystem, the flow measurement subsystem, the continuous gas
measurement subsystem, and the data recording/control subsystem. The
measurement points identified previously are shown inputing to each
of the major subsystems. There are additional locations which have
not been identified in this now diagram where temperature, pressure,
and humidity are measured continuously as individual parameters.
Some of these locations are discussed in subsequent text.
Gas concentrations are measured on a time-shared basis at seven
different locations by the time-shared gas measurement subsystem.
One of these locations is the stack (point 14) where a second gas
measurement subsystem is dedicated to measuring gas concentrations
at this one point alone. There is a common time slot when both
the time-shared subsystem and the dedicated subsystem both measure
gas concentrations in the stack simultaneously so that data from
both subsystems can be correlated.
721
-------
KJ
TO
TIME-SHARED GAS
MEASUREMENT
SUBSYSTEM
1
F
S
FLOW MEASUREMENT
SUBSYSTEM
DATA RECORDING &
CONTROL SUBSYSTEM
*
CONTINUOUS GAS
MEASUREMENT SUBSYSTEM
FIGURE 17
OVERALL INSTRUMENTATION SYSTEM
-------
With the exception of point 6 which permits access within the
converter, flow measurements are made at every point where gas con-
centration measurements are made so that mass flow rates can be computed.
In addition, there are several points where continuous flow measurements
are made without corresponding continuous gas measurements. At these
points flow measurements are combined with manually measured gas con-
centrations or are used of themselves for subsystem evaluation.
Data from the gas and flow measurement subsystems are automatically
recorded on strip charts, by printer, and on magnetic tape. Strip
chart and printer data are utilized in real-time to assure that the
instrumentation is functioning properly. Data recorded on magnetic
tape is used for subsequent computer processing eliminating the tedious
task of transferring data from strip charts to punched cards. In
addition, the system is designed to perform certain automatic control
functions such as switching of the time-shared gas subsystem, gas
analyzer calibration, and blowback of sampling lines.
Time-Shared Gas Measurement Subsystem
Figure 18 shows the flow diagram of the time-shared gas measurement
subsystem. Flue gas is drawn into the analyzers through a filtered
probe which may be either an in-the-duct filter or an external heated
filter. The gas from the probe is passed through a heated water trap
and then through a heated teflon gas line (Dekoron line). The gas
lines are heated to prevent condensation of water vapor and hydrocarbons.
723
-------
KJ
CALIBRATE £
HEATED
SAMPLE
HANDLING
I—fa?
: T W
. )
H20
VAPOR
FIGURE 18
TIME-SHARED GAS MEASUREMENT SUBSYSTEM
-------
Gas is sampled from seven different locations by a multi-point
sequential sampler. The gas is aspirated through the seven lines
continuously except during blowback. Each line is then selected
sequentially by switching pneumatic valves drawing a fraction of the
gas into the analyzers. The gas lines from the sequential sampler
to the combined S09-N0_/N0 analyzer and the THC and water vapor analyzers
*• £ X
are also heated for the same reasons given above. The flue gas to the
C02 and 02 analyzers is first passed through a refrigerator-condenser
in order to remove the water vapor, thereby preventing water vapor
interference in the CCL analyzer and corrosion in the 02 analyzer.
Operation of the time-shared gas measurement subsystem is shown
in Figure 19. The subsystem is sequenced automatically on a one hour
time base by the control subsystem. Flue gas is drawn into the ana-
lyzers from one particular line for a period approximately 7 minutes,
and then that same line is blown back by high pressure air for approxi-
mately 1 minute to remove particulates from the ceramic filter. Sub-
sequently, each of the remaining lines are sampled in succession in
the same manner until the sequence is completed. Then all of the
analyzers are automatically zeroed against nitrogen (except for the
SO--NO./NO analyzer), and then spanned against a calibration gas.
w fc X
The SCL-N00/NO analyzer provides the blowback air and is also designed
£, £» 3C
to zero on the blowback air which is passed through the sample cells
of the analyzer.
725
-------
TIME PERIOD
POINT
1'
4
5
6
8
10
14
ANALYZERS
OPERATION
SAMPLE
BLOWBACK
SAMPLE
BLOWBACK
SAMPLE
BLOWBACK
SAMPLE
BLOWBACK
SAMPLE
BLOWBACK
SAMPLE
BLOWBACK
SAMPLE
BLOWBACK
ZERO
SPAN
(MINS . )
7
1
7
1
7
1
7
1
7
1
7
1
7
1
2
2
60
FIGURE 19
OPERATION OF TIME-SHARED SUBSYSTEM
726
-------
Continuous Gas Measurement Subsystem
Figure 20 shows the flow diagram of the continuous gas measurement
subsystem which samples .gas continuously from the stack. The principle
gas sampled is SO. so that a continuous record of SO. emission from
the stack is obtained. For purposes of illustration, an external
heated filter is shown as opposed to an in-the-duct filter which was
shown in the flow diagram of the time-shared subsystem. In addition,
a multi-point gas probe is shown which will be used at several of
sampling locations where gas concentration is expected to be non-uniform.
The advantage of the external filter is that the filter element can be
removed for cleaning without removal of the probe itself. The continuous
gas measurement subsystem is synchronized with the time-shared gas
measurement subsystem so that both subsystems overlap in time while
sampling gas from the stack permitting correlation of data from both
systems.
Flow Measurement Subsystem
The flow measurement subsystem is shown in Figure 21. Gas flow
is determined by measuring differential pressure, static pressure, and
temperature which are combined in an analytical relationship to calcu-
late flow. As the crossection of ducting is very large throughout the
steam generator and Cat-Ox process, it is necessary to measure these
parameters at a number of points within any particular duct in order
to obtain a representative measurement. Therefore, the crossection of
the duct is divided into a number of sampling points based on the ASMA
727
-------
ro
CO
MULTI-POINT GAS PROBE
FLUE GAS
FIGURE 20
CONTINUOUS GAS MEASUREMENT SUBSYSTEM
-------
tsj
IO
TEMP/PRES RAKE
IU iu iu iu
FLUE GAS
PRESSURE
L
PRESSURE
C= TRANSMITTER
^•^n
••—•to.
••*••
L^
— ^
— »•
— *.
»
»
— >•
>
— *•
— K
»
»
»
>
— *.
»
»
»
>
— »•
TEMPERATURE
TRANSMITTER
T
E
M
P
E
R
A
T
U
R
E
S
C
A
N
N
E
R
TEM
AMP
r
PERj
PEN
LIFI
, VOLUME
FLOW
CONVERTER
BAROMETRIC
PRESSURE
TRANSDUCER
ftTURE
ER
fc CHART
^ RECORDER
"|
. A-D CONVERTER
FIGURE 21
ROW MEASUREMENT SUBSYSTEM
-------
power test codes, and an array of combined temperature/pressure rakes
are used to sense differential pressure, static pressure and temperature
at the sampling points selected.
The magnitudes of the differential pressure and static pressure
are measured by pressure transmitters and are normally recorded
directly on strip charts and magnetic tape. As temperature is measured
at many more locations in the system than is required for flow measure-
ments, the temperature sensor, which is an iron-constantan thermocouple,
is inputed through a constant temperature enclosure to a scanner. The
scanner acts as a switch to connect thermocouples from various locations
to a temperature-compensated amplifier which amplifies, linearizes, and
temperature compensates the signal prior to entry into the A-D converter.
The constant temperature enclosure, which is not shown in the flow
diagram, maintains the connector junctions at constant temperature so
that emfs which are generated cancel out.
In addition to recording the three parameters directly, a specialized
analog computer identified as the volume flow converter is used to calcu-
late the volume flow in real-time. The volume flow converter can be
utilized at the output of any of the flow measurement instrumentation.
Gas volume flow is of particular interest because it is one of the
design parameters of the Cat-Ox process.
There are nine flow measurement locations which are nearly identical.
However, at two of the locations, the economizer and stack, where dif-
ferential pressures are particularly low, it has been necessary to use
an electronic manometer in place of the differential pressure transmitter.
730
-------
Data Recording and Control Subsystem
The outputs of the analyzers, transmitters, and other sensors
are recorded on magnetic tape as shown in Figure 22. The data
acquisition system proper has a capacity of fifty channels, which has
been expanded, as discussed previously, with an additional twenty
channels by means of a low noise temperature scanner. Eight channels
are assigned to gas concentration measurements, 10 to static pressure,
9 to differential pressure, 1 to gas volume flow, 3 to channel identi-
fication and ambient measurements, and 14 to additional temperature
measurements. The last 14 channels are tentatively assigned depending
on time available to integrate the necessary electronics.
The scanner connects the analog signal from each channel in
sequence to the A-D converter which digitizes the analog signal. The
data from the A-D converter is transmitted to the coupler which formats
it for recording on magnetic tape. A printer is utilized for visual
recording of selected data. In addition, a teletypewriter (TTY) is
employed as an input/output device. As an input device, the TTY
writes directly onto the magnetic tape via the keyboard.
The scanner will normally be operated at 1 scan/minute but is
capable of being operated at better than 1 scan/5 seconds. The
digital clock will generate time in days, hours, and minutes and will
provide reference signals to the function controller. The function
controller will Initiate start—stop commands to perform remote control
functions such as sequencing of the time-shared gas measurement sub-
system, calibration of the analyzers, and blowback of the probes.
731
-------
GAS (8}
SJ
STATIC PRESSURE (10)
DIFF. PRESSURE (9)
VOLUME FLOW
& SPARE <6)
I.D.&
AMBIENT (3)
TEMPERATURE (14)
SCANNER
INPUT FROM
TEMPERATURE
SCANNER
CONTROL
SIGNALS
SYNC TO
TEMPERATURE
SCANNER
FIGURE 22
DATA RECORDING AND CONTROL SUBSYSTEM
-------
Status of the Instrumentation
Figures 23» 24, 25, and 26 list the equipment previously discussed
for each of the major subsystems and identify the manufacturer and
model number. If the equipment is specialized to MITRE specifications,
model numbers have not been provided by the manufacturer.
Installation of the continuous measurement instrumentation is
near completion and will be on line prior to. start of the one-year
demonstration test program. Some of the hardware is shown in the
following photographs. Figure 27 shows two types of gas probes. The
probe on the left uses a heated water trap with an external heated
filter, whereas the probe on the right uses an in-the-duct filter
with an external heated water trap. Figure 28 shows the interior
of the multi-point sequential gas sampler. The gas from each of the
7 Dekoron lines is passed through teflon water traps shown in the
center of the photograph and then from the top of the traps through
pneumatic valves to the analyzer. The center cabinet is heated to
prevent condensation of water vapor. Figure 29 shows the interior
of the S92-H02/NOx analyzer. The center cabinet is heated and
contains the N02/N0x gas cell. The S02 cell is attached to the side
of the cabinet and is not in view. Figure 30 shows the refrigerator-
condenser in the foreground with parallel gas pumps. The rack in the
rear contains the 02 and C02 analyzers with sample handling and
flow balancing instrumentation.
733
-------
u>
FUNCTION
SEQUENTIAL SAMPLER
S02 - N02/N0x ANALYZER
THC ANALYZER
GENERATOR
H20 VAPOR ANALYZER
HEATED SAMPLE HANDLING
REFRIGERATOR-CONDENSER
SAMPLE HANDLING
ANALYZER
C02 ANALYZER
EQUIPMENT
DUPONT
DUPONT 461C
BECKMAN 400
MILTON ROY ELHYGEN R 8320
MSA LIRA M202 (MODIFIED)
MSA
BENDIX
BENDIX
BECKMAN F-3
BENDIX UNOR-6
FIGURE 23
TIME-SHARED GAS MEASUREMENT SUBSYSTEM
-------
FUNCTION EQUIPMENT
S02 ANALYZER DUPONT 460
REFRIGERATOR-CONDENSER BENDIX
*j
LO
01 SAMPLE HANDLING BENDIX
ANALYZER BECKMAN F3
FIGURE 24
CONTINUOUS GAS MEASUREMENT SUBSYSTEM
-------
FUNCTION
TEMP./PRES. RAKES
DIFFERENTIAL PRESSURE TRANSMITTERS
STATIC PRESSURE TRANSMITTERS
TEMPERATURE TRANSMITTER
VOLUME FLOW CONVERTER
TEMPERATURE SCANNER
TEMPERATURE COMPENSATED AMPLIFIER
BAROMETRIC PRESSURE TRANSDUCER
ELECTRONIC MANOMETER
EQUIPMENT
UNITED SENSOR & CONTROL
LEEDS & NORTHRUP 1912
LEEDS & NORTHRUP 1912/1970
LEEDS & NORTHRUP 1992
LEEDS & NORTHRUP
MONITOR LABS 1100
IRCON DATA SYSTEMS 3J16F
ROSEMOUNT ENGINEERING 1331
CGS DATAMETRICS 1023
FIGURE 25
FLOW MEASUREMENT SUBSYSTEM
-------
-•J
CO
FUNCTION
DATA ACQUISITION SYSTEM
FUNCTION CONTROLLER
STRIP CHART RECORDERS
EQUIPMENT
DATA GRAPHICS CAT-12
DATA GRAPHICS DGC-100
MFE M26 CAHA
L & N SPEEDOMAX M MARK II
FIGURE 26
DATA RECORDING AND CONTROL SUBSYSTEM
-------
A
..
I
I
-
• .
FIGURE 27
GAS PROBES WITH EXTERNAL AND
IN-THE-DUCT FILTERS
738
-------
F1:::
[-POINT ! .., ' [P
739
-------
URE 29
SO -NO /NO ANALYZI-K
£- *— A
7/10
-------
FIGURE 30
TOR-CONDENSER, O? ANALYZER, AND C02 ANALYZER
741
-------
Figure 31 shows a photograph of the temperature/pressure rake.
The support for the temperature/pressure sensors is aerodynamically
shaped to minimize turbulence effects at the pressure sensor. This
particular rake shows two sets of sensors. The longer of the two
sensors is the pressure sensor and is a conventional pitot-static
probe. The temperature sensor is an iron-constantan thermocouple
encased in stainless steel protective tubing. Figure 32 shows the rakes
mounted in six inch ports with manifolding to the pressure transmitters.
Figure 33 shows the differential pressure transmitter on the left and
the static pressure transmitter on the right with the zero/blocking
valve and drip pots below.
Figure 34 shows the data acquisition subsystem and function
controller. The TTY is on the extreme left. The data acquisition
subsystem is in the rack next to the TTY. The lowest panel of this
rack is the temperature compensated amplifier. The next rack is the
function controller and the last rack shows strip chart recorders
and the volume flow converter mounted in the top panel.
742
-------
FIGURE 31
TEMPERATURE/PRESSURE RAKE
-------
-
FIGURE 32
6 INCH PORTS SHOWING MANIFOLDING OF RAKES
-------
FIGURE 33
') I PFF.RENTIAL AND STATIC PRESSURE TRANSMITTERS
745
-------
.
' -•-!$.-*
* m-**!*-**-^* -
iL -
DATA ACQUISITION SI'BSY -
-------
DISPOSAL AND USE OF BYPRODUCTS
FROM FLUE GAS DESULFURIZATION PROCESSES
INTRODUCTION AND OVERVIEW
by
A. V. Slack and J. M. Potts
Office of Agricultural and Chemical Development
Tennessee Valley Authority
Muscle Shoals, Alabama
747
-------
DISPOSAL AND USE OF BYPRODUCTS FROM FLUE CAS DESULFURIZATION PROCESSES
INTRODUCTION AND OVERVIEW
By
A. V. Slack and J. M. Potts
Office of Agricultural and Chemical Development
Tennessee Valley Authority
Muscle Shoals, Alabama
The removal of sulfur oxides from flue gases poses several very
difficult problems to the owner of the emitting plant--whether the operation
involved is power production, smelting, sulfuric acid production or a Glaus
plant. The overriding consideration in many situations may be the diffi-
culty in disposing of the sulfur once it has been gathered from the gas
and concentrated in some solid or liquid material. Whatever the product--
a waste solid or a salable material such as sulfuric acid, elemental sulfur,
or a fertilizer—there must be the assurance that it can be moved away
from the plant as fast as it is produced, except to the extent that surge
storage may allow fluctuations. Since it is a byproduct rather than the
primary one, production cannot be planned in accordance with market demand.
Instead, the byproduct flow must be disposed of in the best way possible,
even though it may vary widely with production of the main product.
For a new plant the problem is not as troublesome because byproduct
disposal is part of the overall planning involved in determining project
feasibility. For an existing plant, however, finding a way to dispose of
a byproduct that was never considered in the original project planning can
be quite difficult or perhaps in some cases impossible.
One of the difficulties, both from the country-wide viewpoint
and that of the local plant, is the very large tonnage involved. Various
estimates have been published on the amount of sulfur oxide emission in
the United States; a reasonable order-of-magnitude figure for 1980 appears
to be 50 million tons of sulfur dioxide from sources amenable to stack gas
cleaning--and if it is assumed that emission from the other sources will
be reduced by fuel cleaning, about 10 million tons would be added to the
total recoverable sulfur dioxide. Since complete removal is unlikely, a
factor must be applied; assuming 80% removal, the total is 48 million tons
as S02--or 24 million tons of sulfur.
Obviously this amount of sulfur, or anything approaching it,
would produce an almost overwhelming disposal problem no matter what the
product. If converted to a waste solid by CaO-CaC03 scrubbing the resulting
tonnage (wet basis) would be on the order of JQQ million tons per year or,
more importantly, the sludge accumulation over the next 20 years would cover
an area of almost 5000 square miles a foot deep. Or if converted to salable
materials, the amount of sulfur involved would be about 1.7 times the expected
1980 consumption in the United States.
748
-------
The local situation is even more difficult. Assuming a modern
large plant such as TVA's Paradise station (2550 raw; about 4.2$ S in coal),
sludge production could be as high as 8500 tons per day and the disposal
volume about 1600 acre-feet per year (at 70$ capacity factor). If the
sulfur were recovered in useful form the annual tonnage (as H2S04) would
be on the order of 600,000 tons, which would generate such a sales problem
that for the particular location involved it might be necessary to ship
the acid so far that shipping cost would be higher than sales revenue.
Thus the problem of byproduct disposal looms as a major obstacle
in the future of sulfur oxide emission control.' In this symposium an effort
will be made to explore the problem from all angles and in depth. One paper
will describe in detail the situation regarding disposal or use of the waste
solids from CaO-CaC03 scrubbing and others are concerned with the future
market for sulfur products, including new uses that are under development.
Finally, one paper will describe the overall situation in Japan, where both
throwaway and recovery methods have been carried farther than perhaps any-
where else in the world.
The present paper introduces the problem and presents an overview
of the current status of the technology, particularly from the viewpoint
of a utility such as TVA that is faced with the problem of process selection.
Most of the information presented has been abstracted from reports on research
work carried out by the TVA Office of Agricultural and Chemical Development
for EPA and/or the TVA Office of Power. The work of other organizations is
also summarized to the extent that it has been made available to the public.
Most of the producers who operate sulfur dioxide-emitting plants,
particularly the utilities, favor discarding a waste solid over recovering
a product for sale. Ash disposal is an old technique to power producers
whereas sale of chemical products is not. For this reason, the majority of
the full-scale projects now under way are of the throwaway type. Full-scale
tests of recovery-type processes are also being carried out, however, mainly
in the United States and Japan. It is not yet clear which type will be the
most economical. The situation should be clarified somewhat in a major
process evaluation and cost study being funded by EPA; the study, to be
carried out by TVA, win compare those processes in the United States on
which full-scale design and operating cost is now available (or soon will
be). Three recovery and two throwaway processes will be included.
The ideal waste solid is one that is concentrated in sulfur content,
resistant to pile erosion, relatively dense (to reduce disposal volume), and
water insoluble. Only three materials meet these qualifications well enough
to have received serious consideration—calcium sulfite, calcium sulfate, and
elemental sulfur. Of these, sulfur has been considered only in a minor way
as a waste solid because it is expected that large quantities can be marketed.
There are areas, however, particularly in the Southwest, where the market
prospect is so bleak that the waste solid route is being considered. Moreover,
if any major portion of the sulfur oxide emitted were recovered as a useful
product the resulting market depression could well push elemental sulfur into
the waste solid category.
749
-------
In comparing discard as a waste solid with recovery as a useful
product, it can be said as a general observation that recovery does not
necessarily avoid the solid waste problem because a large part of the
sulfur consumed ends up eventually as a waste, usually in the solid form.
Over half of the U.S. sulfur consumption is in the fertilizer industry,
where the principal use is in making phosphoric acid. In this process the
sulfur is converted to calcium sulfate, a waste solid that is discarded
in settling ponds. Some sulfur, much less than in past years, is left in
fertilizers such as ordinary superphosphate and ammonium sulfate, and thus
ends up on the farmer's fields as a waste solid (since it is either already
in the form of calcium sulfate or may become so in calcareous soils) except
to the extent that it supplies nutrient calcium and sulfur.
Thus recovering the sulfur in a useful form does not eliminate
the waste solid problem but merely diffuses it. It is quite true, however,
that recovery reduces the overall quantity of waste solid since otherwise
both the fertilizer industry and the power industry would produce a waste.
Waste Sludge from Lime-Limestone Scrubbing
Most of the sulfur dioxide removal systems operating or under
construction in the United States are based on CaO-CaC03 scrubbing with
disposal of the product sludge as a waste. In planning these systems, dis-
posal of the sludge is a major consideration along with cost and reliability
of operation. There are still several unsolved problems in these areas but
a major effort is under way aimed at solving them. The utilities that have
pioneered in CaO-CaC03 scrubbing are quite active in this, as well as EPA
with its funding of research projects. It is expected that many of the
needed answers will be forthcoming from the EPA-TVA-Bechtel test program at
TVA' s Shawnee station.
Sludge Tonnage
The amount of sludge produced is governed mainly by the excess
lime or limestone used, the amount of fly ash collected from the gas in the
scrubber, and the degree of sludge dewatering.
Since lime is more reactive than limestone, less excess is normally
needed to achieve the same degree of sulfur dioxide removal. The actual
excess of each required for good operation, however, is not yet known with
any certainty. In the Mitsui Aluminum CaO system in Japan (designed by Chemico),
the stoichiometry (mole ratio of CaO to S02 in entering streams) apparently
has ranged from 0.95 to 1-05 with 80-85$ removal. In the TVA limestone slurry
pilot plant a stoichiometry of 1.5 has given about the same removal but there
is some indication (from the TVA pilot, the EPA pilot at Durham, and the
750
-------
EPA-TVA operation at Shawnee) that the system might be operated at 1.2, or
possibly lower, without much reduction in efficiency. So much depends on
type of scrubber, gas velocity, liquor circulation rate, and other factors
that no conclusions can be drawn at present regarding the minimum excess
of absorbent.
The weights of dry solids obtained under various conditions are
shown in Figure 1. The values given are approximate only since in actual
operation several factors can vary and change the tonnage to some degree.
The increase in weight of solids resulting from fly ash inclusion
can vary over a wide range. When large-scale application of CaO-CaC03
scrubbing first began in the United States a few years ago the emphasis
was on injecting limestone into the boiler, in which case all the fly ash
went with the lime. This approach has now been well nigh abandoned, however,
in favor of lime or limestone introduction directly into the scrubber circuit.
For such a method of operation, the situation differs between existing plants
and those started after the EPA emission regulations came into effect. Since
most existing plants have fairly efficient particulate removal systems already,
or will have before sulfur dioxide removal equipment is installed, the amount
of ash entering the scrubber would be relatively low. (There are exceptions
to this, e.g., TVA's limestone scrubbing installation at Widows Creek, where
the existing precipitator removes only about 50$ of the ash.) For new plants,
however, the regulations are such that in most cases scrubbers must be installed
initially. In practically all cases so far, dust removal is being combined
with sulfur dioxide absorption in either one scrubber or two in series, on
the basis that the incremental cost for dust removal is lower than for installing
an electrostatic precipitator ahead of the scrubber. There is some difference
of opinion as to whether or not this course actually is the most economical
but the consensus seems to favor it; much depends on the dust characteristics
and on the sulfur dioxide content of the inlet gas, each of which can favor
either, the precipitator or the scrubber. Another consideration is that some
utilities would like to bypass the scrubbers to maintain power production if
the scrubber system fails, and consider that this would be more acceptable
if the dust could be removed in an independent unit. Some producers also
want to retain the option of marketing fly ash and therefore must provide
for dry collection.
The type of boiler is also a factor, since the usual types—cyclone,
tangentially fired, and front-fired--introduce, in that order, increasing
proportions of the coal ash into the stack gas.
Thus the tonnage of ash accompanying the calcium-based solids can
vary all the way from an insignificant amount up to as much as 11 tons of
ash per ton of calcium solids (for coal containing 20$ ash and 0.6% s).
The amount of water remaining in the sludge has an important effect
on total tonnage, and is significant mainly when the sludge is to be dewatered
and transported to the disposal area as a solid rather than as a slurry. In
filtration tests at TVA, vacuum filtration gave a moisture content ranging
from about tfO to ^% for various pilot plant samples. Centrifuging at
1000 x gravity gave about Zjjo.
751
-------
200
150
o
_j
o
CO
a:
o
CD
100
50
STOICHIOMETRY
1.5
4
.3
.2
. I
.0
1
1
1
2
1
3
1
4
1
5
% S IN COAL
FIGURE I
EFFECT OF SULFUR CONTENT OF COAL ON
AMOUNT OF WASTE MATERIALS PRODUCED
J
752
-------
The degree of dewataring depends to a considerable extent on the
ratio of calcium aulfate to calcium sulfite, since the normally larger and
more blocky nature of the sulfate crystals makes them easier to dewater.
For example, in the Chiyoda process (water scrubbing to give sulfurous acid,
which is then oxidized to sulfuric acid and reacted with limestone to produce
calcium sulfate), centrifuging reduces the water content to 10-15$.
Thua the wet tonnage of dewatered sludge is likely to be l.J to
1.8 times the combined dry tonnage of fly ash and calcium solids.
For the TVA Widows Creek installation.. (550 mw, 4-5$ S in coal),
which may or may not be typical, the dry tonnage of waste solids is expected
to be about 2000 tons per day at full load and 500,000 tons per year at 70$
capacity factor. Assuming 40$ moisture in filtered solids, this would amount
to 5500 tons per day to transport away from the plant each day at full load
if a solids handling disposal system were used. Added to this would be the
incoming limestone, about 1300 tons per day, or a total solids handling re-
quirement of 4600 tons per day for a 550-mw boiler.
Sludge Volume
In waste solid disposal systems of the type normally operated by
utilities and mining operations, the solids are slurried and sluiced to a
settling pond from which the sluice water is either overflowed or recycled.
Since the pond must be abandoned when it becomes full of settled solids, a
key factor in planning is the ultimate pond volume required per ton of solids
(dry basis). This varies widely depending on the size, shape, density, and
gelling characteristics of the solid particles.
The utility industry is fortunate in regard to fly ash disposal
because most ash types settle compactly, requiring only about 20 cubic feet
of pond volume per ton of ash (dry basis). Probably one of the worst
situations is in the phosphate mining industry, where the clay-laden gangue
forms a gel structure in waste ponds and may occupy as much as 125 cubic
feet per ton. The situation most analogous to lime-limestone scrubbing is
in the phosphoric acid industry, where byproduct calcium sulfate (gypsum)
is ponded in large quantities; the pond volume required is about 28 cubic
feet per ton.
Unfortunately, the pond volume requirement for CaO:CaC03 scrubber
sludge will likely be relatively high since the calcium sulfite tends to
crystallize in small, thin platelets that settle to a loose bulky structure
and occlude a relatively large amount of water because of their gelling
tendency. The crystal form is shown in Figure 2, which is from an electron
microscope study by McClellan and Mills (TVA).
In other studies U), Davenport (TVA) noted three phases in the
settling of the slurry—(l) an induction period (slow settling while floes
were forming), (2) free settling, and (3) compression settling. The average
753
-------
Figure 2. Single and stacked
calcium sulfite crystals shoving
the thickness of some of the
crystals.
Figure 5. A rosette aggregate
of calcium sulfite crystals
formed by interpenetration
during crystal growth.
Figure 4. Gypsum from Chiyoda
Chemical Engineering & Construction
Co., Ltd. (japan)
Figure 5. A fly ash
particle with calcium sul-
fite on and next to it.
-------
settling rate over the first two phases was about 5 centimeters per hour.
Phase 3 was reached when the floes began to touch each other, at which
point a continuous gel was formed and settling proceeded at a much slower
and rapidly decreasing rate. There was very little settling after about
48 hours, even over a period of several months.
Measurements with a torsion wire gelometer showed that a rela-
tively strong gel was gradually formed. A l6$> slurry had no gel strength
immediately after stirring but developed a strength of 6 g-cm after 50
minutes and >25 g-cm after 18 hours.
Reported values for degree of compaction have varied over a fairly
wide range. Many of the data on settling are reported on the basis of water
content in the settled sludge, whereas the value needed is weight of the
solid phase per unit volume of fully settled sludge; water content is a
factor in this but is not necessarily proportional. From calculations and
estimates based on the various reports, the packing volume appears to range
between 45 and 75 cubic feet per ton of dry solids. This covers a range of
roughly 30 to 60$ solids in the settled sludge.
The factors affecting degree of compaction have not been identified
with any great degree of certainty and very little quantitative information
is available. The following considerations are perhaps significant.
1. Hydraulic head. The TVA work mentioned earlier indicated that
increasing the height of the slurry column—from 13 centimeters
to 100 centimeters--increased compaction by about 15$* Thus a
20-foot head; say, in an actual pond might be quite helpful.
2. Ash content. Since the ash particles are relatively large in
comparison with the sulfite, sludges containing a large pro-
portion of ash might settle more compactly. Not much data
appear to be available on the point, however, and the high gel
strength would indicate that a large proportion of ash would be
needed for help in compaction.
3- Degree of oxidation. Since sulfate crystals are normally larger,
levels of scrubber operating parameters that increase oxidation
might be advantageous. However, both EPA (2) and TVA (%) data
indicate that a high degree of oxidation is necessary if compaction
is to be increased very much. Presumably the gel structure per-
sists unless the content of sulfite crystals is reduced to a very
low level, and thus the situation is analogous to the effect of
fly ash content as described in 2.
4. Agglomeration* Agglomeration of sulfite crystals, which presumably
should help compaction, has been noted in some tests; a typical
agglomerate is shown in Figure 3 /from the work of McClellan and
Mills (If)/. However, the scrubber conditions that promote agglom-
eration have not been identified.
755
-------
5. Lime versus limestone. There has been some indication that
lime scrubbing may produce a more compact settled sludge than
does limestone scrubbing. The Chemico-Mitsui Aluminum unit
in Japan, for example, apparently produces a settled sludge
containing 50 to 60$ solids whereas the range for limestone
scrubbing has been more on the order of 30 to 50$. There are
indications that limestone type and grind may affect settling
and filtration characteristics of slurry.
6. Double alkali operation. In the "double alkali" type of
operation—scrubbing with a clear water solution of an alkali
followed by reaction with lime or limestone to precipitate
sulfite and sulfate—the calcium sulfite and sulfate are
crystallized under somewhat different conditions as compared
with slurry scrubbing and thus might have different settling
properties. Exploratory studies indicate that they may be
somewhat improved.
In the Chiyoda process oxidation of the sulfurous acid to sul-
furic before reaction with limestone produces large crystals
of gypsum as the end product (Fig. 4)* In Japan this is used
in construction products but if the process were used in the
United States the sulfate would probably be discarded, in which
case the packing volume should be relatively low.
From the planning standpoint, therefore, a relatively large pond
volume must be provided for CaO-CaC03 scrubber solids unless the situation
can be improved or unless some process is used (Chiyoda, for example) that
gives low packing volume. For Widows Creek, the pond being provided contains
about 100 acres and it is expected that the ultimate depth of the stored
sludge.will be about 37 feet. The initial storage capacity of the pond is
about 4.5 MM cubic yards, with the dikes designed so that they can be increased
in height by 10 feet to give an increase in capacity to 5.8 MM cubic yards.
The estimated total scrubber effluent ponding requirements over the remaining
life of the plant (about 25 years) is 9-3 MM cubic yards based on projected
load factors and current knowledge of settling characteristics.
Pond volume requirement is an important factor in overall emission
control economics because of the high pond cost. The cost per ton of solids
varies widely, of course, with pond size, topography, depth of fill, pond
type, and many other factors. The Widows Creek pond, which will accommodate
fly ash and sludge in separate sections, is expected to cost on the order of
$0.50-0.75 per cubic yard of capacity.
Methods for Increasing Compaction
The discussion thus far has been concerned with the degree of com-
paction under normal operation of the CaO-CaC03 scrubbing system. There
are various special measures that might be taken to get greater compaction.
756
-------
Slow Stirring; Since the calcium sulfite crystals form a gel,
breaking up the gel occasionally by slow stirring allows further settling
before the gel reforms. In TVA tests (l) intermittent stirring for short
periods was fairly effective. A slurry that had settled to 3!$ solids was
made to settle finally to 48$ solids by intermittent stirring (about 1
minute stirring followed by 8 hours or so of settling). Continuous stirring
was not effective.
These results indicate that pond capacity might be increased by
occasionally removing the supernatant layer completely and stirring the
sludge by some method, perhaps by air injection (which might help oxidize
sulfite to sulfate) or by towing across the pond a raft equipped with sus-
pended stirring elements. Thickening from 50 to 50$ solids, for example,
would increase storage capacity by about
Flocculation; Addition of agents to promote flocculation should
improve settling rate but not necessarily the degree of compaction; in
earlier TVA work on phosphate slimes (ore washing tailings) flocculation was
not effective. In tests of various flocculants added to limestone scrubber
slurry (jj), settling rate was increased to a major degree but final compaction
was not improved. It should be noted that the normal, unassisted settling
rate is adequate for providing clear supernatant liquor (for recycle) in a
large pond.
Dewatering: if the sludge is dewatered by filtering or centri-
fuging and transported to the disposal point as a solid, less disposal volume
will be required than for settling in a pond. TVA filtration tests have
given a packing volume of 39 cubic feet per ton (59$ solids by filtration
versus 38$ by settling). EPA filtration tests on sludge made from another
type of limestone gave a solids content of about 70$. Other exploratory
tests (TVA) have shown that the "dry" solids do not expand to larger volume
when submerged in water or exposed to rainfall.
Disposal as a solid rather than a slurry will be discussed further
in a later section.
Increase in Crystal Size; It may be possible to increase the size
of the sulfite crystals, or to promote agglomeration, by some technique
applied either in the scrubber or in a special vessel through which the
slurry would flow on the way to the pond. Apparently no work has been done
in this area. It would be necessary, presumably, to increase the size to
the extent that gel formation would not occur.
Oxidation; If the sulfite could be oxidized to sulfate at
reasonable cost this might be the best answer to the problem. Such oxi-
dation is an established practice in Japan (see paper in this symposium
by Jumpei Ando) where the objective is production of gypsum for sale. The
method involves blowing air through the slurry after it leaves the scrubber;
the main problem is the slow rate of oxidation, which has dictated in many
plants the use of special (and expensive) equipment to increase the rate of
oxygen absorption.
/ J I
-------
Extensive work has been done at TVA on oxidizing limestone
scrubber slurry (jj, 6). The results can be summarized as follows:
1. Air introduction into the scrubber. In one pilot plant
test., air was drawn into the scrubber along with the stack
gas. Oxidation was increased to 90$ and both settling
rate and degree of compaction were greatly improved. Similar
tests made since did not increase oxidation as much, and
compaction was not improved.
2. Oxidation in separate unit. A small grid-packed scrubber
(6-inch diameter) in which air was blown through product
slurry (02:S03= mole ratio of 10) was built at the pilot
plant. Oxidation was quite slow and the tests were generally
unsuccessful.
3. Spinning cup oxidizer. A laboratory-scale "spinning cup"
oxidizer such as used in some of the Japanese plants was con-
structed and tested in the laboratory. Approximately 80$> of
the sulfite could be oxidized in about 1 hour if the pH were
first reduced to about 5 by adding an acid. (The reduced pH
is apparently necessary in order to get an adequate supply
of sulfite in solution.) Nearly, complete oxidation was
necessary to achieve any significant improvement in com-
paction, presumably because this was required to avoid gel
formation by residual sulfite. The preferred acid was sul-
furous (which perhaps could be obtained by passing part of
the stack gas through the product slurry and then into the
main scrubber).
4. Use of catalysts. The oxidation could be speeded up, at a
given pH, by adding a small amount of catalyst to the slurry.
Manganese and iron compounds such as the sulfate and carbonate
were the most effective. At a pH of ^.0 and with 0.1% of
manganese oxide added, 100$ oxidation of a 1% calcium sulfite
slurry was attained in 30 minutes.
In general, the work on oxidation to increase compaction has been
discouraging. An analysis should be made to compare the cost of pond capacity
increase by oxidation as compared with building a larger pond or dewatering.
Pond Management
The way a pond is managed over its lifetime has an important bearing
on costs and must be planned in advance. Several questions should be con-
sidered.
Will it be operated as a single unit or divided into sections?
•
Will the original depth be the limit or can the walls be built
up by using the settled material?
758
-------
Shall the pond be partially filled with water before operation
begins?
Can the pond be filled to the top of the dike or must some
freeboard be allowed for periods in which rainfall exceeds
evaporation?
The usual method of ash pond management is to operate the pond
(built by excavating and throwing up a low dike wall of earth) as a single
unit with the ash slurry entering at one end, flowing as a stream over
settled ash to a low point at the other end where a pool forms, and over-
flowing supernatant water over a weir or through a standpipe. This is a
relatively simple method, feasible only because the ash settles rapidly
and compactly. Ash is seldom used to build up the pond wall because the
spherical form (Fig. 5) gives it a low angle of repose.
In the phosphate industry, waste gypsum is ponded somewhat
differently. The pond is often divided into two or more sections by dikes
and operated independently, one filling while the other is drying and
hardening. Supernatant liquor is drained from the one being filled by
overflow into a standpipe that carries the liquor down through the settled
solids and out to a collecting ditch extending around the pond (from which
the liquor is recycled to the plant). When the active section is filled
with solids almost to the top of the dike wall, the slurry flow is trans-
ferred to another section. After a time the solids dry and harden to the
extent that excavating equipment can be used to build the wall up a few
feet higher with excavated solids. The cycle is then repeated, resulting
in some gypsum piles as much as 100 feet high.
Whether this system can be used to reduce the acreage requirement
for lime-limestone scrubber slurry ponds remains to be seen. Small-scale
tests indicate that under some climatic conditions the sludge might harden
adequately for such a technique and that the angle of repose and stability
of the wall should be acceptable.
The necessity for recycling may make it desirable to start with
some water in the pond to provide recycle until the supernatant overflow
can take over. Otherwise the water balance and pumping situation at the
scrubber could be a problem. This will affect the composition of any seepage,
as discussed later.
In most areas evaporation and seepage from the pond should exceed
the rainfall collected but the local situation should be evaluated in planning.
Seasonal variations in the balance could make it necessary to keep some surge
capacity for supernatant liquor in the pond, in which case the effective pond
volume would be reduced.
In some situations it may be feasible to use a mined-out area as
the pond, as is done in some phosphate plants in Florida. The important
considerations are distance from the main plant and suitability of the area
for retaining liquid.
759
-------
In other cases it may be possible to use existing waste ponds
for the calcium solids--for example, ash ponds in power plants, tailing
ponds at smelters, and gypsum ponds at phosphoric acid plants. The last
of these would be especially appropriate because liquor-recycling facilities
will already be available in most cases. A complication, however, is that
the calcium sulfite (produced by absorbing sulfur dioxide from the sulfuric
acid plant tail gas) could decompose and give off sulfur dioxide if mixed
with the acidic supernatant liquor in the phosphogypsum pond.
Landfill Disposal
Since waste ponds are expensive and sometimes infeasible because
real estate is not available, the possibility of landfill disposal must be
considered. The term "landfill" as used here includes both (l) dumping into
suitable excavated areas (mines, quarries) or natural depressions, and (2)
piling solids at some suitable point near the main plant. All require de-
watering the sludge and getting the material in suitable physical form for
transport as a solid. The discarded solids may or may not be covered with
earth later, depending on the situation.
The attainment of suitable handling characteristics depends in the
first place on the degree of dewatering. As noted earlier, about 25-40$
water is typical for filtration and centrifuging tests carried out so far.
With some sludges, there are indications that this degree of dewatering
would result in acceptable handling properties, but this no doubt depends
on factors such as proportion of fly ash, CaO versus CaC03, degree of oxi-
dation, crystal size, and type of limestone. At any rate there is some
question at present whether simple dewatering will be adequate. Some of the
utilities, Commonwealth Edison, for example \Jj], plan to add dry materials
to the sludge to improve its properties. Dry fly ash is an obvious possi-
bility since it acts somewhat like a cement.
The cost of dewatering, plus any treatment to improve handling
properties, is likely to be quite high. Filtration rate has been fairly
good in TVA tests, typically 50-55 gal/hr/ft2. Exploratory centrifuging
tests have indicated that lower moisture content can be obtained as com-
pared with filtering, and if satisfactory centrifuging techniques and rates
can be developed, this method of dewatering might be utilized if solids are
to be transported without reslurrying. The optimum type of filter or centri-
fuge has not been established.
For operation in which a relatively low solids content is carried
in the scrubber slurry, it may be economical to operate a thickener before
the final dewatering device. Use of a thickener as the sole dewatering unit
would not seem indicated, however, since the relatively low degree of de-
watering would increase the amount of dry material necessary to give a
manageable solid.
760
-------
A further possibility is to "thicken" the water in the sludge.
Relatively small amounts of certain commercially available materials thicken
water to a solid or semisolid and thus could convert a soupy sludge to a
form more easily handled.
Again, oxidation to sulfate should give a product with good
handling properties directly from the filter or centrifuge. The Japanese
processes, both sulfurous acid oxidation (Chiyoda) and calcium sulfite
oxidation (Mitsubishi, Bahco, Ishikawajima-Harima Heavy Industries Company,
Ltd.) produce a relatively dry solid by centrifuging or filtering and very
likely could do so by thickener settling.
Assuming production of a solid with good handling properties, the
next problem is planning the transport system--truck, rail, barge, belt,
overhead cableway, or other. The cost of any of these systems will normally
be much higher than for sluicing to a waste pond, which is the reason that
phosphogypsum disposal seldom involves dry transport; in only two or three
of the world1s phosphoric acid plants, located in Europe, are the waste
solids transported in the solid state (by truck or conveyor belt).
If the dewatered sludge can be piled satisfactorily, it may be
attractive to transport to a suitable area near the plant and there build
a mound something like the coal pile. The saving in pond construction cost
and liquor recycle should be considerable. For the relatively short distance
involved, it might be feasible to pump the dewatered sludge with a positive-
displacement pump.
From the long-range standpoint, a good case probably can be made
for returning the sludge to the limestone quarry or the coal mine—thus
returning the nonusable constituents of the original raw material back to
the place from which they came. There are obvious problems but the very
large tonnages may make such a course imperative.
For both landfill and ponding, wind and water erosion of the
exposed sludge surfaces could become a problem. If so, covering with earth
may be an acceptable solution but an expensive one. Or established practices
for stabilizing the surfa * of waste piles may be applicable.
One consideration ...• landfill is whether the dumped solids will
bear the weight of people or animals. As noted earlier, ponded phospho-
gypsum will eventually bear the weight of even excavating equipment. From
preliminary tests it appears that some CaO-CaC03 scrubbing solids may
eventually behave similarly. It should be noted, however, that differing
process conditions produce waste solids of different characteristics. Type
of limestone seems to have an effect, some producing waste solids much more
thixotropic than others.
761
-------
Water Pollution
Although calcium sulfite and sulfate are relatively insoluble
compounds, small amounts dissolve in the liquid phase of the scrubber
slurry; moreover, magnesium in the limestone produces soluble salts and
soluble materials such as chlorides and nitrates are introduced with the
gas. Also, in double alkali processes some of the alkali will usually be
left in the sludge, dissolved in the liquid phase.
The nature of the water pollution problem resulting from these
dissolved impurities depends mainly on whether the system is open or closed
loop in regard to water. By closed loop is meant operation without any
purging or "blowdown" of liquor other than that remaining in the discarded
solids. In contrast, full open loop is similar to current ash pond practice
in which all the sluice water is overflowed to a watercourse.
In open-loop operation, fresh water must be added to the scrubber
loop at a rate sufficient to maintain the desired solids concentration in
the slurry to the pond. Since this is usually 5-15$ (unless a thickener
is used), a large amount of water must be introduced--which, incidentally,
makes scrubber operation much easier. As a result, the chlorides and other
impurities are purged rapidly and the concentration of these constituents
remains at a low level. Sulfite and sulfate concentrations are not affected
as much because the solution is in contact with excess crystals all the time
and thus can become saturated. The concentration is still below that for
closed-loop operation, however, because.eliminating the recycle from the
pond prevents supersaturation from building up to the high levels encountered
in closed-loop systems.
Thus open-loop operation produces a relatively low concentration
in the overflow, low enough to meet many water pollution regulations. A
typical liquid phase analysis for open-loop operation, taken from runs in
the TVA pilot plant, is given in Table I. For comparison, the overflowing
sluice water from a fly ash pond (TVA at Widows Creek) contains about
250 mg/1 of dissolved solids.
The situation in regard to water pollution regulations is quite
complex, making it quite difficult to predict the degree to which open-loop
operation can be tolerated. In some cases the regulations are based on the
degree to which the impurity concentration in the receiving watercourse is
increased, which depends, of course, on the ratio of watercourse flow to
effluent flow, the impurity concentration already in the watercourse, and
the amount of pollutants introduced by the effluent. Thus each situation
must be evaluated separately in regard to open-loop acceptability.
As an example of the complexity, the Illinois regulation applicable
to the Commonwealth Edison's limestone scrubbing system originally was such
as to limit the dissolved solids in the effluent to 750 ppm (7). This was
changed recently to read that "total dissolved solids shall not be increased
more than 750 mg/1 above background concentration unless caused by recycling
or other pollution abatement practices, and in no event shall exceed 3500 mg/1
at any time." The latter is much more lenient and reportedly would allow
open-loop operation.
-------
TABLE I
Composition of Pond Liquor (Open-Loop Operation;
Constituent Concentration, tng/1
Calcium 815
Magnesium 85
Sulfate 1450
Sulfite 70
Chloride 675
Sodium plus potassium 59
Iron 0.17
Barium 0.15
Cyanide
-------
situations and that closed-loop operation would unquestionably eliminate
any concern. This is not necessarily true, however, because regulatory
attitudes differ so much that, although in some areas full pond outflow
is allowed, in others there is concern even about seepage from the pond.
The latter has become so much in question, in fact, that major research
programs are being mounted to determine the magnitude of the problem, if
any, and possible means for solving it.
In closed-loop operation, all constituents build up in the liquid
phase until the purge afforded by the liquor remaining with the solids
(which must be replaced with fresh water) brings the concentration to a
steady-state level. Thus the final liquor content of the fully settled
solids will determine the actual steady-state concentration, low residual
liquor in the solids giving high dissolved solids concentration in the
liquor. If any water is placed in the pond before startup (because of the
pumping arrangement, as discussed earlier), this also will affect liquor
composition, making it less concentrated in the beginning of the pond
operation and more concentrated toward the end.
Liquor composition will also vary widely with the composition of
both the limestone and the coal (or the oil, the ore, or other source of
sulfur dioxide). The best situation is in scrubbing tail gas from a sul-
furic acid plant, for which the impurities come only from the limestone.
The coal and limestone compositions expected for TVA* s Widows Creek operation
are given in Table II. Under closed-loop operation, it is expected that the
composition of the pond liquid phase will approximate that given in Table III.
Although the concentrations of some of the major constituents are higher than
for open-loop operation, the increase is not enough to cause a significant
effect on river water quality even if major seepage occurred.
TABLE II
Expected Composition of Coal and Limestone
for TVA's Widows Creek Operation
Limestone, % by wt
Acid
CaO MgO NaP0 KP0 CljA* FegOa _S _ COg insoluble
51.1 2.1 o.oi 0.17 0.66 o.4o o.o4 4o.8 4.0
Coal, as fired
_ _ % by wt
Sulfur Ash Moisture Heating valve, Btu/Ib
4.5 25 5 10,000
764
-------
TABLE III
Composition of Pond Liquor (Closed-Loop Operation)
Constituent _ Concentre tion, mg/1
Calcium 830
Magnesium 230
Sulfate 11*00
Sulfite 1^5
Chloride 1200
Sodium plus potassium 50
Iron 0.07
Barium 0.2
Cyanide
-------
5- In some areas, sandy soil conditions might make an impervious
pond difficult to construct. (However, a lining of clay or
other impervious material could be used.)
It should be noted that there are many large waste ponds over the
world that in some instances contain liquors much more impure than that
from lime-limestone scrubbing. The liquor in phosphogypsum ponds, for
example, has low pH and a high concentration of fluoride. Seepage from
such ponds does not appear to be a problem.
For landfill the situation is somewhat different since the disposal
area would not ordinarily be prepared as carefully as a pond. A primary
consideration would be protection against erosion of solids* Dissolution
in rain water, however, would not appear to be a problem. If it should
prove to be, the methods described earlier for improving the handling proper-
ties by a cementing or water thickening action should also reduce any leaching
in landfill disposal.
Waste Calcium Sulfate from Acid Neutralization
There are at least two situations that might arise in which sul-
furic acid made in the process of reducing sulfur dioxide emission could be
justifiably neutralized with limestone to make waste calcium sulfate. One
is the situation in many smelters, where the stack gas is rich enough in
sulfur dioxide to make acid production by the standard method the most
appropriate way for reducing emission. However, in some areas, particularly
in the western part of the United States, there is little or no market for
acid beyond that already being produced and therefore the acid would have
to be neutralized. The second is the situation in which a system has been
installed for acid recovery and sale but the market fails or becomes inade-
quate, in which case the acid must either be stored or neutralized. Since
acid storage is expensive (requiring 17.6 cubic feet per ton), neutralization
would appear to be more economical. Thus it may be desirable to install
neutralization facilities even though sale of the acid is the principal method
of disposal.
Work is under way on determining the best method of neutralization.
The preferred method, if feasible, is to react strong acid (such as produced
in a standard plant) directly with limestone to make a solid product, thus
avoiding the need for a solids separation step. The problem is similar to
that in making "ordinary superphosphate" in the fertilizer industry, where
phosphate ore (mainly calcium phosphate) is reacted with strong sulfuric acid.
The product slurry sets up in a short time to a solid that can be handled
and piled, but several weeks are required before complete reaction of the
acid is attaiitpd. In the fertilizer industry, the "curing" is done in a
covered building. For the sulfur dioxide removal situation, transfer of the
solid to a landfill disposal area could pose an "acid leach" problem until
reaction was completed.
766
-------
Another approach is to dilute the acid, react with limestone,
and separate the solid. Complete reaction can be attained in the dilute
system, and the solids can be washed, if desired, to remove occluded acid.
Research is also being carried out on this method.
Elemental Sulfur as a Waste Material
At the moment it seems unlikely that production of elemental sul-
fur as a waste material will become significant. However, it is conceivable
that, with further development, such a system would be more economical in
some situations than production of calcium sulfite-sulfate or calcium sulfate
as the waste product.
The closest analogy would appear to be the current situation in
western Canada, where large quantities of sulfur have been produced from
sour natural gas and stockpiled because of marketing problems. The problems
are the cost of reclaiming if the sulfur is ever marketed and also the
emission of hydrogen sulfide from the pile. Sulfur produced from sour natural
gas by the Glaus process may contain up to about 0.01% hydrogen sulfide which
causes an odor but is not considered .very hazardous in outside storage;
hydrogen sulfide evolution could present a problem in inside storage and the
odor in outside storage would be quite objectionable.
There is also the possibility of wind and water erosion, plus
oxidation that would produce acid constituents subject to leaching. It
appears that these problems have not been adequately evaluated for the
situation in which sulfur is discarded permanently in the vicinity of plants
in populated areas.
Assuming that these problems can be handled satisfactorily and
that the economics are acceptable, sulfur as a waste material has some
attractive features. The volume requirement is only about 23 cubic feet
per ton, and since there are no diluting constituents the waste storage
volume required is only about 10% of that for lime-limestone scrubber sludge.
The potential (long tern) value of the material is also important; even if
discarded, it should remain available for use at relatively low reclaiming
expense, if at some future time the demand were such as to make marketing
attractive.
Use of Lime-Limestone Scrubber Sludge
Although the sludge from lime-limestone scrubbing is a somewhat
unattractive material—wet, impure, and containing unreacted limestone--
there has been a considerable amount of research on finding a use for it.
Companies such as Combustion Engineering, G. and W. H. Corson, Dravo, and
others are involved in this, and EPA has funded work by West Virginia University
and by the Aerospace Corporation.
767
-------
The main approach has been to find bulk uses for the material
that require very little processing. One such use is as a soil amendment,
analogous to use of waste gypsum from the phosphate Industry as "land
plaster" in the Far West. The main value for such use, however, is to
improve the physical properties of alkali soils, a problem that is of
little significance in the areas where most of the sludge is likely to be
produced. Fanners would have little incentive to accept the material, and
even if they did the net result would appear to be a relatively expensive
type of landfill operation. And the leaching problem would be magnified.
Other efforts have been aimed toward converting the sludge to a
material suitable for use as a highway base material. It has been demon-
strated that some quantity of sludge can be sold for such use but it seems
highly unlikely that a market could be developed for any significant pro-
portion of the potential sludge tonnage. Special treatment would be
necessary since gypsum used as such has been known to dissolve slowly and
collapse under weight.
Other uses, such as for structural products, mineral wool pro-
duction, and beneficiation to produce minerals, would appear to be almost
hopeless. The situation can be compared with that of fly ash, which is an
excellent additive to cement—used in .large quantities (10,000 tons in 19J2)
by TVA in power plant foundations and in dam construction. Notwithstanding
an intense promotional effort, TVA has been able to move less than 1% of the
annual fly ash production. The situation would thus appear quite difficult
for moving scrubber sludge, which is much less useful and economical than
fly ash.
Gypsum for Construction Use
The usefulness of lime-limestone scrubber product solids can be
improved if fly ash is kept out, unreacted calcium oxide or calcium carbonate
is eliminated, and calcium sulfite is oxidized to sulfate. The result is
a fairly clean grade of gypsum, suitable for use in making wallboard or as
an additive to cement.
The only use of this approach has been in Japan, where the general
use of oil as fuel eliminates the fly ash problem, acid is added to react
with any excess absorbent, and the sulfite is oxidized in special equipment
to sulfate. One of the papers in this symposium describes this practice in
detail.
Sulfite oxidation equipment is also being installed at the Mitsui
Aluminum plant in Japan, where coal is used as boiler fuel. However, the
plant is equipped with very efficient electrostatic precipitators that keep
the ash content of the sludge at a minimum.
768
-------
It should be noted that even with oil firing there is some
residual soot that darkens the product gypsum. Moreover, the cost of
oxidizing and of drying the product solid is not insignificant. In the
United States, where there are large natural deposits of dry, relatively
pure gypsum, costs of oxidizing and drying scrubber sludge would appear
to be unacceptable. Moreover, the dark color (in coal-burning plants,
from fly ash and other impurities)—as compared with the white color of
wallboard made of natural gypsum—could be a major obstacle to marketing.
For the Chiyoda and Hitachi processes, the gypsum is made from
sulfuric acid and therefore the expensive oxidation step is not required.
This would improve the economics.
Sulfuric Acid
Several of the proposed recovery methods give sulfuric acid as the
end product. Catalytic oxidation (Monsanto) and some of the carbon processes
(Lurgi, Hitachi) are restricted to acid as the product, and the magnesia
methods (Chemico-Basic, Grillo, United Engineers), plus some of the carbon
processes (Reinluft, Bergbau-Forschung, Sumitomo), are better suited to acid
than to elemental sulfur. Others, such as copper oxide (Shell, Esso-B & w),
sodium scrub (Wellman-Lord, lonics-S & w), and ammonia scrub (EPA-TVA),
usually give an essentially pure stream of sulfur dioxide that can be con-
verted either to acid or elemental sulfur as the situation dictates.
Acid production methods can be said to be farther along than sulfur
processes, on the basis that acid production from sulfur dioxide is a well-
established technology whereas sulfur dioxide reduction to sulfur has been
operated commercially in only one plant. The main problem with acid, other
than operability of the gathering processes for producing the sulfur dioxide
(or of the Monsanto and the carbon methods that produce acid in situ in the
main gas stream), is marketing of the product.
One problem is that the acid from some processes (Monsanto, Hitachi,
Lurgi) is not commercial grade, either because of lower concentration than
normal or because of impurities. There would be more difficulty in selling
such acid than for the standard 92-98$ water-white product.
Even if the acid produced is high grade, however, the very large
tonnage, the high cost of storage and shipping, and the vagaries of the acid
market are major problems.
This is particularly true in the Southwest and Mountain areas,
where the large sulfur dioxide emission from smelters and power plants, the
limited acid consumption in these areas, and the long distance to an adequate
market combine to make acid production for sale a very dubious proposition.
769
-------
In a study made by the Arthur G. McKee Company for EPA (lo), it was esti-
mated that only about 60-65^ of the sulfur dioxide emitted from western
smelters could be sold as acid (5-0-5-5 million tons per year) and that
this could be done only if the acid were priced at $4 per ton. This leaves
the remainder of the smelter acid capacity and all of that from the power
plants and other sources as essentially unsalable.
Probably the best situation for acid production is that in which
a process for using the acid is operated contiguous to the power plant, thus
avoiding the cost of shipping and marketing the acid. A phosphate fertilizer
plant is the most likely prospect because most of the sulfur consumption is
in the fertilizer industry. Phosphoric acid, triple superphosphate, and
ammonium phosphate are the logical end products.
An appropriate location for such a joinder of processes is the
upper Midwest, where over half of the phosphate fertilizer produced in the
United States is consumed and where many of the power plants burning high-
sulfur coal are located. It should be noted, however, that there are many
drawbacks to such an Arrangement, the main one being that the sulfuric acid
must be used as it is produced (unless expensive surge storage is installed)
and thus the fertilizer facility would have to be operated even at times
when otherwise it would not be economical to do so
The overall problem of marketing sulfuric acid is quite complicated,
too much so for any full treatment here. EPA is funding a market study at
TVA which should be useful in evaluating the situation; the TVA power system
will be used as an example and an analysis made of the potential quantity of
acid, shipping cost to various points, and the amount of acid that could be
produced and marketed for various levels of netback (including zero and minus
levels). The study is in the beginning stages.
Elemental Sulfur
Sulfur has several advantages over acid as a product for marketing,
including lower cost of storage, higher concentration for shipping, better
marketing flexibility, and broader spectrum of use. It seems likely, however,
that costs will be higher because of the need for a reducing agent. Moreover,
the basic economics are questionable since the sulfur is in the oxidized form
in the stack gas and this is the form (as sulfuric acid) in which most of it
is used. It does not seem reasonable to back and fill by oxidizing the sul-
fur compounds in the fuel during combustion, reducing the resulting sulfur
dioxide back to sulfur, and then oxidizing it again to acid before use—unless,
of course, storages and shipping costs are overriding considerations. It is
for these reasons that installing an acid-using facility at the power plant
seems economically desirable.
770
-------
The future of sulfur versus acid as a recovery product is un-
certain at this time. Much will depend on the process economics as they
finally develop--plus, of course, local considerations that may favor one
product over another.
Some of the processes ^Westvaco (carbon), Institut Francaise du
Petrole (ammonia scrub), U.S. Bureau of Mines (sodium citrate scrubby
produce sulfur only. Others, as noted above, can convert the gathered
sulfur dioxide either to sulfur or to acid.
Fertilizer Products
Some processes produce a fertilizer material directly without
sulfuric acid or sulfur being involved as an intermediate. An example is
ammonium sulfate, which is made in Japan either by ammonia scrubbing (with
ammonium sulfite oxidation) or by the Mitsubishi manganese absorption process
(reaction of manganese sulfate with ammonia to give ammonium sulfate). Bufete
Industrial, an engineering and construction firm in Mexico, also has a process
in which ammonia scrubbing is followed by ammonium sulfite crystallization
and oxidation of the sulfite to sulfate in the dry state.
There is some question as to the marketability of ammonium sulfate
in large quantities. Even in Japan, where ammonium sulfate has been a more
important fertilizer than in the United States, the material is losing ground.
The relatively low nitrogen content, as compared with fertilizers such as
ammonium nitrate and urea, is a major handicap. It is considered, however,
that if ammonium sulfate were made in power plants located on the Mississippi
River system, barge transportation could be used effectively to reduce the
adverse effect of low concentration on shipping cost. By pricing the material
below the present level, fairly large tonnages probably could be moved.
One factor offsetting the low nitrogen content of ammonium sulfate
is that sulfur is also a plant nutrient. In the past the natural supply of
sulfur compounds in the soil, plus that supplied incidentally in fertilizers,
made primary application unnecessary. This situation is changing because of
soil depletion and general use of fertilizers containing little sulfate. Thus
the sulfur in byproduct ammonium sulfate is likely to have increasing value
as soils become further depleted.
Other fertilizers, including some nitrogen-phosphate combinations,
can be made in scrubbing processes that do not involve intermediate production
and shipping of sulfuric acid. These have been evaluated in a TVA study
carried out for EPA (ll). It was concluded that a fair tonnage of such
products could be sold even in direct competition (same netback) with present
commercial fertilizers.
771
-------
Future Market for Sulfur Products
Whether sulfur products will be recovered and marketed in large
quantities depends on the cost of recovery processes versus the throwaway
type of operation (see earlier discussion of forthcoming EPA-TVA study).
This in turn depends, in part at least, on whether the recovered products
can be marketed and what netback they will provide.
A major question in this is how well the demand for sulfur-based
products will keep up with the supply from byproduct sources. This is a
complicated question, covered in a recent study for EPA by Esso Research
and Engineering and summarized by .a paper in this symposium. Up until a
few years ago, the main production was from natural deposits and the supply
could therefore be adjusted more or less to comply with the demand. With
the growing production from desulfurization of sour gas and now perhaps
from fuel and stack gas cleaning, the supply from byproduct sources is be-
ginning to outstrip the primary production. As a result, the market situation
has become more complex, more difficult to analyze, and more discouraging to
those who need a reasonably stable market prospect if they are to invest
billions of dollars in stack gas cleaning installations.
772
-------
References
1. Davenport, J. E. Tennessee Valley Authority unpublished report.
2. Borgwardt, R. H. Environmental Protection Agency unpublished report
(November 1972).
3. Potts, J. M., Jordan, J. E., et al. Tennessee Valley Authority
unpublished reports.
4. McClellan, G. H., and Mills, M. E. Journal of the Air Pollution Control
Association 2£, No. 2, 122-2? (February 1975).
5. Nason, M. C. Tennessee Valley Authority unpublished report.
6. Kelso, T. M. , Schultz, J. J. , et al. Tennessee Valley Authority unpublished
reports.
7. Gifford, D. C. Proceedings of Sulfur in Utility Fuels: The Growing
Pi lemna, a technical conference sponsored by Electrical World, pp. 2%5-87 (1972).
8. Tennessee Valley Authority. Environmental Statement; Experimental SOg
Removal System and Waste Disposal Pond Widows Creek Steam Plant
(January 15, 1973 J. "
9. Goldschmidt, K. Steinkohlen-Elektrizitat Aktiengesellschaf t (STEAG),
private communication.
10. Arthur G. McKee and Company. Systems Study for Control of Emissions
Primary Npnferrous Smelting Industry (June 1969). Vol. I, Report No.
PB lb^-88^; Vol. II, Report No. PB 18^-885; Vol. Ill, Report No. PB 1&-886,
Clearinghouse for Scientific and Technical Information, 5285 Port Royal
Road, Springfield, Virginia 22151.
11. Tennessee Valley Authority. Sulfur Oxide Removal from Power Plant Stack
Gas: Conceptual Design and Cost Study. Ammonia Scrubbing Cl970jl
Report No. PB 196 -Bo4, Clearinghouse for Scientific and Technical Information,
5285 Port Royal Road, Springfield, Virginia 22151.
773
-------
STUDY OF DISPOSAL AND UTILIZATION
OF BY-PRODUCTS FROM THROW AW AY
DESULFURIZATION PROCESSES
by
J. Rossoff
R. C. Rossi
J. Meltzer
The Aerospace Corporation
El Segundo, California
775
-------
STUDY OF DISPOSAL AND UTILIZATION OF BY-PRODUCTS
FROM THROWAWAY DESULFURIZATION PROCESSES
by
J. Ross off
R. C. Rossi
J. Meltzer
The Aerospace Corporation
El Segundo, California
ABSTRACT
This paper summarizes an on-going Aerospace Corporation study
for the Environmental Protection Agency concerning the ecologically sound
disposal of sludges produced by the limestone scrubbing of coal-fired uility
boiler flue gases. Related utility disposal problems concerning lime-coal
and limestone-oil sludges are also considered. In addition, the usage of
sludge in commercial products is briefly summarized. A review of critical
factors which affect disposal requirements and techniques is given. These
include items such as the chemical characteristics of process materials and
sludges, sludge physical properties, water quality criteria, trace element
constituents, and liquor recirculation. Disposal methods described include
ponding and landfill-type operations. The impact of input parameters on
disposal methods is briefly discussed, as well as features of ponding and
landfill requirements.
776
-------
1. Introduction
With the expanding use of coal by the electric power utilities, and
the incorporation of lime or limestone scrubbers to desulfurize the flue
gases, the disposition of and resultant effects of the sludges produced by
these processes has become a major concern. The concern for the dis-
position of the sludges is based on two principal factors: (1) the sludges
contain soluble salts and some compounds containing trace elements --
all or part of which may leach or drain from ponds, landfills or other us-
age areas to ground or surface waters and produce problems of water pol-
lution, and (2) many lime/limestone/sulfur sludges are highly water reten-
tive and are therefore not structurally acceptable in their natural form as
landfill materials. At this time, little is known of the potential effects, if
any, that may result from these sludges; however, there has been consid-
erable effort expended in the determination of technology to support the
utilization of the sludges and to condition them for use as a landfill material.
Since the sulfur scrubbing technology is just emerging from its pilot plant
stage, representative power plant operational disposal data are not abundant.
Additionally, current water quality standards are not directly applicable in
many cases, or are very difficult to apply to the leaching of sludges. There-
fore, adequate data do not exist to verify whether the sludges can or will be
used or disposed of in any ecologically sound manner.
The EPA Control Systems Laboratory, Research Triangle Park,
North Carolina, has designed a program to establish a valid understanding
as to whether the sludges pose a water quality problem, to assess solutions
or approaches to solutions in the event the sludges do pose a problem, and
to assess the industrial status of sludge usage and disposal in light of this
analysis. As part of this program, a contract has been let to The Aerospace
Corporation to perform analyses and bench scale tests of lime stone-coal
sludges to study the potential effects just mentioned, and to assess the status
of industrial usage and disposal technology. Background for this study is
fully described in an EPA position paper {Ref. 1).
The Aerospace study is now in the early stages of a program directed
toward an understanding of limestone sludges from power plants burning
eastern coal and western coal. That program* is the subject of this paper
A potential expansion of the program to include additional samplings from
other power plants so that a broader data base can be obtained, and to in-
clude a more detailed study of material conditioned for usage or disposal
is being planned by the EPA (Ref. 1).
777
-------
wnich is a status report that essentially summarizes the program's objec-
tives and progress to date.
2. Rationale
Much has been written a'bout the projected abundance of sulfurous
sludges and the technical feasibility of their use as a constituent of commer-
cial products or as a landfill material. For example, in Reference 2, Brackett
presents data taken in a 1971 survey of the electric utilities industry which
estimates a sludge plus bottom and fly ash production of 117 million tons in
1976. And, in Reference 1, it is noted that a government interagency Sul-
fur Oxide Control Technology Panel has made a preliminary lime/lime stone
sludge (50% solids) production forecast of 86-112 million tons annually by
1980. That value is based on 75% of coal burning power plants using a
desulfurization system, 75% of which will be lime or limestone scrubbers
(see Figure 1).
Although these values are necessarily "soft," they highlight an
acute by-product disposition problem. It is significant to note that the plants
committed to lime or limestone scrubbers are located in widely spaced
areas such as Pennsylvania and Nevada as well as in the North, Mid-West and the
South. Thus it is evident that disposition is not a matter of local concern, but
definitely one of national interest.
The concern for potential toxicity problems is emphasized by some trace
element data presented by Aerospace in Reference 3. These data, produced
for the EPA by the Shell Development Company and the Oak Ridge National Labor-
atories, were taken from samples of ash modified by the dry injection of lime-
stone into the boiler. The elements of concern show little tendency to be con-
centrated in any size fractions, five of which ranged from an average diameter
of 28. 7 //m to 1. 5 #m. In all size fractions, however, heavy metals were re-
ported at levels high enough to cause concern, e.g., As: 50 to 200 ppm; Ba
200 ppm; Pb: 200 to 500 ppm; V: 200 ppm. If these elements are combined
such that they are chemcially inert, there would be little concern over their
reaching ground or surface waters. However, application of wet lime/lime-
stone scrubbing of flue gases from coal-fired boilers is just emerging and the
possibility of producing toxic water-soluble compounds has not been determined.
The most often considered methods of sludge disposal are ponding ahd landfilling
778
-------
Figure 1
COMPARISON OF ESTIMATED ELECTRIC
UTILITY COAL GENERATION CAPACITY WITH CURRENT
AND FORECAST CONTROL BY LIME/LIMES TONE SCRUBBING
CAPACITY, KNOWN CONTROL
COMMITMENTS
1972
1974
1975 1976
YEAR
1977 1978 1979 1980
Courtesy of EPA Control
Systems Laboratory
(See Reference 1)
779
-------
and, for commercial applications, as an additive in road base construction
among other technically feasible products (Refs. 8 through 12). Sludges
used in these applications, whether treated or not, can possibly leach or
drain toxic trace elements (as well as other contaminants) to ground or
surface waters. It is on this basis that a determination is considered nec-
essary as to whether the sludges pose an ecological problem.
3. Aerospace Study Plan
The plan on which the Aerospace study is based was derived from
information such as that previously mentioned (Ref. 1). Supporting data re-
garding the production of lime/limestone sludges are given in an EPA paper
"Control of Sulfur Oxide Pollution from Power Plants" (Ref. 4), an article,
"Removing SO2 from Stack Gases, " (Ref. 5), and a previous Aerospace
study, "Technical and Economic Factors Associated with Fly Ash Utilization, "
(Ref. 3). From these and other data, the EPA Control Systems Laboratory-
defined a one-year program, "Lime stone-Coal Sludge Characterization and
Toxicity Study, " which The Aerospace Corporation is now conducting. This
study is currently limited to analyses of limestone-coal sludges from one
plant burning eastern coal and one burning western coal; however, surveys
of disposal techniques and utilization effects are not restricted to these two
plants. An EPA plan for the potential expansion of this effort to other plants
including some which use lime as the absorbent material for eastern and
weatern coal-burning plants, and limestone scrubbing at an oil-burning plant
is given in Reference 1. This potential expansion is designed to produce a
reasonably broad data base such that adequate results can be obtained to apply
generally to lime/limestone scrubbers.
Following is a listing of the program objectives of the current Aero-
space study and descriptions of each phase:
a. Chemical Characterization - chemical characterization (includ-
ing trace element identifications) and crystalline phase determina-
tions are being made of sludge, process liquors, coal, limestone,
make-up water, and regular fly ash. Special emphasis will be
placed on the chemical and physical state of toxic elements.
These chemical analyses will determine possible elemental
losses that occur in the combustion or collection processes;
they will determine the possible toxic hazards associated with
fresh materials that have been dewatered and with materials
that have experienced subsequent treatment and they will identify
780
-------
soluble components and sublimation components. At this time,
samples are being analyzed from the Prototype Turbulent Con-
tact Absorber (TCA) scrubber at TVA Shavmee Unit #10 in
Paducah, Kentucky (Ref. 6) and the pilot TCA scrubber at the
Mohave Station Unit #1 in Nevada (Ref. 7).
b. Potential Toxicity Determination - A literature study is being
conducted to determine any potentially toxic effects of the com-
pounds identified in the chemical analyses. This will include a
determination of the means of inhalation or ingestion in humans,
and an assessment of the potential intake of toxic elements in
plants and animals.
c. Physical Properties - Physical property studies are being con-
ducted on sludges that have experienced aging, i. e., drying,
leaching, exposure to sunlight, to assess permeability, com-
pressive strength, pozzolanic properties, and other factors
which may have an effect on the use of the sludge as a landfill
material.
d. Detoxification - An assessment of the potential for detoxification
of the sludge will be made. This will include schemes such as
oxidation, fixation, and heating. Positive indications will result
in process definitions, estimated costs, and the potential impact
on handling and disposal or usage.
e. Disposal Methods - In light of any potentially toxic effects identi-
fied, evaluations will be made as to the feasibility of various
types of disposal such as ponding, landfilling, and mine filling.
Also, a limited bench scale program is being conducted tc define
the properties that will affect the economics of transporting the
sludge to disposal sites. This includes corrosive properties,
abrasion, adhesion, thixotropic properties, viscosities of sus-
pension, and bulk densities as a function of water content.
f. Disposal Costs - An engineering study will be conducted to define
the relative costs for the ecologically sound disposal of the sludge
as a function of variables such as transport type and distance,
treatment processes, receiving site preparation, operations, and
maintenance.
781
-------
g. Water Quality and Solid Waste Disposal Criteria Review - A
review of federal and state solid waste disposal and water
quality criteria, is being made. The findings of this program
will be incorporated and a reasonable interpretation will be
made as to the impact of the disposal of sludges on the existing
or proposed criteria.
The potential EPA expansion of this program will include additional
sampling sources representing a broader cross-section of disposal and
treatment processes as well as other sorbent/fuel classes. Particular em-
phasis will be given to the assessment and evaluation of both on-going and
developing treatment processes now being employed by industry and govern-
ment agencies.
4. Water Quality and Solid Waste Disposal Criteria
In conforman.ee with the requirements of the Water Quality Act of 1965
which amended the Federal Water Pollution Control Act of 1965, all of the states.
the District of Columbia, and the territories of Guam, Puerto Rico and Virgin
Islands established or are establishing water quality standards* for interstate
(including coastal) waters. In December 1970, the responsibility for adminis-
tering the Water Quality Act of 1965 was transferred from the Secretary of the
Interior to the Administrator of the EPA. Most of the state standards have
now been written, and accepted by the EPA. The state standards are therefore
the major sources of criteria by which the power plant scrubber effluents are
to be judged at this time and they deal with the quality of the receiving surface
waters only. Further, the state standards when citing criteria for domestic
water supplies or for food processing, generally repeat or refer to the Public
Health Service Drinking Water Standards (Ref. 18) which apply to water dis-
tribution systems.
Other legislation, e. g., the Federal Water Pollution Control Act A-
mendments of 1972 (PL92-500) which applies to both surface and ground waters,
establishes a goal of zero pollution discharge by 1985. While calling for interim
guidelines and standards to regulate pollution discharges, it establishes the
applicability of two definitions of particular interest to future consideration in
sludge studies. They are:
*
Significant background information for the state standards can be found in
"California Water Quality Criteria, " McKee and Wolfe (Ref. 16); "Federal
Water Quality Criteria, " FWPCA, Dept. of the Interior, (Ref. 17); and
"Public Health Drinking Water Standards," PHS Publication 956 (Ref. 18).
782
-------
a. "The term 'navigable waters' means the waters of the United
States including the territorial seas. "
b. "It is the national policy that the discharge of toxic pollutants
in toxic amounts be prohibited. "
Summarizing regulations, we find the following:
a. Current standards are established by the states which specify
water quality criteria for interstate (including coastal) waters.
b. Future standards will regulate pollution discharges to all waters
of the United States including ground waters. And the discharge
of toxic elements in non-toxic amounts may be allowed.
The route by which potential pollutants can reach various water systems
is either by direct discharge or runoff, or by leaching through the soil to
ground waters or surface waters. Because of the high concentration of dis-
solved solids (generally sulfates, carbonates, sulfites and chlorides)regard-
less of other possible pollutants, the process liquors generally cannot be
discharged without additional treatment and are therefore recycled. Disposal
sites can be constructed to prevent drainage. This leaves leaching as the only
other route (barring accidents) to potential pollution.
The effect of a leachate from a disposal pond or landfill on the quality
of a receiving body of water is generally a variable factor depending on local
weather, soil conditions, topography, the chemical and physical characteristics
of the leachate, and the flow and quality of the ground and surface receiving
waters. Because this effect is so difficult to determine, regulatory bodies
ordinarily will prohibit the earth disposal of an untreated sludge if it contains
heavy metals considered toxic in concentrations exceeding the safe limits set
for drinking water. This of course assumes that the toxic element(s) exists in
a. soluble compound within the sludge and that after leaching through the soil it
will appear in the receiving water in toxic concentrations. Such an approach
is undoubtedly safe, but the knowledge of when, where and under what condi-
tions these precautions are necessary is not known.
It can be argued that the concern for toxicity in the sludges may not be
critical since it will be necessary to treat the material to cause a fixation
condition which will eliminate the formation of a bog. This would permit the
landfill site to be reclaimed for some useful purpose and it may effectively
encapsulate the undesirable constituents of the sludge. This is no doubt true
in some instances; however, a fixation treatment isn't necessarily required
for all sludges and furthermore, it is not known what the impact of the leaching
783
-------
of a conditioned sludge will be. If it is considered harmful, then a problem is
defined. If it is not considered harmful, then the conditions which are accept-
able should be identified so that those who have both the problems of disposal
and regulation can satisfy their requirements.
Although the foregoing applies specifically to water quality, a few com-
ments regarding solid waste disposal are in order. The EPA Office of Solid
Waste Management Programs is in the process of issuing guidelines for land
disposal and thermal processing operations. This is in response to the Re-
source Recovery Act of 1970 {PL91-513) which amended the Solid Waste Dis-
posal Act of 1965 (PL89-272). The presently proposed guidelines do not apply
to hazardous materials because of a lack of sufficient information; however,
hazardous wastes and sludges containing free moisture are considered special
wastes which under certain circumstances may be accepted for disposal at a
disposal site but under the authority of the responsible agency. Most state
standards for solid waste disposal which may be applicable to sludges are
similar, i. e., special permission is required of the responsible office, and
finite criteria are not given.
In light of the foregoing, the current Aerospace study effort has foc-
used on determining the potential toxic hazard posed by the sludge as a con-
sequence of the variations that arise in alternative disposal schemes and
selected processing variables. A potential toxic hazard from a pond or dis-
posal site may exist in the leachate, the run-off, a purge flow, from vapors
arising from a dry pond, or fine dust particles blown from a dry pond surface,
or from anaeorobic attack from the bacteria in the containment soils. Toxicity
in any of these cases is defined by existing state and federal standards but is
subject to the chemical considerations discussed in the following section.
5, Toxicity
The determination of the toxicity of a sludge is dependent upon the
factors relating to what kinds of elements it contains, how much are present,
and in what form they exist. Many elements and substances that are acceptable
or even essential to the body can be toxic when consumed in large doses. The
question of what constitutes a large dose is often never well defined in physio-
logical terms but is instead Hie consequence of arbitarily defined factors. In
a few cases* such as mercury, arsenic, lead and cadmium, where the metal
784
-------
serves no known biological function, a toxic dose and threshold has been
defined from physiological experience. In most other cases, the concentra-
tion threshold is less well definable and "safe" dosage levels tend to be de-
fined by standards. In this study an attempt will be made to correlate toxicity
with specific physiological effects when possible and also with existing state
or federal standards.
An additional factor that determines the toxicity of a specific element
besides its concentration, is the chemical form in which it is fouund. Even
the most toxic material can be non-toxic when it is in a form that is unavail-
able to the body. However, caution must be taken to ascertain that it does
not exist in a form that could be assimilated by plant or animal life and
thereby become available to humans. Thus, a determination of a chemical
analysis, in itself, can not define toxicity but instead defines only concentra-^
tion. In addition to chemical analyses, this study will determine the chemical
state of each potentially toxic element and its subsequent availability to humans.
The study will start with a precise determination of the chemical
analysis of the sludge and its liquor, and the coal, limestone and process
waters constituting the raw materials in the process. A composite listing
of sampling points is given in Figure 2. In Table I is an analysis of a com-
posite sample of western coal (Ref. 19) and an analysis of a specific sample
of an eastern coal from Aerospace experimental data. Only those elements
that could be potentially toxic were selected. Also in Table I is a specific
analysis of fly ash and bottom ash from a western coal (Ref. 19).
A comparison of the coal samples analyses reveals specific
differences. Such differences can be found between coal samples within the
same region and do not necessarily indicate specific differences between
eastern and western coal. In general, the western coals are associated with
more basic minerals; the eastern coals with more acidic minerals. The
western coals tend to contain a broader spectrum of rare earth elements.
The trace elements are found usually as contaminants in minerals codeposited
with the coal or in the organic matter from which the coals originated. In
this regard, general statements referring to relative toxicity between the two
types is not possible.
The coal for which the fly ash and bottom ash were derived contained
approximately 10% ash. As a consequence, the ash samples could be expected
785
-------
Figure 2
COMPOSITE SAMPLING OF TCA SCRUBBER SYSTEMS
1. LIMESTONE
2. COAL
3. FLUE GAS PARTICULATES AT SCRUBBER INLET
4. FLUE GAS PARTICULATES AT SCRUBBER OUTLET
5. BOTTOM ASH
6. SLUDGE AT SCRUBBER EXIT
7. CLARIFIER UNDERFLOW
8. RECYCLE LIQUOR
9. CENTRIFUGE PRECIPITATE
-------
Table 1. Selected Elements in Coals and Ash (ppm)
Eastern Coal
Western Coal
Composite
Element
Arsenic
Mercury
Antimony
Selenium
Cadmium
Zinc
Manganese
Boron
Barium
Beryllium
Nickel
Chromium
Lead
Vanadium
Aerospace
Data
N. D.
<0. 01
<0. 05
N.D.
N. D.
180
350
46
1800
<0. 01
N. D.
310
30
180
Sample
3
0. 05
0.17
1.6
<0. 5
0. 56
15
15
400
N. D.
25
5
4
9
Fly Ash
(Ref. 19)
15
0. 03
2.1
18
<0. 5
70
150
300
5000
3
70
150
30
150
Bottom Ash
fRef. 19)
3
<0. 01
0.26
1
<0. 5
25
150
70
1500
<2
15
70
20
70
N. D. - Not detected
-------
to have a concentration of a specific element 10 times that of the coal sample.
(Since the western coal analysis was from a composite sample a direct com-
parison between the coal data and ash data is not valid in this case). Compar-
ing the fly ash and bottom ash reveals a. major difference in heavy metal con-
centrations. Relative to the fly ash, the bottom ash is low in refractory metals
and metals having high vapor pressures at combustion temperatures. In con-
trast, those base elements that easily slag, are more predominant in the
bottom ash.
In Table 2 is an analysis of slew water that may be typically found in
an ash pond (Ref. 20) and an analysis of a scrubber liquor from Aerospace
data. The analysis represents the solubility of some of the elements specifi-
cally found in a boiler ash and a sludge. Whereas only one of these elements
(manganese) poses a health hazard at the given levels in those cases, they
may not represent the actual case. In addition, factors existing in the sludge
processing technology may serve to concentrate other elements thereby posing
additional hazards.
As an example of the concentration effect, the case of fixation technology
will be used. In general, this technology requires the introduction of additives
to the sludge which react with the dissolved sulfates to form a new crystalline
phase. If toxic elements are accepted within this new phase, they become
eliminated from solution and no longer become available as a contaminant in
leaching or run-off waters. On the other hand, if they are rejected by the
new phase, they could become concentrated within the remaining water
and thereby pose a health hazard. Details of the specific chemistry of each
potentially toxic element have not yet been determined and the toxic hazard of
sludges in the various forms they may be found has not yet been assessed.
6. Disposal
Of the many million tons of sludge that will be produced each year, the
possibility of economic utilization exists for only a small portion of the total.
The major quantities of sludge produced will require disposal in an ecologically
sound manner. Presently, the alternatives being considered are ponding and
landfilling; both disposal operations will require the establishment of procedures
that avoid hazards related to health and safety or land use.
788
-------
Table 2. Selected Elements in Solution {ppm)
Element
Lead
Antimony
Barium
Manganese
Mercury
Beryllium
Boron
Nickel
Cadmium
Selenium
Zinc
Arsenic
Ash Pond
(Ref. 20)
.01
.015
.07
.075
<.00l
.002
.5
.015
.01
.035
.03
.01
Scrubber Liquor
Aerospace Data
<. 01
N. D.
<. 05
1.6
N. D.
N. D.
11
.05
N. D.
N. D.
N. D.
N. D.
PHS
Drinking
Water
Standards
. 05
-
1.
.05
-
-
-
-
.01
.01
5.
.05
N. D. - Not detected
789
-------
The requirements for disposal will be strongly dependent upon whether
or not the sludges are found to be toxic. If the sludges are found not to be toxic,
the problem of disposal is simplified, but the problem of safety and land use
must still be addressed. An untreated sludge will retain a considerable
quantity of water. Laboratory tests on a limestone sludge containing a rela-
tively high sulfate content will settle only to 45% solids content if under -
drainage is not provided, but if drainage to sub soil is allowed, the settled
sludge will reach a content of 50% solids. A maximum particle packing
equivalent to a value of about 70% solids content is reached upon air drying.
When a dried sludge is re-wetted, the sludge will bloat and decrease its bulk
density depending upon the quantity of water absorbed by the sludge. When a
sludge contains a high sulfite content, the sludge will settle to a value of about
35% solids whether underdrainage is provided or not. A sulfite sludge does
not dry as readily or settle as much during drying as a sulfate sludge.
A sulfate sludge for which underdrainage is provided will produce
savings relative to one with no such provision because the higher packing
density allows for the disposal of about 20% more sludge in the same ponding
volume. However, if the sub-soil is the only means of underdrainage, the
possibility for soil plugging exists especially in the case of sulfate sludges
containing high concentrations of dissolved solids. Depending upon soil type,
salts will precipitate within the soil, filling pores and preventing further
water passage. Thus, the advantage of underdrainage would be lost. The
more reliable system is one in which underdrainage is provided. Although
in the case considered the sludge is assumed not to be toxic, it is very likely
that the leachate would not be of a quality to be acceptable for discharge to a
water course.
Ponding can be used not only for the final disposal of the sludge, but
also as an interim measure to allow settling before removing the material to
a landfill site. The most fundamental problem in the ecologically sound dis-
posal of the sludge arises when the hazard exists for leaching toxic elements or
soluble salts to ground or surface waters, or drainage to surface waters.
The immediate solution to this potential problem may require dewatering
and/or fixation of the sludge and storing it in a clay lined pond so that per-
meation to the sub-soil is minimized. Overdrainage and underdrainage
790
-------
must return waters to the scrubber system thereby eliminating direct discharge
to surface waters. Such a pond can serve as a final disposal site, but in most
cases because of the large volumes of sludge and water to be handled, the pond
can serve as primary settling basin from which the sludge can be periodically
removed and placed in a landfill. A fixation treatment applied to the sludge
prior to placing it in the pond is an alternative technique generally applicable
only if the sludge is to be removed to a landfill after curing. This treatment
has a dual purpose, one, the removal of water and increasing compaction
qualities and, two, the minimization of leaching by decreasing the perme-
ability of the material.
If a sulfate sludge is found to be toxic, it may be adequate to store it
untreated in a pond having an impervious base. The use of underdrainage and
overdrainage would be required to prevent the formation of a bog and the
possibility of overflow, especially in areas where rainfall exceeds evaporation.
This, of course, requires the lining of a large area pond, monitoring, and
maintenance as necessary. A further problem would be faced when a point in
time is reached where the generating plant can no longer accept all drainage
originating from rainfall from all its disposal ponds, and direct discharge to
streams would not be allowed, or where the plant is to be abandoned. Under
either of those conditions, encapsulation by covering with an impervious
material may be the best solution.
Numerous alternatives for ponding with or without pond lining or sludge
conditioning may be possible. A cursory survey of the industry has shown
that many of these alternatives are presently being considered and employed.
Alternatives, such as treatment processes mentioned earlier, produce a high
solids content sludge material having acceptable structural qualities for a
landfill usage; a process for producing aggregate has been developed. The
details of these processes will be discussed by the various process develop-
ers. Future concern by Aerospace for processes of this sort will be in the
assessment of alternative disposal systems regarding potential ecological
problems.
Some very rough approximations of costs for interim pond and fixation
disposal have been put at about $7 to $10 per ton of dry sludge, (Refs. 15, 21)
which includes labor, trucking, additives, supervision, and fixation process-
ing. The cost of lining a pond varies considerably. For ponds in the 5 to 10
791
-------
acre size, estimates have been put at $5, 000 to $20, 000 per acre for clay
linings and stablized pozzolan base linings, respectively, without piping -
the clay lining will undoubtedly require maintenance to repair damage
when the "dried" sludge is removed. A soil covered plastic lining complete
with all drainage can cost $25, 000 to $30, 000 per acre. At this stage of
our study, these values, though only approximate, underline the need for the
determination of whether lined ponds with their attendant maintenance and
monitoring (equipment) are necessary and where they should be located.
7. Utilization
The utilization of sludges will be discussed by several speakers at
this symposium, therefore it is mentioned here only in terms of what the
ecological implications may be and how that relates to the Aerospace study.
Numerous programs directed toward sludge utilization have been conducted
through government sponsorship and by industry alone (see Table 3 ). Some
of the more significant of these are: 1) the successful development of the
technology for autoclaved calcium silicate products, mineral wool and soil
amendments at the Coal Research Bureau, West Virginia University under
EPA contract (Refs. 10, 12); parking lot construction, Federal Highway
Administration (Ref. 8) and I. U. C. S. (Ref. 11); landfill, road bases,
aggregate, and structural shapes, I. U. C. S. (Refs. 9, 11); characterization
and multiple use investigations at Combustion Engineering; and landfill and
sludge properties by the Dravo Corporation (Ref. 13).
A review of the many possible utilization schemes indicates the fol-
lowing significant factors regarding their potential impact on ecologically
sound disposal or utilization:
1. Applications to be considered for potential pollution effects are:
a. Fixation treatments for landfill
b. Aggregate for either landfill storage for future use, or
construction usage in road building
c. Road base construction
d. Soil amendment
The applications listed above are all subject to various degress of
leaching. For item l.a above, permeability values ranging from 2 x
^ C *J
10 cm/sec to 10~ cm/sec have been reported. In some industrial tests
of the treatment of acid waste sludges, it has been claimed that the permeability
—5 —7
has been educed from 10" to 10" cm/sec and the concentration of trace ele-
ments in the leachate was additionally reduced by about two orders of magnitude.
792
-------
TABLE 3
POTENTIAL UTILIZATION OF SLUDGE
Technology
Calcium-silicate Products; Mineral
Wool; Soil Amendments
Research Organization
Coal Research Bureau,
West Virginia University,
Under EPA Contract
Reference
10, 12
Parking Lot Construction
Federal Highway Administration, Wash. D. C. 8
Research Review
Coal Research Bureau
West Virginia University
14
VD
Sanitary Structural Land Reclamation;
Compacted Road Base; Binder for
Stabilized Road Base; Aggregate;
Structural Shapes
I. U. Conversion Systems, Inc.
Plymouth Meeting, Pa.
9,
Utilization Survey
The Aerospace Corporation
El Segundo, California
Sludge Properties
Dravo Corp., Pittsubrgh, Pa.
13
Characterization, Multiple Utilization
Combustion Engineering, Windsor, Conn.
Lightweight Aggregate
Michigan Institute of Technology
-------
Thus, as a consequence of chemically combining trace elements into new
crystalline phases and the reduction in leaching ratio, soluble salts or toxic
contaminants that may be available to ground or surface waters are reduced
by about 10, 000 times over sludge that is not treated.
2. Aside from using the sludge as a landfill, the commercial appli-
cations listed in item 1 above constitute a potential outlet for the usage of vast
tonnages of power plant sludges. A survey of the potential fly ash market by
The Aerospace Corporation (Ref. 3} in 1970-71 indicated that the maximum
potential marketability of fly ash in the near term was approximately 25 per-
cent of the supply. Major inhibitions identified were transportation economics
coupled with competition from other materials and the lack of control of fly
ash quality and supply. A growing market for bottom ash in road construc-
tion brightens the total ash utilization picture somewhat (Ref. 2). However,
with the flue gas scrubbing which increases the by-product tonnage by a factor
between two and three, a simple deduction identifies "disposal" as the major
outlet for the sludge, particularly in the near term.
A summary of potential utilization schemes for the sulfate sludges
(other than sulfur or sulfur products), the researchers, and related liter-
ature is given in Table 3. Sulfur and sulfur product recovery are considered
in other presentations at this symposium.
794
-------
8. Summary
The laboratory results presented herein represent only the initial
phase of experimentation. Since these data were derived from the analyses
of samples taken from one plant on a given day, they should in no way be
considered necessarily representative or typical. They do serve as a base
point, however, to which other experimental and analytical data will be
added so that a reasonable interpretation will be made of the potential im-
pact of the disposal of scrubber sludges on water quality and solid waste
disposal standards.
795
-------
BIBLIOGRAPHY
1. "Waste Product from Throwaway Flue Gas Cleaning Processes -
Ecologically Sound Treatment and Disposal, " J. W. Jones, R. D.
Stern, Control Systems Laboratory, EPA, Research Triangle Park,
North Carolina, for Presentation at Flue Gas Desulfurization Sym-
posium, New Orleans, La., May 14-17, 1973.
2. "Production and Utilization of Ash in The United States, " C. E.
Brackett, Southern Electric Generating Company, Birmingham,
Alabama, March 13-14, 1973.
3. "Final Report Technical and Economic Factors Associated with Fly
Ash Utilization, " Aerospace Corporation Report No. TOR-0059(6781)-1,
El Segundo, Calif., July 26, 1971.
4. "Control of Sulfur Oxide Pollution from Power Plants, " Frank T.
Princiotta, Environmental Protection Agency, March 14, 1972.
5. "Removing SO, from Stack Gases, " A. V. Slack, Environmental
Science & Technology, Vol. 7, Number 2, February 1973.
6. "Test Program for the EPA Alkali Scrubbing Test Facility at the
Shawnee Power Plant, " M. Epstein, F. Princiotta, R. M. Sherwin,
L. Szeibert, and I. A. Raben, Presented at the Second International
Lime/Lime stone Wet Scrubbing Symposium, New Orleans, La. ,
November 8-12, 1971.
7. "The Mohave/Navajo Pilot Facility for Sulfur Dioxide Removal, "
J. L. Shapiro and W. L. Kuo, Presented at Second International
Lime/Lime stone Wet Scrubbing Symposium, New Orleans, La.,
November 8-12, 1971.
796
-------
8. "Use of Waste Sulfate on Transpo '72 Parking Lot, " Russell H.
Brink, Federal Highway Administration, Washington, D. C., Third
International Ash Utilization Svmoosium, March 13-14, 1973,
9. "Structural Compositions Prepared from Inorganic Waste Products, "
%
L. John Minnick, G. & W. H. Corson, Inc., Plymouth Meeting, Pa.,
Presented at the Annual Meeting of the American Association of
State Highway Officials, Miami Beach, Fla., December 5-10, 1971.
10. "Pilot Scale Up of Processes to Demonstrate Utilization of Pulverized
Coal Fly Ash Modified by the Addition of Limestone-Dolomite Sulfur
Dioxide Removal Additive, " Final Report Contract CPA 70-66, Coal
Research Bureau, West Virginia University, Morgantown, West
Virginia, October 1971.
11. "Multiple By-Product Utilization, " L. John Minnick, IU Conversion
Systems, Inc., Plymouth Meeting, Pa., Presented at Third Interna-
tional Ash Utilization Symposium, Pittsburgh, Pa., March 13-14, 1973.
12. "Potential Utilization of Solid Waste from Lime/Lime stone Wet-
Scrubbing of Flue Gases, ".Linda Z. Condry, Richard B. Muter, and
William F. Lawrence, Prepared for presentation before the Second
International Lime/Limestone Wet-Scrubbing Symposiu, New Orleans,
La., November 8-12, 1971.
13. "Properties of Power Plant Waste Sludges," Joseph G. Selmeczi,
and R. Gordon Knight, Dravo Corporation, Pittsburgh, Pa. -
March 13-14, 1973.
14. "Review of Current Research on Coal Ash in the United States, "
John F. Slonaker, and Joseph W. Leonard, Coal Research Bureau,
School of Mines, West Virginia University, Morgantown, West
Virginia, Presented at Third International Ash Utilization Sym-
posium (TIAUS), Pittsburgh, Pa., March 13-14, 1973.
797
-------
15. "Will County Unit 1 Limestone Wet Scrubber, " D. C. Gifford,
Commonwealth Edison Company, Chicago, Illinois, Presented at
Second International Lime/Limestone Wet Scrubbing Symposium,
New Orleans, La., November 8-12, 1971.
16. "Water Quality Criteria, " Second Edition, J. E. McKee and H. W.
Wolf, The Resources Agency of California State Water Resources
Control Board, 1963.
17. "Report of the Committee on Water Quality Criteria, " Federal Water
Pollution Control Administration,U. S. Department of the Interior,
April 1, 1968.
18. "Public Health Service Drinking Water Standards, " PHS Publication
956, dated 1962.
19. "Southwest Energy Study, " Report of the Coal Resources Work Group,
February 1972.
20. EPA Survey Data
21. "Will County Unit 1 - Limestone Wet Scrubber, " D. C. Gifford,
Presented at Electrical World Sulfur in Utility Fuels Conference,
Chicago, Illinois, October 25-26, 1972.
798
-------
EXPERIENCE IN THE DISPOSAL AND UTILIZATION
OF SLUDGE FROM LIME-LIMESTONE SCRUBBING PROCESSES
by
W. C. Taylor, Supervisor
Materials Management Research and Development
Kreisinger Development Laboratory
C-E Combustion Division
Combustion Engineering, Inc.
Windsor, Connecticut
799
-------
INTRODUCTION
The C-E waste disposal program was initiated in 1969
with the primary objective of providing ecologically safe
methods for the disposal of solid and liquid waste materials
discharged from the C-E air pollution control system. A
second objective was to investigate and develop beneficiation
methods capable of producing useful products that would offset
the disposal cost.
The program was formulated after considering a number
of factors relating to the feasibility of beneficiation
processes and its effect on the smooth operation of the power
plant. One of the major factors considered was the industry
experience in promoting the utilization of fly ash. Although
fuel ash was first used as an admixture to concrete in 1936
with favorable results, its utilization in 1971 amounted to
only about 201 of that produced. This low utilization persists
despite the formation of a national organization with the
single objective of promoting the use of ash from utility
boilers. This is in contrast to the relatively high utiliza-
tion in other industrialized countries, for example, Germany --
80%, France -- 65%, United Kingdom -- 55%, etc. Another
consideration was the desire of the utilities that any utiliza-
tion or disposal scheme could not interfere with their primary
role of producing power. A third consideration was the require-
ment that waste material from the APCS would have to be disposed
of within the ecological limits set by the Federal Environmental
Protection Agency as well as local control agencies.
800
-------
Using the above factors as major criteria, a program
was formulated to develop ecologically safe and economically
sound disposal and/or utilization methods for APCS sludge.
The program was divided into five phases and are discussed
below.
PHASE 1. THE ENVIRONMENTAL SURVEY
The Procedure
The addition of a limestone wet scrubbing air pollution
control system to a coal-fired steam generating unit more
than doubles the amount of solids (modified ash) to be
disposed of by the utility. In order to approximate this
effect on the ash handling and disposal capabilities presently
used and to gather other pertinent environmental data, a survey
of the present practices was made in September 1970. The
survey also gathered available data concerning pollution con-
trol limits of local regulatory agencies which were more
strict than Federal regulations.
Forty-two utilities covering the continental U. S.
were reviewed and from these 22 were selected to be surveyed.
The companies were chosen on the basis of:
a. Geographic location
b. Population center
c. System capacity
d. Type of coal
These companies account for about 301 of the electricity
generated by coal in continental U. S, A. Questionnaires were
delivered to the utilities by C-E sales personnel at which time
801
-------
an appointment was made for future discussions between
C-E's engineers and the utility's environmental engineer.
During this latter visit, the questionnaire was discussed
and filled out.
Data Analyses
General plant site data -- About 25% of the stations had
multiple means of fuel delivery. The data showed that railroads
were used 67.51,,barge 39.6%, trucks 141, and conveyor 7% for
the transportation of coal. Railroad delivery was not further
defined as to unit train or regular rail transportation.
Furnace type data were reported on 215 units which represented
83.2% of the units surveyed. Balanced draft units represented
71.7% and pressurized furnaces represented 10.7%.
Firing method data were improperly stated on 29 of the
units. The data indicated the following for the firing methods
reported:
Tangential 40.5%
Stoker 15.8$
Horizontal 8.41
Vertical 1.9*
Cyclone 4.71
Improper response
Pulverized coal 28.8%
Air data -- Air pollutant data (stack, wind, and emission
rates) were requested to permit determination of ground level
concentration. However, insufficient emission data were
obtained to determine concentration by calculations. It may
802
-------
have been more fruitful to request ground level concentrations
directly as well as emission rates. (One of the stations did
report ground level concentrations when emission rates were
requested.) It is significant to note that only a very few
stations reported NO emission rates. This indicates that
J\
very few stations were determining NO stack emission levels
J\.
at the time of the survey. (The regulations on NO emissions
X
had not yet been promulgated and may explain this minimal
response.)
The waste solids were transported from the dust collector
hopper to the disposal area by sluicing through pipes
by 74% or by trucking by 28%, respectively, for the
stations reporting. Approximately 24% of the stations were
planning future improvements in their equipment for control
of atmospheric emissions.
Land Waste Disposal Data Plant waste solids disposal rates
varied from 80 to 440 tons per day. The average solid waste
disposal rate of the 41 reporting plants was 800 tons per day
Fourteen plants reported sales of waste solids of from 1 to
741 of the ash produced. Actual sales tonnage based on all
41 reporting plants was 1200 tons per day of 33,000 tons produced.
This represents the sale of less than 4% of the ash produced.
Maximum daily sales for any plant was 400 tons per day.
Nearly 16% of the ash sales were made by stations using
plant acres for disposal. Over 841 of the sales were made
by plants using off-site disposal areas. Since plant site
solid waste disposal is used by 79% of the plants in the
survey it appeared that the economic pressure, cost of the off
803
-------
site disposal, may be a major factor in the need for developing
sales outlets for solid wastes.
Solid waste disposal costs were reported by 81% of the
plants in the survey. The cost varied from $0.03 to $1.10
per ton for sluicing compared to $0.11 to $1.33 per ton for
trucking the solid waste to an off-site disposal area. The
average sluicing rate cost for 25 plants was $0.45 per ton
while the average trucking rate for 10 plants was $0.51 per
ton.
Water Waste Data -- The boiler blowdown and demineralizer waste
disposal data obtained were not amenable to specific statistical
analysis. This was also true for cooling tower and ash sluicing
water data.
Once-through condenser cooling water temperature rise data
were obtained from 58% of the reporting power plants. The
average temperature rise was 16 F for cooling water flow rates
varying from 0.5 to 7 million gallons per day.
More than one-half of the utilities did not report the
composition of the waste water streams from their stations
This type of data had not been obtained, in most cases, at
the time of the survey.
Conclusions
The electric power industry's response to the C-E environmental
survey was very good with 86% of the canvassed companies com-
pleting the questionnaire. The high rate of response is a
direct indication of utility interest and concern with environ-
mental problems. As a result of the survey, a greater appreciation
for the complexity and diversity of problems associated with
804
-------
waste disposal was acquired by both C-E and the cooperating
utilities.
PHASE 2, THE CHARACTERIZATION OF APCS SLUDGE
Methods of Study
Since there was very little information available on
the physical, chemical, Land engineering properties of the
modified ash produced by the APCS, it was decided that this
should be one of the first areas to be investigated. Hence,
the second phase of the program was designed to completely
characterize the APCS sludge. The laboratory program initially
concentrated on the solids APCS discharge which was to be
followed by a study of the treatment methods for any waste
water stream associated with the system.
Properties relating to the permeability, solubility,
compactibility, drying, etc. of sludge are important considerations
in the direct disposal of this material. Thermal properties such
as melting points, decomposition temperatures, sinterability,
etc. are important in many areas where utilization as building
products are considered.
In order to supplement the classical analytical methods
and the X-ray fluorescence and diffraction spectrometers used
in the characterization of the sludge, a thermal analyzer was
obtained. The equipment shown in Fig. 1 is capable of performing
simultaneously thermogravimetric analysis CTGD), differential
thermal analysis (DTA), and differential thermogravimetric
analysis (DTG). Some of the engineering and rheological studies
805
-------
that require specialized equipment were contracted to labora-
tories specializing in establishing these properties.
Analytical Procedure
Initially, four sludges were selected as standards for
the characterization studies. The sludges were selected to
represent the material that would be produced by the different
types of air pollution control systems using the lime or limestone
scrubbing method. This number was eventually increased to
ten standard samples when the variations in the scrubbing
processes increased. The chemical composition of these ten
standards, identified in Table I, are listed in Table II.
X-ray analyses were also performed on these ten standards and
the results are listed in Table III. Each of the samples,
which weighed about 50 Ibs., were dried to a constant weight
at 100 F and stored for future analysis. Instructions were
also given to the contracting laboratories to treat the sludges
that they received in the same manner so that interlaboratory
data pertaining to a particular sludge could be compared.
Table IV gives the range of the specific gravity of the
as they varied with the water content. A summary of the
leaching studies using water at two pH levels is given in Table V.
Thermal Analysis
In order to accurately characterize the changes that
occur when the sludges are heated to various temperatures,
an extensive program of thermal analysis was set up. The
program started with the determination of the thermal proper-
ties of some pure compounds found in the APCS sludges. This
was followed by a study of the effect of fly ash on these
806
-------
thermal properties. The study was then extended to synthetic
mixtures of compounds formulated to simulate APCS sludge.
TGA curves of the known materials were used to identify and
analyze these components in the actual sludges. A summary
of the work on pure compounds and with fly ash added is
shown in Tables VI and VII and represented in Fig. 2, while
Table VIII compares the analytical results obtained with the
differential thermal analyzer to those found using wet chemistry
Figure 3 shows the rmogravime trie curve of some synthetic sludge
samples formulated in the lavatory, whiie Fig. 4 shows a differential
thermalanalysis curve of a representative sludge.
PHASE 3, DIRECT DISPOSAL METHODS
Based on the assumption that most utilities would
prefer the minimum modification of their present waste handling
practices, this program investigated methods of direct disposal
immediately after the characterization step had been completed.
Since direct disposal could take the form of either ponding
or recoverable land fill operations, C-E's study investigated
both of these.
Ponding Studies
The ponding studies began by investigating the permeability
of settled and compacted sludge to determine its ability to
act as its own sealant. This was followed up by a field study
that is currently in progress. Two consultant hydrologists
were interviewed and one was selected to draw up a test program
to determine the effect of a sludge pond on the surrounding
ground water. Oft_
807
-------
After preparation of the pond site and before the introduction
of the sludge, a series of 14 wells were placed around the
perimeter of a new APCS installation. Weekly samples were
taken from wells for two months prior to the introduction of
sludge and analyzed to establish the initial quality of the
water. Sampling will continue for at least one year and any
change and rate of change in the quality of the water will
be tabulated and analyzed. A diagram of the ponding site and
wells is shown in Fig. 5. A description of the wells is con-
tained in Table IX.
Land Fill Studies
Laboratory studies have indicated that in order to
produce a stable land fill, an air pollution control system
sludge would have to be dewatered to about 70% solids. Studies
have also indicated that conventional vacuum filters
probably will not dewater the sludge to that extent and other
equipment or procedures will have to be used. These other
procedures may involve the mixing of the sludge with dry fly
ash if enough of it is available to reduce the moisture con-
tent to the required level. Other equipment under consideration
includes a press type filter, various drying equipment, and
mixing apparatus. An extensive program involving laboratory
characterization, pilot plant testing, and full-scale field
demonstration of processes that would produce a solid suitable
for land fill application is planned. Figure 6 shows one such
scheme.
808
-------
Road Construction
In addition to the work with stable land fill material,
an extensive program dealing with the formulation and testing
of sulfate sludge for use in road construction is being carried
out. During these studies, APCS sludges were dried and compacted
into test cylinders that have reached compressive strengths
of 5,000 psia after 28 days of curing. Other sludges, however,
have exhibited compressive strengths of only 250 psi.
Another phase of the study on the use of APCS sludge as
a highway construction material was C-E's participation in
a project at Dulles Airport in Washington D. C. This project
is described in a later section.
PHASE 4, BENEFICIATION METHODS
FOR SLUDGE UTILIZATION
Procedure
Consideration was given to the possibility that in some
metropolitan areas a combination of high disposal costs and
a shortage of raw material for which APCS sludge could be sub-
stituted might lead to a situation where it would be economical
to develop some beneficiation process that would lead to the
development of products. We believe that due to the volume
of the sludge involved, any by-product utilization schemes
would be similar to those with the normal fly ash. Further,
it was believed advantageous to the disposal and/or utilization
of both the modified and unmodified ash if a systematic inves-
tigation was made of where the APCS sludge could be substituted
for fly ash and where the use of sludge would be detrimental
809
-------
to areas where fly ash is being used. In this study, data
collected by other organizations, such as the Bureau of Mines
and the West Virginia Coal Research Bureau, were considered.
Table X is a partial list of the areas where investigations
in varying degrees were made for the utilization of the APCS
sludge.
The Most Promising Schemes
A number of the more promising utilization methods were
investigated either in our Kreisinger Development Laboratory
or by laboratories already working with fly ash in these areas.
Figure 7 and 8 show some of these processes. A number of
products made from the fly ash involve sintering. Since
laboratory studies have shown that with APCS sludge sintering
is accompanied by the decomposition of the calcium sulfate
and subsequent release of SC>2, a study of the economics of
the joint process involving the sintered product and sulfur
recovery was made by two sulfur-producing companies. One
of these companies developed processes for the production
of sulfur from Gypsum. Even while the study was in progress,
the price of sulfur changed from about $40 per ton to about
$10 per ton making any process depending on the sale of
this material uneconomical at this time.
LARGE SCALE DEMONSTRATION STUDIES
C-E participated in a project with the Research and
Development Division of the Federal Highway Department to study
the use of sludge as a highway construction material. The
program was part of the U. S. Governments' International
Transportation Exposition (Transpo '72) that was held from
810
-------
May 24 to June 4, 1972 at the Dulles Airport in Washington, D. C.
In order to accommodate the 50,000 automobiles and 600
buses expected daily at the exposition, a 120-acre parking lot
was constructed. The Research and Development Division of
the Federal Highway Administration used this opportunity
to demonstrate the recycling of waste products as highway
construction materials. The major porti-pn of the parking
lot was paved with a mixture of fly ash and sulfate sludge
from an acid plant. Other portions of the parking lot were
paved wtih waste products such as air pollution control system
sludge and acid mine drainage sludge.
Combustion Engineering transported 75 tons of sludge from
Lawrence, Kansas to the Dulles site - a distance of about 1300
miles. Since this was the first time that APCS sludge had
been transported by any vehicle, the project was designed to
obtain the maximum amount of information on possible materials
handling problems. The sludge was dredged from a pond at
Kansas Power and Light where it had been stored for 6 months.
It was allowed to drain for 24 hours and loaded on flat and
round-bottom trucks for the non-stop trip. Two C-E engineers
monitored the sludge during the entire trip taking measurements
and samples for laboratory analyses at various intervals.
In general, there was no problem of sludge leaking from
the trucks while they were on the road. Excess water did leak
from the track tailgate during the trip. At Dulles, the
sludge was removed from the truck.
811
-------
It was considerably easier to unload the round bottom
trailers than it was those with square beds. All of the sludge
slid out readily from the former while the latter required
manipulation with the backhoe to complete removal of the sludge.
The APCS sludge was slurried with water, mixed with fly
ash and lime, and placed in a test section of the parking lot.
Evaluation of the suitability of APCS sludge as a highway
construction and repair material is continuing. In a recent
report to the third International ash association, the
highway administration indicated that so far the performance
of the material was satisfactory. An analysis of the densities
of the sludges at the bottom of the truck is given in Table XI.
Figure 9 shows the loading, hauling, and unloading of the sludge
material.
CURRENT STATUS
At present, C-E is applying its efforts to the area of
direct disposal. Beneficiation methods and studies are limited
to those that are necessary to dewater the slurry from 10%
solids to the 70% for land fill operation. To support this
study, APCS sludge in 200-gallon lots has been shipped to
the laboratories of manufacturers of commercial drying equipment
for tests. One spray drying test has proved to be very satis-
factory.
812
-------
The more we get involved in the study of waste disposal
the more we learn the enormity of the problem and the effort
that must be applied before a solution is reached. A portion
of the waste disposal dilemma is a result of our attempt to
supply the nation with energy without an adverse effect on our
environment. C-E is not in the waste disposal business and
has nothing to gain from the sale of any products produced from
the sludge or from the various contracts for the sludge handling
that the utility may enter. Utilities, suppliers, and the
federal government should share the expense of waste disposal
utilization studies.
813
-------
TABLE I
IDENTIFICATION OF APCS SLUDGE STANDARDS
STD I -- Flyash from Connecticut Light and Power Company's
Devon Station
STD II -- C-E sludge - CaC03, 1501 stoichiometry, 2000 ppm S02
STD III -- Kansas Power and Light sludge
STD IV -- C-E sludge - Ca(OH)2, 38% to 50% stoichiometry,
50 to 60% SO, removal, slurry feed 220 gpm, recycle
165 gpm with 55 gpm blowdown
STD V -- Union Electric sludge
STD VI -- C-E sludge - CaC03, 150% stoichiometry, 45 to 55%
removal, no recycle
STD VIA -- STD VI plus 50% STD I (flyash)
STD VII -- C-E sludge - 300 to 325% stoichiometry, 64% S02
removal, 300 Ib/hr flyash, 550 Ib/hr CaC03
STD VIII -- C-E sludge - 120 to 130% stoichiometry Ca(OH)2,
90.8% removal, 120 gpm (Ca(OH)2 slurry underbed,
inlet S02 860 to 840 ppm, outlet S02 80 ppm,
145 Ib/hr Ca(OH)2, no flyash addition
STD IX -- C-E sludge - 220 gpm H90 spray, 275 Ib/hr lime
£*
feed, 300 F reaction temperature
814
-------
TABLE II
WET CHEMICAL ANALYSIS OF SLUDGE STANDARDS
STD I STD II STD III STD IV STD V STD VI STD VIA STD VII STD VIII STD IX
sio2
A12°3
Fe2°3
CaO
MgO
Na.O
K20
Ti°2
P2°5
C02
so2
so3
CaCO,
46.7
23.2
13.7
4.7
0.9
0.3
2.6
1.5
0.3
2.6
0.8
5.9
1.5
0.32
0.27
49.6
0.54
0.04
0.17
<0.02
O.OS
29.2
11.7
3.5
65.7
30.7
6.6
8.6
22.7
1.5
0.50
1.1
0.26
0.11
5.3
5.8
6.5
12.0
0.79
0.05
0.18
42.5
0.10
0.03
0.05
<0.02
0.06
3.7
38.8
3.3
8.4
19.4
6.8
5.4
27.6
3.2
0.08
0.24
0.32
0.08
7.2
2.2
12.3
16.3
1.1
0.01
0.09
52.5
0.52
0.02
0.14
<0.02
0.13
36.6
6.3
0.5
80.6
27.7
14.7
8.3
24.2
0.70
0.16
1.2
0.79
0.19
15.3
3.4
<0.1
34.7
4.6
2.3
1.6
40.1
0.20
0.05
0.29
0.11
0.08
13.6
S.4
24. P
30.9
1.2
0.48
0. 72
42.5
0.90
0.05
0.07
<0.02
0.06
11.5
24.1
8.4
26.1
2.0
0.45
0.72
46.2
0.40
0.04
0.21
<0.02
0.07
24.4
13.7
4.4
55.4
-------
TABLE III
X-RAY ANALYSES OF APCS SLUDGES
III
IV
IV
VIA
VII
VIII
IX
Major
sio2
3Al203-2Si02
sio2
2CaS03'H20
Si02
CaC03
CaC03
CaCOj
2CaS03'H20
2CaS03-H20
CaCOj
Minor Trace
Fe2°3 CaC03
Fe3°4
CaS04
Fe304 3A1203
CaC03 Ca(OH)
2CaS03'H20 CaS04
MgO
Si02
CaS04-2H20 Ca(OH)
Fe3°4
2CaS03 H20 CaS04
Si02
Si02 2CaS03
2A1203-2S102 CaS04
CaC03 2CaS03
Si02
CaC03 Ca(OH)
Si02
i
2
2
' 2Si(
'H2°
'H2°
2
2CaS03-H20
Si02
CaS04'2R20
Fe2°3
Ca(OH)
-------
00
TABLE IV
THE VARIATION OF THE SPECIFIC GRAVITIES OF STANDARD SLUDGE WITH WATER CONTENT
STANDARD
H20
0
5
10
20
30
40
50
60
70
80
90
I
2.28
2.10
2.06
1.83
1.62
1.50
1.38
1.29
1.21
1.13
1.06
II
2.64
2.27
2.09
1.91
1.70
1.54
1.43
1.32
1.22
1.14
1.07
III
2.18
2.16
2.06
1.92
1.S8
1.47
1.40
1.28
1.21
1.12
1.06
IV
2.48
2.00
1.91
1.72
1.51
1.44
1.34
1.25
1.18
1.13
1.07
V
2.18
1.86
1.80
1.81
1.62
1.44
1.38
1.28
1.18
1.13
1.06
VI
2.62
2.41
2.27
2.00
1.78
1.57
1.43
1.32
1.22
1.14
1.06
VIA
2.33
2.21
2.09
1.92
1.66
1 54
1.43
1.30
1.21
1.12
1.07
VII
2.62
2.42
2.22
1.94
1.84
1.56
1.42
1.31
1.25
1.12
1.07
VIII
2.34
2.20
1.98
1.86
1.70
1.57
1.37
1.31
1.20
1.14
1.06
2
2
2
1
1
1
1
1
1
1
1
IX
.19
.19
.01
.84
.65
.48
.40
.30
.21
.13
.07
-------
TABLE V
* IEACHING PROPERTIES OF APCS SLUDGE STANDARDS
00
STD
I
II
III
IV
V
VI
VIA
VII
VIII
IX
Wt.
Loss,
g
0.8178
0.5515
3.2688
0.6599
4.6063
0.5258
0.5491
4.3096
1.9332
0.9464
*Air
pH
% Wt.
Loss
2.64
2.94
16.21
6.53
33.60
2.95
2.36
23. 77
11.91
6.55
5.90
Vol-
ume .
ml
6400
6250
5825
6810
8915
6870
6650
6890
8330
6655
Loss,
g/ml
0.127 x 10"3
0.088 x 10"3
0.561 x 10'3
0.097 x 10~3
0.516 x 10'3
0.077 x IO"3
0.083 x 10~3
0.625 x IO"3
0.232 x IO"3
0.142 x IO"3
dried uncompacted materials
Wt.
Loss,
g
0.2525
0.5821
1.8784
0.8218
3.1613
0.6936
0.3009
4.1662
1.3071
1.1391
PH
% Wt.
Loss
1.06
3.71
10.54
8.68
27.92
3.64
1.59
30.01
9.90
7.93
4.15
Volume ,
ml
6495
6750
6885
5175
6625
7295
5095
6315
7285
7155
Loss
g/ml
0.038 x
0.086 x
0.272 x
0.158 x
0.477 x
0.095 x
0.059 x
0.659 x
0.179 x
0.159 x
•
ID'3
ID'3
ID'3
ID'3
!0-3
io-3
io-3
io-3
io-3
io-3
-------
TABLE VI
THERMAL ANALYSIS OF SOME PURE COMOPUNDS FOUND IN APCS SLUDGE
CO
_4
10
Temperature and
Events
Compounds
CaS04 • 2H20
CaS04 - 1/2H20
CaSO,
0-500 C
Dehydrations,
Lit.
20.9
6.2
Exper.
18.7
3.5
500-700 C
Oxidation of
? Weight Gain
Lit.
Exper.
700-900 C
Decomposition of
Limestone, % C0~
1200-1400 C
Decomposition of
Calcium Sulfate,
% SOT
Lit. Exper. Lit.
46.5
55.2
58.8
Exper,
46.7
48.2
CaS03 • 1/2H20 7.0
CaCO,
6.8
13.3
5.0
44.0
43.5
62.0
64.2
Ca(OH)2
24.3
22.5
-------
TABLE VII
DECOMPOSITION TEMPERATURE OF COMPOUNDS WITH AND WITHOUT FLY ASH
COMPOUND
Without
Literature
CaS04 • 2H20 128
CaS04 • 1/2 H20 162
00
K>
o
CaS04 1200
CaSO, • 1/2 H,0 400
•j £
CaC03 825
Ca(OH)- 580
Decomposition Temperature, C
Flay Ash
Experimental
130
145
1400
400
730
460
With Fly
Experime
130
145
1200
400
730
460
-------
TABLE VIII
ANALYTICAL RESULTS OBTAINED BY WET CHEMISTRY AND THERMAL ANALYSIS
Total SO.
Equivalent
Calcium Sulfate
CO,
Chemical
Standard I
III
CO
V
VI
VII
VIII
IX
0.
12.
13.
5.
32.
39.
24.
8
5
5
1
3
7
2
Thermal
1
14
14
6
34
29
22
.9
.5
.7
.5
.4
*
.3
.8
Chemical
1
21
22
8
55
65
41
.4
.1
.8
.6
.0
.7
.2
Thermal
2
27
27
10
58
*
49
38
.9
.0
.5
. 2
.4
.3
.7
Chemcial Thermal
2
4
4
34.6 34
13.5 11
8.8 6
23.8 21
.6
.1
.7
.9
.3
.2
.9
Equivalent
Calcium Carbonate
Chemical Thermal
5.9
9.1
79.5
30.8
54.3
80.0
25.8
14.2
49.82
Did not reach maximum temperature during the thermal analysis
-------
TABLE IX
SAMPLING WELLS AT THE FULL SCALE PONDING STUDY SITE
Well Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Depth*
(Feet)
28
32
29
34
23
29
27
31
27
30
28
32
28
32
Water Level*
(Feet)
23
23.
24.
24
20
20
23.
23
22
19.
23
23
25
24.
2
8
4
8
5
* These values include 2 feet of pipe above the ground level
822
-------
TABLE X
UTILIZATION AREAS FOR FLY ASH AND MODIFIED ASH
1. Airport pavement mixture
2. Asphaltic filler and wear-course aggregate
3. Cement manufacture
4. Concrete admixture for construction
5. Grouting agent in wells
6. Filler for glass
7. Filler for fertilizer
8. Filler for paint
9. Filler for plastic
10. Filler for rubber
11. Fill material for land recovery
12. Fill material in abandoned mines for fire control
13. Fill material in abandoned mines for subsidance control
14. Neutralization of acid mine drainage
15. Manufacture of sinter bricks
16. Manufacture of autoclaved bricks
17. Manufacture of light weight aggregate
18. Manufacture of mineral aggregate
19. Production of cenosphere
20. Soil amendment
21. Soil stabilizer for road base, dam, embankments, etc.
22. Recovery of gypsum
23. Recovery of sulfur
24. Sealant in sanitary land fill operation
25. Sludge dewatering
26. Reclamation of polluted lakes
27. Sand blasting
28. Production of mineral wool
29. Recovery of magnesium oxide
30. Recovery of calcium oxide
823
-------
TABLE XI
DENSITY MEASUREMENT MADE ON THE SLUDGE IN TRANSPORT
Miles
0
5-1/2 (Weigh Station)
11 (KPL-Weigh
Station Round
Trip)
(Kansas Tpke
Toll Station)
35
103
214
431
657
931
1375
(Between K.C.
§ Mexico, Mo.)
(Mexico, Mo.)
(Effingham, 111.)
(Richmond, Ind.)
(Wheeling, W. Va.)
(Dullas Airport)
Dike Truck 1
51.5
54.6
54.2
I Solids [All readings for bottom
of bed except as noted]
Truck 2 Truck 3
,54.3 top
156.8 bottom
57.1
54.2
54.2
55.4
56.2
47.8
42.7
64.2
48.4 top
53.1 bottom
55.8
56.7
55.4
53.6
60.9
824
-------
at
•
ig- 1: Simultaneous analyses equipment
-------
00
to
150
120
110
100
9(X
8o
70
60
50
Decomposition Of Calcium Sulfate In
Sludges Without Fly Ash
Decomposition Of CaS04 In Sludge With Fly Ash
Phase Change In Calcium Sulfate
CaS04
CaC •*• S02 + 1/2 02
Decomposition Of Calcium Carbonate CaCCs
Phase Change In The Calcium Oxide (Endothermic)
\ Slight Wt. Gain Due To Oxidation Of CaSOa
CaO 4
J
30C _
20C .
100
Phase Change In The Calcium Carbonate (Endothermic)
, ^composition Of Calcium Hydroxide Ca(CH)s —y CaC •+ H20
Decomposition Of Calcium Sulfite
CaS03 1/2H2C
1/2 H20
Decomposition Of The Hemihydrate CaS°4 1//2 Ha°
Decomposition Of Gypsu-n CaSc4 2H2°—>
Slight Wt. Loss Miy Occur Due To Loss Of Surface Water
1/2 H20
1/2 H£0 + 1 1/2
j Slight Wt. Gain May Occur Due To Oxidation Of Some Components In The Sample
THERMAL EVENTS TTJ A TYPICV. APCr i-MJDGE ?..Wi\£
Fig. 2: Summary of the work on pure compounds with fly ash added
-------
oo
10
CO 20
o
X
O 30
Ul
40
50
a
tr
>•
••••••••• »^
/b
\.
1
/c
..— .y
••••••^L«»t
i.
•
•
•
•••••••••«*
— — 33%CaC03l33%Ca(OH)2,33%CaSCV2-H2O
^— 50%FLYASH,IO%CaS042H20,30%CoS09,IO*CaCO,
••••••50%FLYASH,25%CaSO42H20,25%CaS{V2-H20
— 50%FLYASH, 30%CaS042H20,IO%CaS03,IO%CaC03
-.. ,d
— ,\
A
Vs
*
K
•
•
•
•
•
\-r
s^
\
•*-**>
X
^
200
400
600 800 1000
TEMPERATURE-°C
1200
1400
160O
Fig. 3: Thermogravimetric curve of some synthetic sludge samples
-------
00
NJ
oo
2OO
400
600 800 IOOO I2OO
TEMPERATURE-°C
1400
1600
1800
Fig. 4: Differential th^rmalanalysis curve of representative sludge
-------
EVEN NUMBERED WELLS ARE DEEP, ODD NUMBERED WELLS ARE SHALLOW
DISTANCES ARE APPROXIMATE IN FEET
CO
KJ
350'
8 -*•
NORTH
POWER PLANT
Fig 5: Ponding site and wells diagram
-------
* /ROTARYX
ARCS
SCRUBBER
ROTARY DRYER
STORAGE
HOPPER
TRANS PORT
SYSTEM
LAND FILL AREA
ROAD BASE
Fig. 6: Scheme for producing suitable land fill
-------
CO
u>
WATER TO
* CLARIFIER™^
|
WET 1
m-£±-*\ VACUUM FILTER
SLUDGE ]
-»
•*
+
-»
^^
^^^
nci i CT17PD
rcLLcTlZcK
CEMENT K ILN
DRYING KILN
UIVCB
IWIAtfl
IMI^H
•»sc
—
_J
••w
>2^
•*
AIITOn AVF
SINTERING
MACHINE
+S02
S02
PROCESSING
FACILITY
SOzf
ELECTRIC
FURNACE
KILN
—
"""*
— *•
MINERAL
AGGREGATE
LIGHT
WEIGHT
AGGREGATE
SULFURIC ACID
OR SULFUR
MINERAL WOOL
ADDITIVE FOR
ASPHALT
1 AMD Pll L
Fig. 7: Sludge disposal schemes for a single scrubber system
-------
CD
U>
to
CALCIUM CARBONATE
S02
REMOVAL
^ CALCIUM
W CALCIUM CA
L >•
•^WATERHH
ICOO BACK TO SCRUBBER
^^•^
^^^^
1
1
•^
SINTERING
MACHINE
POZZOLAN
FOR
CONCRETE
t
A
C02
Kl
700 °C
1292 °F
S02
PROCESSING
FACILITY
s|2
.N
1200 °C
2192 °F
1
LIGHT
U/PI ftUT
wc.1 \yr\ \
AGGREGATE
MINE
LANDFILL DRAINAGE
NEUTRALIZATION
t t
Fig. 8: Sludge disposa^. schemes for a dual scrubber system
-------
•
(a)
Cb)
(d)
Ce)
Fig. 9: Sludge material handling
833
-------
FIXATION AND DISPOSAL OF FLUE GAS
WASTE PRODUCTS:
TECHNICAL AND ECONOMIC ASSESSMENT
by
L. John Minnick, Executive Vice President
IU Conversion Systems, Inc.
1500 Walnut Street
Philadelphia, Pennsylvania
835
-------
FIXATION AND DISPOSAL OF FLUE GAS
WASTE PRODUCTS -
TECHNICAL AND ECONOMIC ASSESSMENT
L. John Minnick-/
The three primary by-products produced in the burning of
coal in modern power plants consist of fly ash, bottom ash,
and sulfur oxides. Over the years considerable technology has
emerged with respect to the disposal and utilization of fly
ash and bottom ash, and a number of commercially viable pro-
cesses are available. Unfortunately, however, the large
tonnages require that the major portion of these ash constit-
uents must be sent to dumps, which presents a disposal prob-
lem to the power industry. As the industry complies with the
regulations relating to control of sulfur oxide emissions, a
similar situation will exist if an attempt is made to reclaim
large tonnages of sulfur. Opportunities for utilization will,
of course, be available: some sulfuric acid can be produced
and sold, and some elemental sulfur or liquid S02 may find a
market, but the majority of the sulfur will unquestionably be
faced with a lack of marketing opportunities and therefore
may have to be discharged as waste to disposal sites.
The lime and limestone scrubbing processes which are
JL/ Executive Vice President, IU Conversion Systems, Inc.,
1500 Walnut Street, Philadelphia, Pa.
836
-------
receiving considerable attention at "the present time result
in a by-product slurry or sludge and because of the lack of
any commercial market for this material, it is necessary to
transport this waste to disposal sites. Therefore these
scrubber operations are currently referred to as "throwaway"
systems.
The prime purpose of this paper is to outline the oppor-
tunities that are available to convert power plant waste mate-
rials through suitable processing steps to stabilized struc-
tural fill material or, if desired, to usable manufactured by-
products. This concept converts the "throwaway" waste into
products that can be placed or used and recycled in an ecologi-
cally acceptable manner.
The main concept of the system is based on modifications
of existing technology dealing with the stabilization or bene-
ficiation of fly ash. To some degree fly ash is being used in
the construction industry as an additive to Portland cement,
lime, and asphaltic compositions. Millions of tons have also
been used throughout the world in the construction of struc-
tural embankments, reservoirs, dams, road base materials and
the like. Large scale utilization programs have been carried
out by governmental agencies in England, France, and by a
number of organizations in the United States; for example,
National Ash Association, University of West Virginia, and
837
-------
by private companies such as Chicago Fly Ash Company and
G. & W. H. Corson, Inc.
Several years ago the Corson company discovered that the
addition of sulfur oxide salts beneficially improved the struc-
tural properties of stabilized fly ash. While the original work
was done with waste sulfate sludges from the chemical and coal
industry (from the neutralization of acid mine drainage), it was
also determined that the underflow sludges from lime and lime-
stone desulfurization scrubbers worked equally well. The net
effect of this activity has been to provide the coal burning
plants with a practical and economical tail end conversion pro-
cess which includes the sulfur oxide sludges as an important
and integral part of stabilized structural materials.
DESCRIPTION OF THE CONVERSION SYSTEM
Basically the tail end process is a modification of
Poz-0-Pac®, which has been used for many years in structural
applications such as road bases, reservoirs, dams, and embank-
ments. The basic chemistry involved in Poz-O-Pac® is a com-
bination of (1) the pozzolanic reaction that occurs between
the silica content of the fly ash and hydrated lime, which is
added in small amounts (2 to 3% by weight), to form cementi-
tious hydrated calcium silicate compounds; (2) the more rapid
reactions that occur between the soluble salts present in fly
ash with lime and the alumina that is found in the fly ash
838
-------
glass; and (3) the mechanical support derived from the addi-
tion of various aggregate materials which serve as bulk fillers
in the fly ash-lime matrix. Of considerable importance in the
design of Poz-O-Pac® is the water content of the mixture. In
fact, whenever fly ash is used as a structural composition,
it is essential that optimum moisture contents be utilized and,
in addition, it is important to specify adequate compaction when
the material is placed.
In contrast to Poz-O-Pac®, the stabilization of fly ash
using sulfur oxide sludge - commercially referred to as the
Poz-0-Tec* process - is based on additions of large quantities
of calcium and/or magnesium sulfite, calcium sulfate, and, in
certain instances, magnesium oxide which results when dolomitic
lime is used in the treatment of the sulfur oxide acids or
waste gases. The presence of the sulfites and sulfates are
beneficial to the cementitious reactions and, in addition,
allow the reactions to proceed at a much more rapid rate. The
engineering properties of the end products are improved to the
point where it is possible to place the stabilized fill compo-
sitions without the addition of aggregate. This does not mean
that aggregate cannot be used; in fact, Poz-0-Tec* mixtures
have been produced with various types of aggregate including
bottom ash from coal burning power plants, by-product slags,
*A service mark owned by IU Conversion Systems, Inc.
839
-------
culm, or mine refuse materials, as well as conventional con-
struction aggregates. Work has also been carried out with
these mixtures to encapsulate potentially toxic materials such
as crushed storage battery cases.
Much of the chemistry involved in the Poz-0-Tec* process
can be compared with reactions which have been described many
times in the literature by investigators dealing with portland
cement. Calcium sulfate reacts preferentially with calcium
aluminates to form hydrated calcium sulfoaluminates. Hydrated
calcium sulfoferrites, analogues of the sulfoaluminates, per-
form equally well in the cementing process. The Poz-0-Tec*
process involves a special and inventive application of these
established chemical concepts used by the portland cement
industry, although the conditions under which these reactions
proceed and the ultimate performance criteria are different.
Of considerable interest to the utilities is the effect
of sulfite ions in the mixtures since calcium sulfite is pro-
duced in large quantities in the scrubber slurries. The pres-
ence of sulfite provides several advantages in the conversion
process, an important example being the acceleration of the
cementitious reactions. Figure 1 shows scanning electron
micrographs which illustrate visually the characteristic of
a fly ash sulfur oxide sludge mixture as freshly prepared and
a similar composition after processing and aging for a period
840
-------
of 6 months. The original platelets that are made up primar-
ily of calcium sulfite (or sulfate) are substantially depleted
after curing. The crystalline material which is formed assists
in the hardening or cementation process. Figure 2 shows X-ray
diffraction profiles and demonstrates the rapid rate with which
available lime is used up in these compositions during the
aging process.
Figure 3 shows a series of patterns developed using an
electron probe scan of a fly ash-calcium sulfite-lime mixture
which has been cured at room temperature for 10 days. The
presence of aluminum in the scan indicates early formation of
calcium sulfoaluminate crystals which are developing within
the calcium sulfite platelets.
Figure 4 is a transmission electron micrograph showing
the hydrated calcium silicate structures (tobermorite-like
material) that results in both Poz-O-Pac® and Poz-O-Tec*
compositions. This cementitious product contributes sub-
stantially to the ultimate strength and durability of the
structure.
Figure 5 is a flow diagram of the tail end conversion
process. In the processing of the fly ash and sulfur oxide
sludge, it is usually necessary to modify the water content
of the discharge from the power plant which in some cases
can be a fairly major operation. As indicated in the sche-
-------
matic drawing, the fly ash may be added into the primary de-
watering vessel as a wet slurry or, if the fly ash is dis-
charged from an electrostatic precipitator, it can be added
in a dry or damp state, thereby assisting in the dewatering
process. Depending on the type of scrubber that is used, and
the particular discharge system which is adopted, dewatering
may or may not require a secondary state (usually a vacuum
filtering step). Dewatering devices of this type have been
used conventionally in many large scale applications related
to the coal and steel industries.
The mixing and conditioning step is basically a modifi-
cation of a conventional Poz-O-Pac® plant. Figure 6 shows one
of these plants of recent design. This plant has a production
capacity of 600 tons per hour. After applying minor changes
to a plant of this type, the fly ash and sulfur oxide sludge
can be supplied to the mixer and adjusted in composition with
make-up additives which may include additional lime, limestone,
fly ash, bottom ash, other sulfur oxide salts, and, if desired,
aggregate or other waste products.
As indicated in the flow sheet, Sulf-O-Poz®, suitable for
placement as a stabilized fill material, is discharged from the
mixing and conditioning device.
Although Sulf-0-Poz® is primarily a disposal material, it
has certain minimal structural qualities which would allow its
842
-------
application in some cases as structural embankments, land rec-
lamation projects, liners for reservoirs, etc.
As also shown in the flow sheet, Sulf-0-Poz® can be
further processed to produce other compositions such as syn-
thetic aggregate which are suitable for use in producing struc-
tural shapes, Portland cement or asphaltic concrete, or may be
supplied directly for use as a high strength monolithic road
base composition equivalent to stabilized road base required
in "primary" highways, airport runways, trucking terminals,
and the like. This supplementary processing is of interest
since it offers opportunity to convert the power plant waste
material from a "throwaway" to a "layaway" disposal system.
In this connection aggregate can be stockpiled and made avail-
able for commercial projects as they develop in the future.
It has been well publicized that shortages of aggregates are
already of serious concern in many areas of the country.
Poz-0-Tec* aggregate is therefore one approach to solving
this critical problem.
PROPERTIES OF THE CONVERSION PRODUCTS
The two areas in establishing criteria for products pro-
duced in the conversion process consist of, first, the eco-
logical, and, secondly, the engineering properties of the
materials. Both of these properties relate to the mix design
and are involved with the ratios of fly ash to sulfur oxide
sludge, concentrations of lime and water, and also to the
843
-------
basic chemical properties of the fly ash and lime. For ex-
ample, fly ash made from lignite coal which contains high CaO
works quite differently than fly ash produced from bituminous
coal. Underflows from limestone scrubbers present substan-
tially different design requirements than underflows from
hydrated lime systems.
Ecological Properties
The evaluation of the various compositions as related to
effect on the environment is dependent primarily on measurements
of permeability and leachability, whether the compositions are
simply placed as a stabilized fill or used as a construction
material. Where massive quantities of Sulf-0-Poz® are placed
as a stabilized fill, land improvement, or road base construc-
tion project, the specific ecological concern is with perme-
ability measurements and possibly with surface runoff.
Table 1 includes some results of permeability tests made on
compositions associated with the recent Transpo 72 project at
Dulles International Airport using standard falling head perme-
ability procedures. Figure 7 is a graph that shows the reduc-
tion in permeability in sulfur oxide containing mixtures when
compared with those that have been previously made with stan-
dard Poz-O-Pac® compositions. Permeability values of freshly
-5
prepared compositions usually run in the range of 10 cm/sec,
which quickly changes during the hardening process to values
844
-------
•~ ~
ranging from 10 to 10 cm/sec. Obviously a road base or
embankment with this low permeability provides excellent
protection against leaching of deleterious materials through
the mass.
From the standpoint of making leaching measurements, while
no "standard" tests are available, it is recognized that all
materials can be subjected to special laboratory procedures
that will exhibit the presence of small quantities of soluble
salts. Accelerated tests run on specimens weighing 500 grams
which were placed in plastic containers covered with two
liters of distilled water then shaken continuously for 48
hours are given in Table 2. The table includes limits for
potable water as described in U. S. Health Service, HEW
Bulletin No. 956 on Drinking Water Standards. Figure 8 is a
photograph of the equipment used in the shaking test. For the
sake of comparison, materials other than stabilized fly ash
compositions are included in this table. It will be noted
that raw, loose materials are subject to substantially higher
quantities of leachate in a test of this type. This obser-
vation is a further indication of the beneficial effect of
forming a monolythic structural Sulf-0-Poz® mass.
The question of leaching of heavy metals from surface run-
off is one that has been studied extensively by the United States
Potters Association and the International Lead Zinc Research
845
-------
Organization, Inc. The object of these studies was to eval-
uate the effect of temperature, time, and pH on the amount of
heavy metal that could be leached from ceramic surfaces.
Recent studies at Rutgers University School of Ceramics using
a scanning electron probe has shown the pattern of lead con-
centration across the surface of the ceramic ware. Figure 9
is an example of this type of investigation. Careful exami-
nation of the lead profile shows that leaching has occurred
to a depth of approximately 1/2 micron. Techniques of this
type may be of interest for stabilized fly ash materials since
this affords a good method of determining the total impact of
consolidated masses on the environment. It is evident from
laboratory leaching tests of the type described that stabilized
structures can provide virtually non-significant quantities of
leachable ions to the environment.
Runoff tests have also been carried out by preparing slabs
of the fly ash-sulfur oxide compositions. Two liters of dis-
tilled water were poured over the surface of the specimens to
provide a runoff that could be tested by atomic absorption
techniques. Table 3 shows some of the results of these tests.
Figure 10 illustrates the technique that has been used.
It is beyond the scope of this paper to make detailed
recommendations as to the most appropriate method of test for
these stabilized compositions or to suggest absolute limits
846
-------
that might be acceptable from an environmental standpoint.
However, the values that have been obtained fall well within
limits that have been established by a number of governmental
organizations which have addressed themselves to this type of
situation.
The Poz-0-Tec* compositions that have been studied not
only meet rigorous environmental requirements, but due to the
high pH of these compositions they, in fact, can serve to
"protect" the environment from attack. It has been noted,
for example, that when spoil bank materials which contain
pyrites are encapsulated within Poz-0-Tec* mixtures that
leaching of sulfuric acid is avoided and the attack on vege-
tation in the spoil bank area is therefore eliminated. This
type of chemical encapsulation is a phenomenon analogous to
studies that have been made in Portland cement compositions.
The technical literature provides numerous references which
describe the method by which soluble components migrate into
the lattice complexes of a cement structure becoming chemi-
cally bound, thus preventing the leaching of soluble ions.
A well known example of this is the lime-alkali-silica complex
resulting from the solid state diffusion of soluble NaOH or
KOH into the lime hydrous silica complex. When in this form,
the Na+ and K+ ions are chemically entrapped thus preventing
potentially harmful leaching.
847
-------
Engineering Properties
The most important engineering factor which must be eval-
uated is the strength of the reacted mixtures. A correlation
exists between the formation of cementitious compounds of the
types described above and the development of strength and other
engineering properties in fly ash-sulfur oxide mixtures.
Figure 11 shows the effect of aging on the strength of a
Sulf-O-Poz® composition. The method of measuring penetration
resistance is based on a modified vicat procedure, the pene-
tration needle having a cross section area of one-fortieth of
a square inch. The measured penetration resistance represents
the pressure that must be applied in pounds per square inch to
cause a penetration of one inch.
Figure 12 provides information on a similar composition
when cast into 2 x 2 x 2" cubes and tested for unconfined
compressive strength and Figure 13 presents California Bearing
Ratio values. Based on results of this type, it is possible
to design compositions with strength properties commensurate
with intended applications.
Table 1 presents compressive strength tests that were made
for design purposes at Transpo 72. Table 4 shows results that
were obtained from a field project and is based on cores that
were removed after periods of 4 and 6 weeks.
The major criteria which relate to the ultimate strength
848
-------
of these mixtures are composition, moisture content, and com-
pacted density. Where a Sulf-O-Poz® fill or a land rehabil-
itation project is contemplated, strengths may not be required
to exceed those that are conventionally developed in embank-
ments, earthwork dams, and the like. However, where the product
is to be used under heavy loads such as for foundations of
bridges or for road base, strengths must not only be higher but
should be required to reach specified values at earlier ages.
The moisture contents which can be used may range from 20% to
50%, depending on the end use which is contemplated and on
other factors. In certain applications additives may be used
to accelerate the early strength.
In addition to strength, a second factor that is of con-
siderable importance from an engineering standpoint is struc-
tural integrity. Procedures that have been used to evaluate
this characteristic are usually of two types; first, laboratory
measurements of dimensional stability, and, secondly, field
performance. Data from a few of the mixtures that have been
evaluated in the laboratory are given in Figure 14. These
measurements were made on moist cured bars (1 x 1 x 10") cured
at 73°F. In each case slight expansion is developed at early
ages, although the degree of expansion reaches a constant
value after several weeks and does not change significantly
from that point on. It is believed that this type of expansion
849
-------
is beneficial particularly where the material is used in mass
applications since it helps to control shrinkage cracks (a
concept that is used when expansive portland cements are used
in mass structures).
The structural integrity of the sulfur oxide-fly ash mix-
tures as observed in field applications to date has indicated
that these mixtures perform quite well. By making comparisons
with early Poz-0-Pac® evaluation programs, it is indicated that
mass structures such as have been placed in reservoirs and dams
can perform for many years with little detectable deterioration.
Poz-0-Tec* mixtures which have been used for road base work have
proven to be superior to the Poz-0-Pac® compositions, although
like Poz-O-Pac® they should be covered with an adequate wearing
course, as the stabilized mixtures per se are not capable of
withstanding the action of excessive traffic.
ECONOMIC CONSIDERATIONS
The costs of converting the by-product wastes of a power
plant depend on whether or not the end product is produced in
a marketable form. The least conversion costs result when
Sulf-0-Poz® is produced since this product may simply be dis-
posed of as stabilized fill which is designed to meet full
ecological, but only minimum structural and engineering criteria.
In many cases, depending on factors to be elaborated below,
Sulf-O-Poz® can be economically competitive with ponding and
850
-------
ecologically unsound disposal methods.
In certain markets, it may be economical to produce ana
sell or stockpile for future sale a usable end product such
as Poz-0-Tec* road base or aggregate. Revenues from the sale
of these products might appropriately be credited to the
additional costs of converting the waste by-products to
stabilized structural materials.
Since both the nonmarketable and the marketable approaches
result in an ecologically sound material, the decision as to
which approach to take is primarily an economic one.
A second area of importance in developing cost estimates
is related to the nature of the installation, thus if the con-
version process is added on to an existing plant, the capital
costs and operating costs may be quite different than would be
realized in the case of a completely new installation. In an
operating plant, fly ash may already be recovered in electro-
Static precipitators and this would obviously affect the capital
cost of the tail end system.
The major factors which influence the cost of the conversion
may be summarized as follows:
- Annual tonnages to be handled by the conversion
process.
- New boiler installation versus existing facilities.
- The type of equipment selected for fly ash removal
— for example, electrostatic precipitators versus
wet scrubbers.
851
-------
- The chemical analysis of coal — sulfur, CaO,
and ash contents.
- Location of plant — on site versus off site.
- Transportation costs — to and from conversion
plant.
- Redundancy factor — duplication of equipment
versus emergency holding basins, etc.
- Type of scrubber — limestone versus lime.
- Acquisition and cost of land.
- Type of end product selected.
In view of the fact that each of these items considered
alone or grouped together can cause major variations in both
the capital and operating costs, it should be recognized that
specific estimates cannot be given unless each situation is
reviewed on an individual basis. On the other hand, IU
Conversion Systems, Inc. has had an opportunity to make numer-
ous cost evaluations based on the programs which are currently
underway, and it is at least possible at this time to suggest
a few examples which may be helpful to the utilities in eval-
uating the conversion system approach in individual cases.
In several installations in which lime scrubbers have
been specified for new plants in the general size range of
1000 to 2000 megawatts, the cost of conversion to Sulf-O-Poz®
fill material will range between about $1.50 per ton to
$2.50 per ton, the variation depending in large measure on
items such as listed previously. Converting this to a basis
852
-------
of cost per million Btu for coal of approximately 3 to 4%
sulfur and 10 to 15% ash represents a range of about 2 to
3-1/3 cents per million Btu. Adding the necessary equipment
and operating staff to convert from a simple Sulf-0-Poz®
stabilized fill process to one in which commercial products
are fabricated would increase these costs by approximately
20 to 50% before allowing for any offsetting credits from
the sale of commercial products.
IU Conversion Systems, Inc. can provide the power plant
with the necessary hardware on a turnkey basis and will under-
take to operate the conversion plant on the basis of a long-
term contract. In general, the power industry prefers to"
contract with an outside organization to take care of the
disposal of its waste materials since it is not normally
geared up to operate this type of plant. What the industry
requires then is a reliable process which will give them the
assurances of low-cost disposal with adequate warranties and
guarantees. It is also essential that the conversion system
shall not interfere with the operation of the power plant.
At the presert time the cost estimates given above appear
to be quite attractive since they are, in fact, competitive
in many instances to what the power plant pays simply to hire
a trucking organization to haul the waste material to a dump
or lagoon. In view of the concern of the environmentalists
853
-------
and the difficulty in obtaining permits for disposal of the
waste material, it is clearly indicated that a system which
converts the waste to ecologically safe compositions is to
the advantage of a power plant operation.
SUMMARY
The extension of existing technology dealing with the
stabilization of fly ash into the area where waste sulfur
oxide sludges can be utilized provides a practical solution
to the question of fixation and disposal of flue gas waste
products. The real effectiveness of the fixation process
depends on the concept of converting the waste materials into
compositions that comply with ecological and engineering
criteria that allow for the disposal or utilization of the
product in a manner which is beneficial or at least not de-
leterious to the environment.
The performance of the stabilized mixtures as demonstrated
in laboratory tests and field evaluation programs has been
quite favorable. The ratios of ash to sulfur oxide sludges
that are produced in the average coal-burning operation fall
within the ranges that can be utilized for the manufacture of
the various compositions.
The conversion system is adaptable to power plants as a
retrofit application or to new installations. Since the pro-
cess depends on the availability of fly ash, it is not appli-
854
-------
cable to oil fired operations unless such operations are
carried out in the vicinity of stockpiled waste fly ash.
Cost evaluations which have been made to date indicate
that for full sized plants the conversion system is usually
competitive with costs that would be incurred simply by
transporting the waste material to dumps or lined lagoons.
In specific cases commercial opportunities may be available
which would allow the utilization and sale of products
produced in the process.
Based on the findings to date, it is indicated that a
realistic solution to the disposal of sulfur oxide sludges
from lime and limestone scrubbers can be attained.
855
-------
TABLE 1
RESULTS OF TESTS OF SELECTED STABILIZED ROAD BASE MIXTURES
PREPARED AT DULLES AIRPORT TRANSPO 72 PROJECT
CO
Moisture
Content
19.5
19.4
20.0
19.8
19.7
20.0
19.1
Dry
Density
(pcf)
98.8
98.1
98.3
98.2
100.6
98.8
100.4
Falling Head
Permeabilities
(cm/sec)
7 Days
2.4 x 10~6
N.D.*
2 . 9 x 10"6
6.5 x 10~6
5.7 x 10~6
1.0 x 10~6
N.D.
Compress ive Strength
at 100° F (psi)
2 Days
301
267
369
196
333
290
200
14 Days
732
586
630
458
772
761
868
28 Days
881
622
889
490
861
789
1091
*N. D.
Not Determined
-------
TABLE 2
ATOMIC ABSORPTION TESTS FOR LEACHABLE IONS ON SELECTED
SPECIMENS SUBJECTED TO 48 HOUR SHAKING TEST
00
en
Total
Dissolved
pH* Solids
FEDERAL SPECIFICATIONS-MAX.
Individual Solid Specimen
Dulles Cylinder (13 Days)
Dulles Cylinder (22 Days)
Poz-O-Tec* Test Road Core
Poz-O-Tec* Test Road
Cylinder
Poz-O-Pac® Cylinder
Fly Ash Concrete
Cinder Block
Clay Brick
Asphalt Roofing Shingle
Aggregate
Argillite
Do lorn i tic Limestone
Calcitic Limestone
Steel Slag Aggregate
Pumice
Fly Ash-Sludge Aggregate
Cement Mortar Balls
Mine Tailings
Loose Powdered Materials
Fly Ash
Portland Cement
Water Samples
Tap Water
Snow Sample from Pittsburgh
Water Supply (Peggs Creek)
10.6
9.5
9.5
6.7
9.2
9.3
10.7
8.2
7.3
7.1
6.9
9.75
8.4
10.8
7.1
11.7
9.0
3.95
9.8
12.0
7.5
6.45
7.25
500
840
620
90
250
150
440
410
110
150
120
96
180
840
120
700
530
130
2900
3700
180
40
316
Sulfate Cl
250
100
120
16
136
44
170
60
28
46
28
8
8
16
< 1
< 1
27
6
1500
200
36
< 1
"
250
8
12
14
16
26
46
6
12
22
22
18
—
28
10
16
8
2
8
20
76
6
"
Al
None
.38
.37
.03
.05
.10
.22
.01
.03
.01
,07
.02
.02
.05
.06
.03
.04
.05
. 11
.05
.02
.06
Total
Iron
.3
.08
.08
.06
.10
.25
.01
.04
.10
. 12
,06
.36
1.8
. 15
2.2
.26
.17
. 15
.26
.44
< .01
.46
2.9
Mn
.05
. 18
. 16
<.05
.10
<.05
<.05
<.05
<.05
<.05
<.05
<.05
<.05
<.05
<.05
<.05
<.05
<.05
<.O5
<.05
<.05
<.05
<.05
Cu
1.0
.08
.08
< .01
.08
.08
.04
< .01
.01
< .01
.08
< .01
< .01
< .01
< .01
< .01
< .01
. 16
< .01
< .01
.08
< .01
.05
Zn Cd Cr+3
5.0 .01 .05
.02 <.01 .02
< .01 <.01 <.01
< .01 <.01 <.01
< .01 <.01 <.01
< .01 <.01 <.01
< .01 <.01 <.01
< .01 <.01 <.01
< .01 <.01 <.01
< .01 <.01 <.01
< .01 <.01 <,01
.02 <.01 <.01
.04 <.01 <.01
.02 .01 <.01
.03 <.01 <.01
.02 <.01 <.01
. 01 <. 01 <. 01
.02 <.01 <.01
. 01 <. 01 <. 01
< .01 <.01 <.01
.05 <.01 <.01
< .01 <.01 <.01
.02 <.01 <.01
As Hg
.01 .001
.02 <.001
.02 <.001
<.01 <.001
.01 <.001
.01 <.001
<.01 <.001
•c.Ol <.001
<.01 <.001
<.01 <.001
<.01 <.001
<.01 <.001
<.01 <.001
<.01 <.001
<.01 <.001
<.01 «.001
<.01 <.001
<.01 <.001
<.01 <.001
<.01 <.001
<.01 <.001
<.01 <.001
<.01 <.001
Pb
.05
.08
.09
<_ oi
.01
.02
<, 01
.02
<.01
<-01
<.oi
!o7
.03
.01
.06
.06
.03
.07
.06
.04
.04
.02
.02
Sn
None
. 10
. 10
< oi
<. 01
•c.Ol
<. 01
<. 01
<. 01
•S.01
<.01
<. 5
<. 5
-------
TABLE 3
ATOMIC ABSORPTION TESTS MADE ON SURFACE RUNOFF OF A
STABILIZED FLY ASH-SLUDGE MIXTURE
in
CO
pH*
Dulles Cylinder (13 Days) 7.0
Dulles Cylinder (23 Days) 6.9
Sulfite Beam 7.2
Total
Dissolved
Solids
100
96
85
Total
Sulfate Cl Al Iron Mn Cu Zn Cd Cr
26 12 .10 .22 <.05 <.01 <.01 <.01 <.01
32 18 .15 .06 <.05 <.01 <.01 <.01 <.01
8 18 .13 .06 .60 .12 <.01 <.01 <.01
As Hg Pb Sn
<.01 <.O01 <.01 <.01
<.01 <.001 c.Ol <.01
<.01 <.001 .03 <.01
+With the exception of pH, all values are reported in parts per million.
-------
TABLE 4
RESULTS OF FIELD TESTS SHOWING COMPARISON
OF POZ-O-PAC® AND POZ-O-TEC* FORMULATIONS
Description
Dry
Density
(pcf)
Compressive Strength
at 100°F (psi)
2 Days 7 Days
Strength of Cores
from Road1 (psi)
4 Weeks 6 Weeks
CO
in
<£>
Standard Fly Ash Mix 121.2
Fly Ash-Sludge Blend A 121.4
Fly Ash-Sludge Blend B 120.8
66
348
318
770
729
746
NCP2 NCP
NCP 1034
756 1089
Average temperature during curing period was 50°F.
2
NCP - No core possible due to insufficient strength.
-------
FIGURE 1 — Freshly prepared mixture of fly ash, sulfur oxide
and sludge (Left View).
Ettringite crystals formed after aging
{Right View).
860
-------
FIGURE 2 — Comparison of fly ash-calcium sutfite-Vime
mixture showing depletion of free Ca(OH)2
after aging. Upper profile — imaged.
Lower profile — aged 2 weeks.
8
8
I
»
8
i
£
M
I
_c
'i
*-•
3
&
Ca(OH)2
16
18
20
22
24
Angle of Diffraction (26)
-------
2000X
FIGURE 3 — Electron probe patterns showing early reaction
involving AI2O3 and calcium sulfite.
Al
Ca
862
-------
FIGURE 4 — Tobermorite crystals resulting from reaction
between SiO, in fly ash with lime.
863
-------
I
I
i
i.
00
SULFUR
OXIDES
UNDERFLOW
t mill
MAKE-UP PROCESS
ADDITIVES
Alllllllttt
t %
PRIMARY DEWATERING *'*»***
A ^r
• ••••ttlllllllllMllllllllllllllllllllllMIIIIIMIIIIIIlTlllllllllllllltiiiuu
• ^ •
SECONDARY
DEWATERING
MIXING &
CONDITIONING
Sulf-O-Poz®Products
Dams
Reservoir!
Impermeable Liners
Road Base
Structural Fill
•••••••••••••••i
Structural Products
Aggregate
Concrete
Structural Shapes
FIGURE 5 - Schematic diagram of Poz-O-Tec* process.
-------
FIGURE 6 - Fly Ash stabilization (Poz-O-Pac®) plant.
865
-------
100
b
^
X
U
00
<
111
5
DC
LU
Q.
Standard Fly Ash Mix
Sludge - Fly Ash Mix
468
AGE-DAYS OF CURING AT 100°F
FIGURE 7 - Comparison of Poz-0-Pac® and Poz-O-Tec*
permeability values.
866
-------
****-
id
FIGURE 8 - Shaking apparatus used for laboratory shaking test.
867
-------
FIGURE 9 — Electron probe scan showing lead profile in
ceramic ware subjected to long time outside
exposure.
-------
FIGURE 10 - Runoff test as conducted on sulfite
stabilized fly ash specimens.
869
-------
6000
4000
-
3
8
I
I
2
c
I
2000
Age (Weeks)
FIGURE 11 - Penetration resistance for a typical fly
ash-calcium sulfate-lime mixture.
-------
2000
f
I
w
5
's
I
o
u
3
1000
0
8
Age (Weeks)
FIGURE 12 — Compressive strength fora composition similar
to that given in Figure 11. Moisture
content of composition is 35%.
-------
200--
Sludge- Fly Ash Mix
CD
z
S 150
z
oc
O
Standard Fly Ash Mix
100-
1
AGE-DAYS OF CURING AT 70°F
FIGURE 13 - California Bearing Ratio for Poz-0-Tec* road base.
872
-------
.012
s.
00
^J
Ul
.000
0
8 12
Age. Weeks at 73° F
20
FIGURE 14 — Early expansion of fly ash-sludge mixtures.
-------
UTILIZING AND DISPOSING OF SULFUR PRODUCTS
FROM FLUE GAS DESULFURIZATION PROCESSES IN JAPAN
by
Jumpei Ando
Faculty of Science and Engineering
Chuo University
Kasuga, Bunkyo-ku, Tokyo, Japan
875
-------
Utilizing and Disposing of Sulfur Products from
Flue Gas Desulfurization Processes in Japan
Jumpei Ando
Japan's energy supply has doubled in the past five years depending
mainly on growing amounts of Imported oil. Even though nuclear energy
may eventually supplant petroleum, Imported oil and sulfur derived from
it are likely to keep increasing for some time to come. Regulations on
SOn emissions are becoming more and more stringent, inducing greater
efforts for the desulfurization of fuel oil and waste gas. While virtually
all of the by-products from the desulfurization—elemental sulfur, sodium
sulfite, sulfuric acid, gypsum, etc.—have been utilized so far, it will
not be long before supply of by-products runs far ahead of demand. Gypsum
is generally regarded as the most rational by-product because of its
several uses including land filling and also because of its ease of aban-
donment. Elemental sulfur is considered to be another logical by-product
because it may be stored for some time for future use.
1. Characteristics of desulfurization in Japan
Of the total electric power (386 billion Kwhr) produced in Japan in
1971, about 62% depended upon oil, 12% upon coal, 22% upon hydraulic
power, and 2% upon nuclear energy.1 Heavy oil, a residue of atmospheric
distillation of crude oil, has been used as a major fuel in recent Japan.
Heavy oil consumption in 1971 reached 745 million barrels. About 30% of
the heavy oil was burned in power plants and the rest in other industries
and buildings.1 Heavy oil from high-sulfur crude contains 4% sulfur.
In 1971 nearly one-fourth of the total heavy oil was subjected to hydro-
desulfurization giving 287,000 tons of elemental sulfur as by-product.
Still nearly 6 million tons of SO2 was emitted by burning the heavy oil,
constituting the chief source of SC^ emissions.
At present in Japan, more than 100 commercial and prototype plants
for waste-gas desulfurization are on stream. Most of them treat either
flue gas from oil-fired industrial boilers or waste gas from chemical
plants, smelters, etc. Nine major power companies which produce more
than 70% of tbtal power in Japan have recently decided to build many
flue gas desulfurization plants. The total capacity of the plants of the
power companies will increase from 357 Mw in 1972 to 2,700 Mw in 1974 and
to 4,800 Hw in 1976.
A salient-feature of the desulfurization efforts in Japan is that
they are oriented toward processes that yield salable by-products such
as sodium sulfite for paper mills, sulfuric acid, gypsum for retarder
of cement' setting and for gypsum board, etc. The feature is predicated
on the following .circumstances: (1) Japan is subject to limitations In
domestic 'supply of sulfur and its compounds as well as in land space
876
-------
available for disposal of useless by-products. (2) Most desulfurization
plants treat flue gas from oil-fired boilers, waste gases from chemical
plants, etc. which contain less dust than does coal-fired flue gas and
therefore are suited to the production of by-products of high purity.
(3) Most steam power plants are located close to chemical plants; hence
it is easy to utilize the desulfurization by-products in chemical plants.
However, since so many desulfurization plants are to be built, it
is considered that in future the supply of the by-products will far
exceed the demand, necessitating the abandonment of a substantial portion
of them. Even now there are many small plants, (less than about 30,000
scfm) which produce waste solutions of sodium sulfate, sodium sulfite, and
ammonium sulfate; it is uneconomical to recover by-products at small
plants.
2. By-production of sodium salts
Since 1966 more than 50 units have been built to recover S02 to
by-produce sodium sulfite for paper mills. Some of the plants are listed
in Table 1. In most of them, S02 is absorbed with a sodium sulfite
solution to form sodium bisulfite which is then treated with sodium
hydroxide to produce sodium sulfite. (By treating flue gas with a sodium
hydroxide solution, C02 is also absorbed.) One-half of the sulfite solu-
tion is recycled to the absorber and the rest is concentrated to produce
anhydrous sodium sulfite. An example of the process is shown in Figure 1.
Flue gas is first washed with water in a conventional scrubber for cool-
ing and dust removal; 60 to 80% of dust which is contained in flue gas
from oil-fired boiler to an extent of about 0.1 grain/scf is removed.
The purity of the product ranges from 90 to 97%; small amounts of sodium
sulfate, dust, etc. are contained in the product. Sodium sulfite solu-
tion of about 20% concentration is also produced in some plants and used
for paper mills.
A new semi-wet process for sodium scrubbing has been developed
recently by Hitachi, Ltd. (Figure 2). Flue gas from oil-fired boiler
at about 350°F after passing through a dust eliminator is introduced
into a reactor into which sodium hydroxide solution is fed. By the
heat of the gas, moisture is removed and powdery product consisting of
sodium sulfite, sulfate and carbonate (for example, Na2SOo-60%,
Na2S04-20%, and Na~C03-20%) is formed which is caught by a dust separator.
The product is usable for kraft pulp production.
877
-------
CO
^•J
CO
Table 1. Major Waste-Gas Desulfurization Plants that By-Produce Sodium Salts
(Charge: NaOH)
Unit capacity
Process developer
Oji Paper
Oji Jinkoshi
Kureha Chemical
Kureha Chemical
Showa Denko
Showa Denko
Bah co -Ts uk i s h ima
Banco -Tsukishima
Hitachi, Ltd.
IHI-TCA
Mitsubishi (MKK)
Kurabo Ind.
Toyobo Co.
IHI-TCA
Product
Na2S03
Na2S03
Na2S03
Na2S03
Na2S03
Na2S03
Na2S03
Na2S03
Na2S03
Na2S04
Na2S04
Waste
Na2S04
Waste
NaoSO*
i •**
Waste
Na2S04
User
Oji Paper
Oji Paper
Kureha Chemical
Konan Utility
Aj inomoto
Asia Oil
Daishowa Paper
Daio Paper
Jujo Paper
Mitsui S. 0.
Asahi Glass
Bridgestone Tire
Toyobo Co.
Nissan Motor
Plant site
Kasugai
Tomakomai
Nishiki
Konan
Kawasaki
Yokohama
Yoshinaga
lyomishima
Miyako j ima
Sakai
Amagasaki
Tokyo
Shogawa
Oppama
(1,000 scfm)
805 (in 12 units)a'b
400 (in 4 units )a
176a, 176a
123a
159a
142a
129b, 65b, 26b
88a, 70a
57a
88a, 88a, 88a
130C
71a
24a
67a
Date of completion
1966 - 1972
1971 - 1972
1968
1972
1971
1972
1971
1972
1972
1973
1973
1972
1971
1972
a: Oil-fired boiler
b: Kraft recovery boiler
c: Glass melting furnace
-------
00
^J
<£>
Water
(27Vhr)
Flue gas
(I?6f000scfm)
(S02 l.^OOppm)
To wastewater
treatment
W&ste gas
-------
00
00
o
t
Reactor
Dust \.
collector
To stack
t
NaOH
i
i^^.i
I By-product
Figure 2 Hitachi semi-wet sodium process
-------
These sodium processes are simple and are operated with ease.
Investment cost is low. Demand for sodium sulfite, however, is limited.
As shown in Table 2, production of the sulfite has increased rapidly
with the progress of desulfurization and has resulted in a considerable
price drop for the sulfite due to oversupply.
2
Table 2. Production and Price of Anhydrous Sodium Sulfite
1967 1968 1969 1970 1971
Production (1,000 t) 113 135 171 289 330
Price ($/t) 64 62 61 58 60
Under such situation, the following ways of sodium scrubbing have
been developed recently: (1) Several plants have been built recently
to by-produce salable solid sodium sulfate. The sulfate is produced by
air oxidation of sodium sulfite solution. Demand for the sulfate is
also limited. (2) Many smaller plants have started to produce waste
solution of sodium sulfate or sulfite. Some of these plants are listed
in Table 2. (3) Showa Denko as well as Kureha Chemical jointly with
Kawasaki Heavy Industries have developed sodium-limestone double alkali
processes which are described in the present author's separate paper
for the symposium entitled "Flue Gas- Desulfurization Technology in
Japan."
881
-------
Table 3. Major Waste-Gas Desulfurizatlon Plants that By-Produce Gypsum
CO
KJ
Process developer
Mitsubishi-JECCO
Mitsubishi-JECCO
Mitsubishi-JECCO
Mitsubishi-JECCO
Mitsubishi-JECCO
Mitsubishi-JECCO
Babcock -Hitachi
Chiyoda
Chiyoda
Chiyoda
Chiyoda
Chiyoda
Hitachi, Ltd.
Showa Denko
Kureha-Kawasaki
Kureha- Kawasaki
Nippon Kokan
Chemi co -Mitsui
Absorbent
Ca(OH)2
Ca(OH)2
Ca(OH}2
Ca(OH)2
Ca(OH)2
CaC03
CaC03
H2S04, CaC03
H2S04, CaC03
H,,SQ. , CaCOr
Z 4 ^
H_SO . . CaCO,;
2. 4 -S
H2S04, CaC03
Carbon , CaCO „
KaOH, CaC03
NaOH, CaCO 3
NaOH, CaC03
NH3,Ca(OH)2
Ca(OH)2
User
Nippon Kokan
Kansai Electric
Onahama Smelting
Kansai Electric
Tohoku Electric
Tokyo Electric
Chugoku Electric
Fuji Kosan
Mitsubishi Rayon
Daicel
Hokuriku Electric
Mitsubishi
Petrochem.
Tokyo Electric
Showa Denko
Shikoku Electric
Tohoku Electric
Nippon Kokan
Mitsui Aluminum
Plant site
Koyasu
Amagasaki
Onahama
Kainan
Hachinoe
Yokosuka
Mizushima
Kainan
Otake
Aboshi
Shinminato
Yokkaichi
Kashima
Chiba
Shintokushima
Sendai
Keihin
Omuta*
Capacity
(1,000 scfm)
37d
59a
54e
235a, 221a
224a
220a
170a
93a
S33
S9a
420a
413a
250a
340a
250a
220a
88b
226C
Date of completion
1964
1972
1972
1974
1974
1974
1974
1972
1973
1973
1974
1974
1972
1973
1974
1974
1972
1972
*Producing waste calcium sulfite; gypsum will be produced in near future.
a: Oil-fired boiler d: Sulfuric acid plant
b: Iron-ore sintering plant e: Smelting furnace
c: Coal-fired boiler
-------
3. Gypsum and calcium sulfite
3.1 Uses
Gypsum is considered the ideal by-product for the time being.
Many processes have been developed in Japan recently that recover
salable gypsum with sufficiently good quality to make it available as
retarder of cement setting and for production of gypsum board (Table 3).
Six of the processes which seem to be of more interest for possible
application in the U. S. A. are described in the author's paper referred
to above. Demand for and supply of gypsum in Japan are illustrated in
Figure 3. Prices of by-produced gypsum (mainly phosphogypsum from
wet-process phosphoric acid production) are shown in Table 4.
2
Table 4. Prices of By-Produced Gypsum ($/t)
All of the phosphogypsum and most of the other by-products have been
used for cement and board production. As demand for gypsum has increased
substantially, supply is on the short side at present, resulting in some
price increase.
In order to use by-product gypsum for wall board, proper crystal
size and shape and less impurity are required to ensure high strength.
Phosphogypsum from "the conventional dihydrate process usually has a
small crystal size (10-30 microns) with much impurity and gives less
strength. Japanese phosphoric acid producers have developed hemihydrate-
dihydrate processes by which good crystals of gypsum (50-150 microns)
with less impurity are obtained. Good phosphogypsum from these processes
has a strength equal to or better than natural gypsum.
Such background has helped find outlets for by-product gypsum from
desulfurization. For example, a wall board production plant of
Onahama-Yoshino Gypsum Co., Ltd., using gypsum (450 t/day) recovered
from the Mitsubishi-JECCO process is under construction at Onahama works,
Onahama Smelting and Refining Co. Another example has been presented
by Mitsui Aluminum Co. which has produced waste calcium sulfite by
treating coal-fired flue gas (226,000 scfm) with the Chemico-Mltsui
process. Mitsui Aluminum, jointly with Mitsui Toatsu Chemicals, Inc.,
has succeeded in tests to manufacture wall board from gypsum obtained
from the calcium sulfite. The gypsum has a little less than 90% purity,
containing fly ash and other impurities. Mitsui Aluminum plans to
install a prototype reactor within 1973 to convert about one-fourth of
its by-product calcium sulfite to gypsum. It is likely that a full-
scale reactor will be installed in 1974.
883
-------
CO
CO
W
O
•P
O
CO
o
•H
rH
Q
8 r
6 -
—
—
•§•
B
Q
OU
B
C
rH
CL
Q.
W
OS
R
P
OT1
B
C
OS
H
P
at
1
O
OTT
C
a.
to
OS
R
OU
B
c
OS
H
p
•
2 -
196?
1973
Demand
Supply
1976
1970
C: Cement B: Board OU: Other uses
Pi Phospho-gypaura R: Recovered OS: Other sources
Figure 3 Demand for and supplv of gypsum in Japan
-------
By-product gypsum is available also as retarder of cement setting.
Gypsum is added to Portland cement clinker to the extent of 3-4% at
grinding. Nippon Kokan (Mitsubishi-JECCO lime-gypsum process) has
sold the by-product gypsum for cement since several years ago. Gypsum
should be charged continuously into the cement mill. Powdery gypsum
with more than 10% moisture tends to form a "bridge" in the hopper and
cannot be charged smoothly. By the Mitsubishi-JECCO process and the
Nippon Kokan ammonia-lime process, the product gypsum from a centrifuge
has a low moisture content of about 10%. Gypsum of about 90% purity
containing fly ash is useful, but alkali-rich gypsum by-produced from
the double alkali process would impair cement setting. New uses of
gypsum and calcium sulfite have been developed recently in Japan.
Mitsui Toatsu Chemicals, Inc., jointly with Taisei Construction Co. has
started producing a new material "gypsum polymer composite" with gypsum
and methyl methacrylate. Lion Fat and Oil Co. jointly with Idemitsu
Kosan has commenced production of synthetic paper from calcium sulfite
and polyethylene at a weight ratio of about 70:30. By-produced gypsum
and calcium sulfite could be used if the dust in flue gas is removed
reasonably well so that the color is not too dark.
3.2 Discarding
Gypsum is produced from calcium sulfite by oxidation. Oxidizer is
virtually unnecessary in certain plants—such as Amagasaki plant,
Kansai Electric Power (Mitsubishi-JECCO process) and Rumagaya plant,
Chichibu Cement (IHI-TCA process)--which treat flue gas with relatively
low S02 and high 02 concentrations; essentially all of calcium sulfite
is converted to gypsum in the scrubber. But normally an oxidizer (for
oxidation with air) or a reactor (by the above mentioned Mitsui gypsum
process in which a catalyst is used) is required which adds some cost
to desulfurization. Gypsum has advantages over calcium sulfite even
for land filling or discarding. Gypsum can be grown into fairly large
crystals (50 to 300 microns); moisture content of the centrifuge dis-
charge can be made as low as 10%. On the other hand, the crystal of
calcium sulfite is normally very small (1-10 microns); centrifuge
discharge contains about 60% moisture and is like a paste. The calcium
sulfite may not suit land filling because of the high moisture which is
not easily removed, while gypsum would be useful. Moreover, calcium
sulfite has some danger of consuming oxygen in ambient water. In
discharging in slurry form to a waste pond, gypsum precipitates much
more easily in smaller volumes than does calcium sulfite, thus reducing
the required pond size. In case of truck transportation, gypsum can be
handled with greater ease. By the sodium-limestone processes developed
recently by Showa Denko and also by Kureha Chemical jointly with
Kawasaki Heavy Industries, calcium sulfite grows into much larger crystals
than usual so that the above mentioned problems might be solved.3
885
-------
4. Other by-products
4.1 Sulfuric acid and sulfur
There are many processes that by-produce sulfuric acid and sulfur
(Table 5). By these processes SO^ in waste gases is first absorbed
with various absorbents and then recovered as SC>2 gas of 7 to 95%
concentration which is used for sulfuric acid or sulfur production.
Among these processes, the Wellman-Lord process is well known. In
Japan plants based on this process were constructed respectively by
Mitsubishi Chemical Machinery Mfg. (MKK) and Sumitomo Chemical Engi-
neering Co. (SCEC) in 1972. The former produces sulfuric acid and the
latter returns the recovered S02 gas to a Glaus furnace to produce
elemental sulfur. By magnesium scrubbing processes, magnesium sulfite
formed by the reaction of magnesium hydroxide and S0_ is calcined to
recover SCL for sulfuric acid production. Some magnesium sulfate is
also formed which is not as easy as magnesium sulfite to be thermally
decomposed. The Mitsui Mining process features the by-production of
some solid magnesium sulfate for fertilizer and other uses. Demand
for magnesium sulfate, however, is limited. The Onahama Smelting
process treats waste gas from a copper smelter containing about 20,000
ppm S02 to produce 6,600 t/month sulfuric acid. By the Chemico process,
a new plant to treat 294,000 scfm waste gas from a Claus furnace is to
be constructed by 1974. By the Sumitomo Shipbuilding process, the SC^
absorbed on activated carbon is expelled by heating it in a reducing
gas, to release SC^ gas of 10-20% concentration which is used for
sulfuric acid production.
By the Shell process, SC>2 is absorbed with copper oxide to form
copper sulfate, which is then treated with reducing gas to expel SC^-^
A commercial plant is scheduled to come on-stream in a few months at
Yokkaichi. The MHI-IFP process uses ammonia scrubbing with thermal
decomposition of ammonium sulfite and sulfate to regenerate S0«
(Figure 4).6
A commercial plant based on the MHI-IFP process will come on-stream
in 1974. In these two processes, the recovered SO* is reacted with ELS
to produce elemental sulfur.
Supply of sulfur and sulfuric acid in Japan is shown in Table 6.
Their prices are shown in Table 7. Demand for sulfur and sulfuric acid
is nearly equal to the supply. There has been little import or export
of sulfur and sulfuric acid. Sulfur mines in Japan are small and
sulfuric acid has been produced mainly from pyrite and smelter gas.
Since elemental sulfur recovered from hydrodesulfurization of heavy oil
has increased resulting in the lowering of sulfur price, the use of
sulfur, for sulfuric acid production has been started recently.
An annual increase in the demand for sulfuric acid is estimated at
about 5%; some more flue gas desulfurization plants by-producing sulfuric
acid will be built but not very many of them are expected.
886
-------
Table 5. Major Waste Gas Desulfurization Plants that Produce Other By-Products
Process developer
Wellman-MKK
Wellman-MKK
Wellman-MKK
Wellman-SCEC
Wellman-SCEC
Mitsui Mining
Onahama Smelting
Chemico-Mitsui
a, Sumitomo Shipbuilding
OS
•vj
Shell
MHI-IFP
Nippon Kokan
Mitsubishi (MHI)
Kurabo Ind.
Kurabo Ind.
Absorbent
NaOH
NaOH
NaOH
NaOH
NaOH
MgO
MgO
MgO
Carbon
CuO
NH3
NH3
MnO , NH..
A. J
NH3
NH3
Product
H2S04
H2S04
H2S°4
S
H2S04
H2S°4
H2S04
S
S
S
-------
Voter
Ammonia
Recovery
Tower
Absorber
Cooler
1
Waste
$as
X
3
X
TTTK
Make-up
oo
OS
00
NH3
Evaj
ratoi
c
5O-
*
T
>
X
s*
J
t
S02,
/
Dust
Reactor
1
I
~,I.F.P I
t reactor I
T
I
I
I
I
r
Fuel
After-burner
Sulfur
Hydrogen sulflde
I
L.
Foul acid gas
Figure
Fl°w sheet of M.H.I.- I. P.P. process
-------
Table 6. Supply of Sulfur and Its Compounds2
(1,000 tons of material)
1968 1969 1970 1971 1972
Sulfur
Mined 283 180 120 47 18
Recovered 97 179 297 414 526
Total 380 354 417 461 544
Sulfuric acid
from pyrite 4,576 4,524 4,303 3,348 2,747
from smelter gas 2,719 2,972 3,242 3,770 4,343
from sulfur 0 0 17 210 295
Total 7,295 7,496 7,562 7,328 7,385
2
Table 7. Price of Sulfur and Its Compounds ($/t)
1967- 1968
Sulfur 73
Sulfuric acid 26 25
Elemental sulfur can be produced from recovered SO, easily at oil
refineries which have H2S. For plants which have no H S, the production
may be costly because it is required to reduce two-thirds of the SO2 into
H2S; the desulfurization cost including production of elemental sulfur may
be close to that of hydrodesulfurization of heavy oil by the topped-crude
process which ranges from $1.1 to 1.5/bl oil to reduce sulfur from 4 to
1%. When a large oversupply of sulfur and its compounds occurs due to
the development of desulfurization, the sulfur production process would
assume greater importance because elemental sulfur has the smallest
volume among sulfur and its compounds and requires the least transpor-
tation and storage space.
4.2 Ammonium sulfate
There were several ammonia scrubbing plants a few years ago in
Japan treating tail gas from sulfuric acid plants to by-produce ammonium
sulfate. All of the plants have been shut down because of the oversupply
of ammonium sulfate and also because double contact processes for
889
-------
production of the acid has been Introduced to reduce S02 emissions. Now
there are three ammonia scrubbing plants. Two of them produce dilute
ammonium sulfate solution to be discarded, and the last produces solid
ammonium sulfate (Nippon Kokan process) although in this plant ammonium
sulfite and sulfate are being converted to gypsum at present. In
addition, ammonium sulfate has been produced at Yokkaichi Station,
Chubu Electric Power, using the Mitsubishi manganese process. Due to
the worldwide increase in the demand for nitrogen fertilizers, it is
possible that ammonium sulfate will be produced in future from S0~ in
flue gas at a considerable number of plants in several countries. One
of the problems in ammonia scrubbing has been plume formation, but the
problem has been nearly solved in Japan. For plants that produce waste
solution, ammonia scrubbing.is less expensive than sodium scrubbing
because ammonia is much cheaper than sodium hydroxide. It is likely,
however, that the emission of ammonium sulfate solution will be restricted
because it could cause a eutrophication problem.
REFERENCES
1. Enerugi Tokei (Energy Statistics), Ministry of International Trade
and Industry (MITI) Japan, 1972.
2. Sekiyu to Sekiyu Kagaku (Petroleum and Petrochemistry), Vol. 16,
No. 8, 1972.
3. H.W. Elder, F.T. Princiotta, G.A. Hollinden, and S.T. Gage,
Sulfur Oxide Control Technology, Visits in Japan—August 1972,
Interagency Technical Committee U.S.A., Oct. 1972.
4. F. Nishimi and Y. Ikeda, 24th Technical Meeting, the Sulfuric
Acid Association of Japan, Oct. 1972.
5. K. Kishi and R.F. Bauman, Kogai Boshlsangyo (Pollution Control
Industry), Dec. 1972.
6. Haiendatsuryu no subete (All about Waste-Gas Desulfurization),
Jukogyo Shimbunsha, Nov. 1972.
890
-------
LONG RANGE MARKET PROJECTIONS FOR BY-PRODUCTS
OF REGENERABLE FLUE GAS DESULFURIZATION PROCESSES
by
M. H. Farmer
Esso Research and Engineering
Linden, New Jersey
891
-------
ABSTRACT
LONG RANGE MARKET PROJECTIONS FOR BY-PRODUCTS OF
REGENERABLE FLUE GAS DESULFURIZATION PROCESSES
If all sulfur emitted from utility stacks, smelters, etc.
were to be recovered in useful form, the quantity would be so large
as to have a major impact on sulfur's price and markets, with the
possibility of causing severe dislocations'. This paper desribes a
quantitative study of this proposition, using a computer model, with
projections extending to the year 2020. The purpose was to provide
the EPA with a planning tool for establishing the relative priorities
that should be given to development of technology for recovery of
'.'abatement sulfur" in (a) marketable and (b) non-marketable forms.
The paper describes how various abatement schedules were
simulated, and the supply/demand/price implications of these schedules.
Because sulfur is an international commodity, the domestic abatement
schedules and their implications are placed in an international
context and related to world supply/demand/price relationships. In
addition, the relative values of sulfur recovered in acid and elemental
form are discussed.
Note: The work on which this paper is based was performed pursuant to
Contract No. EHSD 71-13 with The Environmental Protection Agency,
892
-------
INTRODUCTION
In its 1970 report on "Abatement of Sulfur Oxide Emissions from
Stationary Combustion Sources," a joint panel of the National Academy of
Engineering/National Research Council remarked that it would be desirable
to conduct a study of the long range supply and demand situation with
regard to the several alternative (sulfur) by-products to aid in establish-
ing priorities for support of control and abatement technology.
NAE/NRC's suggestion was implemented by the EPA through Contract
No. EHSD 71-13 with the Government Research Laboratory of Esso Research
and Engineering Company.
The full report of this study has been issued as Document No.
PB 208993 by the National Technical Information Service, Springfield, Va.
22151.
PURPOSE OF STUDY
To provide the EPA with a planning tool for establishing the
relative priorities that should be given to development of technology for
recovery of "abatement sulfur" in marketable and non-marketable forms.
TO BE DISCUSSED
• How various abatement schedules were simulated.
• The modeling of domestic and foreign sulfur supply and demand to
permit projections to the year 2020.
• The effect of stockpiling elemental sulfur.
• The price implications of the supply and demand projections, as
derived by linear programming (computer) calculations.
• The relative values of sulfur recovered in acid and elemental forms
893
-------
SIMULATION OF ABATEMENT SCHEDULES
Abatement refers to removal of sulfur so that it has not emitted
to the atmosphere. The degree of abatement, as treated in this study,
refers to the percentage reduction from the total quantity of sulfur
oxides emitted from utility stacks in 1970. This was the base year for
the study. The degree of abatement does not include sulfur removed from
fossil fuels, e.g., by desulfurization of petroleum, prior to combustion.
This is because such sulfur was treated separately in the supply/demand
calculations, so that the effects of flue gas desulfurization could be
isolated.
The term "useful form" is used to distinguish between sulfur
removed in elemental form or as acid from that recovered as a "throwaway"
by-product. Theoretically, a given degree of abatement could be achieved
with no sulfur in useful form (if only throwaway by-products were produced),
or with 1007. recovery in useful form, or with an infinite number of inter-
mediate cases. However, what matters in the computer calculations is
the quantity of sulfur in useful form, rather than the percentage that
this quantity represents of the total amount abated or of the total abate-
ment potential.
The total abatement potential in 1970 was the amount of sulfur
actually emitted from utility stacks in that year. For the other years for
which calculations were made (1975, 1980, 1985, 1990, 2000, 2010, 2020),
it was necessary to project both the amounts and sulfur contents of the
fossil fuels that domestic electric utilities would burn. This led to the
894
-------
abatement potential without any stack-gas treatment. From these
quantities, it was then possible to assume varying degrees of abatement,
e.g.., so that 1970emissions would not be exceeded! and other cases in
which various percentage improvements over the 1970 level would be achieved.
Having done this, further assumptions were made about how much of the
abated sulfur would be in useful form. These are the quantities of useful
sulfur that were included in the supply/demand calculations.
It is important to recognize that the study did not forecast that
any particular quantity of sulfur would be abated in any future year.
What" was done was to simulate a wide range of possibilities, and then to
assess the implications of these possibilities in terms of the calculated
value for recovered sulfur in different parts of the United States. The
whole purpose was to provide the Environmental Protection Agency with a
quantitative tool for exploring the consequences of different policies and
technological approaches for the abatement of sulfur oxide emissions.
MODELING OF DOMESTIC AND FOREIGN SUPPLY AND DEMAND
For convenience, the study has been referred to as a "long range
sulfur supply and demand model." In practice, several models and sub-models
were developed, and one of these models was "computerized" for the purpose
of calculating sulfur values in different locations. Two types of model
were used:
- a simulation or morphological model to represent the sulfur
industry and the structural way in which abatement sulfur
could impact on it.
895
-------
- a forecasting model to project the demand/supply/price framework
of the industry at various times into the distant future.
In preparing the simulation model, it was necessary to take into
account not only all of the major supply/demand regions under specific
study, but also the external sources of supply and demand that could inter-
act with the former. Thus, the starting point was the "World model"
shown in Figure 1. The next step was to construct the "North Americal model"
shown in Figure 2. The numbered boxes in this figure are the U.S. regions
under specific study. The cross-hatched boxes are the extra-regional
North American suppliers to the United States. For example, Calgary
is the center for Canadian sulfur recovered from sour natural gas,
Coatzocoalcos represents Mexican Frasch production, while Aruba represents
production of sulfur from all Caribbean refineries. The cross-hatched
boxes with the underscoring represent sources of sulfuric acid, rather than
elemental sulfur, that is potentially available for shipment to other
regions. Whether such shipment of acid will occur in practice involves
economic factors, particularly transportation costs, that will be discussed
later.
Figure 3, an exploded map of the continental U.S., shows the numbered
regions of Figure 2 in their normal geographical perspective and also the
way in which the regions coincide approximately with the Federal Power
Commission's regionalization of electricity generating capacity.
The elements of the forecasting model are listed in Table 1.
Unlike many long range projections that may be little more tihan 10-15 year
extrapolations, the sulfur study is concerned with the next five decades.
896
-------
For this reason, it was considered necessary to start by making economic
projections, and then to use correlations to convert the levels of
industrial and agricultural activity into projections of sulfur supply
and demand. The principal steps are indicated in Table 1. It should be
noted that a key step was to translate the projections of economic activity
into energy demand. This is necessary because it is from fossil fuels
that the greater part of the by-product sulfur ia coming, and will
continue to come, during the next several decades. In addition, the demand
for sulfur, whether industrial or fertilizer demand, also correlates with
economic activity and economic growth. Consequently, by basing the overall
projections on correlations with constant dollar G.N.F., there is some
self-compensation In the net statistics. For example, if economic develop-
ment is overestimated this would have the effect both of overestimating
the potential for sulfur recovery and of demand for sulfur.
Separate forecasts were developed for nuclear and hydro energy
that, when subtracted from projected total energy demand, gave the fossil
fuel energy requirements. The latter were broken down into individual
projections for natural gas, petroleum and coal. These quantities, in
conjunction with data concerning the sulfur content of these natural
resources in different parts of the world, led to the amounts of sulfur
potentially recoverable from each. Separate projections were made of the
sulfur recoverable from smelters, pyrites and miscellaneous sources. And,
finally, estimates were made of the future availability of elemental
sulfur from domestic and foreign Frasch mines.
897
-------
On the demand side, a set of GNP/Industrial sulfur correlations
was used to project industrial sulfur demand in different areas, while
separate correlations were used to project fertilizer sulfur demand.
Worldwide, the latter accounts for about half of the total demand for
sulfur, but, because of differences in agricultural acreage in different
countries and other factors, does not necessarily move in step with
industrial demand for sulfur.
The sum of industrial and fertilizer demand for a particular
region* is the region's'total demand. The difference between this quantity
and the region's internal supplies of sulfur represents either a net demand
for sulfur or the capability of the region to supply sulfur to another
region. For the purpose of the computer calculations, a net foreign
demand was located in Northern Europe (symbolized by Rotterdam) while net
supply potentials were located in Canada, Mexico and the Caribbean.
These four foreign regions, together with the eleven regions of the
continental U.S., comprised the "North American model" used for the
computer calculations.
EFFECT OF STOCKPILING ELEMENTAL SULFUR
By 1970, stockpiles of elemental sulfur in Western Canada had
reached several million tons and now approximate 8 million long tons. This
sulfur is coming from sour natural gas, and must be removed before the gas
is delivered to a commercial pipeline. Thus, it is the demand for natural
gas, not the demand for sulfur, that has been a major factor in the rate
if as defined in Figures 1 and 2.
B98
-------
at which this by-product sulfur has been produced. The reason for the
stockpiles is that production has greatly exceeded the demand of the
markets to which the Western Canadian sulfur can be delivered economically.
Recently, there have been additional restraints on movements of sulfur
through the port of Vancouver and on rail deliveries to the Midwest.
These factors have accentuated the stockpiling in Alberta, and have con-
tributed to a temporary tightness in worldwide sulfur supplies. The word
temporary is used advisedly because the transportation bottlenecks can be
relieved and, once this happens, the Canadian stockpiles will again exert
a major influence on world sulfur supplies and prices.
Transportation costs are a major factor in the sale of Western
Canadian sulfur. Competition takes place at the point of delivery, e.g.,
Northern Europe. The price obtainable here minus the transportation cost
gives the net-back to Western Canadian producer. Clearly, it is not in
the interest of the producers to move sulfur into world markets that would
net little or nothing back to Alberta. This is why Canadian sour gas sulfur
competes in some world markets but not in others . In turn, this is why
the Canadian stockpiles of recovered sulfur are growing. On the other
hand, the existence of the stockpiles means that this sulfur could enter
world markets on comparatively short notice provided that the price obtainable
for it was sufficient to cover transportation costs and give a tolerable net-
back to the producer. This statement begs the question of what is meant
by "tolerable return." An elliptical answer to this very difficult question
is that an incremental sale may cause the average price of all sales to be
depressed thereby lowering the overall net-back. Thus, the optimum strategy
899
-------
from the standpoint of a Western Canadian sulfur producer may require a
combination of stockpiling and export sales.
The Canadian stockpiles differ in function from those of the
Frasch sulfur producers. At present, domestic Frasch stocks are probably
just below the level of 4 million long tons and have been declining slowly
as sulfur demand has grown and various production and transportation
bottlenecks have occurred. In principle, however, the Frasch stocks are
merely a working inventory. When this inventory rises, the Frasch
producers will tend to cut production; when inventory falls, there will be
attempts to expand production. As discussed already, such flexibility
of production is not enjoyed by producers of by-product sulfur.
What has been happening in Canada illuminates what may occur in
the U.S. as and when air pollution controls force the recovery of sulfur
from utility stacks. If a surplus of sulfur from this source is to be
avoided, some combination of the following approaches may be necessary:
(a) recovery of large amounts of sulfur in the form of throw-away by-
products .
(b) recovery and stockpiling of elemental sulfur.
Singly, or in combination, these approaches could be applied so as to limit
the quantity of sulfur recovered in useful form to the amount that could
be marketed without causing major dislocations to the industry.
There is some evidence that the application of throw-away systems
may be limited either by logistics or by the possibility of water pollution
from the discarded by-products. This is the reason, why the E.P.A. asked
for the implications of national stockpiles of elemental sulfur to be
900
-------
examined. As treated in the computer calculations, a national stockpile
of sulfur is a source of "negative supply" since it removes sulfur from
the market. Additionally, a national stockpile would permit the re-entry
of sulfur in later years on the assumption that the general condition
of potential oversupply will eventually be corrected. Conceptually, this
condition could result from a continuing increase in the demand for sulfur
and from an eventual decline in the use of fossil fuels for the generation
of electricity, general industrial purposes, etc.
LINEAR PROGRAM FOR CALCULATION OF SULFUR VALUES
The computer calculations were made with IBM's linear program
"LP-MOSS." This is a variant of the familiar transportation linear program.
In addition, a number of specially developed programs and sub-routines were
needed to get the information in and out of the computer. A schematic
diagram of the entire procedure is given in Figures 4.
The linear program calculates the way that each of the net
demand regions in Figure 2 can secure the necessary supplies at lowest
cost from other domestic and foreign regions that are in a net supply position.
The net demand in a particular region is the region's gross demand
minus whatever supplies of sulfur are available from within the region.
The base case of net demand excludes regional abatement sulfur from the
region's internal supplies. Other cases then consider varying quantities
of abatement sulfur in useful form. Such sulfur is conceived to reduce the
net demand of the pertinent region before the possibility of shipment to
another region is considered (i.e., permitted by the computer program).
The LP program also requires the following input data;
901
-------
- Quantity of supply available from each extra-regional supplier.
- F.O.B. price for each source of extra-regional supply.
- Transportation costs for each linkage between the net demand
regions and the extra-regional supply points.
In addition, certain special feastures were incorporated:
- Upper bounds on the amount of acid that may be shipped into a
given region (in order to simulate the capacity of the region
to accept merchant acid). The upper bounds were raised with
time in order to simulate an increasing ability to accept sulfur
in the form of acid.
- An acid equivalency credit, recognizing that the cost of manufactur-
ing acid is avoided by a customer who purchases acid rather than
elemental sulfur.
Lower bounds on certain supply likages in order to simulate the
effects of marketing strengths (e.g., captive terminals in the
net demand regions) and captive use of sulfur by the suppliers (e.g.,
the manufacture of P fertilizers by the producers of Frasch sulfur).
The output data from the LP program include a matrix of extra-
regional suppliers and net demand regions which shows how demand was
filled (i.e., who sold to whom), and what the delivered cost was. If no
sales were made by a particular supplier to a particular demand region,
the required reduction in minimum delivered cost (and, hence, in F.O.B.
value) for a sale to occur is printed in the matrix. Other output programs
tabulate the sales by and netback to each supplier, and the calculated
delivered value of S in each region.
902
-------
Each LF case calculates the value of an incremental unit of
supply rather than of supply increments of any size. However, parametric
cases involving substantially different quantities of abatement S. supply
make it possible to obtain value estimates by interpolation. The calculated
values are considered to be maxima because new suppliers would have to
shave their price in order to "buy into" an existing market. These maxima
are referred to later as "maximum delivered values" (M.D.V.). AS a first
approximation, it is postulated that a recoverer of abatement S in
elemental form might expect a plant netback of about $10/LT less than the
M.D.V. The $10/LT represents a composite of transportation and marketing
costs (including price-shaving to buy into the market). Actual transporta-
tion and marketing costs are likely to vary greatly from one specific
situation to another. Thus, the deduction of $10/LT from M.D.V. to obtain
the F.O.B. value (F.O.B.V.) of abatement S should be used with caution.
PRICE IMPLICATIONS OF SUPPLY/DEMAND PROJECTIONS
The computer calculations permitted the construction of correla-
tion charts for the individual regions of the continental U.S. Examples
are given in Figures 5 and 6. Each chart indicates:
(a) The variation in net regional demand with increasing regional
supply of abatement sulfur in useful form.
(b) The maximum delivered value (M.D.V.) of abatement sulfur
corresponding to a given level of supply in useful form.
In addition, a rough estimate of the F.O.B. value (F.O.B.V.) of abatement
S in useful form is shown on the charts. In most cases, F.O.B.V. is
M.D.V.-10 ($/LT), i.e., it is merely the M.D.V. minus an arbitrary delivery
903
-------
cost of $10/LT. In a few cases, the F.O.B.V. refers to net-back esti-
mated for abatement S shipped to another region. There are also cases
in which it was not possible to estimate either an M.D.V. or F.O.B.V.
No charts were prepared for the year 2020 because the results
obtained, while similar to those for 2010, are judged to be too sensitive
to the assumptions made. Each figure is a composite of six charts
covering the years 1975 through 2010. All of the charts have certain
features in common:
• Regional abatement 8 in useful form is plotted as an abscissa.
• Net regional demand is plotted as an ordinate, with the scale at
the left.
• M.D.V. and F.O.B.V. are plotted as ordinates, with a common scale
at the right.
• The numbers 0, +5, +10 and +15 refer to parametric demand (as
discussed in the complete report, but not here).
• All volumes are expressed in million LT of S equivalent.
• All values are expressed in $ per LT of S equivalent.
Chicago Region (Figure 5)
This region has a significant demand for sulfur but it has an even
greater potential for producing abatement S. Markets for all of this
potential cannot be conceived. Hence, all of the M.D.V. and F.O.B.V.
curves have sharp downturns. A calculated F.O.B.V. curve is shown for 1990,
based upon the assumption that Chicago region sulfur will be moving to
extra-regional markets by then. In retrospect, the contrast between the
904
-------
1985 and 1990 charts suggests that a turn around may not come so quickly.
However, a national stockpile in the Chicago region could be a spur to
exports and could support the conditions projected in the 1990 chart.
Tampa Region (Figure 6)
The net demand in the Tampa region greatly exceeds its abatement
potential. But most of the demand involves captive production of acid
from purchased elemental S. The charts in Figure 6 suggest that reason-
able F.O.B.V.'s may be possible for local deliveries of acid in the
Tampa/Bartow area.* The downtrend in M.D.V. for the Tampa region with
increasing supplies of abatement S in useful form is due to the general
level of recovery in the U.S. not to the impact of the Tampa region's own
abatement potential.
Implications
The years 1980 and 2000 may be used to illustrate the results of
the computer calculations. The choice permits significantly different
situations to be contrasted. In 1980, production of W. Canadian sulfur is
expected to be at peak levels and backs topped by huge stockpiles, world
markets will be under intense supply pressure, and the fitting (and
retrofitting) of abatement systems in the U.S. should be.beyond the
demonstration stage and into widespread use. In contrast, two decades
later, W. Canadian S is expected to be a minor factor in world markets,
S recovered from petroleum refining will be of major importance while, in
the U.S., the policies applied to abatement S in prior years will have
determined the shape of the domestic sulfur and sulfuric acid industries
and their relationship to world markets.
* Direct negotiations between potential producers of abatement acid acid
users of fertilizer acid would seem to be required.
905
-------
Results that illustrate the conditions projected for the years
1980 and 2000 are given in Tables 2 and 3. The overall sulfur demand
of the continental U.S. is summarized in the top segment of each table.
This is followed by estimates of demand in two of the most important
regions of the Model, Chicago and Tampa. In both cases, sulfur supplies
from within the region are deducted to give the net regional demand.
These sulfur supplies exclude abatement S in useful form.
The third segment of each table shows cases that illustrate
the effects of different levels of abatement combined with different
percentage recoveries in useful form and, hence, different quantities of
abatement S in useful form.
The fourth segment of each table shows the net regional demand
for the Chicago and Tampa regions, after subtracting the pertinent quanti-
ties of abatement S in useful form from the base case of net demand. The
reason that the net demand in cases (B) and (C) is the same as in case (A)
is that the two former cases make the assumption of no recovery of S in
useful form. Even though cases (A), (B) and (C) represent different levels
of abatement, they are identical in terms of the assumption that no useful
sulfur is recovered. This implies different levels of recovery in throw-
away form. However, the latter has no direct effect on the value estimated
for abatement sulfur.
The fifth segment of each table shows the estimated value of
abatement S corresponding to each quantity of abatement S in useful form.
Here, it must be pointed out that, for both years, cases (A), (B) and (C)
lie outside the area actually investigated by computer cases. This is
because it is considered an unrealistic assumption that no abatement S
at all will be recovered in useful form. With no useful abatement S in
906
-------
the year 2000, the M.D.V.'s of $31+ and $30+ (obtained by extrapolation from
calculated cases) might be somewhat higher. Additionally, some switching
to processes not requiring sulfur would be expected since cases (A), (B)
and (C) correspond to a net deficit in supplies for the continental U.S.
It may seem somewhat of a contradiction to have an M.D.V. for
abatement S if none is assumed to be recovered in useful form, as in cases
(A), (B) and (C). However, the M.D.V. applies, conceptually, to the first
units of abatement S that would be recovered in useful form.
For 1980, case (D) suggests an M.D.V. only slightly less than for
cases (A) - (C) in both regions. However, a marked difference should be
noted for case (E). Here, a marginally lower value is estimated for the
Tampa region while an indeterminately low value is estimated for the
Chicago region. The explanation is that abatement sulfur has hardly changed
the net demand of the Tampa region whereas, in case (E), it has produced
a condition of gross oversupply in the Chicago region, namely a net demand
of minus 0.14 million LT. Because of general conditions of oversupply
corresponding to case 1980 E, no extra-regional outlet is envisaged for the
Chicago region's surplus of 0.14 million LT.
The drop off in estimated M.D.V.'s for the Tampa region is attri-
butable not to the small quantities of abatement S in useful form assumed
to be produced within the region but to the corresponding quantities produced
in other regions.
For the year 2000, it will be seen that both cases (D) and (E)
represent a surplus of supply over demand in the Chicago region. In spite
of this, significant M.D.V.'s are estimated for both cases. The explanation
is that, in contrast to the 1980 cases, extra-regional outlets for Chicago
907
-------
region sulfur are anticipated, i.e., the Chicago region is expected to be
an "extra-regional supplier" by the year 2000. Nevertheless, if the
percentage of abatement S recovered in useful form were to be only slightly
greater than in case (E)— e.g., 60% instead of 50% in useful form — then
the Chicago region would become grossly oversupplied and the estimated
value of abatement S in useful form would be indeterminately low.
RELATIVE VALUES OF SULFUR RECOVERED IN ACID AND ELEMENTAL FORMS
U.S. consumption of sulfur is mostly in the form of acid. Thus,
recovered sulfur has a different value depending on whether it is recovered
as acid or as elemental sulfur. Theoretically, acid has the higher value
because it means that the customer does not have to bear the cost of manu-
facturing it. However, the cost of transporting acid is at least three times
agreater than for elemental S per ton of S-value. In consequence, the
greater value of recovering acid rather than elemental sulfur is critically
dependent on assured, local markets for all of the acid produced. Unless
these conditions are fulfilled, the value of acid may be low and may even
be negative. Elemental sulfur can be stockpiled; in most cases at small
cost. Sulfuric acid can not be stockpiled.
At the end of 1965, just over half of U.S. sulfuric acid capacity
was in 6 states: Florida, Texas, New Jersey, Illinois, California and
Louisiana. Today, the industry is even more concentrated on the Gulf
Coast and in Florida, with a relative loss of capacity in the Midwest. The
geographical trends reflect the importance of Florida's pebble phosphate
deposits and the development of chemical industry on the Gulf Coast.
Most of the acid used to manufacture P fertilizers is produced
captively. This is true of a significant percentage of industrial acid
908
-------
as well. Some of the acid manufacturers also have captive production of
S values. This applies not only to the special case of acid recovered
by smelters but also to combinations of:
• Frasch S production and P fertilizer manufacture.
• S recovery from oil-and-gas operations and F fertilizer
manufacture.
• S recovery by chemical companies and manufacture of industrial
acid.
The structures and geography of the elemental sulfur and acid
industries will make it difficult for abatement acid to enter the market.
The willingness of existing acid marketers and captive users to offtake
abatement acid is necessary if a significant outlet is to be developed.
The incentives for such offtake have not been clearly established. Currently
the acid manufacturers, particularly those who merchant industrial acid,
stand to benefit if abatement S were to enter the market in elemental
form but to lose if entry were to be as acid. On the other hand, a
significant amount of old acid plant capacity is being shut down and this
may provide opportunities for abatement acid to enter the market.
The acid equivalency credit, discussed in the computer program,
is a device making it possible to deal with both acid and elemental sulfur
in the same calculation. The credit was changed with time to reflect the
difference between:
(a) having to shut down existing acid manufacturing capacity
in order to be able to purchase merchant acid
and (b) purchasing acid incrementally rather than building new acid
manufacturing capacity.
909
-------
The structure of the U.S. sulfuric acid industry suggests that,
near term, it will be essential for many power stations to recover or
remove sulfur in forms other than acid. The market for S02 is small,
and that for ammonium sulfate is declining. Hence, apart from using low
sulfur fuels (if available), the only broadly applicable choices appear
to be elemental sulfur and waste gypsum. This is not to say that acid
recovery systems will not be useful, but it does say that the larger
part of the current problem will require another solution.
910
-------
TABLE 1
FORECAST BASIS
• Supply
(1) PROJECTIONS OF I (2) CONSTANT $ GNP
(3) POPULATION
(2) CNP/ENERGY CORRELATION + (2) —*»(5) ENERGY FORECAST
(6) NUCLEAR AND HYDRO ENERGY FORECASTS
(5) - (6) —» (7) FOSSIL FUEL ENERGY FORECAST
(8) BREAKDOWN OF FOSSIL FUEL ENERGY BY SOURCE
(7) & (8) —M 9) NATURAL GAS FORECAST
—V(10) PETROLEUM FORECAST
—+(11) COAL FORECAST
(12) NATURAL GAS SULFUR CONTENT & (9)—*(13) NAT. GAS S RECOVERY POTENTIAL
(14) PETROLEUM SULFUR CONTENT & (10) —»(15) PETROLEUM S RECOVERY POTENTIAL
(16) COAL SULFUR CONTENT & (11) —*(17) COAL S RECOVERY POTENTIAL
(18) % RECOVERY IN USEFUL FORM FROM NATURAL GAS
(19) 7. RECOVERY IN USEFUL FORM FROM PETROLEUM
(20) 7. RECOVERY IN USEFUL FORM FROM COAL
(13) & (18) —+(21) S RECOVERED FROM NATURAL GAS
(15) & (19) --*(22) S RECOVERED FROM PETROLEUM
(17) & (20) —*(23) S RECOVERED FROM COAL
(24) PROJECTION OF S RECOVERY FROM SMELTERS, PYRITE AND OTHER SOURCES
(25) PROJECTION OF FRASCH S AVAILABILITY
(21) & (22) & (24) & (25) —*(26) TOTAL S SUPPLY (EXCLUDING STOCKPILES)
• Demand
(27) POPULATION & GNP PER CAPITA/FERTILIZER S DEMAND/RELATIONSHIPS
(2) & (3) & (27) —» (28) FERTILIZER S DEMAND
(29) GNP/INDUSTRIAL S DEMAND CORRELATION
(2) & (29) "* (30) INDUSTRIAL S DEMAND
(28) + (30) —* (31) TOTAL S DEMAND
• Supply/Demand Balance
(31) - (21) - (22) - (23) - (24) —* (32) NET FOREIGN DEMAND, INCLUDING ABATEMENT S BUT
EXCLUDING FRASCH SUPPLY
(31) - (21) - (22) - (24) ---%(33) U.S. NET REGIONAL DEMAND, EXCLUDING FRASCH
SUPPLY AND ABATEMENT S
Note: The symbol & signifies the joint consideration of one factor
with another, not a simple arithmetic sum of the factors.
911
-------
TABLE 2
ILLUSTRATION OF U.S. SULFUR SITUATION IN 1980
• Continental U.S.
Fertilizer S Demand
Industrial S Demand
Total S Demand
Million LT
7.1
6.7
13.8
• Demand in Representative Regions
Gross Demand
Regional Supply (excl. Abate. S)
(A) Net Regional Demand
Chicago
1.66
0.59
1.07
TampA
3.19
0.10
3.09
Illustrative Cases of Abatement Supply
Total SOX
Emitted
(B) Same as 1970
(C) 40% Less Than 1970
(D) Same as 1970
(E) 40% Less Than 1970
• Net Regional Demand After
(B), (C)
(D)
(E)
% of Abate. S
In Useful Form
None
None
50
50
Inclusion of Abate.
• Estimated Maximum Delivered Value (M.D.V.) for
(A), (B), (C)
(D)
(E)
In de terrain
Quantity of Abate. S
In Useful Form (10<)
Nil
Nil
0.49
1.21
S in Useful Form
1.07
0.58
-0.14
Abate. S ($/LT)
26
25
ately Low
Nil
Nil
0.02
0.06
(106 LT)
3.09
3.07
3.03
23
22+
21
Note: Estimates of Value are in 1970 constant dollars.
912
-------
TABLE 3
ILLUSTRATION OF U.S. SULFUR SITUATION IN 2000
• Continental U.S.
Fertilizer S Demand
Industrial S Demand
Total s Demand
Million LT
11.0
14.2
25.2
• Demand in Representative Regions (Million LT)
(A)
Gross Demand
Regional Supply (excl. Abate. S)
Net Regional Demand
Chicago
3.53
2.92
0.61
Tampa
4.20
0.30
3.90
• Illustrative Cases of Abatement Supply
Total SO,
Emitted
(B) Same as 1970
(C) 40% Less than 1970
(D) Same as 1970
(E) 40% Less than 1970
% of Abate. S
In Useful Form
None
None
50
50
Quantity of Abate S
In Useful Form (106LT)
Nil
Nil
2.35
3.50
Nil
Nil
0.13
0.20
Net Regional Demand After Inclusion of Abate. S in Useful Form (106 LT)
(B), (C) Same as Base Case (A)
(D)
(E)
0.61
-1.74
-2.89
3.90
3.77
3.60
Estimated Maximum Delivered Value (M.D.V.) for Abate. S ($/LT)
(A), (B), (C)
(D)
(E)
31+
26
21
30+
29
28
Note: Estimates of Value are in 1970 constant dollars.
913
-------
Figure 1
WORLD MODEL
y////////
UAPAN/
tCOMMUNIST ASIAE
FAR EAST
EXCLUDING JAPAN
CANADA:
^UNITED STATES
LATIN
AMERICA
EUROPEAN O.E.C. D/
:u. s. s. R.:
EASTERN EUROPE^
MIDDLE EAST
AFRICA
LEGEND
ORGANIZATION FOR ECONOMIC CO-OPERATION AND DEVELOPMENT
COMMUNIST COUNTRIES
"DEVELOPING FREE WORLD"
-------
Figure 2
VI
(7)
SEATTLE
(8)
SAN FRANCISCO
LOS ANGELES
(8)
NORTH AMERICAN MODEL
(fi]
OMAHA
(1)
BOSTON
(4)
CHICAGO (DETROIT)
(5)
MEMPHIS'
NEW ORLEANS/
(2)
(BUFFALO) NEWARK
TAMPA
(3)
NORFOLK
ROTTERDAM
/
COATZACOALCOS
/
LEGEND
(/)-(II) DESIGNATE U.S. REGIONS
# SOURCE OF "EXTRA-REGIONAL SUPPLY"
IN LATER YEARS OF FORECAST
( ) SUB-REGIONS
EXTRA-REGIONAL
ACID SUPPLIER
-------
Figure 3
U.S. REGIONS
vo
)
\
VTRO.
EAST NORTH
CENTRAL OHI
WEST NORTH
CENTRALS __
~]uTAH"~f
I
/ MOUNTAIN
i
-------
Figure 4
COMPUTER CALCULATION
FT,
l_
^
> INPUT PROGRAMS
SUPPLY
FILE
1
PROGRAM: RRB
i
f SUPPLY
J CARDS
1
L.
16
— ii
•^
DEMAND
FILE
PROGRAM: RRB is
i
C DEMAND
]r CARDS
' CALCULATION PROGRAM
I
1—
^^
"III
JT
TRANSPORTATION
COST FILE
.
FOB
COST FILE
— h-
PROGRAM: RRB 15
r ACTIVITY
CARDS
1
JU
PROGRAM
L
f1
I^LP/CONTROL
| CARDS
1
III
:LP-MOSS
P SOLUTION
CARDS
LP OUTPUT OPERATIONAL
PROGRAMS
\
P
ROGRAM:RRBOT
i
LP OUTPUT
MATRIX
J
1
1
0
]
SALES
FILE
i
.
«
.
VALUE
FILE
1
ACID
EQUIVALENCY
VALUE FILE
1
("" COST
J CARDS
1
1
I
1
J-
1
1
1
NET- BACK
FILE
1
1
-------
•£>
O
UJ
c;
o
o
o
u.
o
LJ
(3
Figure 5
CHICAGO REGION
1,4
1.2
1.0
0.8
0.6
0.4
0.2
NIL
-0.2
-0.4
-0.6
-0.8
-1.0
-1.2
-1.4
2
1
NIL
-1
-2
-3
-4
(1975)
(1950)
M.D.V. _-_
• — — _ "" """"*•»»
^ "v M.D.V. N
^F.O.B.V. \ ^ ^
"^"""^ F.O.B.V.N
*
1 (I
T I
C1985)
I I
(1990)
F.O.B.V.
25
20
15
10
5
0
0 0.2 0.4 0.6 0.8 0.5 1.0 1.5 2-0 2.5
I
25
20
15
10
5
0
_)
>
2
NIL
-2
-4
-6
-8
-10
+10
+5J
-0
F.O.B.V.
\ \
12000) +15
+10
0
riii
(2010)
30
25
20
15
10
5
0
8 10
REGIONAL ABATEMENT SULFUR RECOVERED
IN USEFUL FORM (106 LT>
918
-------
sO
o
o
LU
Q.
Figure 6
TAMPA REGION
5 o
I I
(1975)
1 I
(1980)
(1985)
(1990)
+5
F.O.B.V.
1 1 1 1 1
1
1
0 0.1 0 0.1 0 0.1 0.2
0.1 0.2
U_
O
UJ
o
§ 5
LL
S 4
f-
LU
2 3
0
I T
(2000)
(2010)
- M.D.V.
- F.O.B.V.
M.D.V.
F.O.B.V.
1
1
0.2 0.4
1
0 0.2 0.4 0.6 0.8
REGIONAL ABATEMENT SULFUR RECOVERED
IN USEFUL FORM (106 LT)
25
20
15
10
5
0 P
LJ
35
30
25
20
15
10
5
0
919
-------
NEW USES FOR SULFUR-
THEIR STATUS AND PROSPECTS
by
H. L. Fike, Director of Industrial Research
J. S. Platou, Director of Information
The Sulphur Institute
1725 K Street, N. W.
Washington, D. C.
921
-------
Hew Uses for Sulfur - Their Status and Prospects
H. L. Fike and J^v S^ JPlatou2
Introduction
Prom statements and forecasts on the energy needs of our nation
and the means available to fill these needs, it appears certain that
high-sulfur coal will continue to pla^ an important part in the
foreseeable future. The Air Quality Criteria which are now being
enacted and which are due to be fully implemented by 1977 make it
equally certain that all of the sulfur content of this coal will riot
be allowed to escape into the atmosphere. Thus, it appears highly-
probable that large tonnages of sulfur will be recovered in one way
or another from high-sulfur coal.
Two approaches are presently considered possible to prevent sulfur
oxide emissions from coal: stack gas scrubbing and coal gasification.
The majority of the stack gas scrubber technologies near commerciali-
zation recover the sulfur values as gypsum or sulfuric acid. While
these are acceptable intermediate term solutions, the problems
associated with spent gypsum disposal and the general transportation
and marketing problems with sulfuric acid make them unattractive and
even unacceptable over the long run. Recovery as elemental sulfur
appears to be the ultimate solution, from the point of view of
marketing, storage, and transport of the product. Also, the processes
under development for coal gasification generally recover sulfur in
the elemental form. The following discussion is therefore restricted
to the question of finding a market for the elemental sulfur recovered
from coal as a result of air pollution abatement regulations.
The Magnitude of the Problem
Let us first put the recovered sulfur in perspective in relation to
tonnage, time, and cost. It is unlikely that large tonnages of sulfur
from stack gas scrubbers or coal gasification will reach the market
before the 1980's. The first demonstration plants for recovery of
elemental sulfur have recently been completed or are still under.
construction. It is likely that they will have to operate successfully
for a year or two before the technology achieves general commercial
acceptance. This will then be followed by an estimated three-year
period between placing the order and full operation. However, by the
mid-eighties it is quite conceivable that 500 million tens of high-
sulfur coal per year will be consumed in the U.S. in power ntations
and/or coal gasification plants equipped with scrubbers recovering
sulfur. Assuming an average sulfur content of 3$, about 15 mill?'-::..!
tons sulfur per year could theoretically be recovered.
The disposal of this amount of sulfur becomes of considerable economic
importance. If it can be sold at a price of $25/tcri f .o.b. the
Paper presented at the Flue Gas Desulfurization Symposium, New Orleans,
Louisiana, May lk-17, 1973.
o
Directors of Industrial Research and Information, respectively, She
Sulphur Institute, 1J25 K Street, N.W., Washington, D.C. 2COOu.
922
-------
power plant*, the U.S. utilities industry could realize an annual
income of about $375 million, to set off against the capital and
operating costs of desulfurization equipment. Conversely, if no market
exists for this sulfur, the storage of large quantities in heavily
populated areas may cause its own pollution problems. Furthermore,
the costs associated with storage of sulfur could amount to $2-$5 par
ton per year—representing an additional cost to the industry.
The Sulfur Market
What are the possibilities of marketing the large quantities of
sulfur which will be recovered from stack gases or coal gasification?
To answer this, let us look at the present and probable future situation.
In 1972, Free World production of elemental sulfur reached 21 million
tons. This exceeded consumption by over 1 million tons. In addition,
nearly 2 million tons were imported from communist countries and,
as a result, sulfur inventories increased by over 3 million tons.
Over half of the sulfur presently produced is recovered from natural
gas or petroleum. The demand for natural gas and low-sulfur petroleum
products will accelerate the recovery of sulfur from these sources.
The traditional uses of sulfur, that is, in the fertilizer, chemical,
and other industries, are not expected to grow at an average of more
than U-55W annually. Most forecasts agree that production of recovered
sulfur will grow at rates considerably in excess of this. Thus, even
without any sulfur from coal, the oversupply situation can be expected
to continue in the 1980*3. If recovery from stack gas or coal gasifi-
cation reaches the estimated levels, sulfur production in North America
in the mid-1980's would outstrip consumption by a substantial margin.
Under these circumstances, the majority of utilities would be unable
to sell substantial quantities of their recovered sulfur in the market.
New Uses
There are new, potentially high-tonnage uses for sulfur which,
if developed commercially, can absorb the sulfur that will have to be
produced if the current air quality standards are to be met. In the
remainder of this paper, we will discuss with you the more promising
of these new uses.
A word here about the agricultural uses of sulfur. Sulfur is essential
in plant and animal nutrition, and is valuable as a soil amendment and
pesticide. The Sulphur Institute is working actively with agricultural
scientists and the fertilizer industry to promote and develop the uses
of sulfur in agriculture. The potential market for these uses in the
U.S. is estimated at between 1.5 and 2 million tons per year above the
present use, but as this is, strictly speaking, not a "new" use but an
expansion of an existing one, we will not discuss it further in this paper.
*This figure is given for illustrative purposes only and does not imply
a recommendation or forecast of the sulfur selling price.
923
-------
To alleviate the supply/demand imbalance and to "be commercially
acceptable, a new use for sulfur should ideally satisfy the following
criteria:
1. The potential tonnage use, realistically assessed, must be
large, at a minimum several hundred thousand tons per year.
2. The time and cost needed to develop the use must be reasonable.
3- The economics should be favorable, to attract industry
interest and capital.
U. The new use must be ecologically acceptable, i.e., it must
not create any pollution problems of its own.
The Sulphur Institute has carefully considered a great number of possible
new uses for sulfur and its compounds, including sulfuric acid and
sulfur dioxide. Those uses which in one way or another utilize the
mechanical properties of elemental sulfur appear to have merits which
match these criteria most closely. Our discussion here will be limited
to these uses.
Pure elemental sulfur is familiar to most people as a rather brittle,
yellow solid. However, it has a number of interesting mechanical
properties, not all of which are, as yet, fully understood. These
properties are dependent upon the time-temperature history of the
sulfur and can be greatly modified by additives. Conversely, when
sulfur is added to or incorporated into other structural materials, the
mechanical properties of these materials are often improved.
The following examples will demonstrate some of the commercially
interesting utilizations of these properties.
Sulfur-Asphalt Paving Materials
Processes for improving the properties of asphalts by treatment with
sulfur were first introduced over 100 years ago, but have generally
not become commercially successful. Most of this early work had as
its objective to chemically react sulfur with asphalt. The new
technology described below relates to the physical addition of sulfur
to asphalt-aggregate mixtures. Chemical reaction between the sulfur
and asphalt is deliberately minimized to prevent the evolution of
which generally accompanies the reaction of sulfur with hydrocarbons.
The composition of sulfur-asphalt paving materials varies considerably,
but a typical material consists of about 13$ sulfur, 6% asphalt, and
81$ sand, by weight. This material is interesting for two reasons:
1. It has certain technical advantages over conventional paving
materials. In general, addition of sulfur to asphalt paving
material can increase durability, strength, and resistance to
water. Also, there are indications that the material may be
particularly suitable for direct application on weak subgrades,
without prior excavation and backfilling.
2. In many areas of the country (Gulf Coast region, Pacific
Northwest, Great Lakes region) good quality aggregate is
becoming scarce and increasingly expensive. Sulfur-asphalt
permits the use of low-cost, widely available sand instead
of aggregate.
924
-------
The composition and coat of sulfur-asphalt paving materials compares
with those of conventional asphalt paving materials approximately
aa follows:
Ibs/cu ft % by wt ^ "by vol cost (£/cu ft
Conventional asphalt material*
Asphalt 8 6 13 14
Aggregate, crushed & screened 135 94 81 27
Air voids — -- 6 zz.
Total ilj.3 100 100 kL
Sulfur-asphalt material*
Asphalt 8 6 13 14
Sulfur 17 13 iU 21
Sand 105 8l $5 2,5
Air voids — — 8 —
130 100 100 3T-5
^Assuming asphalt at $35/ton, aggregate at $4/ton, sand at $0.50/ton,
and sulfur at $25/ton.
There are indications that it may be feasible to use lower grade asphalts
in the sulfur-asphalt mix. This would further improve the economics
of this material.
Approximately 22 million tons asphalt are used each year in the United
States for road building. For the non-communist world as a whole, the
corresponding figure is about 45 million tons. Any sizable penetration
of sulfur-asphalt into this market clearly could consume multi-million
tonnages of sulfur.
Further work is needed in several areas before sulfur-asphalt paving
materials can become a commercial reality. Although HXS formation
during manufacturing and transport does not appear to be a problem, it
must be studied further before the technology is released for general
use. Durability comparisons of sulfur-asphalt with conventional materials
require considerable time, and the introduction of new paving materials
into general use is often a slow process.
The most extensive work on sulfur-asphalt paving materials has been
carried out by Shell Canada, Ltd., in British Columbia and Ontario.
Trials have also taken place in Prance, Japan, and Scandinavia. The
Sulphur Institute has recently concluded an agreement with Shell whereby
the development of sulfur-asphalt pavements in the U.S. will be carried
out under a Joint government-private industry program. Initial develop-
ment vork is now under way at the Texas Transportation Institute,
Texas A & M University. Following preliminary familiarization vork
this ; ear, road tests are scheduled for 1974.
The sulfur-asphalt paving materials so far developed do not result in
a decreased consumption of asphalt per unit of road. The existing
and projected shortage of domestically produced petroleum may make it
necessary to limit the consumption of petroleum products, including
925
-------
asphalt, where satisfactory substitutes exist or can "be developed. As
large quantities of sulfur become available from increased use of
doaestic coal., a partial substitution of sulfur for asphalt would seem
to make sense both from the national economy and ecology points of view.
There are indications that satisfactory sulfur-asphalt materials containing
less asphalt may be possible. If the asphalt is eliminated completely,
the result is a sulfur-sand concrete, and we will discuss this next.
Sulfur Concretes
Mixes of molten sulfur with sand and/or aggregate set on. cooling to
form concretes with properties equal or superior to those of Portland
cement-aggregate concretes. Sulfur concretes set and reach full
strength in a matter of minutes, as opposed to one or more days for
the conventional materials. There is also the advantage that the large
quantities of water required for conventional concretes are not needed.
Preparation of the mix is simple: the sand or aggregate (70$ by weight)
is pre-heated to about 325° P and sulfur (30$) is added. Mixing is
continued until all the sulfur is melted and the mix is then poured,
tamped, and left to harden.
Using this technique, tiles, blocks, bricks, and other structural elements
have been produced. In 1972, part of an experimental house was built
near Montreal by McGill University staff, using interlocking sulfur-
concrete bricks, without mortar or other joining material. Reports
indicate that the walls built using this technique are structurally
sound and weatherproof.
Sulfur concretes may be particularly attractive in countries with no
indigenous supply of Portland cement, but with a readily available supply
of sand and sulfur. Northern Canada, western U»3., and several Middle
Eastern countries fall into this category. Sulfur concretes may also
have application in remote areas where the cost of casting and curing
Portland cement concretes is excessive. Another potential use is as an
emergency patching material for highways.
Sulfur concretes are at an early stage of development. Their eventual
acceptance and the size of the market are matters for speculation.
However, the potential market is very large if the technique is
basically acceptable in housing and road construction. For example,
building a 2-lane highway suitable for heavy traffic would require
about 1,800 tons sulfur per mile.
Surface Bond Construction
Low-cost easily built structures -can be- erected l>y -aurface-bonding
concrete blocks with a sulfur formulation. This technique has
interesting potential applications in low-cost housing, storage
buildings, silos, etc.
The sulfur-surface bond construction technique-ia simple both in
concept and practice. It is applicable to structures utilizing
lightweight concrete blocks- or other- blocJt or -block-like materials.
976
-------
No mortar or Joining material is used between the blocks; they are
simply stacked dry, one upon the other, until the desired wall con-
figuration is achieved. A thin single coating of molten mixture,
consisting primarily of sulfur vith small percentages of glass fibers
and other additives, is applied over the outside and inside of the
wall. Within seconds, the molten coating solidifies and forms a
hard, impervious surface. If a co'lor other than yellow is desired,
pigments can be added to the formulation or the coating may be painted
over with conventional house paint. Walls constructed using this
technique are much stronger than walls using conventional masonry
construction. In particular, sulfur surface bond-constructed walls
have considerable strength in tension, whereas conventional masonry
walls primarily show strength in compression and very little in tension.
Using this technique, a building was constructed in 1963 at the South-
west Research Institute, San Antonio, Texas. Unskilled labor with
no previous experience in applying hot coatings of this type carried
out the work, using regular paint brushes to apply the sulfur.
Concrete block walls are highly porous and must be properly sealed
after installation. At the Southwest Research Institute, there are
conventional concrete block buildings, some of which are more than
15 years old, which have required sealing, caulking, and re-sealing
from time to time, and still leak when subjected to a hard driving
rain. No leaks have ever occurred in the sulfur building, nor have
any cracks developed. The building has not presented any problems and
as a consequence has not received any maintenance. No cracking,
spelling, or other deterioration of the sulfur coating has occurred.
The performance of the sulfur building over a period of ten years
under the climatic conditions prevailing in San Antonio has convincingly
demonstrated the practicability of the technique. Another sulfur-surface
bond building was recently completed at the U.S. Bureau of Mines
research facility at Boulder City, Nevada, using recently developed
spray equipment to apply the sulfur-fiber formulation.
The difficulty and cost involved in changing building codes and
practices probably preclude the near-term use of the technique for
human dwellings in the U.S. and other developed countries. However,
it would appear to have considerable potential in the construction of
warehouses, utility buildings, silos, and similar structures. A
proposal to evaluate the technique in Latin America and Africa has
been accepted in principle by U.S.A.I.D. The economics of the technique
appear favorable, and the advantages of higher strength and improved
water resistance should make it attractive in many applications.
Sulfur as a Coating Material
Sulfur formulations similar to those used in surface bond construction
are being tested as low-cost coating and repair materials. Sulfur
coatings can provide both chemical (corrosion) and mechanical (erosion)
protection. For coatings of maximum mechanical strength, glass or
other fibers are added to the formulation, but in many cases this
can be dispensed with and coatings consisting essentially of sulfur
vith small amounts of organic modifiers can be used.
927
-------
Sulfur coatings are being evaluated or considered for a number of
applications: protection of concrete structures in corrosive environ-
ments, stabilization and erosion control of earthworks and storage
ponds, and stabilization of mineral tailings piles.
Sulfur is also being evaluated as a repair material for cracks
in Portland cement and asphalt concretes, for example, in roads,
airport runways, and swimming pools. Several federal agencies recognize
that inexpensive coatings and repair materials can fill an important
need, and are investigating various applications.
A different, but potentially very important application of sulfur
coating relates to its use in fertilizer technology. In particular,
sulfur-coated urea (SOU) shows signs of becoming commercially accepted
for use on a number of agricultural crops in several regions. By
coating urea with sulfur, the dissolution rate of this nitrogen fertilizer
is considerably reduced. This results in more efficient utilization
of the nitrogen by plants and reduced loss of nitrogen to the environ-
ment by leaching and run off. Many other methods of providing
slow-release nitrogen have been suggested, but SOU appears to be the
only one that is cheap enough for general use. The Tennessee Valley
Authority has been pioneering the development and testing of SCU,
and is presently considering the construction of a full-scale plant.
A typical SCU product contains about 10-15$ sulfur. With world production
of urea for fertilizer use at about 15 million tons annually, this
clearly represents a considerable outlet for sulfur.
Sulfur Impregnation
Ceramic and other porous materials can be impregnated with sulfur,
thereby improving the chemical resistance and the mechanical properties
of the material. Bonded abrasive grinding wheels impregnated with
sulfur have improved strength; the sulfur also acts as a lubricant and
coolant during grinding. Sulfur impregnation of ceramic tiles results
in lower water adsorption, making the tiles frost resistant for external
applications. Impact and compressive strengths are also improved.
Corrosion of concrete sewer pipes is a serious problem in many localities.
Impregnating the pipes with sulfur has been estimated to extend the
life of the pipe by a factor of 10, and to double the flexural strength.
Impregnation of ceramic and concrete materials is generally carried
out by immersion in molten sulfur. Use of vacuum to remove air from
the pores of the material prior to immersion in the sulfur speeds
the rate of impregnation.
Research and development on sulfur impregnation is in its infancy.
Potential applications are numerous, particularly in building and construction.
Sulfur Foams
Using small amounts of additives, sulfur can be turned into a foam with
interesting properties. The foamed sulfur is lighter than water (typical
densities are in the range 0.2-0.5); the compreasive strength is in
the range 50 to several hundred psi, higher than typical organic
polymer foams. The foam shows excellent thermal insulation properties.
928
-------
One promising application of foamed sulfur is as a sub-base for highways
and airport runways in cold climates. Polystyrene foams are being used
experimentally to prevent freeze and thaw damage to highways and
runways, replacing costly and unreliable methods like deep sub-base
beds of stone or gravel. Foamed sulfur is being evaluated as an
alternative to polystyrene foams in this application. One inch of
polystyrene or sulfur foam is equivalent in insulating value to 24
inches of gravel. Sulfur foams have higher compressive strength than*
polystyrene, they are less expensive and can be foamed in place.
Construction of a 1,000-mile, four-lane highway would require about
too,000 tons of sulfur.
Several other potential applications of sulfur foams are being investigated.
Conclusions
By the early 1980's, it is probable that large tonnages of sulfur will
become available from utilization of high-sulfur coal through stack
gas scrubbing, coal gasification, or both. These tonnages, when added
to those resulting from other sources, will be considerably in excess
of sulfur demand. Because the electric utilities will be producing
sulfur at a large number of locations often unfavorably situated for
the sulfur markets, disposing of this sulfur in an oversupplied market
can be expected to be initially difficult, and eventually Impossible.
Stockpiling the recovered sulfur would impose an added financial burden
on the industry. However, if new uses were developed, this sulfur
could be marketed, resulting in added income for the industry,
partly offsetting the cost of desulfuriaation.
The potential new uses outlined above, could, if developed commercially,
absorb the foreseeable supply of sulfur. They fulfill the criteria
stated earlier to the extent that they are large-volume uses and they
are ecologically acceptable. They also appear to have favorable
economics, and the time and cost needed for commercial development do
not appear excessive.
It would be wishful thinking, however, to expect these new end-use
markets to materialize automatically, or that their development will be
eagerly undertaken by private industry. In most cases, the potential
financial rewards will be insufficient to entice firms which do not
have an interest in sulfur or fuels to commit funds. Up to now, the
burden of supporting the development of new uses for sulfur has been
carried by the main suppliers of sulfur, i.e., the Frasch producers and
the oil and gas companies. As new sources of sulfur become increasingly
important, the originators of the new supply will have to bear e. portion
of the burden to develop the new uses needed to absorb this
increased production. Several federal agencies, notably the U.S. Bureau
of Mines, have recognized, the implications of cutting down the
emission of sulfur oxides and are actively sponsoring or cooperating
in sulfur development programs. The electric utilities industry,
potentially the largest sulfur supplier in the country, has a special
responsibility in this regard. Although large-scale sulfur production
by the utilities may be several years in the future, now is the time
to research and develop the needed markets. We invite the utilities
industry, in its own interest, to actively support sulfur development
work, either through the Edison Institute, through cooperative programs
vith Federal agencies, or through The Sulphur Institute.
929
-------
PANEL DISCUSSION:
DISPOSAL AND USES OF BY-PRODUCTS
FROM FLUE GAS DESULFURIZATION PROCESSES
931
-------
PANEL DISCUSSION
A. V. Slack (TVA)— We have had an extended discussion of about all
the problems we can think of in. regard to disposing of the products that
we will be making when we go extensively to recovery or to ridding stack
gases of the sulfur compounds. We now have our panel assembled, including
Mr. Fike (K. L. Fike - The Sulphur Institute) co-author of the last paper;
so we are ready for any questions that either the panel or the audience
may wish to pose. I would like to ask Dr. Minnick (L. J. Minnick-IU
Conversion Systems) and Bill Taylor (W. C. Taylor - Combustion Engineering)
a question. If we are able to convert the sludge to useful construction
products (and it seems quite obvious that we can) what proportion of our
sludge production do you think that we could ultimately (say, within 10
years) convert to this useful purpose?
L. J. Minnick— I do not think I can project the finished statistics.
I think that I would like to comment to this extent. The utilization of
the technology which was described and illustrated is something that is here.
It is our mission, as far as IUCS is concerned, to take on projects as they
develop and really turn out to be the garbage man, to take the problem
off the hands of the utility entirely. That is our mission so that they
are not involved with the running of chemical plant or process. Now the
thing that dictates whether or not 1t is a good idea to do 1t, of course,
ts the economics; and the answer to your question really 1s that each
situation has Its own logistics. If you are going to consider stockpiling
aggregate and you have a place to stockpile it in an ecologically safe
way, fine. If there is a market for aggregate, you may sell enough
932
-------
of it to justify some costs, If It.1s simply disposal of a product that's
going to be put Into a landfill or land improvement project, what land
improvement project or what kind of land is available? If you are near
a strip mine and you want to rehabilitate it, then it makes a lot of
sense. We have looked at a lot of projects across the country. Now
as I say, I can not give you a statistical number; but I can assure you
that many, many of the situations that are opening up in the southwest
and eastern United States do lend themselves to the conversion system
approach. The quantities of materials that are involved, the very fact
that there is lot of it, make it much more economical. If someone says
that we have one million or two million tons of material to convert a year,
it makes a lot more sense than if you are just handling a small quantity.
So what the exact amount is, I can not answer that.
A. V. Slack— All I was looking for was a guess.
W. C. Taylor— Well, I agree with John (Minnick) in that the question
bears on many factors; but I see no reason why you could not
use TOO percent of the sludge, I said in my talk that in Germany now
they use 80 percent of all the flyash that they manufacture. About 80
percent of all of the construction of say less than 8-story buildings in
Europe now has a lot of the flyash in it; they are now digging up some of
the flyash that they buried to use in construction, But there are many
factors in this country that will determine whether or not we use large
quantities of the stuff. In my slide, I mentioned my one brief moment
of glory when we had thought that in the laboratory we had proved the
technical feasibility of utilizing this material in a number of areas.
Several different useful products pertained to building and highway con-
933
-------
struction. If you used it in sub-base for highways, you could use all
the material. If you look at the highway building program, you can see
that we use about 15 billion tons of stone in this country every year
for construction, But, like we said, the flyash brick that the University
of West Virginia came up with was supposed to be economic and it was
cheaper than the clay, It was lighter and cheaper to ship. Structurally,
it was sounder. Everything was better than the clay brick; but, you
could not get anyone to buy it. You could not get anyone to put up a plant
to make it. So this is the problem you have — the old inertia. A guy
has been getting his clay brick from a distributor for years. He has estab-
lished his rapport. He is not going to change all of a sudden just because
you say, "Look, I've got some bricks and I can give you 2 cents off
per brick." I think it is a big marketing problem. Frankly, we are now
looking at some throwaway processes because we feel that, until the marketing
problem is solved, we will be throwing it away in the beginning. But if
someone has a lot of money and wants to make a lot more money, we have a lot
technically proven processes that he can use and he can start buying some sludge.
A. V. Slack-" Are there any questions that the panel members now
would like to put to each other?
J. Ando(Chuo University)— May I ask a question of Mr. Minnick?
I think the use of the sludge for road construction is a very good thing.
I must say I have English trouble and I am afraid I may not understand you,
but would you say it Is mainly a matter of our economy? I think, it is
rather that if the material Is very cheap and if its use 1s technically
feasible, it is economically feasible.
1. J. Minnjck— Well, first of all, as we look at it, we do not attempt
to take credit for the sale of the product. We feel that the cost tif
934
-------
conversion is a disposal cost. Now one of the reasons that the aggregate
is useful is that It can be stockpiled. It is important that the power
plant stay on stream; it must run; and every minute of the day the
sludge is coming out and you have got to do something about it. Now if
you convert it to aggregate and stockpile it, in the construction
season the stockpiles are depleted; when the construction
season is oft, like in the wintertime or something of that sort,
you can still operate and just pile up the material. But whether or no,
you sell It is a function of the market. If you are in an area where
there is lots of cheap aggregate, you're not going to sell it. If you
are in an area where there is a depletion or shortage of aggregate, yes,
you can sell and sell it at a profitable price. And again, it's the
logistics of it, the individual situation. But in our approach to it,
we are looking at it as a disposal charge and then these other things
are credits that can come back and give you some economic trade-offs.
A. V. Slack— I have a question here from Bill Richardson of Bechtel who
is concernedi with about 6 mill ton tons of gypsum that we import each year
into this country. So the question
-------
to get this 95 or >90 percent gypsum. Here you can mine 99 percent pure
gypsum and you can float it down on a barge for practically nothing. It
would cost, it seems to me, in this country about 5 times as much to get
90 percent gypsum from sludge as it would be to float 1t down at 99 percent.
And unless we are talking reverse economics, I do not think even (though
we are now looking at wall board) that so far it looks like 1t is much more
expensive to try to recover it than to just mine it.
J. Ando—Yes, I believe we agree that byproduct gypsum cannot com-
pete with your pure cheap natural gypsum, it is true; but I think that
in certain areas byproduct gypsum may be used because you are still im-
porting some, And in Japan, until about 10 years ago, we had plenty of
byproduct gypsum from wet process phosphoric acid production, No one
used it and we were still importing because they had no experience using
ft; they thought the use of byproduct gypsum might give bad results for
cement and plaster board, I worked in the field and made some reliability
tests and showed It could be used, but no one believed 1t. Now all of
them are using it and now we have a shortage of gypsum, Every bit of
gypsum is now used. I think there is a similar situation in this country,
A. V. Slack—Of course, it does depend on what it costs to make
tfiis gypsum. Since there are various processes in Japan for making it,
I suppose we really need an economic comparison. Now may we have questions
from the floor.
John Salm.(Pioneer Service & Engineering)—On the matter of gypsum,
even the naturally occurring gypsum in the United States has various
degrees of effectiveness when used as a moderator or regulator in the
setting of Portland cement. I was wondering how much work has been done
936
-------
on testing the gypsums that are formed in these various processes to see
how reactive they are, how effective they are in actual use of Portland
cement? I talked with Dr. Ando in the intermission; he indicated that it
has been used successfully in Japan, but there have also been some
problems with the use of the artificially produced gypsum. I just wanted
to throw that out to find if there had been any actual research done in
that regard over here?
J. Ando—In Japan, phosphogypsum has been the problem if it
can be used for cement or not. Until about several years ago, cement
companies did not want to use it. The main reason is that some small
amount of P205 (phosphoric acid) and also a fluorine hindered the
cement setting. Since that time they have improved the process of the
production of gypsum so that now the P205 in the gypsum can be reduced
to below 12,1 percent and it can be used without any problem. Also/many
cement companies recently have tried to use byproduct gypsum from coal-
fired boilers produced by Mitsui in Omuta by the Chemico and other
processes: the Chiyoda process and the Mitsubishi-JECCO process. All
of the byproduct gypsum can be used v/ithout any problem. The only problem
is, if it contains too much moisture, it cannot be put into the cement
mill; but if the moisture is about 11-12 percent or less it can be used
without any problem. In the Mitsui case, if the flyash content in the
byproduct gypsum is less than about 10 percent it can be used in cement.
Ken O'Brien (R. W. Beck)—I have two questions. The first one is
«
to Dr. Minnick. I am wondering about the ratio of flyash to calcium sul-
937
-------
fite or sulfate. Mould you say that what comes out of an average coal-
fired power plant 1s compatible to make these building products or 1s
there likely to be a large surplus or a lack of one or the other? Jfy
other question Is to anyone who can answer It. Assuming flyash is one
of the products necessary to make a building product, what do you do on
an oil-fired plant? Involved in a particular project, we are oxidizing
to sulfate. It probably wouldn't do the job'because the landfill would
be under water.
L. J. Minnick^Both of those points were covered in the written
portion of my paper. But very briefly, the answer is "yes" to your first
question. For most coals (those with 3, 4, even 5 percent sulfur),
assuming you have 8, 10, 12 percent ash, you have the right proportion*
The only place where you get out of line is when you have a boiler that
produces a very low flyash; such as the centrifugal or,what do you call
it, cyclone boiler which produces most of the ash in the form of bottom
ash. In this case, you are short of flyash. In connection with oil,
the system that I presented depends on flyash; that's the magic ingredient.
And if you don't have flyash available somewhere, you don't get the
chemistry that was described.
Max Schmidt (The university of Wurzburg, Germany)—I want to
ask Dr. Platou (0. S. Platou, The Sulphur Institute), the last speaker.
Just one question. You showed us very nice slides on the use of sulfur
in the construction business. You showed us a house which was built
in Te*as about 10 years ago and another one just under construction in
Montreal. Now you are talking about sulfur, talking about millions of
tons of sulfur, but what do you mean by sulfur? It 1s not pure sulfur.
Can you tell us something about the additives, the cost of the additives
938
-------
compared with the elemental sulfur itself?
J. S. Platou^The sulfur mixtures, the sulfur formulations that we
used in the surface bond construction used in constructing these houses
are approximately 90 percent elemental sulfur, The remaining 10 percent
i.s composed largely of fibers; glass fibers have been used mostly up
to now and a very small amount of organic additives in an order of a
few percent. Does that answer your question?
M. Schmidt —I did ask for the cost average of the organic additives
or thiocols. What is the price of sulfur? Not what is the price of the
additives.
J. S. Platou—Harold (Fike), do you want to comment on that? I would think
that probably thiocols are not exactly what we would want to talk about now.
H. L. Fike — Over the years, Prof. Schmidt, we have looked at a
good many additives. As you recall, we did use many of the thiocols,
which cost roughly $1.00 per pound 8 or 10 years ago. Now we have prf-
marily concentrated on some unsaturated hydrocarbons, particularly
the dipentene, the terpene-type products, the dicyclopentadienes. We
have looked at various other crude aromatic unsaturated compounds which
come from refineries, some of which are available for 8-10 cents per
pound, we feel that if these additives are used in the amount of 3, 4,
or 5 percent they will still represent a rather large portion of the
costs of the sulfur coating. It will still be a coating which will be
perhaps slightly less than most of the standard asphalt coatings and much
less than the asphalt emulsion coatings. In some of the cases where we
have used this, we have been more or less competing against the epoxy
materials and, of course, there it was only a fraction of the cost.
939
-------
A. V. Slack—Two of our panelists may have to leave early and
I have a couple of questions I would like for them to cover. One to
Jerry Rossoff (Aerospace Corporation) and again I am concerned with
this dewatered sludge that we may produce and what happens to it when it
rains? Have you gotten into that problem much, Jerry?
J. Rossoff—Well, we have some limited data and we are working with
one sludge from the TVA Shawnee plant that shows it will absorb and re-
absorb water and bloat and hold it. So as far as rain goes, if
the sludge is not thick and you do not have any way to let it drain through
and carry the water off, it will absorb the water again,
J, W. Jones (EPA) —I would like to make one comment on that.
That particular material was allowed to settle naturally and then
air-dried; isn't that right, Jerry?
J. Rossoff— Yes,
J. W. Jones — There was no compaction or any treatment, any physical
treatment of it?
J. Rossoff—Right, there was no treatment whatsoever. It was allowed
to dry out. And it will reach a saturation point and will hold so much
and that is it, but it will regain water. I think that was the question;
if it rains, it will pick it up and bloat.
A. V. Slack — And let me get one more question in to Mr. Farmer (M. H.
Farmer, Esso Research) who may have to leave. In your model, did you
include what we mentioned earlier, the possibility of building a ferti-
lizer plant at a power plant or smelter and thereby affecting the economics
in the market?
M. H. Farmer— Well, in a way, yes. Of course, there were many things
940
-------
covered in the full report, a very heavy thing, that is not even in
the handout material that most of you probably have, The Western smelters
were considered in some detail and some of them, of course, particularly
those in the more northern copper mining states, are tied in with
fertilizer production. But those in the southern part of Arizona have
a particular problem. But the study did not ge't into local marketing
situations. We wish that we had been able to do so, but it was outside
of the scope of the program.
Fred Grozinger-(Potomac Electric Power Company}—I would like to ask
Mr, Farmer a question, I am somewhat confused about the overall practi-
cality and objectives of the report that you did for EPA. Could you
possibly explain if there's anything in this country in general? Or
is it going to be specifically a tool for EPA?
M. H. Farmer — Well, I hope that it will be of some benefit;to the
utility industry; but of course the terms of work, the scope of work,
was decided by the EPA, Working with Norman Plaks (EPA), we were able
to expand the original terms somewhat and I think we made it more useful
as the project developed. But it had a specific purpose to kelp them
in establishing their research priorities; I believe that it has had
some impact on those decisions already. But also it was my hope at least
that, by placing the industry in its world context, it would make it
easier for a utility perhaps or any other individual to fit his piece
of the puzzle into a broader context. Now, a lot of times a suggestion
is made, "Well, surely such and such a thing could be done," and the
suggestor does not realize what the problem really is in accomplishing
what he wants to do. Just one quick example, I read a suggestion
941
-------
some tfme ago that phosphogypsum should be used as a soil conditioner.
Well, of course, most of the fertilizer gypsum is produced in Florida
and most of the soil conditioning goes on in California, So the
transportation cost is a very, very important part in the utilization
of any low unit value material, I am sorry if I have not answered your
question.
F. Grozinger-~-That is fine. I was just a little confused on the
actual objective.
J. $alm—-0n the subject of utilizing sludge as a construction material,
what experience has there been with regard to the resistance of this
material to freezing and thawing cycles? Or to, say, gully washing
type rains in the Southwest? Have these materials actually been
tested for all of these requirements?
L. J. Hinnick"-Hen t yes. As I mentioned, this is sort of an
outgrowth that was done with Poz-o-Pac, The test procedures have been
established by Federal highway departments and ASTM and so on. They
are very well known, freezing and thawing being a case in point, and
the answer is "yes." They have been fully evaluated from that standpoint.
J. Salm — One other question, Mr. Minnick. On the expansiveness
of this material, is that a controlled expansion, like that of
cements that can either be expanded or . , ?
L. J. Minnick — Exactly. That Is the way to look at it, as an expan-
sive cement. Ettringite formation in Portland cement can cause dele-
terious reactions if the ettringlte forms after the cemfent sets. In the
case that we have here, the ettringite Is the setting process and it
forms during the hardening and therefore does not cause deleterious
expansions. The expansions are beneficial, just like you have with an
942
-------
expansive cement. Same idea.
J. Salm—Could I ask one question of Mr. Platou, please? On
the use of sulfur in the construction materials, you are talking about
yery high percentages of sulfur. What is the fire hazard? Would
OSHA agree with it?
J. S. Platou — It is a good question, sir. There are several aspects
of this, of course. Before going any further, I would just say that
additives have been developed, they are even patented, that will make
sulfur non-sustaining of combustion. In other words, they comply with
one of the ASTM standards. Self-extinguishing. Another aspeet is that
when you add a certain amount of aggregate or other non-combustible
material to the sulfur, this will of course also act as a heat sink
and again contribute to making the material self-extinguishing. This
is a problem that we are aware of, It is, of course, being evaluated
for these various uses. Do you want to add to that, Harold?
H. L. Fike — About the only thing I might add to this is the
thought that where many people think that they are going to be just
storing sulfur out in the country, it does pose a potential danger to
just store tremendous piles of it in the country. It could catch fire
there. And this is one of the reasons where I think the beneficial use
(for example, in a sulfur-asphalt road or a sulfur foam as an underlay
for a pavement material) would not really create any problem as far as
being combustible. But it is a problem which we would like to face up to
tnore directly than we have. In fact, we would like to come up with some-
thing which would actually make the sulfur non-burning, We have not done
this as yet. And it is not really a very promising endeavor.
943
-------
Jacques Dulin (Industrial Resources)—We have heard today discussions
of taking care of the waste products from the calcium systems (specifi-
cally, scrubbers) and also what to do with the sulfuric acid and the
sulfur. Tomorrow we are going to hear something in connection with
sodium systems and ammonia systems, I wonder if the panel could jump
ahead just a little bit and perhaps give a few moments of discussion
to the state of the art in connection with the disposal of sodium sulfite
or sulfate from those types of systems, There is one aspect that I would
like to mention for them to direct their attention to. We have heard
here specifically of scrubbers and not of baghouses and the use of reactive
filter aids in connection with the baghouses to produce (rather to re-
act with) S02» and to take it out of the air along with the particulates.
I wonder if they could direct their attention to the disposal of that
type of baghouse waste product material as well,
A. V. Slack — Well the question of sodium sulfate as a throwaway
product I suppose is best discussed by Dr, Ando and he covered that to
some extent but , . ,
J, Ando—Again, there is a big difference in circumstances.
Japan is a small country and surrounded by sea, We can just put it in
the sea. We can deliver what is going directly to the sea, so sodium
sulfate is no problem,
A. V. Slack — Would any member of the panel or any member of the
audience like to discuss the throwing away of sodium sulfate in the
United States?
944
-------
W. C. Taylor-»»Hell. I would like to express Combustionvs apologies
for not having considered that at all, He seem to be pretty well tied
up, looking at the calcium sulfate sludges. While I am speaking, I
jnfght clarify Combustion*s point of view on this in that we at CE are
really not In the waste disposal business. We thought that we had to
do this because at the ttme we began there was no one else worrying
about the disposal. And there was a lot of talk going on that we had
a potential problem, but no one seemed to be solving It. We thought
we had to do something, Once some of these other companies seem to
be coming up with, processes for disposing of the sludge, we would be
yery glad to let them handle our problem. We were only offering this
in-between service.
ft. L. Fike-—Mav I make one brief comment on that as well? The
Sulphur Institute is very interested in finding uses for sulfuric add
and S02. We just concentrated on the one effort today; but I think I
would have to say that over the longer term we have not been able to
see any future market for sodium sulfate in the United States* As
you know multi-million ton quantities are used today by the pulp and
paper industry. From what I can see from the general direction that
this industry is going and from the people I talk with who are quite
knowledgeable in this, I certainly get the suggestion that they are
going to non-sulfur using pulp and paper processes. Probably about the
time the utility Industries might have this available, it will not be
used to any great extent in North America.
Don 6ylfe-{ Atomics International)—I'd like to answer part of the
question of disposal of sodium sulfate in this country, a paatial
answer, anyway. We have a system that uses sodium carbonate that
945
-------
we are trying to promote. It turns out that in the western states where
the sources of the trona are located ( in general in such areas as dry
lakes in the California area), you can bring the spent product, the sodium
sulfate, back to such lakes, It is being done currently in Owens Lake
in Owens Valley, in connection v/tth an operation there. The California
Water Resources Board has told us that a similar process could be done
at Searles Lake. That is not a general answer on what you do with sodium
sulfate in all locations, but at least in the western U.S. It is one
possibility for an interim solution.
G. A. Hollinden— I would like to ask Dr. Ando one question. When
the team was over there in Japan last August, we were told by MKK that
they were going to use sodium sulfate as a soap binder and they showed
us a product of that. Do you know anything further on that, Dr. Ando?
J. Ando — Yes, they may use it, but actually they are
wasting their sodium sulfate to the sea. MKK is discarding the ....
The Wellman-Lord process waste water containing sodium sulfate is still. , .
G. A. Hollinden —Are they discarding any other products over
there too, in the sea, Dr. Ando?
J. Ando—Yes, MKK recently built another plant in which they
recover sodium sulfate. It is sodium scrubbing of waste gas.
H. L. Fike—• I believe there is a sizable quantity if you talk
in terms of millions of pounds of sodium sulfate added to detergent to
protect the aluminum in washing machines, but this would not utilize
the types of tonnages which the utility industry will produce. But I
think most detergent formulations in the United States contain 3 or 4
percent sodium sulfate if I recall correctly.
946
-------
J. Dulin — Seeing as to how we have heard a few plugs from people,
our company is of course very deeply interested in sodium systems since
we have a couple hundred million tons of the stuff to sell out in Colorado,
specifically a sodium bicarbonate mineral called nahcolite, We have
addressed ourselves to that problem and have found ways in which to in-
solubilize the sodium sulfate as an easily dewaterable granular precipi-
tate, which is a double salt of sodium sulfate and ferric sulfate. And
the material, you are not talking about 60 percent solids sludge,
you are talking about really a material that is 100 percent solids—the
water just drains fight out of it. The water will then be suitable for
the very simplest type of landfill. We can not say to you that it is
100 percent available today, because we are still doing research on it and
answering the hundreds of questions that people have in connection with it.
But it can be done. We are also handling a couple of other methods as well.
I think this should alert the industry to the fact that whereas there has
been a focus on calcium systems because of the end-product theoretical
insolubility, there has been an earlier recognition of the advantages of a
clearwater liquor type of scrubber or dry baghouses for use with sodium
systems. The problem with the sodium systems has always been, "What do you
do with the sodium sulfate or sulfite that you get at the back end?" We
think we have an answer to that and I would like to alert you to the
fact that sodium systems should be considered more carefully than they
have in the past.
H. W. Elder (TVA, Muscle Shoals)— I have a question for the Sulphur
Institute. Have you considered how there can be equitable competition in
a market where there's an oversupply between a regulated and a non-regu-
lated industry?
947
-------
J. S. Platou-—Since the Sulphur Institute has as its members both
the Frasch industry and the industry which recovers sulfur, such as natural
gas producers, I cannot comment on that question. I am sorry.
J. Ando— May I ask one question to Dr. Taylor? You showed me that sludge
is usable for the production of lightweight aggregate. How could you make it?
W. C. Taylor—- Yes, well, any product made from sulfate sludge that
requires sintering would have to have a sulfur recovery process with it.
We have considered that. At the time we were looking at beneficiation
processes, we looked at everything for which flyash had been used,
Among these were making lightweight aggregate, making sintered brick, or
anything like that. That is one of the reasons for our extensive thermal
studies of the material. And we found that you could not sinter the
material without recovering the sulfur. Again, at that time we were
considering a sulfur recovery process with a number of thermal sintering
processes for utilization of sludge, Once you get off. the S02 you are
right back to something like a lignite flyash. And you can make the
same type of sintered products that you could from the lignite ash,
Russ Eberhart (John Hopkins Applied Physics Lab)—Question for
the gentlemen from The Sulphur Institute. What kind of landfill
preparation, pitlining, or whatever would be necessary to stockpile
sulfur? That is, to protect (if any protection is necessary), ground-
water, and so on.
J. S. Platou^-Are you referring to possible oxidation of the elemental
sulfur into sulfate and getting it into groundwater?
R. Eberhart—Anything you can think of; that is, if a power company
were required to stockpile this sulfur because the market were not
available, what would they have to do in terms of preparing the ground
948
-------
or the pit in which the sulfur would be stored?
J. S. Platou— I do not know to what extent I am qualified to answer
that question as of now. It would of course depend very greatly upon
the location of the sulfur stockpile, whether under average conditions
or under conditions of high rainfall. Probably the people who would have
most experience in that field would be the Frasch industry who, of course,
maintained large stockpiles of sulfur. I do not really think I can answer
that question too well. Harold, do you have any comment?
H. L. Fike —Probably all I can do is to add to the problem rather
than to solve it. Sulfur is very susceptible to both chemical and bio-
logical oxidation, and as a result could cause a problem in certain areas
of the country. I think that we should also recognize that practically
all the sulfur sold in any large quantities in North America, particularly
the United States, is now delivered in liquid form. What they do now,
of course, is just put it in forms, vat it, and let it sit there. If it
were going to be sold, it would have to be remelted and put into tank
cars. Now to avoid that up in Canada, because they are no longer allowed
to ship solid sulfur through the port of Vancouver, they have gone into
a slating process. In other words, as the liquid sulfur comes, they put
it on a roll and actually make small slates out of it. They can load
this in the port of Vancouver and this minimizes the danger from dust
explosions which they had a real problem with in the past, I only point
this out because I think your question, as one gets down to it, would
require a good many hours of thought and conjecture and certainly a great
deal more knowledge than I possess.
A. V. Slack —WeTl. Harold, could you continue with that discussion
and cover the HgS emission problem? I have heard about 1t, but I do not
949
-------
know much about the details. I have heard that on some of the, what
do you call it, blocks of sulfur in Canada, they are considering having
to put hoods over them to recover the H2S. Is this true, or what is
the extent of the problem?
H. L. Fike—I do not know. I am trying to check into that.
You, as with many of these problems, perhaps get a different answer from
different individuals. However, I think it is generally accepted that
as it comes out of a Claus kiln, liquid sulfur will contain fairly
sizable amounts of hydrogen sulfide. In the liquid sulfur, it is more
soluble hot that it is cold, so when it does cool down and crystallize
you now have a potential problem of its emitting hydrogen sulfide. I
do not think this will be a problem in rural areas. I do not think the
concentrations will perhaps be so large. In metropolitan areas I imagine
it could be a problem, but I could not quantify it closer than that, Archie.
A. V. Slack —How does sludge smell? I am not sure.
John Raslle (Ebasco Services) —We have done some work on the disposal
of these waste sludges from a number of power plants. We have found that
there are in most cases, in fact all cases, concentrations of boron,
cadmium, nickel, and silver which far exceed drinking water standards.
In all of effluents from both wet particulate scrubbers and sulfur
dioxide absorbers, there seems to be a certain lack of experience in these
things. I just thought I might add that. Somebody asked a question
about lining. In the few plants that have either proposed or put in
wet scrubbers or sulfur dioxide absorbers, synthetic materials have been
required for the lining. Clay has been found to disintegrate under
the action of high calcium and magnesium salts contained in the scrubbing
liquor and it's not considered an acceptable lining material. There-
fore, you normally have to go to something synthetic. The Bureau of
950
-------
Reclamation has actually designed and 1s requiring the Installation of an
extensive monitoring system at the Mohave plant, just to make sure that
the synthetic lining does not fall or, 1f It does fall, to observe
when it fails. What I am trying to say, I guess, 1s that I do not think
there 1s going to be any acceptable disposal procedure which 1s not
going to be associated with lining and extensive monitoring.
V, C. Taylor —I would like to make on comment on that, I have
heard of this problem of potential hazards associated with sludge. But
I would like to mention that you have been ponding flyash now for a number
of years and we are now talking about maybe 25-30 million tons a year
flyash in ponds. Now the use of sulfur from air pollution control sludge...
in other words, when you added the limestone or the calcium sulfate to it,
you have not increased these potential hazardous elements. Any hazardous
elements that were there, were probably associated with the coal and
it's been with the coal long before you started adding limestone to it.
So what? You are not aggravating any problem here; you may be accentuating
it or thinking about it.
J. Rasile — He have not found that to be the case. We have run
leachate tests on flyash that was collected dry. We have stirred vigorously
for 18 hours and determined how much leaching of the same trace
elements comes out and I think the highest we ever got on anything was
boron and that was 1 part per million, on trying to leach
ft out of dry flyash. As a contrast, when you scrub it out of a wet
participate scrubber, when you are scrubbing the particulates out, we
have reached 60 parts per million of boron with no trouble at all.
This has been observed in about three plants, So there seems to be some
chemistry involved. In other words, it is not the same procedure.
951
-------
W. C. Taylor— May I ask what are you adding to the flyash now?
You are saying the scrubbing process itself makes the flyash mor6
soluble?
J. Rasile— No, I am saying, what I think is happening. What perhaps
should be determined is whether we are removing elements which used
to go up the stack.
W. C. Taylor— Well, it was a hazard in any case, wasn't it?
J. Rasile — Well, except we are talking here about the disposal
of waste products. And it is too bad it was a hazard. Yes, but right
now somebody is going to be sitting there monitoring and if they did not
monitor the stack well, somebody got away with a lot a while ago. But
it is not going to be the situation any longer.
A, V. Slack — I might point out again, as I did earlier, that tests
in Lunen, Germany, have not, I believe, established any leaching at all
in test wells. Tests from Ontario Hydro indicate, I believe, that the
liquid phase pretty well stays occluded in the sludge and doesn't get
washed out. These are just bits of data.
J. Rossoff-—Well, I have my own observations on that. I would not
want to doubt your fears, but I have to agree with Bill. You know that
you are not adding any new elements. They may be in a different form.
I would just like to ask you whether you are basing your feelings on actual
leaching test data to the soil or whether you are basing it on concen-
trations within the solid sludge itself. Or are you talking about con-
centrations in the liquid form, which are now in solution?
952
-------
J. Rasile—The concentrations I referred to are those
which were in the supernatant scrubbing liquor, the clarified
scrubbing liquor. The same trace elements exist, for example,
in flyash, in roughly the same proportions they did in the original
coal. They stay there. I have no proof of this but my feeling at
this point in time is that we are scrubbing the gases or vapors,
these volatilized materials, out of the gas. Me are not adding
anything. But now that it is in the water (and it is in the waste
water) you are trying to discharge it at the rate at which waste
water will leach.
A.V. Slack —Well, of course. May I ask, though, that if we
did agree that there is no leaching downward, are you saying that
this waste water could leach sidewise because we planned to recycle
it?
J. Rasile—-By what definition is it not going to leach down-
ward?
A.V. Slack — I said tests have indicated that. Now, of course,
we need to corroborate those. But if you assume this were true,
then would there be a problem?
J. Rasile-"Well, about the tests, not knowing anything about the
tests, I would say you would have to know something about the soil
953
-------
characteristics to determine whether it is going to leach downward or
not.
A. V, Slack — Well, this liquor is staying in the solids. It does not
make any difference about the soil underneath. Well, I think perhaps
we are exhausting the subject.
J. Rossoff —-May I ask one more? When Dick Stern (R. D. Stern, EPA)
spoke here Monday, he made a plea for cooperation and help. Since
that time I have had a lot of offers, gracious offers, to send me carloads
and 55 gallon drums. If you would like to part with that data we
would be happy to use it and compare with ours and include it. I think
it would be very helpful,
Jim Henderson(ASARCO)^ I don't quite understand the opinion
expressed that we are not adding anything, because we are certainly
adding lime rock. And in looking at various lime rocks in the
vicinities of our smelters, we see a wide variation in the trace
elements present in the lime rock itself. So I think we truly are
adding something, as compared to flyash.
W. C. Taylor —Well, that is why I asked what were you adding? The
limes that we have been adding and have been testing have not had any
higher quantity of trace elements than we had in the ash. So I imagine
you can get limestone in any purity that you want. And maybe some of
them are contaminated. But I guess you would have to look at your
limestone then if you were concentrating trace elements,
J. Rasile — It is not the quantity of concentration I'm speaking
of. The total gross quantity of trace elements is going to be increased
as compared to the flyash depending on what's in the limestone.
954
-------
Howard Hesketh(Southern Illinois University)— It kind of surprises
me in one respect that we are talking about a problem here with not
only the sludge or the sulfur storage. You realize that here in the
same discussion we have been talking about the fact that you can use the
sludge as a retainer for city water supplies, you can use the sulfur
as a waterproof barrier, and so on, I really see that we have the
answer right here, should we want to use it,
A. V. Slack — If we can get somebody to buy the stuff, There is
a practical problem here.
H. Hesketh— I am talking about the power pi ants utilizing their
own material to solve their own problem.
A. V. Slack— I think your material balance might be a little out
of perspective there.
R. H. Borgwardt(EPA)"- I have a question for Bill Taylor or
anyone else who might know, concerning the magnesium that enters the
scrubber with the limestone. All limestones contain magnesium in some
degree or another. In our pilot plant, the evidence seems to Indicate
that the magnesium leaves primarily in the solid (rather than in the
liquid) purge. We do not know; however, whether it is in the form of
magnesium carbonate (that is, the carbonate entering Just didn't dissolve
and leaves with the solid) or whether it dissolved and was precipitated
as possibly the magnesium sulfite and leaves as magnesium sulfite,
VI. C. Taylor — Well, I'm sorry you asked me that. But maybe there 1s
someone here from Combustion. Is there anyone here from Combustion that
could help me out on this? I do know that the magnesium hydroxide 1s
fairly Insoluble and 1f you had a high enough pH, you would not have
considerable magnesium in your. . . ,
955
-------
R. H. Borgwardt — I am talking about the limestone scrubbing system.
Robert Van Ness (Louisville Gas & Electric)— One of the reasons
we are using carbide lime is because of this particular point. Carbide
lime slurry contains probably the range we are using, probably from 0.04
to 0.07 percent MgO. Now the lime that the carbide was originally made
from was about 1 percent MgO and I believe the same thing would happen
in Japan because MgO is vaporized in the furnace, and it was already in
the atmosphere. So you are not adding again. If you are using limestone,
then you do have the problem. But that is one of the reasons we chose
carbide lime. Because it was cheaper also. But this is another point.
We would not have high MgO levels.
Joe Se1meczi(Dravo Corporation)— As a "has-been" geologist I
would like to make a stipulation relative to the magnesium content.
It's one of the paradoxes of geology that calcium and magnesium carbonate
has a lower solubility than calcium carbonate or magnesium carbonate alone.
Yet, you cannot synthetically make dolomite. Now when somebody is trying
to use dolomitic limestone (where the combination is not a camouflaged
dispersion of the magnesium in a calcium structure, but a dolomite
structure) then the magnesium will go through undissolved because it
will be a lot less soluble. In other words, equilibrium and kinetic-
wise, one way a field geologist determines the difference between calcium
carbonate and dolomite is to drip 4 percent hydrochloric add on a
stone. Now if he observes a fast evolution of carbon dioxide, 1t 1s
956
-------
limestone, If the evolution 1s very slow, 1t 1s dolomite. I do not
know how much this 1s worth but probably it has something to do with
the way magnesium 1s leaving a system of limestone scrubbing. Now
relative to the trace elements. We have been working with western coals
(I am sorry, eastern coals) primarily at Dravo and we have run several
Teachability tests on flyashes and the sludges generated primarily
by lime scrubbing. Now we Pound that in all cases the trace elements
were lower than you would get from flyash leaching alone. In one case
only the arsenic content was in excess of Federal regulations for
drinking water. And even in the case of arsenic, the flyash leachate
(the leachate from flyash) contained higher arsenic content than the
sludge. One reason for that is perhaps that the sludge 1s creating Its
own reducing condition and arsenic is not thermodynamically present as
arsenate in a reducing condition. Now if you are slurrylng and a
dissolution experiment is done in a highly oxidizing condition, you may
dissolve more arsenic from your sludge. Now western coals do contain
a great deal more arsenic than eastern coals. As a matter of fact, mine
leaching (mine discharge waters) in the West contain arsenic In excess
of what's permissible by drinking water standards. Pedple we talked to
at EPA are not concerned about mine acid waters in the East, as far as
arsenic content is concerned.
957
-------
Steve Smith (Koch Engineering Co)— I have a question for Bill
Taylor regarding the spray drying of the sludges from the scrubber
with the waste heat of the boiler gas. We think there may be some
trend for the unreacted limestone in the sludge to reactive SOp and
also it looks like there might possibly be some oxidation of the sulfite
to sulfate. Has your work turned up anything on this?
U. C. Taylor — We have gone to people like you and other spray dryers.
But address that to Rao again. Rao, did you hear the question?
M. R. Gogineni — I did not quite understand the question. Could
you repeat the question? Maybe I can help.
S. Smith — Yes, in spray drying the sludge from a limestone
scrubber, have you found any reaction between the SCL and the limestone,
the unreacted limestone in the sludge? And have you found any oxidation
from the oxygen, the flue gas of the sulfite to sulfate?
M. R. Gogineni ->-I think in saying we have done work, I think it is
not experimental work, It is still in the paper evaluation and we
have done some pilot plant tests using the, I think, air at the tempera-
ture of the flue gas. But there was no S02 removed in the spray dryer as
a result of reacting with the unreacted calcium carbonate that was present
in the sludge.
S. Smith-"Okay; I think, depending on the amount of limestone in
the sludge, of course, that this will vary. I think there may be a little
958
-------
bit of removal,
M.R. Gogineni — Maybe,
S. Smith-Okay.
M.R. Gogineni —Or maybe not, if it is solid.
W. V. Botts (Atomic International)— I might just comment on
spray drying, We have some data that suggest, even with the lime,
that you do get S02 removal. We are selling systems of a spray
drying nature using sodium carbonate. A spray dryer is probably one of the
most efficient contactors that you can come up with for
contacting flue gas with a reactant, and we are using sodium carbonate.
In our tests at the Mohave station, we ran some lime dilute solutions and
got quite good S02 removal. So if there is any lime in the sludge, you
do indeed get $02 removal,
J. Selmeczj — i do not think I have to introduce myself again, but
I remember what I wanted to say, As far as the heavy metals leaching
into the soil and the probes, valve points, or whatever devices you
try to determine, their transferral in the groundwater depends entirely
on the soil conditions, As geochemists know, humus is a very good ion
exchanger which you probably find close to the surface. As a matter
of fact, this is how we detect buried deposits of valuable minerals.
Clays also have fantastic capacities for heavy metals in the cationic
form; not in the anionic form, however, Now, for example, with cesium
(which was conducted under the supervision or sponsorship of AEC), a
959
-------
radioactive cesium was found to be most economically removed by clay or
just natural soil, mixing radioactively contaminated water with
natural soil. So just to sound a little bit of a hopeful note, heavy
metals if they do leach into the ground and if the ground is not porous
sandy soil, but contains even a small amount of clays, the clays will
remove the heavy metals.
J. Ando—Let me just mention just one thing about the spray
drying of the sludge. I don't think that desulfurization takes place
fry the calcium carbonate temperature which is not very high, But I am
sure that considerable oxidation of calcium sulfite will occur in
the spray drying.
M.R. Gogineni—Regarding the spray drying and S02 removal in the
spray dryer. I think the gentleman commented about the fact that he was
getting some SO^ removal when he tried to dry sodium carbonate. Am I
right? It may be due to what's in the solution rather than what's in the solid
form.
960
-------
REMOVAL OF SULFUR DIOXIDE FROM STACK GASES
BY SCRUBBING WITH AMMONIACAL SOLUTIONS:
PILOT SCALE STUDIES AT TVA
by
Gerald A. Hollinden and Neal D. Moore
Power Research Staff
Tennessee Valley Authority
Chattanooga, Tennessee
P. C. Williamson
Division of Chemical Development
Tennessee Valley Authority
Muscle Shoals, Alabama
D. A. Denny
Control Systems Laboratory
Environmental Protection Agency
Research Triangle Park, North Carolina
961
-------
REMOVAL OF SULFUR DIOXIDE FROM STACK GASES
BY SCRUBBING WITH AMMONIACAL SOLUTIONS:
PILOT SCALE STUDIES AT TVA
ABSTRACT
The Tennessee Valley Authority and the Environmental Protection
Agency have pursued an ammonia scrubbing program at TVA's Colbert
pilot plant since 1969. This effort began as a wholly funded EPA
program with the objective to fully characterize ammoniacal liquor
scrubbing of S02-laden flue gases. Since 1971, the program has
become a jointly funded effort in which the operational phase of
ammonia scrubbing was coupled with a regeneration scheme for producing
a concentrated stream of sulfur dioxide. The principal advantage in
the EPA-TVA process is economic regeneration by acidulation with
ammonium bisulfate produced from thermal decomposition of ammonium
sulfate. This paper highlights the activities during the continuing
test program.
In general, S02 recovery is excellent and NH3 losses are low
using ammoniacal scrubbing solutions. However, the following problem
areas have been established:
•Fume formation in the scrubber and after discharge
from the stack
•Fly ash separation
•Ammonium sulfate separation
Under proper operation of the system, it is possible to control
fume formation inside the scrubber while operating at relatively high
salt concentrations. Avoiding fume formation on days of low tempera-
ture and high relative humidity may be impractical to achieve. While fly
ash and ammonium sulfate separation have not been satisfactorily demon-
strated, several promising leads will be investigated in a modified
process sequence.
The double alkali approach is another possible scheme to utilize
ammonia in recovering oxides of sulfur from stack gases. This approach
also will be studied.
962
-------
Introduction
The Tennessee Valley Authority and the Environmental Protection
Agency embarked on an ammonia pilot plant scrubbing project 4 years
ago. The pilot plant treats 2600 standard ftVmin*slip stream from a
TVA coal-fired steam generating station located in northwestern Alabama.
The objectives of the study were to establish the ability of ammoniacal
solutions to efficiently remove dilute quantities of sulfur oxides from
flue gas with extremely low gaseous ammonia losses. Trouble-free
operation of equipment during the first 1-1/2 years of testing
promises a highly reliable, low-maintenance scrubbing system.
Because of the cost of ammonia and the extreme solubility of the
ammonium salts, the process is not amenable to throwaway operation.
In this light, a regeneration scheme ("bisulfate process") was adopted
that appears to have several advantages over other regenerative systems.
This paper describes the pilot plant operation incorporating that
portion of the total system—absorption plus regeneration—that is
currently available from a 1-year period of testing.
Background
Several regenerable or semi-regenerable ammonia scrubbing processes
have been developed and are, or have been, in full-scale operation. The
earliest of these was pioneered by The Consolidated Mining and Smelting
Company (Cominco) in Trail, British Columbia.1 The process consists of
aqueous ammonia scrubbing followed by acidification of the sulfite
liquor with sulfuric acid to evolve S02 and produce ammonium sulfate.
The S02 is sent to a sulfuric acid plant and the ammonium sulfate is
further processed for sale as a fertilizer. The process has been
operating continuously and reliably on smelter gas since the mid-1930's
and is still in operation. Work on adapting the method to power plant
stack gas was carried out by TVA (pilot plant scale) in 1953-54.2 The
absence of a large market for ammonium sulfate, however, severely
limits the applicability of this otherwise excellent process in the
United States.
Quite similar processes have been developed and are in full-scale
operation on sulfuric acid plant tail gases in Czechoslovakia and
Romania.3*4 In these methods, the effluent scrubber liquor is acidu-
lated with nitric or phosphoric acid; and the process, therefore, must
be integrated with fertilizer plants to provide an outlet for the
ammonium nitrate or ammonium phosphate formed. Again, the limited
marketability of the fertilizer products and constraints as to the
location of the SC^-emitting plant near a fertilizer manufacturing
center limit their wide use in this country.
*See English units and metric equivalents on page 996.
963
-------
A regeneration process producing SCL only was postulated and tested
in the 1930 's by H. F. Johnstone. Steam stripping was employed to
recover S02 and to regenerate the ammoniacal solution for reuse in the
scrubber. Since concentrated S02 is the major product, either sulfuric
acid, liquid S02, or elemental sulfur can be the final product, depend-
ing on the need of the user. Development of the Johnstone process has
been vigorously pursued in the U.S.S.R. A 140,000 standard ft3/min
(60 Mw) unit was installed near Moscow in 1952 and operated
continuously until 1967 at which time it was disassembled because the
power plant was converted from coal to natural gas. It is reported
that the cyclic ammonia process will be installed on a 200 Mw coal-
fired utility boiler in the U.S.S.R. in 1973. Although the Johnstone
process has been shown to be feasible and presupposes no link to a
complex fertilizer plant, it possesses some undesirable characteristics.
Energy requirements for the process are high, approximately 12 Ib steam
per pound of S02. Oxidation products may be difficult to purge from
the system without loss of active species thereby resulting in a higher
effective oxidation. The occurrence of undesirable disproportionation
reactions in the steam stripper further aggravates the oxidation problem.
TVA-EPA Ammonium Bisulfate Process
In the decomposition of sulfites with an acid, it may be possible
to regenerate the acid, thereby avoiding the need for disposal of the
acid salt. Such an acid regeneration process*) has been known in the
fertilizer industry since the 1920 's. In this process, the ammonium
sulfate is heated to drive off ammonia and produce acidic ammonium
bisulfate, which is then used as an acidulant to release
In more recent work, an engineering company has incorporated the
bisulfate technique in various fertilizer flowsheets? and has carried
out pilot plant work for converting ammonium sulfate to bisulfate.
The process involves direct heating of the ammonium sulfate with
combustion gas. The decomposition reaction is highly endothermic.
This work is being expanded currently to a large-scale test program,
on a scale equivalent to the size required for a 30 Mw power plant burn-
ing 3.5 percent sulfur coal. Several other research organizations have
also worked on the conversion step but in a less extensive way.
Application of the bisulfate technique to regeneration of ammonia
scrubber liquor in an S02 removal process was proposed by Hixson and
Miller in 1944. 8 Ammonium sulfite-bisulfite liquor is pumped from the
scrubber to an acidifier-stripper where the following reactions take
place:
(NH4)2S03 + 2NH4HS04 -»• 2(NH4)2S(>4 + H2<> + SO + (1)
NH4HS03 + NH4HS04 -»• (NH4)2S04 + H20 + S02 t (2)
964
-------
Essentially all of the SO. is released from the concentrated
liquor in the acidifier-stripper. If a condenser is employed on the
off gas, virtually a 100 percent stream of SC>2 can be obtained. The
resulting stripped solution, containing mainly (NH^oSO^, is sent to
an evaporator-crystallizer where the water is removed. Ammonium
sulfate crystals are then transported to the decomposer where the
(NH^oSO^ is thermally dissociated into NH.HSO, and gaseous NH3. The
bisulfate is returned to the acidifier and the ammonia to the scrubber.
Pilot Plant Equipment and Operation
The pilot plant is designed to treat 4000 actual ft3/min flue gas
at 300°F. Stainless steels (304 and 316} and rubber are used in all
wetted sections for corrosion protection. The gas ducts are all
mild steel. For simplicity, the plant is divided into the absorption
and regeneration sections.
Absorption: The absorption section is shown in Figure 1. The
scrubber is a three-stage, marble-bed unit which was available from
a previous ammonia-scrubbing study. The three scrubber stages can
be operated to maximize the S02:NH3 ratio in the product liquor while
limiting the S02 and NH3 content of the scrubbed gases to 250 and
50 ppm, respectively.
The source of the pilot plant flue gas is downstream of the
electrostatic precipitator on Unit No. 4 at TVA's Colbert Power Plant.
Most of the fly ash is removed in the precipitators. However, the
remaining fly ash (up to 0.5 gr/ft3) enters the pilot plant scrubber
where a portion of the ash is removed and accumulates in the scrubber
solution at a rate of 1 to 10 pounds per hour, depending on the
efficiency of the precipitator
Most of the fly ash settles out and accumulates in the scrubber
product storage tanks, F-5 and F-6. (Each of the storage tanks has
a capacity for approximately 24 hours' absorber operation.) Periodi-
cally, the settled material is purged from the system. In later work,
the settled material will be sent to a tub filter and the cake washed
in a batch-wise operation. The wash from the filter will enter the
absorber cycle as make-up water.
An alternative method of fly ash removal is to scrub the flue
gas with water ahead of the absorption sections. Wj.th this scheme,
a small amount of fly ash in the gas stream, as well as entrained mist,
will get through the fly ash scrubber even though a high efficiency
scrubber is used. Thus, a portion of the carry-over fly ash will be
collected in the absorber, necessitating a purge cycle. A portion of
the make-up water will be used to wash the fly ash cake from the purge
cycle.
965
-------
WET ELECTROSTATIC
PRECIPITATOR
SCRUBBER
TO STACK
FLUE
GAS
.
V.
r
MAK
WAI
N
E-UP
PER
i
J
G-3
G-2
G-l
: 7
i i r i' v
VARIABLE SPEED BLOWER
J-23
f
•\ - /-
*v /•
J-21
J-22
AT/1MONIA FROM
REGENERATION
SECTION
VARIABLE SPEED PUMP VARIABLE SPEED PUMP VARIABLE SPEED PUMP
J-l J-2 J-3
REC1RCULATION TANK RECIRCULATION TANK RECIRCULATIONTANK
PRODUCT LIQUOR TO
REGENERATION SECTION
J-12
Figure 1. Absorption section ammonia scrubbing-ammonium bisulfate regeneration process.
966
-------
Regeneration: The regeneration section shown in Figure 2 is
designed to process liquor produced in the absorber section.
Originally, the pilot plant regeneration process was designed to
include the following steps:
1. Acidulation of the absorber product with ammonium bisulfate
melt to release S02 and form ammonium sulfate.
2. Stripping of the released S02 from the ammonium sulfate
solution.
3. Crystallization of ammonium sulfate from the mother liquor.
4. Separation of ammonium sulfate from the mother liquor.
5. Generation of ammonium bisulfate for the acidulation step
by heating a 1:1 mole ratio of H2S04 and ammonium sulfate.
The acidulation is accomplished in a vessel made from a 6-foot
section of a 12-inch stainless steel pipe. The vessel is coated
internally with DuPont's TFE Teflon for corrosion protection. Mixing
of the incoming scrubber liquor and ammonium bisulfate is accomplished
in a cone mixer in the upper part of the acidulator. Released S02 flows
from the acidulator to the stack aftei: joining the effluent from the
stripper vessel. The acidulator is mounted so that the point of gravity
overflow from the acidulator to the stripper can be raised or lowered to
vary the retention time of material in the acidulator from zero to a
maximum of 4 minutes.
The Teflon-lined stripping vessel is 1 foot in diameter by 6 feet
high and contains a 4-foot section of 1/2-inch ceramic Raschig rings.
Acidulated liquor enters the top of the vessel and flows countercurrent
to a stream of stripping gas entering the vessel near the bottom. In
the pilot plant, provisions are made to use either steam, air or
scrubbed flue gas as the stripping gas.
Ammonium sulfate solution from the stripping vessel flows by gravity
to crystallization tank F-7 where water is evaporated. The concentrated
solution then flows to F-8 where cooling and crystallization take place. The
saturated solution, containing crystals of ammonium sulfate, is centrifuged
to separate ammonium sulfate crystals from the mother liquor. The filtrate
is returned to the absorber section.
In a completely closed-loop ammonium bisulfate regeneration scheme,
all the ammonium sulfate produced in the acidulation step would be
thermally decomposed. The resulting ammonium bisulfate is fed to the
acidulation step and the ammonia is returned to the absorption section
(to recirculation tank F-2).
967
-------
c
SULFURIC ACID
ETANK
f~X
I
Jj
10 I)
II /
n"
U '
HOT
COMBUSTION
GASES I
T
^
/^
Ji
AMMON
AIR BISULF
HIK GENER/
^
AMMONIUM SULFATE
FEEDER
J-ll
AMMONIA STORAGE TANK
PROPANE
GAS
TO STACK
(I
AMMONIA TO
ABSORPTION
SECTION
STORAGE AND SETTLING TANKS
ACIDULATCR
FLY ASH TO WASH AND DISCARD
STEAM
STRIPPING
GAS
F-7
AMMONIUM SULFATE J*10
CRYSTALLIZATION TANKS
CENTRIFUGE
|
PRODUCT LIQUOR FROM
ABSORPTION SECTION
J-8
AMMONIUM SULFATE
Figure 2. Regeneration section ammonia scrubbing-ammonium bisulfate regeneration process.
968
-------
In the present pilot plant study, the thermal decomposition stage
was omitted and the acid ion was furnished as sulfuric acid. Ammonium
bisulfate will be the source of the acid ion later. The required
ammonium bisulfate will be generated by heating a 1:1 molar mixture of
sulfuric acid and ammonium sulfate. The sulfuric acid and the ammonium
sulfate (from the centrifuge) will be fed separately and continuously
to the ammonium bisulfate generator," D-l, a Teflon-lined stainless
steel vessel. Here, sufficient heat will be added to the acid-ammonium
sulfate mixture to evaporate the moisture added with the acid and to
bring the mixture to about 350*F. The ammonium bisulfate melt will then
be fed to the acidulator.
Since ammonium sulfate is not decomposed in the present study, no
ammonia is available from the regeneration section. The makeup ammonia
is fed to the absorption section from the ammonia storage tank, F-9.
During the present pilot plant operation, the ammonia is fed undiluted
directly to the points of use in the absorption section. Provisions
will be made later to simulate ammonia recovery from a thermal decom-
position process.
Vent System: The scrubbed flue gas is vented to the atmosphere for
plume observation. The recovered S02 from the acidulator and stripper
can be vented into the power plant stack. The storage tanks containing
ammoniated solutions can be vented to the absorber or to the stack.
Instrumentation: The pilot plant is instrumented throughout so
that all pertinent liquid and gas flows are monitored and values are
recorded. All signals are electrically transmitted from the sensing
element to the recorder-controller.
The gas flow through the scrubber system is monitored using a
flange orifice in the duct leaving the scrubber. A differential
pressure cell senses the pressure differential across the
orifice and sends a signal to a recorder-controller on the
pilot plant instrumentation board. Any deviation from the preset values
on the recorder-controller causes a signal to be sent to the variable-
speed drive mechanism on the induced draft blower (J-23) to correct the
deviation. This arrangement assures that a constant gas flow through
the scrubber system is maintained even though the pressure drop across
the system may change.
Sulfur dioxide levels in the gas to the scrubber and after each
scrubber stage is monitored using an ultra-violet analyzer^
The analyzer has three ranges of S02 values: zero-4000, zero-
1000, and zero-100 parts per million full-scale reading. The
sample point is changed manually from station to station to avoid the
possibility of leaks from an automatic sample sequencing system.
Periodic checks by wet-chemical methods rarely differ from the analyzer
reading by more than 5 percent.
969
-------
An NOX analyzer has been installed in the pilot plant but has not
been activated. It is expected that this instrument will be available
to monitor the inlet and outlet NO and NO- levels in future pilot plant
runs.
A smoke detector is used to monitor the intensity of the plume at
the stack exit. The instrument uses a light source and a photocell
to measure the plume intensity. The digital readout is in Ringelmann
units.
Gaseous ammonia is metered to the system as required by the use of
a differential pressure cell coupled with a recorder-controller
and a flow control valve. Liquid flows are sensed by magnetic flow-
meters which send electronic signals to recorder-controllers. The
required flows of recirculating liquor to the scrubber stages are
controlled by variable-speed pumps (J-l, J-2, and J-3). Variable-
speed pumps are used for flow control instead of valves because fly ash
removed in the scrubber could cause plugging and erosion of control
valves. Automatic flow control valves are used to control the flow of
the remaining liquid streams.
Temperatures throughout the system are sensed with thermocouples
and are recorded on strip charts in the control room.
Problem Areas
Fume: The most apparent problem surfacing during the past and the
current TVA-EPA work with the ammonia-scrubbing process is that of fume
formation. Figure 3 shows a typical ammonia-based plume including steam
emitted during routine pilot plant operations with no gas reheat (exit
temperature is approximately 125°F). The plume was formed during opera-
tion using three scrubbing stages, no prior humidification, and while
producing a liquor with a C of about 10 and an S/C of 0.8 (C = moles of
ammonia as sulfite and bisulfite per 100 moles water*, S = moles of
sulfite and bisulfite sulfur per 100 moles water). Sulfur dioxide
removal efficiency under these conditions is typically about 90 percent.
Efforts to reduce or eliminate the plume through manipulation of
the pH of the scrubber liquor and the addition of a humidification
step ahead of the scrubber did not reduce the plume opacity signifi-
cantly. However, under these test conditions, the steam plume may have
masked any reduction in the ammonia-based plume. Adding a wet electro-
static precipitator after the scrubber failed to produce significant
results at the normal gas flow rate of 2600 standard ft3/min. (The
precipitator was found to be 50 percent efficient in removing the
particulate mass at this flow rate--15 ft/s shell velocity through the
precipitator.)
*combined and uncombined
970
-------
ami
Figure 3. Typical plume during routine operation; high scrubber liquor concentration (c = 10); no reheat.
-------
Chemical and petrographic analyses of the plume collected in an
impaction sampler (Brinks mass sampler] indicate that the major fraction
of the ammonia-sulfur salt is ammonium sulfate. It is probable that the
particulate was formed in the vapor phase as ammonium sulfite and then
oxidized to the sulfate form in the sampler. A portion of the particulate
has been analyzed as ammonium chloride. (The coal used during these tests
normally contained 0.1-0.2 percent chlorides.) Tests showed that a water
wash ahead of the scrubber materially reduced the chloride content of the
gas entering the scrubber.
Personnel from the Southern Research Institute of Birmingham, ~
Alabama, determined the fume particle size to be 0.25 micron with 10
particles in the size range of 0.005 to 0.5 micron per cubic centimeter.
Reheating the scrubbed gas was effective in reducing plume opacity.
These data also show that plume opacity is greater at higher salt con-
centration (high C's) than at lower salt concentration (low C's).
Solids Separation: Other problems identified in the pilot plant
program are not as visible as the plume problem but may be difficult to
solve. These are the problems of fly ash separation and ammonium
sulfate separation. As stated earlier, the fly ash is expected to
separate by gravity in the absorber product storage and settling tanks.
These tanks, which have capacity for about 24 hours' retention time, are
effective in removing most of the fly ash. However, the small quantity
in the supernatant liquor from the tanks may be sufficient to adversely
affect the production of crystalline ammonium sulfate further downstream
in the process. Also the fly ash which contains iron may catalyze the
decomposition of ammonia during thermal decomposition of ammonium
sulfate. The cracking of ammonium caused by iron and other contaminants
in the fly ash during the ammonium sulfate decomposition step will be
examined by the Eugene Kuhlmann Company in France. The work will be
done in a pilot plant sized to handle the production from a 30 Mw power
plant.
Attempts to remove the product liquor from the settled fly ash by
filtration in a tub filter failed because of blinding of the filter
media. The fly ash was settled from scrubber product liquor with C's
of approximately 10 and containing from 10 to about 30 percent by weight
ammonium sulfate. The blinding was caused by gelatinous, thixotropic
material composed of finely divided fly ash and tiny needle-like crystals
of ferrous ammonium sulfite which precipitated from the scrubber solution.
The iron in this compound is believed to come from fly ash. The quantity
of iron dissolved in the solution is small—on the order of 0.01 gram
per liter. Use of filter aids and precoats on various types of filter
media failed to prevent blinding.
An equipment company specializing in solids separation made filtra-
tion tests on the scrubber product liquor containing fly ash and also
on the mud from the bottom of the settling tanks. They recommended use
of drum filters and filter aids on both materials.
972
-------
Separation of ammonium sulfate crystals from the acidulated and
concentrated liquor in the regeneration loop, though difficult, may
be easier than fly ash separation. As stated earlier, the liquor
concentration and crystal growth take place by evaporating water from
the solution in a tank at atmospheric pressure. Crystal growth cannot
be controlled in this equipment and attempts to separate the small
crystals in the pilot plant centrifuge and drum filter have failed.
Petrographic analysis indicated that ferrous ammonium sulfite was
present in the material blinding the separation equipment as was the
case in the fly ash separation tests.
Bench-scale work using solutions from the pilot plant operation
produced adequate crystals for filtration. Filtration rates of over
200 gallons per hour per square foot were obtained in these tests which
produced much larger crystals than those produced in the pilot plant.
The ferrous ammonium sulfite did not cement together the larger crystals
produced in the bench-scale work. The use of an evaporator-crystallizer
is expected to produce the required crystals although the small amount of
fly ash in the solution may influence the design (and cost) of the unit.
The equipment manufacturers contacted concerning the sulfate crystal
growth and separation were of the opinion that the standard sulfate
crystallizer equipment would be sufficient. However, all stated that
they could not guarantee this equipment, without making tests using the
actual solution containing fly ash.
Once adequate crystals of proper size are produced, separation
would be accomplished by the standard centrifugation method used in the
ammonium sulfate industry.
Pilot Plant Test Results
Routine Operation: Aside from the aforementioned problems, per-
formance of the pilot plant in removing S02 from the flue gas and in
subsequently releasing the absorbed S02 has been good. Table I shows
data from a typical pilot plant run when producing an absorber liquor
with a C of 10 and an S/C of 0.8. The ammonia required to react with
the S02 was added to the second-stage absorber loop (tank F-2). Under
these conditions, the S02 removal was approximately 90 percent. When
the S/C was raised to 0.85, the removal efficiency, as expected,
dropped (to 85 percent). Approximately 13 percent of the absorbed S02
was oxidized to ammonium sulfate. The Murphree tray efficiency for
the marble-bed scrubber was about 0.90.
The regeneration loop was operated using sulfuric acid (93 percent)
to acidify the liquor from the absorber section. Typical data from
these tests are shown in Table II. Sulfuric acid was used for acidu-
lation instead of ammonium bisulfate because no ammonium sulfate crystals
were available from the pilot plant for production of the bisulfate.
Clarified liquor from the absorber section was pumped to the acidulator at
approximately 0.4 gaVmin, the rate the liquor is produced in the absorber.
973
-------
TABLE I
Typical Absorber Loop Data
Test Conditions
Gas to scrubber
Flow rate, ft3/min at 32°F
Temperature, °F
S02, ppm
Gas leaving first stage
Temperature, °F
SO2> Ppm
Gas leaving second stage
Temperature, °F
SC»2, ppm
Gas leaving scrubber
Temperature, °F
SO2, ppm
862 removal, %
Stoichiometrya
Forward feed flow rate,*5
Liquor to top stage
C
S/C
PH
Specific gravity
Liquor to middle stage
C
S/C
PH
Specific gravity
Liquor to bottom stage
C
S/G
pH
Specific gravity
Product from scrubber
C
S/C
PH
Specific gravity
Flow rate, gaL/min
NH3 to F-2 tank
(Normal Operation!
2650
285
2360
127
1440
125
360
122
280
88
1.45
1.3
1.4
0.78
6.1
1.036
5.1
0.62
6.8
1,098
10.8
0.81
5.7
1.200
10.4
0.84
5.7
1.202
0.41
a
Stoichiometry is the ratio of moles of. gaseous ammonia added to the
moles of S02 in the gas stream.
3The only forward feed added to the system was water.
974
-------
TABLE II
Typical Regeneration Loop Data
Test Conditions
Acidulation with
Only
Acidulator
Liquor feed in
C
S/C
pH
Specific gravity
Flow rate, gaL/min
Sulfuric acid
Flow rate, gaL/min
Percent su If uric acid
St oi chiome t ry a
Liquor flow out
C
S/C
PH
Specific gravity
Percent 862 release
Stripper
Stripping gas
Type gas
Flow rate, ffVmin at 70°F
Liquor flow out
C
S/C
pH
Specific gravity
Percent S0 release
Overall
S02 release
9.8
0.83
6.0
1.193
0.36
0.11
93
1.2
1.3
1.0
2.0
1.194
85
Air
12
0.2
1.0
2.1
.188
83
97
Stoichioraetry is the ratio of moles of acid ion to moles NH3 as
ammonium bisulfite and ammonium sulfite in the liquor to the
acidulator.
975
-------
Sulfuric acid was pumped to the acidulator at a rate to give a ratio
of 1.2 moles of acid (hydrogen) ion per mole of sulfite and bisulfite
sulfur in the feed liquor to the acidulator. (The 1.2 ratio was used
for convenience in these preliminary tests; the theoretical require-\
ment is 1.0.) Approximately 85 percent of the absorbed sulfur as
sulfite and bisulfite is released from the acidulator. The remaining
15 percent of the S02 flowed to the stripper where air at 12 ft3/rain
was added to remove the residual S02. Approximately 83 percent of
the residual S&2 to the stripper was removed by the stripping gas to
give an overall removal efficiency of 97 percent in the combined
acidulator and stripper.
The acidulated and stripped liquor was concentrated to about 45
percent by weight as ammonium sulfate and then cooled to near 100°F to
precipitate crystalline ammonium sulfate. However, as noted earlier,
the ammonium sulfate could not be separated because of blinding of the
equipment by extremely fine crystals of ammonium sulfate and ferrous
ammonium sulfite.
Double Alkali Tests: Production of liquors with low C's is
uneconomical for use in the ammonium bisulfate regeneration process
because of the energy requirements to concentrate the solutions in the
regeneration sections. However, scrubbing with solutions of low C's
has possible application in the double alkali process being considered
as a backup system for TVA's Widow's Creek Power Plant located in
northeast Alabama. In the double alkali process based on sodium as
the absorbent, the concentration of sodium in the absorber loop is
limited to about 0.17 mole per 100 moles of water because of solubility
limitations in the regeneration section. Ammoniacal solutions, which
are not limited to low concentrations in the regeneration loop, would
offer economic advantages because the size of the regeneration equipment
could be reduced. When using an ammoniacal solution with a C of 0.5,
the regeneration equipment size could be reduced to about one-third
that required when scrubbing with sodium solutions.
A test series was made to determine whether a plume would be
present under scrubber conditions similar to the proposed Widow's Creek
operation. The test series was made using the flow configuration shown
in Figure 4. Water was used on the bottom stage for humidification and
fly ash removal. Fresh makeup water was added to this loop at the rate
required to maintain the temperature of the gas leaving the bottom (G-l)
stage at 120°F. Forward feed liquor (ammoniacal solution) was pumped
to the middle (G-2) stage at 30 gal/min on a once-through basis and
then drained to the product liquor hold tank. In one test, two-stage
scrubbing was used to increase S02 removal. The forward feed liquor
was pumped at a rate of 30 gal/min to each of the middle and top stages.
The gas leaving the scrubber was reheated to 150, 175, and 200°F during
most of the test series by direct heating with a propane torch.
976
-------
CHEVRON-TYPE
MIST ELIMINATOR
G-3
(MARBLE BED)
G-2
(MARBLE BED)
G-l
(MARBLE BED)
INLET GAS
EXIT GAS TO ATMOSPHERE
MAKE-UP
WATER
REHEAT
«
ONCE-THROUGH
LIQUOR
TO
STORAGE
TO SEWER
Figure 4. Flow configuration for once-through liquor on G-2 and water on G-1,
977
-------
A steam plume, as well as a minor residual particulate plume, was
present when scrubbing flue gas with water only (Figure 5). The
steam component of the plume was dissipated when the exit gas was
reheated to 175°F, but a light particulate plume remained (Figure 6).
An additional plume component was present when an ammoniacal solution
was used. The plume produced when scrubbing with a solution having a
C of 0.5 or 1.0 was reduced to an opacity of 10 percent or less, below
the maximum specification of 20 percent, by reheating the stack gas to
175°F. Figure 7 shows the severe plume while scrubbing with a solution
having a C of 1.0 and with no gas reheat. Figure 8 shows the reduced
plume from this operation while reheating to 175°F. The SC>2 removal
efficiencies varied from 75 to 85 percent during these tests.
A second absorber stage was added in one test sequence to increase
S02 removal efficiency. An average removal of 92 percent resulted when
operating with solutions having a C of 1.8 to both scrubbing stages.
Reheating to 200°F was required in this test sequence to lower the
plume opacity to 20 percent or less (Figure 9).
From these tests it was concluded that the plume produced when
scrubbing with solution having low ammonia content is significantly
lower than the plume produced during routine scrubber operation. Also,
the plume produced under these low salt conditions could be made
acceptable by reheating the exit gas to temperatures necessary to
dissipate the steam component of the plume. This temperature will
vary according to the ambient air conditions—temperature and relative
humidity.
Modified Operation: The absorption section of the ammonia scrubbing-
ammonium bisulfate regeneration process pilot plant was operated from
April 16 to May 4, 1973, to study additional methods of minimizing the
plume from the absorber. Two basic flow configurations were used: a
recirculation system using one, two, or three stages of absorption
(Figure 10); and a once-through system using one stage of absorption
(Figure 11). This operation was carried out in cooperation with Air
Products and Chemicals, Inc., Allentown, Pennsylvania.
The tests were made to determine whether the plume leaving the
absorber could be reduced to an acceptable level by operating under the
following conditions:
• Insulated absorber
• A water wash ahead of the first absorber stage
• A water wash after the absorber stage
• Reheating the scrubbed flue gas
978
-------
Figure 5. Plume emitted while scrubbing with water only; no reheat.
-------
kfi
oo
o
Figure 6. Particulate plume emitted while scrubbing with water only; 175° F reheat.
-------
to
00
Figure 7. Plume from operation with scrubber solution having C =1.0; no reheat.
-------
1C
DC
Figure 8. Plume from operation with scrubber solution having C= 1-0; 1750 F reheat.
-------
Figure 9. Plume from operation with solution having C= 1.8; two scrubber stages; 200" F reheat.
-------
EX IT GAS TO ATMOSPHERE
CHEVRON-TYPE
MIST ELIMINATOR
G3
(MARBLE BED)
G-2
(MARBLE BED)
G-l
(KOCH TRAY)
INLET GAS
PRODUCT
TO SEWER
Figure 10. Normal recirculating flow configuration.
984
-------
EX IT GAS TO ATMOSPHERE
CHEVRON-TYPE
MIST ELIMINATOR
G-3
(MARBLE BED)
G2
iMARBLE BED)
G-l
(KOCH TRAY)
INLET GAS
ONCE THROUGH
LIQUOR
TO
STORAGE
MAKE-UP
WATER
TO SEWER
TO SEWER
Figure 11. Once-through flow configuration.
985
-------
An acceptable plume for the pilot plant was defined as one having
a maximum of 5 percent opacity. The opacity readings were made by
qualified observers who successfully completed the EPA-prescribed
visual emissions school. The absorber and all downstream ductwork
were insulated for reasons that Air Products and Chemicals, Inc.,
consider proprietary. The water wash ahead of the first absorption
stage was used to humidify and cool the hot flue gas; to remove
chlorides and, thereby, prevent formation of an ammonium chloride
plume; and to remove fly ash. It was expected that cooling the flue
gas would reduce the ammonia-based plume containing sulfur. The
moving marble bed of the bottom stage (G-l) used in previous work was
replaced with a multi-venturi FlexiTray manufactured by Koch Engineering
Company, Inc. No downtime was required when this type of tray was
deactivated during absorber operation. (The glass marbles had to be
removed from the marble bed scrubber when the unit was deactivated to
prevent damage from thermal shock should the scrubber liquor accidentally
come in contact with the hot bed.)
The water wash after the absorption stage was to decrease the salt
content of the entrained mist leaving the absorber. Reheating would
have evaporated the mist and produced an increased concentration of
ammonia and sulfur dioxide in the exit flue gas which could have then
reacted to form a plume. The top stage of the absorber (G-3) was used
for this wash. The gas leaving the'absorber was directly reheated with
a propane torch to destroy the particulate matter present and eliminate
a steam plume. The retention time of the heated flue gas in ductwork
was 1.5 seconds.
A summary of the test conditions and the opacity readings for each
test are given in Table III. These tests show that an acceptable plume
was obtained during production of liquor with high C's when:
• Water wash was used ahead of the first absorber stage.
• Absorber and all ducts were insulated.
• Reheat was applied as required to dissipate the steam
plume.
Comparison of Figures 12 and 13 shows the effect of the
water wash ahead of the first absorber stage. In Figure 12, the water
wash was not employed and the opacity was 5 percent. The water wash
was activated and the opacity dropped to 0 percent (Figure 13).
The concentration of ammonia salts in the water wash (G-3) after
the absorption stage did not show any significant effect upon the plume
formed in the range studied (C = 0.1 to 3),
986
-------
Summary
TABLE
III
of Operating Data for
the
Test Series, April - May 1973
Test No.
Date
Time
Liquor flow configuration
Flow
Liquor, gaL/m'.n
To G-l (bottom stage)
To G-2 (middle stage)
To G-3 (top stage)
Gas, ft3/miu
Liquor Ca
G-l-lb
G-l-OC
G-2-lb
G-2-OC
G-3-Ib
G-3-Oc
Liquor S/C^
G-l- fb
G-1-0C
G-2-fb
G-2-Oc
G-3-Ib
G-3-OC
Liquor pll
G-l-lb
G-1-0C
G-2-Ib
G-2-OC
G-3-fi
G-3-Oc
SO,, ppm
Entering G-l
Leaving G-l
Leaving G-2
Leaving G-3
Percent removal
KH3 leaving G-3, ppm
Temperatures, °F
Liquor
From G-,1
From G-2
From G-3
Gas
To G-l
From G-l
From G-2
From G-3
Exit gas
Ambient
Relative humidity, %
Predicted temperature at
which steam plume forms,
Reheat temperature, °F
Percent opacity
0
4/19
11:00 a.m.
•4— •
30
30
15
2700
28.1
-
9.52
9.48
1.35
1.88
0.76
-
0.58
0.59
0.62
O.dO
5.90
-
6 , 70
b,70
0.70
6.80
2600
1700
80
40
98.5
47
141
120
125
282
144
130
126
200
73
80
"Fe 152
200
25
1-1
4/20
1:00 p.m.
I-2A
4/20
1:45 p.m.
> Da*-* i T»r-i 1 1 n
I-2B
4/20
2:15 p.m.
I-2C
4/20
3:15 p.m.
-o
30
30
15
2700
13.21
13.09
4.90
4.73
-
0.10
0.82
0.83
0.55
0.57
-
0.86
5.75
5.75
7.25
6.95
7.20
7.10
2320
1860
400
100
95,6
47
131
127
126
290
132
127
125
150
84
49
120
150
65
33
30
15
2700
12.6
13.0
5.2
5.2
0.1
0.6
0.85
0.84
0.55
0.57
0.84
0.58
5.80
5.75
7.10
6.95
7.25
7.15
2440
2000
300
40
98.4
51
131
127
126
289
133
127
125
200
85
47
120
200
15
31
30
15
2700
14,0
13,8
4,9
5.0
0.1
0.4
0.81
0.82
0.56
0.57
0.85
0.61
5.80
5.75
7.10
6.95
7.25
7.15
2600
1880
16C
60
97.7
51
130
127
126
286
132
126
125
220
85
47
120
220
20
31
30
15
2700
.
_
_
.
_
.
_
.
.
.
.
.
5.80
5.90
6.80
6.70
6.50
6.50
2400
2040
160
60
97.5
51
131
126
125
290
133
127
125
225
82
42
124
22S
10
I-2D
4/20
3:40 p.m.
•^
30
30
15
2700
13.0
13.6
4.5
4.5
1.2
1.4
0.79
0.79
0.60
0.62
0.72
0.70
5.80
5.90
6.80
6.70
6. SO
6. SO
2400
1640
240
160
93.3
SO
131
128
125
290
133
127
126
200
84
42
120
200
15
w>les of active ammonia as ammonium sulfite-bisulfite per 100 moles of water.
blnlet sample.
C0utlet sample.
aMole ratio of S02:NH3.
Calculated temperature at which steam plume can form regardless of dilution; does
not include a solid particulate plume, that is, an ammonia salt plume.
987
-------
TABLE in (continued)
Summary 'of
Operating
Test Series, April
Test No.
Date
Time 5
Liquor flow
1-3
4/20
:5S p.m.
1-4
4/20
6:05 p.m.
configuration 4 — Recirculating— *
Flow
Liquor, gaL/ndn
To G-l (bottom stage)
To G-2 (middle stage)
To G-3 (top stage)
Gas, ft3/nin
Liquor Ca
G-l-lb
G-l-OC
G-2-lb
G-2-Oc
G-3-Ib
G-3-Oc
Liquor S/Cd
G-l-lb
G-1-0C
G-2-lb
G-2-Oc
G-3-Ib
C-3-QC
Liquor pH
G-l-Ib
G-l-OC
G-2-Ib
G-2-Oc
G-3-Ib
G-3-Oc
S02, ppm
Entering G-l
Leaving G-l
Leaving G-2
Leaving G-3
Percent removal
NH3 leaving G-3, ppm
Temperatures, °F
Liquor
From G-l
From G-2
From G-3
Gas
To G-l
From G-l
From G-2
From G-3
Exit gas
Ambient
Relative humidity, %
Predicted temperature
at which steam plume
forms, °Fe
Reheat temperature, °F
Percent opacity
29
30
15
2700
0.85
1.50
10.6
10.4
0.14
0.31
0.95
0.93
0.78
0.79
0.99
0.95
5,10
5.70
5.90
5.90
5.70
4.90
2520
1760
920
920
63. S
87
12J
129
122
270
124
129
121
160
83
42
124
160
Nil
30
30
15
2700
0.85
1.50
10.6
10.4
0.14
0,31
0.95
0.93
0.78
0.79
0.99
0.95
5.10
S.70
5.90
5.90
5.70
4.90
2520
1760
920
920
63. S
-
121
129
122
270
124
129
121
-
83
42
124
None
5
1-5
4/23
2:15 p.m.
31
31
15
2700
0.30
0.50
10.1
10.1
0.21
0.38
0.93
0.88
0.77
0.78
0.72
0.81
4.70
5.90
6.00
6.00
4.70
5.00
2560
2600
880
680
73.4
47
125
130
130
290
127
130
125
ISO
72
84
159
150
20
Data for the
- May 1973
I-6A
4/23
1:45 p
32
32
15
2700
0.30
0.50
10.1
10.1
0.21
0.38
0.93
0.88
0.77
0.78
0.72
0,81
4.70
5.30
6.00
6.00
4.70
5,00
2480
2520
880
720
70.9
47
127
ISO
127
290
128
130
Z25
200
70
84
160
200
5
1-7
4/23
.m. 2:40 p.n.
0
30
IT
2700
.
_
9.1 '
8.9
' 0.30
0.41
_
0.77
0.79
0.93
0.90
_
_
5. SO
5.80
5.10
4.20
2480
2480
020
780
68.5
79
_
128
124
290
274
138
124
150
72
80
154
ISO
30
I-8A
4/23
3:05 p.m.
0
30
17
2700
_
_
9.1
8.9
0.30
0,41
0.77
0.79
0.93
0.90
5.80
5.80
5.10
4.20
2480
2480
920
780
68.5
_
128
124
288
285
128
124
200
72
SO
154
200
5
1-13
4/24
3:15 p.m.
fc
30
30
15
2700
0.12
0,11
1.9
1.9
0.33
0.29
0.50
0.70
0.76
0.81
0.78
0.91
2.40
2.30
6.00
5.80
3.10
3.10
2480
2520
480
560
77.4
37
125
126
125
288
125
126
125
150
65
93
181
150
90/15 at lip
aMoles of active ammonia as ammonium sulfite-bisulfite per 100 moles of water.
^Inlet sample.
C0utlet sample.
dMole ratio of S02:NH3.
Calculated temperature at which steam plume can form regardless of dilution; does
not include a solid participate plume, that is, an ammonia salt plume.
988
-------
TABLE III (continued)
Summary of
Operating Data for the
Test Series, April - May 1973
Test No.
Date
Time 3
Liquor flow
configuration 4
Flow
Liquor, ga.l/min
To G-l (bottom stage)
To G-2 (middle stage)
To G-3 (top stage)
Gas, ft3/min
Liquor Ca
G-1-IU
G-1-0C
G-2-Ib
G-2-Oc
G-3-Ib
G-3-Oc
Liquor S/Cd
G-l-Ib
G-1-0C
G-2-Ib
G-2-QC
G-3-Ib
G-3-QC
Liquor pH
G-l-Ib
G-l-0=
G-2-Ib
G-2-Oc
G-3-Ib
G-3-QC
SO,, ppm
Entering G-l
Leaving G-l
Leaving G-2
Leaving G-3
Percent removal
NH- leaving G-3, ppm
Temperatures, °F
Liquor
From G-l
From G-2
From G-3
Gas
To G-l
From G-l
From G-2
From G-3
txit gas
Ambient
Relative humidity, %
Predicted temperature
at which steam plume
forms, *F«
I-14A
4/24
:45 p.
30
30
IS
2700
0.12
0,11
1.9
1.9
0.33
0.29
0.50
0.70
0.76
0.81
0.78
0.91
2.40
2.30
6.00
5.80
3.10
3.10
2600
2600
440
560
78.5
39
125
126
125
286
126
127
126
200
66
93
181
200
Reheat temperature, °F 0
Percent opacity
aMoles of active ammonia as
Inlet sample.
C0utlet sample.
"We ratio of S02:NH.
era 1 MI 1 a t «d temnerat 111
,.
re &t •
1-15
4/24
m. 4:30 p.m. 4
0
30
15
2700
_
1.8
1.7
0.23
0.24
_
0.78
0.86
0.95
0.95
_
_
6.00
5.70
3.10
2.90
2600
2540
440
600
76.9
34
_
128
126
286
268
127
125
150
66
92
181
150
80/30
i8T. Up
I-16A
4/24
:10 p.m.
>•
0
30
15
2700
.
1.8
1.7
0.23
0.24
_
0.78
0.86
0.95
0.9S
_
^
6,00
5.70
3.10
2.90
2600
2540
580
64
75.4
34
.
126
125
284
255
127
124
200
66
92
181
20(1
5
1-17
4/26
3:05 p.m.
30
30
25
3000
0.03
0.03
11.6
11.8
0.29
0.38
_
0.74
0.75
0.93
0.93
2.30
2.20
6.00
5.90
5.30
5. .50
2620
2440
600
680
74.0
89
124
128
124
280
124
128
124
200
60
96
198
200
40/5
aTlio
ammonium sulfitc-bisulfite per
jhieh steam nil
LUBA can \
rrvrm raaar
1-17 1-18
4/27
1:45 p.m. 1
S/l
:00 p.m.
» m « f.
1-18
5/2
4:00 p.m.
^
31
30
26
3000
0.03
0.03
11.6
11.8
0.29
0.38
_
0.74
0.75
0.93
0.93
2.30
2.20
6.00
5.90
5.30
5.50
2840
2560
680
400
85.9
70
120
128
121
272
120
128
120
160
57
51
170
160
15/10
aTliD
100 moles of
dlnss of dill
30
30
25
2900
_
11.62
11.72
2.06
2.39
.
0.75
0.75
0.93
0.89
2.20
2.20
6.00
5.90
5.60
5.60
2640
2480
560
440
83.3
93
125
134
129
295
125
133
127
200
74
64
141
200
10/5
aT lip
water.
nt ion* dot
30
30
25
2700
0.04
0.04
7.61
8.39
2.94
3.21
_
0.82
0.79
0.95
0.92
2.40
2.30
5.80
5.70
S.40
5.40
2520
2400
560
800
68.3
85
110
125
122
294
115
124
121
160
80
56
131
160
5
is
not include a solid particulate plume, that is, an ammonia salt plume.
989
-------
.
Figure 12. Operation with no water wash ahead of first absorber stage.
-------
c
-
Figure 13. Operation with water wash ahead of first absorber stage.
-------
The chloride concentration of the inlet flue gas was approximately
30 ppm and in the exit gas was about 3 ppm. The effect of chloride
removal on plume reduction could not be separated from the effects of
cooling and humidifying the gas ahead of the first absorber stage.
Heat losses were minimized by insulating the absorber system. The
difference between the liquor temperature and the outside absorber skin
temperature averaged about 1 °F. In operation without insulation,
differences as great as 30 °F have been measured. No definite conclu-
sions were drawn about how effective the insulation was in preventing
plume formation.
The reheat required to dissipate the steam plume is a function of
the ambient temperature and relative humidity. The reheat temperature
required to avoid formation of a steam plume can be predicted for a
given gas composition and known ambient conditions.9 The predicted
reheat temperature for each test is included in Table III. Reheating
the gas to a temperature above the predicted reheat temperature resulted
in an acceptable plume when a water wash was used ahead of the absorber.
In one test (1-4), an acceptable plume was obtained without reheating
(Figure 14). In this case, the gas leaving the absorber had a higher
temperature than the predicted reheat temperature. In those tests
without a water wash, reheating decreased the plume opacity and in two
tests (1-8 and 1-16) the decreased opacity reached an acceptable level.
Tests 1-17 and 1-18 were extended runs designed to show the reheat
required as a function of the ambient conditions (temperature and
relative humidity). The plume opacity was maintained at a constant
5 percent by adjusting the reheat temperature of the scrubbed flue gas
(maximum temperature set at 200°F). Data from these tests show that
the reheat requirements for an acceptable plume increase with increased
relative humidity. These data, as well as the predicted reheat tempera-
ture necessary to avoid formation of a steam plume and the opacity
reading of the stack,are given below.
Predicted Reheat
Temp Required Opacity Reading, %
Relative Ambient to Eliminate Reheat At 10 Feet
Humidity. % Temp, °F Steam Plume. °F Temp, °F Stack Exit Above"Stack
94 62 189 125 50 60
(No Reheat)
94 62 189 200 5 30
80 65 166 193 5 10
69 63 162 196 5 10
62 68 152 197 5 5
53 57 170 180 5 5
42 60 154 175 5 5
32 82 124 158 5 5
992
-------
\\ V '
r" ,
Figure 14. Acceptable plume with no reheat.
-------
Formation of a high-opacity plume several feet from the discharge
of the stack occurred on days when the relative humidity was high (see
above data). Figure 15 shows a plume reforming downwind from a clear
stack. Reheat temperatures required to avoid formation of a plume on
days of high relative humidity and low temperature ar.e impractical
to achieve.
Conclusions
The ammonia bisulfate process is a promising candidate for second
generation SC^ removal systems. The following conclusions are
reasonable for the bisulfate process:
• Under proper operation of the scrubber, fume formation
can be controlled in the scrubber while producing a
liquor having a high salt concentration (C > 12).
• For low salt concentrations, there is much greater
flexibility in the manner of scrubber operation.
However, such low concentration is prohibitively
expensive in the bisulfate process, but very reasonable
in the double alkali approach.
• Avoiding steam plume formation outside the stack on days
of high relative humidity and low temperature may be
impractical to achieve.
• Adequate separation of fly ash and ammonium sulfate
crystals has not been effective in the pilot plant
equipment; with proper equipment, adequate separation
is expected.
994
-------
-
Figure 15. Plume reforming downwind of stack.
-------
REFERENCES
1. Lepsoe, R. and Kirkpatrick, W.S. "SO- Recovery at Trail ,"
Trans. Can. Inst. Mining Met. XL, 399-404 (1937).
2. Hein, L.B., Phillips, A.B., and Young, R.D. "Recovery of
SCU from Coal Combustion Stack Gases." In Problems and
Control of Air Pollution (Frederick S. Mallatte, ed) ,
Reinhold, New York (1955) pp. 155-69.
3. Sulfur 80 (1), 36-37 (Jan. -Feb. 1969).
4. Rumanian Ministry of Petroleum Industry and Chemistry.
"Ammonium Sul fate" Brit. Pat. 1,097,257 (Jan. 3, 1968).
5. Johnstone, H.F. "Recovery of S02 from Waste Gases."
Ind. Eng. Chem. 29 (12), 1396-98 (Dec. 1937).
6. Alabama Power Company, "New Process of Fertilizer Manufacture
Announced." Mfr. Rec. 92(26), 53 (Dec. 29, 1927).
7. Ruben, Allen G. (Bohna Engineering $ Research Inc.). Private
Communication.
8. Hixson, A. W. and Miller, R. "Recovery of Acidic Gases." U. S. Pat
2,405,747 (Aug. 13, 1946).
9. EPA study by L. I. Griffin, Jr., (to be published).
Conversion of English Units to Metric Equivalents
Multiply To Obtain
English Unit B^ Metric Equivalent
ft /min 0.0283 m3/min
lb 0.454 kg
°F °C = |- (°F - 32) °C
3 3
gr/ft 2.288 g/m
in. 2.54 cm
ft 3.281 m
gal/rain 3.785 1/min
996
-------
AN EPA OVERVIEW OF SODIUM-BASED DOUBLE ALKALI PROCESSES
PART I. A VIEW OF THE PROCESS
CHEMISTRY OF IDENTIFIABLE AND ATTRACTIVE
SCHEMES
by
Dean Draemel
Research Branch
Control Systems Laboratory
National Environmental Research Center
Office of Research and Monitoring
U. S. Environmental Protection Agency
Research Triangle Park, North Carolina
997
-------
INTRODUCTION
"Double" or "dual" alkali scrubbing involves circulating a clear liquor
solution of a soluble alkali salt (Na, K or NH-), with scrubbing taking place
by absorption and reaction to form the bisulfite from the sulfite. The spent
scrubbing liquor is treated with limestone and/or lime which precipitates the
absorbed sulfur as CaSO^ (and possibly some CaSOJ and regenerates the alkali
scrubbing solution. Refinements may involve additional regeneration of the
sulfate which is formed by oxidation.
Double alkali flue gas desulfurization processes have received increased
attention in recent years because of some potential advantages over competing
scrubber systems.
The circulation of a clear liquor removes many errosion, corrosion and sol it
deposition problems. The regeneration using limestone and/or lime is relatively
cheap and simple and the solid formed is acceptable as a throwaway product. The
system appears to be versatile in terms of modes of operation and may even be
used to produce a salable product. Both installed and operating costs for
such systems appear to be very competitive with other flue gas desulfurization
systems at comparable levels of development. '
A number of organizations, including the Environmental Protection Agency,
have become involved with major double alkali development efforts. General
Motors has been heavily involved in the development and design of double alkali
i?\
systems for use in its industrial facilities/ ' Other major double alkali
development efforts have been conducted by the Tennessee Valley Authority,
Arthur D. Little, Inc., FMC Corporation, Envirotech Corporation and others. The
998
-------
Environmental Protection Agency 1s funding a contract with Arthur D. Little, Inc.
to aid the development of this promising flue gas desulfuHzatlon technology.
Major aspects of the process chemistry and the attractive operating schemes
which are under development by various organizations will be discussed 1n this
paper. The particular systems under development by the various organizations
are discussed 1n greater detail 1n Part II of this paper.
The mention of company or product names 1s not to be considered as
endorsement or recommendation for use by the Environmental Protection Agency.
999
-------
PROCESS DESCRIPTION
Absorption
A simple flow scheme for a double alkali process is shown in Figure 1.
The clear alkali solution is circulated through a scrubber where SOp is absorbed
and some oxidation takes place. The type of scrubber used depends somewhat on the
concentration of the alkali scrubbing solution, but the clear scrubbing liquor
circulated allows flexibility in scrubber selection. Some of the effluent liquor
from the scrubber is recirculated and the remainder is sent to the regeneration
sys tern.
Regeneration
The liquor from the scrubber is treated with either limestone or lime
(Ca(OH)2) to precipitate the absorbed sulfur as calcium solids. After separating
the precipitated solids, the regenerated scrubbing liquor, which has a high
sulfite to bisulfite ratio, is sent back to the scrubber loop.
Oxidation and Sulfate Control
Oxidation of absorbed sulfur may require additional regeneration of the
sulfate formed. A separate stream from the primary bisulfite/sulfite regeneration
system may be treated specifically for sulfate regeneration to remove sulfate from
the system as an insoluble solid product.
1000
-------
FLUE GAS OUT
1
1
FLUE
GAS
IN
OPTIONAL SOFTENING
I
EFFLUENT
RECIRCULATIONTANK
»
SOLIDS
OPTIONAL SULFATE
REGENERATION
SOLIDS
REGENERATION
Figure 1. Double-alkali process flow scheme.
-------
PROCESS CHEMISTRY
Generally, sodium-based double alkali systems are the only systems receiving
major development emphasis. Since sodium-based systems are under major active
development in this country and appear to have promise with respect to flue gas
desulfurization, this paper will deal only with sodium-based systems.
Absorption
The scrubbing step involves absorption of SCL in a scrubbing solution of
sodium sulfite, bisulfite and sulfate. The absorbed S02 reacts with the sulfite
in solution to form bisulfite (reaction 1). The scrubbing solution is actually
a buffer system of sulfurous acid (reaction 2). The pH range over which the
scrubber systems may operate is from * 6 to 8.5. Figure 2 shows the major species
in solution as a function of pH. Solutions of higher pH, with sulfite as the
major species, are circulated into the scrubber while the lower pH (higher
bisulfite concentration) spent solution from the effluent redrculatlon tank Is
sent to the regeneration system.
Na^SO., + SO, + H,0 * 2 NaHSO, (1)
t J c. £. \ * j
S02(aq) + H?0 -—* H2$03 —^ H+ + HSOg —^ 2H+ + SO* (2)
Regeneration
Spent scrubbing solution, with a low sulfite to bisulfite ratio, which is
sent to the regeneration system is treated with limestone or lime to precipitate
the absorbed S02 as calcium solids by reactions 3 and 4. The bisulfite is
neutralized and sulfite is regenerated as the "active" scrubbing agent.
1002
-------
o
o
2- Distribution of aqueous sulfite species as a function of pH.
-------
CaC03 + 2 NaHSOg -—* CaS03 + Na2$03 + H20 + C02 (3)
Ca(OH)2 + 2NaHS03 -—*• CaS03 + Na2S03 + 2 HgO (4)
The EPA has performed batch laboratory experiments to characterize reactions
3 and 4 in simulated scrubber effluent solutions. ' The results of these
experiments are consistent with laboratory and pilot plant results of other
organizations involved in double alkali development programs. Hold times and
utilizations using limestone and lime are mainly dependent on solution concentra-
tions, temperature, agitation level and reactant stoichiometry. Generally, using
limestone, hold times on the order of an hour or more are needed and utilizations
on the order of 75-85% may be realized. Generally, using lime, hold times on
the order of around 10 minutes may be used and utilizations of 90% or more may
be realized. The benefits of using lime for the bisulfite/sulfite regeneration
may be offset by higher chemical costs, slaking requirements, scaling potential
and possible pH control problems.
In addition to the bisulfite/sulfite regeneration, lime may be used to
react further with the sulfite and sulfate present in solution by reactions
5 and 6. The reaction between lime -and sulfate is limited by the equilibrium
hydroxide ion concentration of approximately 0.15M.* '
Ca(OH)? + Na?S07 * CaSO, + 2NaOH (5)
£ £ 0 ^~'"' • • j
Ca(OH)2 + Na2S04 -—^ CaS04 + 2NaOH (6)
Sulfite is the more active species and if more hydroxide is formed by reaction
with sulfite (reaction 5) than the approximately 0.15M equilibrium value for
reaction with sulfate (reaction 6), the reaction between lime and sulfate will
not occur. The effective regeneration of sulfate using lime thus depends on the
1004
-------
sulfite ion concentration present and the hydroxide ion concentration formed
from reaction 5. Batch scale laboratory experiments conducted by the
Environmental Protection Agency have indicated that increased sodium sulfate
concentrations tend to favor reaction with sulfate over sulfite when solutions
are treated with an equilibrium amount of lime. With 1.78M (^ 20 wt?) sodium
sulfate, roughly 50% of the final (3-hour batch reaction) hydroxide is from
reaction with sulfate at 0.066M initial sulfite. With 0.67M HO wt£) sodium
sulfate, roughly 50% of the final (3-hour batch reaction) hydroxide is from
reaction with sulfate at less than .055M initial sulfite. Figure 3 shows the
relative amount of reaction between lime and sulfite and sulfate in solution
as a function of initial sulfite concentrations. Batch reactions were conducted
at 52°C (125°F) with 1.78M (•>, 20 wU) sodium sulfate solutions and roughly
equilibrium (0.078M/1) lime addition. This indicates that lime can be used to
regenerate sulfate in the presence of dilute sulfite concentrations.
Oxidation and Sulfate Control
Oxidation of bisulfite and sulfite in the scrubber liquors occurs by
reactions 7 and 8. The sulfate formed by oxidation must be removed from the
system either as a solid product by- regeneration or as soluble salts by a purge.
2 NaHS03 + 1/2 02 -—* Na2S04 + S02 + HgO (7)
Na,SO, + 1/2 0, _—* Na,S04 (8)
L. 3 <_ V~~ L. *r
The effective removal of sulfate as a solid product is highly desirable
from the standpoint of environmental acceptability. Completely closed loop
operation is an ideal situation, while the opposite extreme would Involve a
balance between sulfate formed by oxidation and sulfate lost as soluble salts
1005
-------
o
o
en
U.1D
0.10
.", moles/1 itei
° 0.05
0
0
I
•• •*
i
j
j ;
- 7
//
«
.02 0.04
1
TOTAL OH'
. — —
•
OH'FROMSOj2' ,
/OH"FROMS042' J
/ ii
i «
|l
li
0.06 0.
1
— — '
1
1
1
1
1
i !
i ii
08 0.10 .0.1
S032' CHARGED, moles/i900 cc
Figure 3. Batch reactions-Ca(OH)2/Na2S03, Na2S04[equilibrium OH".
1.60 moles/900 cc Na2S04 @ 52° C (125° F)T,
-------
leaving the system with the solids or possibly as a simple purge. Considering
the volume of sulfur removed and the fractional oxidation levels expected, the
loss of all oxidized sulfur from the system as soluble salts represents a
serious water pollution potential. Attempts must be made to regenerate the
sulfate formed and remove it as an insoluble solid product.
Oxidation by reactions 7 and 8 is a function of solution concentration,
oxygen mass transfer to the scrubbing solutions and probably a number, of other
factors such as traces of catalysts or inhibitors that may be present from sources
such as fly ash. At least one oxidation inhibitor for this type of process is
(4\
presently available/ ; The oxidation reactions are not well defined or well
understood. So far, most efforts have attempted to cope with oxidation problems
without developing a complete understanding of the factors involved. The EPA
recently funded a grant (No. 800303) with the University of Illinois to determine
the mechanisms and kinetics of oxidation in soluble alkali scrubbing systems.
Sulfate regeneration with lime is possible in the presence of low sulfite
ion concentrations (<0.08M) subject to the limitation of the equilibrium hydroxide
ion concentration mentioned previously. The low sulfite ion concentrations
necessary for effective sulfate regeneration using lime are a disadvantage because
of the large volume of the dilute scrubbing liquor necessary for effective S02
removal. As an alternative to sulfate regeneration using lime, other sulfate
removal techniques are under development which may be used with concentrated
scrubbing liquors (sulfite concentrations of ^ 0.5M).
Sulfate removal would generally be carried out by treating a side stream
from the scrubber liquor loop. An example of this approach is shown by the
fa\
following reactions.v/
1007
-------
H2S04 + CaS03 * 1/2 H20 + 1/2 HgO •* S02 + CaS04 ' 2 HgO (9)
S02 + CaS03 ' 1/2 H20 + Na2S04 + 5/2 H20 + CaS04 ' 2H20 + 2 NaHS03 (10)
A side stream from the scrubber discharge is mixed with calcium sulfite
solids and sulfuric acid is added. The net effect is that sodium sulfate
in the scrubbing liquor is converted to solid calcium sulfate and sodium
bisulfite is generated for recycle to the main regeneration system. Additional
limestone or lime is required to remove the sulfur added as sulfuric acid, i.e.
the bisulfite generated by reaction 10. Roughly 125-150% of the theoretical
sulfuric acid is required, based on the amount of sodium sulfate reacted. It
is estimated that with 7% oxidation, roughly 9% of the solids produced would be
derived from the sulfuric acid. This would add to the operating cost of a
throwaway system.
Another alternative for sulfate removal being considered is the selective
crystallization of sodium sulfate decahydrate by cooling of scrubbing liquors,
in which the decahydrate will crystallize out at about 32°C (90°F). The crystals
could then be separated and somehow treated to regenerate active sodium and
precipitate a relatively insoluble solid product. Alternative methods of sulfate
control are being considered.
Scaling
Scaling may occur if conditions develop in the scrubber which cause super-
saturation of calcium sulfate. The solubility product, K , for calcium sulfite
is roughly 2 orders of magnitude less than the K value for calcium sulfate
although steady state sulfite and sulfate concentrations may develop for which
either solid may precipitate. Supersaturation and solids precipitation is desired
1008
-------
in the regeneration section but calcium 1on concentrations must be controlled
in the liquor returning to the scrubber to guard against supersaturation and
precipitation in the scrubber. The relative solubilities of the various
calcium compounds are shown in Figure 4.
The presence of solid calcium hydroxide and/or solid calcium sulfate at
equilibrium with the liquid in the regeneration system produces high calcium
ion concentrations on the order of 300-400 ppm. These high calcium ion con-
centrations must be reduced considerably to guard against calcium sulfate
supersaturation and scaling in the scrubber.
Calcium ion concentrationsmay be controlled by "softening" reactions
11 and 12.
Na2C03 + Ca(OH)2 -—* 2 NaOH + CaCOg (11)
C02 + H20 + Ca"1"1' -—* CaC03 + 2H* (12)
Treatment of scrubbing liquor with sodium carbonate (reaction 11) or
carbon dioxide (reaction 12) before being sent to the scrubber leads to precipita-
tion of calcium carbonate and thus removes calcium ions from solution. The
reduction of calcium Ion concentrations in this manner ensures against scaling
in the scrubber. Addition of sodium carbonate has the advantage of both
softening and replacing sodium losses from the system. Addition of Na2C03 has
the disadvantage of requiring a possible sodium purge if the amount required
for softening exceeds the amount required to replace sodium losses.
Addition of carbon dioxide, which is acidic, also requires Increased use
of regeneration chemicals for neutralization. The net effect of adding either
sodium carbonate or carbon dioxide 1s to reduce calcium ion concentrations
1009
-------
o
o
0.28
0.24
0.20
o
s
0.16
ro
>^"
5 0-12
0.08
0.04
i i—r
Ca(OH)2
— -CaS03»2H20
-^— CaCCU
I
I
10
20
30
40 SO 60 70 80
TEMPERATURE,0 C
Figure 4. Relative solubilities of calcium compounds.
90 100
-------
entering the scrubber to a level below the saturation value for calcium
sulfate.
Process Chemistry—Summary
Before going on to discuss possible modes of operation for sodium-based
double alkali systems, a brief summary of the process chemistry discussion
will be made. Sulfur oxides are absorbed into a sulfite/bisulfite buffer solution
This absorption shifts the pH down and increases the bisulfite concentration.
The liquor from the scrubber may then be treated with limestone to precipitate
calcium sulfite and neutralize the bisulfite. The liquor from the scrubber
or the limestone reaction vessel may be treated with lime to precipitate calcium
sulfite and possibly calcium sulfate. Oxidation of absorbed sulfur requires
the regeneration of sulfate; a few possible methods have been discussed. Calcium
ion concentrations in the scrubber can be controlled by softening steps in the
liquor loop.
1011
-------
DOUBLE ALKALI OPERATING SCHEMES
Five relatively distinct modes of operation have been identified for
sodium/calcium double alkali scrubbing systems. These modes will be discussed
very briefly with some advantages and disadvantages noted for each. More
detailed discussion of these modes, with emphasis on commercial developments,
is presented in Part II of this paper. The five operating schemes to be
discussed are:
1. Limestone and lime regeneration, dilute active alkali, with sulfite
softening.
2. Lime regeneration, dilute active alkali, with sulfite softening.
3. Lime regeneration, concentrated active alkali with side stream
sulfate treatment (removal).
4. Limestone regeneration, concentrated active alkali with side
stream sulfate treatment (removal).
5. Lime regeneration, dilute active alkali, with carbonate softening.
Limestone and Lime Regeneration, Dilute Active Alkali,with Sulfite Softening
The first mode identified is a double-loop (limestone/lime), dilute alkali
system. Dilute alkali indicates that the sulfite or alkali concentration entering
the scrubber will be less than approximately 0.08M. Dilute alkali scrubbing
solutions allow sulfate regeneration with lime in this mode of operation.
"Double loop" refers to the regeneration method. Spent scrubbing liquor is
treated with limestone to neutralize the bisulfite and precipitate calcium
sulfite. A stream from this limestone reaction vessel is then treated with lime
to precipitate calcium sulfite and sulfate and generate sodium hydroxide.
1012
-------
The relative amount of reaction between lime and the sulfite and sulfate
is determined mainly by sulfite and sulfate concentrations and lime stoichiometry.
The liquor from the lime reaction vessel is returned to the limestone reaction
vessel to ensure unsaturated calcium ion concentrations going to the scrubber.
Desaturation is accomplished because of the presence of sulfite ion in the
limestone reactor. The reduction of calcium ion concentrations by calcium
sulfite precipitation is referred to as sulfite softening. Sodium carbonate
may be added to provide additional softening and make up any sodium losses.
This scheme has the disadvantage of requiring relatively long hold times
for the reaction with limestone, requiring two major reactors, and having to
circulate large volumes of relatively dilute scrubbing liquor. The advantages
appear to be the use of relatively inexpensive limestone and lime for regeneration
of both bisulfite and sulfate, and relatively simple equipment requirements.
Lime Regeneration. Dilute Active Alkali, with Sulfite Softening
The second mode of operation is a double-loop (lime only), dilute alkali
system. This system arrangement is very similar to the first system discussed.
Spent scrubbing liquor is treated with lime to neutralize the bisulfite only.
A stream from the "lime/bisulfite reaction vessel is then treated with additional
lime in a separate reaction vessel to react with sulfite and sulfate, forming
calcium solids and sodium hydroxide. The liquor from this second reaction vessel is
returned to the lime/bisulfite reaction vessel to ensure unsaturated calcium ion
concentrations with respect to calcium sulfate by "sulfite"softening of the
liquors before entering the scrubber.
This flow scheme has the disadvantage of difficult pH and calcium ion control
in the regeneration loop because of the addition of a relatively concentrated
1013
-------
strong base (lime) to a dilute solution of a weak acid. The 'regeneration
loop requires two reactors, and large volumes of relatively dilute scrubbing
liquor must be circulated to the scrubber. The advantages of this system
appear to be the relatively short hold times necessary for regeneration using
lime and the relatively simple equipment requirements.
Lime Regeneration, Concentrated Active Alkali, with Side Stream Sulfate Treatment
The third mode of operation involves lime regeneration, concentrated alkali
solutions and side stream treatment for sulfate removal and control. Lime is
used for the bisulfite neutralization of the concentrated (^ 0.5M) alkali
scrubbing solutions. A side stream is taken from the liquor loop and treated
specifically for sulfate removal. The major difference between dilute (<0.08M)
and concentrated (% 0.5M) alkali systems is that lime cannot be used to regenerate
sulfate in concentrated (greater than ^O.OSMJ sulfite solutions because of the
equilibrium hydroxide ion concentration mentioned previously. Side stream treatment
(3)
has been accomplished with proven technology (reactions 9 and 10)v ' but sulfate
removal and control is the subject of intense development efforts since proven
technology is limited in this critical area.
This third flow scheme has the disadvantage of possible sulfate regeneration
complications. The advantages of this flow scheme are: the relatively short
hold times necessary for the bisulfite/sulfite regeneration With lime, high
reactant utilizations, an advanced state of development and high S02 removal
efficiencies with low flows of the concentrated alkali scrubbing liquors. This
third mode of operation appears to be very promising but universal large-scale
successful application depends on effective treatment for sulfate removal and
control.
1014
-------
Limestone Regeneration, Concentrated Active Alkali, with Side Stream Sulfate
Treatment
The fourth mode of operation involves limestone regeneration, concentrated
alkali solutions and a side stream treatment for sulfate removal and control.
Limestone is used for the bisulfite neutralization of the concentrated (^0.5M)
alkali scrubbing solutions. A side stream is taken from the liquor loop and
treated specifically for sulfate removal. This scheme, like the previous one,
has the possible disadvantage of sulfate regeneration complications. This scheme
also has the disadvantage of requiring long hold times for the bisulfite/sulfite
regeneration using limestone. Advantages of this flow scheme appear to be high
S0« 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.
Lime Regeneration, Dilute Active Alkali, with Carbonate Softening
The fifth mode of operation is lime regeneration, dilute alkali with
carbonate softening for calcium ion (scaling) control. The spent scrubbing
liquor is treated with lime to neutralize bisulfite and react with sulfite and
sulfate. The dilute active alkali allows simultaneous reaction with .both the
sulfite and sulfate by reactions 6 and 7 even though the equilibrium hydroxide
ion concentration for sulfate regeneration (reaction 7) limits the extent of the
reactions. The calcium ion concentration in the liquor from the lime treatment
tank would be high and represent serious scaling potential unless it is reduced
considerably below saturation values for calcium sulfate and hydroxide. Carbon
dioxide may be added to the system to precipitate calcium carbonate (reaction 12)
and significantly reduce calcium 1on concentrations in the liquor before entering
the scrubber. Sodium carbonate may be used to soften (reaction 11) and make
1015
-------
up sodium losses. This scheme has the disadvantages of possible Inadequate
sulfate regeneration, possible scaling potential and having to circulate large
volumes of relatively dilute scrubbing liquors. The advantages appear to be
the ability to deal with high oxidation levels, relatively simple equipment
requirements and an advanced state of development.
Modes 2 and 5 are very similar except the double-loop regeneration method
of mode 2 would probably provide both sulfate regeneration and calcium ion
control without a major softening step involving the addition of chemicals
such as sodium carbonate or carbon dioxide. All advantages and disadvantages
discussed serve mainly to compare these five modes relative to each other and
not to other scrubbing systems. The advantages and disadvantages of these double
alkali systems relative to other scrubber systems have been discussed in the
introduction.
1016
-------
SUMMARY
The major process chemistry for sodium/calcium double alkali scrubbing
of SO 1s well defined and reasonably well understood. A number of relatively
A
distinct modes of operation have been Identified and are at various stages
of development. Development efforts are presently being directed at effective
sulfate regeneration In concentrated sulflte solutions, control of scaling, and
control and understanding of the oxidation reactions. SolIds characteristics
such as settleabllity, cake moisture and cake "fixing" are also being studied.
Many minor variations are possible in the flow scheme configurations discussed;
however, the description of these schemes was meant mainly to show the versatility
of double alkali systems and the range of development efforts.
1017
-------
REFERENCES
1. Rochelle, G., Economics of Flue Gas Desulfurization, paper, Flue Gas
Desulfurization Symposium, May 14-17, 1973, New Orleans, La.
2. Phillips, R. J., Sulfur Dioxide Emission Control for Industrial
Power Plants, paper, Second International Lime/Limestone Wet Scrubbing
Symposium, Nov. 8-12, 1972, New Orleans, La.
3. Draemel, D. C., Regeneration Chemistry of Sodium-Based Double-Alkali
Scrubbing Process, Environmental Protection Technology Series,
EPA-R2-73-186, March 1973.
4. Elder, H. W., Princiotta, F. T., Hollinden, D. G. A. and Gage, Dr. S. J.,
Sulfur Oxide Control Technology Visits in Japan - August 1973, U. S.
Government Interagency Report, Muscle Shoals, Ala., Oct. 30, 1972.
1018
-------
AN EPA OVERVIEW OF SODIUM-BASED
DOUBLE ALKALI PROCESSES
PART II
STATUS OF TECHNOLOGY AND DESCRIPTION
OF ATTRACTIVE SCHEMES
by
Norman Kaplan
Control Systems Laboratory
Office of Research and Monitoring
Environmental Protection Agency
Research Triangle Park, North Carolina
1019
-------
ABSTRACT
AN EPA OVERVIEW OF SODIUM-BASED
DOUBLE ALKALI PROCESSES
PART II. STATUS OF TECHNOLOGY AND
DESCRIPTION OF ATTRACTIVE SCHEMES
Important criteria for evaluation of double alkali schemes are
given.
Flow sheets for potentially attractive schemes of operation
are presented and described with emphasis on evaluation of each with
respect to the criteria given.
A brief description of selected pilot plant and prototype
experience and future plans for further development of double alkali
systems is given. In addition, an effort is made to categorize each
of the known commercially available double alkali systems presently
being marketed, in accordance with the identified schemes.
The EPA/A. D. Little double alkali development program plan and
general philosophy are described.
Double alkali technology, although less advanced, is presented
as being potentially more reliable and less costly than lime/limestone
technology.
1020
-------
ACKNOWLEDGEMENTS
The author wishes to express appreciation for assistance in
preparation of this paper to Frank T. Princiotta, Chief, Engineering
Test Section, for assistance with technical presentation, to
Charlotte Bercegeay for typing and to Beverly Tilton for assistance
in preparation of diagrammatic material; all of these personnel are
assigned to Environmental Protection Agency components in the
Research Triangle Park area of North Carolina.
Appreciation is also expressed toward the many personnel repre-
senting Arthur D. Little, Inc., Chemical Construction Corporation
(Chemico), Combustion Equipment Associates, Envirotech Corporation,
FMC Corporation, General Motors Corporation, The Southern Company,
Utah Power S Light, and Zurn Industries (Zum Air Systems) for their
informative input and cooperation, without which a presentation of
this type would not be possible, and to Dr. Ando for his summary of
some of the technology developed by Japanese companies.
1021
-------
NOTES
1. Company Names and Products.
The mention of company names or products is not to be considered
an endorsement or recommendation for use by the U. S. Environmental
Protection Agency.
2. Units of Measure.
EPA policy is to express all measurements in Agency documents in
metric units. When implementing this practice will result in
undue cost or difficulty in clarity, NERC/RTP is providing con-
version factors for the particular non-metric units used in the
document. Generally, this paper uses British units of measure.
For conversion to the Metric system, use the following equivalents:
British
5/9 (°F-32)
1 ft
1 ft3
1 grain
1 in.
1 in. 2
1 in.3
1 Ib (avoir.)
1 ton (long)
1 ton (short)
1 gal
Metric
°C
0.3048 meter
0.0929 meters2
0.0283 meters3
0.0648 gram
2.54 centimeters
6.452 centimeters2
16.39 centimeters3
0.4536 kilogram
1.0160 metric tons
0.9072 metric tons
3.7853 liters
1022
-------
INTRODUCTION AND BACKGROUND
Based on presently known firm commitments of utilities, the prevalent
method of stack gas cleaning for reduction of sulfur oxide emissions fr
fossil fueled power plants is the wet lime/limestone scrubbing process.
Unfortunately, this method of control is not the perfect solution to the
problem. Problems have been identified with wet lime/limestone scrubbing
of flue gas which are mainly due to the fact that the absorbent and the
products of the absorption reaction are present in the scrubbing equipment
in the form of a slurry which has a tendency to cause solids build-up,
scaling and erosion in the equipment; these can reduce the overall relia-
bility of the system.
In order to circumvent the problems inherent with the use of an
absorbent slurry, scrubbing with a soluble alkali is a logical first
consideration; however, by itself, this method is not satisfactory since
it would necessitate the disposal of large quantities of liquid containing
dissolved sulfur compounds, thus creating a water pollution problem. The
"Double Alkali" process, on the other hand, combines the advantages of
wet lime/limestone scrubbing with those of soluble alkali scrubbing by
scrubbing with soluble alkali and following with regeneration of the
soluble alkali and precipitation of the absorbed sulfur oxides as insoluble
calcium salts. The regeneration reaction involving the use of calcium
compounds, which are considered to be potentially scale producing, is caused
to occur outside the scrubber system where scaling tendencies can be more
adequately controlled. In addition--whereas with lime/limestone systems,
S02 removal is limited by solids dissolution rate--no such limitation is
inherent with soluble alkali scrubbing systems. The net effect of the use
of a double alkali process is the same as that of a wet lime/limestone
process in that limestone and/or lime is consumed and calcium sulfite and
sulfate are generated as a disposable waste product. An important excep-
tion to the previous statement is the fact that the disposable waste may
be moderately contaminated with soluble salts which can potentially cause
a water pollution problem.
The term "double alkali" is used generally to describe systems which
employ a soluble aqueous alkali (e.g., Na+, K+, NH4+ based) to scrub
acidic sulfur oxides from flue gas and then produce an insoluble throwaway
product by reacting the scrubber effluent with limestone and/or lime.
This paper will address only sodium-calcium systems as these are the
prevalent schemes under study and commercially available in the United
States. An overview of these systems is presented with only some reference
to the chemical complexities associated with them. The chemistry of these
systems has been discussed in detail in part I of this paper; in summary,
however, the principal scrubbing reaction in the many variations of the
double alkali process discussed involves the absorption of sulfur dioxide
by an aqueous solution of sodium sulfite, converting the sulfite to
bisulfite.
1023
-------
IMPORTANT CONSIDERATIONS
In the discussion of various double alkali schemes, it is considered
helpful to compare these in relation to lime/limestone scrubbing systems.
The double alkali process has been considered a second generation version
of the wet lime/limestone process and therefore comparison is considered
reasonable.
Sulfate Removal
As in the operation of a lime/limestone system, some of the sulfur
absorbed in double alkali processes appears in the system in the oxidized
form as sulfate. In order to operate a double alkali system in a steady
state continuous manner, provisions must be made to remove sulfate from
the system at the rate at which it is formed in or absorbed into the
system. This is not the case with lime/limestone systems since the
oxidation product is insoluble gypsum. Failure to allow for sulfate
removal in double alkali systems will ultimately result in precipitation
of sodium sulfate somewhere in the system: quite possibly in the scrubber
circuit. The known possible methods of removing sulfate from the system
are:
1. purging a waste liquid stream containing dissolved
sodium sulfate;
2. removing wet calcium sulfite/sulfate/flyash sludge
containing dissolved sodium sulfate in the occluded
water;
3. removing solid sodium sulfate from the system for
sale or other disposal by selective crystallization
or operation of another concentration process; and
4. precipitating the sulfate as insoluble calcium sulfate
to be disposed of with the waste product sludge.
The first method is ecologically undesirable due to its contributing
to water pollution by emission of dissolved sodium salts. In addition,
depending on the concentration of active alkali (sulfite/bisulfite), the
liquid discharge may have a high COD value.
The second method disposes of the sulfate in a more subtle manner
than the first. Here a liquid stream is actually purged; however, it
is much less obvious since the liquid is contained in the solid waste
product, which appears to be relatively dry even when containing up to
40 percent moisture. The potential contribution to water pollution in
this case is due to the leachability of the waste product sludge and
water run-off from the sludge disposal area. A highly leachable sludge
will result in slow contamination of the ground water table in the
disposal area while water run-off would lead to contamination of surface
water by soluble salts. An acceptable alternative to direct disposal of
1024
-------
wet solids as outlined above, would be disposal of a treated sludge
which contains the required amount of soluble sulfates but in a "fixed"
solid, unleachable, innocuous form. Although at present there is no
demonstrated commercially available system to accomplish this, inclusion
of double alkali sludge as part of a current program concerned with
ecological treatment and disposal of lime/limestone wet scrubbing sludge
is under consideration by EPA.(2)
The third method cited for sulfate removal involves the separation
of sodium sulfate crystal from the scrubber liquor by a heating-cooling
cycle or possibly by operation of a vapor compression cycle such as the
system being marketed by Resources Conservation Corporation as a brine
concentrator. To date, these methods have not been applied to double
alkali systems and therefore more evaluation is necessary to determine
their applicability. One possible problem created in removal of
sodium sulfate crystal is the ultimate disposal of this soluble salt.
Since the market for large quantities of sodium sulfate is sparse at
best, dissolution and treatment with lime to produce insoluble calcium
sulfate (gypsum) may be the acceptable short-term solution to the
problem.
The fourth method, precipitating calcium sulfate directly from
the scrubber liquor, can only be accomplished under carefully controlled
conditions. Basically, calcium sulfate and sulfite will not be pre-
cipitated in appreciable quantities, simultaneously, in neutral or
basic solution with lime treatment, unless the [S04=]/[S03=] concentration
ratio in solution is relatively high. A simple explanation for this is
that CaSOj is much less soluble than CaS04 and thus is preferentially
precipitated from solution. Since the concentration of sulfate is
limited by the solubility of sodium sulfate (approximately 3 molar) the
active alkali, sulfite, must be relatively dilute in order to maintain
a high [S04=]/[S03=] ratio and thus precipitate calcium sulfate upon
treatment with lime. (See part I of this paper.)
A variation of the fourth cited method, precipitating calcium
sulfate from the scrubber liquor, was developed by the Japanese. In
this variation, a small portion of the scrubber liquor (a "slip stream")
is treated with sulfuric acid and calcium sulfite to precipitate calcium
sulfate. Adjustment of the pH with sulfuric acid allows the calcium
sulfite to go into solution as soluble calcium bisulfate, thus increasing
the calcium ion concentration in solution until the solubility of calcium
sulfate is exceeded and calcium sulfate is precipitated from solution.
The major advantage of this variation is that concentrated active alkali
can be used in the scrubber loop. The obvious disadvantage is that
sulfur is added to the system as sulfuric acid, procured from another
source, and must be removed along with the sulfur in the flue gas.
1025
-------
Scale Prevention
One of the primary reasons, and probably the most important, for
development of double alkali processes was to circumvent the scaling
problems associated with lime/limestone wet scrubbing systems.
Therefore, an obvious consideration in any double alkali system is
the ability of the system to operate in a non-scaling manner.
In order to eliminate scaling tendencies, the calcium concentration
in the scrubber liquor must be held to a minimum. Recirculation to the
scrubber loop of supernatant liquor from a lime reaction vessel which
is controlled to precipitate sulfite and sulfate is not acceptable. The
calcium concentration of this liquor would be high enough to cause
scaling problems since it is saturated with respect to calcium sulfate
and calcium hydroxide. Figure 1 shows (with straight line interpolation
between two data points) relative solubilities in water of the various
calcium compounds present in a typical double alkali system. The
figure shows that CaSO, and Ca(OH)2, although only slightly soluble,
are about two orders of magnitude more soluble than CaCOj and CaSC^.
In order to minimize the calcium concentration in the scrubber loop, the
liquor returned to the scrubber from the regeneration process should be
saturated with respect to CaSO- or CaCQ$ but not with respect to CaSO^
or Ca(OH)2.
It should be noted that Figure 1 is not meant to specify the exact
concentrations to be expected in double alkali systems but only to
point out the direction to go in order to minimize scaling tendencies.
The actual solubilities of these calcium species in real systems will
be greatly affected by such important factors as pH, ionic strength
and the presence of other chemical species.
In many cases scaling problems are not only due to high soluble
calcium values related to saturated solutions of CaS04 and Ca(OH)2 but
are also possibly due to supersaturation with respect to any of the
calcium compounds. In order to reduce scaling due to supersaturation,
provisions to maintain high concentrations of suspended solids in the
reaction zones of the regeneration equipment are necessary. This can
be accomplished by recirculating precipitated solids to the reaction
zones. This technique is stressed in the Envirotech system which is
described under "Status of Technology" later in this paper. It has
also been referred to as a crystal seeding.
In general, efforts to minimize scaling rely upon:
1. carbonate softening (i.e., reduction of soluble calcium),
2. sulfite softening, and
3. crystal seeding techniques.
1026
-------
0.30
0 '
40 60
TEMPERATURE, °C
Figure 1. Solubility of calcium compounds.
1027
-------
Water Balance
In order to operate a closed system to avoid potential water
pollution problems, system water balance is a primary concern. Water
cannot be added to the system at a rate greater than the normal water
losses from the system.
Generally there is a tendency to add fresh water to a scrubbing
system to serve many purposes. These include:
1. saturation of flue gas,
2. pump seal needs,
3. demister washing needs,
4. slurry make-up needs, and
5. waste product washing.
On the other hand water should only leave the system in the following
ways:
1. evaporation by the hot flue gas,
2. water occluded with solid waste product, and
3. water of crystallization in solid waste product.
Careful water management, part of which is the use of recycled
rather than fresh water wherever possible, is necessary in order to
operate a closed system.
Waste Product Washing
As previously indicated, disposal of wet solid waste containing
soluble salts is ecologically undesirable. In addition, allowing
active alkali or sodium salts to escape from the system is an important
operating cost factor. Sodium make-up to double alkali systems is
usually accomplished by adding soda ash (recently quoted at $42 per ton)
at some point in the system. Thus, both ecological and economic con-
siderations dictate that waste product washing is desirable. With lime/
limestone systems, there is no need to wash the solid waste product.
S0? Removal
Generally, the concentrated active alkali systems have a greater
capability for S02 removal than the dilute systems. Based on pilot
testing of concentrated systems in Japan, S02 removals of up to 98 percent
1028
-------
have been reported. J With the dilute systems the range of S02 removal
to be expected is probably in the 80-90 percent range. In general, St^
removal with double alkali systems is more efficient than with lime/
limestone systems, since the active alkali is soluble.
Costs
Based on cursory design and cost analysis, with the assumption
that a single scrubber device (e.g., a single stage venturi) can be
used to remove particulates and 862 to the extent required to meet new
source performance standards, the capital and operating costs of double
alkali systems appear to be significantly less than those for a lime
or limestone system designed for the same requirements.^ '
Generally, the dilute active alkali systems tend to be higher in
capital costs than the concentrated systems since both equipment size
and process flows are required to be larger to accomplish the same
degree of desulfurization. Methods used for sulfate removal and for
reduction of scaling tendencies will affect both capital and operating
costs.
1029
-------
SCHEMES OF OPERATION
Five basic schemes of operation considered to be potentially
attractive, based on design considerations discussed previously, are
described below and illustrated with schematic diagrams. It should
be noted that many variations on these five basic schemes can be produced
by varying the many parameters associated with the systems such as:
1. concentration of active alkali;
2. water make-up addition points;
3. regeneration with lime, limestone or lime plus limestone;
4. sulfate removal method; and
5. method employed to reduce scaling tendency.
In addition, although the schematics shown are specific with respect
to the type of equipment employed, this is done only for the sake of
establishing the equipment function at a glance. Thus, the systems
presented are not limited to the specific equipment shown, namely:
1. scrubber type,
2. solids separation equipment,
3. settling and thickening equipment, and
4. reaction vessel type.
Limestone and Lime Regeneration, Dilute Active
Alkali with Sulfite Softening (Scheme 1)
This scheme of operation is typified by Figure 2. In this system
sulfur dioxide is absorbed by the active alkali, aqueous sodium sulfite,
in the scrubber to form sodium bisulfite. Limestone is used to regen-
erate sodium sulfite, by reacting with bisulfite and precipitating
calcium sulfite in the limestone reaction vessel. Lime is used to
precipitate sulfate as gypsum with regeneration of active hydroxide ion
in the lime reaction vessel. The hydroxide is recirculated to the
limestone reaction tank where it reacts with bisulfite to regenerate
sulfite as does the limestone.
The product from the limestone reaction vessel is split into a
clarified liquor stream for recirculation to the scrubber loop and a
slurry stream to feed the lime reaction vessel for sulfate removal.
1030
-------
CLEANED
FLUE
GAS
EFFLUENT
HOLD TANK
WASH
WATER ROTARY
VACUUM
FILTER
REGENERATED RECYCLE
LIMESTONE
SLURRY
MAKE-UP
TANK
LIMESTONE
REACTION
TANK
Figure 2. Limestone and lime regeneration, dilute, sulfite softening (Scheme 1).
-------
Supernatant liquor (containing hydroxide ion) from the lime reaction
vessel is recirculated to the limestone reaction vessel while the
remaining slurry is processed to remove waste product solids from the
system.
Sulfate is removed from the system as gypsum, by precipitation
with lime; but in order to accomplish this, the system must be operated
with dilute active alkali in order to maintain the needed high
[S04=]/[SOj=] ratio. Dilute active alkali is somewhat of a disadvantage
with double alkali systems in that higher liquor flow rates are needed
and scrubbing efficiency is generally lower than with concentrated
active alkali systems.
The scaling problem is minimized through the use of a sulfite
softening technique. This is accomplished by recirculating (to the
scrubber loop) only supernatant liquor which is unsaturated with respect
to CaS04 and CafOH)? from the limestone reaction system. Sulfite
regenerated in the limestone reaction system tends to limit the calcium
ion concentration to the low value established by the CaS03 solubility
product. Sulfite softening is also advantageous when compared to car-
bonate softening in that a carbonate supplying compound (e.g., C02 or
Na2CO_) need not be added to the system at additional cost (except as
needed to make up for sodium losses).
Fresh water addition to the system is indicated only for:
1. lime slurry make-up, and
2. waste product washing.
Note that the limestone slurry is prepared with recycled water, and the
gas-is saturated in the venturi with recycled water. It has been
reported(5) that lime cannot be slaked successfully with recycled water
containing high concentrations of sulfate ion.
Operating cost for this system may be lower than for some of the
others to be discussed in that limestone, a very inexpensive calcium
compound, is used as the major source of calcium. On the other hand,
capital costs for this type of system may be higher than for some of
the others since separate systems must be installed for handling lime
and limestone. Limestone handling generally includes crushing and
grinding equipment in addition. Also, regeneration with limestone, as
opposed to that with lime, requires a much greater residence time to
attain good utilization, usually on the order of 1-2 hours.(3)
It is anticipated that most of the water addition to this system
will be done via a waste product wash step, thus reducing solubles in
the solid waste product and conserving expensive sodium in the system.
1032
-------
It is interesting to note that the use of both lime and limestone
for regeneration in this scheme is limited by the rate of sulfate
formation (sometimes referred to as oxidation rate) in the system. At
a 50 percent oxidation rate, no limestone would be consumed, since the
hydroxide produced by the sulfate removal step will consume all of the
bisulfite in the limestone reaction system. As a corollary, at oxida-
tion rates greater than 50 percent, the system could be prone to scaling
unless other provisions are made to reduce scaling tendencies. Obviously,
if the expected oxidation rate is 40 percent or more, scaling could be
a major problem and the advantages of using both limestone and lime are
greatly diminished.
Lime Regeneration, Dilute Active Alkali
with Sulfite Softening (Scheme 2)
This scheme of operation is illustrated in Figure 3. Basically,
this scheme is the same as scheme 1 except that only lime is used for
both regeneration of sulfite from bisulfite and for removing sulfate.
The sulfite softening is shown to take place in a reactor clarifier,
a piece of equipment designed to combine the functions of a reaction
vessel with those of a clarifier.
Comparison of this scheme with the previous one shows:
1. greater use of fresh water tor lime slaking; and
2. an overall simpler system, probably leading to
lower capital cost and possibly higher operating
costs due to the use of lime only as opposed to
lime and limestone.
Lime (Scheme 5) or Limestone (Scheme 4) Regeneration,
Concentrated Active Alkali with Sulfuric Acid Sulfate Removal
These schemes of operation are illustrated in Figure 4. The schematic
is taken from the flow sheets developed and pilot tested by two
Japanese companies, Showa Denko K. K. and Kureha Chemical Industry
Company.
Since there appears to be a good market for gypsum in Japan, the
processes developed by the two Japanese companies include an oxidation
step to covert all of the calcium sulfite to gypsum for sale. Limestone
has been selected as the source of calcium by both Japanese companies,
In the United States, the market for gypsum is relatively smaller
than in Japan. In addition, there is a tendency to avoid production
of a salable sulfur waste product due to the inherent problems associated
1033
-------
FLUE
VENTURI
SCRUBBER
CLEANED
FLUE
GAS
t-
DEM1STER
o
o
LU
0=
K
Ul
to
CO
K
o
EFFLUENT
HOLD TANK
REGENERATED RECYCLE
LIME
LIME
SLAKING
TANK
WATER
LIME
REACTION
TANK
WASH
WATER
ROTARY
VACUUM
FILTER
WASTE
PRODUCT
Figure 3. Lime regeneration, dilute, sulfite softening (Scheme 2).
-------
CLEANED
FLUE
GAS
LIME
OR
LIMESTONE
VENTURI
SCRUBBER
SLURRY
TANK
CENTRIFUGE
OR FILTER
WASTE PRODUCT
(CaS03)
A SULFURIC
CONVERSION
TANK
REACTION
TANK
WASTE
PRODUCT
(GYPSUM)
CENTRIFUGE
OR FILTER
EFFLUENT
HOLD TANK
Figure 4. Lime (Scheme 3) or limestone (Scheme 4) regeneration, concentrated,
sulfate removal.
-------
with marketing a by-product. Due to these considerations, Figure 4
does not show the oxidation step designed for gypsum production.
In these schemes sulfate is removed by precipitation as gypsum
from an acidic solution in which CaSO, is soluble. The overall sulfate
removal reaction is represented as follows:
Na2S04 + 2 CaSOj + H2S04 + 2 CaS04 4- + 2 NaHSOj
(water of hydration not shown)
The sulfuric acid used to acidify the reaction mixture must be
procured from an outside source and added to the system. Scaling
tendency is circumvented again in these schemes by sulfite softening.
(There is excess sulfite present in solution in the liquor recycled
to the scrubber circuit, thus maintaining a low calcium concentration.)
The fact that these systems can be operated with concentrated
active alkali tends to reduce the capital costs. Liquor flows can be
greatly reduced and most of the equipment can generally be smaller in
size than that required by the dilute active alkali systems. On the
other hand, anticipated operating costs for these systems would be
higher than for dilute systems due to the cost of sulfuric acid and
the additional cost of calcium required to remove the sulfur added by
the sulfuric acid. In the United States, where there does not appear
to be a large market for gypsum, these processes become prohibitively
costly if the oxidation rate is high. Based on data from the Kureha
pilot plant, however, only about 9 percent of the sulfur in the waste
product is due to sulfuric acid added to the system. This figure is
primarily affected by rate of oxidation in the system. It should be
noted that the factors affecting oxidation rate in any double alkali
system are not clearly defined and various systems have reported
oxidation rates varying over a large range.
Lime Regeneration, Dilute Active Alkali with Carbonate
Softening (Scheme 5)
This scheme is illustrated in Figure 5. Although this scheme is
similar to schemes 1 and 2, there is a subtle difference. Here clarified
liquor from the sulfate precipitation reaction is fed back to the scrubber
loop, but only after its calcium content is significantly reduced by
reaction with carbonate and after another clarification step. The
carbonate is generally supplied to the system in the form of soda ash
(Na2C03) or carbon dioxide.
The advantage of using CO? is that no excess sodium is added to the
system. A major disadvantage is the need to provide a complex system
for adding the C02> thus increasing capital cost and possibly leading to
complications in the operation of the system (chance of CaCOj scaling).
1036
-------
VENTURI
SCRUBBER
EFFLUENT
HOLD TANK
Figure 5. Lime regeneration, dilute, carbonate softening (Scheme 5).
-------
Using soda ash, on the other hand, requires much simpler equipment
and, therefore, would probably lead to lower capital cost. The potential
problem, however, is that the amount of sodium carbonate added for
softening purposes is in excess of that required as make-up for sodium
losses. As a result, this excess sodium would have to be purged from
the system, as sodium sulfate in order to establish steady state concen-
tration. Obviously, a high purge rate is unacceptable from the stand-
point of potential for water pollution by these emissions. A good
compromise may be to use both C02 and soda ash, thus achieving the
advantages of each.
The operating cost for this scheme would probably be slightly
higher than for schemes 1 and 2 due to the additional cost of the
carbonate-supplying reagent.
An advantage of this system over the other dilute systems discussed
is that the active alkali recycled to the scrubber loop would be
hydroxide rather than sulfite as is necessary in the systems employing
sulfite softening. Feeding hydroxide rather than sulfite to the scrubber
loop results in a higher scrubber liquor pH, and thus more efficient
scrubbing.
Comparison of Schemes
Table 1 presents a qualitative comparison as a summary of the
various double alkali schemes described. The ratings used are very
subjective rather than absolute since some of the systems described
are hypothetical. A comparison of actual systems would be greatly
dependent on the specific equipment used.
Capital costs are shown to be higher for dilute systems only
because it is assumed that the dilute systems will generally require
larger size equipment and greater flow rates. Operating costs are
shown to be higher or lower, based only on the need to supply additional
calcium or other reagents exclusive of normal sodium make-up requirements.
Oxidation limitation of the systems, although tabulated as a
specific oxidation rate, should only be taken to represent an attempt
at quantification of this factor based on the chemistry of the system;
this is commonly known as a "Guestimate."
Sulfur dioxide removal capability is relatively higher with the
concentrated active alkali systems since more SO can be removed per
given volume of absorbent liquor than with dilute active alkali systems.
In theory, however, both concentrated and dilute active alkali systems
should be able to attain about the same S02 removal given the right
conditions. In practice, cost becomes the controlling factor.
Additional reagents required is meant to include reagents in
addition to those required to make up for minimal unavoidable sodium
losses.
1038
-------
Table 1. COMPARISON OF ATTRACTIVE DOUBLE ALKALI SCHEMES
SCHEME
1
2
3
4
5
Ca"
CONTROL
(SOFTENING)
SULFITE
SULFITE
SULFITE
SULFITE
CARBONATE
ACTIVE
ALKAU
DILUTE
DILUTE
CONC.
CONC.
DILUTE
S04= REMOVAL
AGENT
UME
UME
•W
H.SO,
UME
CALCIUM
SOURCE
UME AND
UMESTONE
UME
UME
UMESTONE
UME
ADDITIONAL
REAGENTS
REQUIRED
NONE
NONE
H2S04
H2S°i,
Na2C03
AND/OR
CO,
ESTIMATED
OXIDATION
UMITATION
40%
50%
30%
30%
NONE
RELATIVE COST
CAPITAL
HIGHER
HIGHER
LOWER
LOWER
HIGHER
OPERATING
LOWER
LOWER
HIGHER
HIGHER
HIGHER
RELATIVE
SO2 REMOVAL
CAPABIUTY
LOWER
LOWER
HIGHER
HIGHER
MEDIUM
-------
STATUS OF TECHNOLOGY
A number of organizations in the United States and abroad have
studied and tested various double alkali schemes on systems ranging
in size from the laboratory bench scale to 30,000 CFM. A summary of
the status of development of these various systems will be presented
in order to familiarize potential users of these systems with what is
commercially available and what is planned for larger scale installation.
Every effort has been made to present a status report that is
accurate. Information presented in an overview report of this type is
necessarily "second hand," and obtained from a multiplicity of sources
including process and equipment vendors, engineering companies and
potential system users, all of whom did not always yield consistent
and/or specific information due in part to the complexities and changing
conditions associated with development programs.
FMC Corporation(6)
The requirement to control SO- emissions from large reduction kilns
operated by FMC in Modesto, California, led to the development of a
workable double alkali system. This full scale system has been in
operation since December 1971. The double alkali system processes a
combined 30,000 ACFM off-gas stream containing 5000-8000 ppm S02 and
about 5-8 percent oxygen from two kilns. The scrubber at the Modesto
plant consists of a vertical column packed with 9 feet of Intalox
saddles. A wire mesh entrainment separator is used in series with the
packed column. The absorbent liquor contains a high concentration of
active alkali (Na2S03/NaHS03) and sodium sulfate. Sulfate formed in
the system as a result of oxidation is purged from the system, dissolved
in the liquor adhering to the CaS03 waste product. A rotary vacuum
filter effects the final solids separation and produces a waste product
containing about 50 percent moisture, CaS03, dissolved Na2S04/Na2S03 and
kiln ash which is disposed of in an on-site pond.
The Modesto system is reported to operate continuously without
scaling problems. The system is capable of attaining S02 removal in
excess of 95 percent with simultaneous removal of kiln ash.
Figure 6 is a schematic representation of the FMC system. Comparing
the FMC scheme to the five schemes previously presented, it is most
closely related to scheme 3. The major difference is that the FMC scheme
makes no provision for removal of sulfate as a solid innocuous waste
product. Sulfate is removed from the system as a dissolved solid in the
liquor entrained with the CaS03 waste product.
Although the FMC scheme makes no provisions for removal of sulfate
as an insoluble solid, FMC reportsC6) that tests indicate a reduction in
1040
-------
o
FILTER PLANT
VACUUM IATER
PUMP
LIME
STORAGE
BIN
REHEATER
CONDENSATE
FILTRATE
RECEIVER
SODA ASH
STORAGE
BIN
ROTARY FILTER
PLANT
WATER
ENTRAPMENT
SEPARATOR
LIME
REACTOR
SOLIDS TO
LANDFILL
THICKENER
UNDERFLOW
RECIRCULATION
TANK
OVERFLOW
PUMP
RECIRCULATION
PUMP
REGENERATION
CIRCUIT PUMP
_ -FILTRATE
0 1 RETURN
PUMP
Figure 6. FMC schematic.
-------
sulfate formation when the system is operated with high ionic strength
absorbent liquor. As a result, provisions have been made for washing
the solid waste product to the extent consistent with maintaining a
steady state sulfate level.
In an effort to develop their system for application to industrial
and utility boilers, FMC has continued testing their system at the
pilot plant level.
A 2,000 CFM pilot plant was operated at an FMC chemical plant in
South Charleston, W. Va., for a period in excess of 6 months starting
in June 1972. The flue gas supply was taken from an underfed stoker
type steam boiler rated at 80,000 Ib of steam per hour and burning
3.5 percent sulfur coal. The oxygen content of this flue gas was varied
up to 13 percent in order to obtain conditions conducive to oxidation
and to observe effects on the system.
FMC has also recently operated a 3,500 CFM pilot plant housed in a
40-foot trailer at the Mossville engine plant of the Caterpillar Tractor
Company. The flue gas supply for this testing was taken from a steam
boiler at the plant which burns typical high sulfur Illinois coal. This
plant was equipped for filter cake washing to reduce sodium losses from
the system. The waste product from this system contained about 40 percent
moisture under optimum conditions and appeared to be relatively dry and
easy to handle.
After initial testing in March 1972, in a 5,000 CFM unit at the
Modesto Plant, all of the FMC pilot plant programs have employed the
"FMC-Link Belt Dual Throat Variable Scrubber." This is a single stage
venturi type scrubber coupled with a cyclonic type separator.
Typical operating parameters for the FMC system are:
S02 removal efficiency 70-95%
Approximately alkali composition 15-20% Na2S04
5-7% Na2S03/
NaHS03
Ca:S stoichiometry 1.0
Scrubber pH range 6-7
Na.CO^ requirements 8 moles Na2COj
moles Ca(OH)
Ca++ in scrubber liquor < 5 ppm
1042
-------
Capital and operating costs for the FMC system are estimated by
FMC to be:
Generating Capacity, MW
8 20 200
Capital costs, $/KW 184 100 30
Operating costs, $/ton of coal 11 8 5
FMC is presently negotiating with a large industrial manufacturer
in Illinois for a contract to install the FMC process to control SOX
and particulates from a system of steam boilers approximately equivalent
in steam generating capacity to a 40 MW power plant. It is presently
anticipated that this system will be in operation by June 1974.
Envirotech Corporation(5)
Envirotech has operated a 3,000 CFM double alkali pilot plant using
various equipment arrangements for approximately a year, since early
1972. The pilot plant is located at the .Gadsby station of Utah Power
and Light.
After testing various equipment arrangements, and fighting the
scaling problem with some, Envirotech developed the scheme shown in
Figure 7. This scheme might be the typical scheme which would be
proposed by Envirotech for control of SOX and particulates from a
utility or industrial boiler.
Examination of this flow sheet indicates that it is related to
scheme 3 or 5, depending on the concentration of active alkali in the
liquor. The concentration range of sulfite, bisulfite and sulfate in
the liquor is considered by Envirotech to be proprietary information
and, therefore, has not been disclosed. Operated with concentrated
active alkali, this system is similar to scheme 3 without provisions
for removal of sulfate (with sulfuric acid) as an insoluble solid.
Operated with dilute active alkali, the system is similar to scheme 5,
using soda ash for carbonate softening. In this case there is the
necessity to purge from the system a certain amount of sodium dissolved
in the liquor adhering to the waste product filter cake.
In the pilot plant, a perforated plate and an expanded metal tray
were tested in the tray scrubber. The preferred arrangement was a
2-stage expanded metal tray scrubber. The pilot plant is reported to
have operated in a scale-free manner for a period of 6 months with
1043
-------
RAN H20
i CLEAN GAS
* rrsf
KENER } k
CLARIFIER LIQUOR
STORAGE
COOLING
TONER
BLON-
DONN
PROCESS
NEEDS
NASTE
Figure 7. Envirotech schematic.
-------
provisions for soda ash softening, inlet gas quenching with fresh water,
and recycle_of precipitated solids to the reaction tanks and reactor
clarifiers." Quenching with fresh water, as pointed out previously,
could lead to water balance problems and thus require purge. Use of
cooling tower blowdown and sending demister wash water to plant use
are examples of good water conservation attempts.
Envirotech lists the problems they consider significant in their
operation of the system as follows (not necessarily in order):
1. scaling;
2. gas distribution at low pressure drop;
3. over-liming (relating to pH control);
4. mechanical problems - valves, pumps, piping;
S. flow regulation; and
6. inlet wet/dry interface accumulation.
A list of important process variables and their typical values from
various pilot tests is given below:
S02 in inlet flue gas 350-400 ppm
Scrubber pressure drop 4-9 in. H20
Calcium content of scrubber liquor 400 ppm
pH of scrubber liquor at exit 7
Scrubber L/G ratio 12 to 36 gal/
1,000 ft3
S0^=/S03= concentration in scrubber liquor not disclosed
S02 removal efficiency 90%
Additional pilot plant testing, 24 hours/day, 5 days/week, for
2-3 months will be geared to investigate:
1. simultaneous removal of SCL and particulates; and
2. system operation at inlet S02 levels up to 1500 ppm.
Envirotech will attempt to demonstrate their system next at the
100-200 MW level.
1045
-------
General MotorsC7»8)
The double alkali system developed by GM is»schematically illus-
trated in Figure 8. After considering other possible alternatives, GM
decided that a double alkali system was most practical for their S02
control requirements in that it fit the criteria of being relatively
economical and uncomplex.
Pilot plant testing was conducted at the Cleveland Chevrolet plant
using a 2,800 CFM cross-flow packed scrubber supplied and operated by
Ceilcote for the joint development effort. Flue gas for testing was
an isokinetic sample representing 10 percent of the total flue gas flow
from a boiler having a steaming capacity of 80,000 Ib/hour and burning
2 percent sulfur coal with 100 percent excess air.
In the pilot plant, it was found that a 1 molar sodium solution
would give reasonably good SO? absorption, while also regenerating
caustic and precipitating sulfate as gypsum. It was found that a maximum
concentration of 0.1 M hydroxide could be regenerated in a 1 M sodium
solution. Increasing the sodium ion concentration above 1 M did not
give appreciable increase in regenerated hydroxide concentration. Lowering
the sodium ion concentration, however, gave a decrease in regenerated
hydroxide concentration.
Optimum lime utilization was found to occur with high speed mixing,
using near stoichiometric quantities of lime. Eighty percent conversion
of lime was attained in 5 minutes.
Early in the pilot plant program, GM ran into the calcium plugging
problem; but later, using soda ash softening, they were able to alleviate
it. Reportedly, using soda ash for sodium make-up, they were able to
reduce the calcium content in the scrubber loop from 400 to 250 ppm.
A high degree of confidence in the double alkali system has prompted
GM to construct: a full-scale system at the Cleveland Chevrolet plant.
The plant presently has four stoker-fired boilers, equivalent in steaming
capacity to a 32 MW electric generating plant. The double alkali system,
however, is designed to handle the flue gas equivalent of a 40 MW plant
in order to accommodate possible future expansion. The system is being
built at an approximate cost of $3,000,000 and is scheduled to start up
in December 1973.
The system in the full scale plant will consist of four parallel
Koch tray (valve tray) scrubbers. Provisions for softening with soda
ash and/or carbon dioxide will be included. Sodium make-up will be as
sodium hydroxide and/or soda ash. The active alkali will be dilute to
ensure removal of sulfate as insoluble gypsum by lime treatment. The
anticipated sodium ion concentration will be in the range of 1-2 molar.
10/16
-------
SULFUR-FREE
FLUE GAS
o
4?
-j
(flETGAS
))SCRUBBER
Na2S03 SOLUTION
REACTOR
CLARIFIERS
HIGH-SULFUR
BOILER
FLUE GAS
FILTER |
FILTER CAKE
(TO LANDFILL)
SLURRY
CaCOs FOR _
NEUTRALIZATION-*-
Figure 8. General Motors schematic.
-------
Depending upon control of the operation, this system should be capable
of producing an environmentally acceptable solid waste product, at least
with respect to contamination by dissolved sodium salts.
This design is comparable to scheme 5, previously discussed.
Zurn Air SystemsC9»10)
Zurn contracted with Southern Research Institute to conduct a
laboratory study of the regeneration process for a lime/sodium, dilute
double alkali system. The study, completed in 1972, basically confirmed
the information reported by General Motors. This study was conducted in
order to obtain a firmer basis for the design of a marketable double
alkali system. Figure 9 is a schematic representation of the type system
Zurn is prepared to market. This flow diagram is comparable to scheme 5,
previously discussed, with soda ash softening.
The Zurn system would employ a "Dustraxtor" scrubber, briefly
described as a multitube entrainment contactor capable of achieving high
internal liquid/gas ratios.
It should be pointed out that the system would be capable of
removing sulfate as insoluble gypsum by precipitation with lime; however,
due to the lack of provisions for softening with carbon dioxide, a
certain amount of sodium and sulfate would necessarily have to be purged
from the system in the moisture associated with the waste product filter
cake. Provisions for filter cake washing are shown. The system is
anticipated to operate at a sodium concentration Of 1-2 molar.
A Zurn operating cost estimate, exclusive of amortization of equip-
ment, was given as $4.86 per ton of coal burned. Zurn is presently
negotiating with a large industrial manufacturer for a contract to
install the equivalent of a 20 MW system on a coal-fired industrial
boiler system.
ChemicoC11)
Chemico tested a double alkali system in a 1,500 CFM pilot plant
located at the Mitchell Power Station of Allegheny Power Service Company
Testing lasted approximately 6 weeks, and Chemico was able to operate
the pilot unit in a scale-free manner, using either lime or limestone
as the calcium source. Figure 10 is a schematic representation of that
pilot plant, but not necessarily of the system which might be marketed
by Chemico.
In the pilot plant, there were no waste product cake washing
provisions, so dissolved solids were purged from the system in the
moisture associated with the cake. In addition, since make-up sodium
was added to the system as a mixture of sodium sulfite and bisulfite
1048
-------
o
vo
CLEAN GAS
AND H20
S02, S03, AND
PARTICULATES
I
MECHANICAL
DUSTRAXTOR
SCRUBBER
DUST
COLLECTO
DILUTE
NaOH
NaHS03,
Na2S03,
Na2S04, AND
PARTICULATES
WASTE
PRODUCT
VACUUM
FILTER
LIME
MIX
TANK I
Na2C03
MIX TANK
SLUDGE
STORAG
^VACUUM
TO \/ # PUMP
LANDFILLT SEAL H20
-------
Na SALTS
VENTURI
H20
PRODUCT
LIQUOR
TANK
FUE. •SE
Figure 10. CHEMICO Pilot Rant schematic.
-------
on a regular schedule, it necessarily follows that there must have been
a significant purge from the system. Testing in this fashion was
attributed by Chemico to be due to the fact that testing at such a
small scale with separate pieces of equipment which were not sized to be
consistent with each other tended to necessitate a somewhat open system.
Typically, in small unit testing, leakage from the system is generally
enough to simulate open system operation.
Based on testing in the pilot plant, however, Chemico developed
other flow sheets for application to utility and other boiler systems
which are apparently similar to schemes 1 and 2, previously discussed,
provided that the systems are intended for operation with dilute active
alkali. If these systems are operated with concentrated alkali, a purge
of sulfate from the system in some form would be necessary.
Arthur D. Little/Combustion Equipment Associates^12-*
ADL and CEA are involved, per a joint venture agreement, in developing
and marketing double alkali systems. Generally speaking, CEA is an equip-
ment and systems designer and manufacturer, and ADL participates in a
developmental and engineering capacity, in the joint venture agreement.
A 2,000 CFM pilot plant has been in operation since January 1973,
testing various double alkali configurations at the ADL facility in
Cambridge, Massachusetts. The pilot plant unit was built by CEA for
testing in connection with any of the ADL/CEA joint ventures. This pilot
unit is supplied with flue gas generated by combustion of natural gas.
There are provisions for addition of S02, flyash and excess air. The
scrubber system consists of a'variable throat venturi, tray separator
and radial vane demister. The solids handling facilities include a
circular clarifier or settling tank with revolving bottom rake arm
mechanism and a rotary vacuum filter with provisions for filter cake
washing and drying. The plant is equipped with seven miscellaneous
tanks, four of which are provided with mechanical agitation. These tanks
serve as scrubber recycle tanks and reaction vessels as necessary.
ADL/CEA have contracted with The Southern Company (a utility combine
producing electricity in the southeastern states, consisting of Alabama
Power Co., Georgia Power Co., Gulf Power Co., Mississippi Power Co.,
Southern Electric Generating Co. and Southern Services, Inc.) for
installation of a 20 MW S02 and particulate scrubber prototype system.C13)
Flue gas supply for the prototype will be half the flue gas produced from
the Unit No. 1 boiler of the Scholz plant operated by Gulf Power Company.
The scrubber systems to be tested will be various combinations of a
venturi, spray tower, tray tower and packed tower. The system will be
designed to test double alkali operation with lime only or lime and
limestone.(14) in addition, the system is designed to test lime and
limestone slurry scrubbing. The exact plans for operation of this system
1051
-------
with respect to alkali concentration are not known; however, in order
to produce an environmentally acceptable waste product, the system
would probably have to operate, as previously described, by schemes 1,
2, or 5, since no provisions for sulfuric acid addition were revealed,
as would be required for operation as depicted by schemes 3 and 4.
It is anticipated that construction of this prototype facility
will be initiated in late 1973, and that the system will be in operation
in June 1974, or earlier.
Kureha Chemical Industry Co. (Japan)
Kureha has recently developed a sodium/calcium double alkali system
which operates in a manner similar to that described by scheme 4,
previously discussed. After testing in a small pilot plant, a larger
pilot plant (3,000 SCFM) was built in a joint development effort with
Kawasaki Heavy Industries and has been in operation since July 1972.
Two commercial plants using this scheme of operation will be completed
in 1974 to treat flue gas from oil-fired boilers at two different
electric power plants each having a 150 MW generating capacity.
The proposed Kureha system consists of a venturi for particulate
removal, followed by a grid-packed scrubber for S02 removal. The SC>2
absorber operates in the pH range 6.0 - 6.5. With an inlet concentration
of 1,500 ppm S02, 98 percent removal is achieved. The scrubber liquid/
gas ratio is 7 gal/1,000 ft3 of gas.
The limestone reactor operates at a temperature somewhat higher than
the scrubber temperature and is designed to provide a 2-hour residence
time. Centrifuges are used for the solids separation steps.
Composition of the absorber feed and discharge liquor is reported
below:
Absorber feed -
Sodium sulfite 20-25%
Calcium 30 ppm
pH 7-8
Absorber discharge -
4
Sodium sulfite 10%
Sodium bisulfite 10%
Sodium sulfate 2-5%
1052
-------
Since there is a good market for gypsum in Japan, an oxidation
step is added to the process described previously by scheme 4. Calcium
sulfite is reacted with air at atmospheric pressure in an oxidizer
developed by Kureha. The product gypsum, removed by centrifuge from
the oxidizer liquor, is suitable for use in wallboard and cement.
Operating costs associated with this process include additional
costs for sulfuric acid and the incremental cost for additional
limestone required for precipitation of the sulfur added in the form
of sulfuric acid. The cost penalty is aggravated if gypsum cannot be
sold. As pointed out previously, the sulfuric acid requirement for
decomposition of the sulfate is 125 percent of the theoretical amount;
thus, about 9 percent of the product gypsum is derived from sulfuric
acid.
Capital costs for this process can be split as follows:
Absorption system 30%
Limestone reaction system 30%
Sulfate removal 10%
Oxidation 30%
100%
The system is reported to operate reliably with no scaling problem,
while removing sulfate in an environmentally acceptable manner.
Showa Denko K. K. CJapan)
The Showa Denko process is very similar to the process described
for Kureha and again comparable to scheme 4 with the additional provision
for oxidation of all the CaS03 to gypsum (for sale) in a separate oxida-
tion step. This company uses a vertical cone type absorber which
operates with liquid/gas ratios ranging from 7-14 gal/1,000 ft3 and with
a pressure drop range from 8-15 inches of water.
Showa Denko has operated a 5,900 SCFM pilot plant using this process
since 1971 to desulfurize flue gas from an oil-fired boiler. A commercial
plant is currently under construction and expected to be in operation in
June 1973. This plant (Chiba plant of Showa Denko) will treat flue gas
equivalent to that emitted by a 34 MW electric generating plant.
Summary of Commercial Systems
A comparison of the double alkali systems developed by the various
organizations discussed in this paper is presented in Table 2. The
systems described by the table generally represent the systems developed,
1053
-------
Table 2. COMPARISON OF COMMERCIAL DOUBLE ALKALI SYSTEMS
COMMERCIAL
SYSTEM
FMC
ENVIROTECH
GENERAL
MOTORS
ZURN AIR
SYSTEMS
CHEMICO
A.D. LITTLE/
COMBUSTION
EQUIPMENT
ASSOCIATES
SHOWA DENKO
K.K.
KUREHA
ACTIVE
ALKALI
CONC.
UNKNOWN
DILUTE
DILUTE
UNKNOWN
DILUTE
CONC.
CONC.
SULFATE
REMOVAL
PURGE
WITH WASTE
PURGE
WITH WASTE
UME
UME
TREATMENT.
SOME PURGE
LIME OR
PURGE
UME OR
OTHER
„,*>,
H2SO,
CALCIUM
CONTROL
*>*'
<»3=
co3=
cof
S03=
S03= OR
co3=
—
»,'
CALCIUM
SOURCE
L
L
L
L
L AND/OR
LS
L AND/OR
LS
LS
LS
ADDITIONAL
REAGENTS
H.,00,
Na2C03
COj AND/OR
*.,«,,
DEPENDENT ON
OPERATION
DEPENDENT ON
OPERATION
H2S04
H2S04
LARGEST UNIT
TESTED, cfm
30,000
3,000
2,800
LAB SCALE
.
1,500
2,000
3,000
5,900
PLANNED
INSTALLATIONS
SIZE
40 MW
-
40 MW
20 MW
-
20 MW
34 MW
150 MW
150 MW
DATE
7/74
-
12/73
'
-
6/74
6/73
1974
1974
o
en
-------
tested and/or available for marketing by each of the listed companies.
In some cases several systems are implied. The reason for this is that
most of the companies have tested several double alkali schemes during
their development efforts. In some cases, the schemes tested are not
necessarily the ones which would be marketed by the company. Information
is presented in the table to allow the reader to identify each of the
commercially developed systems with the attractive schemes of operation
previously discussed in detail.
1055
-------
THE EPA DOUBLE ALKALI DEVELOPMENT PROGRAM
In an effort to more fully test and characterize double alkali
systems, EPA has contracted with Arthur D. Little, Inc., (contract
value approximately $500,000) to conduct a two-task study. The
proposed study is entitled, "A Laboratory and Pilot Plant Study of
the Dual Alkali Process for S02 Control." Technical administration
for this study will be split between the Research Laboratory and the
Development Engineering Branches of the Control Systems Laboratory -for
administration, respectively, of the laboratory and pilot plant tasks.
The test program and study are anticipated to last for 15 months. One
of the main objectives of the program will be to develop several methods
of producing environmentally acceptable waste while operating in a
reliable trouble-free manner.
An outline of the program by task is given below:
Task 1 - Laboratory Study
1. A comprehensive literature survey will be performed stressing
kinetics and thermodynamics of regeneration chemistry, S02
absorption, and solids separation operations related to
calcium/sulfur compounds.
2. A laboratory study of the regeneration chemistry will be
conducted, using lime and/or limestone.
3. A complete laboratory bench scale simulation of the absorption
and regeneration systems will be constructed and tested in
various attractive configurations to investigate S02 removal,
scaling problems, residence times, optimum solution concentra-
tions and solids settling and filtration rates.
Task 2 - Pilot Plant Study
1. Attractive schemes of operation will be tested at the
ADL/CEA 2,000 CFM pilot plant in Cambridge.
2. A 9,000 CFM pilot plant at the Reid-Gardner station of Nevada
Power Company will be modified to accommodate testing of
attractive schemes of operation. (This pilot plant was
formerly used to test once-through sodium carbonate scrubbing
to firm up the design of a 250 MW desulfurization system now
under construction at the Reid-Gardner plant by CEA and
anticipated to start up in mid-1973.) The scrubber system
in this pilot plant is the same as in the 2,000 CFM system
located at the ADL facility.
1056
-------
3. Attractive schemes will be tested at the 9,000 CFM pilot
level.
4. Data from the pilot plant operations will be.assembled
to provide input for an economic study to estimate capital
and operating costs for various full scale systems.
5. Particular emphasis will be given to the important considera-
tions previously discussed:
• Sulfate removal methods
• Scale prevention
• Water balance
• Waste product washing
• SC>2 removal
• Costs
Figure 11 is a simplified schematic of the "Double Alkali" process
which is used by ADL to illustrate the system in a very general way.
It should be noted that sulfate treatment is shown in a general manner,
illustrating the possible application'of any of a number of techniques
to be tested to cope with the problem of removing sulfate in an
environmentally acceptable way. There is additional significance to
the general philosophy of treating for sulfate removal in a slip stream
off the main liquor regeneration loop: methods developed to handle
this problem would be applicable to other double alkali systems.
Conduct of the Test Program and Objectives
Testing will be conducted at three successive levels:
1. laboratory bench scale,
2. 2,000 CFM pilot plant, and
3. 9,000 CFM pilot plant.
The successive levels of testing will be used to screen and select
the most attractive schemes from the many initially showing potential.
Testing in this manner should allow testing of many schemes at the bench
scale level, fewer at the 2,000 CFM pilot level and still fewer at the
9,000 CFM level where long-term, reliable, closed-loop operation will be
stressed. Equipment at each level of testing will be designed with
maximum flexibility consistent with budgetary constraints of the program.
1057
-------
I
FLUE
GAS
BY-PASS
- FLUE
S GAS '
cS FEED
(*OHk
SCRUBBED
GAS
SCRUBBER
JL
n
MIXING
TANK
SCRUBBER
FEED
MAKE-UP
SCRUBBER
EFFLUENT
U
CAUSTIC1ZER
H
H
THICKENER
NASH
HATER
SULFATE
TREATMENT
(GENERAL)
WASTE
CALCIUM
SALTS
HOLDING
TANK
Figure 11. Arthur D. Little general schematic.
-------
Pilot plant capabilities will include provisions for:
1. solid waste product washing,
2. control of SC>2 content of the inlet flue gas,
3, flyash addition,
4. sodium make-up as sodium sulfate,
5. recirculation of precipitated solids,
6. flue gas quenching, and
7. testing of at least five attractive schemes.
Following is a list of important program objectives:
1. demonstrate reliable system operation,
2. demonstrate high S02 removal, 95 percent desirable,
3. demonstrate high particulate removal,
4. demonstrate environmentally acceptable sulfate
removal schemes,
5. minimize disposal of soluble material,
6. demonstrate closed-loop operation (or close approach
to this ideal), and
7. obtain complete material balances.
1059
-------
BIBLIOGRAPHY
1. F. T. Princiotta and N. Kaplan, "Control of Sulfur Oxide Pollution
from Power Plants." Presented at LASCON '72, Washington, D. C.,
October 18, 1972.
2. J. W. Jones, R. D. Stern, F. T. Princiotta, "Waste Products from
Throwaway Flue Gas Cleaning Processes - Ideologically Sound
Treatment and Disposal." Presented at F.PA Flue Gas Desulfuri zation
Symposium, New Orleans, La., May 14-17, 1973.
3. J, Ando, "Recent Developments in Desulfurization of Fuel Oil and
Waste Gas in Japan (1973)." !:PA report EPA-R2-73-229, May 1973.
4. G. T. Rochelle, "Economics of Flue Has Desulfurization." Presented
at EPA Flue Gas Desulfurization Symposium, New Orleans, La.,
May 14-17, 1973.
5. D. Dahlstrom and others, L;imco Division, Fmvirotech Corp.,
personal communication, 1973.
6. J. Brady and others, FMC Corp., personal communication,
November 1972-May 1973.
7. R. J. Phillips, General Motors, personal communication, 1973.
8. R. J. Phillips, "Sulfur Dioxide Emission Control for Industrial
Power Plants." Presented at Second International Lime/Limestone
Wet Scrubbing Symposium, New Orleans, La., November 8-12, 1971.
9. G. G. Langlcy, Zurn Industries, personal communication, 1972-1973.
10. "Sulfur Oxide and Particulate Control System." Zurn Industries,
July 1972.
11. I. S. Shah and S. Sawyer, Chemico, personal communication, 1972.
12. C. R. LaMantia, ADL, personal communication, 1972-1973.
13. J. M. Craig, The Southern Company, personal communication, 1973.
14. I. S. Shah, CEA, personal communication, 1973.
1060
-------
STONE & WEBSTER/IONICS
SO2 REMOVAL AND RECOVERY PROCESS
by
N. L. Foskett and E. G. Lowrance
Stone & Webster Engineering Corporation
Boston, Massachusetts
Wayne A. McRae
Ionics, Incorporated
Watertown , Massachusetts
1061
-------
1. IMTRODUCTION
The Stone & Webster /Ionics S02 Removal and Recovery Process is based on the
absorption of sulfur dioxide in a caustic solution which then is regenerated
using Ionics' electrolytic cells (electrolyzers), the key component of the
process. This patented process -is applicable to gaseous effluents from
stationary power plants burning fuels containing sulfur and to tail gases
from sulfur recovery plants, smelters and sulfuric acid plants.
The process consits of three essential steps arranged in a closed loop:
1. Sulfur dioxide (SOg) is absorbed from flue gas in a solution
of caustic soda (NaOH) in water to produce a solution of
sodium bisulfite (NaHS03) containing some sodium sulfite
2. The sulfite-bisulfite solution is mixed with sulfuric acid
(H2SOl|), causing release of 802 ft*011 the solution. The SOg
is recovered for sale as such, or converted to sulfur or
sulfuric acid. The solution is converted to sodium sulfate
(Na2SOif ) .
3. The sulfate solution is converted by an electrolyzer into two
solutions, one containing caustic, and the other sulfuric
acid. The caustic solution is recycled for use in step 1;
the sulfuric acid solution is recycled for use in step 2.
Under a joint program co-sponsored by the Environmental Protection Agency
and Wisconsin Electric Power Company, the process will be tested at
Wisconsin Electric 's Valley Station in Milwaukee. Phase I of the potential
three-phase demonstration program is currently underway and consists of
design, installation and operation of an integrated pilot plant, development
of prototype scale electrolyzer system, preliminary design of a prototype
system and development of test programs and operating schedules. Phases II
and -III of the program would involve the design, procurement and installation
of a prototype facility treating the flue gas from one of the four 75 MW
coal fired boilers at the station, and its start-up and long-term operation,
respectively.
2. HISTORY OF THE PROCESS
Several years ago, Stone & Webster, because of its involvement in the design
and construction of power plants, undertook to examine the various sugges-
tions which had been advanced for 802 removal from flue gas and came to the
early conclusion that scrubbing with an alkali solution was a promising and
probably enconomic approach, provided that appropriate methods could be
developed for regenerating and reusing the solution without excessive cost
and without the production of additional pollutants. The advantages of
electrolytic regeneration as proposed by Ionics, Incorporated of Watertown,
Massachusetts, appeared particularly interesting.
1062
-------
Joint activity by Stone & Webster and Ionics commenced in 1966, and has
been directed toward the engineering development of an economic system for
an electric generating plant to remove S02 from flue gas by absorption
in a caustic solution, the caustic being electrically regenerated and
reused in a closed circuit. A major commercial installation of an
electrolyzer system employing technology similar to that practiced in
the Stone & Webster/Ionics S02 Recovery System was installed in a plant
of Monsanto Textile Corporation in Decatur, Alabama, in 1967, for the
reduction of acrylonitrile to adiponitrile, an intermediate in nylon
manufacture.
An SOg removal pilot plant was installed at the Gannon Station of Tampa
Electric Company in Florida in 1967 and was operated for approximately
6 months. It treated approximately 200 cfm of flue gas from a boiler
operating on pulverized coal. The flue gas contained about 0.25 vol
percent sulfur dioxide. The pilot equipment included a flue gas quench
which cooled the gas to about 120° F and removed most of the fly ash.
This was followed by an absorber for removal of the sulfur dioxide, and
an electrolyzer system for regeneration of the caustic. The pilot program
confirmed the soundness and operability of the process, and pointed the
way toward further development of both the electrolyzer and the flue gas
absorber.
As result of the pilot work, electrolyzer materials of construction were
revised by Ionics to improve reliability and to reduce costs. In addition
some design modifications were introduced to reduce electrical energy
requirements. Subsequently, the electrolyzer was scaled to its full
commercial height, and a six month test period of operation was carried
out in a second electrolyte pilot plant at Ionics' Watertown Laboratory
using this "tall cell model." A third pilot plant has been built and
successfully operated in Japan. Its purpose is to demonstrate the
regeneration of caustic for removal of S02 from the flue gas of oil fired
boiler under Japanese conditions.
Following the Tampa pilot work and recognizing that considerable effort
had to be expended on the design of absorbers for large steam plant
installations, the National Air Pollution Control Administration (NAPCA)
engaged Stone & Webster to make a study of existing absorber concepts.
The study was based on removing sulfur dioxide from flue gas produced in
a coal-fired power station rated at 1,200 MW and consisting of four boilers,
with special emphasis on minimizing the oxidation of sulfites to sulfate in
the absorption system. It was concluded that a packed tower was best suited
to the Stone & Webster/Ionics Process.
Having carried through the laboratory experiments at Ionics, scale-up
engineering and seni-commercial studies for the various system components
within the process, the next logical step would be to undertake a demon-
stration program of the entire process using each significant component in
1063
-------
its presently developed state. However, considering the small size
(200 CFM) of the Tampa Pilot Plant and the large thruput of even a modest
sized prototype plant, say 75 MW, which would have stack gas volume of
about 225,000 CFM, it was decided to install an intermediate sized pilot
plant of 2,000 or 3,000 ACFM. This should provide data less subject to
gross error than a scale up from 200 to 225,000 CFM.
Such a program is now under way at Wisconsin Electric's Valley Station in
Milwaukee.
3. PROCESS DESCRIPTION^1*
3.1 General
The process involves three basic steps. These are:
1. Absorption
2. Sulfur Dioxide Recovery
3. Absorbent Regeneration and Oxidation Product Rejection
The flow scheme is shown on Pigs. 1-1 and 1-2.
3.1.1 Absorption
The flue gas is first cooled by direct water quench in the bottom of the
absorber tower, after which it is contacted with a dilute (&%) caustic
solution containing sodium sulfate. At the top of the tower, the flue
gas, with most of the S02 removed, is reheated if necessary and returned
to the stack. The absorbing solution is converted to a sodiun bisulfite-
sodium sulfite-sodium sulfate mixture which contains the 802 and 803
removed from the flue gas.
3.1.2 Sulfur Dioxide Recovery
The bisulfite-sulfite mixture is reacted with a dilute mixture of sulfuric
acid and sodium sulfate recycled front the electrolytic cells (electrolyzers).
This reaction forms sodium sulfate, S02 and water, and is carried out prior
to entering a stripping tower. In the latter, SOg is removed as the over-
head, leaving as tower bottoms a solution of sodium sulfate.
3.1.3 Absorbent Regeneration and Oxidation Product Rejection
The sodiun sulfate solution is sent to two types of electrolyzers which are
described in detail later in this section. In both electrolyzers, sodium
hydroxide (caustic) containing sodium sulfate is generated. This caustic
is recycled to the absorption tower for reuse as the absorbing medium.
Also in both electrolyzers a mixture of sodium sulfate and sulfuric acid is
generated and recycled as the dilute acidic solution required in the Sulfur
Dioxide Recovery step mentioned above.
(1) Description is general but relates more specifically to
prototype plant rather than pilot plant.
1064
-------
STACK
STRIPPER
STRIPPER
REFLUX
CONDENSER
FLY ASM SLOWDOWN TO ASH TRANSFER SYSTEM
QUENCH
CIRCULATION
PUMP
ABSORBER CIRCULATION
PUMPS
STRIPPER
REFLUX PUMP
FIGURE 1-1
S02 REMOVAL SECTION
STONES WEBSTER/IONICS S02 REMOVAL G RECOVERY PROCESS
-------
ELECTROLYTIC CELL SYSTEM
"A" BANK
CELL FEED TANK
CCLL FCCO
•»
5
r»
P>
•CCTCLE
RECYCLE
p
J
ACID 1ULFATE
CAUSTIC
r
~?—
k
f
ANODE CATI
1 1
D M
• i • 1 1
V B
ANOI
OR
f
ANK
.YTE
UM
,
«••__
I-'
f
J
WOE CATI
-
ANODE CATHODE
0 M
I 1
J
—
ANODE CATHODE
1 1
D M
— 11
.1
"A"/'B"BANK
CATHOLYTE
DRUM
V BANK
ANODE
COOLER
(
c
I
T-'
f
CATf
CATt
•
«H>C | UK
1 1 1
It D M
1 1 1
J L-
K>DE I AN(
1 1 1
M D M
1 1 1
J L_
IOOC f AW
1 f 1
MOM
1 1 1
^1 1^
IDE
>OE
»DE
—
V/'B'BANK "B'BANK
CATHODE ANODE
COOLER COOLER
;
f
VB
ANOL
OR
}
' — r
ANK
YTE
JU
)
—
ACID SULFATE
RECYCLE f-
DRUM '
( )
^
W
^
t RECYCLE
DRUM
••Ten
10%
CELL FEED
PUMP
CIRCULATION
PUMP
"AW CELL CATHODE
CIRCULATION
PUMP
B' CELL ANODE ACID SULFATE
CIRCULATION RECYCLE
PUMP PUMP
CAUSTIC
RECYCLE
PUMP
FIGURE 1-2
REGENERATION SECTION
STONE & WEBSTER /IONICS SO, REMOVAL & RECOVERY PROCESS
-------
Oxygen is generated at the electrolyzer anodes as a pure gas. Hydrogen
is generated at the cathodes as a pure gas* If not recovered as
"byproducts, the hydrogen is burned and the oxygen is vented to the
atmosphere.
In the four compartment electrolyzers , and anode product is dilute
equivalent to about 6 to 8 percent of the SO^ absorbed. Being commercially
pure, it is vithdrawn from the system for marketing, or other disposal.
The sulfate ion so removed as dilute acid represents the method by which
S03 contained in the flue gas from the boiler plus the S02 oxidized in the
absorber to 803 are removed from the system. The total amount of SO^
determines the required number of four compartment electrolyzers and the
amount of dilute acid produced.
3.2 Technical Considerations
3.2.1 Absorption
In the absorption step, capital investment and operating cost are influenced
by causxic utilization and by sulfite oxidation. The following sections
explain and describe the influence of each.
A. Caustic Utilization
The first important process variable is the caustic utilization in the
absorption system, that is, the mols of SOa reacted per mol of caustic
(NaOH). The 8 percent caustic at the top of the absorber will first react
with the C02 in the flue gas to form sodium carbonate. The latter will
then react with the S02 in the flue gas to form sodium sulfite (NajjSO^).
The partial pressure of S02 above a solution containing sodium sulfite is
essentially zero, so it is possible to absorb virtually all of the SOg from
the flue gas.
In the lower part of the absorber, the entering flue gas reacts with the
sodium sulfite as follows:
Na2S03 + S02 + H20 - >• 2NaHS03 (sodium bisulfite)
As the conversion of sodium sulfite to sodium bisulfite occurs, the
absorption work accomplished by each mol of caustic fed to the absorber
increases. For example, if the fluid removed from the tower is a solution
of sodium sulfite in water (neglecting the sodium sulfate recirculated and
any formed in the tower), one mol of S02 can be removed from the flue gas
for each two raols of caustic entering the tower. If all of the sodium
sulfite is converted to sodium bisulfite, one mol of S02 can be removed from
the flue gas for each mol of caustic entering the tower.
There is, however, an inherent constraint that limits the extent to which
sodium sulfite may be converted to sodiura bisulfite in a practical situation;
it is that the partial pressure of S02 in the flue gas entering the absorber
must exceed the equilibrium partial pressure of S02 exerted by the sodiursi
1067
-------
bisulfite solution. The partial pressure of 802 in the flue gas entering
the absorber depends on the sulfur content of the boiler fuel and the
amount of excess air fired. The equilibrium partial pressure of S02
exerted by the sodium bisulfite solution depends on the concentration of
bisulfite in the solution and the absorption temperature, both of which
variables can be influenced by the engineer designing the system. Addi-
tionally, the design of the absorber tower, which the engineer controls,
influences the rate of transfer of S02 from the flue gas to the solution.
Accordingly, the absorption system can be designed to optimize the production
of sodium bisulfite, taking into consideration absorber capital investment,
solution recirculation expense, and electrolyzer system investment and opera-
ting cost.
B. Oxidation
The second major consideration in the absorber design is the extent of
oxidation of sulfite to sulfate by oxygen absorbed from the flue gas in
the absorber. 502 in the sulfate form cannot be released by neutraliza-
tion with recycled dilute acid solution. Sulfate generated by oxidation
in the absorber poses no technical problem but imposes some economic
penalty, reflected in both the investment and operating costs of the
entire process.
Operating costs are increased because more caustic is required to remove
S02 as sulfate than as bisulfite. Capital investment also is increased by
the need to include in the process a four compartment electrolyzer for
rejecting the excess sulfate as dilute pure I^SOi^. This four compartment
electrolyzer, described in section 3.2.3, achieves rejection of sulfate
without loss by the overall system of valuable sodium ion but with some
additional sodium ion recirculation.
Since absorbed oxygen reacts slowly to form sulfate, the engineer's objec-
tive is to design an absorber that will not absorb much oxygen, but will
rapidly absorb a substantial portion of the S02 as bisulfite. Based on
performance data from packed tower in S02 removal service, a system
removing S02 to the 150 to 300 ppm range should experience an oxidation to
sulfate of approximately 6 to 8 percent of the S02 transferred fron the
flue gas to the liquid.
3.2.2 Sulfur Dioxide Recovery
The liquid effluent from the absorber is mixed with a recycled dilute acid
solution before entering the stripper. The following reactions take place:
2NaIIS03
1068
-------
The S02 is taken overhead from the stripper, sufficient trays being
provided to insure that 99.5 percent of the S(>2 generated in the above
reactions is removed from the liquid.
The stripped sodium sulfate solution is cooled, adjusted in pH, filtered,
and fed to the electrolyzers. It is desirable to maintain the feed slightly
alkaline in order to keep the cation membrane resistance at a minimum.
Therefore, the pH of the stripped sodium sulfate solution is adjusted vith
caustic. Addition of the caustic causes heavy metal ions (leached out of
fly ash in the absorber) to precipitate from solution as hydroxides, which
are removed by filtration and ion exchange.
Absorber output varies with plant load, and average electrolyzer operation
is a function of average caustic production. In a plant with a large peak
demand, it is possible to shut off cell power during the peak and run at a
slightly higher rate when off peak. This can be done by providing storage
for sodium sulfate solution, caustic and acid, thus allowing absorber and
stripper operation to follow power plant loading, and the electrolyzers to
operate on the spinning reserve power available from the generators during
off-peak periods.
3.2.3 Absorbent Regeneration and Oxidation Product Rejection
The unique feature of the Stone & Webster/Ionics Process is the electrolyzer
system in which the caustic is continuously regenerated. Reference has been
made above to the use of two different designs, a three compartment elec-
trolyzer and a four compartment electrolyzer. The former is the basic
design which converts sodium sulfate into two streams, sulfuric acid and
caustic soda. The four compartment design represents a further refinement
and is the means by which excess sulfate is removed from the recirculating
liquid system as dilute, pure sulfuric acid.
A schematic diagram of the three compartment (Type A) electrolyzer is
presented on Figure 2—1. The main components are an anode, a microporous
diaphragm, a cation-transfer membrane, ar.4 a cathode. These components are
separated from each other by flow directing spacers which also provide the
required gasket ing. The sodium sulfate is fed to the central compartment
at a concentration of about 20 wt percent. Sodium ions migrate through the
cation-transfer membrane toward the cathode under the influence of a direct
current voltage impressed across the electrolyzer. At the cathode, water
Is electrolyzed to hydrogen gas and hydroxide ion as follows:
HO + e~ - » ?# («) * OH~ (cathode)
The hydroxide ions are electrically balanced by sodium ions entering the
cathode compartment through the cation-transfer membrane. Thus, the catho-
lyte effluent consists of NaOU plus I^SOfc which, after disengaging hydrogen,
is sent to the caustic storage tank and then to the absorption tower.
1069
-------
The function of the cation-transfer membrane is to prevent physical mixing
of the catholyte and center compartment feed streams. Essentially, only
sodium ions from the center compartment pass through the membrane to combine
with hydroxide ions produced at the cathode. In industrial applications
other than S02 removal by the Stone & Webster/ Ionics Process water may be fed
to the catholyte compartment in sufficient quantity to produce 6 to 10 per-
cent caustic. However, the Stone & Webster /Ionics Process is an essentially
closed system. Therefore, to maintain water balance in the system it will
generally be necessary to feed recycled sodium solution to the catholyte
compartment instead. This procedure does not introduce any problems.
The center compartment feed stream passes through the microporous diaphragm.,
into the anode compartment. At the anode, water is electrolyzed to hydrogen
ions and oxygen gas:
H0 - * %Q (g) + 2H+ + 2e~ (anode)
The hydrogen ions combine with sulfate ions to form sulfuric acid. The
diaphragm flow is designed to prevent hydrogen ions from migrating across
the cat ion- transfer membrane. Such flow must give enough linear velocity
through the diaphragm to sweep hydrogen ions back into the anolyte compart-
ment. The diaphragm should have an hydraulic resistance adequate to insure
that such flow is substantially uniform over the entire surface. A flow
which is sufficient to sweep back most of the hydrogen ions carries with it
about half of the sodium in the center compartment. Thus at the anode, only
part of the sodium sulfate in the feed stream is electrolyzed to sulfuric
acid. The anode product is therefore a mixed solution containing both
sulfuric acid and sodium sulfate. This product is the dilute acid solution
that is recycled to release S02 in the stripper.
In the four compartment (Type B) electrolyzer, a schematic diagram of whicn
is presented on Figure 2-2, the reactions and operation of the catholyte
compartment are exactly those described for the three compartment electro-
lyzer. The catholyte effluents from both electrolyzer types are combined
before being pumped to the absorber. The center compartment feed flows
through the porous diaphragm, into the "mid-anolyte compartment," from
whence most is discharged from the electrolyzer. This stream contains some
sulfuric acid due to the inability of the membrane to exclude completely the
hydrogen ion. It is combined with the anolyte stream from the three compart-
ment electrolyzers .
The important reactions of the four compartment electrolyzer occur at the
anode, where water is electrolyzed to hydrogen ions and gaseous oxygen.
These hydrogen ions combine with sulfate ions entering the anode compartment
to form a sulfuric acid anolyte. The function of the anion-transfer membrane
is to exclude sodium from the anode compartment while allowing sulfate to
enter. It also, to a large extent, prevents hydrogen ions from leaving the
anode compartment. Thus, the anode reactions permit removing sulfate ions
without losing sodium.
1070
-------
1/2 02
H2S04
D CM
* Na2S04
D-POROUS DIAPHRAGM
CM* CATION SELECTIVE MEMBRANE
M
Na2S04
NaOH
Na2S04
FIGURE 2-1
SCHEMATIC DIAGRAM OF "A" ELECTROLYTIC CELL
H2S04
1/2 02 Na2S04
H2S04
* J >
NaOH
1
© AM D CM 0
1/2 02
j fc
f —
4. "*"
-»*2H
t
+ Na2S04
-S04« •*— S04« -
1
>
20H-*-
It
1
Na2S04
t
t
POROUS DIAPHRAGM Na2S04
H20
CM • CATION SELECTIVE MEMBRANE
AM • ANION SELECTIVE MEMBRANE
FIGURE 2-2
SCHEMATIC DIAGRAM OF "B" ELECTROLYTIC CELL
1071
-------
A large number of materials of construction have been tested in order to
optimize the life and performance of the electrolyzer components. Anodes
are a lead alloy; cathodes are steel. The microporoua diaphragms and ion
transfer membranes are polymeric organic materials.
Each electrolyzer consists of an anode, a cathode, at least one membrane,
and one diaphragm vith appropriate separators and internal flov distributors.
A second membrane is placed between the diaphragm and the anode in the Type B
electrolyzers. Fluid is internally manifolded and is in parallel. The
electrolyzers are arranged into a number of modules within a single struc-
tural frame. The electrolyzers within a module are in parallel electrically
and the modules are in series. In order to conserve space and minimize field
piping, three frames are arranged in a vertical array to form a bank. The
internal fluid manifolding is brought to headers in the frames. The headers
are provided with auick disconnect fittings to the electrolyzer fluid distri-
bution and collection mains. Typical arrangements of modules, frames and
banks are shown schematically in Figures 3-1 and 3-2.
3.2.U Electrolyzer System Auxiliaries
Hydrogen produced by the cathodes is gathered in standpipes and passed to
gas collection mains. Caustic is collected from the bases of the stand-
pipes and pumped to the caustic surge tank. Oxygen is similarly collected
from anode fluids.
Since Ohmic losses must be dissipated from the electrolyzers, heat exchangers
are provided for the recirculating anolyte and catholyte streams. The anolyte
coolers have Karbate tubes and the catholyte coolers carbon steel tubes.
The pure, dilute acid, produced in the Type B electrolyzers, will be approxi-
mately 9 percent acid.
3.2.5 Potential Problems and Process Limitations
A number of problem areas have been identified and potential solutions
postulated. These include:
1. Oxidation of S02 to 803} amount and affect on economics.
2. Effect of oxidation inhibitors if used, on system and cells.
3. Effect of NOX on system and cells.
4. Effect of fly ash on system; which minerals, if any, affect
oxidation, absorption, filtration, cell performance and
cell life.
5. Effectiveness of quench section and absorption section in
removing fly ash.
1072
-------
6. Effectiveness of ion exchange in removing heavy metals from
solution.
7. Deleterious impurities in recovered S02; nature, amount,
and effect.
8. Efficiency of absorption, stripping, and electrolysis.
9. Materials of construction.
10. Disposal of weak acid from "B" cells.
None of the above problem areas appear to be without a reasonable solution.
The potential process limitations mainly lie in the area of efficiency in
absorption, stripping, mimization of cell power, and the ability of the
system to produce a high bisulfite to sulfite ratio.
Pilot plant work in Tampa and at Watertown has given indications of what
may be expected but long run testing on commercial sized equipment will be
required to give reliable answers to many of these areas. Corrosion test
coupons that are being installed in the WEPCO pilot plant will aid in
selection of materials for the prototype plant.
U. PILOT PLANT
The pilot plant at Wisconsin Electric- is being installed primarily to
verify operability of the Stone & Webster/Ionics process in an integrated
operation under Wisconsin Electric conditions, and to confirm the process
design of the 75 MW prototype plant. The pilot plant will process a slip
stream of about 2,900 ACFM of flue gas at 300 P containing about 1,000 ppm
of sulfur dioxide from a coal fired boiler. It is designed to remove a
maximum 95 percent of the S02*
As in any coal fired boiler, it is to be expected that the sulfur content
in the flue gas will vary with the quality of the coal. At Wisconsin, the
amount of sulfur is known to vary from as low as 1/2 percent to as high as
5 percent. Since it is desirable to operate the pilot plant at reasonably
uniform conditions during any one test, a by-pass from the discharge of
the absorber to the suction of the forced draft fan has been installed to
permit adjusting the S02 content of the flue gas to a fixed value for each
test.
The absorber contains three stages of contact with provision to operate on
two only if desired.
1073
-------
H2S04
>
I"'
T;
k JV *|
ft
JJ
'T
i
i
i
i
n!
J
y
fl
i
LI
IT,
r
11
ft
|
Lj
i
i
i
i
i
u
r
i
It
u
u
ft
J(
•I
1
NoOH
M'MEMBRANE
D-DIAPHRAGM
FIGURE 3-1
MODULE
,so.
II20W4 "•— *
^ 1± t
NoOH
TOP
FRAME
MIDDLE
FRAME
i
__c
__/
BOTTOM
FRAME
No!
^
4
1
i
I
i
M
^
i
M
ft
^^
^
L
M
t t
> ^ ^
••
^
t
M
ii
\ \
>
t
M
I
V
j
^
1
n*T
M
f
>
k
4
M
J
>i>
M
1*
4 * 4t A t
\
k
M
^ h
M
>lf-
— «
T t r
' k , _ _^ ^ ^
>t >
.
^ k
M
'A ^
M
h
J
J-*-H2S04
' ? **'
S
M
1
^
<
<
~^~s
3
c s
M
T T T
r
IY
«
T^^^"T^^.
t
]
k ^
-n*
M
^
t ^
-n*
M
t
J
k
M
M
T '
J1+ l-n*
M
k ' M >
M
M
j
l
M
T * * *
FIGURE 3-2
/
t
^ ,
t
No2SC
\ —
4
\
i_-_
U
V
»
t
M*MODULE
BANK
1074
-------
The pilot plant contains all the elements that present knowledge indicates
will be required in the 75 MW prototype plant. The equipment, in so far
as possible, is of a type that can "be obtained f.or the larger plant.
Materials of construction have been selected based on previous experience
to be suitable for commercial operation. In addition, test coupons of
various materials are installed at strategic locations so as to provide
information on alternate materials of construction.
The testing program is expected to begin in May 1973 and extend for about
six months. Data will be obtained on:
a. Mass transfer coefficients in the absorber and stripper;
b. Extent of S02 oxidation to 803;
c. Effect of operating variables on S02 removal and on
sulfite-bisulfite ratio in the rich liquor;
d. Performance data on electrolyzers, including power
consumption for both types of electrolyzers under
field conditions and over a range of operating
variables.
Concurrent with the pilot plant testing, prototype electrolyzer modules
similar in design to that intended for installation in commercial plants,
will be operated at Ionics1 tfatertown facility to obtain data for the
design of the full-sized electrolytic "cell system.
5. 75 MW DEMONSTRATION (PROTOTYPE) FACILITY
Following successful demonstration of the pilot plant, Stone & Webster will
proceed with design and installation of the 75 MW Demonstration Facility,
assuming that the estimated costs of the system are acceptable.
5.1 Basis of Design
The basis of design is:
1. Process all the flue gas from one 75 MW boiler at 100#
load factor; approximately 225,000 ACFM at 300° F
2. Average sulfur content at 2,5 wt % in the coal.
3. Flue gas to be desulfurized to 200 ppm or lower.
k. Oxidation of S02 to be a maximum of 10 percent.
5. Minimum ratio of mols SO2 absorbed per atom of recirculated
sodium equal to 0.87.
1075
-------
Figs. 1-1 and 1-2, referred to earlier, show the major equipment and flow
schemes for the 75 Mtf unit.
The Facility will be designed to remove a minimum of 90 percent of the S02
from the stack gas. A minimum of 90 percent of the absorbed SOg will\be
recovered as S02 and the balance removed as dilute silfuric acid.
Anticipated pressure drop across the absorber, including duct losses, is
eight inches of water.
The recovered S02 vill be dried, compressed, and liquefied for sale. It is
believed that during the test program an outlet can be found for the 9 wt %
sulfuric acid purge stream, amounting to about five tons per day as lOOjJ
H2SOJ|.
5.2 Operating Program
The primary purpose of the prototype plant is to provide information on the
operating and maintenance characteristics of the process, and to develop
conclusions with respect to its technical and economic feasibility. The
program includes:
1. Systematically varying the primary process variables over
the practical operating ranges, and collecting the data
required to characterize the process;
2. Developing correlations to assist with scale-up of the
process equipment and estimating performance under a
variety of operating conditions;
3. Determining the electric power consumption of the
electrolyzers;
kt Evaluation of the operator work load and maintenance in
the electrolyzer and non-electrolyzer sections.
5. Evaluating the effects of corrosion, to serve as a basis
for selecting materials of construction for future
installations.
6. PROCESS ECONOMICS
Since many large boilers supplying steam for power generation are in the
UOO to 500 MW range, we have developed process economics for this paper
for a 1*00 MW unit instead of the smaller 75 MW unit planned for Wisconsin.
We believe this will be more representative of the average power plant
installation. Cost of utilities do not necessarily reflect WEPCO costs
but are those generally used in assessing costs of competitive processes.
1076
-------
An operating load factor of 80 percent for the UoO MW unit has been used
for the utility and a k percent sulfur coal has been assumed. Credit
taken for the byproduct sulfuric acid is deemed to be reasonable for the
sale of the acid by the utility. Fuel value only is assumed for the
hydrogen even though there may be particular locations at vhich hydrogen
could be sold at a much higher value. No credit is taken for the byproduct
oxygen.
Approximate installed costs and operating costs for Stone & Webster/Ionics
S02 removal unit for a UOO Mw coal fired steam generating unit are given
in Table 1. Included is the cost of facilities to convert the S02 to
commercial grade sulfuric acid. Alternatively, the S02 could be dried and
sold as liquid product or could be converted to sulfur.
The fixed charge rate assumes 100 percent financing by industrial revenue
bonds at 6 percent interest, straight line pay backs of principal,
accelerated depreciation under IRS Section l69» and a 7 percent discount
rate for all cash flows. Such industrial revenue bonds and accelerated
depreciation have become very popular for financing pollution control
plants. It is assumed that these provisions will remain in effect. They
permit a fixed charge rate of 7.1 percent per annum after taxes on the
20 year life assumed for this project.
The indicated annual cost of about $2 million for S02 removal is influenced
by the rate the station charges for power used in the electrolyzers and by
the credit assigned to the by-product sulfuric acid. As shown in the foot-.
note of Table 1, the net annual cost of the plant may be defrayed by (a) an
average increase of 0.73 mills in the price of sold power; (b) an increase
of $15.00 per ton in the price of sold sulfuric acid; (c) a decrease of
9 mills in the station service electric rate; (d) a decrease of $1.83 in
the price of purchased coal; or (e) any combination of these. Each of
these alternatives includes a provision for income taxes of 50 percent.
As indicated by the footnote, the trade-off point between S02 removal and
the purchase of low sulfur coal is a premium of $1.86 per ton for the latter,
The energy consumption of the electrolyzers is a unique type of station
service account. The absorption section must process flue gases as
generated, but the sulfur dioxide recovery and the absorbent regeneration
and oxidation product rejection may be decoupled from a real-time, opera-
tional standpoint. (They may be geographically removed as well, connected
with the absorption section by relatively small, liquid-handling pipelines.)
Electrolyzers, when on line, can be instantly interrupted without damage,
either in blocks or all at once at the sole control of the utility dispatch
center. Power pool agreements as well as plant operating principles provide
for spinning reserve capacity which can be made available from generators
Vhich are spinning and generating power at a rate less than nameplate
Ratings. Such reserve is usually sufficient to cover a predetermined per-
1077
-------
centage of total system load or a percentage of the load on the largest
generator in the system. If the generator capacity represented by this
reserve is under the sole control of the utility dispatch center, there
should be no bar to using this reserve for performance of useful work.
Such an interruptible station service account is instantly available in
comparison vith the usual "interruptible" customer and should be regarded
as loaded spinning reserve. It therefore seems reasonable to charge the
removal system only for the marginal cost of fuel, labor, and maintenance
materials vithout capital charges. If this approach is followed and the
electrolyzers are operated on spinning reserve power, the S02 removal cost
might be reduced by about 0.3 mills per kwhr.
The regeneration subsystem is amenable to evolutionary development.
Significant reductions in energy and maintenance costs of the electrolyzers
may be confidently predicted, based on pilot plant results to date. Such
improvements should result in the electrolyzer characteristics shown in
Table 2. The 1975 components should result in 15 to 20 percent saving in
electrical energy consumption.
1078
-------
APPROXIMATE ANNUAL OIERATING COST
STONE & WEBSTER/IONICS 302 REMOVAL PROCESS
Basis
o
^j
IO
Installed Cost
302 Removal Unit
Sulfuric Acid Unit
Annual Cost
Fixed Charges after Taxes
Operating Cost - S0% Removal Unit
Steam
Electricity (cells, pumps, blowers & lighting)
Cooling Water
Process Water
Deionized Water
Operating Labor
Maintenance
Subtotal
Operating Cost - H2S04
Credits
Sulfuric Acid
Hydrogen
Subtotal
Total Annual Cost
400 Mw Station Burning 4.0 Wt
Operating Load Factor
$14,000,000
2,500,000
$16,500,000
Unit Cost
& per annum
$0.60 per 1,000 Ib
$0.007 per kwhr
$0.02 per 1,000 gal
$0.05 per 1,000 gal
$1.00 per 1,000 gal
$10.00 per hrU)
4^> per annum
$2.00 per ton
$15.00 per ton
$0.60 per MM Btu (LHV)
Notes
Equivalent Cost per Sold Kwhr
Equivalent Increase in Sold Sulfuric Acid Price
Equivalent Decrease in Station Service Rate
Equivalent Decrease in Cost of Purchased Coal (l,083,OOO ton)
(l) Includes overhead, and supervision
(2) All costs including maintenance are represented.
S Coal
Quantity
70,000 Ib per hr
31,900 kw
8,500 gpm
650 gpm
40 gpm
2 oper/shift
465 tons per day
465 tons per day
24.7 MM Btu per hr
0.73 mills or
$15.00/ton or
iiO.OOQ/Kwhr or
$1.88/ton
Annual Cost
$1,174,000
313,000
1,570,000
76,000
14,000
18,000
175,000
560.000
$2,726,000
271.000
4,171,000
(2,035,000)
(io4tooo)
(2,139,000)
$2,032,000
I
sr
-------
ELECTRQLYZER CHARACTERISTICS
STOME & WEBSTER/IONICS S02 REMOVAL PROCESS
Basis: Plant of Table I
Table 2
Elect rolyzer
Type
1973 "A"
1973 "Bn
1975 "A1
1975 "Bf
Voltage
4.6
3.8
4.2
Current
Efficiency
95
95
Kilowatts AC
24,400
Total 1973 2
21,300
2.600
Total 1975 23^900
(l) First number is overall current efficiency.
Second number is current efficiency for the pure
(2) Cell requirement only. Added 3,000 few in Table 1 for
pumps, blowers, etc.
1080
-------
FW-BF DRY ABSORPTION SYSTEM
FOR
FLUE GAS CLEAN UP
by
W. F. Bischoff
Foster Wheeler Corporation
110 South Orange Avenue
Livingston, New Jersey
1081
-------
ABSTRACT
The information contained in this paper is directed
to those individuals in private industry and governmental agencies
concerned with the reduction of S02 in flue gas emissions. It is
intended to serve as a means of presenting information on the history
of the dry adsorption process developed by Bergbau-Forschung, GmbH,
process description, technical details, pilot plant experience,
demonstration plant design and economic factors.
1082
-------
FW-BF Dry Adsorption System
for
Flue Gas Clean up
by
W. F. Bisehoff
Foster Wheeler Corporation
I. Introduction
Foster Wheeler is a major manufacturer of steam generation
equipment and also a major process plants contractor. In both of these
roles we have been actively concerned with pollution problems and their
proper solutions. To this end the Corporation has participated in consid-
erable research and development work in order to solve the pollution pro-
blems of our customers.
Our research program has included) in part, the extent*j..~
technical and economic evaluation of various pollution control systems.
One of the systems on which we have expended considerable time is the
subject of this paper.
II. Background of Bergbau-Forschung* GmbH
Bergbau-Forschung, GmbH, (BF) is the central research institute
for the German coal mining industry. They are industry supported and there-
fore analagous to our own Bituminous Coal Research Institute..
In 1963 Bergbau became interested in the application of chars
produced from coal to remove sulfur dioxide from stack gases. The chemical
reactions involved for adsorption were well known but, the means for doing
it successfully on a continuous basis were not. Bergbau therefore determined
to produce a char from non-coking coals which would meet the system require-
ments they had conceived. The char had to have the following characteristicst
1. High mechanical strength
2. High S02 adsorption capability
3. High ignition point
li. Low pressure drop when put in a travelling bed reactor
1083
-------
•L'iK^fSy^f^f:'.- -;•-.* ;-*K,
vH,:*.•;>•?>?'>: *>j<* -^ > ».' - ••-• -.'t*' ;-~- X -Jinf* .-
...•v."o -'i- » '>-M> - -,^-<---. :,•- ->.;*--V^-:-..
••v::.;^: x-^w^-^-':
.•-X^MV-; ^'^-S- •:•'-.-:
-f " • ' ^: -i \ •''--,- ^-' •'• ^' ' •
A6 R AM^::a«^SSS:^:"
** " -'. ^ -"' • *•?-, *^*V-^' 'j--** • " '•' --r '• ' '• J '*- '- fc-'i1'' -'if^^r^ \ '.'-' •'• -r' - * -
J" -..•.."••••• ' " "*-%'-i:^l»" .j7 - vv * •", " Tj.rj,* • •", " • '"-'"'V Ji ^'ii *'"<"" ;i-r:.'3 " . ^" •••'.-.-." ' •':".
o
00
BOILER
CHAR FROM
REGENERATOR
DUST
COLLECTOR
w
ADSORBER
I
CHAR TO
REGENERATOR
/STACK
FW-I3I-004
Fig. 1
-------
Bergbau was successful in this endeavor and then built a laboratory scale
pilot unit to test their system. After the successful operation of the
laboratory unit Bergbau built a pilot plant of about 1.$ MW equivalent which
was operated in a slip stream from a coal fired utility boiler. The gas flow
through the pilot plant was 3000 m^/hr. (10J>,000 SCFH). This pilot unit was
run continuously for two years during which time all of the correlations
developed during the laboratory work were confirmed. Although the system was
basically developed for removing S02, it was discovered in the pilot plant
work that the system also has capability to significantly reduce NO^ and
particulars in the flue gas stream.
As a result of the work done by Bergbau as well as our own develop-
ment efforts on the system Foster Wheeler is convinced that the Bergbau system
is the most feasible for application to a wide variety of boiler flue gas pollu-
tion problems. Because of this conviction Foster Wheeler has become fully
licensed under the Bergbau technology and patents.
During the period that Foster Wheeler has been licensed for the
process, development work has continued both in Germany and in the United
States. The development work has resulted in what is known as the Ftf-BF Dry
Adsorption System.
III. Process Description
Figure 1 shows how the FW-BF System fits into a typical utility
boiler system. The adsorber is the key to this system. It consists of
parallel louvre beds which support and contain the char during the adsorption
phase. The char moves slowly downward in plug flow while the pollutant laden
gases pass through the adsorber bed in a cross flow. The fan indicated at the
gas discharge of the adsorber is utilized to make up the pressure losses which
occur during the passage of the gas through the adsorber bed. Clean gases
1085
-------
o
oo
FLUE GAS (250-300 °F)
S0
I
CHAR PELLET
3/8" DIA. X3/4"LG.
FW-131-005
Fig. 2
-------
leaving the fan are directed to the stack for disposal. Once the char has
progressed through the adsorber it is taken off and sent to the regeneration
section. Once regenerated the char is recycled to the top of the adsorber.
This process scheme allows for the continuous adsorption of pollutants and
the continuous on site regeneration of the adsorbent media.
The reactions that are taking place in the adsorber are shown
in part in Figure 2. The rectangular block represents a single char pellet
which is approximately 3/8" in diameter x 3/8" - 5/8" long. Boiler flue
gases containing S02, NOX, water vapor and particulate matter come into contact
with the char pellet. The S02, oxygen and water vapor are adsorbed into the
char pellet. Once in the char pellet the S02 is oxidized to SO^ and subsequently
catalytically converted to sulfuric acid which is firmly held in the interior
pore system of each char pellet.
Nitrous oxides are likewise adsorbed in the char pellet and held
until regeneration occurs. Particulate matter is collected on the surface of
the char pellets.
The reactions in the adsorber are designed to take place through a
range of temperatures above 2£0*F. The lower limit of 2$0°F is specified
because of dew point considerations. The relationships between adsorber size,
adsorption efficiency and operating temperatures indicate that optimum
efficiency is achieved at temperatures around 27£° to 300°F range.
As a result of the adsorption phenomena and the dust collection
capabilities of the system, it is possible to reduce pollution emission levels
as indicated below:
1087
-------
Pollutants Removal Efficiency
S02 80-9#
NOx UO-60J6
Particulate matter 9
The S02 removal efficiency is listed as a range to indieate\
flexibility of design. In some cases, dependent on the sulfur in the fuel,
it may not be required to achieve a 95% removal efficiency to be in accordance
with the required codes. Therefore economies in the design can be realized
by virtue of the lower removal efficiency requirement. NO^ efficiencies are
indicated as a range for the same reasons. The particulate removal efficiency
shown above is meant to indicate that the system is capable of removing from
90 to 95/5 of the particulate matter received in the inlet of the adsorber
after prior treatment with high efficiency electrostatic precipitators.
The regeneration of the saturated char can be accomplished by wash-
ing (wet regeneration) or by contacting it with hot sand (thermal regeneration),
Wet regeneration produces dilute sulfuric acid (lS>% by weight) as a by-product
material. Because of the limited use-for this material thermal regeneneration
is the more practical method of regeneration.
In thermal regeneration an enclosed refractory lined vessel is
utilized to contact the saturated char pellets with hot sand which has been
heated to approximately 15>00°F. Regeneration is a function of vessel size,
retention time, sand to char ratio, and sand temperature. As the char pellets
become heated the reactions that took place in the adsorber are reversed pro-
ducing a concentrated stream of S02, 1^0, C02, and N2- In other words the
I^SO^ produced in adsorption is converted to H^O + SO?. The SO^ is reduced
to SC>2 in the presence of the hot char and produces C02 by the combination of
the liberated oxygen with the carbon in the char. In a like manner the oxides
1088
-------
:.-^>JV^
l$&&.-&'
"i^*^^
i$t's|
:^--^H'-':^i;^;?wr\*^.:-'i?^ "'*-£•- ':'• "-.••:'' ' . *:"-r.-^J^-< ^Af^K'^' If-'- ' < .- •'•• '-'^1'^VA^
isaiPSiii^
o
00
vo
FROM
DUST
COLLECTOR
CHAR FROM
REGENERATION
N.
HAR TO
REGENERATION
FLUE GAS
" TO STACK
FW-I3I-008
Fig. 3
-------
of nitrogen are reduced to N2 with the resultant production of additional
003.
The concentrated stream of off-gas is then directed to a proprie-
tary Foster Wheeler OGT system for conversion of the S02 to elemental sulfur.
IV. Technical Details
The FW-BF System consists of three basic sections:
1. Adsorption
2. Regeneration
3. Off Gas Treatment
The adsorption and regeneration sections of the system are those
which were developed and designed originally by Bergbau. The Off Gas Treat-
ment section is a proprietary FW design for which we have made patent
application.
A. Adsorption
The adsorption section, as stated previously, is the heart of the
system. It is here that the reduction of pollutants take place. As can be
seen in Figure 3 the adsorber is a vertical bed in which the char moves slowly
downward at a rate of from 0.5 to 1.0 ft .per hour. This slow but variable
movement allows for operational flexibility as well as guaranteeing low abra-
sion rates in the adsorber. The louvres are designed to not only support and
contain the moving char bed but also to give direction to the incoming and
outgoing flue gases. Flow distribution of the gases across the bed is main-
tained by means of adequate pressure drop. The pressure drops that have been
measured for these types of adsorber beds vary from k" K2Q to li;" H20 depend-
ing upon the adsorber configuration.
The adsorber has been designed based on the information generated
during pilot plant operation. During this pilot plant operation several diff-
erent bed cross sectional areas were utilized varying from 0.£ meter x O.J> meter
1090
-------
to 1 meter x 1 meter. Commercially sized beds vary from U1 x li1 to 6' x 6*
thus keeping the scale-up factor for adsorption beds within reasonable lijnits.
The height of adsorber beds is varied to achieve the necessary frontal bed
surface area to bed volume ratio. For large gas flows a series of beds are
operated in parallel so as to achieve the desired removal efficiency, ty
designing large adsorbers as a series of modules it allows for maximization
of shop fabrication and minimizes field erection.
The flow of char through the adsorber beds is controlled at the
bottom of each bed by a vibratory feeder which is controlled as a function of
the amount of S(>2 entering the adsorber. This mode of operation gives flexi-
bility to the system not only from the standpoint of variations in load but
also variations in the percent of sulfur in the fuel. As load decreases the
amount of S(>2 entering the adsorber also decreases and therefore it is possible
to reduce the flow rate of char so as to maintain a consistent removal effi-
ciency. The reduction of char flow then results in a lowering of operating
costs for the unit. For a given boiler load with an increasing percent sulfur
in the fuel the char flow rate can be increased so as to achieve a reduction
efficiency compatable with the fuel being fired.
Upon leaving the feeder at the bottom of each adsorber bed section
the char is discharged onto a natural frequency or oscillating type conveying
system. The end of this conveyor is fitted with a screening section which
screens off the majority of the fly ash collected in the adsorber. The fly ash
removed is in a dry form. In order to eliminate fly ash losses from the system
the entire system is totally enclosed. Once the char and fly ash have been
separated the char is elevated by means of an enclosed bucket elevator and sent
to the regeneration section. Upon completion of regeneration the char is
returned to the adsorber section, elevated in an enclosed bucket elevator and
distributed evenly across the top of the adsorber beds.
The entire adsorber and its associated distribution system is
1091
-------
o
10
N»
SATURATED
CHAR FROM
ADSORBER
REGENERATOR
REGENERATED
CHAR TO
ADSORBER ""
SAND
HEATER
!
C02 I
H20 f
N2 J
TO OFF-GAS
TREATMENT
FW-I3I-009
Fig. U
-------
enclosed so as to provide a sealed system and eliminate any possible second-
ary pollution sources.
B. Regeneration
In Figure k the saturated char from the adsorber enters the
regenerator vessel where it is mixed with hot sand. The char temperature is
raised as a result of this contact to a level where the reactions that have
occurred in the adsorber are reversed as previously discusseo.
Sand is utilized as an inert heat transfer media and as such does
not take part in the reactions occurring within the regenerator. Its sole
function is to supply heat so that the reactions may take place.
The mixture of hot sand and hot char pellets flow slowly downward
through the regenerator and their flow is controlled at the discharge of the
regenerator by a char-sand feeder and separator. This feeder-separator not
only controls the flow rate of the materials through the regenerator but also
acts to separate these two materials after regeneration has taken place. The
char pellets, being considerably larger than the sand, flow overhead on the
char-sand separator and are subsequently cooled and are then returned to the
adsorber for reuse. The sand passes through the screen section and is returned
via an enclosed hot bucket elevator to a fluidized bed sand heater. The
fluidized bed sand heater restores the heat lost to the char and recycles the
reheated sand to the top of the regenerator.
The production of C02 in the regenerator results in a chemical
consumption of the char pellets which accounts for approximately 90% of the
char makeup requirement. The remaining 10$ char makeup is as a result of
mechanical or abrasion losses. As the char is consumed chemically or reduced
in size mechanically it eventually becomes too small to pass over the top of
the screening section in the char-sand separator. At this point the small
quantity of small char particles enter the hot sand side of the cycle and
1093
-------
make their way to the fluidized bed sand heater. There, in the presence
of excess oxygen, they are consumed as a fuel. However, the quantity of these
particles represent a negligible savings to the fuel requirement for the sand
heater.
The hot flue gases eminating from the top of the sand heater
contain S02 as a result of fuel combustion and also contain a source of
recoverable heat. These hot gases are piped back to the inlet of the boiler
air preheater where they are mixed with the boiler flue gases. The heat in
this stream is utilized in the air preheater to preheat incoming boiler air.
The additional S02 is now a part of the incoming flue gases to the adsorber
and is adsorbed in the same manner as the S02 from the boiler flue gases.
That portion of the fly ash which was not removed by screening at
the discharge of the adsorber is removed from the char in the regenerator by
the gentle interaction of the sand with the char. This material passes through
the char sand separator and enters the hot sand loop where it too acts as a
heat transfer media. The quantities of fly ash removed in this manner are
extremely small but eventually constitute a buildup in the inventory of material
in the sand loop. It is therefore necessary to periodically remove a portion
of the sand fly ash mixture and replace it with a smaller quantity of new sand.
The regenerator vessel utilized in this section is identical to
a vessel utilized in the production of the adsorbent char. This regenerator
has been in operation for five years in a commercially sized char production
facility in Germany. Therefore the problems associated with its design and
operation have already been solved.
The fluidized sand bed heater utilized in the regeneration section,
while unfamiliar to many in the utility field, has been used for many years
by the process industries in applications identical to those contemplated here.
C. Off-Qas Treatment Section
Figure 5> schematically illustrates Foster Wheeler's proprietary
1094
-------
OFF-GAS TREATMENT
CRUSHED COAL
OFF-GAS
FROM
REGENERATOR
to
in
REACTOR
TO
ADSORBER
FW-131-010
Fig. 5
-------
•'-.-•-• x •••• *"'•'; '-'-' •. i^^i-i-^T, ••% ^*'--V." ' '"
•'.'V^iCv-'^1 ;4\ "••:"*''''•*?•''J. :'*'•'' "'
^^^^•^
o
o
o»
»yii.vj^ .•*
DUST COLLECTOR
•••••••••••••••••••
^:^r?^^j;-;^^.,.::'l f/^'JC "
.'-•• •. &. v-:--jj••* '••''^•Oi^a*'%iiii
^K'Svi- .Vr\^0|5^^p^t-^
c^ii^>Si
.K. ^; .JI4...TT
••• .•' • 'j- ^ '--' ,^-**S>SL.".3*
ASH
-- . : >;• j-^1""•?''?.*' •-"??-^ •-- 'V^-'- |"-t-«*-:
• - •jK »;./.
•
REGENERATION
OFF GAS TREATMENT
- ^ ,- •
iMi.
Fig. 6
-------
design for the off-gas treatment section. In this section concentrated
off-gas, rich in SC^, is received from the regenerator and converted to
elemental sulfur. The reaction takes place in a reactor vessel which is
filled with crushed coal. The SC^ stream is reduced to gaseous elemental
sulfur and the liberated oxygen combines with a portion of the coal tc forn
carbon dioxide. The gases leaving the reactor enter a sulfur condenser where
the sulfur is condensed and carried off as molten elemental sulfur. The
sulfur is collected and stored in an insulated tank which is equipped with
steam heating to make up for heat losses through the insulation system and
thereby maintaining the sulfur in a molten form pending shipment via heated
tank car or heated truck. These two methods of sulfur transfer are common in
the sulfur industry and have been utilized for many years. As an alternative
the sulfur could be pumped to cooling pits and allowed to harden for shipment
as a dry solid.
The gases leaving the condenser consist of carbon dioxide, water
vapor, elemental nitrogen and those remaining sulfur values not converted to
elemental sulfur in the Off-Gas Treatment section. These gases are recycled
to the front of the adsorber where the sulfur values are adsorbed on the char.
The C02, H20 and N2 are non-reactive materials at 33>0°F and therefore pass
through the adsorber bed and enter the atmosphere as non-polluting gases.
D. System Recap
The three sections of the FW-BF Dry Adsorption System are shown
together in Figure 6. As can be seen the system is composed of a series of
closed loops which allows only cleaned gases or by-product materials to leave
the system.
The requirements of the system are relatively simple. Cooling
water is required to cool the char and also condense the sulfur. The major
portion of this water is recoverable and recycleable. In addition tc cooling
water the system requires char, dand, clashed jp.ant ocal ar.o aurf lifcj-y DTWS -.
1097
-------
V. Maf1or Areas of Application
The FW-BF system is applicable to all fossil fuel fired utility
applications. Primary application would be jji the area of large coal fired
control station utility boilers because- of the system's inherent ability to
handle all three of the pollutants associated with coal fired boilers. Size is
another factor. It is felt at least for the present that applications above
150 MW size range will show superior economic advantages. Of course it is possible
to apply the system to smaller units which could be connected in parallel to a
single FW-BF unit thereby achieving the economics of larger sized units.
Similar units in refinery or chemical process plants could also
be handled provided that similar size considerations are taken into account.
Work is presently underway to determine application to metals
smelting operation. Initial studies, particularly those involving wet regenera-
tion, indicate good potential application in this area.
VI. Demonstration Plant Design
At the present time two demonstration plants for the system are in
the design stages.
One contract is being executed in Germany which has been financed in
part by the West German government. This contract involves the installation of
a BF unit as a 35 MW equivalent slipstream off of a large central station boiler.
The system will use the thermal regeneration scheme and will be ready for
operation in fourth quarter of 1973.
At the same time Foster Wheeler is designing an FW-BF unit to be
installed at a major southeastern utility. This unit is designed to meet State
and EPA code for S02 removal when burning 3% S coal. The adsorber will be
capable of handling a 50/6 slipstream from a nominal UO MW coal fired boiler.
The thermal regeneration and off-gas treatment sections are being sized for
the full 1*0 MW flow. The combination of half sized front end and full sized
1098
-------
back-end will allow a comprehensive testing program utilizing fuels varying
in ash and sulfur contents while minimizing capital expenditures during the
demonstration program. At the completion of testing it is anticipated that a
second 20 MW adsorber will be installed thereby converting the unit to a full
sized lj.0 MW unit.
High efficiency electrostatic precipitators capable of meeting
state codes for particulate emissions will be installed upstream of the FW-BF
system. During the testing program sections of these precipitators will be
selectively de-energized thereby imposing higher dust loadings on the FW-BF
unit. Measurement of dust loadings will be taken at the inlet and outlet on
the FW-BF unit to determine its actual efficiency as a particulate removal
system. It is anticipated that the test work previously conducted by Bergbau-
Forschung will be re-confirmed on this large scale test and that the particulate
removal efficiency of the FW-BF unit will be firmly established.
VII Capital and Operating Costs
The capital and operating costs of the system vary as a function
of the unit size (expressed as the MW rating of the boiler) and the sulfur con-
tent in the fuel. Capital costs represented as dollars per kilowatt can be
plotted against the megawatt rating of the boiler and the percent sulfur in the
fuel. This plotting will result in a family of curves which stabilize and become
relatively level after a boiler size of 150 to 200 megawatts is reached. Beyond
the 200 megawatt level relatively little decrease in the curves is witnessed.
This indicates that the most economic application of the system is in a range of
boiler sizes above l£o to 200 megawatts. Below this level the fixed engineering
costs cause the cuves to rise more sharply so that the cost for units below
50 megawatts are uneconomic. This uneconomic situation can be overcome, however,
by the connection of several smaller boilers to a single treatment system which
would then have a rating equivalent to the sum of the individual boilers connected
to it.
1099
-------
The capital costs in the range above 200 megawatts vary from $20 to
$UO depending on the percent sulfur in the fuel. $20 per kLllowat represents a
system utilizing ,9% sulfur fuel and $UO per killowat represents a system utilizing
U.3JC sulfur fuel.
Operating costs vary from 80# per ton of fuel fired to $2.60 per
ton as a function of the sulfur content in the fuel and within the limits of
those sulfur contents described above.
VIII. Conclusions
In conclusion we can say that the FW-BF Dry Adsorption System for
Flue Gas Clean-up will:
1. Meet or exceed EPA and State codes for S02 removal
2. Significantly reduce NOX emissions
3. Significantly reduce particulate emissions
Further we can summarize the advantages of the FW-BF Dry Adsorption
System as follows:
1. No stack plume
2. No reheating of stack gases
3. Minimum water consumption
k» Low auxiliary power
5>. Minimum waste disposal
6. No handling of wet material
7. Handles coals of variable ash and sulfur contents
8. High S02 removal efficiency and the ability to deal with S02, NOX
and particulates in one system
9. Low operating costs
10. Low capital costs
We therefore say that the FV-BF Dry Adsorption System is a second generation
gas cleanup system of superior capability.
1100
-------
THE ATOMICS INTERNATIONAL MOLTEN CARBONATE
PROCESS FOR SO 2 REMOVAL FROM STACK GASES
by
W. V. Botts, Program Manager, Environmental & Utility Systems
R. D. Oldenkamp, Manager, Environmental Systems
Atomics International Division
Rockwell International Corporation
Canoga Park, California
1101
-------
THE ATOMICS INTERNATIONAL MOLTEN CARBONATE
PROCESS FOR SO^ REMOVAL FROM STACK GASES
I. INTRODUCTION
The molten carbonate process (MCP) was invented at Atomics International
in 1965, and has been under continuous development there since then. In
the first phase of the process development program, emphasis was placed
on process chemistry and materials corrosion studies and on bench scale
engineering tests. This phase was completed in 1971; the results were
encouraging, and the process was developed to the point where a pilot plant
test under actual power plant operating conditions was both warranted and
required for further progress. The initial process development work was
funded by Rockwell International Corporation; subsequent first-phase work
was done under contract to the U. S. Environmental Protection Agency.
Con Edison has undertaken to help finance the pilot plant program, in con-
junction with Northeast Utility Services Company (Hartford, Connecticut)
and Rockwell International Corporation, United Engineers and Constructors
(Philadelphia, Pennsylvania) are construction managers for the pilot plant.
The pilot plant phase of the process development program is now underway.
A pilot plant has been designed and is now being built at the Consolidated
Edison Company's Arthur Kill Power Generating Station. This pilot plant
will treat a side stream of 20, 000 scfm of flue gas from an oil-fired steam
generator (equivalent to about 10 mw of generating capacity). This paper
describes the molten carbonate process in general and the pilot plant design
in particular, and gives a status report on the pilot plant program.
1102
-------
II. PROCESS DESCRIPTION
A. BASIC PROCESS FEATURES
In the molten carbonate process, a molten eutectic mixture of lithium, sodrum,
and potassium carbonates is used to scrub the power plant gas stream. The
sulfur oxides in the gas stream react with the molten carbonates to form
sulfites and sulfates, which remain dissolved in an unreacted excess of
carbonate melt. The molten carbonate (sulfite-sulfate) mixture is then regenerated
chemically, converting the sulfite and sulfate back to carbonate and recovering
the sulfur as hydrogen sulfide. The regenerated carbonate is recirculated to
the scrubber to repeat the process cycle, and the hydrogen sulfide is converted
to elemental sulfur in a Claus Plant.
The regeneration of the carbonate is done in two steps: first the sulfite and
sulfate are reduced to sulfide, and then the sulfide is converted to carbonate
plus hydrogen sulfide. The reduction is accomplished by reaction with a form
of carbon, such as petroleum coke. The conversion of the sulfide to carbonate
is accomplished by reacting the melt with steam and carbon dioxide.
The basic process fluid is the molten- eutectic mixture of 32 wt. % Li£
33 wt. % Na2CO3, and 35 wt. % K2 CO3. This mixture melts at 747 °F
(397 °C) for form a clear, mobile, non- volatile liquid. At 800 CF, the melt
has the physical properties listed in Table I.
TABLE I
PHYSICAL PROPERTIES OF MELT AT 800°F
Viscosity 12 cp
Specific Gravity 2. 0
Specific Heat 0. 40
Thermal Conductivity ~ 0. 3 Btu'/hr. , ft. , °F (estimated)
The melt can be pumped and sprayed just like any other liquid.
1103
-------
In the process, the melt composition (and freezing point) changes as sulfur
compounds are formed and reacted. The process will be controlled so that
melts with freezing points above 850°F are avoided. This limitation corres-
ponds to melts containing about 30 wt. % sulfur compounds.
The freezing point limitation makes it necessary to keep the temperature
above 850°F in all process equipment in which melt is handled, including the
scrubber. The flue gas being treated must also be above 850*F when it enters
the scrubber. The ramifications of this requirement are discussed further in
the following sections of the paper.
B. PROCESS FLOW DIAGRAM
The basic process flow diagram is shown in Figure 1. Each step in the flow
diagram is numbered, and discussed in the corresponding section below.
1. Gas Preparation
The gas to be treated is removed from the boiler at a temperature above
850°F, or else it is reheated to above 850T. In a new plant designed for this
process and in some retrofit situations, the gas will be removed from the
boiler at the superheater or reheater outlet or economizer inlet at a nominal
temperature of 850°F. If the boiler is burning coal, the flue gas will be passed
through an 850°F, high-efficiency electrostatic precipitator where nearly all
of the fly ash will be removed. Such electrostatic precipitators with
efficiencies of greater than 99% are presently available. If hot gas cannot be
removed from the boiler, or if the heat needed to reheat the flue gas can be
used or recovered, the flue gas can be removed from the power plant air
heater (or low temperature precipitator if coal is burned) and reheated to
a nominal 850°F. The reheat can be done by direct firing with gas or oil,
or by indirect heat exchange. In any case, the gas preparation step provides
the flue gas stream at 850°F, with as much fly ash removed as possible.
2. Scrubbing
The flue gas stream then enters the scrubber, where it is contacted intimately
with the molten salt. The sulfur oxides react chemically with the molten
1104
-------
GAS TREATMENT-*-1—MELT PROCESSING AND SULFUR RECOVERY
FLUE GAS 1 CARBONATE
RETURN I MAKEUP
TO BOILER 850oF |
S CLAUS
PLANT
SULFUR
SCRUBBER
850°F
PRECIPITATOR
FLUE GAS
FROM BOILER
REGENERATOR
CARBONATE
ANDSULFIDE
QUENCH TANK
EXCHANGER
COKE FILTER
ASH FILTER
CARBONATE,
SULFITE
AND SULFATE
CAKE TO
REPROCESSING
^^
10
AIR
PETROLEUM
COKE
MELT FLOW
GASES
SOLIDS
Fig. 1. Process Flow Diagram
-------
carbonate to produce sulfur compounds and evolve carbon dioxide:
M2C03
-------
burning coal, the ash is removed by filtering the side stream of melt being
processed. If the power plant is burning oil, the particulate matter in the
flue gas is mainly carbonaceous. In this case, it should be possible to
eliminate the ash filter, since the carbon will be consumed in the reducer
and the oil ash can be removed in the coke filter (see Figure 1). The need
to filter oil particulate s out of the melt upstream of the reducer will be
determined in the pilot plant.
The ash filtration step is carried out in centrifugal basket-type filters, which
produce a continuous "dry" cake discharge. A filter aid, such as petroleum
coke particles, may be used to enhance the filtration. The filter cake is
collected and treated to recover its lithium content, using a recovery process
described below (Item 10).
4. Reduction
The melt next enters the reducer, where the sulfur compounds are reduced to
sulfide with carbon. The sulfite disproportionates to sulfate and sulfide as
the melt enters the reducer (or even before) according to the reaction:
(3)
The reduction reaction is thus actually that for sulfate only:
M2S04(I) + 2C (s) - —M2S(£) + 2C02(g) (4)
The melt temperature is raised from 850 °F to about 1600'F to enhance the
rate of reaction 4. The heat required for this, plus the endothermic heat of
reaction, is obtained by re-oxidizing some of the product sulfite to sulfate
with air:
M2S(£) + 202(g) - »-M2S04 (5)
The net result of the heat generation reaction is the sum of reactions 4 and 5:
2C(s) + 202(g) - *-2C02(g) (6)
The heat is thus actually generated by the indirect combustion of excess carbon.
1107
-------
In actual practice the reduction step will be carried out in a two-zone vessel,
using petroleum coke as the reducing agent. The coke will be consumed in
the first (reduction) zone, and air will be injected into the second (oxidation)
zone to generate heat. Internal circulation will carry heat and reactants be-
tween the two zones. This is shown schematically in Figure 2. The vessel
will be lined with corrosion resistant high-density alumina refractory, since
the melt is very corrosive at these elevated temperatures.
The reducer produces a molten mixture of carbonate and sulfide, plus carbon
dioxide. The carbon dioxide is sent to the regenerator (Step 8) where it is
used as a reactant.
5. Quenching
The reducer produces a molten mixture of carbonate and sulfide, at about
1600°F. This mixture must be cooled back to about 900°F before it under-
goes further reaction. The major part of this cooling is done by mixing the
hot melt stream with a pool of cooler melt (about 900°F) in the quench tank.
This procedure makes it possible to cool the melt to below 1000°F without
using a direct-contact heat exchanger.
The quench tank is a large tank containing a melt pump; it is also used as a
drain tank for melt storage during temporary shut-downs.
6. Coke Filtration
Cooled melt from the quench tank is filtered to remove coke ash and coke
particles carried over from the reducer. Equipment similar to that described
for ash filtration (Item 3) is used.
The filter cake from this step is combined with that from the ash filter (if one
is used) and treated to recover its lithium content.
7. Cooling
The filtered melt is then cooled further, to about 850° F, in a heat exchanger.
The cooled melt discharge stream from the heat exchanger is divided; about
90% is recycled to the quench tank to cool melt from the reducer, and the
remainder goes to the regenerator.
lino
-------
FEED MELT
FROM SCRUBBER
850°F
C02 TO
REGENERATOR
WASTE GAS
TO STACK
REDUCTION
M2SO4 + 2C-+M2S + 2CO2
1400°F,
t
OXIDATION
2O2-»M2SO4
600°F
AIR AND
COKE
REDUCED MELT TO
QUENCH TANK
AND FILTER
Fig. 2. Reducer Schematic
1109
-------
The heat exchanger is a shell and tube design, with melt on the tube side and
air or water on the shell side. Provision must be made to prevent the melt
from freezing, even under upset conditions.
8. Regeneration
In the regenerator, the sulfide-containing melt is reacted with a gaseous
mixture of carbon dioxide (from the reducer) and steam. The reaction is
M2S(I) + C02(g) + H20(g)^^lM2C03(i) + H2S(g) (7)
The reaction is exothermic, and reaches equilibrium rapidly, with the
formation of hydrogen sulfide being favored by lower temperatures. To
attain a high degree of regeneration, the reaction is carried out in a multi-
stage countercurrent gas-liquid reactor, such as a tray column.
9. Sulfur Recovery
The hydrogen sulfide from the regenerator is sent to a Glaus unit, where it
is converted to elemental sulfur. A conventional Glaus unit is suitable;
however, the feed gas stream has a high water content, so it is sent to a
condenser-cooler before undergoing the Glaus reactions. The tail gas
from the Glaus plant is incinerated and ducted to the scrubber inlet and
scrubbed along with the flue gas, so there is no Glaus plant air pollution.
1 0. Lithium Recovery
The melt used in the process is relatively inexpensive except for lithium
carbonate, which it is desirable to recover from the process filter cake.
An aqueous process has been developed for this purpose. The filter cake
is slurried with water and filtered, to extract the very soluble sodium and
potassium carbonates; lithium carbonate remains with the ash since it is
relative insoluble (1 wt. %) under these conditions. The ash-lithium carbonate
cake is then re-slurried in water and the lithium is solubilized by treating it
with carbon dioxide to convert it to the bicarbonate. The ash is removed by
filtration and the soluble lithium bicarbonate in the filtrate is precipitated
as the insoluble carbonate. The lithium carbonate is separated by filtration
and returned to the process stream; the saturated lithium carbonate filtrate
is recycled to conserve lithium. Laboratory tests have demonstrated that
over 90% of the lithium can be recovered from fly ash filter cake by this technique.
1110
-------
C. PROCESS MATERIALS AND EQUIPMENT
1. Materials of Construction
Extensive testing has been done to determine the rate of corrosion of steels
and other materials in the different process melts. The results of many tests,
including one-year dynamic tests in rotating capsules under various conditions,
have shown that the 300-series austenitic stainless -steels are suitable up to
temperatures of about lOOOT. Type 347 stainless steel is most corrosion-
resistant, followed in order by 321, 304, and 316. Type 304 was chosen for
the pilot plant, because of its availability and its well-known welding and
fabricating technology.
The corrosion resistance of the 300-series stainless steels in the carbonate
melts is due to a protective LiCrO_ film which forms a compact, tenacious,
and self-healing layer. This film forms in about 500 hours, and decreases
the corrosion rate to a few mils per year below 1000'F. However, at
temperatures of 1500 to 1600°F (corresponding to the reducer internals) the
protective layer breaks down and steels are rapidly and severely attacked.
At these temperatures, the only corrosion-resistant material available is
dense, high-purity alumina.
Corrosion testing is continuing in a forced-circulation test loop which is in
operation at Atomics International.
2. Pumps^
Three pumps are required to circulate the process melt in the flow scheme
of Figure 1. These pumps are vertical cantilever-shaft centrifugal pumps,
designed to operate with no seals or bearings in contact with the melt. The
design features of this type of pump are shown in Figure 3. One of these
pumps has been operating satisfactorily for over 3200 hours in the test loop
mentioned above.
3. Valves
Conventional stainless steel valves can be used, employing a corrosion-
resistant stem packing material such as Union Carbide's "Grafoil". Both
hand and pneumatically actuated valves have been tested successfully in the
test loop.
1111
-------
SUCTION
o
o
JISCHARGE
Fig. 3. Molten Salt Pump
1112
-------
4. Instrumentation
Process instrumentation is needed for temperature, pressure, flow and
liquid level measurement and control. Temperature measurement is done
with stainless steel or inconel-sheathed thermocouples, protected with
alumina thermowells in the reducer. Pressure and flow measurement
is done with strain gage or variable reluctance transformer sensors,
designed for use at high temperatures. Liquid level measurement is done
with conventional float displacement instruments designed for high temperature
operation. These components are currently being tested in the test loop.
5. Trace Heating
The process equipment and piping must be heated up to about 800°F for
startup, and then kept at this temperature during operation. This heating
is done electrically, with resistance heaters fastened to the vessel walls and
pipe walls, under the thermal insulation. SCR-type heater controllers are
used to maintain the proper temperatures.
1113
-------
III. PILOT PLANT PROGRAM
The process described above will be tested extensively in a pilot plant being
built at the Arthur Kill Power Station of the Consolidated Edison Company,
on Staten Island, New York. The pilot plant program has the following goals:
1) Demonstrate the integrated behavior of the process in
a real environment;
2) Provide a test facility to check out and modify equipment
or internals of equipment in order to evaluate design options
or improve operation;
3) Obtain equipment design criteria, performance, and reaction
yield data for scale-up;
4) Establish maintenance, reliability, and availability data; and
5) Provide data for an economic assessment of the process.
The pilot plant will process a side stream of stack gases from a 335 Mw
boiler at the Arthur Kill Station. This station will be burning 0. 3% S oil
during all of the test program. It is planned to vary the SO2 concentration
at the inlet to the scrubber in order to map process performance over a
wide range of conditions.
A. PILOT PLANT DESIGN
The overall size of the pilot plant was fixed as small as possible, com-
mensurate with the primary consideration that each component be large
enough to yield meaningful data for scale-up to full-scale systems. The
limiting component is the regenerator column, where an 18-in. ID tower
was determined to be the minimum size. To operate a column of this
diameter at its optimum capacity will require a throughput of 100 Ib/hr of sulfur
-------
as sulfide in the melt. This requirement then set the throughput for the reducer,
and its size. The sulfur throughput required for optimum operation will be
obtained from the flue gas and partially (up to 85%) from recycling sulfur
dioxide from the Glaus plant incinerator.
1. Flow Diagram
The flow diagram for the pilot plant is shown as Figure 4. This flow diagram
differs from that of Figure 1 only in the following features;
1) The flue gas temperature is raised to 850°F by reheat;
2) Sulfur dioxide will be intentionally recycled to the scrubber
from the Claus plant; and
3) No provision is made for on-site lithium recovery from the
filter cake.
2. Flue Gas Supply
The pilot plant treats 16, 700 scfm of flue gas from the oil-fired boiler,
after the gases have left the boiler and passed through the electrostatic
precipitator. The gases are nominally at 250°F; they are reheated to
about 850°F with an inline burner firing the same No. 6 fuel oil as the
boiler. The total quantity of gas from the boiler is increased by 3400 scfm
from the>burner, to a total flow of 20, 100 scfm (equivalent to about 10 Mw).
A blower upstream of the burner provides the pressure head to force the
gas through the ducts, burner, and scrubber. A damper is used to adjust
the flow rate to the test requirements.
The flue gases produced by the boiler will contain '—200 ppm sulfur oxides.
This concentration is very low compared to the 2000 ppm typical of gases
produced by boilers burning fuel containing 3% sulfur. In order to provide
the flexibility needed to operate over a range of sulfur oxide concentrations,
the pilot plant is designed to recycle sulfur dioxide from its Claus plant
incinerator.
-------
CLEAN
GAS TO
ATM.
t
SULFUR
FLUE
GAS
BLOWER
SCRUBBER
N.
NO. 6OIL-*
cO
M2CO3 + M2SO3
A
BURNER
M2SO3
PUMP
FILTER
REDUCER
M2CO3
M2S
PUMP
TANK
CAKE
COKE
HOPPER
QUENCH
TANK
PUMP
TANK
AIR '• ' ' CHEATER
COMPRESSOR *™ NO> 6 OIL
Fig. 4. Pilot Plant Flow Diagram
-------
The flue gases produced by burning the fuel oil contain about 0. 05 grains/scf
of particulates after passing through the electrostatic precipitator. In high
temperature electrostatic precipitators, such as would be used in a full-
scale molten carbonate process installation, the particulate removal
efficiency will be much higher. The quantity of ash to be handled in the
pilot plant, therefore, is conservatively high. It is possible that the
particulates can be burned in the reducer, thereby allowing future simplification
of the process by elimination of one filtration step.
3. Component Design
a. Gas Preparation System
The side stream of flue gas is removed from the power plant ducting through
a 4-ft duct, and passed through a blower fan and direct-fired reheater. A
recycle line also brings SC^-rich gas from the Glaus plant to the blower inlet.
The blower fan is of conventional design, with a 200 HP, 1800 RPM motor.
The reheater is a refractory-lined unit which burns the same No. 6 fuel oil
as the power plant. The burner discharges its 850°F gas stream directly into
the scrubber.
b. Scrubber System
The pilot plant scrubber is shown in Figure 5. It is a simple spray chamber,
10 ft. in diameter and 25 ft. tall. The gas enters tangentially near the
bottom, and is discharged from the top after passing through a 1 ft. thick
wire mesh mist eliminator. The gas velocity is nominally 11 ft. /sec.
The melt is sprayed in through 3 tiers of spray nozzles, and drains out
through the bottom to the scrubber pump tank. The scrubber pump has a
capacity of 33 gpm, allowing a recycle of up to 31 gpm of melt to the scrubber
inlet. Flow control is provided by pump drive speed control.
The scrubber design is flexible, allowing variation in the melt and gas flow
rates, the number, type and configuration of the spray nozzles, and the
type and thickness of the mist eliminator.
1117
-------
GAS
OUTLET
SPRAY NOZZLE
MANIFOLDS
INLETS
MIST
ELIMINATOR
VIEWPORT
GAS
INLET
MELT
OUTLET
Fig. 5. Pilot Plant Scrubber
1118
-------
c. Ash Filters
Continuous filters with dry cake discharge are desired for the process; how-
ever, development work will be required to make them available. For the
pilot plant, batch-type cartridge filters using wire-wound filter elements were
selected. Two filters are provided; one is in operation while the second is
being cleaned. A filter is shown in Figure 6. Each filter has an active area
of 160 sq. ft. , provided by 25, 2-3/4 in. diameter by 9-ft. long filter tubes
in each 2 ft. diameter shell.
d. Reduction Systems
The pilot plant reducer is shown in Figure 7. It is a steel vessel 9 ft. in
diameter and 16 ft. tall, lined with blocks of high density alumina. An inner
concentric wall of alumina blocks divides the vessel into an inner cylindrical
oxidation region and an outer annular reduction region. Holes in the inner
wall allow melt circulation between the two regions. The melt is introduced
into the reduction region from the top, and leaves the same region from the
bottom through an underflow-overflow weir. Air and petroleum coke are
introduced into the bottom of the oxidation region, and the spent gas from this
region is collected at the top and carried back to the scrubber inlet, where
it is scrubbed before entering the atmosphere. The carbon-dioxide-rich gas from
the reduction region is collected at the top in a separate manifold, cooled to
850°F by liquid water injection, and then ducted to the regenerator.
An air compressor provides the reducer air supply. The air is preheated
to 600°F in a gas-fired indirect heater, and then picks up coke particles
introduced through a star feeder.
e. Quench Tank
The pilot plant quench tank is a horizontal cylindrical tank, 6 ft. 6 in. ID
by 15 ft. long. It is of conventional stainless steel construction, and has
sufficient excess capacity to serve as a reducer drain tank during shutdown.
It is mounted in a pit directly below the reducer.
S. Coke Filters
The coke filters are identical in design and operation to the ash filters.
1119
-------
2 in. OUTLET
THERMOCOUPLE
PENETRATION
1/2 in. LOWER VENT
Fig. 6. Pilot Plant Filter
1120
-------
PREHEA
INLET
OVERFLOVy
WIER
STEEL
SHELL
SPENT GAS
OUTLET
PREHEAT
EXHAUST
INNER
WALL
REFRACTORY
LINING
REDUCTION
REGION
MELT
CIRCULATION
PORTS
AIR AND COKE
INLET
MELT
DISCHARGE
Fig. 7. Pilot Plant Reducer
1121
-------
g. Melt Cooler, The melt cooler is shown in Fig. 8. It is a 750,000
Btu/hr, air-blast type heat exchanger, with recirculation of the air stream
to provide temperature control.
h. Re g e ne rat or. The regenerator column is shown in Fig. 9. It is
a sieve-tray column, 18 in. in diameter by 36 ft. tall. It contains 15 trays
spaced 2 ft. apart. Provision has been made to vary the number and
spacing of the trays during the test program.
i. Glaus Plant. Sulfur recovery from sour gases by the Glaus process
represents commercial technology and does not require testing or demon-
stration in the pilot plant. The Glaus unit, however, had to be specially
engineered as its capacity of 100 Ib. of sulfur per hour is smaller than any
built commercially. It provides a convenient way of recovering the sulfur
removed by the pilot plant rather than burning the H S in the regenerator
off-gas to SO and returning it to the power plant stack. The Glaus plant
is shown in Fig. 10.
4. Component Arrangement
The pilot plant component arrangement plan is shown in Fig. 11. The pilot
plant is built on a 68-ft. by 100-ft. concrete slab, and has a high-bay filter
room and reducer enclosure and a control room and laboratory. The quench
tank and make-up tank are mounted in a pit, and serve as drain tanks
during shut-down. There is also a large, 40-ft. tall coke hopper to hold
the reducer coke supply.
B. PILOT PLANT STATUS REPORT
As of November 1, 1972, the pilot plant program was well underway. The
design and engineering work was completed, and all major components had
been fabricated, delivered, and set in place at the site. The steelwork
-------
Fig, 8. Melt Cooler
1123
-------
— ^
MELT INLET j[
THERMOCOUPLE NOZZLE "^
4
GAS IN LET
NOZZLE _-f
rr^'
—t
(
L
y.1
^j
V
f
to-
p1
I
O
®
©
©
@
@
0
7^
^
I
J
e
Vj>
S
i-
w— »
*» UUII.CI
1
1t
THERMAL
"~ INSULATION
6 in. THICK 10
\ 1(
ft
ft
28ft
{141
)ft
MELT OUTLET
3ft
_ 1-
FRAYS)
"\^
-*- 1 ft 6 in.
Fig. 9. Regenerator
1124
-------
NJ
in
SULFUR
CONDENSERS
WATER
CONDENSER
HEAT EXCHANGER
SNUFFING
STEAM
SULFUR
TANK
ACID
GAS
BLOWER
AIR BLOWER
C
<£%^
SALT BATH
HEATER
REACTION
FURNACE
Fig. 10. Glaus Plant
-------
EXISTING
ELECTRICAL SUBSTATION
V
o
n
8ft-
NEW 4 ft x 4 ft DUCT-
60ft-
J
TRUCK DOOR
)
GRATING
I
\
v^
t—fc
c
LT^
NK
^
\s
urriuc
AREA
LAB
/
HIGH
BAY
r
1
^L
L_U
v ;
PUMP
10^
o£
ol:
o"-
u
u
^HMH^^W
F
\ f
©
SCRUBBEI
~*T
tEHEATER
• j
-, U
REGENE
BLOWER
1
i
CLAUS
PLANT
ORATOR
QUENCH TANK
PIT /
~» D(
_ /~*
Df
. m
Vp- REDUCER
1001 ^"" 1
\
\
COKE
HOPPER
TRUCK /^v
DOOR f J
68ft
0
O BLOWI
D
1
=R
S>- CONCRETE PAD 20 ft
EXISTING
FLUE DUCT
."NEW SUPPLY
DUCT TIE-IN
20ft
ASPHALT
100ft
CONCRETE PAD
SITE FENCE -
25ft
ASPHALT
Fig. 11. Pilot Plant Equipment Arrangement Plan
1126
-------
for the platforms, supports, and building was being installed, and the refractory
lining of the reducer vessel was about to start. Electrical heaters were also
being mounted on tanks and vessels. All of the major subcontracts had been let,
and construction was scheduled to be completed by February 1, 1972.
Photographs showing the construction work in progress are presented in
Figures 12 to 15 inclusive.
C. TEST PLAN
The test plan covers a one-year test program. The schedule for this program
is shown in Figure 16. The test period is divided up into four phases: 1) Start-
up and Stabilization, 2) Design Verification, 3) System and Component Parametric
Studies, and 4) Reliability and Operability Studies.
The startup and stabilization phase will require about 4. 5 months to complete
and will consist of a thermal test, a hydraulic simulation, fill and circulation
of the molten carbonate, and activation and stabilization of the process
chemical reactions.
The second phase of the test program will be the design verification, which
will consist of recycling SO£ from the Glaus plant to the scrubber and operating
at the design sulfur throughput of the pilot plant. About two months is scheduled
for this phase, the results of which will be used to verify the design point of
the pilot plant and establish system and component performance characteristics
and operating cost data at the design conditions.
The third phase will involve system and component parametric studies. It is
scheduled for 3. 5 months. This will consist of general parameter studies of
the performance of the system as a whole and of each major component. These
studies will establish performance characteristics over a range of operating
conditions, maximize and establish capabilities of the system and individual
components, and establish design and operating criteria for a large demonstration
plant.
The fourth and last phase of the currently planned test program will be reliability
and operability studies. Reliability and operability of the pilot plant will, of
course, be under study throughout the one-year test program. This phase will
be a special two-month test at the end of the year to determine minimum
operator supervision and maintenance requirements.
1127
-------
Fig. 12. Pilot Plant Foundation Under Construction
-------
(SJ
10
Fig. 13. Glaus Plant
-------
co
o
Fig. 14. Pilot Plant Scrubber Being Installed
-------
Fig. 15. Pilot Plant Equipment Being Installed
-------
CO
NJ
I 1 I 2 I 3 I 4 1
Months
6 j 7 i 8 [ 9
10
11
12
STARTUP & STABILIZATION
DESIGN
VERIFICATION
, SYSTEM & COMPONENT
PARAMETER STUDIES
RELIABILITY/
OPERABILITY
STUDIES
THERMAL
TEST
3 WEEKS
HYDRAULIC
SIMULATION
4 WEEKS
C03 FILL fc
CIRCULATION
4 WEEKS
PROCESS ACTIVATION
AND STABILIZATION
8 WEEKS
Fig. 16. MCP Pilot Plant Test Plan and Schedule
-------
SUMMARY OF FLUE GAS
DESULFURIZATION SYMPOSIUM
1133
-------
SYMPOSIUM SUMMARY
E. L. Plyer (EPA)—.The last part of our symposium is the
summary. As you can imagine this has to be a very awesome and formidable
job and I think we have a very capable person in Frank Princiotta who
has agreed to do this. I would like to make a few remarks concerning
this. Frank, as you might realize, has had to put this-together sort of
on the run here. So even in summarizing something like a symposium of
this type, there's bound to be some subjectivity that comes in. And
I'm sure that there will be a little bit of it here. You might not
agree with everything Frank is going to say here. And what Frank has to
say I don't think you can take as being EPA policy. But we do think
that it would be quite helpful to you to get a summary of the
highlights of the symposium. Necessarily, he will not be able to
cover, even briefly, everything that has been given here. And he will have
to concentrate his remarks, I think, on the large-scale demonstrations and
some of the commercial units which we have tried to concentrate on in
our symposium. Since Frank has already been introduced as the session
chairman for the lime/limestone wet scrubbing processes, I don't think
I need to repeat any biographical information on him. I will just
mention that he is the Chief of the Engineering Test Section of the
Control Systems Laboratory. He's had about 2 years experience there
and due to his enthusiasm and drive he has become, I think, one of
our real experts in flue gas cleaning. I think he will do a good
job in summarizing our symposium for us.
1134
-------
Frank Princiotta (EPA)—Thanks, Bill, for the introduction
and disclaimer. I can't get in trouble now. As Bill indicated,
keeping in mind the difficulty of summarizing a symposium of the
comprehensiveness that took 4 days in half an hour. I'm bound to
miss some important highlights, so please bear with me. First of
all, we've heard that sulfur oxides are among the most dangerous
air pollutants to human health. Dr. Newill described how morbidity
and mortality can be correlated to ambient SO^ levels. He
also pointed out some relatively new information that can have
important repercussions. Namely, that particulate sulfates appear
to be even more dangerous to human health than S02 alone. He stated
that in cities where the primary ambient levels of S02 are met, high
daily sulfate levels have been associated with aggravation of symptoms
of heart and lung disease patients. Although further data are necessary,
the sulfate particulate problem could lead to more stringent ambient
SOX levels and therefore SOX emission control. Time will tell.
I think Dick Harrington (EPA) had some relevant
ooints about the importance of controlling industrial boilers.
He indicated that despite the fact that area sources and industrial
boilers represent a relatively small percentage of total SOY emitted,
/\
they contribute due to meteorological factors. He stressed the
importance of trying to utilize stack gas technology developed for
power plants on these smaller sources. It certainly seems like a
1135
-------
reasonable suggestion and we're hopeful that systems like double
alkali, lime scrubbing, and others could fulfill this role.
Steve Gage (Council on Environmental Quality) presented
a very interesting paper on the alternatives to flue gas
cleaning—the technological alternatives. And really, in summary,
there are none in the short term. He reviewed coal liquefaction,
gasification, and advanced combustion processes. He indicated
that these are in an embryonic stage with years of development
necessary. Probably they will make an impact somewhere between
the 1980 and 1983 time period. All of these processes, though,
are exoensive and, as most of you know, as you approach true
hardware, they tend to get more expensive.
Therefore, flue gas desulfurization is the only alternative to
flue switching now and will remain competitive for many years in the
future, particularly on retrofit applications and in the important
industrial boiler size applications.
We then heard from Dr. Ando (J. Ando, Chuo University) who
presented the Japanese situation and I think it's of great relevance
to the U.S. situation. They have a severe air pollution problem
aggravated by a high population density and concentrated industrialization,
The great majority of their power plants are oil-fired and, therefore,
there are differences that have to be remembered. In the past, S0y
A
regulations in Japan were met primarily by hydrodesulfurization of
1136
-------
fuel oil. They import high sulfur oil and then desulfurize it.
However, the recent tighter regulations, equivalent to less than
1 percent sulfur flue gas desulfurization, have led to plans for
at least 100 air pollution control units on power plants, Glaus
plants, sulfuric acid gas treatment, industrial boilers, etc.
Most of these units produce sulfur products and are either presently
installed or will be installed in the near future. Sulfur products
produced by these systems in Japan include sodium sulfite, sulfuric
acid, sulfur, and (most importantly in recent months) gypsum. However,
Dr. Ando points out that by 1976 he expects that significantly more
gypsum will be produced than can be used; so they, too, may start worrying
about throwaway processes. The most important Japanese systems appear
to be: the Mitsui lime scrubbing unit - I will mention it briefly
later; the Mitsubishi Heavy Industry-JECCO lime scrubbing process, which
this unit has operated with high reliability for 9 months on a 35-Mw, oil-
fired, closed-loop unit; and the Mitsubishi Chemical Machinery - Wellman
Lord process, which has operated reliably for 1.5 years on a 75-Mw,
oil-fired power plant. Also, there are three double-alkali systems
under active development in Japan: the Kureha/Kawasaki and Showa Denko
sodium double-alkali systems, and the Nippon Kokan ammonium double-
alkali system. These are considered to be very important and should
be followed closely.
1137
-------
Moving on to costs, I think Gary Rochelle (EPA) presented a very
useful scheme for the rough estimation of capital and operating costs
for flue gas desulfurization systems. Briefly, the following are
some of the operating costs he came up with. (I will just talk about
total annualized operating costs that are perhaps more important be-
cause they include capital charge considerations.) For lime/limestone
and magnesium oxide processes, he found costs for a new 500-Mw, 3.5-
percent sulfur coal unit to be very close for both processes at about
2.5 mills/kw-hr. Costs for the Wellman Lord system and the Stone and
Webster/Ionics System under the same groundrules were about 2.7 mills/
kw-hr. Operating cost for the Cat-Ox system was about 2.75 mills/kw-
hr. He indicated that about 75 percent of existing plants can be retro-
fitted within 3 mills/kw-hr. Comparisons of this estimating technique
with actual costs for six full scale installations, most of them first
of a kind, indicate that he's about 9-21 percent low in his capital cost
estimates, ffy own feeling is that this is due primarily to the fact
they are first of a kind and that these estimates will come even more in
line as there's more widespread application of these systems. However,
there were discrepancies, You heard them and I heard them. The Will
County unit and the Widow's Creek unit in particular have had recent
cost escalations which make Gary's estimates look even a little bit
farther off.
1138
-------
Most of these systems, reqenerable and throwaway, look pretty
similar to Gary as far as operating costs are concerned and he feels
the key parameter which can turn the tide is sludge disposal costs.
And he feels these operating costs are equal at $3 or $4 per ton of
wet sludge. Additionally, interestingly, he also indicates what's
obvious I guess when one thinks about it: the low load factor systems
strongly favor the throwaway processes economically. And this has
obvious implications for industrial boilers which ordinarily operate
at low load factors.
Getting into lime/limestone scrubbing systems, I think Irv Raben
(Bechtel Corporation) presented an excellent paper and I recommend that
everybody review it carefully. He pointed out that 21 full—scale units,
now comprising 9600 Mw of capability, will be or have been installed in
this country. In particular the Ohio Edison Bruce Mansfield, the Mohave,
Navajo, and Northern States units are the real biggies that contribute
most to that 9600-Mw figure. His costs were in line with Gary's costs,
and he estimates for a hew 500-Mw unit on a similar basis for Gary,
about 2.3 to 2.5 mills/kw-hr. Irv feels relatively confident, hopeful
I guess 1s the word, that reliability will be demonstrated soon in the
United States based on all the units that are going to come on to line
relatively soon.
1139
-------
Talking about some limestone systems in particular, the Chemico
lime-Mitsui aluminum plant has to be mentioned. Of coarse this
paper was presented by Mr. Sakanishi (Mitsui Aluminum Co.). Bob Quig
(Chemico) made perhaps a controversial introductory statement.
when he expressed his belief that the operation to date has indicated
this technology is now demonstrated by anyone's definition, including
the National Academy of Sciences. As you know this system has removed
90 percent of the S02 on a coal-fired plant for 14 months with reliable
operation; no problems of any significance were reported. Briefly
mentioning, though, there were people on our panel, assembled to discuss
the relevance of this system to the U.S., who felt that they have to be
careful in trying to extrapolate this experience to the U.S., primarily
because the system apparently operated open-loop for at least part of
its time in Japan. It's a base-loaded unit with few excursions, and it's also
a 2 percent sulfur coal system, which is somewhat lower than many eastern
U.S. applications. Other panel members, however, felt that it is a very
relevant system and has many characteristics in common with many potential
eastern U. S. coal applications.
Moving on to Dr. Weir's (A. Weir, Jr,, Southern California Edison)
paper on the Mohave power plant results. This ties in to full-scale units.
These results, like those at Shawnee and other places, indicate that the
lower the S02 inlet in the flue gas, the easier things are as far as
1140
-------
removal is concerned. He was able to get down to 20 ppm and less in
outlet for many of the eight scrubbers he looked at under many
conditions. He was also able to get 0.001 to 0.002 gr/scf outlet
for several of his scrubbers (keep in mind that this is downstream
from a relatively efficient precipitator). As you know this information
(and perhaps data from two big modules which will be built
at Mohave) will influence the selection of the very large Mohave and
Navajo plants. They add up, by the way, to over 3000 Mw of capability.
So that's a real slug of capability.
Talking a little about the B&W systems, the Will County unit was
described to us. It initially was a 175-Mw unit designed with a limestone
tail end system for 80 percent removal. It started up about a year ago and
has had many operating problems, primarily minor problems with apparently
no inherent system-type problems. Perhaps the most serious of the
problems has been demister pluggage. However, there
appears to be relatively high probability in the near term that thts
system will be made to operate reliably., The next B&W unit we heard
about was the Kansas City P&L La Cygne unit, which is kind of a version of
the Commonwealth unit multiplied by about 7. There are seven
parallel circuits in this 840-Mw system, which just recently started up.
I think the boilers are now at about half load and I'm personally going
to keep an eye out for that system. I suspect the unit could be
very reliable since I think many of the problems they had at Will
County have been corrected on that particular unit.
1141
-------
I think we should mention some of the Combustion Engineering
units as well. But I think we should mention the Kansas City Power &
Light units briefly. They have been changed to a tail end
configuration; one of the two Hawthorne units (Unit No. 3) has been
changed from boiler injection to a tail end limestone system. I think
Irv Raben indicated that he felt this was a trend in the industry; I
think this change is indicative of this trend. There has not been
sufficient operating time on this tail end limestone Hawthorne unit
to get much information yet. Our understanding is that they've been
burning some low-sulfur coal lately. The Louisville Gas and Electric
Paddy's Run Station, a 70-Mw carbide sludge tail end system, though,
has started up. It has about 1 month's operation now. And, as you
heard, things look pretty good. Reliability so far looks promising.
502 removal nas been reported as very high, although I haven't heard
a number yet.
Let's move on briefly now to the EPA Shawnee lime/limestone program.
We heard from Dr. Epstein (M. Epstein, Bechtel Corporation) about all
the SOg data that's been generated as a function of system parameters;
this should certainly help design these systems in the future. There's
information also on particulate removal; its range is from about 0.01 to
0.03 gr/scf for all three scrubbers. You might remember these are three
10-Mw scrubbers in parallel to test three different types of scrubbers.
-------
I'll summarize reliability information and perhaps even add a
bit to what Bill Elder (TVA, Muscle Shoals) said. This
is hot off the press. In general, reliability tests started about
2 months ago. And we are very pleased so far with the reliability
seen to date. We have had little evidence of scaling or demister
problems. We feel the trick is low pH operation, high L/G's, and
backing off somewhat on gas velocity. The venturi after-spray scrubber
has operated for 1 month at 75 percent removal without any
problems of any type. The TCA has also operated for about a month with
from 85 to 90 percent removal; no demister or scaling problems occurred,
although we have seen some erosion of balls and grids and some
solids buildup in the inlet duct. We feel confident these problems
will be resolved shortly. The hydrofilter had very few problems over
about a month's tests until lately when the nozzles apparently have
eroded and led to some bed and demister problems.
Moving on now to the first of the regenerate systems or saleable
product systems. We heard quite a bit about the magnesia oxide process.
Gerry McGlamery (TVA) described the system in detail. He
compared the costs to those of wet limestone systems and found them relatively
close although the magnesia oxide process was perhaps somewhat more
expensive; again, sludge costs were very important for any cost
comparison. He also indicated that magnesium oxide systems mak-
ing sulfuric acid makes most sense in metropolitan areas where
sludge is difficult to handle and sulfuric acid can be marketed.
-------
Of course the most important Mag-Ox system is the Boston
Edison system. This unit started up in April 1972 at Boston
Edison's Mystic Station, a 150-Mw facility producing 50 tons per day of
sulfuric acid, which is processed at a separate location in Rumford,
Rhode Island. Operations to date have demonstrated each process step.
Greater than 90 percent S0« removal has been achieved. But, as you
heard, there have been several "nitty gritty" mechanical problems
which have prohibited a long-term test which is really needed to check
out the system. Such problems have included dryer problems, calciner
seal leakage, calcined material loss of reactivity, solids handling
problems, and so forth. And Chemico has indicated that top priority
is going to be given to performance of some long-term testing relatively soon
to really home in on the reliability of this important system.
Other Mag-Ox systems were mentioned. The Philadelphia Electric
Eddystone system designed by United Engineers is a 120-Mw system that
will start up in the fall of 1973. And the PEPCO Dickerson No. 3
coal-fired station, a 100-Mw station, will start up in several months.
Perhaps the most reliable of the regenerable systems,
in terms of operation to date, is the Wellman Lord process of Davy Powergas.
This process, at least based on what we heard today, has demonstrated the
greatest reliability, I think it's safe to say, of any flue gas desulfuri-
zation system to date. There are five units presently operating reliably;
-------
eight are under construction on power plants, sulfuric acid treatment
systems, and Claus treatment systems. Perhaps the most significant
is Japan Synthetic Rubber's Chiba plant, the 75-Mw, oil-fired system
that has operated for over 1.5 years at a reported availability of 95 percent.
Sulfur dioxide removal has been at about 90 percent. EPA is quite
interested in the process and has co-funded with Northern Indiana
Public Service Co. for installation of the Wellman unit at the NIPSCO
station, a 115-Mw system. It's a coal-fired boiler that will produce
sulfur directly using the Allied SOg reduction process which uses nat-
ural gas as a reductant. This sulfur-production technology was success-
fully demonstrated at the Falconbridge Nickel Plant near Sudbury, Canada.
It should be noted, though, that for about a 1000-Mw system you need
about 20 Mw worth of natural gas. Of course this is a problem in many
locations; however, Allied indicated they are working very intensely for
substitutes for natural gas as the reductant. Another problem attendant
to that system is the requirement for a sulfate purge based on a 4,5,
or 6 percent oxidation of sulfute/bisulfate to sulfate.
The Monsanto Cat-Ox process should be mentioned. This is another
regenerate process somewhat unique in that it's based on the well-known
contact sulfuric acid process. Flue gas is heated to 800°F and oxidized
in the catalyst bed to 850°F where the sulfur dioxide is oxidized to sul-
fur trioxide to produce about 78 percent sulfuric acid. You heard the
story about the Wood River station of Illinois Power. This is where
the system has been installed; the system started up back in September
of 1972. I should add that a 99.6 percent efficient ESP, which is
1145
-------
necessary upstream of the system, was installed in February
of 1972 and has been operating well since then. Initially the system
worked well when it was using natural gas as the reheating combustion
medium; however, after about 17 days of successful operation, reheater
problems with oil associated with soot buildup (with potential bed
pluggage problems) shut down the system. As you heard, there will be
modifications and it's expected the unit will be back on line in late
summer of this year. One of the problems with the system is the sale
of dilute sulfuric acid.
Other processes that appear to be very well along in their develop-
ment and very important are the double alkali processes. I was pleased to
hear the summary by Norm Kaplan and Dean Draemel (EPA) of all the varia-
tions of this process, and there are quite a few of them. And I think
it's wort,h everyone's while, who's interested in these systems, to care-
fully look at a summary of these systems and see just what these differences
are — dilute versus concentrated, various ways of handling the sodium sul-
fate, etc. But they all seemed to have the potential for low capital and
operating costs and for high reliability. Gary Rochelle indicated that
the costs were approximately 15-20 percent lower than an equivalent lime/
limestone system. This was based on an assumption which will be validated.
1146
-------
Norm Kaplan mentioned some of the corporations involved ~ FMC, Envirotec,
General Motors, A. D. Little, Showa Denko, and Kureha. Let me just
mention some of the more significant full-scale units that are planned.
The 20-Mw unit at Southern Services Scholz station of Gulf
Power is of real significance; this will be an A. D. Little/Combustion
Equipment Associates unit. General Motors' 40-Mw double alkali system
for four stoker boilers at the Cleveland Chevrolet plant will
start up late this year; this will be the first of the U.S. full-scale
installations. In addition, there are full-scale units planned in
Japan, one of which will start up in June, the Showa Denko 110-Mw,
system on an oil-fired boiler. Two additional 150-Mw units will begin opera-
tion in 1974. There is quite a bit of activity in the double alkali
area and it looks like it will be a real viable alternative to lime/
limestone scrubbing as far as the throwaway processes are concerned.
Let me just briefly mention other important processes
which were mentioned today. I'm sure they are probably pretty fresh in
your mind. The Foster Wheeler-Bergbau Forschung process, TVA-EPA
ammonium bisulfate, and Stone & Webster/Ionics molten carbonate processes
were described and could play an important role several years
from now when these systems are able to be applied to full-scale commercial
units.
Let me briefly summarize some of the things we heard about sulfur
product problems. One can easily write a book on this, so I'll try to
be brief. The large quantities of sulfur
1147
-------
inherent in flue gas must end up as a low-value sulfur product.
Generally this is either a throwaway sludge, which you can call a
zero value, or negative value material as opposed to a low-value
sulfur or sulfuric acid type of material. Quantities are somewhat
staggering. Dick Stern and Julian Jones (EPA) indicated that a
100,000-Mw capacity for 20 years of sludge storage to
a depth of 10 feet would require 150 square miles of area for settled
sludge. They also made some very provocative comparisons between the
affected areas associated with sludge based on predictions of lime/
limestone,utilization and areas associated with coal strip mining.
For sulfuric acid production, the same 100,000-Mw capacity will
produce acid at a rate equivalent to the present total U.S. production,
about 28 million tons per year. Sulfur, as far as volume and quantities
are concerned, is certainly the most desirable end product.
The EPA/Aerospace program will hopefully present a complete, com-
prehensive evaluation of the toxicity and water pollution problems
associated with throwaway sludges, and will also evaluate some sludge
treatment processes. We'll look at lime/limestone and double alkali
sludges, and Dick Stern has asked all relevent parties for information.
Also to be evaluated are the sludge fixation (treatment) processes.
We've heard quite a few of them mentioned; IUCS and Dravo perhaps being
the most important processes. These fixation processes basically
1148
-------
involve a recipe requiring the inclusion of flyash. The recipe
calls for flyash, calcium sulfite and sulfate sludge, additional
lime-based material, and perhaps an accelerator to make a cementitious
type material which might have attractive landfill application. And
perhaps, at least in lUCS's opinion, a potential for sale as aggregate
and cementitious materials. Preliminarily, data presented by IUCS indicate
that these processes lead to decreased Teachability and permeability and
therefore reduce water pollution potential. We hope to validate
such claims during our Aerospace program.
Let me mention quickly a little bit about sulfuric acid. This has
been the product of choice for the regenerable systems so far in this
country. But it appears that only selective applications will be possible
due to difficulty in selling large quantities of sulfuric acid. Each
1000 Mw produces approximately 1 percent of the total U.S. production.
However, there seem to be obvious urban applications where such systems
make quite a bit of sense.
Let me briefly discuss sulfur as the end product. As
I mentioned, the storage volumes are certainly attractive; there's no
question about that* Sulfur can also be stored for eventual use as opposed to
sludges, which must just lie there with no potential use except as
landfill. However, some problems associated with the storage of sulfur
were mentioned in this symposium which were new at least to me.
1149
-------
They include potential flammabi.lity problems, potential
HgS evolution, and perhaps wind and water erosion and other related
problems. Hopefully this will be evaluated because sulfur certainly
seems (at least preliminarily) to be a very important end product.
Also we heard some potential uses of sulfur. They include sulfur in
asphalt materials, sulfur concretes, etc., although it should be pointed
out there's quite a bit of technical and marketing work ahead before
these use any great quantities of sulfur material.
As far as the symposium conclusion is concerned, I would like to
be a little bit subjective now. I have tried to be objective during
this presentation although somebody may disagree with that. Let me now
quote from the SOCTAP report. You might be familiar with this. This
is the Sulfur Oxide Control Technology Assessment Panel, a
governmental interagency group which reported to Mr. Ruckelshaus
(W.D. Ruckelshaus, former EPA Administrator). Its charter was basically
to look into flue gas desulfurization and come up with a reasonable, objective
assessment. This report is available, I might add from the Air Pollu-
tion Technical Information Center, Research Triangle Park, N.C. Let
me just read briefly the most relevant conclusions. Now keep in mind,
these are the conclusions of the report that I feel accuarately reflect
the sum total of what we have heard at this symposium:
"We (the SOCTAP group) have examined the status of stack gas cleaning
technology in the United States and Japan and have concluded that
sulfur dioxide removal from stack gases is technologically feasible
in commercial-sized installations. We have concluded that technologi-
cal feasibility should not now be considered a decisive element in
utilization of these systems and that a large fraction of the nation's
coal-fired steam electric plants can ultimately be fitted with com-
mercially available stack gas cleaning systems... The reliability
of currently available systems has been the subject of some question.
1150
-------
We concur that SOX control systems must exhibit the high degree of
reliability required by the utility industry. We believe that the
required reliability will be achieved with early resolution of a
number of applications engineering problems related to specific hard-
ware components and system design parameters.... In view of the
fact that a number of large scale plants scheduled for operation in
the U.S. in the near future will provide additional 18 months opera-
ting experience (or by 1974) should effectively remove engineering
barriers to the application of stack gas cleaning to many facilities1.'
The report emphasized the need for additional R & D primarily in ad-
vanced processes to cut back costs and help produce less noxious solid waste
problems and for a solid waste sludge disposal evaluation program. Thank you,
1151
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing}
REPORT NO.
EPA-650/2-73-038
2.
3. RECIPIENT'S ACCESS!ON>NO.
4. TITLE AND SUBTITLE
Proceedings, Flue Gas Desulfurization Symposium-
1973
5. REPORT DATE
December 1973
6. PERFORMING ORGANIZATION CODE
. AUTHOR(S)
Miscellaneous
8. PERFORMING ORGANIZATION REPORT NO.
10. PROGRAM ELEMENT NO.
1AB013
. PERFORMING ORGANIZATION NAME AND ADDRESS
Miscellaneous
11. CONTRACT/GRANT NO.
ROAP 21ACY-30
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
NERC-RTP, Control Systems Laboratory
Research Triangle Park, N. C. 27711
13. TYPE OF REPORT AND PERIOD COVERED
Proceedings •
14. SPONSORING AGENCY CODE
16. SUPPLEMENTARY NOTES
. ABSTRACT
Tne proceedings document the 30 presentations made during the symposium,
attended by 430 representatives of electric utilities , pollution control system supp-
liers , engineering and construction firms , state and local pollution control agencies ,
and EPA Headquarters, Regional Offices, and NERCs. The objective was to present
the current status of flue gas desulfurization technology for full-scale power plants ,
the only near-term alternative to the use of low-sulfur fuels in meeting air quality
standards. It emphasized lime/limestone scrubbing, magnesia scrubbing, sodium-
based scrubbing with thermal regeneration, and catalytic oxidation. Disposal and
uses of SOx control process by-products and the second-generation or advanced
SOx control processes were also discussed. The symposium filled the need for up-
to-date information in support of federal, state, and local air pollution control
activities.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
e. COSATI Field/Group
Air Pollution
Flue Gases
Desulfurization
Scrubbers
Sludge Disposal
Economic Analysis
Marketing
Sulfur
Sulfur Oxides
Electric Power PlantsjSodium
Calcium Oxides
Limestone
18. DISTRIBUTION STATEMENT
Unlimited
Air Pollution Control
Stationary Sources
Utility Boilers
-Based Scrubbing
Catalytic Oxidation
Molten Carbonate
orption
19. SECURITY CLASS (ThisReport}
Unclassified
13B, 14A, 2IB
21. NO. OF PAGES
1161
20. SECURITV CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (»-73)
1153
-------
INSTRUCTIONS
1. REPORT NUMBER
Insert the EPA report number as it appears on the cover of the publication.
2. LEAVE BLANK
3. RECIPIENTS ACCESSION NUMBER
Reserved for use by each report recipient.
4. TITLE AND SUBTITLE
Title should indicate clearly and briefly the subject coverage of the report, and be displayed prominently. Set subtitle, if used, in smauer
type or otherwise subordinate it to main title, when a report is prepared in more than one volume, repeat the primary title, add volume
number and include subtitle for the specific title.
5. REPORT DATE
Each report shall carry a date indicating at least month and year. Indicate the basis on which it was selected (e.g., date of issue, date of
approvcl, date of preparation, etc.).
6. PERFORMING ORGANIZATION CODE
Leave blank.
7. AUTHOR(S)
Give name(s) in conventional order (John R. Doe, J. Robert Doe, etc.). List author's affiliation if it differs from the performing organi-
zation.
8. PERFORMING ORGANIZATION REPORT NUMBER
Insert if performing organization wishes to assign this number.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Give name, street, city, state, and ZIP code. List no more than two levels of an organizational hirearchy.
10. PROGRAM ELEMENT NUMBER
Use the program element number under which the report was prepared. Subordinate numbers may be included in parentheses.
11. CONTRACT/GRANT NUMBER
Insert contract or grant number under which report was prepared.
12. SPONSORING AGENCY NAME AND ADDRESS
Include ZIP code.
13. TYPE OF REPORT AND PERIOD COVERED
Indicate interim final, etc., and if applicable, dates covered.
14. SPONSORING AGENCY CODE
Leave blank.
15. SUPPLEMENTARY NOTES
Enter information not included elsewhere but useful, such as: Prepared in cooperation with, Translation of, Presented at conference of,
To be published in. Supersedes, Supplements, etc.
16. ABSTRACT
Include a brief (200 words or leaf factual summary of the most significant information contained in the report. If the report contains a
significant bibliography or literature survey, mention it here.
17. KEY WORDS AND DOCUMENT ANALYSIS
(a) DESCRIPTORS - Select from the Thesaurus of Engineering and Scientific Terms the proper authorized terms that identify the major
concept of the research and are sufficiently specific and precise to be used as index entries for cataloging.
(b) IDENTIFIERS AND OPEN-ENDED TERMS - Use identifiers for project names, code names, equipment designators, etc. Use open-
ended terms written in descriptor form for those subjects for which no descriptor exists.
(c) COSATI FIELD GROUP - Field and group assignments are to be taken from the 1965 COSATI Subject Category List. Since the ma-
jority of documents are multidisciplinary in nature, the Primary Field/Group assignment(s) will be specific discipline, area of human
endeavor, or type of physical object. The application(s) will be cross-referenced with secondary Field/Group assignments that will follow
the primary posting(s).
18. DISTRIBUTION STATEMENT
Denote reusability to the public or limitation for reasons other than security for example "Release Unlimited." Cite any availability to
the public, with address and price.
19.8.20. SECURITY CLASSIFICATION
DO NOT submit classified reports to the National Technical Information service.
21. NUMBER OF PAGES
Insert the total number of pages, including this one and unnumbered pages, but exclude distribution list, if any.
22. PRICE
Insert the price set by the National Technical Information Service or the Government Printing Office, if known.
EPA Form 2220-1 (9.73) (RBVim) 115U *U.S. G.P.O. ; 1974—747-791/340, Region No. 4
------- |