5/2-76-010
THIRD PROGRESS REPORT-
EPA
TECHNOLOGY
TRANSFER
LIME/LIMESTONE
WET-SCRUBBING
TEST RESULTS
AT THE
EPA ALKALI
SCRUBBING
TEST FACILITY
US. EPA
OFFICE OF
RESEARCH AND
DEVELOPMENT
PROTOTYPE
DEMONSTRATION
FACILITY
-------
THIRD PROGRESS REPORT:
EPA
TECHNOLOGY
TRANSFER
LIME/LIMESTONE
WET-SCRUBBING
TEST RESULTS
AT THE
EPA ALKALI
SCRUBBING
TEST FACILITY
U.S. EPA
OFFICE OF
RESEARCH AND
DEVELOPMENT
PROTOTYPE
DEMONSTRATION
FACILITY
EPA-625/2-76-010
-------
Wet Scrubbing Test Facility with Shawnee Power Station in Background
-------
This is the third of a series of capsule reports de-
scribing a program conducted by the Environmen-
tal Protection Agency (EPA) to test prototype lime
and limestone wet-scrubbing systems for removing
sulfur dioxide (SO2) and particulate matter (fly
ash) from coal-fired boiler flue gases. The program
is being conducted in a test facility which is inte-
grated into the flue gas ductwork of a 160 MW
coal-fired boiler at the Tennessee Valley Authority
(TVA) Shawnee Power Station, Paducah, Ken-
tucky. Bechtel Corporation of San Francisco is the
major contractor and test director, and TVA is the
constructor and facility operator. This report de-
scribes the initial results of an advanced program of
lime and limestone tests, with and without the
addition of magnesium oxide, conducted from
January 1975 to April 1976. Earlier capsule re-
ports described the results from the inception of
testing in March 1972 until January 1975.
In a lime or limestone wet-scrubbing system, the
flue gas is contacted (scrubbed) with a slurry of
lime or limestone. SO2 is absorbed into the liquor,
where it reacts with the dissolved alkali, precipitat-
ing waste products of calcium sulfite and calcium
sulfate (gypsum). Particulates are removed in the
scrubber by impact with the slurry droplets.
At the Shawnee Test Facility, two parallel wet-
scrubber systems are in operation: a venturi/spray
tower system and a Turbulent Contact Absorber
(TCA) system. Each system is capable of treating
approximately 10 MW equivalent (up to 35,000
acfm at 300°F) of flue gas containing 1,500 to
4,500 ppm of SC»2 and 2 to 5 grains/scf of particu-
lates. The following tests were conducted during
this part of the advanced test program:
• Limestone utilization tests on both scrubber
systems to determine the effects of scrubber
inlet liquor pH, effluent hold tank residence
time, and hold tank design on limestone utili-
zation (moles SO2 absorbed/mole Ca added)
• Lime and limestone reliability tests on both
scrubbers to define methods of operating with
a mist eliminator free of scale and soft solids
• Lime variable-load tests on the venturi/spray
tower to evaluate this system's ability to
handle a variable gas rate and composition
resulting from daily boiler load cycles
• Short (6 to 8 hours each) factorial tests using
lime, limestone, and limestone with magne-
sium oxide addition in both the venturi/spray
tower and TCA systems to observe the inter-
action of the systems' major variables
Limestone utilization test results showed that it
is easier to prevent soft solids deposits in the mist
eliminator when operating at high alkali utilization
(moles SO2 removed/mole Ca added). Above
about 85 percent utilization, an intermittent un-
derside and topside wash was sufficient to keep the
mist eliminator clean. At lower utilization, a con-
tinuous underside wash combined with intermit-
tent topside wash was required. High limestone
utilization was achieved by reducing the scrubber
inlet liquor pH, but operation at low pH also re-
duced percent SO2 removal. The use of three
stirred hold tanks in series (instead of a single efflu-
ent hold tank) improved limestone utilization by
5 to 10 percent, with no loss in SC»2 removal. High
limestone utilization combined with high SO2 re-
moval was achieved by increasing the magnesium
ion concentration in the scrubber slurry liquor.
During variable-load tests using lime in the ven-
turi/spray tower system, the gas rate to the scrub-
ber was varied hourly to follow the normal boiler
load cycle of 60 to 160 MW. These tests demon-
strated reliable operation with good control for
over 1,000 operating hours, with the gas rate vary-
ing between 17,000 and 35,000 acfm and with
inlet SC>2 concentration varying between 1,500
and 4,400 ppm. The mist eliminator remained
clean throughout the tests.
The factorial tests provided, over a wide range
of variables, steady-state data useful for design and
for mathematical modeling.
-------
GASiOUT
Chevron Mist
Eliminator
SPRAY TOWER
MIST
ELIMINATOR
WASH WATER
INLET '
SLURRY 4
Throat
Adjustable
Plug
VENTURI
SCRUBBER
OUT
INLET SLURRY
APPROX. SCALE
EFFLUENT SLURRY
Figure 7. Schematic of VenturifSpray Tower
MIST
ELIMINATOR
WASH WATER
Chevron Mist '
Eliminator
Retaining
Bar-Grids
°o
g °0°0°
L°_Q_ O _
GAS IN
MIST
ELIMINATOR
LIQUOR
INLET SLURRY
Mobile Packing
Spheres
5'
APPROX. SCALE
EFFLUENT ''SLURRY
•N^,
Figure 2. Schematic of Three-Bed TCA
Two parallel scrubber systems — a venturi/spray
tower system and a TCA system — are operated at
the test facility. Each has its own slurry-handling
system. The scrubbers are of prototype size and
are each capable of treating up to 35,000 acfm (at
300°F) of flue gas from the 160 MW TVA Shaw-
nee coal-fired Boiler No. 10. This corresponds to
approximately 10 MW of power plant generating
capacity. The equipment selected was sized for
minimum cost, consistent with the ability to ex-
trapolate results to commercial scale. Boiler No. 10
burns medium- to high-sulfur bituminous coal,
which results in SO2 concentrations of 1,500 to
4,500 ppm and particulate inlet loadings of about
2 to 5 grains/scf in the flue gas.
The TCA was manufactured by Universal Oil
Products. It operates with beds of low-density
spheres that are free to move between retaining
grids. As the incoming flue gas contacts the slurry
in these beds, both SO2 and particulates are
removed.
The adjustable-throat venturi scrubber in the
venturi/spray tower system was manufactured by
Chemical Construction Company. The venturi
scrubber removes the bulk of the particulates. But
because the residence time in the venturi scrubber
is low, the scrubber with lime/limestone slurry re-
moves less than half of the SO2- The spray tower
that follows the venturi scrubber provides suffi-
cient residence time for the removal of most of the
remaining SO2-
Figures 1 and 2, drawn roughly to scale, show
the venturi/spray tower and the TCA, along with
the mist eliminators selected for deentraining
slurry in the gas streams.
The test facility was designed to allow a number
of different scrubber internals and piping configu-
rations to be used with each scrubber system. For
example, the TCA can be operated as a one-, two-,
or three-bed unit, and spent solids dewatering can
be achieved with a clarifier alone or with a clarifier
in combination with a filter or a centrifuge. Either
-------
scrubbing system can operate in combination with
one to three stirred hold tanks in series. A typical
venturi/spray tower system configuration used dur-
ing lime testing and a typical TCA system configu-
ration used during limestone testing are shown in
Figures 3 and 4, respectively. Process details, such
as flue gas cooling sprays, are not shown.
SAMPLE POINTS
® SLURR Y OR SOLIDS
COMPOSITION
SETTLING POND
Figure 3. EPA Test Facility — Typical Process Flow Diagram for Venturi/Spray Tower System in
Lime Service.
SAMPLE POINTS
SCRUBBER EFFLUENT HOLD TANKS
SETTLING POND
Figure 4. EPA Test Facility — Typical Process Flow Diagram for TCA System in Limestone Service
-------
Scrubber performance is judged by the potential
for removing both SO2 and participates at high
efficiencies (defined for the Shawnee facility as
SO2 removal greater than 80 percent and particu-
late removal greater than 99 percent). Other fac-
tors to be considered are an ability to handle slur-
ries without plugging or excessive scaling, reason-
able cost and maintenance, ease of control, and
reasonable pressure drop.
For all configurations, gas is withdrawn from the
boiler ahead of the power plant paniculate removal
equipment so that entrained fly ash can be intro-
duced into the scrubber. The gas flow rate to each
scrubber is measured by venturi flow meters and
controlled by dampers on the induced-draft (I.D.)
fans. Concentration of SO2 in the inlet and outlet
gas is monitored continuously by Du Pont photo-
metric analyzers. Inlet and outlet gas particulate
concentrations are measured periodically using a
modified EPA particulate train.
Control of scrubbing systems is carried out
from a central graphic panel board. Onsite digital
display of selected information is available in the
control room. Important process control variables
are continuously recorded, and trend recorders are
provided for periodic monitoring of selected data
sources.
To meet the formidable analytical requirements
of the facility at reasonable costs, the onsite ana-
lytical laboratory was furnished with equipment
selected to minimize manpower; e.g., an X-ray
fluorescence unit is used for comprehensive slurry
analyses. Samples of slurry, flue gas, limestone,
and coal are taken periodically for chemical analy-
ses and particulate mass loading. The locations of
slurry and gas sample points for the venturi/spray
tower system are shown in Figure 3 and for the
TCA in Figure 4. All analytical computations and
recording of results are handled by an onsite
minicomputer.
Views of the venturi, control room, TCA, and
spray tower are shown in Figures 5, 6, 7, and 8,
respectively.
Figure 5. The Venturi
-------
Figure 6. Control Room
Figure 7. TCA
Figure 8. Spray Tower
-------
The original Shawnee test program was schedul-
ed from March 1972 through June 1974. Tests in-
cluded on the original program were:
• Air/water tests to determine pressure drop
model coefficients and to observe fluid
hydrodynamics
• Sodium carbonate tests to determine gas-
phase mass-transfer coefficients in mathemati-
cal models for predicting SO2 removal
• Limestone wet-scrubbing tests to identify im-
portant variables affecting scrubber perfor-
mance, to resolve operating problems while
maintaining satisfactory SO2 removal and
economic operation, to demonstrate long-
term reliability, and to provide scale-up data
• Lime wet-scrubbing tests with the same objec-
tives as for limestone tests
Results of the original test program are outlined
in two previous capsule reports and are discussed in
more detail in EPA report EPA-650/2-75-047.
A need for additional development of lime/lime-
stone wet-scrubbing technology led to the establish-
ment of an advanced test program to extend past
June 1974. The entire advanced test program
schedule from July 1974 through December 1977
is shown in Figure 9. Detailed advanced test re-
sults through mid-February 1976 have been pre-
sented in EPA Progress Reports EPA-600/2-75-050
and EPA-600/7-76-008. A third detailed progress
report will be issued in early 1977.
The current capsule report covers the advanced
test program from January 1975 through April
1976. During this period, the following tests were
conducted:
• Limestone utilization tests
• Mist eliminator reliability tests
• Variable-load tests
• Factorial tests
LIMESTONE UTILIZATION TESTS
The effects of scrubber inlet liquor pH and hold
tank residence time on limestone utilization were
investigated for both the venturi/spray tower and
the TCA systems. In addition, the effect of the
number of hold tanks in the TCA system was
examined. Results are given in Section 4.
MIST ELIMINATOR OPERATION
In alkali wet-scrubbing, and particularly in lime-
stone wet-scrubbing, solids-free operation of the
mist eliminator section has been difficult to
achieve. There has, however, been significant prog-
ress and this progress is discussed in Section 5.
VARIABLE-LOAD TESTS
The way a power station's normal daily fluctuat-
ing load will affect the scrubber's reliability and
controllability must be determined. This relation-
ship was investigated for the venturi/spray tower
system in lime service and is described in Section 6.
FACTORIAL TESTS
For each of the scrubber systems, a series of
short factorial tests (6 to 8 hours each) was run to
determine the effect of the following major vari-
ables on SO2 removal:
Gas flow rate
Recirculating slurry flow rate
Scrubber inlet liquor pH
Magnesium ion concentration
Venturi pressure drop
Height of spheres per bed in the TCA
Spray tower header configuration and nozzle
pressure drop
Some of the major results apparent from these
tests are presented in Section 7.
QUALITY ASSURANCE
A quality assurance program was initiated to
assure that all procedures meet the high standards
of accuracy demanded by the advanced test pro-
gram and to maintain these standards under an in-
creased work load. A laboratory procedures man-
ual is available which documents all primary and
secondary procedures used at Shawnee.
DATA BASE
A wide selection of data, describing runs from
September 1972 to the present, is being stored in a
-------
computerized data base, which includes (1) sum-
maries of run and operating conditions, and
(2) solid and liquid analytical data from the recir-
culating slurry stream, usually collected three times
daily. This data base, together with plotting and
file manipulation programs (all described in a user's
manual), will be made available for public access at
a later date.
TEST BLOCK
JJJF_ M A I* I j\ AJS Oil*, 9
I F M AIM I J AlS O N 0
TCA SYSTEM
Limesto™ RHiJbility
LimeMoiw Utilization
Lime/Limestone Factorial
Limestone with MfO Addition
Lime with and without HiO Addition
Limestone Fly Ash Free
Limestone Automatic Control
LimeMoneVariabKLoad
Limntont Altcmatne Scrubber Internals
Limestone Energy Minimization
Limeitone Reliability Demonstration
VENTURJ»I>RA¥ TOWtd SYSTEM
Lime Reliability
Limt Variable Load
Limestone Utilization
Lime/Limestone Factorial
Lime with MjO Addition
Lime Fly A* Free
Lime Flue Ga.Charaaerization
Limestone Sulfite Oxidation with Fly A*
Lime Sulfite Oxidation with Fly Ash
Lime Sulfite Ox idation without fly Ash
Ltmei«CMie Sulfite Oxidation wrlnoMt Fl y Ash
Lime Energy Minimization
Lime Reliability Demonstration
tnr
Figure 9. Shawnee Advanced Test Schedule
-------
From October 1975 through mid-February 1976,
limestone utilization tests were conducted on both
scrubber systems. These tests were designed to
substantiate the findings of a TVA economic study
which showed that the economics of limestone scrub-
bing could be improved by increasing limestone utili-
zation (moles SO2 absorbed/mole Ca added). In-
creased limestone utilization not only reduces the
limestone feed requirements but also reduces waste
sludge production.
Tests were conducted to correlate alkali utiliza-
tion with scrubber inlet liquor pH, effluent hold
tank residence time, and hold tank configuration.
VENTURI/SPRAY TOWER UTILIZATION
TESTS
In the venturi/spray tower system, the relation-
ship between limestone utilization and scrubber in-
let liquor pH was determined for three different
hold tank residence times. Major test conditions
and results are shown in Figure 10, where limestone
utilization is plotted as a function of scrubber inlet
liquor pH, and in Figure 11, where percent SO2 re-
moval is plotted against limestone utilization. As
expected, limestone utilization decreased with in-
creasing scrubber inlet liquor pH and increasing
SO2 removal. Limestone utilization was similar
at 12 and 20 minutes' residence time. At 6 min-
utes' residence time, however, limestone utiliza-
tion was adversely affected at the higher pH
values. Scatter in the data was too great to observe
a corresponding reduction in SO2 removal with a
decrease in residence time.
TCA UTILIZATION TESTS
In the TCA system, three stirred hold tanks in
series were compared with a single stirred hold
tank. Kinetic theory shows that for a continuous
system where the reaction order is greater than
zero, raw materials are more completely converted
in a plug flow reactor than in a backmixed reactor
with the same residence time. A plug flow reactor
can be approximated by three stirred tanks in
series. Hold tank configurations tested in the TCA
system were:
• Single hold tank at 12 minutes
• Three hold tanks at 14.4 minutes total
• Three hold tanks at 10.8 minutes total
100
II
SPKA Y TOWER GAS VELOCITY = 9.4 ft/sec
VENTURI LIQUID-TO-GAS KA T/O = 2I gal/mcf
SPKA Y TOWEK LIQUID-TO-GAS KA TIO - 50 gal/mcf
VENTURI PRESSURE DROP'9 in H2O
PERCENT SOLIDS RECIRCULATED = 75
INLET SO2 CONCENTRA TION = 2.500 to 3,500ppm
SCRUBBER INLET LIQUOK pH
Figure 10.
Scrubber Inlet Liquor pH and Hold
Tank Residence Time versus Limestone
Utilization in the Venturi/Spray Tower
System
I
Is
—J (J-l
RESIDENCE TIME = 6 to 20 mm
SPRA Y TOWER GAS VELOCITY = 94 ft/sec
VENTURI LIQUID-TO-GAS KA TIO = 21 gal/mcf
SPRAY TOWER LIQUID-TO-GAS RATIO = 50 gal/mcf
VENTURI PRESSURE DROP = 9 in Hf>
PERCENT SOLIDS RECIRCULA TED = 15
INLETSO2 CONCENTRA TION = 2,500 to 3,500 ppm
70 SO
PERCENT SO2 REMO VA L
Figure 77.
SO2 Removal versus Limestone
Utilization in the Venturi/Spray
Tower System
-------
Total residence times for the three stirred tanks
in series were intended to be 12 and 9 minutes, but
because of a flow meter error the residence times
were actually 14.4 and 10.8 minutes, respectively.
As seen in Figure 12, limestone utilization was
improved for a given scrubber inlet liquor pH with
the use of three stirred tanks in series — especially
at higher pH values. No effect of residence time on
utilization was observed over the range tested. As
seen in Figure 13, SO2 removal at a given lime-
stone utilization was improved with three tanks in
series. SO2 removal with three tanks in series at
10.8 minutes' residence time is not shown in this
figure because the replacement of worn spheres
with new ones immediately before these tests
caused an enhancement of SO2 removal.
It is apparent from these tests that operation at
reduced scrubber inlet liquor pH has the greatest
effect on increasing limestone utilization, but that
it also results in a corresponding reduction in SO2
removal. To increase SO2 removal at reduced pH,
one can increase the liquid-to-gas ratio, the pressure
drop across the scrubber, or the magnesium ion con-
centration in the slurry liquor. The effects of these
variables on SO2 removal are discussed in Section 7.
3 TANKS
RESIDENCE TIME = 10 8 to 144 mm
1 TANK
RESIDENCE TIME = 12 mm
TCA SUPERFICIAL GAS VELOCITY =125 ft/sec
LIQUID-TO-GAS RATIO = 42 to 50 gal/mcf
SPHERE BED HEIGHT = 5 in /bed, 3 beds
PERCENT SOLIDS RECIRCULA TED - 15
INLET SO2 CONCENTRA TION = 2,500 to 3,500ppm
54 56
SCRUBBER INLET LIQUOR pH
Figure 12.
Scrubber Inlet Liquor pH and Hold
Tank Configuration versus Limestone
Utilization in the TCA System
Uj O
i'
3 TANKS
RESIDENCE TIME =144 mm
TCA SUPERFICIAL GAS VELOCITY' 125 ft/sec '
LIQUID-TO-GAS RA TIO = 42 to SOgal/mcf
SPHERE BED HEIGHT =5m I bed, 3 beds
PERCENT SOLIDS RECIRCULA TED => 15
INLETSO2 CONCENTRA TION = 2,500 to 3,500ppm '
60
70 SO
PERCENT SO2 REMO VA L
Figure 13.
SO2 Removal and Hold Tank
Configuration versus Limestone
Utilization in the TCA System
-------
Removing the entrained mist from the scrubbed
flue gas without fouling the mist eliminator system
is one of the key factors in successful scrubber
operation. Early in the testing program, more
difficulty was experienced in keeping the mist
eliminator clean on the TCA in limestone service
than on the spray tower in lime service. This diffi-
culty was initially attributed to differences in
scrubber design, which might have caused a finer
mist in the TCA. Later it was found that the mist
eliminator was much easier to keep clean at high
alkali utilization (above about 85 percent) than at
lower utilization. Lime systems operate at an in-
herently high utilization; limestone systems can
operate over a range of utilization.
CHEVRON MIST ELIMINATOR IN
LIME SERVICE
The venturi/spray tower mist eliminator system
consists of a three-pass, open-vane, 316L stainless-
steel chevron mist eliminator with provision for
underside and topside washing (see Figure 1). In
earlier tests in lime service with underside wash
only, scale formed on the top vanes of the mist
eliminator. This problem was eliminated by the
additional infrequent topside wash. At the end of
a series of tests lasting a total of 1,075 hours using
an 8 to 15 percent slurry solids concentration and
an 8.0 to 9.4 ft/sec spray tower gas velocity, the
mist eliminator was essentially clean. The under-
side was washed with makeup water at 1.5 gpm/ft2
for 6 minutes every 4 hours; the topside was
washed with makeup water on an 8-hour sequential
cycle with one of the six nozzles activated for 4
minutes every 80 minutes. The water consumption
was about half the makeup water permissible for
closed-liquor-loop operation (effluent sludge con-
taining at least 38 percent solids and no liquor
bleed).
CHEVRON MIST ELIMINATOR WITH
WASH TRAY IN LIMESTONE SERVICE
In the TCA system in limestone service at low al-
kali utilization (65 percent), a wash tray followed
by a six-pass, closed-vane, 316L stainless-steel
chevron mist eliminator was successfully demon-
strated in 1,835 hours of operation using an 8.6 ft/
sec superficial scrubber gas velocity (see previous
capsule report). In tests at higher gas velocities, ac-
cumulation of soft solids could not be prevented.
After limited tests with other configurations, a
three-pass, open-vane, 316L stainless-steel chevron
mist eliminator similar to the one being used
successfully in the spray tower in lime service was
installed.
CHEVRON MIST ELIMINATOR IN
LIMESTONE SERVICE
During limestone utilization testing, both scrub-
bers operated with three-pass, open-vane, 316L
stainless-steel chevron mist eliminators. Alkali
utilization during this period was varied from 50 to
100 percent. Above about 85 percent utilization,
the mist eliminator could be kept free of solids
deposits in both scrubber systems with the same
topside and underside wash used in lime service.
Below about 85 percent utilization, this wash
scheme did not limit solids accumulation. How-
ever, in a TCA run at 60 to 70 percent alkali utili-
zation that lasted 811 hours, intermittent topside
wash combined with continuous underside wash at
0.4 gpm/ft2 using diluted clarified liquor was suc-
cessful in limiting restriction by soft solids to a
steady-state level of less than 10 percent. There
was some evidence that a partially fouled mist
eliminator may be cleaned up if the alkali utiliza-
tion is raised to a high value.
Figures 14 and 15 show the spray tower mist
eliminator at the end of limestone runs at low and
at high alkali utilization, respectively. Both runs
were made with intermittent topside and under-
side wash. In Figure 14, the mist eliminator was
50 to 60 percent restricted after 73 hours at 70
percent utilization. In Figure 15, the mist elimina-
tor was only 1 percent restricted after 319 hours at
93 percent utilization. These values of mist elimi-
nator restriction are based on visual estimates by
inspectors during and at the end of runs.
-------
Figure 14. Venturi/Spray Tower Mist Eliminator A fter 73 Hours at 70 Percent Alkali Utilization
Figure 15. Venturi/Spray Tower Mist Eliminator A fter 319 Hours at 93 Percent Alkali Utilization
-------
A 7-week variable-load (cycling gas rate) test
series with lime was conducted on the venturi/
spray tower system from August to October 1975.
These tests were designed to evaluate the ability of
the scrubber system to handle the variable gas rate
and composition resulting from a daily boiler load
cycle. A summary of the operating conditions for
this test series is given in Table 1.
The test series was divided into two parts. The
first part was a 717-hour test in which the venturi
plug position was adjusted according to the varying
gas flow rate to maintain a constant 9-inch H2O
pressure drop across the venturi. The second part
was a 426-hour test with the venturi plug fixed at a
position giving a 9-inch H2O pressure drop across
the venturi at the maximum gas flow rate of
35,000acfm.
During the test, the gas flow rate was varied be-
tween 17,000 and 35,000 acfm so as to approxi-
mately follow the-actual boiler load, which fluctu-
ated between 60 and 160 MW. The inlet SO2 con-
centration varied between 1,500 and 4,400 ppm.
Because of the changing gas rate, the total liquid-
to-gas ratio for the venturi and spray tower varied
between 71 and 161 gal/mcf.
The mist eliminator was washed intermittently
on both the underside and the topside with make-
up process water. The wash water requirement of
2.3 gpm on a continuous basis was at all times less
than the permissible makeup water in closed-liquor-
loop operation, even under the most constraining
conditions of low gas rate and high inlet SO2
concentration.
Throughout the 1,143 hours of operation, the
system ran reliably and had good control. The
mist eliminator remained clean: at the conclusion
of the tests, there was only a 2 percent restriction
of the cross-sectional area by solids.
Table 1
OPERATING CONDITIONS FOR VENTURI/SPRAY TOWER VARIABLE LOAD TEST
-------
From mid-February to April 1976, lime and
limestone factorial tests were performed at the test
facility. A partial factorial set of 262 runs, each
lasting 6 to 8 hours, was made to determine the
effects of the following independent variables on
percent SO2 removal by the venturi (only), the
spray tower (operated with minimum flow to the
venturi section), and the TCA:
• Gas flow rate
• Recirculating slurry flow rate
• Scrubber inlet liquor pH
• Spray tower header configuration and nozzle
pressure drop
• Venturi pressure drop
• TCA sphere height per bed
• Magnesium ion concentration
For these factorial tests, SO2 inlet concentration
was controlled by the use of coal from a single
source in Boiler No. 10. Because of the short dura-
tion of each test, no attempt was made to control
alkali utilization. The effluent hold tank was held
at a constant level while residence time fluctuated
with changes in slurry rate. The percent solids in
100
90
80
3
ce
£
Uj
70
60
SO
40
SLURR Y FLOW RA TE = 30 gal/min-ft2
15 gal/min-ft2
D
SO2 INLET CONCENTRA TION = 2,540 to 3,160 ppm
SCRUBBER IN LET LIQUOR pH = 5.7 to 5 8
SPRA Y TOWER CROSS-SECTIONAL AREA = 50 ft2
56 7 8 9 10
SPRA Y TOWER GAS VELOCITY, ft/sec
Figure 16. Effect of Gas Velocity and Slurry
Flow Rate on SO2 Removal — Spray
Tower with Limestone
the recirculated slurry was held to 15 ±1 percent.
During spray tower tests, the slurry flow to the
venturi was kept to the minimum required for flue
gas cooling (approximately 140 gal/min), and the
venturi plug position was held wide open. This
procedure kept percent SO2 removal by the ven-
turi to less than 10 percent.
FACTORIAL TESTS - NO MAGNESIUM
OXIDE ADDED
With both lime and limestone, percent SO2 re-
moval in the venturi (only) and in the spray tower
decreased as the gas rate increased from 20,000 to
35,000 acfm at a constant liquor rate (decreasing
liquid-to-gas ratio). In the TCA, percent SO2 re-
moval did not change or increased slightly as the
flue gas flow rate increased. In Figures 16 and 17,
the percent SO2 removal by limestone slurry is
plotted against gas velocity, with slurry flow rate as
a parameter, for the spray tower and TCA scrub-
bers, respectively. Similar data were obtained with
lime. Figures 16 and 17 also show the expected
strong effect of the recirculated slurry flow rate
on the percent SO2 removal observed in all of the
factorial tests.
700
90
O 80
70
60
50
SLURR Y FLOW RATE = 38 gal/mm-ft2
SO2 INLET CONCENTRA TION = 2,360 to 2,690ppm
SCRUBBER INLET LIQUOR pH = 5.8 to 5.9
SPHERE BED HEIGHT = 5 in./bed, 3 beds
TCA CROSS-SECTIONAL AREA = 32 ft2
9 10 11 12
TCA GAS VELOCITY, ft/sec
13
Figure 77.
Effect of Gas Velocity and Slurry
Flow Rate on SO2 Removal - TCA
with Limestone
-------
700
90
2
o
5 70
a:
IN
O
60
50
40
30
20
TCA
(60ga//mcf)
SPRA Y TOWER
(51 gal/mcf)
VENTURI
(27 gal I me f)
I
678
SCRUBBER INLET LIQUOR pH
Figure 18.
Effect of Scrubber Inlet Liquor pH on
SO2 Removal
Figure 18 shows the increase in percent SO? re-
moval with increasing pH for all three scrubbers us-
ing lime or limestone. Run conditions for these
tests are shown in Table 1. In all three scrubbers,
percent SO? removal was greater at a scrubber
liquor pH of 6 with limestone than with lime.
However, excess limestone is required to maintain
this pH (alkali utilization below 80 percent). This
excess limestone is lost in the discharged waste
solids. In a lime system, because of the relatively
high solubility and high alkalinity of lime, the alka-
li utilization is essentially 100 percent at a pH of 6
and SO2 removal is poor. Lime systems normally
operate at higher pH for good SO2 removal with
alkali utilization above 95 percent.
Percent SO? removal in the spray tower in-
creased with spray nozzle pressure drop. At a gas
rate of 7.4 ft/sec, a slurry flow rate of 22.5 gal/
min-ft2, and a pH of 5.8 (limestone system), SO2
removal increased from 66 percent with four head-
ers (nozzle pressure drop 7.9 psi) to 76 percent
with three headers (nozzle pressure drop 14.3 psi).
SO2 removal in the venturi increased with in-
creasing pressure drop, although liquid-to-gas ratio
had a greater effect. For example, at a pressure
drop of 9 inches H2O, SO2 removal increased from
20 to 45 percent as liquid-to-gas ratio increased
from 10 to 37 gal/mcf. For both lime and lime-
stone, these removal efficiencies were virtually inde-
Table 2
RUN CONDITIONS FOR FACTORIAL TESTS SHOWING pH EFFECT
Gas Flow Rate, acfm @ 300°F
Gas Velocity, ft/sec
Slurry Flow Rate,gal/min
Liquid-to-Gas Ratio, gal/mcf
Scrubber Pressure Drop, in. H2O
Slurry Solids Concentration, wt %
Inlet SO2 Concentration, ppm
*Sphere Bed Height = 5 inches/bed, 3
**3 Sorav Ranks in Oneratinn
TCA*
25,000
10.4
1,200
60
6
15
2,300-2,800
beds
Spray Tower**
21,500
7.4
1,125
51
1
15
2,500-3,100
Venturi
21,500
—
600
21
9
15
2,400-2,900
-------
pendent of gas rate. During other tests, raising the
venturi pressure drop from 6 to 12 inches H2O in-
creased SO2 removal from 37 to only 46 percent.
Percent SO2 removal increased when the TCA
sphere bed height was increased. This is shown for
limestone in Figure 19. Similar behavior was ob-
served with lime. In choosing bed height in a com-
mercial TCA scrubber, however, the increasing
pressure drop across the TCA as the bed height is
increased must be considered. At a gas velocity of
10.4 ft/sec and a slurry flow rate of 38 gal/min-ft2,
the pressure drop across the TCA varied from 1.5
inches H2O to 9 inches H2O as the sphere bed
height was changed from 0 to 7.5 inches.
FACTORIAL TESTS - MAGNESIUM
OXIDE ADDED
When magnesium ion concentration in the scrub-
ber liquor exceeded the equivalent chloride ion
concentration, percent SO2 removal increased with
increasing magnesium ion concentration. One
gram-mole of magnesium (Mg++) requires two
gram-moles of chloride (CI-) for charge neutrality.
Additional Mg++ (called effective Mg++ in this re-
port) is associated with sulfate (SO4=) and sulfite
SO3=), which in turn affects SO2 removal. Mg++
concentration in the scrubber slurry liquor was in-
creased by adding magnesium oxide.
Percent SO2 removal increased sharply with in-
creasing Mg++ ion concentration in the recirculated
slurry for the spray tower, TCA, and venturi scrub-
bers. Spray tower results are shown in Figure 20.
TCA results were similar. For the venturi, SO2
removal increased from 35 percent (no Mg++) to
85 percent (9,000 ppm effective Mg++) with a
9-inch pressure drop and a pH of 5.8. Thus, at
high Mg++ ion concentration, the venturi scrubber
can remove a large percentage of the SO2 as well as
the particulates. Such flexibility should be useful
in the design and operation of commercial SO2
scrubbing systems.
700
90
O
80
£70
S
o
£60
ki
40
30
SLURR Y FLOW RA TE = 38 gal/m/n-ft?
Figure 19.
SO2 INLET CONCENTRA TION = 2,360 to 2,740 ppm
TCA GAS VELOCITY = 10.4 ft/sec
SCRUBBER INLET LIQUOR pH = 5.8
TCA CROSS-SECTIONAL AREA = 32 ft?
2468
STA TIC SPHERE HEIGHT/BED, in.
10
Effect of Sphere Bed Height and
Slurry Flow Rate on SO2 Removal
— Three-Bed TCA with Limestone
100
90
80
70
60
50
40
30
8,000 to 10,400 ppm
EFFECTIVE Mg++ CONC
= 2,800 to 4,200 ppm
500 to J,250pp,
0 ppm
•EFFECTIVE Mg++ = Mg++ - Cl72.92
= OforMg++
-------
HOT-GAS/LIQUID INTERFACE
Both the TCA and the venturi/spray tower
systems have operated with no solids buildup prob-
lems at the hot-gas/liquid interface. Normally,
daily soot blowing in the direction of gas flow has
been used to remove accumulated solids. Intervals
between soot blowing of up to 3 days have been
successfully employed.
REHEATERS
Reheaters fired with fuel oil directly in the flue
gas stream were originally installed, but their per-
formance proved unsatisfactory owing to frequent
flameouts. An external combustion chamber was
added to the venturi/spray tower reheater in March
1974 and to the TCA reheater (Figure 21) in May
1975. These units have performed satisfactorily with
high reliability for over 13,700 and 4,400 operating
hours, respectively.
Figure 21. TCA Reheater
PUMPS
The rubber-lined, variable-speed centrifugal
pumps used in slurry service have operated reliably.
Air-flushed packings, although a high-maintenance
item, have been used with satisfactory results since
February 1973 to avoid seal water dilution in the
slurry system. Installation of hardened shaft
sleeves has improved packing life. Two mechanical
seals, installed on slurry bleed pumps as an alterna-
tive to air-flushed packing, failed after 1,500 and
6,000 hours of service.-
FANS
Erosion, corrosion, pitting, and scaling have
been negligible on the I.D. fans in both systems.
The fans normally operate with flue gas reheated
to 2500 F.
LININGS
The neoprene rubber linings on the agitator
blades and in the spray tower, TCA, process water
hold tanks, pumps, and circulating slurry piping
continue to be in excellent condition, except for
occasional small tears and bubbles. These tears and
bubbles have been successfully repaired. Some
erosive wear was noted in the glass flake lining on
the effluent hold tanks and clarifiers, but no pene-
tration of the lining has been observed after 4 years
of service. Field repairs of glass flake lining are
possible.
NOZZLES
The 316 stainless-steel, full-cone, open-type
nozzles in the TCA have operated 9,000 hours with
no significant wear at 5 psi pressure drop. Stellite-
tipped, full-cone, spiral nozzles in the spray tower
have operated 11,700 hours at 10 psi pressure drop
with approximately 40 percent total weight loss of
the tips. Wear rate during operation with lime-
stone was greater than with lime. The 316
stainless-steel inlet reducers, which connect the
spray nozzles to headers, have been so eroded by
the high-velocity slurry that nozzles have occasion-
ally broken loose and dropped into the effluent
hold tank below.
SPHERES
Evaluation of TCA sphere life has continued.
Six-gram thermoplastic rubber (TPR) spheres were
tested. The failure rate from seam splitting was
-------
linear with time and about 11 percent after 3,800
hours of operation. Average weight loss of the un-
split spheres was about 10 percent. These spheres
were replaced with 6.5-gram, solid nitrile foam
spheres. The performance of the foam spheres has
been encouraging thus far.
WASTE SOLIDS HANDLING
Alternative waste solids dewatering equipment
includes separate clarifiers for each scrubber
system, plus a rotary-drum vacuum filter and hori-
zontal solid-bowl centrifuge common to the two
systems.
The underflow solids concentration of the larger
(30-foot diameter) clarifier serving the TCA has
normally exceeded 38 weight percent, but the
underflow solids from the smaller (20-foot dia-
meter) venturi/spray tower clarifier has averaged
only about 25 weight percent solids. To achieve
closed-liquor-loop operation (effluent sludge con-
taining at least 38 percent solids and no liquor
bleed), the smaller clarifier has to be used in series
with the filter or centrifuge. Downward extension
of improperly designed feed wells in both clarifiers
has improved the clarity of the overflow.
The rotary-drum vacuum filter was modified
from a roll discharge to a deflector-assisted, blow-
back discharge scheme in February 1975. Under
normal operations, the rotary-drum vacuum filter
(Figure 22) produced a filter cake containing 50 to
60 weight percent solids from either lime or lime-
stone slurries. Filter operation continues to be
hampered by the short life (usually less than 600
hours) of the filter cloth. It is suspected that mal-
adjustment of the discharge deflector has contrib-
uted to the short cloth life.
The continuous solid-bowl centrifuge (Figure 23)
produced a cake with 55 to 65 weight percent
solids from limestone slurry, and the centrate
solids averaged 0.5 to 1.0 weight percent. Al-
though the unit has operated successfully as a
solids dewatering device, major resurfacing of the
wear surfaces was necessary after 1,400 hours of
operation in limestone service with fly ash and
again after 3,500 hours in predominately lime ser-
vice with fly ash.
Figure 22. Rotary-Drum Vacuum Filter
Figure 23. Solid-Bowl Centrifuge
INSTRUMENTS
Submersible electrode pH meters have been used
in slurry service and, except for occasional scale
formation which can be removed with hydro-
chloric acid, these meters have been reliable. A
short test indicates that a continuous ultrasonic
cleaner would alleviate this scale-formation
problem.
Two types of density meters have been tried in
slurry service: radiation and vibrating U-tube
meters. The radiation meter exhibited calibration
shifts which were accelerated by scale formation.
The performance of two vibrating U-tube meters
installed in September 1973 has been excellent.
-------
Slurry flow rates have been measured by both
magnetic and orifice flow meters. Orifice flow
meters are accurate only if the diaphragm pressure
taps are frequently cleaned — usually about every
2 weeks. Magnetic flow meter performance has
been good except for liner failure experienced with
the 1-inch and "P/2-inch sizes. This problem was
traced to the tapered design of the liner. Conver-
sion to an untapered liner appears to have elimi-
nated these failures.
LIME FEED SYSTEM
The lime feed system has performed well. There
have been minor problems of grit-screen blinding
in the slaker and a moderately high replacement
rate for the injection pump rotor/stator assembly.
COMPONENTS TEST PROGRAM
A mechanical components test program, initi-
ated during the advanced program, has continued
with plastic pipe, butterfly and knife gate valves,
line strainers, mechanical pump seals, a coated ori-
fice plate, and several liner materials under test.
Results will be given in later reports.
A Third Interim Report on the continuing
corrosion/erosion study by TVA is included in the
Advanced Program's Second Progress Report
(EPA-600/7-76-008).
For further information:
Detailed progress reports, EPA-600/2-75-050 and EPA-600/7-76-008 are available from the National
Technical Information Service, Springfield, Virginia, 22151. Also available is a report, EPA-650/2-75-
047, on the earlier test program.
A further detailed progress report and a final report on the advanced program will be prepared. If
you wish to be notified when these reports are available, write:
Technology Transfer
Environmental Protection Agency
Cincinnati, Ohio 45268
------- |