EPA
TECHNOLOGY
TRANSFER
FIRST PROGRESS REPORT:
LIMESTONE U.S. EPA
WET-SCRUBBING OFFICE OF
TEST RESULTS RESEARCH AND
AT THE DEVELOPMENT
EPA ALKALI PROTOTYPE
SCRUBBING DEMONSTRAT ON
TEST FACILITY FACILITY
-------
EPA
TECHNOLOGY
TRANSFER
FIRST PROGRESS REPORT:
LIMESTONE
WET-SCRUBBING
TEST RESULTS
AT THE
EPA ALKALI
SCRUBBING
TEST FACILITY
U.S. EPA
OFFICE OF
RESEARCH AND
DEVELOPMENT
PROTOTYPE
DEMONSTRATION
FACILITY
-------
TVA Shawnee Steam Plant with Wet-Scrubbing Facility to Left of Boiler House
-------
The Clean Air Act of 1970 has now given increased
emphasis to programs directed to decreasing SOs
and participate emissions from new and existing
power plants and other facilities. At the present state
of technology there are only three methods for sub-
stantially reducing oxide emissions: (1) switching to a
low sulfur fuel; (2) desulfurizing the fuel; and (3) de-
sulfurizing the gases produced. Among the more
than 50 gas desulfurization control concepts which
have been proposed and studied, the wet lime/lime-
stone systems are considered to be the most
advanced.
In a lime/limestone wet-scrubbing system, the flue
gas is contacted (scrubbed) with a slurry of lime/
limestone and water. The particulate matter (fly ash)
is captured by liquid droplets and the sulfur dioxide
(SOa) is absorbed into the liquor where it reacts
with the dissolved lime/limestone, forming the waste
products of calcium sulfite and calcium sulfate
(gypsum).
In June of 1968, the Environmental Protection
Agency (EPA), through its Office of Research and
Development (OR&D) and Control Systems Lab-
oratory, initiated a program to test a prototype wet
lime and limestone scrubbing system for removing
sulfur dioxide and particulates from flue gases.
Bechtel Corporation of San Francisco is the major
contractor and test director, and the Tennessee Valley
Authority (TVA) is the constructor and facility operator.
To minimize the time required for the technology
from this project to become available to engineers
and technical managers, a series of four technical
reports are scheduled as work progresses. This cap-
sule report discusses the highlights of the first detailed
engineering progress report. It describes the test
facility and test program and presents the results to
date of the limestone wet-scrubbing testing. In addi-
tion, the reliability and operability of the test facility
during long-term (2+ weeks) closed liquor loop
operation is discussed.
The test facility consists of three parallel scrubber
systems: a venturi followed by a spray tower; a Turbu-
lent Contact Absorber (TCA): and a Marble-Bed
Absorber. Each system is capable of treating approx-
imately 10 Mw equivalent (30,000 acfm) of flue gas
containing 2300 to 3300 ppm SC>2, and is integrated
into the flue gas ductwork of a coal-fired boiler at
the TVA Shawnee Power Station, Paducah, Kentucky.
The following sequential test blocks were defined
for the program: (1) air/water testing; (2) sodium
carbonate testing; (3) limestone wet-scrubbing test-
ing; and (4) lime wet-scrubbing testing. The air/water
and sodium carbonate tests have been completed.
As of June 1973, short-term (less than 1 day) lime-
stone wet-scrubbing factorial tests were 95 percent
complete and longer-term (2+ weeks) reliability veri-
fication tests were approximately 50 percent
complete. Long-term (4 to 10 months ) limestone
testing and lime testing are scheduled to begin in
September, 1973.
The short-term factorial limestone tests were con-
ducted at high scrubber inlet liquor pH (6.0-6.2).
Series operation of the venturi and spray tower pro-
duced SC>2 removals of up to 80 percent at a total
liquid-to-gas ratio of 80 gal/rncf and pressure drop
of 10 in. FbO. The three stage TCA scrubber obtained
up to 96 percent SOs removal at a liquid-to-gas ratio
of 64 gal/mcf and pressure drop of 7 in. HsO.
Removals of 80 percent were achieved with a single
stage Marble-Bed Absorber at a liquid-to-gas ratio of
40 gal/mcf and a pressure drop of 11 in. HaO.
The three initial long-term reliability verification
tests with closed liquor loop operation have been
run at reduced scrubber inlet liquor pH (5.7-5.9)
and, consequently, at reduced stoichiometric ratio,
in order to increase system reliability and limestone
utilization (moles SOa absorbed/moles CaCOa
added). For the TCA, limestone utilization was 83
percent with an SO:? removal of 80 to 85 percent
and a pressure drop of 7 in. HaO.
The operability and reliability of the scrubber
systems for the initial reliability verification tests have
been good. There has been little evidence of sulfate
or sulfite scale after approximately 500 hours of
operation on all three systems, with effluent resi-
dence times greater than 20 minutes and percent
solids recirculated greater than 10 percent (40
percent of solids is fly ash). Presently, more severe
operating conditions (e.g., lower effluent residence
times) are being tested to determine the regions of
reliable operation for the three systems.
-------
The test facility 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
theTVA Shawnee coal boiler No. 10. Therefore.
each circuit is handling the equivalent of approxi-
mately 10 Mw of power plant generating 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 SCh
concentrations of 2300-3300 ppm and particulate
inlet loadings of about 2 to 5 grains/scf in the flue gas.
The major criterion for scrubber selection was the
potentiality for removing both sulfur dioxide and
particulates at high efficiencies (sulfur dioxide re-
moval greater than 80 percent and particulate
removal greater than 99 percent). Other factors
considered in the selection of the scrubbers were:
(1) ability to handle slurries without plugging or
excessive scaling: (2) reasonable cost and mainten-
ance: (3) ease of control: and (4) reasonable
pressure drop.
Based on the information available in the literature,
the following scrubbers were selected:
Venturi followed by a spray tower after-absorber
Turbulent Contact Absorber (TCA)
Marble-Bed Absorber
The venturi, manufactured by Chemical Con-
struction Co.. contains an adjustable throat that
permits control of pressure drop under a wide range
of flow conditions. Although a venturi is ordinarily
an effective particulate removal device, gas absorp-
tion is limited in limestone wet-scrubbing systems by
low slurry residence time. For this reason the after-
absorber was included for additional absorption
capability. The TCA, manufactured by Universal Oil
Products, uses a fluidized bed of low density plastic
spheres that are free to move between retaining
grids. The Marble-Bed Absorber, supplied by Com-
bustion Engineering Co., uses a packing of:1 i-inch
glass spheres (marbles). A "turbulent layer" of liquid
and gas above the glass spheres enhances mass
transfer and particulate removal. Figures 1, 2. and 3,
drawn roughly to scale, show the three scrubber
types along with the demisters selected for de-
entraining slurry in the gas streams.
The test facility was designed so that a number of
different scrubber internals and piping configurations
can be used with each scrubber system. For example,
the TCA can be operated as a one; two- or three-
CHEVRON
DEMISTER
DEM1STER
WASH
AFTER
SCRUBBER
INLET
SLURRY
THROAT
ADJUST-
ABLE
PLUG
VENTURI
SCRUBBER
EFFLUENT
SLURRY
APPROX. SCALE
Figure 1. Schematic of Venturi Scrubber and
After-Absorber
stage unit, and solids separation can be achieved
with any combination of clarifier. filter, centrifuge,
and pond.
A typical TCA system configuration used during
limestone testing is shown schematically in Figure 4.
Process details, such as flue gas slurry saturation
sprays and demister or Koch tray wash sprays, are
not shown.
For all configurations, gas is withdrawn from the
boiler ahead of the power plant particulate removal
equipment so that entrained dust (fly ash) can be
introduced into the scrubber. The gas flow rate to
each scrubber is measured by venturi flow tubes
and controlled by dampers on the induced-draft
fans. Concentration of sulfur dioxide in the inlet and
outlet gas is determined continuously by DuPont
photometric analyzers.
Control of the scrubbing systems is carried out
from a central graphic panelboard. An electronic
data acquisition system is used to record the operat-
ing data. The system is hard-wired for data output
-------
.GAS OUT
/ CHEVRON .>>»- \ /
DEMISTER "~~-jg£l ' \ 1
INLET KOCH
TRAY WASH *»*'
LIQUOR \r
Jr\
STEAM WASH
X
RETAINING /
GRIDS
GAS IN
A A A
°o ° o
00 °0
o ° o o
0 0°
0 0 0 ^
?a°^ll
^
V
i| EFFLUENT
i*^^-KOCH TRAY
i^ WASH LIQUOR
U/NLfT
SLURRY
MOB/LE
/ PACKING
/ SPHERES
APPROX. SCALE
GAS OUT
/^H ^\ CHEVRON
/ ^-^^ DEMISTER
7$^-^^$
DEMISTER ^^^^ v v v v
WMSH
w; FT ^^^ ^ TURBULENT
SLURRY Q-^ LAYER
-"« p"1^^ CLASS
/rVLET ^_^^ v V v cpHFRFS
SLUKKV '^^
^T ^^- EFFLUENT
GAS IN ^ \ / SLURRY
EFFLUENT
SLURRY
5'
APPROX. SCALE
\ EFFLUENT^T / \
V SLURRY y V
Figure 2. Schematic of Three-Stage TCA
Figure 3. Schematic of Marble-Bed Absorber
»«> 0I | |S
c Prwus
i>f -So/lrfs Composjfion
Gas Composition & Paniculate Loading
GoS N.'rram
Liquor Sirea^j
srrri/NG fo\n
Figure 4. EPA Test FacilityTypical Process Flow Diagram for TCA System
-------
directly on magnetic tape, and on-site display of
selected information is available. Also, important
process control variables are continuously recorded,
and trend recorders are provided for periodic mon-
itoring of selected data sources.
The Shawnee facility contains five major areas:
(1) the scrubber area (including tanks and pumps),
(2) the operations building (including laboratory
area, electrical gear, centrifuge, and filter), (3) the
thickener area (including tanks and pumps), (4) the
utility area (including air compressors, air dryer,
limestone storage silos, mix tanks, gravimetric-feeder,
and pumps), and (5) the pond area.
The scrubber area (looking toward the power-
house), scrubber control room, and the operations
building and thickener area are shown in
Figures 5, 6, and 7, respectively.
Figure 5. Scrubber Area
-------
Figure 6. Control Room
Figure 7. Operations Building & Thickener Area
-------
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 shown in Figure 8.
As indicated, the air/water and sodium carbonate
tests have been completed. As of June 1973, lime-
stone wet-scrubbing short-term factorial tests were
approximately 95 percent complete and longer-term
reliability verification tests were approximately 50
percent complete. In Table 1, a description of the
reports which are presently scheduled for general
distribution is presented.
AIR/WATER TESTING
These experiments, which use air to simulate flue
gas and water to simulate alkali slurry, were designed
to determine pressure drop model coefficients and
observe fluid hydrodynamics for all three scrubbers
under clean conditions.
SODIUM CARBONATE TESTING
These experiments, which utilized both high- and
low-concentration sodium carbonate solutions to
absorb SOa from flue gas, were designed to deter-
mine coefficients within the mathematical models
for predicting SOi removal.
LIMESTONE WET-SCRUBBING TESTING
The primary objectives of this test sequence are:
(1) To characterize, as completely as practicable,
the effect of important independent variables
on particulate removal and SOa removal.
(2) To identify and resolve operating problems,
such as scaling and demister plugging.
(3) To identify areas or regions for reliable opera-
tion of the three scrubber systems, consistent
with reasonable SCh removal, and to choose
economically attractive operating configura-
tions from within these regions.
(4) To determine long-term operating reliability
with attractive configurations for one or more
of the scrubber systems and to develop more
definitive process economics data and scale-
up factors.
To accomplish the first objective a large number
of short-term (4+ hours) factorial tests have been
made for each scrubber system. The test sequences
were full or partial factorial designs based upon the
chosen independent variables, their levels, and the
restraints of time as outlined in Figure 8. The choice
of the independent variables and their levels was
based on pilot plant test results, the restraints of the
system, and results from mathematical models that
relate the dependent and independent variables.
To accomplish the second and third objectives, a
relatively small number of longer-term (2 H- weeks)
reliability verification tests are being made on each
scrubber system. The variables being investigated
are: (1) percent solids recirculated, (2) effluent resi-
dence time, (3) gas rate, (4) scrubber inlet liquor
pH, and (5) demister types (e.g., plastic chevron vs.
stainless chevron). Solids separation tests for the
clarifier, filter, and centrifuge are also being made
on the three systems throughout the test period.
The fourth objective will be accomplished by
running reliability tests, lasting from 4 to 10 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 gas
rate, particulate loading, and SO-2 inlet concentra-
tions will also be studied.
LIME WET-SCRUBBING TESTING
This test series, which involves introduction of
hydrated lime (calcium hydroxide) directly into the
scrubber circuit, will resemble the limestone wet-
scrubbing test program. Again, tests will be divided
into three general categories (see Figure 8): short-
term factorial tests, longer-term reliability verification
tests, and long-term reliability tests.
ANALYTICAL PROGRAM
Samples of slurry, flue gas, limestone, and coal are
taken periodically for chemical analyses, particulate
size sampling, and limestone reactivity tests. Loca-
tions of slurry and gas sample points for theTCA are
shown in Figure 4.
To meet the formidable analytical requirements of
the facility at reasonable costs, equipment has been
selected that minimizes manpower. For example, an
x-ray fluorescence unit is used for comprehensive
slurry analyses. All analytical computations and
recording of results are handled by an on-site
minicomputer.
-------
TEST PROGRAM FUNCTIONS
SYSTEM CHECK- OUT
AIR WATER & SODIUM CARBONATE
TESTING
LIMESTONE WET SCRUBBING TESTING:
ShortTerm Factorial Tests
Reliability Verification Tests
Short-Term Factorial Tests
Re liability Tests
LIME WET-SCRUBBING TESTING:
Short-Term Factorial Tests
Reliability Verification Tests
Reliability Tasts
ENGINEERING & COST ESTIMATE
STUDIES
1972
MAMJJASOND
1 23456789 10
MMI
BOILER OUTAGE &
1973
J FMAMJ JASOND
11 12 13 14 15 16 17 18 19 20 21 22
"1 1
J
SYSTEM MODIFICATIONS
mmm
MM
1974
J I F M A M J
23J24 25 26 27 28
__
Figure 8. Shownee Test Schedule
-------
Table 1
PROGRESS AND FINAL REPORT DESCRIPTION*
Report Title
1. EPA Alkali Scrubbing
Facility: Sodium
Carbonate and Lime-
stone Test Results
2. EPA Alkali Scrubbing
Facility: Limestone
Wet-Scrubbing Test
Results
3. EPA Alkali Scrubbing
Test Facility: Lime
Wet-Scrubbing Test
Results
4. EPA Alkali Scrubbing
Test Facility: Final
Report
Information to be Included
Estimated
General
Publication
Date
August 1973
November 1973
Summary of operational problems and resolutions,
planned and actual test designs, results of air-water
and NaaCOs testing, utilization of data for model
development, results of factorial limestone testing
with interpretation of data.
Summary of operating problems and resolutions
associated with reliability verification testing, planned
and actual test designs, interpretation of data, status
of process model development, and selection of
parameters 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
reliability testing.
Summary of total test program with particular emphasis July 1974
on lime and limestone reliability test results, mathe-
matical models, scale-up design, and economic studies.
March 1974
* It is planned that EPA Capsule Reports will be issued which summarize each of the progress and final
reports. This Capsule Report summarizes the August 1973 progress report.
-------
In this section some of the significant SOa re-
moval results from the limestone short-term factorial
test sequences are presented graphically. As men-
tioned previously, the objective of the factorial tests
was to characterize, as completely as practical, the
effect of independent variables (e.g., liquid-to-gas
ratio) on paniculate and SOa removal.
A majority of the short-term (4-hours) factorial
tests were made at "high" scrubber inlet liquor pH's
(6.0-6.2) and, consequently, high stoichiometric
ratios (> 1.75 moles CaCCh added per mole SOa
absorbed). The data collected indicated that stoi-
chiometric ratio has an insignificant effect upon SOa
removal at values of inlet liquor pH greater than 6.0.
The data also indicate that SOa absorption increases
with decreasing inlet gas SOa concentration and/or
with decreasing scrubber liquor temperature. Care,
therefore, has been exercised in segregating these
noncontrolled independent variables in the
presentation of the data.
For the venturi scrubber, the SOa removals varied
between 30 and 40 percent for gas rates from 15.000
to 30,000 acfm and for liquid-to-gas ratios from 10
to 60 gal/mcf. SOa removal was not significantly
affected by pressure drop within the region inves-
tigated (6-12 in. HaO).
In Figure 9. the effect of gas velocity and L/G on
SOa removal for the four-header spray tower is
shown. For these data, the SOa removals are outside
the range of interest for commercially acceptable
gas velocities (> 7 ft/sec). An analysis of the data has
shown that an L/G of approximately 110 gal/mcf
would be required to achieve 80 percent SOa
removal at gas velocities greater than 7 ft/sec.
In Figure 10. the effect of gas velocity and L/G
on SOa removal for a three-stage TCA scrubber is
shown. For 95 percent removal, at a gas velocity of
9.8 ft/sec and an L/G ratio of 64 gal/mcf, the pres-
sure drop was approximately 7 in. HsO (excluding
Koch tray).
80
70
60 --
50--
30--
20
SOs INLET CONCENTRATION 2.5003 000 ppm
STOKHIOMETKIC RATIO 1 75 MOLES CoCQi MOLES SO?
ABSORBED
PERCENT SOLIDS 59%
HOLD TANK RESIDENCE TIME
SCRUBBER OUTLET LIQUOR TEMP = 120° F
2.5 FT SEC
3.7 f-T SEC
5.0 FT. SEC
75 !-'!' SEC
100
10
30 40 50 60 70
LIQUID-TOGAS RATIO, gal me/
80
90
Figure 9. Effect of Liquid-To-Gas Ratio and Gas
Velocity on SOa Removal in the Four-
Header Spray Tower
95 -
90- -
S5--
~j
| SO--
z
a
S 75--
fe
S 70- -
60--
55
9 8 FT SEC
\
78FT'SEC/ N 5.9 FT/PEC
.SO.' /MET CONCENTRATION 1.800-2.500 ppm
STOICHIOMETRIC RATIO 1 75 MOLES CoCO.i MOLES
SO.. ABSORBED
PERCENT SOLIDS 6 1H,
HOLD TANK RESIDENCE TIME 18 35 MIN
M 'K( WKEK OUTLETUQUOR TEMP 111 12F F
ll!:IGIITOFSI'HEHt.S :'i INCi H'S STAGE
20 30 40 50 60 70
UQUIDTOGAS RATIO. gaVmcf
80
90
Figure 10. Effect of Liquid-To-Gas Ratio and Gas
Velocity on SO2 Removal in the Four-
Grid Three-Stage TCA
-------
In Figure 11, the effect of gas velocity and L/G on
SOa removal for the Marble-Bed Absorber with five
inches of marbles is shown. For 80 percent removal
at a gas velocity of 7.7 ft/sec and an L/G ratio of 40
gal/mcf, the pressure drop was approximately 11 in.
H^O.
The data presented in Figures 9, 10 and 11 have
been replorted in Figure 12, where the effect of
slurry flow rate and gas velocity on SO-2 removal
for the spray tower. TCA and Marble-Bed Absorber
is shown. The data indicate that SOa removal is a
strong function of liquor rate and is only slightly
affected by gas velocity within the region
investigated.
KIM
S 604-
I
20
111
77I'I SEC
5.1 FT/SEC
si) /.\; rr n >.\< 7 .NTKAYTON 2.4003.200 ppm
ST( JI< V //< )M/ ( M<' KAT/0 -1 75 MOJ.KS CoCO < MOLES
SO: AKSOKKI fi
;/1\'( / .v/1.so;.//).s 5 7%
nor n /A.VK w..s/f w M :E T/M;-: rio M/N
M MV /{/!; /,' i)(/; ;:r LIQUOR TEMP n~>r25° F
-4-
-I-
H 1-
-4-
4-
30 40 50
I IQl 111) TO GAS RATIO.
-------
As mentioned previously, the objectives of the
limestone reliability verification tests are to: (1) iden-
tify and resolve operating problems (e.g., demister
plugging and scaling), and (2) identify areas or
regions for reliable operation, consistent with reason-
able SO2 removal and choose economically
attractive operating configurations from within
these regions.
A majority of these tests will be made at reduced
scrubber inlet liquor pH's (5.7-5.9) to increase system
reliability and limestone utilization. System reliability
can be improved at reduced pH's because of higher
oxidation rates, resulting in a larger percentage of
"seed" CaSO-4 crystals within the process slurry. An
increase in limestone utilization (decrease in stoichio-
metric ratio) results, of course, in a reduction in waste
mass solids and in limestone requirements. The
penalty for operating at reduced inlet liquor pH is a
modest reduction in SO-2 removal from high pH
performance.
A summary of the data from the initial limestone
reliability verification tests of the venturi, TCA, and
Marble-Bed Absorber systems is given in Table 2.
The highest limestone utilization (83 percent) and
SO removal (80-85 percent) was obtained with the
TCA system. The operability and reliability of all
systems were good during the duration of the testing
(e.g., no scale buildup), and the Koch Flexitray wash
tray (see Figure 2) was particularly effective in de-
entraining slurry from the exiting gas stream. Overall
material balances for sulfur and calcium were in
agreement to better than 12 percent for the three runs.
Reliability verification tests will be conducted next
for all systems at increased gas rates, decreased
effluent tank residence times, and decreased percent
solids recirculated in the process liquor. Also, the
venturi system has been modified to allow for in-
creased liquor flow to the spray tower. This will result
in higher SOa removal efficiencies for that system.
Table 2
A SUMMARY OF INITIAL RELIABILITY VERIFICATION RUNS
Parameters
Venturi and
SprayTower
Run501-lA
TCA
Run501-2A
Marble-Bed
Absorber
Run501-3A
Operating Time, hr 410 550
GasVelocity, ft/sec 5t 7.8
L/G,gal/mcf 8(T 80
Pressure Drop, in. HteO 10.5** 6
Percent Solids Recirculated 15 15
Eff luent Tank Residence Time, min 20 20
Percent SO2 Removal 70-75 80-85
Stoichiometric Ratio, moles CaCOs added/moles
SO2 absorbed 1.5 1.20
Limestone Utilization, 100 x moles SC>2 absorbed/moles
CaCOs added 67% 83%
Scrubber Inlet Liquor pH 5.8-5.9 5.8
Percent Oxidation of Sulfite to Sulfate 15 20-30
Dissolved Solids, ppm 7000 7500
520
5
53
9
11
30
65-70
1.25
80%
5.8
30
8000
t Spray tower.
*L/G's of 40 for spray tower and 40 for venturi.
*9 inches across venturi and 1.5 inches across spray tower.
-------
Generally, the operability and reliability of the three
scrubber systems has been good throughout the
short-term factorial testing and the initial phase of
the limestone verification testing. It should be em-
phasized that there is little evidence of sulfate
(CaSO-O scale for any of the tests to date. Sulfite
(CaSCh) scale was encountered during one series
of TCA tests and has been attributed to scrubber
operation at inlet liquor pH's> 6.3. Presently, more
severe operating conditions (e.g., lower effluent resi-
dence times) are being tested to determine the
regions of reliable operation for the three systems.
This section will highlight the most significant
results, to date, affecting the operability and reliability
of the systems. An evaluation of spray nozzle and
material reliability has not yet been completed and,
consequently, will not be reported here.
CLOSED LIQUOR LOOP OPERATION
Early in the short-term factorial test period it be-
came apparent that it was not feasible to operate
the test facility in a totally closed liquor loop without
modifications. For a closed liquor loop, the raw water
input to the system is nearly equal to the water
normally exiting the system in the humidified flue
gas and in the waste sludge transferred to the pond.
In an open liquor 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
water balance. In a commercial system, such a dis-
charge 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 dilutes liquors
returning to the scrubber, thereby tending to reduce
scaling and plugging. Open-loop operation was not
considered to be a serious problem during the short-
term factorial testing, since, at a specified scrubber
inlet liquor pH, SO-2 removal is not significantly
affected by liquor composition.
In order to allow for closed liquor loop operation
on all three systems, the following modifications
were made to the facility during the 5-week boiler
outage in February and March, 1973: (1) water seals
on the pumps were converted to mechanical seals
supplemented with air purge, (2) quench spray
systems using circulating slurry were provided for
the TCA and Marble-Bed Absorber, and (3) mist
eliminator and Koch tray wash systems using process
liquor plus raw water makeup were provided for the
scrubber systems.
DEMISTER OPERATION
During the short-term factorial test period, the
demister sections on the scrubber systems were pro-
vided with top (downstream) wash sprays only and,
consequently, there was a continual accumulation
of soft mud-like deposits on the demister blades.
In order to remedy this problem, the following
modifications were made to the systems:
(1) In November, 1972, a Koch Flexitray wash tray
was installed in the TCA scrubber between the
inlet liquor spray header and the chevron
demister (see Figure 2). At first, irrigation was
obtained with raw water. A subsequent modi-
fication in February, 1973, allowed for-irrigation
with process liquor, diluted with the available
raw water makeup.
(2) During the boiler outage in early 1973, the
spray tower and Marble-Bed Absorber demister
systems were modified to allow for washing
from both the upstream (underside) and
downstream directions with process liquor,
diluted with the available raw water makeup
(see Figures 1 and 3).
The Koch Flexitray wash tray has been successful,
to date, in eliminating solids buildup on the TCA
demister blades for gas velocities up to 8.8 ft/sec
and 15 percent solids recirculated. For the spray
tower, washing the demister from both the topside
and underside has been successful, to date, in
eliminating solids buildup for gas velocities up to
5 ft/sec and 10 percent solids recirculated. Some
difficulty is still being experienced with solids build-
up on the Marble-Bed Absorber demister blades.
This maybe attributed to "channelling" of the gas
through the marble-bed and the resultant high gas
velocities, due to partial pluggage of the bed.
REHEATER OPERATION
Flue gas is reheated after evolving from the
scrubber to increase plume buoyancy, prevent
condensation in the exhaust system, facilitate
isokinetic and analytical sampling, and protect the
induced draft fans from solid deposits and droplet
erosion. The reheaters employed are fuel oil com-
bustion units with a separate combustion air supply
and with combustion occurring in the flue gas flow
stream. During the short-term factorial testing, it had
been difficult to start and keep the reheaters operat-
ing and, during operation, combustion had been
incomplete leading to a visible plume containing
significant quantities of soot and oil. This led to
difficulties in interpreting outlet particulate data and
-------
affected gas sampling by the DuPont SCh photo-
metric analyzers. The difficulty appeared to result
from quenching of the flame due to the flue gas
flow before complete combustion could occur.
Modifications were made to the reheater systems
during the boiler outage in early 1973. Internal stain-
less steel sleeves were installed that provide approx-
imately 50 ft3 of isolated combustion zone for each
reheater. Also, the turbulent mixing type nozzles
supplied originally were replaced with mechanical
atomizing nozzles.
To date, the above modifications appear to have
been effective. Essentially no soot is visible in the
stack gas and the outlet particulate samples have
shown no evidence of carbon from the reheaters.
HOT GAS LIQUID INTERFACE
The hot flue gas must be cooled (humidified)
before entering the neoprene rubber lined spray
tower, TCA, and Marble-Bed scrubbers. During the
early stages of reliability verification testing, there
was a continual problem of solids buildup at the gas
humidification sections on the TCA and Marble-Bed
scrubbers. There has never been evidence of any
solids buildup within the venturi scrubber, which is
an extremely reliable gas humidifying device.
By careful selection of soot-blowing schedules and
humidification spray locations and flow patterns, the
solids buildup problem has been brought under
control. During the current series of limestone reli-
ability verification testing, there has been no evidence
of solids buildup at the hot gas-liquid interfaces for
the TCA and Marble-Bed scrubbers.
-------
For further information:
The detailed Progress Report on this project
is available from the National Technical Information
Service, Springfield, Va. 22151, as EPA Report
PH 22-68-67, first Progress Report: Limestone
Wet-Scrubbing Test Results at the EPA Alkali
ScrubbingTest Facility"
Or write:
Technology Transfer
Environmental Protection Agency
Washington, D.C., 20460
-------
\
LU
TECHNOLOGY
TRANSFER
------- |