United States
Environmental Protection
Agency
Industrial Environmental Research
Laboratory
Research Triangle Park NC 27711
Research and Development
EPA-600/S7-81-140 Sept. 1981
 Project  Summary
Demonstration  of  Wellman-
Lord/Allied  Chemical  FGD
Technology:  Final  Report and
Demonstration  Test Second
Year  Results
R. C. Adams, S. W. Mulligan, and R. R. Swanson
  Performance of a full-scale flue gas
desulfurization unit to demonstrate
the Wellman-Lord/Allied Chemical
process was evaluated for a period in
excess of 2 years. The Wellman-Lord/
Allied Chemical process is a regener-
able process employing sodium sulfite
wet scrubbing, thermal regeneration
of the spent scrubber solution, and
reduction to elemental sulfur of the
recovered SO2.
  Test program results indicate that
89 to 90% of the SO2 can be readily
removed from the flue gas in a long-
term dependable manner. Reliability
of the coupled absorber and regenera-
tion system  for the second year was
61%.  For the last 7 months, it was
74%. The major operating limitations
were  reduction unit problems, but
unscheduled outages of the evaporator
and the booster blower and start-ups
and shutdowns also contributed to
down time.
  As expected, the energy require-
ments  of the process, primarily for
thermal regeneration of the scrubber
solution and subsequent recovery of
SOa, were quite large, amounting to
12% of the boiler heat input derived
from fuel. Actualannualized operating
cost was 14.9 mills/kWh, using 1978
prices for raw materials and utilities.
Credits for the sale of byproduct sulfur
amounted to only 0.2 mills/kWh.
  The reported operation and per-
formance occurred after some modifi-
cation to the boi|er to increase inlet
flue gas temperature and after needed
improvements to the FGD plant
identified during initial operation were
implemented. Design limitations af-
fecting overall performance were lack
of redundancy, regeneration area
capacity of only 80% of full load,
underdesign of the purge solids dryer,
and limited turndown capability.
  This Project Summary  was devel-
oped by EPA's Industrial Environmen-
tal Research Laboratory, Research
Triangle Park, NC, to announce key
findings of the research project that is
fully documented in a separate report
of the same title (see Project Report
ordering information at back).

Introduction
  In 1972, the EPA entered into a cost-
shared contract with Northern Indiana
Public Service  Company (NIPSCO) to
design, construct, and operate a flue gas
desulfurization (FGD) plant that uses
the Wellman-Lord/Allied Chemical
(WL/A) FGD process. NIPSCO entered
into contracts with Davy Powergas (now
Davy McKee) to design and construct
the unit and with Allied Chemical (now
Allied) to operate the plant. The FGD
unit was retrofitted to NIPSCO's Mitchell
No. 11 boiler in Gary, IN.

-------
  The WL/A process developed by the
two design organizations is a regener-
able process, based on the recovery of
concentrated SOz and its  subsequent
reduction to elemental sulfur. The
product is sold to partially offset the
process costs.  This was  the first coal-
fired WL application, as well as the first
joint WL/A installation.
  To ensure that potential users are
fully aware of the commercial practical-
ity  of  the WL/A process, the EPA
conducted a test program that included:
  • A baseline characterization of the
    host boiler.
  • A performance  test required  by
    contract to demonstrate compliance
    with the performance guarantees.
  • Testing for a  2-year period to
    demonstrate long-term perform-
    ance and dependability.
This summary highlights  results and
conclusions of the test program.
  Principal objectives of the test pro-
gram were:
1. Verification of the reduction in pol-
   lutants achieved by  the  WL/A
   process FGD unit.
2. Validation of the estimated technical
   and economic  performance of the
   demonstration  unit.
3. Assessment of  the applicability of
   the WL/A  process to the general
   population of utility boilers.
The test program was designed to attain
these objectives to the maximum extent
possible.  Emphasis  on the various
objectives was sometimes redirected
due to test program  findings and
operating difficulties but, in general, the
program goals remained  unchanged.
The test design featured  continuous
monitoring of SO2 removal performance.
Evaluation of the data was in response
to the test objectives and  focused on the
dependability and  economics  of SOz
removal capability while operating the
boiler to provide the expected range and
variability of the flue gas properties with
a single coal type.

Overall Process Design
  Mitchell  No. 11 is a  115-MW pul-
verized-coal-fired, balanced-draft boiler
with cold-end  electrostatic precipitator
(ESP) particle  control. The boiler was
designed to use a coal with a nominal
sulfur content slightly  above 3% by
weight. The FGD  unit was designed to
accept flue gas at SOz concentrations
equivalent to this sulfur level in the coal.
Flue gas was fed to the FGD plant by the
boiler's two induced draft (ID) fans.
Before retrofit of the FGD plant, the flue
gas went to a stack shared with another
boiler.  A quick opening'damper  was
installed  in  the duct to that stack to
bypass the FG D pla nt when not operati ng
and to protect the boiler from damage
during upsets.  Normally, the FGD plant
operated  with the bypass  damper
closed.
  The WL/A FGD process removes S02
from the flue gas  stream by scrubbing
with an aqueous sodium sulfite solution
and subsequent thermal  regeneration
to recover the SO2. The solution is free
of solid material. The liberated SO2 is
then reduced to elemental sulfur which
is sold. The  FGD unit was designed to
remove 90% of the SOz delivered with
the flue gas at flue gas rates equivalent
to a boiler load of 92 MW (80% of full
boiler load).  The processes are propri-
etary designs of Davy McKee and Allied.
Logical separation of the  various
process steps are:
  • SOz absorption - Davy.
  • SOz  recovery and  scrubber solu-
    tion regeneration - Davy.
  • Purge treatment - Davy.
  • SO2 reduction - Allied.
The following description  is  based on
Davy and Allied non-proprietary design
data.

  Figure 1 shows the process steps. The
FGD plant accepts the  total flue gas
stream from  the  discharge of the
boiler's ID fans using a  booster fan to
overcome the flow resistance through a
prescrubber and an absorber. The pre-
scrubber  is  a  single-stage  orifice
contactor designed to remove paniculate
matter and cool the flue gas before the
S02 absorption step. A  pump recircu-
lates the scrubber water from a sump
back to the contactor. In order to control
solids buildup in  the  liquid  stream, a
purge stream is withdrawn and makeup
water is added to the prescrubber to
compensate for this loss as  well as to
humidify the flue gas. The purge stream
is sent to the power station's fly ash
settling ponds. Particles not removed in
the prescrubber are removed with a
filter in the spent absorbing solution line
leaving the  absorber. The wash water
from periodic  washing of this filter is
also discharged to the power station's
fly  ash settling ponds. These are the
only waste streams  expected to be
discharged from the FGD plant.
  The cooled, humidified flue gas leaves
the prescrubber and enters the bottom
of a three-stage absorber where the gas
is contacted with the  sulfite solution
flowing  countercurrent  to  the gas
stream.  The solution absorbs the SO2
and the treated flue gas, saturated at
about 54.4°C (130°F), is then discharged
to the atmosphere through an integrally
mounted stack. Direct-fired reheat of
the flue gas with natural gas as fuel was
provided  but never  used  because of
limited gas supplies.
  The absorber is a three-tray column.
An absorber demister pad above the top
contact  stage  prevents  entrained  ab-
sorber solution from being exhausted
with the treated gas. Each contact stage
of the absorber hasa separate recircula-
tion system to promote good gas/liquid
contact.  The absorbing process is con-
ducted at about 54.4°C (130°F), and
spent absorbing solution is withdrawn
from the  bottom contact stage.  Re-
generated absorbing  solution is added
to the top contact stage. The process
chemistry for absorption  is:

   SOz + Na2S03 + H20 - 2NaHSO3

  The bisulfite-rich  solution is then
passed to the evaporator-crystallizer
unit. The evaporator-crystallizer is a
single-effect forced circulation unit. The
heat exchanger employs  steam and the
clean condensate is discharged  for
reuse by  the  host  boiler.  The heat  ,
supplied to the liquor decomposes the |
sodium  bisulfite  solution  to sodium
sulfite (which crystallizes out), SO2, and
water:
           heat
  2NaHSO3
Na2S03 + S02 + H20
The SO2 and water vapor are discharged
overhead from the evaporator. The wet
gas stream is cooled to condense and
separate the water vapor from the gas
stream, providing a S02 feed stream of
low humidity for the SO2 reduction unit.
The condensate that was removed is
stripped of dissolved S02 a nd is added to
the sodium sulfite slurry discharged
from the evaporator salt leg. The sodium
sulfite and condensate, along with
makeup sodium carbonate, are mixed to
provide a solids-free  sodium  sulfite
solution suitable for reuse  in the
absorber.
  Oxidation of some  of the  sodium
sulfite during the absorption  step  to
form sodium sulfate is  unavoidable:
     Na2SO3 + O2
       2Na2SO4
A portion of the spent absorber solution
leaving the absorber is passed through a

-------
                                                                           Treated Flue Gas
 Figure  1.
                           \FGD ProceddBoundary

Block flow diagram of major process steps.
                                                                                                                 J
purge treatment system to separate and
dry the NasSCh for eventual sale. The
purge treatment unit was designed to
minimize the amount of purge and to
produce a salable byproduct of sodium
sulfate. Purge requirements are reduced
by  subjecting part  of the  absorber
product stream to a fractional crystalli-
zation process. Sodium sulfate crystals
are produced by  chilling  the solution
under  very specific  conditions. This
treatment allows most of  the active
sodium compounds (sulfite and bisulfite)
to remain in the mother liquor which is
separated  from the  crystals by a
centrifuge.  This mother liquor is then
returned to  the main process and the
crystals  are dehydrated  in a drying
system.  The final product is a dry
sodium sulfate suitable for  marketing.
Purge treatment requirements depend
on the amount of sodium sulfate and,
possibly, sodium thiosulfate formed in
the process  and on the allowable con-
centration in the absorbing solution of
these inactive components. The amount
of sodium  sulfate formed  may be a
function of the amount of excess air in
the flue gas.
  The  S02  discharged  overhead from
the evaporator-crystallizer is reduced to
                           elemental sulfur during the S02 reduc-
                           tion step, a two-stage process employ-
                           ing a primary catalytic reaction of SOa
                           with natural gas to produce sulfur and
                           some hydrogen sulfide(H2S) followed by
                           a secondary Claus conversion system in
                           which the H2S is reacted with residual
                           S02 to produce additional sulfur. The
                           primary reaction system consists of two
                           packed-bed regenerative heat exchangers
                           and a catalyst-packed reduction reactor.
                           These  vessels  and their connecting
                           flues are refractory-lined for protection
                           against high temperatures and corrosive
                           gases.
                             The regenerative heat exchangers
                           remove the heat from the gases leaving
                           the reactor, and utilize this heat to raise
                           the temperature of the gases entering
                           the reactor. At appropriate intervals, the
                           duties  of the  two regenerative  heat
                           exchangers are alternated;  i.e., the
                           packed bed heat exchanger that was
                           heating the entering  gases becomes a
                           cooler for the gases leaving the reactor.
                             Essentially half of the S02 in the feed
                           is reduced to elemental sulfur by direct
                           reaction with the reducing gas:


                             2S02 + CH4 —»• C02 + 2H20 +  2S
Simultaneously, most of the remaining
S02 is converted into H2S and additional
sulfur by a similar reaction:
SSOs + 2CH* —*• 2C02 + 2H20 + 2H;
The hot gases from the primary reaction
system pass through condensers where
the sulfur is  removed and sent to
storage. The gas  flows and  operating
conditions in the primary reaction
system are carefully controlled so that
the mixture of H2S and unconverted SOa
in the product gases from the primary
reaction system  closely approximates
the ideal volumetric ratio (two parts H2S
to one part SO2) required for the sub-
sequent Claus reaction.
     2H2S + S02
At the same time, maximum utilization
of the reducing agent is achieved and
the formation of undesirable side-reac-
tion products is minimized.
  The Claus con version system is a two-
stage unit. Associated with the Claus
unit  is an  interstage  condenser from
which elemental sulfur also passes to
storage. A  final condenser follows the
second stage converter to recover the

-------
last portion of the elemental sulfur
product.
  The tail gases from the SOz reduction
unit pass through an incinerator where
natural gas is burned in air to oxidize the
remaining H2S to SOz. The hot gases
from the incinerator are admixed with
the untreated flue gases at the booster
blower inlet, thus avoiding having to
discharge a stream of small volume but
containing a relatively high concentra-
tion of S02.

Design Limitations
  The FGD demonstration would  have
benefited from a more conservative
design. Deficient performance was due
partially to design related causes. The
most  significant design decisions
affecting performance were:
  • The  booster  blower and  the ab-
    sorber were designed for flue gas
    flows in excess of boiler full-load
    flow. The absorber was designed to
    remove the expected  amount of
    SO2 at boiler full load. The effect on
    performance was positive.
  • Capacity for recovery  of the SOa
    (evaporator, purge treatment, re-
    duction) was based on an expected
    load factor of 80% of full  load.
    Surge capacity provided by storage
    tanks for absorber feed solution
    and  absorber spent solution was
    limited to about 4 hours. The effect
    was to  limit either the  amount of
    S02 removed or maximum utiliza-
    tion  of  the boiler.  The  latter pre-
    vailed  because  of a  policy  to
    operate with  a closed  bypass and
    the boiler limited to 80% load factor
    (92 MW).
  • Baseline testing revealed that flue
    gas flow at a given load exceeded
    the design flue  gas flow signifi-
    cantly. High excess air levels in the
    flue gas and  additional flue gas
    due to higher-than-expected boiler
    heat rates were identifiable causes.
  • The  FGD plant was designed with
    virtually no redundancy as installed
    spares.
  • Initially, the evaporator circulating
    pump had  a  steam-turbine  drive.
    Loss of high pressure steam during
    boiler shutdowns delayed  start-
    ups  because slurry in the evapo-
    rator  had to be  removed and
    diluted. The  deficiency was cor-
    rected after 1 year of demonstra-
    tion with the installation  of an
    electric-motor drive.
  • The purge solids dryer was under-
    designed or a misapplication and
     this prevented  full recovery of
     sodium  sulfate from  the  purge
     stream.
  • The reduction unit had less turn-
     down capability than the absorber/
     evaporator.
  • The FGD plant  was designed to
     take flue gas at 149°C  (300°F).
     Initially, actual temperatures were
     substantially lower due in part to
     cooling of the flue gas by in leakage
     air. Some of the baskets  were
     removed from the air  heaters to
     provide flue gas at temperatures of
     149°C(300°F) and above.
Test results that were design-related, as
opposed to process related causes, have
been identified as such in the discussion
of test results that follows.

Conclusions
  The WL/A  demonstration test pro-
gram consisted  of  three major test
phases:
  •  Baseline  testing.
  •  Acceptance tests to verify per-
     formance guarantees.
  •  Two-year demonstration test pro-
     gram.

Baseline Testing
  Major baseline  testing occurred in
two time-separated stages. The baseline
test was conducted prior to completion
of construction of the FGD plant. Flue
gas  characterizations correlated  with
boiler operating settings showed that
some of the flue  gas properties, par-
ticularly flow and temperature, differed
from those used  to  design the  FGD
plant. The differences had a profound
effect on  the criteria for acceptance
testing, and the baseline data were very
useful for defining those criteria. The
data also were useful for defining the
range of testing.
  Following completion of the first year
of the  FGD  process demonstration,
baseline tests  were repeated with
several objectives:
  •  Establish an  updated  boiler per-
     formance baseline for comparison
     with boiler performance when the
     FGD plant is operating.
  •  Obtain an updated characterization
     of the flue gas leaving the boiler.
The  FGD  plant was down  and com-
pletely  isolated from the boiler. The
collection  of  additional baseline data
was necessary  to establish to what
extent a substantial rebuild of the boiler
since the first baseline test and a
change in the type of  coal being burned
had  affected  the  performance of the
boiler.  Flue  gas volumes, flue gas
temperatures, and coal quality were of
particular  interest  because of their
impact  on FGD  operation and per-
formance.

Flue Gas Volume
  Flue gas volumes were 27%  higher
than that used for FGD plant design,
151  mVs at  148.9°C  (320,000  cfm at
300°F). An excess of inleakage  air
appears to be the major contributor to
the high flue gas volumes, but heat
rates that were higher than the new
boiler design heat rates probably had
some effect. Factors contributing to the
higher heat rates were high turbine
steam rates and, at low load factors,
combustion air in excess of the operating
set point. Both the high excess air and
the high steam rates add to the volume
of flue gas that the FGD plant must treat.
The high steam rates would be expected
to add to the  amount of  SOa  to  be
removed.

Flue Gas Temperature
  Flue gas temperature  averaged
148.9°C (300°F) at all load levelsduring
the second baseline  test. This was
substantially higher than the tempera-
tures measured during the first baseline
test. The higher temperatures were due
to the removal of  part  of the heat
transfer surfaces of the air preheaters in
a  deliberate  attempt to  raise the
temperatures above the dew point and
thus prevent the scaling and corrosion
that was occurring at  the FGD booster
fan. The temperatures  were  well above
the sulfuric acid dew point.

Coal Quality
  The  high sulfur (—3%)  coal  used
during  FGD  operation was burned
during the second baseline test and
during subsequent demonstration test-
ing. The coal sulfur content  was 0.2 to
0.3 wt. % lower than  that of the coal
burned during the first baseline test.
Except for expected  differences in
variability of the ash and moisture
contents, the quality of the two coals
was about the same.

Acceptance Test
  Process  performance guarantees
were met or exceeded  as confirmed by
acceptance testing.  The boiler was
operated to provide the flue gas flows
used to design the FGD plant. Despite
flue gas dilution from  high  excess  air
levels that was found  during baseline
testing, the SOa feed rates equalled or

-------
exceeded  the design expectation  of
0.6101 kg/s(4842 Ib/h).
  Specific performance criteria were
met or exceeded:
  1.  S02  removal  of 90%,  2-hour
     average, or better was achieved
     for 261 hours at flue gas flows of
     151 mVs (320,000 cfm) or higher
     and was achieved for 84 hours at
     about 183 mVs (388,000 cfm) or
     higher.
  2.  Paniculate  emissions did not
     exceed 40 ng/J (0.1 lb/106Btu)of
     boiler heat  input at either the
     lower (design load) or higher load
     conditions.
  3.  The consumption of steam, natural
     gas, and electrical power averaged
     76% of the performance guarantee
     requirements at design load con-
     ditions.
  4.  Soda ash consumption averaged
     less than 0.069 kg/s (6.6 tons of
     sodium carbonate per day) which
     was the limit set in  the perform-
     ance guarantees that was not to
     be exceeded during the  design
     load period.  Sodium  carbonate
     consumption  of 0.069 kg/s (6.6
     tons/day)  is equivalent to 0.152
     mol Na/mol S removed from the
     flue  gas at  a SO2  feed rate of
     0.6101 kg/s (4842 Ib/h) and 90%
     removal.
  5. Sulfur product purity was greater
     than 99.5%  at both design- and
     high-load conditions.

Two-Year Demonstration
Program
  The test program, as originally
planned, was to monitor performance of
the FGD plant for 1 year. The FGD plant
achieved only 90 days of  accumulated
operation during the first year due in
part  to boiler operating problems. The
principal boiler problems that prevented
FGD operation were unstable flue gas
flows and steam pressures. They were
the  result  of  poor coal  quality, that
exacerbated some problems with the
coal  feeding equipment, and of boiler
feedwater quality problems. Problems
were also encountered at the boiler/FGD
interface, in particular, booster blower
and  damper problems.  A midyear
review resulted in the initiation  of a
plant improvement program with  the
goal  of correcting the major problems.
The program was targeted for comple-
tion during a scheduled boiler shutdown
which coincided with  the end of the 1 -
year demonstration. The test program
was continued for another full year to
evaluate the effects of the boiler and
FGD plant modifications.
  Table 1  lists the major improvement
projects. The  need for these  improve-
ments are identified with the following
FGD plant process and design  limita-
tions:
  • Steam and flue  gas fluctuations
    limited  or prevented  reliable
    operation.
  • Flue gas temperatures below the
    acid dew point caused unbalancing
    of the booster blower and (ulti-
    mately) severe corrosion/erosion
    damage to the fan blades.
  • The guillotine-type damper installed
    to isolate the FGD plant from the
    boiler during  shutdowns became
    inoperable, after  binding from the
    accumulation of aggregated fly ash
    along the tracks.
  • Energy supply to drive the evapora-
    tor Circulating pump  must  not be
    interrupted. Originally, this drive
    was a steam turbine operating on
    high  pressure steam supplied by
    the boiler. During a  boiler shut-
    down and with  no evaporator
    circulation, slurry in the evaporator
Table 1.     Plant Improvement Projects

      Item          Date completed
                          had to be drained  and diluted at
                          considerable penalty in additional
                          startup time required.
                        • Sulfur condenser  leaks  were a
                          recurring problem throughout the
                          2-year demonstration.
                      Test  program  results and  operating
                      experience during the second year
                      showed a  substantial improvement in
                      FGD performance; since boiler utilization
                      was  high,  the results more  nearly
                      duplicated  those expected during com-
                      mercial application. Flue gas character-
                      istics were essentially unchanged from
                      those  of the second baseline test. The
                      conclusions that follow are based on
                      second-year results.

                      Dependability
                        Reliability of the FGD  unit (hours
                      operated/hours called upon to operate)
                      was 61%.  However, FGD plant utiliza-
                      tion  was  consistent for  the  last 7
                      months: reliability averaged 74% for
                      that  period. Only full operation was
                      counted as hours  operated, full opera-
                      tion being integrated operation of the
                      absorber/evaporator loop with  the
                      reduction unit when the flue gas bypass
                      was closed. The purge treatment unit
                                   Action
 Coal supply


 Air heater



 Duct insulation


 Blanks
Completed June 78
During September 78
shutdown
After September 78
shutdown

During September 78
shutdown
 Booster blower   Not completed
 Evaporator pump


 Absorber


 Booster blower
  turbine

 Sulfur condenser
During September 78
shutdown

During September 78
shutdown

After September 78
shutdown

During September 78
shutdown
Provide an uninterrupted supply of
Captain coal for Mitchell No. 11 use.

Remove part of baskets which provide
heat storage, to raise inlet duct
temperature.

Insulate duct before and after booster
blower.

Provide way to install blanks rapidly
at inlet of booster fan as an alternative
to the isolation damper.

Install a sparger pipe in the booster
blower to periodically steam clean
blades  while running.

Install an electric motor as an
alternative to steam turbine drive.

Recoat and repair leaks.
Provide enclosure to protect against
SOz and weak acid attack.

Plug leaking tubes.
 Note: Dates indicated  were during first  year  of demonstration test  program,
      September 1977-September 1978.

-------
may or may not have been operating.
The reliability record was established
with virtually no redundancy built into
the FGD unit. Also, the evaporator was
designed for the equivalent of only 80%
of full boiler load. With limited surge
capacity in the regeneration loop, the
FGD plant was unable to effect complete
SOz recovery during  evaporator or
reduction  unit shutdowns.  The 'FGD
plant was utilized (hours operated/hours
in period) 56% of the time. The downtime
was due to:
  • FGD repair - 33%.
  • Start-up and shutdown - 5%.
  • Boiler down - 6%.
The S02  absorber was  essentially
trouble free. The reduction unit required
the most downtime for repairs, followed
by the evaporator circulating pump and
the booster blower (Table 2).  Called
upon time  is defined as the time the
boiler operated to provide flue gas and
utilities and the time other feed streams
were  available  within the specific
design criteria of the FGD plant. These
include:
  • Flue gas at rates not less than 46
    MW equivalent.
  • Stable steam pressures within an
    operable range around a  design
    pressure of 3,790 kPa (550 psig).
  • Electricity.
  • Natural gas.
  • Soda  ash.
  • Boiler stable  within  limits of
    greater than 46 gross MWe, and
    coal sulfur content greater than 2.8
    and less than 3.5 wt. %.

SO2 Removal
  The process was controlled to remove
89% of the inlet SO2. Thirty-day average
removal efficiencies varied from 88% to
93% (Figure 2). The 24-hour and 1 -hour
average data  for  removal efficiency
were:
                      Table 2.    Reasons for Interruption of Operations
       SOz
     removal
    Percent of
  time operated
 24-h      1-h
average  average
  90% and greater   60       52
  89% and greater   84       78
  85% and greater   97       97

 On  a boiler  heat  input  basis, SO2
 emissions were controlled in the range
 of 110 to 400 ng/J (0.25 to 0.94 lb/106
 Btu).
   S02 removal was attained at electrical
 generating outputs in the range of 53 to
 85 MW of the 115 MW boiler. The lower
Equipment or
reasons
Reduction unit
Evaporator circulating pump
Booster blower
Start-up and shutdown
Other equipment, including absorber
Evaporator
Days of
interruption
59
28
18
17
10
8
% of called
upon time
17
8
5
5
3
2
                      limit was set by the limiting turndown
                      capability of the reduction  unit.  The
                      upper limit was set by the 80% capacity
                      limitation of the evaporator, as designed.
                      Because a substantial amount of energy
                      (primarily as boiler main  steam)  was
                      consumed by  the FGD plant,  the
                      generating potential of the boiler  was
                      actually  about  95 MW at the FGD
                      maximum capacity limit  of 85 MW.
                      "Generating potential" refers to the
                      gross megawatts  that the boiler  is
                      capable  of generating  during FGD
                      operation but cannot attain due to the
                      boiler main  steam consumed by the
                      FGD plant.

                      Confirmation of Design
                      Limitations
                        Design limitations were identified  in
                      the Introduction. The  effect on FGD
                      plant performance was:
                        • Performance at maximum design
                          flue  gas flow rates and SO2 feed

                       WO
                        90
                                        o

                                       I

                                       §«>
                        70
                                            05
3
c
.0)
.G
1
                 rates  could not be adequately
                 tested because of the design
                 limitations  of  the  regeneration
                 system. Flue gas volumes  up to
                 95% of full boiler load were treated
                 successfully during the acceptance
                 test. During the demonstration test
                 program, the maximum  SOz feed
                 rate that was sustainable for 24
                 hours was 0.750 kg/s (5950 Ib/h),
                 exceeding design expectations.
                 Capacity of the recovery unit met
                 design expectations. The recovery
                 unit successfully processed 90% of
                 an SOz feed rate of 0.588  kg/s
                 (4670  Ib/h) at a flue gas flow rate
                 of 152 mVs (323,000 acfm). The
                 boiler generating potential was 89
                 MW and actual gross  electrical
                 generating output  was  80  MW.
                 This output  was sustained for 65
                 hours  during maximum load tests|
                 Baseline flue gas flows  were
                 confirmed. At a generating poten-
                                     £_	Desigt
                           Oct  Nov  Dec  Jan  Feb  Mar  Apr May  Jun  Jul  Aug   Sep Oct
                                     1978 1979
                                                        Period

                       Figure 2.    Thirty-day average SOz removal.

-------
    tial of 92 MW, the design point, the
    flue  gas flow rate was 170 mVs
    (360,000  acfm). This was 13%
    higher than design expectations.
    Oxygen in  the flue  gas averaged
    8.0% by volume, compared to an
    expected oxygen level  of 5.6%.
  • Installation of an electric drive for
    the  evaporator circulating  pump
    reduced start-up time significantly.
  • Tests of absorber turndown, with
    the  reduction  unit  not operating,
    established a minimum throughput
    equivalent to a generating poten-
    tial of 50 MW. The limiting factor
    was a flue gas flow rate of 100
    mVs (220,000 acfm), limited by
    the  minimum  governor setting of
    the  steam turbine drive  of the
    booster blower. The minimum
    sustainable load at full  operation of
    the FGD plant, with the reduction
    unit operating,  was  a 62  MW
    generating  potential.  Ninety per-
    cent S02 removal was achieved for
    4 days at an average S02 feed rate
    of 0.426 kg/s (3380 Ib/h).

Process Economics
  The projected annual operating cost
in 1978  dollars was 13.2  mills/kWh.
Actual  annualized cost for  the second
year of the demonstration,  using 1978
prices for- raw materials and utilities,
was 14.9 mills/kWh. Byproduct sulfur
production  averaged 1.81  kg/s (17.3
ton/day).  Credits  for the  sale  of the
sulfur amounted to only 0.2 mills/kWh.
Actual  annualized cost  was based on
the 82.2  MW of generator  output that
was possible at the FGD design capacity
of 92 MW, assuming 351 days of boiler
operation per year. On this basis, the
actual  boiler capacity in total kilowatt
hours was 92% of the projected capacity.
Annualized costs are quite  sensitive to
lower-than-projected capacities because
fixed costs,  about 50% of  the annual
costs,  and  labor costs continue to
accrue whether or not the boiler or the
FGD plant is operating at full capacity.

Energy and Raw Material
Consumption
  A significant  amount of  steam pro-
duced by the boiler was consumed by
the FGD plant,  used primarily  by the
evaporator to recover S02  and  regen-
erate the scrubber solution (Figure 3).
Electrical power consumption amounted
to about 1  MW after the evaporator
circulating pump had been converted
from steam-turbine to electrical drive
for improved operability. Actual average
steam consumption  was 105%  of
design expectations. The energy equiv-
alent of this steam was 11% of the boiler
input  energy derived  from fuel. Since
the average generating output of the
boiler was 77 MW, the equivalent loss
in electrical  generating capacity
amounted to an 8% derating  of the
boiler from a nameplate capacity of 115
MW.  Including 1  MW of electricity
consumed, the total energy requirement
was 1*2% of the boiler heat input derived
from fuel at an average boiler load of 77
MW.
  Soda ash was  used  as  makeup
sodium carbonate for  the scrubbing
process. Makeup is made necessary by
the buildup of  inactive  constituents in
the absorber/evaporator loop, such  as
sulfate and thiosulfate, that must  be
purged. Any loss from the system due to
leaks also would  require soda ash
makeup.  High  soda ash consumption
during the first demonstration year was
due to leaks at the bottom collector tray
of the absorber that were  repaired
before commencing the second demon-
stration year. Average daily consump-
tion of soda ash for the last 7 months of
operation was 0.091 kg/s (8.7 tons/day),
using the total operating days of the
absorber/evaporator as the time base.
Soda  ash consumption as a function of
S02 removed was 0.217 mol Na/mol S
removed. The performance  guarantee
for  acceptance  was 0.069  kg/s (6.6
tons/day) at the design levels for flue
gas flow and  inlet SOj. Soda ash
consumption of 0.069  kg/s  (6.6 tons/
day) is equivalent to 0.152 mol Na/mol
S removed at a design feed rate of S02 of
0.6101  kg/s (4842  Ib/h)  and 90%
removal.
  Natural gas was used as the reductant
for  converting the SO2  to  elemental
sulfur. It also was the  fuel  used to
incinerate the tail gas emitted from the
reduction process.  The  tail gas was
returned to the  inlet of  the absorber
after incineration. It was necessary to
continue  incinerator operation during
shutdowns to  destroy  the reduced
sulfur forms that desorb  from the
reduction unit  refractory materials.
Thus, there  was a  corresponding
improvement  in unit consumption of
natural gas with improvement in reli-
ability. About 0.2 m3 (7 ft3) of natural gas
was  consumed  per pound  of sulfur
produced which was in accordance with
the design expectations. During opera-
tion, the incinerator consumed 7.5% of
the natural gas. In contrast, the inciner-
ator consumed  over  12% of  the gas
overall because  it continued  to operate
during shutdowns, demonstrating the
                                                                     Design
                                                                  11.5 Ib/lb
     Oct  Nov  Dec   Jan  Feb Mar  Apr May  Jun  Jul  Aug  Sep  Oct
              1978 1979
                                  Period


Figure 3.    Comparison of actual steam consumed as a function of SOz feed rate
            with design value.
                                                                       U. S. GOVERNMENT PRINTING OFFICE: 198I/559-092/33!0

-------
    result of a 61% FGD plant  reliability
    factor

    Purge  Treatment
    Considerations
      The purge unit, as initially designed,
    was to  have  treated a small purge
    stream removed from the regeneration
    loop, to separate sodium sulfate  from
    most of the sulfite/bisulfite components,
    and to dry the sodium sulfate to produce
    a marketable product. The "wet" end of
    this purge treatment system performed
    satisfactorily. The limitation of the
    purge dryer already has  been men-
    tioned.
      The amount of purge to be treated is a
    function of the formation of sulfate and
    possibly thiosulfate. Attempts to deter-
    mine the amount of sulfate formation
    during absorption were frustrated by an
    inability to obtain correct flow measure-
    ments and  uncertainties about the
    specific  water  balance  across the
    absorber. However, the data  seem to
    indicate  that  sulfate formation  is a
    function of the oxygen concentration in
    the flue gas. Thus, higher  than design
    purge rates might have been due to the
    high excess air levels in the flue gas.
    The average purge rate* for the last 7
    months of operation has been estimated
    to be 18.2%  to 24.8%,  substantially
    higher than expected. The estimate was
    determined from soda ash consumption
    and  the calculated  amount of S02
    removed. A purge rate of about 10% was
    the value indicated during the design
    phase of the project. In summary, the
    information  seems to  indicate actual
    purge rates much higher than design.
          Purge rates of this magnitude put a
        further load on the purge solids dryer.
        Dryer tests performed by Davy McKee
        determined that the dryer did not have
        the  needed capacity,  even at design
        rates.  The  maximum  dryer capacity
        achieved during the test, approximately
        66% of the design heat duty, could not
        be sustained because of a buildup of
        solids at the discharge end of the dryer.
        Maximum capacity that could be sus-
        tained without this buildup was only 45-
        50% of design.
          Davy McKee is investigating the use
        of an antioxidant that shows promise for
        reducing the oxidation rate and, thus,
        the amount of sodium values that must
        be purged as sodium sulfate.


        Recommendations
          Overall  performance of  the  FGD
        demonstration unit was affected signifi-
        cantly by design limitations that were
        known but not eliminated for various
reasons. Typically, capital cost saving
funding limitations, and need for furthi
development are the incentives for le;
conservative design. For installatior
demonstrating  new FGD technolog
design criteria should be established .
the start of the program that focus c
the advantages and limitations  of th
process rather  than having to  repo
poor performance solely because <
design limitations.  The WL/A demor
stration plant was design-limited by lac
of regeneration capacity and by almos
complete lack of installed  spare;
Improved  performance would be e>
pected with a more conservative desigr
It  is recommended that full or exces
capacity and redundant equipment b
designed into future demonstrations t
the  maximum  extent possible.  Th
demonstration test and evaluatio
would  have to  indicate  the degree c
overdesign and associated costs, if an\
so  that installed costs relative t
performance are demonstrated.
    •Purge rate was determined as the ratio of moles
    sodium consumed to moles SO? removed from flue
    gas, expressed as a percentage
           R. C. Adams andS. W. Mulligan are with TRW, Inc., Research Triangle Park, NC
             27709; R. R. Swanson was with Aide,  Richmond, VA 23225.
           Norman Kaplan is the EPA Project Officer (see below).
           The complete report, entitled "Demonstration of Wellman-Lord/Allied Chemical
             FGD Technology: Final Report and Demonstration Test Second Year Results,"  (
             (Order No. PB 81-246 316; Cost: $29.00, subject to change) will be available
             only from:
                  National Technical Information Service
                  5285 Port Royal Road
                  Springfield, VA 22161
                   Telephone: 703-487-4650
           The EPA Project Officer can be contacted at:
                  Industrial Environmental Research Laboratory
                  U.S. Environmental Protection Agency
                  Research Triangle Park, NC 27711
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
                Postage and
                Fees Paid
                Environmental
                Protection
                Agency
                EPA 335
Official Business
Penalty for Private Use $300

RETURN  POSTAGE GUARANTEED

                                                                   e30  S  Ut
                                                                   CHICAUU  IL
                                                       srhEer

-------