United States
                   Environmental Protection
                   Agency
                   Research and Development
  Air and Energy Engineering
  Research Laboratory
  Research Triangle Park, NC 27711
  EPA/600/SR-94/184  November 1994
sorEPA      Project  Summary

                   Demonstration  of  Sorbent
                   Injection Technology  on  a
                   Tangentially  Coal-Fired  Utility
                   Boiler (Yorktown  LIMB
                   Demonstration)
                   J. P. Clark, R. W. Koucky, M. R. Gogineni, A. F. Kwasnik
                    Limestone  Injection  Multistage
                  Burner (LIMB) technology has been
                  successfully demonstrated on a tan-
                  gentially fired coal-burning utility boiler,
                  Virginia Power's  180 MWe Yorktown
                  Unit No. 2. This document summarizes
                  the  activities  conducted, and results
                  achieved,  under this EPA-sponsored
                  demonstration program. LIMB  com-
                  bines furnace injection  of a calcium-
                  based sorbent for moderate reductions
                  of sulfur  dioxide  (SO2)  with  a
                  low-nitrogen-oxide (NOX) firing system
                  for NOX emissions reduction. The pro-
                  cess is attractive for retrofit of existing
                  coal-burning utility boilers, since the
                  capital equipment requirements and
                  overall sulfur reduction costs per ton
                  of sulfur oxide removed are less than
                  for most other options, such as wet
                  flue gas desulfurization.
                    Testing was conducted on an east-
                  ern bituminous coal with a typical sul-
                  fur content of 2.3%. Five sorbents were
                  tested: commercial hydrated lime, with
                  and  without calcium lignosulfonate
                  treatment, each from two suppliers, and
                  finely pulverized limestone.  Short-term
                  parametric testing showed full-load SO2
                  removals of up to 56  and  63% at
                  calcium-to-sulfur (Ca/S) ratios of 2:1
                  and 2.5:1,  respectively. Intermediate
                  load performance was higher, with SO2
                  removals of up to 60 and 67% at Ca/S
                  ratios of 2:1 and  2.5:1, respectively.
                  Results varied with specific LIMB oper-
                  ating conditions, boiler load, and sor-
                  bent  used.  Results of both extensive
                  parametric testing and continuous long-
                  term operation of the LIMB system are
 presented. Results of performance test-
 ing of the Low-NOx Concentric Firing
 System (LNCFS) Level II firing system
 are also  presented. Typically, under
 comparable boiler operating conditions,
 a 42% reduction in NOX at full load and
 a 33% reduction in NOX at intermediate
 load,  relative to  baseline levels, were
 achieved with the LNCFS Level II Sys-
 tem.
  The effects of LIMB  operation on
 boiler, electrostatic precipitator (ESP),
 and ash handling system performance
 are also discussed. The most signifi-
 cant impact on boiler performance was
 the collection rate of LIMB solids plus
 fly ash on boiler convective surfaces
 during continuous operation, resulting
 in poorer boiler  heat  transfer perfor-
 mance and higher temperatures leav-
 ing the boiler. Continuous operation of
 the sootblowing system minimized this
 effect. The results of two ESP perfor-
 mance tests that  were conducted dur-
 ing continuous  LIMB operation are
 discussed  and compared to results
 from similar testing conducted without
 LIMB operation. Ash conditioning and
 disposal during  LIMB operation at
 Yorktown were significantly affected by
 the unreacted lime in  the ash. These
 problems,  as  well as  suggested pre-
 cautions to avoid them, are discussed.
  Recommendations for LIMB  com-
 mercialization and an evaluation of
the economics of the technology in
comparison to a conventional flue gas
desulfurization  system  are dis-
cussed.
                                                               Printed on Recycled Paper

-------
  This Project Summary was developed
by EPA's Air and Energy Engineering
Research Laboratory, Research Triangle
Park, NC, to announce key findings of
the research project that Is fully docu-
mented In a separate report of the same
tltlo (sea Project Report ordering Infor-
mation at back).
Introduction
  Increasing concern for the environmen-
tal effects of acid rain has led to active
efforts to control emissions of sulfur diox-
ide and oxides of nitrogen, SO2 and  NOX,
from the fossil-fuel-burning electric  utility
industry. The choice of control techniques
for limiting SO2 and NOX emissions from
utility boilers will most likely include a mix
of technologies to achieve the desired re-
ductions at minimum cost. The  primary
method for controlling SO2 from new coal-
burning boilers is flue gas scrubbing.  Of-
ten, however, this technique may not be a
viable option for existing power plants.
Space may not  be available for installing
scrubbers  and disposing of  waste from
the  scrubbers. In addition, the installation
of scrubbers may not be a viable eco-
 nomic choice for older units with less than
 10-20 years of remaining useful life. The
 need for a  method of providing a moder-
 ate degree of  emission control at  lower
 cost and with minimal retrofit to an exist-
 ing boiler and  minimal space has  led to
 the development of the LIMB process. In
 this process for simultaneous SO2  and
 NOX control, a  calcium-containing  mate-
 rial, such as limestone (CaCO3) or hy-
 drated lime [Ca(OH)A, is injected into the
 furnace at appropriate temperatures.  The
 material, called the sorbent, breaks down
 to form lime (CaO). The CaO reacts  with
 SO2 in trie gas to  form calcium sulfate
 (CaSO4). CaSO4 is collected in the boiler's
 partfeulate  removal device, along with the
 coal ash, and transferred to a landfill  site.
 Concurrent with SO2 control by  injection
 of the calcium-containing material, NOX is
 controlled by installation of low-NOx burn-
 ers that provide staged introduction of air
 to  control flame temperature and  gas
 chemistry in the flame region.
    The LIMB process had its origin in the
 observation that significant sulfur is re-
 tained during the burning of coals that are
 high in alkali-containing  ash and  low in
 sulfur. This observation led to several in-
 vestigations (in Europe, Japan, and the
  U.S. during the 1960s) of the feasibility of
  injecting calcium-based sorbents directly
  Into the furnace of  a coal-fired  boiler  as
  the primary SO2 control technique. Early
  field demonstrations of furnace  injection,
  using limestone, produced  disappointing
levels of SO2 reduction.  During the late
1970s, interest in direct sorbent injection
for SO2 control waned as emphasis shifted
to scrubber development. However, dur-
ing the  early 1980s, development of im-
proved firing systems  with  lowered
combustion zone temperatures, provided
the impetus for the development of LIMB
as a low-cost retrofittable system  for the
simultaneous  reduction of SO2 and  NOX.
The EPA established  a major research
and development program, both at their
laboratories and with contractors, to es-
tablish LIMB as a viable, low-cost, alter-
native  to  conventional  flue  gas
desulfurization (FGD) systems for attain-
ing  moderate reductions in SO2 on ex-
isting coal-burning boilers. Through early
research efforts, LIMB eventually evolved
into the more generic "furnace  sorbent
injection"  as sorbents that were  more
 reactive than limestone, such as Ca(OH)2,
 and injection locations  that were  more
consistent with the optimum SO2  capture
 "window"  of  2300 to 1650°F (1260  to
 890°C), such as into the upper furnace,
 were identified. Because tangential firing
 introduces special considerations for the
 design of sorbent injection systems due to
 the unique aerodynamics in the  sorbent
 mixing region of tangentially fired boilers,
 the EPA established parallel development
 programs for LIMB application to both wall-
 fired and tangentially fired boilers. The
 Yorktown  demonstration project was the
 culmination of a multiphase program, which
 included both pilot- and prototype-scale
 testing, to develop LIMB for tangentially
 fired boiler applications. The EPA selected
 Combustion  Engineering (C-E) as prime
 contractor for the demonstration project in
  1987. Virginia Power's Yorktown  Unit No.
 2, a 180-MWe coal-fired boiler located at
 the Yorktown Power Station in Yorktown,
  Virginia, was selected  as the host unit.
  C-E,  Virginia Power, Stone & Webster,
  and the  U.S. Department of Energy all
  helped fund the project.
    The demonstration project objectives
  were established as at least 50%  reduc-
  tion in SO2 at a Ca/S of 2.5:1  and a
  reduction in NOX emissions to 0.4 lb/106
  Btu (0.17 kg/GJ) or below. These emis-
  sions would be reduced with no significant
  adverse impact on boiler pperability and
  reliability. An eastern bituminous  coal with
  a  sulfur content of 2.0  to 2.5%  was
  selected  as the demonstration coal.
    The overall program  was conducted in
  two phases: site-specific studies to sup-
•  port the system design, and the field dem-
  onstration itself. The site-specific studies
  included  a boiler characterization  test to
  identify the  location of  the ideal sulfation
zone through extensive mapping  of the
boiler temperatures profile, identification
of promising injection  schemes via cold-
flow testing on  a plastic model  of the
Yorktown boiler, and pilot-scale combus-
tion testing to verify performance of both
the most promising injection schemes and
the proposed low-NOx firing system. Those
injection schemes that performed best
during pilot-scale testing were incorporated
into the LIMB system design. Computa-
tional fluid  dynamics computer modeling
of the ESP inlet ducts, where the flue gas
humidification system would be located,
showed that there was insufficient resi-
dence time in the ducts to achieve a close
approach to the adiabatic saturation tem-
perature  for  enhanced  SO2 capture.
Humidification for ESP performance main-
tenance during LIMB operation had been
demonstrated to be effective at two other
EPA-sponsored LIMB  projects:  the
Edgewater wall-fired demonstration project
and the tangentially fired  prototype  test
program at Richmond Power & Light. En-
hanced SO2 capture with a close approach
to saturation had already been  demon-
strated at  Edgewater. To avoid incurring
 additional costs to extensively modify the
 Yorktown  ducts  for enhanced SO2 cap-
 ture, the Yorktown humidification system
 was designed only to maintain ESP  per-
 formance.
   While the final system design  was es-
 sentially complete by June 1990, construc-
 tion did not begin until May 1991 due, in
 large part, to a  1-year delay in  the  Unit
 No. 2 outage,  which was rescheduled to
 allow installation of a new steam turbine/
 generator set.
    A 1-month baseline test to establish the
 performance of Unit No. 2 with its original
 firing  system  was  conducted in March
  1991, prior to installation of the LIMB sys-
 tem. An extensive ESP performance test
 was also conducted. The medium-sulfur
 eastern bituminous demonstration coal was
  used for this test.
    The boiler was returned to service in
  December 1991, with the LNCFS Level II
  low-NOx firing system in place and instal-
  lation of the sorbent injection equipment
  nearly complete. Performance testing of
  the  LNCFS  system  was conducted in
  March 1992. LIMB system start-up/shake-
  down  testing  started  in May 1992  and
  continued through  August 1992. Perfor-
  mance testing of the LIMB system, includ-
  ing three long-term demonstration tests of
  over 22 days of continuous operation each,
  extended from September 1992 to Octo-
  ber 1993. The results of the LIMB and
  low-NOx testing, plus the effects of  LIMB

-------
 on Unit No. 2 operation, are presented
 below.

 NOX Emission Reduction
   A LNCFS Level II system was installed
 to meet the project objective of 0.4 lb/106
 Btu (0.17 kg/GJ)  NOX  emissions. The
 LNCFS  Level  II system  features  sepa-
 rated overfire air (SOFA) for staged com-
 bustion, flame attachment tips for early
 fuel ignition, and offset air nozzles to fur-
 ther minimize NOX formation. During early
 testing  of the LNCFS Level II system, it
 was determined that the new Unit  No. 2
 steam turbine had  introduced substantial
 changes in  the  boiler  waterwall and
 reheater  heat  absorption requirements.
 The boiler controls responded by  raising
 burner tilts, which, in turn,  adversely af-
 fected the effectiveness of the LNCFS fir-
 ing system in reducing NOX. To make the
 LNCFS performance data fully comparable
 to baseline data with the original firing
 system, all NOX emissions were adjusted
 to the 0° burner tilt used during baseline
 testing  and to a common  fuel  nitrogen
 content.  The  LNCFS  system  without
 overfire air performed at essentially the
 same level as  the  original baseline sys-
 tem at full boiler load, suggesting that the
 concentric firing  system,  in combination
 with the flame attachment  tips, did not
 contribute to overall NOX reduction. A 15%
 reduction in NOX was measured at inter-
 mediate boiler  load. Adding SOFA pro-
 duced a 42% reduction in NOX at full load
 and a 33% reduction at intermediate load.
 Carbon-in-ash levels for the baseline sys-
 tem and the LNCFS system with  SOFA
 nozzles closed were comparable at ap-
 proximately 7%, under similar test  condi-
 tions. The effect of fully  opening all three
 SOFA nozzles was to increase carbon-in-
 ash  by about 85%. Comparable  effects
 have been noted on other LNCFS  Level
 ll-equipped units burning similar coals.

 LIMB System for Yorktown
  The Yorktown sorbent injection system
 was designed to provide a great deal of
 operational flexibility to  permit operation
 of the system over a wide range of boiler
 and sorbent injection conditions. Three el-
 evations for injection into the boiler were
provided.  The elevations  were selected
based on the results of site-specific stud-
 ies, including field and pilot-scale testing.
The middle and lower elevations each pro-
vided for tangential injection of the sor-
bent through  16 injectors  located at the
eight corners of the divided furnace. The
lower elevation was designated as the
primary injection location for reduced boiler
load operation.  The middle elevation, lo-
 cated approximately at the "nose" of the
 furnace, was designated as the primary
 injection elevation for  full boiler load op-
 eration. The third elevation, with 16 injec-
 tors  across  the front  wall of the boiler,
 above  the tangential  injector elevations,
 was  an alternate  full-load injection loca-
 tion.
   Parallel sorbent feed systems supplied
 sorbent to both sides  of the divided  fur-
 nace. Feed rates of from 4 to 15 tons/hr
 (3.6 to 13.6 tonnes/hr)  (total) allowed sub-
 stantial flexibility in evaluating LIMB sys-
 tem  performance  over a range of  boiler
 loads and Ca/S molar ratios. In addition,
 either sorbent feed-system was capable
 of feeding both sides of the furnace, if the
 other feed system was  not available. Each
 feed  system consisted of a day bin  (65
 tons—58.9 tonnes—of  hydrate), plus feed-
 ers  and transport blowers. A  large  fan
 provided additional air  to achieve sorbent
 injection velocities into the furnace  of up
 to 450 ft/sec (137 m/sec).
   Sorbent for  LIMB testing at Yorktown
 was delivered by rail to an unloading site
 approximately 9 miles  (14.5 km) from the
 Yorktown Station. At this site, the sorbent
 was off-loaded to pneumatic trucks, typi-
 cally  holding 20 tons  (18 tonnes) of  hy-
 drated  lime, for delivery to the Station
 where  the sorbent was pumped  into a
 long-term storage bin  designed to hold
 390 tons (354 tonnes)  (roughly  2 days
 supply) of  hydrate. The truck  unloading
 system worked well  throughout the opti-
 mization and demonstration test periods.
 As many as 13 trucks were unloaded in 1
 day during demonstration testing.
 SO2 Removal Performance:
 Optimizing Testing
  A parametric test program was con-
 ducted  to evaluate the  effectiveness of
 the LIMB System  over a wide range of
 configuration and operating variables. The
 primary objective of the  program was to
 identify the LIMB configuration and oper-
 ating  variables that provided the greatest
 reduction  in  SO2,  consistent with  good
 boiler performance, and identify the opti-
 mum  sorbent system operation for  long-
 term demonstration testing.
  A  total  of 147 tests were conducted
 from September 1992 through September
 1993. Full-load tests were conducted at
 all three injection elevations. One alter-
 nate  injection yaw  configuration  was
 tested. Intermediate load tests were con-
 ducted  at two  injection  elevations. The
 effect of Ca/S on SO2 capture was evalu-
 ated. The effect of sorbent mixing on SO2
capture was  evaluated by varying injec-
tion air plus sorbent discharge velocity
 from 150 to 450 ft/sec (45.7 to  137.2 m/
 sec) (approximately 2.5 to 7.5% of total
 airflow to the boiler) and injector tilt from
 -35 to +35° relative to horizontal (0° tilt)
 injection. Commercial  hydrated lime from
 two suppliers was tested. Tests were also
 conducted   using   calcium-lignosul-
 fonate-treated hydrate from the same two
 suppliers. In addition,  limited testing was
 conducted  using finely  pulverized
 limestone as the sorbent.
   SO2 capture did not improve significantly
 when testing with calcium-lignosulfonate-
 treated hydrate. Data for pulverized lime-
 stone  (95% through  325  mesh)  were
 limited and  showed considerable variabil-
 ity.  The limestone tests, which  included
 low- (down to 90 MWe), intermediate- (128
 MWe), and full-load (169 MWe) operation,
 showed a  clear  advantage for the in-
 creased residence time available at low
 load. In general, the SO2 removal perfor-
 mance for pulverized limestone was about
 60% of the performance for hydrated lime.
   Optimized full-load  SO2 removal effi-
 ciency on hydrated lime was 56% at Ca/S
 = 2:1  and 63% at Ca/S = 2.5:1.  At  inter-
 mediate load, SO2 removal was 60% at
 Ca/S = 2:1 and 67% at Ca/S = 2.5:1. The
 increase in SO2 capture at intermediate
 boiler load is attributed to an increase in
 residence time in the 2300 to 1650°F (1260
 to 890°C) sulfation "window" at intermedi-
 ate  load. SO2 capture improved with in-
 creasing injector discharge velocity up to
 300 ft/sec (91 m/sec). Tests conducted at
 450 ft/sec (137  m/sec) showed  no  addi-
 tional  improvement in SOZ capture. No
 significant effect of injector tilt on SO2 cap-
 ture was measured during  this optimiza-
 tion testing.  Boiler operation, however, as
 measured by the ability to. maintain super-
 heat and reheat outlet  temperatures, was
 improved during  injection through the
 middle elevation when the injectors  were
 tilted downward. Full load boiler operation
 was further improved by injecting through
 the lowest elevation with the injectors tilted
 upward.
  The overall performance of the sorbent
 injection system was good throughout the
 optimization test period. The primary op-
 erational problem during this period was
 the  unreliability  of  the solids pumps in
 both the long term and day bin areas. In
 the long-term bin area, the solids pump
 was able to  supply only 11  tons/hr (10
 tonnes/hr) of sorbent to the day bins ver-
 sus  the design  rate of 20 tons/hr (18.1
tonnes/hr). By removing the solids pump
 and  discharging  from  the existing rotary
feeder directly into a  "pick-up tee," the
 required capacity was achieved and main-
tained  through the remainder of  the test

-------
program. The solids pumps in the day bin
area also proved troublesome. While the
pumps were able to achieve the required
maximum capacity of 7.5 tons (6.8 tonnes)/
hr/pump, ft was necessary to operate the
pumps at the unusually high screw speed
of 1500 rpm to attain this feed rate. The
pumps were not able to operate continu-
ously at this speed. Since this was unac-
ceptable for  long-term   continuous
demonstration testing,  the solids  pumps
were removed  and replaced with  rotary
airlock feeders which, in turn, discharged
into "pick-up tees." This arrangement op-
erated well during both optimization and
continuous demonstrating testing, although
the rotary feeders repeatedly experienced
an unanticipated  problem: hard deposits
on the feeder drum surfaces. This required
occasional cleaning  but did not interrupt
the demonstration tests.
   At the discharge of  the injectors, hard
deposits formed inside  the nozzle tips
which restricted, but did not  completely
close, the nozzles. This necessitated peri-
odic removal and cleaning of the injectors
to permit design  air/sorbent flow  rates.
The mechanism for formation of these de-
 posits is not yet understood.
 SO2 Removal Performance:
 Demonstration Testing
   A primary objective  of the Tangentially
 Fired LIMB Demonstration  Project was to
 evaluate the long-term  effectiveness of the
 integrated LIMB  system in  achieving the
 SO2 and NOX emissions reduction objec-
 tives and establish  the long-term effects
 on boiler, ESP, and ash handling system
 operation. Three long-term demonstration
 tests were conducted. Test duration and
 gaseous-emission data  collection  objec-
 tives for these tests were established in
 accordance  with Performance Testing re-
 quirements contained  in Vol. 44, No. 113
 of the  Federal Register. These require-
 ments may  be satisfied in  as few as  22
 days (out of 30) if at least 18 hours of
 acceptable continuous emissions monitor-
 ing (GEM)  data are produced each day.
 Because of the high quality of GEM data,
 H was possible to complete each of the
 demonstration tests after  only  22 to  24
 days, having satisfied test duration require-
  ments.
    During the demonstration tests, the
  boiler was  operated  in its normal duty
  cycle. Sootblowing  and ash handling sys-
  tem operation were adjusted as required
  by LIMB operation. ESP  operation and
  humidification of the gas and solids enter-
  ing the ESP were adjusted to conform to
  opacity requirements, except during ESP
  performance tests when additional require-
  ments were specified.
  The sorbent injection configurations and
operating conditions were selected for the
demonstration tests to maintain good boiler
performance  and operability  as  well as
achieve good SO2 capture performance.
Thus,  when  injection  was through the
middle elevation, the injectors were tilted
downward to improve boiler temperature
control. The first demonstration was con-
ducted with lignosulfonate-treated hydrate
as the sorbent.  Untreated commercial hy-
drate was used for the second and third
demonstration tests. The SO2 removal ef-
ficiencies at the Ca/S molar ratios of com-
mercial interest (Ca/S = 2:1  and 2.5:1)
were essentially the same for all three
tests, averaging 44% at  Ca/S =  2:1 and
50% at Ca/S = 2.5:1. These levels are
lower than those measured during optimi-
zation testing. Two factors that adversely
impacted  the SO2 removal performance
were sootblowing  requirements and data
interval selection.  During continuous  in-
jection of   sorbent,  the Unit  No.  2
sootblowing  system was operated almost
continuously to maintain critical tempera-
tures  within limits. Continuous  cleaning
 may reduce secondary  SO2 capture by
 deposited sorbent  material, thereby reduc-
 ing overall. SO2 capture. Sootblowers were
 not  operated during  individual optimiza-
 tion tests. The  second  factor-data interval
 selection-was also affected by sootblower
 operation. While it was possible to deter-
 mine  maximum  SO2  capture (minimum
 CEM  SO2) for  individual optimization test
 points without sootblower operation, it was
 not possible to obtain  continuous demon-
 stration test periods of sufficient duration
 without including at least one sootblowing
 cycle. This resulted in  the requirements to
 use average, rather than minimum, CEM
 SO2 levels to define SO2 capture. There-
 fore,  the  impact  of  the  continuous
 sootblowing requirement tends to be con-
 siderable.
   Much of the first demonstration test was
 conducted  at  minimum sorbent feed in
 response to concerns over indications of
 deposit buildup on the floors and turning
 vanes of the ESP inlet ducts. The ducts
 were cleaned  during  a subsequent non-
 LIMB-related outage. The second and third
 demonstration tests  were conducted at
 design sorbent feed  rates  with minimal
 deposit problems, reflecting  the effective-
 ness of improved  humidification lances that
 were installed for these tests.
  ESP and Humidifier
 Performance
    The Yorktown Unit No. 2 ESP has a
  specific collection  area (SCA) of  720
 ft2/1000 acfm (14 m2/1000 m3/sec) of
 treated gas.  This exceptionally  large ESP
provided an opportunity to evaluate the
impact of LIMB on ESP performance over
a wide  range of conditions. Prior to the
start of  optimization  testing, a test was
conducted with design sorbent feed rates
but without humidification. The ESP per-
formance deteriorated rapidly after only a
few hours of operation. This was consis-
tent with results obtained at Edgewater.
With the humidification system in opera-
tion, it was possible to maintain ESP per-
formance  at or near its pre-LIMB levels.
Throughout the optimization and first dem-
onstration test periods, water distribution
was poor, resulting in repeated problems
with deposits  in the ESP  ducts, on  the
ESP  inlet turning vanes,  and  on  the
humidification lances themselves.  Destruc-
tive examination of two  lances revealed
weld cracks that allowed the air and water
to mix within the lances, producing severe
maldistributions of water. Lances fabricated
with revised  procedures significantly  im-
proved  water distribution with a dramatic
 reduction in duct and turning vane depos-
 its. Opacity was always maintained within
compliance limits throughout the three con-
tinuous demonstration tests.
   ESP performance tests were conducted
 during  the baseline test and the first two
 demonstration tests. Total particulates en-
 tering and leaving the ESP, particle  size
 distributions, and particle resistivities were
 measured under  various ESP  operating
 conditions. Fields were taken out of  ser-
 vice  to simulate  smaller ESPs,  with the
 objective of identifying  the smallest ESP
 that  could operate  with  LIMB without
 modification. Performance test results
 showed that  essentially no performance
 losses were seen during  ESP operation
 With  SCAs of 720 to 480 ft2/1000 acfm (14
 to 9  mz/100 m3/sec) of treated gas, under
 the same LIMB and humidifier operating
 conditions. Further SCA  reductions re-
 sulted  in increases  in opacity.  Collection
 efficiencies were comparable during the
 three performance tests. The average col-
 lection efficiency  with LIMB was 99.21%,
 versus 99.42% without LIMB. Flyash re-
 sistivity  varied   over  a  narrow  range.
 Baseline resistivity without LIMB was 3.7
 x 1010 ohm-cm. Resistivity during the first
 performance test  with LIMB was 1.4 x 10"
 ohm-cm. During the second performance
 test with LIMB, resistivity was slightly lower
 at 1.4  x 1010 ohm-cm, suggesting improved
 humidification effectiveness with the modi-
 fied lances.
    An unexpected ESP performance effect
 was noted during LIMB testing with pul-
 verized  limestone.  The  humidifier was
 turned off during the final 2 days of test-
  ing.  Opacity did  not increase above
  non-LIMB levels. No  ESP performance

-------
measurements, including flyash resistiv-
ity, were made during limestone testing.

Ash Handling Effects
  The Yorktown Ash Handling Facility fea-
tures a single ash silo that receives flyash
from all three units. When Units No. 1 and
2 were both in operation, LIMB ash from
Unit No. 2 was mixed with, and diluted by,
ash  from  Unit No.  1. Under  these
conditions, the mixed  ash was easier to
handle and produced less steaming than
in previous LIMB field demonstrations. It
was, nevertheless, more difficult to entirely
remove the ash from ash truck beds than
when  no LIMB ash was present. When
Unit No. 1 was not in operation, the LIMB
ash was  somewhat  more difficult  to
process through the  rotary conditioner,
which received the ash frorri the silo, added
water while  mixing, and discharged  the
mixture into trucks. Steaming was greater
for LIMB-only ash than for the mixed Unit
No.1/Unit No. 2 ash.
  Long-term  continuous testing of  the
LIMB system placed increased strain on
the  Yorktown ash  handling system.  The
concentration of LIMB  ash in the ash  silo
increased  over that experienced  during
optimization testing. This  resulted in a
greater degree of steaming of the condi-
tioned LIMB ash in the ash  trucks and at
the disposal site. The LIMB  ash exhibited
a much greater tendency to set up in the
truck beds than did non-LIMB ash or a
mixture of Unit No. 1  ash with no LIMB
and Unit No. 2 ash with LIMB. Frequent
problems of material setting up in the truck
beds were experienced. Decreasing  the
water to the rotary conditioner (mixer) ap-
peared to reduce problems of setting up
and steaming.
  The quantity of material produced dur-
ing continuous testing taxed the Yorktown
ash handling personnel and  equipment. A
substantial increase in  the amount of time
Unit No. 2 was operated at full load,  fol-
lowing the substantial  heat  rate improve-
ment  as a  result  of the steam  turbine
replacement, contributed to this problem.
  All of the LIMB ash  was placed in one
area of a single ash cell at  the Yorktown
ash disposal site. No significant problems
were experienced due to the cementitious
properties of the LIMB ash. It was pos-
sible to spread the LIMB ash by bulldozer
with only occasional steaming problems.
Boiler Performance and
Operability
  During LIMB operation, lime, ash, and
LIMB products accumulated  on boiler heat
transfer surfaces  at a greater rate than
during  normal non-LIMB operation. This
reduced the heat absorption  effectiveness,
shifted the gas temperature profile in the
boiler downstream,  and reduced overall
boiler efficiency. Reheat outlet tempera-
ture, for example, typically dropped about
30 F° (16.7 C°) during continuous full-load
sorbent injection at Ca/S = 2:1. Frequent
sootblowing  recovered  essentially all  of
the temperature loss, suggesting friable,
easily removable deposits. The  increase
in the temperature of the gas entering the
ESP inlet ducts (and the humidifier)  as a
result of the shift in heat absorption pat-
tern during sorbent injection was typically
75 to 80°F (23.9 to 26.7°C). This shift in
absorption and the concurrent increase in
gas heat loss  resulting from the increase
in gas flow and deposits on heat transfer
surfaces typically resulted  in a reduction
of 1.5 points  in boiler efficiency  during
continuous sorbent injection relative to the
efficiency level during continuous testing
without  sorbent injection. The net boiler
efficiency reduction was about  1  point
after adjusting for carbon-in-ash and cal-
cination/sulfation  effects. Optimized
sootblower coverage  and operation
would likely reduce this penalty.
  During injection of finely pulverized  lime-
stone, boiler effects were significantly dif-
ferent than those observed during injection
of hydrated lime. Superheater and reheater
outlet temperatures did  not drop as much
as during hydrate injection, suggesting a
decreased tendency of deposits to  build
up  on the heat transfer surface.  As a
consequence,  sootblowing  requirements
were significantly reduced.
  No LIMB-related  boiler  outages were
experienced during any part of the LIMB
test program at Yorktown,  including the
three demonstration tests. Boiler operabil-
ity was,  however, adversely affected  by
the demonstration coal and  operation  of
the LIMB system, in addition to the effects
of the new Unit No. 2 steam turbine.  In-
creased  slagging at the burner corners
was noted throughout the LIMB test pro-
gram.  However, slag formation on the
waterwalls was, generally, not severe and
did  not require  operating  the  wall
sootblowers. Ash/sorbent deposits built  up
around many of the sorbent injection ports.
These deposits would break away and fall
into the boiler  ash  hopper  and were ulti-
mately removed from the  boiler via the
wet bottom  ash removal system. Occa-
sional problems  were experienced in the
bottom  ash  removal system due  to the
high lime content of this material.
Economics and
Commercialization
  Capital and  operating costs associated
with a 180-MWe "base case" LIMB  sys-
tem were generated, based on  actual
Yorktown costs, and used as a basis for a
sensitivity study on the effects of unit size
and various process parameters on LIMB
economics. The Yorktown system, plus a
new dedicated ash silo, had a total capital
requirement of $81/kW, with 2.5% S  coal
at a Ca/S of 2.5:1. Levelized cost for the
base case was $781/ton ($861/tonne) SO2
removed. These costs are lower than those
reported for previous LIMB demonstrations,
indicating a maturing of this SO2 control
technology.
  Capital investment increased 46%, to
$118/kW, for  a 100-MWe unit  but de-
creased 23%, to $62/kW, for a 300-MWe
unit, clearly showing the economy of scale
for a retrofit LIMB system.
  LIMB has  been demonstrated  as a vi-
able low-cost option for achieving moder-
ate levels of SO2 reduction with capital
and operating costs that are substantially
below  those of  conventional flue  gas
desulfurization systems. The 1990 Clean
Air Act Amendments (CAAAs) focus on
high SO2 removal efficiency technologies.
They also make SO2 allowances (credits)
available at a fraction of the cost of FGD
systems. Consequently, the domestic mar-
ket for LIMB, with its moderate  SO2 re-
moval, is currently  best suited  to older,
small to  intermediate units, which might
otherwise be retired, and also to combina-
tions of LIMB with back-end SO2 removal
processes.  An  example would  be LIMB
combined with the ADVACATE process to
achieve high overall -SO2 removals. Thus,
in the contemporary domestic utility mar-
ket, LIMB alone is postured as a "niche"
technology  for  units requiring  moderate
levels  of SO2  control. However, if com-
bined with a complementary back-end sys-
tem, LIMB can offer  the potential for
competitive (i.e., 90 +%) SO2 removals at
favorable capital and operating costs.

Summary and Conclusions
  The  viability of  LIMB as a low-cost op-
tion for achieving  moderate levels of SO2
reduction on tangentially  coal-fired utility
boilers was  established  under  the
EPA-sponsored LIMB Demonstration  Pro-
gram  at Virginia  Power's  180-MWe
Yorktown Unit No. 2. Extensive  paramet-
ric testing to evaluate various sorbent in-
jection configurations  and   process
variables was conducted over a range of
boiler loads. Full-load SO2 removals of up
to 56 and 63% at Ca/S molar ratios of 2:1
and  2.5:1,  respectively, were  achieved,
substantially above the project  objective
of 50% SO2 removal at a Ca/S molar  ratio
of 2.5:1.  Higher SO2  removals were
achieved at intermediate boiler load. Test-
ing was  conducted on calcitic  hydrated
lime, with and  without calcium lignosul-

-------
fonata treatment, each from two  suppli-
ers. SO2 removal did not increase signifi-
cantly during testing  with  the  treated
hydrates. Three continuous long-term tests
of the LIMB system, each of from 22 to 24
days in duration, were conducted over the
normal boiler operating range. Lower SO2
removal levels were seen during this con-
tinuous testing, due, largely, to the need
to continuously operate the sootblowers
to remove accumulated LIMB solids from
boiler heat transfer surfaces. Commercial
LIMB installations would require increased
sootblower coverage to  maintain boiler
surface cleanliness. No boiler outages at-
tributable to LIMB were experienced dur-
ing the entire LIMB test program.
   ESP performance was  maintained dur-
ing LIMB operation  by humidification of
the gas stream at the inlet to the ESP. No
deep humidification for enhanced  SO2 re-
moval was attempted. Early problems with
deposition of LIMB solids on duct  sur-
faces were resolved by improvements in
humidifier lance fabrication  procedures.
Opacity was maintained within compliance
requirements throughout the test program.
  LIMB ash was collected in, and pro-
cessed from,  a common ash  silo that
received material from both  coal-burning
units.  Problems were  experienced with
material  setting up in ash  truck  beds.
Steaming of material in  the ash trucks
varied with the concentration of LIMB ash,
decreasing as dilution from non-LIMB Unit
No. 1 ash increased. A dedicated ash silo
would be a necessary component of a
commercial LIMB installation.
  The  LNCFS  Level II  firing  system
achieved  a 42% reduction in NOX  at full
boiler load and 33% reduction in NOX at
intermediate boiler load, relative to baseline
levels with the original tangential firing sys-
tem,  under comparable boiler operating
conditions. These  levels  were below the
program NOX objective of 0.4 lb/106 Btu
(172 mg/J).
  Capital and levelized costs associated
with the 180 MWe Yorktown LIMB system
were determined to be $81/kW and $7817
ton ($859/tonne)  SO2 removed,  respec-
tively, based on actual equipment and op-
erating costs for the Yorktown  system,
plus the cost of  a new ash silo. Sensitivity
studies showed the cost  effectiveness  of
LIMB on tangentially fired units from 100
to 300 MWe operating on coals with 2  to
3% sulfur.
  The commercial applicability of a stand-
alone LIMB system  is currently consid-
ered to be that of a "niche" technology for
certain older units as a result of the 1990
CAAAs. In  combination  with a  comple-
mentary "back-end" SO2 removal process,
however,  LIMB offers the  potential  of
achieving  compliance levels  of  SO2 re-
moval at costs that are significantly below
those of conventional FGD systems.

-------