SUMMARY OF OPERATION
                        AND TESTING AT THE
                  SHAWNEE PROTOTYPE LIME/LIMESTONE
                          TEST FACILITY
INDUSTRIAL ENVIRONMENTAL RESEARCH LABORATORY - RTF
                   Environmental Protection Agency
               Research Triangle Park, North Carolina 27711

-------
            SUMMARY OF OPERATION
             AND TESTING AT THE
      SHAWNEE PROTOTYPE LIME/LIMESTONE
                TEST FACILITY
              John E. Williams
    Emissions/Effluent Technology Branch
   Utilities and Industrial Power Division
Industrial Environmental Research Laboratory
    U.S. Environmental Protection Agency
Research Triangle Park, North Carolina  27711
                 April 1977

-------
                              INTRODUCTION
EPA's Industrial Environmental Research Laboratory-RTF and its predecessors
over approximately the past 15 years have undertaken a program of research
and development to establish the technical and economic feasibility of
promising processes for removing sulfur dioxide and particulates from
coal-fired boiler flue gases.  An important part of this effort involves
the operation and testing of a prototype wet lime and limestone scrubbing
test facility located at the Tennessee Valley Authority's (TVA) Shawnee
power plant near Paducah, Kentucky.  This versatile facility allows
comprehensive testing of up to three scrubber types under a variety of
operating conditions.  Bechtel Corporation of San Francisco, as the major
contractor, designed the test facility and directs the test program.  TVA
constructed and operates the facility.

Purposes of this report are to describe the Shawnee test facility, to
discuss the test program objectives, and to summarize major test results
to date.  More detailed information is presented in published EPA reports
listed in the reference section at the end of this report.  Two additional
reports are presently being reviewed and should be published by July 1977.
These are:  1) the third progress report covering testing from February
through November 1976 under the advanced program; and 2) the second annual
report covering the Shawnee field evaluation of the disposal of flue gas
cleaning wastes.

The test facility consists of three parallel scrubber systems:  a venturi
followed by a spray tower, a Turbulent Contact Absorber (TCA), and a
Marble-Bed Absorber.  Each system is capable of treating approximately
10 MW equivalent (30,000 acfm @ 300°F) of flue gas from Shawnee unit 10
(nominally 150 MW total) containing 1000 to 5000 ppm sulfur dioxide
(S0?) and 0.2-11 gms/m^ of particulates (fly ash).  Operation of the
Marble-Bed Absorber was discontinued in July 1973 (see Reference 1 for a
discussion of the reasons), and subsequent operation has been limited to
the two remaining scrubber systems.  Each of the systems can be operated
in a variety of configurations.  Typical system configurations depicting
lime testing with the venturi/spray tower and limestone testing with the
TCA scrubber are illustrated schematically in Figures 1 and 2, respectively.^.
                                                                           «•
Major goals of the original two-year test program were:

   • Characterization of the effect of important process variables on
     sulfur dioxide and particulate removal.

 ,, • Development of mathematical models to allow scale-up to full-size
     scrubber facilities.

-------
    VENTURI
   SCRUBBER
           ®
LIME
                                                 MAKEUP
XA A AN
wa»»\\"ft
v y v

A A A
A /\ A
A A A
AA A
V


SCRUBBER
EFFLUENT
HOLD '
TANK





SPR
TOW

-«
/
"*N




AY
ER
*r
rn—
(
•MB
)

. WATER \

r^~
  SAMPLE POINTS

  O  Gas Composition
  ®  Particulate Composition & Loading
  0  Slurry or Solids Composition
Gas Stream
Liquor Stream
                                                                                                                                       I'M
                                                                                                              SETTLING POND
                   Figure !•[  Venturi/Spray  Tower-Typical  System Configuration

-------
                TCA
              SCRUBBER
   FLUE GAS J>- O - - +
                       SCRUBBER  REACTION  TANKS
SAMPLE POINTS

O  Gas Composition
®  Particulate Composition & Loading
©  Slurry or Solids Composition
_ _  Gas Stream
___  Liquor Stream
                                                            2^4.
                                                           I. D. FAN

A /\ /\
5-7-5
.y.v.
% .T/
.v.v.


,



,


                                                                            I
                                                                            I
                                                                            I

                                                                            I
                                                                            L
                                                                                                                     STACK
                                                                                                             SETTLING POND
              Figure  2.  Turbulent  Contact  Absorber-Typical  System Configuration

-------
   • Development of information to study the technical and economic
     feasibility of lime/limestone scrubbing.

   • Demonstration of long-term reliability.

Although for the most part these goals were met at the completion of the
original test program, it became apparent that additional Shawnee testing
was needed to supply further information and to optimize lime and limestone
systems in the areas of:  sludge disposal, improved reliability, variable
load operation, and improved process economics.  This conclusion was based
on comments of need from utilities and flue gas desulfurization system ven-
dor representatives, and particularly the encouraging results from IERL-RTP
pilot plant support studies which indicated several potentially attractive
new scrubber operating concepts and process variations that could favorably
affect both the economics and the operating reliability of scrubbing.
Therefore, in 1974 the Shawnee program was extended and the scope was
expanded.  Additional objectives of the expanded program were:

   • Investigation of advanced process and equipment design variations for
     improving system reliability and economics.

   • Evaluation of process variations for a substantial increase in alkali
     utilization for limestone systems.

   • Evaluation of the effect of increased magnesium ion concentration or
     other additives on reducing gypsum saturation and increasing SO-
     removal efficiency.

   • Determination of the efficiency and reliability of lime and limestone
     scrubbers under conditions of widely varying flue gas flow rate and
     inlet SO- concentration.

   • Evaluation of system performance and reliability without fly ash in
     the flue gas.

   • Assessment of the effectiveness of forced oxidation or other process
     variations for producing an improved throwaway sludge product.

   • Determination of the practical upper limit of SO- removal efficiency
     for both lime and limestone scrubbing systems.

   • Development of a design/economic study computer program for comparison
     of both initial investment and lifetime operating costs for full-scale
     lime and limestone systems.

Two smaller scrubbing systems (0.1 megawatt each) have also been operated
at the IERL-RTP pilot plant in support of the Shawnee program.  These
small, pilot-scale scrubber systems are capable of simulating the Shawnee
scrubber systems with excellent agreement in the lime/limestone wet-scrub-
bing chemistry.  Preliminary data are generated on the pilot-scale systems
to verify and guide the selection of those promising concepts that should
logically be investigated on the larger size Shawnee units.  Indeed, results
obtained from operation of the IERL-RTP pilot plant were very instrumental
in the decision to continue testing at Shawnee under the advanced test
program.

-------
                          DISCUSSION OF RESULTS
The test program schedules based on the defined objectives for the
initial program and the advanced concept testing are presented in Figures
3 and 4, respectively.  As shown by the schedules, much of the early
effort was devoted to identifying and learning how to control two separate
and distinct operating reliability problems—scaling and soft, mud-type
solids deposition, especially in the relatively restricted area of the
mist eliminators.  Much of the early work was also conducted at a re-
duced gas velocity using various combinations of mist eliminator washing
techniques and hardware configurations, including a wash tray upstream
of the mist eliminator.  In fact, it was not recognized early in the
program that scaling and mud-type solids deposition were indeed separate
problems, each with separate and distinct solutions.  This fact was not
really apparent until methods were found to adequately control both.

In a lime or limestone based system, the most frequently encountered
scaling problem is due to sulfate (gypsum) crystals which precipitate on
scrubber internals rather than outside the scrubber in the reaction
tank.  Selection of operating parameters (such as liquid-to-gas ratio
(L/G), reaction tank residence time, pH, or percent solids in the circu-
lating slurry to maintain a gypsum saturation level in the scrubber
below about 130%) will prevent this type of scaling.  The use of very
infrequent (once every 4-8 hours) top sequential washing with fresh
make-up water and low pressure drop nozzles proved adequate at Shawnee
for control of scaling in the mist eliminators.  However, identification
of those operating parameters affecting scaling potential, safe oper-
ating ranges, and mist eliminator washing techniques proved time con-
suming.  This was especially true with the problem of mud-type solids
deposition which was superimposed over the scaling problem.

Solution of the problem of mud-type solids deposition proved even more
difficult than control of scaling, especially in the restricted area of
the mist eliminators.  The use of a wash tray upstream of the mist
eliminator to transfer the mud-type deposits to an area where they could
more easily be removed by washing was only marginally successful and
required reduced gas flow rates.  Washing techniques for the mist elimi-
nator were ultimately found without requiring a wash tray and at design
gas velocities which limited restriction from these mud-type solids to
less than 10% of the open area.  These solid deposits did not hinder
operation or cause shut-down for periodic cleaning.  However, the real
key to successful control of these mud-type solid deposits was not found
until well into the advanced test program.

-------
TEST PROGRAM FUNCTIONS
SYSTEM CHECK-OUT
AIR/ WATER TESTS
SODIUM CARBONATE TESTS
LIMESTONE WET-SCRUBBING TESTS:
Factorial Tests
Reliability Verification Tests
Reliability Tests (TCA)
LIME WET-SCRUBBING TESTS:
Reliability Tests (Venturi/Spray Tower)
1972
MAMJJASOND
123456789 10
—
—
—



BOILER OUTAGE
SYSTEM MODIFICATK

1973
JFMAMJJASOND
(111213141516171819202122

-


—1 1
i i
III
J
DNS


1974
J F M A M J J A S 0
23242526272829303132










Figure 3.  Shawnee Test Program Schedule

-------
	munm Muni m°i O«I°A'
   MMCitOMI AlIEIWMWi tCHUMIH IMTiHMAU
      Ul HIE IMU1DOWM .COMTMACI COMTLEIED.
      CM »WAt fllraflT MAPI.
                                     Figure  4.    Shawnee  Advanced Test  Program Schedule
   SCHEDULE FOR
ADVANCED TEST PROGRAM
WITH FDOPOSED SIX MONTH
EXTENSION-NOVEMBER 1976

-------
During tests to study methods to improve alkali utilization, a correlation
was found between alkali utilization and the accumulation of these mud-type
solids.  Above about 85% alkali utilization, the solids can easily be
removed even with very infrequent (once per 8-hour shift) washing with
fresh make-up water.  In fact, an entirely clean system was maintained over
an extended period using only approximately 25% of the available fresh
make-up water.  (The available make-up water is defined as that added to
maintain a closed-loop water balance and represents the water lost through
evaporation and that which leaves the system with the purged solids.)  This
correlation was subsequently confirmed at Shawnee using three different
methods for improving alkali utilization and has also been confirmed on a
different system at TVA's 1 MW Colbert pilot plant.  The ability to improve
alkali utilization then becomes extremely important for limestone scrubbing
because:

   • It provides a means for existing installed full-scale systems, at
     minimal cost, to solve or improve an operating reliability problem
     that persisted in many of the early applications.

   • It permits the use of limestone costing about $4-6/metric ton rather
     than lime costing about $30-40/metric ton at greater than 90% utiliza-
     tion (or about l.lx stoichiometric quantity), which is comparable to
     that normally obtained with lime.

   • It further reduces the overall costs of scrubbing by substantially
     reducing the quantity of by-product sludge produced.

Another major area being studied at Shawnee during the advanced test
program is that of forced oxidation of sulfite to sulfate (gypsum).  This
concept was developed in the IERL-RTP pilot plant and includes both a
staged scrubber configuration and forced oxidation within a single scrubber.
Both configurations have the potential for obtaining essentially complete
oxidation of the solids to gypsum and an accompanying improvement in solids
settling characteristics and dewatering properties.  However, each method
has possible advantages and disadvantages compared with the other, and both
will be evaluated further on the larger units at Shawnee.  Forced oxidation
in a single scrubber is certainly a more simplified approach requiring less
equipment than staged scrubbing and can probably be more easily controlled
from an operational point of view.  However, this method applies only to
limestone systems while the staged scrubbing configuration can be used with
either lime or limestone.  The staged scrubber configuration has the added
advantage of also being able to obtain very high (about 95%) alkali utili-
zation simultaneously with the forced oxidation to gypsum.  The benefits
and significance of high alkali utilization, especially for limestone
operation, were pointed out in the preceding paragraph.

It has been estimated that the application of forced oxidation in the
staged scrubber configuration (compared to conventional limestone scrub-
bing) can reduce limestone requirements by roughly one-third and can
reduce the volume of by-product sludge generated by more than 50%.  An

-------
economic sensitivity analysis made earlier in the program by TVA indicated
that sludge disposal costs represent a significant portion of the overall
costs for lime/limestone scrubbing, and the large quantities of waste
sludge that must be disposed of have long been one of the major objections
of the utility industry to these types of flue gas desulfurization (FGD)
processes.  Obviously, reductions of this magnitude in both limestone
requirements and waste sludge generation would have a large impact on the
economics of scrubbing and, hopefully, on wider acceptance and application
by industry of lime/limestone scrubbing.

Other promising areas* which have been or will be studied at Shawnee during
the advanced test program include:

     1.   Addition of MgO to increase the SO™ removal efficiency and to
          force operation into the gypsum unsaturated mode.  From the test
          results to date, it appears virtually certain that the SO-
          removal efficiency can be substantially improved by MgO addition
          using either lime or limestone feed.  Higher MgO addition rates
          are required for limestone than for lime, but the S0_ removal
          efficiency can be increased using either alkali from a normal 80%
          removal to 95% removal or greater, with other operating para-
          meters (L/G, percent solids recirculated, scrubber type, pressure
          drop, etc.) remaining the same.  However, MgO addition did not
          always result in a gypsum subsaturated operation, and under some
          conditions the scaling potential was actually increased.  The
          reasons for this, certain other apparent inconsistencies in the
          data, and how these types of systems can be controlled are not
          yet fully understood.  Efforts are being continued to satisfac-
          torily resolve these important questions.

     2.   Demonstration of variable load operation on the venturi/spray
          tower system with lime.  The gas rate was adjusted hourly to
          follow the boiler load over a range of 60-150 MW while the inlet
          SO- also varied from 1400 to 5000 ppm.  Sustained runs were made
          at conditions probably much more severe than would be encountered
          at most operating plants with no problems of reliability or
          control of the system.  Variable load operation will also be
          included during the long term reliability runs planned on each
          scrubber system (one with lime and one with limestone) in which
          the most promising concepts will be combined for the extended
          reliability demonstration runs near the end of the current program.

     3.   Operation with and without fly ash in the system has also beer-
          made with both lime and limestone.  The major differences noted
          were in the solids settling rate and the solids dewatering
          properties.  This was true for both lime and limestone operation
*These are discussed in more detail in References 1-3.

-------
                              10
     with filter cake solids averaging 40-50% without fly ash compared
     to 50-60% with fly ash in the system.   When operating without fly
     ash, there was also a greater tendency for the clarifier overflow
     to be turbid,  and on occasion flocculent addition to the clari-
     fier was even required to maintain a clear overflow for return to
     the scrubber loop.  Some difference was also noted in the pH and
     SO- removal efficiency at a given stoichiometric feed rate.   The
     direction of the difference was opposite to that expected based
     on the high pH of the effluent from the regular plant fly ash
     pond which normally has a pH of 10.5-11.0.  Without fly ash in
     the system the pH averaged about 0.2 pH units higher, and there-
     fore the S0? removal was also about 5% higher.  This is attributed
     to a higher leaching rate for the acidic components in the fly
     ash compared to the basic components.   In separate tests it was
     shown that upon standing, Shawnee fly ash/water mixtures first
     exhibit a lower pH, and with time the pH gradually increases.  It
     is probable that operation without fly ash would lead to lower
     erosion rates of nozzles, piping, TCA spheres, etc.  However,
     operation has not yet been for extended periods to properly
     assess this possibility.

4.   Factorial tests conducted with both lime and limestone have been
     very useful in developing and checking the accuracy of the design
     models for scale-up to full—scale units.  The models have also
     been useful in developing the computerized Shawnee data base and
     the design/economic study computer program.  When completed the
     design/economic study computer program will be one of the most
     useful tools for industry, vendors, and engineering firms pro-
     duced from the Shawnee program.  It will readily permit compari-
     sons of capital and lifetime operating costs for essentially all
     of the equipment and process variations studied during the Shawnee
     program.  The computer program may also be modified to include
     other types of scrubbers and processes to even further expand its
     applicability and usefulness.

5.   Field evaluation studies to assess techniques for the disposal
     of power plant flue gas cleaning wastes^ are also being conducted
     at Shawnee in parallel with the regular program.  Waste material
     (sludge) produced during operation of the test facility using
     both lime and limestone as absorbents has been placed in six
     small disposal ponds on the plant site.  Three of the ponds
     contain untreated wastes; each of the three remaining ponds
     contain wa^te that was chemically treated by one of the three
     contractors (Dravo, Chemfix, and IU Conversion Systems) who
     presently offer such commercial fixation processes.  Test samples
     of treated and untreated wastes, ground water, surface water,
     leachate, and soil cores are being analyzed to evaluate the
     environmental acceptability of current disposal technology.

-------
                              11
     Results to date indicate that the total dissolved solids in
     leachate from untreated wastes increase after filling,  reaching a
     peak value equivalent to the input liquor,  and then steadily
     diminish.  Leachates from treated wastes exhibit similar charac-
     teristics, except the peak value is approximately half  the level
     of the total dissolved solids (TDS) in the  input liquor.  In
     addition to a reduction in the leachate TDS,  the treated materials
     also exhibit greatly improved structural stability and, for at
     least one of the processes, greatly reduced permeability.  (These
     improved physical properties would allow management of  disposal
     operations such that leachate generation can be minimized or
     eliminated.)  Ground water monitoring has shown no change at-
     tributable to the ponds.  Monitoring will be continued  during the
     balance of the Shawnee program.

6.    Peripheral types of studies are also underway or are planned
     during the Shawnee program.  These will not be discussed in
     detail but include the following (reports listed in the reference
     section are available for more detail):

     a.   A corrosion/erosion study and a mechanical components
          evaluation program conducted by TVA.  (ongoing)

     b.   Flue gas characterization study to determine the effect of
          changes in major operating variables on particulate mass
          loading and particle size distribution,  gaseous and solid
          sulfates, and on mist eliminator efficiency.  (ongoing)

     c.   Future studies of alternative scrubber internals,  automatic
          feed control, minimum energy requirements, maximum SO-
          removal efficiency, and variation of limestone type and
          size.  (planned)

-------
                                   12
                         SUMMARY AND CONCLUSIONS
With dramatic increases in the use of high sulfur coal reserves being a.
near certainty in the Administration's developing energy policy, particu-
larly in the eastern half of the country, the widespread use of FGD will
play an important role in near term SO  control strategies.  Expanded coal
use must be accompanied by sound air and water pollution control measures
to minimize threats to human health and to avoid widespread environmental
damage.

As the FGD technology most commonly accepted and used by domestic utilities
for coal-fired boiler applications, lime/limestone processes are especially
important to-the success of a policy to rapidly expand coal use.  Based on
the most recent estimate, 30 FGD systems are presently installed and an
additional 94 are in various stages of planning or construction.  These
124 units represent almost 50,000 MW of electric generating capacity, most
of which will be in service by the early 1980's.  Of these, 78 have already
selected lime/limestone processes and an additional 25 have not yet selected
a process type.

The major concerns of the utility industry to date regarding lime/limestone
scrubbing have centered around scaling and plugging potential, the large
quantities of by-product sludge generated, and the high costs (capital and
operating) of scrubbing.  It is toward these areas of concern that the
Shawnee program has been directed.

The Shawnee program has made major contributions toward improvement of lime
and limestone scrubbing technology in the areas of reliability, variable
load operation, system control, sludge disposal techniques, and process
economics.  The most significant results to date include:

   • Demonstration has shown that conventional lime/limestone systems can
     be operated reliably.  Two separate reliability problems have been
     identified—scaling and soft, mud-type solids deposits, and methods to
     control each have been demonstrated.

   • Soft, mud-type solids deposition was shown to be a strong function of
     alkali utilization.  At high alkali utilization (greater than about
     85 percent) these solids are much more easily removed, and very infre-
     quent intermittent fresh water wash is adequate for their complete
     removal in restricted areas such as the mist eliminator where accumu-
     lation can lead to plugging problems.

-------
                                       13
   • Several equipment or process variations were demonstrated to improve
     alkali utilization.  This is particularly significant for limestone
     where alkali utilization is typically about 60-70% and can be
     increased to 85-95%, or comparable to that normally obtained with
     lime.  This not only improves reliability, but also reduces costs
     by permitting the use of a much less expensive alkali feed material
     and by substantially reducing the quantity of by-product sludge
     produced.

   • The ability to operate during widely varying boiler load and inlet
     S0_ concentrations for extended periods was demonstrated on the
     venturi/spray tower system using lime with no reliability or system
     control problems.  This has long been a major concern of utilities
     for lime/limestone FGD systems.

   • Addition of MgO to either lime or limestone systems has shown that
     a substantial increase in S0~ removal efficiency can be obtained,
     and also indicates a good potential for forcing operation into' the
     gypsum unsaturated mode.  However, further work is needed to fully
     understand how to design and control such a system for subsaturated
     operation to avoid potential problems.

   • The initial results of staged scrubber forced oxidation studies at
     Shawnee appear very promising.  The ability to simultaneously
     obtain both high alkali utilization (over 90%) and essentially
     complete oxidation of sulfite to sulfate (gypsum) opens up numerous
     possibilities for further reducing costs and quantity of sludge
     produced while maintaining or even improving operating reliability.

   • Data generated during the factorial testing have been very useful
     in developing and checking the accuracy of the design models for
     scale-up to full-scale lime/limestone systems.  The models have
     also been useful in developing other valuable tools for industry
     such as the computerized Shawnee data base and the design/economic
     study computer program.  Furthermore, all of these will have increased
     value as they are expanded to include additional data and other
     equipment and process variations.

Although substantial progress and significant improvement have been made
over the past several years in lime/limestone scrubbing, to be of practical
use to industry, results of research and development efforts must be
accepted and applied by the utilities and FGD system vendors.  Acceptance
and application of the Shawnee results by several of the major FGD
system vendors has recently been more apparent, but this area of techno-
logy transfer to commercial application by utilities has, unfortunately,
been sluggish.  Consequently, a positive, applications-oriented program
involving a higher degree of participation and coordination by EPRI, the
utility industry, and FGD system vendors is now being considered as a
means of overcoming the apparent reluctance to accept and apply pilot
plant and prototype results to commercial size units.  The Shawnee pro-
totype test facility will in all likelihood play an important role in
this program continuation.

-------
                                   14
                                  REFERENCES
1.   Bechtel Corporation, "EPA Alkali Scrubbing Test Facility:  Summary of
     Testing Through October 1974."  EPA-650/2-75-047 (NTIS No. PB 244-901),
     June 1975.  The results of the initial test program from March 1972 to
     October 1974 are presented in this report.

2.   Bechtel Corporation, "EPA Alkali Scrubbing Test Facility:  Advanced
     Program - First Progress Report."  EPA-600/2-75-050 (NTIS No. PB-245-
     279/AS), September 1975.  The results of advanced testing from October
     1974 through April 1975 are presented in this report.

3.   Bechtel Corporation, "EPA Alkali Scrubbing Test Facility:  Advanced
     Program - Second Progress Report."  EPA-600/7-76-008 (NTIS No. PB
     258-783/AS), September 1976.  The results of advanced testing from
     June 1975 through mid-February 1976 are presented in this report.

4.   Aerospace Corporation, "Disposal of Flue Gas Cleaning Wastes: EPA
     Shawnee Field Evaluation Initial Report."  EPA-600/2-76-070  (NTIS No.
     PB 251-876/AS), March 1976.  The results of the field evaluation
     program from September 1974 through July 1975 are presented  in this
     report.

5.   EPA Technology Transfer Capsule Report, "First Progress Report:
     Limestone Wet-Scrubbing Test Results at the EPA Alkali Scrubbing Test
     Facility."

6.   EPA Technology Transfer Capsule Report, "Second Progress Report:
     Lime/Limestone Wet-Scrubbing Test Results at the EPA Alkali  Scrubbing
     Test Facility."

7.   EPA Technology Transfer Capsule Report, "Third Progress Report:
     Lime/Limestone We^t-Scrubbing Test Results at the EPA Alkali  Scrubbing
     Test Facility,"  EPA-625/2-76-010.

-------