-------
SO-> is removed by an aqueous solution at 50-6Q°C in a counter cur rent
absorber with 10-20 feet of packing or 3-6 trays.  The lean and rich
solutions arc cross-exchanged for optimum heat recovery.  Intermediate
storage of the lean and rich solutions permits decoupling of scrubber
operation from solution regeneration.  The SO^-rich solution is strip-
ped with live steam at 90-100°C in a counter cur rent packed or tray
column.  Water is condensed from the stripper overhead, stripped with
a small amount of steam in a separate column, and vaporized to produce
live steam for the main stripper.  The SC>2 product is normally a gas
containing about 95% SC>2 and 5% H20.  It would be converted by further
processing to sulfuric acid, liquid SC>2, or elemental sulfur.  Con-
tinuous purge treatment is necessary to remove any accumulation of
, ':">.•!• ite or chloride from the scrubbing solution.

     Three general types of aqueous absorbents have been proposed for
use with this flowsheet - buffers, alkali sulf ite/bisulf ite and alde-
hydes.  Buffers in the pH range of 4 to 7, such as sodium citrate,
provide for reversible S0« absorption as bisulfite by instantaneous
acid/base reaction:

                  S02(g) + H20 J  HS03" + H+

In a well buffered solution the S02 partial pressure will be pro-
portional to total dissolved SOo.  Alkali sulf ite/bisulf ite solution
is a special case of buffers where the net absorption reaction is
given by:
S02(g) + S03= + H20  j  2HS0
                                             3
In this case, the SC>2 vapor pressure is a nonlinear function of  total
dissolved SC>2 as given by:


                  p        [IISO ~]2
                  *SO  - K - - -
                           [S03~]

Aldehydes such as glyoxalic acid absorb S02 reversibly at low pH as a
hydroxysul f ona te (Green and Hine, 1974):

                  RC=0 + S02  + H20 •*  RCOH-S03~ + H+

Because pH is not constant, the SC>2 vapor pressure will be a non-
linear function of SO- absorption.

     Absorber/stripper design and performance for most steam regenera
tion alternatives is dominated by liquid/vapor equilibrium.  Four
simplifications can be made to idealize system steam requirements:

     1.  Assume an infinite number of stages or infinite height  of
         packing in the absorber 'and stripper.
     2.  Assume perfect cross-exchange of the lean and rich solutions
     3.  Assume a linear equilibrium relationship.
     4.  Assume no change of the ratio of H20 to SC^ vapor pressures
         with temperature.

                                  905

-------
Under these conditions the amount of steam required in the stripper is
equal to the total amount of water vapor contained in the saturated
flue gas.  In other words the ratio of 1*2° to S02 in the stripper
overhead is equal to the ratio of Ii20 to S02 in inlet flue gas satu-
rated to water.  Thus a typical flue gas containing 2000 ppm SC>2
and saturated with 15% 1^0 would require about 75 moles steam/mole
S02 for steam stripping.  Note with these simplifications that the
level of S02 removal does not affect steam requirements.

     In practice a finite number of contacting stages must be used
in the absorber and stripper.  Depending on the relative costs of
steam and equipment, this consideration would increase steam require-
ments by a factor of 1.2 to 1.5.  The optimum number of stages would
also be directly related to the required level of S(>2 removal.

     The assumption of perfect cross-exchange results in a system
where the steam requirements are independent of solution capacity fox
S02 absorption.  In practice the solution capacity for S02 absorption
should be greater than 0.05-0.10 moles/liter in order to minimize the
size and cost of cross-exchange.  With acid/base absorption of S02
this criteria implies a solution of pH greater than 4-4.5.  With
high solution capacities the assumption of perfect cross-exchange is
generally acceptable.

     The sodium sulfite/bisulfite system gives a highly nonlinear
equilibrium relationship of liquid and gas 862 concentrations.  As a
result, in an optimized system gas and liquid are nearly at equili-
brium in the top and bottom of the absorber and the middle of the
stripper and far from equilibrium in the top and bottom of the strip-
per and the middle of the absorber.  Ideally, the approach to gas/
liquid equilibrium should be about the same throughout the absorber
and stripper.  At 90% S02 removal in the absorber, the nonlinearity
of sodium sulfite/bisulfite equilibria increases the steam require-
ments by a factor of 1.5.  At 97% 862 removal, nonlinearity in-
creases the steam requirements of this system by a factor of 3.0.
Steam requirements vary with S02 removal in the absorber because
it becomes Increasingly more difficult to strip S02 from the scrub-
ber feed.  As S02 is stripped from sulfite/bisulfite solution, the
pH increases and the next increment of S02 is even more difficult
to remove.

     Johnstone found for most solutions of weak acid buffers or
alkali sulfite/bisulfite that the ratio of I^O to S02 vapor pressure
was only a weak function of temperature (Johnstone, 1935;  Johnstone
et al, 1938),  Therefore, the fourth simplification is valid when
scrubbing with sodium sulfite/bisulfite solution.  As a result steam
requirements with sodium sulfite/bisulfite solution will be 1.8 to
2.2 times more than the ratio of H20 to S02 in the scrubber inlet,
because of the need for finite contactors and the nonlinearity of
the absorption stripping equilibria.

Gas Cooling

     The steam requirements of a simple absorption/stripping process
with any aqueous absorbent can be reduced by scrubbing at temperatures

                                  906

-------
below the adiabatic saturation temperature of the flue gas.  This
would be achieved by direct or indirect nonadiabatic cooling of the
flue gas.  Reduced scrubbing temperature results in a lower concen-
tration of H20 in the flue gas and thereby reduces the ratio of i\2®
to S(>2 vapor pressure in the solution from the absorber.  As a result,
with weak acid buffers or alkali sulf ite/bisulf ite , absorber cooling
from 55°C to 35°C will reduce steam requirements almost a factor of
three.  Gas cooling may also enhance the rates of S02 mass transfer
and reduce the volatility of solution components such as NH3 and
organic acids.

     Such a concept has been practiced by the Russians at Niiogaz.
They cooled the flue gas to 35°C by direct contact with once-through
cooling water which was neutralized and disposed of.  S02 was ab-
sorbed by ammonium sulf ite/bisulf ite solution.  With an inlet SC>2 gas
concentration of 3000 ppm, the steam requirement for stripping was
27 moles H20/mole S02 (Slack and Hollinden, 1975).

     Commercial practice in the U.S. would not allow for direct con-
tact cooling with disposal of waste water.  Cooling would probably
be achieved, as in Figure 2, by direct countercurrent contact of
flue gas with cold recirculating solution.  The recirculating solu-
tion would be indirectly cooled by water in a heat exchanger made of
corrosion-resistant materials such as Incoloy 625.  The temperature
driving force available for heat exchange could be increased by per-
mitting high concentrations of soluble additives such as CaCl2 to
accumulate in the recirculating solution.  Clearly, such a cooling
system would not be inexpensive nor maintenance-free.  Therefore,
non-adiabatic gas cooling is an alternative for reducing steam re-
quirements that has not generally been exploited.

Alleviating Equilibrium Nonlinearity

     Steam requirements could be reduced as much as 35% by alleviat-
ing the nonlinear characteristics of the gas-liquid equilibrium of
sodium sulf ite/bisulf ite solution.  Johnstone recognized the value of
using an additional buffer to give a more nearly linear equilibrium
relationship.  At a constant pH in the range of 4-6, his data show
that S02 vapor pressure is directly proportional to total dissolved
S0:
                               M


Thus a solution buffered at a constant pH would give a linear equili-
brium relationship.  Johnstone (1935) tested phosphate buffer.  The
Bureau of Mines and Peabody have tested sodium citrate buffer.  Other
potentially-effective buffers include phthalate, adipate, and
succinate (Rochelle, 1977).

     With a sulf ite/bisulf ite' buffer, the effect of nonlinearity can
also be essentially eliminated Cat 90% SO, removal) by using a split

                                 907

-------
o,
150UC
FLUE
GAS
       35°C
                             LEAN
                             SOLUTION
                 35°C     RICH
                 	» SOLUTION
                 35°C

                             (CAO)
C,W,
        L/G=30 GAL/MSCF
       Figure 2. Nonadiabatic gas cooling
                908

-------
feed to the absorber as in Figure 3.  Solution is withdrawn from  the
middle of the stripper and fed to the middle of  the absorber.  Thus,
there is a higher rate of liquid circulation through  the bottom of
the absorber and the top of the stripper, where  the pH and solution
capacity is generally lower.  This flowsheet is  generally used in
absorption/stripping systems for CC>2 concentration  (Kohl and
Riesenfeld, 1960).

Temperature Effects

     Johnstone  (1938) found for most weak acid buffers and sulfite/
bisulfite solutions that the ratio of SC^ to H20 vapor pressure was
only a weak function of temperature.  Steam requirements for strip-
ping will be lower if the ratio of SC>2 to 1^0 vapor pressure increases
with temperature.  Table 1 gives the factor by which  this ratio in-
creases in going from 55°C to 100°C for several  absorbent solutions.
For sulf ite/bisulf ite solutions, the factor varies  from 0.84 (Na+)
to 1.15 (NH4+) .  The systems using NH^+ and CH3NH3+ give a somewhat
greater temperature factor because they have some weak base chara-
cteristics.  The temperature factors for weak acid  buffers appear
to be in the range of 0.9 to 1.4.

     Johnstone  recognized that weak base buffers such as dissolved
aluminum and aromatic amines give higher temperature  factors, in  the
range of 1,8 to 5.8 for 55°C to 100°C.  The pKa  values of these
buffers decrease rapidly with temperature.  Therefore, the pH of  a
given solution  will decrease with heating, thereby  increasing the
862 vapor pressure.  Aniline and other aromatic  amines such as tolui-
dine (Weidmann  and Roessner, 1936) and dimethylaniline (Fleming and
Fitt, 1950) have been proposed for and used in absorption/stripping
systems at low  gas temperature (25°C) .  However, their volatility
makes them unacceptable alternatives for use with large gas volumes
at 55°C.  Alo(SO^)^ and A1C13 solutions are effective only at pH
less than 4-4.5 because of the limited solubility of aluminum hydrox-
ide and basic aluminum sulfate at higher pH.  Therefore, to get
adequate solution capacity the flue gas must be scrubbed at 30-40°C.
Nevertheless, with flue gas cooling the steam requirements of systems
using these solutions would be quite attractive.  The polyamines,
ethylenediamine and diethylenetriamine , were briefly characterized
by Roberson and Marks (1938).  They should be nonvolatile under
scrubber conditions because of the multiple hydrophilic groups which
will usually be partially protonated.  Ethylenediamine has a first
pKa value of 7.0 at 20°C which decreases to 6.07 at 60°C and 5.34
at 100°C (Mclntyre et al, 1959).   It is therefore suitable for use
at 50-60°C, especially if dissolved chloride or sulfate is allowed
to accumulate so that the stripped solution at pH 6 contains little
dissolved sulfite.

     Johnstone also recognized that systems containing acids as
separate liquid or solid phases could give greater temperature
factors.  If the acid is more soluble at higher temperatures it will
dissolve and thereby reduce the pH of the solution as it is heated.
The reduced pH increases SC>2 vapor pressure.  Johnstone (1938) identi-
fied valeric acid as a possible alternative, but it is too volatile for
                                909

-------
 LU
 Q_
 CL-
 h-
00
 LU
 o:
 o
 CQ
                                                                          (LI
                                                                          M-
                                                                          O
                                                                          CO
                                                                          Q.
                                                                          co
                                      910

-------
Table 1:   TEMPERATURE FACTORS
                              (Fso2/'pii2o)ioo0c
                                   / P
                                   '
Absorbent






Sulfite/bisulfite



Na+



K+



NH.+
  4


CH3NH3+



Monethanolamine



Ethylenediamine



Weak Acid



Citrate, Na+





Phosphate



Sulfosuccinate





Weak Base
                 Reference
Al+3, Cl-



Al+3, SO,"



Aniline



Ethylenediamine, Cl*



Diethylenetriamine



Separate Acid Phase



Valeric



Adipic, pH-4.5
K2S2°5
  0.84



  1.01



  1.15



  1.10



0.7-1.2



  1.1







0.9-1.4





1.0-1.2



  1.2
  3.7



1.8-3.0



  3,3



  6*



2.4-5.8







1.8-3.2



  5-6*



  5-6*



  3,5*
Johnstone et al, 1938



Linek and Hala, 1967



Johnstone, 1935



Johnstone et al, 1938



Roberson and Marks, 1938



Roberson and Marks, 1938
Rosenbaum et al, 1971

Oestreich, 1976



Johnstone, 1935



Keller and Wiseman,

1950
Johnstone, 1935



Applebey, 1937



Johnstone et al, 1938



Rochelle, 1977



Roberson and Marks, 1938







Johnstone  et al, 1938



Rochelle, 1977



Rochelle, 1977



Rochelle, 1977
*Estimated from pK  values and solubilities
                  Si

                                   911

-------
practical use,  Adipic acid, K2S205, and KI19P04 are all acid solids
whose solubilities increase rapidly with temperature.  Careful de-
sign would be required to avoid the crystallization of the acid
solids in the scrubber.

     Aldehydes such as glyoxalic acid (Marcheguet and Garden, 1967)
and glyoxal (Marcheguet, 1961) have totally different characteristics.
The reversible reaction of 862 with an aldehyde is very sensitive to
temperature.  The temperature factor from 55° to 100°C of a typical
aldehyde, such as benzaldehyde, will be 15-20 (Stewart and Donnally,
1932).  Therefore, steam requirements are determined more by sensible
heating than by stripping needs.  The limitation of aldehydes is
their lack of effectiveness at a normal scrubbing temperature of 50-
60°C.  In order to get sufficient solution capacity and/or adequate
rates of mass transfer, developers of the process using glyoxalic
acid have found it desirable to cool the flue gas to 35°C (Stark
et al, 1976).  In the future, other nonvolatile aldehydes may be
identified which do not have this limitation.

SOLUTION EVAPORATION

     With a system configuration as in Figure 1, water condensed from
the S02 product is used to produce live steam for the stripper.  A
more common arrangement proposed by system developers is indirect
evaporation of the stripper bottoms solution to produce stripping
steam.  Condensate from the.SOo product is usually returned to the
top of the main stripper, thereby avoiding the need for a condensate
stripper.  This configuration is simpler than that of Figure 1, but
requires an evaporator constructed of corrosion resistant material.
Furthermore, under some conditions return of condensate to the
stripper will significantly increase steam requirements.

     If the amount of solution evaporated during regeneration is a
large fraction of the total circulating solution, then the solution
can be significantly more concentrated in the evaporator than in the
rest of the system.  In the extreme, such evaporation can result in
the crystallization of sulfite, bisulfite, sulfate, or buffer salts.
Furthermore, H20 condensed from the S02 product must be recycled to
some point in the system" where it will significantly dilute the
working solution.  The condensate can also be used to redissolve
crystalized solids.  As a result of concentration/dilution or
crystallization/dissolution, the SC^ vapor pressure can be signifi-
cantly affected at the respective points in the system.  In the
absence of additional liquid or solid phases (besides aqueous solu-
tion and gas) or with the crystallization/dissolution of basic solids
such as Na2SOj or buffer salts,solution concentration will increase
the S0£ vapor pressure.  Dilution by H^O condensate will reduce the
SOo vapor pressure.  With the crystallization/dissolution of acid
solids such as 1^28205 or buffer acids, evaporative concentration will
reduce the S02 vapor pressure, while H20 dilution will increase it.

     These dilution/concentration effects can be effectively used to
reduce steam requirements and/or the number of 'stages in the stripper.
The Wellman-Lord process evaporates sodium sulfite/bisulfite solution
with the crystallization of ^2803.  The effect of concentrating the
solution is so great that only one stage of evaporation/stripping is

                                 912

-------
adequate to strip out the SC>2'  Dilution water is used to make up
scrubber feed solution by redissolving the Na2SC>3 solids.  The actual
steam requirement is directly related to the solubility of the sulfite
salt.  Lower steam requirements could potentially be achieved by
crystallizing more soluble sulfite salts such as ammonium sulfite or
methylammonium sulfite, thereby reducing the amount of evaporation
required to crystallize the salts.

     With high capacity buffer solutions where as much as 50% of the
solution would be evaporated to produce steam for the stripper,
dilution water should be added to the scrubber feed.  This permits
the stripper to operate with fewer stages.  Such placement of dilution
water would necessarily require the use of a condensate stripper.
Dilution water should not be added to the stripper feed or at the top
of the stripper because this would reduce the S02 vapor pressure of
the solution and thereby increase the steam requirements.

     If steam stripping is used in the presence of acid solids such
as 1^28205, the dilution water should be added to the top of a counter-
current stripper where it will dissolve additional 1^28205 and thereby
increase the S02 vapor pressure and reduce the steam requirements.
Stripping steam would be provided by evaporation of bottoms solution
from the stripper.  The concentration/dilution effects should serve
to reduce the net steam requirements.  It is important that this
stripping be carried out in a countercurrent stripper, because
solution evaporation and crystallization of K2S205 in the bottom stage
will significantly reduce the S(>2 vapor pressure over that stage. '

HEAT RECOVERY

     Almost all systems with steam regeneration include condensation
of water from the product S02«  Heat recovery from this condenser can
result in substantial energy savings.  At atmospheric pressure 90%
of the H20 vapor can be condensed at 90-95°C.  Possible methods of
heat recovery include multiple-effect evaporators/strippers, vapor
compression cycles, use or production of low pressure steam, and other
methods of system integration.

     Multiple-effect evaporation or stripping as in Figure 4 is
particularly attractive.  Condensation of H20 from the product of
the first-effect, higher-pressure stripper is used to produce steam
from the second-effect stripper,  which operates at lower temperature
and pressure.   The number of effects is limited by the maximum temper-
ature, usually 100°C to avoid sulfite disproportionation, and by the
need to avoid  excessively low pressure in the final effect.   Practi-
cally, 'two or  three effects should be feasible.   The system steam
requirements would essentially be inversely proportional to the
number of effects.   However, multiple effects necessarily increase
the operating  and design complexity of the system.  The need for
multiple parallel strippers can result in capital cost increases,
but in the case of large systems,  multiple strippers may be required
to handle the  full stripping load in any gas.

     Steam generated during condensation of H20 from the product SC>2
could be utilized in a single effect if its pressure is increased by

                                913

-------
                                                         D1
                                                         Q.
                                                         a.
                                                         O
                                                         O)
                                                         14-
                                                         q-
                                                         QJ

                                                         a>

                                                         -Q
                                                         ^
                                                         O
                                                         O)
                                                         S-
                                                         cr>
914

-------
mechanical compression or by a steam ejector,  Mechanical compression
would permit the stripper to operate on a minimum amount of electric
power, rather than relying on steam generated in the main boiler.
Assuming that steam could be generated at 85°C/0.6 atm in the con-
denser, approximately 15 kwh of power would be required to recycle
one million Btu's of steam,  A steam ejector would permit compression
to be accomplished at minimal capital cost without rotating equip-
ment.  About 0.4 Ibs of steam at 600 psia would be required to provide
1.0 Ibs of steam heat to a single-effect evaporator at 90°C with an
allowance of 15°£ for driving force and boiling point elevation.
     Many steam regeneration systems can be effectively operated below
atmospheric pressure at temperatures as low as 70-80°C.  Under these
conditions it is attractive to operate the stripper or evaporator with
atmospheric or vacuum steam exhausted from a turbine.  Thus the
effective cost of energy for the steam stripping system can be sub-
stantially reduced.  Integration with exhaust from a power turbine
could be difficult because low pressure steam would have to be trans-
ported over a large distance.  However it is sometimes attractive to
use moderate-or high-pressure steam to drive fans and pumps in the
flue gas desulfurizat ion-system , and then to use exhaust steam from
the turbine drivers in the stripping system.

     Other miscellaneous methods of heat recovery could also be signi-
ficant.  Boiler feedwater could be preheated by condensing water from
the S02 product,  However, this would require extensive integration
with the power production system.  The ^0 condenser could be used
to preheat feed to the stripper, thereby eliminating the need for a
cross-exchanger,

SULFATE FORMATION AND REMOVAL

     In most steam regeneration systems some sulfate is formed
irreversibly by the oxidation and/or disproportionation of bisulfite.
Unless intentionally removed by other means, dissolved sulfate x^ill
accumulate in the system to the point of uncontrolled crystallization
of the sulfate salt.   The accumulation of dissolved sulfate reduces
the effective solubilities of other salts by the common ion effect.
In systems where solubilities determine steam requirements, such as
the Wellman-Lord process, sulfate accumulation will degrade system
performance.  Furthermore any sulfate formation ultimately results
in a sulfate byproduct and usually requires alkali makeup.   Therefore
it is desirable to minimize sulfate formation and provide for effec-
tive sulfate removal.

     As much as 5 to 10% of the S02 absorbed by a system such as the
Wellman-Lord process may be oxidized to sulfate in the absorber.
However, there is little information on the variables affecting bi-
sulfite oxidation.  The physical absorption and mass transfer of
oxygen through the liquid phase is probably important.  Therefore,
oxidation rates should vary directly with oxygen concentration in the
flue gas and should be greater in absorbers with larger amounts of
mass transfer capability.  It is also expected that 'flue gas impuri-
ties such as iron (from the flyash) and nitrogen oxides may act as
catalysts for the oxidation (Graefe et al, 1970).  There is some

                                  915

-------
evidence that any NC>2 in the flue gas can react stoichiornetr ically
           ite to give sulfate and N£ or NH4+ (Sawai and Gorai, 1977),
with bisulfite to g
Prescrubbing of the flue gas could remove flyash and possibly NC>2
and thereby reduce rates of oxidation in absorber.  Prescrubbing could
also remove HC1 and SO-j which would otherwise be absorbed and accumu-
late as dissolved chloride and sulfate.  Oxidation inhibitors have
been tried with limited success in the Wellman-Lord process
(Pedroso, 1976).

     Sulfate may also be formed by the disproport ionation of bisulfite
to thiosulfate  (S203=) and sulfate.  The reaction probably proceeds
as the formation of trithionate followed by hydrolysis  (Battaglia
and Miller, 1968;  Foerster and Hornig, 1922):


          2H* + 4HSO ~ + S00,,= -* 2S-0,=! + 3H00
                    J     L J  •*•   o o      /

                 2S..O,3 + 2H20 ->_ 250^" + 2S203~ + 4H+


                        4HS03~ J 2S04=< + S203" + 2H+ +  H2<)

The rate expression for trithionate formation  is given  by:

          d[HSO "]             m 3
            d(.  J --- k[H+]J[S203 r[HS03 ]


                 k=2.2«106M~6 sec"1 at 70°C

           (Battaglia and Miller, 1963)

Since  the net disproportionat ion produces thiosulfate,  which also
shows  up in the rate expression, the reaction  is autocataly tic .  As
more thiosulfate accumulates the disproportionat ion occurs  even
faster.  Low pH also appears to give faster rates of disproportiona-
tion.  Significant dispropor tionation has been observed in  the
Wellman-Lord Process (Bailey, 1974) and the  NH-j-based  steam-stripping
process when regeneration is carried out at temperature greater  than
100°C.  Dispropor tionation has been essentially eliminated  in  the
Wellman-Lord process by operating a vacuum evaporator at 90-100°C
with a liquid purge to minimize thiosulfate accumulation.   When
using  an aluminum sulfate scrubbing solution,  I.C.I, avoided dis-
proportionat ion by boiling a portion of the liquor with copper
sulfate in the  presence of SC>2 to eliminate thiosulfate from the
system (Applebey, 1937).  Potentially, disproport ionation could  be
reduced by minimizing solution holdup in the stripper and evaporator.
Thus the stripper should use low-holdup packing rather  than  trays and
solution contained in the evaporator should either be pure  water or
scrubbing  solution containing no dissolved S02-

     Oxidation  and disproportionation may be substantially  reduced  in
systems using aldehydes such as glyoxalic acid.  These  systems nor-
mally  operate at very low pH conditions where  there is  no dissolved
bisulfite  and very little dissolved S02-  All  of the S02 is  absorbed

                                  916

-------
as the hydvoxy-sulfouac,e complex,  therefore, rates of, oxidation and
disproportionation of the dissolved SC>2 species  should be greatly
inhibited.

     Sulfate can be removed from steam regeneration systems by solu-
tion purge or by Che controlled crystallization  of salts such as
Glauber's salt  (NaSO,•10H-0) or gypsum (CaSO  »2H20).  Direct solution
purge is attractive only ror sys'tems with inexpensive absorbents con-
sisting primarily of dissolved sulfate, such  as  aluminum sulfate.
Absorbent systems that use sodium alkali, such as sodium citrate or
sodium sulfite/bisulfite, can selectively remove sulfate by the re-
frigerated crystallization of Glauber's salt.  Buffered systems can
usually crystallize gypsum from stripped solution by adding lime or
limestone (Applebey, 1937).  Crystallization  of  gypsum from alkali
sulfite/bisulfite systems usually requires complicated processing
with the addition of both lime and S02 (Johnstone and Singh, 1940).

MORPHOLOGICAL PROCESS ALTERNATIVES

     A number of evolutionary alternatives have  been generated for
reducing steam  consumption and handling sulfate, as discussed in
the preceding sections.  Many of these innovations can be used
simultaneously, so a large number of combinations are possible.
Table 2 represents independent groups of alternatives.  A complete
process can be  defined by selecting one alternative from each group.
The alternative groups are distinguished as being reactant alterna-
tives or flowsheet alternatives.

Reactant Alternatives

     In general the scrubbing solution must use  at least one alkali
and may or may  not use an acid buffer, except in systems which use
only aldehydes.  The alkali additives are divided into three groups.
The strong alkalis - Na+, K+, Mg++ - are usually the least expensive
and are nonvolatile.  Na+ has been the dominant  alkali selected in
development work, but K+ and Mg++ have unique solubility characteris-
tics which could make them more or less attractive.  The weak alkalis
NH^+,  methylamine, ethanolamine - act as buffers at pH 8-10 and are
therefore suitable for sulfate removal as CaSO^.  They also generally
have large solubilities.  However, NH3 has problems with volatility
in the scrubber and other amines tend to be quite expensive.  The
buffering alkalis - Al+  , ethylenediamine - potentially reduce steam
requirements by straightening the equilibrium cruve and by increasing
the temperature coefficient of S02 vapor pressure.

     The acid buffer alternatives include the use of no acid additive
or the use of several possible acids with significant buffer capa-
cities in the pH range 4.5-6..0.   With no acid additive, the buffer
system must be sulfite/bisulfite or a buffering base.   Thus, only
sulfite or pyrosulfite salts are likely to crystallize in the case of
solution evaporation.   The use of buffering acids straightens the
equilibrium curve and adds a degree of freedom to the solids that can
be present in saturated solution.  The buffers would also permit
effective use of gypsum crystallization for sulfate removal.  Lower
pH buffers such as glycolate" (pKa^S.SS) are probably not attractive
because of low solution capacity for S02 at the low pH,

                                  917

-------
Table 2:
Reactants
MORPHOLOGICAL ALTERNATIVES
     1.   Alkali  —  (a)  strong - Na+, K+, Mg
     2.  Acid
     3.  Aldehyde--
          (,b)  weak - NH, , methylamine, ethanolamine


                                            + 3
          (c)  buffering-ethylenediamine, Al



          Ca)  none



          (b)  high pK  - phosphoric, adipic, citric,
                      3,


                          phthalic, sulfosuccinic



               glyoxalic acid, glyoxal
Flowsheets
     4.  Evaporation/dilution  —  none, stripper bottoms/scrubber



                                   feed, stripper bottoms/stripper



                                   feed



     5.  Number of scrubber feeds  —  one, two



     6.  Additional phases  —  none, one, two



     7.  Heat recovery  --  single-effect, multiple-effect, vapor



                            compression, low-pressure steam



     8.  Sulfate removal  --  purge, Glauber's salt, gypsum,



                              other sulfate
                                 918

-------
     Aldehyde reactants - glV'Qxa,lic a,cid, glyoxal - are unique in
requiring no alkali additives.  Tlxey should generally result in
excellent steam requirements, but may* have to be used with lower gas
temperatures, resulting in high costs for flue gas cooling.

Flowsheet Alternatives

     There are five important gr'oups of flowsheet alternatives.  These
can interact quite strongly with the selection of reactants and/or
the selection of other specific flowsheet alternatives.

     If solution evaporation is used for steam generation, the result-
ing dilution water should not be added to the stripper feed unless
pyrosulfite or buffer acid solids are present.  Avoiding solution
evaporation can be advantageous to the extent that it avoids corrosion-
resistant materials in the evaporator and eliminates boiling point
elevation (important with some methods of heat recovery).

     A dual scrubber feed eliminates the detrimental effect of non-
linear equilibria.  Since buffers accomplish the same purpose, the
combination of buffer additives with a dual scrubber feed is not
attractive.

     A scrubber system can be operated with additional solid or liquid
phases with or without solution evaporation.  The effect of an addit-
ional phase interacts with reactant selection and the mode of
evaporation/dilution.

     Heat recovery options can usually be incorporated with any
reactant or flowsheet combination.  However, multiple-effect evapora-
tion or low-pressure steam would be less attractive with systems us-
ing temperature e'ffects to reduce steam requirements, since portions
of the regeneration would occur at lower temperature.

     The optimum method of sulfate removal interacts strongly with
the reactant selection.  It may also vary depending on the mode of
evaporation/dilution,

CONCLUSIONS

     1.  It will be difficult to develop a steam regeneration system
which is superior to the Wellman-Lord process in steam requirements
and simplicity, though there are many possibilities.

     2.  Without gas cooling or solids crystallization, ethylene-
diamine sulfate or chloride solution may be effectively used with
lower steam requirements than Wellman-Lord.

     3.  A properly designed system based on the crystallization/
dissolution of 1(28205 solids should give steam requirements compe-
titive with Wellman-Lord.

     4.  Aluminum salt solution could be a competitive scrubbing
solution when used in combination with gas cooling.

     5.  Nonvolatile aldehydes have a definite potential for reduced
steam requirements, although they may also require non-adiabatic flue
gas cooling.

-------
     6.  The use of evaporation and/or crystallization generates a
number of possible alternatives that have not been fully  explored,

ACKNOWLEDGEMENTS

     The author was supported during the period of this work by a
graduate fellowship from the National Science Foundation,

REFERENCES

1.  Applebey, M. P., J. Soc. Chem. Ind. Trans., 56,  139  (.1937) .

2.  Bailey, E. E,, "Proceedings:  Symposium on Flue  Gas Desulfurization
  .  - Atlanta", EPA-650/2-74~136b, p. 745  (1974).

3.  Battaglia, C. J, and W. J. Miller, Phot. Sci. Eng.. 12, 46  (1968).

4.  Davis, J. C., Chem. Eng.. 7jK27),43 (1971).

5.  Fleming, E. P. and T, C, Fitt, Ind. Eng. Chem..  42. 2253  (1950).

6.  Foerster, F. and A. Hornig, A.norg. Allgem. Chem....  125,  86  (1922).
    l
7.  Graefe, A. F., L. E. Gressingh, and F. E. Miller,  PB  196-781,  U.S.
    EPA, U970),

8,  Green, L. R. and J. Hine, J.._  Org. Chem. . 39.  3896  (1974).

9.  Johnstone, H. F., Ind. Eng. Chem.. 27, 587 (1935).

10. Johnstone, H. F., H. F. Read, and H. C. Blankmeyer, Ind. Eng.
    Chem. . 3jO, 101 U938) ,

11. Johnstone, H. F. and A. D. Singh, Univ. of 111.  Eng.  Expt.  Sta.
    Bull., 324 C1940).

12, Kohl, A. L. and F. C. Riesenfeld, "Gas Purification", McGraw-Hill
    Book Co,, New York, C1960).

13. Marcheguet, H. G. L., U.S. Patent 2,994,585 (1961).

14. Marcheguet, H. G. L. and L. Gardon, U. S. Patent 3,350,165  (1967).

15. Mclntyre, G. H. , B. P. Block, and W. C. Fernelius, J_._ Amer. Chem.
    Soc. , 8.1, 529 (1959) .

16, Nissen, W. I.,  D. A. Elkins, and W. A. McKinney,  "Proceedings:
    Symposium on Flue Gas Desulfurization - New Orleans", EPA-600/
    2-76-136b, p. 843, 1976,

17. Pedroso, R. I., "Proceedings:  Symposium on Flue Gas  Desulfuri-
    zation - New Orleans",' EPA-600/2-76-136b, p.  719,  C1976) .

18. Roberson, A. H,, G. W, Marks, U,S. Bureau of  Mines Rep, of  Invest.
    3415 (1938).

                                  920

-------
19.  Rochelle, G, T,, Electric Power Research Institute Report No.
    FP-463-SR, July, (1977).

20.  Rochelle, G. T. and C. J. King,"Proceedings of the Second Pacific
    Chemical Engineering Congress", p. 261 (1977a).

21.  Rochelle, G. T. and C. J, King, "Process Alternatives for Stack
    Gas Desulfurization with H2S' Regeneration to Produce Sulfur",
    presented at AIChE 70th Annual Meeting, New York, November, 1977b.

22,  Sawai, K and T. Gorai,"Proceedings of the Second Pacific Chemical
    Engineering Congress", p. 340 (1977),

23.  Schneider, R. T. and C, B. Earl, "Proceedings:  Flue Gas Desul-
    furization Symposium", EPA-650/2-73-038, p. 641, (1973).

24.  Slack, A. V. and G. A. Hollinden, "Sulfur Dioxide Removal from
    Waste Gases", Noyes Data Corp., Park Ridge, New Jersey  (1975).

25.  Stark, W. H., A. A. Syme, and J. C. H. Chu, "Proceedings:
    Symposium on Flue Gas Desulfurization - New Orleans", EPR-600/
    2-76-136b, p. 981 (1976).

26.  Stewart, T. D. and L. J.  Donnally, J. Am. Chem. Soc. , 54, 3555
    (1932) .

27.  Weidmann, H., G, Roesner, Metallges. Periodic Rev., February (1936)
                                  921

-------
APPLICATION  OF DRY  SORBENT  INJECTION  FOR SO2
               AND  PARTICULATE  REMOVAL
             N. D. Shah, D. P. Teixeira, and R. C. Carr
                         Air Quality Control
                 Fossil Fuel Power Plants Department
                         Palo Alto, California
  ABSTRACT

      Integrated processes designed to remove two or more pollutants
  simultaneously from  a  coal-fired boiler flue gas  are of considerable
  interest from  both an economic and operational standpoint. One such
  approach is through the injection of dry sorbent in the flue gas duct
  followed by collection of spent material and fly ash in the electrostatic
  precipitator or baghouse.
      A review  of literature indicates that most of the work conducted to
  date is confidential and concentrated in specific areas for sales purposes
  without due  consideration to understanding  the basic parameters
  involved in the  process. Previous investigations have confirmed that
  alkaline material containing calcium and  magnesium are relatively
  ineffective in removing S02,  while the  alkaline materials containing
  sodium have been identified as attractive dry sorbents. A higher flue gas
  temperature generally results in better S02 removal, but basic informa-
  tion concerning  kinetics and  thermodynamics of the heterogeneous
  reaction is lacking. It has been concluded that a systematic study of the
  process under well-characterized and controlled conditions simulating
  practical utility boiler applications is  necessary before the technology
  can be considered for constructing commercial installations.
      EPRI's current efforts in dry SO2 removal research are focused on
  bench-scale research to provide a data base that will define the range of
  operating parameters for future  pilot- and full-scale  installations. The
  effects of sorbent type, residence time, temperature, particle size, and
  stoichiometric ratio on  S02 removal  efficiency and sorbent utilization
  will be evaluated.
      Future plans include formulation of a complete program at a larger
  scale based on the data obtained from the bench-scale study. Additional
  future work will  concentrate on the problem concerning the effect of
  sorbent injection on air preheaters, waste disposal of the spent material,
  availability  and  supply of various  sodium  based sorbents, and  an
  economic evaluation.
                                  922

-------
                   APPLICATION OF DRY SORBENT INJECTION FOR
                         S02 AND PARTICULATE REMOVAL
INTRODUCTION

To comply with the emission standards promulgated by EPA for coal-fired
boilers, a considerable effort towards the development of sulfur oxide removal
equipment has been mounted.  To date, efforts have focused on wet
lime/limestone scrubbing of Eastern high-sulfur coals.  However, interest in
developing a practical and low cost 502 removal system for Western low sulfur
coal has increased significantly in recent months due to pressures from local
regulatory bodies and passage of the 1977 Clean Air Act Amendments.  Dry
sorption of 502 is a relatively new technology but indicates promise of
becoming a commercially viable process.

Dry 502 cemova^ possess features which show promise for lower overall
capital/operating costs and greater reliability compared to wet scrubbing.
The advantages of dry scrubbing are:

•    Lower capital cost since particulate and 802 removal can be achieved in a
     single device.
•    Anticipate greater reliability and lower maintenance due to the
     simplicity of the process.
•    3-5% energy savings compared to wet scrubbers.
•    Savings of 1 GPM/MW wet scrubber water consumption.
                                    923

-------
While the process sounds attractive, it does have some disadvantages:

•    Moderate S02 removal efficiency (60-70%).
•    High reagent cost
•    Problems in connection with availability, supply and transportation of
     the sorbents.
•    Lack of operating experience.
•    Spent sorbent disposal requirements unknown/lack of commercially
     available regeneration process.

It appears from the above analysis that dry scrubbing shows promise for the
western part of the country where only moderate S02 removal is needed and
reagents are generally available.

LITERATURE SURVEY

EPRI's report No. FP-207^3^ "Evaluation of Dry Alkalis for Removing Sulfur
Dioxide from Boiler Flue Gases" summarized the various known aspects of dry
scrubbing processes.  From the review of this report, it appears that raost of
the development work^ »''»»'' conducted to date is oriented towards
sales promotion without due consideration to providing specific design
criteria relative to the process.  Basic technical information pertaining to
energy consumption, resource requirements, capital and operating costs is
lacking.

Tests conducted by Air Preheater Co.,  ' at the Public Service of New Jersey's
Mercer station are probably the best to date in terms of identifying and
quantifying the effect of the major controllable variables in the process.
These tests (see Table 1) and work conducted by others^ »^ have shown
significantly higher reactivity of alkaline material containing sodium
compared to calcium and magnesium.  For this reason, future work is expected
to be concentrated in investigating various sodium based dry alkaline
materials.
                                      924

-------
Temperature and stoichiometric ratio were identified in the Mercer study^ '  as
important process variables.   Table
                                     ^ '
       shows that
                                                        removal efficiency
increases as the temperature and stoichiometric ratio are increased.   However,
no attempt was made to explain these results in terms of rate controlling
steps for the heterogeneous reaction.  Since both duct and baghouse were
operated at high temperatures, extrapolation of the data presented in Table 1
to full scale where collection occurs at low temperature, is questionable.
Data presented in Table 1 for sorbent utilization should be used with caution
because of the questionable validity of the assumptions^ ' made.
                                   Table 1
                          Results from Mercer Tests
                          Stoich. Ratio = 1
                 Stoich. Ratio = 3

Additive
Sodium Bicarbonate-270°F
Sodium Bicarbonate-350°F
Sodium Bicarbonate-600°F
Nahcolite -350°F
Nahcolite -600°F
Conv.
Eff .*
%
32
48
90
65
94
Util.
Eff.*
%
32
48
90
65
94
Conv.
In Gas*
%
—
26
72
11
60
Conv.
Eff.
%
48
76
—
85
—
Util.
Eff.
%
16
26
—
30
—
Conv.
in Gas
%
12
12
—
24
—
Hydrated Dolomite
  Lime            -350°F
Hydrated Dolomite
  Lime            -600°F   20
20
                                         95
20
38
 7
13
75
80
     *Nonienclatures:
      Conv. Eff. = SC>2 Conversion Efficiency
      Util. Eff. = Additive Utilization Effectiveness
      Conv. in Gas =  SCU Conversion in the gas stream or in suspension,
Tests by Wheelabrator-Frye Inc.^ '  at the Nucla Station of Colorado Ute
Electric Association showed SC>2  removal efficiency in the range of  50 to 70%.
Since these tests were conducted on a confidential basis, details of the
tests — such as particle size,  temperature-time history — are not readily
available.
                                     925

-------
                         to   oo
    oo
    o
                         o

                         ii
                                        oo  «—
                                        co  M
&   «-

n   ii
                O  I'-


                ll    II
a:
in
cc   cc   en
co   vi   oo
                                            cc
                                            co
                        (g)
(O



LL
in — o
o


CO
O
Z
O
U rH
LU
*f LU
LU §
5 3
p ,"- ,-
-uj-g U-
Z "»
LU
O
CO
(O LU
tr


0.
to
LU
ffLU
U. H2
c"- <"J
r- — g «9
^ 01 CO
i- a. LU
5* oc
UJ ^n
LU —
II
X ^
LL. O LU

o §I|
8 
OQ
EFFECT
SO2 REM
Z
O














(£
5
INDARY ,
IY
°%

Log
£iu
Z i_3
Q:H
DO
lECTED INTO B
E SUPERIOR
iB
10 ~<
si
8"














1VAOIAI3U :OSlN30U3d

                       926

-------
The results of tests conducted by The Superior Oil Co-^ '  in conection with
Nahcolite injection in a flue gas duct are presented in Figure 1.   It is
evident from this figure that as residence time, temperature and stoichio-
metric ratio are increased, SCK removal efficiency also increases.  The effect
of temperature is more pronounced between the temperature range 1400°F to
2500°F.  Residence time above 1 second does not result in an appreciable
increase in S02 removal efficiency.  It should be noted that these data were
obtained on flue gas from an oil-fired furnace with synthetic S02  gas
injection and hence might not be applicable to flue gas from a coal-fired
boiler.

Figure 2 shows the comparison of various sodium alkalis tested by  Superior.^ '
Results clearly show the superior reactivity of biocarbonate containing
compounds compared to carbonates.  The higher reactivity of. bicarbonate may be
due to the porous structure of the material caused by the decomposition
reaction:
             2 NaHC03->Na2C03 + H20 + C02
The formation of C02 and H20 leaves a porous Na2C03 which allows easier
passage of S02 resulting in an improved rate of diffusion of S02-   The final
reaction between Na-jCO-, and S02 is assumed to take place as follows:
             Na2C03 + S02 + 1/2 02->Na2S04 + C02
It is evident from the above reactions that the kinetics involved  in the dry
scrubbing process are much more complex.

Recently, tests were conducted by Wheelabrator-Frye Inc. ^   ' at Stanton
Station of Basin Electric to investigate the use of fabric filter  for fly ash
and S02 removal by nahcolite injection.  Unfortunately, no information
concerning the results from these proprietary tests is available at present.

In summary, the presently available data on the dry sorbents are encouraging,
but there is a need to obtain more data to better characterize and optimize
the removal process in terms of the pertinent parameters.
                                      927

-------
O
o
                                                                                              rvi
                                                                             cv-
                                                                             .3
                                                                             o
                                                                             u.
o
O)
o
00
S
                                                                                                   2  <
                                                                                                   o  >
                                                                                                   o  o

                                                                                                   (/)  UJ


                                                                                                   *  %
                                                                                                   C5  UJ   K

                                                                                                   Z  g   P

                                                                                                       UJ
                                                                                                   U
                                                                                                   u
                                                                                               Q.

                                                                                               z.
                                                                                                   z  o

                                                                                                   li.
                                                                                                   o

                                                                                                   u
                                                                                                   u
                                                                                                   u.
                                                                                                   u.
                                                                                                   LJ
                                                                           w

                                                                           D
                                                                           w

                                                                           W
                                                                           a
                                        :OSlN30d3d

                                                 928

-------
EPRI'S ACTIVITIES

Before going into the details of EPRI's plan in the dry scrubbing area, it is
necessary to examine various methods proposed to remove SC^ with solid
alkaline materials.

These methods are:

•    Addition to  fuel
•    Injecting the agent into the furnace separate from the fuel
•    Contacting the gas containing SC^ with a fixed or fluidized bed of
     sorbent
•    Injecting the agent in the flue gas duct followed by collection in an
     electrostatic precipitator or baghouse

One of the serious drawbacks of adding the agent to fuel or combustion zone
injection is that it could interfere with the operation of the boiler.
Problems associated with high pressure drop, mechanical complexity, cost and
ash handling are yet to be solved before the fixed and fluidized bed processes
could become commercial.  Accordingly, it seems that the only practical dry
process application in the near future is through injection of sorbent in the
flue gas duct followed by collection in an electrostatic precipitator or
fabric filter.   However, before, this technology can be applied to a
commercial installation, a systematic study quantifying the effects of
controllable parameters and examining application related problems, is
necessary.

For this reason,  EPRI's current efforts are focused on a bench-scale research
designed to quantify the effect of various significant parameters on SC>2
removal efficiency and sorbent utilization.   The bench-scale study will be
conducted at a  scale of 725 SCFM under well characterized and controlled
conditions simulating a utilty boiler application.   The program (see Table 2)
is designed to  obtain data pertaining to the effect of residence time,
temperature, particle size, stoichiometric ratio,  sorbent type on SC^ removal
                                     929

-------
efficiency and sorbent utilization.  Electrical resistivity of the spent
material will be determined to investigate the feasibility of using dry
scrubbing process to enhance particulate collection in electrostatic
precipitators.  The effect of sorbent injection on NO  and particulate
                                                     X
emissions will also be evaluated.  The data relating to the degree of reaction
occurring in suspension versus on the filter cake of the bags or precipitator
plates will also be gathered.  Assuming successful completion of the bench-
scale research, future work will be aimed at pilot and full-scale
demonstrations.
                                   Table  2
                  Test Program  for EPRI's Bench-Scale Study

   No.         Variable                         Range
    1.     Duct temperature              300, 450, 600, 800,  and 1100°F
    2.     Stoichiometric ratio          0, 1, 2
    3.     Residence time in the duct    0, 1, 2, and 4 sec.
    4.     Type of sorbent               Nahcolite*, NaHCo^*, Na2C03, trona*
    5.     Sorbent particle size         -10 mesh, -65 mesh,  -100 mesh and -200 mesh

 Including calcined form.
WASTE DISPOSAL/REGENERATION

No  study on dry SQj removal would be complete without evaluating the disposal
aspects of spent sodium sorbents.  A number of investigators^ *' have
proposed several chemical and mechanical fixation methods for disposal of
spe t sorbent based on theoretical considerations.  To present a clear picture
to  utility industries, a detailed economic analysis of various options is
necessary.  EPRI's future plans will be concentrated in examining and
screening these options for utility application.

Another alternative to waste disposal is the regeneration^ '' of feed
sorbent material from spent sodium compounds.  This would obviate the problem
in  connection with availability and supply of various sodium containing
compounds.  To present an overall picture, the regeneration process will have
to  be integrated with dry scrubbing.  A description of sodium regeneration
processes as applied to dry scrubbing is presented by Dalton.^  '
                                       930

-------
SUMMARY

Dry scrubbing appears to be an attractive alternative in some cases, but
further research is needed before the technology can be considered for
commercial application.  Sodium based sorbents have been identified as
superior reactants in dry state compared to sorbents containing calcium and
magnesium.  The bench-scale study sponsored by EPRI will provide answers to
questions related to the effect of controllable parameters involved in the dry
scrubbing processes.  It is anticipated that follow-on efforts will include
construction and operation of a pilot plant, and subsequently, a full-scale
plant.  The problem of disposal of waste sodium sorbents must be solved before
the technology can be further considered for commercial application.
                                     931

-------
REFERENCES

1.   Dulin, J. M. and Rosar, E. C., Environmental Science and Technology,
     Vol. 9, p. 627, July 1975.

2.   Genco, J. M., Rosenberg, H. S., Anastas, M. Y.,  Rosar, E. C. and Dulin,
     J. M., Journal of Air Pollution Control Association,  Vol.  25,  No.  12,
     p. 1244, December 1975.

3.   EPRI report No. FP-207, Palo Alto, California, October 1976.

4.   Levelspiel, 0., Chemical Reaction Eng., Wiley Eastern Pvt- Ltd., New
     Delhi, 1969.

5.   Han Liu, et al., "Evaluation of Fabric Filter as Chemical Contactor for
     Control of Sulfur Dioxide from Flue Gas," Air Preheater Co.   NTIS report
     No. PB 194 196, Durham, North Carolina, 1969.

6.   Genco, J. M. and Rosenberg, H. S., Journal of Air Pollution Control
     Association, Vol. 26, No. 10, p. 989, October 1976.

7.   Doyle, D. J.,  Electrical World, February 15, 1977.

8.   Dulin, J. M. and Rosar, E. C., paper presented at 104th Annual AIME
     meeting, New York, New York, February 1975.

9.   Genco, J. M. and Rosenberg, H. S., paper presented at AICHE meeting,
     Chicago, Illinois, August 1976.

10.  Cook,  W. W., and Maitland, J. A., U.S. Patent No. 3, 823, 676, July 16,
     1974.

11.  Veazie, et al., "Feasibility of Fabric Filter as a Gas-Solid Contactor to
     Control Gaseous Pollutants," NTIS report No. PB 195 884, August 1970.
                                     932

-------
12.   Philips, T.,  Soot,  P.,  and Niman,  S.,  report  on  "Evaluation of Dry
     Alkalis for Sulfur  Dioxide Removal  Both Retrofit and New Construction,"
     Pacific Power and Light Co.,  April  1977.

13.   Slack, A. V., "Sulfur Dioxide Removal  from Waste Gases," Noyes Data
     Corp., Park Ridge,  New  Jersey,  p. 20,  1971.

14.   Environmental Science and Technology,  p.  856,  Vol.  11,  No. 9, September
     1977.

15.   Dalton, S. M., "Sub-system Combination for Recovery Processes—Addressing
     the Problems," paper presented at Symposium on FGD, Hollywood, Florida,
     November 11,  1977.
                                      933

-------
UNPRESENTED PAPERS
        935

-------
           OPERATING EXPERIENCES WITH KAWASAKI
MAGNESIUM-GYPSUM  FLUE  GAS  DESULFURIZATION  PROCESS
                             Hajimu Tsugeno,
                           Takashi Mashita, and
                               Tadaharu Itoh
                        Kawasaki Heavy Industries, Ltd.
                      Chemical  Plant Engineering Division
                               Akashi, Japan
       ABSTRACT

           Construction of the first and the second commercial plants using a
       Magnesium  Gypsum Flue  Gas  Desulfurization  Process  recently
       developed by KAWASAKI HEAVY INDUSTRIES, LTD. (KHI) were com-
       pleted and the trial runs were finished in the beginning of 1976. An
       outline of the processing equipment, and performance data in the trial
       runs of the flue  gas desulfurization plant for Japan Exlan Co., one of
       these commercial plants, are summarized in this paper.
           Construction of the Japan Exlan plant was completed in the end of
       1 975, and it has remained in operation, after a 2-month trial run, since
       March 1976. In this plant, lime is used as the absorbent agent (with
       addition  of small amounts of magnesium  hydroxide) and gypsum is
       recovered as a byproduct from the plant. A mixed slurry of calcium and
       magnesium solids  is used in  the absorber,  in which sulfur dioxide
       removal  efficiency is more than 93 percent. Stable continuous opera-
       tion, with no trouble of scaling, has been maintained since the trial run.
           We have found a possibility to refine this process to a high degree,
       improving on the present Magnesium-Gypsum Process, based on  our
       experiences obtained in the runs at the above commercial plants and on
       the results of several investigations of our own related to the process.
           The  results  of the above  investigations  and  an  outline of  the
       improved process are also summarized in this paper.
                                     936

-------
                         OPERATING EXPERIENCES WITH
       KAWASAKI  MAGNESIUM-GYPSUM FLUE GAS  DESULFURIZATION PROCESS
                                  INTRODUCTION
 KHI  constructed two commercial plants of the newly developed Magnesium Gypsum
 Flue Gas Desulfurization Process at the end of 1975,  which  have been in stable
 operation after the trial run since the beginning of  1976.

 The  KHI Magnesium-Gypsum Flue Gas Desulfurization Process was developed and
 brought into practical use based on our own experimental study on a pilot
 plant in 1972 and 1973 with a magnesium absorbent in  which  we at KHI have a
 rich experience.

 We obtained an order of the first commercial plant by this  process from Unitika
 Co., Okazaki Works, and that of the second one from Japan Exlan Co., Saidaiji
 Works, 1974, and both of these plants were constructed in 1975.

 This process is suitable for the flue gas of coal firing as well as that of oil
 firing, and is expected to be a main process of KHI's flue  gas desulfurization
 plant in the future.

 Lime is used in the plant for Japan Exlan Co., as in  the case of our pilot
 plant, meanwhile, limestone is used as a main absorbent agent with a little
 addition of lime in the plant for Unitika Co.    Gypsum is  recovered as a by-
 product in the both plants.    Results of the trial run and successive com-
 mercial run of the flue gas desulfurization plant for Japan Exlan Co. and our
 several investigations related to this process, improvements of the process
.according to the above results and an outline of a proposal of the flue gas
 desulfurization plant applying the improved process are summarized in this
 paper.
                                     937

-------
         J,  RESULTS  OF RUN OF THE FLUE GAS  DESULFURI-
             ZATION PUNT  FOR JAPAN EXLAN  CO,
The flue gas desulfurization  process adopted to Japan Exlan plant is a standard
Magensium-Gypsum Process,  in  which  lime is used as an absorbent and gypsum is re-
covered as a by-product.

This plant was ordered  in  August  1974, constructed in December 1975, and has
been in commercial run  since  March  1976  (after  the two-month trial run begin-
ning in January 1976).

A.  OUTLINE OF EQUIPMENT

1.  Design Specification

    The absorption section  is  divided into two units for the convenience of opera-
    tion, but the other sections  such as oxidation, gypsum separation, raw mate-
    rial feeding and magnesium  hydroxide regeneration are each common units which
    serve the two absorption  units.
         Location             : Japan Exlan Co., Saidaiji Works,  Japan
         Use
:  Desulfurization of the flue gas from the boilers
  for  power  generation and for supply of process
  steam
         Fuel of boiler
  8,2 — 3 wt%, Bunker C Oil
         Gas flow rate
  No.l  Absorber
  No.2  Absorber
:  160,000  NrnVh
:  140,000  Nm3/h
         SOx concentration
         (dry)
  Absorber  Inlet
  Absorber  Outlet
  1,412  ppm
     90  ppm
         Particulates con-
         centration (dry)
  Absorber  Inlet
  Absorber  Outlet
  less  than  0.2  g/Ntn3
  less  than  0.1  g/Nm3
  (after  reheating)
                                    938

-------
     Absorbent agent      :  Ca(OH)2  :  953 kg/h
                            Mg(OH)2  :    6 kg/h
     By-product           : Gypsum :  2,214 kg/h
                            (not including the moisture)
     Exhanust gas heating :  Afterburner
     system
The general view and the general arrangement of this plant are shown in

Figures 1 and 2, respectively.


The specifications of the main equipment are shown in Table 1.
                               939

-------
Figure 1.   General View of the FGD Plant for
            Japan Exlan Co.
                         940

-------
                                                           o
                                                          u
                                                           X
                                                          w

                                                           c
                                                           (C
                                                           cu
                                                           n)
                                                           M
                                                           O
                                                          Q
                                                          0
                                                           (1)
                                                          JC
                                                          4J
                                                           O

                                                           4-'
                                                           U
                                                           cu
                                                           e
                                                           OJ
                                                          OJ


                                                           0)
                                                           5-1

                                                           3
941

-------
W

P-i
Cf


z
X
 cfl
H












^
4-J
•H
O
cfl
CX
cfl
U



















QJ
CX
^i
H




















B
01
4-1
M
























X X
"" — """-.
rO (T>
B B
Z Z

0 0
0 O
0 0
*. r>
o o
v£3 - £>

1 1

•H -H
4-J 4J
rH rH
3 3
S S

















S-i S-J
QJ QJ
X Xt
S-i S-i
0 0
cn co
x x
^ ^C

rH CN
• •
o o
z z
cr
<£
B
E

0
m
01

X

X
-" —
CO
B


0
o
0
rs
o
\^)
rH
C
0
•H
4-1
O
3
cn

0)
rH
O
3
0


X
4-1
•H
£g

•V
o
X
S-l
3
H









C
cfl
1JH

4-1
14_|
cfl
S_i
-o

"-TJ
0)
U
S-i
o


rH
•
o
z
cr
^
ro 4—1
B 3
o
o
CO CM
















rH
CO

3
MH
•H
S_i
4-1
C
0)
0
cx
B
3
CX
C
O
•H
4J
Cfl
rH
3
O
S-i
•H
0

4-1
C
0)
X
S-i
0
cn
x


rH
•
O
z
c
o
•H
4J
CO to
S_i
0)
X CX
0
B
c
m -H
m
CO
X
MH
E 0

ro ±J
B 3
O
0
CO CXI
m ^















, — i
cfl
60
3
14-1
•H
S_i
4-1
C
0)
u
a
B
3
CX
c
0
•H
4-J
cfl
rH
3
O
S_i
•H
0

4J
C
0)
X
S-i
o
cn
x
<3^

CN
.
O

























S-i
0)
J5
0
4-J

60
C
•H
rH
X
X
3
X

rH
Cfl
CJ
•H S-l
(-1 -H
-a cfl
c
•H X
rH 4-1
>. -H
O S

















S-i
0)
N
•H
TJ
•H
X
o

rH CXI
• •
O 0
z z


























S-i
o
cn
cn
CD

cx
B
O
o

O
X
S-i
3
4-1

0)
60
CO
4_J
cn
I
CN




QJ
N
•H
TJ
•H
X
O

S-
o
14-1

^_l
O
cn
M
0)
S_i
cx
e
0
CJ

S-i
•H



















m
B

CO
•
CO






rH
CO
CJ
•H
4-J
S-i
QJ
J>

rH
Cfl
CJ
•H
S-i
TJ
C
•H
rH
^1














V
C
cfl
4-1

C
o
•H
4J
Cfl
S-i
QJ
C
0)
60
QJ




















fO
B

00







rH
CO
CJ
•H
4-1
S-i
0)
[>

rH
CO
CJ
•H
S-l
"O
c
•H
rH
^













Sf?
C
cfl
4-1

K*~»
S-i
S-l
3
rH
cn

•H
t — 1
CO

rH



















CO
B

0
oo
rH






rH
CO
CJ
•H
4-J
S-
0)
^>

rH
CO
O
•H
S-i
-a
c
•H
rH
^>













^
C
cfl
4-J

!H
O
3
cr
•H
rH

S-i
QJ
X
4-J
O




















ro
S

CXI
rH

















MH
O
O
S-i

QJ
C
O
u












V1
c
cfl
4-1

•a
•H
o
cfl

CJ
•H
J_4
3
M-l
, — 1
3

                                           942

-------
2.  Description of the Process

    The process flow diagram is shown in Figure 3.
    The process consists of the following three sections;

    a.  Raw material feeding and magnesium hydroxide regeneration section.

    b.  Absorption section.

    c.  Oxidation and gypsum separation section.


    a.  Raw Material Feeding and Magnesium Hydroxide Regeneration Section;

        Raw materials of lime and magnesium hydroxide are supplied from the
        hoppers to a mixing tank, in which they are mixed with water to form a
        slurry and are sent into a magnesium hydroxide regeneration tank.

        In  the regeneration, tank, mother liquor (solution of MgSOi,) which
        is circulating in the system and has no ability of absorbing the sul-
        furdioxide gas is regenerated as an absorbent, changing into Mg(OH)2
        by the reaction (1).

            MgSOi, + Ca(OH)2  -»•  Mg(OH)2  + CaSCK ......... (1)

        The slurry of raw materials which contains lime and magnesium hydroxide
        is supplied to the absorber, after being stored in a alkali slurry
        tank.

    b.  Absorption Section: Flue gas from the boilers is sent to the absorber
        through a fan, is quenched by cooling water in the absorber, and then
        is contacted efficiently with a  slurry of magnesium and calcium solids.
         The  SOa  is absorbed and removed according to the reactions (2) and (3)

            MgS03  + S02  + H20  *  Mg(HS03)2  .............. (2)

            CaS03  + S02  + H20  •*  Ca(HS03)2  .............. (3)

                                    943

-------
                                                          o
                                                         u
                                                          (0
                                                         r-H
                                                          X
                                                         w

                                                          c
                                                          (0
                                                          Du
                                                          (d
                                                          -P
                                                          C
                                                          m
                                                          Q
                                                          O
                                                          En

                                                          (U
                                                          x;
                                                          -P
                                                          E
                                                          (U
                                                          (tJ
                                                          •H
                                                          ro


                                                          
-------
    After  the elimination  of  mist  and being reheated to prevent a steam
    plume,  the  clean gas  is emitted to atmosphere.

    Meanwhile,  part  of absorbent slurry  is  pumped to the oxidation and
    gypsum  separation section ;     the remainder is  neutralized by the
    reactions (4) — (7).

        Mg(HS03)2 +  Mg(OH)2   -*•  2MgS03 + 2H20 ...............  (4)

        Mg(HS03>2 -I-  Ca(OH)2   •* MgSOa  + CaSO, +  2H20 ........  (5)

        Ca(HS03)2 +  Ca(OH)2   -*• 2CaS03 + 2H20  ...............  (6)
        Ca(HS03)2 + Mg(OH)2  •*•   CaS03  + MgS03 + 2H20 ..
    Then,  the neutralized slurry is circulated to the absorber to be
    re-used.
c.  Oxidation and Gypsum Separation Section: The absorbent  liquor  which has
    been consumed to absorb S02 , is supplied to an oxidizer from the  absorp
    tion section and oxidized by air according to the  reactions  (8) ^ (11).
    Consequently it becomes possible to separate magnesium  from calcium,
    as the solid phase crr.sists of only CaSOi» and liquid phase consits  of
    only MgSOi».

        CaSOs +  -f-°2 + 2R2°   *  CaSOi,. 2H20 ............... (8)

        Ca(HS03)2 + -y-°2 + H20  •*  CaSO^. 2H20 + S02  ........ (9)

        MgS03 +   -02 -»• MgSOv ............................... (10)
        Mg(HS03)2 + -y-02 * MgSOn + H20 + S02 ............... (11)

    Alkali remaining in the slurry are neutralized through the reactions (12)
    and (13).

        Ca(OH)2 + S02 + -4-02+ H20  •*  CaSOi, . 2H20  ......... (12)

        CaC03  + S02 + --Oz + 2H20  •*•  CaSOw . 2H20 + C02 ...(13)
                               945
-------
         At first we planned to supply HaSCK in order to neutralize the excess
         of Ca(OH)2, but Ha SO i, is not used at present as it proved to be un-
         necessary.

         Gypsum slurry from the oxidizer is dewatered by a thickener to a favour-
         able concentration for a centrifuge (about 15%), and,  separated into
         gypsum and  mother liquor through a centrifuge.

         The gypsum  is taken out from the system as a by-product,  and the rest
         liquor is pumped to the Mg(OH)2 regeneration section and  is re-cycled
         as absorbent.

         The quantity of make-up of magnesium hydroxide  is equal to the loss of
         magnesium(as MgSOu)contained in the  moisture of gypsum,  and it is
         only a very small quantity.
B.  RESULTS OF TRIAL RUN

1.  Progress and Outlook of Trial Run

    Progress of the trial run is shown in Table 2.
    The trial run was started in December 1975 and was completed in the middle of
    March 1976 with no major trouble.
    We delivered the plant to the client after the trial run,  and the plant pass-
    ed an official test with no problem at the end of March 1976.

    After an adjustment run (a test run in which an absorbent  liquor is adjusted
    to be a designed condition), we had a minimum load run (15 — 20%) in the
    No.l absorber and a maximum load run (100%) in the No.2 absorber, respective-
    ly, in which we could confirm the high efficiency with no  problem.

    In addition to the above, we confirmed the effects of such factors as pH of
    absorbent liquor, liquid-to-gas ratio, etc., on the absorption efficiency.
                                     946
-------
                          to
                          S
                                                                                                        o
                                                                                                        CM
                                                                                                                                           r
                    vO
                                                                                     CO

                                                                                     in
                          CD
                          tn
                                                                                                                o
                                                                                                                iH

                                                                                                                &
                                                                                                                          m

                                                                                                                          CM
                                                                                                                0
                                                                                                        OT   0
                                                                                                        CM
                                                                                                                &
                                                                                                                ON
00
H
H
Pi
Pi
H
CO
00
H
Pi
Pi
PH
CM


 CU
                          C
                          to
                          <->
                       1
m
       O
       su
      o
                                                                                             o


                                                                                             
                                      en
                                      CU
                                      H
                                            0)
                                            4-1
                                            CO
                        I
                                          C
                                          3
 C
 cu
 s

 CD
 3

T)
                 C
                 3
                 M

                -O
                                                                              §
                                                                             3
                                                                                              C
                                                                                             en
                                                                                             cu
                                                                                             4J
 CO
 t*
 0)


 cu
CO
                                                                                       £
                                                                                       O
                                                                                       cfl
                                                                                       CO
                                                                                       O
 CD
 CO
O
                                                    O
                                                    z
                                                                                                           W)
                                                                                                           •H
                                                                                                           0)
                                                                                                           3
                                                                                                           0)
                                                                                                           3
                                                                                                                               C
                                                                                                                              •H
                                                                                                           3
                                                                                                           •4-1
                                                                  3
                                                                 OO
                                                                                CO
                                                                                                                                             cu
4-1
 O

 01
 60
 cfl
 &
 CX
 O

CO
                                                                   947
-------
2.  Results of the Performance Tests
    a.  Results of Maximum Load Test;  The results of the maximum load test was as
        follows, in which we obtained very good results surpassing the design
        value.
        Flue gas conditions (at the inlet of the No.2 absorber);

             Gas flow rate     :  146,000 Nm3/h
             Gas temperature   :  170°C
             Sulfur in oil     :  2.5 - 3.0 wtZ

        Operating conditions:
             Mole ratio cf alkali agents supplied
             to the absorber
             Flow rate of circulating absorbent liquor
             pH of absorbent liquor
             pH in the  oxidizer (1st stage)
             pH in the  oxidizer (2nd stage)
             Moisture of by-product gypsum
  Ca2+/Mg2+ .
  1,260 m3/h
  5.3
  3.5
  2.5
  9 wtZ
3.5 -4
        Result of measurement of efficiency:
             SOx concentration at the inlet of the
             absorber (dry)
             SOx concentration at the outlet of the
             absorber (dry)
             Desulfurization efficiency
             Guarantee value:
                 SOx concentration at the inlet
                 SOx concentration at the outlet
                 Desulfurization efficiency
             Particulate concentration at the inlet
             of the absorber (dry)
             Particulate concentration after re-
             heating (dry):
             Guarantee value:
                 Particulate concentraion at
                 the inlet
                                    948
:  1,570 ppm

:     79 ppm
:  95.OZ

:  1,412 ppm
:    100 ppm
:  93Z

:  0.190 g/Nm3

:  0.051 g/Nm3


:  0.2 g/Nm3
-------
              Particulate concentration after
              re-heating                            :  0.1 g/Nm3

    We found no scaling in an inspection to the inside of absorber during a
    stoppage of the plant, and so we could confirm that the scaling trouble
    could be mitigated by the presence of MgSO<,.

b.  Results of Several Tests:
    Relation between pH of Absorbent Liquor and Desulfurization  Efficiency;
    Relation between pH of absorbent liquor and desulfurization  efficiency is
    shown in Figure 4 with a parameter of liquid-to-gas ratio.
    Above data are for a 75% of gas load factor,  and the desulfurization ef-
    ficiency can be expected to be higher in a 100% load, because the contact
    of gas and liquid is expected to be more efficient.

    Relation between Liquid-to-Gas Ratio and Desulfurization Efficiency;
    Relation between the liquid-to-gas ratio and   the  desulfurization effi-
    ciency, in the condition that pH of the absorbent  liquor is  5.5 and the
    gas load factor is 100%, is shown in Figure 5.

    It is recognized that a high efficiency of desulfurization can be obtain-
    ed with a small value of liquid-to-gas ratio  in the Magnesium-Gypsum
    Process, compared with  a  lime scrubbing process in which a  liquid-to-gas
    ratio of 10 — 20£/Nm3 is necessary in general.

    The velocity of the reaction of desulfurization depends largely on a
    solubility of absorbent agent,  and the higher the  solubility of SOs  in
    the absorbent liquor is, the higher absorption  efficiency can be obtained.

    Relation between the solubility of Sol~ and the MgSO., concentration in
    the  absorbent  slurry  is  shown  in  Figure  6.    In  this  process,  the  solu-
    bility  of  SOl" is  high,  because
    at   a   concentration  of  50  g/i.
bility of SOl" is high, because MgSOi, is contained in the absorbent liquor
    This  is  exactly  the  reason why  there  exists an  apparent  difference  in
    desulfurization  efficiency between Magnesium-Gypsum  Process, which  contains
    MgSOi*  in the absorbent,  and  lime  scrubbing process which does not apply
-------
dP
                            gas  load  factor  :  75%
                        5.5                 6.0

                           Absorbent pH
  Figure 4.   Relation  between  Desulfurization Efficiency and
              Absorbent pH
                               950
-------
  100
e-
c
0)
1-1
u
W


o
•H
•P
id
N
•H
M
(0
Q>
Q
   90
80
   70
                                          absorbent pH :  5.5,


                                      gas load factor :  100%
                    Liquid-to-Gas Ratio  L/G U/Nm3)
      Figure 5.   Relation between Desulfurization Efficiency and

                  Liquid-to-Gas  Ratio
                                  951
-------
       10
    O
    CO

   I
   NO
    O
    CO
50
                                                 100
                 MgSO*. Concentration
Figure 6.   Relation  between MgSO. Concentration  and


            S0?~  Solubility
                              952
-------
Relation between the Gas Load Factor and Desulfurization Efficiency:
Relation between the gas load factor and desulfurization efficiency with
a parameter of liquid-to-gas ratio is shown in Figure 7, where pH of the
absorbent liquor is 5.5.

Desulfurization efficiency tends downward as the gas load factor becomes
lower if all other operating conditions are held constant.
However, in practice, the desulfurization efficiency becomes higher,
since the flow rate of circulated absorbent liquor is kept constant
irrespective of the gas load factor, resulting in the liquid-gas ratio
becoming relatively larger as the gas load factor becomes smaller.
                             953
-------
    50
   70-         80
Gas Load Factor  (%)
90
100
Figure 7.   Relation between Desulfurization Efficiency and
            Gas Load Factor
                             954
-------
C.  PROBLEMS AND SOLUTIONS IN THE TRIAL RUN

    The trial run was successful and we had not so many major problems.
    Some problems we experienced in the trial run were as follows:

1.  Scaling Trouble of the Mist Eliminator

    When we inspected the inside of the equipment after 1,000 hours of operation,
    no scaling was found in the absorbers  proper,  in the piping around the absorb
    er, in the raw material slurry section and in the mother liquor section,
    however, some scale mixed with carbon  dust was found in the No. 2 mist elimi-
    nator and in the No. 2 drain tank including the attached piping.

    We made an investigation of the factor of the scaling and the  solution  to
    avoid the scaling,  which may disturb the continuous operation.
    The mist eliminator was washed with mother liquor in order to prevent the
    adherence of mist.

    According to our investigation,  we found the factor of scaling as follows;

    a.  When the gypsum contained in a mother liquor as a saturated solution is
        mixed with the mist of absorbent liquor,  solidification of gypsum occurs
        because a solubility of gypsum becomes lower.

    b.  The gypsum is generated by the following reaction between the sulfuric
        acid contained in mother liquor for washing and CaSOa contained in the
        mist of absorbent liquor.
            CaS03  + HiSCK   -»•  CaSO.,  + H20 + S02

        Considering the above phenomena,  we tried to change a condition of opera-
        tion,  to make some modification of the equipment,  etc.,  but we could not
        get a  final solution.

        We finally   gave  up the washing with mother liquor in the No. 2 absorption
        section, and an intermittent washing method  with industrial water was
        adopted.
                                    955
-------
        Studying the optium quantity of washing water and the optimum inter-
        val of washing,  we found it possible to wash the mist eliminator within
        the.  limits of the water balance of the process, which keeps  no discharge
        of effluent liquor.

        Good operation without any trouble of scaling has been maintained since
        then.   On the other hand,  in the No.l absorption section,  washing with
        mother liquor has been kept in operation without any trouble  of scaling,
        because the composition of  absorbent liquor in the No.l absorption section
        is different from that of the No.2 absorption section due to  a low gas
        load factor.
2.  Clogging in a Chute of Pulverized Material Feeding Equipment

    The chute  provided between a screw conveyor and a bucket elevator and those
    provided between a bucket elevator and a service hopper were often clogged.
    However, the absorption section which is an essential part of the whole equip-
    ment, has never been stopped in operation, partly because the enough quantity
    of raw material slurry for several hours operation was stored in the alkali
    slurry tank.

    It proved that the clogging was caused by inadequacy of the shape of the chute
    such as too large curvature, too small area of its cross section, etc.

    Trouble of the clogging was solved by the following counter-plan.

    a.  Enlargement of the area of the cross section of the chute.

    b..  Decreasing the number and inclination of the curvatures.

    c.  Lining the chute with Teflon resin.


3.  Clogging of a Filter-Cloth in a Centrifuge

    Drop of the separation efficiency of centrifuges occured in two months from
    the beginning of the trial run.   This process applies non effluent liquor
    system, therefore, the only way for the carbon dust to be taken out of the
    process is the mixing with the gypsum.
                                       956
-------
The drop of the separation efficiency of a centrifuges is caused by clogging
of a filter-cloth with the carbon dusts.
In a normal operation, by-product gypsum has a liquid content of about 102
and can be handled easily, but the separation efficiency becomes lower gradu-
ally as the operation of the centrifuge repeats its batchwise operation,
resulting in the increase of liquid content of the gypsum far over 10Z.

In order to this problem, we tried to use several kinds of filter-cloths and
compared their durability.

Fortunately we  did  find a filter-cloth which was superior in durability and
could be re-used after a light cleaning, outside the centrifuge.

We  solved      the problem of clogging by using the new type of filter-cloth
as mentioned above.
                                957
-------
D.  RESULTS OF COMMERCIAL RUN

    This plant has been maintained a commercial run in a good condition since
    March in 1976.
    There were two times of stoppage for an annual inspection of the boilers, and
    the operability factor (hours the FGD plant was operated/boiler operating
    hours) of the FGD.plant is 98 percent.

    Other conditions of the operation are as follows:

         Gas load factor (average)                    : No.l   20%
                                                        No.2   80Z

         SOx concentration, Absorber inlet            : 1,400 ppm (dry)
         SOx concentration, Absorber outlet           :   100 ppm (dry)

         pH-value of absorbent liquor                 : 5 — 6

         The number of operators                      : 1 person/1 shift
                                                        3 shifts a day

    This process is a closed cycle with no effluent liquor from the system.
    Only a leakage of sealing water from the stuffing box of the pumps or a
    drain water from the gas ducts are discharged outside without recovery,
    which are almost the same as fresh water in quality.

    Cl  concentration in the system is about one thousand ppm.

    By-product gypsum is used for cement additives.

    As an electrostatic precipitator is not equipped in this plant, most of the
    carbon dusts contained in flue gas are mixed into the by-product gypsum.
    Therefore, there are some problems in the quality of gypsum when it is applied
    to the other uses such as wall board.

    After long run, some worn areas have been  observed  in piping made  of resin
    or in a  control valve  for slurry.
                                    958
-------
E.  ECONOMICS

    The operating cost indicated in Table 3 is based on the tentative design of
    Magnesium-Gypsum Process for 500 MW coal fired power unit.
    The capital investment is calculated on the basis of the domestic cost of
    Japan in 1976.
    Basis:
           Coal consumption,  1,372,000 tons/yr - 9,340 Btu/kwh.
           Power unit on-stream time,  7,000 hr/yr.
           Capital investment,  $22,860,000 (without stack gas reheat).
           Disposed gypsum in solid 160,000 tons/yr.
           Cost of utility supplied from power plant  is for the  case of full lo, d
           operation.
                                    959
-------















2
D
53
W

I
2 ^
| S
•s w
oJ w M
O w
r*. U *•*
CO 8 1
£ « M
8 * ^
i S *
S 0. 41
H >. M

2 ^
U rt
O °
^
j S
I §
z £.

u
n
«j
u
<


ro
0)
1— 1
H





x
S *
\*~4
• £
u
d
o
•of
ll
« o
« u
o ^
0
O
-H
a
•J
rt
** O
o u
H


««•
n
O
u

I



X
*l
i
a
3
0-
_
fl)
3
1















fO ^J* ON ~« ON NO ^
*•• rO NO O O ON «-l
^i o o o co ^H o
do d do do




* IN- in ro t- o ON
m co t- o co o rj
o o -< or- in o

— o o o o o o





o o o o o o o
O O O NO O O
m oo o NO o oo o
in" oo" d" *f d" in o
^* "* "f 00 00 ^
^* -H IM O NO

NO ro ro ro oo
00 ro f~ O O
m o ON -HO
d •- i d do




^J1 in ro r\j ON
ON ro 00 NO -H
•* NO •* fN) 0

— ' fNJ N O O





O 0 O •* 0
N rj •*> no
•# r~ •< oo •*
d •* NO co NO
in ••< o m N
O NO f ro
(M ro (O

00 >-t C- ON
O -H (NJ ro
-• (NJ d rvi




•* O> O ON
NO ON O ON
r- ro r- o

N m o NO





NO oo o oo
ro O O O
(Ml NO O* NO*
ON O NO NO
ro |-T oo

Je
a H a •°_ 1
22 E a <
\ ^X "^N. ^ »O
o o 5 -M
m o \ o
^ o oo o o
(M (NJ (\J
I—I

1^
Jj
' ~* jd
II s 1 1

O O o •* O >•
o tr* o <*> *•< o i^
OO>0«M -g ""v^O
*" °" » 1 ! i s
fs. — X

— — - - a i
a a o. " * ^
§• J I * I 2 | -
n » 1;55»5!5
a - i«" S-9«So°0
•3 2 S ^ * 81? 5S 81-
T3 »Q 2 C ^4» C** 0 ,.3^H'***«-<'*
TT! ** ., ^5 Qv 7i« ® *^HO^§«
a* -a j? •« 3s ^^ yStt -T'-'S'S^S'"
-^'C X 2 «'C C n»rt 5? ^?<«rt£-i * -O
u[? ^ ^0 «!** U** **** *!,1 *>** *rt«_*
I | | m * s 1 s || || g | * | | j||| ||
w \ « I '5 -S I " ? ,^5 . '2
8 2 i & H : 3 3 ^ ° s- i i
^ 1 jj 1 * g 2 * s : 1 3 3
f% M |^ [i] 00 ^
w > > 2i 2 •£
S 2 S 9 > >
M V O ^ ' ?
S Q o 5 < 6
960
-------
   II,   INVESTIGATIONS  RELATED TO  DESULFURIZATION  PROCESS
Since practical applications of desulfurization plants  have  been put  into prac-
tice, many experiences of flue gas desulfurization  processes have  been  obtained
and the studies about  the flue gas desulfurization process  have been concent-
rated on some special problems.

Recently, the problems on the waste water  and  the quality of by-product gypsum
have been regarded as more important problems  than  those on  the desulfurization
efficiency which was an important problem at  first stage.
A.  TREATMENT OF THE EFFLUENT LIQUOR

1.  Quantity of Effluent Liquor

    It depends on a balance of water in  the  FGD  process and an allowable level of
    concentration of the harmful  impurities  whether a discharge of effluent liquor
    from the sytem is necessary or  not,  and  how  much the quantity of effluent
    liquor is, if necessary.

    The harmful impurities said above are  classified into the following two kinds:

    a.  The harmful substances for  the process,  which may effect on a desulfuriza-
        tion efficiency,  etc., when they reach a certain level of concentration...
                ~   ~        , etc.
    b.   The harmful  substances  for  the material, which may effect on a corrosion
        when they  reach a certain level of concentration . ......... Cl , F .

    In  general,  it is  seldom  that the discharge of effluent liquor is necessary in
    order to maintain  a balance of  water in the system, except for the case when
    a  pre-scrubber  is completely separated from an absorber in the process.

    It  is possible to  operate without discharging effluent liquor from a full load
    to  a 1/3 load  in the Magnesium-Gypsum Process, where a pre-scrubber  is not
    provided at  the  upstream  of an  absorber.
                                   961
-------
    Therefore,  the factor to decide a quantity of effluent liquor is a level of
    concentration of the abovementioned harmful substances,  especially that of
    Cl ,  which  flows much into the system and is accumulated to a high concentra-
    tion in the case of coal-fired power plant.

    The effect  of the accumulated Cl  is explained as follows:
2.   Effect of Cl  on Desulfurization Efficiency

    It is known that Cl  lowers the desulfurization efficiency in a lime scrubb-
    ing process.
    We have made a basic experiment to confirm an effect of Cl  on the desulfur-
    ization efficiency in a magnesium scrubbing process compared with that in a
    lime scrubbing process.

    The results are shown in Figure 8.
    The conditions of the experiment were as follows:
    a.  Method of test
    b.  Supplied S02 gas
    c.  Initial quantity of the absorbent liquor
        before SOa gas is supplied

    d.  Initial composition of the absorbent liquor
:  Batch reaction

:  Concentration :  1,050 ppm
  Flow rate     :  4.5£/min

:  330 g

Cl
MgS03.6H20
MgS04.7H20
CaSOi,.2E20
r -i en TT c\
L.clbU3 . ^ Ii2^
(wt.%)
(wt.%)
(wt.%)
(wt.%)
(wt.%)
Magnesium- Scrubbing
0
0.8
8.8
10.9
0
2
0.8
8.8
10.9
0
Lime-Scrubbing
0
0
0
10.9
0.5
2
0
0
10.9
0.5
                                    962
-------
lUU'

Oft


0 80-
c
OJ
•H
o
•H
 9A-
	 T^IUO




80-





60-
—
Cl 0%


	 Cl 2%

40-
'~r5:^— --^^
v ^s>\
\ N.
\ ^^
\ N
\
\


\
\
\
\
__
Cl 0^
\

	 Cl 2%


Magnesium-Scrubb- Lime-Scrubbing
ing







 +J
 q
 o
 to
 XI
 M-t
 o
       0  0.5  1.0  1.5   0   0.5  1.0  1.5
                      hr                hr
                   S02 supply time

      Relation between S02 supply time and

      desulfurization efficiency
2-
         Magnesium-Scrub
         ing
                    Lime-Scrubbing
       0  0.5  1.0  1.5   0  0.5  1.0  1.5
                      iir               hr

                   S02 supply  time

      Relation  between SO2  supply time and

      absorbent pH
>Figure  8
          Effect of  Cl  on Desulfurization Efficiency
                          963
-------
According to the results of the experiment, SOz absorption efficiency was
conspicuously lowered by the effect of Cl  in the absorbent liquor of a lime
scrubbing process.    On the other hand, in the absorbent liquor of a magnesium
scrubbing process,  the absorption efficiency was hardly lowered in spite of
the presence of Cl .

In addition, the fall of pH-value of the absorbent liquor owing to the pres-
ence of Cl  in a magnesium scrubbing process is much smaller than that in a
lime scrubbing process.

The fall of absorption efficiency and pH-value of absorbent liquor by the
effect of Cl  may be explained by the common ion effect, that is, Cl  sur-
presses the dissolution of SOa , which carries out the absorption of
The solubility of calcium compounds is very small and it is sensitive to an
effect of Cl  , so a small amount of Cl  has a remarkable effect on the de-
sulfurization efficiency.

Meanwhile, the solubility of magnesium compoundsis very large (for example,
the ratio of the solubility of MgSOs to that of CaSOs is over 100:1), so it
is insensitive to a dissolution surpressing effect by Cl ,  and Cl  has not a
practical effect on the desulfurization efficiency as is shown in the results
of the experiment said above.

Therefore, in case that the flue gas from coal-firing which contains a lot
of HC1 is treated, it is necessary to install a  pre-scruhhpr  at the upstream
of an absorber in order to absorb the HC1, or it is necessary to discharge a
large amount of effluent liquor to prevent a rise of Cl  concentration in the
absorbent liquor in a lime scrubbing process.

However, such countermeasures are not necessary in a magnesium scrubbing
process, so it is more advantageous in this respect.

In addition, F has the same effect as Cl~ on the desulfurization efficiency.
However the effect of F is not so large as that of Cl , since the quantity
of HF contained in a flue gas is not so much as that of HCl.

Concerning the effects of Cl~ on materials, problems of pitting corrosion and
stress corrosion cracking are important.
                                 964
-------
    Recently, a new metallic material has been developed and put into practical
    use, which is suitably applicable to the handling of the liquid with high
    concentration of Cl  in a FGD plant. -

    On planning of some FGD plants for coal fired boiler, it is practically studi-
    ed to keep Cl~ concentration as high as 10,000 ^ 20,000 ppm in the system in
    order to minimize the quantity of effluent liquor as far as possible.

    As for the materials of the processing equipment and parts which require
    corrosive-proofing in such FGD plant, non-metallic materials such as resin,
    flake-glass lining, rubber lining, may be used for the parts
    contacting the liquor in the system, and special metallic materials may be
    used only for the parts where non-metallic materials are not applicable, such
    as a mechanical seal and shaft sleeve of pumps.
3.  COD Value and 820$  in Effluent Liquor

    In relation to the harmful substance for the process,  it is necessary to take
    account of the trace substances generated in the process itself besides those
    such as Cl  flowing into the FGD process from the outside of the system.

    These trace substances may have a harmful effect on the process occasionally
    when they accumulate to some degree.

    The following phenomena have become clear recently.

    a.  As an example of the substances generated in the process,  SaOe   is gene-
        rated in every FGD process:   sodium-scrubbing process, magnesium-scrubb-
        ing process and calcium-scrubbing process.

    b.  SaOe  is generated in the case when SO3   is oxidized by the air in the
                      I j |
        presence of Fe   .

    c.  The quantity of $20$  to be generated depends both on the  quantity of SOT"
        oxidized and a concentration of Fe
                                     965
-------
SaOe  is inert for the reaction of desulfurization, and has seldom an effect
on the process unless it is accumulated to a high concentration.

In case of discharging the effluent liquor, high COD value by SaOe   comes
into question, because SzOe  is a substance which can not be oxidized easily
by an ordinary method.

The folloiwng methods shown in Table 4 have been developed for the treatment
of SaOe  to lower the COD value of effluent liquor.
Each method has some merits and demerits and can not be said to be completely
satisfactory.

Meanwhile, we found that the quantity of SaOs  generated in the process
decreases greatly when a reaction of oxidation of SO3  takes place under such
a high pH-value as 5.0 —6.0, because the dissolved ferrous ion which plays
a role of a catalyzer, is hardly present in the liquor under such a high
pH-value.

Therefore, in the improved process of Magnesium-Gypsum process, which is ex-
plained later, the generation of SaOg  is surpressed by controlling pH-value
to high degree.   That is, most of SOs  are oxidized in an absorber in which
the pH-value is high, and the rest of 80s  is oxidized by the air in an
auxiliary oxidation tank in which the pH-value is kept 5 — 6.

Under the above condition, the quantity of SaOs generated in the system can
be controlled to a very small quantity, so it becomes possible to keep the
COD value of effluent liquor lower than a specified regulation value in most
case of discharging of the effluent liquor.
                                  966
-------
o
w
X
o

w
O
H



§
-U
w
o
H
c/>

CQ

P
01


W

33

H
H
IZ
W
c/3
O

s
H
W
 CO

 H















CO
CJ
•H
4-1
en
•H
S-l
0>
4-1
CJ
Cfl
S-i
CO
U
•a
c
cfl

CO
0)
iH
O
S-i
(X













CO
4-1
r!
Contei








c
o
•H
4-1
cfl
CJ
•H
4-4
•H
cn
w
cfl
1— 1
U
O
2


Tl
. i
O

j
0) CJ
> -H
•H S-l
4J 3
Cfl 4-1
S-i * — 1
cfl 3
CX CO
O M-J
•• CJ 0
4-1 o; c
•H S-l O •-
CJ cfl -H Ol
CO 4-1 00
CX C CX S-i
cfl O g cfl
0 -H 3 iH
4J en
i— I cfl C co
iH S-i O -H
CO 0) CJ
g CX cfl
en o cu 13
43 O
CO 13 4-1 CO
c
O Cfl 4-1 CJ
4-J 3 -H
4-1 43 4-1
cu c tn
iH O) "3
43 g 0) cfl
Cfl CX rH CJ
4-1 -H CX
3 cr -H c
c/} w en cfl
1
cfl

4_)
C CJ
0) o
cj ••• O
C 13 O
O *H i — 1
U CJ

«* 3
4-1 CJ O
Cfl -H 43
S-J Cfl
C 3
O M-J ••>
•H rH 0)
4-13 S-i
•H CO 3
en "4-i
O 4-1 6^2 cfl
cx o o s-i
g iH 0)
o c p a
o o £5
CU -H CU
O 4-1 m 4-i

cu

4-1

(>•>
f\

C T3
O -H
•H CJ
4-1 Cfl
•H
CO CJ
O -H
CX S-i
g 3
O U-l
U rH
Ol 3
Q en
iH






, — 1
cfl
4-1

0) rH
g cu
0) en
rH CO
CU 0)
n |
0 T3
cu
•H -H g
Cfl S-i CO
030)
CX CO 4-1
cu co en
O Ol
S-l 4-1
•• cx o
4-1 Cfl C
•H 0
> 4-1 -H
•H O 4J
cn cx
o >, e
S-i 4J 3
S-i -ri en
O en G.
a co o
0) CJ
oo o
c o) cu
O !Z 00
S-i S-J
4-1 •" CO
CO GO J


cfl

4-1
cfl C
cfl
C 43
O 4-J
•H

•H CU
CO 43
o oc
CX -H
g 43
o
cj CU
01 S-J
13 3
4-1
I— 1 Cfl
cfl S-i
g 0) 0
cu g o
43 0) 1 — •
H 4-1 rH
c
o
•H
4-J
•H
Cfl
o
cx
g
o
o
01


rH
cfl
g
M
0)
43
H
CN
CU
1 4-1 43
•H O 4J 4-4
M-l O
M-l >, C
CU 4-1 -H 4-1
en -H C
•H 4-i cn ;>, cu
cn C en s-i g
>-, 0) 0) CU 4J
•H S-l O p> Cfl
O S-i 0) O CU
S-i 3 Z CJ S-i
4-1 o a) 4-j
0 •- S-i
0) 4-1 C 0)
i— 1 O O CD 43
0) -H 43 4-J
CU 4-1 4-1
3 CJ CO O
O C S-i O 4-i
1-4 01 4-1 4-1
T) C 0)
... C CU CU iH
>> 0) CJ rH 43
U CX C 43 cfl
C 0) O Cfl 4J
CU 13 CJ 4-1 -H
•H -H 3
cj 0)1 3 co
•H 00 rH C/5
M-l S-l O 4-1 g
M-i cfl •" O cfl
o) rJ cu 4-) c cu
43 C S-i
4-1 ..-4-1 0) 4-1 4-1
C P^ g 3 en
0) O C 4-1 43
S-i C O CO 4-1
s-i 0) cu - c
3 -H >^ S-i g 01
cj o a 4-1 cu 3
•H C 1 4-1 rH
3 M-I 01 0) en M-I
O 4-1 -H S-l >-, M-l
_] cu o cx cn cu
cu
43 C
4-1 Cfl
•H S-i
3 43
g
c cu
0 g
•H
4-i en
Cfl -H
T3 CO
•H >>
X rH
O CO
•H
O 13
•H 0
4-1 S-l
>^ 4-J
rH CJ
O 0)
S-l rH
4-1 01
o
0) T3
rH C
W CO


1
CO
13
•H
X
o

CJ
•H
4-1
^Xj
rH
o
S-l
4-1
CJ C
0) O
r-i -H
W 4J
ro



T3
0)
^~, 4-1
4-1 Cfl
c cu a
0) 00 -H
g S-l rH
4J Cfl CX
CO rH g
01 O
S-l 4-J O
4-1 O
cu e c"
S-i cfl O
dj CU -H
^— ' S-l 4-J
00 4J CO
S en s-i
4-J
M-l 4-1 C
o c cu
0) cj
rH 3 C
CO rH O
> M-l CJ
O M-<
g oi 3
0) O
S-l C rH
CO
4-1 13
o o c
4-J CO
^
4-J CU >.
•H rH 4-1
Cfl 43 -H g
CO Cfl 4-1 0)
0) 4-> C J-J
CJ -H Cfl CO
CU 3 3 >,
JS C/J D' CO
4-1 1
C 43 "O
0) 4J 01
43. • cu cn
S-i O g C
O J-J CU
cn cu a) 13
13 43 C
CO - 4-1 O
c o
C -H 43
cfl en 4-j cfl
CU -H
-C >-i 3 s-i
4-1 O
•H CO S-i M-l
3 cfl cu
C 43 4J -
O O 0) CO
•H 3 00 "
4-J en o CN
CX 4-1 ' S-l
S-l 4-1 rH O
O C 13 3
en cu 0) cn cj*
13 oo en t3 -H

43 cfl
13 00
C -H 4-J
cfl 33 0
c
oo •••
C 4-1 en
H rH -H
iH Cfl
13 cn 4-1
C 1 cfl
CO CO CU
43 & -C

4-1 0)
iH CX 4-J
3 3 cn
cj cO
H -O 3
4-1 0)
14-1 -H 0)
H S-i 43
Q 13 4-J
.
1 0
cn 4-J
3 Q)
C 43
0 g -
•HOC
4-J CJ O
cfl -H
S-i 13 4-J
o cu cn
CX 00 3
CO S^ 43
>> CU g
cu g o
43 CJ
43 CO >-,
CO
CX cfl S-i
3 CX
1 43 CO
oo oo
e 3 -
•HOC
;>> s-i o
S-i 43 -H
Q 4-1 4-J












cx
3
1
00
c
•H
^
S-i
o
m
                                                       967
-------
B.  IMPROVEMENT OF THE QUALITY OF BY-PRODUCT GYPSUM

1.  Use and Quality of Gypsum

    As the flue gas desulfurization plant has made progress in its practical use,
    it has become a nation-wide problem to secure the use of the by-product gypsum
    from a desulfurization plant in Japan.

    Main use of the gypsum by-produced from FGD plant are for cement additives and
    for the raw material of wall board at present, and a percentage of the lat-
    ter is anticipated to increase in future.

    In order to use the by-product gypsum from FGD plant as the raw material of
    wall board,  the  values which regulate the quality of gypsum for board, such
    as tensile strength at wet state, mixing water quantity, coagulation time,
    etc., are necessary to be within a certain range.
    The main factors which have a remarkable effect on the values related to the
    quality are the size of crystal and the concentration of impurities.

    In the Magnesium-Gypsum Process,  Mg(OH)2 is regenerated by the following re-
    action of regeneration, adding Ca(OH)2 to the MgSOi, generated in the system.

         MgS04 + Ca(OH)2 + 2H20   ->  Mg(OH)2 + CaS04.2H20

    In the above reaction, only the gypsum like mud which consists of the crystal
    of very small size is generated in the case when a mole ratio of MgSOit/Ca(OH) 2
    is larger than 1.0, which is called "Perfect Regeneration".

    In the standard Magnesium-Gypsum Process, the regeneration of Mg(OH)2 under
    a condition where the mole ratio of MgSOit/Ca(OH) 2 is smaller than 1.0, which
    is called "Partial Regeneration", is adopted in the basic design in order to
    prevent the generation of the gypsum with the small-sized crystal, which is
    troublesome  to  handle with and is inadequate to any use.

    However, we have found a method to control the crystal size of gypsum generat-
    ed in "Perfect Regeneration" process, where the mole ratio MgSOi*/Ca(OH) 2 is
    larger than 1.0, by an addition of some amount of the crystal of gypsum as a
    "seed crystal" to a regeneration tank and by regulation of a reaction
    temperature of regeneration.
                                     968
-------
    An outline of our experiment on the above method is shown in the followings.


2.   Experiment to Control the Size of Crystal of Gypsum

    a.  Condition and Results of the Experiment: Conditions of the experiment in
        a continuous reaction tank of small scale are shown in Table 5.
        Parameters regulated are as follows:

        1)   Feed rate of the "see crystal"
            (weight of seed crystal/weight of gypsum to be regenerated)  x 100(%)

        2)   Temperature of reaction (°C)

        3)   Mole ratio of MgSOi4/Ca(OH) 2 in regeneration reaction

        The concentration of MgSOi* in the mother liquor (the filtrate of separa-
        tion of gypsum after the oxidation) supplied into the Mg(OH)2 regeneration
        tank was regulated to be 5% for all the test run except for No.7, in
        which MgSOu concentration was regulated to be 10%.

        The "seed crystal" was prepared for the test run by feeding into a mother
        liquor by a prescribed quantity.

        Ca(OH)2 to be supplied to Mg(OH)2 regeneration tank was adjusted to be
        a slurry of 7% mixed with fresh water before the test.
                                    969
-------
An analysis of the particle size of "seed crystal" and the gypsum generat-
ed in each test run is shown in Table 6.
                 Table 5.
CONDITION OF TEST RUN
Test Run
No.
1
2
3
4
5
6
7
Feed rate of the
"seed crystal"
87%
0
87
87
0
87
50
Temperature of
reaction
58°C
58
68
77
77
77
77
Mole ratio of
MgS01|/Ca(OH)2
1.25
1.25
1.25
1.25
1.25
0.67
2.50 ;
        Table 6.     ANALYSIS OF THE PARTICLE SIZE OF "SEED
                     CRYSTAL"  AND THE GYPSUM GENERATED
Particle size
(micron)
0 ^ 8
8 ^ 16
16 ^ 24
24 ^ 32
32 % 44
44 % 53
53 % 74
74 ^ 105
105 ^ 210
210 ^ 297
297 ^ 350
over 350
Test run No.
1
0.5
8.5
10.7
14.2
11.8
17.5
17.4
12.5
5.1
1.4
0.4
nil
2
1.8
10.6
12.2
5.4
4.5
13.8
21.8
14.0
14.5
0.9
0.5
nil
3
0.1
3.5
8.8
4.5
15.2
17.2
22.4
13.9
6.1
0.9
0.4
nil
4
0.2
2.4
5.4
8.4
9.9
19.5
25.9
17.4
9.1
1.3
0.5
nil
5
0.9
9.6
7.1
4.5
3.8
11.2
15.5
18.6
22.4
5.8
0.6
nil
6
0.6
7.9
6.6
6.7
6.3
14.6
14.3
8.0
9.7
4.3
2.8
18.2
7
0.2
4.3
11.9
9.4
9.8
16.5
24.4
15.2
7.5
0.5
0.3
nil
seed
crystal
1.2
14.4
13.3
11.1
6.1
24.8
14.6
7.7
5.0
0.9
0.2
0.7
                             970
-------
b.  Examination of the Results of Experiment;

    Effect of Feed Rate of "Seed Crystal"; The microscopic photographs of
    the crystal of gypsum generated in each test run are shown in
    Figures 9—13.

    It is obvious that the crystal of gypsum supplied with "seed crystal"
    grows to a larger size than that without a supply of "seed crystal",
    making a comparison between the photographs for test run No.l and
    No.2 or No.4 and No.5, which are the cases    with and without a sup-
    ply of "seed crystal" under the same condition of a mole-ratio and a
    reaction temperature.

    In addition, the number of small particles of gypsum in the case with
    a supply of "seed crystal" proved much less than that in case without
    "seed crystal".

    Effect of Reaction Temperature; An effect of a reaction temperature
    on the particle size of gypsura under a constant feed rate of "seed
    crystal" and a constant mole-ratio of the regeneration is shown in
    Table 7, where the particle size is found to become larger as a reac-
    tion temperature rises higher.
                                971
-------
Figure 9  Crystal of Gypsum in Test Run No.l




          Feed rate of the "seed crystal" : 87 %




          Temperature of Reaction         : 58°C




          Mole ratio of MgSO4/Ca(OH)2     : 1.25
Figure 10  Crystal of Gypsum in Test Run No.2



           Feed rate of the "seed crystal"  :   0  %




           Temperature of Reaction          : 58°C



           Mole ratio of MgS04/Ca(OH)2      : 1-25





                     972
-------
Figure 11  Crystal of Gypsum in Test Run No.4



           Feed rate of the "seed crystal"  :  87  %



           Temperature of Reaction          :  77°C



           Mole ratio of MgSO4/Ca(OH)2      :  1.25
Figure 12  Crystal of Gypsum in Test Run No.5



           Feed rate of the "seed crystal"  :  0 %




           Temperature of Reaction          : 77°C



           Mole ratio of MgS04/Ca (OH.) 2      : 1.25
                     973
-------
Figure 13  Crystal of Gypsum in Test Run No.6



           Feed rate of the "seed crystal"  : 87 %



           Temperature of Reaction          : 77 °C



           Mole ratio of MgS04/Ca(OH)2      : 0.67
                     974
-------
 Table  7.    RELATION BETWEEN THE REACTION TEMPERATURE
              AND THE PARTICLE SIZE OF GYPSUM
Test run No.
"seed crystal"
1
3
4
Reaction
Temperature

58°C
68°C
77°C
Weight ratio
of particles
less than
20 micron
24%
16%
8%
5%
Diameter of
particle at
weight ratio
of 50%
41 micron
46 micron
50 micron
58 micron
Effect of Mole-Ratio of Regeneration: It is found that the number of
particles of small size in the case of a smaller mole-ratio of regenera-
tion is more than that in the case of a larger mole-ratio, making a com-
parison between the photograph for test run No.4 and No.6, in which the
mole-ratio is 1.25 and 0.67 under the same condition of a feed rate of
"seed crystal" and a reaction temperature.
                            975
-------
      Ill,   NEW MAGNESIUM-BASE  DOUBLE ALKALI  PROCESS

A.   DESCRIPTION OF THE NEW PROCESS

     We have recently established a  new magnesium-base double alkali process,
     improving the standard Mg-Gypsum Process.
     The new process is characterized by "Perfect Regeneration" in the process,
     that is, in the following reaction of regeneration in which Mg(OH)2 is
     regenerated by addition of Ca(OH)2 to MgSOi* generated in the system,

         MgSOi* + Ca(OH)2 + 2H20   ->  Mg(OH)2 + CaS04.2H20

     a mole-ratio of the reactants supplied into a regeneration tank is re-
     gulated to be MgSOit/Ca(OH)2 >ll.O, therefore, all the Ca(OH)2 supplied
     into a regeneration tank are changed into Mg(OH)2 and CaSOi*.2H20, so,
     Ca(OH)2 is not present in the liquor supplied to an absorber from the
     regeneration tank.

     On the contrary, "Partial Regeneration" in which a mole-ratio of regene-
     ration is regulated to be MgSOit/Ca(OH) 2 < 1.0, and the liquor or alkali
     material supplied to an absorber from the regeneration tank contains
     Ca(OH)2 as well as Mg(OH)2, has been adopted in the standard Mg-Gypsum
     Process, and the mole-ratio of  regeneration has been regulated to be
     MgS04/Ca(OH)2 = 1/3 ^ 1/4 in order to obtain the by-product gypsum with
     a favorable size of particles.

     A comparison between "Perfect Regeneration" and "Partial Regeneration"
     is shown in Table 8.   According to the comparison, it is found that
     "Perfect Regeneration" is superior to "Partial Regeneration".

     As shown above, "Perfect Regeneration" is a process in which the most of
     the advantages of a double alkali FGD process are fully made, and it can
     be applied  in practice at once, if only its conventional disadvantage of
     generating  the gypsum with a very small particle size is overcome.
                                976
-------
o
M


I
W

w
o
w
cu


o
o
M
H
w
z
w
o
w
Pi

H
O
w
h
Pi
w
PH
W
w
pq

z
o
en
M
PS
O
CJ
CO




 QJ
 cfl
H















z
o
M
H
^^
CM
w
^
w
o
w
pi'

J
<3j
H
H
Pi
•^
P-c




















£2
O
h- 1
f— i
<^
PH
W
Z
W
o
w
p^

H
CJ
w
rW
Pi
W
PH



































O
•
.— 1

V
CM
X- N
pd
o
s 	 '
cfl
CJ
1 — ^
J-
0
w
IP
















o
•
rH

/\ |j

CM
X"N
X
O

cfl
CJ
-* —
j-
0
C/3
60
*£










U-l
0 C
o
0 -H
•H 4-1
4-1 Cfl
Cfl S-l
S-l CU
1 (3
Q) CU
rH 60
J2 ^
g M

4-1
S-i C
QJ M-l QJ
SO n
0 S-i
rH Q) S O
s-i O en
QJ 3 
4_1 rf~\ |
E^ (3 S ffl
O QJ CX
C 60 4-1
QJ Cfl rH QJ
•H cfl H
o 4-1 en 4J
•H C 1 S-I
M-l CU CO T3 QJ
M-l 43 CJ f3 &
QJ S-i CO O
O T3 rH
C en c >-,
O 42 cfl 4-i en
•H CO -H -H
4-1 4-1 rH
CX 0) rH -H S-l
S-i 43 CO 4D O
o 4-1 en 3 3
en 1 rH cr
42 en 60 o -H
< co S en rH
QJ
43 S-i
4-1 QJ
43
en 4-1 QJ 60
cfl rH 43 -H
cfl 4-1 43
~ en
s-i I -a en
QJ 60 C -H
4J S CO
4-1 S-l
QJ 0) >> O
rt f? 4_i p
4-> -H cr
en IH -H
•H ^ -H iH
> — 1 43
>, C 3 4J
O O rH (3
C O QJ
QJ en en 42
•H -H S-l
O 43 O
•H 4-1 60 en
M-l C3 -H 43
LH QJ 43 CO
QJ 60
cfl co m
C O
O 4-1 en
•H C Cfl CU
4-J QJ 43 3
CX 43 rH
S-i S-i 43 CO
0 0 0 >
en co -H i
43 43 43 33

co en co i 4-1
43 T3 33 C
" CO -H CX QJ
60 X >
(3 (3 O S QJ
•H cfl O S-i
rH f^, rH CX
cfl O 43
CJ 4-1 cfl O
en C O 4-1
•H CM ^-i
M-i 33 QJ >,
O T) CN 13 S-i
QJ • C CO
s-i -H j- 3 en
co rH O en
QJ cx eo c a)
M-I ex eg o o
3 U -H QJ
cu en we
S-i 0 cfl
O QJ 4-J M en
B S-i C QJ -H
Cfl -H CX
en o S-J
•H CM T3 QJ 60
^ QJ (3 43 C
QJ 33 60 cfl S-i -H
M O C O rH
Q) ^ cfl en cfl
43 cfl 43 O 43 CJ
H CJ cj c/) cfl en
0
CM en cu
rH 33 TJ CU 3
rH CN C O rH
CO • cfl T3 CO
* >
en o " O I
cfl c/j S-i CM 32
cfl QJ 33 CX
' CJ 43 csl
60 S-i 43
f3 O O 60
•H 4-1 en J -H
-H C 43 O -43
cfl -H cfl CO S^
O cfl CU cfl
en QJ C O 43 CU
60 cfl s-i S-i i-H
M-l (3 M-l O QJ 43
O cfl O O en T3 -H
43 4-1 43 C3 en
s-i cj C C cfl 3 en
cfl -H o O
QJ en -H Q) c cx
M-l Td 60 4-1 43 O
(3 c co 4-1 -H en
en 3 -H s-i 4-1 -H
en o s-i QJ c cfl
QJ CX QJ (3 -H (J S-l
rH 6 4J QJ Q) QJ
O C 60 S-l CX 43
eQ CJ QJ 3 O S-i
•HI U 0 O
CS QJ QJ CJ C CO
Q) O S-i C O cfl 43
S-i O cfl
CU QJ U-l QJ 4-J •>
43 43 QJ 43 0 0 C
H 4J 43 4-1 C C/l -H





60
C
•H
rH
Ct)
a
CO
                                                     977
-------
         As has been previously mentioned,  we have found a method to obtain the
         gypsum with a favorable size of particles even in "Perfect Regeneration"
         by addition of "seed crystal" and  by control of the temperature in the
         regeneration.

         So, it has become possible to establish a new magnesium-base double alkali
         process applied with "Perfect Regeneration1'.
         A typical example of the new process is shown in the followings.
B.  DESIGN EXAMPLE OF THE NEW PROCESS

    This example is for a flue gas desulfurization plant for a coal-fired boiler
    in a power station in Japan.
    The "Perfect Regeneration" is applied in the Process.

1.  Flow Diagram of the Process

    The flow diagram of the process is shown in Figure 14.
    All the Ca(OH)a change into Mg(OH)z in the regeneration tank "9" by a reac-
    tion with MgSOit, so the contents of the alkali slurry tank "10" are Mg(OH)2,
          and CaS0lt.2H20.
    The oxidation tank "11" is much smaller than an oxidizer in the standard
    Mg-Gypsum Process, because there is no CaSOs in- the absorber "2", and it is
    not necessary to oxidize the calcium compounds.
    Besides, most of MgSOs and Mg(HS03)2 are oxidized in the absorber and the
    oxidizing load is expected to be much lighter.
                                     978
-------
                                                o:
                                                o
                                                CO
                                                
                                                LU
                                                <->
                                                o:
                                          Oj X
                                            2:  o
 U3
 w
 o
 o
 o
•H

rH

 fC








 0)
                                                                 C
                                                                O
                                                                 ft
                                                                tt
                                                                 I
                                                                 fc,
                                                                 Oi
                                                                 c
                                                                 c.
                                                                 (C
                                                                 0)
                                                                2

                                                                 0)
                                                                Xi
                                                                •p
                                                                e
                                                                (0
                                                                (C

                                                                •H

                                                                P




                                                                O
                                                                rH

                                                                PL,
                                                                0)
                                                                ^
                                                                3
                                                                &»
                                                               •H
                                                                PL,
979
-------
2.   Design Conditions

    a.   Out put              :  175 MW

    b.   Fuel                 :  Coal

    c.   Absorber inlet gas condition
            Gas flow rate (wet  base)                 :  610,000 Nm3/h
            Temperature                              :  137°C
            S02 concentration (dry base)             :  350 ppm
            Particulate concentration (wet base)      :  40 mg/Nm3
            Cl concentration in coal                 :  290 ppm

    d.   Absorber outlet gas condition
          S02 concentration (dry base)
            design value     :  10 ppm
            guarantee value  :  20 ppm
          Desulfurization efficiency
            design value     :  97%
            guarantee value  :  94%

    e.   Gas temperature after reheating              :  137°C

    f.   Effluent liquor flow rate                    :  1,500 kg/h

    g.   Effluent liquor COD                          :  90 ppm
                                   980
-------
   TECHNICAL AND  ECONOMIC  FEASIBILITY  OF
     SODIUM-BASED  SO2  SCRUBBING SYSTEMS


         L. K. Legatski, J. E. Makar, and A. A. Ramirez
                       FMC Corporation
                         Itasca, Illinois
ABSTRACT

    Sodium-based scrubbing utilizing commercial soda ash, trona ore,
and process waste liquors containing soda ash has been demonstrated
to be an effective  means of sulfur dioxide removal. It is especially
applicable  to  utility  boiler applications  in the Eastern states  where
stringent state regulations require low S02 and particulate emission
levels. A soluble salt scrubbing system offers the advantages of very
high removal efficiency combined with superior entrainment removal
capabilities to minimize the contribution of  entrainment to particulate
emissions.
    The Company's extensive experience with single and double-alkali
sodium-based flue gas desulfurization (FGD) systems is reviewed. The
operating advantages of these systems have resulted in demonstrably
higher  availabilities  than  exhibited  by conventional lime/limestone
systems. Comparative economics are presented and indicate that the
sodium systems are competitive and can be less costly than the conven-
tional systems for many cases. Costs per ton of coal for the various FGD
alternatives indicate significant potential savings when compared to the
use of low  sulfur coal.
                             981
-------
            TECHNICAL AND ECONOMIC FEASIBILITY OF
        SODIUM BASED SULFUR DIOXIDE SCRUBBING SYSTEMS
Introduction

     The passage of the Clean Air Act Amendments in 1971 and the
     increasingly high cost and questionable availability of oil
     have provided industry substantial motivation to develop tech-
     nologies to deal with sulfur dioxide emissions.  Of the many
     technologies developed, the most widely used systems utilize
     lime and limestone as chemical reagents.  The many operating
     problems of these conventional lime/limestone systems are well
     publicized.  These problems motivated FMC to explore the use
     of sodium chemicals for sulfur dioxide absorption.  The tech-
     nical and economic feasibility of these systems is the subject
     of this paper.

Process Development

     The original system was developed at FMC's plant in Modesto,
     California and uses a sodium solution to react with the SO-
     to produce sodium sulfite, bisulfite and sulfate for disposal.
     This "single-alkali" system is further described subsequently.
     The single-alkali system is particularly applicable to small
     installations, installations which have low concentrations of
     SO^, or those plants which have caustic waste streams readily
     available.  The initial work at Modesto formed the basis for
     the single-alkali sodium scrubbing system which has been
     installed at FMC's soda ash plant in Green River, Wyoming.  At
     this plant there are waste streams available with sufficient
     sodium values to remove the desired S02 from the flue gas.  The
     spent liquor is ponded with the rest of the plant blowdown.
     This type of installation is the simplest and least capital
     cost intensive of the FGD systems available.

     The high cost of sodium chemicals at Modesto prompted FMC to
     investigate ways to transfer the absorbed sulfur dioxide from
     a sodium chemical to a less expensive carrier, calcium.  Exper-
     imentation led to the development of a process in which lime
     is added to the spent liquor causing a reaction to occur which
     regenerates most of the sodium values for reuse.  This patented
     "double-alkali" process decreases chemical costs significantly
     and minimizes potential disposal problems.  The double-alkali
     process has been increasingly accepted as a technical and eco-
     nomically feasible solution to the sulfur dioxide emission
     problem.
                                   982
-------
Process Description

     Single-Alkali

     A schematic of the single-alkali process is presented in
     Figure I.  Sulfur dioxide is absorbed in a solution containing
     sodium sulfite (Na^O^), sodium bisulfite (NaHSO.,), and sodium
     sulfate (Na2SO.).  As S02 is absorbed, the sodium sulfite is
     converted to sodium bisulfite and a very small percentage is
     oxidized to sodium sulfate.  A bleed stream is taken from the
     recirculation system at the same rate that the SO- is being
     collected.  The bleed stream can then be further processed
     through neutralization to eliminate the sodium bisulfite and
     aeration to convert the sodium sulfite to sodium sulfate if
     it's necessary to lower the chemical oxygen demand.  Neutrali-
     zation and aeration may be necessary to allow disposal into
     other wastewater treatment facilities or waste ponds.

     Double-Alkali

     A process schematic of the double-alkali system is presented in
     Figure II.  As can be seen by comparing Figure I and Figure II,
     the SO- absorption loops for single and double-alkali systems
     are virtually identical.  In a double-alkali system, however,
     the bleed stream is further processed in the regeneration loop
     where the sodium bisulfite is reacted with slaked lime
     (Ca(OH)2) in a low residence time agitated tank.  The reaction
     of lime and sodium bisulfite regenerates sodium sulfite which
     is returned to the recirculation loop for reuse.  Calcium sul-
     fite (CaSO.,) which is precipitated is fed from the lime reac-
     tor to the thickener where the solids are concentrated.  The
     slurry from the thickener is pumped to vacuum filters where
     filter cake of approximately 60% solids is formed and washed
     to minimize entrained sodium values.  Sodium consumption is
     thus reduced to approximately 2% to 5% of the S02 collected
     for many applications.

     The scrubbing solution in the S02 absorption loop is normally
     controlled at a pH of 6 to 7 witn 6.5 as a design point.  Above
     a pH of 7, carbon dioxide absorption becomes significant and
     can lead to formation of calcium carbonate scale.  Below a pH
     of 6, the increase in S02 vapor pressure reduces the system's
     ability to absorb S02«  PMC uses a setpoint of 6.5 where the
     sodium sulfite-sodium bisulfite solution is highly buffered and
     can readily adapt to rapid changes in S02 concentrations while
     maintaining constant collection efficiency.

     The formation of sodium sulfate, which cannot be readily regen-
     erated to recover sodium values, is inhibited through the use
     of a high ionic strength scrubbing solution that contains a
                              983
-------
    FLUE
     GAS
   BY-PASS
  FLUE GAS



_
y\
DISC
CONTACTOR
SCRUBBER



Na2SOj
NaHSOj
NazS04



•~
         S02+Nfl2S03+H20-«-2NaHS03
                                     TO EXHAUST
                                     'STACK
                                                                TO  ATMOSPHERE
                                      —Na2C03
                                             NdjSO,
                                             Na2S04
                              NEUTRALIZATION
                                  TANK

                              NfljCOj-l- ZNoHSO,— -


                                    FIGURE  I
                                  SINGLE- ALKALI
                               PROCESS  SCHEMATIC
                                                                          NoiS04
LOW C O. D.
LIQUID TO
DISPOSAL
                                                                  i—• 2NozS04
    FLUE
    GAS
  BY-PASS
FLUE
GAS


m-l
/
Dl
L
sc
CONTACTOR
SCRUBBER



Na,SO,
NaHSO,
Na2S04



                                    TO EXHAUST
                                    STACK
                                                            Ca(OH)z
                                                                      SOLID TO DISPOSAL
                                            Co(OH)2+2NoHS03-»CaSOJ'|H2O^Na,SOj-l-Ij

                                      FIGURE  IT
                             CONCENTRATED  DOUBLE-ALKALI
                              PROCESS   SCHEMATIC
                                              984
-------
     high sulfate concentration.  This inhibits oxidation from sul-
     fite to sulfate and minimizes the consumption of sodium chemi-
     cals.  FMC's scrubber designs have also been optimized to pro-
     vide for minimal oxidation.

     The regeneration of sodium values takes place in the lime reac-
     tor where the reaction of sodium bisulfite with lime is con-
     trolled at a pH of 8.5.  This is effectively the titrimetric
     endpoint for sodium bisulfite and operating at this pH insures
     excellent response of the control system.  This greatly reduces
     the possibility of introducing excess lime.  In addition to
     poor chemical utilization, excess lime would result in poor
     filter cake quality and substantially higher operating costs.

     Maintaining high dissolved sulfite concentrations in the scrub-
     bing liquid results in dissolved calcium concentration consider-
     ably below saturation levels, thereby eliminating the formation
     of calcium scale.  Calcium scaling is the major contributor to
     the operating problems experienced by conventional lime/limestone
     systems.

Technical Advantages

     Sodium scrubbing systems have significant technical advantages
     which result in lower cost for many applications.  These advan-
     tages are described below.

     1.   Reliability.  A soluble scrubbing system is inherently more
          reliable than a calcium slurry system.  FMC's experience
          to date clearly illustrates this fact.  Reliabilities have
          in almost all cases exceeded the 90% requirement often spe-
          cified.  The prime reason for this is the elimination of
          the calcium scale which has caused the majority of the
          problems for conventional systems.  To date no FMC system
          has been shut down due to calcium scaling.

     2.   Simultaneous SOg and particulate removal.  Sodium scrubbing
          systems provide the flexibility to simultaneously control
          sulfur dioxide and particulate.  Alternatively, particulate
          control systems can be provided which can be subsequently
          adapted for SO- control.

     3.   Lower maintenance.   The use of a buffered solution rather
          than a slurry for scrubbing greatly reduces corrosion and
          erosion in the scrubber and associated equipment and piping.
          This combined with the no-scaling characteristics previously
          discussed leads to considerably reduced maintenance costs.
                             985
-------
     4.    Manageable waste.   The filter cake produced by the double-
          alkali system differs significantly from that produced by
          lime or limestone  processes.   As can be seen in the illus-.
          trations (Figures  III to VI)  the double-alkali precipitate
          is a granular material,  rather than a sludge.  It is moved
          by conveyor belts  from the vacuum filters to transfer
          points.  It can be conveyed directly to a disposal con-
          tainer or allowed  to fall on the ground for future removal
          by front-end loader and  truck.  The material does not go
          into a slurry form upon  mechanical agitation.  The mechan-
          ical stability allows the precipitate to be moved and
          landfilled using conventional wide-track equipment.  Suc-
          cessful landfill operations are presently being conducted
          in two states.

          Local regulations  for landfilling the filter cake may or
          may not require the area to be lined.  Linings may be
          required in locations where the soil is porous or the
          ground water level high.  Fixation is not required to pro-
          vide mechanical stability.

     5.    Low power requirements.   Sodium systems provide the oppor-
          tunity to operate  at comparable pressure drops and at sig-
          nificantly lower liquid-to-gas ratios than those utilized
          in conventional lime/limestone systems.  Total power con-
          sumption can be as low as 40% of other systems.

     6-    High S02 collection efficiencies.  S02 collection effi_
          ciencies of up to  99% have been demonstrated.  High collec-
          tion efficiencies  often  permit partial bypass of the flue
          gas for reheat.

     7.    Ease of operation.  A highly satisfactory control system
          package has been developed and demonstrated.  The design
          of this package is predicated on the system being operated
          by "typical" boiler room operators and not engineers.

Operating Experience

     Figure VII summarizes major installations by FMC.  Installations
     thus far range up to 265 Mw equivalent.  FMC's experience in
     sulfur dioxide control  began  with the completion of the original
     installation at FMC's Modesto chemical plant.  The installation
     of this system was dictated by adverse ambient conditions exist-
     ing in the area.  The unit is a 30 Mw equivalent system which
     has been in operation since late 1971.  While not completely
     representative of the design  FMC would utilize today, the system
     has demonstrated an extremely high degree of reliability in an
     application which experiences significant fluctuations in SO-
     conditions.  Since December of 1971, this system has been avail-
     able in excess of 95% of the  time.  FMC's largest non-utility
                                    986
-------
                                      FIGURE III


                                Conveyor discharge into
                                disposal container at
                                Firestone Demonstration
                                Plant.
                 FIGURE IV

Sample of filter cake at Firestone landfill,

                  987
-------
                    FIGURE V

  Landfill  area at Firestone  (note dozer tracks
              in right foreground).
                     FIGURE VI

Climbing a recently dumped three-foot deep pile of
  waste material demonstrates physical stability.
                    988
-------
           3
           •H
            id
         CN>
        O
        cn
       r-
       X
       CN
                      4-1
                   rH -rl
8-9
 CN
O  i
en
        cn m     o  o
        cn cn     r^  en
rH  4-1
   i-H
 O  -H
 E  ja
 CD  rd
OS  rH
    •rl
 CN rd
O  >
en  <

dP  dP
o  ^3*
cn  cn
                             IT)
                  VD
 0

 
<0
rH
rH
C8
4->
in
c
H
C
0)
^
(8
>
•H
3
O"1
W

>
2 E
>i 1 (8
4J E  X
(8 CO
SH J2
3 rH
4-1
fd
cn




3
2

o
CO

1

S
HH
U
o
o
0
H
o
m







3
2

o
i-H

1

MH
O
O
o
o
K
o
CO







3
g

cn


I

g
MH
U
O
O
O
v
O
i-H






3
2

in
in
j^i
I A
X
g co
MH &
0 rH

0 O
O O
0 O
O O
^ VO
CN ^





3
2

O
O
i-H
j^
1 rj
*\
g CO
4H XI
0 rH

O O
0 0
O O
O O
CN r-
no co





3
2

O
O SH

x^
i co

E rH
U 0
o
0 0
o »
O 0
- o
O CO
^r «.
H> rH





3
2

O

rH £
\
1 CO

g I-M
U 0
0
0 0
o »
o o
> o
O co

10 i — 1





3
2

in
^0 SH
CN JZ
\
1 CO

E rH
U 0
o
0 0
o »
O 0
» in
O CO
^r ^
00 rH



ft
co
3
Q

c
rH
•H
^

T)
J^
id

CN
o
en


JZ
eo
rd
^s,
rH
r^4

<^J
C
<8

CN
O
en


A
U)
id
^i
i-H
fa

13
C
(8

CN
O
en


r^
CO
rd
^i
rH
fa

Tl
C
rd

CN
0
en






rH
id
0
u

en
dP
CN

ro


A
CO
rd

rH
fa

Tl
c
rd

CN
O
en






rH
rrj
0
CJ

en
dP
CN

CO


A
CO
(d
^s,
rH
fa

T3
C
id

CN
0
en






rH
id
O
CJ

cn
dP
CN

CO





rH
O
SH
4J
C
0
CJ

CN
O
en
rH
id
o
CJ

cn
dP
CO
•
rH

0
4J

ft
D
i-H
id
0
U

rH en
0 dP
SH m
4-1 •

0
CJ 0
4-1
CN
O ft
en D
c
0
•H
4-1
(8
rH
H
(8
4J
CO
q
H
Q
H
U
•rl
Q)
CJ
rH
•rl
SH
4-1
CO
3
•O
fi
H

rj
12
fa
CJ
o
4-1
co
0)
•0
Q
S


0)
rH
•H 
a cn
IH co
0) 0
4-> 2
id
CJ
H
•H
SH
0
0)


i
W


H
q
0
4-1
OJ
rH
ft
rrt
5J


trial Chemical Di
in
3
""O
C
M

O
2
fa
SH
0)
•rl
(V,

C
0)
0)
SH
CJ


Indiana Gas and

c
SH
0)

4-1
3
0
en
Company
lie, IN
.,_!
u >
•H CO
SH C
4-i id
U >

rH
W
                                              989
-------
     louble-alkali installations to date are for the Caterpillar
     Tractor Co.   The initial unit is located in Mossville, Illinois,
     and has been in operation since October, 1975.  A smaller instal
     lation has been operating at Firestone Tire & Rubber Company's
     Pottstown, Pennsylvania plant since January, 1975.  Both of the
     above installations have demonstrated the disposability of the
     filter cake  generated by the double-alkali process.

     In spring of 1976,  FMC started up its largest single-alkali
     rcrubbing system thus far.  It consists of two 330,000 acfm
     disc contactor scrubbers (Figure VIII) on the new pulverized
     7oal-fired boilers  installed at FMC's soda ash plant in
     5reen Fiver, Wyoming.  The boilers are two 650,000 Ibs/hr units
     followed by  high efficiency electrostatic precipitators.  The
     scrubb -rs are designed to remove in excess of 90% of the sulfur
     dioxide at a low pressure drop.  These scrubbers are represen-
     '-ative of those which are being used by FMC in utility appli-
     In late 1976,  FMC was awarded its first contract for a utility
     double-alkali  flue gas desulfurization system.  This system will
     be installed at- Southern Indiana Gas and Electric Company's
     ' . '   Brrwn Un t No. 1 Station.  The award of this contract was
      igr.ificant in that it was based on the lowest evaluated cost
     .3 compared to conventional lime and limestone systems.

      ne of the key detriments to the further commercialization of
     sodium based sy: terns, especially double-alkali systems, is the
     mistaken contention that the higher chemical cost of thesw sys-
     tems will make them noncompetitive when compared to lime/lime-
     stone systems.  FMC believes that for many applications the
     PC nomics of sodium based systems are competitive when all
     as ects of c^sts are considered.  A complete cost evaluation
     nu; t include factors such as maintenance, chemical storage and
     handling cost, operability, system availability, and disposal
     cost.  In the  following section, the costs of sodium based sys-
     tems as compared to limestone systems are presented.  These
     costs are also related to coal costs since a majority of utility
     and non-utility applications consider low sulfur coal as an
     alternative to flue gas desulfurization systems.

Economic Comparisons

     Annualized costs of single-alkali, double-alkali, and limestone
     scrubbing systems for a 500 Mw utility power station are presented
     and compared with the higher cost of low sulfur coal.  Assump-
     tions made in  these calculations include:

          Load Factoi :  65%

          HHV Coal:  12,000 Btu/lb
                     1.18 MM TPY coal fired

                             990
-------
                 FIGURE VIII
 Two 330,000 acfm Disc Contactor Scrubbers
    controlling SO- emissions from
two 650,000 Ibs/hr pulverized coal boilers,
                        991
-------
     Outlet S02 Concentration Required:  1.2 Ibs/MM Btu

     Capital Charge:  15%

     Lime:  $40/ton

     Limestone:  $5/ton

     Soda Ash:  $85/ton

     Power at 2 cents/Kwh

Chemical consumption and power requirements of the various
systems were estimated based on published data or internally
generated information.  Disposal costs were assumed to be
$2.50/ton of double-alkali filter cake, $3.00/ton for limestone
sludge (40% solids), and $4.75/ton of S02 removed for a single-
alkali system.  It should be noted that direct disposal of the
bleed liquor from a single-alkali sodium system is usually not
feasible.

The above assumptions can vary significantly for different sites;
however, the assumptions are reasonable, and the resulting eval-
uation suggests that double-alkali system costs will be either
competitive or lower than those of conventional limestone systems,

In Figures IX-XI, the cost for the FGD systems evaluated are pre-
sented for varying levels of sulfur content.  For both double-
alkali and limestone systems, the cost per ton of coal varies
from approximately $4/ton to $9/ton.  Given that the cost of
low sulfur compliance coal in the East is often more than $10/ton
great r than for high sulfur coal, both double-alkali and lime-
stone systems often provide economic alternatives to the import-
ing of western coal or foreign oil.
                         992
-------
                          FIGURE IX
                   CAPITAL COST PER KW  ($'S)
Coal Sulfur Content           2%        3%         4%         5%
Double-Alkali                 $32.50    $35.00     $37.50     $40.00

Non-regenerated
Sodium                        $21.00    $22.50     $24.00     $25.50

Limestone                     $45.00    $50.00     $55.00     $60.00
                                   993
-------
                           FIGURE  X

        ANNUALIZED COST FOR FLUE GAS DESULFURIZATION


                    Single-Alkali       Double-Alkali       Limestone

2% Sulfur Coal
Capital Cost         10,500,000          16,250,000         22,500,000

Reagent Cost          4,203,000           1,263,000            297,400
Capital Charge (15%)  1,575,000           2,437,500          3,375,000
Power Cost              484,000             569,500          1,139,000
Maintenance             105,000             325,000          1,125,000
Labor                   300,000             300,000            700,000
Disposal                143,000             261,000            590,000
  Total               7,760,000           5,156,000          7,226,000


3% Sulfur Coal
Capital Cost         11,250,000          17,500,000         25,000,000

Reagent Cost          7,480,000           2,248,000            529,000
Capital Charge        1,688,000           2,625,000          3,750,000
Power Cost              484,000             569,000          1,140,000
Maintenance             112,500             350,000          1,250,000
Labor                   300,000             300,000            700,000
Disposil                255,000             464,000          1,050,000
  Total              10,320,000           6,556,000          8,419,000


4% Lullur Coal
Capital Cost         12,000,000          18,750,000         27,500,000

Reagent Cost         10,759,000           3,236,000            758,000
Cap ta1 Charge        1,800,000           2,813,000          4,125,000
Pow r Cost              484,000             569,000          1,139,000
Maintenance             120,000             375,000          1,375,000
Lab r                   300,000             300,000            700,000
Disposal                376,000             668,000          1,512,000
  Total              13,830,000           7,960,000          9,610,000


5% Sulfur Coal
Capital Cost         12,750,000          20,000,000         30,000,000

Reagent Cost         14,121,000           4,240,000          1,000,000
Capital Charge        1,912,000           3,000,000          4,500,000
Power Cost              484,000             569,000          1,140,000
Maintenance             127,000             400,000          1,500,000
Labor                   300,000             300,000            700,000
Disposal                481,000             876,000          1,982,000
  Total              17,430,000           9,390,000         10,820,000

                             994
-------
   15-


   14-


   13-


   12-


   II -


   10-

Z  8-
j-  6H
(O
5 -


4 -


3 -


2 -
                                              D-SINGLE ALKALI

                                              O-DOUBLE ALKALI

                                              A-LIMESTONE
                    % SULFUR IN COAL
                         FIGURE 31*

             COST OF FLUE GAS DESULFURIZATION
                PER TON  OF COAL  CONSUMED
                                 995
-------
SULFUR  RECOVERED FROM SO2 EMISSIONS
AT NIPSCO'S DEAN H. MITCHELL STATION
               Howard A. Boyer

            Allied Chemical Corporation
             Morristown, New Jersey

               Roberto I. Pedroso

            Davy Powergas Incorporated
                Lakeland, Florida
                      996
-------
                SULFUR RECOVERED FROM SO- EMISSIONS
               AT NIPSCO'S DEAN H. MITCHELL STATION
           HOWARD A. BOYER, ALLIED CHEMICAL CORPORATION
                      MORRISTOWN, NEW JERSEY
          ROBERTO I. PEDROSO, DAVY POWERGAS INCORPORATED
                         LAKELAND, FLORIDA
                              SUMMARY

Performance tests were completed and FGD operations are being continued
by Allied Chemical Corporation for Northern Indiana Public Service Co.
(NIPSCO) on a regenerable FGD facility for Unit No. 11, a 115 MW boiler
burning 3% sulfur coal at NIPSCO's Dean H. Mitchell Station in Gary,
Indiana.  This fully integrated demonstration facility successfully
removes 90% of the SO- from the entire flue gas output of Unit No. 11
and converts the recovered SO.- to elemental sulfur at rates over 20
tons per day.  The FGD system uses the Wellman-Lord SO- Recovery Process
of Davy Powergas Inc. (DPG) and SO. Reduction to Sulfur Process of
Allied Chemical.

Jointly funded by the U.S. Environmental Protection Agency (EPA) and
NIPSCO, this fully integrated demonstration facility represents the
first application of DPG's Wellman-Lord process in coal fired electric
utility service and the first combination of the Wellman-Lord/Allied
Chemical systems in flue gas desulfurization.  It is also the first time
that an experienced chemical firm has been engaged by an electric utility
to manage its entire emission control facility and market the recovered
products.

Performance tests were successfully completed on September 14, 1977,
following a virtually uninterrupted period of operation which included
12 days at a flue gas volume equivalent to 92 MW and 84 hours at an
equivalent of 110 MW.  During this period, 91% of the S0_ in the flue
gas was removed while burning coal averaging 2.9% sulfur.  About 204
long tons of elemental sulfur were recovered during this period.

Results were better than the performance criteria established for the
acceptance test period.   These criteria included:

1.   Minimum 90% S02 removal.
2.   Particulate emission not to exceed the Federal New Source Perfor-
     mance Standard (NSPS) for fossil fuel fired steam generators.
3.   Daily average soda ash feed to make up for sulfite oxidation not to
     exceed 6.6 STPD.
4.   Aggregate cost of steam, electricity, and natural gas at pre-
     determined unit costs not to exceed $56 per hour.
5.   Minimum 99.5% sulfur purity suitable for conversion to quality
     sulfuric acid by standard production practice.

                                   997
-------
Operation of the facility will continue through a demonstration year
during which operating reliability will be tested over a wide range of
conditions.  As of September 19, 1977, 608 long tons of high purity
sulfur have been recovered and sold to a nearby Allied Chemical plant
where it is being consumed in the manufacture of sulfuric acid.  Since
the beginning of the performance tests, the reliability of the S0_
emission control system has been very high, greater than 99%.

Performance at NIPSCO, to date and throughout the demonstration period,
should firmly establish the commercial availability of a reliable and
efficient regenerable FGD system in the U.S.  There is growing evidence
that electric utilities have a viable FGD option with conversion of
stack gas SO- into a useful product.   The concentrated S0_ recovered
by the Wellman-Lord process may be converted to either sulfur or
sulfuric acid - avoiding the burdens  and uncertainties of long-term
sludge disposal.
                            BACKGROUND

                           Introduction

At the inception of this project, the Wellman-Lord S0~ recovery process
of Davy Powergas had demonstrated remarkable SO,, removal efficiency and
operating reliability on large oil-fired boilers in Japan.  Allied
Chemical had successfully recovered high quality elemental sulfur from
S0~ emitted by a Canadian metallurgical operation on a scale equivalent
to a 2000 MW station burning 3% sulfur coal.

The 115 MW FGD facility now in operation at NIPSCO's Dean H. Mitchell
Station was conceived by NIPSCO and EPA as an opportunity to significantly
advance the state-of-the-art of regenerable FGD by combining two com-
mercially proven processes to recover useful products from the emissions
of a coal-fired boiler.

Under the basic project terms, EPA and NIPSCO jointly funded the in-
stallation while NIPSCO assumed all costs of operation.  Davy Powergas
undertook engineering, procurement, construction and performance testing
of the complete system under contract to NIPSCO.  Allied Chemical pro-
vided Davy Powergas with the process design for the conversion of S0~ to
elemental sulfur, as well as start-up services for the entire system.

In addition, NIPSCO engaged Allied to operate and maintain the facility
and to market the chemical products during the demonstration period and
on a continuing basis thereafter.

Construction was completed in the summer of 1976 and the first flue
gas was accepted into the absorber on July 19, 1976.  Following shakedown
of the component parts in the FGD system, the FGD facility was idled
for five months while repairs were made to NIPSCO's No. 11 boiler.

Start-up resumed on May 27, 1977, and performance test were successfully
concluded on September 14, 1977, over the prescribed 12-day and 3-1/2
day test periods at 92 MW equivalent and 110 MW equivalent flue gas
rates of 320,000 ACFM and 390,000 ACFM respectively.
                                   998
-------
                         Project Structure

At an early date, NIPSCO the EPA recognized the need to assemble the
specialized resources necessary to assure success of a chemical processing
installation.  An optimum combination of skills and resources were brought
together in contractual arrangements between:

1.   NIPSCO and DPG;
2.   DPG and Allied Chemical;
3.   Allied Chemical and NIPSCO.

                       Operating Organiztion

With decades of sulfur pollution control services experience to client
companies in the petroleum, detergent and other industries, Allied's
participation as system manager at NIPSCO was a logical extension of its
commitment to the recovery of useful sulfur products from potentially
polluting wastes.

At the Mitchell Station, highly qualified professionals were selected
from within Allied's organization for permanent assignment to all the
supervisory functions in the facility.  Operations and mechanics were
recruited by Allied and trained at the site.

Allied's responsibilities include the full range of technical, maintenance,
tests and inspection and process control functions (Exhibit 1) required
for successful operation of the facility on a continuing basis.

Additional professionals were assigned to the project during the training
period and throughout start-up to insure addquate round-the-clock super-
vision of initial operations through performance testing.  The Allied
permanent staff and start-up group supplemented by a Davy Powerpas start-up
team is illustrated in Exhibit 2.

During the demonstration year, which began September 16, 1977, technical
specialists will assist the permanent Allied staff to optimize operations
and make adjustments and modifications to assure efficient performance
under changing load, fuel and other conditions.  The permanent on-site
organization is shown in Exhibit 3.  It should be noted that the manage-
ment and supervisory staff are sufficient to support a 500 MW FGD
facility.  Some additional maintenance men would be required.

                         Product Marketing

Sale and distribution of useful products, whether sulfur or sulfuric
acid, can be assured as an integral part of Allied Chemical's FGD
system management commitment.

At NIPSCO, over 20 tons per day of sulfur are produced.  It is shipped
from the site as a liquid.  The sulfur produced (608 long tons as of
September 19, 1977) provides a fraction of the sulfur requirements of
a nearby Allied Chemical plant where it is consumed in the manufacture
of sulfuric acid.  Likewise, the sodium purge salt, recovered from the
Wellman-Lord system as dry granular product, wi/Ll be channeled into such
markets as the pulp and paper industry.        \
                                 999
-------
                        PROCESS DESCRIPTION

                     Wellman-Lord SO,., Recovery

At NIPSCO, the Wellman-Lord process is composed of three major operating
sections - SO. absorption, purge treatment and S0« regeneration.

In the SO  absorption section, residual fly ash in the flue gas is
removed by water scrubbing.  S0_ is then removed by scrubbing with a
solution of sodium sulfite.  The chemicals contained in this solution
remain completely dissolved throughout the absorber.  Flue gas scrubbing
with a clear solution, free from suspended solidsj plugging and scaling,
is a fundamental reason underlying the exceptional on-stream reliability
experienced in the commercial operations of the Wellman-Lord process.

The purge treatment section at NIPSCO selectively removes inactive
oxidized sodium compounds from a sidestream of the absorbing solution
and converts this material into a dry granular product which is marketed.

The third section of the Wellman-Lord process involves thermal regeneration
of the absorbing solution to release the absorbed SO  as a concentrated
gas stream and return of the reconstituted solution to the absorber.

The concentrated SO- gas may be converted to either elemental sulfur
by Allied Chemical's S0~ Reduction Process or to sulfuric acid by the
Well-known contact process.

              Allied Chemical SO. Reduction to Sulfur
              	2	

At NIPSCO, sulfur is recovered by Allied's S0_ reduction process which
consists of two principal operating sections.

In the primary reduction section, more than one-half the entering S0?
is converted to elemental sulfur.  A key feature of this section is the
effective control of chemical reactions between S0» and natural gas over
a low cost catalyst developed by Allied for this purpose.  Heat generated
by these chemical reactions is recovered and utilized to preheat the
SO,, gas stream entering this section.

Packed bed regenerative heaters provide a rugged and efficient means
for achieving this heat exchange function.  The process gas flow through
the regenerators is periodically reversed to alternately store and remove
heat from the packing; hence, the overall section is thermally self-
sustaining.

Automatic control of the flow reversing cycles and other process conditions
achieves optimum performance in the system, with high sulfur recovery
efficiency and reductant utilization at all operating rates.

The secondary reaction section uses another catalyst which converts the
remainder of the sulfur values in the process stream to elemental sulfur.

Sulfur produced in each section is removed from the process, stored and
shipped as a hot liquid product.

                                   1000
-------
After the sulfur has been removed, the gas stream is incinerated and
returned to the Wellman-Lord absorber for recovery and recycle of the
residual SO-.   There are no exi±s or effluents requiring separate
disposal.

A flow diagram showing the combined Wellman-Lord/Allied Chemical
processes is shown in Figure 1.
                   SYSTEM PERFORMANCE AT NTPSCO

Results were better than the performance criteria established for the
acceptance test period.  The criteria and results obtained are detailed
below.

                       92 MW Equivalent Test

S00 Removal
 L l£ ' "'" " "'"

1.   Required:      Minimum S0_ removal of 90%, measured continuously
                    and averaged every 2 hours.

2.   Results:       SO  removal averaged 91% over the 12-day test
                    period.  In only two 2-hour periods (out of 144)
                    was the SO  removal less than 90%, and for those
                    periods, it averaged 88% and 89%.  Daily results
                    are shown in Figure 2.

Particulate Removal

1.   Required:      Particulate emission measured once daily will
                    not exceed the Federsl NSPS for fossil fuel
                    fired steam generators of 0.1 Ib/million Btu
                    heat input.

2.   Results:       Particulate emission averaged 0.04 Ib/million
                    Btu, or  40% of the maximum allowable.  Of
                    the 12 days, tests could not be run on four days
                    due to inclement weather.  On one day, the test
                    data was not valid.  Daily results are shown
                    in Figure 2.

Soda Ash Consumption

1.   Required:      Average over the 12-day test period not to
                    exceed 6.6 STPD.

2.   Results:       Soda ash consumption determined by daily
                    inventory obtained from storage bin measurement
                    (official result) averaged 6.2 STPH, or 94%
                    of the maximum allowable.  Consumption determined
                    by manually weighing the feeder output every
                    two hours throughout the 12-day test period
                    averaged 5.7 STPD, or 86% of the maximum
                    allowable.  Daily results are shown in Figure 2.
                                 1001
-------
Aggregate Cost of Steam, Electricity and Natural Gas

1.   Required:
2.
     Results:
Sulfur Purity

1.   Required:
2.   Results:
S00 Removal
1.   Required:
2.   Results:
Aggregate cost not to exceed $56 per hour based
on predetermined unit cost.

Hourly cost averaged $43 per hour over the 12-day
test period, or 77% of the maximum allowable.
Daily results are shown in Figure 2.
                    Minimum sulfur purity 99.5%,  suitable for con-
                    version to quality sulfuric acid by standard
                    production practice.

                    Sulfur purity determined from a composite sample
                    collected over the 12-day test period was 99.9%,
                    easily exceeding the  required purity.

                      110 MW Equivalent Test
                    Minimum SO- removal of 90%,  measured continuously
                    and averaged every 2 hours.

                    SO,, removal averaged 91% over the 3-1/2 day
                    test period.  In only one 2-hour period (out
                    of 42) was the SO- removal less than 90%,
                    and for that period, it averaged 89%.   Daily
                    results are shown in Figure  3.
Particulate Removal
1.   Required:
                    Particulate emission measured once daily will
                    not exceed the Federal NSPS for fossil fuel fired
                    steam generators of 0.1 Ib/million Btu heat input.

2.   Results:       Particulate emission averaged 0.04 Ib/million
                    Btu, or 40% of the maximum allowable.  Of the
                    3-1/2 days, a test could not be run on one day
                    due to inclement weather.  Daily results are
                    shown in Figure 3.

             Viability of the Wellman-Lord FGD System

During the 12-day test period at the 92 MW equivalent rate, there was
a total of 26 hours in interruptions in the fully integrated operation
of the FGD system.  Of the 26 hours, 18 were related to boiler problems
and 8 were related to problems in the sulfur plant.  In addition, there
was a 4-hour period in which the S0? removal averaged "only" 88.5%;
this 4-hour period was added to the acceptance test at the end of the
12-day test period.  It should be mentioned that outages in the sulfur
plant did not interrupt SO  removal.  Furthermore, an S0? removal of

                                  1002
-------
88.5% at the NIPSCO site falls well below an emission of 1.2 Ib/million
Btu heat input.  During the acceptance test the parameters used by EPA
for judging the viability of a FGD system were:  availability, 94%;
reliability, 100%; operability, 100%; utilization factor, 94%.  Similar
results are expected during the demonstration year.
                             ECONOMICS

                    Capital and Operating Costs

The estimated capital investment and operating costs of a Wellman-Lord
FGD installation for a typical midwestern 500 MW power station burning
3% sulfur coal are shown in Tables 1 and 2.   Both the elemental sulfur
and sulfuric acid options are offered.

          Energy Requirements of the Wellman-Lord System

The Wellman-Lord S0? Recovery System has been portrayed by others as
being a high energy user when compared to other FGD systems.  However,
it can be shown that this portrayal is not justified.

The Wellman-Lord FGD Process requires low pressure steam for re-
generation of the absorbing solution.  Turbine extraction or exhaust
steam is ideally suited for the requirements.  However, it is not always
possible to modify the power plant cycle in order to provide this type
steam, especially when the application is a retrofit.

In these cases, the steam required is provided directly from the
steam drum after a suitable pressure reduction.  Main steam provided
to the Wellman-Lord System is arbitrarily assigned a cost equivalent to
that of producing this steam in the power plant.  This is a higher
energy consumption than is actually incurred.

A typical modern power plant will operate with steam to the turbine at
2400 psi and 1000°F.  The turbine exhaust will be at 2.5 in. Hg absolute,
and the corresponding saturated condensate will be at a temperature of
109°F.  After various reheat stages, the boiler feedwater will be at a
temperature of about 475°F.  The corresponding enthalpies of water at
the various stages are:
1.   Boiler feedwater:
2.   Main steam:
3.   Turbine exhaust steam:
4.   Condensate:
 459 Btu/lb
1462 Btu/lb
1020 Btu/lb (75% efficient turbine)
  77 Btu/lb
Of the 1000 Btu/lb of energy provided to water in the boiler, only
440 Btu/lb - or 44% - is actually used in a turbine to produce electric
power.  The remainder - 56% - is rejected to the environment via a
condenser and cooling tower!

Conversely, the Wellman-Lord FGD system will expand steam through
turbine(s) driving the booster blower(s) to greatly reduce electric
power consumption.  The energy remaining in the turbine exhaust steam

                                 1003
-------
will then be utilized to regenerate the absorbing solution.  Furthermore,
uncontaminated condensate will be returned to the power plant at or near
its normal boiling point, thereby reducing the amount of energy required
to heat the water back up to the normal boiler feed temperature.

The steam cost charged to the Wellman-Lord system should, therefore, be
corrected by a factor corresponding to the much more efficient utilization
of energy.  Furthermore, credit should be given the Wellman-Lord system
for the return of hot, uncontaminated condensate.

Purely from an energy standpoint, about 7 Ib/hr of main steam are required
to produce 1 kw of electric power in a modern power plant.  In a Wellman-
Lord FGD system applied to a 500 MW power plant burning 3% sulfur coal,
about 200,000 Ib/hr of steam will be required for absorbing solution
regeneration.  This amount of steam is equivalent to about 28.6 MW of
electric power generation.  In addition, 5000 HP - or 3.4 MW - are
required for the Wellman-Lord system in question.

The total energy consumption of the Wellman-Lord system then is 32.0 MW,
or 6.4% of the generated power.  Recent applications of..lime/limestone
FGD systems show energy requirements of the same order.
                    WELLMAN-LORD VS. THROWAWAY

Why should a utility select the Wellman-Lord FGD system instead of a
throwaway system?  There are several reasons for making this decision.

                               Cost

It has been previously reported that a Wellman-Lord FGD system requires
more capital investment than a throwaway system.  This difference is
rapidly becoming non-existent.  In order to ensure reliability, throw-
away processes are becoming more expensive.  Furthermore, when the
cost of waste sludge disposal is added to the cost of the FGD system,
throwaway processes often become more expensive than Wellman-Lord.

                            Reliability

While the reliability of lime/limestone scrubbing systems has shown
some improvement, one basic fact about these systems remains unchanged.
As S0~ is absorbed into the scrubbing medium, the latter becomes more
insoluble.  Small departures from narrow operating margins usually result
in scaling and/or plugging in the absorber with consequent downtime.
After ten months of operation at NIPSCO, the historical performance
of the Wellman-Lord system remains unchanged.  A Wellman-Lord absorber
has never been shut down due to scaling and/or plugging.  Furthermore,
it requires little operator attention.
      Forsythe, R. C., "Experiences with Flue Gas Desulfurization at
the Bruce Mansfield Plant," Conference on Coal Gasification, Liquefaction
and Conversion to Electricity, Pittsburgh, August 1977.
                                  1004
-------
                          Waste Disposal

Sulfuric acid or elemental sulfur are useful products that can be
marketed.  Large quantities of sulfuric acid are used in the chemical,
agricultural and oil industries.  Elemental sulfur is used primarily
in the manufacture of sulfuric acid.  Sulfuric acid or elemental
sulfur users will pay a competitive price for the product from a
Wellman-Lord FGD system.  This will partially offset the cost of
flue gas desulfurization via Wellman-Lord.

The sulfate purge from a Wellman-Lord system can be marketed.  Sodium
sulfate is also a byproduct of the rayon fiber industry.  However, the
amount of sodium sulfate available from this source is declining and
will continue to do so in the foreseeable future.  The pulp and paper
industry has - and will continue to have - a demand for sodium sulfate.
DPG believes that marketing of this commodity is by no means an im-
possible task.

Conversely, at present there is no market for the sludge produced
by a throwaway system, and the economic and environmental impacts
of a sludge disposal system are uncertain.  These have been described
recently.
                QUALIFICATIONS FOR FUTURE PROJECTS

                           Davy Powergas

The experience of Davy Powergas in SO- recovery began with the start-up
in 1970 of the first Wellman-Lord installation.  Development work by
Davy on predecessor processes dates back to 1965.  Since 1970, 25
installations have been completed and are in regular commercial service,
and five more are in various stages of engineering and construction.

While the NIPSCO project represents the first application of Wellman-
Lord technology to a coal-fired utility boiler, pilot plant tests made
prior to design work indicated that no major difficulties were to be
expected in applying the process to coal-fired boilers.  This expecta-
tion has been confirmed in operations at NIPSCO to date.  Construction
of Wellman-Lord FGD facilities for two 350 MW coal-fired units will
be completed early in 1978, and an installation on a 550 MW boiler is
in the early construction stages.

The Wellman-Lord process offers more experience in commercial service
than any other regenerable system.  Its operating reliability, flexi-
bility, and SO. removal efficiency are well proven, and the fact that
a valuable resource is recovered rather than creating a waste product
should make the system more attractive to many power plants.
      Forsyhte, R. C., "Experiences with Flue Gas Desulfurization at
the Bruce Mansfield Plant," Conference on Coal Gasification, Liquefaction
and Conversion to Electricity, Pittsburgh, August 1977.
                                  1005
-------
Davy Powergas offers this long experience in regenerable FGD systems in
the form of process and design engineering, procurement services, full
construction services, and start-up assistance.

                          Allied Chemical

Allied Chemical is one of the largest producers of sulfuric acid for
the merchant market and consumes more than 700,000 tons of sulfur each
year.

Today, eight of Allied's fifteen North American sulfuric acid locations
can reprocess sulfur-containing waste material from petroleum refineries,
detergents manufacturers and other industries.  Allied Chemical's partici-
pation in the FGD project at Mitchell Station is a logical extension of
sulfur pollution control services provided to others for many years.  Many
companies not experienced in chemical operations and chemical markets
call upon Allied to provide specialized technology and/or services in the
recovery of sulfur values from polluting wastes.

Allied Chemical is prepared to provide to electric utilities a full range
of technical, operating, maintenance and marketing services necessary
for the long-term implementation of a sound regenerable FGD program
yielding either sulfur or sulfuric acid as the principal useful product.
Allied Chemical is prepared to manage a regenerable FGD facility as
it would its own chemical plants and bring to bear on-site experience
and in-depth support of a national sulfur products network to assure
successful execution.
                            CONCLUSION

There is every expectation that the NIPSCO regenerable FGD installation
will continue to demonstrate the efficiency and reliability sought by
the electric utility industry for the control of S02 emissions from
coal fired boilers.

The management of NIPSCO with the encouragement and assistance of the EPA
took a progressive and forward looking step in their decision to proveed
with this regenerable FGD project.  Transformation of SO- in flue gas
to useful chemical products rather than worthless sludge is an objective
worthy of serious consideration by the power industry.

Concentrated S0« from the Wellman-Lord SO- Recovery Process may be con-
verted to either sulfur or sulfuric acid.

Davy Powergas is prepared to design and build a FGD system to recover
sulfur or sulfuric acid from the flue gas of any size fossil fuel fired
boiler in the power industry.

Allied Chemical is prepared to operate such a regenerable FGD facility
for the electric utility producing either useful product and is prepared
to undertake marketing arrangements to assure long-term distribution
of either sulfuric acid or sulfur.


                                 1006
-------
                              TABLE 1

                   NOVEMBER 1977 COST PROJECTION
                     WELLMAN-LORD FGD PROCESS
                      SULFURIC ACID RECOVERY
BASIS:         500 MW Coal Fired Power Plant
               Midwestern 3% Sulfur Coal, 10,600 Btu/lb
               Capacity Factor - 80%
               Fourth Quarter 1977 U.S. Dollars
INSTALLED CAPITAL COST BATTERY LIMITS:                 $35,500.000


ANNUAL OPERATING COSTS:                                     $1,000

     Soda Ash ($89/ton dlvd. Midwest)                          910
     Steam ($1.30/1000 Ibs.)                                 1,820
     Cooling Water ($0.01/1000 gal.)                            80
     Process Water ($0.05/1000 gal.)                            10
     Electricity (15 mills/kwh)                                530
     Adra. & Supervision $14/hr incl. fringes)                  170
     Operating Labor ($10/hr incl. fringes)                    390
     Overhead (75% of labor & supervision)                     420
     Maintenance (3.5% of capital)                           1,220

          Total Direct Operating Expense                     5,550

     Credit from Sale of Acid ($15/ton)                     (1,660)

          Net Direct Operating Expense                       3,890

     Capital Charges (18% per year)                          6,300

          TOTAL                                             10,190
               Capital Cost = $71/kw
               Annual Operating Cost = 2.9 mills/kwh
                                  1007
-------
                              TABLE 2

                   NOVEMBER 1977 COST PROJECTION
                     WELLMAN-LORD FGD PROCESS
                      WITH SULFUR RECOVERY BY
               ALLIED CHEMICAL S07 REDUCTION PROCESS
BASIS:         500 MW Coal Fired Power Plant
               Midwestern 3% Sulfur Coal, 10,600 Btu/lb
               Capacity Factor - 80%
               Fourth Quarter 1977 U.S. Dollars
INSTALLED CAPITAL COST BATTERY LIMITS;                 $37,500,000


ANNUAL OPERATING COSTS:                                     $1,000

     Soda Ash ($89/ton dlvd. Midwest)                          910
     Natural Gas ($2.00/MCF)                                   970
     Steam ($1.30/1000 Ibs.)                                 1,820
     Cooling Water ($0.01/1000 gal.)                            60
     Process Water ($0.05/1000 gal.)                            10
     Electricity (15 mills/kwh)                                530
     Adm. & Supervision ($14/hr incl. fringes)                 170
     Operating Labor ($10/hr incl. fringes)                    390
     Overhead (75% of labor & supervision)                     420
     Maintenance (3.5% of capital)                           1,310

          Total Direct Operating Expense                     6,590

     Credit from Sale of Sulfur ($40/long ton)

          Net Direct Operating Expense

     Capital Charges (18% per year)
          TOTAL                                             12,040
               Capital Cost = $75/kw
               Annual Operating Cost = 3.4 mills/kwh
                                    1008
-------
CO
CD
I—
sr
CD
h-
CD
CD
C/3
-------
Figure 2 - 92 MW Performance Test Parameters
                567
                   TEST DAY
10    11    12    13
-------
                 Figure  3  -  110 MW Performance Test Parameters
                                                     S;^
        91
             '    s. Actual
                                 Aseraige
_i_—
                                               ^-nfirtr
      Oi.lO
        .09
      0
.08
      £0.07
        ,06
      6.03
    tch
                                     TEST DAY
                                                           1011
-------
1-1
LL
K
CD
  O
60
           1012
-------
 ^ g>
 co I
CM

H
M
oa
i—i
ffi
X
w
i|
   £
 CO
 CO
 o
 CO
    •c
    2
    55



£5
^2
1«
£2


i_ 52
guuj
















Za.
01 3
il
£E 5
CO £

_l
cj uu
^^ ^u
c3 S
LU Z
d1"

LU UJ
C3 tg
££
                                n LU

                                = gg

                                 I S^

Is
< co

5 =9
I *»




I— uu a:
ocz tf
sis
~l
1
1
1
i
1



UJ
o . 	 .
-------
   UJ



   CO  t=
         cu


         CO
   o   i
   u  ts
w
   CO  CD

   CO  "•
   LLJ
         o
   CO  0
                                   1014
-------
                         Exhibit 4
Regenerable FGD Facility Next to NIPSCO's Mitchell Station
  Loading Sulfur Recovered from the SO  Emission Control
           Facility at NIPSCO's Mitchell Station

                      1015
-------
SO2  SPRAY ABSORPTION WITH DRY WASTES
                  K. Felsvang,
                    K. Gude,
                   S.  Kaplan
                Niro Atomizer, Inc.
                Columbia, Maryland
                     1016
-------
N1R.O ATOMIZER INC
                   SO2  SPRAY ABSORPTION WITH DRY WASTES
                              K.  Felsvang
                              K.  Gude
                              S.  Kaplan

                            NIRO  ATOMIZER
    Technology which  has  been  used  within the field of industrial
    spray drying  for  more than forty years has now been adapted  for
    solving  the problems  associated with flue gas desulfurization.

    The  Niro Atomizer spray  absorption  system is designed  for  SO,,
    removal  from  boiler exhaust gases.   It can further be  utilized
    for  removal of HCL or other harmful gases.   The process, which
    is very  similar to conventional spray drying technology, is  out-
    lined in Figure 1.  It consists of  the feed preparation  system,
    which will vary depending  upon  the  absorbant to be utilized, a
    transportation system for  delivering the  feed from the prepara-
    tion site to  the  atomization assembly, a  rotary or spinning  wheel
    atomizer for  absorbant atomization  in the process  chamber  and
    a gas distribution system  for obtaining the correct gas-liquid
    mixing within the chamber.   This, along with the dust  collection
    equipment, will remove all  sulfur oxide and particulate pollu-
    tants prior to the flue  gas emission into the atmosphere and
    will produce  a dry residue  which can be disposed of along  with
    fly  ash  wastes, thereby  requiring no further treatment.

    The  preparation system for  the  absorbing  slurry must be designed
    very carefully, and varying conditions during the  mixing stages
    can  have substantial  effects on the absorbant utilization.   Exten-
    sive test work has shown that in the case of lime,  for instance,
    the  type of quick lime used and the method  of mixing it with
    water can effect  S02  removal  efficiencies significantly.   By
    selecting optimum operating conditions, up  to 90%  S02  removal has
    been achieved.  Other materials, such as  sodium carbonate  based
    minerals, have not shown as  drastic variations due  to  preparation
    parameters, but close  control must  be maintained nevertheless
    to assure that the proper concentration of  solids  necessary  for
    the  desired S02 removal will  be  present.

    The  transportation system for delivering  the  absorbant will  con-
    sist  simply of a  pump capable of moving a high solids' slurry.
    With  the Niro Atomizer rotary atomizer, there are no pressure
    requirements at which the feed must enter the  wheel; therefore,
    the  feed will be pumped to  a  small  constant  pressure head  tank
    on the top of the  reaction  chamber  and  will  then flow  to the
    atomizer by gravity.   By utilizing  this set-up,  all  high pressure
    components have been eliminated  completely as  simple centrifugal
    pumps will be utilized for  slurry delivery.

                                 1017
-------
N1RO ATOMIZER INC.
    The key to the entire system is the atomization of the absorbing
    slurry and the intermixing of this atomized slurry with the boiler
    exhaust gases.  Niro Atomizer is capable of supplying single
    rotary atomizers which can handle up to 200 metric tons per
    hour oi slurry when required.  Systems to treat up to 500,000
    ACFM in a single chamber have been designed where absorbant
    feed rates in the order of 20 to 35 metric tons per hour are
    fed.  With the larger atomizer capability, higher gas flow rates
    per chamber are possible.  The absorbant is atomized by the forces
    exerted on it as it is thrown from the spinning disk or wheel.
    The wheel contains up to 24 orifices, about one-half inch in
    liameter, and as the slurry is accelerated through the
    -nternal section of the wheel, it is thrown out through these
    orifices, oreaking into millions of small droplets.  This special
    patented abrasion resistant wheel is currently used throughout
    the world for atomizing highly abrasive slurries such as in
    mineral concentrates, cement and kaolin.  The speed of the
    rfheel can be in the order of ten thousand revolutions per minute
    and by changing this speed, the size of the droplets produced can
    be modified.  As the droplets exit from the wheel, a cloud is
    formed which extends horizontally throughout much of the chamber
    (Ser Figure 2) .

    It is critical that the droplets in this cloud mix completely with
    incoming boiler flue gases to achieve the necessary S02 removal,
     nd equally important to have complete water evaporation take
    place before the slurry droplets come in contact with the chamber
    Optimum mixing is achieved by means of the flue gas distribution
    system shown on Figure 2.  The flue gas leaves the air preheater
    an* is transported to a roof gas disperser which funnels the gas
    ev nly around the atomizer as it enters the chamber.  A series of
    adjustable vanes in the gas disperser give the gas a swirling
    motion to provide for complete interaction of the SC>2 laden gas
    with the atomized absorbing droplets.

    As this interaction occurs, the SO2 in the gas reacts with the
    active material in the atomized slurry while at the same time water
    is evaporated to form a dry product.  Part of this dried powder
    can then be collected in the absorption chamber which, through its
    design, acts as a cyclone separator, or the powder can be collected
    in the same dust collection equipment used for fly ash removal.
    It is possible to split the percentage of dust collected in the
    absorber and precipitator or baghouse so that by collecting a
    larger quantity in the absorber, the requirements of the dust
    collection equipment down stream can be reduced significantly.

    Temperatures within the absorption system are closely controlled,
    as in normal spray dryer operation.  The outlet temperature is held
                                  1018
-------
NIRO ATOMIZER INC.
    enough above  the  dew point to allow for a temperature drop across
    the  dust collection  equipment without causing any operational
    problems.   This is especially important when a baghouse,  which
    is sensitive  to both high and low temperatures,  is used.   Since
    the  temperature at the  absorber outlet can be controlled  to within
    two  degrees,  it is possible to utilize low temperature bag
    materials  within  the baghouse which could reduce both capital and
    operational costs.

    Keeping a  close temperature control over the system is also
    important  when considering that the amount of active material to
    be utilized will  vary as  inlet SCK concentrations vary.   Test work
    has  indicated that stoichiometric ratios approaching those
    achieved in conventional  wet scrubbers can be obtained.   It is,
    therefore,  important to maintain a very tight control over tem-
    perature,  from a  material consumption viewpoint  as well as to
    avoid  operational problems.

    Whereas in a  wet  scrubbing process the wet sludges represent a
    major  pollution problem in themselves,  the residue from the dry
    absorption system is a  dried,  free-flowing powder which can be
    disposed of in a  manner similar to normal fly ash disposal.   It
    will not be necessary to  provide any further treatment to these
    wastes and the volume of  material which must be  disposed  of is
    substantially less than that produced by a wet scrubber.   Also,
    since  there is no further treatment of the wastes,  equipment such
    as thickeners, centrifuges,  and filters can be eliminated,  which
    will greatly  reduce  the amount of land necessary for a pollution
    control system.

    From a maintenance point  of view,  the dry SO2 absorption  process
    offers many advantages  over conventional wet scrubbing systems.
    Since  there is no wet-dry interface,  problems usually occurring,
    such as scaling within  the scrubber,  have been eliminated.   Alsof
    since  the  feed transportation system is simpler  in the dry pro-
    cess,  the  likeliness of mechanical problems in transporting the
    absorbant  is  reduced.   And,  as already discussed,  the removal of
    the  dry residue from the  absorber presents no problems while
    sludge from a wet process will result in material handling as well
    as disposal problems.

    The  Niro Atomizer spray absorption system offers a new method for
    removal of  SC>2 from  boiler off gases  while employing concepts and
    operations  which  have been used in conventional  spray drying plants
    for  over forty years.   High tonnage spray dryers have consistently
    shown  the  high reliability and continuous service which is essential
    for  operation in  the utility industry.   It is, therefore,  possible
    to eliminate  the  problems with existing flue gas desulfurization
    systems by  utilizing this unique spray absorption process  with
    dry  waste  disposal.
                                  1019
-------
0)
M
3
cn
                                                 1020
-------

                                                                                      Q:
                                                                                      -*v
u n
3e
u. rr
-------
CIRCUMSTANCES OF FGD AT
CHUBU ELECTRIC POWER CO.
       Masato Miyajima
   Chubu Electric Power Co., Ltd.
        Nagoya, Japan
            1022
-------
    Circumstances  of FQD  at  Chubu  Electric Power  Co.

                                            Masato  Miyajima
                                            Chubu Electric Power Co., Ltd.
                                            Nagoya, Japan

      In Japan,  where  its  national  territory, particulary the inhabitable area
is quite limited,  both population and industry are much too overcrowded.
      For this reason,  residential  districts are not quite separated from  in-
dustrial districts,  and in the  past we experienced problems where atmospheric
pollution had resulted in  damages to the health of the inhabitants.
      In the city  of Yokkaichi,  in  spite of the fact that six factories  in-
cluding our power  plant were operating by abiding by the emission  standards
respectively, we were  sued for  the  alleged damages to the health of the local
inhabitants owing  to the exhaust gas, and the trials have resulted to impose
exceedingly severe control on us, the enterprises.
      In Japan,  as regards emission of SOx, in accordance with the agreement
concluded with the local governments, it is strictly specified to be equivalent
of 0.1 to 0.3 percent  in terms  of sulphur content. Since long before the problem
of such atmospheric  pollution came  to the fore (the first half of the  1960s),
this company has been  studying  problems relative to air pollution.
      It is an opinion of  the electric power suppliers as the users of  the
fuel that efforts  towards  supply of oil with less sulphur content should be
made on the part of  petroleum companies, suppliers of the fuel. However, in
consideration of the demand-supply  situation of low-sulphur crude oil, avail-
able technology for  oil desulphurization, the capacity of existing  necessary
facilities and other reasons, it has become necessary to adopt the method of
desulphurizing the  flue  gas.    For such reasons that heavy oil with  low
sulphur content be  used with priority for boilers operated by medium and small-
size companies,  some of the boilers belonging to power supply companies  have
been equipped with  flue   gas  desulphurizing systems so as to cooperate  in
                                       1023
-------
meeting the low sulphurization plan target which is being sought throughout




the country.



      The first process of desulphurization ever to be adopted by Chubu  Elec-




tric Power Co. was the dry activated manganese oxide process.  Based on this




process, various test plants on different scales were constructed  at  our




Yokkaichi Thermal Power Station as follows!




      DAP - 3,000 (capacity  3,000 Mm3/hr)




      DAP -    55 (capacity     55 MW)



      DAP -   110 (capacity    110 MW, (220 MW x 1/2))




            (Note - DAP: dry absorption process)




      Through researches conducted from 1964 to 1974, a large number of im-



portant data were obtained.  This process, however, because of its own char-




acterristics, failed to match the situation being prevalent then, and there-



fore no additional expansion of the process was effected.



      In parallel with the dry process, this company was proceeding  with



research on wet processes, and having a prospect that the wet process  is




commercializable earlier than the dry process, as the first step, FGD  of




the Wellman-Lord process'on full scale was erected as No. 1 power  plant




(220 MW) at Nishi-Nagoya in September 1973.




      The Wellman-Lord process was  selected for the following reasons?



l)   The quantity of raw materials required and the quantity of by-products



  turned out are small.



2)   No possibility of scaling in the absorption tower unlike  in the case of



  the lime gypsum process.




3)   Japan Synthetic Rubber Co. has constructed one with a capacityt




  100,000 NmVhr.



4)   The  by-products H2S04 is  effectively usables as a chemical industry




  raw material.  Further, it finds a rather large market and consequently



  problem is limited as to its disposal.
                                      1024
-------
      Operation results of the Wellman-Lord process at the Nishi-Nagoya Power




Station during the past four years are as follows:




 Operation time:  26,000 hours (Sept. 1973 - Sept. 1977)




 Operation rate:  91$




 Efficiency:   3^2 at inlet:  1,500 ppm




                            (equivalent to 2.5$ sulphur content.)




              S02 at outlet:   110 ppm




                            (equivalent to 0.2$ sulphur content)




              Mean desulphurization rate 92$




 By-produced sulphuric acid:  60,200 tons




 Principal troubles experienced and counter measures taken therefor:




(l)   Clogging of fine pipes of the heater of the regeneration system evapo-




  ration boiler




      Clogging and corrosion of the fine pipes of the heater due to depositing




  of such substances as sodium sulphite, and Glauber's salt.  To eliminate the




  trouble, periodical washing with water (once monthly) and jet cleaning ( 3




  to 4 times annually) are being conducted.






(2)   Cracking in the interior of the absorption tower




      Cracking was observed in the tray inside the absorption tower during the




  initial period following the startup.




      As countermeasure, various parts were reinforced and as the result no




  similar troubles have occurred thereafter.






(3)   Leak through the fine pipes of the surface condenser




      Leak took place resulting from corrosion of the interior of the  fine




  pipes (seawater side).  As the solution, the material of the fine pipes was




  changed from stainless steel ( epoxy lined) to titanium.




             We are almost content with the results obtained.




      At No.l and No.2 plants at the Owaae Mita Power Station (with a capacity




of 375 MW respectively), FGD of the lime gypsum process for  total  flue gas






                                       1025
-------
treatment was erected in the beginning of 1976.



      The lime gypsum process was adopted for the reasons to be enumerated in



the following:




l)   This process, that has already been employed by other power supply com-




 panies, is now offering higher reliability.



2)   The process itself is simple.




3)   The raw material of lime is abundantly available in Japan.




4)   Even in case where a large number of PGDs are set up in Japan in  the



 future with resultant oversupply of the by-product of gypsum, it will be




 easy, unlike sulphuric acid, to store and discard in some cases, because of



 its chemical stability.  What is more, gypsum is being consumed in  large



 quantities for such purposes as making cement and gypsum boards, and it




 offers greater possibility of new applications and uses in the future.



     Stepped up efforts are being made in that direction too.






      Operation results of the lime gypsum process at the Owashe Mita  Power



Station are:




 Operation time:  No.l  11,000 hrs (April 1976-Sept. 1977)




                  No.2   9,600 hrs (June 1976-Sept. 1977)



 Operation rate:  No.l      96 %




                  No.2      99 %



 Efficiency:  S02 at inlet - 1,600 ppm




                             (equivalent to 2.9 % sulphur content),



              S02 at outlet -  120 ppm



                             (equivalent to 0.25 % sulphur content),



              mean desulphurization rate 92 %



 Fuel consumption:  No.l  860,000kl




                    No.2  760,000kl




 Gypsum production:   No.l   128,000 tons



                     No.2   133,000 tons
                                      1026
-------
 Principal troubles and countermeasures therefor:




(l)   Clogging of the mist eliminator



      There were instances where the trains were stopped alternately  to  be




  cleaned in turns because the differential pressure tended to increase owing




  to gypsum scale that deposited on the elements of the mist eliminator during




  the initial period following the startup.






(2)   Slurry piping




      Leakage through the slurry piping and valves due to local abrasion was




  detected particularly in the parts that were subjected to high temperature.




  Thus, the pipeline was modified in order to facilitate maintenance.






(3)   Absorbing tower




      Although partial clogging of the internal grid was seen, it did not lead




  to malfunction in particular.  Cleaning of the grid was performed once  an-




  nually at the time of periodical inspection.






(4)   While clogging of the filter element of the gypsum separator (centrifu-




  gal machine) occurred frequently during the initial period following  the




  startup, its occurrence was brought to a minimum by  employing  filter




  element of a different type and changing the operation method.




      We are virtually satisfied with the foregoing results.






Re economics




      Although the economics of desulphurization of flue gas that are re-




lated to the operations of this company cannot be stated definitely,  they




register at 6,COO to 8,000 yen per kl by a 7-year depreciation and a 70 %




operation rate, and this is nearly equivalent to the fuel coat difference in




Japan between high-sulphur oil and low-sulphur oil.




      FGD in Japan, however, is of very high efficiency, including its after-




treatment system in consideration of the various strict environmental stand-




ards, with resultant high costs, but in cases where very high efficiency is




not called for depending on the sitting conditions of power plants, the cost





                                       1027
-------
is expected to be somewhat lower.




      On the other hand,  in Japan although primary emphasis used to be placed




on SOx in the past by way of environmental control of thermal power stations,




the control standards are becoming increasingly severer of late extending over




to all areas such as NOx, smoke dust, water quality and wastes, and FGD, too,




in consideration of such control standards, has come to be placed in a delicate




situation.  In other words, use of HS consequent upon installation of the FGD




entails increase in NOx generated, and as measures to be taken therefor more




sophisticated techniques and greater costs are required.  Use of the  FGD,




furthermore, incurs other problems which are associated with wastewater treat-




ment and disposal of by-products.




      In view of the above-mentioned background cf the FGD, use of fuel with




higher quality as well as use of fuels of diversified types and kinds,  even




LNG and naphtha have come to be employed.  What is more, FGD has rarely been




installed recently because the citing conditions of thermal power stations




call for use of ultra low sulphur fuel (such as LNG and equivalent of naphtha)




and  flue   gas  with minimum NOx content.  However, two units of coal-burning




500 MW FGDs are said to be installed at the Matsushima Power Station  of




the Electric Power Development Company which is scheduled to be completed




in I960.
                                       1028
-------
FLUE GAS DESULPHURIZATION PLANT
 ON OWASE-MITA POWER STATION
           Masato Miyajima
       Chubu Electric Power Co., Ltd.
            Nagoya, Japan
                1029
-------
                     FLUE GAS DESULPHURIZATION PLANT


                       ON OWASE-MITA POWER STATION
                                              M,  Miyajima,

                                              Chubu Electric Power Co.,  Ltd.

                                              Nagoya,  Japan
Introduction

      Owase-Mita Power Station's No.l and 2 machines having a total output of

      750,000 kW (375,000 kW x 2 units)  was started in operation in July and

      September, 1964, respectively.

      Since then we have made a maximum  effort in lowering the sulphur content

      of the fuel to keep steps with  the requirement of age.   Then, in  order
                                                                        *
      to improve the environment further and taking into account the totalized

      conditions such as the relationship of combination with the adjacent Toho

      Petroleum Oil Co., Ltd. and fuel affairs,  we decided to install a flue

      gas desulphurization plant belonging to the largest class in Japan  as

      an effective means to lowering  the sulphur content.  Installation work

      was started on Oct. 17, 1974, and  building work was continued for more

      than   year and a half since then.  The No.l unit passed the   per -

      operation test on April 8, 1976 and No.2 unit on June 25, 1976.

      Both units have been and are in favorable operation.

      The units will be described briefly here.
                                     1030
-------
1.    Outline of facility
     Kind of unit
     Amount of gas processed
     Desulphurization factor        :
     Absorbent source material      :
     Amount of source material used!

     Byproduct                     !
     Amount of byproduct produced  :

     Flue gas reheating system     :
     Installation area             :
     Total cost of construction    :
     Manufacturer
Wet type lime-gypsum method
 1,200 x ID? Nm3/h
(Total amount of flue gas for 375,000 kW
 is processed)
 90 % or more
 Quicklime

 About 60,000 tons/year
 (Nos.l and 2 units)
 Gypsum
 About 180,000 tons/year
 (Nos.l and 2 units)
 After burner system
 About 29,000 m2
 About 16,000,000,000 yen
 (Aprox.64 million dollars)
 (Nos.l and 2 units)
 Mitsubishi Heavy Industries Co., Ltd.
2.   Process
     This unit accepts quicklime (CaO) as the source material and prepares
 from it the absorbent (Ca (011)2) solution.  This solution is used to rinse
 the flue gas from the boiler to remove SOx content and recover gypsum
 (CaS04.2H20) as the byproduct.
     The flow chart of the unit is attached at the end of this booklet
                                 1031
-------
                                Jl
        Owase-Mita Power Station and its surrounding viewed from
air.
2-1   Material accepting storing facility
      Since the material lime is dangerous it is carried by special dump
  trucks,  delivered directly to the hopper,  sent from the bottom of  the
  hopper to the lime storing tank  by using Various  conveyors  and  stored
  there temporarily,
                                    1032
-------
      The storing capacity is 1100 m3.  The tank capacity was designed  as



  the capacity for 5 day's use, taking into account the holidays and  road




  accidents.






2-2   Absorbing liquid preparation process



      The quick lime is picked up from the storage tank, weighed, and  then




  supplied to the lime mixing tank.  There the lime is mixed with supernatant




  supplied from the gypsum thickener and digested to form milk of lime,  The




  milk of lime is sent to the milk-lime thickener to classify the  digestion



  residue.  The over flowing liquid is taken out continuously and sent to the




  second absorbent reservoir as an absorbent 5 % in concentration.  The di-



  gestion residue is taken out from the bottom of the thickener, crushed



  vertical tower mill, and all is used as the absorbent.






2-3   Flue  gas cooling process



      The  flue  gas of the boiler undergoes humidification, cooling,  and




  dust removal in the cooling tower.  The gas temperature, which is  some



  150 °C at the inlet, decreases to 55 °C  at the outlet.




      This process makes the desulphurizing reaction in absorbing  process



  advantageous by decreasing the gas temperature and the same time,  by



  saturating the moisture, plays a role of preventing the local condensation




  of slurry which is caused by the vaporization of the moisture  in  the



  absorbing unit.






2-4   Absorbing process



      The gas that has passed through the cooling tower is washed with  ab-



  sorbent and removed of S02 here.




      Grid filling type of simple cooling tower is used because of its advan-




  tage of having a large capacity factor of 5^2 absorption, small pressure



  loss, high performance stability, and being easily scaled up.



      Two absorbing towers, installed in series with the gas flow, are employed




  in this unit in order to satisfy the SO? absorbing factor and Ca utility



                                      1033
-------
  factor.
           GAS
1st
ABSORBING
TOWER
=£>

< —
2nd
ABSORBING
TOWER
                                                        ABSORBENT
      The main purpose of the 2nd absorbing tower is the removal of 302 .



  The pH of the absorbent is set at 6.5 to remove most of the 30^.



      The purpose of the 1st absorbing tower is the role of making the  un-




  reacted Ca component in the absorbent used in the 2nd absorbing tower react



  again rather than S02 removal.



      In the case where lime slurry is used to wash the gas, the most important




  practical problem is the prevention of sc.ale adhesiont  Scale is defined as



  the adherence of gypsum formed by the reaction between a portion of calcium




  sulphite and the oxigen contained in the waste gas to the grid and inner



  walls of the tower.  Since in this unit various countermeasures are being



  taken, a smooth operation is continued.




      Although calcium sulphite is formed by the main reaction of 302 absorption




  as shown in formula (l), one portion is oxidized to gypsum as shown in formula




  (2).
Ca(OH)2 + S02 + H20




            *• H2°
H20
                                                             (l)
                                   CaS04 2H20 + H20 ............... — (2)
2-5   Oxidation Process



      Calcium sulphite slurry is fed and compressed air is blown into  the




  oxidation tower, where calcium sulphite is oxidized to form gypsum as shown



  in the reaction formula (2).




      To cause oxidation to take place with the highest efficiency a  rotary



  atomizer is installed on the bottom of the oxidation tower in order  to




  float up very minute air bubbles through the slurry.




      The gypsum liquid is controlled to below pH4.  Sulphuric acid is added
                                    1034
-------
  when the pH is higher.  Since two absorbing towers are used in this unit
  the Ca utility factor is almost 10C$>, and the pS adjusting system is almost
  unused•
2-6   Filtration process
      The gypsum slurry extracted from the bottom of the oxidation tower is
  condensed in the thickener sent to the squirrel-cage centrifugal separator
  as a 25^ concentration fluid to be dehydrated.  The moisture content  of
  the gypsum is 10$ or less, pH value is neutral, and the purity is 98/'°  or
  more.
      The overflown liquid of the thickener and the filtrate of the centrifu-
  gal separator are sent to the absorbent preparating process and reused in
  the digestion between it and lime.
2-7   Gypsum storing and delivering process
      The gypsum that has been produced is carried to and stored in the gypsum
  warehouse by means of conveyors.  The storing capacity was designed to  be
  for 7 day's amount taking into account climatic condition etc.  Marine
  transport of gypsum is being planned, and the quay allows up to a maximum
  of 1,500 ton vessels to approach.
      Since the cargo-working is made in the day time in principle a  ship
  loader having a maximum loading capacity of 230 tons/hour is installed on
  the quay and the conveyor carrying the gypsum to the loader is run through
  the inside a cylinder type closed duct.
2-8   Utilizing water, waste   water treating
      The amount of water used by this unit is about 3»500 nr/day (2 units),
  the supply source is the surplus water from the deep well of the adjacent
  plant-TOHO PETROLEUM Co., Ltd. which is stored in a 3,500 kl tank installed
  in our yard.  Abut 2,400 m-Vunit of liquid is circulating through the system
  and its control must be made with considerable care.
      Spending a large amount of water this unit was so designed as to  keep
  the water balance in the system by reusing as much water as possible.
                                     1035
-------
   Most of the waste water comes from the cooling system and it is treated by

   the waste water treating facility established within the system.


3.   Installation schedule

     Starting in February, 1974 we removed and moved 3 warehouses, 2 heavy oil

 tanksflight oil tank, modified appended pipings, and moved the ammonia tank

 to secure the installation site of the desulphurization plant.  We also made

 geological survey and preparation work for electric power facility for  en-

 gineering work.

     This engineering schedule is as shown below.



                                No.l  unit          No.2  unit

      Foundation work          Oct. 17, '74        Same as left

      Machine installation     Feb. 10, '75        Mar. 28, '75

      Electric power           Sep. 18, '75        Same as left
                reception

      Flue gas passing         Dec. 26, '75        Mar.  4, '76

      Commercial running       Apr.  8. '76        June 25» '76



     A period of about 10.5 months was spent between the installation of the

 units and the time of passing the flew gas .   The scale of the unit was large,

 the kinds of work were many and the number of subcontractors was numerous.

 Moreover, almost all works were concentrated around the cooling and absorbing

 tower and the schedule was made severer by the following additional require-

 ments.

  (a)   The lining of the  cooling and absorbing tower and the inner path of the

   flue duct were made being always influenced by weather during the work.

  (b)   Since the lining pipe was the main subject of piping, installation  of

   on-site fitting pipes  such as the piping of coupling section could not be

   made until the completion of the lining of the main body.

  (c)   There were many rainy days.


                                       1036
-------
4.   Electrical facilities

     The power consumption of two desulphurization units is about 20,000 kW,

 and it was necessary to secure highly reliable electric wires.   However,  since

 the already existing power supply facility was insufficient in  capacity  it

 was determined to receive power from the 77 kV bus of the substation addition-

 ally installed.   A transformer was installed on the desulphurization unit and

 a 200 mm2 CV triplex cable is used,  buried underground,  to connect  them.

     Since the protection pipe has many bendings and its total length  was

 530 m, the EFLBX pipe,  which is easy to lay,  was used.

     One unit of  transformer was used common to Nos.l and 2 units.   Employment

 of 27/13.5 + 13.5 MVA 3 wire wound type made  the short-circuit  capacity small

 and MSB of small cut-off capacity (350 MVA)  was used.   No special stand-by

 power supply was installed because of  the  high reliability of the power supply.

 The emergency power  was designed to  be supplied from the existing diesei

 generator.

     As for the DC power supply  only  a  125  V  1000 AH  capacity one was used to

 operate those devices necessary to be  operated  assuredly immediately after

 the power failure.
                    »   t #»;

                      UlLi;  •  Ll
                          ta  S       *"* * "W» *
                        .11  -J'-H  (  II  >«
                          ;             «
                      Control room of the FGD plants

                                      1037
-------
5.   Control method




     A control room common to Nos.l and 2 desulphurization units having facili-




 ties capable of conducting concentrated operation and supervision of Nos.l and




 2 units by using 4 operators.  The central control board was a vertical divided




 board arranged for each reaction group which is economically advantageous.




     To make the performance of desulphurization unit to 450 x ICr Nm^/hour




(equivalent to 125,000 kW) to 1200 x 105 Nm5/nour (equivalent to 375,000 kW)




 in waste gas amount and enable to follow the change in gas amount  of 40 x 10^




 NmVmin. (equivalent to 10,000 kW), the coordination of the control with the




 boiler load, a method in which the pressure at the outlet of the electrical




 dust precipitator, which is the junction point, is controlled to the pressure




 setting value determined by the load, was employed.








£,   Actual results of running




     Total duration of operation of No. 1 and No. 2 is 11,090 hours (No. l)




 and 9,680 hours (No. 2) as of the end of September 1977 since their startup




 in April (No. l) and in June (No. 2) in 1976.




     Even though during the initial period following the startup, troubles




 occurred resulting from such causes as abrasion of the slurry piping and




 clogging -f the mist eliminator, such troubles have recently been eliminated




 through operation improvements thereby reaching an operation rate of almost




 100 percent.
                                      1038
-------
-------
-------
     The operation conditions may be outlined as follows«



           (as of the end of September 1977)
~~— — -___
Operation time of boiler
(hr)
Operation time of FGD
(nr)
Operation rate of FGD
(£)
Shutdown due to FGD trouble
I, times)
Duration of shutdown due
to FGD trouble (hr)
HS fuel consumption
(kl)
Gypsum production
(ton)


Principal cause of
shutdown due to
trouble



No. 1
No. 2
No. 1
No. 2
No. 1
No. 2
No. 1
Ho. 2
No. 1
No. 2
No. 1
No. 2
No. 1
No. 2



No. 1



No. 2
Startup -
March 1977
7,384
5,375
6,977
5,290
94
98
2
0
24i
0
550,000
420,000
85,000
84,000
1 ) Leakage
due to abra-
sion of the
alurrypiping
2) Clogging
of the mist
eliminator.

April 1977-
Sept. 1977
4,122
4,386
4,106
4,366
99.6
100
1
0
6
0
310,000
340,000
43,000
49,000
1 ) Leakage
due to abra-
sion of the
. slurry piping




Total
11,506
9,761
11,083
9,676
96
99
3
0
248
0
860 , 000
760,000
128,000
133,000








    Principal problems so far experienced were as  follows:




(.1)   Clogging of the mist eliminator




      There were instances where the trains were stopped alternately to be




  cleaned in turns because the differential pressure  tended  to  increase owing
                                      1039
-------
   to  gypsum scale  that  deposited  on  the  elements  of  the  mist  eliminator during




   the initial period  following the startup*




 (2)    Slurry piping




       Leakage through the  slurry  piping  and  valves due  to  local abrasion  was




   detected particularly in  the  parts that  were  subjected to high temperatures.




   Thus,  the pipeline  was modified in order to facilitate maintenance.




 (3)    Absorbing tower




       Although partial  clogging of the internal grid was seen,  it did  not lead




   to  malfunction in particular.  Cleaning of the  grid was  performed once




   annually at the  time  of  perjodical inspection.




 (4)    While clogging of the filter element of the gypsum separator (centrifugal




   machine) occurred frequently during the initial period following the startup*




   its occurrence was  brought to a minimum by employing filter element  of  a




   different type and  changing the operation method.
Summary




         The unit,  which was completed by investing a large capital, is different




   from the conventional power generating plant and appears to be a large chemical




   plant.  In order to operate new devices with a few operators,  there may arise




   many problematic points.   However,  if the desulphurization unit continues




   stable operation over a long period of time it will do its duty as a public




   pollution preventive unit and serve to accelerate the site planning of power




   supply facilities.   This  may be the best desired result.  We are intending to




   make effort for maintaining the stable operation.




         Although the utilization of gypsum is still low, since its utilization




   is being investigated eagerly in the related fields,  we are expecting  good




   results and, at the same  time, hoping those concerned will give as further




   understanding and aid.
                                       1040
-------
a
o
~  c
                                1041
-------
                                TECHNICAL REPORT DATA
                         (Please read Iiiuructions on the reverse bcjore completing)
1. REPORT NO.
 EPA-600/7-78-oA8b
                           2.
                                                      3. RECIPIENT'S ACCESSION" NO.
4. TITLE AND SUBTITLE
Proceedings: Symposium on Flue Gas Desulfurization-
   Hollywood, FL, November 1977  (Volume H)
                                5. REPORT DATE
                                  March 1978
                                6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)

Franklin A. Ayer,  Compiler
                                                      8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Research Triangle Institute
P.O. Box 12194
Research Triangle Park, North Carolina 27709
                                                      10. PROGRAM ELEMENT NO.
                                E HE 62 4 A
                                11. CONTRACT/GRANT NO.

                                 68-02-2612, Task 38
12. SPONSORING AGENCY NAME AND ADDRESS
 EPA, Office of Research and Development
 Industrial Environmental Research Laboratory
 Research Triangle  Park, NC 27711
                                                      13. TYPE OF REPORT AND PERIOD COVERED
                                                       Proceedings; 11/8-11/77
                                 14. SPONSORING AGENCY CODE
                                  EPA/600/13
15. SUPPLEMENTARY NOTES IERL-RTP project officer is Julian W.  Jones, Mail Drop 61, 919/
541-2489.
16. ABSTRACT
           The proceedings document presentations made during the symposium,
which dealt with the status of flue gas desulfurization technology in the United States
and abroad. Subjects considered included: regenerable, non-regenerable, and
advanced processes: process costs: and by-product disposal, utilization,  and
marketing. The purpose of the symposium was to provide developers, vendors, users
and those concerned with regulatory guidelines with a current review of progress
made in applying processes for the reduction of sulfur dioxide emissions at the full-
and semi-commercial scale.
17.
                             KEY WORDS AND DOCUMENT ANALYSIS
                DESCRIPTORS
                    b.IDENTIFIERS/OPEN ENDED TERMS  C. COSATI Field/Group
Pollution
Flue Gases
Sulfur Dioxide
Desulfurization
Regeneration
Cost Analysis
Byproducts
Disposal
Marketing
Pollution Control
Stationary Sources
13 B
21B
07 B
07A,07D
14 B
13. DISTRIBUTION STATEMENT

 Unlimited
                    19. SECURITY CLASS (This Report)
                     Unclassified
                                                                   21. NO. OF PAGES
                             614
                    20. SECURITY CLASS (Thispage)
                     Unclassified
                                             22. PRICE
EPA Form 2220-1 (9-73)
                 1042
            ->U.S. GOVERNMENT PRINTING OFFICE: 19 78 -7HO-261/ 338 REGIONNO.4
-------